2014-04-18 12:41:34

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 00/28] nios2 Linux kernel port

This patchset adds the Linux kernel port for Nios II processor from Altera.
The nios2 Linux port follows the guidance for new architecture ports using
generic headers (including unistd.h).

About Nios II Cores
-------------------
Nios II is a 32-bit embedded-processor architecture designed specifically for the
Altera family of FPGAs.
More information is available at http://www.altera.com/devices/processor/nios2/ni2-index.html

Instruction set and architecture overview documents can be found on the
following page:
http://www.altera.com/literature/lit-nio2.jsp

Nios2 GCC port is in mainline and and will be in the FSF 4.9 release.

These patches are also available on this branch:
git://git.rocketboards.org/linux-socfpga-next.git nios2-upstream

Regards,
Ley Foon

Ley Foon Tan (28):
nios2: Build infrastructure
nios2: Assembly macros and definitions
nios2: Kernel booting and initialization
nios2: Exception handling
nios2: Traps exception handling
nios2: Memory management
nios2: I/O Mapping
nios2: MMU Fault handling
nios2: Page table management
nios2: Process management
nios2: Cache handling
nios2: TLB handling
nios2: Interrupt handling
nios2: DMA mapping API
nios2: ELF definitions
nios2: System calls handling
nios2: Signal handling support
nios2: Library functions
nios2: Device tree support
nios2: Time keeping
nios2: Cpuinfo handling
nios2: Miscellaneous header files
nios2: Nios2 registers
nios2: Module support
nios2: ptrace support
Add ELF machine define for Nios2
MAINTAINERS: Add nios2 maintainer
Documentation: Add documentation for Nios2 architecture

Documentation/nios2/README | 23 ++
MAINTAINERS | 7 +
arch/nios2/Kconfig | 213 +++++++++++
arch/nios2/Kconfig.debug | 17 +
arch/nios2/Makefile | 78 ++++
arch/nios2/boot/Makefile | 51 +++
arch/nios2/boot/dts/3c120_devboard.dts | 205 +++++++++++
arch/nios2/boot/linked_dtb.S | 19 +
arch/nios2/configs/3c120_defconfig | 77 ++++
arch/nios2/include/asm/Kbuild | 66 ++++
arch/nios2/include/asm/asm-macros.h | 309 ++++++++++++++++
arch/nios2/include/asm/asm-offsets.h | 20 ++
arch/nios2/include/asm/cache.h | 36 ++
arch/nios2/include/asm/cacheflush.h | 48 +++
arch/nios2/include/asm/checksum.h | 78 ++++
arch/nios2/include/asm/cmpxchg.h | 61 ++++
arch/nios2/include/asm/cpuinfo.h | 58 +++
arch/nios2/include/asm/delay.h | 92 +++++
arch/nios2/include/asm/dma-mapping.h | 113 ++++++
arch/nios2/include/asm/elf.h | 101 ++++++
arch/nios2/include/asm/entry.h | 152 ++++++++
arch/nios2/include/asm/gpio.h | 23 ++
arch/nios2/include/asm/io.h | 227 ++++++++++++
arch/nios2/include/asm/irq.h | 34 ++
arch/nios2/include/asm/irqflags.h | 69 ++++
arch/nios2/include/asm/linkage.h | 25 ++
arch/nios2/include/asm/mmu.h | 18 +
arch/nios2/include/asm/mmu_context.h | 66 ++++
arch/nios2/include/asm/mutex.h | 1 +
arch/nios2/include/asm/page.h | 111 ++++++
arch/nios2/include/asm/pci.h | 25 ++
arch/nios2/include/asm/pgalloc.h | 86 +++++
arch/nios2/include/asm/pgtable-bits.h | 35 ++
arch/nios2/include/asm/pgtable.h | 304 ++++++++++++++++
arch/nios2/include/asm/processor.h | 102 ++++++
arch/nios2/include/asm/prom.h | 26 ++
arch/nios2/include/asm/ptrace.h | 40 +++
arch/nios2/include/asm/registers.h | 65 ++++
arch/nios2/include/asm/setup.h | 38 ++
arch/nios2/include/asm/signal.h | 22 ++
arch/nios2/include/asm/string.h | 24 ++
arch/nios2/include/asm/switch_to.h | 31 ++
arch/nios2/include/asm/syscalls.h | 25 ++
arch/nios2/include/asm/thread_info.h | 120 +++++++
arch/nios2/include/asm/timex.h | 27 ++
arch/nios2/include/asm/tlb.h | 34 ++
arch/nios2/include/asm/tlbflush.h | 46 +++
arch/nios2/include/asm/traps.h | 19 +
arch/nios2/include/asm/uaccess.h | 233 ++++++++++++
arch/nios2/include/asm/ucontext.h | 34 ++
arch/nios2/include/uapi/asm/Kbuild | 14 +
arch/nios2/include/uapi/asm/byteorder.h | 22 ++
arch/nios2/include/uapi/asm/elf.h | 67 ++++
arch/nios2/include/uapi/asm/ptrace.h | 123 +++++++
arch/nios2/include/uapi/asm/sigcontext.h | 30 ++
arch/nios2/include/uapi/asm/signal.h | 23 ++
arch/nios2/include/uapi/asm/stat.h | 23 ++
arch/nios2/include/uapi/asm/statfs.h | 24 ++
arch/nios2/include/uapi/asm/swab.h | 37 ++
arch/nios2/include/uapi/asm/unistd.h | 25 ++
arch/nios2/kernel/Makefile | 14 +
arch/nios2/kernel/asm-offsets.c | 88 +++++
arch/nios2/kernel/cpuinfo.c | 201 +++++++++++
arch/nios2/kernel/entry.S | 551 ++++++++++++++++++++++++++++
arch/nios2/kernel/head.S | 175 +++++++++
arch/nios2/kernel/insnemu.S | 592 +++++++++++++++++++++++++++++++
arch/nios2/kernel/irq.c | 90 +++++
arch/nios2/kernel/misaligned.c | 319 +++++++++++++++++
arch/nios2/kernel/module.c | 135 +++++++
arch/nios2/kernel/nios2_ksyms.c | 35 ++
arch/nios2/kernel/process.c | 262 ++++++++++++++
arch/nios2/kernel/prom.c | 67 ++++
arch/nios2/kernel/ptrace.c | 198 +++++++++++
arch/nios2/kernel/setup.c | 214 +++++++++++
arch/nios2/kernel/signal.c | 364 +++++++++++++++++++
arch/nios2/kernel/sys_nios2.c | 72 ++++
arch/nios2/kernel/syscall_table.c | 29 ++
arch/nios2/kernel/time.c | 150 ++++++++
arch/nios2/kernel/traps.c | 185 ++++++++++
arch/nios2/kernel/vmlinux.lds.S | 75 ++++
arch/nios2/lib/Makefile | 1 +
arch/nios2/lib/memcpy.c | 199 +++++++++++
arch/nios2/lib/memmove.c | 82 +++++
arch/nios2/lib/memset.c | 81 +++++
arch/nios2/mm/Makefile | 12 +
arch/nios2/mm/cacheflush.c | 255 +++++++++++++
arch/nios2/mm/dma-mapping.c | 186 ++++++++++
arch/nios2/mm/extable.c | 25 ++
arch/nios2/mm/fault.c | 242 +++++++++++++
arch/nios2/mm/init.c | 152 ++++++++
arch/nios2/mm/ioremap.c | 188 ++++++++++
arch/nios2/mm/mmu_context.c | 116 ++++++
arch/nios2/mm/pgtable.c | 73 ++++
arch/nios2/mm/tlb.c | 274 ++++++++++++++
arch/nios2/mm/uaccess.c | 162 +++++++++
arch/nios2/platform/Kconfig.platform | 129 +++++++
arch/nios2/platform/Makefile | 1 +
arch/nios2/platform/platform.c | 69 ++++
include/uapi/linux/elf-em.h | 1 +
99 files changed, 10186 insertions(+)
create mode 100644 Documentation/nios2/README
create mode 100644 arch/nios2/Kconfig
create mode 100644 arch/nios2/Kconfig.debug
create mode 100644 arch/nios2/Makefile
create mode 100644 arch/nios2/boot/Makefile
create mode 100644 arch/nios2/boot/dts/3c120_devboard.dts
create mode 100644 arch/nios2/boot/linked_dtb.S
create mode 100644 arch/nios2/configs/3c120_defconfig
create mode 100644 arch/nios2/include/asm/Kbuild
create mode 100644 arch/nios2/include/asm/asm-macros.h
create mode 100644 arch/nios2/include/asm/asm-offsets.h
create mode 100644 arch/nios2/include/asm/cache.h
create mode 100644 arch/nios2/include/asm/cacheflush.h
create mode 100644 arch/nios2/include/asm/checksum.h
create mode 100644 arch/nios2/include/asm/cmpxchg.h
create mode 100644 arch/nios2/include/asm/cpuinfo.h
create mode 100644 arch/nios2/include/asm/delay.h
create mode 100644 arch/nios2/include/asm/dma-mapping.h
create mode 100644 arch/nios2/include/asm/elf.h
create mode 100644 arch/nios2/include/asm/entry.h
create mode 100644 arch/nios2/include/asm/gpio.h
create mode 100644 arch/nios2/include/asm/io.h
create mode 100644 arch/nios2/include/asm/irq.h
create mode 100644 arch/nios2/include/asm/irqflags.h
create mode 100644 arch/nios2/include/asm/linkage.h
create mode 100644 arch/nios2/include/asm/mmu.h
create mode 100644 arch/nios2/include/asm/mmu_context.h
create mode 100644 arch/nios2/include/asm/mutex.h
create mode 100644 arch/nios2/include/asm/page.h
create mode 100644 arch/nios2/include/asm/pci.h
create mode 100644 arch/nios2/include/asm/pgalloc.h
create mode 100644 arch/nios2/include/asm/pgtable-bits.h
create mode 100644 arch/nios2/include/asm/pgtable.h
create mode 100644 arch/nios2/include/asm/processor.h
create mode 100644 arch/nios2/include/asm/prom.h
create mode 100644 arch/nios2/include/asm/ptrace.h
create mode 100644 arch/nios2/include/asm/registers.h
create mode 100644 arch/nios2/include/asm/setup.h
create mode 100644 arch/nios2/include/asm/signal.h
create mode 100644 arch/nios2/include/asm/string.h
create mode 100644 arch/nios2/include/asm/switch_to.h
create mode 100644 arch/nios2/include/asm/syscalls.h
create mode 100644 arch/nios2/include/asm/thread_info.h
create mode 100644 arch/nios2/include/asm/timex.h
create mode 100644 arch/nios2/include/asm/tlb.h
create mode 100644 arch/nios2/include/asm/tlbflush.h
create mode 100644 arch/nios2/include/asm/traps.h
create mode 100644 arch/nios2/include/asm/uaccess.h
create mode 100644 arch/nios2/include/asm/ucontext.h
create mode 100644 arch/nios2/include/uapi/asm/Kbuild
create mode 100644 arch/nios2/include/uapi/asm/byteorder.h
create mode 100644 arch/nios2/include/uapi/asm/elf.h
create mode 100644 arch/nios2/include/uapi/asm/ptrace.h
create mode 100644 arch/nios2/include/uapi/asm/sigcontext.h
create mode 100644 arch/nios2/include/uapi/asm/signal.h
create mode 100644 arch/nios2/include/uapi/asm/stat.h
create mode 100644 arch/nios2/include/uapi/asm/statfs.h
create mode 100644 arch/nios2/include/uapi/asm/swab.h
create mode 100644 arch/nios2/include/uapi/asm/unistd.h
create mode 100644 arch/nios2/kernel/Makefile
create mode 100644 arch/nios2/kernel/asm-offsets.c
create mode 100644 arch/nios2/kernel/cpuinfo.c
create mode 100644 arch/nios2/kernel/entry.S
create mode 100644 arch/nios2/kernel/head.S
create mode 100644 arch/nios2/kernel/insnemu.S
create mode 100644 arch/nios2/kernel/irq.c
create mode 100644 arch/nios2/kernel/misaligned.c
create mode 100644 arch/nios2/kernel/module.c
create mode 100644 arch/nios2/kernel/nios2_ksyms.c
create mode 100644 arch/nios2/kernel/process.c
create mode 100644 arch/nios2/kernel/prom.c
create mode 100644 arch/nios2/kernel/ptrace.c
create mode 100644 arch/nios2/kernel/setup.c
create mode 100644 arch/nios2/kernel/signal.c
create mode 100644 arch/nios2/kernel/sys_nios2.c
create mode 100644 arch/nios2/kernel/syscall_table.c
create mode 100644 arch/nios2/kernel/time.c
create mode 100644 arch/nios2/kernel/traps.c
create mode 100644 arch/nios2/kernel/vmlinux.lds.S
create mode 100644 arch/nios2/lib/Makefile
create mode 100644 arch/nios2/lib/memcpy.c
create mode 100644 arch/nios2/lib/memmove.c
create mode 100644 arch/nios2/lib/memset.c
create mode 100644 arch/nios2/mm/Makefile
create mode 100644 arch/nios2/mm/cacheflush.c
create mode 100644 arch/nios2/mm/dma-mapping.c
create mode 100644 arch/nios2/mm/extable.c
create mode 100644 arch/nios2/mm/fault.c
create mode 100644 arch/nios2/mm/init.c
create mode 100644 arch/nios2/mm/ioremap.c
create mode 100644 arch/nios2/mm/mmu_context.c
create mode 100644 arch/nios2/mm/pgtable.c
create mode 100644 arch/nios2/mm/tlb.c
create mode 100644 arch/nios2/mm/uaccess.c
create mode 100644 arch/nios2/platform/Kconfig.platform
create mode 100644 arch/nios2/platform/Makefile
create mode 100644 arch/nios2/platform/platform.c

--
1.8.3.2


2014-04-18 12:27:51

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 09/28] nios2: Page table management

This patch adds support for page table management.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/pgalloc.h | 86 ++++++++++
arch/nios2/include/asm/pgtable-bits.h | 35 ++++
arch/nios2/include/asm/pgtable.h | 304 ++++++++++++++++++++++++++++++++++
arch/nios2/mm/pgtable.c | 73 ++++++++
4 files changed, 498 insertions(+)
create mode 100644 arch/nios2/include/asm/pgalloc.h
create mode 100644 arch/nios2/include/asm/pgtable-bits.h
create mode 100644 arch/nios2/include/asm/pgtable.h
create mode 100644 arch/nios2/mm/pgtable.c

diff --git a/arch/nios2/include/asm/pgalloc.h b/arch/nios2/include/asm/pgalloc.h
new file mode 100644
index 0000000..6e2985e0a7b90284641d41a839b34df22d8864fd
--- /dev/null
+++ b/arch/nios2/include/asm/pgalloc.h
@@ -0,0 +1,86 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 1994 - 2001, 2003 by Ralf Baechle
+ * Copyright (C) 1999, 2000, 2001 Silicon Graphics, Inc.
+ */
+
+#ifndef _ASM_NIOS2_PGALLOC_H
+#define _ASM_NIOS2_PGALLOC_H
+
+#include <linux/mm.h>
+
+static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd,
+ pte_t *pte)
+{
+ set_pmd(pmd, __pmd((unsigned long)pte));
+}
+
+static inline void pmd_populate(struct mm_struct *mm, pmd_t *pmd,
+ pgtable_t pte)
+{
+ set_pmd(pmd, __pmd((unsigned long)page_address(pte)));
+}
+#define pmd_pgtable(pmd) pmd_page(pmd)
+
+/*
+ * Initialize a new pmd table with invalid pointers.
+ */
+extern void pmd_init(unsigned long page, unsigned long pagetable);
+
+extern pgd_t *pgd_alloc(struct mm_struct *mm);
+
+static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd)
+{
+ free_pages((unsigned long)pgd, PGD_ORDER);
+}
+
+static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm,
+ unsigned long address)
+{
+ pte_t *pte;
+
+ pte = (pte_t *) __get_free_pages(GFP_KERNEL|__GFP_REPEAT|__GFP_ZERO,
+ PTE_ORDER);
+
+ return pte;
+}
+
+static inline pgtable_t pte_alloc_one(struct mm_struct *mm,
+ unsigned long address)
+{
+ struct page *pte;
+
+ pte = alloc_pages(GFP_KERNEL | __GFP_REPEAT, PTE_ORDER);
+ if (pte) {
+ if (!pgtable_page_ctor(pte)) {
+ __free_page(pte);
+ return NULL;
+ }
+ clear_highpage(pte);
+ }
+ return pte;
+}
+
+static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
+{
+ free_pages((unsigned long)pte, PTE_ORDER);
+}
+
+static inline void pte_free(struct mm_struct *mm, struct page *pte)
+{
+ pgtable_page_dtor(pte);
+ __free_pages(pte, PTE_ORDER);
+}
+
+#define __pte_free_tlb(tlb, pte, addr) \
+ do { \
+ pgtable_page_dtor(pte); \
+ tlb_remove_page((tlb), (pte)); \
+ } while (0)
+
+#define check_pgt_cache() do { } while (0)
+
+#endif /* _ASM_NIOS2_PGALLOC_H */
diff --git a/arch/nios2/include/asm/pgtable-bits.h b/arch/nios2/include/asm/pgtable-bits.h
new file mode 100644
index 0000000..ce9e7069aa969e54021b2fbbf4b019e2333d24eb
--- /dev/null
+++ b/arch/nios2/include/asm/pgtable-bits.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_PGTABLE_BITS_H
+#define _ASM_NIOS2_PGTABLE_BITS_H
+
+/*
+ * These are actual hardware defined protection bits in the tlbacc register
+ * which looks like this:
+ *
+ * 31 30 ... 26 25 24 23 22 21 20 19 18 ... 1 0
+ * ignored........ C R W X G PFN............
+ */
+#define _PAGE_GLOBAL (1<<20)
+#define _PAGE_EXEC (1<<21)
+#define _PAGE_WRITE (1<<22)
+#define _PAGE_READ (1<<23)
+#define _PAGE_CACHED (1<<24) /* C: data access cacheable */
+
+/*
+ * Software defined bits. They are ignored by the hardware and always read back
+ * as zero, but can be written as non-zero.
+ */
+#define _PAGE_PRESENT (1<<25) /* PTE contains a translation */
+#define _PAGE_ACCESSED (1<<26) /* page referenced */
+#define _PAGE_DIRTY (1<<27) /* dirty page */
+#define _PAGE_FILE (1<<28) /* PTE used for file mapping or swap */
+
+#endif /* _ASM_NIOS2_PGTABLE_BITS_H */
diff --git a/arch/nios2/include/asm/pgtable.h b/arch/nios2/include/asm/pgtable.h
new file mode 100644
index 0000000..9aedd3408350d5fe8f24bfc7fa4f05f275381a86
--- /dev/null
+++ b/arch/nios2/include/asm/pgtable.h
@@ -0,0 +1,304 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ *
+ * Based on asm/pgtable-32.h from mips which is:
+ *
+ * Copyright (C) 1994, 95, 96, 97, 98, 99, 2000, 2003 Ralf Baechle
+ * Copyright (C) 1999, 2000, 2001 Silicon Graphics, Inc.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_PGTABLE_H
+#define _ASM_NIOS2_PGTABLE_H
+
+#include <linux/io.h>
+#include <linux/bug.h>
+#include <asm/page.h>
+#include <asm/cacheflush.h>
+#include <asm/tlbflush.h>
+
+#include <asm/pgtable-bits.h>
+#include <asm-generic/pgtable-nopmd.h>
+
+#define FIRST_USER_ADDRESS 0
+
+#define VMALLOC_START CONFIG_KERNEL_MMU_REGION_BASE
+#define VMALLOC_END (CONFIG_KERNEL_REGION_BASE - 1)
+
+struct mm_struct;
+
+/* Helper macro */
+#define MKP(x, w, r) __pgprot(_PAGE_PRESENT | _PAGE_CACHED | \
+ ((x) ? _PAGE_EXEC : 0) | \
+ ((r) ? _PAGE_READ : 0) | \
+ ((w) ? _PAGE_WRITE : 0))
+/*
+ * These are the macros that generic kernel code needs
+ * (to populate protection_map[])
+ */
+
+/* Remove W bit on private pages for COW support */
+#define __P000 MKP(0, 0, 0)
+#define __P001 MKP(0, 0, 1)
+#define __P010 MKP(0, 0, 0) /* COW */
+#define __P011 MKP(0, 0, 1) /* COW */
+#define __P100 MKP(1, 0, 0)
+#define __P101 MKP(1, 0, 1)
+#define __P110 MKP(1, 0, 0) /* COW */
+#define __P111 MKP(1, 0, 1) /* COW */
+
+/* Shared pages can have exact HW mapping */
+#define __S000 MKP(0, 0, 0)
+#define __S001 MKP(0, 0, 1)
+#define __S010 MKP(0, 1, 0)
+#define __S011 MKP(0, 1, 1)
+#define __S100 MKP(1, 0, 0)
+#define __S101 MKP(1, 0, 1)
+#define __S110 MKP(1, 1, 0)
+#define __S111 MKP(1, 1, 1)
+
+/* Used all over the kernel */
+#define PAGE_KERNEL __pgprot(_PAGE_PRESENT | _PAGE_CACHED | _PAGE_READ | \
+ _PAGE_WRITE | _PAGE_EXEC | _PAGE_GLOBAL)
+
+#define PAGE_COPY MKP(0, 0, 1)
+
+#define PGD_ORDER 0
+#define PTE_ORDER 0
+
+#define PTRS_PER_PGD ((PAGE_SIZE << PGD_ORDER) / sizeof(pgd_t))
+#define PTRS_PER_PTE ((PAGE_SIZE << PTE_ORDER) / sizeof(pte_t))
+
+#define USER_PTRS_PER_PGD (CONFIG_KERNEL_MMU_REGION_BASE / PGDIR_SIZE)
+
+#define PGDIR_SHIFT 22
+#define PGDIR_SIZE (1UL << PGDIR_SHIFT)
+#define PGDIR_MASK (~(PGDIR_SIZE-1))
+
+/*
+ * ZERO_PAGE is a global shared page that is always zero: used
+ * for zero-mapped memory areas etc..
+ */
+extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
+#define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page))
+
+extern pgd_t swapper_pg_dir[PTRS_PER_PGD];
+extern pte_t invalid_pte_table[PAGE_SIZE/sizeof(pte_t)];
+
+/*
+ * (pmds are folded into puds so this doesn't get actually called,
+ * but the define is needed for a generic inline function.)
+ */
+static inline void set_pmd(pmd_t *pmdptr, pmd_t pmdval)
+{
+ pmdptr->pud.pgd.pgd = pmdval.pud.pgd.pgd;
+}
+
+/* to find an entry in a page-table-directory */
+#define pgd_index(addr) (((addr) >> PGDIR_SHIFT) & (PTRS_PER_PGD - 1))
+#define pgd_offset(mm, addr) ((mm)->pgd + pgd_index(addr))
+
+static inline int pte_write(pte_t pte) \
+ { return pte_val(pte) & _PAGE_WRITE; }
+static inline int pte_dirty(pte_t pte) \
+ { return pte_val(pte) & _PAGE_DIRTY; }
+static inline int pte_young(pte_t pte) \
+ { return pte_val(pte) & _PAGE_ACCESSED; }
+static inline int pte_file(pte_t pte) \
+ { return pte_val(pte) & _PAGE_FILE; }
+static inline int pte_special(pte_t pte) { return 0; }
+
+#define pgprot_noncached pgprot_noncached
+
+static inline pgprot_t pgprot_noncached(pgprot_t _prot)
+{
+ unsigned long prot = pgprot_val(_prot);
+
+ prot &= ~_PAGE_CACHED;
+
+ return __pgprot(prot);
+}
+
+static inline int pte_none(pte_t pte)
+{
+ return !(pte_val(pte) & ~(_PAGE_GLOBAL|0xf));
+}
+
+static inline int pte_present(pte_t pte) \
+ { return pte_val(pte) & _PAGE_PRESENT; }
+
+/*
+ * The following only work if pte_present() is true.
+ * Undefined behaviour if not..
+ */
+static inline pte_t pte_wrprotect(pte_t pte)
+{
+ pte_val(pte) &= ~_PAGE_WRITE;
+ return pte;
+}
+
+static inline pte_t pte_mkclean(pte_t pte)
+{
+ pte_val(pte) &= ~_PAGE_DIRTY;
+ return pte;
+}
+
+static inline pte_t pte_mkold(pte_t pte)
+{
+ pte_val(pte) &= ~_PAGE_ACCESSED;
+ return pte;
+}
+
+static inline pte_t pte_mkwrite(pte_t pte)
+{
+ pte_val(pte) |= _PAGE_WRITE;
+ return pte;
+}
+
+static inline pte_t pte_mkdirty(pte_t pte)
+{
+ pte_val(pte) |= _PAGE_DIRTY;
+ return pte;
+}
+
+static inline pte_t pte_mkspecial(pte_t pte) { return pte; }
+
+static inline pte_t pte_mkyoung(pte_t pte)
+{
+ pte_val(pte) |= _PAGE_ACCESSED;
+ return pte;
+}
+
+static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
+{
+ const unsigned long mask = _PAGE_READ | _PAGE_WRITE | _PAGE_EXEC;
+ pte_val(pte) = (pte_val(pte) & ~mask) | (pgprot_val(newprot) & mask);
+ return pte;
+}
+
+static inline int pmd_present(pmd_t pmd)
+{
+ return (pmd_val(pmd) != (unsigned long) invalid_pte_table)
+ && (pmd_val(pmd) != 0UL);
+}
+
+static inline void pmd_clear(pmd_t *pmdp)
+{
+ pmd_val(*pmdp) = (unsigned long) invalid_pte_table;
+}
+
+#define pte_pfn(pte) (pte_val(pte) & 0xfffff)
+#define pfn_pte(pfn, prot) (__pte(pfn | pgprot_val(prot)))
+#define pte_page(pte) (pfn_to_page(pte_pfn(pte)))
+
+/*
+ * Store a linux PTE into the linux page table.
+ */
+static inline void set_pte(pte_t *ptep, pte_t pteval)
+{
+ *ptep = pteval;
+}
+
+static inline void set_pte_at(struct mm_struct *mm, unsigned long addr,
+ pte_t *ptep, pte_t pteval)
+{
+ unsigned long paddr = page_to_virt(pte_page(pteval));
+ flush_dcache_range(paddr, paddr + PAGE_SIZE);
+ set_pte(ptep, pteval);
+}
+
+static inline int pmd_none(pmd_t pmd)
+{
+ return (pmd_val(pmd) ==
+ (unsigned long) invalid_pte_table) || (pmd_val(pmd) == 0UL);
+}
+
+#define pmd_bad(pmd) (pmd_val(pmd) & ~PAGE_MASK)
+
+static inline void pte_clear(struct mm_struct *mm,
+ unsigned long addr, pte_t *ptep)
+{
+ pte_t null;
+
+ pte_val(null) = (addr >> PAGE_SHIFT) & 0xf;
+
+ set_pte_at(mm, addr, ptep, null);
+ flush_tlb_one(addr);
+}
+
+/*
+ * Conversion functions: convert a page and protection to a page entry,
+ * and a page entry and page directory to the page they refer to.
+ */
+#define mk_pte(page, prot) (pfn_pte(page_to_pfn(page), prot))
+
+#define pte_unmap(pte) do { } while (0)
+
+/*
+ * Conversion functions: convert a page and protection to a page entry,
+ * and a page entry and page directory to the page they refer to.
+ */
+#define pmd_phys(pmd) virt_to_phys((void *)pmd_val(pmd))
+#define pmd_page(pmd) (pfn_to_page(pmd_phys(pmd) >> PAGE_SHIFT))
+#define pmd_page_vaddr(pmd) pmd_val(pmd)
+
+#define pte_offset_map(dir, addr) \
+ ((pte_t *) page_address(pmd_page(*dir)) + \
+ (((addr) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)))
+
+/* to find an entry in a kernel page-table-directory */
+#define pgd_offset_k(addr) pgd_offset(&init_mm, addr)
+
+/* Get the address to the PTE for a vaddr in specfic directory */
+#define pte_offset_kernel(dir, addr) \
+ ((pte_t *) pmd_page_vaddr(*(dir)) + \
+ (((addr) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)))
+
+#define pte_ERROR(e) \
+ pr_err("%s:%d: bad pte %08lx.\n", \
+ __FILE__, __LINE__, pte_val(e))
+#define pgd_ERROR(e) \
+ pr_err("%s:%d: bad pgd %08lx.\n", \
+ __FILE__, __LINE__, pgd_val(e))
+
+/*
+ * Encode and decode a swap entry (must be !pte_none(pte) && !pte_present(pte)
+ * && !pte_file(pte)):
+ *
+ * 31 30 29 28 27 26 25 24 23 22 21 20 19 18 ... 1 0
+ * 0 0 0 0 type. 0 0 0 0 0 0 offset.........
+ *
+ * This gives us up to 2**2 = 4 swap files and 2**20 * 4K = 4G per swap file.
+ *
+ * Note that the offset field is always non-zero, thus !pte_none(pte) is always
+ * true.
+ */
+#define __swp_type(swp) (((swp).val >> 26) & 0x3)
+#define __swp_offset(swp) ((swp).val & 0xfffff)
+#define __swp_entry(type, off) ((swp_entry_t) { (((type) & 0x3) << 26) \
+ | ((off) & 0xfffff) })
+#define __swp_entry_to_pte(swp) ((pte_t) { (swp).val })
+#define __pte_to_swp_entry(pte) ((swp_entry_t) { pte_val(pte) })
+
+/* Encode and decode a nonlinear file mapping entry */
+#define PTE_FILE_MAX_BITS 25
+#define pte_to_pgoff(pte) (pte_val(pte) & 0x1ffffff)
+#define pgoff_to_pte(off) __pte(((off) & 0x1ffffff) | _PAGE_FILE)
+
+#define kern_addr_valid(addr) (1)
+
+#include <asm-generic/pgtable.h>
+
+#define pgtable_cache_init() do { } while (0)
+
+extern void __init paging_init(void);
+extern void __init mmu_init(void);
+
+extern void update_mmu_cache(struct vm_area_struct *vma,
+ unsigned long address, pte_t *pte);
+
+#endif /* _ASM_NIOS2_PGTABLE_H */
diff --git a/arch/nios2/mm/pgtable.c b/arch/nios2/mm/pgtable.c
new file mode 100644
index 0000000..21e865f1079aba32ed8964b82ff1979c473b119a
--- /dev/null
+++ b/arch/nios2/mm/pgtable.c
@@ -0,0 +1,73 @@
+/*
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/mm.h>
+#include <linux/sched.h>
+
+#include <asm/pgtable.h>
+#include <asm/cpuinfo.h>
+
+/* pteaddr:
+ * ptbase | vpn* | zero
+ * 31-22 | 21-2 | 1-0
+ *
+ * *vpn is preserved on double fault
+ *
+ * tlbacc:
+ * IG |*flags| pfn
+ * 31-25|24-20 | 19-0
+ *
+ * *crwxg
+ *
+ * tlbmisc:
+ * resv |way |rd | we|pid |dbl|bad|perm|d
+ * 31-24 |23-20 |19 | 20|17-4|3 |2 |1 |0
+ *
+ */
+
+/*
+ * Initialize a new pgd / pmd table with invalid pointers.
+ */
+static void pgd_init(pgd_t *pgd)
+{
+ unsigned long *p = (unsigned long *) pgd;
+ int i;
+
+ for (i = 0; i < USER_PTRS_PER_PGD; i += 8) {
+ p[i + 0] = (unsigned long) invalid_pte_table;
+ p[i + 1] = (unsigned long) invalid_pte_table;
+ p[i + 2] = (unsigned long) invalid_pte_table;
+ p[i + 3] = (unsigned long) invalid_pte_table;
+ p[i + 4] = (unsigned long) invalid_pte_table;
+ p[i + 5] = (unsigned long) invalid_pte_table;
+ p[i + 6] = (unsigned long) invalid_pte_table;
+ p[i + 7] = (unsigned long) invalid_pte_table;
+ }
+}
+
+pgd_t *pgd_alloc(struct mm_struct *mm)
+{
+ pgd_t *ret, *init;
+ ret = (pgd_t *) __get_free_pages(GFP_KERNEL, PGD_ORDER);
+ if (ret) {
+ init = pgd_offset(&init_mm, 0UL);
+ pgd_init(ret);
+ memcpy(ret + USER_PTRS_PER_PGD, init + USER_PTRS_PER_PGD,
+ (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t));
+ }
+
+ return ret;
+}
+
+void __init pagetable_init(void)
+{
+ /* Initialize the entire pgd. */
+ pgd_init(swapper_pg_dir);
+ pgd_init(swapper_pg_dir + USER_PTRS_PER_PGD);
+}
--
1.8.3.2

2014-04-18 12:28:26

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 12/28] nios2: TLB handling

This patch adds the TLB maintenance functions.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/tlb.h | 34 +++++
arch/nios2/include/asm/tlbflush.h | 46 +++++++
arch/nios2/mm/tlb.c | 274 ++++++++++++++++++++++++++++++++++++++
3 files changed, 354 insertions(+)
create mode 100644 arch/nios2/include/asm/tlb.h
create mode 100644 arch/nios2/include/asm/tlbflush.h
create mode 100644 arch/nios2/mm/tlb.c

diff --git a/arch/nios2/include/asm/tlb.h b/arch/nios2/include/asm/tlb.h
new file mode 100644
index 0000000..d3bc648e08b5dad86e5e9c449e655fe291381c91
--- /dev/null
+++ b/arch/nios2/include/asm/tlb.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_TLB_H
+#define _ASM_NIOS2_TLB_H
+
+#define tlb_flush(tlb) flush_tlb_mm((tlb)->mm)
+
+extern void set_mmu_pid(unsigned long pid);
+
+/*
+ * NiosII doesn't need any special per-pte or per-vma handling, except
+ * we need to flush cache for the area to be unmapped.
+ */
+#define tlb_start_vma(tlb, vma) \
+ do { \
+ if (!tlb->fullmm) \
+ flush_cache_range(vma, vma->vm_start, vma->vm_end); \
+ } while (0)
+
+#define tlb_end_vma(tlb, vma) do { } while (0)
+#define __tlb_remove_tlb_entry(tlb, ptep, address) do { } while (0)
+
+#include <linux/pagemap.h>
+#include <asm-generic/tlb.h>
+
+#endif /* _ASM_NIOS2_TLB_H */
diff --git a/arch/nios2/include/asm/tlbflush.h b/arch/nios2/include/asm/tlbflush.h
new file mode 100644
index 0000000..e19652fca1c6826404f4abca6581ce7f42b8ed03
--- /dev/null
+++ b/arch/nios2/include/asm/tlbflush.h
@@ -0,0 +1,46 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_TLBFLUSH_H
+#define _ASM_NIOS2_TLBFLUSH_H
+
+struct mm_struct;
+
+/*
+ * TLB flushing:
+ *
+ * - flush_tlb_all() flushes all processes TLB entries
+ * - flush_tlb_mm(mm) flushes the specified mm context TLB entries
+ * - flush_tlb_page(vma, vmaddr) flushes one page
+ * - flush_tlb_range(vma, start, end) flushes a range of pages
+ * - flush_tlb_kernel_range(start, end) flushes a range of kernel pages
+ */
+extern void flush_tlb_all(void);
+extern void flush_tlb_mm(struct mm_struct *mm);
+extern void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
+ unsigned long end);
+extern void flush_tlb_kernel_range(unsigned long start, unsigned long end);
+extern void flush_tlb_one(unsigned long vaddr);
+
+static inline void flush_tlb_page(struct vm_area_struct *vma,
+ unsigned long addr)
+{
+ flush_tlb_one(addr);
+}
+
+#endif /* _ASM_NIOS2_TLBFLUSH_H */
diff --git a/arch/nios2/mm/tlb.c b/arch/nios2/mm/tlb.c
new file mode 100644
index 0000000..80f0d6e4
--- /dev/null
+++ b/arch/nios2/mm/tlb.c
@@ -0,0 +1,274 @@
+/*
+ * Nios2 TLB handling
+ *
+ * Copyright (C) 2009, Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/init.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/pagemap.h>
+
+#include <asm/tlb.h>
+#include <asm/mmu_context.h>
+#include <asm/pgtable.h>
+#include <asm/cpuinfo.h>
+
+#define TLB_INDEX_MASK \
+ ((((1UL << (cpuinfo.tlb_ptr_sz - cpuinfo.tlb_num_ways_log2))) - 1) \
+ << PAGE_SHIFT)
+
+/* Used as illegal PHYS_ADDR for TLB mappings
+ */
+#define MAX_PHYS_ADDR 0
+
+static void get_misc_and_pid(unsigned long *misc, unsigned long *pid)
+{
+ *misc = RDCTL(CTL_TLBMISC);
+ *misc &= (TLBMISC_PID | TLBMISC_WAY);
+ *pid = *misc & TLBMISC_PID;
+}
+
+/*
+ * All entries common to a mm share an asid. To effectively flush these
+ * entries, we just bump the asid.
+ */
+void flush_tlb_mm(struct mm_struct *mm)
+{
+ if (current->mm == mm)
+ flush_tlb_all();
+ else
+ memset(&mm->context, 0, sizeof(mm_context_t));
+}
+
+/*
+ * This one is only used for pages with the global bit set so we don't care
+ * much about the ASID.
+ */
+void flush_tlb_one_pid(unsigned long addr, unsigned long mmu_pid)
+{
+ unsigned int way;
+ unsigned long org_misc, pid_misc;
+
+ pr_debug("Flush tlb-entry for vaddr=%#lx\n", addr);
+
+ /* remember pid/way until we return. */
+ get_misc_and_pid(&org_misc, &pid_misc);
+
+ WRCTL(CTL_PTEADDR, (addr >> PAGE_SHIFT) << 2);
+
+ for (way = 0; way < cpuinfo.tlb_num_ways; way++) {
+ unsigned long pteaddr;
+ unsigned long tlbmisc;
+ unsigned long pid;
+
+ tlbmisc = pid_misc | TLBMISC_RD | (way << TLBMISC_WAY_SHIFT);
+ WRCTL(CTL_TLBMISC, tlbmisc);
+ pteaddr = RDCTL(CTL_PTEADDR);
+ tlbmisc = RDCTL(CTL_TLBMISC);
+ pid = (tlbmisc >> TLBMISC_PID_SHIFT) & TLBMISC_PID_MASK;
+ if (((((pteaddr >> 2) & 0xfffff)) == (addr >> PAGE_SHIFT)) &&
+ pid == mmu_pid) {
+ unsigned long vaddr = CONFIG_IO_REGION_BASE +
+ ((PAGE_SIZE * cpuinfo.tlb_num_lines) * way) +
+ (addr & TLB_INDEX_MASK);
+ pr_debug("Flush entry by writing %#lx way=%dl pid=%ld\n",
+ vaddr, way, (pid_misc >> TLBMISC_PID_SHIFT));
+
+ WRCTL(CTL_PTEADDR, (vaddr >> 12) << 2);
+ tlbmisc = pid_misc | TLBMISC_WE |
+ (way << TLBMISC_WAY_SHIFT);
+ WRCTL(CTL_TLBMISC, tlbmisc);
+ WRCTL(CTL_TLBACC, (MAX_PHYS_ADDR >> PAGE_SHIFT));
+ }
+ }
+
+ WRCTL(CTL_TLBMISC, org_misc);
+}
+
+void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
+ unsigned long end)
+{
+ unsigned long mmu_pid = get_pid_from_context(&vma->vm_mm->context);
+
+ while (start < end) {
+ flush_tlb_one_pid(start, mmu_pid);
+ start += PAGE_SIZE;
+ }
+}
+
+void flush_tlb_kernel_range(unsigned long start, unsigned long end)
+{
+ while (start < end) {
+ flush_tlb_one(start);
+ start += PAGE_SIZE;
+ }
+}
+
+/*
+ * This one is only used for pages with the global bit set so we don't care
+ * much about the ASID.
+ */
+void flush_tlb_one(unsigned long addr)
+{
+ unsigned int way;
+ unsigned long org_misc, pid_misc;
+
+ pr_debug("Flush tlb-entry for vaddr=%#lx\n", addr);
+
+ /* remember pid/way until we return. */
+ get_misc_and_pid(&org_misc, &pid_misc);
+
+ WRCTL(CTL_PTEADDR, (addr >> PAGE_SHIFT) << 2);
+
+ for (way = 0; way < cpuinfo.tlb_num_ways; way++) {
+ unsigned long pteaddr;
+ unsigned long tlbmisc;
+
+ tlbmisc = pid_misc | TLBMISC_RD | (way << TLBMISC_WAY_SHIFT);
+ WRCTL(CTL_TLBMISC, tlbmisc);
+ pteaddr = RDCTL(CTL_PTEADDR);
+ tlbmisc = RDCTL(CTL_TLBMISC);
+
+ if ((((pteaddr >> 2) & 0xfffff)) == (addr >> PAGE_SHIFT)) {
+ unsigned long vaddr = CONFIG_IO_REGION_BASE +
+ ((PAGE_SIZE * cpuinfo.tlb_num_lines) * way) +
+ (addr & TLB_INDEX_MASK);
+
+ pr_debug("Flush entry by writing %#lx way=%dl pid=%ld\n",
+ vaddr, way, (pid_misc >> TLBMISC_PID_SHIFT));
+
+ tlbmisc = pid_misc | TLBMISC_WE |
+ (way << TLBMISC_WAY_SHIFT);
+ WRCTL(CTL_PTEADDR, (vaddr >> 12) << 2);
+ WRCTL(CTL_TLBMISC, tlbmisc);
+ WRCTL(CTL_TLBACC, (MAX_PHYS_ADDR >> PAGE_SHIFT));
+ }
+ }
+
+ WRCTL(CTL_TLBMISC, org_misc);
+}
+
+void dump_tlb_line(unsigned long line)
+{
+ unsigned int way;
+ unsigned long org_misc;
+
+ pr_debug("dump tlb-entries for line=%#lx (addr %08lx)\n", line,
+ line << (PAGE_SHIFT + cpuinfo.tlb_num_ways_log2));
+
+ /* remember pid/way until we return */
+ org_misc = (RDCTL(CTL_TLBMISC) & (TLBMISC_PID | TLBMISC_WAY));
+
+ WRCTL(CTL_PTEADDR, line << 2);
+
+ for (way = 0; way < cpuinfo.tlb_num_ways; way++) {
+ unsigned long pteaddr;
+ unsigned long tlbmisc;
+ unsigned long tlbacc;
+
+ WRCTL(CTL_TLBMISC, TLBMISC_RD | (way << TLBMISC_WAY_SHIFT));
+ pteaddr = RDCTL(CTL_PTEADDR);
+ tlbmisc = RDCTL(CTL_TLBMISC);
+ tlbacc = RDCTL(CTL_TLBACC);
+
+ if ((tlbacc << PAGE_SHIFT) != (MAX_PHYS_ADDR & PAGE_MASK)) {
+ pr_debug("-- way:%02x vpn:0x%08lx phys:0x%08lx pid:0x%02lx flags:%c%c%c%c%c\n",
+ way,
+ (pteaddr << (PAGE_SHIFT-2)),
+ (tlbacc << PAGE_SHIFT),
+ ((tlbmisc >> TLBMISC_PID_SHIFT) &
+ TLBMISC_PID_MASK),
+ (tlbacc & _PAGE_READ ? 'r' : '-'),
+ (tlbacc & _PAGE_WRITE ? 'w' : '-'),
+ (tlbacc & _PAGE_EXEC ? 'x' : '-'),
+ (tlbacc & _PAGE_GLOBAL ? 'g' : '-'),
+ (tlbacc & _PAGE_CACHED ? 'c' : '-'));
+ }
+ }
+
+ WRCTL(CTL_TLBMISC, org_misc);
+}
+
+void dump_tlb(void)
+{
+ unsigned int i;
+ for (i = 0; i < cpuinfo.tlb_num_lines; i++)
+ dump_tlb_line(i);
+}
+
+void flush_tlb_pid(unsigned long pid)
+{
+ unsigned int line;
+ unsigned int way;
+ unsigned long org_misc, pid_misc;
+
+ /* remember pid/way until we return */
+ get_misc_and_pid(&org_misc, &pid_misc);
+
+ for (line = 0; line < cpuinfo.tlb_num_lines; line++) {
+ WRCTL(CTL_PTEADDR, line << 2);
+
+ for (way = 0; way < cpuinfo.tlb_num_ways; way++) {
+ unsigned long pteaddr;
+ unsigned long tlbmisc;
+ unsigned long tlbacc;
+
+ tlbmisc = pid_misc | TLBMISC_RD |
+ (way << TLBMISC_WAY_SHIFT);
+ WRCTL(CTL_TLBMISC, tlbmisc);
+ pteaddr = RDCTL(CTL_PTEADDR);
+ tlbmisc = RDCTL(CTL_TLBMISC);
+ tlbacc = RDCTL(CTL_TLBACC);
+
+ if (((tlbmisc>>TLBMISC_PID_SHIFT) & TLBMISC_PID_MASK)
+ == pid) {
+ tlbmisc = pid_misc | TLBMISC_WE |
+ (way << TLBMISC_WAY_SHIFT);
+ WRCTL(CTL_TLBMISC, tlbmisc);
+ WRCTL(CTL_TLBACC,
+ (MAX_PHYS_ADDR >> PAGE_SHIFT));
+ }
+ }
+
+ WRCTL(CTL_TLBMISC, org_misc);
+ }
+}
+
+void flush_tlb_all(void)
+{
+ int i;
+ unsigned long vaddr = CONFIG_IO_REGION_BASE;
+ unsigned int way;
+ unsigned long org_misc, pid_misc, tlbmisc;
+
+ /* remember pid/way until we return */
+ get_misc_and_pid(&org_misc, &pid_misc);
+ pid_misc |= TLBMISC_WE;
+
+ /* Map each TLB entry to physcal address 0 with no-access and a
+ bad ptbase */
+ for (way = 0; way < cpuinfo.tlb_num_ways; way++) {
+ tlbmisc = pid_misc | (way << TLBMISC_WAY_SHIFT);
+ for (i = 0; i < cpuinfo.tlb_num_lines; i++) {
+ WRCTL(CTL_PTEADDR, ((vaddr) >> PAGE_SHIFT) << 2);
+ WRCTL(CTL_TLBMISC, tlbmisc);
+ WRCTL(CTL_TLBACC, (MAX_PHYS_ADDR >> PAGE_SHIFT));
+ vaddr += 1UL << 12;
+ }
+ }
+
+ /* restore pid/way */
+ WRCTL(CTL_TLBMISC, org_misc);
+}
+
+void set_mmu_pid(unsigned long pid)
+{
+ WRCTL(CTL_TLBMISC, (RDCTL(CTL_TLBMISC) & TLBMISC_WAY) |
+ ((pid & TLBMISC_PID_MASK) << TLBMISC_PID_SHIFT));
+}
--
1.8.3.2

2014-04-18 12:28:31

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 15/28] nios2: ELF definitions

This patch adds definitions for the ELF format

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/elf.h | 101 ++++++++++++++++++++++++++++++++++++++
arch/nios2/include/uapi/asm/elf.h | 67 +++++++++++++++++++++++++
2 files changed, 168 insertions(+)
create mode 100644 arch/nios2/include/asm/elf.h
create mode 100644 arch/nios2/include/uapi/asm/elf.h

diff --git a/arch/nios2/include/asm/elf.h b/arch/nios2/include/asm/elf.h
new file mode 100644
index 0000000..6c5c653
--- /dev/null
+++ b/arch/nios2/include/asm/elf.h
@@ -0,0 +1,101 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_ELF_H
+#define _ASM_NIOS2_ELF_H
+
+#include <uapi/asm/elf.h>
+
+/*
+ * This is used to ensure we don't load something for the wrong architecture.
+ */
+#define elf_check_arch(x) ((x)->e_machine == EM_ALTERA_NIOS2)
+
+#define ELF_PLAT_INIT(_r, load_addr)
+
+#define USE_ELF_CORE_DUMP
+#define ELF_EXEC_PAGESIZE 4096
+
+/* This is the location that an ET_DYN program is loaded if exec'ed. Typical
+ use of this is to invoke "./ld.so someprog" to test out a new version of
+ the loader. We need to make sure that it is out of the way of the program
+ that it will "exec", and that there is sufficient room for the brk. */
+
+#define ELF_ET_DYN_BASE 0xD0000000UL
+
+/* regs is struct pt_regs, pr_reg is elf_gregset_t (which is
+ now struct_user_regs, they are different) */
+
+#define ARCH_HAS_SETUP_ADDITIONAL_PAGES 1
+struct linux_binprm;
+extern int arch_setup_additional_pages(struct linux_binprm *bprm,
+ int uses_interp);
+#define ELF_CORE_COPY_REGS(pr_reg, regs) \
+{ do { \
+ /* Bleech. */ \
+ pr_reg[0] = regs->r8; \
+ pr_reg[1] = regs->r9; \
+ pr_reg[2] = regs->r10; \
+ pr_reg[3] = regs->r11; \
+ pr_reg[4] = regs->r12; \
+ pr_reg[5] = regs->r13; \
+ pr_reg[6] = regs->r14; \
+ pr_reg[7] = regs->r15; \
+ pr_reg[8] = regs->r1; \
+ pr_reg[9] = regs->r2; \
+ pr_reg[10] = regs->r3; \
+ pr_reg[11] = regs->r4; \
+ pr_reg[12] = regs->r5; \
+ pr_reg[13] = regs->r6; \
+ pr_reg[14] = regs->r7; \
+ pr_reg[15] = regs->orig_r2; \
+ pr_reg[16] = regs->ra; \
+ pr_reg[17] = regs->fp; \
+ pr_reg[18] = regs->sp; \
+ pr_reg[19] = regs->gp; \
+ pr_reg[20] = regs->estatus; \
+ pr_reg[21] = regs->ea; \
+ pr_reg[22] = regs->orig_r7; \
+ { \
+ struct switch_stack *sw = ((struct switch_stack *)regs) - 1; \
+ pr_reg[23] = sw->r16; \
+ pr_reg[24] = sw->r17; \
+ pr_reg[25] = sw->r18; \
+ pr_reg[26] = sw->r19; \
+ pr_reg[27] = sw->r20; \
+ pr_reg[28] = sw->r21; \
+ pr_reg[29] = sw->r22; \
+ pr_reg[30] = sw->r23; \
+ pr_reg[31] = sw->fp; \
+ pr_reg[32] = sw->gp; \
+ pr_reg[33] = sw->ra; \
+ } \
+} while (0); }
+
+/* This yields a mask that user programs can use to figure out what
+ instruction set this cpu supports. */
+
+#define ELF_HWCAP (0)
+
+/* This yields a string that ld.so will use to load implementation
+ specific libraries for optimization. This is more specific in
+ intent than poking at uname or /proc/cpuinfo. */
+
+#define ELF_PLATFORM (NULL)
+
+#endif /* _ASM_NIOS2_ELF_H */
diff --git a/arch/nios2/include/uapi/asm/elf.h b/arch/nios2/include/uapi/asm/elf.h
new file mode 100644
index 0000000..a5b91ae
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/elf.h
@@ -0,0 +1,67 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+
+#ifndef _UAPI_ASM_NIOS2_ELF_H
+#define _UAPI_ASM_NIOS2_ELF_H
+
+#include <linux/ptrace.h>
+
+/* Relocation types */
+#define R_NIOS2_NONE 0
+#define R_NIOS2_S16 1
+#define R_NIOS2_U16 2
+#define R_NIOS2_PCREL16 3
+#define R_NIOS2_CALL26 4
+#define R_NIOS2_IMM5 5
+#define R_NIOS2_CACHE_OPX 6
+#define R_NIOS2_IMM6 7
+#define R_NIOS2_IMM8 8
+#define R_NIOS2_HI16 9
+#define R_NIOS2_LO16 10
+#define R_NIOS2_HIADJ16 11
+#define R_NIOS2_BFD_RELOC_32 12
+#define R_NIOS2_BFD_RELOC_16 13
+#define R_NIOS2_BFD_RELOC_8 14
+#define R_NIOS2_GPREL 15
+#define R_NIOS2_GNU_VTINHERIT 16
+#define R_NIOS2_GNU_VTENTRY 17
+#define R_NIOS2_UJMP 18
+#define R_NIOS2_CJMP 19
+#define R_NIOS2_CALLR 20
+#define R_NIOS2_ALIGN 21
+/* Keep this the last entry. */
+#define R_NIOS2_NUM 22
+
+typedef unsigned long elf_greg_t;
+
+#define ELF_NGREG \
+ ((sizeof(struct pt_regs) + sizeof(struct switch_stack)) / \
+ sizeof(elf_greg_t))
+typedef elf_greg_t elf_gregset_t[ELF_NGREG];
+
+typedef unsigned long elf_fpregset_t;
+
+/*
+ * These are used to set parameters in the core dumps.
+ */
+#define ELF_CLASS ELFCLASS32
+#define ELF_DATA ELFDATA2LSB
+#define ELF_ARCH EM_ALTERA_NIOS2
+
+#endif /* _UAPI_ASM_NIOS2_ELF_H */
--
1.8.3.2

2014-04-18 12:28:47

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 23/28] nios2: Nios2 registers

This file contains constants for the instruction macros, cpu registers, fields and bits.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/registers.h | 65 ++++++++++++++++++++++++++++++++++++++
1 file changed, 65 insertions(+)
create mode 100644 arch/nios2/include/asm/registers.h

diff --git a/arch/nios2/include/asm/registers.h b/arch/nios2/include/asm/registers.h
new file mode 100644
index 0000000..0e6ffef08b20e072480ca3780e2626eb3f1de2bd
--- /dev/null
+++ b/arch/nios2/include/asm/registers.h
@@ -0,0 +1,65 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_REGISTERS_H
+#define _ASM_NIOS2_REGISTERS_H
+
+/* control register numbers */
+#define CTL_STATUS 0
+#define CTL_ESTATUS 1
+#define CTL_BSTATUS 2
+#define CTL_IENABLE 3
+#define CTL_IPENDING 4
+#define CTL_CPUID 5
+#define CTL_RSV1 6
+#define CTL_EXCEPTION 7
+#define CTL_PTEADDR 8
+#define CTL_TLBACC 9
+#define CTL_TLBMISC 10
+#define CTL_RSV2 11
+#define CTL_BADADDR 12
+#define CTL_CONFIG 13
+#define CTL_MPUBASE 14
+#define CTL_MPUACC 15
+
+/* access control registers using GCC builtins */
+#define RDCTL(r) __builtin_rdctl(r)
+#define WRCTL(r, v) __builtin_wrctl(r, v)
+
+/* status register bits */
+#define STATUS_PIE (1 << 0) /* processor interrupt enable */
+#define STATUS_U (1 << 1) /* user mode */
+#define STATUS_EH (1 << 2) /* Exception mode */
+
+/* estatus register bits */
+#define ESTATUS_EPIE (1 << 0) /* processor interrupt enable */
+#define ESTATUS_EU (1 << 1) /* user mode */
+#define ESTATUS_EH (1 << 2) /* Exception mode */
+
+/* tlbmisc register bits */
+#define TLBMISC_PID_SHIFT 4
+#define TLBMISC_PID_MASK ((1UL << cpuinfo.tlb_pid_num_bits) - 1)
+#define TLBMISC_WAY_MASK 0xf
+#define TLBMISC_WAY_SHIFT 20
+
+#define TLBMISC_PID (TLBMISC_PID_MASK << TLBMISC_PID_SHIFT) /* TLB PID */
+#define TLBMISC_WE (1 << 18) /* TLB write enable */
+#define TLBMISC_RD (1 << 19) /* TLB read */
+#define TLBMISC_WAY (TLBMISC_WAY_MASK << TLBMISC_WAY_SHIFT) /* TLB way */
+
+#endif /* _ASM_NIOS2_REGISTERS_H */
--
1.8.3.2

2014-04-18 12:28:36

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 17/28] nios2: Signal handling support

This patch adds support for signal handling.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/signal.h | 22 ++
arch/nios2/include/asm/ucontext.h | 34 +++
arch/nios2/include/uapi/asm/sigcontext.h | 30 +++
arch/nios2/include/uapi/asm/signal.h | 23 ++
arch/nios2/kernel/signal.c | 364 +++++++++++++++++++++++++++++++
5 files changed, 473 insertions(+)
create mode 100644 arch/nios2/include/asm/signal.h
create mode 100644 arch/nios2/include/asm/ucontext.h
create mode 100644 arch/nios2/include/uapi/asm/sigcontext.h
create mode 100644 arch/nios2/include/uapi/asm/signal.h
create mode 100644 arch/nios2/kernel/signal.c

diff --git a/arch/nios2/include/asm/signal.h b/arch/nios2/include/asm/signal.h
new file mode 100644
index 0000000..bbcf11eecb0194b7c7b0f9352b0154d67b2dc06a
--- /dev/null
+++ b/arch/nios2/include/asm/signal.h
@@ -0,0 +1,22 @@
+/*
+ * Copyright Altera Corporation (C) 2013. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+#ifndef _NIOS2_SIGNAL_H
+#define _NIOS2_SIGNAL_H
+
+#include <uapi/asm/signal.h>
+
+#endif /* _NIOS2_SIGNAL_H */
diff --git a/arch/nios2/include/asm/ucontext.h b/arch/nios2/include/asm/ucontext.h
new file mode 100644
index 0000000..5870ef4e92fbaec9137142534ce3c5b23705d43e
--- /dev/null
+++ b/arch/nios2/include/asm/ucontext.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * derived from m68knommu
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_UCONTEXT_H
+#define _ASM_NIOS2_UCONTEXT_H
+
+typedef int greg_t;
+#define NGREG 32
+typedef greg_t gregset_t[NGREG];
+
+struct mcontext {
+ int version;
+ gregset_t gregs;
+};
+
+#define MCONTEXT_VERSION 2
+
+struct ucontext {
+ unsigned long uc_flags;
+ struct ucontext *uc_link;
+ stack_t uc_stack;
+ struct mcontext uc_mcontext;
+ sigset_t uc_sigmask; /* mask last for extensibility */
+};
+
+#endif
diff --git a/arch/nios2/include/uapi/asm/sigcontext.h b/arch/nios2/include/uapi/asm/sigcontext.h
new file mode 100644
index 0000000..6bfd8807597a7d6abc0ae32f030e44cb4f77809f
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/sigcontext.h
@@ -0,0 +1,30 @@
+/*
+ * Taken from the m68knommu.
+ *
+ * Copyright (C) 2004, Microtronix Datacom Ltd.
+ *
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT. See the GNU General Public License for more
+ * details.
+ */
+
+#ifndef _ASM_NIOS2_SIGCONTEXT_H
+#define _ASM_NIOS2_SIGCONTEXT_H
+
+#include <asm/ptrace.h>
+
+struct sigcontext {
+ struct pt_regs regs;
+ unsigned long sc_mask; /* old sigmask */
+};
+
+#endif
diff --git a/arch/nios2/include/uapi/asm/signal.h b/arch/nios2/include/uapi/asm/signal.h
new file mode 100644
index 0000000..f29ee63144817e9798965c0b8279c08844c44e6a
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/signal.h
@@ -0,0 +1,23 @@
+/*
+ * Copyright Altera Corporation (C) 2013. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+#ifndef _ASM_NIOS2_SIGNAL_H
+#define _ASM_NIOS2_SIGNAL_H
+
+#define SA_RESTORER 0x04000000
+#include <asm-generic/signal.h>
+
+#endif /* _ASM_NIOS2_SIGNAL_H */
diff --git a/arch/nios2/kernel/signal.c b/arch/nios2/kernel/signal.c
new file mode 100644
index 0000000..1c5cbe9c55dde1c3467d727ec66c36b96f3cb9cd
--- /dev/null
+++ b/arch/nios2/kernel/signal.c
@@ -0,0 +1,364 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2011-2012 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ * Copyright (C) 1991, 1992 Linus Torvalds
+ *
+ * This file is based on kernel/signal.c from m68knommu.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file COPYING in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/signal.h>
+#include <linux/errno.h>
+#include <linux/ptrace.h>
+#include <linux/uaccess.h>
+#include <linux/unistd.h>
+#include <linux/personality.h>
+#include <linux/tracehook.h>
+
+#include <asm/ucontext.h>
+#include <asm/cacheflush.h>
+
+#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
+
+static int do_signal(struct pt_regs *regs, sigset_t *oldset, int in_syscall);
+
+/*
+ * Do a signal return; undo the signal stack.
+ *
+ * Keep the return code on the stack quadword aligned!
+ * That makes the cache flush below easier.
+ */
+
+struct sigframe {
+ char retcode[12];
+ unsigned long extramask[_NSIG_WORDS-1];
+ struct sigcontext sc;
+};
+
+struct rt_sigframe {
+ char retcode[12];
+ struct siginfo info;
+ struct ucontext uc;
+};
+
+static inline int rt_restore_ucontext(struct pt_regs *regs,
+ struct switch_stack *sw,
+ struct ucontext *uc, int *pr2)
+{
+ int temp;
+ greg_t *gregs = uc->uc_mcontext.gregs;
+ int err;
+
+ err = __get_user(temp, &uc->uc_mcontext.version);
+ if (temp != MCONTEXT_VERSION)
+ goto badframe;
+ /* restore passed registers */
+ err |= __get_user(regs->r1, &gregs[0]);
+ err |= __get_user(regs->r2, &gregs[1]);
+ err |= __get_user(regs->r3, &gregs[2]);
+ err |= __get_user(regs->r4, &gregs[3]);
+ err |= __get_user(regs->r5, &gregs[4]);
+ err |= __get_user(regs->r6, &gregs[5]);
+ err |= __get_user(regs->r7, &gregs[6]);
+ err |= __get_user(regs->r8, &gregs[7]);
+ err |= __get_user(regs->r9, &gregs[8]);
+ err |= __get_user(regs->r10, &gregs[9]);
+ err |= __get_user(regs->r11, &gregs[10]);
+ err |= __get_user(regs->r12, &gregs[11]);
+ err |= __get_user(regs->r13, &gregs[12]);
+ err |= __get_user(regs->r14, &gregs[13]);
+ err |= __get_user(regs->r15, &gregs[14]);
+ err |= __get_user(sw->r16, &gregs[15]);
+ err |= __get_user(sw->r17, &gregs[16]);
+ err |= __get_user(sw->r18, &gregs[17]);
+ err |= __get_user(sw->r19, &gregs[18]);
+ err |= __get_user(sw->r20, &gregs[19]);
+ err |= __get_user(sw->r21, &gregs[20]);
+ err |= __get_user(sw->r22, &gregs[21]);
+ err |= __get_user(sw->r23, &gregs[22]);
+ /* gregs[23] is handled below */
+ err |= __get_user(sw->fp, &gregs[24]); /* Verify, should this be
+ settable */
+ err |= __get_user(sw->gp, &gregs[25]); /* Verify, should this be
+ settable */
+
+ err |= __get_user(temp, &gregs[26]); /* Not really necessary no user
+ settable bits */
+ err |= __get_user(regs->ea, &gregs[27]);
+
+ err |= __get_user(regs->ra, &gregs[23]);
+ err |= __get_user(regs->sp, &gregs[28]);
+
+ regs->estatus = (regs->estatus & 0xffffffff);
+ regs->orig_r2 = -1; /* disable syscall checks */
+
+ err |= restore_altstack(&uc->uc_stack);
+ if (err)
+ goto badframe;
+
+ *pr2 = regs->r2;
+ return err;
+
+badframe:
+ return 1;
+}
+
+asmlinkage int do_rt_sigreturn(struct switch_stack *sw)
+{
+ struct pt_regs *regs = (struct pt_regs *)(sw + 1);
+ /* Verify, can we follow the stack back */
+ struct rt_sigframe *frame = (struct rt_sigframe *) regs->sp;
+ sigset_t set;
+ int rval;
+
+ if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
+ goto badframe;
+
+ if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set)))
+ goto badframe;
+
+ sigdelsetmask(&set, ~_BLOCKABLE);
+ spin_lock_irq(&current->sighand->siglock);
+ current->blocked = set;
+ recalc_sigpending();
+ spin_unlock_irq(&current->sighand->siglock);
+
+ if (rt_restore_ucontext(regs, sw, &frame->uc, &rval))
+ goto badframe;
+
+ return rval;
+
+badframe:
+ force_sig(SIGSEGV, current);
+ return 0;
+}
+
+static inline int rt_setup_ucontext(struct ucontext *uc, struct pt_regs *regs)
+{
+ struct switch_stack *sw = (struct switch_stack *)regs - 1;
+ greg_t *gregs = uc->uc_mcontext.gregs;
+ int err = 0;
+
+ err |= __put_user(MCONTEXT_VERSION, &uc->uc_mcontext.version);
+ err |= __put_user(regs->r1, &gregs[0]);
+ err |= __put_user(regs->r2, &gregs[1]);
+ err |= __put_user(regs->r3, &gregs[2]);
+ err |= __put_user(regs->r4, &gregs[3]);
+ err |= __put_user(regs->r5, &gregs[4]);
+ err |= __put_user(regs->r6, &gregs[5]);
+ err |= __put_user(regs->r7, &gregs[6]);
+ err |= __put_user(regs->r8, &gregs[7]);
+ err |= __put_user(regs->r9, &gregs[8]);
+ err |= __put_user(regs->r10, &gregs[9]);
+ err |= __put_user(regs->r11, &gregs[10]);
+ err |= __put_user(regs->r12, &gregs[11]);
+ err |= __put_user(regs->r13, &gregs[12]);
+ err |= __put_user(regs->r14, &gregs[13]);
+ err |= __put_user(regs->r15, &gregs[14]);
+ err |= __put_user(sw->r16, &gregs[15]);
+ err |= __put_user(sw->r17, &gregs[16]);
+ err |= __put_user(sw->r18, &gregs[17]);
+ err |= __put_user(sw->r19, &gregs[18]);
+ err |= __put_user(sw->r20, &gregs[19]);
+ err |= __put_user(sw->r21, &gregs[20]);
+ err |= __put_user(sw->r22, &gregs[21]);
+ err |= __put_user(sw->r23, &gregs[22]);
+ err |= __put_user(regs->ra, &gregs[23]);
+ err |= __put_user(sw->fp, &gregs[24]);
+ err |= __put_user(sw->gp, &gregs[25]);
+ err |= __put_user(regs->ea, &gregs[27]);
+ err |= __put_user(regs->sp, &gregs[28]);
+ return err;
+}
+
+static inline void push_cache(unsigned long vaddr)
+{
+ flush_dcache_range(vaddr, vaddr + 12);
+ flush_icache_range(vaddr, vaddr + 12);
+}
+
+static inline void *get_sigframe(struct k_sigaction *ka, struct pt_regs *regs,
+ size_t frame_size)
+{
+ unsigned long usp;
+
+ /* Default to using normal stack. */
+ usp = regs->sp;
+
+ /* This is the X/Open sanctioned signal stack switching. */
+ if ((ka->sa.sa_flags & SA_ONSTACK) && (current->sas_ss_sp != 0)) {
+ if (!on_sig_stack(usp))
+ usp = current->sas_ss_sp + current->sas_ss_size;
+ }
+
+ /* Verify, is it 32 or 64 bit aligned */
+ return (void *)((usp - frame_size) & -8UL);
+}
+
+static void setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
+ sigset_t *set, struct pt_regs *regs)
+{
+ struct rt_sigframe *frame;
+ int err = 0;
+
+ frame = get_sigframe(ka, regs, sizeof(*frame));
+
+ if (info)
+ err |= copy_siginfo_to_user(&frame->info, info);
+
+ /* Create the ucontext. */
+ err |= __put_user(0, &frame->uc.uc_flags);
+ err |= __put_user(0, &frame->uc.uc_link);
+ err |= __save_altstack(&frame->uc.uc_stack, regs->sp);
+ err |= rt_setup_ucontext(&frame->uc, regs);
+ err |= copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set));
+
+ /* Set up to return from userspace. */
+ regs->ra = (unsigned long) &frame->retcode[0];
+
+ /* movi r2,__NR_rt_sigreturn */
+ err |= __put_user(0x00800004 + (__NR_rt_sigreturn << 6),
+ (long *)(frame->retcode));
+ /* trap */
+ err |= __put_user(0x003b683a, (long *)(frame->retcode + 4));
+
+ if (err)
+ goto give_sigsegv;
+
+ push_cache((unsigned long) &frame->retcode);
+
+ /* Set up registers for signal handler */
+ regs->sp = (unsigned long) frame;
+ regs->r4 = (unsigned long) (current_thread_info()->exec_domain
+ && current_thread_info()->exec_domain->signal_invmap
+ && sig < 32
+ ? current_thread_info()->exec_domain->signal_invmap[sig]
+ : sig);
+ regs->r5 = (unsigned long) &frame->info;
+ regs->r6 = (unsigned long) &frame->uc;
+ regs->ea = (unsigned long) ka->sa.sa_handler;
+ return;
+
+give_sigsegv:
+ force_sigsegv(sig, current);
+}
+
+static inline void handle_restart(struct pt_regs *regs, struct k_sigaction *ka,
+ int has_handler)
+{
+ switch (regs->r2) {
+ case ERESTART_RESTARTBLOCK:
+ case ERESTARTNOHAND:
+ regs->r2 = EINTR;
+ regs->r7 = 1;
+ break;
+ case ERESTARTSYS:
+ if (has_handler && !(ka->sa.sa_flags & SA_RESTART)) {
+ regs->r2 = EINTR;
+ regs->r7 = 1;
+ break;
+ }
+ /* fallthrough */
+ case ERESTARTNOINTR:
+ regs->r2 = regs->orig_r2;
+ regs->r7 = regs->orig_r7;
+ regs->ea -= 4;
+ break;
+ }
+}
+
+/*
+ * OK, we're invoking a handler
+ */
+static void handle_signal(int sig, struct k_sigaction *ka, siginfo_t *info,
+ sigset_t *oldset, struct pt_regs *regs)
+{
+ /* set up the stack frame */
+ if (ka->sa.sa_flags & SA_SIGINFO)
+ setup_rt_frame(sig, ka, info, oldset, regs);
+ else
+ setup_rt_frame(sig, ka, NULL, oldset, regs);
+
+ if (!(ka->sa.sa_flags & SA_NODEFER)) {
+ spin_lock_irq(&current->sighand->siglock);
+ sigorsets(&current->blocked, &current->blocked,
+ &ka->sa.sa_mask);
+ sigaddset(&current->blocked, sig);
+ recalc_sigpending();
+ spin_unlock_irq(&current->sighand->siglock);
+ }
+}
+
+/*
+ * Note that 'init' is a special process: it doesn't get signals it doesn't
+ * want to handle. Thus you cannot kill init even with a SIGKILL even by
+ * mistake.
+ */
+static int do_signal(struct pt_regs *regs, sigset_t *oldset, int in_syscall)
+{
+ struct k_sigaction ka;
+ siginfo_t info;
+ int signr;
+
+ current->thread.kregs = regs;
+
+ if (!oldset)
+ oldset = &current->blocked;
+
+ signr = get_signal_to_deliver(&info, &ka, regs, NULL);
+ if (signr > 0) {
+ /*
+ * Are we from a system call? If so, check system call
+ * restarting.
+ */
+ if (in_syscall)
+ handle_restart(regs, &ka, 1);
+ /* Whee! Actually deliver the signal. */
+ handle_signal(signr, &ka, &info, oldset, regs);
+ return 1;
+ }
+
+ /*
+ * No signal to deliver to the process - restart the syscall.
+ */
+ if (in_syscall) {
+ /* Did the syscall return an error code */
+ if (regs->r7 == 1) {
+ if (regs->r2 == ERESTARTNOHAND ||
+ regs->r2 == ERESTARTSYS ||
+ regs->r2 == ERESTARTNOINTR) {
+ regs->r2 = regs->orig_r2;
+ regs->r7 = regs->orig_r7;
+ regs->ea -= 4;
+ } else if (regs->r2 == ERESTART_RESTARTBLOCK) {
+ regs->r2 = __NR_restart_syscall;
+ regs->ea -= 4;
+ }
+ }
+ }
+
+ return 0;
+}
+
+asmlinkage void do_notify_resume(struct pt_regs *regs, sigset_t *oldset,
+ int in_syscall)
+{
+ pr_debug("--> ENTERING %s\n", __func__);
+ /*
+ * We want the common case to go fast, which is why we may in certain
+ * cases get here from kernel mode. Just return without doing anything
+ * if so.
+ */
+ if (!user_mode(regs))
+ return;
+
+ if (test_thread_flag(TIF_SIGPENDING))
+ do_signal(regs, oldset, in_syscall);
+
+ if (test_and_clear_thread_flag(TIF_NOTIFY_RESUME))
+ tracehook_notify_resume(regs);
+}
--
1.8.3.2

2014-04-18 12:29:36

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 28/28] Documentation: Add documentation for Nios2 architecture

Signed-off-by: Ley Foon Tan <[email protected]>
---
Documentation/nios2/README | 23 +++++++++++++++++++++++
1 file changed, 23 insertions(+)
create mode 100644 Documentation/nios2/README

diff --git a/Documentation/nios2/README b/Documentation/nios2/README
new file mode 100644
index 0000000..33eec4d
--- /dev/null
+++ b/Documentation/nios2/README
@@ -0,0 +1,23 @@
+Linux on the Nios II architecture
+=================================
+
+This is a port of Linux to Nios II (nios2) processor.
+
+In order to compile for Nios II, you need a version of GCC with support for the generic
+system call ABI. Please see this link for more information on how compiling and booting
+software for the Nios II platform:
+http://www.rocketboards.org/foswiki/Documentation/NiosIILinuxUserManual
+
+For reference, please see the following link:
+http://www.altera.com/literature/lit-nio2.jsp
+
+What is Nios II?
+================
+Nios II is a 32-bit embedded-processor architecture designed specifically for the
+Altera family of FPGAs. In order to support Linux system, Nios II needs to configure
+with MMU and hardware multiplier enabled.
+
+Nios II ABI
+===========
+Please refer to chapter "Application Binary Interface" in Nios II Processor Reference
+Handbook.
--
1.8.3.2

2014-04-18 12:42:14

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 04/28] nios2: Exception handling

This patch contains the exception entry code (kernel/entry.S) and misaligned exception.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/kernel/entry.S | 551 +++++++++++++++++++++++++++++++++++++++++
arch/nios2/kernel/misaligned.c | 319 ++++++++++++++++++++++++
2 files changed, 870 insertions(+)
create mode 100644 arch/nios2/kernel/entry.S
create mode 100644 arch/nios2/kernel/misaligned.c

diff --git a/arch/nios2/kernel/entry.S b/arch/nios2/kernel/entry.S
new file mode 100644
index 0000000..cda0854d1d29450325fa90ab62ced2c215035f3c
--- /dev/null
+++ b/arch/nios2/kernel/entry.S
@@ -0,0 +1,551 @@
+/*
+ * linux/arch/nios2/kernel/entry.S
+ *
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2009, Wind River Systems Inc
+ *
+ * Implemented by [email protected] and [email protected]
+ *
+ * Based on:
+ *
+ * linux/arch/nios2/kernel/entry.S
+ *
+ * Copyright (C) 1999-2002, Greg Ungerer ([email protected])
+ * Copyright (C) 1998 D. Jeff Dionne <[email protected]>,
+ * Kenneth Albanowski <[email protected]>,
+ * Copyright (C) 2000 Lineo Inc. (http://www.lineo.com)
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * Based on:
+ *
+ * linux/arch/m68knommu/kernel/entry.S
+ *
+ * Copyright (C) 1991, 1992 Linus Torvalds
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Linux/m68k support by Hamish Macdonald
+ *
+ * 68060 fixes by Jesper Skov
+ * ColdFire support by Greg Ungerer ([email protected])
+ * 5307 fixes by David W. Miller
+ * linux 2.4 support David McCullough <[email protected]>
+ */
+
+#include <linux/sys.h>
+#include <linux/linkage.h>
+#include <asm/asm-offsets.h>
+#include <asm/asm-macros.h>
+#include <asm/thread_info.h>
+#include <asm/errno.h>
+#include <asm/setup.h>
+#include <asm/entry.h>
+#include <asm/unistd.h>
+#include <asm/processor.h>
+
+.macro GET_THREAD_INFO reg
+.if THREAD_SIZE & 0xffff0000
+ andhi \reg, sp, %hi(~(THREAD_SIZE-1))
+.else
+ addi \reg, r0, %lo(~(THREAD_SIZE-1))
+ and \reg, \reg, sp
+.endif
+.endm
+
+.macro kuser_cmpxchg_check
+ /*
+ * Make sure our user space atomic helper is restarted if it was
+ * interrupted in a critical region.
+ * ea-4 = address of interrupted insn (ea must be preserved).
+ * sp = saved regs.
+ * cmpxchg_ldw = first critical insn, cmpxchg_stw = last critical insn.
+ * If ea <= cmpxchg_stw and ea > cmpxchg_ldw then saved EA is set to
+ * cmpxchg_ldw + 4.
+ */
+ /* et = cmpxchg_stw + 4 */
+ movui et, (KUSER_BASE + 4 + (cmpxchg_stw - __kuser_helper_start))
+ bgtu ea, et, 1f
+
+ subi et, et, (cmpxchg_stw - cmpxchg_ldw) /* et = cmpxchg_ldw + 4 */
+ bltu ea, et, 1f
+ stw et, PT_EA(sp) /* fix up EA */
+ mov ea, et
+1:
+.endm
+
+.section .rodata
+.align 4
+exception_table:
+ .word unhandled_exception /* 0 - Reset */
+ .word unhandled_exception /* 1 - Processor-only Reset */
+ .word external_interrupt /* 2 - Interrupt */
+ .word handle_trap /* 3 - Trap Instruction */
+
+ .word instruction_trap /* 4 - Unimplemented instruction */
+ .word handle_illegal /* 5 - Illegal instruction */
+ .word handle_unaligned /* 6 - Misaligned data access */
+ .word handle_unaligned /* 7 - Misaligned destination address */
+
+ .word handle_diverror /* 8 - Division error */
+ .word protection_exception_ba /* 9 - Supervisor-only instr. address */
+ .word protection_exception_instr /* 10 - Supervisor only instruction */
+ .word protection_exception_ba /* 11 - Supervisor only data address */
+
+ .word unhandled_exception /* 12 - Double TLB miss (data) */
+ .word protection_exception_pte /* 13 - TLB permission violation (x) */
+ .word protection_exception_pte /* 14 - TLB permission violation (r) */
+ .word protection_exception_pte /* 15 - TLB permission violation (w) */
+
+ .word unhandled_exception /* 16 - MPU region violation */
+
+trap_table:
+ .word handle_system_call /* 0 */
+ .word instruction_trap /* 1 */
+ .word instruction_trap /* 2 */
+ .word instruction_trap /* 3 */
+ .word instruction_trap /* 4 */
+ .word instruction_trap /* 5 */
+ .word instruction_trap /* 6 */
+ .word instruction_trap /* 7 */
+ .word instruction_trap /* 8 */
+ .word instruction_trap /* 9 */
+ .word instruction_trap /* 10 */
+ .word instruction_trap /* 11 */
+ .word instruction_trap /* 12 */
+ .word instruction_trap /* 13 */
+ .word instruction_trap /* 14 */
+ .word instruction_trap /* 15 */
+ .word instruction_trap /* 16 */
+ .word instruction_trap /* 17 */
+ .word instruction_trap /* 18 */
+ .word instruction_trap /* 19 */
+ .word instruction_trap /* 20 */
+ .word instruction_trap /* 21 */
+ .word instruction_trap /* 22 */
+ .word instruction_trap /* 23 */
+ .word instruction_trap /* 24 */
+ .word instruction_trap /* 25 */
+ .word instruction_trap /* 26 */
+ .word instruction_trap /* 27 */
+ .word instruction_trap /* 28 */
+ .word instruction_trap /* 29 */
+ .word instruction_trap /* 30 */
+ .word handle_breakpoint /* 31 */
+
+.text
+.set noat
+.set nobreak
+
+ENTRY(inthandler)
+ SAVE_ALL
+
+ kuser_cmpxchg_check
+
+ /* Clear EH bit before we get a new excpetion in the kernel
+ * and after we have saved it to the exception frame. This is done
+ * wheter it's trap, tlb-miss or interrupt. If we don't do this
+ * estatus is not updated the next exception.
+ */
+ rdctl r24, status
+ movi r9, %lo(~STATUS_EH)
+ and r24, r24, r9
+ wrctl status, r24
+
+ /* Read cause and vector and branch to the associated handler */
+ mov r4, sp
+ rdctl r5, exception
+ movia r9, exception_table
+ add r24, r9, r5
+ ldw r24, 0(r24)
+ jmp r24
+
+
+/***********************************************************************
+ * Handle traps
+ ***********************************************************************
+ */
+ENTRY(handle_trap)
+ ldw r24, -4(ea) /* instruction that caused the exception */
+ srli r24, r24, 4
+ andi r24, r24, 0x7c
+ movia r9,trap_table
+ add r24, r24, r9
+ ldw r24, 0(r24)
+ jmp r24
+
+
+/***********************************************************************
+ * Handle system calls
+ ***********************************************************************
+ */
+ENTRY(handle_system_call)
+ /* Enable interrupts */
+ rdctl r10, status
+ ori r10, r10, STATUS_PIE
+ wrctl status, r10
+
+ /* Reload registers destroyed by common code. */
+ ldw r4, PT_R4(sp)
+ ldw r5, PT_R5(sp)
+
+ /* Check that the requested system call is within limits */
+ movui r1, __NR_syscalls
+ bgeu r2, r1, ret_invsyscall
+ slli r1, r2, 2
+ movhi r11, %hiadj(sys_call_table)
+ add r1, r1, r11
+ ldw r1, %lo(sys_call_table)(r1)
+ beq r1, r0, ret_invsyscall
+
+ /* Get thread info pointer */
+ movi r11, %lo(0xfffff000)
+ and r11, sp, r11
+ ldw r11, TI_FLAGS(r11)
+
+ /* If someone is ptrace:ing us, take the long way. */
+ BTBNZ r11, r11, TIF_SYSCALL_TRACE, traced_system_call
+
+ /* Execute the system call */
+ callr r1
+
+ /* If the syscall returns a negative result:
+ * Set r7 to 1 to indicate error,
+ * Negate r2 to get a positive error code
+ * If the syscall returns zero or a positive value:
+ * Set r7 to 0.
+ * The sigreturn system calls will skip the code below by
+ * adding to register ra. To avoid destroying registers
+ */
+translate_rc_and_ret:
+ movi r1, 0
+ bge r2, zero, 3f
+ sub r2, zero, r2
+ movi r1, 1
+3:
+ stw r2, PT_R2(sp)
+ stw r1, PT_R7(sp)
+end_translate_rc_and_ret:
+
+ret_from_exception:
+ ldw r1, PT_ESTATUS(sp)
+ /* if so, skip resched, signals */
+ TSTBNZ r1, r1, ESTATUS_EU, Luser_return
+
+restore_all:
+ rdctl r10, status /* disable intrs */
+ andi r10, r10, %lo(~STATUS_PIE)
+ wrctl status, r10
+ RESTORE_ALL
+ eret
+
+ /* If the syscall number was invalid return ENOSYS */
+ret_invsyscall:
+ movi r2, -ENOSYS
+ br translate_rc_and_ret
+
+ /* This implements the same as above, except it calls
+ * syscall_trace before and after the syscall in order
+ * for utilities like strace and gdb to work.
+ */
+traced_system_call:
+ SAVE_SWITCH_STACK
+ call syscall_trace
+ RESTORE_SWITCH_STACK
+
+ /* Create system call register arguments. The 5th and 6th
+ arguments on stack are already in place at the beginning
+ of pt_regs. */
+ ldw r2, PT_R2(sp)
+ ldw r4, PT_R4(sp)
+ ldw r5, PT_R5(sp)
+ ldw r6, PT_R6(sp)
+ ldw r7, PT_R7(sp)
+
+ /* Fetch the syscall function, we don't need to check the boundaries
+ * since this is already done.
+ */
+ slli r1, r2, 2
+ movhi r11,%hiadj(sys_call_table)
+ add r1, r1, r11
+ ldw r1, %lo(sys_call_table)(r1)
+
+ callr r1
+
+ /* If the syscall returns a negative result:
+ * Set r7 to 1 to indicate error,
+ * Negate r2 to get a positive error code
+ * If the syscall returns zero or a positive value:
+ * Set r7 to 0.
+ * The sigreturn system calls will skip the code below by
+ * adding to register ra. To avoid destroying registers
+ */
+translate_rc_and_ret2:
+ movi r1, 0
+ bge r2, zero, 4f
+ sub r2, zero, r2
+ movi r1, 1
+4:
+ stw r2, PT_R2(sp)
+ stw r1, PT_R7(sp)
+end_translate_rc_and_ret2:
+ SAVE_SWITCH_STACK
+ call syscall_trace
+ RESTORE_SWITCH_STACK
+ br ret_from_exception
+
+Luser_return:
+ GET_THREAD_INFO r11 /* get thread_info pointer */
+ ldw r10, TI_FLAGS(r11) /* get thread_info->flags */
+ ANDI32 r11, r10, _TIF_WORK_MASK
+ beq r11, r0, restore_all /* Nothing to do */
+ BTBZ r1, r10, TIF_NEED_RESCHED, Lsignal_return
+
+ /* Reschedule work */
+ call schedule
+ br ret_from_exception
+
+Lsignal_return:
+ ANDI32 r1, r10, _TIF_SIGPENDING | _TIF_NOTIFY_RESUME
+ beq r1, r0, restore_all
+ mov r4, sp /* pt_regs */
+ SAVE_SWITCH_STACK
+ mov r5, r0 /* oldset = 0 */
+ movi r6, 1 /* in_syscall = 1 */
+ call do_notify_resume
+ RESTORE_SWITCH_STACK
+ br restore_all
+
+
+
+/***********************************************************************
+ * Handle external interrupts.
+ ***********************************************************************
+ */
+/*
+ * This is the generic interrupt handler (for all hardware interrupt
+ * sources). It figures out the vector number and calls the appropriate
+ * interrupt service routine directly.
+ */
+external_interrupt:
+ rdctl r12, ipending
+ rdctl r9, ienable
+ and r12, r12, r9
+ /* skip if no interrupt is pending */
+ beq r12, r0, ret_from_interrupt
+
+ movi r24, -1
+ stw r24, PT_ORIG_R2(sp)
+
+ /*
+ * Process an external hardware interrupt.
+ */
+
+ addi ea, ea, -4 /* re-issue the interrupted instruction */
+ stw ea, PT_EA(sp)
+2: movi r4, %lo(-1) /* Start from bit position 0,
+ highest priority */
+ /* This is the IRQ # for handler call */
+1: andi r10, r12, 1 /* Isolate bit we are interested in */
+ srli r12, r12, 1 /* shift count is costly without hardware
+ multiplier */
+ addi r4, r4, 1
+ beq r10, r0, 1b
+ mov r5, sp /* Setup pt_regs pointer for handler call */
+ call do_IRQ
+ rdctl r12, ipending /* check again if irq still pending */
+ rdctl r9, ienable /* Isolate possible interrupts */
+ and r12, r12, r9
+ bne r12, r0, 2b
+ /* br ret_from_interrupt */ /* fall throught to ret_from_interrupt */
+
+ENTRY(ret_from_interrupt)
+ ldw r1, PT_ESTATUS(sp) /* check if returning to kernel */
+ TSTBNZ r1, r1, ESTATUS_EU, Luser_return
+
+#ifdef CONFIG_PREEMPT
+ GET_THREAD_INFO r1
+ ldw r4, TI_PREEMPT_COUNT(r1)
+ bne r4, r0, restore_all
+
+need_resched:
+ ldw r4, TI_FLAGS(r1) /* ? Need resched set */
+ BTBZ r10, r4, TIF_NEED_RESCHED, restore_all
+ ldw r4, PT_ESTATUS(sp) /* ? Interrupts off */
+ andi r10, r4, ESTATUS_EPIE
+ beq r10, r0, restore_all
+ movia r4, PREEMPT_ACTIVE
+ stw r4, TI_PREEMPT_COUNT(r1)
+ rdctl r10, status /* enable intrs again */
+ ori r10, r10 ,STATUS_PIE
+ wrctl status, r10
+ PUSH r1
+ call schedule
+ POP r1
+ mov r4, r0
+ stw r4, TI_PREEMPT_COUNT(r1)
+ rdctl r10, status /* disable intrs */
+ andi r10, r10, %lo(~STATUS_PIE)
+ wrctl status, r10
+ br need_resched
+#else
+ br restore_all
+#endif
+
+/***********************************************************************
+ * A few syscall wrappers
+ ***********************************************************************
+ */
+/*
+ * int clone(unsigned long clone_flags, unsigned long newsp,
+ * int __user * parent_tidptr, int __user * child_tidptr,
+ * int tls_val)
+ */
+ENTRY(sys_clone)
+ SAVE_SWITCH_STACK
+ addi sp, sp, -4
+ stw r7, 0(sp) /* Pass 5th arg thru stack */
+ mov r7, r6 /* 4th arg is 3rd of clone() */
+ mov r6, zero /* 3rd arg always 0 */
+ call do_fork
+ addi sp, sp, 4
+ RESTORE_SWITCH_STACK
+ ret
+
+ENTRY(sys_rt_sigreturn)
+ SAVE_SWITCH_STACK
+ mov r4, sp
+ call do_rt_sigreturn
+ RESTORE_SWITCH_STACK
+ addi ra, ra, (end_translate_rc_and_ret - translate_rc_and_ret)
+ ret
+
+/***********************************************************************
+ * A few other wrappers and stubs
+ ***********************************************************************
+ */
+protection_exception_pte:
+ rdctl r6, pteaddr
+ slli r6, r6, 10
+ call do_page_fault
+ br ret_from_exception
+
+protection_exception_ba:
+ rdctl r6, badaddr
+ call do_page_fault
+ br ret_from_exception
+
+protection_exception_instr:
+ call handle_supervisor_instr
+ br ret_from_exception
+
+handle_breakpoint:
+ call breakpoint_c
+ br ret_from_exception
+
+#ifdef CONFIG_ALIGNMENT_TRAP
+handle_unaligned:
+ SAVE_SWITCH_STACK
+ call handle_unaligned_c
+ RESTORE_SWITCH_STACK
+ br ret_from_exception
+#else
+handle_unaligned:
+ call handle_unaligned_c
+ br ret_from_exception
+#endif
+
+handle_illegal:
+ call handle_illegal_c
+ br ret_from_exception
+
+handle_diverror:
+ call handle_diverror_c
+ br ret_from_exception
+
+/*
+ * Beware - when entering resume, prev (the current task) is
+ * in r4, next (the new task) is in r5, don't change these
+ * registers.
+ */
+ENTRY(resume)
+
+ rdctl r7, status /* save thread status reg */
+ stw r7, TASK_THREAD + THREAD_KPSR(r4)
+
+ andi r7, r7, %lo(~STATUS_PIE) /* disable interrupts */
+ wrctl status, r7
+
+ SAVE_SWITCH_STACK
+ stw sp, TASK_THREAD + THREAD_KSP(r4)/* save kernel stack pointer */
+ ldw sp, TASK_THREAD + THREAD_KSP(r5)/* restore new thread stack */
+ movia r24, _current_thread /* save thread */
+ GET_THREAD_INFO r1
+ stw r1, 0(r24)
+ RESTORE_SWITCH_STACK
+
+ ldw r7, TASK_THREAD + THREAD_KPSR(r5)/* restore thread status reg */
+ wrctl status, r7
+ ret
+
+ENTRY(ret_from_fork)
+ call schedule_tail
+ br ret_from_exception
+
+ENTRY(ret_from_kernel_thread)
+ call schedule_tail
+ mov r4,r17 /* arg */
+ callr r16 /* function */
+ br ret_from_exception
+
+/*
+ * Kernel user helpers.
+ *
+ * Each segment is 64-byte aligned and will be mapped to the <User space>.
+ * New segments (if ever needed) must be added after the existing ones.
+ * This mechanism should be used only for things that are really small and
+ * justified, and not be abused freely.
+ *
+ */
+
+ /* Filling pads with undefined instructions. */
+.macro kuser_pad sym size
+ .if ((. - \sym) & 3)
+ .rept (4 - (. - \sym) & 3)
+ .byte 0
+ .endr
+ .endif
+ .rept ((\size - (. - \sym)) / 4)
+ .word 0xdeadbeef
+ .endr
+.endm
+
+ .align 6
+ .globl __kuser_helper_start
+__kuser_helper_start:
+
+__kuser_helper_version: /* @ 0x1000 */
+ .word ((__kuser_helper_end - __kuser_helper_start) >> 6)
+
+__kuser_cmpxchg: /* @ 0x1004 */
+ /*
+ * r4 pointer to exchange variable
+ * r5 old value
+ * r6 new value
+ */
+cmpxchg_ldw:
+ ldw r2, 0(r4) /* load current value */
+ sub r2, r2, r5 /* compare with old value */
+ bne r2, zero, cmpxchg_ret
+
+ /* We had a match, store the new value */
+cmpxchg_stw:
+ stw r6, 0(r4)
+cmpxchg_ret:
+ ret
+
+ kuser_pad __kuser_cmpxchg, 64
+
+ .globl __kuser_helper_end
+__kuser_helper_end:
diff --git a/arch/nios2/kernel/misaligned.c b/arch/nios2/kernel/misaligned.c
new file mode 100644
index 0000000..f8845ab72af03d946561c84c125fc8ab4f6fce1b
--- /dev/null
+++ b/arch/nios2/kernel/misaligned.c
@@ -0,0 +1,319 @@
+/*
+ * linux/arch/nios2/kernel/misaligned.c
+ *
+ * basic emulation for mis-aligned accesses on the NIOS II cpu
+ * modeled after the version for arm in arm/alignment.c
+ *
+ * Brad Parker <[email protected]>
+ * Copyright (C) 2010 Ambient Corporation
+ * Copyright (c) 2010 Altera Corporation, San Jose, California, USA.
+ * Copyright (c) 2010 Arrow Electronics, Inc.
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License. See the file COPYING in the main directory of
+ * this archive for more details.
+ */
+
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/proc_fs.h>
+#include <linux/init.h>
+#include <linux/sched.h>
+#include <linux/uaccess.h>
+#include <linux/seq_file.h>
+
+#include <asm/traps.h>
+#include <asm/unaligned.h>
+
+/* instructions we emulate */
+#define INST_LDHU 0x0b
+#define INST_STH 0x0d
+#define INST_LDH 0x0f
+#define INST_STW 0x15
+#define INST_LDW 0x17
+
+static unsigned long ma_user, ma_kern, ma_skipped, ma_half, ma_word;
+
+static unsigned int ma_usermode;
+#define UM_WARN 0x01
+#define UM_FIXUP 0x02
+#define UM_SIGNAL 0x04
+#define KM_WARN 0x08
+
+/* see arch/nios2/include/asm/ptrace.h */
+static u8 sys_stack_frame_reg_offset[] = {
+ /* struct pt_regs */
+ 8, 9, 10, 11, 12, 13, 14, 15, 1, 2, 3, 4, 5, 6, 7, 0,
+ /* struct switch_stack */
+ 16, 17, 18, 19, 20, 21, 22, 23, 0, 0, 0, 0, 0, 0, 0, 0
+};
+
+static int reg_offsets[32];
+
+static inline u32 get_reg_val(struct pt_regs *fp, int reg)
+{
+ u8 *p = ((u8 *)fp) + reg_offsets[reg];
+ return *(u32 *)p;
+}
+
+static inline void put_reg_val(struct pt_regs *fp, int reg, u32 val)
+{
+ u8 *p = ((u8 *)fp) + reg_offsets[reg];
+ *(u32 *)p = val;
+}
+
+/*
+ * (mis)alignment handler
+ */
+asmlinkage void handle_unaligned_c(struct pt_regs *fp, int cause)
+{
+ u32 isn, addr, val;
+ int in_kernel;
+ u8 a, b, d0, d1, d2, d3;
+ u16 imm16;
+ unsigned int fault;
+
+ /* back up one instruction */
+ fp->ea -= 4;
+
+ if (fixup_exception(fp)) {
+ ma_skipped++;
+ return;
+ }
+
+ in_kernel = !user_mode(fp);
+
+ isn = *(unsigned long *)(fp->ea);
+
+ fault = 0;
+
+ /* do fixup if in kernel or mode turned on */
+ if (in_kernel || (ma_usermode & UM_FIXUP)) {
+ /* decompose instruction */
+ a = (isn >> 27) & 0x1f;
+ b = (isn >> 22) & 0x1f;
+ imm16 = (isn >> 6) & 0xffff;
+ addr = get_reg_val(fp, a) + imm16;
+
+ /* do fixup to saved registers */
+ switch (isn & 0x3f) {
+ case INST_LDHU:
+ fault |= __get_user(d0, (u8 *)(addr+0));
+ fault |= __get_user(d1, (u8 *)(addr+1));
+ val = (d1 << 8) | d0;
+ put_reg_val(fp, b, val);
+ ma_half++;
+ break;
+ case INST_STH:
+ val = get_reg_val(fp, b);
+ d1 = val >> 8;
+ d0 = val >> 0;
+
+ pr_debug("sth: ra=%d (%08x) rb=%d (%08x), imm16 %04x addr %08x val %08x\n",
+ a, get_reg_val(fp, a),
+ b, get_reg_val(fp, b),
+ imm16, addr, val);
+
+ if (in_kernel) {
+ *(u8 *)(addr+0) = d0;
+ *(u8 *)(addr+1) = d1;
+ } else {
+ fault |= __put_user(d0, (u8 *)(addr+0));
+ fault |= __put_user(d1, (u8 *)(addr+1));
+ }
+ ma_half++;
+ break;
+ case INST_LDH:
+ fault |= __get_user(d0, (u8 *)(addr+0));
+ fault |= __get_user(d1, (u8 *)(addr+1));
+ val = (short)((d1 << 8) | d0);
+ put_reg_val(fp, b, val);
+ ma_half++;
+ break;
+ case INST_STW:
+ val = get_reg_val(fp, b);
+ d3 = val >> 24;
+ d2 = val >> 16;
+ d1 = val >> 8;
+ d0 = val >> 0;
+ if (in_kernel) {
+ *(u8 *)(addr+0) = d0;
+ *(u8 *)(addr+1) = d1;
+ *(u8 *)(addr+2) = d2;
+ *(u8 *)(addr+3) = d3;
+ } else {
+ fault |= __put_user(d0, (u8 *)(addr+0));
+ fault |= __put_user(d1, (u8 *)(addr+1));
+ fault |= __put_user(d2, (u8 *)(addr+2));
+ fault |= __put_user(d3, (u8 *)(addr+3));
+ }
+ ma_word++;
+ break;
+ case INST_LDW:
+ fault |= __get_user(d0, (u8 *)(addr+0));
+ fault |= __get_user(d1, (u8 *)(addr+1));
+ fault |= __get_user(d2, (u8 *)(addr+2));
+ fault |= __get_user(d3, (u8 *)(addr+3));
+ val = (d3 << 24) | (d2 << 16) | (d1 << 8) | d0;
+ put_reg_val(fp, b, val);
+ ma_word++;
+ break;
+ }
+ }
+
+ addr = RDCTL(CTL_BADADDR);
+ cause >>= 2;
+
+ if (fault) {
+ if (in_kernel) {
+ pr_err("fault during kernel misaligned fixup @ %#lx; addr 0x%08lx; isn=0x%08x\n",
+ fp->ea, (long unsigned int)addr,
+ (unsigned int)isn);
+ } else {
+ pr_err("fault during user misaligned fixup @ %#lx; isn=%08x addr=0x%08x sp=0x%08lx pid=%d\n",
+ fp->ea,
+ (unsigned int)isn, addr, fp->sp,
+ current->pid);
+
+ _exception(SIGSEGV, fp, SEGV_MAPERR, fp->ea);
+ return;
+ }
+ }
+
+ /*
+ * kernel mode -
+ * note exception and skip bad instruction (return)
+ */
+ if (in_kernel) {
+ ma_kern++;
+ fp->ea += 4;
+
+ if (ma_usermode & KM_WARN) {
+ pr_err("kernel unaligned access @ %#lx; BADADDR 0x%08lx; cause=%d, isn=0x%08lx\n",
+ fp->ea,
+ (long unsigned int)addr, cause,
+ (long unsigned int)isn);
+ /* show_regs(fp); */
+ }
+
+ return;
+ }
+
+ ma_user++;
+
+ /*
+ * user mode -
+ * possibly warn,
+ * possibly send SIGBUS signal to process
+ */
+ if (ma_usermode & UM_WARN) {
+ pr_err("user unaligned access @ %#lx; isn=0x%08lx ea=0x%08lx ra=0x%08lx sp=0x%08lx\n",
+ (unsigned long)addr, (unsigned long)isn,
+ fp->ea, fp->ra, fp->sp);
+ }
+
+ if (ma_usermode & UM_SIGNAL)
+ _exception(SIGBUS, fp, BUS_ADRALN, fp->ea);
+ else
+ fp->ea += 4; /* else advance */
+}
+
+#ifdef CONFIG_PROC_FS
+static const char * const usermode_action[] = {
+ "ignored", /* 0 */
+ "warn", /* 1 */
+ "fixup", /* 2 */
+ "fixup+warn", /* 3 */
+ "signal", /* 4 */
+ "signal+warn", /* 5 */
+ "signal+fixup",
+ "signal+fixup+warn"
+};
+
+static int misaligned_proc_show(struct seq_file *m, void *v)
+{
+ seq_printf(m, "User:\t\t%lu\n", ma_user);
+ seq_printf(m, "Kernel:\t\t%lu\n", ma_kern);
+ seq_printf(m, "Skipped:\t%lu\n", ma_skipped);
+ seq_printf(m, "Half:\t\t%lu\n", ma_half);
+ seq_printf(m, "Word:\t\t%lu\n", ma_word);
+ seq_printf(m, "User faults:\t%i (%s)\n", ma_usermode,
+ usermode_action[ma_usermode & 7]);
+
+ return 0;
+}
+
+static int proc_misaligned_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, misaligned_proc_show, NULL);
+}
+
+static ssize_t proc_misaligned_write(struct file *file,
+ const char __user *buffer,
+ size_t count, loff_t *ppos)
+{
+ char mode;
+
+ if (count > 0) {
+ if (get_user(mode, buffer))
+ return -EFAULT;
+ if (mode >= '0' && mode <= '5')
+ ma_usermode = mode - '0';
+ }
+
+ return count;
+}
+
+static const struct file_operations misalign_fops = {
+ .open = proc_misaligned_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+ .write = proc_misaligned_write,
+};
+#endif /* CONFIG_PROC_FS */
+
+static void __init misaligned_calc_reg_offsets(void)
+{
+ int i, r, offset;
+
+ /* pre-calc offsets of registers on sys call stack frame */
+ offset = 0;
+
+ /* struct pt_regs */
+ for (i = 0; i < 16; i++) {
+ r = sys_stack_frame_reg_offset[i];
+ reg_offsets[r] = offset;
+ offset += 4;
+ }
+
+ /* struct switch_stack */
+ offset = -sizeof(struct switch_stack);
+ for (i = 16; i < 32; i++) {
+ r = sys_stack_frame_reg_offset[i];
+ reg_offsets[r] = offset;
+ offset += 4;
+ }
+}
+
+
+static int __init misaligned_init(void)
+{
+#ifdef CONFIG_PROC_FS
+
+ struct proc_dir_entry *res;
+
+ res = proc_create("misalign", S_IWUSR | S_IRUGO, NULL, &misalign_fops);
+ if (!res)
+ return -ENOMEM;
+#endif
+
+ /* default mode - silent fix */
+ ma_usermode = UM_FIXUP | KM_WARN;
+
+ misaligned_calc_reg_offsets();
+
+ return 0;
+}
+
+fs_initcall(misaligned_init);
--
1.8.3.2

2014-04-18 12:42:29

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 10/28] nios2: Process management

This patch adds support for thread creation and context switching.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/mmu_context.h | 66 +++++++++
arch/nios2/include/asm/processor.h | 102 ++++++++++++++
arch/nios2/include/asm/switch_to.h | 31 +++++
arch/nios2/include/asm/thread_info.h | 120 ++++++++++++++++
arch/nios2/kernel/process.c | 262 +++++++++++++++++++++++++++++++++++
arch/nios2/mm/mmu_context.c | 116 ++++++++++++++++
6 files changed, 697 insertions(+)
create mode 100644 arch/nios2/include/asm/mmu_context.h
create mode 100644 arch/nios2/include/asm/processor.h
create mode 100644 arch/nios2/include/asm/switch_to.h
create mode 100644 arch/nios2/include/asm/thread_info.h
create mode 100644 arch/nios2/kernel/process.c
create mode 100644 arch/nios2/mm/mmu_context.c

diff --git a/arch/nios2/include/asm/mmu_context.h b/arch/nios2/include/asm/mmu_context.h
new file mode 100644
index 0000000..294b4b1
--- /dev/null
+++ b/arch/nios2/include/asm/mmu_context.h
@@ -0,0 +1,66 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 1996, 1997, 1998, 1999 by Ralf Baechle
+ * Copyright (C) 1999 Silicon Graphics, Inc.
+ *
+ * based on MIPS asm/mmu_context.h
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_MMU_CONTEXT_H
+#define _ASM_NIOS2_MMU_CONTEXT_H
+
+#include <asm-generic/mm_hooks.h>
+
+extern void mmu_context_init(void);
+extern unsigned long get_pid_from_context(mm_context_t *ctx);
+
+/*
+ * For the fast tlb miss handlers, we keep a pointer to the current pgd.
+ * processor.
+ */
+extern pgd_t *pgd_current;
+
+static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk)
+{
+}
+
+/*
+ * Initialize the context related info for a new mm_struct instance.
+ *
+ * Set all new contexts to 0, that way the generation will never match
+ * the currently running generation when this context is switched in.
+ */
+static inline int init_new_context(struct task_struct *tsk,
+ struct mm_struct *mm)
+{
+ mm->context = 0;
+ return 0;
+}
+
+/*
+ * Destroy context related info for an mm_struct that is about
+ * to be put to rest.
+ */
+static inline void destroy_context(struct mm_struct *mm)
+{
+}
+
+void switch_mm(struct mm_struct *prev, struct mm_struct *next,
+ struct task_struct *tsk);
+
+static inline void deactivate_mm(struct task_struct *tsk,
+ struct mm_struct *mm)
+{
+}
+
+/*
+ * After we have set current->mm to a new value, this activates
+ * the context for the new mm so we see the new mappings.
+ */
+void activate_mm(struct mm_struct *prev, struct mm_struct *next);
+
+#endif /* _ASM_NIOS2_MMU_CONTEXT_H */
diff --git a/arch/nios2/include/asm/processor.h b/arch/nios2/include/asm/processor.h
new file mode 100644
index 0000000..cff9b2da0cdd3f9ba392f782492dd3b8d435bb9a
--- /dev/null
+++ b/arch/nios2/include/asm/processor.h
@@ -0,0 +1,102 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ * Copyright (C) 2001 Ken Hill ([email protected])
+ * Vic Phillips ([email protected])
+ *
+ * based on SPARC asm/processor_32.h which is:
+ *
+ * Copyright (C) 1994 David S. Miller
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_PROCESSOR_H
+#define _ASM_NIOS2_PROCESSOR_H
+
+#include <asm/ptrace.h>
+#include <asm/registers.h>
+#include <asm/page.h>
+
+#define NIOS2_FLAG_KTHREAD 0x00000001 /* task is a kernel thread */
+
+#define NIOS2_OP_NOP 0x1883a
+#define NIOS2_OP_BREAK 0x3da03a
+
+#ifdef __KERNEL__
+
+#define STACK_TOP TASK_SIZE
+#define STACK_TOP_MAX STACK_TOP
+
+#endif /* __KERNEL__ */
+
+/* Kuser helpers is mapped to this user space address */
+#define KUSER_BASE 0x1000
+#define KUSER_SIZE (PAGE_SIZE)
+#ifndef __ASSEMBLY__
+
+/*
+ * Default implementation of macro that returns current
+ * instruction pointer ("program counter").
+ */
+#define current_text_addr() ({ __label__ _l; _l: &&_l; })
+
+# define TASK_SIZE 0x7FFF0000UL
+# define TASK_UNMAPPED_BASE (PAGE_ALIGN(TASK_SIZE / 3))
+
+/* The Nios processor specific thread struct. */
+struct thread_struct {
+ struct pt_regs *kregs;
+
+ /* Context switch saved kernel state. */
+ unsigned long ksp;
+ unsigned long kpsr;
+};
+
+#define INIT_MMAP \
+ { &init_mm, (0), (0), __pgprot(0x0), VM_READ | VM_WRITE | VM_EXEC }
+
+# define INIT_THREAD { \
+ .kregs = NULL, \
+ .ksp = 0, \
+ .kpsr = 0, \
+}
+
+extern void start_thread(struct pt_regs *regs, unsigned long pc,
+ unsigned long sp);
+
+struct task_struct;
+
+/* Free all resources held by a thread. */
+static inline void release_thread(struct task_struct *dead_task)
+{
+}
+
+/* Free current thread data structures etc.. */
+static inline void exit_thread(void)
+{
+}
+
+/* Return saved PC of a blocked thread. */
+#define thread_saved_pc(tsk) ((tsk)->thread.kregs->ea)
+
+extern unsigned long get_wchan(struct task_struct *p);
+
+/* Prepare to copy thread state - unlazy all lazy status */
+#define prepare_to_copy(tsk) do { } while (0)
+
+#define task_pt_regs(p) \
+ ((struct pt_regs *)(THREAD_SIZE + task_stack_page(p)) - 1)
+
+/* Used by procfs */
+#define KSTK_EIP(tsk) ((tsk)->thread.kregs->ea)
+#define KSTK_ESP(tsk) ((tsk)->thread.kregs->sp)
+
+#define cpu_relax() barrier()
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* _ASM_NIOS2_PROCESSOR_H */
diff --git a/arch/nios2/include/asm/switch_to.h b/arch/nios2/include/asm/switch_to.h
new file mode 100644
index 0000000..c47b3f4afbcddcea3a89f44569141339412d6cf3
--- /dev/null
+++ b/arch/nios2/include/asm/switch_to.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+#ifndef _ASM_NIOS2_SWITCH_TO_H
+#define _ASM_NIOS2_SWITCH_TO_H
+
+/*
+ * switch_to(n) should switch tasks to task ptr, first checking that
+ * ptr isn't the current task, in which case it does nothing. This
+ * also clears the TS-flag if the task we switched to has used the
+ * math co-processor latest.
+ */
+#define switch_to(prev, next, last) \
+{ \
+ void *_last; \
+ __asm__ __volatile__ ( \
+ "mov r4, %1\n" \
+ "mov r5, %2\n" \
+ "call resume\n" \
+ "mov %0,r4\n" \
+ : "=r" (_last) \
+ : "r" (prev), "r" (next) \
+ : "r4", "r5", "r7", "r8", "ra"); \
+ (last) = _last; \
+}
+
+#endif /* _ASM_NIOS2_SWITCH_TO_H */
diff --git a/arch/nios2/include/asm/thread_info.h b/arch/nios2/include/asm/thread_info.h
new file mode 100644
index 0000000..1f266575beb51c459432ac918e96e0b3b810ae85
--- /dev/null
+++ b/arch/nios2/include/asm/thread_info.h
@@ -0,0 +1,120 @@
+/*
+ * NiosII low-level thread information
+ *
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * Based on asm/thread_info_no.h from m68k which is:
+ *
+ * Copyright (C) 2002 David Howells <[email protected]>
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_THREAD_INFO_H
+#define _ASM_NIOS2_THREAD_INFO_H
+
+#ifdef __KERNEL__
+
+/*
+ * Size of the kernel stack for each process.
+ */
+#define THREAD_SIZE_ORDER 1
+#define THREAD_SIZE 8192 /* 2 * PAGE_SIZE */
+
+#ifndef __ASSEMBLY__
+
+typedef struct {
+ unsigned long seg;
+} mm_segment_t;
+
+/*
+ * low level task data that entry.S needs immediate access to
+ * - this struct should fit entirely inside of one cache line
+ * - this struct shares the supervisor stack pages
+ * - if the contents of this structure are changed, the assembly constants
+ * must also be changed
+ */
+struct thread_info {
+ struct task_struct *task; /* main task structure */
+ struct exec_domain *exec_domain; /* execution domain */
+ unsigned long flags; /* low level flags */
+ __u32 cpu; /* current CPU */
+ int preempt_count; /* 0 => preemptable,<0 => BUG */
+ mm_segment_t addr_limit; /* thread address space:
+ 0-0x7FFFFFFF for user-thead
+ 0-0xFFFFFFFF for kernel-thread
+ */
+ struct restart_block restart_block;
+ struct pt_regs *regs;
+};
+
+/*
+ * macros/functions for gaining access to the thread information structure
+ *
+ * preempt_count needs to be 1 initially, until the scheduler is functional.
+ */
+#define INIT_THREAD_INFO(tsk) \
+{ \
+ .task = &tsk, \
+ .exec_domain = &default_exec_domain, \
+ .flags = 0, \
+ .cpu = 0, \
+ .preempt_count = INIT_PREEMPT_COUNT, \
+ .addr_limit = KERNEL_DS, \
+ .restart_block = { \
+ .fn = do_no_restart_syscall, \
+ }, \
+}
+
+#define init_thread_info (init_thread_union.thread_info)
+#define init_stack (init_thread_union.stack)
+
+/* how to get the thread information struct from C */
+static inline struct thread_info *current_thread_info(void)
+{
+ register unsigned long sp asm("sp");
+
+ return (struct thread_info *)(sp & ~(THREAD_SIZE - 1));
+}
+#endif /* !__ASSEMBLY__ */
+
+/*
+ * thread information flags
+ * - these are process state flags that various assembly files may need to
+ * access
+ * - pending work-to-be-done flags are in LSW
+ * - other flags in MSW
+ */
+#define TIF_SYSCALL_TRACE 0 /* syscall trace active */
+#define TIF_NOTIFY_RESUME 1 /* resumption notification requested */
+#define TIF_SIGPENDING 2 /* signal pending */
+#define TIF_NEED_RESCHED 3 /* rescheduling necessary */
+#define TIF_MEMDIE 4 /* is terminating due to OOM killer */
+#define TIF_SECCOMP 5 /* secure computing */
+#define TIF_SYSCALL_AUDIT 6 /* syscall auditing active */
+#define TIF_RESTORE_SIGMASK 9 /* restore signal mask in do_signal() */
+
+#define TIF_POLLING_NRFLAG 16 /* true if poll_idle() is polling
+ TIF_NEED_RESCHED */
+
+#define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE)
+#define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME)
+#define _TIF_SIGPENDING (1 << TIF_SIGPENDING)
+#define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED)
+#define _TIF_SECCOMP (1 << TIF_SECCOMP)
+#define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT)
+#define _TIF_RESTORE_SIGMASK (1 << TIF_RESTORE_SIGMASK)
+#define _TIF_POLLING_NRFLAG (1 << TIF_POLLING_NRFLAG)
+
+/* work to do on interrupt/exception return */
+#define _TIF_WORK_MASK 0x0000FFFE
+
+/* work to do on any return to u-space */
+# define _TIF_ALLWORK_MASK 0x0000FFFF
+
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_NIOS2_THREAD_INFO_H */
diff --git a/arch/nios2/kernel/process.c b/arch/nios2/kernel/process.c
new file mode 100644
index 0000000..162b3f0
--- /dev/null
+++ b/arch/nios2/kernel/process.c
@@ -0,0 +1,262 @@
+/*
+ * Architecture-dependent parts of process handling.
+ *
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * based on arch/m68knommu/kernel/process.c which is:
+ *
+ * Copyright (C) 2000-2002 David McCullough <[email protected]>
+ * Copyright (C) 1995 Hamish Macdonald
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/export.h>
+#include <linux/sched.h>
+#include <linux/tick.h>
+#include <linux/uaccess.h>
+
+#include <asm/unistd.h>
+#include <asm/traps.h>
+#include <asm/cpuinfo.h>
+
+asmlinkage void ret_from_fork(void);
+asmlinkage void ret_from_kernel_thread(void);
+
+void (*pm_power_off)(void) = NULL;
+EXPORT_SYMBOL(pm_power_off);
+
+void arch_cpu_idle(void)
+{
+ local_irq_enable();
+}
+
+/*
+ * The development boards have no way to pull a board reset. Just jump to the
+ * cpu reset address and let the boot loader or the code in head.S take care of
+ * resetting peripherals.
+ */
+void machine_restart(char *__unused)
+{
+ pr_notice("Machine restart (%08x)...\n", cpuinfo.reset_addr);
+ local_irq_disable();
+ __asm__ __volatile__ (
+ "jmp %0\n\t"
+ :
+ : "r" (cpuinfo.reset_addr)
+ : "r4");
+}
+
+void machine_halt(void)
+{
+ pr_notice("Machine halt...\n");
+ local_irq_disable();
+ for (;;)
+ ;
+}
+
+/*
+ * There is no way to power off the development boards. So just spin for now. If
+ * we ever have a way of resetting a board using a GPIO we should add that here.
+ */
+void machine_power_off(void)
+{
+ pr_notice("Machine power off...\n");
+ local_irq_disable();
+ for (;;)
+ ;
+}
+
+void show_regs(struct pt_regs *regs)
+{
+ pr_notice("\n");
+ show_regs_print_info(KERN_DEFAULT);
+
+ pr_notice("r1: %08lx r2: %08lx r3: %08lx r4: %08lx\n",
+ regs->r1, regs->r2, regs->r3, regs->r4);
+
+ pr_notice("r5: %08lx r6: %08lx r7: %08lx r8: %08lx\n",
+ regs->r5, regs->r6, regs->r7, regs->r8);
+
+ pr_notice("r9: %08lx r10: %08lx r11: %08lx r12: %08lx\n",
+ regs->r9, regs->r10, regs->r11, regs->r12);
+
+ pr_notice("r13: %08lx r14: %08lx r15: %08lx\n",
+ regs->r13, regs->r14, regs->r15);
+
+ pr_notice("ra: %08lx fp: %08lx sp: %08lx gp: %08lx\n",
+ regs->ra, regs->fp, regs->sp, regs->gp);
+
+ pr_notice("ea: %08lx estatus: %08lx\n",
+ regs->ea, regs->estatus);
+}
+
+void flush_thread(void)
+{
+ set_fs(USER_DS);
+}
+
+int copy_thread(unsigned long clone_flags,
+ unsigned long usp, unsigned long arg, struct task_struct *p)
+{
+ struct pt_regs *childregs = task_pt_regs(p);
+ struct pt_regs *regs;
+ struct switch_stack *stack;
+ struct switch_stack *childstack =
+ ((struct switch_stack *)childregs) - 1;
+
+ if (unlikely(p->flags & PF_KTHREAD)) {
+ memset(childstack, 0,
+ sizeof(struct switch_stack) + sizeof(struct pt_regs));
+
+ childstack->r16 = usp; /* fn */
+ childstack->r17 = arg;
+ childstack->ra = (unsigned long) ret_from_kernel_thread;
+ childregs->estatus = STATUS_PIE;
+ childregs->sp = (unsigned long) childstack;
+
+ p->thread.ksp = (unsigned long) childstack;
+ p->thread.kregs = childregs;
+ return 0;
+ }
+
+ regs = current_pt_regs();
+ *childregs = *regs;
+ childregs->r2 = 0; /* Set the return value for the child. */
+ childregs->r7 = 0;
+
+ stack = ((struct switch_stack *) regs) - 1;
+ *childstack = *stack;
+ childstack->ra = (unsigned long)ret_from_fork;
+ p->thread.kregs = childregs;
+ p->thread.ksp = (unsigned long) childstack;
+
+ if (usp)
+ childregs->sp = usp;
+
+ /* Initialize tls register. */
+ if (clone_flags & CLONE_SETTLS)
+ childstack->r23 = regs->r8;
+
+ return 0;
+}
+
+/*
+ * Generic dumping code. Used for panic and debug.
+ */
+void dump(struct pt_regs *fp)
+{
+ unsigned long *sp;
+ unsigned char *tp;
+ int i;
+
+ pr_emerg("\nCURRENT PROCESS:\n\n");
+ pr_emerg("COMM=%s PID=%d\n", current->comm, current->pid);
+
+ if (current->mm) {
+ pr_emerg("TEXT=%08x-%08x DATA=%08x-%08x BSS=%08x-%08x\n",
+ (int) current->mm->start_code,
+ (int) current->mm->end_code,
+ (int) current->mm->start_data,
+ (int) current->mm->end_data,
+ (int) current->mm->end_data,
+ (int) current->mm->brk);
+ pr_emerg("USER-STACK=%08x KERNEL-STACK=%08x\n\n",
+ (int) current->mm->start_stack,
+ (int)(((unsigned long) current) + THREAD_SIZE));
+ }
+
+ pr_emerg("PC: %08lx\n", fp->ea);
+ pr_emerg(KERN_EMERG "SR: %08lx SP: %08lx\n",
+ (long) fp->estatus, (long) fp);
+
+ pr_emerg("r1: %08lx r2: %08lx r3: %08lx\n",
+ fp->r1, fp->r2, fp->r3);
+
+ pr_emerg("r4: %08lx r5: %08lx r6: %08lx r7: %08lx\n",
+ fp->r4, fp->r5, fp->r6, fp->r7);
+ pr_emerg("r8: %08lx r9: %08lx r10: %08lx r11: %08lx\n",
+ fp->r8, fp->r9, fp->r10, fp->r11);
+ pr_emerg("r12: %08lx r13: %08lx r14: %08lx r15: %08lx\n",
+ fp->r12, fp->r13, fp->r14, fp->r15);
+ pr_emerg("or2: %08lx ra: %08lx fp: %08lx sp: %08lx\n",
+ fp->orig_r2, fp->ra, fp->fp, fp->sp);
+ pr_emerg("\nUSP: %08x TRAPFRAME: %08x\n",
+ (unsigned int) fp->sp, (unsigned int) fp);
+
+ pr_emerg("\nCODE:");
+ tp = ((unsigned char *) fp->ea) - 0x20;
+ for (sp = (unsigned long *) tp, i = 0; (i < 0x40); i += 4) {
+ if ((i % 0x10) == 0)
+ pr_emerg("\n%08x: ", (int) (tp + i));
+ pr_emerg("%08x ", (int) *sp++);
+ }
+ pr_emerg("\n");
+
+ pr_emerg("\nKERNEL STACK:");
+ tp = ((unsigned char *) fp) - 0x40;
+ for (sp = (unsigned long *) tp, i = 0; (i < 0xc0); i += 4) {
+ if ((i % 0x10) == 0)
+ pr_emerg("\n%08x: ", (int) (tp + i));
+ pr_emerg("%08x ", (int) *sp++);
+ }
+ pr_emerg("\n");
+ pr_emerg("\n");
+
+ pr_emerg("\nUSER STACK:");
+ tp = (unsigned char *) (fp->sp - 0x10);
+ for (sp = (unsigned long *) tp, i = 0; (i < 0x80); i += 4) {
+ if ((i % 0x10) == 0)
+ pr_emerg("\n%08x: ", (int) (tp + i));
+ pr_emerg("%08x ", (int) *sp++);
+ }
+ pr_emerg("\n\n");
+}
+
+unsigned long get_wchan(struct task_struct *p)
+{
+ unsigned long fp, pc;
+ unsigned long stack_page;
+ int count = 0;
+ if (!p || p == current || p->state == TASK_RUNNING)
+ return 0;
+
+ stack_page = (unsigned long)p;
+ fp = ((struct switch_stack *)p->thread.ksp)->fp; /* ;dgt2 */
+ do {
+ if (fp < stack_page+sizeof(struct task_struct) ||
+ fp >= 8184+stack_page) /* ;dgt2;tmp */
+ return 0;
+ pc = ((unsigned long *)fp)[1];
+ if (!in_sched_functions(pc))
+ return pc;
+ fp = *(unsigned long *) fp;
+ } while (count++ < 16); /* ;dgt2;tmp */
+ return 0;
+}
+
+/*
+ * Do necessary setup to start up a newly executed thread.
+ * Will startup in user mode (status_extension = 0).
+ */
+void start_thread(struct pt_regs *regs, unsigned long pc, unsigned long sp)
+{
+ memset((void *) regs, 0, sizeof(struct pt_regs));
+ regs->estatus = ESTATUS_EPIE | ESTATUS_EU;
+ regs->ea = pc;
+ regs->sp = sp;
+}
+
+#include <linux/elfcore.h>
+
+/* Fill in the FPU structure for a core dump. */
+int dump_fpu(struct pt_regs *regs, elf_fpregset_t *r)
+{
+ return 0; /* Nios2 has no FPU and thus no FPU registers */
+}
diff --git a/arch/nios2/mm/mmu_context.c b/arch/nios2/mm/mmu_context.c
new file mode 100644
index 0000000..45d6b9c58d677a93b3bbe38148693e02394aac87
--- /dev/null
+++ b/arch/nios2/mm/mmu_context.c
@@ -0,0 +1,116 @@
+/*
+ * MMU context handling.
+ *
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/mm.h>
+
+#include <asm/cpuinfo.h>
+#include <asm/mmu_context.h>
+#include <asm/tlb.h>
+
+/* The pids position and mask in context */
+#define PID_SHIFT 0
+#define PID_BITS (cpuinfo.tlb_pid_num_bits)
+#define PID_MASK ((1UL << PID_BITS) - 1)
+
+/* The versions position and mask in context */
+#define VERSION_BITS (32 - PID_BITS)
+#define VERSION_SHIFT (PID_SHIFT + PID_BITS)
+#define VERSION_MASK ((1UL << VERSION_BITS) - 1)
+
+/* Return the version part of a context */
+#define CTX_VERSION(c) (((c) >> VERSION_SHIFT) & VERSION_MASK)
+
+/* Return the pid part of a context */
+#define CTX_PID(c) (((c) >> PID_SHIFT) & PID_MASK)
+
+/* Value of the first context (version 1, pid 0) */
+#define FIRST_CTX ((1UL << VERSION_SHIFT) | (0 << PID_SHIFT))
+
+static mm_context_t next_mmu_context;
+
+/*
+ * Initialize MMU context management stuff.
+ */
+void __init mmu_context_init(void)
+{
+ /* We need to set this here because the value depends on runtime data
+ * from cpuinfo */
+ next_mmu_context = FIRST_CTX;
+}
+
+/*
+ * Set new context (pid), keep way
+ */
+static void set_context(mm_context_t context)
+{
+ set_mmu_pid(CTX_PID(context));
+}
+
+static mm_context_t get_new_context(void)
+{
+ /* Return the next pid */
+ next_mmu_context += (1UL << PID_SHIFT);
+
+ /* If the pid field wraps around we increase the version and
+ * flush the tlb */
+ if (unlikely(CTX_PID(next_mmu_context) == 0)) {
+ /* Version is incremented since the pid increment above
+ * overflows info version */
+ flush_cache_all();
+ flush_tlb_all();
+ }
+
+ /* If the version wraps we start over with the first generation, we do
+ * not need to flush the tlb here since it's always done above */
+ if (unlikely(CTX_VERSION(next_mmu_context) == 0))
+ next_mmu_context = FIRST_CTX;
+
+ return next_mmu_context;
+}
+
+void switch_mm(struct mm_struct *prev, struct mm_struct *next,
+ struct task_struct *tsk)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+
+ /* If the process context we are swapping in has a different context
+ * generation then we have it should get a new generation/pid */
+ if (unlikely(CTX_VERSION(next->context) !=
+ CTX_VERSION(next_mmu_context)))
+ next->context = get_new_context();
+
+ /* Save the current pgd so the fast tlb handler can find it */
+ pgd_current = next->pgd;
+
+ /* Set the current context */
+ set_context(next->context);
+
+ local_irq_restore(flags);
+}
+
+/*
+ * After we have set current->mm to a new value, this activates
+ * the context for the new mm so we see the new mappings.
+ */
+void activate_mm(struct mm_struct *prev, struct mm_struct *next)
+{
+ next->context = get_new_context();
+ set_context(next->context);
+ pgd_current = next->pgd;
+}
+
+unsigned long get_pid_from_context(mm_context_t *context)
+{
+ return CTX_PID((*context));
+}
--
1.8.3.2

2014-04-18 12:42:39

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 20/28] nios2: Time keeping

Add time keeping code for nios2.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/delay.h | 92 +++++++++++++++++++++++++
arch/nios2/include/asm/timex.h | 27 ++++++++
arch/nios2/kernel/time.c | 150 +++++++++++++++++++++++++++++++++++++++++
3 files changed, 269 insertions(+)
create mode 100644 arch/nios2/include/asm/delay.h
create mode 100644 arch/nios2/include/asm/timex.h
create mode 100644 arch/nios2/kernel/time.c

diff --git a/arch/nios2/include/asm/delay.h b/arch/nios2/include/asm/delay.h
new file mode 100644
index 0000000..8070a09
--- /dev/null
+++ b/arch/nios2/include/asm/delay.h
@@ -0,0 +1,92 @@
+/*
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_DELAY_H
+#define _ASM_NIOS2_DELAY_H
+
+#include <asm/param.h>
+
+static inline void __delay(unsigned long loops)
+{
+ int dummy;
+
+ __asm__ __volatile__(
+ "1:\n\t"
+ " beq %0,zero,2f\n\t"
+ " addi %0, %0, -1\n\t"
+ " br 1b\n\t"
+ "2:\n\t"
+ : "=r" (dummy) /* Need output for optimizer */
+ : "0" (loops)); /* %0 Input */
+}
+
+/*
+ * Note that 19 * 226 == 4294 ==~ 2^32 / 10^6, so
+ * loops = (4294 * usecs * loops_per_jiffy * HZ) / 2^32.
+ *
+ * The mul instruction gives us loops = (a * b) / 2^32.
+ * We choose a = usecs * 19 * HZ and b = loops_per_jiffy * 226
+ * because this lets us support a wide range of HZ and
+ * loops_per_jiffy values without either a or b overflowing 2^32.
+ * Thus we need usecs * HZ <= (2^32 - 1) / 19 = 226050910 and
+ * loops_per_jiffy <= (2^32 - 1) / 226 = 19004280
+ * (which corresponds to ~3800 bogomips at HZ = 100).
+ * -- paulus
+ */
+#define __MAX_UDELAY (226050910UL/HZ) /* maximum udelay argument */
+#define __MAX_NDELAY (4294967295UL/HZ) /* maximum ndelay argument */
+
+extern unsigned long loops_per_jiffy;
+
+static inline void __udelay(unsigned int x)
+{
+ unsigned int loops;
+
+ /*
+ * Note, if this is compiled with -mhw-mulx it will produce a "mulxuu"
+ * (at least in toolchain 145) so there is no need for inline
+ * assembly here anymore, which might in turn be emulated if unsupported
+ * by the design.
+ */
+ loops = (unsigned int)((((unsigned long long)(x) *
+ (unsigned long long)(loops_per_jiffy * 226))) >> 32);
+
+/*
+ __asm__("mulxuu %0,%1,%2" : "=r" (loops) :
+ "r" (x), "r" (loops_per_jiffy * 226));
+*/
+ __delay(loops);
+}
+
+static inline void __ndelay(unsigned int x)
+{
+ unsigned int loops;
+
+ /* see comment in __udelay */
+ loops = (unsigned int)((((unsigned long long)(x) *
+ (unsigned long long)(loops_per_jiffy * 5))) >> 32);
+
+/*
+ __asm__("mulxuu %0,%1,%2" : "=r" (loops) :
+ "r" (x), "r" (loops_per_jiffy * 5));
+*/
+ __delay(loops);
+}
+
+extern void __bad_udelay(void); /* deliberately undefined */
+extern void __bad_ndelay(void); /* deliberately undefined */
+
+#define udelay(n) (__builtin_constant_p(n) ? \
+ ((n) > __MAX_UDELAY ? __bad_udelay() : __udelay((n) * (19 * HZ))) : \
+ __udelay((n) * (19 * HZ)))
+
+#define ndelay(n) (__builtin_constant_p(n) ? \
+ ((n) > __MAX_NDELAY ? __bad_ndelay() : __ndelay((n) * HZ)) : \
+ __ndelay((n) * HZ))
+
+#endif /* _ASM_NIOS2_DELAY_H */
diff --git a/arch/nios2/include/asm/timex.h b/arch/nios2/include/asm/timex.h
new file mode 100644
index 0000000..524b47d79d3dad53b98713f31517e28ac0c99d98
--- /dev/null
+++ b/arch/nios2/include/asm/timex.h
@@ -0,0 +1,27 @@
+/*
+ * Copyright (C) 2010 Thomas Chou <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_TIMEX_H
+#define _ASM_NIOS2_TIMEX_H
+
+/* Supply dummy tick-rate. Real value will be read from devicetree */
+#define CLOCK_TICK_RATE (HZ * 100000UL)
+
+#include <asm-generic/timex.h>
+
+#endif
diff --git a/arch/nios2/kernel/time.c b/arch/nios2/kernel/time.c
new file mode 100644
index 0000000..0677ebaa1dffd8213f52a704c6a8366d34c05ca2
--- /dev/null
+++ b/arch/nios2/kernel/time.c
@@ -0,0 +1,150 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/init.h>
+#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <linux/time.h>
+#include <linux/timex.h>
+#include <linux/profile.h>
+#include <linux/clocksource.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/io.h>
+
+#define TICK_SIZE (tick_nsec / 1000)
+#define NIOS2_TIMER_PERIOD (timer_freq / HZ)
+
+#define ALTERA_TIMER_STATUS_REG 0
+#define ALTERA_TIMER_CONTROL_REG 4
+#define ALTERA_TIMER_PERIODL_REG 8
+#define ALTERA_TIMER_PERIODH_REG 12
+#define ALTERA_TIMER_SNAPL_REG 16
+#define ALTERA_TIMER_SNAPH_REG 20
+
+#define ALTERA_TIMER_CONTROL_ITO_MSK (0x1)
+#define ALTERA_TIMER_CONTROL_CONT_MSK (0x2)
+#define ALTERA_TIMER_CONTROL_START_MSK (0x4)
+#define ALTERA_TIMER_CONTROL_STOP_MSK (0x8)
+
+static u32 nios2_timer_count;
+static void __iomem *timer_membase;
+static u32 timer_freq;
+
+static inline unsigned long read_timersnapshot(void)
+{
+ unsigned long count;
+
+ outw(0, timer_membase + ALTERA_TIMER_SNAPL_REG);
+ count =
+ inw(timer_membase + ALTERA_TIMER_SNAPH_REG) << 16 |
+ inw(timer_membase + ALTERA_TIMER_SNAPL_REG);
+
+ return count;
+}
+
+static inline void write_timerperiod(unsigned long period)
+{
+ outw(period, timer_membase + ALTERA_TIMER_PERIODL_REG);
+ outw(period >> 16, timer_membase + ALTERA_TIMER_PERIODH_REG);
+}
+
+/*
+ * timer_interrupt() needs to keep up the real-time clock,
+ * as well as call the "xtime_update()" routine every clocktick
+ */
+irqreturn_t timer_interrupt(int irq, void *dummy)
+{
+ /* Clear the interrupt condition */
+ outw(0, timer_membase + ALTERA_TIMER_STATUS_REG);
+ nios2_timer_count += NIOS2_TIMER_PERIOD;
+
+ profile_tick(CPU_PROFILING);
+
+ xtime_update(1);
+
+ update_process_times(user_mode(get_irq_regs()));
+
+ return IRQ_HANDLED;
+}
+
+static cycle_t nios2_timer_read(struct clocksource *cs)
+{
+ unsigned long flags;
+ u32 cycles;
+ u32 tcn;
+
+ local_irq_save(flags);
+ tcn = NIOS2_TIMER_PERIOD - 1 - read_timersnapshot();
+ cycles = nios2_timer_count;
+ local_irq_restore(flags);
+
+ return cycles + tcn;
+}
+
+static struct clocksource nios2_timer = {
+ .name = "timer",
+ .rating = 250,
+ .read = nios2_timer_read,
+ .shift = 20,
+ .mask = CLOCKSOURCE_MASK(32),
+ .flags = CLOCK_SOURCE_IS_CONTINUOUS,
+};
+
+static struct irqaction nios2_timer_irq = {
+ .name = "timer",
+ .flags = IRQF_TIMER,
+ .handler = timer_interrupt,
+};
+
+static void __init nios2_time_init(struct device_node *timer)
+{
+ int irq;
+ unsigned int ctrl;
+
+ timer_membase = of_iomap(timer, 0);
+ if (!timer_membase)
+ panic("Unable to map timer resource\n");
+
+ if (of_property_read_u32(timer, "clock-frequency", &timer_freq))
+ panic("Unable to get timer clock frequency\n");
+
+ irq = irq_of_parse_and_map(timer, 0);
+ if (irq < 0)
+ panic("Unable to parse timer irq\n");
+
+ if (setup_irq(irq, &nios2_timer_irq))
+ panic("Unable to setup timer irq\n");
+
+ write_timerperiod(NIOS2_TIMER_PERIOD - 1);
+
+ /* clocksource initialize */
+ nios2_timer.mult = clocksource_hz2mult(timer_freq, nios2_timer.shift);
+ clocksource_register(&nios2_timer);
+
+ /* interrupt enable + continuous + start */
+ ctrl = ALTERA_TIMER_CONTROL_ITO_MSK | ALTERA_TIMER_CONTROL_CONT_MSK |
+ ALTERA_TIMER_CONTROL_START_MSK;
+ outw(ctrl, timer_membase + ALTERA_TIMER_CONTROL_REG);
+}
+
+void read_persistent_clock(struct timespec *ts)
+{
+ ts->tv_sec = mktime(2007, 1, 1, 0, 0, 0);
+ ts->tv_nsec = 0;
+}
+
+void __init time_init(void)
+{
+ clocksource_of_init();
+}
+
+CLOCKSOURCE_OF_DECLARE(nios2_timer, "ALTR,timer-1.0", nios2_time_init);
--
1.8.3.2

2014-04-18 12:43:18

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 18/28] nios2: Library functions

Add optimised library functions for nios2.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/checksum.h | 78 +++++++++++++++
arch/nios2/include/asm/string.h | 24 +++++
arch/nios2/lib/memcpy.c | 199 ++++++++++++++++++++++++++++++++++++++
arch/nios2/lib/memmove.c | 82 ++++++++++++++++
arch/nios2/lib/memset.c | 81 ++++++++++++++++
5 files changed, 464 insertions(+)
create mode 100644 arch/nios2/include/asm/checksum.h
create mode 100644 arch/nios2/include/asm/string.h
create mode 100644 arch/nios2/lib/memcpy.c
create mode 100644 arch/nios2/lib/memmove.c
create mode 100644 arch/nios2/lib/memset.c

diff --git a/arch/nios2/include/asm/checksum.h b/arch/nios2/include/asm/checksum.h
new file mode 100644
index 0000000..6bc1f0d
--- /dev/null
+++ b/arch/nios2/include/asm/checksum.h
@@ -0,0 +1,78 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS_CHECKSUM_H
+#define _ASM_NIOS_CHECKSUM_H
+
+/* Take these from lib/checksum.c */
+extern __wsum csum_partial(const void *buff, int len, __wsum sum);
+extern __wsum csum_partial_copy(const void *src, void *dst, int len,
+ __wsum sum);
+extern __wsum csum_partial_copy_from_user(const void __user *src, void *dst,
+ int len, __wsum sum, int *csum_err);
+#define csum_partial_copy_nocheck(src, dst, len, sum) \
+ csum_partial_copy((src), (dst), (len), (sum))
+
+extern __sum16 ip_fast_csum(const void *iph, unsigned int ihl);
+extern __sum16 ip_compute_csum(const void *buff, int len);
+
+/*
+ * Fold a partial checksum
+ */
+static inline __sum16 csum_fold(__wsum sum)
+{
+ __asm__ __volatile__(
+ "add %0, %1, %0\n"
+ "cmpltu r8, %0, %1\n"
+ "srli %0, %0, 16\n"
+ "add %0, %0, r8\n"
+ "nor %0, %0, %0\n"
+ : "=r" (sum)
+ : "r" (sum << 16), "0" (sum)
+ : "r8");
+ return (__force __sum16) sum;
+}
+
+/*
+ * computes the checksum of the TCP/UDP pseudo-header
+ * returns a 16-bit checksum, already complemented
+ */
+#define csum_tcpudp_nofold csum_tcpudp_nofold
+static inline __wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
+ unsigned short len,
+ unsigned short proto,
+ __wsum sum)
+{
+ __asm__ __volatile__(
+ "add %0, %1, %0\n"
+ "cmpltu r8, %0, %1\n"
+ "add %0, %0, r8\n" /* add carry */
+ "add %0, %2, %0\n"
+ "cmpltu r8, %0, %2\n"
+ "add %0, %0, r8\n" /* add carry */
+ "add %0, %3, %0\n"
+ "cmpltu r8, %0, %3\n"
+ "add %0, %0, r8\n" /* add carry */
+ : "=r" (sum), "=r" (saddr)
+ : "r" (daddr), "r" ((ntohs(len) << 16) + (proto * 256)),
+ "0" (sum),
+ "1" (saddr)
+ : "r8");
+
+ return sum;
+}
+
+static inline __sum16 csum_tcpudp_magic(__be32 saddr, __be32 daddr,
+ unsigned short len,
+ unsigned short proto, __wsum sum)
+{
+ return csum_fold(csum_tcpudp_nofold(saddr, daddr, len, proto, sum));
+}
+
+#endif /* _ASM_NIOS_CHECKSUM_H */
diff --git a/arch/nios2/include/asm/string.h b/arch/nios2/include/asm/string.h
new file mode 100644
index 0000000..14dd570d64f70bfd19e62fe373650fd32cdf88a9
--- /dev/null
+++ b/arch/nios2/include/asm/string.h
@@ -0,0 +1,24 @@
+/*
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_STRING_H
+#define _ASM_NIOS2_STRING_H
+
+#ifdef __KERNEL__
+
+#define __HAVE_ARCH_MEMSET
+#define __HAVE_ARCH_MEMCPY
+#define __HAVE_ARCH_MEMMOVE
+
+extern void *memset(void *s, int c, size_t count);
+extern void *memcpy(void *d, const void *s, size_t count);
+extern void *memmove(void *d, const void *s, size_t count);
+
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_NIOS2_STRING_H */
diff --git a/arch/nios2/lib/memcpy.c b/arch/nios2/lib/memcpy.c
new file mode 100644
index 0000000..ac7eec4
--- /dev/null
+++ b/arch/nios2/lib/memcpy.c
@@ -0,0 +1,199 @@
+/* Extracted from GLIBC memcpy.c and memcopy.h, which is:
+ Copyright (C) 1991, 1992, 1993, 1997, 2004 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Torbjorn Granlund ([email protected]).
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
+
+#include <linux/types.h>
+
+/* Type to use for aligned memory operations.
+ This should normally be the biggest type supported by a single load
+ and store. */
+#define op_t unsigned long int
+#define OPSIZ (sizeof(op_t))
+
+/* Optimal type for storing bytes in registers. */
+#define reg_char char
+
+#define MERGE(w0, sh_1, w1, sh_2) (((w0) >> (sh_1)) | ((w1) << (sh_2)))
+
+/* Copy exactly NBYTES bytes from SRC_BP to DST_BP,
+ without any assumptions about alignment of the pointers. */
+#define BYTE_COPY_FWD(dst_bp, src_bp, nbytes) \
+do { \
+ size_t __nbytes = (nbytes); \
+ while (__nbytes > 0) { \
+ unsigned char __x = ((unsigned char *) src_bp)[0]; \
+ src_bp += 1; \
+ __nbytes -= 1; \
+ ((unsigned char *) dst_bp)[0] = __x; \
+ dst_bp += 1; \
+ } \
+} while (0)
+
+/* Copy *up to* NBYTES bytes from SRC_BP to DST_BP, with
+ the assumption that DST_BP is aligned on an OPSIZ multiple. If
+ not all bytes could be easily copied, store remaining number of bytes
+ in NBYTES_LEFT, otherwise store 0. */
+/* extern void _wordcopy_fwd_aligned __P ((long int, long int, size_t)); */
+/* extern void _wordcopy_fwd_dest_aligned __P ((long int, long int, size_t)); */
+#define WORD_COPY_FWD(dst_bp, src_bp, nbytes_left, nbytes) \
+do { \
+ if (src_bp % OPSIZ == 0) \
+ _wordcopy_fwd_aligned(dst_bp, src_bp, (nbytes) / OPSIZ);\
+ else \
+ _wordcopy_fwd_dest_aligned(dst_bp, src_bp, (nbytes) / OPSIZ);\
+ src_bp += (nbytes) & -OPSIZ; \
+ dst_bp += (nbytes) & -OPSIZ; \
+ (nbytes_left) = (nbytes) % OPSIZ; \
+} while (0)
+
+
+/* Threshold value for when to enter the unrolled loops. */
+#define OP_T_THRES 16
+
+/* _wordcopy_fwd_aligned -- Copy block beginning at SRCP to
+ block beginning at DSTP with LEN `op_t' words (not LEN bytes!).
+ Both SRCP and DSTP should be aligned for memory operations on `op_t's. */
+/* stream-lined (read x8 + write x8) */
+static void _wordcopy_fwd_aligned(long int dstp, long int srcp, size_t len)
+{
+ while (len > 7) {
+ register op_t a0, a1, a2, a3, a4, a5, a6, a7;
+ a0 = ((op_t *) srcp)[0];
+ a1 = ((op_t *) srcp)[1];
+ a2 = ((op_t *) srcp)[2];
+ a3 = ((op_t *) srcp)[3];
+ a4 = ((op_t *) srcp)[4];
+ a5 = ((op_t *) srcp)[5];
+ a6 = ((op_t *) srcp)[6];
+ a7 = ((op_t *) srcp)[7];
+ ((op_t *) dstp)[0] = a0;
+ ((op_t *) dstp)[1] = a1;
+ ((op_t *) dstp)[2] = a2;
+ ((op_t *) dstp)[3] = a3;
+ ((op_t *) dstp)[4] = a4;
+ ((op_t *) dstp)[5] = a5;
+ ((op_t *) dstp)[6] = a6;
+ ((op_t *) dstp)[7] = a7;
+
+ srcp += 8 * OPSIZ;
+ dstp += 8 * OPSIZ;
+ len -= 8;
+ }
+ while (len > 0) {
+ *(op_t *)dstp = *(op_t *)srcp;
+
+ srcp += OPSIZ;
+ dstp += OPSIZ;
+ len -= 1;
+ }
+}
+
+/* _wordcopy_fwd_dest_aligned -- Copy block beginning at SRCP to
+ block beginning at DSTP with LEN `op_t' words (not LEN bytes!).
+ DSTP should be aligned for memory operations on `op_t's, but SRCP must
+ *not* be aligned. */
+/* stream-lined (read x4 + write x4) */
+static void _wordcopy_fwd_dest_aligned(long int dstp, long int srcp,
+ size_t len)
+{
+ op_t ap;
+ int sh_1, sh_2;
+
+ /* Calculate how to shift a word read at the memory operation
+ aligned srcp to make it aligned for copy. */
+
+ sh_1 = 8 * (srcp % OPSIZ);
+ sh_2 = 8 * OPSIZ - sh_1;
+
+ /* Make SRCP aligned by rounding it down to the beginning of the `op_t'
+ it points in the middle of. */
+ srcp &= -OPSIZ;
+ ap = ((op_t *) srcp)[0];
+ srcp += OPSIZ;
+
+ while (len > 3) {
+ op_t a0, a1, a2, a3;
+ a0 = ((op_t *) srcp)[0];
+ a1 = ((op_t *) srcp)[1];
+ a2 = ((op_t *) srcp)[2];
+ a3 = ((op_t *) srcp)[3];
+ ((op_t *) dstp)[0] = MERGE(ap, sh_1, a0, sh_2);
+ ((op_t *) dstp)[1] = MERGE(a0, sh_1, a1, sh_2);
+ ((op_t *) dstp)[2] = MERGE(a1, sh_1, a2, sh_2);
+ ((op_t *) dstp)[3] = MERGE(a2, sh_1, a3, sh_2);
+
+ ap = a3;
+ srcp += 4 * OPSIZ;
+ dstp += 4 * OPSIZ;
+ len -= 4;
+ }
+ while (len > 0) {
+ register op_t a0;
+ a0 = ((op_t *) srcp)[0];
+ ((op_t *) dstp)[0] = MERGE(ap, sh_1, a0, sh_2);
+
+ ap = a0;
+ srcp += OPSIZ;
+ dstp += OPSIZ;
+ len -= 1;
+ }
+}
+
+void *memcpy(void *dstpp, const void *srcpp, size_t len)
+{
+ unsigned long int dstp = (long int) dstpp;
+ unsigned long int srcp = (long int) srcpp;
+
+ /* Copy from the beginning to the end. */
+
+ /* If there not too few bytes to copy, use word copy. */
+ if (len >= OP_T_THRES) {
+ /* Copy just a few bytes to make DSTP aligned. */
+ len -= (-dstp) % OPSIZ;
+ BYTE_COPY_FWD(dstp, srcp, (-dstp) % OPSIZ);
+
+ /* Copy whole pages from SRCP to DSTP by virtual address
+ manipulation, as much as possible. */
+
+ /* PAGE_COPY_FWD_MAYBE (dstp, srcp, len, len); */
+
+ /* Copy from SRCP to DSTP taking advantage of the known
+ alignment of DSTP. Number of bytes remaining is put in the
+ third argument, i.e. in LEN. This number may vary from
+ machine to machine. */
+
+ WORD_COPY_FWD(dstp, srcp, len, len);
+
+ /* Fall out and copy the tail. */
+ }
+
+ /* There are just a few bytes to copy. Use byte memory operations. */
+ BYTE_COPY_FWD(dstp, srcp, len);
+
+ return dstpp;
+}
+
+void *memcpyb(void *dstpp, const void *srcpp, unsigned len)
+{
+ unsigned long int dstp = (long int) dstpp;
+ unsigned long int srcp = (long int) srcpp;
+
+ BYTE_COPY_FWD(dstp, srcp, len);
+
+ return dstpp;
+}
diff --git a/arch/nios2/lib/memmove.c b/arch/nios2/lib/memmove.c
new file mode 100644
index 0000000..c65ef517eb80c05e337c079ad49e709bbe847188
--- /dev/null
+++ b/arch/nios2/lib/memmove.c
@@ -0,0 +1,82 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/types.h>
+#include <linux/string.h>
+
+#ifdef __HAVE_ARCH_MEMMOVE
+void *memmove(void *d, const void *s, size_t count)
+{
+ unsigned long dst, src;
+
+ if (!count)
+ return d;
+
+ if (d < s) {
+ dst = (unsigned long) d;
+ src = (unsigned long) s;
+
+ if ((count < 8) || ((dst ^ src) & 3))
+ goto restup;
+
+ if (dst & 1) {
+ *(char *)dst++ = *(char *)src++;
+ count--;
+ }
+ if (dst & 2) {
+ *(short *)dst = *(short *)src;
+ src += 2;
+ dst += 2;
+ count -= 2;
+ }
+ while (count > 3) {
+ *(long *)dst = *(long *)src;
+ src += 4;
+ dst += 4;
+ count -= 4;
+ }
+restup:
+ while (count--)
+ *(char *)dst++ = *(char *)src++;
+ } else {
+ dst = (unsigned long) d + count;
+ src = (unsigned long) s + count;
+
+ if ((count < 8) || ((dst ^ src) & 3))
+ goto restdown;
+
+ if (dst & 1) {
+ src--;
+ dst--;
+ count--;
+ *(char *)dst = *(char *)src;
+ }
+ if (dst & 2) {
+ src -= 2;
+ dst -= 2;
+ count -= 2;
+ *(short *)dst = *(short *)src;
+ }
+ while (count > 3) {
+ src -= 4;
+ dst -= 4;
+ count -= 4;
+ *(long *)dst = *(long *)src;
+ }
+restdown:
+ while (count--) {
+ src--;
+ dst--;
+ *(char *)dst = *(char *)src;
+ }
+ }
+
+ return d;
+}
+#endif /* __HAVE_ARCH_MEMMOVE */
diff --git a/arch/nios2/lib/memset.c b/arch/nios2/lib/memset.c
new file mode 100644
index 0000000..65e97802f5cc83c15703c427e61aadf28b4d5b0c
--- /dev/null
+++ b/arch/nios2/lib/memset.c
@@ -0,0 +1,81 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/types.h>
+#include <linux/string.h>
+
+#ifdef __HAVE_ARCH_MEMSET
+void *memset(void *s, int c, size_t count)
+{
+ int destptr, charcnt, dwordcnt, fill8reg, wrkrega;
+
+ if (!count)
+ return s;
+
+ c &= 0xFF;
+
+ if (count <= 8) {
+ char *xs = (char *) s;
+
+ while (count--)
+ *xs++ = c;
+ return s;
+ }
+
+ __asm__ __volatile__ (
+ /* fill8 %3, %5 (c & 0xff) */
+ " slli %4, %5, 8\n"
+ " or %4, %4, %5\n"
+ " slli %3, %4, 16\n"
+ " or %3, %3, %4\n"
+ /* Word-align %0 (s) if necessary */
+ " andi %4, %0, 0x01\n"
+ " beq %4, zero, 1f\n"
+ " addi %1, %1, -1\n"
+ " stb %3, 0(%0)\n"
+ " addi %0, %0, 1\n"
+ "1: mov %2, %1\n"
+ /* Dword-align %0 (s) if necessary */
+ " andi %4, %0, 0x02\n"
+ " beq %4, zero, 2f\n"
+ " addi %1, %1, -2\n"
+ " sth %3, 0(%0)\n"
+ " addi %0, %0, 2\n"
+ " mov %2, %1\n"
+ /* %1 and %2 are how many more bytes to set */
+ "2: srli %2, %2, 2\n"
+ /* %2 is how many dwords to set */
+ "3: stw %3, 0(%0)\n"
+ " addi %0, %0, 4\n"
+ " addi %2, %2, -1\n"
+ " bne %2, zero, 3b\n"
+ /* store residual word and/or byte if necessary */
+ " andi %4, %1, 0x02\n"
+ " beq %4, zero, 4f\n"
+ " sth %3, 0(%0)\n"
+ " addi %0, %0, 2\n"
+ /* store residual byte if necessary */
+ "4: andi %4, %1, 0x01\n"
+ " beq %4, zero, 5f\n"
+ " stb %3, 0(%0)\n"
+ "5:\n"
+ : "=r" (destptr), /* %0 Output */
+ "=r" (charcnt), /* %1 Output */
+ "=r" (dwordcnt), /* %2 Output */
+ "=r" (fill8reg), /* %3 Output */
+ "=r" (wrkrega) /* %4 Output */
+ : "r" (c), /* %5 Input */
+ "0" (s), /* %0 Input/Output */
+ "1" (count) /* %1 Input/Output */
+ : "memory" /* clobbered */
+ );
+
+ return s;
+}
+#endif /* __HAVE_ARCH_MEMSET */
--
1.8.3.2

2014-04-18 12:43:33

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 06/28] nios2: Memory management

This patch contains the initialisation of the memory blocks, MMU
attributes and the memory map.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/mmu.h | 18 +++
arch/nios2/include/asm/page.h | 111 +++++++++++++++++++
arch/nios2/include/asm/uaccess.h | 233 +++++++++++++++++++++++++++++++++++++++
arch/nios2/mm/init.c | 152 +++++++++++++++++++++++++
arch/nios2/mm/uaccess.c | 162 +++++++++++++++++++++++++++
5 files changed, 676 insertions(+)
create mode 100644 arch/nios2/include/asm/mmu.h
create mode 100644 arch/nios2/include/asm/page.h
create mode 100644 arch/nios2/include/asm/uaccess.h
create mode 100644 arch/nios2/mm/init.c
create mode 100644 arch/nios2/mm/uaccess.c

diff --git a/arch/nios2/include/asm/mmu.h b/arch/nios2/include/asm/mmu.h
new file mode 100644
index 0000000..96ab052
--- /dev/null
+++ b/arch/nios2/include/asm/mmu.h
@@ -0,0 +1,18 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * Taken from m68knommu.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_MMU_H
+#define _ASM_NIOS2_MMU_H
+
+/* Default "unsigned long" context */
+typedef unsigned long mm_context_t;
+
+#endif /* _ASM_NIOS2_MMU_H */
diff --git a/arch/nios2/include/asm/page.h b/arch/nios2/include/asm/page.h
new file mode 100644
index 0000000..4f25e0fe6cb9666a713dd2738871915f73f9a1d7
--- /dev/null
+++ b/arch/nios2/include/asm/page.h
@@ -0,0 +1,111 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * MMU support based on asm/page.h from mips which is:
+ *
+ * Copyright (C) 1994 - 1999, 2000, 03 Ralf Baechle
+ * Copyright (C) 1999, 2000 Silicon Graphics, Inc.
+ *
+ * NOMMU support based on asm/page.h from m68knommu.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_PAGE_H
+#define _ASM_NIOS2_PAGE_H
+
+#include <linux/pfn.h>
+
+/*
+ * PAGE_SHIFT determines the page size
+ */
+#define PAGE_SHIFT 12
+#define PAGE_SIZE 4096
+#define PAGE_MASK (~(PAGE_SIZE - 1))
+
+/*
+ * PAGE_OFFSET -- the first address of the first page of memory.
+ */
+#define PAGE_OFFSET (CONFIG_MEM_BASE + CONFIG_KERNEL_REGION_BASE)
+
+#ifndef __ASSEMBLY__
+
+/*
+ * This gives the physical RAM offset.
+ */
+#define PHYS_OFFSET CONFIG_MEM_BASE
+
+/*
+ * It's normally defined only for FLATMEM config but it's
+ * used in our early mem init code for all memory models.
+ * So always define it.
+ */
+#define ARCH_PFN_OFFSET PFN_UP(PHYS_OFFSET)
+
+#define clear_page(page) memset((page), 0, PAGE_SIZE)
+#define copy_page(to, from) memcpy((to), (from), PAGE_SIZE)
+
+struct page;
+
+extern void clear_user_page(void *addr, unsigned long vaddr, struct page *page);
+extern void copy_user_page(void *vto, void *vfrom, unsigned long vaddr,
+ struct page *to);
+
+extern unsigned long shm_align_mask;
+
+/*
+ * These are used to make use of C type-checking.
+ */
+typedef struct page *pgtable_t;
+typedef struct { unsigned long pte; } pte_t;
+typedef struct { unsigned long pgd; } pgd_t;
+typedef struct { unsigned long pgprot; } pgprot_t;
+
+#define pte_val(x) ((x).pte)
+#define pgd_val(x) ((x).pgd)
+#define pgprot_val(x) ((x).pgprot)
+
+#define __pte(x) ((pte_t) { (x) })
+#define __pgd(x) ((pgd_t) { (x) })
+#define __pgprot(x) ((pgprot_t) { (x) })
+
+extern unsigned long memory_start;
+extern unsigned long memory_end;
+extern unsigned long memory_size;
+
+extern struct page *mem_map;
+
+#endif /* !__ASSEMBLY__ */
+
+# define __pa(x) \
+ ((unsigned long)(x) - PAGE_OFFSET + PHYS_OFFSET)
+# define __va(x) \
+ ((void *)((unsigned long)(x) + PAGE_OFFSET - PHYS_OFFSET))
+
+#define page_to_virt(page) \
+ ((((page) - mem_map) << PAGE_SHIFT) + PAGE_OFFSET)
+
+# define pfn_to_kaddr(pfn) __va((pfn) << PAGE_SHIFT)
+# define pfn_valid(pfn) ((pfn) >= ARCH_PFN_OFFSET && \
+ (pfn) < (max_mapnr + ARCH_PFN_OFFSET))
+
+# define virt_to_page(vaddr) pfn_to_page(PFN_DOWN(virt_to_phys(vaddr)))
+# define virt_addr_valid(vaddr) pfn_valid(PFN_DOWN(virt_to_phys(vaddr)))
+
+# define VM_DATA_DEFAULT_FLAGS (VM_READ | VM_WRITE | VM_EXEC | \
+ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
+
+# define UNCAC_ADDR(addr) \
+ ((void *)((unsigned)(addr) | CONFIG_IO_REGION_BASE))
+# define CAC_ADDR(addr) \
+ ((void *)(((unsigned)(addr) & ~CONFIG_IO_REGION_BASE) | \
+ CONFIG_KERNEL_REGION_BASE))
+
+#include <asm-generic/memory_model.h>
+
+#include <asm-generic/getorder.h>
+
+#endif /* _ASM_NIOS2_PAGE_H */
diff --git a/arch/nios2/include/asm/uaccess.h b/arch/nios2/include/asm/uaccess.h
new file mode 100644
index 0000000..3e36f536afaba24c308a8416a66b8a2573e042a5
--- /dev/null
+++ b/arch/nios2/include/asm/uaccess.h
@@ -0,0 +1,233 @@
+/*
+ * User space memory access functions for Nios II
+ *
+ * Copyright (C) 2010-2011, Tobias Klauser <[email protected]>
+ * Copyright (C) 2009, Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ *
+ * Based on asm/uaccess.h from m68knommu
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_UACCESS_H
+#define _ASM_NIOS2_UACCESS_H
+
+#include <linux/errno.h>
+#include <linux/thread_info.h>
+#include <linux/string.h>
+
+#include <asm/page.h>
+
+#define VERIFY_READ 0
+#define VERIFY_WRITE 1
+
+/*
+ * The exception table consists of pairs of addresses: the first is the
+ * address of an instruction that is allowed to fault, and the second is
+ * the address at which the program should continue. No registers are
+ * modified, so it is entirely up to the continuation code to figure out
+ * what to do.
+ *
+ * All the routines below use bits of fixup code that are out of line
+ * with the main instruction path. This means when everything is well,
+ * we don't even have to jump over them. Further, they do not intrude
+ * on our cache or tlb entries.
+ */
+struct exception_table_entry {
+ unsigned long insn;
+ unsigned long fixup;
+};
+
+extern int fixup_exception(struct pt_regs *regs);
+
+/*
+ * Segment stuff
+ */
+#define MAKE_MM_SEG(s) ((mm_segment_t) { (s) })
+#define USER_DS MAKE_MM_SEG(0x80000000UL)
+#define KERNEL_DS MAKE_MM_SEG(0)
+
+#define get_ds() (KERNEL_DS)
+
+#define get_fs() (current_thread_info()->addr_limit)
+#define set_fs(seg) (current_thread_info()->addr_limit = (seg))
+
+#define segment_eq(a, b) ((a).seg == (b).seg)
+
+#define __access_ok(addr, len) \
+ (((signed long)(((long)get_fs().seg) & \
+ ((long)(addr) | (((long)(addr)) + (len)) | (len)))) == 0)
+
+#define access_ok(type, addr, len) \
+ likely(__access_ok((unsigned long)(addr), (unsigned long)(len)))
+
+# define __EX_TABLE_SECTION ".section __ex_table,\"a\"\n"
+
+/*
+ * Zero Userspace
+ */
+
+static inline unsigned long __must_check __clear_user(void __user *to,
+ unsigned long n)
+{
+ __asm__ __volatile__ (
+ "1: stb zero, 0(%1)\n"
+ " addi %0, %0, -1\n"
+ " addi %1, %1, 1\n"
+ " bne %0, zero, 1b\n"
+ "2:\n"
+ __EX_TABLE_SECTION
+ ".word 1b, 2b\n"
+ ".previous\n"
+ : "=r" (n), "=r" (to)
+ : "0" (n), "1" (to)
+ );
+
+ return n;
+}
+
+static inline unsigned long __must_check clear_user(void __user *to,
+ unsigned long n)
+{
+ if (!access_ok(VERIFY_WRITE, to, n))
+ return n;
+ return __clear_user(to, n);
+}
+
+extern long __copy_from_user(void *to, const void __user *from,
+ unsigned long n);
+extern long __copy_to_user(void __user *to, const void *from, unsigned long n);
+
+static inline long copy_from_user(void *to, const void __user *from,
+ unsigned long n)
+{
+ if (!access_ok(VERIFY_READ, from, n))
+ return n;
+ return __copy_from_user(to, from, n);
+}
+
+static inline long copy_to_user(void __user *to, const void *from,
+ unsigned long n)
+{
+ if (!access_ok(VERIFY_WRITE, to, n))
+ return n;
+ return __copy_to_user(to, from, n);
+}
+
+extern long strncpy_from_user(char *__to, const char __user *__from,
+ long __len);
+extern long strnlen_user(const char __user *s, long n);
+
+#define __copy_from_user_inatomic __copy_from_user
+#define __copy_to_user_inatomic __copy_to_user
+
+/* Optimized macros */
+#define __get_user_asm(val, insn, addr, err) \
+{ \
+ __asm__ __volatile__( \
+ " movi %0, %3\n" \
+ "1: " insn " %1, 0(%2)\n" \
+ " movi %0, 0\n" \
+ "2:\n" \
+ " .section __ex_table,\"a\"\n" \
+ " .word 1b, 2b\n" \
+ " .previous" \
+ : "=&r" (err), "=r" (val) \
+ : "r" (addr), "i" (-EFAULT)); \
+}
+
+#define __get_user_unknown(val, size, ptr, err) do { \
+ err = 0; \
+ if (copy_from_user(&(val), ptr, size)) { \
+ err = -EFAULT; \
+ } \
+ } while (0)
+
+#define __get_user_common(val, size, ptr, err) \
+do { \
+ switch (size) { \
+ case 1: \
+ __get_user_asm(val, "ldbu", ptr, err); \
+ break; \
+ case 2: \
+ __get_user_asm(val, "ldhu", ptr, err); \
+ break; \
+ case 4: \
+ __get_user_asm(val, "ldw", ptr, err); \
+ break; \
+ default: \
+ __get_user_unknown(val, size, ptr, err); \
+ break; \
+ } \
+} while (0)
+
+#define __get_user(x, ptr) \
+ ({ \
+ long __gu_err = -EFAULT; \
+ const __typeof__(*(ptr)) __user *__gu_ptr = (ptr); \
+ unsigned long __gu_val; \
+ __get_user_common(__gu_val, sizeof(*(ptr)), __gu_ptr, __gu_err);\
+ (x) = (__typeof__(x))__gu_val; \
+ __gu_err; \
+ })
+
+#define get_user(x, ptr) \
+({ \
+ long __gu_err = -EFAULT; \
+ const __typeof__(*(ptr)) __user *__gu_ptr = (ptr); \
+ unsigned long __gu_val = 0; \
+ if (access_ok(VERIFY_READ, __gu_ptr, sizeof(*__gu_ptr))) \
+ __get_user_common(__gu_val, sizeof(*__gu_ptr), \
+ __gu_ptr, __gu_err); \
+ (x) = (__typeof__(x))__gu_val; \
+ __gu_err; \
+})
+
+#define __put_user_asm(val, insn, ptr, err) \
+{ \
+ __asm__ __volatile__( \
+ " movi %0, %3\n" \
+ "1: " insn " %1, 0(%2)\n" \
+ " movi %0, 0\n" \
+ "2:\n" \
+ " .section __ex_table,\"a\"\n" \
+ " .word 1b, 2b\n" \
+ " .previous\n" \
+ : "=&r" (err) \
+ : "r" (val), "r" (ptr), "i" (-EFAULT)); \
+}
+
+#define put_user(x, ptr) \
+({ \
+ long __pu_err = -EFAULT; \
+ __typeof__(*(ptr)) __user *__pu_ptr = (ptr); \
+ __typeof__(*(ptr)) __pu_val = (__typeof(*ptr))(x); \
+ if (access_ok(VERIFY_WRITE, __pu_ptr, sizeof(*__pu_ptr))) { \
+ switch (sizeof(*__pu_ptr)) { \
+ case 1: \
+ __put_user_asm(__pu_val, "stb", __pu_ptr, __pu_err); \
+ break; \
+ case 2: \
+ __put_user_asm(__pu_val, "sth", __pu_ptr, __pu_err); \
+ break; \
+ case 4: \
+ __put_user_asm(__pu_val, "stw", __pu_ptr, __pu_err); \
+ break; \
+ default: \
+ /* XXX: This looks wrong... */ \
+ __pu_err = 0; \
+ if (copy_to_user(__pu_ptr, &(__pu_val), \
+ sizeof(*__pu_ptr))) \
+ __pu_err = -EFAULT; \
+ break; \
+ } \
+ } \
+ __pu_err; \
+})
+
+#define __put_user(x, ptr) put_user(x, ptr)
+
+#endif /* _ASM_NIOS2_UACCESS_H */
diff --git a/arch/nios2/mm/init.c b/arch/nios2/mm/init.c
new file mode 100644
index 0000000..4c4538dbbea7e461ae049122957765ea89fa9ded
--- /dev/null
+++ b/arch/nios2/mm/init.c
@@ -0,0 +1,152 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * based on arch/m68k/mm/init.c
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/pagemap.h>
+#include <linux/bootmem.h>
+#include <linux/slab.h>
+#include <linux/binfmts.h>
+
+#include <asm/setup.h>
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/sections.h>
+#include <asm/tlb.h>
+#include <asm/mmu_context.h>
+#include <asm/cpuinfo.h>
+#include <asm/processor.h>
+
+pgd_t *pgd_current;
+
+/*
+ * paging_init() continues the virtual memory environment setup which
+ * was begun by the code in arch/head.S.
+ * The parameters are pointers to where to stick the starting and ending
+ * addresses of available kernel virtual memory.
+ */
+void __init paging_init(void)
+{
+ unsigned long zones_size[MAX_NR_ZONES];
+ unsigned long start_mem, end_mem;
+
+ memset(zones_size, 0, sizeof(zones_size));
+
+ /*
+ * Make sure start_mem is page aligned, otherwise bootmem and
+ * page_alloc get different views of the world.
+ */
+ start_mem = PHYS_OFFSET;
+ end_mem = memory_end;
+
+ pagetable_init();
+ pgd_current = swapper_pg_dir;
+
+ /*
+ * Set up SFC/DFC registers (user data space).
+ */
+ zones_size[ZONE_DMA] = ((end_mem - start_mem) >> PAGE_SHIFT);
+
+ /* pass the memory from the bootmem allocator to the main allocator */
+ free_area_init(zones_size);
+}
+
+void __init mem_init(void)
+{
+ unsigned long end_mem = memory_end; /* this must not include
+ kernel stack at top */
+
+ pr_debug("mem_init: start=%lx, end=%lx\n", memory_start, memory_end);
+
+ end_mem &= PAGE_MASK;
+ high_memory = __va(end_mem);
+
+ max_mapnr = ((unsigned long)end_mem) >> PAGE_SHIFT;
+
+ /* this will put all memory onto the freelists */
+ free_all_bootmem();
+ mem_init_print_info(NULL);
+}
+
+void __init mmu_init(void)
+{
+ flush_tlb_all();
+}
+
+#ifdef CONFIG_BLK_DEV_INITRD
+void __init free_initrd_mem(unsigned long start, unsigned long end)
+{
+ free_reserved_area((void *)start, (void *)end, -1, "initrd");
+}
+#endif
+
+void __init_refok free_initmem(void)
+{
+ free_initmem_default(-1);
+}
+
+#define __page_aligned(order) __aligned(PAGE_SIZE << (order))
+pgd_t swapper_pg_dir[PTRS_PER_PGD] __page_aligned(PGD_ORDER);
+pte_t invalid_pte_table[PTRS_PER_PTE] __page_aligned(PTE_ORDER);
+static struct page *kuser_page[1];
+
+static int alloc_kuser_page(void)
+{
+ extern char __kuser_helper_start[], __kuser_helper_end[];
+ int kuser_sz = __kuser_helper_end - __kuser_helper_start;
+ unsigned long vpage;
+
+ vpage = get_zeroed_page(GFP_ATOMIC);
+ if (!vpage)
+ return -ENOMEM;
+
+ /* Copy kuser helpers */
+ memcpy((void *)vpage, __kuser_helper_start, kuser_sz);
+
+ flush_icache_range(vpage, vpage + KUSER_SIZE);
+ kuser_page[0] = virt_to_page(vpage);
+
+ return 0;
+}
+arch_initcall(alloc_kuser_page);
+
+int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
+{
+ struct mm_struct *mm = current->mm;
+ int ret;
+
+ down_write(&mm->mmap_sem);
+
+ /* Map kuser helpers to user space address */
+ ret = install_special_mapping(mm, KUSER_BASE, KUSER_SIZE,
+ VM_READ | VM_EXEC | VM_MAYREAD |
+ VM_MAYEXEC, kuser_page);
+
+ up_write(&mm->mmap_sem);
+
+ return ret;
+}
+
+const char *arch_vma_name(struct vm_area_struct *vma)
+{
+ return (vma->vm_start == KUSER_BASE) ? "[kuser]" : NULL;
+}
diff --git a/arch/nios2/mm/uaccess.c b/arch/nios2/mm/uaccess.c
new file mode 100644
index 0000000..4a38ad37c18b4e10792df094d81087a3f32cd9e9
--- /dev/null
+++ b/arch/nios2/mm/uaccess.c
@@ -0,0 +1,162 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2009, Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ */
+
+#include <linux/export.h>
+#include <linux/uaccess.h>
+
+asm(".global __copy_from_user\n"
+ " .type __copy_from_user, @function\n"
+ "__copy_from_user:\n"
+ " movi r2,7\n"
+ " mov r3,r4\n"
+ " bge r2,r6,1f\n"
+ " xor r2,r4,r5\n"
+ " andi r2,r2,3\n"
+ " movi r7,3\n"
+ " beq r2,zero,4f\n"
+ "1: addi r6,r6,-1\n"
+ " movi r2,-1\n"
+ " beq r6,r2,3f\n"
+ " mov r7,r2\n"
+ "2: ldbu r2,0(r5)\n"
+ " addi r6,r6,-1\n"
+ " addi r5,r5,1\n"
+ " stb r2,0(r3)\n"
+ " addi r3,r3,1\n"
+ " bne r6,r7,2b\n"
+ "3:\n"
+ " addi r2,r6,1\n"
+ " ret\n"
+ "13:mov r2,r6\n"
+ " ret\n"
+ "4: andi r2,r4,1\n"
+ " cmpeq r2,r2,zero\n"
+ " beq r2,zero,7f\n"
+ "5: andi r2,r3,2\n"
+ " beq r2,zero,6f\n"
+ "9: ldhu r2,0(r5)\n"
+ " addi r6,r6,-2\n"
+ " addi r5,r5,2\n"
+ " sth r2,0(r3)\n"
+ " addi r3,r3,2\n"
+ "6: bge r7,r6,1b\n"
+ "10:ldw r2,0(r5)\n"
+ " addi r6,r6,-4\n"
+ " addi r5,r5,4\n"
+ " stw r2,0(r3)\n"
+ " addi r3,r3,4\n"
+ " br 6b\n"
+ "7: ldbu r2,0(r5)\n"
+ " addi r6,r6,-1\n"
+ " addi r5,r5,1\n"
+ " addi r3,r4,1\n"
+ " stb r2,0(r4)\n"
+ " br 5b\n"
+ ".section __ex_table,\"a\"\n"
+ ".word 2b,3b\n"
+ ".word 9b,13b\n"
+ ".word 10b,13b\n"
+ ".word 7b,13b\n"
+ ".previous\n"
+ );
+EXPORT_SYMBOL(__copy_from_user);
+
+asm(
+ " .global __copy_to_user\n"
+ " .type __copy_to_user, @function\n"
+ "__copy_to_user:\n"
+ " movi r2,7\n"
+ " mov r3,r4\n"
+ " bge r2,r6,1f\n"
+ " xor r2,r4,r5\n"
+ " andi r2,r2,3\n"
+ " movi r7,3\n"
+ " beq r2,zero,4f\n"
+ /* Bail if we try to copy zero bytes */
+ "1: addi r6,r6,-1\n"
+ " movi r2,-1\n"
+ " beq r6,r2,3f\n"
+ /* Copy byte by byte for small copies and if src^dst != 0 */
+ " mov r7,r2\n"
+ "2: ldbu r2,0(r5)\n"
+ " addi r5,r5,1\n"
+ "9: stb r2,0(r3)\n"
+ " addi r6,r6,-1\n"
+ " addi r3,r3,1\n"
+ " bne r6,r7,2b\n"
+ "3: addi r2,r6,1\n"
+ " ret\n"
+ "13:mov r2,r6\n"
+ " ret\n"
+ /* If 'to' is an odd address byte copy */
+ "4: andi r2,r4,1\n"
+ " cmpeq r2,r2,zero\n"
+ " beq r2,zero,7f\n"
+ /* If 'to' is not divideable by four copy halfwords */
+ "5: andi r2,r3,2\n"
+ " beq r2,zero,6f\n"
+ " ldhu r2,0(r5)\n"
+ " addi r5,r5,2\n"
+ "10:sth r2,0(r3)\n"
+ " addi r6,r6,-2\n"
+ " addi r3,r3,2\n"
+ /* Copy words */
+ "6: bge r7,r6,1b\n"
+ " ldw r2,0(r5)\n"
+ " addi r5,r5,4\n"
+ "11:stw r2,0(r3)\n"
+ " addi r6,r6,-4\n"
+ " addi r3,r3,4\n"
+ " br 6b\n"
+ /* Copy remaining bytes */
+ "7: ldbu r2,0(r5)\n"
+ " addi r5,r5,1\n"
+ " addi r3,r4,1\n"
+ "12: stb r2,0(r4)\n"
+ " addi r6,r6,-1\n"
+ " br 5b\n"
+ ".section __ex_table,\"a\"\n"
+ ".word 9b,3b\n"
+ ".word 10b,13b\n"
+ ".word 11b,13b\n"
+ ".word 12b,13b\n"
+ ".previous\n");
+EXPORT_SYMBOL(__copy_to_user);
+
+long strncpy_from_user(char *__to, const char __user *__from, long __len)
+{
+ int l = strnlen_user(__from, __len);
+ int is_zt = 1;
+
+ if (l > __len) {
+ is_zt = 0;
+ l = __len;
+ }
+
+ if (l == 0 || copy_from_user(__to, __from, l))
+ return -EFAULT;
+
+ if (is_zt)
+ l--;
+ return l;
+}
+
+long strnlen_user(const char __user *s, long n)
+{
+ long i;
+
+ for (i = 0; i < n; i++) {
+ char c;
+ if (get_user(c, s + i) == -EFAULT)
+ return 0;
+ if (c == 0)
+ return i + 1;
+ }
+ return n + 1;
+}
--
1.8.3.2

2014-04-18 12:43:55

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 16/28] nios2: System calls handling

This patch adds support for system calls from userspaces. It uses the
asm-generic/unistd.h definitions with architecture spcific syscall. The sys_call_table
is just an array defined in a C file and it contains pointers to the syscall functions.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/syscalls.h | 25 +++++++++++++
arch/nios2/include/uapi/asm/unistd.h | 25 +++++++++++++
arch/nios2/kernel/sys_nios2.c | 72 ++++++++++++++++++++++++++++++++++++
arch/nios2/kernel/syscall_table.c | 29 +++++++++++++++
4 files changed, 151 insertions(+)
create mode 100644 arch/nios2/include/asm/syscalls.h
create mode 100644 arch/nios2/include/uapi/asm/unistd.h
create mode 100644 arch/nios2/kernel/sys_nios2.c
create mode 100644 arch/nios2/kernel/syscall_table.c

diff --git a/arch/nios2/include/asm/syscalls.h b/arch/nios2/include/asm/syscalls.h
new file mode 100644
index 0000000..0245d780351b35a11bfe132377c9f5230cbebd8e
--- /dev/null
+++ b/arch/nios2/include/asm/syscalls.h
@@ -0,0 +1,25 @@
+/*
+ * Copyright Altera Corporation (C) 2013. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+#ifndef __ASM_NIOS2_SYSCALLS_H
+#define __ASM_NIOS2_SYSCALLS_H
+
+int sys_cacheflush(unsigned long addr, unsigned long len,
+ unsigned int op);
+
+#include <asm-generic/syscalls.h>
+
+#endif /* __ASM_NIOS2_SYSCALLS_H */
diff --git a/arch/nios2/include/uapi/asm/unistd.h b/arch/nios2/include/uapi/asm/unistd.h
new file mode 100644
index 0000000..c4bf7951046153b3a39f5c11a21dadfe3de0f74a
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/unistd.h
@@ -0,0 +1,25 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+ #define sys_mmap2 sys_mmap_pgoff
+
+/* Use the standard ABI for syscalls */
+#include <asm-generic/unistd.h>
+
+/* Additional Nios II specific syscalls. */
+#define __NR_cacheflush (__NR_arch_specific_syscall)
+__SYSCALL(__NR_cacheflush, sys_cacheflush)
diff --git a/arch/nios2/kernel/sys_nios2.c b/arch/nios2/kernel/sys_nios2.c
new file mode 100644
index 0000000..994bf8ad53c4a1c68a658bc936a44812235e46c3
--- /dev/null
+++ b/arch/nios2/kernel/sys_nios2.c
@@ -0,0 +1,72 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2011-2012 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/export.h>
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/slab.h>
+#include <linux/syscalls.h>
+
+#include <asm/cacheflush.h>
+#include <asm/traps.h>
+
+asmlinkage long sys_mmap(unsigned long addr, unsigned long len,
+ unsigned long prot, unsigned long flags,
+ unsigned long fd, unsigned long offset)
+{
+ if (offset & ~PAGE_MASK)
+ return -EINVAL;
+
+ return sys_mmap_pgoff(addr, len, prot, flags, fd, offset >> PAGE_SHIFT);
+}
+
+/* sys_cacheflush -- flush the processor cache. */
+asmlinkage int sys_cacheflush(unsigned long addr, unsigned long len,
+ unsigned int op)
+{
+ struct vm_area_struct *vma;
+
+ if (len == 0)
+ return 0;
+
+ /* Check for overflow */
+ if (addr + len < addr)
+ return -EFAULT;
+
+ /*
+ * Verify that the specified address region actually belongs
+ * to this process.
+ */
+ vma = find_vma(current->mm, addr);
+ if (vma == NULL || addr < vma->vm_start || addr + len > vma->vm_end)
+ return -EFAULT;
+
+ /* Ignore the scope and cache arguments. */
+ flush_cache_range(vma, addr, addr + len);
+
+ return 0;
+}
+
+asmlinkage int sys_getpagesize(void)
+{
+ return PAGE_SIZE;
+}
+
+#if defined(CONFIG_FB) || defined(CONFIG_FB_MODULE)
+#include <linux/fb.h>
+unsigned long get_fb_unmapped_area(struct file *filp, unsigned long orig_addr,
+ unsigned long len, unsigned long pgoff, unsigned long flags)
+{
+
+ struct fb_info *info = filp->private_data;
+ return info->screen_base;
+}
+EXPORT_SYMBOL(get_fb_unmapped_area);
+#endif /* CONFIG_FB */
diff --git a/arch/nios2/kernel/syscall_table.c b/arch/nios2/kernel/syscall_table.c
new file mode 100644
index 0000000..06e6ac1835b2ed93f420a4ab0d1e06911144486b
--- /dev/null
+++ b/arch/nios2/kernel/syscall_table.c
@@ -0,0 +1,29 @@
+/*
+ * Copyright Altera Corporation (C) 2013. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include <linux/syscalls.h>
+#include <linux/signal.h>
+#include <linux/unistd.h>
+
+#include <asm/syscalls.h>
+
+#undef __SYSCALL
+#define __SYSCALL(nr, call) [nr] = (call),
+
+void *sys_call_table[__NR_syscalls] = {
+#include <asm/unistd.h>
+};
--
1.8.3.2

2014-04-18 12:44:11

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 19/28] nios2: Device tree support

Add device tree support to arch/nios2.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/boot/dts/3c120_devboard.dts | 205 +++++++++++++++++++++++++++++++++
arch/nios2/boot/linked_dtb.S | 19 +++
arch/nios2/include/asm/prom.h | 26 +++++
arch/nios2/kernel/prom.c | 67 +++++++++++
arch/nios2/platform/platform.c | 69 +++++++++++
5 files changed, 386 insertions(+)
create mode 100644 arch/nios2/boot/dts/3c120_devboard.dts
create mode 100644 arch/nios2/boot/linked_dtb.S
create mode 100644 arch/nios2/include/asm/prom.h
create mode 100644 arch/nios2/kernel/prom.c
create mode 100644 arch/nios2/platform/platform.c

diff --git a/arch/nios2/boot/dts/3c120_devboard.dts b/arch/nios2/boot/dts/3c120_devboard.dts
new file mode 100644
index 0000000..cad29a9
--- /dev/null
+++ b/arch/nios2/boot/dts/3c120_devboard.dts
@@ -0,0 +1,205 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ * This file is generated by sopc2dts.
+ */
+
+/dts-v1/;
+
+/ {
+ model = "ALTR,qsys_ghrd_3c120";
+ compatible = "ALTR,qsys_ghrd_3c120";
+ #address-cells = < 1 >;
+ #size-cells = < 1 >;
+
+ cpus {
+ #address-cells = < 1 >;
+ #size-cells = < 0 >;
+
+ cpu: cpu@0x0 {
+ device_type = "cpu";
+ compatible = "ALTR,nios2-1.0";
+ reg = < 0x00000000 >;
+ interrupt-controller;
+ #interrupt-cells = < 1 >;
+ clock-frequency = < 125000000 >;
+ dcache-line-size = < 32 >;
+ icache-line-size = < 32 >;
+ dcache-size = < 32768 >;
+ icache-size = < 32768 >;
+ ALTR,implementation = "fast";
+ ALTR,pid-num-bits = < 8 >;
+ ALTR,tlb-num-ways = < 16 >;
+ ALTR,tlb-num-entries = < 128 >;
+ ALTR,tlb-ptr-sz = < 7 >;
+ ALTR,has-div = < 1 >;
+ ALTR,has-mul = < 1 >;
+ ALTR,reset-addr = < 0xc2800000 >;
+ ALTR,fast-tlb-miss-addr = < 0xc7fff400 >;
+ ALTR,exception-addr = < 0xd0000020 >;
+ ALTR,has-initda = < 1 >;
+ ALTR,has-mmu = < 1 >;
+ }; //end cpu@0x0 (cpu)
+ }; //end cpus
+
+ memory@0 {
+ device_type = "memory";
+ reg = < 0x10000000 0x08000000
+ 0x07FFF400 0x00000400 >;
+ }; //end memory@0
+
+ sopc@0 {
+ device_type = "soc";
+ ranges;
+ #address-cells = < 1 >;
+ #size-cells = < 1 >;
+ compatible = "ALTR,avalon", "simple-bus";
+ bus-frequency = < 125000000 >;
+
+ pb_cpu_to_io: bridge@0x8000000 {
+ compatible = "simple-bus";
+ reg = < 0x08000000 0x00800000 >;
+ #address-cells = < 1 >;
+ #size-cells = < 1 >;
+ ranges = < 0x00400000 0x08400000 0x00000020
+ 0x00004D40 0x08004D40 0x00000008
+ 0x00004D50 0x08004D50 0x00000008
+ 0x00004000 0x08004000 0x00000400
+ 0x00004400 0x08004400 0x00000040
+ 0x00004800 0x08004800 0x00000040
+ 0x00002000 0x08002000 0x00002000
+ 0x00004C80 0x08004C80 0x00000020
+ 0x00004CC0 0x08004CC0 0x00000010
+ 0x00004CE0 0x08004CE0 0x00000010
+ 0x00004D00 0x08004D00 0x00000010 >;
+
+ timer_1ms: timer@0x400000 {
+ compatible = "ALTR,timer-1.0";
+ reg = < 0x00400000 0x00000020 >;
+ interrupt-parent = < &cpu >;
+ interrupts = < 11 >;
+ clock-frequency = < 125000000 >;
+ }; //end timer@0x400000 (timer_1ms)
+
+ sysid: sysid@0x4d40 {
+ compatible = "ALTR,sysid-1.0";
+ reg = < 0x00004D40 0x00000008 >;
+ id = < 0 >;
+ timestamp = < 1364882880 >;
+ }; //end sysid@0x4d40 (sysid)
+
+ jtag_uart: serial@0x4d50 {
+ compatible = "ALTR,juart-1.0";
+ reg = < 0x00004D50 0x00000008 >;
+ interrupt-parent = < &cpu >;
+ interrupts = < 1 >;
+ }; //end serial@0x4d50 (jtag_uart)
+
+ tse_mac: ethernet@0x4000 {
+ compatible = "ALTR,tse-1.0";
+ reg = < 0x00004000 0x00000400
+ 0x00004400 0x00000040
+ 0x00004800 0x00000040
+ 0x00002000 0x00002000 >;
+ reg-names = "control_port", "rx_csr", "tx_csr", "s1";
+ interrupt-parent = < &cpu >;
+ interrupts = < 2 3 >;
+ interrupt-names = "rx_irq", "tx_irq";
+ rx-fifo-depth = < 8192 >;
+ tx-fifo-depth = < 8192 >;
+ address-bits = < 48 >;
+ max-frame-size = < 1518 >;
+ local-mac-address = [ 02 00 00 00 00 00 ];
+ phy-mode = "rgmii-id";
+ ALTR,mii-id = < 0 >;
+ phy-handle = < &phy0 >;
+ tse_mac_mdio: mdio {
+ compatible = "altr,tse-mdio";
+ #address-cells = < 1 >;
+ #size-cells = < 0 >;
+ phy0: ethernet-phy@18 {
+ reg = < 18 >;
+ device_type = "ethernet-phy";
+ };
+ };
+ }; //end ethernet@0x4000 (tse_mac)
+
+ uart: serial@0x4c80 {
+ compatible = "ALTR,uart-1.0";
+ reg = < 0x00004C80 0x00000020 >;
+ interrupt-parent = < &cpu >;
+ interrupts = < 10 >;
+ current-speed = < 115200 >;
+ clock-frequency = < 62500000 >;
+ }; //end serial@0x4c80 (uart)
+
+ user_led_pio_8out: gpio@0x4cc0 {
+ compatible = "ALTR,pio-1.0";
+ reg = < 0x00004CC0 0x00000010 >;
+ width = < 8 >;
+ resetvalue = < 255 >;
+ #gpio-cells = < 2 >;
+ gpio-controller;
+ }; //end gpio@0x4cc0 (user_led_pio_8out)
+
+ user_dipsw_pio_8in: gpio@0x4ce0 {
+ compatible = "ALTR,pio-1.0";
+ reg = < 0x00004CE0 0x00000010 >;
+ interrupt-parent = < &cpu >;
+ interrupts = < 8 >;
+ width = < 8 >;
+ resetvalue = < 0 >;
+ edge_type = < 2 >;
+ level_trigger = < 0 >;
+ altr,interrupt_type = < 3 >;
+ #gpio-cells = < 2 >;
+ gpio-controller;
+ }; //end gpio@0x4ce0 (user_dipsw_pio_8in)
+
+ user_pb_pio_4in: gpio@0x4d00 {
+ compatible = "ALTR,pio-1.0";
+ reg = < 0x00004D00 0x00000010 >;
+ interrupt-parent = < &cpu >;
+ interrupts = < 9 >;
+ width = < 4 >;
+ resetvalue = < 0 >;
+ edge_type = < 2 >;
+ level_trigger = < 0 >;
+ altr,interrupt_type = < 3 >;
+ #gpio-cells = < 2 >;
+ gpio-controller;
+ }; //end gpio@0x4d00 (user_pb_pio_4in)
+ }; //end bridge@0x8000000 (pb_cpu_to_io)
+
+ cfi_flash_64m: flash@0x0 {
+ compatible = "cfi-flash";
+ reg = < 0x00000000 0x04000000 >;
+ bank-width = < 2 >;
+ device-width = < 1 >;
+ #address-cells = < 1 >;
+ #size-cells = < 1 >;
+
+ partition@800000 {
+ reg = < 0x00800000 0x01E00000 >;
+ label = "JFFS2 Filesystem";
+ }; //end partition@800000
+ }; //end flash@0x0 (cfi_flash_64m)
+ }; //end sopc@0
+
+ chosen {
+ bootargs = "debug console=ttyJ0,115200";
+ }; //end chosen
+}; //end /
diff --git a/arch/nios2/boot/linked_dtb.S b/arch/nios2/boot/linked_dtb.S
new file mode 100644
index 0000000..071f922db338e2cb4064bc77bf346f50e584d04f
--- /dev/null
+++ b/arch/nios2/boot/linked_dtb.S
@@ -0,0 +1,19 @@
+/*
+ * Copyright (C) 2011 Thomas Chou <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+.section .dtb.init.rodata,"a"
+.incbin "arch/nios2/boot/system.dtb"
diff --git a/arch/nios2/include/asm/prom.h b/arch/nios2/include/asm/prom.h
new file mode 100644
index 0000000..7a71e4c50a42134ffceabd233f83b5d1c48917c1
--- /dev/null
+++ b/arch/nios2/include/asm/prom.h
@@ -0,0 +1,26 @@
+/*
+ * Copyright (C) 2010 Thomas Chou <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_PROM_H
+#define _ASM_NIOS2_PROM_H
+
+extern int __dtb_start;
+
+extern unsigned long early_altera_uart_or_juart_console(void);
+
+#endif /* _ASM_NIOS2_PROM_H */
diff --git a/arch/nios2/kernel/prom.c b/arch/nios2/kernel/prom.c
new file mode 100644
index 0000000..a2d8308b939c33496e1416253b37c018d88c0c7f
--- /dev/null
+++ b/arch/nios2/kernel/prom.c
@@ -0,0 +1,67 @@
+/*
+ * Device tree support
+ *
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2010 Thomas Chou <[email protected]>
+ *
+ * Based on MIPS support for CONFIG_OF device tree support
+ *
+ * Copyright (C) 2010 Cisco Systems Inc. <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/bootmem.h>
+#include <linux/of.h>
+#include <linux/of_fdt.h>
+#include <linux/io.h>
+
+#include <asm/prom.h>
+#include <asm/sections.h>
+
+void __init early_init_dt_add_memory_arch(u64 base, u64 size)
+{
+ u64 kernel_start = (u64)virt_to_phys(_text);
+
+ if (!memory_size &&
+ (kernel_start >= base) && (kernel_start < (base + size)))
+ memory_size = size;
+
+ return;
+}
+
+void * __init early_init_dt_alloc_memory_arch(u64 size, u64 align)
+{
+ return __alloc_bootmem(size, align, __pa(MAX_DMA_ADDRESS));
+}
+
+void __init early_init_devtree(void *params)
+{
+ if (params && be32_to_cpup((__be32 *)params) == OF_DT_HEADER)
+ initial_boot_params = params;
+#if defined(CONFIG_NIOS2_DTB_AT_PHYS_ADDR)
+ else if (be32_to_cpup((__be32 *)CONFIG_NIOS2_DTB_PHYS_ADDR) ==
+ OF_DT_HEADER)
+ initial_boot_params = (void *)CONFIG_NIOS2_DTB_PHYS_ADDR;
+#endif
+ else if (be32_to_cpu((__be32)__dtb_start) == OF_DT_HEADER)
+ initial_boot_params = (void *)&__dtb_start;
+ else
+ return;
+
+ early_init_dt_scan(initial_boot_params);
+}
diff --git a/arch/nios2/platform/platform.c b/arch/nios2/platform/platform.c
new file mode 100644
index 0000000..2e5b8f624f4d593a2d9bf544ea7cc06b076b44b4
--- /dev/null
+++ b/arch/nios2/platform/platform.c
@@ -0,0 +1,69 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2011 Thomas Chou
+ * Copyright (C) 2011 Walter Goossens
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License. See the file COPYING in the main directory of this
+ * archive for more details.
+ */
+
+#include <linux/init.h>
+#include <linux/of_platform.h>
+#include <linux/of_address.h>
+#include <linux/of_fdt.h>
+#include <linux/err.h>
+#include <linux/slab.h>
+#include <linux/sys_soc.h>
+#include <linux/io.h>
+
+#define NIOS2_ID_DEFAULT (0x1)
+#define NIOS2_REVISION_DEFAULT (0x1)
+
+static struct of_device_id altera_of_bus_ids[] __initdata = {
+ { .compatible = "simple-bus", },
+ { .compatible = "altr,avalon", },
+ {}
+};
+
+static void __init nios2_soc_device_init(void)
+{
+ struct soc_device *soc_dev;
+ struct soc_device_attribute *soc_dev_attr;
+ const char *machine;
+
+ machine = of_flat_dt_get_machine_name();
+ if (!machine)
+ return;
+
+ soc_dev_attr = kzalloc(sizeof(*soc_dev_attr), GFP_KERNEL);
+ if (!soc_dev_attr)
+ return;
+
+ soc_dev_attr->soc_id = kasprintf(GFP_KERNEL, "%u", NIOS2_ID_DEFAULT);
+ soc_dev_attr->revision = kasprintf(GFP_KERNEL, "%d",
+ NIOS2_REVISION_DEFAULT);
+ soc_dev_attr->machine = kasprintf(GFP_KERNEL, "%s", machine);
+ soc_dev_attr->family = "Nios II";
+
+ soc_dev = soc_device_register(soc_dev_attr);
+ if (IS_ERR_OR_NULL(soc_dev)) {
+ kfree(soc_dev_attr->soc_id);
+ kfree(soc_dev_attr->machine);
+ kfree(soc_dev_attr->revision);
+ kfree(soc_dev_attr);
+ return;
+ }
+
+ return;
+}
+
+static int __init nios2_device_probe(void)
+{
+ nios2_soc_device_init();
+
+ of_platform_bus_probe(NULL, altera_of_bus_ids, NULL);
+ return 0;
+}
+
+device_initcall(nios2_device_probe);
--
1.8.3.2

2014-04-18 12:44:05

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 13/28] nios2: Interrupt handling

This patch adds the support for IRQ handling.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/irq.h | 34 +++++++++++++++
arch/nios2/include/asm/irqflags.h | 69 ++++++++++++++++++++++++++++++
arch/nios2/kernel/irq.c | 90 +++++++++++++++++++++++++++++++++++++++
3 files changed, 193 insertions(+)
create mode 100644 arch/nios2/include/asm/irq.h
create mode 100644 arch/nios2/include/asm/irqflags.h
create mode 100644 arch/nios2/kernel/irq.c

diff --git a/arch/nios2/include/asm/irq.h b/arch/nios2/include/asm/irq.h
new file mode 100644
index 0000000..f7e94b9
--- /dev/null
+++ b/arch/nios2/include/asm/irq.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_IRQ_H
+#define _ASM_NIOS2_IRQ_H
+
+#define NIOS2_CPU_NR_IRQS 32
+/* Reserve 32 additional interrupts for GPIO IRQs */
+#define NR_IRQS (NIOS2_CPU_NR_IRQS + 32)
+
+#ifndef NO_IRQ
+#define NO_IRQ (-1)
+#endif
+
+#include <asm-generic/irq.h>
+#include <linux/irqdomain.h>
+
+#endif
diff --git a/arch/nios2/include/asm/irqflags.h b/arch/nios2/include/asm/irqflags.h
new file mode 100644
index 0000000..e91f4f0b7c69454f7dd89590ddebccff73aad62a
--- /dev/null
+++ b/arch/nios2/include/asm/irqflags.h
@@ -0,0 +1,69 @@
+/*
+ * Copyright (C) 2010 Thomas Chou <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+#ifndef _ASM_IRQFLAGS_H
+#define _ASM_IRQFLAGS_H
+
+#include <asm/registers.h>
+
+static inline unsigned long arch_local_save_flags(void)
+{
+ return RDCTL(CTL_STATUS);
+}
+
+/*
+ * This will restore ALL status register flags, not only the interrupt
+ * mask flag.
+ */
+static inline void arch_local_irq_restore(unsigned long flags)
+{
+ WRCTL(CTL_STATUS, flags);
+}
+
+static inline void arch_local_irq_disable(void)
+{
+ unsigned long flags;
+ flags = arch_local_save_flags();
+ arch_local_irq_restore(flags & ~STATUS_PIE);
+}
+
+static inline void arch_local_irq_enable(void)
+{
+ unsigned long flags;
+ flags = arch_local_save_flags();
+ arch_local_irq_restore(flags | STATUS_PIE);
+}
+
+static inline int arch_irqs_disabled_flags(unsigned long flags)
+{
+ return (flags & STATUS_PIE) == 0;
+}
+
+static inline int arch_irqs_disabled(void)
+{
+ return arch_irqs_disabled_flags(arch_local_save_flags());
+}
+
+static inline unsigned long arch_local_irq_save(void)
+{
+ unsigned long flags;
+ flags = arch_local_save_flags();
+ arch_local_irq_restore(flags & ~STATUS_PIE);
+ return flags;
+}
+
+#endif /* _ASM_IRQFLAGS_H */
diff --git a/arch/nios2/kernel/irq.c b/arch/nios2/kernel/irq.c
new file mode 100644
index 0000000..e9265bd
--- /dev/null
+++ b/arch/nios2/kernel/irq.c
@@ -0,0 +1,90 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2008 Thomas Chou <[email protected]>
+ *
+ * based on irq.c from m68k which is:
+ *
+ * Copyright (C) 2007 Greg Ungerer <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/of.h>
+
+asmlinkage void do_IRQ(int hwirq, struct pt_regs *regs)
+{
+ struct pt_regs *oldregs = set_irq_regs(regs);
+ int irq;
+
+ irq_enter();
+ irq = irq_find_mapping(NULL, hwirq);
+ generic_handle_irq(irq);
+ irq_exit();
+
+ set_irq_regs(oldregs);
+}
+
+static void chip_unmask(struct irq_data *d)
+{
+ u32 ien;
+ ien = RDCTL(CTL_IENABLE);
+ ien |= (1 << d->hwirq);
+ WRCTL(CTL_IENABLE, ien);
+}
+
+static void chip_mask(struct irq_data *d)
+{
+ u32 ien;
+ ien = RDCTL(CTL_IENABLE);
+ ien &= ~(1 << d->hwirq);
+ WRCTL(CTL_IENABLE, ien);
+}
+
+static struct irq_chip m_irq_chip = {
+ .name = "NIOS2-INTC",
+ .irq_unmask = chip_unmask,
+ .irq_mask = chip_mask,
+};
+
+static int irq_map(struct irq_domain *h, unsigned int virq,
+ irq_hw_number_t hw_irq_num)
+{
+ irq_set_chip_and_handler(virq, &m_irq_chip, handle_level_irq);
+
+ return 0;
+}
+
+static struct irq_domain_ops irq_ops = {
+ .map = irq_map,
+ .xlate = irq_domain_xlate_onecell,
+};
+
+void __init init_IRQ(void)
+{
+ struct irq_domain *domain;
+ struct device_node *node;
+
+ node = of_find_compatible_node(NULL, NULL, "ALTR,nios2-1.0");
+ BUG_ON(!node);
+
+ domain = irq_domain_add_linear(node, NIOS2_CPU_NR_IRQS, &irq_ops, NULL);
+ BUG_ON(!domain);
+
+ irq_set_default_host(domain);
+ of_node_put(node);
+}
--
1.8.3.2

2014-04-18 12:44:35

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 24/28] nios2: Module support

This patch adds support for loadable modules.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/kernel/module.c | 135 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 135 insertions(+)
create mode 100644 arch/nios2/kernel/module.c

diff --git a/arch/nios2/kernel/module.c b/arch/nios2/kernel/module.c
new file mode 100644
index 0000000..360c53ab29f64f4e44c8cfe27f7176d793578569
--- /dev/null
+++ b/arch/nios2/kernel/module.c
@@ -0,0 +1,135 @@
+/*
+ * Kernel module support for Nios II.
+ *
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ * Written by Wentao Xu <[email protected]>
+ * Copyright (C) 2001, 2003 Rusty Russell
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License. See the file COPYING in the main directory of this
+ * archive for more details.
+ */
+
+#include <linux/moduleloader.h>
+#include <linux/elf.h>
+#include <linux/mm.h>
+#include <linux/vmalloc.h>
+#include <linux/slab.h>
+#include <linux/fs.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+
+#include <asm/pgtable.h>
+#include <asm/cacheflush.h>
+
+void *module_alloc(unsigned long size)
+{
+ if (size == 0)
+ return NULL;
+ return kmalloc(size, GFP_KERNEL);
+}
+
+/* Free memory returned from module_alloc */
+void module_free(struct module *mod, void *module_region)
+{
+ kfree(module_region);
+}
+
+int apply_relocate_add(Elf32_Shdr *sechdrs, const char *strtab,
+ unsigned int symindex, unsigned int relsec,
+ struct module *mod)
+{
+ unsigned int i;
+ Elf32_Rela *rela = (void *)sechdrs[relsec].sh_addr;
+
+ pr_debug("Applying relocate section %u to %u\n", relsec,
+ sechdrs[relsec].sh_info);
+
+ for (i = 0; i < sechdrs[relsec].sh_size / sizeof(*rela); i++) {
+ /* This is where to make the change */
+ uint32_t word;
+ uint32_t *loc
+ = ((void *)sechdrs[sechdrs[relsec].sh_info].sh_addr
+ + rela[i].r_offset);
+ /* This is the symbol it is referring to. Note that all
+ undefined symbols have been resolved. */
+ Elf32_Sym *sym
+ = ((Elf32_Sym *)sechdrs[symindex].sh_addr
+ + ELF32_R_SYM(rela[i].r_info));
+ uint32_t v = sym->st_value + rela[i].r_addend;
+ pr_debug("reltype %d 0x%x name:<%s>\n",
+ ELF32_R_TYPE(rela[i].r_info),
+ rela[i].r_offset, strtab + sym->st_name);
+
+ switch (ELF32_R_TYPE(rela[i].r_info)) {
+ case R_NIOS2_NONE:
+ break;
+ case R_NIOS2_BFD_RELOC_32:
+ *loc += v;
+ break;
+ case R_NIOS2_PCREL16:
+ v -= (uint32_t)loc + 4;
+ if ((int32_t)v > 0x7fff ||
+ (int32_t)v < -(int32_t)0x8000) {
+ pr_err("module %s: relocation overflow\n",
+ mod->name);
+ return -ENOEXEC;
+ }
+ word = *loc;
+ *loc = ((((word >> 22) << 16) | (v & 0xffff)) << 6) |
+ (word & 0x3f);
+ break;
+ case R_NIOS2_CALL26:
+ if (v & 3) {
+ pr_err("module %s: dangerous relocation\n",
+ mod->name);
+ return -ENOEXEC;
+ }
+ if ((v >> 28) != ((uint32_t)loc >> 28)) {
+ pr_err("module %s: relocation overflow\n",
+ mod->name);
+ return -ENOEXEC;
+ }
+ *loc = (*loc & 0x3f) | ((v >> 2) << 6);
+ break;
+ case R_NIOS2_HI16:
+ word = *loc;
+ *loc = ((((word >> 22) << 16) |
+ ((v >> 16) & 0xffff)) << 6) | (word & 0x3f);
+ break;
+ case R_NIOS2_LO16:
+ word = *loc;
+ *loc = ((((word >> 22) << 16) | (v & 0xffff)) << 6) |
+ (word & 0x3f);
+ break;
+ case R_NIOS2_HIADJ16:
+ {
+ Elf32_Addr word2;
+
+ word = *loc;
+ word2 = ((v >> 16) + ((v >> 15) & 1)) & 0xffff;
+ *loc = ((((word >> 22) << 16) | word2) << 6) |
+ (word & 0x3f);
+ }
+ break;
+
+ default:
+ pr_err("module %s: Unknown reloc: %u\n",
+ mod->name, ELF32_R_TYPE(rela[i].r_info));
+ return -ENOEXEC;
+ }
+ }
+
+ return 0;
+}
+
+int module_finalize(const Elf_Ehdr *hdr, const Elf_Shdr *sechdrs,
+ struct module *me)
+{
+ flush_cache_all();
+ return 0;
+}
+
+void module_arch_cleanup(struct module *mod)
+{
+}
--
1.8.3.2

2014-04-18 12:44:43

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 22/28] nios2: Miscellaneous header files

This patch introduces a few nios2-specific header files.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/cmpxchg.h | 61 +++++++++++++++++++++++++++++++++
arch/nios2/include/asm/gpio.h | 23 +++++++++++++
arch/nios2/include/asm/linkage.h | 25 ++++++++++++++
arch/nios2/include/asm/mutex.h | 1 +
arch/nios2/include/asm/pci.h | 25 ++++++++++++++
arch/nios2/include/uapi/asm/byteorder.h | 22 ++++++++++++
arch/nios2/include/uapi/asm/stat.h | 23 +++++++++++++
arch/nios2/include/uapi/asm/statfs.h | 24 +++++++++++++
arch/nios2/include/uapi/asm/swab.h | 37 ++++++++++++++++++++
9 files changed, 241 insertions(+)
create mode 100644 arch/nios2/include/asm/cmpxchg.h
create mode 100644 arch/nios2/include/asm/gpio.h
create mode 100644 arch/nios2/include/asm/linkage.h
create mode 100644 arch/nios2/include/asm/mutex.h
create mode 100644 arch/nios2/include/asm/pci.h
create mode 100644 arch/nios2/include/uapi/asm/byteorder.h
create mode 100644 arch/nios2/include/uapi/asm/stat.h
create mode 100644 arch/nios2/include/uapi/asm/statfs.h
create mode 100644 arch/nios2/include/uapi/asm/swab.h

diff --git a/arch/nios2/include/asm/cmpxchg.h b/arch/nios2/include/asm/cmpxchg.h
new file mode 100644
index 0000000..8593871
--- /dev/null
+++ b/arch/nios2/include/asm/cmpxchg.h
@@ -0,0 +1,61 @@
+/*
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_CMPXCHG_H
+#define _ASM_NIOS2_CMPXCHG_H
+
+#include <linux/irqflags.h>
+
+#define xchg(ptr, x) \
+ ((__typeof__(*(ptr)))__xchg((unsigned long)(x), (ptr), sizeof(*(ptr))))
+
+struct __xchg_dummy { unsigned long a[100]; };
+#define __xg(x) ((volatile struct __xchg_dummy *)(x))
+
+static inline unsigned long __xchg(unsigned long x, volatile void *ptr,
+ int size)
+{
+ unsigned long tmp, flags;
+
+ local_irq_save(flags);
+
+ switch (size) {
+ case 1:
+ __asm__ __volatile__(
+ "ldb %0, %2\n"
+ "stb %1, %2\n"
+ : "=&r" (tmp)
+ : "r" (x), "m" (*__xg(ptr))
+ : "memory");
+ break;
+ case 2:
+ __asm__ __volatile__(
+ "ldh %0, %2\n"
+ "sth %1, %2\n"
+ : "=&r" (tmp)
+ : "r" (x), "m" (*__xg(ptr))
+ : "memory");
+ break;
+ case 4:
+ __asm__ __volatile__(
+ "ldw %0, %2\n"
+ "stw %1, %2\n"
+ : "=&r" (tmp)
+ : "r" (x), "m" (*__xg(ptr))
+ : "memory");
+ break;
+ }
+
+ local_irq_restore(flags);
+ return tmp;
+}
+
+#include <asm-generic/cmpxchg.h>
+#include <asm-generic/cmpxchg-local.h>
+
+#endif /* _ASM_NIOS2_CMPXCHG_H */
diff --git a/arch/nios2/include/asm/gpio.h b/arch/nios2/include/asm/gpio.h
new file mode 100644
index 0000000..e726bfc946d4fcf1b54ce9dd46a59cf507171536
--- /dev/null
+++ b/arch/nios2/include/asm/gpio.h
@@ -0,0 +1,23 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_GPIO_H
+#define _ASM_NIOS2_GPIO_H
+
+#include <linux/gpio.h>
+
+#endif /* _ASM_NIOS2_GPIO_H */
diff --git a/arch/nios2/include/asm/linkage.h b/arch/nios2/include/asm/linkage.h
new file mode 100644
index 0000000..e0c6decd7d58db8b37e0f4e060e6394b15eac76c
--- /dev/null
+++ b/arch/nios2/include/asm/linkage.h
@@ -0,0 +1,25 @@
+/*
+ * Copyright (C) 2009 Thomas Chou <[email protected]>
+ *
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT. See the GNU General Public License for more
+ * details.
+ */
+
+#ifndef _ASM_NIOS2_LINKAGE_H
+#define _ASM_NIOS2_LINKAGE_H
+
+/* This file is required by include/linux/linkage.h */
+#define __ALIGN .align 4
+#define __ALIGN_STR ".align 4"
+
+#endif
diff --git a/arch/nios2/include/asm/mutex.h b/arch/nios2/include/asm/mutex.h
new file mode 100644
index 0000000..ff6101a
--- /dev/null
+++ b/arch/nios2/include/asm/mutex.h
@@ -0,0 +1 @@
+#include <asm-generic/mutex-dec.h>
diff --git a/arch/nios2/include/asm/pci.h b/arch/nios2/include/asm/pci.h
new file mode 100644
index 0000000..69c86faffe17dfd3ba69db90787f859d88f55ea8
--- /dev/null
+++ b/arch/nios2/include/asm/pci.h
@@ -0,0 +1,25 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+
+#ifndef __ASM_NIOS2_PCI_H__
+#define __ASM_NIOS2_PCI_H__
+
+/* We don't support PCI yet, but some drivers require this file anyway */
+
+#endif /* __ASM_NIOS2_PCI_H__ */
diff --git a/arch/nios2/include/uapi/asm/byteorder.h b/arch/nios2/include/uapi/asm/byteorder.h
new file mode 100644
index 0000000..3ab5dc20d757c3abe89775d0ba585f47d567d055
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/byteorder.h
@@ -0,0 +1,22 @@
+/*
+ * Copyright (C) 2009 Thomas Chou <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#ifndef _ASM_NIOS2_BYTEORDER_H
+#define _ASM_NIOS2_BYTEORDER_H
+
+#include <linux/byteorder/little_endian.h>
+
+#endif
diff --git a/arch/nios2/include/uapi/asm/stat.h b/arch/nios2/include/uapi/asm/stat.h
new file mode 100644
index 0000000..009f505da848ff388d55582984ff83f95ab395ad
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/stat.h
@@ -0,0 +1,23 @@
+/*
+ * Copyright Altera Corporation (C) 2014. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _UAPI_ASM_NIOS2_STAT_H
+#define _UAPI_ASM_NIOS2_STAT_H
+
+#include <asm-generic/stat.h>
+
+#endif
diff --git a/arch/nios2/include/uapi/asm/statfs.h b/arch/nios2/include/uapi/asm/statfs.h
new file mode 100644
index 0000000..625c86418dc807f196553bff2f25f8ac40474303
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/statfs.h
@@ -0,0 +1,24 @@
+/*
+ * Copyright Altera Corporation (C) 2014. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _UAPI_ASM_NIOS2_STATFS_H
+#define _UAPI_ASM_NIOS2_STATFS_H
+
+#define __statfs_word __s32
+#include <asm-generic/statfs.h>
+
+#endif
diff --git a/arch/nios2/include/uapi/asm/swab.h b/arch/nios2/include/uapi/asm/swab.h
new file mode 100644
index 0000000..b4e22ebaeb1792c2a703ebb235c5448f054685aa
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/swab.h
@@ -0,0 +1,37 @@
+/*
+ * Copyright (C) 2012 Tobias Klauser <[email protected]>
+ * Copyright (C) 2011 Pyramid Technical Consultants, Inc.
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License. See the file COPYING in the main directory of this
+ * archive for more details.
+ */
+
+#ifndef _ASM_NIOS2_SWAB_H
+#define _ASM_NIOS2_SWAB_H
+
+#include <linux/types.h>
+#include <asm-generic/swab.h>
+
+#ifdef CONFIG_NIOS2_CI_SWAB_SUPPORT
+#ifdef __GNUC__
+
+#define __nios2_swab(x) \
+ __builtin_custom_ini(CONFIG_NIOS2_CI_SWAB_NO, (x))
+
+static inline __attribute__((const)) __u16 __arch_swab16(__u16 x)
+{
+ return (__u16) __nios2_swab(((__u32) x) << 16);
+}
+#define __arch_swab16 __arch_swab16
+
+static inline __attribute__((const)) __u32 __arch_swab32(__u32 x)
+{
+ return (__u32) __nios2_swab(x);
+}
+#define __arch_swab32 __arch_swab32
+
+#endif /* __GNUC__ */
+#endif /* CONFIG_NIOS2_CI_SWAB_SUPPORT */
+
+#endif /* _ASM_NIOS2_SWAB_H */
--
1.8.3.2

2014-04-18 12:45:52

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 25/28] nios2: ptrace support

Add ptrace support for nios2.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/ptrace.h | 40 +++++++
arch/nios2/include/uapi/asm/ptrace.h | 123 ++++++++++++++++++++++
arch/nios2/kernel/ptrace.c | 198 +++++++++++++++++++++++++++++++++++
3 files changed, 361 insertions(+)
create mode 100644 arch/nios2/include/asm/ptrace.h
create mode 100644 arch/nios2/include/uapi/asm/ptrace.h
create mode 100644 arch/nios2/kernel/ptrace.c

diff --git a/arch/nios2/include/asm/ptrace.h b/arch/nios2/include/asm/ptrace.h
new file mode 100644
index 0000000..ca8588ed8c32eff5ed081c21c55fc142bb16f151
--- /dev/null
+++ b/arch/nios2/include/asm/ptrace.h
@@ -0,0 +1,40 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * based on m68k asm/processor.h
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_PTRACE_H
+#define _ASM_NIOS2_PTRACE_H
+
+#include <uapi/asm/ptrace.h>
+
+#ifndef __ASSEMBLY__
+
+/* Arbitrarily choose the same ptrace numbers as used by the Sparc code. */
+#define PTRACE_GETREGS 12
+#define PTRACE_SETREGS 13
+
+/*
+ * Supervisor mode
+ */
+
+# define user_mode(regs) (((regs)->estatus & ESTATUS_EU))
+
+
+#define instruction_pointer(regs) ((regs)->ra)
+#define profile_pc(regs) instruction_pointer(regs)
+#define user_stack_pointer(regs) ((regs)->sp)
+extern void show_regs(struct pt_regs *);
+
+#define current_pt_regs() \
+ ((struct pt_regs *)((unsigned long)current_thread_info() + THREAD_SIZE)\
+ - 1)
+#endif /* __ASSEMBLY__ */
+#endif /* _ASM_NIOS2_PTRACE_H */
diff --git a/arch/nios2/include/uapi/asm/ptrace.h b/arch/nios2/include/uapi/asm/ptrace.h
new file mode 100644
index 0000000..e37ca15350ea60e8eb7f73b5dad331886abb5e31
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/ptrace.h
@@ -0,0 +1,123 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * based on m68k asm/processor.h
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _UAPI_ASM_NIOS2_PTRACE_H
+#define _UAPI_ASM_NIOS2_PTRACE_H
+
+#ifndef __ASSEMBLY__
+
+/*
+ * Register numbers used by 'ptrace' system call interface.
+ */
+
+/* GP registers */
+#define PTR_R0 0
+#define PTR_R1 1
+#define PTR_R2 2
+#define PTR_R3 3
+#define PTR_R4 4
+#define PTR_R5 5
+#define PTR_R6 6
+#define PTR_R7 7
+#define PTR_R8 8
+#define PTR_R9 9
+#define PTR_R10 10
+#define PTR_R11 11
+#define PTR_R12 12
+#define PTR_R13 13
+#define PTR_R14 14
+#define PTR_R15 15
+#define PTR_R16 16
+#define PTR_R17 17
+#define PTR_R18 18
+#define PTR_R19 19
+#define PTR_R20 20
+#define PTR_R21 21
+#define PTR_R22 22
+#define PTR_R23 23
+#define PTR_R24 24
+#define PTR_R25 25
+#define PTR_GP 26
+#define PTR_SP 27
+#define PTR_FP 28
+#define PTR_EA 29
+#define PTR_BA 30
+#define PTR_RA 31
+/* Control registers */
+#define PTR_PC 32
+#define PTR_STATUS 33
+#define PTR_ESTATUS 34
+#define PTR_BSTATUS 35
+#define PTR_IENABLE 36
+#define PTR_IPENDING 37
+#define PTR_CPUID 38
+#define PTR_CTL6 39
+#define PTR_CTL7 40
+#define PTR_PTEADDR 41
+#define PTR_TLBACC 42
+#define PTR_TLBMISC 43
+
+/* Text/data offsets, needed by gdbserver */
+#define PT_TEXT_ADDR (44*4)
+#define PT_TEXT_END_ADDR (45*4)
+#define PT_DATA_ADDR (46*4)
+
+/* this struct defines the way the registers are stored on the
+ stack during a system call.
+
+ There is a fake_regs in setup.c that has to match pt_regs.*/
+
+struct pt_regs {
+ unsigned long r8; /* r8-r15 Caller-saved GP registers */
+ unsigned long r9;
+ unsigned long r10;
+ unsigned long r11;
+ unsigned long r12;
+ unsigned long r13;
+ unsigned long r14;
+ unsigned long r15;
+ unsigned long r1; /* Assembler temporary */
+ unsigned long r2; /* Retval LS 32bits */
+ unsigned long r3; /* Retval MS 32bits */
+ unsigned long r4; /* r4-r7 Register arguments */
+ unsigned long r5;
+ unsigned long r6;
+ unsigned long r7;
+ unsigned long orig_r2; /* Copy of r2 ?? */
+ unsigned long ra; /* Return address */
+ unsigned long fp; /* Frame pointer */
+ unsigned long sp; /* Stack pointer */
+ unsigned long gp; /* Global pointer */
+ unsigned long estatus;
+ unsigned long ea; /* Exception return address (pc) */
+ unsigned long orig_r7;
+};
+
+/*
+ * This is the extended stack used by signal handlers and the context
+ * switcher: it's pushed after the normal "struct pt_regs".
+ */
+struct switch_stack {
+ unsigned long r16; /* r16-r23 Callee-saved GP registers */
+ unsigned long r17;
+ unsigned long r18;
+ unsigned long r19;
+ unsigned long r20;
+ unsigned long r21;
+ unsigned long r22;
+ unsigned long r23;
+ unsigned long fp;
+ unsigned long gp;
+ unsigned long ra;
+};
+
+#endif /* __ASSEMBLY__ */
+#endif /* _UAPI_ASM_NIOS2_PTRACE_H */
diff --git a/arch/nios2/kernel/ptrace.c b/arch/nios2/kernel/ptrace.c
new file mode 100644
index 0000000..5db61796b2c11532dd6219a972331d9c09f5c30a
--- /dev/null
+++ b/arch/nios2/kernel/ptrace.c
@@ -0,0 +1,198 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ *
+ * based on arch/m68knommu/kernel/ptrace.c
+ *
+ * Copyright (C) 1994 by Hamish Macdonald
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License. See the file COPYING in the main directory of this
+ * archive for more details.
+ */
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/errno.h>
+#include <linux/ptrace.h>
+#include <linux/user.h>
+
+#include <linux/uaccess.h>
+
+/*
+ * does not yet catch signals sent when the child dies.
+ * in exit.c or in signal.c.
+ */
+
+/* determines which bits in the SR the user has access to. */
+/* 1 = access 0 = no access */
+#define SR_MASK 0x00000000
+
+/* Find the stack offset for a register, relative to thread.ksp. */
+#define PT_REG(reg) ((long)&((struct pt_regs *)0)->reg)
+#define SW_REG(reg) ((long)&((struct switch_stack *)0)->reg \
+ - sizeof(struct switch_stack))
+
+/* Mapping from PT_xxx to the stack offset at which the register is
+ * saved.
+ */
+static int regoff[] = {
+ -1, PT_REG(r1), PT_REG(r2), PT_REG(r3),
+ PT_REG(r4), PT_REG(r5), PT_REG(r6), PT_REG(r7),
+ PT_REG(r8), PT_REG(r9), PT_REG(r10), PT_REG(r11),
+ PT_REG(r12), PT_REG(r13), PT_REG(r14), PT_REG(r15), /* reg 15 */
+ SW_REG(r16), SW_REG(r17), SW_REG(r18), SW_REG(r19),
+ SW_REG(r20), SW_REG(r21), SW_REG(r22), SW_REG(r23),
+ -1, -1, PT_REG(gp), PT_REG(sp),
+ PT_REG(fp), PT_REG(ea), -1, PT_REG(ra), /* reg 31 */
+ PT_REG(ea), -1, -1, -1, /* use ea for pc */
+ -1, -1, -1, -1,
+ -1, -1, -1, -1 /* reg 43 */
+};
+
+/*
+ * Get contents of register REGNO in task TASK.
+ */
+static inline long get_reg(struct task_struct *task, int regno)
+{
+ unsigned long *addr;
+
+ if (regno >= ARRAY_SIZE(regoff) || regoff[regno] == -1)
+ return 0;
+
+ addr = (unsigned long *)((char *)task->thread.kregs + regoff[regno]);
+ return *addr;
+}
+
+/*
+ * Write contents of register REGNO in task TASK.
+ */
+static inline int put_reg(struct task_struct *task, int regno,
+ unsigned long data)
+{
+ unsigned long *addr;
+
+ if (regno >= ARRAY_SIZE(regoff) || regoff[regno] == -1)
+ return -1;
+
+ addr = (unsigned long *)((char *)task->thread.kregs + regoff[regno]);
+ *addr = data;
+ return 0;
+}
+
+/*
+ * Called by kernel/ptrace.c when detaching..
+ *
+ * Nothing special to do here, no processor debug support.
+ */
+void ptrace_disable(struct task_struct *child)
+{
+}
+
+long arch_ptrace(struct task_struct *child, long request,
+ unsigned long addr, unsigned long data)
+{
+ unsigned long tmp;
+ unsigned int i;
+ int ret;
+
+ switch (request) {
+ /* read the word at location addr in the USER area. */
+ case PTRACE_PEEKUSR:
+ pr_debug("PEEKUSR: addr=0x%08lx\n", addr);
+ ret = -EIO;
+ if (addr & 3)
+ break;
+
+ addr = addr >> 2; /* temporary hack. */
+ ret = -EIO;
+ if (addr < ARRAY_SIZE(regoff))
+ tmp = get_reg(child, addr);
+ else if (addr == PT_TEXT_ADDR / 4)
+ tmp = child->mm->start_code;
+ else if (addr == PT_DATA_ADDR / 4)
+ tmp = child->mm->start_data;
+ else if (addr == PT_TEXT_END_ADDR / 4)
+ tmp = child->mm->end_code;
+ else
+ break;
+ ret = put_user(tmp, (unsigned long *) data);
+ pr_debug("PEEKUSR: rdword=0x%08lx\n", tmp);
+ break;
+ /* write the word at location addr in the USER area */
+ case PTRACE_POKEUSR:
+ pr_debug("POKEUSR: addr=0x%08lx, data=0x%08lx\n", addr, data);
+ ret = -EIO;
+ if (addr & 3)
+ break;
+
+ addr = addr >> 2; /* temporary hack. */
+
+ if (addr == PTR_ESTATUS) {
+ data &= SR_MASK;
+ data |= get_reg(child, PTR_ESTATUS) & ~(SR_MASK);
+ }
+ if (addr < ARRAY_SIZE(regoff)) {
+ if (put_reg(child, addr, data))
+ break;
+ ret = 0;
+ break;
+ }
+ break;
+ /* Get all gp regs from the child. */
+ case PTRACE_GETREGS:
+ pr_debug("GETREGS\n");
+ for (i = 0; i < ARRAY_SIZE(regoff); i++) {
+ tmp = get_reg(child, i);
+ if (put_user(tmp, (unsigned long *) data)) {
+ ret = -EFAULT;
+ break;
+ }
+ data += sizeof(long);
+ }
+ ret = 0;
+ break;
+ /* Set all gp regs in the child. */
+ case PTRACE_SETREGS:
+ pr_debug("SETREGS\n");
+ for (i = 0; i < ARRAY_SIZE(regoff); i++) {
+ if (get_user(tmp, (unsigned long *) data)) {
+ ret = -EFAULT;
+ break;
+ }
+ if (i == PTR_ESTATUS) {
+ tmp &= SR_MASK;
+ tmp |= get_reg(child, PTR_ESTATUS) & ~(SR_MASK);
+ }
+ put_reg(child, i, tmp);
+ data += sizeof(long);
+ }
+ ret = 0;
+ break;
+ default:
+ ret = ptrace_request(child, request, addr, data);
+ }
+
+ return ret;
+}
+
+asmlinkage void syscall_trace(void)
+{
+ if (!test_thread_flag(TIF_SYSCALL_TRACE))
+ return;
+ if (!(current->ptrace & PT_PTRACED))
+ return;
+ current->exit_code = SIGTRAP;
+ current->state = TASK_STOPPED;
+ ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD)
+ ? 0x80 : 0));
+ /*
+ * this isn't the same as continuing with a signal, but it will do
+ * for normal use. strace only continues with a signal if the
+ * stopping signal is not SIGTRAP. -brl
+ */
+ if (current->exit_code) {
+ send_sig(current->exit_code, current, 1);
+ current->exit_code = 0;
+ }
+}
--
1.8.3.2

2014-04-18 12:51:14

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 27/28] MAINTAINERS: Add nios2 maintainer

Signed-off-by: Ley Foon Tan <[email protected]>
---
MAINTAINERS | 7 +++++++
1 file changed, 7 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index 80399ff..389f3d5315b69660060f34617df402e27323c543 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -6247,6 +6247,13 @@ S: Maintained
F: Documentation/scsi/NinjaSCSI.txt
F: drivers/scsi/nsp32*

+NIOS2 ARCHITECTURE
+M: Ley Foon Tan <[email protected]>
+L: [email protected] (moderated for non-subscribers)
+T: git://git.rocketboards.org/linux-socfpga.git
+S: Maintained
+F: arch/nios2/
+
NTB DRIVER
M: Jon Mason <[email protected]>
S: Supported
--
1.8.3.2

2014-04-18 12:51:41

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 26/28] Add ELF machine define for Nios2

Signed-off-by: Ley Foon Tan <[email protected]>
---
include/uapi/linux/elf-em.h | 1 +
1 file changed, 1 insertion(+)

diff --git a/include/uapi/linux/elf-em.h b/include/uapi/linux/elf-em.h
index 01529bd..7f06c0d 100644
--- a/include/uapi/linux/elf-em.h
+++ b/include/uapi/linux/elf-em.h
@@ -33,6 +33,7 @@
#define EM_M32R 88 /* Renesas M32R */
#define EM_MN10300 89 /* Panasonic/MEI MN10300, AM33 */
#define EM_BLACKFIN 106 /* ADI Blackfin Processor */
+#define EM_ALTERA_NIOS2 113 /* Altera Nios II soft-core processor */
#define EM_TI_C6000 140 /* TI C6X DSPs */
#define EM_AARCH64 183 /* ARM 64 bit */
#define EM_FRV 0x5441 /* Fujitsu FR-V */
--
1.8.3.2

2014-04-18 12:42:59

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 01/28] nios2: Build infrastructure

This patch adds Makefile and Kconfig files required for building a
nios2 kernel.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/Kconfig | 213 +++++++++++++++++++++++++++++++++++
arch/nios2/Kconfig.debug | 17 +++
arch/nios2/Makefile | 78 +++++++++++++
arch/nios2/boot/Makefile | 51 +++++++++
arch/nios2/configs/3c120_defconfig | 77 +++++++++++++
arch/nios2/include/asm/Kbuild | 66 +++++++++++
arch/nios2/include/uapi/asm/Kbuild | 14 +++
arch/nios2/kernel/Makefile | 14 +++
arch/nios2/kernel/nios2_ksyms.c | 35 ++++++
arch/nios2/kernel/vmlinux.lds.S | 75 ++++++++++++
arch/nios2/lib/Makefile | 1 +
arch/nios2/mm/Makefile | 12 ++
arch/nios2/platform/Kconfig.platform | 129 +++++++++++++++++++++
arch/nios2/platform/Makefile | 1 +
14 files changed, 783 insertions(+)
create mode 100644 arch/nios2/Kconfig
create mode 100644 arch/nios2/Kconfig.debug
create mode 100644 arch/nios2/Makefile
create mode 100644 arch/nios2/boot/Makefile
create mode 100644 arch/nios2/configs/3c120_defconfig
create mode 100644 arch/nios2/include/asm/Kbuild
create mode 100644 arch/nios2/include/uapi/asm/Kbuild
create mode 100644 arch/nios2/kernel/Makefile
create mode 100644 arch/nios2/kernel/nios2_ksyms.c
create mode 100644 arch/nios2/kernel/vmlinux.lds.S
create mode 100644 arch/nios2/lib/Makefile
create mode 100644 arch/nios2/mm/Makefile
create mode 100644 arch/nios2/platform/Kconfig.platform
create mode 100644 arch/nios2/platform/Makefile

diff --git a/arch/nios2/Kconfig b/arch/nios2/Kconfig
new file mode 100644
index 0000000..2bd0d0d7160f31a5c5890cf8a059b9f53378be6e
--- /dev/null
+++ b/arch/nios2/Kconfig
@@ -0,0 +1,213 @@
+config NIOS2
+ def_bool y
+ select OF
+ select OF_EARLY_FLATTREE
+ select USB_ARCH_HAS_HCD if USB_SUPPORT
+ select ARCH_WANT_OPTIONAL_GPIOLIB
+ select GENERIC_IRQ_PROBE
+ select GENERIC_IRQ_SHOW
+ select GENERIC_CPU_DEVICES
+ select GENERIC_ATOMIC64
+ select MODULES_USE_ELF_RELA
+ select IRQ_DOMAIN
+ select SOC_BUS
+ select CLKSRC_OF
+
+config GENERIC_CSUM
+ def_bool y
+
+config GENERIC_FIND_NEXT_BIT
+ def_bool y
+
+config GENERIC_HWEIGHT
+ def_bool y
+
+config GENERIC_CALIBRATE_DELAY
+ def_bool y
+
+
+config NO_IOPORT
+ def_bool y
+
+config HAS_DMA
+ def_bool y
+
+config ZONE_DMA
+ def_bool y
+
+config FPU
+ def_bool n
+
+config SWAP
+ def_bool n
+
+config RWSEM_GENERIC_SPINLOCK
+ def_bool y
+
+config TRACE_IRQFLAGS_SUPPORT
+ def_bool n
+
+source "init/Kconfig"
+
+menu "Kernel features"
+
+source "kernel/Kconfig.preempt"
+
+source "kernel/Kconfig.freezer"
+
+source "kernel/Kconfig.hz"
+
+source "mm/Kconfig"
+
+config FORCE_MAX_ZONEORDER
+ int "Maximum zone order"
+ range 9 20
+ default "11"
+ help
+ The kernel memory allocator divides physically contiguous memory
+ blocks into "zones", where each zone is a power of two number of
+ pages. This option selects the largest power of two that the kernel
+ keeps in the memory allocator. If you need to allocate very large
+ blocks of physically contiguous memory, then you may need to
+ increase this value.
+
+ This config option is actually maximum order plus one. For example,
+ a value of 11 means that the largest free memory block is 2^10 pages.
+
+endmenu
+
+source "arch/nios2/platform/Kconfig.platform"
+
+menu "Processor type and features"
+
+config MMU
+ def_bool y
+
+config ALIGNMENT_TRAP
+ bool "Catch alignment trap"
+ default y
+ help
+ Nios II CPUs cannot fetch/store data which is not bus aligned,
+ i.e., a 2 or 4 byte fetch must start at an address divisible by
+ 2 or 4. Any non-aligned load/store instructions will be trapped and
+ emulated in software if you say Y here, which has a performance
+ impact.
+
+comment "Boot options"
+
+config CMDLINE_BOOL
+ bool "Default bootloader kernel arguments"
+ default y
+
+config CMDLINE
+ string "Default kernel command string"
+ default ""
+ depends on CMDLINE_BOOL
+ help
+ On some platforms, there is currently no way for the boot loader to
+ pass arguments to the kernel. For these platforms, you can supply
+ some command-line options at build time by entering them here. In
+ other cases you can specify kernel args so that you don't have
+ to set them up in board prom initialization routines.
+
+config CMDLINE_FORCE
+ bool "Force default kernel command string"
+ depends on CMDLINE_BOOL
+ help
+ Set this to have arguments from the default kernel command string
+ override those passed by the boot loader.
+
+config CMDLINE_IGNORE_DTB
+ bool "Ignore kernel command string from DTB"
+ depends on !CMDLINE_FORCE
+ default y
+ help
+ Set this to ignore the bootargs property from the devicetree's
+ chosen node and fall back to CMDLINE if nothing is passed.
+
+config PASS_CMDLINE
+ bool "Passed kernel command line from u-boot"
+ default n
+ help
+ Use bootargs env variable from u-boot for kernel command line.
+ will override "Default kernel command string".
+ Say N if you are unsure.
+
+config BOOT_LINK_OFFSET
+ hex "Link address offset for booting"
+ default "0x00500000"
+ help
+ This option allows you to set the link address offset of the zImage.
+ This can be useful if you are on a board which has a small amount of
+ memory.
+
+endmenu
+
+menu "Advanced setup"
+
+config ADVANCED_OPTIONS
+ bool "Prompt for advanced kernel configuration options"
+ help
+
+comment "Default settings for advanced configuration options are used"
+ depends on !ADVANCED_OPTIONS
+
+config KERNEL_MMU_REGION_BASE_BOOL
+ bool "Set custom kernel MMU region base address"
+ depends on ADVANCED_OPTIONS
+ help
+ This option allows you to set the virtual address of the kernel MMU region.
+
+ Say N here unless you know what you are doing.
+
+config KERNEL_MMU_REGION_BASE
+ hex "Virtual base address of the kernel MMU region " if KERNEL_MMU_REGION_BASE_BOOL
+ default "0x80000000"
+ help
+ This option allows you to set the virtual base address of the kernel MMU region.
+
+config KERNEL_REGION_BASE_BOOL
+ bool "Set custom kernel region base address"
+ depends on ADVANCED_OPTIONS
+ help
+ This option allows you to set the virtual address of the kernel region.
+
+ Say N here unless you know what you are doing.
+
+config KERNEL_REGION_BASE
+ hex "Virtual base address of the kernel region " if KERNEL_REGION_BASE_BOOL
+ default "0xc0000000"
+
+config IO_REGION_BASE_BOOL
+ bool "Set custom I/O region base address"
+ depends on ADVANCED_OPTIONS
+ help
+ This option allows you to set the virtual address of the I/O region.
+
+ Say N here unless you know what you are doing.
+
+config IO_REGION_BASE
+ hex "Virtual base address of the I/O region " if IO_REGION_BASE_BOOL
+ default "0xe0000000"
+
+endmenu
+
+menu "Executable file formats"
+
+source "fs/Kconfig.binfmt"
+
+endmenu
+
+source "net/Kconfig"
+
+source "drivers/Kconfig"
+
+source "fs/Kconfig"
+
+source "arch/nios2/Kconfig.debug"
+
+source "security/Kconfig"
+
+source "crypto/Kconfig"
+
+source "lib/Kconfig"
diff --git a/arch/nios2/Kconfig.debug b/arch/nios2/Kconfig.debug
new file mode 100644
index 0000000..8d4e6bacd997b3f687d1262631830a3c73c1f0f6
--- /dev/null
+++ b/arch/nios2/Kconfig.debug
@@ -0,0 +1,17 @@
+menu "Kernel hacking"
+
+config TRACE_IRQFLAGS_SUPPORT
+ def_bool y
+
+source "lib/Kconfig.debug"
+
+config DEBUG_STACK_USAGE
+ bool "Enable stack utilization instrumentation"
+ depends on DEBUG_KERNEL
+ help
+ Enables the display of the minimum amount of free stack which each
+ task has ever had available in the sysrq-T and sysrq-P debug output.
+
+ This option will slow down process creation somewhat.
+
+endmenu
diff --git a/arch/nios2/Makefile b/arch/nios2/Makefile
new file mode 100644
index 0000000..89840a5763797a26bcbb723d32db64826024689c
--- /dev/null
+++ b/arch/nios2/Makefile
@@ -0,0 +1,78 @@
+#
+# This file is subject to the terms and conditions of the GNU General Public
+# License. See the file "COPYING" in the main directory of this archive
+# for more details.
+#
+# Copyright (C) 2013 Altera Corporation
+# Copyright (C) 1994, 95, 96, 2003 by Wind River Systems
+# Written by Fredrik Markstrom
+#
+# This file is included by the global makefile so that you can add your own
+# architecture-specific flags and dependencies. Remember to do have actions
+# for "archclean" cleaning up for this architecture.
+#
+# Nios2 port by Wind River Systems Inc trough:
+# [email protected] and [email protected]
+
+UTS_SYSNAME = Linux
+
+export MMU
+
+cflags-y :=
+LDFLAGS :=
+LDFLAGS_vmlinux :=
+
+LIBGCC := $(shell $(CC) $(KBUILD_CFLAGS) -print-libgcc-file-name)
+
+KBUILD_AFLAGS += $(cflags-y)
+KBUILD_CFLAGS += -pipe -D__linux__ -D__ELF__ $(cflags-y)
+KBUILD_CFLAGS += $(if $(CONFIG_NIOS2_HW_MUL_SUPPORT),-mhw-mul,-mno-hw-mul)
+KBUILD_CFLAGS += $(if $(CONFIG_NIOS2_HW_MULX_SUPPORT),-mhw-mulx,-mno-hw-mulx)
+KBUILD_CFLAGS += $(if $(CONFIG_NIOS2_HW_DIV_SUPPORT),-mhw-div,-mno-hw-div)
+KBUILD_CFLAGS += $(if $(CONFIG_NIOS2_FPU_SUPPORT),-mcustom-fpu-cfg=60-1,)
+
+KBUILD_CFLAGS += -fno-optimize-sibling-calls
+KBUILD_CFLAGS += -DUTS_SYSNAME=\"$(UTS_SYSNAME)\"
+KBUILD_CFLAGS += -fno-builtin
+KBUILD_CFLAGS += -G 0
+
+head-y := arch/nios2/kernel/head.o
+libs-y += arch/nios2/lib/ $(LIBGCC)
+core-y += arch/nios2/kernel/ arch/nios2/mm/
+core-y += arch/nios2/platform/
+
+INSTALL_PATH ?= /tftpboot
+boot := arch/$(ARCH)/boot
+BOOT_TARGETS = vmImage zImage
+PHONY += $(BOOT_TARGETS) install
+KBUILD_IMAGE := $(boot)/vmImage
+
+ifneq ($(CONFIG_DTB_SOURCE),"")
+ core-y += $(boot)/
+endif
+
+all: vmImage
+
+archclean:
+ $(Q)$(MAKE) $(clean)=$(boot)
+
+%.dtb:
+ $(Q)$(MAKE) $(build)=$(boot) $(boot)/$@
+
+dtbs:
+ $(Q)$(MAKE) $(build)=$(boot) $(boot)/$@
+
+$(BOOT_TARGETS): vmlinux
+ $(Q)$(MAKE) $(build)=$(boot) $(boot)/$@
+
+install:
+ $(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(KBUILD_IMAGE) install
+
+define archhelp
+ echo '* vmImage - Kernel-only image for U-Boot (arch/$(ARCH)/boot/vmImage)'
+ echo ' install - Install kernel using'
+ echo ' (your) ~/bin/$(CROSS_COMPILE)installkernel or'
+ echo ' (distribution) PATH: $(CROSS_COMPILE)installkernel or'
+ echo ' install to $$(INSTALL_PATH)'
+ echo ' dtbs - Build device tree blobs for enabled boards'
+endef
diff --git a/arch/nios2/boot/Makefile b/arch/nios2/boot/Makefile
new file mode 100644
index 0000000..00aef5e266f9cc629f126445535f9aae6ca8962f
--- /dev/null
+++ b/arch/nios2/boot/Makefile
@@ -0,0 +1,51 @@
+#
+# arch/nios2/boot/Makefile
+#
+# This file is subject to the terms and conditions of the GNU General Public
+# License. See the file "COPYING" in the main directory of this archive
+# for more details.
+#
+
+UIMAGE_LOADADDR = $(shell $(NM) vmlinux | awk '$$NF == "_stext" {print $$1}')
+UIMAGE_ENTRYADDR = $(shell $(NM) vmlinux | awk '$$NF == "_start" {print $$1}')
+UIMAGE_COMPRESSION = gzip
+
+OBJCOPYFLAGS_vmlinux.bin := -O binary
+$(obj)/vmlinux.bin: vmlinux FORCE
+ $(call if_changed,objcopy)
+
+$(obj)/vmlinux.gz: $(obj)/vmlinux.bin FORCE
+ $(call if_changed,gzip)
+
+$(obj)/vmImage: $(obj)/vmlinux.gz
+ $(call if_changed,uimage)
+ @$(kecho) 'Kernel: $@ is ready'
+
+# Rule to build device tree blobs
+DTB_SRC := $(subst ",,$(CONFIG_DTB_SOURCE))
+
+# Make sure the generated dtb gets removed during clean
+extra-$(CONFIG_DTB_SOURCE_BOOL) += system.dtb
+
+$(obj)/system.dtb: $(DTB_SRC) FORCE
+ $(call cmd,dtc)
+
+# Ensure system.dtb exists
+$(obj)/linked_dtb.o: $(obj)/system.dtb
+
+#ifneq ($(CONFIG_DTB_SOURCE),"")
+obj-$(CONFIG_DTB_SOURCE_BOOL) += linked_dtb.o
+#endif
+
+targets += $(dtb-y)
+
+# Rule to build device tree blobs with make command
+$(obj)/%.dtb: $(src)/dts/%.dts FORCE
+ $(call if_changed_dep,dtc)
+
+$(obj)/dtbs: $(addprefix $(obj)/, $(dtb-y))
+
+clean-files := *.dtb
+
+install:
+ sh $(srctree)/$(src)/install.sh $(KERNELRELEASE) $(BOOTIMAGE) System.map "$(INSTALL_PATH)"
diff --git a/arch/nios2/configs/3c120_defconfig b/arch/nios2/configs/3c120_defconfig
new file mode 100644
index 0000000..0147082994473779eeb138e2cfd0a17be73242c4
--- /dev/null
+++ b/arch/nios2/configs/3c120_defconfig
@@ -0,0 +1,77 @@
+CONFIG_SYSVIPC=y
+CONFIG_BSD_PROCESS_ACCT=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_SYSCTL_SYSCALL=y
+# CONFIG_ELF_CORE is not set
+# CONFIG_FUTEX is not set
+# CONFIG_EPOLL is not set
+# CONFIG_SIGNALFD is not set
+# CONFIG_TIMERFD is not set
+# CONFIG_EVENTFD is not set
+# CONFIG_SHMEM is not set
+# CONFIG_AIO is not set
+CONFIG_EMBEDDED=y
+CONFIG_SLAB=y
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_MEM_BASE=0x10000000
+CONFIG_NIOS2_HW_MUL_SUPPORT=y
+CONFIG_NIOS2_HW_DIV_SUPPORT=y
+CONFIG_CUSTOM_CACHE_SETTINGS=y
+CONFIG_NIOS2_DCACHE_SIZE=0x8000
+CONFIG_NIOS2_ICACHE_SIZE=0x8000
+# CONFIG_CMDLINE_IGNORE_DTB is not set
+CONFIG_PASS_CMDLINE=y
+CONFIG_BOOT_LINK_OFFSET=0x00800000
+CONFIG_NET=y
+CONFIG_PACKET=y
+CONFIG_UNIX=y
+CONFIG_INET=y
+CONFIG_IP_MULTICAST=y
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+CONFIG_IP_PNP_BOOTP=y
+CONFIG_IP_PNP_RARP=y
+# CONFIG_INET_XFRM_MODE_TRANSPORT is not set
+# CONFIG_INET_XFRM_MODE_TUNNEL is not set
+# CONFIG_INET_XFRM_MODE_BEET is not set
+# CONFIG_INET_LRO is not set
+# CONFIG_IPV6 is not set
+# CONFIG_WIRELESS is not set
+CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug"
+CONFIG_DEVTMPFS=y
+CONFIG_DEVTMPFS_MOUNT=y
+# CONFIG_FW_LOADER is not set
+CONFIG_MTD=y
+CONFIG_MTD_CMDLINE_PARTS=y
+CONFIG_MTD_BLOCK=y
+CONFIG_MTD_CFI=y
+CONFIG_MTD_CFI_INTELEXT=y
+CONFIG_MTD_CFI_AMDSTD=y
+CONFIG_MTD_PHYSMAP_OF=y
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_NETDEVICES=y
+CONFIG_ALTERA_TSE=y
+CONFIG_MARVELL_PHY=y
+# CONFIG_WLAN is not set
+# CONFIG_INPUT_MOUSE is not set
+# CONFIG_SERIO_SERPORT is not set
+# CONFIG_VT is not set
+CONFIG_SERIAL_ALTERA_JTAGUART=y
+CONFIG_SERIAL_ALTERA_JTAGUART_CONSOLE=y
+CONFIG_SERIAL_ALTERA_UART=y
+# CONFIG_HW_RANDOM is not set
+# CONFIG_HWMON is not set
+# CONFIG_USB_SUPPORT is not set
+CONFIG_NEW_LEDS=y
+CONFIG_LEDS_CLASS=y
+CONFIG_LEDS_TRIGGERS=y
+CONFIG_LEDS_TRIGGER_HEARTBEAT=y
+# CONFIG_DNOTIFY is not set
+# CONFIG_INOTIFY_USER is not set
+CONFIG_JFFS2_FS=y
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3_ACL=y
+CONFIG_SUNRPC_DEBUG=y
+CONFIG_DEBUG_INFO=y
+# CONFIG_ENABLE_WARN_DEPRECATED is not set
diff --git a/arch/nios2/include/asm/Kbuild b/arch/nios2/include/asm/Kbuild
new file mode 100644
index 0000000..848a67f
--- /dev/null
+++ b/arch/nios2/include/asm/Kbuild
@@ -0,0 +1,66 @@
+include include/asm-generic/Kbuild.asm
+
+header-y += ucontext.h
+header-y += traps.h
+
+generic-y += atomic.h
+generic-y += auxvec.h
+generic-y += barrier.h
+generic-y += bitops.h
+generic-y += bitsperlong.h
+generic-y += bug.h
+generic-y += bugs.h
+generic-y += cputime.h
+generic-y += current.h
+generic-y += device.h
+generic-y += div64.h
+generic-y += dma.h
+generic-y += emergency-restart.h
+generic-y += errno.h
+generic-y += exec.h
+generic-y += fb.h
+generic-y += fcntl.h
+generic-y += ftrace.h
+generic-y += hardirq.h
+generic-y += hash.h
+generic-y += hw_irq.h
+generic-y += ioctl.h
+generic-y += ioctls.h
+generic-y += ipcbuf.h
+generic-y += irq_regs.h
+generic-y += kdebug.h
+generic-y += kmap_types.h
+generic-y += kvm_para.h
+generic-y += local.h
+generic-y += mcs_spinlock.h
+generic-y += mman.h
+generic-y += module.h
+generic-y += msgbuf.h
+generic-y += param.h
+generic-y += percpu.h
+generic-y += poll.h
+generic-y += posix_types.h
+generic-y += preempt.h
+generic-y += resource.h
+generic-y += scatterlist.h
+generic-y += sections.h
+generic-y += segment.h
+generic-y += sembuf.h
+generic-y += serial.h
+generic-y += shmbuf.h
+generic-y += shmparam.h
+generic-y += siginfo.h
+generic-y += signal.h
+generic-y += socket.h
+generic-y += sockios.h
+generic-y += spinlock.h
+generic-y += stat.h
+generic-y += statfs.h
+generic-y += termbits.h
+generic-y += termios.h
+generic-y += topology.h
+generic-y += trace_clock.h
+generic-y += types.h
+generic-y += unaligned.h
+generic-y += user.h
+generic-y += xor.h
diff --git a/arch/nios2/include/uapi/asm/Kbuild b/arch/nios2/include/uapi/asm/Kbuild
new file mode 100644
index 0000000..6d9f2e8e6798d254dee8dd32a8a963ddd70bbe07
--- /dev/null
+++ b/arch/nios2/include/uapi/asm/Kbuild
@@ -0,0 +1,14 @@
+include include/uapi/asm-generic/Kbuild.asm
+
+header-y += byteorder.h
+header-y += elf.h
+header-y += kvm_para.h
+header-y += processor.h
+header-y += ptrace.h
+header-y += setup.h
+header-y += sigcontext.h
+header-y += signal.h
+header-y += stat.h
+header-y += statfs.h
+header-y += swab.h
+header-y += unistd.h
diff --git a/arch/nios2/kernel/Makefile b/arch/nios2/kernel/Makefile
new file mode 100644
index 0000000..c46ddee
--- /dev/null
+++ b/arch/nios2/kernel/Makefile
@@ -0,0 +1,14 @@
+#
+# Makefile for the nios2 linux kernel.
+#
+
+extra-y := head.o vmlinux.lds
+
+obj-y += entry.o
+
+obj-y += cpuinfo.o insnemu.o irq.o nios2_ksyms.o process.o prom.o ptrace.o \
+ setup.o signal.o sys_nios2.o syscall_table.o time.o traps.o
+
+obj-$(CONFIG_MODULES) += module.o
+obj-$(CONFIG_CONSOLE) += console.o
+obj-$(CONFIG_ALIGNMENT_TRAP) += misaligned.o
diff --git a/arch/nios2/kernel/nios2_ksyms.c b/arch/nios2/kernel/nios2_ksyms.c
new file mode 100644
index 0000000..eaf7f0cee5facf67e9a7f191ecbb8bc34e387b23
--- /dev/null
+++ b/arch/nios2/kernel/nios2_ksyms.c
@@ -0,0 +1,35 @@
+/*
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License. See the file COPYING in the main directory of this
+ * archive for more details.
+ */
+
+#include <linux/export.h>
+#include <linux/string.h>
+
+/* string functions */
+
+EXPORT_SYMBOL(memcpy);
+EXPORT_SYMBOL(memset);
+EXPORT_SYMBOL(memmove);
+
+/*
+ * libgcc functions - functions that are used internally by the
+ * compiler... (prototypes are not correct though, but that
+ * doesn't really matter since they're not versioned).
+ */
+#define DECLARE_EXPORT(name) extern void name(void); EXPORT_SYMBOL(name)
+
+DECLARE_EXPORT(__gcc_bcmp);
+DECLARE_EXPORT(__divdi3);
+DECLARE_EXPORT(__divsi3);
+DECLARE_EXPORT(__moddi3);
+DECLARE_EXPORT(__modsi3);
+DECLARE_EXPORT(__udivdi3);
+DECLARE_EXPORT(__udivmoddi4);
+DECLARE_EXPORT(__udivsi3);
+DECLARE_EXPORT(__umoddi3);
+DECLARE_EXPORT(__umodsi3);
+DECLARE_EXPORT(__muldi3);
diff --git a/arch/nios2/kernel/vmlinux.lds.S b/arch/nios2/kernel/vmlinux.lds.S
new file mode 100644
index 0000000..ec43c9f72b18a6ce79559d0da7ab8483a8aaba45
--- /dev/null
+++ b/arch/nios2/kernel/vmlinux.lds.S
@@ -0,0 +1,75 @@
+/*
+ * Copyright (C) 2009 Thomas Chou <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+#include <asm/page.h>
+#include <asm-generic/vmlinux.lds.h>
+#include <asm/cache.h>
+#include <asm/thread_info.h>
+
+OUTPUT_FORMAT("elf32-littlenios2", "elf32-littlenios2", "elf32-littlenios2")
+
+OUTPUT_ARCH(nios)
+ENTRY(_start) /* Defined in head.S */
+
+jiffies = jiffies_64;
+
+SECTIONS
+{
+ . = CONFIG_MEM_BASE | CONFIG_KERNEL_REGION_BASE;
+
+ _text = .;
+ _stext = .;
+ HEAD_TEXT_SECTION
+ .text : {
+ TEXT_TEXT
+ SCHED_TEXT
+ LOCK_TEXT
+ IRQENTRY_TEXT
+ KPROBES_TEXT
+ } =0
+ _etext = .;
+
+ .got : {
+ *(.got.plt)
+ *(.igot.plt)
+ *(.got)
+ *(.igot)
+ }
+
+ EXCEPTION_TABLE(L1_CACHE_BYTES)
+
+ . = ALIGN(PAGE_SIZE);
+ __init_begin = .;
+ INIT_TEXT_SECTION(PAGE_SIZE)
+ INIT_DATA_SECTION(PAGE_SIZE)
+ PERCPU_SECTION(L1_CACHE_BYTES)
+ __init_end = .;
+
+ _sdata = .;
+ RO_DATA_SECTION(PAGE_SIZE)
+ RW_DATA_SECTION(L1_CACHE_BYTES, PAGE_SIZE, THREAD_SIZE)
+ _edata = .;
+
+ BSS_SECTION(0, 0, 0)
+ _end = .;
+
+ STABS_DEBUG
+ DWARF_DEBUG
+ NOTES
+
+ DISCARDS
+}
diff --git a/arch/nios2/lib/Makefile b/arch/nios2/lib/Makefile
new file mode 100644
index 0000000..f9c5bc02ba357adb37d812f0bd0a30fc76d909b8
--- /dev/null
+++ b/arch/nios2/lib/Makefile
@@ -0,0 +1 @@
+lib-y = memset.o memcpy.o memmove.o
\ No newline at end of file
diff --git a/arch/nios2/mm/Makefile b/arch/nios2/mm/Makefile
new file mode 100644
index 0000000..6ef15aae70ef0937bd53574997fcae82116f9245
--- /dev/null
+++ b/arch/nios2/mm/Makefile
@@ -0,0 +1,12 @@
+#
+# Makefile for the Nios2-specific parts of the memory manager.
+#
+
+obj-y := init.o
+
+obj-y += cacheflush.o
+
+obj-y += pgtable.o tlb.o uaccess.o fault.o
+obj-y += ioremap.o extable.o mmu_context.o
+
+obj-y += dma-mapping.o
\ No newline at end of file
diff --git a/arch/nios2/platform/Kconfig.platform b/arch/nios2/platform/Kconfig.platform
new file mode 100644
index 0000000..e767c51afd9f4b588a256ff2c75ac273fe920fa7
--- /dev/null
+++ b/arch/nios2/platform/Kconfig.platform
@@ -0,0 +1,129 @@
+menu "Platform options"
+
+comment "Memory settings"
+
+config MEM_BASE
+ hex "Memory base address"
+ default "0x00000000"
+ help
+ This is the physical address of the memory that the kernel will run
+ from. This address is used to link the kernel and setup initial memory
+ management. You should take the raw memory address without any MMU
+ or cache bits set.
+ Please not that this address is used directly so you have to manually
+ do address translation if it's connected to a bridge.
+
+comment "Device tree"
+
+config NIOS2_DTB_AT_PHYS_ADDR
+ bool "DTB at physical address"
+ default n
+ help
+ When enabled you can select a physical address to load the dtb from.
+ Normally this address is passed by a bootloader such as u-boot but
+ using this you can use a devicetree without a bootloader.
+ This way you can store a devicetree in NOR flash or an onchip rom.
+ Please note that this addres is used directly so you have to manually
+ do address translation if it's connected to a bridge. Also take into
+ account that when using an MMU you'd have to ad 0xC0000000 to your
+ address
+
+config NIOS2_DTB_PHYS_ADDR
+ hex "DTB Address"
+ depends on NIOS2_DTB_AT_PHYS_ADDR
+ default "0xC0000000"
+ help
+ Physical address of a dtb blob.
+
+config DTB_SOURCE_BOOL
+ bool "Compile and link device tree into kernel image"
+ default n
+ help
+ This allows you to specify a dts (device tree source) file
+ which will be compiled and linked into the kernel image.
+
+config DTB_SOURCE
+ string "Device tree source file"
+ depends on DTB_SOURCE_BOOL
+ default ""
+ help
+ Absolute path to the device tree source (dts) file describing your
+ system.
+
+comment "Nios II instructions"
+
+config NIOS2_HW_MUL_SUPPORT
+ bool "Enable MUL instruction"
+ default n
+ help
+ Set to true if you configured the Nios II to include the MUL
+ instruction. This will enable the -mhw-mul compiler flag.
+
+config NIOS2_HW_MULX_SUPPORT
+ bool "Enable MULX instruction"
+ default n
+ help
+ Set to true if you configured the Nios II to include the MULX
+ instruction. Enables the -mhw-mulx compiler flag.
+
+config NIOS2_HW_DIV_SUPPORT
+ bool "Enable DIV instruction"
+ default n
+ help
+ Set to true if you configured the Nios II to include the DIV
+ instruction. Enables the -mhw-div compiler flag.
+
+config NIOS2_FPU_SUPPORT
+ bool "Custom floating point instr support"
+ default n
+ help
+ Enables the -mcustom-fpu-cfg=60-1 compiler flag.
+
+config NIOS2_CI_SWAB_SUPPORT
+ bool "Byteswap custom instruction"
+ default n
+ help
+ Use the byteswap (endian convertor) Nios II custom instruction provided
+ by Altera and which can be enabled in SOPC builder. This accelerates
+ endian conversions in the kernel (e.g. ntohs).
+
+config NIOS2_CI_SWAB_NO
+ int "Byteswap custom instruction number" if NIOS2_CI_SWAB_SUPPORT
+ default 0
+ help
+ Number of the instruction as configured in SOPC Builder.
+
+comment "Cache settings"
+
+config CUSTOM_CACHE_SETTINGS
+ bool "Custom cache settings"
+ help
+ This option allows you to tweak the cache settings used during early
+ boot (where the information from device tree is not yet available).
+ There should be no reason to change these values. Linux will work
+ perfectly fine, even if the Nios II is configured with smaller caches.
+
+ Say N here unless you know what you are doing.
+
+config NIOS2_DCACHE_SIZE
+ hex "D-Cache size" if CUSTOM_CACHE_SETTINGS
+ range 0x200 0x10000
+ default "0x800"
+ help
+ Maximum possible data cache size.
+
+config NIOS2_DCACHE_LINE_SIZE
+ hex "D-Cache line size" if CUSTOM_CACHE_SETTINGS
+ range 0x10 0x20
+ default "0x20"
+ help
+ Minimum possible data cache line size.
+
+config NIOS2_ICACHE_SIZE
+ hex "I-Cache size" if CUSTOM_CACHE_SETTINGS
+ range 0x200 0x10000
+ default "0x1000"
+ help
+ Maximum possible instruction cache size.
+
+endmenu
diff --git a/arch/nios2/platform/Makefile b/arch/nios2/platform/Makefile
new file mode 100644
index 0000000..46364f1d9352d97547e8b8fc336c2fea06adb5b8
--- /dev/null
+++ b/arch/nios2/platform/Makefile
@@ -0,0 +1 @@
+obj-y += platform.o
--
1.8.3.2

2014-04-18 12:52:13

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 11/28] nios2: Cache handling

This patch adds functionality required for cache maintenance.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/cache.h | 36 +++++
arch/nios2/include/asm/cacheflush.h | 48 +++++++
arch/nios2/mm/cacheflush.c | 255 ++++++++++++++++++++++++++++++++++++
3 files changed, 339 insertions(+)
create mode 100644 arch/nios2/include/asm/cache.h
create mode 100644 arch/nios2/include/asm/cacheflush.h
create mode 100644 arch/nios2/mm/cacheflush.c

diff --git a/arch/nios2/include/asm/cache.h b/arch/nios2/include/asm/cache.h
new file mode 100644
index 0000000..2293cf57e307a6b84e3972c2d27221dda54af500
--- /dev/null
+++ b/arch/nios2/include/asm/cache.h
@@ -0,0 +1,36 @@
+/*
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT. See the GNU General Public License for more
+ * details.
+ */
+
+#ifndef _ASM_NIOS2_CACHE_H
+#define _ASM_NIOS2_CACHE_H
+
+#define NIOS2_DCACHE_SIZE CONFIG_NIOS2_DCACHE_SIZE
+#define NIOS2_ICACHE_SIZE CONFIG_NIOS2_ICACHE_SIZE
+#define NIOS2_DCACHE_LINE_SIZE CONFIG_NIOS2_DCACHE_LINE_SIZE
+#define NIOS2_ICACHE_LINE_SHIFT 5
+#define NIOS2_ICACHE_LINE_SIZE (1 << NIOS2_ICACHE_LINE_SHIFT)
+
+/* bytes per L1 cache line */
+#define L1_CACHE_SHIFT NIOS2_ICACHE_LINE_SHIFT
+#define L1_CACHE_BYTES NIOS2_ICACHE_LINE_SIZE
+
+#define ARCH_DMA_MINALIGN L1_CACHE_BYTES
+
+#define __cacheline_aligned
+#define ____cacheline_aligned
+
+#endif
diff --git a/arch/nios2/include/asm/cacheflush.h b/arch/nios2/include/asm/cacheflush.h
new file mode 100644
index 0000000..ce2605a8c44de7e33686f9021b2b6736e94455a2
--- /dev/null
+++ b/arch/nios2/include/asm/cacheflush.h
@@ -0,0 +1,48 @@
+/*
+ * Copyright (C) 2003 Microtronix Datacom Ltd.
+ * Copyright (C) 2000-2002 Greg Ungerer <[email protected]>
+ *
+ * Ported from m68knommu.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_CACHEFLUSH_H
+#define _ASM_NIOS2_CACHEFLUSH_H
+
+#include <linux/mm_types.h>
+
+struct mm_struct;
+
+extern void flush_cache_all(void);
+extern void flush_cache_mm(struct mm_struct *mm);
+extern void flush_cache_dup_mm(struct mm_struct *mm);
+extern void flush_cache_range(struct vm_area_struct *vma, unsigned long start,
+ unsigned long end);
+extern void flush_cache_page(struct vm_area_struct *vma, unsigned long vmaddr,
+ unsigned long pfn);
+#define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 1
+extern void flush_dcache_page(struct page *page);
+
+extern void flush_icache_range(unsigned long start, unsigned long end);
+extern void flush_icache_page(struct vm_area_struct *vma, struct page *page);
+
+#define flush_cache_vmap(start, end) flush_dcache_range(start, end)
+#define flush_cache_vunmap(start, end) flush_dcache_range(start, end)
+
+extern void copy_to_user_page(struct vm_area_struct *vma, struct page *page,
+ unsigned long user_vaddr,
+ void *dst, void *src, int len);
+extern void copy_from_user_page(struct vm_area_struct *vma, struct page *page,
+ unsigned long user_vaddr,
+ void *dst, void *src, int len);
+
+extern void flush_dcache_range(unsigned long start, unsigned long end);
+extern void invalidate_dcache_range(unsigned long start, unsigned long end);
+
+#define flush_dcache_mmap_lock(mapping) do { } while (0)
+#define flush_dcache_mmap_unlock(mapping) do { } while (0)
+
+#endif /* _ASM_NIOS2_CACHEFLUSH_H */
diff --git a/arch/nios2/mm/cacheflush.c b/arch/nios2/mm/cacheflush.c
new file mode 100644
index 0000000..0c4d0d416f8f603f92b66017d100d3388f3d04c1
--- /dev/null
+++ b/arch/nios2/mm/cacheflush.c
@@ -0,0 +1,255 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2009, Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ */
+
+#include <linux/export.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/fs.h>
+
+#include <asm/cacheflush.h>
+#include <asm/cpuinfo.h>
+
+static void __flush_dcache(unsigned long start, unsigned long end)
+{
+ unsigned long addr;
+
+ start &= ~(cpuinfo.dcache_line_size - 1);
+ end += (cpuinfo.dcache_line_size - 1);
+ end &= ~(cpuinfo.dcache_line_size - 1);
+
+ if (end > start + cpuinfo.dcache_size)
+ end = start + cpuinfo.dcache_size;
+
+ for (addr = start; addr < end; addr += cpuinfo.dcache_line_size) {
+ __asm__ __volatile__ (" flushda 0(%0)\n"
+ : /* Outputs */
+ : /* Inputs */ "r"(addr)
+ /* : No clobber */);
+ }
+}
+
+static void __flush_dcache_all(unsigned long start, unsigned long end)
+{
+ unsigned long addr;
+
+ start &= ~(cpuinfo.dcache_line_size - 1);
+ end += (cpuinfo.dcache_line_size - 1);
+ end &= ~(cpuinfo.dcache_line_size - 1);
+
+ if (end > start + cpuinfo.dcache_size)
+ end = start + cpuinfo.dcache_size;
+
+ for (addr = start; addr < end; addr += cpuinfo.dcache_line_size) {
+ __asm__ __volatile__ (" flushd 0(%0)\n"
+ : /* Outputs */
+ : /* Inputs */ "r"(addr)
+ /* : No clobber */);
+ }
+}
+
+static void __invalidate_dcache(unsigned long start, unsigned long end)
+{
+ unsigned long addr;
+
+ start &= ~(cpuinfo.dcache_line_size - 1);
+ end += (cpuinfo.dcache_line_size - 1);
+ end &= ~(cpuinfo.dcache_line_size - 1);
+
+ if (end > start + cpuinfo.dcache_size)
+ end = start + cpuinfo.dcache_size;
+
+ for (addr = start; addr < end; addr += cpuinfo.dcache_line_size) {
+ __asm__ __volatile__ (" initda 0(%0)\n"
+ : /* Outputs */
+ : /* Inputs */ "r"(addr)
+ /* : No clobber */);
+ }
+}
+
+static void __flush_icache(unsigned long start, unsigned long end)
+{
+ unsigned long addr;
+
+ start &= ~(cpuinfo.icache_line_size - 1);
+ end += (cpuinfo.icache_line_size - 1);
+ end &= ~(cpuinfo.icache_line_size - 1);
+
+ if (end > start + cpuinfo.icache_size)
+ end = start + cpuinfo.icache_size;
+
+ for (addr = start; addr < end; addr += cpuinfo.icache_line_size) {
+ __asm__ __volatile__ (" flushi %0\n"
+ : /* Outputs */
+ : /* Inputs */ "r"(addr)
+ /* : No clobber */);
+ }
+ __asm__ __volatile(" flushp\n");
+}
+
+static void flush_aliases(struct address_space *mapping, struct page *page)
+{
+ struct mm_struct *mm = current->active_mm;
+ struct vm_area_struct *mpnt;
+ pgoff_t pgoff;
+
+ pgoff = page->index;
+
+ flush_dcache_mmap_lock(mapping);
+ vma_interval_tree_foreach(mpnt, &mapping->i_mmap, pgoff, pgoff) {
+ unsigned long offset;
+
+ if (mpnt->vm_mm != mm)
+ continue;
+ if (!(mpnt->vm_flags & VM_MAYSHARE))
+ continue;
+
+ offset = (pgoff - mpnt->vm_pgoff) << PAGE_SHIFT;
+ flush_cache_page(mpnt, mpnt->vm_start + offset,
+ page_to_pfn(page));
+ }
+ flush_dcache_mmap_unlock(mapping);
+}
+
+void flush_cache_all(void)
+{
+ __flush_dcache_all(0, cpuinfo.dcache_size);
+ __flush_icache(0, cpuinfo.icache_size);
+}
+
+void flush_cache_mm(struct mm_struct *mm)
+{
+ flush_cache_all();
+}
+
+void flush_cache_dup_mm(struct mm_struct *mm)
+{
+ flush_cache_all();
+}
+
+void flush_icache_range(unsigned long start, unsigned long end)
+{
+ __flush_icache(start, end);
+}
+
+void flush_dcache_range(unsigned long start, unsigned long end)
+{
+ __flush_dcache(start, end);
+}
+EXPORT_SYMBOL(flush_dcache_range);
+
+void invalidate_dcache_range(unsigned long start, unsigned long end)
+{
+ __invalidate_dcache(start, end);
+}
+EXPORT_SYMBOL(invalidate_dcache_range);
+
+void flush_cache_range(struct vm_area_struct *vma, unsigned long start,
+ unsigned long end)
+{
+ __flush_dcache(start, end);
+ if (vma == NULL || (vma->vm_flags & VM_EXEC))
+ __flush_icache(start, end);
+}
+
+void flush_icache_page(struct vm_area_struct *vma, struct page *page)
+{
+ unsigned long start = (unsigned long) page_address(page);
+ unsigned long end = start + PAGE_SIZE;
+
+ __flush_icache(start, end);
+}
+
+void flush_cache_page(struct vm_area_struct *vma, unsigned long vmaddr,
+ unsigned long pfn)
+{
+ unsigned long start = vmaddr;
+ unsigned long end = start + PAGE_SIZE;
+
+ __flush_dcache(start, end);
+ if (vma->vm_flags & VM_EXEC)
+ __flush_icache(start, end);
+}
+
+void flush_dcache_page(struct page *page)
+{
+ struct address_space *mapping = page_mapping(page);
+
+ /* Flush this page if there are aliases. */
+ if (mapping) {
+ if (!mapping_mapped(mapping)) {
+ clear_bit(PG_arch_1, &page->flags);
+ } else if (mapping) {
+ unsigned long start = (unsigned long)page_address(page);
+ __flush_dcache(start, start + PAGE_SIZE);
+ flush_aliases(mapping, page);
+ }
+ }
+}
+EXPORT_SYMBOL(flush_dcache_page);
+
+void update_mmu_cache(struct vm_area_struct *vma,
+ unsigned long address, pte_t *pte)
+{
+ unsigned long pfn = pte_pfn(*pte);
+ struct page *page;
+
+ if (!pfn_valid(pfn))
+ return;
+
+ page = pfn_to_page(pfn);
+
+ if (!PageReserved(page) && !test_bit(PG_arch_1, &page->flags)) {
+ unsigned long start = page_to_virt(page);
+ struct address_space *mapping;
+
+ __flush_dcache(start, start + PAGE_SIZE);
+
+ mapping = page_mapping(page);
+ if (mapping)
+ flush_aliases(mapping, page);
+ set_bit(PG_arch_1, &page->flags);
+ }
+}
+
+void copy_user_page(void *vto, void *vfrom, unsigned long vaddr,
+ struct page *to)
+{
+ __flush_dcache(vaddr, vaddr + PAGE_SIZE);
+ copy_page(vto, vfrom);
+ __flush_dcache((unsigned long)vto, (unsigned long)vto + PAGE_SIZE);
+}
+
+void clear_user_page(void *addr, unsigned long vaddr, struct page *page)
+{
+ __flush_dcache(vaddr, vaddr + PAGE_SIZE);
+ clear_page(addr);
+ __flush_dcache((unsigned long)addr, (unsigned long)addr + PAGE_SIZE);
+}
+
+void copy_from_user_page(struct vm_area_struct *vma, struct page *page,
+ unsigned long user_vaddr,
+ void *dst, void *src, int len)
+{
+ flush_cache_page(vma, user_vaddr, page_to_pfn(page));
+ memcpy(dst, src, len);
+ __flush_dcache((unsigned long)src, (unsigned long)src + len);
+ if (vma->vm_flags & VM_EXEC)
+ __flush_icache((unsigned long)src, (unsigned long)src + len);
+}
+
+void copy_to_user_page(struct vm_area_struct *vma, struct page *page,
+ unsigned long user_vaddr,
+ void *dst, void *src, int len)
+{
+ flush_cache_page(vma, user_vaddr, page_to_pfn(page));
+ memcpy(dst, src, len);
+ __flush_dcache((unsigned long)dst, (unsigned long)dst + len);
+ if (vma->vm_flags & VM_EXEC)
+ __flush_icache((unsigned long)dst, (unsigned long)dst + len);
+}
--
1.8.3.2

2014-04-18 12:52:10

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 21/28] nios2: Cpuinfo handling

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/cpuinfo.h | 58 +++++++++++
arch/nios2/kernel/cpuinfo.c | 201 +++++++++++++++++++++++++++++++++++++++
2 files changed, 259 insertions(+)
create mode 100644 arch/nios2/include/asm/cpuinfo.h
create mode 100644 arch/nios2/kernel/cpuinfo.c

diff --git a/arch/nios2/include/asm/cpuinfo.h b/arch/nios2/include/asm/cpuinfo.h
new file mode 100644
index 0000000..eafd7267e46b422ff1acc8efb6a771fb383cc5c4
--- /dev/null
+++ b/arch/nios2/include/asm/cpuinfo.h
@@ -0,0 +1,58 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_CPUINFO_H
+#define _ASM_NIOS2_CPUINFO_H
+
+#include <linux/types.h>
+#include <asm/prom.h>
+
+struct cpuinfo {
+ /* Core CPU configuration */
+ char cpu_impl[12];
+ u32 cpu_clock_freq;
+ u32 mmu;
+ u32 has_div;
+ u32 has_mul;
+ u32 has_mulx;
+
+ /* CPU caches */
+ u32 icache_line_size;
+ u32 icache_size;
+ u32 dcache_line_size;
+ u32 dcache_size;
+
+ /* TLB */
+ u32 tlb_pid_num_bits; /* number of bits used for the PID in TLBMISC */
+ u32 tlb_num_ways;
+ u32 tlb_num_ways_log2;
+ u32 tlb_num_entries;
+ u32 tlb_num_lines;
+ u32 tlb_ptr_sz;
+
+ /* Addresses */
+ u32 reset_addr;
+ u32 exception_addr;
+ u32 fast_tlb_miss_exc_addr;
+};
+
+extern struct cpuinfo cpuinfo;
+
+extern void setup_cpuinfo(void);
+
+#endif /* _ASM_NIOS2_CPUINFO_H */
diff --git a/arch/nios2/kernel/cpuinfo.c b/arch/nios2/kernel/cpuinfo.c
new file mode 100644
index 0000000..c7b677db5c3d6b4ae0404bee0fbc8881dace086d
--- /dev/null
+++ b/arch/nios2/kernel/cpuinfo.c
@@ -0,0 +1,201 @@
+/*
+ * Copyright (C) 2013 Altera Corporation
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ *
+ * Based on cpuinfo.c from microblaze
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/seq_file.h>
+#include <linux/string.h>
+#include <linux/of.h>
+#include <asm/cpuinfo.h>
+
+struct cpuinfo cpuinfo;
+
+#define err_cpu(x) \
+ pr_err("ERROR: Nios II " x " different for kernel and DTS\n");
+
+static inline u32 fcpu(struct device_node *cpu, const char *n)
+{
+ u32 val = 0;
+
+ of_property_read_u32(cpu, n, &val);
+
+ return val;
+}
+
+static inline u32 fcpu_has(struct device_node *cpu, const char *n)
+{
+ return of_get_property(cpu, n, NULL) ? 1 : 0;
+}
+
+void __init setup_cpuinfo(void)
+{
+ struct device_node *cpu;
+ const char *str;
+ int len;
+
+ cpu = of_find_node_by_type(NULL, "cpu");
+ if (!cpu)
+ panic("%s: No CPU found in devicetree!\n", __func__);
+
+ if (!fcpu_has(cpu, "ALTR,has-initda"))
+ panic("initda instruction is unimplemented. Please update your "
+ "hardware system to have more than 4-byte line data "
+ "cache\n");
+
+ cpuinfo.cpu_clock_freq = fcpu(cpu, "clock-frequency");
+
+ str = of_get_property(cpu, "ALTR,implementation", &len);
+ if (str)
+ strlcpy(cpuinfo.cpu_impl, str, sizeof(cpuinfo.cpu_impl));
+ else
+ strcpy(cpuinfo.cpu_impl, "<unknown>");
+
+ cpuinfo.has_div = fcpu_has(cpu, "ALTR,has-div");
+ cpuinfo.has_mul = fcpu_has(cpu, "ALTR,has-mul");
+ cpuinfo.has_mulx = fcpu_has(cpu, "ALTR,has-mulx");
+
+#ifdef CONFIG_NIOS2_HW_DIV_SUPPORT
+ if (!cpuinfo.has_div)
+ err_cpu("DIV");
+#endif
+#ifdef CONFIG_NIOS2_HW_MUL_SUPPORT
+ if (!cpuinfo.has_mul)
+ err_cpu("MUL");
+#endif
+#ifdef CONFIG_NIOS2_HW_MULX_SUPPORT
+ if (!cpuinfo.has_mulx)
+ err_cpu("MULX");
+#endif
+
+ cpuinfo.tlb_num_ways = fcpu(cpu, "ALTR,tlb-num-ways");
+ if (!cpuinfo.tlb_num_ways)
+ panic("ALTR,tlb-num-ways can't be 0. Please check your hardware "
+ "system\n");
+ cpuinfo.icache_line_size = fcpu(cpu, "icache-line-size");
+ cpuinfo.icache_size = fcpu(cpu, "icache-size");
+ if (CONFIG_NIOS2_ICACHE_SIZE != cpuinfo.icache_size)
+ pr_warn("Warning: icache size configuration mismatch "
+ "(0x%x vs 0x%x) of CONFIG_NIOS2_ICACHE_SIZE vs "
+ "device tree icache-size\n",
+ CONFIG_NIOS2_ICACHE_SIZE, cpuinfo.icache_size);
+
+ cpuinfo.dcache_line_size = fcpu(cpu, "dcache-line-size");
+ if (CONFIG_NIOS2_DCACHE_LINE_SIZE != cpuinfo.dcache_line_size)
+ pr_warn("Warning: dcache line size configuration mismatch "
+ "(0x%x vs 0x%x) of CONFIG_NIOS2_DCACHE_LINE_SIZE vs "
+ "device tree dcache-line-size\n",
+ CONFIG_NIOS2_DCACHE_LINE_SIZE, cpuinfo.dcache_line_size);
+ cpuinfo.dcache_size = fcpu(cpu, "dcache-size");
+ if (CONFIG_NIOS2_DCACHE_SIZE != cpuinfo.dcache_size)
+ pr_warn("Warning: dcache size configuration mismatch "
+ "(0x%x vs 0x%x) of CONFIG_NIOS2_DCACHE_SIZE vs "
+ "device tree dcache-size\n",
+ CONFIG_NIOS2_DCACHE_SIZE, cpuinfo.dcache_size);
+
+ cpuinfo.tlb_pid_num_bits = fcpu(cpu, "ALTR,pid-num-bits");
+ cpuinfo.tlb_num_ways_log2 = ilog2(cpuinfo.tlb_num_ways);
+ cpuinfo.tlb_num_entries = fcpu(cpu, "ALTR,tlb-num-entries");
+ cpuinfo.tlb_num_lines = cpuinfo.tlb_num_entries / cpuinfo.tlb_num_ways;
+ cpuinfo.tlb_ptr_sz = fcpu(cpu, "ALTR,tlb-ptr-sz");
+
+ cpuinfo.reset_addr = fcpu(cpu, "ALTR,reset-addr");
+ cpuinfo.exception_addr = fcpu(cpu, "ALTR,exception-addr");
+ cpuinfo.fast_tlb_miss_exc_addr = fcpu(cpu, "ALTR,fast-tlb-miss-addr");
+}
+
+#ifdef CONFIG_PROC_FS
+
+/*
+ * Get CPU information for use by the procfs.
+ */
+static int show_cpuinfo(struct seq_file *m, void *v)
+{
+ int count = 0;
+ const u32 clockfreq = cpuinfo.cpu_clock_freq;
+
+ count = seq_printf(m,
+ "CPU:\t\tNios II/%s\n"
+ "MMU:\t\t%s\n"
+ "FPU:\t\tnone\n"
+ "Clocking:\t%u.%02u MHz\n"
+ "BogoMips:\t%lu.%02lu\n"
+ "Calibration:\t%lu loops\n",
+ cpuinfo.cpu_impl,
+ cpuinfo.mmu ? "present" : "none",
+ clockfreq / 1000000, (clockfreq / 100000) % 10,
+ (loops_per_jiffy * HZ) / 500000,
+ ((loops_per_jiffy * HZ) / 5000) % 100,
+ (loops_per_jiffy * HZ));
+
+ count += seq_printf(m,
+ "HW:\n"
+ " MUL:\t\t%s\n"
+ " MULX:\t\t%s\n"
+ " DIV:\t\t%s\n",
+ cpuinfo.has_mul ? "yes" : "no",
+ cpuinfo.has_mulx ? "yes" : "no",
+ cpuinfo.has_div ? "yes" : "no");
+
+ count += seq_printf(m,
+ "Icache:\t\t%ukB, line length: %u\n",
+ cpuinfo.icache_size >> 10,
+ cpuinfo.icache_line_size);
+
+ count += seq_printf(m,
+ "Dcache:\t\t%ukB, line length: %u\n",
+ cpuinfo.dcache_size >> 10,
+ cpuinfo.dcache_line_size);
+
+ count += seq_printf(m,
+ "TLB:\t\t%u ways, %u entries, %u PID bits\n",
+ cpuinfo.tlb_num_ways,
+ cpuinfo.tlb_num_entries,
+ cpuinfo.tlb_pid_num_bits);
+
+ return 0;
+}
+
+static void *cpuinfo_start(struct seq_file *m, loff_t *pos)
+{
+ unsigned long i = *pos;
+
+ return i < num_possible_cpus() ? (void *) (i + 1) : NULL;
+}
+
+static void *cpuinfo_next(struct seq_file *m, void *v, loff_t *pos)
+{
+ ++*pos;
+ return cpuinfo_start(m, pos);
+}
+
+static void cpuinfo_stop(struct seq_file *m, void *v)
+{
+}
+
+const struct seq_operations cpuinfo_op = {
+ .start = cpuinfo_start,
+ .next = cpuinfo_next,
+ .stop = cpuinfo_stop,
+ .show = show_cpuinfo
+};
+
+#endif /* CONFIG_PROC_FS */
--
1.8.3.2

2014-04-18 12:53:15

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 02/28] nios2: Assembly macros and definitions

This patch add assembly macros and definitions used in
the .S files across arch/nios2/ and together with asm-offsets.c.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/asm-macros.h | 309 +++++++++++++++++++++++++++++++++++
arch/nios2/include/asm/asm-offsets.h | 20 +++
arch/nios2/kernel/asm-offsets.c | 88 ++++++++++
3 files changed, 417 insertions(+)
create mode 100644 arch/nios2/include/asm/asm-macros.h
create mode 100644 arch/nios2/include/asm/asm-offsets.h
create mode 100644 arch/nios2/kernel/asm-offsets.c

diff --git a/arch/nios2/include/asm/asm-macros.h b/arch/nios2/include/asm/asm-macros.h
new file mode 100644
index 0000000..29fa2e4d7b009da93a362731a278bc5296c70f25
--- /dev/null
+++ b/arch/nios2/include/asm/asm-macros.h
@@ -0,0 +1,309 @@
+/*
+ * Macro used to simplify coding multi-line assembler.
+ * Some of the bit test macro can simplify down to one line
+ * depending on the mask value.
+ *
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+ * NON INFRINGEMENT. See the GNU General Public License for more
+ * details.
+ *
+ */
+#ifndef _ASM_NIOS2_ASMMACROS_H
+#define _ASM_NIOS2_ASMMACROS_H
+/*
+ * ANDs reg2 with mask and places the result in reg1.
+ *
+ * You cannnot use the same register for reg1 & reg2.
+ */
+
+.macro ANDI32 reg1, reg2, mask
+.if \mask & 0xffff
+ .if \mask & 0xffff0000
+ movhi \reg1, %hi(\mask)
+ movui \reg1, %lo(\mask)
+ and \reg1, \reg1, \reg2
+ .else
+ andi \reg1, \reg2, %lo(\mask)
+ .endif
+.else
+ andhi \reg1, \reg2, %hi(\mask)
+.endif
+.endm
+
+/*
+ * ORs reg2 with mask and places the result in reg1.
+ *
+ * It is safe to use the same register for reg1 & reg2.
+ */
+
+.macro ORI32 reg1, reg2, mask
+.if \mask & 0xffff
+ .if \mask & 0xffff0000
+ orhi \reg1, \reg2, %hi(\mask)
+ ori \reg1, \reg2, %lo(\mask)
+ .else
+ ori \reg1, \reg2, %lo(\mask)
+ .endif
+.else
+ orhi \reg1, \reg2, %hi(\mask)
+.endif
+.endm
+
+/*
+ * XORs reg2 with mask and places the result in reg1.
+ *
+ * It is safe to use the same register for reg1 & reg2.
+ */
+
+.macro XORI32 reg1, reg2, mask
+.if \mask & 0xffff
+ .if \mask & 0xffff0000
+ xorhi \reg1, \reg2, %hi(\mask)
+ xori \reg1, \reg1, %lo(\mask)
+ .else
+ xori \reg1, \reg2, %lo(\mask)
+ .endif
+.else
+ xorhi \reg1, \reg2, %hi(\mask)
+.endif
+.endm
+
+/*
+ * This is a support macro for BTBZ & BTBNZ. It checks
+ * the bit to make sure it is valid 32 value.
+ *
+ * It is safe to use the same register for reg1 & reg2.
+ */
+
+.macro BT reg1, reg2, bit
+.if \bit > 31
+ .err
+.else
+ .if \bit < 16
+ andi \reg1, \reg2, (1 << \bit)
+ .else
+ andhi \reg1, \reg2, (1 << (\bit - 16))
+ .endif
+.endif
+.endm
+
+/*
+ * Tests the bit in reg2 and branches to label if the
+ * bit is zero. The result of the bit test is stored in reg1.
+ *
+ * It is safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTBZ reg1, reg2, bit, label
+ BT \reg1, \reg2, \bit
+ beq \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bit in reg2 and branches to label if the
+ * bit is non-zero. The result of the bit test is stored in reg1.
+ *
+ * It is safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTBNZ reg1, reg2, bit, label
+ BT \reg1, \reg2, \bit
+ bne \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bit in reg2 and then compliments the bit in reg2.
+ * The result of the bit test is stored in reg1.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTC reg1, reg2, bit
+.if \bit > 31
+ .err
+.else
+ .if \bit < 16
+ andi \reg1, \reg2, (1 << \bit)
+ xori \reg2, \reg2, (1 << \bit)
+ .else
+ andhi \reg1, \reg2, (1 << (\bit - 16))
+ xorhi \reg2, \reg2, (1 << (\bit - 16))
+ .endif
+.endif
+.endm
+
+/*
+ * Tests the bit in reg2 and then sets the bit in reg2.
+ * The result of the bit test is stored in reg1.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTS reg1, reg2, bit
+.if \bit > 31
+ .err
+.else
+ .if \bit < 16
+ andi \reg1, \reg2, (1 << \bit)
+ ori \reg2, \reg2, (1 << \bit)
+ .else
+ andhi \reg1, \reg2, (1 << (\bit - 16))
+ orhi \reg2, \reg2, (1 << (\bit - 16))
+ .endif
+.endif
+.endm
+
+/*
+ * Tests the bit in reg2 and then resets the bit in reg2.
+ * The result of the bit test is stored in reg1.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTR reg1, reg2, bit
+.if \bit > 31
+ .err
+.else
+ .if \bit < 16
+ andi \reg1, \reg2, (1 << \bit)
+ andi \reg2, \reg2, %lo(~(1 << \bit))
+ .else
+ andhi \reg1, \reg2, (1 << (\bit - 16))
+ andhi \reg2, \reg2, %lo(~(1 << (\bit - 16)))
+ .endif
+.endif
+.endm
+
+/*
+ * Tests the bit in reg2 and then compliments the bit in reg2.
+ * The result of the bit test is stored in reg1. If the
+ * original bit was zero it branches to label.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTCBZ reg1, reg2, bit, label
+ BTC \reg1, \reg2, \bit
+ beq \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bit in reg2 and then compliments the bit in reg2.
+ * The result of the bit test is stored in reg1. If the
+ * original bit was non-zero it branches to label.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTCBNZ reg1, reg2, bit, label
+ BTC \reg1, \reg2, \bit
+ bne \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bit in reg2 and then sets the bit in reg2.
+ * The result of the bit test is stored in reg1. If the
+ * original bit was zero it branches to label.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTSBZ reg1, reg2, bit, label
+ BTS \reg1, \reg2, \bit
+ beq \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bit in reg2 and then sets the bit in reg2.
+ * The result of the bit test is stored in reg1. If the
+ * original bit was non-zero it branches to label.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTSBNZ reg1, reg2, bit, label
+ BTS \reg1, \reg2, \bit
+ bne \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bit in reg2 and then resets the bit in reg2.
+ * The result of the bit test is stored in reg1. If the
+ * original bit was zero it branches to label.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTRBZ reg1, reg2, bit, label
+ BTR \reg1, \reg2, \bit
+ bne \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bit in reg2 and then resets the bit in reg2.
+ * The result of the bit test is stored in reg1. If the
+ * original bit was non-zero it branches to label.
+ *
+ * It is NOT safe to use the same register for reg1 & reg2.
+ */
+
+.macro BTRBNZ reg1, reg2, bit, label
+ BTR \reg1, \reg2, \bit
+ bne \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bits in mask against reg2 stores the result in reg1.
+ * If the all the bits in the mask are zero it branches to label.
+ *
+ * It is safe to use the same register for reg1 & reg2.
+ */
+
+.macro TSTBZ reg1, reg2, mask, label
+ ANDI32 \reg1, \reg2, \mask
+ beq \reg1, r0, \label
+.endm
+
+/*
+ * Tests the bits in mask against reg2 stores the result in reg1.
+ * If the any of the bits in the mask are 1 it branches to label.
+ *
+ * It is safe to use the same register for reg1 & reg2.
+ */
+
+.macro TSTBNZ reg1, reg2, mask, label
+ ANDI32 \reg1, \reg2, \mask
+ bne \reg1, r0, \label
+.endm
+
+/*
+ * Pushes reg onto the stack.
+ */
+
+.macro PUSH reg
+ addi sp, sp, -4
+ stw \reg, 0(sp)
+.endm
+
+/*
+ * Pops the top of the stack into reg.
+ */
+
+.macro POP reg
+ ldw \reg, 0(sp)
+ addi sp, sp, 4
+.endm
+
+
+#endif /* _ASM_NIOS2_ASMMACROS_H */
diff --git a/arch/nios2/include/asm/asm-offsets.h b/arch/nios2/include/asm/asm-offsets.h
new file mode 100644
index 0000000..5b9f5e04a058056e0aedcfe35b1a3512ac74e989
--- /dev/null
+++ b/arch/nios2/include/asm/asm-offsets.h
@@ -0,0 +1,20 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Thomas Chou <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include <generated/asm-offsets.h>
diff --git a/arch/nios2/kernel/asm-offsets.c b/arch/nios2/kernel/asm-offsets.c
new file mode 100644
index 0000000..3852f5c
--- /dev/null
+++ b/arch/nios2/kernel/asm-offsets.c
@@ -0,0 +1,88 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include <linux/stddef.h>
+#include <linux/sched.h>
+#include <linux/kernel_stat.h>
+#include <linux/ptrace.h>
+#include <linux/hardirq.h>
+#include <linux/thread_info.h>
+#include <linux/kbuild.h>
+
+int main(void)
+{
+ /* struct task_struct */
+ OFFSET(TASK_THREAD, task_struct, thread);
+ BLANK();
+
+ /* struct thread_struct */
+ OFFSET(THREAD_KSP, thread_struct, ksp);
+ OFFSET(THREAD_KPSR, thread_struct, kpsr);
+ BLANK();
+
+ /* struct pt_regs */
+ OFFSET(PT_ORIG_R2, pt_regs, orig_r2);
+ OFFSET(PT_ORIG_R7, pt_regs, orig_r7);
+
+ OFFSET(PT_R1, pt_regs, r1);
+ OFFSET(PT_R2, pt_regs, r2);
+ OFFSET(PT_R3, pt_regs, r3);
+ OFFSET(PT_R4, pt_regs, r4);
+ OFFSET(PT_R5, pt_regs, r5);
+ OFFSET(PT_R6, pt_regs, r6);
+ OFFSET(PT_R7, pt_regs, r7);
+ OFFSET(PT_R8, pt_regs, r8);
+ OFFSET(PT_R9, pt_regs, r9);
+ OFFSET(PT_R10, pt_regs, r10);
+ OFFSET(PT_R11, pt_regs, r11);
+ OFFSET(PT_R12, pt_regs, r12);
+ OFFSET(PT_R13, pt_regs, r13);
+ OFFSET(PT_R14, pt_regs, r14);
+ OFFSET(PT_R15, pt_regs, r15);
+ OFFSET(PT_EA, pt_regs, ea);
+ OFFSET(PT_RA, pt_regs, ra);
+ OFFSET(PT_FP, pt_regs, fp);
+ OFFSET(PT_SP, pt_regs, sp);
+ OFFSET(PT_GP, pt_regs, gp);
+ OFFSET(PT_ESTATUS, pt_regs, estatus);
+ DEFINE(PT_REGS_SIZE, sizeof(struct pt_regs));
+ BLANK();
+
+ /* struct switch_stack */
+ OFFSET(SW_R16, switch_stack, r16);
+ OFFSET(SW_R17, switch_stack, r17);
+ OFFSET(SW_R18, switch_stack, r18);
+ OFFSET(SW_R19, switch_stack, r19);
+ OFFSET(SW_R20, switch_stack, r20);
+ OFFSET(SW_R21, switch_stack, r21);
+ OFFSET(SW_R22, switch_stack, r22);
+ OFFSET(SW_R23, switch_stack, r23);
+ OFFSET(SW_FP, switch_stack, fp);
+ OFFSET(SW_GP, switch_stack, gp);
+ OFFSET(SW_RA, switch_stack, ra);
+ DEFINE(SWITCH_STACK_SIZE, sizeof(struct switch_stack));
+ BLANK();
+
+ /* struct thread_info */
+ OFFSET(TI_TASK, thread_info, task);
+ OFFSET(TI_FLAGS, thread_info, flags);
+ OFFSET(TI_PREEMPT_COUNT, thread_info, preempt_count);
+ BLANK();
+
+ return 0;
+}
--
1.8.3.2

2014-04-18 12:53:50

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 08/28] nios2: MMU Fault handling

This patch adds support for the handling of the MMU faults (exception
entry code introduced by a previous patch, kernel/entry.S).

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/mm/extable.c | 25 +++++
arch/nios2/mm/fault.c | 242 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 267 insertions(+)
create mode 100644 arch/nios2/mm/extable.c
create mode 100644 arch/nios2/mm/fault.c

diff --git a/arch/nios2/mm/extable.c b/arch/nios2/mm/extable.c
new file mode 100644
index 0000000..4d2fc5a589d09492630ddff2a52a7a6a96ab6486
--- /dev/null
+++ b/arch/nios2/mm/extable.c
@@ -0,0 +1,25 @@
+/*
+ * Copyright (C) 2010, Tobias Klauser <[email protected]>
+ * Copyright (C) 2009, Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/module.h>
+#include <linux/uaccess.h>
+
+int fixup_exception(struct pt_regs *regs)
+{
+ const struct exception_table_entry *fixup;
+
+ fixup = search_exception_tables(regs->ea);
+ if (fixup) {
+ regs->ea = fixup->fixup;
+ return 1;
+ }
+
+ return 0;
+}
diff --git a/arch/nios2/mm/fault.c b/arch/nios2/mm/fault.c
new file mode 100644
index 0000000..96258b18fc32204b98a70b2dcadc3cb8661539bc
--- /dev/null
+++ b/arch/nios2/mm/fault.c
@@ -0,0 +1,242 @@
+/*
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ *
+ * based on arch/mips/mm/fault.c which is:
+ *
+ * Copyright (C) 1995-2000 Ralf Baechle
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/uaccess.h>
+#include <linux/ptrace.h>
+
+#include <asm/mmu_context.h>
+#include <asm/traps.h>
+
+#define EXC_SUPERV_INSN_ACCESS 9 /* Supervisor only instruction address */
+#define EXC_SUPERV_DATA_ACCESS 11 /* Supervisor only data address */
+#define EXC_X_PROTECTION_FAULT 13 /* TLB permission violation (x) */
+#define EXC_R_PROTECTION_FAULT 14 /* TLB permission violation (r) */
+#define EXC_W_PROTECTION_FAULT 15 /* TLB permission violation (w) */
+
+/*
+ * This routine handles page faults. It determines the address,
+ * and the problem, and then passes it off to one of the appropriate
+ * routines.
+ */
+asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long cause,
+ unsigned long address)
+{
+ struct vm_area_struct *vma = NULL;
+ struct task_struct *tsk = current;
+ struct mm_struct *mm = tsk->mm;
+ int code = SEGV_MAPERR;
+ int fault;
+ unsigned int flags = 0;
+
+ cause >>= 2;
+
+ /* Restart the instruction */
+ regs->ea -= 4;
+
+ /*
+ * We fault-in kernel-space virtual memory on-demand. The
+ * 'reference' page table is init_mm.pgd.
+ *
+ * NOTE! We MUST NOT take any locks for this case. We may
+ * be in an interrupt or a critical region, and should
+ * only copy the information from the master page table,
+ * nothing more.
+ */
+ if (unlikely(address >= VMALLOC_START && address <= VMALLOC_END)) {
+ if (user_mode(regs))
+ goto bad_area_nosemaphore;
+ else
+ goto vmalloc_fault;
+ }
+
+ if (unlikely(address >= TASK_SIZE))
+ goto bad_area_nosemaphore;
+
+ /*
+ * If we're in an interrupt or have no user
+ * context, we must not take the fault..
+ */
+ if (in_atomic() || !mm)
+ goto bad_area_nosemaphore;
+
+ if (user_mode(regs))
+ flags |= FAULT_FLAG_USER;
+
+ down_read(&mm->mmap_sem);
+ vma = find_vma(mm, address);
+ if (!vma)
+ goto bad_area;
+ if (vma->vm_start <= address)
+ goto good_area;
+ if (!(vma->vm_flags & VM_GROWSDOWN))
+ goto bad_area;
+ if (expand_stack(vma, address))
+ goto bad_area;
+/*
+ * Ok, we have a good vm_area for this memory access, so
+ * we can handle it..
+ */
+good_area:
+ code = SEGV_ACCERR;
+
+ switch (cause) {
+ case EXC_SUPERV_INSN_ACCESS:
+ goto bad_area;
+ case EXC_SUPERV_DATA_ACCESS:
+ goto bad_area;
+ case EXC_X_PROTECTION_FAULT:
+ if (!(vma->vm_flags & VM_EXEC))
+ goto bad_area;
+ break;
+ case EXC_R_PROTECTION_FAULT:
+ if (!(vma->vm_flags & VM_READ))
+ goto bad_area;
+ break;
+ case EXC_W_PROTECTION_FAULT:
+ if (!(vma->vm_flags & VM_WRITE))
+ goto bad_area;
+ flags = FAULT_FLAG_WRITE;
+ break;
+ }
+
+survive:
+ /*
+ * If for any reason at all we couldn't handle the fault,
+ * make sure we exit gracefully rather than endlessly redo
+ * the fault.
+ */
+ fault = handle_mm_fault(mm, vma, address, flags);
+ if (unlikely(fault & VM_FAULT_ERROR)) {
+ if (fault & VM_FAULT_OOM)
+ goto out_of_memory;
+ else if (fault & VM_FAULT_SIGBUS)
+ goto do_sigbus;
+ BUG();
+ }
+ if (fault & VM_FAULT_MAJOR)
+ tsk->maj_flt++;
+ else
+ tsk->min_flt++;
+
+ up_read(&mm->mmap_sem);
+ return;
+
+/*
+ * Something tried to access memory that isn't in our memory map..
+ * Fix it, but check if it's kernel or user first..
+ */
+bad_area:
+ up_read(&mm->mmap_sem);
+
+bad_area_nosemaphore:
+ /* User mode accesses just cause a SIGSEGV */
+ if (user_mode(regs)) {
+ _exception(SIGSEGV, regs, code, address);
+ return;
+ }
+
+no_context:
+ /* Are we prepared to handle this kernel fault? */
+ if (fixup_exception(regs))
+ return;
+
+ /*
+ * Oops. The kernel tried to access some bad page. We'll have to
+ * terminate things with extreme prejudice.
+ */
+ bust_spinlocks(1);
+
+ pr_alert("Unable to handle kernel %s at virtual address %08lx",
+ address < PAGE_SIZE ? "NULL pointer dereference" :
+ "paging request", address);
+ pr_alert("ea = %08lx, ra = %08lx, cause = %ld\n", regs->ea, regs->ra,
+ cause);
+ panic("Oops");
+ return;
+
+/*
+ * We ran out of memory, or some other thing happened to us that made
+ * us unable to handle the page fault gracefully.
+ */
+out_of_memory:
+ up_read(&mm->mmap_sem);
+ if (is_global_init(tsk)) {
+ yield();
+ down_read(&mm->mmap_sem);
+ goto survive;
+ }
+ if (!user_mode(regs))
+ goto no_context;
+ pagefault_out_of_memory();
+ return;
+
+do_sigbus:
+ up_read(&mm->mmap_sem);
+
+ /* Kernel mode? Handle exceptions or die */
+ if (!user_mode(regs))
+ goto no_context;
+
+ _exception(SIGBUS, regs, BUS_ADRERR, address);
+ return;
+
+vmalloc_fault:
+ {
+ /*
+ * Synchronize this task's top level page-table
+ * with the 'reference' page table.
+ *
+ * Do _not_ use "tsk" here. We might be inside
+ * an interrupt in the middle of a task switch..
+ */
+ int offset = pgd_index(address);
+ pgd_t *pgd, *pgd_k;
+ pud_t *pud, *pud_k;
+ pmd_t *pmd, *pmd_k;
+ pte_t *pte_k;
+
+ pgd = pgd_current + offset;
+ pgd_k = init_mm.pgd + offset;
+
+ if (!pgd_present(*pgd_k))
+ goto no_context;
+ set_pgd(pgd, *pgd_k);
+
+ pud = pud_offset(pgd, address);
+ pud_k = pud_offset(pgd_k, address);
+ if (!pud_present(*pud_k))
+ goto no_context;
+ pmd = pmd_offset(pud, address);
+ pmd_k = pmd_offset(pud_k, address);
+ if (!pmd_present(*pmd_k))
+ goto no_context;
+ set_pmd(pmd, *pmd_k);
+
+ pte_k = pte_offset_kernel(pmd_k, address);
+ if (!pte_present(*pte_k))
+ goto no_context;
+
+ flush_tlb_one(address);
+ return;
+ }
+}
--
1.8.3.2

2014-04-18 14:36:04

by Sam Ravnborg

[permalink] [raw]
Subject: Re: [PATCH 01/28] nios2: Build infrastructure

On Fri, Apr 18, 2014 at 08:26:48PM +0800, Ley Foon Tan wrote:
> This patch adds Makefile and Kconfig files required for building a
> nios2 kernel.

It is good practice to add build infrastructure last - so build is not broken.
But for a new architecture it matters only little.

Some comments in the following.

Sam
>
> Signed-off-by: Ley Foon Tan <[email protected]>
> ---
>
> diff --git a/arch/nios2/Kconfig b/arch/nios2/Kconfig
> new file mode 100644
> index 0000000..2bd0d0d7160f31a5c5890cf8a059b9f53378be6e
> --- /dev/null
> +++ b/arch/nios2/Kconfig
> @@ -0,0 +1,213 @@
> +config NIOS2
> + def_bool y
> + select OF
> + select OF_EARLY_FLATTREE
> + select USB_ARCH_HAS_HCD if USB_SUPPORT
> + select ARCH_WANT_OPTIONAL_GPIOLIB
> + select GENERIC_IRQ_PROBE
> + select GENERIC_IRQ_SHOW
> + select GENERIC_CPU_DEVICES
> + select GENERIC_ATOMIC64
> + select MODULES_USE_ELF_RELA
> + select IRQ_DOMAIN
> + select SOC_BUS
> + select CLKSRC_OF
If you keep this list sorted then merge issues are less likely.

> +menu "Advanced setup"
> +
> +config ADVANCED_OPTIONS
> + bool "Prompt for advanced kernel configuration options"
> + help
> +
> +comment "Default settings for advanced configuration options are used"
> + depends on !ADVANCED_OPTIONS
> +
> +config KERNEL_MMU_REGION_BASE_BOOL
> + bool "Set custom kernel MMU region base address"
> + depends on ADVANCED_OPTIONS
> + help
> + This option allows you to set the virtual address of the kernel MMU region.
> +
> + Say N here unless you know what you are doing.

All Nios2 specific options should be prefixed with NIOS2_
This is what is done for all new arch specific symbols.

> diff --git a/arch/nios2/boot/Makefile b/arch/nios2/boot/Makefile
> new file mode 100644
> index 0000000..00aef5e266f9cc629f126445535f9aae6ca8962f
> --- /dev/null
> +++ b/arch/nios2/boot/Makefile
> @@ -0,0 +1,51 @@
> +#
> +# arch/nios2/boot/Makefile
> +#
> +# This file is subject to the terms and conditions of the GNU General Public
> +# License. See the file "COPYING" in the main directory of this archive
> +# for more details.
> +#
> +
> +UIMAGE_LOADADDR = $(shell $(NM) vmlinux | awk '$$NF == "_stext" {print $$1}')
> +UIMAGE_ENTRYADDR = $(shell $(NM) vmlinux | awk '$$NF == "_start" {print $$1}')
> +UIMAGE_COMPRESSION = gzip
> +
> +OBJCOPYFLAGS_vmlinux.bin := -O binary
> +$(obj)/vmlinux.bin: vmlinux FORCE
> + $(call if_changed,objcopy)
> +
> +$(obj)/vmlinux.gz: $(obj)/vmlinux.bin FORCE
> + $(call if_changed,gzip)
> +
> +$(obj)/vmImage: $(obj)/vmlinux.gz
> + $(call if_changed,uimage)
> + @$(kecho) 'Kernel: $@ is ready'

Add:
targets += vmlinux.bin vmlinux.gz vmImage
to make $(call if_changed ... actually work.


> +obj-y += cpuinfo.o insnemu.o irq.o nios2_ksyms.o process.o prom.o ptrace.o \
> + setup.o signal.o sys_nios2.o syscall_table.o time.o traps.o

In new stuff try to avoid using ... \
to contatenate lines.
Use:
obj-y += xxx

This is much more readable.

2014-04-18 16:49:05

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

Hi all

I just found that 4 patches (03/28, 05/28,07/28 and 14/28) are missing
here, because the issue in Git v1.8.3.2. But, the cover letter is correct.
Do I need to resend the whole series (28 patches) or just the missing 4 patches?
Thanks.

Regards.

On Fri, Apr 18, 2014 at 8:26 PM, Ley Foon Tan <[email protected]> wrote:
> This patchset adds the Linux kernel port for Nios II processor from Altera.
> The nios2 Linux port follows the guidance for new architecture ports using
> generic headers (including unistd.h).
>
> About Nios II Cores
> -------------------
> Nios II is a 32-bit embedded-processor architecture designed specifically for the
> Altera family of FPGAs.
> More information is available at http://www.altera.com/devices/processor/nios2/ni2-index.html
>
> Instruction set and architecture overview documents can be found on the
> following page:
> http://www.altera.com/literature/lit-nio2.jsp
>
> Nios2 GCC port is in mainline and and will be in the FSF 4.9 release.
>
> These patches are also available on this branch:
> git://git.rocketboards.org/linux-socfpga-next.git nios2-upstream
>
> Regards,
> Ley Foon
> --
> 1.8.3.2
>

2014-04-18 18:21:31

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 03/28] nios2: Kernel booting and initialization

This patch adds the kernel booting and the initial setup code.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/entry.h | 152 +++++++++++++++++++++++++++++
arch/nios2/include/asm/setup.h | 38 ++++++++
arch/nios2/kernel/head.S | 175 +++++++++++++++++++++++++++++++++
arch/nios2/kernel/setup.c | 214 +++++++++++++++++++++++++++++++++++++++++
4 files changed, 579 insertions(+)
create mode 100644 arch/nios2/include/asm/entry.h
create mode 100644 arch/nios2/include/asm/setup.h
create mode 100644 arch/nios2/kernel/head.S
create mode 100644 arch/nios2/kernel/setup.c

diff --git a/arch/nios2/include/asm/entry.h b/arch/nios2/include/asm/entry.h
new file mode 100644
index 0000000..8b7cddd086b02f1a404198244ca5a21178a0aadc
--- /dev/null
+++ b/arch/nios2/include/asm/entry.h
@@ -0,0 +1,152 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * Based on m68knommu asm/entry.h
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_ENTRY_H
+#define _ASM_NIOS2_ENTRY_H
+
+#ifdef __ASSEMBLY__
+
+#include <asm/processor.h>
+#include <asm/registers.h>
+#include <asm/asm-offsets.h>
+
+/*
+ * Stack layout in 'ret_from_exception':
+ *
+ * This allows access to the syscall arguments in registers r4-r8
+ *
+ * 0(sp) - r8
+ * 4(sp) - r9
+ * 8(sp) - r10
+ * C(sp) - r11
+ * 10(sp) - r12
+ * 14(sp) - r13
+ * 18(sp) - r14
+ * 1C(sp) - r15
+ * 20(sp) - r1
+ * 24(sp) - r2
+ * 28(sp) - r3
+ * 2C(sp) - r4
+ * 30(sp) - r5
+ * 34(sp) - r6
+ * 38(sp) - r7
+ * 3C(sp) - orig_r2
+ * 40(sp) - ra
+ * 44(sp) - fp
+ * 48(sp) - sp
+ * 4C(sp) - gp
+ * 50(sp) - estatus
+ * 54(sp) - status_extension (NOMMU only)
+ * 58(sp) - ea
+ */
+
+/*
+ * Standard Nios2 interrupt entry and exit macros.
+ * Must be called with interrupts disabled.
+ */
+.macro SAVE_ALL
+ rdctl r24, estatus
+ andi r24, r24, ESTATUS_EU
+ beq r24, r0, 1f /* In supervisor mode, already on kernel stack */
+
+ movia r24, _current_thread /* Switch to current kernel stack */
+ ldw r24, 0(r24) /* using the thread_info */
+ addi r24, r24, THREAD_SIZE-PT_REGS_SIZE
+ stw sp, PT_SP(r24) /* Save user stack before changing */
+ mov sp, r24
+ br 2f
+
+1 : mov r24, sp
+ addi sp, sp, -PT_REGS_SIZE /* Backup the kernel stack pointer */
+ stw r24, PT_SP(sp)
+2 : stw r1, PT_R1(sp)
+ stw r2, PT_R2(sp)
+ stw r3, PT_R3(sp)
+ stw r4, PT_R4(sp)
+ stw r5, PT_R5(sp)
+ stw r6, PT_R6(sp)
+ stw r7, PT_R7(sp)
+ stw r8, PT_R8(sp)
+ stw r9, PT_R9(sp)
+ stw r10, PT_R10(sp)
+ stw r11, PT_R11(sp)
+ stw r12, PT_R12(sp)
+ stw r13, PT_R13(sp)
+ stw r14, PT_R14(sp)
+ stw r15, PT_R15(sp)
+ stw r2, PT_ORIG_R2(sp)
+ stw r7, PT_ORIG_R7(sp)
+
+ stw ra, PT_RA(sp)
+ stw fp, PT_FP(sp)
+ stw gp, PT_GP(sp)
+ rdctl r24, estatus
+ stw r24, PT_ESTATUS(sp)
+ stw ea, PT_EA(sp)
+.endm
+
+.macro RESTORE_ALL
+ ldw r1, PT_R1(sp) /* Restore registers */
+ ldw r2, PT_R2(sp)
+ ldw r3, PT_R3(sp)
+ ldw r4, PT_R4(sp)
+ ldw r5, PT_R5(sp)
+ ldw r6, PT_R6(sp)
+ ldw r7, PT_R7(sp)
+ ldw r8, PT_R8(sp)
+ ldw r9, PT_R9(sp)
+ ldw r10, PT_R10(sp)
+ ldw r11, PT_R11(sp)
+ ldw r12, PT_R12(sp)
+ ldw r13, PT_R13(sp)
+ ldw r14, PT_R14(sp)
+ ldw r15, PT_R15(sp)
+ ldw ra, PT_RA(sp)
+ ldw fp, PT_FP(sp)
+ ldw gp, PT_GP(sp)
+ ldw r24, PT_ESTATUS(sp)
+ wrctl estatus, r24
+ ldw ea, PT_EA(sp)
+ ldw sp, PT_SP(sp) /* Restore sp last */
+.endm
+
+.macro SAVE_SWITCH_STACK
+ addi sp, sp, -SWITCH_STACK_SIZE
+ stw r16, SW_R16(sp)
+ stw r17, SW_R17(sp)
+ stw r18, SW_R18(sp)
+ stw r19, SW_R19(sp)
+ stw r20, SW_R20(sp)
+ stw r21, SW_R21(sp)
+ stw r22, SW_R22(sp)
+ stw r23, SW_R23(sp)
+ stw fp, SW_FP(sp)
+ stw gp, SW_GP(sp)
+ stw ra, SW_RA(sp)
+.endm
+
+.macro RESTORE_SWITCH_STACK
+ ldw r16, SW_R16(sp)
+ ldw r17, SW_R17(sp)
+ ldw r18, SW_R18(sp)
+ ldw r19, SW_R19(sp)
+ ldw r20, SW_R20(sp)
+ ldw r21, SW_R21(sp)
+ ldw r22, SW_R22(sp)
+ ldw r23, SW_R23(sp)
+ ldw fp, SW_FP(sp)
+ ldw gp, SW_GP(sp)
+ ldw ra, SW_RA(sp)
+ addi sp, sp, SWITCH_STACK_SIZE
+.endm
+
+#endif /* __ASSEMBLY__ */
+#endif /* _ASM_NIOS2_ENTRY_H */
diff --git a/arch/nios2/include/asm/setup.h b/arch/nios2/include/asm/setup.h
new file mode 100644
index 0000000..a3e05a7f977a5de50439a1595146a280577ba015
--- /dev/null
+++ b/arch/nios2/include/asm/setup.h
@@ -0,0 +1,38 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef _ASM_NIOS2_SETUP_H
+#define _ASM_NIOS2_SETUP_H
+
+#include <asm-generic/setup.h>
+
+#ifndef __ASSEMBLY__
+# ifdef __KERNEL__
+
+extern char exception_handler_hook[];
+extern char fast_handler[];
+extern char fast_handler_end[];
+
+extern void pagetable_init(void);
+
+extern void setup_early_printk(void);
+
+# endif/* __KERNEL__ */
+#endif /* __ASSEMBLY__ */
+
+#endif /* _ASM_NIOS2_SETUP_H */
diff --git a/arch/nios2/kernel/head.S b/arch/nios2/kernel/head.S
new file mode 100644
index 0000000..cff8fa4f610afe4518ed470aecd97169af0272d1
--- /dev/null
+++ b/arch/nios2/kernel/head.S
@@ -0,0 +1,175 @@
+/*
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ * Copyright (C) 2004 Microtronix Datacom Ltd
+ * Copyright (C) 2001 Vic Phillips, Microtronix Datacom Ltd.
+ *
+ * Based on head.S for Altera's Excalibur development board with nios processor
+ *
+ * Based on the following from the Excalibur sdk distribution:
+ * NA_MemoryMap.s, NR_JumpToStart.s, NR_Setup.s, NR_CWPManager.s
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/init.h>
+#include <linux/linkage.h>
+#include <asm/thread_info.h>
+#include <asm/processor.h>
+#include <asm/cache.h>
+#include <asm/page.h>
+#include <asm/asm-offsets.h>
+#include <asm/asm-macros.h>
+
+/*
+ * ZERO_PAGE is a special page that is used for zero-initialized
+ * data and COW.
+ */
+.data
+.global empty_zero_page
+.align 12
+empty_zero_page:
+ .space PAGE_SIZE
+
+/*
+ * This global variable is used as an extension to the nios'
+ * STATUS register to emulate a user/supervisor mode.
+ */
+ .data
+ .align 2
+ .set noat
+
+ .global _current_thread
+_current_thread:
+ .long 0
+/*
+ * Input(s): passed from u-boot
+ * r4 - Optional pointer to a board information structure.
+ * r5 - Optional pointer to the physical starting address of the init RAM
+ * disk.
+ * r6 - Optional pointer to the physical ending address of the init RAM
+ * disk.
+ * r7 - Optional pointer to the physical starting address of any kernel
+ * command-line parameters.
+ */
+
+/*
+ * First executable code - detected and jumped to by the ROM bootstrap
+ * if the code resides in flash (looks for "Nios" at offset 0x0c from
+ * the potential executable image).
+ */
+ __HEAD
+ENTRY(_start)
+ wrctl status, r0 /* Disable interrupts */
+
+ /* Flush all cache lines within the instruction cache */
+ movia r1, NIOS2_ICACHE_SIZE
+ movui r2, NIOS2_ICACHE_LINE_SIZE
+
+text_flush:
+ flushi r1
+ sub r1, r1, r2
+ bgt r1, r0, text_flush
+ br 1f
+
+ /*
+ * This is the default location for the exception handler. Code in jump
+ * to our handler
+ */
+ENTRY(exception_handler_hook)
+ movia r24, inthandler
+ jmp r24
+
+ENTRY(fast_handler)
+ nextpc et
+helper:
+ stw r3, r3save - helper(et)
+
+ rdctl r3 , pteaddr
+ srli r3, r3, 12
+ slli r3, r3, 2
+ movia et, pgd_current
+
+ ldw et, 0(et)
+ add r3, et, r3
+ ldw et, 0(r3)
+
+ rdctl r3, pteaddr
+ andi r3, r3, 0xfff
+ add et, r3, et
+ ldw et, 0(et)
+ wrctl tlbacc, et
+ nextpc et
+helper2:
+ ldw r3, r3save - helper2(et)
+ subi ea, ea, 4
+ eret
+r3save:
+ .word 0x0
+ENTRY(fast_handler_end)
+
+1:
+ /*
+ * After flushing the instruction cache, we must flush the data
+ * cache.
+ */
+ movia r1, NIOS2_DCACHE_SIZE
+ movui r2, NIOS2_DCACHE_LINE_SIZE
+
+data_flush:
+ flushd 0(r1)
+ sub r1, r1, r2
+ bgt r1, r0, data_flush
+
+ nextpc r1 /* Find out where we are */
+chkadr:
+ movia r2, chkadr
+ beq r1, r2,finish_move /* We are running in RAM done */
+ addi r1, r1,(_start - chkadr) /* Source */
+ movia r2, _start /* Destination */
+ movia r3, __bss_start /* End of copy */
+
+loop_move: /* r1: src, r2: dest, r3: last dest */
+ ldw r8, 0(r1) /* load a word from [r1] */
+ stw r8, 0(r2) /* store a word to dest [r2] */
+ flushd 0(r2) /* Flush cache for safety */
+ addi r1, r1, 4 /* inc the src addr */
+ addi r2, r2, 4 /* inc the dest addr */
+ blt r2, r3, loop_move
+
+ movia r1, finish_move /* VMA(_start)->l1 */
+ jmp r1 /* jmp to _start */
+
+finish_move:
+
+ /* Mask off all possible interrupts */
+ wrctl ienable, r0
+
+ /* Clear .bss */
+ movia r2, __bss_start
+ movia r1, __bss_stop
+1:
+ stb r0, 0(r2)
+ addi r2, r2, 1
+ bne r1, r2, 1b
+
+ movia r1, init_thread_union /* set stack at top of the task union */
+ addi sp, r1, THREAD_SIZE
+ movia r2, _current_thread /* Remember current thread */
+ stw r1, 0(r2)
+
+ movia r1, nios2_boot_init /* save args r4-r7 passed from u-boot */
+ callr r1
+
+ movia r1, start_kernel /* call start_kernel as a subroutine */
+ callr r1
+
+ /* If we return from start_kernel, break to the oci debugger and
+ * buggered we are.
+ */
+ break
+
+ /* End of startup code */
+.set at
diff --git a/arch/nios2/kernel/setup.c b/arch/nios2/kernel/setup.c
new file mode 100644
index 0000000..549f391baa75b6dc759e54d77facd929b29cfabb
--- /dev/null
+++ b/arch/nios2/kernel/setup.c
@@ -0,0 +1,214 @@
+/*
+ * Nios2-specific parts of system setup
+ *
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ * Copyright (C) 2001 Vic Phillips <[email protected]>
+ *
+ * based on kernel/setup.c from m68knommu
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/export.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/sched.h>
+#include <linux/console.h>
+#include <linux/bootmem.h>
+#include <linux/initrd.h>
+#include <linux/of_fdt.h>
+
+#include <asm/mmu_context.h>
+#include <asm/sections.h>
+#include <asm/setup.h>
+#include <asm/prom.h>
+#include <asm/cpuinfo.h>
+
+unsigned long memory_start;
+EXPORT_SYMBOL(memory_start);
+
+unsigned long memory_end;
+EXPORT_SYMBOL(memory_end);
+
+unsigned long memory_size;
+
+/* r1 r2 r3 r4 r5 r6 r7 r8 r9 r10 r11*/
+/* r12 r13 r14 r15 or2 ra fp sp gp es ste ea*/
+static struct pt_regs fake_regs = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+ 0, 0, 0, 0, 0, 0,
+ 0};
+
+/* Copy a short hook instruction sequence to the exception address */
+static inline void copy_exception_handler(unsigned int addr)
+{
+ unsigned int start = (unsigned int) exception_handler_hook;
+ volatile unsigned int tmp = 0;
+
+ __asm__ __volatile__ (
+ "ldw %2,0(%0)\n"
+ "stw %2,0(%1)\n"
+ "ldw %2,4(%0)\n"
+ "stw %2,4(%1)\n"
+ "ldw %2,8(%0)\n"
+ "stw %2,8(%1)\n"
+ "flushd 0(%1)\n"
+ "flushd 4(%1)\n"
+ "flushd 8(%1)\n"
+ "flushi %1\n"
+ "addi %1,%1,4\n"
+ "flushi %1\n"
+ "addi %1,%1,4\n"
+ "flushi %1\n"
+ : /* no output registers */
+ : "r" (start), "r" (addr), "r" (tmp)
+ : "memory"
+ );
+}
+
+/* Copy the fast TLB miss handler */
+static inline void copy_fast_tlb_miss_handler(unsigned int addr)
+{
+ unsigned int start = (unsigned int) fast_handler;
+ unsigned int end = (unsigned int) fast_handler_end;
+ volatile unsigned int tmp = 0;
+
+ __asm__ __volatile__ (
+ "1:\n"
+ " ldw %3,0(%0)\n"
+ " stw %3,0(%1)\n"
+ " flushd 0(%1)\n"
+ " flushi %1\n"
+ " addi %0,%0,4\n"
+ " addi %1,%1,4\n"
+ " bne %0,%2,1b\n"
+ : /* no output registers */
+ : "r" (start), "r" (addr), "r" (end), "r" (tmp)
+ : "memory"
+ );
+}
+
+/*
+ * save args passed from u-boot, called from head.S
+ *
+ * @r4: NIOS magic
+ * @r5: initrd start
+ * @r6: initrd end or fdt
+ * @r7: kernel command line
+ */
+asmlinkage void __init nios2_boot_init(unsigned r4, unsigned r5, unsigned r6,
+ unsigned r7)
+{
+ unsigned dtb_passed = 0;
+ char cmdline_passed[COMMAND_LINE_SIZE] = { 0, };
+
+#if defined(CONFIG_PASS_CMDLINE)
+ if (r4 == 0x534f494e) { /* r4 is magic NIOS */
+#if defined(CONFIG_BLK_DEV_INITRD)
+ if (r5) { /* initramfs */
+ initrd_start = r5;
+ initrd_end = r6;
+ }
+#endif /* CONFIG_BLK_DEV_INITRD */
+ dtb_passed = r6;
+
+ if (r7)
+ strncpy(cmdline_passed, (char *)r7, COMMAND_LINE_SIZE);
+ }
+#endif
+
+ early_init_devtree((void *)dtb_passed);
+
+#ifndef CONFIG_CMDLINE_FORCE
+ if (cmdline_passed[0])
+ strncpy(boot_command_line, cmdline_passed, COMMAND_LINE_SIZE);
+#ifdef CONFIG_CMDLINE_IGNORE_DTB
+ else
+ strncpy(boot_command_line, CONFIG_CMDLINE, COMMAND_LINE_SIZE);
+#endif
+#endif
+}
+
+void __init setup_arch(char **cmdline_p)
+{
+ int bootmap_size;
+
+ console_verbose();
+
+ memory_start = PAGE_ALIGN((unsigned long)__pa(_end));
+ memory_end = (unsigned long) CONFIG_MEM_BASE + memory_size;
+
+ init_mm.start_code = (unsigned long) _stext;
+ init_mm.end_code = (unsigned long) _etext;
+ init_mm.end_data = (unsigned long) _edata;
+ init_mm.brk = (unsigned long) _end;
+ init_task.thread.kregs = &fake_regs;
+
+ /* Keep a copy of command line */
+ *cmdline_p = boot_command_line;
+
+ /*
+ * give all the memory to the bootmap allocator, tell it to put the
+ * boot mem_map at the start of memory
+ */
+ pr_debug("init_bootmem_node(?,%#lx, %#x, %#lx)\n",
+ PFN_UP(memory_start), PFN_DOWN(PHYS_OFFSET),
+ PFN_DOWN(memory_end));
+ bootmap_size = init_bootmem_node(NODE_DATA(0),
+ PFN_UP(memory_start),
+ PFN_DOWN(PHYS_OFFSET),
+ PFN_DOWN(memory_end));
+
+ /*
+ * free the usable memory, we have to make sure we do not free
+ * the bootmem bitmap so we then reserve it after freeing it :-)
+ */
+ pr_debug("free_bootmem(%#lx, %#lx)\n",
+ memory_start, memory_end - memory_start);
+ free_bootmem(memory_start, memory_end - memory_start);
+
+ /*
+ * Reserve the bootmem bitmap itself as well. We do this in two
+ * steps (first step was init_bootmem()) because this catches
+ * the (very unlikely) case of us accidentally initializing the
+ * bootmem allocator with an invalid RAM area.
+ *
+ * Arguments are start, size
+ */
+ pr_debug("reserve_bootmem(%#lx, %#x)\n", memory_start, bootmap_size);
+ reserve_bootmem(memory_start, bootmap_size, BOOTMEM_DEFAULT);
+
+#ifdef CONFIG_BLK_DEV_INITRD
+ if (initrd_start) {
+ reserve_bootmem(virt_to_phys((void *)initrd_start),
+ initrd_end - initrd_start, BOOTMEM_DEFAULT);
+ }
+#endif /* CONFIG_BLK_DEV_INITRD */
+
+ unflatten_and_copy_device_tree();
+
+ setup_cpuinfo();
+
+ copy_exception_handler(cpuinfo.exception_addr);
+
+ mmu_init();
+
+ copy_fast_tlb_miss_handler(cpuinfo.fast_tlb_miss_exc_addr);
+
+ /*
+ * Initialize MMU context handling here because data from cpuinfo is
+ * needed for this.
+ */
+ mmu_context_init();
+
+ /*
+ * get kmalloc into gear
+ */
+ paging_init();
+
+#if defined(CONFIG_VT) && defined(CONFIG_DUMMY_CONSOLE)
+ conswitchp = &dummy_con;
+#endif
+}
--
1.8.3.2

2014-04-18 18:21:48

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 07/28] nios2: I/O Mapping

This patch adds several definitions for I/O accessors and ioremap().

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/io.h | 227 ++++++++++++++++++++++++++++++++++++++++++++
arch/nios2/mm/ioremap.c | 188 ++++++++++++++++++++++++++++++++++++
2 files changed, 415 insertions(+)
create mode 100644 arch/nios2/include/asm/io.h
create mode 100644 arch/nios2/mm/ioremap.c

diff --git a/arch/nios2/include/asm/io.h b/arch/nios2/include/asm/io.h
new file mode 100644
index 0000000..46f4a0438870070e50563a06b329e03e552bf83d
--- /dev/null
+++ b/arch/nios2/include/asm/io.h
@@ -0,0 +1,227 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_IO_H
+#define _ASM_NIOS2_IO_H
+
+#include <asm/pgtable-bits.h>
+
+#define IO_SPACE_LIMIT 0xffffffff
+
+#ifndef CONFIG_CC_OPTIMIZE_FOR_SIZE
+# define __IO_USE_DUFFS
+#endif
+
+#ifdef __IO_USE_DUFFS
+
+/* Use "Duff's Device" to unroll the loops. */
+#define __IO_OUT_LOOP(a, b, l) \
+ do { \
+ if (l > 0) { \
+ int _n = (l + 7) / 8; \
+ switch (l % 8) { \
+ case 0: \
+ do { \
+ *a = *b++; \
+ case 7: \
+ *a = *b++; \
+ case 6: \
+ *a = *b++; \
+ case 5: \
+ *a = *b++; \
+ case 4: \
+ *a = *b++; \
+ case 3: \
+ *a = *b++; \
+ case 2: \
+ *a = *b++; \
+ case 1: \
+ *a = *b++; \
+ } while (--_n > 0); \
+ } \
+ } \
+ } while (0)
+
+#define __IO_IN_LOOP(a, b, l) \
+ do { \
+ if (l > 0) { \
+ int _n = (l + 7) / 8; \
+ switch (l % 8) { \
+ case 0: \
+ do { \
+ *b++ = *a; \
+ case 7: \
+ *b++ = *a; \
+ case 6: \
+ *b++ = *a; \
+ case 5: \
+ *b++ = *a; \
+ case 4: \
+ *b++ = *a; \
+ case 3: \
+ *b++ = *a; \
+ case 2: \
+ *b++ = *a; \
+ case 1: \
+ *b++ = *a; \
+ } while (--_n > 0); \
+ } \
+ } \
+ } while (0)
+
+#else /* __IO_USE_DUFFS */
+
+/* Use simple loops. */
+#define __IO_OUT_LOOP(a, b, l) \
+ do { \
+ while (l--) \
+ *a = *b++; \
+ } while (0)
+
+#define __IO_IN_LOOP(a, b, l) \
+ do { \
+ while (l--) \
+ *b++ = *a; \
+ } while (0)
+
+#endif /* __IO_USE_DUFFS */
+
+static inline void io_outsb(unsigned int addr, const void *buf, int len)
+{
+ volatile unsigned char *ap = (volatile unsigned char *)addr;
+ unsigned char *bp = (unsigned char *)buf;
+ __IO_OUT_LOOP(ap, bp, len);
+}
+
+static inline void io_outsw(unsigned int addr, const void *buf, int len)
+{
+ volatile unsigned short *ap = (volatile unsigned short *)addr;
+ unsigned short *bp = (unsigned short *)buf;
+ __IO_OUT_LOOP(ap, bp, len);
+}
+
+static inline void io_outsl(unsigned int addr, const void *buf, int len)
+{
+ volatile unsigned int *ap = (volatile unsigned int *)addr;
+ unsigned int *bp = (unsigned int *)buf;
+ __IO_OUT_LOOP(ap, bp, len);
+}
+
+static inline void io_insb(unsigned int addr, void *buf, int len)
+{
+ volatile unsigned char *ap = (volatile unsigned char *)addr;
+ unsigned char *bp = (unsigned char *)buf;
+ __IO_IN_LOOP(ap, bp, len);
+}
+
+static inline void io_insw(unsigned int addr, void *buf, int len)
+{
+ volatile unsigned short *ap = (volatile unsigned short *)addr;
+ unsigned short *bp = (unsigned short *)buf;
+ __IO_IN_LOOP(ap, bp, len);
+}
+
+static inline void io_insl(unsigned int addr, void *buf, int len)
+{
+ volatile unsigned int *ap = (volatile unsigned int *)addr;
+ unsigned int *bp = (unsigned int *)buf;
+ __IO_IN_LOOP(ap, bp, len);
+}
+
+#undef __IO_OUT_LOOP
+#undef __IO_IN_LOOP
+#undef __IO_USE_DUFFS
+
+#define readb_relaxed(addr) readb(addr)
+#define readw_relaxed(addr) readw(addr)
+#define readl_relaxed(addr) readl(addr)
+
+#define writeb_relaxed(x, addr) writeb(x, addr)
+#define writew_relaxed(x, addr) writew(x, addr)
+#define writel_relaxed(x, addr) writel(x, addr)
+
+#define outsb(a, b, l) io_outsb(a, b, l)
+#define outsw(a, b, l) io_outsw(a, b, l)
+#define outsl(a, b, l) io_outsl(a, b, l)
+
+#define insb(a, b, l) io_insb(a, b, l)
+#define insw(a, b, l) io_insw(a, b, l)
+#define insl(a, b, l) io_insl(a, b, l)
+
+#include <asm-generic/io.h>
+
+/*
+ * make the short names macros so specific devices
+ * can override them as required
+ */
+#define inb(addr) readb(addr)
+#define inw(addr) readw(addr)
+#define inl(addr) readl(addr)
+#define outb(x, addr) ((void) writeb(x, addr))
+#define outw(x, addr) ((void) writew(x, addr))
+#define outl(x, addr) ((void) writel(x, addr))
+
+extern void __iomem *__ioremap(unsigned long physaddr, unsigned long size,
+ unsigned long cacheflag);
+extern void __iounmap(void __iomem *addr);
+
+static inline void __iomem *ioremap(unsigned long physaddr, unsigned long size)
+{
+ return __ioremap(physaddr, size, 0);
+}
+
+static inline void __iomem *ioremap_nocache(unsigned long physaddr,
+ unsigned long size)
+{
+ return __ioremap(physaddr, size, 0);
+}
+
+static inline void __iomem *ioremap_writethrough(unsigned long physaddr,
+ unsigned long size)
+{
+ return __ioremap(physaddr, size, 0);
+}
+
+static inline void __iomem *ioremap_fullcache(unsigned long physaddr,
+ unsigned long size)
+{
+ return __ioremap(physaddr, size, _PAGE_CACHED);
+}
+
+static inline void iounmap(void __iomem *addr)
+{
+ __iounmap(addr);
+}
+
+/* Pages to physical address... */
+# define page_to_phys(page) virt_to_phys(page_to_virt(page))
+# define page_to_bus(page) page_to_virt(page)
+
+/* Macros used for converting between virtual and physical mappings. */
+# define phys_to_virt(vaddr) \
+ ((void *)((unsigned long)(vaddr) | CONFIG_KERNEL_REGION_BASE))
+/* Clear top 3 bits */
+# define virt_to_phys(vaddr) \
+ ((unsigned long)((unsigned long)(vaddr) & ~0xE0000000))
+
+#define virt_to_bus virt_to_phys
+#define bus_to_virt phys_to_virt
+
+#define ioport_map(port, nr) ioremap(port, nr)
+#define ioport_unmap(port) iounmap(port)
+
+/* Macros used for smc91x.c driver */
+#define readsb(p, d, l) insb(p, d, l)
+#define readsw(p, d, l) insw(p, d, l)
+#define readsl(p, d, l) insl(p, d, l)
+#define writesb(p, d, l) outsb(p, d, l)
+#define writesw(p, d, l) outsw(p, d, l)
+#define writesl(p, d, l) outsl(p, d, l)
+
+#endif /* _ASM_NIOS2_IO_H */
diff --git a/arch/nios2/mm/ioremap.c b/arch/nios2/mm/ioremap.c
new file mode 100644
index 0000000..596b0c0
--- /dev/null
+++ b/arch/nios2/mm/ioremap.c
@@ -0,0 +1,188 @@
+/*
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * based on arch/m68knommu/mm/kmap.c
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/export.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/io.h>
+
+#include <asm/cacheflush.h>
+#include <asm/tlbflush.h>
+
+static inline void remap_area_pte(pte_t *pte, unsigned long address,
+ unsigned long size, unsigned long phys_addr,
+ unsigned long flags)
+{
+ unsigned long end;
+ unsigned long pfn;
+ pgprot_t pgprot = __pgprot(_PAGE_GLOBAL | _PAGE_PRESENT | _PAGE_READ
+ | _PAGE_WRITE | flags);
+
+ address &= ~PMD_MASK;
+ end = address + size;
+ if (end > PMD_SIZE)
+ end = PMD_SIZE;
+ if (address >= end)
+ BUG();
+ pfn = PFN_DOWN(phys_addr);
+ do {
+ if (!pte_none(*pte)) {
+ pr_err("remap_area_pte: page already exists\n");
+ BUG();
+ }
+ set_pte(pte, pfn_pte(pfn, pgprot));
+ address += PAGE_SIZE;
+ pfn++;
+ pte++;
+ } while (address && (address < end));
+}
+
+static inline int remap_area_pmd(pmd_t *pmd, unsigned long address,
+ unsigned long size, unsigned long phys_addr,
+ unsigned long flags)
+{
+ unsigned long end;
+
+ address &= ~PGDIR_MASK;
+ end = address + size;
+ if (end > PGDIR_SIZE)
+ end = PGDIR_SIZE;
+ phys_addr -= address;
+ if (address >= end)
+ BUG();
+ do {
+ pte_t *pte = pte_alloc_kernel(pmd, address);
+ if (!pte)
+ return -ENOMEM;
+ remap_area_pte(pte, address, end - address, address + phys_addr,
+ flags);
+ address = (address + PMD_SIZE) & PMD_MASK;
+ pmd++;
+ } while (address && (address < end));
+ return 0;
+}
+
+static int remap_area_pages(unsigned long address, unsigned long phys_addr,
+ unsigned long size, unsigned long flags)
+{
+ int error;
+ pgd_t *dir;
+ unsigned long end = address + size;
+
+ phys_addr -= address;
+ dir = pgd_offset(&init_mm, address);
+ flush_cache_all();
+ if (address >= end)
+ BUG();
+ do {
+ pud_t *pud;
+ pmd_t *pmd;
+
+ error = -ENOMEM;
+ pud = pud_alloc(&init_mm, dir, address);
+ if (!pud)
+ break;
+ pmd = pmd_alloc(&init_mm, pud, address);
+ if (!pmd)
+ break;
+ if (remap_area_pmd(pmd, address, end - address,
+ phys_addr + address, flags))
+ break;
+ error = 0;
+ address = (address + PGDIR_SIZE) & PGDIR_MASK;
+ dir++;
+ } while (address && (address < end));
+ flush_tlb_all();
+ return error;
+}
+
+#define IS_MAPPABLE_UNCACHEABLE(addr) (addr < 0x20000000UL)
+
+/*
+ * Map some physical address range into the kernel address space.
+ */
+void __iomem *__ioremap(unsigned long phys_addr, unsigned long size,
+ unsigned long cacheflag)
+{
+ struct vm_struct *area;
+ unsigned long offset;
+ unsigned long last_addr;
+ void *addr;
+
+ /* Don't allow wraparound or zero size */
+ last_addr = phys_addr + size - 1;
+
+ if (!size || last_addr < phys_addr)
+ return NULL;
+
+ /* Don't allow anybody to remap normal RAM that we're using */
+ if (phys_addr > PHYS_OFFSET && phys_addr < virt_to_phys(high_memory)) {
+ char *t_addr, *t_end;
+ struct page *page;
+
+ t_addr = __va(phys_addr);
+ t_end = t_addr + (size - 1);
+ for (page = virt_to_page(t_addr);
+ page <= virt_to_page(t_end); page++)
+ if (!PageReserved(page))
+ return NULL;
+ }
+
+ /*
+ * Map uncached objects in the low part of address space to
+ * CONFIG_IO_REGION_BASE
+ */
+ if (IS_MAPPABLE_UNCACHEABLE(phys_addr) &&
+ IS_MAPPABLE_UNCACHEABLE(last_addr) &&
+ !(cacheflag & _PAGE_CACHED))
+ return (void __iomem *)(CONFIG_IO_REGION_BASE + phys_addr);
+
+ /* Mappings have to be page-aligned */
+ offset = phys_addr & ~PAGE_MASK;
+ phys_addr &= PAGE_MASK;
+ size = PAGE_ALIGN(last_addr + 1) - phys_addr;
+
+ /* Ok, go for it */
+ area = get_vm_area(size, VM_IOREMAP);
+ if (!area)
+ return NULL;
+ addr = area->addr;
+ if (remap_area_pages((unsigned long) addr, phys_addr, size,
+ cacheflag)) {
+ vunmap(addr);
+ return NULL;
+ }
+ return (void __iomem *) (offset + (char *)addr);
+}
+EXPORT_SYMBOL(__ioremap);
+
+/*
+ * __iounmap unmaps nearly everything, so be careful
+ * it doesn't free currently pointer/page tables anymore but it
+ * wasn't used anyway and might be added later.
+ */
+void __iounmap(void __iomem *addr)
+{
+ struct vm_struct *p;
+
+ if ((unsigned long) addr > CONFIG_IO_REGION_BASE)
+ return;
+
+ p = remove_vm_area((void *) (PAGE_MASK & (unsigned long __force) addr));
+ if (!p)
+ pr_err("iounmap: bad address %p\n", addr);
+ kfree(p);
+}
+EXPORT_SYMBOL(__iounmap);
--
1.8.3.2

2014-04-18 18:21:41

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 05/28] nios2: Traps exception handling

This patch contains traps exception handling.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/traps.h | 19 ++
arch/nios2/kernel/insnemu.S | 592 +++++++++++++++++++++++++++++++++++++++++
arch/nios2/kernel/traps.c | 185 +++++++++++++
3 files changed, 796 insertions(+)
create mode 100644 arch/nios2/include/asm/traps.h
create mode 100644 arch/nios2/kernel/insnemu.S
create mode 100644 arch/nios2/kernel/traps.c

diff --git a/arch/nios2/include/asm/traps.h b/arch/nios2/include/asm/traps.h
new file mode 100644
index 0000000..82a48473280d870e813c6a298e51166f869c6487
--- /dev/null
+++ b/arch/nios2/include/asm/traps.h
@@ -0,0 +1,19 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#ifndef _ASM_NIOS2_TRAPS_H
+#define _ASM_NIOS2_TRAPS_H
+
+#define TRAP_ID_SYSCALL 0
+
+#ifndef __ASSEMBLY__
+void _exception(int signo, struct pt_regs *regs, int code, unsigned long addr);
+#endif
+
+#endif /* _ASM_NIOS2_TRAPS_H */
diff --git a/arch/nios2/kernel/insnemu.S b/arch/nios2/kernel/insnemu.S
new file mode 100644
index 0000000..8a47a77d314538d2a1f82501c205970b2dea1f3c
--- /dev/null
+++ b/arch/nios2/kernel/insnemu.S
@@ -0,0 +1,592 @@
+/*
+ * Copyright (C) 2003-2013 Altera Corporation
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+
+#include <linux/linkage.h>
+#include <asm/entry.h>
+
+.set noat
+.set nobreak
+
+/*
+* Explicitly allow the use of r1 (the assembler temporary register)
+* within this code. This register is normally reserved for the use of
+* the compiler.
+*/
+
+ENTRY(instruction_trap)
+ ldw r1, PT_R1(sp) // Restore registers
+ ldw r2, PT_R2(sp)
+ ldw r3, PT_R3(sp)
+ ldw r4, PT_R4(sp)
+ ldw r5, PT_R5(sp)
+ ldw r6, PT_R6(sp)
+ ldw r7, PT_R7(sp)
+ ldw r8, PT_R8(sp)
+ ldw r9, PT_R9(sp)
+ ldw r10, PT_R10(sp)
+ ldw r11, PT_R11(sp)
+ ldw r12, PT_R12(sp)
+ ldw r13, PT_R13(sp)
+ ldw r14, PT_R14(sp)
+ ldw r15, PT_R15(sp)
+ ldw ra, PT_RA(sp)
+ ldw fp, PT_FP(sp)
+ ldw gp, PT_GP(sp)
+ ldw et, PT_ESTATUS(sp)
+ wrctl estatus, et
+ ldw ea, PT_EA(sp)
+ ldw et, PT_SP(sp) /* backup sp in et */
+
+ addi sp, sp, PT_REGS_SIZE
+
+ /* INSTRUCTION EMULATION
+ * ---------------------
+ *
+ * Nios II processors generate exceptions for unimplemented instructions.
+ * The routines below emulate these instructions. Depending on the
+ * processor core, the only instructions that might need to be emulated
+ * are div, divu, mul, muli, mulxss, mulxsu, and mulxuu.
+ *
+ * The emulations match the instructions, except for the following
+ * limitations:
+ *
+ * 1) The emulation routines do not emulate the use of the exception
+ * temporary register (et) as a source operand because the exception
+ * handler already has modified it.
+ *
+ * 2) The routines do not emulate the use of the stack pointer (sp) or
+ * the exception return address register (ea) as a destination because
+ * modifying these registers crashes the exception handler or the
+ * interrupted routine.
+ *
+ * Detailed Design
+ * ---------------
+ *
+ * The emulation routines expect the contents of integer registers r0-r31
+ * to be on the stack at addresses sp, 4(sp), 8(sp), ... 124(sp). The
+ * routines retrieve source operands from the stack and modify the
+ * destination register's value on the stack prior to the end of the
+ * exception handler. Then all registers except the destination register
+ * are restored to their previous values.
+ *
+ * The instruction that causes the exception is found at address -4(ea).
+ * The instruction's OP and OPX fields identify the operation to be
+ * performed.
+ *
+ * One instruction, muli, is an I-type instruction that is identified by
+ * an OP field of 0x24.
+ *
+ * muli AAAAA,BBBBB,IIIIIIIIIIIIIIII,-0x24-
+ * 27 22 6 0 <-- LSB of field
+ *
+ * The remaining emulated instructions are R-type and have an OP field
+ * of 0x3a. Their OPX fields identify them.
+ *
+ * R-type AAAAA,BBBBB,CCCCC,XXXXXX,NNNNN,-0x3a-
+ * 27 22 17 11 6 0 <-- LSB of field
+ *
+ *
+ * Opcode Encoding. muli is identified by its OP value. Then OPX & 0x02
+ * is used to differentiate between the division opcodes and the
+ * remaining multiplication opcodes.
+ *
+ * Instruction OP OPX OPX & 0x02
+ * ----------- ---- ---- ----------
+ * muli 0x24
+ * divu 0x3a 0x24 0
+ * div 0x3a 0x25 0
+ * mul 0x3a 0x27 != 0
+ * mulxuu 0x3a 0x07 != 0
+ * mulxsu 0x3a 0x17 != 0
+ * mulxss 0x3a 0x1f != 0
+ */
+
+
+ /*
+ * Save everything on the stack to make it easy for the emulation
+ * routines to retrieve the source register operands.
+ */
+
+ addi sp, sp, -128
+ stw zero, 0(sp) /* Save zero on stack to avoid special case for r0. */
+ stw r1, 4(sp)
+ stw r2, 8(sp)
+ stw r3, 12(sp)
+ stw r4, 16(sp)
+ stw r5, 20(sp)
+ stw r6, 24(sp)
+ stw r7, 28(sp)
+ stw r8, 32(sp)
+ stw r9, 36(sp)
+ stw r10, 40(sp)
+ stw r11, 44(sp)
+ stw r12, 48(sp)
+ stw r13, 52(sp)
+ stw r14, 56(sp)
+ stw r15, 60(sp)
+ stw r16, 64(sp)
+ stw r17, 68(sp)
+ stw r18, 72(sp)
+ stw r19, 76(sp)
+ stw r20, 80(sp)
+ stw r21, 84(sp)
+ stw r22, 88(sp)
+ stw r23, 92(sp)
+ /* Don't bother to save et. It's already been changed. */
+ rdctl r5, estatus
+ stw r5, 100(sp)
+
+ stw gp, 104(sp)
+ stw et, 108(sp) /* et containts previous sp value. */
+ stw fp, 112(sp)
+ stw ea, 116(sp)
+ stw ra, 120(sp)
+
+
+ /*
+ * Split the instruction into its fields. We need 4*A, 4*B, and 4*C as
+ * offsets to the stack pointer for access to the stored register values.
+ */
+ ldw r2,-4(ea) /* r2 = AAAAA,BBBBB,IIIIIIIIIIIIIIII,PPPPPP */
+ roli r3, r2, 7 /* r3 = BBB,IIIIIIIIIIIIIIII,PPPPPP,AAAAA,BB */
+ roli r4, r3, 3 /* r4 = IIIIIIIIIIIIIIII,PPPPPP,AAAAA,BBBBB */
+ roli r5, r4, 2 /* r5 = IIIIIIIIIIIIII,PPPPPP,AAAAA,BBBBB,II */
+ srai r4, r4, 16 /* r4 = (sign-extended) IMM16 */
+ roli r6, r5, 5 /* r6 = XXXX,NNNNN,PPPPPP,AAAAA,BBBBB,CCCCC,XX */
+ andi r2, r2, 0x3f /* r2 = 00000000000000000000000000,PPPPPP */
+ andi r3, r3, 0x7c /* r3 = 0000000000000000000000000,AAAAA,00 */
+ andi r5, r5, 0x7c /* r5 = 0000000000000000000000000,BBBBB,00 */
+ andi r6, r6, 0x7c /* r6 = 0000000000000000000000000,CCCCC,00 */
+
+ /* Now
+ * r2 = OP
+ * r3 = 4*A
+ * r4 = IMM16 (sign extended)
+ * r5 = 4*B
+ * r6 = 4*C
+ */
+
+ /*
+ * Get the operands.
+ *
+ * It is necessary to check for muli because it uses an I-type
+ * instruction format, while the other instructions are have an R-type
+ * format.
+ *
+ * Prepare for either multiplication or division loop.
+ * They both loop 32 times.
+ */
+ movi r14, 32
+
+ add r3, r3, sp /* r3 = address of A-operand. */
+ ldw r3, 0(r3) /* r3 = A-operand. */
+ movi r7, 0x24 /* muli opcode (I-type instruction format) */
+ beq r2, r7, mul_immed /* muli doesn't use the B register as a source */
+
+ add r5, r5, sp /* r5 = address of B-operand. */
+ ldw r5, 0(r5) /* r5 = B-operand. */
+ /* r4 = SSSSSSSSSSSSSSSS,-----IMM16------ */
+ /* IMM16 not needed, align OPX portion */
+ /* r4 = SSSSSSSSSSSSSSSS,CCCCC,-OPX--,00000 */
+ srli r4, r4, 5 /* r4 = 00000,SSSSSSSSSSSSSSSS,CCCCC,-OPX-- */
+ andi r4, r4, 0x3f /* r4 = 00000000000000000000000000,-OPX-- */
+
+ /* Now
+ * r2 = OP
+ * r3 = src1
+ * r5 = src2
+ * r4 = OPX (no longer can be muli)
+ * r6 = 4*C
+ */
+
+
+ /*
+ * Multiply or Divide?
+ */
+ andi r7, r4, 0x02 /* For R-type multiply instructions,
+ OPX & 0x02 != 0 */
+ bne r7, zero, multiply
+
+
+ /* DIVISION
+ *
+ * Divide an unsigned dividend by an unsigned divisor using
+ * a shift-and-subtract algorithm. The example below shows
+ * 43 div 7 = 6 for 8-bit integers. This classic algorithm uses a
+ * single register to store both the dividend and the quotient,
+ * allowing both values to be shifted with a single instruction.
+ *
+ * remainder dividend:quotient
+ * --------- -----------------
+ * initialize 00000000 00101011:
+ * shift 00000000 0101011:_
+ * remainder >= divisor? no 00000000 0101011:0
+ * shift 00000000 101011:0_
+ * remainder >= divisor? no 00000000 101011:00
+ * shift 00000001 01011:00_
+ * remainder >= divisor? no 00000001 01011:000
+ * shift 00000010 1011:000_
+ * remainder >= divisor? no 00000010 1011:0000
+ * shift 00000101 011:0000_
+ * remainder >= divisor? no 00000101 011:00000
+ * shift 00001010 11:00000_
+ * remainder >= divisor? yes 00001010 11:000001
+ * remainder -= divisor - 00000111
+ * ----------
+ * 00000011 11:000001
+ * shift 00000111 1:000001_
+ * remainder >= divisor? yes 00000111 1:0000011
+ * remainder -= divisor - 00000111
+ * ----------
+ * 00000000 1:0000011
+ * shift 00000001 :0000011_
+ * remainder >= divisor? no 00000001 :00000110
+ *
+ * The quotient is 00000110.
+ */
+
+divide:
+ /*
+ * Prepare for division by assuming the result
+ * is unsigned, and storing its "sign" as 0.
+ */
+ movi r17, 0
+
+
+ /* Which division opcode? */
+ xori r7, r4, 0x25 /* OPX of div */
+ bne r7, zero, unsigned_division
+
+
+ /*
+ * OPX is div. Determine and store the sign of the quotient.
+ * Then take the absolute value of both operands.
+ */
+ xor r17, r3, r5 /* MSB contains sign of quotient */
+ bge r3,zero,dividend_is_nonnegative
+ sub r3, zero, r3 /* -r3 */
+dividend_is_nonnegative:
+ bge r5, zero, divisor_is_nonnegative
+ sub r5, zero, r5 /* -r5 */
+divisor_is_nonnegative:
+
+
+unsigned_division:
+ /* Initialize the unsigned-division loop. */
+ movi r13, 0 /* remainder = 0 */
+
+ /* Now
+ * r3 = dividend : quotient
+ * r4 = 0x25 for div, 0x24 for divu
+ * r5 = divisor
+ * r13 = remainder
+ * r14 = loop counter (already initialized to 32)
+ * r17 = MSB contains sign of quotient
+ */
+
+
+ /*
+ * for (count = 32; count > 0; --count)
+ * {
+ */
+divide_loop:
+
+ /*
+ * Division:
+ *
+ * (remainder:dividend:quotient) <<= 1;
+ */
+ slli r13, r13, 1
+ cmplt r7, r3, zero /* r7 = MSB of r3 */
+ or r13, r13, r7
+ slli r3, r3, 1
+
+
+ /*
+ * if (remainder >= divisor)
+ * {
+ * set LSB of quotient
+ * remainder -= divisor;
+ * }
+ */
+ bltu r13, r5, div_skip
+ ori r3, r3, 1
+ sub r13, r13, r5
+div_skip:
+
+ /*
+ * }
+ */
+ subi r14, r14, 1
+ bne r14, zero, divide_loop
+
+
+ /* Now
+ * r3 = quotient
+ * r4 = 0x25 for div, 0x24 for divu
+ * r6 = 4*C
+ * r17 = MSB contains sign of quotient
+ */
+
+
+ /*
+ * Conditionally negate signed quotient. If quotient is unsigned,
+ * the sign already is initialized to 0.
+ */
+ bge r17, zero, quotient_is_nonnegative
+ sub r3, zero, r3 /* -r3 */
+ quotient_is_nonnegative:
+
+
+ /*
+ * Final quotient is in r3.
+ */
+ add r6, r6, sp
+ stw r3, 0(r6) /* write quotient to stack */
+ br restore_registers
+
+
+
+
+ /* MULTIPLICATION
+ *
+ * A "product" is the number that one gets by summing a "multiplicand"
+ * several times. The "multiplier" specifies the number of copies of the
+ * multiplicand that are summed.
+ *
+ * Actual multiplication algorithms don't use repeated addition, however.
+ * Shift-and-add algorithms get the same answer as repeated addition, and
+ * they are faster. To compute the lower half of a product (pppp below)
+ * one shifts the product left before adding in each of the partial
+ * products (a * mmmm) through (d * mmmm).
+ *
+ * To compute the upper half of a product (PPPP below), one adds in the
+ * partial products (d * mmmm) through (a * mmmm), each time following
+ * the add by a right shift of the product.
+ *
+ * mmmm
+ * * abcd
+ * ------
+ * #### = d * mmmm
+ * #### = c * mmmm
+ * #### = b * mmmm
+ * #### = a * mmmm
+ * --------
+ * PPPPpppp
+ *
+ * The example above shows 4 partial products. Computing actual Nios II
+ * products requires 32 partials.
+ *
+ * It is possible to compute the result of mulxsu from the result of
+ * mulxuu because the only difference between the results of these two
+ * opcodes is the value of the partial product associated with the sign
+ * bit of rA.
+ *
+ * mulxsu = mulxuu - (rA < 0) ? rB : 0;
+ *
+ * It is possible to compute the result of mulxss from the result of
+ * mulxsu because the only difference between the results of these two
+ * opcodes is the value of the partial product associated with the sign
+ * bit of rB.
+ *
+ * mulxss = mulxsu - (rB < 0) ? rA : 0;
+ *
+ */
+
+mul_immed:
+ /* Opcode is muli. Change it into mul for remainder of algorithm. */
+ mov r6, r5 /* Field B is dest register, not field C. */
+ mov r5, r4 /* Field IMM16 is src2, not field B. */
+ movi r4, 0x27 /* OPX of mul is 0x27 */
+
+multiply:
+ /* Initialize the multiplication loop. */
+ movi r9, 0 /* mul_product = 0 */
+ movi r10, 0 /* mulxuu_product = 0 */
+ mov r11, r5 /* save original multiplier for mulxsu and mulxss */
+ mov r12, r5 /* mulxuu_multiplier (will be shifted) */
+ movi r16, 1 /* used to create "rori B,A,1" from "ror B,A,r16" */
+
+ /* Now
+ * r3 = multiplicand
+ * r5 = mul_multiplier
+ * r6 = 4 * dest_register (used later as offset to sp)
+ * r7 = temp
+ * r9 = mul_product
+ * r10 = mulxuu_product
+ * r11 = original multiplier
+ * r12 = mulxuu_multiplier
+ * r14 = loop counter (already initialized)
+ * r16 = 1
+ */
+
+
+ /*
+ * for (count = 32; count > 0; --count)
+ * {
+ */
+multiply_loop:
+
+ /*
+ * mul_product <<= 1;
+ * lsb = multiplier & 1;
+ */
+ slli r9, r9, 1
+ andi r7, r12, 1
+
+ /*
+ * if (lsb == 1)
+ * {
+ * mulxuu_product += multiplicand;
+ * }
+ */
+ beq r7, zero, mulx_skip
+ add r10, r10, r3
+ cmpltu r7, r10, r3 /* Save the carry from the MSB of mulxuu_product. */
+ ror r7, r7, r16 /* r7 = 0x80000000 on carry, or else 0x00000000 */
+mulx_skip:
+
+ /*
+ * if (MSB of mul_multiplier == 1)
+ * {
+ * mul_product += multiplicand;
+ * }
+ */
+ bge r5, zero, mul_skip
+ add r9, r9, r3
+mul_skip:
+
+ /*
+ * mulxuu_product >>= 1; logical shift
+ * mul_multiplier <<= 1; done with MSB
+ * mulx_multiplier >>= 1; done with LSB
+ */
+ srli r10, r10, 1
+ or r10, r10, r7 /* OR in the saved carry bit. */
+ slli r5, r5, 1
+ srli r12, r12, 1
+
+
+ /*
+ * }
+ */
+ subi r14, r14, 1
+ bne r14, zero, multiply_loop
+
+
+ /*
+ * Multiply emulation loop done.
+ */
+
+ /* Now
+ * r3 = multiplicand
+ * r4 = OPX
+ * r6 = 4 * dest_register (used later as offset to sp)
+ * r7 = temp
+ * r9 = mul_product
+ * r10 = mulxuu_product
+ * r11 = original multiplier
+ */
+
+
+ /* Calculate address for result from 4 * dest_register */
+ add r6, r6, sp
+
+
+ /*
+ * Select/compute the result based on OPX.
+ */
+
+
+ /* OPX == mul? Then store. */
+ xori r7, r4, 0x27
+ beq r7, zero, store_product
+
+ /* It's one of the mulx.. opcodes. Move over the result. */
+ mov r9, r10
+
+ /* OPX == mulxuu? Then store. */
+ xori r7, r4, 0x07
+ beq r7, zero, store_product
+
+ /* Compute mulxsu
+ *
+ * mulxsu = mulxuu - (rA < 0) ? rB : 0;
+ */
+ bge r3, zero, mulxsu_skip
+ sub r9, r9, r11
+mulxsu_skip:
+
+ /* OPX == mulxsu? Then store. */
+ xori r7, r4, 0x17
+ beq r7, zero, store_product
+
+ /* Compute mulxss
+ *
+ * mulxss = mulxsu - (rB < 0) ? rA : 0;
+ */
+ bge r11,zero,mulxss_skip
+ sub r9, r9, r3
+mulxss_skip:
+ /* At this point, assume that OPX is mulxss, so store*/
+
+
+store_product:
+ stw r9, 0(r6)
+
+
+restore_registers:
+ /* No need to restore r0. */
+ ldw r5, 100(sp)
+ wrctl estatus, r5
+
+ ldw r1, 4(sp)
+ ldw r2, 8(sp)
+ ldw r3, 12(sp)
+ ldw r4, 16(sp)
+ ldw r5, 20(sp)
+ ldw r6, 24(sp)
+ ldw r7, 28(sp)
+ ldw r8, 32(sp)
+ ldw r9, 36(sp)
+ ldw r10, 40(sp)
+ ldw r11, 44(sp)
+ ldw r12, 48(sp)
+ ldw r13, 52(sp)
+ ldw r14, 56(sp)
+ ldw r15, 60(sp)
+ ldw r16, 64(sp)
+ ldw r17, 68(sp)
+ ldw r18, 72(sp)
+ ldw r19, 76(sp)
+ ldw r20, 80(sp)
+ ldw r21, 84(sp)
+ ldw r22, 88(sp)
+ ldw r23, 92(sp)
+ /* Does not need to restore et */
+ ldw gp, 104(sp)
+
+ ldw fp, 112(sp)
+ ldw ea, 116(sp)
+ ldw ra, 120(sp)
+ ldw sp, 108(sp) /* last restore sp */
+ eret
+
+.set at
+.set break
diff --git a/arch/nios2/kernel/traps.c b/arch/nios2/kernel/traps.c
new file mode 100644
index 0000000..53ddf8d
--- /dev/null
+++ b/arch/nios2/kernel/traps.c
@@ -0,0 +1,185 @@
+/*
+ * Hardware exception handling
+ *
+ * Copyright (C) 2010 Tobias Klauser <[email protected]>
+ * Copyright (C) 2004 Microtronix Datacom Ltd.
+ * Copyright (C) 2001 Vic Phillips
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License. See the file COPYING in the main directory of this
+ * archive for more details.
+ */
+
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/signal.h>
+#include <linux/export.h>
+#include <linux/mm.h>
+#include <linux/ptrace.h>
+
+#include <asm/traps.h>
+#include <asm/sections.h>
+
+
+static DEFINE_SPINLOCK(die_lock);
+
+void die(const char *str, struct pt_regs *regs, long err)
+{
+ console_verbose();
+ spin_lock_irq(&die_lock);
+ pr_warn("Oops: %s, sig: %ld\n", str, err);
+ show_regs(regs);
+ spin_unlock_irq(&die_lock);
+ /*
+ * do_exit() should take care of panic'ing from an interrupt
+ * context so we don't handle it here
+ */
+ do_exit(err);
+}
+
+void _exception(int signo, struct pt_regs *regs, int code, unsigned long addr)
+{
+ siginfo_t info;
+
+ if (!user_mode(regs))
+ die("Exception in kernel mode", regs, signo);
+
+ info.si_signo = signo;
+ info.si_errno = 0;
+ info.si_code = code;
+ info.si_addr = (void __user *) addr;
+ force_sig_info(signo, &info, current);
+}
+
+/*
+ * The show_stack is an external API which we do not use ourselves.
+ */
+
+int kstack_depth_to_print = 48;
+
+void show_stack(struct task_struct *task, unsigned long *stack)
+{
+ unsigned long *endstack, addr;
+ int i;
+
+ if (!stack) {
+ if (task)
+ stack = (unsigned long *)task->thread.ksp;
+ else
+ stack = (unsigned long *)&stack;
+ }
+
+ addr = (unsigned long) stack;
+ endstack = (unsigned long *) PAGE_ALIGN(addr);
+
+ pr_emerg("Stack from %08lx:", (unsigned long)stack);
+ for (i = 0; i < kstack_depth_to_print; i++) {
+ if (stack + 1 > endstack)
+ break;
+ if (i % 8 == 0)
+ pr_emerg("\n ");
+ pr_emerg(" %08lx", *stack++);
+ }
+
+ pr_emerg("\nCall Trace:");
+ i = 0;
+ while (stack + 1 <= endstack) {
+ addr = *stack++;
+ /*
+ * If the address is either in the text segment of the
+ * kernel, or in the region which contains vmalloc'ed
+ * memory, it *may* be the address of a calling
+ * routine; if so, print it so that someone tracing
+ * down the cause of the crash will be able to figure
+ * out the call path that was taken.
+ */
+ if (((addr >= (unsigned long) _stext) &&
+ (addr <= (unsigned long) _etext))) {
+ if (i % 4 == 0)
+ pr_emerg("\n ");
+ pr_emerg(" [<%08lx>]", addr);
+ i++;
+ }
+ }
+ pr_emerg("\n");
+}
+
+void __init trap_init(void)
+{
+ /* Nothing to do here */
+}
+
+/* Breakpoint handler */
+asmlinkage void breakpoint_c(struct pt_regs *fp)
+{
+ /*
+ * The breakpoint entry code has moved the PC on by 4 bytes, so we must
+ * move it back. This could be done on the host but we do it here
+ * because monitor.S of JTAG gdbserver does it too.
+ */
+ fp->ea -= 4;
+ _exception(SIGTRAP, fp, TRAP_BRKPT, fp->ea);
+}
+
+#ifndef CONFIG_ALIGNMENT_TRAP
+/* Alignment exception handler */
+asmlinkage void handle_unaligned_c(struct pt_regs *fp, int cause)
+{
+ unsigned long addr = RDCTL(CTL_BADADDR);
+
+ cause >>= 2;
+ fp->ea -= 4;
+
+ if (fixup_exception(fp))
+ return;
+
+ if (!user_mode(fp)) {
+ pr_alert("Unaligned access from kernel mode, this might be a hardware\n");
+ pr_alert("problem, dump registers and restart the instruction\n");
+ pr_alert(" BADADDR 0x%08lx\n", addr);
+ pr_alert(" cause %d\n", cause);
+ pr_alert(" op-code 0x%08lx\n", *(unsigned long *)(fp->ea));
+ show_regs(fp);
+ return;
+ }
+
+ _exception(SIGBUS, fp, BUS_ADRALN, addr);
+}
+#endif /* CONFIG_ALIGNMENT_TRAP */
+
+/* Illegal instruction handler */
+asmlinkage void handle_illegal_c(struct pt_regs *fp)
+{
+ fp->ea -= 4;
+ _exception(SIGILL, fp, ILL_ILLOPC, fp->ea);
+}
+
+/* Supervisor instruction handler */
+asmlinkage void handle_supervisor_instr(struct pt_regs *fp)
+{
+ fp->ea -= 4;
+ _exception(SIGILL, fp, ILL_PRVOPC, fp->ea);
+}
+
+/* Division error handler */
+asmlinkage void handle_diverror_c(struct pt_regs *fp)
+{
+ fp->ea -= 4;
+ _exception(SIGFPE, fp, FPE_INTDIV, fp->ea);
+}
+
+/* Unhandled exception handler */
+asmlinkage void unhandled_exception(struct pt_regs *regs, int cause)
+{
+ unsigned long addr = RDCTL(CTL_BADADDR);
+
+ cause /= 4;
+
+ pr_emerg("Unhandled exception #%d in %s mode (badaddr=0x%08lx)\n",
+ cause, user_mode(regs) ? "user" : "kernel", addr);
+
+ regs->ea -= 4;
+ show_regs(regs);
+
+ pr_emerg("opcode: 0x%08lx\n", *(unsigned long *)(regs->ea));
+}
--
1.8.3.2

2014-04-18 18:22:32

by Ley Foon Tan

[permalink] [raw]
Subject: [PATCH 14/28] nios2: DMA mapping API

This patch adds support for the DMA mapping API.

Signed-off-by: Ley Foon Tan <[email protected]>
---
arch/nios2/include/asm/dma-mapping.h | 113 +++++++++++++++++++++
arch/nios2/mm/dma-mapping.c | 186 +++++++++++++++++++++++++++++++++++
2 files changed, 299 insertions(+)
create mode 100644 arch/nios2/include/asm/dma-mapping.h
create mode 100644 arch/nios2/mm/dma-mapping.c

diff --git a/arch/nios2/include/asm/dma-mapping.h b/arch/nios2/include/asm/dma-mapping.h
new file mode 100644
index 0000000..28db1171247453e8c40f439e6752bf2a51937387
--- /dev/null
+++ b/arch/nios2/include/asm/dma-mapping.h
@@ -0,0 +1,113 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License. See the file COPYING in the main directory of this
+ * archive for more details.
+ */
+
+#ifndef _ASM_NIOS2_DMA_MAPPING_H
+#define _ASM_NIOS2_DMA_MAPPING_H
+
+#include <linux/scatterlist.h>
+#include <linux/cache.h>
+#include <asm/cacheflush.h>
+
+static inline void __dma_sync(void *vaddr, size_t size,
+ enum dma_data_direction direction)
+{
+ switch (direction) {
+ case DMA_FROM_DEVICE: /* invalidate cache */
+ invalidate_dcache_range((unsigned long)vaddr,
+ (unsigned long)(vaddr + size));
+ break;
+ case DMA_TO_DEVICE: /* flush and invalidate cache */
+ case DMA_BIDIRECTIONAL:
+ flush_dcache_range((unsigned long)vaddr,
+ (unsigned long)(vaddr + size));
+ break;
+ default:
+ BUG();
+ }
+}
+
+#define dma_alloc_noncoherent(d, s, h, f) dma_alloc_coherent(d, s, h, f)
+#define dma_free_noncoherent(d, s, v, h) dma_free_coherent(d, s, v, h)
+
+void *dma_alloc_coherent(struct device *dev, size_t size,
+ dma_addr_t *dma_handle, gfp_t flag);
+
+void dma_free_coherent(struct device *dev, size_t size,
+ void *vaddr, dma_addr_t dma_handle);
+
+static inline dma_addr_t dma_map_single(struct device *dev, void *ptr,
+ size_t size,
+ enum dma_data_direction direction)
+{
+ __dma_sync(ptr, size, direction);
+ return virt_to_phys(ptr);
+}
+
+static inline void dma_unmap_single(struct device *dev, dma_addr_t dma_addr,
+ size_t size, enum dma_data_direction direction)
+{
+}
+
+extern int dma_map_sg(struct device *dev, struct scatterlist *sg, int nents,
+ enum dma_data_direction direction);
+extern dma_addr_t dma_map_page(struct device *dev, struct page *page,
+ unsigned long offset, size_t size, enum dma_data_direction direction);
+extern void dma_unmap_page(struct device *dev, dma_addr_t dma_address,
+ size_t size, enum dma_data_direction direction);
+extern void dma_unmap_sg(struct device *dev, struct scatterlist *sg,
+ int nhwentries, enum dma_data_direction direction);
+extern void dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle,
+ size_t size, enum dma_data_direction direction);
+extern void dma_sync_single_for_device(struct device *dev,
+ dma_addr_t dma_handle, size_t size, enum dma_data_direction direction);
+extern void dma_sync_single_range_for_cpu(struct device *dev,
+ dma_addr_t dma_handle, unsigned long offset, size_t size,
+ enum dma_data_direction direction);
+extern void dma_sync_single_range_for_device(struct device *dev,
+ dma_addr_t dma_handle, unsigned long offset, size_t size,
+ enum dma_data_direction direction);
+extern void dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg,
+ int nelems, enum dma_data_direction direction);
+extern void dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg,
+ int nelems, enum dma_data_direction direction);
+
+static inline int dma_supported(struct device *dev, u64 mask)
+{
+ /* we fall back to GFP_DMA when the mask isn't all 1s,
+ * so we can't guarantee allocations that must be
+ * within a tighter range than GFP_DMA.
+ */
+ if (mask < 0x00ffffff)
+ return 0;
+
+ return 1;
+}
+
+static inline int dma_set_mask(struct device *dev, u64 mask)
+{
+ if (!dev->dma_mask || !dma_supported(dev, mask))
+ return -EIO;
+
+ *dev->dma_mask = mask;
+
+ return 0;
+}
+
+static inline int dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
+{
+ return 0;
+}
+
+static inline void dma_cache_sync(struct device *dev, void *vaddr, size_t size,
+ enum dma_data_direction direction)
+{
+ __dma_sync(vaddr, size, direction);
+}
+
+#endif /* _ASM_NIOS2_DMA_MAPPING_H */
diff --git a/arch/nios2/mm/dma-mapping.c b/arch/nios2/mm/dma-mapping.c
new file mode 100644
index 0000000..a9bafe9f5b5c65442e5fafa9b043e529cf6b6e54
--- /dev/null
+++ b/arch/nios2/mm/dma-mapping.c
@@ -0,0 +1,186 @@
+/*
+ * Copyright (C) 2011 Tobias Klauser <[email protected]>
+ * Copyright (C) 2009 Wind River Systems Inc
+ * Implemented by [email protected] and [email protected]
+ *
+ * Based on DMA code from MIPS.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+
+#include <linux/types.h>
+#include <linux/mm.h>
+#include <linux/export.h>
+#include <linux/string.h>
+#include <linux/scatterlist.h>
+#include <linux/dma-mapping.h>
+#include <linux/io.h>
+#include <linux/cache.h>
+#include <asm/cacheflush.h>
+
+
+void *dma_alloc_coherent(struct device *dev, size_t size,
+ dma_addr_t *dma_handle, gfp_t gfp)
+{
+ void *ret;
+
+ /* ignore region specifiers */
+ gfp &= ~(__GFP_DMA | __GFP_HIGHMEM);
+
+ /* optimized page clearing */
+ gfp |= __GFP_ZERO;
+
+ if (dev == NULL || (dev->coherent_dma_mask < 0xffffffff))
+ gfp |= GFP_DMA;
+
+ ret = (void *) __get_free_pages(gfp, get_order(size));
+ if (ret != NULL) {
+ *dma_handle = virt_to_phys(ret);
+ flush_dcache_range((unsigned long) ret,
+ (unsigned long) ret + size);
+ ret = UNCAC_ADDR(ret);
+ }
+
+ return ret;
+}
+EXPORT_SYMBOL(dma_alloc_coherent);
+
+void dma_free_coherent(struct device *dev, size_t size, void *vaddr,
+ dma_addr_t dma_handle)
+{
+ unsigned long addr = (unsigned long) CAC_ADDR((unsigned long) vaddr);
+ free_pages(addr, get_order(size));
+}
+EXPORT_SYMBOL(dma_free_coherent);
+
+int dma_map_sg(struct device *dev, struct scatterlist *sg, int nents,
+ enum dma_data_direction direction)
+{
+ int i;
+
+ BUG_ON(!valid_dma_direction(direction));
+
+ for_each_sg(sg, sg, nents, i) {
+ void *addr;
+
+ addr = sg_virt(sg);
+ if (addr) {
+ __dma_sync(addr, sg->length, direction);
+ sg->dma_address = sg_phys(sg);
+ }
+ }
+
+ return nents;
+}
+EXPORT_SYMBOL(dma_map_sg);
+
+dma_addr_t dma_map_page(struct device *dev, struct page *page,
+ unsigned long offset, size_t size,
+ enum dma_data_direction direction)
+{
+ void *addr;
+
+ BUG_ON(!valid_dma_direction(direction));
+
+ addr = page_address(page) + offset;
+ __dma_sync(addr, size, direction);
+
+ return page_to_phys(page) + offset;
+}
+EXPORT_SYMBOL(dma_map_page);
+
+void dma_unmap_page(struct device *dev, dma_addr_t dma_address, size_t size,
+ enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+
+ if (direction != DMA_TO_DEVICE)
+ __dma_sync(phys_to_virt(dma_address), size, direction);
+}
+EXPORT_SYMBOL(dma_unmap_page);
+
+void dma_unmap_sg(struct device *dev, struct scatterlist *sg, int nhwentries,
+ enum dma_data_direction direction)
+{
+ void *addr;
+ int i;
+
+ BUG_ON(!valid_dma_direction(direction));
+
+ if (direction == DMA_TO_DEVICE)
+ return;
+
+ for_each_sg(sg, sg, nhwentries, i) {
+ addr = sg_virt(sg);
+ if (addr)
+ __dma_sync(addr, sg->length, direction);
+ }
+}
+EXPORT_SYMBOL(dma_unmap_sg);
+
+void dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle,
+ size_t size, enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+
+ __dma_sync(phys_to_virt(dma_handle), size, direction);
+}
+EXPORT_SYMBOL(dma_sync_single_for_cpu);
+
+void dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle,
+ size_t size, enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+
+ __dma_sync(phys_to_virt(dma_handle), size, direction);
+}
+EXPORT_SYMBOL(dma_sync_single_for_device);
+
+void dma_sync_single_range_for_cpu(struct device *dev, dma_addr_t dma_handle,
+ unsigned long offset, size_t size,
+ enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+
+ __dma_sync(phys_to_virt(dma_handle), size, direction);
+}
+EXPORT_SYMBOL(dma_sync_single_range_for_cpu);
+
+void dma_sync_single_range_for_device(struct device *dev, dma_addr_t dma_handle,
+ unsigned long offset, size_t size,
+ enum dma_data_direction direction)
+{
+ BUG_ON(!valid_dma_direction(direction));
+
+ __dma_sync(phys_to_virt(dma_handle), size, direction);
+}
+EXPORT_SYMBOL(dma_sync_single_range_for_device);
+
+void dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, int nelems,
+ enum dma_data_direction direction)
+{
+ int i;
+
+ BUG_ON(!valid_dma_direction(direction));
+
+ /* Make sure that gcc doesn't leave the empty loop body. */
+ for_each_sg(sg, sg, nelems, i)
+ __dma_sync(sg_virt(sg), sg->length, direction);
+}
+EXPORT_SYMBOL(dma_sync_sg_for_cpu);
+
+void dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg,
+ int nelems, enum dma_data_direction direction)
+{
+ int i;
+
+ BUG_ON(!valid_dma_direction(direction));
+
+ /* Make sure that gcc doesn't leave the empty loop body. */
+ for_each_sg(sg, sg, nelems, i)
+ __dma_sync(sg_virt(sg), sg->length, direction);
+
+}
+EXPORT_SYMBOL(dma_sync_sg_for_device);
--
1.8.3.2

2014-04-18 18:38:25

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

Okay, I already submitted the remaining 4 patches.
Thanks.

On Sat, Apr 19, 2014 at 12:49 AM, Ley Foon Tan <[email protected]> wrote:
> Hi all
>
> I just found that 4 patches (03/28, 05/28,07/28 and 14/28) are missing
> here, because the issue in Git v1.8.3.2. But, the cover letter is correct.
> Do I need to resend the whole series (28 patches) or just the missing 4 patches?
> Thanks.
>
> Regards.
>
> On Fri, Apr 18, 2014 at 8:26 PM, Ley Foon Tan <[email protected]> wrote:
>> This patchset adds the Linux kernel port for Nios II processor from Altera.
>> The nios2 Linux port follows the guidance for new architecture ports using
>> generic headers (including unistd.h).
>>
>> About Nios II Cores
>> -------------------
>> Nios II is a 32-bit embedded-processor architecture designed specifically for the
>> Altera family of FPGAs.
>> More information is available at http://www.altera.com/devices/processor/nios2/ni2-index.html
>>
>> Instruction set and architecture overview documents can be found on the
>> following page:
>> http://www.altera.com/literature/lit-nio2.jsp
>>
>> Nios2 GCC port is in mainline and and will be in the FSF 4.9 release.
>>
>> These patches are also available on this branch:
>> git://git.rocketboards.org/linux-socfpga-next.git nios2-upstream
>>
>> Regards,
>> Ley Foon
>> --
>> 1.8.3.2
>>

2014-04-18 19:16:19

by Paul Bolle

[permalink] [raw]
Subject: Re: [PATCH 01/28] nios2: Build infrastructure

On Fri, 2014-04-18 at 20:26 +0800, Ley Foon Tan wrote:
> diff --git a/arch/nios2/Kconfig b/arch/nios2/Kconfig
>[...]
> +config GENERIC_FIND_NEXT_BIT
> + def_bool y

This one is not needed, see commit 63e424c84429 ("arch: remove
CONFIG_GENERIC_FIND_{NEXT_BIT,BIT_LE,LAST_BIT}").

> +config NO_IOPORT
> + def_bool y

Did you mean NO_IOPORT_MAP?


Paul Bolle

2014-04-18 19:41:50

by Paul Bolle

[permalink] [raw]
Subject: Re: [PATCH 01/28] nios2: Build infrastructure

On Fri, 2014-04-18 at 20:26 +0800, Ley Foon Tan wrote:
> diff --git a/arch/nios2/kernel/Makefile b/arch/nios2/kernel/Makefile
[...]
> +obj-$(CONFIG_CONSOLE) += console.o

CONFIG_CONSOLE seems never to be defined. And there appears to be no
arch/nios2/console.c or .../console.S anyway.


Paul Bolle

2014-04-18 20:49:12

by H. Peter Anvin

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 04/18/2014 05:26 AM, Ley Foon Tan wrote:
> This patchset adds the Linux kernel port for Nios II processor from Altera.
> The nios2 Linux port follows the guidance for new architecture ports using
> generic headers (including unistd.h).

Did the generic headers ever get updated to match Linus' guidance that
any new architecture ports should use a 64-bit time_t?

-hpa

2014-04-19 15:33:00

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On Friday 18 April 2014, H. Peter Anvin wrote:
> On 04/18/2014 05:26 AM, Ley Foon Tan wrote:
> > This patchset adds the Linux kernel port for Nios II processor from Altera.
> > The nios2 Linux port follows the guidance for new architecture ports using
> > generic headers (including unistd.h).
>
> Did the generic headers ever get updated to match Linus' guidance that
> any new architecture ports should use a 64-bit time_t?

No, unfortunately not. With my rule that every architecture that gets
added needs to clean up asm-generic some more, to make it easier to add
the next one, we should probably do for nios2.

Arnd

2014-04-19 16:05:59

by Pavel Machek

[permalink] [raw]
Subject: Re: [PATCH 09/28] nios2: Page table management

Hi!

> This patch adds support for page table management.
>
> Signed-off-by: Ley Foon Tan <[email protected]>

> +/*
> + * Initialize a new pgd / pmd table with invalid pointers.
> + */
> +static void pgd_init(pgd_t *pgd)
> +{
> + unsigned long *p = (unsigned long *) pgd;
> + int i;
> +
> + for (i = 0; i < USER_PTRS_PER_PGD; i += 8) {
> + p[i + 0] = (unsigned long) invalid_pte_table;
> + p[i + 1] = (unsigned long) invalid_pte_table;
> + p[i + 2] = (unsigned long) invalid_pte_table;
> + p[i + 3] = (unsigned long) invalid_pte_table;
> + p[i + 4] = (unsigned long) invalid_pte_table;
> + p[i + 5] = (unsigned long) invalid_pte_table;
> + p[i + 6] = (unsigned long) invalid_pte_table;
> + p[i + 7] = (unsigned long) invalid_pte_table;
> + }

Umm. Manual loop unrolling. Does it really improve performance? Is the
code hot enough to warrant the uglyness?

Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

2014-04-19 16:09:52

by Pavel Machek

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

On Fri 2014-04-18 20:26:59, Ley Foon Tan wrote:
> This patch adds support for system calls from userspaces. It uses the
> asm-generic/unistd.h definitions with architecture spcific syscall. The sys_call_table
> is just an array defined in a C file and it contains pointers to the syscall functions.
>
> Signed-off-by: Ley Foon Tan <[email protected]>


> +/* Additional Nios II specific syscalls. */
> +#define __NR_cacheflush (__NR_arch_specific_syscall)
> +__SYSCALL(__NR_cacheflush, sys_cacheflush)

I guess you should Cc: Michael Kerrisk on this one.

Also... explanation why you need this syscall while other
architectures live happily without it would be nice.

> +/* sys_cacheflush -- flush the processor cache. */
> +asmlinkage int sys_cacheflush(unsigned long addr, unsigned long len,
> + unsigned int op)
> +{
> + struct vm_area_struct *vma;
> +
> + if (len == 0)
> + return 0;

op is being ignored? You should remove it... or if (op) return -EINVAL
if you want future extensions.

--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

2014-04-19 19:29:36

by Richard Weinberger

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

On Fri, Apr 18, 2014 at 2:27 PM, Ley Foon Tan <[email protected]> wrote:
> This patch adds support for signal handling.
>
> Signed-off-by: Ley Foon Tan <[email protected]>
> ---
> arch/nios2/include/asm/signal.h | 22 ++
> arch/nios2/include/asm/ucontext.h | 34 +++
> arch/nios2/include/uapi/asm/sigcontext.h | 30 +++
> arch/nios2/include/uapi/asm/signal.h | 23 ++
> arch/nios2/kernel/signal.c | 364 +++++++++++++++++++++++++++++++
> 5 files changed, 473 insertions(+)
> create mode 100644 arch/nios2/include/asm/signal.h
> create mode 100644 arch/nios2/include/asm/ucontext.h
> create mode 100644 arch/nios2/include/uapi/asm/sigcontext.h
> create mode 100644 arch/nios2/include/uapi/asm/signal.h
> create mode 100644 arch/nios2/kernel/signal.c
>
> diff --git a/arch/nios2/include/asm/signal.h b/arch/nios2/include/asm/signal.h
> new file mode 100644
> index 0000000..bbcf11eecb0194b7c7b0f9352b0154d67b2dc06a
> --- /dev/null
> +++ b/arch/nios2/include/asm/signal.h
> @@ -0,0 +1,22 @@
> +/*
> + * Copyright Altera Corporation (C) 2013. All rights reserved
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program. If not, see <http://www.gnu.org/licenses/>.
> + *
> + */
> +#ifndef _NIOS2_SIGNAL_H
> +#define _NIOS2_SIGNAL_H
> +
> +#include <uapi/asm/signal.h>
> +
> +#endif /* _NIOS2_SIGNAL_H */
> diff --git a/arch/nios2/include/asm/ucontext.h b/arch/nios2/include/asm/ucontext.h
> new file mode 100644
> index 0000000..5870ef4e92fbaec9137142534ce3c5b23705d43e
> --- /dev/null
> +++ b/arch/nios2/include/asm/ucontext.h
> @@ -0,0 +1,34 @@
> +/*
> + * Copyright (C) 2010 Tobias Klauser <[email protected]>
> + * Copyright (C) 2004 Microtronix Datacom Ltd
> + *
> + * derived from m68knommu
> + *
> + * This file is subject to the terms and conditions of the GNU General Public
> + * License. See the file "COPYING" in the main directory of this archive
> + * for more details.
> + */
> +
> +#ifndef _ASM_NIOS2_UCONTEXT_H
> +#define _ASM_NIOS2_UCONTEXT_H
> +
> +typedef int greg_t;
> +#define NGREG 32
> +typedef greg_t gregset_t[NGREG];
> +
> +struct mcontext {
> + int version;
> + gregset_t gregs;
> +};
> +
> +#define MCONTEXT_VERSION 2
> +
> +struct ucontext {
> + unsigned long uc_flags;
> + struct ucontext *uc_link;
> + stack_t uc_stack;
> + struct mcontext uc_mcontext;
> + sigset_t uc_sigmask; /* mask last for extensibility */
> +};
> +
> +#endif
> diff --git a/arch/nios2/include/uapi/asm/sigcontext.h b/arch/nios2/include/uapi/asm/sigcontext.h
> new file mode 100644
> index 0000000..6bfd8807597a7d6abc0ae32f030e44cb4f77809f
> --- /dev/null
> +++ b/arch/nios2/include/uapi/asm/sigcontext.h
> @@ -0,0 +1,30 @@
> +/*
> + * Taken from the m68knommu.
> + *
> + * Copyright (C) 2004, Microtronix Datacom Ltd.
> + *
> + * All rights reserved.
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation; either version 2 of the License, or
> + * (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful, but
> + * WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
> + * NON INFRINGEMENT. See the GNU General Public License for more
> + * details.
> + */
> +
> +#ifndef _ASM_NIOS2_SIGCONTEXT_H
> +#define _ASM_NIOS2_SIGCONTEXT_H
> +
> +#include <asm/ptrace.h>
> +
> +struct sigcontext {
> + struct pt_regs regs;
> + unsigned long sc_mask; /* old sigmask */
> +};
> +
> +#endif
> diff --git a/arch/nios2/include/uapi/asm/signal.h b/arch/nios2/include/uapi/asm/signal.h
> new file mode 100644
> index 0000000..f29ee63144817e9798965c0b8279c08844c44e6a
> --- /dev/null
> +++ b/arch/nios2/include/uapi/asm/signal.h
> @@ -0,0 +1,23 @@
> +/*
> + * Copyright Altera Corporation (C) 2013. All rights reserved
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program. If not, see <http://www.gnu.org/licenses/>.
> + *
> + */
> +#ifndef _ASM_NIOS2_SIGNAL_H
> +#define _ASM_NIOS2_SIGNAL_H
> +
> +#define SA_RESTORER 0x04000000
> +#include <asm-generic/signal.h>
> +
> +#endif /* _ASM_NIOS2_SIGNAL_H */
> diff --git a/arch/nios2/kernel/signal.c b/arch/nios2/kernel/signal.c
> new file mode 100644
> index 0000000..1c5cbe9c55dde1c3467d727ec66c36b96f3cb9cd
> --- /dev/null
> +++ b/arch/nios2/kernel/signal.c
> @@ -0,0 +1,364 @@
> +/*
> + * Copyright (C) 2013 Altera Corporation
> + * Copyright (C) 2011-2012 Tobias Klauser <[email protected]>
> + * Copyright (C) 2004 Microtronix Datacom Ltd
> + * Copyright (C) 1991, 1992 Linus Torvalds
> + *
> + * This file is based on kernel/signal.c from m68knommu.
> + *
> + * This file is subject to the terms and conditions of the GNU General Public
> + * License. See the file COPYING in the main directory of this archive
> + * for more details.
> + */
> +
> +#include <linux/signal.h>
> +#include <linux/errno.h>
> +#include <linux/ptrace.h>
> +#include <linux/uaccess.h>
> +#include <linux/unistd.h>
> +#include <linux/personality.h>
> +#include <linux/tracehook.h>
> +
> +#include <asm/ucontext.h>
> +#include <asm/cacheflush.h>
> +
> +#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
> +
> +static int do_signal(struct pt_regs *regs, sigset_t *oldset, int in_syscall);
> +
> +/*
> + * Do a signal return; undo the signal stack.
> + *
> + * Keep the return code on the stack quadword aligned!
> + * That makes the cache flush below easier.
> + */
> +
> +struct sigframe {
> + char retcode[12];
> + unsigned long extramask[_NSIG_WORDS-1];
> + struct sigcontext sc;
> +};
> +
> +struct rt_sigframe {
> + char retcode[12];
> + struct siginfo info;
> + struct ucontext uc;
> +};
> +
> +static inline int rt_restore_ucontext(struct pt_regs *regs,
> + struct switch_stack *sw,
> + struct ucontext *uc, int *pr2)
> +{
> + int temp;
> + greg_t *gregs = uc->uc_mcontext.gregs;
> + int err;
> +
> + err = __get_user(temp, &uc->uc_mcontext.version);
> + if (temp != MCONTEXT_VERSION)
> + goto badframe;
> + /* restore passed registers */
> + err |= __get_user(regs->r1, &gregs[0]);
> + err |= __get_user(regs->r2, &gregs[1]);
> + err |= __get_user(regs->r3, &gregs[2]);
> + err |= __get_user(regs->r4, &gregs[3]);
> + err |= __get_user(regs->r5, &gregs[4]);
> + err |= __get_user(regs->r6, &gregs[5]);
> + err |= __get_user(regs->r7, &gregs[6]);
> + err |= __get_user(regs->r8, &gregs[7]);
> + err |= __get_user(regs->r9, &gregs[8]);
> + err |= __get_user(regs->r10, &gregs[9]);
> + err |= __get_user(regs->r11, &gregs[10]);
> + err |= __get_user(regs->r12, &gregs[11]);
> + err |= __get_user(regs->r13, &gregs[12]);
> + err |= __get_user(regs->r14, &gregs[13]);
> + err |= __get_user(regs->r15, &gregs[14]);
> + err |= __get_user(sw->r16, &gregs[15]);
> + err |= __get_user(sw->r17, &gregs[16]);
> + err |= __get_user(sw->r18, &gregs[17]);
> + err |= __get_user(sw->r19, &gregs[18]);
> + err |= __get_user(sw->r20, &gregs[19]);
> + err |= __get_user(sw->r21, &gregs[20]);
> + err |= __get_user(sw->r22, &gregs[21]);
> + err |= __get_user(sw->r23, &gregs[22]);
> + /* gregs[23] is handled below */
> + err |= __get_user(sw->fp, &gregs[24]); /* Verify, should this be
> + settable */
> + err |= __get_user(sw->gp, &gregs[25]); /* Verify, should this be
> + settable */
> +
> + err |= __get_user(temp, &gregs[26]); /* Not really necessary no user
> + settable bits */
> + err |= __get_user(regs->ea, &gregs[27]);
> +
> + err |= __get_user(regs->ra, &gregs[23]);
> + err |= __get_user(regs->sp, &gregs[28]);
> +
> + regs->estatus = (regs->estatus & 0xffffffff);
> + regs->orig_r2 = -1; /* disable syscall checks */
> +
> + err |= restore_altstack(&uc->uc_stack);
> + if (err)
> + goto badframe;
> +
> + *pr2 = regs->r2;
> + return err;
> +
> +badframe:
> + return 1;
> +}
> +
> +asmlinkage int do_rt_sigreturn(struct switch_stack *sw)
> +{
> + struct pt_regs *regs = (struct pt_regs *)(sw + 1);
> + /* Verify, can we follow the stack back */
> + struct rt_sigframe *frame = (struct rt_sigframe *) regs->sp;
> + sigset_t set;
> + int rval;
> +
> + if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
> + goto badframe;
> +
> + if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set)))
> + goto badframe;
> +
> + sigdelsetmask(&set, ~_BLOCKABLE);
> + spin_lock_irq(&current->sighand->siglock);
> + current->blocked = set;
> + recalc_sigpending();
> + spin_unlock_irq(&current->sighand->siglock);
> +
> + if (rt_restore_ucontext(regs, sw, &frame->uc, &rval))
> + goto badframe;
> +
> + return rval;
> +
> +badframe:
> + force_sig(SIGSEGV, current);
> + return 0;
> +}
> +
> +static inline int rt_setup_ucontext(struct ucontext *uc, struct pt_regs *regs)
> +{
> + struct switch_stack *sw = (struct switch_stack *)regs - 1;
> + greg_t *gregs = uc->uc_mcontext.gregs;
> + int err = 0;
> +
> + err |= __put_user(MCONTEXT_VERSION, &uc->uc_mcontext.version);
> + err |= __put_user(regs->r1, &gregs[0]);
> + err |= __put_user(regs->r2, &gregs[1]);
> + err |= __put_user(regs->r3, &gregs[2]);
> + err |= __put_user(regs->r4, &gregs[3]);
> + err |= __put_user(regs->r5, &gregs[4]);
> + err |= __put_user(regs->r6, &gregs[5]);
> + err |= __put_user(regs->r7, &gregs[6]);
> + err |= __put_user(regs->r8, &gregs[7]);
> + err |= __put_user(regs->r9, &gregs[8]);
> + err |= __put_user(regs->r10, &gregs[9]);
> + err |= __put_user(regs->r11, &gregs[10]);
> + err |= __put_user(regs->r12, &gregs[11]);
> + err |= __put_user(regs->r13, &gregs[12]);
> + err |= __put_user(regs->r14, &gregs[13]);
> + err |= __put_user(regs->r15, &gregs[14]);
> + err |= __put_user(sw->r16, &gregs[15]);
> + err |= __put_user(sw->r17, &gregs[16]);
> + err |= __put_user(sw->r18, &gregs[17]);
> + err |= __put_user(sw->r19, &gregs[18]);
> + err |= __put_user(sw->r20, &gregs[19]);
> + err |= __put_user(sw->r21, &gregs[20]);
> + err |= __put_user(sw->r22, &gregs[21]);
> + err |= __put_user(sw->r23, &gregs[22]);
> + err |= __put_user(regs->ra, &gregs[23]);
> + err |= __put_user(sw->fp, &gregs[24]);
> + err |= __put_user(sw->gp, &gregs[25]);
> + err |= __put_user(regs->ea, &gregs[27]);
> + err |= __put_user(regs->sp, &gregs[28]);
> + return err;
> +}
> +
> +static inline void push_cache(unsigned long vaddr)
> +{
> + flush_dcache_range(vaddr, vaddr + 12);
> + flush_icache_range(vaddr, vaddr + 12);
> +}
> +
> +static inline void *get_sigframe(struct k_sigaction *ka, struct pt_regs *regs,
> + size_t frame_size)
> +{
> + unsigned long usp;
> +
> + /* Default to using normal stack. */
> + usp = regs->sp;
> +
> + /* This is the X/Open sanctioned signal stack switching. */
> + if ((ka->sa.sa_flags & SA_ONSTACK) && (current->sas_ss_sp != 0)) {
> + if (!on_sig_stack(usp))
> + usp = current->sas_ss_sp + current->sas_ss_size;
> + }

You can use sigsp() here.

> + /* Verify, is it 32 or 64 bit aligned */
> + return (void *)((usp - frame_size) & -8UL);
> +}
> +
> +static void setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
> + sigset_t *set, struct pt_regs *regs)
> +{
> + struct rt_sigframe *frame;
> + int err = 0;
> +
> + frame = get_sigframe(ka, regs, sizeof(*frame));
> +
> + if (info)
> + err |= copy_siginfo_to_user(&frame->info, info);
> +
> + /* Create the ucontext. */
> + err |= __put_user(0, &frame->uc.uc_flags);
> + err |= __put_user(0, &frame->uc.uc_link);
> + err |= __save_altstack(&frame->uc.uc_stack, regs->sp);
> + err |= rt_setup_ucontext(&frame->uc, regs);
> + err |= copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set));
> +
> + /* Set up to return from userspace. */
> + regs->ra = (unsigned long) &frame->retcode[0];
> +
> + /* movi r2,__NR_rt_sigreturn */
> + err |= __put_user(0x00800004 + (__NR_rt_sigreturn << 6),
> + (long *)(frame->retcode));
> + /* trap */
> + err |= __put_user(0x003b683a, (long *)(frame->retcode + 4));
> +
> + if (err)
> + goto give_sigsegv;
> +
> + push_cache((unsigned long) &frame->retcode);
> +
> + /* Set up registers for signal handler */
> + regs->sp = (unsigned long) frame;
> + regs->r4 = (unsigned long) (current_thread_info()->exec_domain
> + && current_thread_info()->exec_domain->signal_invmap
> + && sig < 32
> + ? current_thread_info()->exec_domain->signal_invmap[sig]
> + : sig);

Does nios2 really need signal translation and supports execution domains?

> + regs->r5 = (unsigned long) &frame->info;
> + regs->r6 = (unsigned long) &frame->uc;
> + regs->ea = (unsigned long) ka->sa.sa_handler;
> + return;
> +
> +give_sigsegv:
> + force_sigsegv(sig, current);
> +}
> +
> +static inline void handle_restart(struct pt_regs *regs, struct k_sigaction *ka,
> + int has_handler)
> +{
> + switch (regs->r2) {
> + case ERESTART_RESTARTBLOCK:
> + case ERESTARTNOHAND:
> + regs->r2 = EINTR;
> + regs->r7 = 1;
> + break;
> + case ERESTARTSYS:
> + if (has_handler && !(ka->sa.sa_flags & SA_RESTART)) {
> + regs->r2 = EINTR;
> + regs->r7 = 1;
> + break;
> + }
> + /* fallthrough */
> + case ERESTARTNOINTR:
> + regs->r2 = regs->orig_r2;
> + regs->r7 = regs->orig_r7;
> + regs->ea -= 4;
> + break;
> + }
> +}
> +
> +/*
> + * OK, we're invoking a handler
> + */
> +static void handle_signal(int sig, struct k_sigaction *ka, siginfo_t *info,
> + sigset_t *oldset, struct pt_regs *regs)
> +{
> + /* set up the stack frame */
> + if (ka->sa.sa_flags & SA_SIGINFO)
> + setup_rt_frame(sig, ka, info, oldset, regs);
> + else
> + setup_rt_frame(sig, ka, NULL, oldset, regs);
> +
> + if (!(ka->sa.sa_flags & SA_NODEFER)) {
> + spin_lock_irq(&current->sighand->siglock);
> + sigorsets(&current->blocked, &current->blocked,
> + &ka->sa.sa_mask);
> + sigaddset(&current->blocked, sig);
> + recalc_sigpending();
> + spin_unlock_irq(&current->sighand->siglock);
> + }
> +}
> +
> +/*
> + * Note that 'init' is a special process: it doesn't get signals it doesn't
> + * want to handle. Thus you cannot kill init even with a SIGKILL even by
> + * mistake.
> + */

Is this comment still needed? Looks like copy&paste...

> +static int do_signal(struct pt_regs *regs, sigset_t *oldset, int in_syscall)
> +{
> + struct k_sigaction ka;
> + siginfo_t info;
> + int signr;
> +
> + current->thread.kregs = regs;
> +
> + if (!oldset)
> + oldset = &current->blocked;
> +
> + signr = get_signal_to_deliver(&info, &ka, regs, NULL);
> + if (signr > 0) {
> + /*
> + * Are we from a system call? If so, check system call
> + * restarting.
> + */
> + if (in_syscall)
> + handle_restart(regs, &ka, 1);
> + /* Whee! Actually deliver the signal. */
> + handle_signal(signr, &ka, &info, oldset, regs);
> + return 1;
> + }
> +
> + /*
> + * No signal to deliver to the process - restart the syscall.
> + */
> + if (in_syscall) {
> + /* Did the syscall return an error code */
> + if (regs->r7 == 1) {
> + if (regs->r2 == ERESTARTNOHAND ||
> + regs->r2 == ERESTARTSYS ||
> + regs->r2 == ERESTARTNOINTR) {
> + regs->r2 = regs->orig_r2;
> + regs->r7 = regs->orig_r7;
> + regs->ea -= 4;
> + } else if (regs->r2 == ERESTART_RESTARTBLOCK) {
> + regs->r2 = __NR_restart_syscall;
> + regs->ea -= 4;
> + }
> + }
> + }
> +
> + return 0;
> +}
> +
> +asmlinkage void do_notify_resume(struct pt_regs *regs, sigset_t *oldset,
> + int in_syscall)
> +{
> + pr_debug("--> ENTERING %s\n", __func__);

Please remove such printk()s.
If you need to know which functions get called, use ftrace.

> + /*
> + * We want the common case to go fast, which is why we may in certain
> + * cases get here from kernel mode. Just return without doing anything
> + * if so.
> + */
> + if (!user_mode(regs))
> + return;
> +
> + if (test_thread_flag(TIF_SIGPENDING))
> + do_signal(regs, oldset, in_syscall);
> +
> + if (test_and_clear_thread_flag(TIF_NOTIFY_RESUME))
> + tracehook_notify_resume(regs);
> +}
> --
> 1.8.3.2

--
Thanks,
//richard

2014-04-19 20:12:44

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

On Fri, Apr 18, 2014 at 2:26 PM, Ley Foon Tan <[email protected]> wrote:
> +/* sys_cacheflush -- flush the processor cache. */
> +asmlinkage int sys_cacheflush(unsigned long addr, unsigned long len,
> + unsigned int op)
> +{
> + struct vm_area_struct *vma;
> +
> + if (len == 0)
> + return 0;
> +
> + /* Check for overflow */
> + if (addr + len < addr)
> + return -EFAULT;
> +
> + /*
> + * Verify that the specified address region actually belongs
> + * to this process.
> + */
> + vma = find_vma(current->mm, addr);
> + if (vma == NULL || addr < vma->vm_start || addr + len > vma->vm_end)
> + return -EFAULT;
> +
> + /* Ignore the scope and cache arguments. */

This function doesn't have scope and cache arguments
(sys_cacheflush() on m68k does have them ;-)

On Sat, Apr 19, 2014 at 6:09 PM, Pavel Machek <[email protected]> wrote:
> Also... explanation why you need this syscall while other
> architectures live happily without it would be nice.

12 our of 28 architectures provide such a syscall...

Is flushing the cache a privileged operation on nios2?

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

2014-04-21 03:03:17

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 01/28] nios2: Build infrastructure

On Jum, 2014-04-18 at 16:35 +0200, Sam Ravnborg wrote:
> It is good practice to add build infrastructure last - so build is not broken.
> But for a new architecture it matters only little.
>
> Some comments in the following.
>
> Sam
Noted. But, I will keep the patch # for this time for consistency.
Thanks.
> >
> > Signed-off-by: Ley Foon Tan <[email protected]>
> > ---
> >
> > diff --git a/arch/nios2/Kconfig b/arch/nios2/Kconfig
> > new file mode 100644
> > index 0000000..2bd0d0d7160f31a5c5890cf8a059b9f53378be6e
> > --- /dev/null
> > +++ b/arch/nios2/Kconfig
> > @@ -0,0 +1,213 @@
> > +config NIOS2
> > + def_bool y
> > + select OF
> > + select OF_EARLY_FLATTREE
> > + select USB_ARCH_HAS_HCD if USB_SUPPORT
> > + select ARCH_WANT_OPTIONAL_GPIOLIB
> > + select GENERIC_IRQ_PROBE
> > + select GENERIC_IRQ_SHOW
> > + select GENERIC_CPU_DEVICES
> > + select GENERIC_ATOMIC64
> > + select MODULES_USE_ELF_RELA
> > + select IRQ_DOMAIN
> > + select SOC_BUS
> > + select CLKSRC_OF
> If you keep this list sorted then merge issues are less likely.
Noted, will sort this list.

> > +config KERNEL_MMU_REGION_BASE_BOOL
> > + bool "Set custom kernel MMU region base address"
> > + depends on ADVANCED_OPTIONS
> > + help
> > + This option allows you to set the virtual address of the kernel MMU region.
> > +
> > + Say N here unless you know what you are doing.
>
> All Nios2 specific options should be prefixed with NIOS2_
> This is what is done for all new arch specific symbols.
Noted, will add NIOS2 prefix to these CONFIGs.
>
> > diff --git a/arch/nios2/boot/Makefile b/arch/nios2/boot/Makefile

> > + $(call if_changed,gzip)
> > +
> > +$(obj)/vmImage: $(obj)/vmlinux.gz
> > + $(call if_changed,uimage)
> > + @$(kecho) 'Kernel: $@ is ready'
>
> Add:
> targets += vmlinux.bin vmlinux.gz vmImage
> to make $(call if_changed ... actually work.
Sure, will change to this way.

>
>
> > +obj-y += cpuinfo.o insnemu.o irq.o nios2_ksyms.o process.o prom.o ptrace.o \
> > + setup.o signal.o sys_nios2.o syscall_table.o time.o traps.o
>
> In new stuff try to avoid using ... \
> to contatenate lines.
> Use:
> obj-y += xxx
>
> This is much more readable.
>
Will fix this.

2014-04-21 03:44:17

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 01/28] nios2: Build infrastructure

On Jum, 2014-04-18 at 21:41 +0200, Paul Bolle wrote:
> On Fri, 2014-04-18 at 20:26 +0800, Ley Foon Tan wrote:
> > diff --git a/arch/nios2/kernel/Makefile b/arch/nios2/kernel/Makefile
> [...]
> > +obj-$(CONFIG_CONSOLE) += console.o
>
> CONFIG_CONSOLE seems never to be defined. And there appears to be no
> arch/nios2/console.c or .../console.S anyway.
>
>
> Paul Bolle
>
Good catch. Will remove this.
Thanks.

Regards
Ley Foon

2014-04-21 05:02:11

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 01/28] nios2: Build infrastructure

On Jum, 2014-04-18 at 21:16 +0200, Paul Bolle wrote:
> On Fri, 2014-04-18 at 20:26 +0800, Ley Foon Tan wrote:
> > diff --git a/arch/nios2/Kconfig b/arch/nios2/Kconfig
> >[...]
> > +config GENERIC_FIND_NEXT_BIT
> > + def_bool y
>
> This one is not needed, see commit 63e424c84429 ("arch: remove
> CONFIG_GENERIC_FIND_{NEXT_BIT,BIT_LE,LAST_BIT}").
Okay, will remove this CONFIG.

>
> > +config NO_IOPORT
> > + def_bool y
>
> Did you mean NO_IOPORT_MAP?
Yes, will change it to NO_IOPORT_MAP

Thanks.

Regards
Ley Foon

2014-04-21 05:23:23

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On Sat, Apr 19, 2014 at 11:30 PM, Arnd Bergmann <[email protected]> wrote:
> On Friday 18 April 2014, H. Peter Anvin wrote:
>> Did the generic headers ever get updated to match Linus' guidance that
>> any new architecture ports should use a 64-bit time_t?
>
> No, unfortunately not. With my rule that every architecture that gets
> added needs to clean up asm-generic some more, to make it easier to add
> the next one, we should probably do for nios2.
>
> Arnd

Can you give me the documentation on this new guidance and point me
any architecture have implemented this?
Thanks.

Ley Foon

2014-04-21 05:32:34

by H. Peter Anvin

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 04/20/2014 10:23 PM, Ley Foon Tan wrote:
> On Sat, Apr 19, 2014 at 11:30 PM, Arnd Bergmann <[email protected]> wrote:
>> On Friday 18 April 2014, H. Peter Anvin wrote:
>>> Did the generic headers ever get updated to match Linus' guidance that
>>> any new architecture ports should use a 64-bit time_t?
>>
>> No, unfortunately not. With my rule that every architecture that gets
>> added needs to clean up asm-generic some more, to make it easier to add
>> the next one, we should probably do for nios2.
>>
>> Arnd
>
> Can you give me the documentation on this new guidance and point me
> any architecture have implemented this?
> Thanks.
>

We implemented it in the x32 ABI for x86. In generic code this is keyed
by COMPAT_USE_64BIT_TIME, but in your case it isn't actually a matter of
compat, so it should be easier.

See this thread including the discussion about time_t:

https://lkml.org/lkml/2011/8/26/415

-hpa

2014-04-21 08:14:14

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 14/4/21 1:31 PM, H. Peter Anvin wrote:
> On 04/20/2014 10:23 PM, Ley Foon Tan wrote:
>> On Sat, Apr 19, 2014 at 11:30 PM, Arnd Bergmann <[email protected]> wrote:
>>> On Friday 18 April 2014, H. Peter Anvin wrote:
>>>> Did the generic headers ever get updated to match Linus' guidance that
>>>> any new architecture ports should use a 64-bit time_t?
>>>
>>> No, unfortunately not. With my rule that every architecture that gets
>>> added needs to clean up asm-generic some more, to make it easier to add
>>> the next one, we should probably do for nios2.
>>>
>>> Arnd
>>
>> Can you give me the documentation on this new guidance and point me
>> any architecture have implemented this?
>> Thanks.
>>
>
> We implemented it in the x32 ABI for x86. In generic code this is keyed
> by COMPAT_USE_64BIT_TIME, but in your case it isn't actually a matter of
> compat, so it should be easier.
>
> See this thread including the discussion about time_t:
>
> https://lkml.org/lkml/2011/8/26/415
>
> -hpa


I'm not sure why we need CONFIG_USE_64BIT_TIME? We don't have any
32-on-64 problems.

My understanding of what Arnd meant, was to clean up
asm-generic/posix_types.h such that for example, time_t should be
defined as a 'long long' type, which should be 64-bit almost everywhere.
At least AFAIK, that should work for all current asm-generic users.

Thanks,
Chung-Lin

2014-04-21 17:23:56

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

On Sun, Apr 20, 2014 at 4:12 AM, Geert Uytterhoeven
<[email protected]> wrote:
>> + */
>> + vma = find_vma(current->mm, addr);
>> + if (vma == NULL || addr < vma->vm_start || addr + len > vma->vm_end)
>> + return -EFAULT;
>> +
>> + /* Ignore the scope and cache arguments. */
>
> This function doesn't have scope and cache arguments
> (sys_cacheflush() on m68k does have them ;-)

Missed out the comment after rename the arguments.:) Will update the
comment above.

> On Sat, Apr 19, 2014 at 6:09 PM, Pavel Machek <[email protected]> wrote:
>> Also... explanation why you need this syscall while other
>> architectures live happily without it would be nice.
>
> 12 our of 28 architectures provide such a syscall...
>
> Is flushing the cache a privileged operation on nios2?
Yes.

Thanks.

Regards
Ley Foon

2014-04-21 17:32:22

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

On Sun, Apr 20, 2014 at 12:09 AM, Pavel Machek <[email protected]> wrote:
>> +/* Additional Nios II specific syscalls. */
>> +#define __NR_cacheflush (__NR_arch_specific_syscall)
>> +__SYSCALL(__NR_cacheflush, sys_cacheflush)
>
> I guess you should Cc: Michael Kerrisk on this one.
CC him this email.

>
> Also... explanation why you need this syscall while other
> architectures live happily without it would be nice.
This syscall exist in nios2 port since old days and some users already
used it for sometimes.
Like mentioned by Greet, 12 our of 28 architectures provide such a syscall.

>> +/* sys_cacheflush -- flush the processor cache. */
>> +asmlinkage int sys_cacheflush(unsigned long addr, unsigned long len,
>> + unsigned int op)
>> +{
>> + struct vm_area_struct *vma;
>> +
>> + if (len == 0)
>> + return 0;
>
> op is being ignored? You should remove it... or if (op) return -EINVAL
> if you want future extensions.
Yes, op argument is being ignored. Okay, will add the return -EINVAL
if op is non-zero.

Thanks.

Regards
Ley Foon

2014-04-21 17:52:45

by Richard Weinberger

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

On Mon, Apr 21, 2014 at 7:32 PM, Ley Foon Tan <[email protected]> wrote:
> On Sun, Apr 20, 2014 at 12:09 AM, Pavel Machek <[email protected]> wrote:
>>> +/* Additional Nios II specific syscalls. */
>>> +#define __NR_cacheflush (__NR_arch_specific_syscall)
>>> +__SYSCALL(__NR_cacheflush, sys_cacheflush)
>>
>> I guess you should Cc: Michael Kerrisk on this one.
> CC him this email.
>
>>
>> Also... explanation why you need this syscall while other
>> architectures live happily without it would be nice.
> This syscall exist in nios2 port since old days and some users already
> used it for sometimes.
> Like mentioned by Greet, 12 our of 28 architectures provide such a syscall.
>
>>> +/* sys_cacheflush -- flush the processor cache. */
>>> +asmlinkage int sys_cacheflush(unsigned long addr, unsigned long len,
>>> + unsigned int op)
>>> +{
>>> + struct vm_area_struct *vma;
>>> +
>>> + if (len == 0)
>>> + return 0;
>>
>> op is being ignored? You should remove it... or if (op) return -EINVAL
>> if you want future extensions.
> Yes, op argument is being ignored. Okay, will add the return -EINVAL
> if op is non-zero.

Be careful. If existing applications already use this and maybe pass a
dummy value
to it you break them.

--
Thanks,
//richard

2014-04-21 20:46:40

by Pavel Machek

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

Hi!

> > Also... explanation why you need this syscall while other
> > architectures live happily without it would be nice.
> This syscall exist in nios2 port since old days and some users already
> used it for sometimes.
> Like mentioned by Greet, 12 our of 28 architectures provide such a syscall.

Would it make sense to create common code for this syscall?

> >> +/* sys_cacheflush -- flush the processor cache. */
> >> +asmlinkage int sys_cacheflush(unsigned long addr, unsigned long len,
> >> + unsigned int op)
> >> +{
> >> + struct vm_area_struct *vma;
> >> +
> >> + if (len == 0)
> >> + return 0;
> >
> > op is being ignored? You should remove it... or if (op) return -EINVAL
> > if you want future extensions.
> Yes, op argument is being ignored. Okay, will add the return -EINVAL
> if op is non-zero.

Thanks!
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

2014-04-21 20:48:39

by Pavel Machek

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

> >>> +/* sys_cacheflush -- flush the processor cache. */
> >>> +asmlinkage int sys_cacheflush(unsigned long addr, unsigned long len,
> >>> + unsigned int op)
> >>> +{
> >>> + struct vm_area_struct *vma;
> >>> +
> >>> + if (len == 0)
> >>> + return 0;
> >>
> >> op is being ignored? You should remove it... or if (op) return -EINVAL
> >> if you want future extensions.
> > Yes, op argument is being ignored. Okay, will add the return -EINVAL
> > if op is non-zero.
>
> Be careful. If existing applications already use this and maybe pass a
> dummy value
> to it you break them.

Well.. of course, but passing ignored integer to syscall, that can be
never used is beyond ugly. NAK-level ugly.

If it is so, you can never ever use the argument, and might as well
remove it (depending on your calling convention, it should not break
anything.)

Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

2014-04-22 08:09:57

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 09/28] nios2: Page table management

Hi

On Sun, Apr 20, 2014 at 12:05 AM, Pavel Machek <[email protected]> wrote:

>> +static void pgd_init(pgd_t *pgd)
>> +{
>> + unsigned long *p = (unsigned long *) pgd;
>> + int i;
>> +
>> + for (i = 0; i < USER_PTRS_PER_PGD; i += 8) {
>> + p[i + 0] = (unsigned long) invalid_pte_table;
>> + p[i + 1] = (unsigned long) invalid_pte_table;
>> + p[i + 2] = (unsigned long) invalid_pte_table;
>> + p[i + 3] = (unsigned long) invalid_pte_table;
>> + p[i + 4] = (unsigned long) invalid_pte_table;
>> + p[i + 5] = (unsigned long) invalid_pte_table;
>> + p[i + 6] = (unsigned long) invalid_pte_table;
>> + p[i + 7] = (unsigned long) invalid_pte_table;
>> + }
>
> Umm. Manual loop unrolling. Does it really improve performance? Is the
> code hot enough to warrant the uglyness?
We can use the memset here because invalid_pte_table is full 32 bit value.
BTW, arch for score and mips have the similar implementation as well.

Regards

2014-04-22 08:24:42

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

On Tue, Apr 22, 2014 at 4:48 AM, Pavel Machek <[email protected]> wrote:
>> >>> +/* sys_cacheflush -- flush the processor cache. */
>> >>> +asmlinkage int sys_cacheflush(unsigned long addr, unsigned long len,
>> >>> + unsigned int op)
>> >>> +{
>> >>> + struct vm_area_struct *vma;
>> >>> +
>> >>> + if (len == 0)
>> >>> + return 0;
>> >>
>> >> op is being ignored? You should remove it... or if (op) return -EINVAL
>> >> if you want future extensions.
>> > Yes, op argument is being ignored. Okay, will add the return -EINVAL
>> > if op is non-zero.
>>
>> Be careful. If existing applications already use this and maybe pass a
>> dummy value
>> to it you break them.
>
> Well.. of course, but passing ignored integer to syscall, that can be
> never used is beyond ugly. NAK-level ugly.
>
> If it is so, you can never ever use the argument, and might as well
> remove it (depending on your calling convention, it should not break
> anything.)
>
I'm prefer to keep the calling convention here for future use and add
the op checking here.
This prevent us to update toolchain if we plan to add in op option later.

Regards
Ley Foon

2014-04-22 10:37:16

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

Hi Arnd and Peter Anvin,

Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
that we don't need to have 64 bit off_t? See detail in link below.
I can submit the patches for 64-bit time changes
(include/asm-generic/posix_types.h and other archs) if everyone is
agreed on this.

Excerpt from https://lkml.org/lkml/2012/11/14/358 :
"Obviously, we want to use 64-bit off_t, but this is achieved already
through loff_t, which is used in all places in the asm-generic
ABI anyway (the syscalls using off_t are stripped out). I don't
think we want to have the other ones set to 64 bit on ARC or Meta,
although I'm not 100% sure about ino_t and nlink_t. "

Thanks.

Regards
Ley Foon

On Mon, Apr 21, 2014 at 1:31 PM, H. Peter Anvin <[email protected]> wrote:
> On 04/20/2014 10:23 PM, Ley Foon Tan wrote:
>> On Sat, Apr 19, 2014 at 11:30 PM, Arnd Bergmann <[email protected]> wrote:
>>> On Friday 18 April 2014, H. Peter Anvin wrote:
>>>> Did the generic headers ever get updated to match Linus' guidance that
>>>> any new architecture ports should use a 64-bit time_t?
>>>
>>> No, unfortunately not. With my rule that every architecture that gets
>>> added needs to clean up asm-generic some more, to make it easier to add
>>> the next one, we should probably do for nios2.
>>>
>>> Arnd
>>
>> Can you give me the documentation on this new guidance and point me
>> any architecture have implemented this?
>> Thanks.
>>
>
> We implemented it in the x32 ABI for x86. In generic code this is keyed
> by COMPAT_USE_64BIT_TIME, but in your case it isn't actually a matter of
> compat, so it should be easier.
>
> See this thread including the discussion about time_t:
>
> https://lkml.org/lkml/2011/8/26/415
>
> -hpa

2014-04-22 10:58:13

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
> Hi Arnd and Peter Anvin,
>
> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
> that we don't need to have 64 bit off_t? See detail in link below.
> I can submit the patches for 64-bit time changes
> (include/asm-generic/posix_types.h and other archs) if everyone is
> agreed on this.

Yes.

> Excerpt from https://lkml.org/lkml/2012/11/14/358 :
> "Obviously, we want to use 64-bit off_t, but this is achieved already
> through loff_t, which is used in all places in the asm-generic
> ABI anyway (the syscalls using off_t are stripped out). I don't
> think we want to have the other ones set to 64 bit on ARC or Meta,
> although I'm not 100% sure about ino_t and nlink_t. "

This is all still true. You should have no syscall using 'off_t',
only loff_t.

I still don't know whether we would want 32 or 64 bit ino_t and nlink_t
for new architectures. It seems it would gain very little, but have
a noticeable overhead.

Arnd

2014-04-22 10:58:21

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 04/28] nios2: Exception handling

Hi Ley Foon,

On Apr 18, Ley Foon Tan wrote:
> +
> + /* Get thread info pointer */
> + movi r11, %lo(0xfffff000)
> + and r11, sp, r11
> + ldw r11, TI_FLAGS(r11)
> +
> + /* If someone is ptrace:ing us, take the long way. */
> + BTBNZ r11, r11, TIF_SYSCALL_TRACE, traced_system_call
> +

Are you sure this is correct? Have you tested this works correctly?

Last time I tested this, it caused the traced_system_call to be taken
for every system call. I already reported and provided a patch a few
months ago:

http://lists.rocketboards.org/pipermail/nios2-dev/2014-January/006968.html

Here's my proposed fix:

diff --git a/arch/nios2/kernel/entry.S b/arch/nios2/kernel/entry.S
index 9cf2bac..3798008 100644
--- a/arch/nios2/kernel/entry.S
+++ b/arch/nios2/kernel/entry.S
@@ -204,13 +204,10 @@ ENTRY(handle_system_call)
ldw r1, %lo(sys_call_table)(r1)
beq r1, r0, ret_invsyscall

- /* Get thread info pointer */
- movi r11, %lo(0xfffff000)
- and r11, sp, r11
- ldw r11, TI_FLAGS(r11)
-
- /* If someone is ptrace:ing us, take the long way. */
- BTBNZ r11, r11, TIF_SYSCALL_TRACE, traced_system_call
+ /* Check if we are being traced */
+ GET_THREAD_INFO r11
+ ldw r11,TI_FLAGS(r11)
+ BTBNZ r11,r11,TIF_SYSCALL_TRACE,traced_system_call

/* Execute the system call */
callr r1
--
1.9.1

--
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-22 11:08:35

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 04/28] nios2: Exception handling

On Tue, Apr 22, 2014 at 6:57 PM, Ezequiel Garcia
<[email protected]> wrote:
> Hi Ley Foon,
>
> Are you sure this is correct? Have you tested this works correctly?
>
> Last time I tested this, it caused the traced_system_call to be taken
> for every system call. I already reported and provided a patch a few
> months ago:
>
> http://lists.rocketboards.org/pipermail/nios2-dev/2014-January/006968.html
>
> Here's my proposed fix:
>
> diff --git a/arch/nios2/kernel/entry.S b/arch/nios2/kernel/entry.S
> index 9cf2bac..3798008 100644
> --- a/arch/nios2/kernel/entry.S
> +++ b/arch/nios2/kernel/entry.S
> @@ -204,13 +204,10 @@ ENTRY(handle_system_call)
> ldw r1, %lo(sys_call_table)(r1)
> beq r1, r0, ret_invsyscall
>
> - /* Get thread info pointer */
> - movi r11, %lo(0xfffff000)
> - and r11, sp, r11
> - ldw r11, TI_FLAGS(r11)
> -
> - /* If someone is ptrace:ing us, take the long way. */
> - BTBNZ r11, r11, TIF_SYSCALL_TRACE, traced_system_call
> + /* Check if we are being traced */
> + GET_THREAD_INFO r11
> + ldw r11,TI_FLAGS(r11)
> + BTBNZ r11,r11,TIF_SYSCALL_TRACE,traced_system_call
>
> /* Execute the system call */
> callr r1
I think I missed out this in nios2-dev ML previously. I will take your fix.
Thanks.

Regards

2014-04-22 11:20:50

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>> Hi Arnd and Peter Anvin,
>>
>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>> that we don't need to have 64 bit off_t? See detail in link below.
>> I can submit the patches for 64-bit time changes
>> (include/asm-generic/posix_types.h and other archs) if everyone is
>> agreed on this.
>
> Yes.
Okay, will doing that.

>
>> Excerpt from https://lkml.org/lkml/2012/11/14/358 :
>> "Obviously, we want to use 64-bit off_t, but this is achieved already
>> through loff_t, which is used in all places in the asm-generic
>> ABI anyway (the syscalls using off_t are stripped out). I don't
>> think we want to have the other ones set to 64 bit on ARC or Meta,
>> although I'm not 100% sure about ino_t and nlink_t. "
>
> This is all still true. You should have no syscall using 'off_t',
> only loff_t.
>
> I still don't know whether we would want 32 or 64 bit ino_t and nlink_t
> for new architectures. It seems it would gain very little, but have
> a noticeable overhead.
Anyone have comment on this?
Chung-Lin (in CC list) is our nios2 toolchain maintainer. Do you have
any comment for 32 or 64 bit ino_t and nlink_t?
We will update the toolchain to support 64-bit time_t, so we hope that
any other toolchain change can happen in one time.

Thanks

Regards
Ley Foon

2014-04-22 12:31:04

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

On Friday 18 April 2014, Ley Foon Tan wrote:

> +
> + #define sys_mmap2 sys_mmap_pgoff
> +

You use sys_mmap_pgoff here, but

> +asmlinkage long sys_mmap(unsigned long addr, unsigned long len,
> + unsigned long prot, unsigned long flags,
> + unsigned long fd, unsigned long offset)
> +{
> + if (offset & ~PAGE_MASK)
> + return -EINVAL;
> +
> + return sys_mmap_pgoff(addr, len, prot, flags, fd, offset >> PAGE_SHIFT);
> +}

define a separate sys_mmap() here. Doing the former should be enough.
so you can just drop the sys_mmap definition.

Arnd

2014-04-22 12:33:28

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 04/28] nios2: Exception handling

On Friday 18 April 2014, Ley Foon Tan wrote:
> +static const struct file_operations misalign_fops = {
> + .open = proc_misaligned_open,
> + .read = seq_read,
> + .llseek = seq_lseek,
> + .release = single_release,
> + .write = proc_misaligned_write,
> +};
> +#endif /* CONFIG_PROC_FS */

We really shouldn't be doing new architecture specific procfs files
any more. I suggest you drop this one for now, and add back the
functionality using perf or ftrace at a later point.

Arnd

2014-04-22 12:38:40

by Tobias Klauser

[permalink] [raw]
Subject: Re: [PATCH 28/28] Documentation: Add documentation for Nios2 architecture

On 2014-04-18 at 14:27:11 +0200, Ley Foon Tan <[email protected]> wrote:
> Signed-off-by: Ley Foon Tan <[email protected]>
> ---
> Documentation/nios2/README | 23 +++++++++++++++++++++++
> 1 file changed, 23 insertions(+)
> create mode 100644 Documentation/nios2/README
>
> diff --git a/Documentation/nios2/README b/Documentation/nios2/README
> new file mode 100644
> index 0000000..33eec4d
> --- /dev/null
> +++ b/Documentation/nios2/README
> @@ -0,0 +1,23 @@
> +Linux on the Nios II architecture
> +=================================
> +
> +This is a port of Linux to Nios II (nios2) processor.
> +
> +In order to compile for Nios II, you need a version of GCC with support for the generic
> +system call ABI. Please see this link for more information on how compiling and booting
> +software for the Nios II platform:
> +http://www.rocketboards.org/foswiki/Documentation/NiosIILinuxUserManual
> +
> +For reference, please see the following link:
> +http://www.altera.com/literature/lit-nio2.jsp
> +
> +What is Nios II?
> +================
> +Nios II is a 32-bit embedded-processor architecture designed specifically for the
> +Altera family of FPGAs. In order to support Linux system, Nios II needs to configure

Minor nit: 'In order to support Linux, Nios II needs to be configured...'

> +with MMU and hardware multiplier enabled.

> +
> +Nios II ABI
> +===========
> +Please refer to chapter "Application Binary Interface" in Nios II Processor Reference
> +Handbook.
> --
> 1.8.3.2
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-arch" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>

2014-04-22 12:39:19

by Tobias Klauser

[permalink] [raw]
Subject: Re: [PATCH 23/28] nios2: Nios2 registers

On 2014-04-18 at 14:27:06 +0200, Ley Foon Tan <[email protected]> wrote:
> This file contains constants for the instruction macros, cpu registers, fields and bits.
>
> Signed-off-by: Ley Foon Tan <[email protected]>
> ---
> arch/nios2/include/asm/registers.h | 65 ++++++++++++++++++++++++++++++++++++++
> 1 file changed, 65 insertions(+)
> create mode 100644 arch/nios2/include/asm/registers.h
>
> diff --git a/arch/nios2/include/asm/registers.h b/arch/nios2/include/asm/registers.h
> new file mode 100644
> index 0000000..0e6ffef08b20e072480ca3780e2626eb3f1de2bd
> --- /dev/null
> +++ b/arch/nios2/include/asm/registers.h
> @@ -0,0 +1,65 @@
> +/*
> + * Copyright (C) 2011 Tobias Klauser <[email protected]>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation; either version 2 of the License, or
> + * (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
> + *
> + */
> +
> +#ifndef _ASM_NIOS2_REGISTERS_H
> +#define _ASM_NIOS2_REGISTERS_H
> +
> +/* control register numbers */
> +#define CTL_STATUS 0
> +#define CTL_ESTATUS 1
> +#define CTL_BSTATUS 2
> +#define CTL_IENABLE 3
> +#define CTL_IPENDING 4
> +#define CTL_CPUID 5
> +#define CTL_RSV1 6
> +#define CTL_EXCEPTION 7
> +#define CTL_PTEADDR 8
> +#define CTL_TLBACC 9
> +#define CTL_TLBMISC 10
> +#define CTL_RSV2 11
> +#define CTL_BADADDR 12
> +#define CTL_CONFIG 13
> +#define CTL_MPUBASE 14
> +#define CTL_MPUACC 15
> +
> +/* access control registers using GCC builtins */
> +#define RDCTL(r) __builtin_rdctl(r)
> +#define WRCTL(r, v) __builtin_wrctl(r, v)
> +
> +/* status register bits */
> +#define STATUS_PIE (1 << 0) /* processor interrupt enable */
> +#define STATUS_U (1 << 1) /* user mode */
> +#define STATUS_EH (1 << 2) /* Exception mode */
> +
> +/* estatus register bits */
> +#define ESTATUS_EPIE (1 << 0) /* processor interrupt enable */
> +#define ESTATUS_EU (1 << 1) /* user mode */
> +#define ESTATUS_EH (1 << 2) /* Exception mode */
> +
> +/* tlbmisc register bits */
> +#define TLBMISC_PID_SHIFT 4
> +#define TLBMISC_PID_MASK ((1UL << cpuinfo.tlb_pid_num_bits) - 1)

You should probably #include <asm/cpuinfo.h> here since struct cpuinfo
is used in the definition.

> +#define TLBMISC_WAY_MASK 0xf
> +#define TLBMISC_WAY_SHIFT 20
> +
> +#define TLBMISC_PID (TLBMISC_PID_MASK << TLBMISC_PID_SHIFT) /* TLB PID */
> +#define TLBMISC_WE (1 << 18) /* TLB write enable */
> +#define TLBMISC_RD (1 << 19) /* TLB read */
> +#define TLBMISC_WAY (TLBMISC_WAY_MASK << TLBMISC_WAY_SHIFT) /* TLB way */
> +
> +#endif /* _ASM_NIOS2_REGISTERS_H */
> --
> 1.8.3.2
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-arch" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>

2014-04-22 13:42:28

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 19/28] nios2: Device tree support

On Friday 18 April 2014, Ley Foon Tan wrote:
> diff --git a/arch/nios2/boot/dts/3c120_devboard.dts b/arch/nios2/boot/dts/3c120_devboard.dts
> new file mode 100644
> index 0000000..cad29a9
> --- /dev/null
> +++ b/arch/nios2/boot/dts/3c120_devboard.dts
> @@ -0,0 +1,205 @@
> +/*
> + * Copyright (C) 2013 Altera Corporation
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation; either version 2 of the License, or
> + * (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
> + *
> + * This file is generated by sopc2dts.
> + */
> +
> +/dts-v1/;
> +
> +/ {
> + model = "ALTR,qsys_ghrd_3c120";
> + compatible = "ALTR,qsys_ghrd_3c120";

You have a mix of "ALTR" and "altr" prefixes. The general recommendation
is to use lower-case letters, which is also what is used on ARM socfpga,
and what is documented in Documentation/devicetree/bindings/vendor-prefixes.txt
for Altera.

> + sopc@0 {
> + device_type = "soc";
> + ranges;
> + #address-cells = < 1 >;
> + #size-cells = < 1 >;
> + compatible = "ALTR,avalon", "simple-bus";
> + bus-frequency = < 125000000 >;
> +
> + pb_cpu_to_io: bridge@0x8000000 {
> + compatible = "simple-bus";
> + reg = < 0x08000000 0x00800000 >;

Are these all synthesized devices, or is there also some hardwired
logic? It often makes sense to split out the reusable parts into
a separate .dtsi file that gets included by every implementation.

> + #address-cells = < 1 >;
> + #size-cells = < 1 >;
> + ranges = < 0x00400000 0x08400000 0x00000020
> + 0x00004D40 0x08004D40 0x00000008
> + 0x00004D50 0x08004D50 0x00000008
> + 0x00004000 0x08004000 0x00000400
> + 0x00004400 0x08004400 0x00000040
> + 0x00004800 0x08004800 0x00000040
> + 0x00002000 0x08002000 0x00002000
> + 0x00004C80 0x08004C80 0x00000020
> + 0x00004CC0 0x08004CC0 0x00000010
> + 0x00004CE0 0x08004CE0 0x00000010
> + 0x00004D00 0x08004D00 0x00000010 >;

A few style comments:

- no whitespace in the after '<' or before '>
- put each entry into its own '<...>' group.
- lower-case characters for hex digits
- The ranges should reflect what the bus actually translates,
which is typically not individual bytes but rather whole
address ranges.
- sort numerically.

The above could look like

ranges = <0x00000000 0x08000000 0x00010000>,
<0x00400000 0x08400000 0x00001000>;

> + timer_1ms: timer@0x400000 {
> + compatible = "ALTR,timer-1.0";
> +
> + sysid: sysid@0x4d40 {
> + compatible = "ALTR,sysid-1.0";
> + reg = < 0x00004D40 0x00000008 >;

> + jtag_uart: serial@0x4d50 {
> + compatible = "ALTR,juart-1.0";
> + reg = < 0x00004D50 0x00000008 >;
> +
> + tse_mac: ethernet@0x4000 {
> + compatible = "ALTR,tse-1.0";


Does each one of these have a binding document in
Documentation/devicetree/bindings?

I've looked only at the tse binding, which you seem to be
violating in a few places:

- compatible string is "ALTR,tse-1.0", not "altr,tse-1.0"

> + reg = < 0x00004000 0x00000400
> + 0x00004400 0x00000040
> + 0x00004800 0x00000040
> + 0x00002000 0x00002000 >;
> + reg-names = "control_port", "rx_csr", "tx_csr", "s1";

- wrong order, missing "tx_desc" and "rx_desc" entries

> + rx-fifo-depth = < 8192 >;
> + tx-fifo-depth = < 8192 >;
> + address-bits = < 48 >;

address-bits is not documented

> + max-frame-size = < 1518 >;
> + local-mac-address = [ 02 00 00 00 00 00 ];
> + phy-mode = "rgmii-id";
> + ALTR,mii-id = < 0 >;

ALTR,mii-id is not documented, and required "phy-addr" is missing.


> diff --git a/arch/nios2/boot/linked_dtb.S b/arch/nios2/boot/linked_dtb.S
> new file mode 100644
> index 0000000..071f922db338e2cb4064bc77bf346f50e584d04f
> --- /dev/null
> +++ b/arch/nios2/boot/linked_dtb.S
> + */
> +.section .dtb.init.rodata,"a"
> +.incbin "arch/nios2/boot/system.dtb"

Linking in the dtb file is really against the point of device trees.
You should require boot loaders to pass the dtb seperately.

> + soc_dev_attr->soc_id = kasprintf(GFP_KERNEL, "%u", NIOS2_ID_DEFAULT);
> + soc_dev_attr->revision = kasprintf(GFP_KERNEL, "%d",
> + NIOS2_REVISION_DEFAULT);

These are hardcoded constants. If there is no way to identify the
hardware from looking at the registers, better don't fill these at
all.

> + soc_dev = soc_device_register(soc_dev_attr);
> + if (IS_ERR_OR_NULL(soc_dev)) {

Never use IS_ERR_OR_NULL().

If an interface can return an error code, you should rely on
never seeing NULL, or treating it as a valid pointer.

> +static int __init nios2_device_probe(void)
> +{
> + nios2_soc_device_init();
> +
> + of_platform_bus_probe(NULL, altera_of_bus_ids, NULL);
> + return 0;
> +}

This function can get merged into nios2_soc_device_init.

Arnd

2014-04-22 13:45:15

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 20/28] nios2: Time keeping

On Friday 18 April 2014, Ley Foon Tan wrote:
> +static inline unsigned long read_timersnapshot(void)
> +{
> + unsigned long count;
> +
> + outw(0, timer_membase + ALTERA_TIMER_SNAPL_REG);
> + count =
> + inw(timer_membase + ALTERA_TIMER_SNAPH_REG) << 16 |
> + inw(timer_membase + ALTERA_TIMER_SNAPL_REG);
> +
> + return count;
> +}
> +
> +static inline void write_timerperiod(unsigned long period)
> +{
> + outw(period, timer_membase + ALTERA_TIMER_PERIODL_REG);
> + outw(period >> 16, timer_membase + ALTERA_TIMER_PERIODH_REG);
> +}

It's wrong to use 'outw' if this device is not on an PC-style ISA bus,
which I assume it is not. Since timer_membase is an __iomem pointer,
I'm sure you want to use writew() instead.

Arnd

2014-04-22 13:52:09

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 14/28] nios2: DMA mapping API

On Friday 18 April 2014, Ley Foon Tan wrote:

> +static inline int dma_supported(struct device *dev, u64 mask)
> +{
> + /* we fall back to GFP_DMA when the mask isn't all 1s,
> + * so we can't guarantee allocations that must be
> + * within a tighter range than GFP_DMA.
> + */
> + if (mask < 0x00ffffff)
> + return 0;
> +
> + return 1;
> +}

> +void *dma_alloc_coherent(struct device *dev, size_t size,
> + dma_addr_t *dma_handle, gfp_t gfp)
> +{
> + void *ret;
> +
> + /* ignore region specifiers */
> + gfp &= ~(__GFP_DMA | __GFP_HIGHMEM);
> +
> + /* optimized page clearing */
> + gfp |= __GFP_ZERO;
> +
> + if (dev == NULL || (dev->coherent_dma_mask < 0xffffffff))
> + gfp |= GFP_DMA;

What is the significance of ZONE_DMA on this architecture?

Do you actually have DMA masters with a 0x00ffffff mask?

Arnd

2014-04-22 13:59:51

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 07/28] nios2: I/O Mapping

On Friday 18 April 2014, Ley Foon Tan wrote:

> +
> +#include <asm/pgtable-bits.h>
> +
> +#define IO_SPACE_LIMIT 0xffffffff

Please use 0xffff here, this should work for almost any PCI bus.

> +
> +#ifndef CONFIG_CC_OPTIMIZE_FOR_SIZE
> +# define __IO_USE_DUFFS
> +#endif
> +
> +#ifdef __IO_USE_DUFFS
> +
> +/* Use "Duff's Device" to unroll the loops. */
> +#define __IO_OUT_LOOP(a, b, l) \
> + do { \
> + if (l > 0) { \
> + int _n = (l + 7) / 8; \
> + switch (l % 8) { \
> + case 0: \
> + do { \
> + *a = *b++; \

I would recommend just doing all of this in out-of-line implementations
rather than macros and inline functions.

> +
> +/*
> + * make the short names macros so specific devices
> + * can override them as required
> + */
> +#define inb(addr) readb(addr)
> +#define inw(addr) readw(addr)
> +#define inl(addr) readl(addr)
> +#define outb(x, addr) ((void) writeb(x, addr))
> +#define outw(x, addr) ((void) writew(x, addr))
> +#define outl(x, addr) ((void) writel(x, addr))

This makes no sense: the legacy PC I/O accessors take a completely
different argument type: I would recommend to make both inline
functions so you can ensure that readl() gets passed an __iomem
pointer, while inl() gets an integer number.

Please see the asm-generic version for an example. You should also
define a non-NULL PCI_IOBASE to which the PCI I/O space gets mapped.

> +#define virt_to_bus virt_to_phys
> +#define bus_to_virt phys_to_virt

Please drop these, and the CONFIG_VIRT_TO_BUS Kconfig option,
and fix all drivers that rely on it.

> +#define ioport_map(port, nr) ioremap(port, nr)
> +#define ioport_unmap(port) iounmap(port)

Use this one instead:

#ifdef CONFIG_PCI
static inline void __iomem *ioport_map(unsigned long port, unsigned int nr)
{
return PCI_IOBASE + port;
}

static inline void __iomem *ioport_unmap(void __iomem *)
{
}
#endif

> +/* Macros used for smc91x.c driver */
> +#define readsb(p, d, l) insb(p, d, l)
> +#define readsw(p, d, l) insw(p, d, l)
> +#define readsl(p, d, l) insl(p, d, l)
> +#define writesb(p, d, l) outsb(p, d, l)
> +#define writesw(p, d, l) outsw(p, d, l)
> +#define writesl(p, d, l) outsl(p, d, l)

These should of course not fall back to the PCI I/O space functions.
You can do it the other way round.

Arnd

2014-04-22 14:25:16

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

Hi Ley Foon,

On Apr 18, Ley Foon Tan wrote:
> +/*
> + * PAGE_SHIFT determines the page size
> + */
> +#define PAGE_SHIFT 12
> +#define PAGE_SIZE 4096
> +#define PAGE_MASK (~(PAGE_SIZE - 1))
> +

How about something like this:

/* PAGE_SHIFT determines the page size */
#define PAGE_SHIFT 12
#define PAGE_SIZE (_AC(1,UL) << PAGE_SHIFT)
#define PAGE_MASK (~((1 << PAGE_SHIFT) - 1))

Otherwise, the PAGE_SIZE macro above produces some warnings, IIRC.

--
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-22 14:29:07

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 05/28] nios2: Traps exception handling

Hello Ley Foon,

On Apr 19, Ley Foon Tan wrote:
> +
> +#ifndef CONFIG_ALIGNMENT_TRAP
> +/* Alignment exception handler */
> +asmlinkage void handle_unaligned_c(struct pt_regs *fp, int cause)
> +{
> + unsigned long addr = RDCTL(CTL_BADADDR);
> +
> + cause >>= 2;
> + fp->ea -= 4;
> +
> + if (fixup_exception(fp))
> + return;

This will throw an undeclared error if you build with CONFIG_ALIGNMENT_TRAP=n.

You need to add the uaccess.h header, or implement some similar fix.
--
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-22 14:31:27

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 08/28] nios2: MMU Fault handling

Hello Ley Foon,

On Apr 18, Ley Foon Tan wrote:
> +
> +bad_area_nosemaphore:
> + /* User mode accesses just cause a SIGSEGV */
> + if (user_mode(regs)) {

I found that it's useful to add some printing here, just as ARM
does. I carry this patch on my kernel:

+ printk(KERN_INFO "%s: unhandled page fault (%d) at 0x%08lx, cause %ld\n",
+ current->comm, SIGSEGV, address, cause);
+ show_regs(regs);

Do you think we could do something like it? Maybe with a compile time option?

> + _exception(SIGSEGV, regs, code, address);
> + return;
> + }
> +
--
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-22 14:50:18

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>> Hi Arnd and Peter Anvin,
>>>
>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>> that we don't need to have 64 bit off_t? See detail in link below.
>>> I can submit the patches for 64-bit time changes
>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>> agreed on this.
>>
>> Yes.
> Okay, will doing that.
>
>>
>>> Excerpt from https://lkml.org/lkml/2012/11/14/358 :
>>> "Obviously, we want to use 64-bit off_t, but this is achieved already
>>> through loff_t, which is used in all places in the asm-generic
>>> ABI anyway (the syscalls using off_t are stripped out). I don't
>>> think we want to have the other ones set to 64 bit on ARC or Meta,
>>> although I'm not 100% sure about ino_t and nlink_t. "
>>
>> This is all still true. You should have no syscall using 'off_t',
>> only loff_t.
>>
>> I still don't know whether we would want 32 or 64 bit ino_t and nlink_t
>> for new architectures. It seems it would gain very little, but have
>> a noticeable overhead.
> Anyone have comment on this?
> Chung-Lin (in CC list) is our nios2 toolchain maintainer. Do you have
> any comment for 32 or 64 bit ino_t and nlink_t?
> We will update the toolchain to support 64-bit time_t, so we hope that
> any other toolchain change can happen in one time.

For ino_t, 32-bit users of linux-generic glibc already use struct
stat64, stat64(), etc. to align with 64-bit targets, so in terms of the
glibc/kernel interface it doesn't matter much. The in-kernel usage of
the ino_t type should be of more concern here.

nlink_t appears to be always defined as u32 in <linux/types.h>, not sure
if changing it to arch-overridable is reasonable.

Chung-Lin


2014-04-22 15:15:07

by Tobias Klauser

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

On 2014-04-22 at 16:24:43 +0200, Ezequiel Garcia <[email protected]> wrote:
> Hi Ley Foon,
>
> On Apr 18, Ley Foon Tan wrote:
> > +/*
> > + * PAGE_SHIFT determines the page size
> > + */
> > +#define PAGE_SHIFT 12
> > +#define PAGE_SIZE 4096
> > +#define PAGE_MASK (~(PAGE_SIZE - 1))
> > +
>
> How about something like this:
>
> /* PAGE_SHIFT determines the page size */
> #define PAGE_SHIFT 12
> #define PAGE_SIZE (_AC(1,UL) << PAGE_SHIFT)
> #define PAGE_MASK (~((1 << PAGE_SHIFT) - 1))
>
> Otherwise, the PAGE_SIZE macro above produces some warnings, IIRC.

AFAIR old nios2 GCC versions (the 4.1 version from Windriver, IIRC)
would complain about something like the above, that's the reason I added
an explicit value for PAGE_SIZE back then.

I assume with the new nios2 GCC this should no longer be necessary and
Ezequiel's fix should be fine.

2014-04-22 15:36:07

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

On Apr 22, Tobias Klauser wrote:
> On 2014-04-22 at 16:24:43 +0200, Ezequiel Garcia <[email protected]> wrote:
> > Hi Ley Foon,
> >
> > On Apr 18, Ley Foon Tan wrote:
> > > +/*
> > > + * PAGE_SHIFT determines the page size
> > > + */
> > > +#define PAGE_SHIFT 12
> > > +#define PAGE_SIZE 4096
> > > +#define PAGE_MASK (~(PAGE_SIZE - 1))
> > > +
> >
> > How about something like this:
> >
> > /* PAGE_SHIFT determines the page size */
> > #define PAGE_SHIFT 12
> > #define PAGE_SIZE (_AC(1,UL) << PAGE_SHIFT)
> > #define PAGE_MASK (~((1 << PAGE_SHIFT) - 1))
> >
> > Otherwise, the PAGE_SIZE macro above produces some warnings, IIRC.
>
> AFAIR old nios2 GCC versions (the 4.1 version from Windriver, IIRC)
> would complain about something like the above, that's the reason I added
> an explicit value for PAGE_SIZE back then.
>

Other than being "pretty" the above fix is to remove a mismatch type
warning. You can get rid of the warning in different ways:

ifdef __ASSEMBLY__
define PAGE_SIZE 4096
else
define PAGE_SIZE 4096UL
endif

> I assume with the new nios2 GCC this should no longer be necessary and
> Ezequiel's fix should be fine.

Speaking of GCC... is there a libc available to test this kernel port?
(which I assume uses the generic syscall ABI)
--
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-22 16:01:36

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

On 2014/4/22 下午 11:35, Ezequiel Garcia wrote:
> On Apr 22, Tobias Klauser wrote:
>> On 2014-04-22 at 16:24:43 +0200, Ezequiel Garcia <[email protected]> wrote:
>>> Hi Ley Foon,
>>>
>>> On Apr 18, Ley Foon Tan wrote:
>>>> +/*
>>>> + * PAGE_SHIFT determines the page size
>>>> + */
>>>> +#define PAGE_SHIFT 12
>>>> +#define PAGE_SIZE 4096
>>>> +#define PAGE_MASK (~(PAGE_SIZE - 1))
>>>> +
>>>
>>> How about something like this:
>>>
>>> /* PAGE_SHIFT determines the page size */
>>> #define PAGE_SHIFT 12
>>> #define PAGE_SIZE (_AC(1,UL) << PAGE_SHIFT)
>>> #define PAGE_MASK (~((1 << PAGE_SHIFT) - 1))
>>>
>>> Otherwise, the PAGE_SIZE macro above produces some warnings, IIRC.
>>
>> AFAIR old nios2 GCC versions (the 4.1 version from Windriver, IIRC)
>> would complain about something like the above, that's the reason I added
>> an explicit value for PAGE_SIZE back then.
>>
>
> Other than being "pretty" the above fix is to remove a mismatch type
> warning. You can get rid of the warning in different ways:
>
> ifdef __ASSEMBLY__
> define PAGE_SIZE 4096
> else
> define PAGE_SIZE 4096UL
> endif
>
>> I assume with the new nios2 GCC this should no longer be necessary and
>> Ezequiel's fix should be fine.
>
> Speaking of GCC... is there a libc available to test this kernel port?
> (which I assume uses the generic syscall ABI)

It's submitted to libc-alpha, also still review in progress.

Chung-Lin


2014-04-22 16:24:48

by Sam Ravnborg

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

On Tue, Apr 22, 2014 at 12:35:22PM -0300, Ezequiel Garcia wrote:
> On Apr 22, Tobias Klauser wrote:
> > On 2014-04-22 at 16:24:43 +0200, Ezequiel Garcia <[email protected]> wrote:
> > > Hi Ley Foon,
> > >
> > > On Apr 18, Ley Foon Tan wrote:
> > > > +/*
> > > > + * PAGE_SHIFT determines the page size
> > > > + */
> > > > +#define PAGE_SHIFT 12
> > > > +#define PAGE_SIZE 4096
> > > > +#define PAGE_MASK (~(PAGE_SIZE - 1))
> > > > +
> > >
> > > How about something like this:
> > >
> > > /* PAGE_SHIFT determines the page size */
> > > #define PAGE_SHIFT 12
> > > #define PAGE_SIZE (_AC(1,UL) << PAGE_SHIFT)
> > > #define PAGE_MASK (~((1 << PAGE_SHIFT) - 1))
> > >
> > > Otherwise, the PAGE_SIZE macro above produces some warnings, IIRC.
> >
> > AFAIR old nios2 GCC versions (the 4.1 version from Windriver, IIRC)
> > would complain about something like the above, that's the reason I added
> > an explicit value for PAGE_SIZE back then.
> >
>
> Other than being "pretty" the above fix is to remove a mismatch type
> warning. You can get rid of the warning in different ways:
>
> ifdef __ASSEMBLY__
> define PAGE_SIZE 4096
> else
> define PAGE_SIZE 4096UL
> endif

The usual way to do this is as follows:
#include <linux/const.h>

#define PAGE_SHIFT 12
#define PAGE_SIZE (_AC(1, UL) << PAGE_SHIFT)
#define PAGE_MASK (~(PAGE_SIZE-1))

_AC(1, UL) expands to 1 for assembler and 1UL for C code.

This is also what was suggested above.

Please follow this scheme for nios2 too.

Sam

2014-04-22 16:28:33

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

On Apr 23, Chung-Lin Tang wrote:
> On 2014/4/22 下午 11:35, Ezequiel Garcia wrote:
> > On Apr 22, Tobias Klauser wrote:
> >> On 2014-04-22 at 16:24:43 +0200, Ezequiel Garcia <[email protected]> wrote:
> >>> Hi Ley Foon,
> >>>
> >>> On Apr 18, Ley Foon Tan wrote:
> >>>> +/*
> >>>> + * PAGE_SHIFT determines the page size
> >>>> + */
> >>>> +#define PAGE_SHIFT 12
> >>>> +#define PAGE_SIZE 4096
> >>>> +#define PAGE_MASK (~(PAGE_SIZE - 1))
> >>>> +
> >>>
> >>> How about something like this:
> >>>
> >>> /* PAGE_SHIFT determines the page size */
> >>> #define PAGE_SHIFT 12
> >>> #define PAGE_SIZE (_AC(1,UL) << PAGE_SHIFT)
> >>> #define PAGE_MASK (~((1 << PAGE_SHIFT) - 1))
> >>>
> >>> Otherwise, the PAGE_SIZE macro above produces some warnings, IIRC.
> >>
> >> AFAIR old nios2 GCC versions (the 4.1 version from Windriver, IIRC)
> >> would complain about something like the above, that's the reason I added
> >> an explicit value for PAGE_SIZE back then.
> >>
> >
> > Other than being "pretty" the above fix is to remove a mismatch type
> > warning. You can get rid of the warning in different ways:
> >
> > ifdef __ASSEMBLY__
> > define PAGE_SIZE 4096
> > else
> > define PAGE_SIZE 4096UL
> > endif
> >
> >> I assume with the new nios2 GCC this should no longer be necessary and
> >> Ezequiel's fix should be fine.
> >
> > Speaking of GCC... is there a libc available to test this kernel port?
> > (which I assume uses the generic syscall ABI)
>
> It's submitted to libc-alpha, also still review in progress.
>

Do you plan to support uclibc?

--
Ezequiel García, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-22 16:36:21

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

On 2014/4/23 上午 12:27, Ezequiel Garcia wrote:
> On Apr 23, Chung-Lin Tang wrote:
>> On 2014/4/22 下午 11:35, Ezequiel Garcia wrote:
>>> On Apr 22, Tobias Klauser wrote:
>>>> On 2014-04-22 at 16:24:43 +0200, Ezequiel Garcia <[email protected]> wrote:
>>>>> Hi Ley Foon,
>>>>>
>>>>> On Apr 18, Ley Foon Tan wrote:
>>>>>> +/*
>>>>>> + * PAGE_SHIFT determines the page size
>>>>>> + */
>>>>>> +#define PAGE_SHIFT 12
>>>>>> +#define PAGE_SIZE 4096
>>>>>> +#define PAGE_MASK (~(PAGE_SIZE - 1))
>>>>>> +
>>>>>
>>>>> How about something like this:
>>>>>
>>>>> /* PAGE_SHIFT determines the page size */
>>>>> #define PAGE_SHIFT 12
>>>>> #define PAGE_SIZE (_AC(1,UL) << PAGE_SHIFT)
>>>>> #define PAGE_MASK (~((1 << PAGE_SHIFT) - 1))
>>>>>
>>>>> Otherwise, the PAGE_SIZE macro above produces some warnings, IIRC.
>>>>
>>>> AFAIR old nios2 GCC versions (the 4.1 version from Windriver, IIRC)
>>>> would complain about something like the above, that's the reason I added
>>>> an explicit value for PAGE_SIZE back then.
>>>>
>>>
>>> Other than being "pretty" the above fix is to remove a mismatch type
>>> warning. You can get rid of the warning in different ways:
>>>
>>> ifdef __ASSEMBLY__
>>> define PAGE_SIZE 4096
>>> else
>>> define PAGE_SIZE 4096UL
>>> endif
>>>
>>>> I assume with the new nios2 GCC this should no longer be necessary and
>>>> Ezequiel's fix should be fine.
>>>
>>> Speaking of GCC... is there a libc available to test this kernel port?
>>> (which I assume uses the generic syscall ABI)
>>
>> It's submitted to libc-alpha, also still review in progress.
>>
>
> Do you plan to support uclibc?

No, we (Mentor Graphics) are only doing glibc work at this point.

Chung-Lin

2014-04-23 02:47:12

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 04/28] nios2: Exception handling

On Tue, Apr 22, 2014 at 8:33 PM, Arnd Bergmann <[email protected]> wrote:
> On Friday 18 April 2014, Ley Foon Tan wrote:
>> +static const struct file_operations misalign_fops = {
>> + .open = proc_misaligned_open,
>> + .read = seq_read,
>> + .llseek = seq_lseek,
>> + .release = single_release,
>> + .write = proc_misaligned_write,
>> +};
>> +#endif /* CONFIG_PROC_FS */
>
> We really shouldn't be doing new architecture specific procfs files
> any more. I suggest you drop this one for now, and add back the
> functionality using perf or ftrace at a later point.
>
> Arnd

Okay, will remove this.
Thanks.

Ley Foon

2014-04-23 02:53:16

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

On Wed, Apr 23, 2014 at 12:24 AM, Sam Ravnborg <[email protected]> wrote:
>> Other than being "pretty" the above fix is to remove a mismatch type
>> warning. You can get rid of the warning in different ways:
>>
>> ifdef __ASSEMBLY__
>> define PAGE_SIZE 4096
>> else
>> define PAGE_SIZE 4096UL
>> endif
>
> The usual way to do this is as follows:
> #include <linux/const.h>
>
> #define PAGE_SHIFT 12
> #define PAGE_SIZE (_AC(1, UL) << PAGE_SHIFT)
> #define PAGE_MASK (~(PAGE_SIZE-1))
>
> _AC(1, UL) expands to 1 for assembler and 1UL for C code.
>
> This is also what was suggested above.
>
> Please follow this scheme for nios2 too.
>
> Sam
Sure, will change to this way.

Thanks.

Ley Foon

2014-04-23 03:21:52

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 20/28] nios2: Time keeping

On Tue, Apr 22, 2014 at 9:44 PM, Arnd Bergmann <[email protected]> wrote:
> On Friday 18 April 2014, Ley Foon Tan wrote:
>> +static inline unsigned long read_timersnapshot(void)
>> +{
>> + unsigned long count;
>> +
>> + outw(0, timer_membase + ALTERA_TIMER_SNAPL_REG);
>> + count =
>> + inw(timer_membase + ALTERA_TIMER_SNAPH_REG) << 16 |
>> + inw(timer_membase + ALTERA_TIMER_SNAPL_REG);
>> +
>> + return count;
>> +}
>> +
>> +static inline void write_timerperiod(unsigned long period)
>> +{
>> + outw(period, timer_membase + ALTERA_TIMER_PERIODL_REG);
>> + outw(period >> 16, timer_membase + ALTERA_TIMER_PERIODH_REG);
>> +}
>
> It's wrong to use 'outw' if this device is not on an PC-style ISA bus,
> which I assume it is not. Since timer_membase is an __iomem pointer,
> I'm sure you want to use writew() instead.
>
> Arnd
Sure, will change all outw() to writew() and inw() to readw().

Ley Foon

2014-04-23 06:52:51

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 19/28] nios2: Device tree support

On Tue, Apr 22, 2014 at 9:42 PM, Arnd Bergmann <[email protected]> wrote:
> On Friday 18 April 2014, Ley Foon Tan wrote:
>> diff --git a/arch/nios2/boot/dts/3c120_devboard.dts b/arch/nios2/boot/dts/3c120_devboard.dts

>> +/dts-v1/;
>> +
>> +/ {
>> + model = "ALTR,qsys_ghrd_3c120";
>> + compatible = "ALTR,qsys_ghrd_3c120";
>
> You have a mix of "ALTR" and "altr" prefixes. The general recommendation
> is to use lower-case letters, which is also what is used on ARM socfpga,
> and what is documented in Documentation/devicetree/bindings/vendor-prefixes.txt
> for Altera.
Yes, the vendor prefix should be changed to lower case. FYI, this dts
file is generated by our dts generator tool called sopc2dts.
Our tool already aware of this requirement, but haven't support this yet.
I can edit this file manually to change 'ALTR' to 'altr', but I will
update nios2 code to match for both "ALTR" and "altr" for backward
compatibility. "ALTR" will be deprecated later. Are you okay with
this?


>
>> + sopc@0 {
>> + device_type = "soc";
>> + ranges;
>> + #address-cells = < 1 >;
>> + #size-cells = < 1 >;
>> + compatible = "ALTR,avalon", "simple-bus";
>> + bus-frequency = < 125000000 >;
>> +
>> + pb_cpu_to_io: bridge@0x8000000 {
>> + compatible = "simple-bus";
>> + reg = < 0x08000000 0x00800000 >;
>
> Are these all synthesized devices, or is there also some hardwired
> logic? It often makes sense to split out the reusable parts into
> a separate .dtsi file that gets included by every implementation.
All these are synthesized devices and the system hierarchy is also not
fixed and it is depends on user hardware design. It is highly
configurable.
So, we can't have a common .dtsi file.

>
>> + #address-cells = < 1 >;
>> + #size-cells = < 1 >;
>> + ranges = < 0x00400000 0x08400000 0x00000020
>> + 0x00004D40 0x08004D40 0x00000008
>> + 0x00004D50 0x08004D50 0x00000008
>> + 0x00004000 0x08004000 0x00000400
>> + 0x00004400 0x08004400 0x00000040
>> + 0x00004800 0x08004800 0x00000040
>> + 0x00002000 0x08002000 0x00002000
>> + 0x00004C80 0x08004C80 0x00000020
>> + 0x00004CC0 0x08004CC0 0x00000010
>> + 0x00004CE0 0x08004CE0 0x00000010
>> + 0x00004D00 0x08004D00 0x00000010 >;
>
> A few style comments:
>
> - no whitespace in the after '<' or before '>
> - put each entry into its own '<...>' group.
> - lower-case characters for hex digits
Okay, I will change this and I will feedback this to our dts generator
tool team to support these format in future.

> - The ranges should reflect what the bus actually translates,
> which is typically not individual bytes but rather whole
> address ranges.
The ranges here reflect the address range translate by each device and
user can assign any base address they desired in our FPGA. The address
ranges might not in continuous regions as well. So, we will keep this
translation way.

> - sort numerically.
>
> The above could look like
>
> ranges = <0x00000000 0x08000000 0x00010000>,
> <0x00400000 0x08400000 0x00001000>;
Okay, I will sort this.

>
>> + timer_1ms: timer@0x400000 {
>> + compatible = "ALTR,timer-1.0";
>> +
>> + sysid: sysid@0x4d40 {
>> + compatible = "ALTR,sysid-1.0";
>> + reg = < 0x00004D40 0x00000008 >;
>
>> + jtag_uart: serial@0x4d50 {
>> + compatible = "ALTR,juart-1.0";
>> + reg = < 0x00004D50 0x00000008 >;
>> +
>> + tse_mac: ethernet@0x4000 {
>> + compatible = "ALTR,tse-1.0";
>
>
> Does each one of these have a binding document in
> Documentation/devicetree/bindings?
jtag_uart and tse drivers are upstreamed. They already have their own
bindings doc:
Documentation/devicetree/bindings/serial/altera_jtaguart.txt
Documentation/devicetree/bindings/net/altera_tse.txt

I will add the binding doc for timer, but sysid driver is not in
mainline kernel yet. I will remove sysid entry from this file.

>
> I've looked only at the tse binding, which you seem to be
> violating in a few places:
>
> - compatible string is "ALTR,tse-1.0", not "altr,tse-1.0"
I will change to "altr,tse-1.0".

>
>> + reg = < 0x00004000 0x00000400
>> + 0x00004400 0x00000040
>> + 0x00004800 0x00000040
>> + 0x00002000 0x00002000 >;
>> + reg-names = "control_port", "rx_csr", "tx_csr", "s1";
>
> - wrong order, missing "tx_desc" and "rx_desc" entries
FYI, TSE driver supports both SGDMA and MSGDMA soft IP and "tx_desc"
and "rx_desc" entries are for MSGDMA (see below). But this 3c120
hardware design is using TSE + SGDMA. So, the expect entries are
"control_port", "rx_csr", "tx_csr", "s1".
I can sort the entries order by their addresses.

Excerpt from Documentation/devicetree/bindings/net/altera_tse.txt
- reg-names: Should contain the reg names
"control_port": MAC configuration space region
"tx_csr": xDMA Tx dispatcher control and status space region
"tx_desc": MSGDMA Tx dispatcher descriptor space region
"rx_csr" : xDMA Rx dispatcher control and status space region
"rx_desc": MSGDMA Rx dispatcher descriptor space region
"rx_resp": MSGDMA Rx dispatcher response space region
"s1": SGDMA descriptor memory


>
>> + rx-fifo-depth = < 8192 >;
>> + tx-fifo-depth = < 8192 >;
>> + address-bits = < 48 >;
>
> address-bits is not documented
Will remove this.

>
>> + max-frame-size = < 1518 >;
>> + local-mac-address = [ 02 00 00 00 00 00 ];
>> + phy-mode = "rgmii-id";
>> + ALTR,mii-id = < 0 >;
>
> ALTR,mii-id is not documented, and required "phy-addr" is missing.
Will remove ALTR,mii-id. "phy-addr" is not required if we provide "phy-handle".

Excerpt from Documentation/devicetree/bindings/net/altera_tse.txt:
- phy-addr: See ethernet.txt in the same directory. A configuration should
include phy-handle or phy-addr.


>
>
>> diff --git a/arch/nios2/boot/linked_dtb.S b/arch/nios2/boot/linked_dtb.S
>> new file mode 100644
>> index 0000000..071f922db338e2cb4064bc77bf346f50e584d04f
>> --- /dev/null
>> +++ b/arch/nios2/boot/linked_dtb.S
>> + */
>> +.section .dtb.init.rodata,"a"
>> +.incbin "arch/nios2/boot/system.dtb"
>
> Linking in the dtb file is really against the point of device trees.
> You should require boot loaders to pass the dtb seperately.
We do support pass the dtb from boot loaders. This is optional feature
that allow user to linking in dtb into kernel image if
CONFIG_DTB_SOURCE_BOOL is enabled. This is very useful for develpment
stage/debugging stage, where we can download vmlinux image directly
into SDRAM and boot without boot loader.
Noticed that other architectures have linking in dtd file as well, eg:
c6x and microblaze.


>
>> + soc_dev_attr->soc_id = kasprintf(GFP_KERNEL, "%u", NIOS2_ID_DEFAULT);
>> + soc_dev_attr->revision = kasprintf(GFP_KERNEL, "%d",
>> + NIOS2_REVISION_DEFAULT);
>
> These are hardcoded constants. If there is no way to identify the
> hardware from looking at the registers, better don't fill these at
> all.
Okay, I will leave them empty.
>
>> + soc_dev = soc_device_register(soc_dev_attr);
>> + if (IS_ERR_OR_NULL(soc_dev)) {
>
> Never use IS_ERR_OR_NULL().
>
> If an interface can return an error code, you should rely on
> never seeing NULL, or treating it as a valid pointer.
Okay, will change it to IS_ERR().

>
>> +static int __init nios2_device_probe(void)
>> +{
>> + nios2_soc_device_init();
>> +
>> + of_platform_bus_probe(NULL, altera_of_bus_ids, NULL);
>> + return 0;
>> +}
>
> This function can get merged into nios2_soc_device_init.
Okay, will merge nios2_device_probe() into nios2_soc_device_init().

Thanks.

Regards
Ley Foon

2014-04-23 07:10:30

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 23/28] nios2: Nios2 registers

Hi Tobias


On Tue, Apr 22, 2014 at 8:39 PM, Tobias Klauser <[email protected]> wrote:
> On 2014-04-18 at 14:27:06 +0200, Ley Foon Tan <[email protected]> wrote:

>> +
>> +/* estatus register bits */
>> +#define ESTATUS_EPIE (1 << 0) /* processor interrupt enable */
>> +#define ESTATUS_EU (1 << 1) /* user mode */
>> +#define ESTATUS_EH (1 << 2) /* Exception mode */
>> +
>> +/* tlbmisc register bits */
>> +#define TLBMISC_PID_SHIFT 4
>> +#define TLBMISC_PID_MASK ((1UL << cpuinfo.tlb_pid_num_bits) - 1)
>
> You should probably #include <asm/cpuinfo.h> here since struct cpuinfo
> is used in the definition.
Okay, but I think we should add #ifndef __ASSEMBLY__ at #include
<asm/cpuinfo.h> and at line #define TLBMISC_PID_MASK.
Assembly source file might include this file as well.

>
>> +#define TLBMISC_WAY_MASK 0xf
>> +#define TLBMISC_WAY_SHIFT 20
>> +
>> +#define TLBMISC_PID (TLBMISC_PID_MASK << TLBMISC_PID_SHIFT) /* TLB PID */
>> +#define TLBMISC_WE (1 << 18) /* TLB write enable */
>> +#define TLBMISC_RD (1 << 19) /* TLB read */
>> +#define TLBMISC_WAY (TLBMISC_WAY_MASK << TLBMISC_WAY_SHIFT) /* TLB way */
>> +
>> +#endif /* _ASM_NIOS2_REGISTERS_H */

>>

2014-04-23 07:20:47

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: [PATCH 04/28] nios2: Exception handling

Hi Ley,

On Wed, Apr 23, 2014 at 4:47 AM, Ley Foon Tan <[email protected]> wrote:
> On Tue, Apr 22, 2014 at 8:33 PM, Arnd Bergmann <[email protected]> wrote:
>> On Friday 18 April 2014, Ley Foon Tan wrote:
>>> +static const struct file_operations misalign_fops = {
>>> + .open = proc_misaligned_open,
>>> + .read = seq_read,
>>> + .llseek = seq_lseek,
>>> + .release = single_release,
>>> + .write = proc_misaligned_write,
>>> +};
>>> +#endif /* CONFIG_PROC_FS */
>>
>> We really shouldn't be doing new architecture specific procfs files
>> any more. I suggest you drop this one for now, and add back the
>> functionality using perf or ftrace at a later point.
>
> Okay, will remove this.

MIPS and powerpc handle this through debugfs, cfr.
arch/mips/kernel/unaligned.c and arch/powerpc/kernel/traps.c.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

2014-04-23 07:24:54

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: [PATCH 06/28] nios2: Memory management

Hi Sam,

On Tue, Apr 22, 2014 at 6:24 PM, Sam Ravnborg <[email protected]> wrote:
> #define PAGE_SIZE (_AC(1, UL) << PAGE_SHIFT)
> #define PAGE_MASK (~(PAGE_SIZE-1))
>
> _AC(1, UL) expands to 1 for assembler and 1UL for C code.

Ideally, PAGE_SIZE should be size_t, not unsigned long, avoiding issues
like https://lkml.org/lkml/2014/4/13/78. But currently it's unsigned long
everywhere, so nios2 should be consistent.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

2014-04-23 07:35:38

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 19/28] nios2: Device tree support

On Wednesday 23 April 2014 14:52:47 Ley Foon Tan wrote:
> On Tue, Apr 22, 2014 at 9:42 PM, Arnd Bergmann <[email protected]> wrote:
> > On Friday 18 April 2014, Ley Foon Tan wrote:
> >> diff --git a/arch/nios2/boot/dts/3c120_devboard.dts b/arch/nios2/boot/dts/3c120_devboard.dts
>
> >> +/dts-v1/;
> >> +
> >> +/ {
> >> + model = "ALTR,qsys_ghrd_3c120";
> >> + compatible = "ALTR,qsys_ghrd_3c120";
> >
> > You have a mix of "ALTR" and "altr" prefixes. The general recommendation
> > is to use lower-case letters, which is also what is used on ARM socfpga,
> > and what is documented in Documentation/devicetree/bindings/vendor-prefixes.txt
> > for Altera.
> Yes, the vendor prefix should be changed to lower case. FYI, this dts
> file is generated by our dts generator tool called sopc2dts.
> Our tool already aware of this requirement, but haven't support this yet.
> I can edit this file manually to change 'ALTR' to 'altr', but I will
> update nios2 code to match for both "ALTR" and "altr" for backward
> compatibility. "ALTR" will be deprecated later. Are you okay with
> this?

You seem to have a large number of "ALTR" strings at the moment.
I don't think it would be good to allow matching both variants
everywhere, that would be a lot of extra code.

If it's only about the root compatible property, I have no objections,
but I don't know if it will help you when all the other strings don't
match.

> >> + sopc@0 {
> >> + device_type = "soc";
> >> + ranges;
> >> + #address-cells = < 1 >;
> >> + #size-cells = < 1 >;
> >> + compatible = "ALTR,avalon", "simple-bus";
> >> + bus-frequency = < 125000000 >;
> >> +
> >> + pb_cpu_to_io: bridge@0x8000000 {
> >> + compatible = "simple-bus";
> >> + reg = < 0x08000000 0x00800000 >;
> >
> > Are these all synthesized devices, or is there also some hardwired
> > logic? It often makes sense to split out the reusable parts into
> > a separate .dtsi file that gets included by every implementation.
> All these are synthesized devices and the system hierarchy is also not
> fixed and it is depends on user hardware design. It is highly
> configurable.
> So, we can't have a common .dtsi file.

Ok, I see.

> >> + #address-cells = < 1 >;
> >> + #size-cells = < 1 >;
> >> + ranges = < 0x00400000 0x08400000 0x00000020
> >> + 0x00004D40 0x08004D40 0x00000008
> >> + 0x00004D50 0x08004D50 0x00000008
> >> + 0x00004000 0x08004000 0x00000400
> >> + 0x00004400 0x08004400 0x00000040
> >> + 0x00004800 0x08004800 0x00000040
> >> + 0x00002000 0x08002000 0x00002000
> >> + 0x00004C80 0x08004C80 0x00000020
> >> + 0x00004CC0 0x08004CC0 0x00000010
> >> + 0x00004CE0 0x08004CE0 0x00000010
> >> + 0x00004D00 0x08004D00 0x00000010 >;
> >

> > - The ranges should reflect what the bus actually translates,
> > which is typically not individual bytes but rather whole
> > address ranges.
> The ranges here reflect the address range translate by each device and
> user can assign any base address they desired in our FPGA. The address
> ranges might not in continuous regions as well. So, we will keep this
> translation way.

If you can support any address, why not just have an empty 'ranges'
property?

> >> + timer_1ms: timer@0x400000 {
> >> + compatible = "ALTR,timer-1.0";
> >> +
> >> + sysid: sysid@0x4d40 {
> >> + compatible = "ALTR,sysid-1.0";
> >> + reg = < 0x00004D40 0x00000008 >;
> >
> >> + jtag_uart: serial@0x4d50 {
> >> + compatible = "ALTR,juart-1.0";
> >> + reg = < 0x00004D50 0x00000008 >;
> >> +
> >> + tse_mac: ethernet@0x4000 {
> >> + compatible = "ALTR,tse-1.0";
> >
> >
> > Does each one of these have a binding document in
> > Documentation/devicetree/bindings?
> jtag_uart and tse drivers are upstreamed. They already have their own
> bindings doc:
> Documentation/devicetree/bindings/serial/altera_jtaguart.txt
> Documentation/devicetree/bindings/net/altera_tse.txt
>
> I will add the binding doc for timer, but sysid driver is not in
> mainline kernel yet. I will remove sysid entry from this file.

Ok. BTW, could the sysid be used for the soc device strings, or
is that something else?

> >> + reg = < 0x00004000 0x00000400
> >> + 0x00004400 0x00000040
> >> + 0x00004800 0x00000040
> >> + 0x00002000 0x00002000 >;
> >> + reg-names = "control_port", "rx_csr", "tx_csr", "s1";
> >
> > - wrong order, missing "tx_desc" and "rx_desc" entries
> FYI, TSE driver supports both SGDMA and MSGDMA soft IP and "tx_desc"
> and "rx_desc" entries are for MSGDMA (see below). But this 3c120
> hardware design is using TSE + SGDMA. So, the expect entries are
> "control_port", "rx_csr", "tx_csr", "s1".
> I can sort the entries order by their addresses.

You can't reorder the registers, as they can be accessed by
index in some OS, the strings are just annotations really.

> Excerpt from Documentation/devicetree/bindings/net/altera_tse.txt
> - reg-names: Should contain the reg names
> "control_port": MAC configuration space region
> "tx_csr": xDMA Tx dispatcher control and status space region
> "tx_desc": MSGDMA Tx dispatcher descriptor space region
> "rx_csr" : xDMA Rx dispatcher control and status space region
> "rx_desc": MSGDMA Rx dispatcher descriptor space region
> "rx_resp": MSGDMA Rx dispatcher response space region
> "s1": SGDMA descriptor memory

The important part is

- reg: Address and length of the register set for the device. It contains
the information of registers in the same order as described by reg-names.

You can either provide dummy entries for tx_desc, rx_desc and rx_resp,
or change the binding to reflect the current usage, and provide two
different lists depending on the compatible string.

> >> + max-frame-size = < 1518 >;
> >> + local-mac-address = [ 02 00 00 00 00 00 ];
> >> + phy-mode = "rgmii-id";
> >> + ALTR,mii-id = < 0 >;
> >
> > ALTR,mii-id is not documented, and required "phy-addr" is missing.
> Will remove ALTR,mii-id. "phy-addr" is not required if we provide "phy-handle".
>
> Excerpt from Documentation/devicetree/bindings/net/altera_tse.txt:
> - phy-addr: See ethernet.txt in the same directory. A configuration should
> include phy-handle or phy-addr.

Ok.

> >> diff --git a/arch/nios2/boot/linked_dtb.S b/arch/nios2/boot/linked_dtb.S
> >> new file mode 100644
> >> index 0000000..071f922db338e2cb4064bc77bf346f50e584d04f
> >> --- /dev/null
> >> +++ b/arch/nios2/boot/linked_dtb.S
> >> + */
> >> +.section .dtb.init.rodata,"a"
> >> +.incbin "arch/nios2/boot/system.dtb"
> >
> > Linking in the dtb file is really against the point of device trees.
> > You should require boot loaders to pass the dtb seperately.
> We do support pass the dtb from boot loaders. This is optional feature
> that allow user to linking in dtb into kernel image if
> CONFIG_DTB_SOURCE_BOOL is enabled. This is very useful for develpment
> stage/debugging stage, where we can download vmlinux image directly
> into SDRAM and boot without boot loader.
> Noticed that other architectures have linking in dtd file as well, eg:
> c6x and microblaze.

Ok.

Arnd

2014-04-23 07:48:47

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 28/28] Documentation: Add documentation for Nios2 architecture

On Tue, Apr 22, 2014 at 8:28 PM, Tobias Klauser <[email protected]> wrote:

>> +What is Nios II?
>> +================
>> +Nios II is a 32-bit embedded-processor architecture designed specifically for the
>> +Altera family of FPGAs. In order to support Linux system, Nios II needs to configure
>
> Minor nit: 'In order to support Linux, Nios II needs to be configured...'
Noted.

>
>> +with MMU and hardware multiplier enabled.
>

2014-04-23 08:00:30

by Tobias Klauser

[permalink] [raw]
Subject: Re: [PATCH 23/28] nios2: Nios2 registers

On 2014-04-23 at 09:10:26 +0200, Ley Foon Tan <[email protected]> wrote:
> Hi Tobias
>
>
> On Tue, Apr 22, 2014 at 8:39 PM, Tobias Klauser <[email protected]> wrote:
> > On 2014-04-18 at 14:27:06 +0200, Ley Foon Tan <[email protected]> wrote:
>
> >> +
> >> +/* estatus register bits */
> >> +#define ESTATUS_EPIE (1 << 0) /* processor interrupt enable */
> >> +#define ESTATUS_EU (1 << 1) /* user mode */
> >> +#define ESTATUS_EH (1 << 2) /* Exception mode */
> >> +
> >> +/* tlbmisc register bits */
> >> +#define TLBMISC_PID_SHIFT 4
> >> +#define TLBMISC_PID_MASK ((1UL << cpuinfo.tlb_pid_num_bits) - 1)
> >
> > You should probably #include <asm/cpuinfo.h> here since struct cpuinfo
> > is used in the definition.
> Okay, but I think we should add #ifndef __ASSEMBLY__ at #include
> <asm/cpuinfo.h> and at line #define TLBMISC_PID_MASK.
> Assembly source file might include this file as well.

Yes, makes sense.

> >
> >> +#define TLBMISC_WAY_MASK 0xf
> >> +#define TLBMISC_WAY_SHIFT 20
> >> +
> >> +#define TLBMISC_PID (TLBMISC_PID_MASK << TLBMISC_PID_SHIFT) /* TLB PID */
> >> +#define TLBMISC_WE (1 << 18) /* TLB write enable */
> >> +#define TLBMISC_RD (1 << 19) /* TLB read */
> >> +#define TLBMISC_WAY (TLBMISC_WAY_MASK << TLBMISC_WAY_SHIFT) /* TLB way */
> >> +
> >> +#endif /* _ASM_NIOS2_REGISTERS_H */
>
> >>
>

2014-04-23 09:52:24

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 19/28] nios2: Device tree support

On Wed, Apr 23, 2014 at 3:35 PM, Arnd Bergmann <[email protected]> wrote:

>> > for Altera.
>> Yes, the vendor prefix should be changed to lower case. FYI, this dts
>> file is generated by our dts generator tool called sopc2dts.
>> Our tool already aware of this requirement, but haven't support this yet.
>> I can edit this file manually to change 'ALTR' to 'altr', but I will
>> update nios2 code to match for both "ALTR" and "altr" for backward
>> compatibility. "ALTR" will be deprecated later. Are you okay with
>> this?
>
> You seem to have a large number of "ALTR" strings at the moment.
> I don't think it would be good to allow matching both variants
> everywhere, that would be a lot of extra code.
>
> If it's only about the root compatible property, I have no objections,
> but I don't know if it will help you when all the other strings don't
> match.
ALTR is used on compatible and dts parameters, but these all are in
arch/nios2. Drivers already can handle both ALTR and altr.
I can write a helper function in arch/nios2 to match both ALTR and
altr. I believe it shouldn't take a lot of extra code.
What do you think?

>> >> + ranges = < 0x00400000 0x08400000 0x00000020
>> >> + 0x00004D40 0x08004D40 0x00000008
>> >> + 0x00004D50 0x08004D50 0x00000008
>> >> + 0x00004000 0x08004000 0x00000400
>> >> + 0x00004400 0x08004400 0x00000040
>> >> + 0x00004800 0x08004800 0x00000040
>> >> + 0x00002000 0x08002000 0x00002000
>> >> + 0x00004C80 0x08004C80 0x00000020
>> >> + 0x00004CC0 0x08004CC0 0x00000010
>> >> + 0x00004CE0 0x08004CE0 0x00000010
>> >> + 0x00004D00 0x08004D00 0x00000010 >;
>> >
>
>> > - The ranges should reflect what the bus actually translates,
>> > which is typically not individual bytes but rather whole
>> > address ranges.
>> The ranges here reflect the address range translate by each device and
>> user can assign any base address they desired in our FPGA. The address
>> ranges might not in continuous regions as well. So, we will keep this
>> translation way.
>
> If you can support any address, why not just have an empty 'ranges'
> property?
We prefer not to add in any invalid address ranges here.


>>
>> I will add the binding doc for timer, but sysid driver is not in
>> mainline kernel yet. I will remove sysid entry from this file.
>
> Ok. BTW, could the sysid be used for the soc device strings, or
> is that something else?
Initially, we have thought to use sysid here. But, sysid can be used
by our ARM SoC device. For some reasons, we decided to make sysid as
separate module and not tight to CPU.

>
>> >> + reg = < 0x00004000 0x00000400
>> >> + 0x00004400 0x00000040
>> >> + 0x00004800 0x00000040
>> >> + 0x00002000 0x00002000 >;
>> >> + reg-names = "control_port", "rx_csr", "tx_csr", "s1";
>> >
>> > - wrong order, missing "tx_desc" and "rx_desc" entries
>> FYI, TSE driver supports both SGDMA and MSGDMA soft IP and "tx_desc"
>> and "rx_desc" entries are for MSGDMA (see below). But this 3c120
>> hardware design is using TSE + SGDMA. So, the expect entries are
>> "control_port", "rx_csr", "tx_csr", "s1".
>> I can sort the entries order by their addresses.
>
> You can't reorder the registers, as they can be accessed by
> index in some OS, the strings are just annotations really.
Okay, will keep the order.
Currently, the reg names and ordering are based on hardware
information and dts generator tool generates this entry based on this
hardware information. These shouldn't change unless hardware is
changed.

>
>> Excerpt from Documentation/devicetree/bindings/net/altera_tse.txt
>> - reg-names: Should contain the reg names
>> "control_port": MAC configuration space region
>> "tx_csr": xDMA Tx dispatcher control and status space region
>> "tx_desc": MSGDMA Tx dispatcher descriptor space region
>> "rx_csr" : xDMA Rx dispatcher control and status space region
>> "rx_desc": MSGDMA Rx dispatcher descriptor space region
>> "rx_resp": MSGDMA Rx dispatcher response space region
>> "s1": SGDMA descriptor memory
>
> The important part is
>
> - reg: Address and length of the register set for the device. It contains
> the information of registers in the same order as described by reg-names.
>
> You can either provide dummy entries for tx_desc, rx_desc and rx_resp,
> or change the binding to reflect the current usage, and provide two
> different lists depending on the compatible string.
We can update the binding doc to have separate lists depending on the
compatible string.
Added our TSE driver maintainer in CC list.


Regards
Ley Foon

2014-04-23 10:07:53

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 05/28] nios2: Traps exception handling

On Tue, Apr 22, 2014 at 10:28 PM, Ezequiel Garcia
<[email protected]> wrote:
> Hello Ley Foon,
>
> On Apr 19, Ley Foon Tan wrote:
>> +
>> +#ifndef CONFIG_ALIGNMENT_TRAP
>> +/* Alignment exception handler */
>> +asmlinkage void handle_unaligned_c(struct pt_regs *fp, int cause)
>> +{
>> + unsigned long addr = RDCTL(CTL_BADADDR);
>> +
>> + cause >>= 2;
>> + fp->ea -= 4;
>> +
>> + if (fixup_exception(fp))
>> + return;
>
> This will throw an undeclared error if you build with CONFIG_ALIGNMENT_TRAP=n.
>
> You need to add the uaccess.h header, or implement some similar fix.
Okay, I will try with CONFIG_ALIGNMENT_TRAP=n and fix it.

2014-04-23 12:24:10

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 04/28] nios2: Exception handling

On Wednesday 23 April 2014, Geert Uytterhoeven wrote:
> On Wed, Apr 23, 2014 at 4:47 AM, Ley Foon Tan <[email protected]> wrote:
> > On Tue, Apr 22, 2014 at 8:33 PM, Arnd Bergmann <[email protected]> wrote:
> >> On Friday 18 April 2014, Ley Foon Tan wrote:
> >>> +static const struct file_operations misalign_fops = {
> >>> + .open = proc_misaligned_open,
> >>> + .read = seq_read,
> >>> + .llseek = seq_lseek,
> >>> + .release = single_release,
> >>> + .write = proc_misaligned_write,
> >>> +};
> >>> +#endif /* CONFIG_PROC_FS */
> >>
> >> We really shouldn't be doing new architecture specific procfs files
> >> any more. I suggest you drop this one for now, and add back the
> >> functionality using perf or ftrace at a later point.
> >
> > Okay, will remove this.
>
> MIPS and powerpc handle this through debugfs, cfr.
> arch/mips/kernel/unaligned.c and arch/powerpc/kernel/traps.c.

Ah, right, that would work too. Being able to use perf from user
space sounds more flexible though. It's certainly something that
can be decided later.

Arnd

2014-04-23 17:59:53

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>> > On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>> >> Hi Arnd and Peter Anvin,
>>> >>
>>> >> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>> >> that we don't need to have 64 bit off_t? See detail in link below.
>>> >> I can submit the patches for 64-bit time changes
>>> >> (include/asm-generic/posix_types.h and other archs) if everyone is
>>> >> agreed on this.
>> >
>> > Yes.
> Okay, will doing that.

I believe that arm64 ILP32 will also be affected. What is the status of
this configuration? Has the glibc/kernel ABI been finalized?

Thanks,
Chung-Lin

2014-04-23 18:30:35

by Andrew Pinski

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port



> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
>
>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>> Hi Arnd and Peter Anvin,
>>>>>>
>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>>>>> that we don't need to have 64 bit off_t? See detail in link below.
>>>>>> I can submit the patches for 64-bit time changes
>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>>>>> agreed on this.
>>>>
>>>> Yes.
>> Okay, will doing that.
>
> I believe that arm64 ILP32 will also be affected. What is the status of
> this configuration? Has the glibc/kernel ABI been finalized?

Not yet. I am still working out the signal handling part. But we already agreed on 64bit time_t, clock_t, and suseconds_t. And we agreed to a 64bit offset_t too.

On a related note suseconds in the timespec in posix is defined to be long. So it would nice if the kernel ignores the upper 32bits so we (glibc developers) can fix this for new targets including x32 and arm64/ilp32.

Thanks,
Andrew

>
> Thanks,
> Chung-Lin
>

2014-04-23 19:23:59

by H. Peter Anvin

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

I think Linuxs said we should just fix POSIX on that front.

On April 23, 2014 11:15:34 AM PDT, "Pinski, Andrew" <[email protected]> wrote:
>
>
>> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang"
><[email protected]> wrote:
>>
>>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]>
>wrote:
>>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>>> Hi Arnd and Peter Anvin,
>>>>>>>
>>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you
>confirm
>>>>>>> that we don't need to have 64 bit off_t? See detail in link
>below.
>>>>>>> I can submit the patches for 64-bit time changes
>>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone
>is
>>>>>>> agreed on this.
>>>>>
>>>>> Yes.
>>> Okay, will doing that.
>>
>> I believe that arm64 ILP32 will also be affected. What is the status
>of
>> this configuration? Has the glibc/kernel ABI been finalized?
>
>Not yet. I am still working out the signal handling part. But we
>already agreed on 64bit time_t, clock_t, and suseconds_t. And we
>agreed to a 64bit offset_t too.
>
>On a related note suseconds in the timespec in posix is defined to be
>long. So it would nice if the kernel ignores the upper 32bits so we
>(glibc developers) can fix this for new targets including x32 and
>arm64/ilp32.
>
>Thanks,
>Andrew
>
>>
>> Thanks,
>> Chung-Lin
>>

--
Sent from my mobile phone. Please pardon brevity and lack of formatting.

2014-04-24 06:02:51

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 07/28] nios2: I/O Mapping

On Tue, Apr 22, 2014 at 9:59 PM, Arnd Bergmann <[email protected]> wrote:
> On Friday 18 April 2014, Ley Foon Tan wrote:
>
>> +
>> +#include <asm/pgtable-bits.h>
>> +
>> +#define IO_SPACE_LIMIT 0xffffffff
>
> Please use 0xffff here, this should work for almost any PCI bus.
Ah, CONFIG_PCI is not enable in nios2. So, I think IO_SPACE_LIMIT
should set to 0.


>> +
>> +/* Use "Duff's Device" to unroll the loops. */
>> +#define __IO_OUT_LOOP(a, b, l) \
>> + do { \
>> + if (l > 0) { \
>> + int _n = (l + 7) / 8; \
>> + switch (l % 8) { \
>> + case 0: \
>> + do { \
>> + *a = *b++; \
>
> I would recommend just doing all of this in out-of-line implementations
> rather than macros and inline functions.
Okay. Will move the macros to out-of-line function.


>
>> +
>> +/*
>> + * make the short names macros so specific devices
>> + * can override them as required
>> + */
>> +#define inb(addr) readb(addr)
>> +#define inw(addr) readw(addr)
>> +#define inl(addr) readl(addr)
>> +#define outb(x, addr) ((void) writeb(x, addr))
>> +#define outw(x, addr) ((void) writew(x, addr))
>> +#define outl(x, addr) ((void) writel(x, addr))
>
> This makes no sense: the legacy PC I/O accessors take a completely
> different argument type: I would recommend to make both inline
> functions so you can ensure that readl() gets passed an __iomem
> pointer, while inl() gets an integer number.
>
> Please see the asm-generic version for an example. You should also
> define a non-NULL PCI_IOBASE to which the PCI I/O space gets mapped.
I think these macros can be removed if PCI is not enabled in Nios2.

>
>> +#define virt_to_bus virt_to_phys
>> +#define bus_to_virt phys_to_virt
>
> Please drop these, and the CONFIG_VIRT_TO_BUS Kconfig option,
> and fix all drivers that rely on it.
Okay. Will remove these.

>
>> +#define ioport_map(port, nr) ioremap(port, nr)
>> +#define ioport_unmap(port) iounmap(port)
>
> Use this one instead:
>
> #ifdef CONFIG_PCI
> static inline void __iomem *ioport_map(unsigned long port, unsigned int nr)
> {
> return PCI_IOBASE + port;
> }
>
> static inline void __iomem *ioport_unmap(void __iomem *)
> {
> }
> #endif
These also can be removed.

>
>> +/* Macros used for smc91x.c driver */
>> +#define readsb(p, d, l) insb(p, d, l)
>> +#define readsw(p, d, l) insw(p, d, l)
>> +#define readsl(p, d, l) insl(p, d, l)
>> +#define writesb(p, d, l) outsb(p, d, l)
>> +#define writesw(p, d, l) outsw(p, d, l)
>> +#define writesl(p, d, l) outsl(p, d, l)
>
> These should of course not fall back to the PCI I/O space functions.
> You can do it the other way round.
Okay, these will change to something like:
#define readsb(p, d, l) io_insb(p, d, l)
#define readsw(p, d, l) io_insw(p, d, l)
(etc)

Regards
Ley Foon

2014-04-24 06:04:42

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 04/28] nios2: Exception handling

On Wed, Apr 23, 2014 at 8:23 PM, Arnd Bergmann <[email protected]> wrote:

>> >> We really shouldn't be doing new architecture specific procfs files
>> >> any more. I suggest you drop this one for now, and add back the
>> >> functionality using perf or ftrace at a later point.
>> >
>> > Okay, will remove this.
>>
>> MIPS and powerpc handle this through debugfs, cfr.
>> arch/mips/kernel/unaligned.c and arch/powerpc/kernel/traps.c.
>
> Ah, right, that would work too. Being able to use perf from user
> space sounds more flexible though. It's certainly something that
> can be decided later.
Yes, we can add this back through debugfs or perf in later.

Regards
Ley Foon

2014-04-24 06:26:40

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 2014/4/24 上午 02:15, Pinski, Andrew wrote:
>
>> > On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
>> >
>>> >> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>>> >> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>>>>> >>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>>> >>>>>> Hi Arnd and Peter Anvin,
>>>>>>> >>>>>>
>>>>>>> >>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>>>>>> >>>>>> that we don't need to have 64 bit off_t? See detail in link below.
>>>>>>> >>>>>> I can submit the patches for 64-bit time changes
>>>>>>> >>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>>>>>> >>>>>> agreed on this.
>>>>> >>>>
>>>>> >>>> Yes.
>>> >> Okay, will doing that.
>> >
>> > I believe that arm64 ILP32 will also be affected. What is the status of
>> > this configuration? Has the glibc/kernel ABI been finalized?
> Not yet. I am still working out the signal handling part. But we already agreed on 64bit time_t, clock_t, and suseconds_t. And we agreed to a 64bit offset_t too.
>
> On a related note suseconds in the timespec in posix is defined to be long. So it would nice if the kernel ignores the upper 32bits so we (glibc developers) can fix this for new targets including x32 and arm64/ilp32.

Hmm, but that means for purely 32-bit architectures like nios2, which
unlike x86_64 or arm64, never has a 64-bit mode, suseconds_t as a 64-bit
type in the kernel is simply wasted.

Chung-Lin

2014-04-24 06:42:52

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 08/28] nios2: MMU Fault handling

On Tue, Apr 22, 2014 at 10:30 PM, Ezequiel Garcia
<[email protected]> wrote:
> Hello Ley Foon,
>
> On Apr 18, Ley Foon Tan wrote:
>> +
>> +bad_area_nosemaphore:
>> + /* User mode accesses just cause a SIGSEGV */
>> + if (user_mode(regs)) {
>
> I found that it's useful to add some printing here, just as ARM
> does. I carry this patch on my kernel:
>
> + printk(KERN_INFO "%s: unhandled page fault (%d) at 0x%08lx, cause %ld\n",
> + current->comm, SIGSEGV, address, cause);
> + show_regs(regs);
>
> Do you think we could do something like it? Maybe with a compile time option?
Yes, this is useful debug message. I can add this.
Prefer not to add new compile time option, I think this shouldn't
happen frequently.
>
>> + _exception(SIGSEGV, regs, code, address);
>> + return;
>> + }
>> +

2014-04-24 07:18:29

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: [PATCH 08/28] nios2: MMU Fault handling

On Tue, Apr 22, 2014 at 4:30 PM, Ezequiel Garcia
<[email protected]> wrote:
> On Apr 18, Ley Foon Tan wrote:
>> +
>> +bad_area_nosemaphore:
>> + /* User mode accesses just cause a SIGSEGV */
>> + if (user_mode(regs)) {
>
> I found that it's useful to add some printing here, just as ARM
> does. I carry this patch on my kernel:
>
> + printk(KERN_INFO "%s: unhandled page fault (%d) at 0x%08lx, cause %ld\n",
> + current->comm, SIGSEGV, address, cause);
> + show_regs(regs);
>
> Do you think we could do something like it? Maybe with a compile time option?

Can this easily be triggered from userspace? If yes, please use
pr_info_ratelimited().

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

2014-04-24 07:44:29

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 07/28] nios2: I/O Mapping

On Thursday 24 April 2014 14:02:47 Ley Foon Tan wrote:
> On Tue, Apr 22, 2014 at 9:59 PM, Arnd Bergmann <[email protected]> wrote:
> > On Friday 18 April 2014, Ley Foon Tan wrote:
> >
> >> +
> >> +#include <asm/pgtable-bits.h>
> >> +
> >> +#define IO_SPACE_LIMIT 0xffffffff
> >
> > Please use 0xffff here, this should work for almost any PCI bus.
> Ah, CONFIG_PCI is not enable in nios2. So, I think IO_SPACE_LIMIT
> should set to 0.

Yes, possibly with a comment added regarding PCI.

> >> +/*
> >> + * make the short names macros so specific devices
> >> + * can override them as required
> >> + */
> >> +#define inb(addr) readb(addr)
> >> +#define inw(addr) readw(addr)
> >> +#define inl(addr) readl(addr)
> >> +#define outb(x, addr) ((void) writeb(x, addr))
> >> +#define outw(x, addr) ((void) writew(x, addr))
> >> +#define outl(x, addr) ((void) writel(x, addr))
> >
> > This makes no sense: the legacy PC I/O accessors take a completely
> > different argument type: I would recommend to make both inline
> > functions so you can ensure that readl() gets passed an __iomem
> > pointer, while inl() gets an integer number.
> >
> > Please see the asm-generic version for an example. You should also
> > define a non-NULL PCI_IOBASE to which the PCI I/O space gets mapped.
> I think these macros can be removed if PCI is not enabled in Nios2.

We have a bunch of ISA device drivers that use these interfaces but
are not built conditionally through a Kconfig dependency. If you
want 'allyesconfig' to build successfully, you have to provide some
dummy implementation.

> >> +#define ioport_map(port, nr) ioremap(port, nr)
> >> +#define ioport_unmap(port) iounmap(port)
> >
> > Use this one instead:
> >
> > #ifdef CONFIG_PCI
> > static inline void __iomem *ioport_map(unsigned long port, unsigned int nr)
> > {
> > return PCI_IOBASE + port;
> > }
> >
> > static inline void __iomem *ioport_unmap(void __iomem *)
> > {
> > }
> > #endif
> These also can be removed.

For this, there is a Kconfig symbol. If you don't provide ioport_map,
you should set CONFIG_NO_IOPORT_MAP. Uwe Kleine-K?nig has recently
started some work in this area. Ideally we would also have a CONFIG_HAS_IOPORT
option that is only set by architectures that actually implement inb/outb.
Getting there still needs work, but any help is appreciated.

Arnd

2014-04-24 09:35:45

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 2014/4/24 02:26 PM, Chung-Lin Tang wrote:
> On 2014/4/24 上午 02:15, Pinski, Andrew wrote:
>>
>>>> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>>>
>>>>>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>>>>>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>>>>>>>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>>>>>>>>>> Hi Arnd and Peter Anvin,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>>>>>>>>>>>>> that we don't need to have 64 bit off_t? See detail in link below.
>>>>>>>>>>>>>> I can submit the patches for 64-bit time changes
>>>>>>>>>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>>>>>>>>>>>>> agreed on this.
>>>>>>>>>>
>>>>>>>>>> Yes.
>>>>>> Okay, will doing that.
>>>>
>>>> I believe that arm64 ILP32 will also be affected. What is the status of
>>>> this configuration? Has the glibc/kernel ABI been finalized?
>> Not yet. I am still working out the signal handling part. But we already agreed on 64bit time_t, clock_t, and suseconds_t. And we agreed to a 64bit offset_t too.
>>
>> On a related note suseconds in the timespec in posix is defined to be long. So it would nice if the kernel ignores the upper 32bits so we (glibc developers) can fix this for new targets including x32 and arm64/ilp32.
>
> Hmm, but that means for purely 32-bit architectures like nios2, which
> unlike x86_64 or arm64, never has a 64-bit mode, suseconds_t as a 64-bit
> type in the kernel is simply wasted.
>
> Chung-Lin

The more I think of this, the more I feel that suseconds_t should jsut
be 'long', not strictly 64-bitified. An ILP32 sub-mode in a 64-bit
kernel should be using compat_* code paths, something like a
COMPAT_USE_32BIT_SUSECONDS case.

suseconds_t is for micro-seconds in struct timeval, 32-bit is more than
enough.

Chung-Lin

2014-04-24 10:01:48

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

On Sun, Apr 20, 2014 at 3:29 AM, Richard Weinberger
<[email protected]> wrote:

>> +
>> + /* Default to using normal stack. */
>> + usp = regs->sp;
>> +
>> + /* This is the X/Open sanctioned signal stack switching. */
>> + if ((ka->sa.sa_flags & SA_ONSTACK) && (current->sas_ss_sp != 0)) {
>> + if (!on_sig_stack(usp))
>> + usp = current->sas_ss_sp + current->sas_ss_size;
>> + }
>
> You can use sigsp() here.
Okay.


>> +
>> + push_cache((unsigned long) &frame->retcode);
>> +
>> + /* Set up registers for signal handler */
>> + regs->sp = (unsigned long) frame;
>> + regs->r4 = (unsigned long) (current_thread_info()->exec_domain
>> + && current_thread_info()->exec_domain->signal_invmap
>> + && sig < 32
>> + ? current_thread_info()->exec_domain->signal_invmap[sig]
>> + : sig);
>
> Does nios2 really need signal translation and supports execution domains?
Nios2 have one default "default_exec_domain" only. So, we can change
this to "regs->r4 = sig;".
BTW, most of the architectures have similar code.



>> +
>> +/*
>> + * Note that 'init' is a special process: it doesn't get signals it doesn't
>> + * want to handle. Thus you cannot kill init even with a SIGKILL even by
>> + * mistake.
>> + */
Will remove this.


>> +
>> +asmlinkage void do_notify_resume(struct pt_regs *regs, sigset_t *oldset,
>> + int in_syscall)
>> +{
>> + pr_debug("--> ENTERING %s\n", __func__);
>
> Please remove such printk()s.
> If you need to know which functions get called, use ftrace.
Okay, will remove this.


Regards
Ley Foon

2014-04-24 10:08:08

by Richard Weinberger

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

Am 24.04.2014 12:01, schrieb Ley Foon Tan:
> On Sun, Apr 20, 2014 at 3:29 AM, Richard Weinberger
>>> +
>>> + push_cache((unsigned long) &frame->retcode);
>>> +
>>> + /* Set up registers for signal handler */
>>> + regs->sp = (unsigned long) frame;
>>> + regs->r4 = (unsigned long) (current_thread_info()->exec_domain
>>> + && current_thread_info()->exec_domain->signal_invmap
>>> + && sig < 32
>>> + ? current_thread_info()->exec_domain->signal_invmap[sig]
>>> + : sig);
>>
>> Does nios2 really need signal translation and supports execution domains?
> Nios2 have one default "default_exec_domain" only. So, we can change
> this to "regs->r4 = sig;".

Nice.

> BTW, most of the architectures have similar code.

Yep, because everyone is copy&pasting from each others without thinking. 8)
There is already a cleanup going on: https://lkml.org/lkml/2014/3/2/198

Thanks,
//richard

2014-04-24 10:14:06

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

On Thu, Apr 24, 2014 at 6:01 PM, Ley Foon Tan <[email protected]> wrote:
> On Sun, Apr 20, 2014 at 3:29 AM, Richard Weinberger
> <[email protected]> wrote:
>
>>> +
>>> + /* Default to using normal stack. */
>>> + usp = regs->sp;
>>> +
>>> + /* This is the X/Open sanctioned signal stack switching. */
>>> + if ((ka->sa.sa_flags & SA_ONSTACK) && (current->sas_ss_sp != 0)) {
>>> + if (!on_sig_stack(usp))
>>> + usp = current->sas_ss_sp + current->sas_ss_size;
>>> + }
>>
>> You can use sigsp() here.
> Okay.
>
I just noticed that sigsp() have struct ksignal argument. We can use
sigsp() for nios2 because it doesn't have struct ksignal.

unsigned long sigsp(unsigned long sp, struct ksignal *ksig);

2014-04-24 10:17:35

by Richard Weinberger

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

Am 24.04.2014 12:13, schrieb Ley Foon Tan:
> On Thu, Apr 24, 2014 at 6:01 PM, Ley Foon Tan <[email protected]> wrote:
>> On Sun, Apr 20, 2014 at 3:29 AM, Richard Weinberger
>> <[email protected]> wrote:
>>
>>>> +
>>>> + /* Default to using normal stack. */
>>>> + usp = regs->sp;
>>>> +
>>>> + /* This is the X/Open sanctioned signal stack switching. */
>>>> + if ((ka->sa.sa_flags & SA_ONSTACK) && (current->sas_ss_sp != 0)) {
>>>> + if (!on_sig_stack(usp))
>>>> + usp = current->sas_ss_sp + current->sas_ss_size;
>>>> + }
>>>
>>> You can use sigsp() here.
>> Okay.
>>
> I just noticed that sigsp() have struct ksignal argument. We can use
> sigsp() for nios2 because it doesn't have struct ksignal.
>
> unsigned long sigsp(unsigned long sp, struct ksignal *ksig);

Did you at look at the struct ksignal definition and the clean series I've
pointed you to?

Thanks,
//richard

2014-04-24 10:29:53

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

On Thu, Apr 24, 2014 at 6:17 PM, Richard Weinberger <[email protected]> wrote:

>>> Okay.
>>>
>> I just noticed that sigsp() have struct ksignal argument. We can use
>> sigsp() for nios2 because it doesn't have struct ksignal.
>>
>> unsigned long sigsp(unsigned long sp, struct ksignal *ksig);
>
> Did you at look at the struct ksignal definition and the clean series I've
> pointed you to?
>

Not yet, this is the full set of patches? I only see partial..
https://lkml.org/lkml/2014/3/2/198

2014-04-24 10:39:47

by Richard Weinberger

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

Am 24.04.2014 12:29, schrieb Ley Foon Tan:
> On Thu, Apr 24, 2014 at 6:17 PM, Richard Weinberger <[email protected]> wrote:
>
>>>> Okay.
>>>>
>>> I just noticed that sigsp() have struct ksignal argument. We can use
>>> sigsp() for nios2 because it doesn't have struct ksignal.
>>>
>>> unsigned long sigsp(unsigned long sp, struct ksignal *ksig);
>>
>> Did you at look at the struct ksignal definition and the clean series I've
>> pointed you to?
>>
>
> Not yet, this is the full set of patches? I only see partial..
> https://lkml.org/lkml/2014/3/2/198

Maybe lkml.org misses a few but this does not matter.
All you need to know is that using ksignal is easy.
e.g. https://lkml.org/lkml/2014/3/2/162

Thanks,
//richard

2014-04-24 10:46:45

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

On Thu, Apr 24, 2014 at 6:39 PM, Richard Weinberger <[email protected]> wrote:
> Am 24.04.2014 12:29, schrieb Ley Foon Tan:
>> On Thu, Apr 24, 2014 at 6:17 PM, Richard Weinberger <[email protected]> wrote:
>>
>>>>> Okay.
>>>>>
>>>> I just noticed that sigsp() have struct ksignal argument. We can use
>>>> sigsp() for nios2 because it doesn't have struct ksignal.
>>>>
>>>> unsigned long sigsp(unsigned long sp, struct ksignal *ksig);
>>>
>>> Did you at look at the struct ksignal definition and the clean series I've
>>> pointed you to?
>>>
>>
>> Not yet, this is the full set of patches? I only see partial..
>> https://lkml.org/lkml/2014/3/2/198
>
> Maybe lkml.org misses a few but this does not matter.
> All you need to know is that using ksignal is easy.
> e.g. https://lkml.org/lkml/2014/3/2/162
>
Thanks. I am looking at your git tree
https://git.kernel.org/cgit/linux/kernel/git/rw/misc.git/log/?h=signal_v2
Are these patches already available in v3.15 rc?
I can update nios2 to use new implementation.

Regards
Ley Foon

2014-04-24 15:23:12

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 08/28] nios2: MMU Fault handling

On Apr 24, Ley Foon Tan wrote:
> On Tue, Apr 22, 2014 at 10:30 PM, Ezequiel Garcia
> <[email protected]> wrote:
> > Hello Ley Foon,
> >
> > On Apr 18, Ley Foon Tan wrote:
> >> +
> >> +bad_area_nosemaphore:
> >> + /* User mode accesses just cause a SIGSEGV */
> >> + if (user_mode(regs)) {
> >
> > I found that it's useful to add some printing here, just as ARM
> > does. I carry this patch on my kernel:
> >
> > + printk(KERN_INFO "%s: unhandled page fault (%d) at 0x%08lx, cause %ld\n",
> > + current->comm, SIGSEGV, address, cause);
> > + show_regs(regs);
> >
> > Do you think we could do something like it? Maybe with a compile time option?
> Yes, this is useful debug message. I can add this.
> Prefer not to add new compile time option, I think this shouldn't
> happen frequently.

No, it shouldn't happen frequently. You can take a look at what ARM does.
They use a DEBUG_USER compile-time option, much more complex than my silly
proposal.
--
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-24 15:29:05

by Catalin Marinas

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On Thu, Apr 24, 2014 at 09:55:25AM +0100, Chung-Lin Tang wrote:
> On 2014/4/24 02:26 PM, Chung-Lin Tang wrote:
> > On 2014/4/24 上午 02:15, Pinski, Andrew wrote:
> >>
> >>>> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
> >>>>
> >>>>>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
> >>>>>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
> >>>>>>>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
> >>>>>>>>>>>>>> Hi Arnd and Peter Anvin,
> >>>>>>>>>>>>>>
> >>>>>>>>>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
> >>>>>>>>>>>>>> that we don't need to have 64 bit off_t? See detail in link below.
> >>>>>>>>>>>>>> I can submit the patches for 64-bit time changes
> >>>>>>>>>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
> >>>>>>>>>>>>>> agreed on this.
> >>>>>>>>>>
> >>>>>>>>>> Yes.
> >>>>>> Okay, will doing that.
> >>>>
> >>>> I believe that arm64 ILP32 will also be affected. What is the status of
> >>>> this configuration? Has the glibc/kernel ABI been finalized?
> >> Not yet. I am still working out the signal handling part. But we
> >> already agreed on 64bit time_t, clock_t, and suseconds_t. And we
> >> agreed to a 64bit offset_t too.
> >>
> >> On a related note suseconds in the timespec in posix is defined to
> >> be long. So it would nice if the kernel ignores the upper 32bits so
> >> we (glibc developers) can fix this for new targets including x32
> >> and arm64/ilp32.
> >
> > Hmm, but that means for purely 32-bit architectures like nios2, which
> > unlike x86_64 or arm64, never has a 64-bit mode, suseconds_t as a 64-bit
> > type in the kernel is simply wasted.
>
> The more I think of this, the more I feel that suseconds_t should jsut
> be 'long', not strictly 64-bitified. An ILP32 sub-mode in a 64-bit
> kernel should be using compat_* code paths, something like a
> COMPAT_USE_32BIT_SUSECONDS case.

ILP32 mode should use LP64 syscalls as much as possible and that's the
aim with arm64 as well (of course, we still have a few that wouldn't be
possible and we route them via compat).

But here if time_t is 64-bit while susecconds_t is 32-bit, the compat
code wouldn't help.

--
Catalin

2014-04-24 15:58:52

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 07/28] nios2: I/O Mapping

On Thu, Apr 24, 2014 at 3:43 PM, Arnd Bergmann <[email protected]> wrote:
> On Thursday 24 April 2014 14:02:47 Ley Foon Tan wrote:

>> >> +#define IO_SPACE_LIMIT 0xffffffff
>> >
>> > Please use 0xffff here, this should work for almost any PCI bus.
>> Ah, CONFIG_PCI is not enable in nios2. So, I think IO_SPACE_LIMIT
>> should set to 0.
>
> Yes, possibly with a comment added regarding PCI.
Sure.

>
>> >> +/*
>> >> + * make the short names macros so specific devices
>> >> + * can override them as required
>> >> + */
>> >> +#define inb(addr) readb(addr)
>> >> +#define inw(addr) readw(addr)
>> >> +#define inl(addr) readl(addr)
>> >> +#define outb(x, addr) ((void) writeb(x, addr))
>> >> +#define outw(x, addr) ((void) writew(x, addr))
>> >> +#define outl(x, addr) ((void) writel(x, addr))
>> >
>> > This makes no sense: the legacy PC I/O accessors take a completely
>> > different argument type: I would recommend to make both inline
>> > functions so you can ensure that readl() gets passed an __iomem
>> > pointer, while inl() gets an integer number.
>> >
>> > Please see the asm-generic version for an example. You should also
>> > define a non-NULL PCI_IOBASE to which the PCI I/O space gets mapped.
>> I think these macros can be removed if PCI is not enabled in Nios2.
>
> We have a bunch of ISA device drivers that use these interfaces but
> are not built conditionally through a Kconfig dependency. If you
> want 'allyesconfig' to build successfully, you have to provide some
> dummy implementation.
Okay, will provide empty write and read with return 0 dummy implementation here.

>
>> >> +#define ioport_map(port, nr) ioremap(port, nr)
>> >> +#define ioport_unmap(port) iounmap(port)
>> >
>> > Use this one instead:
>> >
>> > #ifdef CONFIG_PCI
>> > static inline void __iomem *ioport_map(unsigned long port, unsigned int nr)
>> > {
>> > return PCI_IOBASE + port;
>> > }
>> >
>> > static inline void __iomem *ioport_unmap(void __iomem *)
>> > {
>> > }
>> > #endif
>> These also can be removed.
>
> For this, there is a Kconfig symbol. If you don't provide ioport_map,
> you should set CONFIG_NO_IOPORT_MAP. Uwe Kleine-König has recently
> started some work in this area. Ideally we would also have a CONFIG_HAS_IOPORT
> option that is only set by architectures that actually implement inb/outb.
> Getting there still needs work, but any help is appreciated.
Will set CONFIG_NO_IOPORT_MAP. I noticed that there is
CONFIG_HAS_IOPORT_MAP, is it same as CONFIG_HAS_IOPORT you refer to?
But, seem only arch/sh is using it.

2014-04-24 16:02:33

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 08/28] nios2: MMU Fault handling

On Thu, Apr 24, 2014 at 11:22 PM, Ezequiel Garcia
<[email protected]> wrote:
>> > I found that it's useful to add some printing here, just as ARM
>> > does. I carry this patch on my kernel:
>> >
>> > + printk(KERN_INFO "%s: unhandled page fault (%d) at 0x%08lx, cause %ld\n",
>> > + current->comm, SIGSEGV, address, cause);
>> > + show_regs(regs);
>> >
>> > Do you think we could do something like it? Maybe with a compile time option?
>> Yes, this is useful debug message. I can add this.
>> Prefer not to add new compile time option, I think this shouldn't
>> happen frequently.
>
> No, it shouldn't happen frequently. You can take a look at what ARM does.
> They use a DEBUG_USER compile-time option, much more complex than my silly
> proposal.
For now, we can print it always if this shouldn't happen frequently.
Compile time option can be added later if we have more other debug
stuff to print later.

Thanks.

Regard,
Ley Foon

2014-04-24 16:06:23

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 19/28] nios2: Device tree support

On Apr 22, Arnd Bergmann wrote:
> On Friday 18 April 2014, Ley Foon Tan wrote:
>
> Are these all synthesized devices, or is there also some hardwired
> logic? It often makes sense to split out the reusable parts into
> a separate .dtsi file that gets included by every implementation.
>

In case you are not aware of this, devicetree files for Nios-II
SoCs are produced through an automated tool calle sopc2dts:

http://www.alterawiki.com/wiki/Sopc2dts
http://git.rocketboards.org/sopc-tools.git/

You feed it with a sopcinfo file (AFAIK, Altera's specific format) and
you obtain a full-fledged devicetree source file. Usually it works
out-of-the-box, although I like to go over it and fix ranges, whitespaces,
and do some cleaning.

So I'm wondering -given we have such superb tool- why would we want to
include the devicetree source's in the kernel?

First, we'll be only supporting a *specific* configuration. Second, the
dts is trivially easy to obtain.

The binding documentation should be enough specification, and there's no
need for further reference or examples.
--
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-24 16:17:05

by Richard Weinberger

[permalink] [raw]
Subject: Re: [PATCH 17/28] nios2: Signal handling support

Am 24.04.2014 12:46, schrieb Ley Foon Tan:
> On Thu, Apr 24, 2014 at 6:39 PM, Richard Weinberger <[email protected]> wrote:
>> Am 24.04.2014 12:29, schrieb Ley Foon Tan:
>>> On Thu, Apr 24, 2014 at 6:17 PM, Richard Weinberger <[email protected]> wrote:
>>>
>>>>>> Okay.
>>>>>>
>>>>> I just noticed that sigsp() have struct ksignal argument. We can use
>>>>> sigsp() for nios2 because it doesn't have struct ksignal.
>>>>>
>>>>> unsigned long sigsp(unsigned long sp, struct ksignal *ksig);
>>>>
>>>> Did you at look at the struct ksignal definition and the clean series I've
>>>> pointed you to?
>>>>
>>>
>>> Not yet, this is the full set of patches? I only see partial..
>>> https://lkml.org/lkml/2014/3/2/198
>>
>> Maybe lkml.org misses a few but this does not matter.
>> All you need to know is that using ksignal is easy.
>> e.g. https://lkml.org/lkml/2014/3/2/162
>>
> Thanks. I am looking at your git tree
> https://git.kernel.org/cgit/linux/kernel/git/rw/misc.git/log/?h=signal_v2
> Are these patches already available in v3.15 rc?

Nope, I have to rebase them again.
But some archs have picked up the changes already.

> I can update nios2 to use new implementation.

Would be great. :)

Thanks,
//richard

2014-04-24 16:18:41

by Ezequiel Garcia

[permalink] [raw]
Subject: Re: [PATCH 08/28] nios2: MMU Fault handling

On Apr 25, Ley Foon Tan wrote:
> On Thu, Apr 24, 2014 at 11:22 PM, Ezequiel Garcia
> <[email protected]> wrote:
> >> > I found that it's useful to add some printing here, just as ARM
> >> > does. I carry this patch on my kernel:
> >> >
> >> > + printk(KERN_INFO "%s: unhandled page fault (%d) at 0x%08lx, cause %ld\n",
> >> > + current->comm, SIGSEGV, address, cause);
> >> > + show_regs(regs);
> >> >
> >> > Do you think we could do something like it? Maybe with a compile time option?
> >> Yes, this is useful debug message. I can add this.
> >> Prefer not to add new compile time option, I think this shouldn't
> >> happen frequently.
> >
> > No, it shouldn't happen frequently. You can take a look at what ARM does.
> > They use a DEBUG_USER compile-time option, much more complex than my silly
> > proposal.
> For now, we can print it always if this shouldn't happen frequently.
> Compile time option can be added later if we have more other debug
> stuff to print later.
>

OK, sound sensible.
--
Ezequiel Garc?a, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

2014-04-24 16:23:24

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 16/28] nios2: System calls handling

On Tue, Apr 22, 2014 at 8:30 PM, Arnd Bergmann <[email protected]> wrote:
> On Friday 18 April 2014, Ley Foon Tan wrote:
>
>> +
>> + #define sys_mmap2 sys_mmap_pgoff
>> +
>
> You use sys_mmap_pgoff here, but
>
>> +asmlinkage long sys_mmap(unsigned long addr, unsigned long len,
>> + unsigned long prot, unsigned long flags,
>> + unsigned long fd, unsigned long offset)
>> +{
>> + if (offset & ~PAGE_MASK)
>> + return -EINVAL;
>> +
>> + return sys_mmap_pgoff(addr, len, prot, flags, fd, offset >> PAGE_SHIFT);
>> +}
>
> define a separate sys_mmap() here. Doing the former should be enough.
> so you can just drop the sys_mmap definition.
Okay, will drop the sys_mmap.

Thanks.
Regards
Ley Foon

2014-04-24 18:37:31

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 2014/4/24 11:28 PM, Catalin Marinas wrote:
> On Thu, Apr 24, 2014 at 09:55:25AM +0100, Chung-Lin Tang wrote:
>> On 2014/4/24 02:26 PM, Chung-Lin Tang wrote:
>>> On 2014/4/24 上午 02:15, Pinski, Andrew wrote:
>>>>
>>>>>> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>>>>>
>>>>>>>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>>>>>>>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>>>>>>>>>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>>>>>>>>>>>> Hi Arnd and Peter Anvin,
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>>>>>>>>>>>>>>> that we don't need to have 64 bit off_t? See detail in link below.
>>>>>>>>>>>>>>>> I can submit the patches for 64-bit time changes
>>>>>>>>>>>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>>>>>>>>>>>>>>> agreed on this.
>>>>>>>>>>>>
>>>>>>>>>>>> Yes.
>>>>>>>> Okay, will doing that.
>>>>>>
>>>>>> I believe that arm64 ILP32 will also be affected. What is the status of
>>>>>> this configuration? Has the glibc/kernel ABI been finalized?
>>>> Not yet. I am still working out the signal handling part. But we
>>>> already agreed on 64bit time_t, clock_t, and suseconds_t. And we
>>>> agreed to a 64bit offset_t too.
>>>>
>>>> On a related note suseconds in the timespec in posix is defined to
>>>> be long. So it would nice if the kernel ignores the upper 32bits so
>>>> we (glibc developers) can fix this for new targets including x32
>>>> and arm64/ilp32.
>>>
>>> Hmm, but that means for purely 32-bit architectures like nios2, which
>>> unlike x86_64 or arm64, never has a 64-bit mode, suseconds_t as a 64-bit
>>> type in the kernel is simply wasted.
>>
>> The more I think of this, the more I feel that suseconds_t should jsut
>> be 'long', not strictly 64-bitified. An ILP32 sub-mode in a 64-bit
>> kernel should be using compat_* code paths, something like a
>> COMPAT_USE_32BIT_SUSECONDS case.
>
> ILP32 mode should use LP64 syscalls as much as possible and that's the
> aim with arm64 as well (of course, we still have a few that wouldn't be
> possible and we route them via compat).
>
> But here if time_t is 64-bit while susecconds_t is 32-bit, the compat
> code wouldn't help.

Why not? You can define the arm64 'struct compat_timeval' with
suseconds_t as s32, and add the 32<-->64 case in the
compat_get/put_timeval path, triggered when the process is ILP32 (test
wrapped in the above hypothetical COMPAT_USE_32BIT_SUSECONDS macro).
Similar to how x32 does COMPAT_USE_64BIT_TIME.

Chung-Lin




2014-04-24 18:42:23

by Andrew Pinski

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port



> On Apr 24, 2014, at 11:37 AM, "Chung-Lin Tang" <[email protected]> wrote:
>
>> On 2014/4/24 11:28 PM, Catalin Marinas wrote:
>>> On Thu, Apr 24, 2014 at 09:55:25AM +0100, Chung-Lin Tang wrote:
>>>> On 2014/4/24 02:26 PM, Chung-Lin Tang wrote:
>>>>> On 2014/4/24 $B>e8a(B 02:15, Pinski, Andrew wrote:
>>>>>
>>>>>>> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>>>>>>
>>>>>>>>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>>>>>>>>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>>>>>>>>>>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>>>>>>>>>>>>> Hi Arnd and Peter Anvin,
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>>>>>>>>>>>>>>>> that we don't need to have 64 bit off_t? See detail in link below.
>>>>>>>>>>>>>>>>> I can submit the patches for 64-bit time changes
>>>>>>>>>>>>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>>>>>>>>>>>>>>>> agreed on this.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Yes.
>>>>>>>>> Okay, will doing that.
>>>>>>>
>>>>>>> I believe that arm64 ILP32 will also be affected. What is the status of
>>>>>>> this configuration? Has the glibc/kernel ABI been finalized?
>>>>> Not yet. I am still working out the signal handling part. But we
>>>>> already agreed on 64bit time_t, clock_t, and suseconds_t. And we
>>>>> agreed to a 64bit offset_t too.
>>>>>
>>>>> On a related note suseconds in the timespec in posix is defined to
>>>>> be long. So it would nice if the kernel ignores the upper 32bits so
>>>>> we (glibc developers) can fix this for new targets including x32
>>>>> and arm64/ilp32.
>>>>
>>>> Hmm, but that means for purely 32-bit architectures like nios2, which
>>>> unlike x86_64 or arm64, never has a 64-bit mode, suseconds_t as a 64-bit
>>>> type in the kernel is simply wasted.
>>>
>>> The more I think of this, the more I feel that suseconds_t should jsut
>>> be 'long', not strictly 64-bitified. An ILP32 sub-mode in a 64-bit
>>> kernel should be using compat_* code paths, something like a
>>> COMPAT_USE_32BIT_SUSECONDS case.
>>
>> ILP32 mode should use LP64 syscalls as much as possible and that's the
>> aim with arm64 as well (of course, we still have a few that wouldn't be
>> possible and we route them via compat).
>>
>> But here if time_t is 64-bit while susecconds_t is 32-bit, the compat
>> code wouldn't help.
>
> Why not? You can define the arm64 'struct compat_timeval' with
> suseconds_t as s32, and add the 32<-->64 case in the
> compat_get/put_timeval path, triggered when the process is ILP32 (test
> wrapped in the above hypothetical COMPAT_USE_32BIT_SUSECONDS macro).
> Similar to how x32 does COMPAT_USE_64BIT_TIME.

We would three timeval then. One for aarch32, one for lp64 and one for ilp32. We really don't want three. Two is already one too many in my mind after developing the ilp32 syscall abi.

Thanks,
Andrew


>
> Chung-Lin
>
>
>
>
>

2014-04-25 06:06:08

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 2014/4/25 02:42 AM, Pinski, Andrew wrote:
>
>
>> On Apr 24, 2014, at 11:37 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>
>>> On 2014/4/24 11:28 PM, Catalin Marinas wrote:
>>>> On Thu, Apr 24, 2014 at 09:55:25AM +0100, Chung-Lin Tang wrote:
>>>>> On 2014/4/24 02:26 PM, Chung-Lin Tang wrote:
>>>>>> On 2014/4/24 上午 02:15, Pinski, Andrew wrote:
>>>>>>
>>>>>>>> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>>>>>>>
>>>>>>>>>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>>>>>>>>>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>>>>>>>>>>>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>>>>>>>>>>>>>> Hi Arnd and Peter Anvin,
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>>>>>>>>>>>>>>>>> that we don't need to have 64 bit off_t? See detail in link below.
>>>>>>>>>>>>>>>>>> I can submit the patches for 64-bit time changes
>>>>>>>>>>>>>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>>>>>>>>>>>>>>>>> agreed on this.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Yes.
>>>>>>>>>> Okay, will doing that.
>>>>>>>>
>>>>>>>> I believe that arm64 ILP32 will also be affected. What is the status of
>>>>>>>> this configuration? Has the glibc/kernel ABI been finalized?
>>>>>> Not yet. I am still working out the signal handling part. But we
>>>>>> already agreed on 64bit time_t, clock_t, and suseconds_t. And we
>>>>>> agreed to a 64bit offset_t too.
>>>>>>
>>>>>> On a related note suseconds in the timespec in posix is defined to
>>>>>> be long. So it would nice if the kernel ignores the upper 32bits so
>>>>>> we (glibc developers) can fix this for new targets including x32
>>>>>> and arm64/ilp32.
>>>>>
>>>>> Hmm, but that means for purely 32-bit architectures like nios2, which
>>>>> unlike x86_64 or arm64, never has a 64-bit mode, suseconds_t as a 64-bit
>>>>> type in the kernel is simply wasted.
>>>>
>>>> The more I think of this, the more I feel that suseconds_t should jsut
>>>> be 'long', not strictly 64-bitified. An ILP32 sub-mode in a 64-bit
>>>> kernel should be using compat_* code paths, something like a
>>>> COMPAT_USE_32BIT_SUSECONDS case.
>>>
>>> ILP32 mode should use LP64 syscalls as much as possible and that's the
>>> aim with arm64 as well (of course, we still have a few that wouldn't be
>>> possible and we route them via compat).
>>>
>>> But here if time_t is 64-bit while susecconds_t is 32-bit, the compat
>>> code wouldn't help.
>>
>> Why not? You can define the arm64 'struct compat_timeval' with
>> suseconds_t as s32, and add the 32<-->64 case in the
>> compat_get/put_timeval path, triggered when the process is ILP32 (test
>> wrapped in the above hypothetical COMPAT_USE_32BIT_SUSECONDS macro).
>> Similar to how x32 does COMPAT_USE_64BIT_TIME.
>
> We would three timeval then. One for aarch32, one for lp64 and one for ilp32. We really don't want three. Two is already one too many in my mind after developing the ilp32 syscall abi.
>
> Thanks,
> Andrew

Okay I now see you're already doing that for 32-bit ARM.

Still, you would probably just need to have an arm64-ILP32 specific case
to be careful about the last padding word upon kernel entry/exit.
(accommodating the difference in sizeof(long)) Penalizing all
architectures does not seem like the best solution.

Having suseconds_t as a strictly 64-bit C type in the kernel, while
defined as <= long in user-space may cause other problems.

I'll try to explain a probable situation for Nios II. I'm not sure about
other soft-cores, but nios2 is sort of uncommon in that the maximum
alignment is 4-bytes (32-bits), even for doubles/long-longs.

So if time_t is 64-bits (which makes sense), then struct timeval, which
is time_t+suseconds_t in userspace is 12-bytes/aligned-4 (unlike many
archs where a 64-bit time_t will expand the size to 16-bytes, due to
align-8)

If the kernel suseconds_t is forced to be 64-bits, then nios2 will have
a 16-byte kernel timeval vs. 12-byte userspace timeval situation. Just
this will require us to do something using compat_*, or weird hacks in
glibc, which is unfair. Nios II has no "other-mode". We are just
strictly ILP32, everywhere.

Of course, we can probably still at the end just use a Nios II specific
posix_types.h header to override things, but I'm just stating this as a
matter of which are the most reasonable default settings in the generic
headers. Making pure ILP32 archs diverge from POSIX standards by default
does not seem to be right.

Thanks,
Chung-Lin

2014-04-25 08:38:33

by Andrew Pinski

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port



> On Apr 24, 2014, at 11:06 PM, "Chung-Lin Tang" <[email protected]> wrote:
>
>> On 2014/4/25 02:42 AM, Pinski, Andrew wrote:
>>
>>
>>>> On Apr 24, 2014, at 11:37 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>>>
>>>>> On 2014/4/24 11:28 PM, Catalin Marinas wrote:
>>>>>> On Thu, Apr 24, 2014 at 09:55:25AM +0100, Chung-Lin Tang wrote:
>>>>>>> On 2014/4/24 02:26 PM, Chung-Lin Tang wrote:
>>>>>>> On 2014/4/24 $B>e8a(B 02:15, Pinski, Andrew wrote:
>>>>>>>
>>>>>>>>> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>>>>>>>>
>>>>>>>>>>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>>>>>>>>>>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>>>>>>>>>>>>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>>>>>>>>>>>>>>> Hi Arnd and Peter Anvin,
>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>>>>>>>>>>>>>>>>>> that we don't need to have 64 bit off_t? See detail in link below.
>>>>>>>>>>>>>>>>>>> I can submit the patches for 64-bit time changes
>>>>>>>>>>>>>>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>>>>>>>>>>>>>>>>>> agreed on this.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Yes.
>>>>>>>>>>> Okay, will doing that.
>>>>>>>>>
>>>>>>>>> I believe that arm64 ILP32 will also be affected. What is the status of
>>>>>>>>> this configuration? Has the glibc/kernel ABI been finalized?
>>>>>>> Not yet. I am still working out the signal handling part. But we
>>>>>>> already agreed on 64bit time_t, clock_t, and suseconds_t. And we
>>>>>>> agreed to a 64bit offset_t too.
>>>>>>>
>>>>>>> On a related note suseconds in the timespec in posix is defined to
>>>>>>> be long. So it would nice if the kernel ignores the upper 32bits so
>>>>>>> we (glibc developers) can fix this for new targets including x32
>>>>>>> and arm64/ilp32.
>>>>>>
>>>>>> Hmm, but that means for purely 32-bit architectures like nios2, which
>>>>>> unlike x86_64 or arm64, never has a 64-bit mode, suseconds_t as a 64-bit
>>>>>> type in the kernel is simply wasted.
>>>>>
>>>>> The more I think of this, the more I feel that suseconds_t should jsut
>>>>> be 'long', not strictly 64-bitified. An ILP32 sub-mode in a 64-bit
>>>>> kernel should be using compat_* code paths, something like a
>>>>> COMPAT_USE_32BIT_SUSECONDS case.
>>>>
>>>> ILP32 mode should use LP64 syscalls as much as possible and that's the
>>>> aim with arm64 as well (of course, we still have a few that wouldn't be
>>>> possible and we route them via compat).
>>>>
>>>> But here if time_t is 64-bit while susecconds_t is 32-bit, the compat
>>>> code wouldn't help.
>>>
>>> Why not? You can define the arm64 'struct compat_timeval' with
>>> suseconds_t as s32, and add the 32<-->64 case in the
>>> compat_get/put_timeval path, triggered when the process is ILP32 (test
>>> wrapped in the above hypothetical COMPAT_USE_32BIT_SUSECONDS macro).
>>> Similar to how x32 does COMPAT_USE_64BIT_TIME.
>>
>> We would three timeval then. One for aarch32, one for lp64 and one for ilp32. We really don't want three. Two is already one too many in my mind after developing the ilp32 syscall abi.
>>
>> Thanks,
>> Andrew
>
> Okay I now see you're already doing that for 32-bit ARM.
>
> Still, you would probably just need to have an arm64-ILP32 specific case
> to be careful about the last padding word upon kernel entry/exit.
> (accommodating the difference in sizeof(long)) Penalizing all
> architectures does not seem like the best solution.

Considering the alignment of long long would be 64bits, the struct does not change sizes if suseconds_t is 32bits or 64bits.

>
> Having suseconds_t as a strictly 64-bit C type in the kernel, while
> defined as <= long in user-space may cause other problems.
>
> I'll try to explain a probable situation for Nios II. I'm not sure about
> other soft-cores, but nios2 is sort of uncommon in that the maximum
> alignment is 4-bytes (32-bits), even for doubles/long-longs.

Yes does that include even if users of aligned? If so that seems broken. Also yes nios ii is unstandard when it comes to alignment here.

>
> So if time_t is 64-bits (which makes sense), then struct timeval, which
> is time_t+suseconds_t in userspace is 12-bytes/aligned-4 (unlike many
> archs where a 64-bit time_t will expand the size to 16-bytes, due to
> align-8)

Unlike most other targets where the struct would 16bits no matter what.

Thanks,
Andrew


>
> If the kernel suseconds_t is forced to be 64-bits, then nios2 will have
> a 16-byte kernel timeval vs. 12-byte userspace timeval situation. Just
> this will require us to do something using compat_*, or weird hacks in
> glibc, which is unfair. Nios II has no "other-mode". We are just
> strictly ILP32, everywhere.
>
> Of course, we can probably still at the end just use a Nios II specific
> posix_types.h header to override things, but I'm just stating this as a
> matter of which are the most reasonable default settings in the generic
> headers. Making pure ILP32 archs diverge from POSIX standards by default
> does not seem to be right.
>
> Thanks,
> Chung-Lin
>

2014-04-25 08:49:32

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On Fri, Apr 25, 2014 at 8:06 AM, Chung-Lin Tang <[email protected]> wrote:
> I'll try to explain a probable situation for Nios II. I'm not sure about
> other soft-cores, but nios2 is sort of uncommon in that the maximum
> alignment is 4-bytes (32-bits), even for doubles/long-longs.

FWIW, that's the same as on m32r. And on m68k it's 2 bytes.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

2014-04-25 10:13:36

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 14/28] nios2: DMA mapping API

On Tue, Apr 22, 2014 at 9:52 PM, Arnd Bergmann <[email protected]> wrote:
> On Friday 18 April 2014, Ley Foon Tan wrote:
>
>> +static inline int dma_supported(struct device *dev, u64 mask)
>> +{
>> + /* we fall back to GFP_DMA when the mask isn't all 1s,
>> + * so we can't guarantee allocations that must be
>> + * within a tighter range than GFP_DMA.
>> + */
>> + if (mask < 0x00ffffff)
>> + return 0;
>> +
>> + return 1;
>> +}
>
>> +void *dma_alloc_coherent(struct device *dev, size_t size,
>> + dma_addr_t *dma_handle, gfp_t gfp)
>> +{
>> + void *ret;
>> +
>> + /* ignore region specifiers */
>> + gfp &= ~(__GFP_DMA | __GFP_HIGHMEM);
>> +
>> + /* optimized page clearing */
>> + gfp |= __GFP_ZERO;
>> +
>> + if (dev == NULL || (dev->coherent_dma_mask < 0xffffffff))
>> + gfp |= GFP_DMA;
>
> What is the significance of ZONE_DMA on this architecture?
>
> Do you actually have DMA masters with a 0x00ffffff mask?
Actually, nios2 doesn't have DMA master with 0x00ffffff mask. I will
change it to return 1, same as some other architectures
implementation.
And the ZONE_DMA stuff will remove as well, it shouldn't have ZONE_DMA
limitation.

2014-04-25 10:34:06

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 14/28] nios2: DMA mapping API

On Friday 25 April 2014 18:13:31 Ley Foon Tan wrote:
> Actually, nios2 doesn't have DMA master with 0x00ffffff mask. I will
> change it to return 1, same as some other architectures
> implementation.
> And the ZONE_DMA stuff will remove as well, it shouldn't have ZONE_DMA
> limitation.

Ok, excellent!

Arnd

2014-04-25 10:36:15

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 14/28] nios2: DMA mapping API

On Fri, Apr 25, 2014 at 6:33 PM, Arnd Bergmann <[email protected]> wrote:
> On Friday 25 April 2014 18:13:31 Ley Foon Tan wrote:
>> Actually, nios2 doesn't have DMA master with 0x00ffffff mask. I will
>> change it to return 1, same as some other architectures
>> implementation.
>> And the ZONE_DMA stuff will remove as well, it shouldn't have ZONE_DMA
>> limitation.
>
> Ok, excellent!
>
Thanks. BTW, do you have any chance to review other patches? :)

Regards
Ley Foon

2014-04-25 13:20:16

by Chung-Lin Tang

[permalink] [raw]
Subject: Re: [PATCH 00/28] nios2 Linux kernel port

On 14/4/25 下午4:37, Pinski, Andrew wrote:
>
>
>> On Apr 24, 2014, at 11:06 PM, "Chung-Lin Tang" <[email protected]> wrote:
>>
>>> On 2014/4/25 02:42 AM, Pinski, Andrew wrote:
>>>
>>>
>>>>> On Apr 24, 2014, at 11:37 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>>>>
>>>>>> On 2014/4/24 11:28 PM, Catalin Marinas wrote:
>>>>>>> On Thu, Apr 24, 2014 at 09:55:25AM +0100, Chung-Lin Tang wrote:
>>>>>>>> On 2014/4/24 02:26 PM, Chung-Lin Tang wrote:
>>>>>>>> On 2014/4/24 上午 02:15, Pinski, Andrew wrote:
>>>>>>>>
>>>>>>>>>> On Apr 23, 2014, at 10:59 AM, "Chung-Lin Tang" <[email protected]> wrote:
>>>>>>>>>>
>>>>>>>>>>>> On 2014/4/22 07:20 PM, Ley Foon Tan wrote:
>>>>>>>>>>>> On Tue, Apr 22, 2014 at 6:56 PM, Arnd Bergmann <[email protected]> wrote:
>>>>>>>>>>>>>>>> On Tuesday 22 April 2014 18:37:11 Ley Foon Tan wrote:
>>>>>>>>>>>>>>>>>>>> Hi Arnd and Peter Anvin,
>>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>>> Other than 64-bit time_t, clock_t and suseconds_t, can you confirm
>>>>>>>>>>>>>>>>>>>> that we don't need to have 64 bit off_t? See detail in link below.
>>>>>>>>>>>>>>>>>>>> I can submit the patches for 64-bit time changes
>>>>>>>>>>>>>>>>>>>> (include/asm-generic/posix_types.h and other archs) if everyone is
>>>>>>>>>>>>>>>>>>>> agreed on this.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Yes.
>>>>>>>>>>>> Okay, will doing that.
>>>>>>>>>>
>>>>>>>>>> I believe that arm64 ILP32 will also be affected. What is the status of
>>>>>>>>>> this configuration? Has the glibc/kernel ABI been finalized?
>>>>>>>> Not yet. I am still working out the signal handling part. But we
>>>>>>>> already agreed on 64bit time_t, clock_t, and suseconds_t. And we
>>>>>>>> agreed to a 64bit offset_t too.
>>>>>>>>
>>>>>>>> On a related note suseconds in the timespec in posix is defined to
>>>>>>>> be long. So it would nice if the kernel ignores the upper 32bits so
>>>>>>>> we (glibc developers) can fix this for new targets including x32
>>>>>>>> and arm64/ilp32.
>>>>>>>
>>>>>>> Hmm, but that means for purely 32-bit architectures like nios2, which
>>>>>>> unlike x86_64 or arm64, never has a 64-bit mode, suseconds_t as a 64-bit
>>>>>>> type in the kernel is simply wasted.
>>>>>>
>>>>>> The more I think of this, the more I feel that suseconds_t should jsut
>>>>>> be 'long', not strictly 64-bitified. An ILP32 sub-mode in a 64-bit
>>>>>> kernel should be using compat_* code paths, something like a
>>>>>> COMPAT_USE_32BIT_SUSECONDS case.
>>>>>
>>>>> ILP32 mode should use LP64 syscalls as much as possible and that's the
>>>>> aim with arm64 as well (of course, we still have a few that wouldn't be
>>>>> possible and we route them via compat).
>>>>>
>>>>> But here if time_t is 64-bit while susecconds_t is 32-bit, the compat
>>>>> code wouldn't help.
>>>>
>>>> Why not? You can define the arm64 'struct compat_timeval' with
>>>> suseconds_t as s32, and add the 32<-->64 case in the
>>>> compat_get/put_timeval path, triggered when the process is ILP32 (test
>>>> wrapped in the above hypothetical COMPAT_USE_32BIT_SUSECONDS macro).
>>>> Similar to how x32 does COMPAT_USE_64BIT_TIME.
>>>
>>> We would three timeval then. One for aarch32, one for lp64 and one for ilp32. We really don't want three. Two is already one too many in my mind after developing the ilp32 syscall abi.
>>>
>>> Thanks,
>>> Andrew
>>
>> Okay I now see you're already doing that for 32-bit ARM.
>>
>> Still, you would probably just need to have an arm64-ILP32 specific case
>> to be careful about the last padding word upon kernel entry/exit.
>> (accommodating the difference in sizeof(long)) Penalizing all
>> architectures does not seem like the best solution.
>
> Considering the alignment of long long would be 64bits, the struct does not change sizes if suseconds_t is 32bits or 64bits.
>
>>
>> Having suseconds_t as a strictly 64-bit C type in the kernel, while
>> defined as <= long in user-space may cause other problems.
>>
>> I'll try to explain a probable situation for Nios II. I'm not sure about
>> other soft-cores, but nios2 is sort of uncommon in that the maximum
>> alignment is 4-bytes (32-bits), even for doubles/long-longs.
>
> Yes does that include even if users of aligned? If so that seems broken. Also yes nios ii is unstandard when it comes to alignment here.

You mean using '__attribute__((aligned(8)))'? Sure of course that
enlarges the alignment as expected, but sprinkling that over glibc, or
getting it into the main glibc bits/time.h header is probably not going
to happen...

Thanks,
Chung-Lin

>>
>> So if time_t is 64-bits (which makes sense), then struct timeval, which
>> is time_t+suseconds_t in userspace is 12-bytes/aligned-4 (unlike many
>> archs where a 64-bit time_t will expand the size to 16-bytes, due to
>> align-8)
>
> Unlike most other targets where the struct would 16bits no matter what.
>
> Thanks,
> Andrew
>
>
>>
>> If the kernel suseconds_t is forced to be 64-bits, then nios2 will have
>> a 16-byte kernel timeval vs. 12-byte userspace timeval situation. Just
>> this will require us to do something using compat_*, or weird hacks in
>> glibc, which is unfair. Nios II has no "other-mode". We are just
>> strictly ILP32, everywhere.
>>
>> Of course, we can probably still at the end just use a Nios II specific
>> posix_types.h header to override things, but I'm just stating this as a
>> matter of which are the most reasonable default settings in the generic
>> headers. Making pure ILP32 archs diverge from POSIX standards by default
>> does not seem to be right.
>>
>> Thanks,
>> Chung-Lin
>>

2014-04-25 23:52:16

by Pedro Alves

[permalink] [raw]
Subject: Re: [PATCH 25/28] nios2: ptrace support

Does this support PTRACE_GETREGSET / PTRACE_SETREGSET ?

IMO, the kernel shouldn't accept ports without those anymore.

And IMHO, we shouldn't even allow new ports having PTRACE_GETREGS.

http://sourceware.org/ml/archer/2010-q3/msg00193.html

--
Pedro Alves

2014-04-28 06:03:05

by Ley Foon Tan

[permalink] [raw]
Subject: Re: [PATCH 25/28] nios2: ptrace support

On Sat, Apr 26, 2014 at 7:52 AM, Pedro Alves <[email protected]> wrote:
> Does this support PTRACE_GETREGSET / PTRACE_SETREGSET ?
>
> IMO, the kernel shouldn't accept ports without those anymore.
>
> And IMHO, we shouldn't even allow new ports having PTRACE_GETREGS.
>
> http://sourceware.org/ml/archer/2010-q3/msg00193.html
>

It doesn't support PTRACE_GETREGSET / PTRACE_SETREGSET now.
We will look into this.

Regards
Ley Foon