2012-05-08 09:54:03

by Jonas Bonn

[permalink] [raw]
Subject: [PATCH 0/6] OpenRISC patches for review


Here's a couple of patches that I'm planning on queuing up for 3.5. These
have been through the openrisc mailing list, but there's not much response
on that channel, so I thought I'd run these once by the "big" list.

Comments welcome, otherwise these will go into 3.5.

Thanks,
Jonas

Jonas Bonn (5):
openrisc: implement irqdomains
openrisc: remove unnecessary includes
openrisc: header file cleanups
openrisc: provide dma_map_ops
openrisc: use scratch regs in atomic syscall

Kautuk Consul (1):
openrisc/mm/fault.c: Port OOM changes to do_page_fault

arch/openrisc/Kconfig | 1 +
arch/openrisc/include/asm/Kbuild | 3 +-
arch/openrisc/include/asm/dma-mapping.h | 147 +++++++------------------------
arch/openrisc/include/asm/elf.h | 12 ++-
arch/openrisc/include/asm/ptrace.h | 6 +-
arch/openrisc/kernel/dma.c | 109 +++++++++++++++--------
arch/openrisc/kernel/entry.S | 8 +-
arch/openrisc/kernel/irq.c | 83 ++++++++++--------
arch/openrisc/mm/fault.c | 32 ++++++--
9 files changed, 199 insertions(+), 202 deletions(-)


2012-05-08 09:54:06

by Jonas Bonn

[permalink] [raw]
Subject: [PATCH 2/6] openrisc: remove unnecessary includes


Signed-off-by: Jonas Bonn <[email protected]>
---
arch/openrisc/kernel/irq.c | 4 ----
1 files changed, 0 insertions(+), 4 deletions(-)

diff --git a/arch/openrisc/kernel/irq.c b/arch/openrisc/kernel/irq.c
index 15c5ea3..e935b9d 100644
--- a/arch/openrisc/kernel/irq.c
+++ b/arch/openrisc/kernel/irq.c
@@ -14,15 +14,11 @@
* 2 of the License, or (at your option) any later version.
*/

-#include <linux/ptrace.h>
-#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/init.h>
#include <linux/of.h>
#include <linux/ftrace.h>
#include <linux/irq.h>
-#include <linux/seq_file.h>
-#include <linux/kernel_stat.h>
#include <linux/export.h>
#include <linux/irqdomain.h>
#include <linux/irqflags.h>
--
1.7.0.4

2012-05-08 09:54:15

by Jonas Bonn

[permalink] [raw]
Subject: [PATCH 3/6] openrisc/mm/fault.c: Port OOM changes to do_page_fault


From: Kautuk Consul <[email protected]>

Commit d065bd810b6deb67d4897a14bfe21f8eb526ba99
(mm: retry page fault when blocking on disk transfer) and
commit 37b23e0525d393d48a7d59f870b3bc061a30ccdb
(x86,mm: make pagefault killable)

The above commits introduced changes into the x86 pagefault handler
for making the page fault handler retryable as well as killable.

These changes reduce the mmap_sem hold time, which is crucial
during OOM killer invocation.

Port these changes to openrisc.

Signed-off-by: Mohd. Faris <[email protected]>
Signed-off-by: Kautuk Consul <[email protected]>
Signed-off-by: Jonas Bonn <[email protected]>
---
arch/openrisc/mm/fault.c | 32 ++++++++++++++++++++++++++------
1 files changed, 26 insertions(+), 6 deletions(-)

diff --git a/arch/openrisc/mm/fault.c b/arch/openrisc/mm/fault.c
index a5dce82..40f850e 100644
--- a/arch/openrisc/mm/fault.c
+++ b/arch/openrisc/mm/fault.c
@@ -54,6 +54,7 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long address,
struct vm_area_struct *vma;
siginfo_t info;
int fault;
+ unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;

tsk = current;

@@ -105,6 +106,7 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long address,
if (in_interrupt() || !mm)
goto no_context;

+retry:
down_read(&mm->mmap_sem);
vma = find_vma(mm, address);

@@ -143,6 +145,7 @@ good_area:
if (write_acc) {
if (!(vma->vm_flags & VM_WRITE))
goto bad_area;
+ flags |= FAULT_FLAG_WRITE;
} else {
/* not present */
if (!(vma->vm_flags & (VM_READ | VM_EXEC)))
@@ -159,7 +162,11 @@ good_area:
* the fault.
*/

- fault = handle_mm_fault(mm, vma, address, write_acc);
+ fault = handle_mm_fault(mm, vma, address, flags);
+
+ if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
+ return;
+
if (unlikely(fault & VM_FAULT_ERROR)) {
if (fault & VM_FAULT_OOM)
goto out_of_memory;
@@ -167,11 +174,24 @@ good_area:
goto do_sigbus;
BUG();
}
- /*RGD modeled on Cris */
- if (fault & VM_FAULT_MAJOR)
- tsk->maj_flt++;
- else
- tsk->min_flt++;
+
+ if (flags & FAULT_FLAG_ALLOW_RETRY) {
+ /*RGD modeled on Cris */
+ if (fault & VM_FAULT_MAJOR)
+ tsk->maj_flt++;
+ else
+ tsk->min_flt++;
+ if (fault & VM_FAULT_RETRY) {
+ flags &= ~FAULT_FLAG_ALLOW_RETRY;
+
+ /* No need to up_read(&mm->mmap_sem) as we would
+ * have already released it in __lock_page_or_retry
+ * in mm/filemap.c.
+ */
+
+ goto retry;
+ }
+ }

up_read(&mm->mmap_sem);
return;
--
1.7.0.4

2012-05-08 09:54:13

by Jonas Bonn

[permalink] [raw]
Subject: [PATCH 4/6] openrisc: header file cleanups


elf.h: We can export some of these symbols to userspace. libc needs them
and we just as well provide them as asm/elf.h as copying them into separate
libc headers.

ptrace.h: Having padding in the user_regs_struct isn't of any particular
value and just confuses GDB. spr_defs isn't needed in userspace; libc
has its own copy anyway.

Signed-off-by: Jonas Bonn <[email protected]>
---
arch/openrisc/include/asm/Kbuild | 3 ++-
arch/openrisc/include/asm/elf.h | 12 ++++++++++--
arch/openrisc/include/asm/ptrace.h | 6 ++----
3 files changed, 14 insertions(+), 7 deletions(-)

diff --git a/arch/openrisc/include/asm/Kbuild b/arch/openrisc/include/asm/Kbuild
index dcea5a0..c936483 100644
--- a/arch/openrisc/include/asm/Kbuild
+++ b/arch/openrisc/include/asm/Kbuild
@@ -1,6 +1,7 @@
include include/asm-generic/Kbuild.asm

-header-y += spr_defs.h
+header-y += elf.h
+header-y += ucontext.h

generic-y += atomic.h
generic-y += auxvec.h
diff --git a/arch/openrisc/include/asm/elf.h b/arch/openrisc/include/asm/elf.h
index 2ce603b..a8fe2c5 100644
--- a/arch/openrisc/include/asm/elf.h
+++ b/arch/openrisc/include/asm/elf.h
@@ -20,11 +20,17 @@
#define __ASM_OPENRISC_ELF_H

/*
+ * This files is partially exported to userspace. This allows us to keep
+ * the ELF bits in one place which should assist in keeping the kernel and
+ * userspace in sync.
+ */
+
+/*
* ELF register definitions..
*/
-#include <linux/types.h>
-#include <linux/ptrace.h>

+/* for struct user_regs_struct definition */
+#include <asm/ptrace.h>

/* The OR1K relocation types... not all relevant for module loader */
#define R_OR32_NONE 0
@@ -62,6 +68,8 @@ typedef unsigned long elf_fpregset_t;

#ifdef __KERNEL__

+#include <linux/types.h>
+
/*
* This is used to ensure we don't load something for the wrong architecture.
*/
diff --git a/arch/openrisc/include/asm/ptrace.h b/arch/openrisc/include/asm/ptrace.h
index 4651a73..8555c0c 100644
--- a/arch/openrisc/include/asm/ptrace.h
+++ b/arch/openrisc/include/asm/ptrace.h
@@ -19,8 +19,6 @@
#ifndef __ASM_OPENRISC_PTRACE_H
#define __ASM_OPENRISC_PTRACE_H

-#include <asm/spr_defs.h>
-
#ifndef __ASSEMBLY__
/*
* This is the layout of the regset returned by the GETREGSET ptrace call
@@ -30,13 +28,13 @@ struct user_regs_struct {
unsigned long gpr[32];
unsigned long pc;
unsigned long sr;
- unsigned long pad1;
- unsigned long pad2;
};
#endif

#ifdef __KERNEL__

+#include <asm/spr_defs.h>
+
/*
* Make kernel PTrace/register structures opaque to userspace... userspace can
* access thread state via the regset mechanism. This allows us a bit of
--
1.7.0.4

2012-05-08 09:54:37

by Jonas Bonn

[permalink] [raw]
Subject: [PATCH 6/6] openrisc: use scratch regs in atomic syscall


The function sys_or1k_atomic was using call-saved registers without
restoring their value before returning. This is a faux pas: either
we need to restore their values or use scratch regs; the latter is
less code so that's the route this patch takes.

Thanks to David Hennerström for doing most of the heavy-lifting in
tracking this one down.

Reported-by: Davd Hennerström <[email protected]>
Signed-off-by: Jonas Bonn <[email protected]>
---
arch/openrisc/kernel/entry.S | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/arch/openrisc/kernel/entry.S b/arch/openrisc/kernel/entry.S
index 6e61af8..ddfcaa8 100644
--- a/arch/openrisc/kernel/entry.S
+++ b/arch/openrisc/kernel/entry.S
@@ -1117,10 +1117,10 @@ ENTRY(sys_rt_sigreturn)
ENTRY(sys_or1k_atomic)
/* FIXME: This ignores r3 and always does an XCHG */
DISABLE_INTERRUPTS(r17,r19)
- l.lwz r30,0(r4)
- l.lwz r28,0(r5)
- l.sw 0(r4),r28
- l.sw 0(r5),r30
+ l.lwz r29,0(r4)
+ l.lwz r27,0(r5)
+ l.sw 0(r4),r27
+ l.sw 0(r5),r29
ENABLE_INTERRUPTS(r17)
l.jr r9
l.or r11,r0,r0
--
1.7.0.4

2012-05-08 09:54:55

by Jonas Bonn

[permalink] [raw]
Subject: [PATCH 5/6] openrisc: provide dma_map_ops


This switches OpenRISC over to fully using the generic dma-mapping
framework. This was almost already the case as the architecture's
implementation was essentially a copy of the generic header.

This also brings this architecture in line with the recent changes
to dma_map_ops (adding attributes to ops->alloc).

Signed-off-by: Jonas Bonn <[email protected]>
---
arch/openrisc/include/asm/dma-mapping.h | 147 +++++++------------------------
arch/openrisc/kernel/dma.c | 109 +++++++++++++++--------
2 files changed, 107 insertions(+), 149 deletions(-)

diff --git a/arch/openrisc/include/asm/dma-mapping.h b/arch/openrisc/include/asm/dma-mapping.h
index b206ba4..fab8628 100644
--- a/arch/openrisc/include/asm/dma-mapping.h
+++ b/arch/openrisc/include/asm/dma-mapping.h
@@ -20,150 +20,71 @@
/*
* See Documentation/DMA-API-HOWTO.txt and
* Documentation/DMA-API.txt for documentation.
- *
- * This file is written with the intention of eventually moving over
- * to largely using asm-generic/dma-mapping-common.h in its place.
*/

#include <linux/dma-debug.h>
#include <asm-generic/dma-coherent.h>
#include <linux/kmemcheck.h>
+#include <linux/dma-mapping.h>

#define DMA_ERROR_CODE (~(dma_addr_t)0x0)

+extern struct dma_map_ops or1k_dma_map_ops;

-#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 *or1k_dma_alloc_coherent(struct device *dev, size_t size,
- dma_addr_t *dma_handle, gfp_t flag);
-void or1k_dma_free_coherent(struct device *dev, size_t size, void *vaddr,
- dma_addr_t dma_handle);
-dma_addr_t or1k_map_page(struct device *dev, struct page *page,
- unsigned long offset, size_t size,
- enum dma_data_direction dir,
- struct dma_attrs *attrs);
-void or1k_unmap_page(struct device *dev, dma_addr_t dma_handle,
- size_t size, enum dma_data_direction dir,
- struct dma_attrs *attrs);
-int or1k_map_sg(struct device *dev, struct scatterlist *sg,
- int nents, enum dma_data_direction dir,
- struct dma_attrs *attrs);
-void or1k_unmap_sg(struct device *dev, struct scatterlist *sg,
- int nents, enum dma_data_direction dir,
- struct dma_attrs *attrs);
-void or1k_sync_single_for_cpu(struct device *dev,
- dma_addr_t dma_handle, size_t size,
- enum dma_data_direction dir);
-void or1k_sync_single_for_device(struct device *dev,
- dma_addr_t dma_handle, size_t size,
- enum dma_data_direction dir);
-
-static inline void *dma_alloc_coherent(struct device *dev, size_t size,
- dma_addr_t *dma_handle, gfp_t flag)
+static inline struct dma_map_ops *get_dma_ops(struct device *dev)
{
- void *memory;
-
- memory = or1k_dma_alloc_coherent(dev, size, dma_handle, flag);
-
- debug_dma_alloc_coherent(dev, size, *dma_handle, memory);
- return memory;
+ return &or1k_dma_map_ops;
}

-static inline void dma_free_coherent(struct device *dev, size_t size,
- void *cpu_addr, dma_addr_t dma_handle)
-{
- debug_dma_free_coherent(dev, size, cpu_addr, dma_handle);
- or1k_dma_free_coherent(dev, size, cpu_addr, dma_handle);
-}
+#include <asm-generic/dma-mapping-common.h>

-static inline dma_addr_t dma_map_single(struct device *dev, void *ptr,
- size_t size,
- enum dma_data_direction dir)
-{
- dma_addr_t addr;
-
- kmemcheck_mark_initialized(ptr, size);
- BUG_ON(!valid_dma_direction(dir));
- addr = or1k_map_page(dev, virt_to_page(ptr),
- (unsigned long)ptr & ~PAGE_MASK, size,
- dir, NULL);
- debug_dma_map_page(dev, virt_to_page(ptr),
- (unsigned long)ptr & ~PAGE_MASK, size,
- dir, addr, true);
- return addr;
-}
+#define dma_alloc_coherent(d,s,h,f) dma_alloc_attrs(d,s,h,f,NULL)

-static inline void dma_unmap_single(struct device *dev, dma_addr_t addr,
- size_t size,
- enum dma_data_direction dir)
+static inline void *dma_alloc_attrs(struct device *dev, size_t size,
+ dma_addr_t *dma_handle, gfp_t gfp,
+ struct dma_attrs *attrs)
{
- BUG_ON(!valid_dma_direction(dir));
- or1k_unmap_page(dev, addr, size, dir, NULL);
- debug_dma_unmap_page(dev, addr, size, dir, true);
-}
+ struct dma_map_ops *ops = get_dma_ops(dev);
+ void *memory;

-static inline int dma_map_sg(struct device *dev, struct scatterlist *sg,
- int nents, enum dma_data_direction dir)
-{
- int i, ents;
- struct scatterlist *s;
+ memory = ops->alloc(dev, size, dma_handle, gfp, attrs);

- for_each_sg(sg, s, nents, i)
- kmemcheck_mark_initialized(sg_virt(s), s->length);
- BUG_ON(!valid_dma_direction(dir));
- ents = or1k_map_sg(dev, sg, nents, dir, NULL);
- debug_dma_map_sg(dev, sg, nents, ents, dir);
+ debug_dma_alloc_coherent(dev, size, *dma_handle, memory);

- return ents;
+ return memory;
}

-static inline void dma_unmap_sg(struct device *dev, struct scatterlist *sg,
- int nents, enum dma_data_direction dir)
-{
- BUG_ON(!valid_dma_direction(dir));
- debug_dma_unmap_sg(dev, sg, nents, dir);
- or1k_unmap_sg(dev, sg, nents, dir, NULL);
-}
+#define dma_free_coherent(d,s,c,h) dma_free_attrs(d,s,c,h,NULL)

-static inline dma_addr_t dma_map_page(struct device *dev, struct page *page,
- size_t offset, size_t size,
- enum dma_data_direction dir)
+static inline void dma_free_attrs(struct device *dev, size_t size,
+ void *cpu_addr, dma_addr_t dma_handle,
+ struct dma_attrs *attrs)
{
- dma_addr_t addr;
+ struct dma_map_ops *ops = get_dma_ops(dev);

- kmemcheck_mark_initialized(page_address(page) + offset, size);
- BUG_ON(!valid_dma_direction(dir));
- addr = or1k_map_page(dev, page, offset, size, dir, NULL);
- debug_dma_map_page(dev, page, offset, size, dir, addr, false);
+ debug_dma_free_coherent(dev, size, cpu_addr, dma_handle);

- return addr;
+ ops->free(dev, size, cpu_addr, dma_handle, attrs);
}

-static inline void dma_unmap_page(struct device *dev, dma_addr_t addr,
- size_t size, enum dma_data_direction dir)
+static inline void *dma_alloc_noncoherent(struct device *dev, size_t size,
+ dma_addr_t *dma_handle, gfp_t gfp)
{
- BUG_ON(!valid_dma_direction(dir));
- or1k_unmap_page(dev, addr, size, dir, NULL);
- debug_dma_unmap_page(dev, addr, size, dir, true);
-}
+ struct dma_attrs attrs;

-static inline void dma_sync_single_for_cpu(struct device *dev, dma_addr_t addr,
- size_t size,
- enum dma_data_direction dir)
-{
- BUG_ON(!valid_dma_direction(dir));
- or1k_sync_single_for_cpu(dev, addr, size, dir);
- debug_dma_sync_single_for_cpu(dev, addr, size, dir);
+ dma_set_attr(DMA_ATTR_NON_CONSISTENT, &attrs);
+
+ return dma_alloc_attrs(dev, size, dma_handle, gfp, &attrs);
}

-static inline void dma_sync_single_for_device(struct device *dev,
- dma_addr_t addr, size_t size,
- enum dma_data_direction dir)
+static inline void dma_free_noncoherent(struct device *dev, size_t size,
+ void *cpu_addr, dma_addr_t dma_handle)
{
- BUG_ON(!valid_dma_direction(dir));
- or1k_sync_single_for_device(dev, addr, size, dir);
- debug_dma_sync_single_for_device(dev, addr, size, dir);
+ struct dma_attrs attrs;
+
+ dma_set_attr(DMA_ATTR_NON_CONSISTENT, &attrs);
+
+ dma_free_attrs(dev, size, cpu_addr, dma_handle, &attrs);
}

static inline int dma_supported(struct device *dev, u64 dma_mask)
diff --git a/arch/openrisc/kernel/dma.c b/arch/openrisc/kernel/dma.c
index f1c8ee2..0b77ddb 100644
--- a/arch/openrisc/kernel/dma.c
+++ b/arch/openrisc/kernel/dma.c
@@ -21,13 +21,16 @@

#include <linux/dma-mapping.h>
#include <linux/dma-debug.h>
+#include <linux/export.h>
+#include <linux/dma-attrs.h>

#include <asm/cpuinfo.h>
#include <asm/spr_defs.h>
#include <asm/tlbflush.h>

-static int page_set_nocache(pte_t *pte, unsigned long addr,
- unsigned long next, struct mm_walk *walk)
+static int
+page_set_nocache(pte_t *pte, unsigned long addr,
+ unsigned long next, struct mm_walk *walk)
{
unsigned long cl;

@@ -46,8 +49,9 @@ static int page_set_nocache(pte_t *pte, unsigned long addr,
return 0;
}

-static int page_clear_nocache(pte_t *pte, unsigned long addr,
- unsigned long next, struct mm_walk *walk)
+static int
+page_clear_nocache(pte_t *pte, unsigned long addr,
+ unsigned long next, struct mm_walk *walk)
{
pte_val(*pte) &= ~_PAGE_CI;

@@ -67,9 +71,19 @@ static int page_clear_nocache(pte_t *pte, unsigned long addr,
* cache-inhibit bit on those pages, and makes sure that the pages are
* flushed out of the cache before they are used.
*
+ * If the NON_CONSISTENT attribute is set, then this function just
+ * returns "normal", cachable memory.
+ *
+ * There are additional flags WEAK_ORDERING and WRITE_COMBINE to take
+ * into consideration here, too. All current known implementations of
+ * the OR1K support only strongly ordered memory accesses, so that flag
+ * is being ignored for now; uncached but write-combined memory is a
+ * missing feature of the OR1K.
*/
-void *or1k_dma_alloc_coherent(struct device *dev, size_t size,
- dma_addr_t *dma_handle, gfp_t gfp)
+static void *
+or1k_dma_alloc(struct device *dev, size_t size,
+ dma_addr_t *dma_handle, gfp_t gfp,
+ struct dma_attrs *attrs)
{
unsigned long va;
void *page;
@@ -87,20 +101,23 @@ void *or1k_dma_alloc_coherent(struct device *dev, size_t size,

va = (unsigned long)page;

- /*
- * We need to iterate through the pages, clearing the dcache for
- * them and setting the cache-inhibit bit.
- */
- if (walk_page_range(va, va + size, &walk)) {
- free_pages_exact(page, size);
- return NULL;
+ if (!dma_get_attr(DMA_ATTR_NON_CONSISTENT, attrs)) {
+ /*
+ * We need to iterate through the pages, clearing the dcache for
+ * them and setting the cache-inhibit bit.
+ */
+ if (walk_page_range(va, va + size, &walk)) {
+ free_pages_exact(page, size);
+ return NULL;
+ }
}

return (void *)va;
}

-void or1k_dma_free_coherent(struct device *dev, size_t size, void *vaddr,
- dma_addr_t dma_handle)
+static void
+or1k_dma_free(struct device *dev, size_t size, void *vaddr,
+ dma_addr_t dma_handle, struct dma_attrs *attrs)
{
unsigned long va = (unsigned long)vaddr;
struct mm_walk walk = {
@@ -108,16 +125,19 @@ void or1k_dma_free_coherent(struct device *dev, size_t size, void *vaddr,
.mm = &init_mm
};

- /* walk_page_range shouldn't be able to fail here */
- WARN_ON(walk_page_range(va, va + size, &walk));
+ if (!dma_get_attr(DMA_ATTR_NON_CONSISTENT, attrs)) {
+ /* walk_page_range shouldn't be able to fail here */
+ WARN_ON(walk_page_range(va, va + size, &walk));
+ }

free_pages_exact(vaddr, size);
}

-dma_addr_t or1k_map_page(struct device *dev, struct page *page,
- unsigned long offset, size_t size,
- enum dma_data_direction dir,
- struct dma_attrs *attrs)
+static dma_addr_t
+or1k_map_page(struct device *dev, struct page *page,
+ unsigned long offset, size_t size,
+ enum dma_data_direction dir,
+ struct dma_attrs *attrs)
{
unsigned long cl;
dma_addr_t addr = page_to_phys(page) + offset;
@@ -147,16 +167,18 @@ dma_addr_t or1k_map_page(struct device *dev, struct page *page,
return addr;
}

-void or1k_unmap_page(struct device *dev, dma_addr_t dma_handle,
- size_t size, enum dma_data_direction dir,
- struct dma_attrs *attrs)
+static void
+or1k_unmap_page(struct device *dev, dma_addr_t dma_handle,
+ size_t size, enum dma_data_direction dir,
+ struct dma_attrs *attrs)
{
/* Nothing special to do here... */
}

-int or1k_map_sg(struct device *dev, struct scatterlist *sg,
- int nents, enum dma_data_direction dir,
- struct dma_attrs *attrs)
+static int
+or1k_map_sg(struct device *dev, struct scatterlist *sg,
+ int nents, enum dma_data_direction dir,
+ struct dma_attrs *attrs)
{
struct scatterlist *s;
int i;
@@ -169,9 +191,10 @@ int or1k_map_sg(struct device *dev, struct scatterlist *sg,
return nents;
}

-void or1k_unmap_sg(struct device *dev, struct scatterlist *sg,
- int nents, enum dma_data_direction dir,
- struct dma_attrs *attrs)
+static void
+or1k_unmap_sg(struct device *dev, struct scatterlist *sg,
+ int nents, enum dma_data_direction dir,
+ struct dma_attrs *attrs)
{
struct scatterlist *s;
int i;
@@ -181,9 +204,10 @@ void or1k_unmap_sg(struct device *dev, struct scatterlist *sg,
}
}

-void or1k_sync_single_for_cpu(struct device *dev,
- dma_addr_t dma_handle, size_t size,
- enum dma_data_direction dir)
+static void
+or1k_sync_single_for_cpu(struct device *dev,
+ dma_addr_t dma_handle, size_t size,
+ enum dma_data_direction dir)
{
unsigned long cl;
dma_addr_t addr = dma_handle;
@@ -193,9 +217,10 @@ void or1k_sync_single_for_cpu(struct device *dev,
mtspr(SPR_DCBIR, cl);
}

-void or1k_sync_single_for_device(struct device *dev,
- dma_addr_t dma_handle, size_t size,
- enum dma_data_direction dir)
+static void
+or1k_sync_single_for_device(struct device *dev,
+ dma_addr_t dma_handle, size_t size,
+ enum dma_data_direction dir)
{
unsigned long cl;
dma_addr_t addr = dma_handle;
@@ -205,6 +230,18 @@ void or1k_sync_single_for_device(struct device *dev,
mtspr(SPR_DCBFR, cl);
}

+struct dma_map_ops or1k_dma_map_ops = {
+ .alloc = or1k_dma_alloc,
+ .free = or1k_dma_free,
+ .map_page = or1k_map_page,
+ .unmap_page = or1k_unmap_page,
+ .map_sg = or1k_map_sg,
+ .unmap_sg = or1k_unmap_sg,
+ .sync_single_for_cpu = or1k_sync_single_for_cpu,
+ .sync_single_for_device = or1k_sync_single_for_device,
+};
+EXPORT_SYMBOL(or1k_dma_map_ops);
+
/* Number of entries preallocated for DMA-API debugging */
#define PREALLOC_DMA_DEBUG_ENTRIES (1 << 16)

--
1.7.0.4

2012-05-08 09:54:04

by Jonas Bonn

[permalink] [raw]
Subject: [PATCH 1/6] openrisc: implement irqdomains


This moves OpenRISC to using the irqdomain infrastructure. This doesn't
fundamentally change anything other than that it will be easier to have
multiple interrupt controllers in the future.

Signed-off-by: Jonas Bonn <[email protected]>
---
arch/openrisc/Kconfig | 1 +
arch/openrisc/kernel/irq.c | 79 ++++++++++++++++++++++++++------------------
2 files changed, 48 insertions(+), 32 deletions(-)

diff --git a/arch/openrisc/Kconfig b/arch/openrisc/Kconfig
index a478719..7589051 100644
--- a/arch/openrisc/Kconfig
+++ b/arch/openrisc/Kconfig
@@ -7,6 +7,7 @@ config OPENRISC
def_bool y
select OF
select OF_EARLY_FLATTREE
+ select IRQ_DOMAIN
select HAVE_MEMBLOCK
select ARCH_WANT_OPTIONAL_GPIOLIB
select HAVE_ARCH_TRACEHOOK
diff --git a/arch/openrisc/kernel/irq.c b/arch/openrisc/kernel/irq.c
index 4bfead2..15c5ea3 100644
--- a/arch/openrisc/kernel/irq.c
+++ b/arch/openrisc/kernel/irq.c
@@ -24,7 +24,7 @@
#include <linux/seq_file.h>
#include <linux/kernel_stat.h>
#include <linux/export.h>
-
+#include <linux/irqdomain.h>
#include <linux/irqflags.h>

/* read interrupt enabled status */
@@ -98,6 +98,7 @@ static void or1k_pic_mask_ack(struct irq_data *data)
#endif
}

+#if 0
static int or1k_pic_set_type(struct irq_data *data, unsigned int flow_type)
{
/* There's nothing to do in the PIC configuration when changing
@@ -107,43 +108,64 @@ static int or1k_pic_set_type(struct irq_data *data, unsigned int flow_type)

return irq_setup_alt_chip(data, flow_type);
}
+#endif
+
+static struct irq_chip or1k_dev = {
+ .name = "or1k-PIC",
+ .irq_unmask = or1k_pic_unmask,
+ .irq_mask = or1k_pic_mask,
+ .irq_ack = or1k_pic_ack,
+ .irq_mask_ack = or1k_pic_mask_ack,
+};
+
+static struct irq_domain *root_domain;

static inline int pic_get_irq(int first)
{
- int irq;
+ int hwirq;

- irq = ffs(mfspr(SPR_PICSR) >> first);
+ hwirq = ffs(mfspr(SPR_PICSR) >> first);
+ if (!hwirq)
+ return NO_IRQ;
+ else
+ hwirq = hwirq + first -1;

- return irq ? irq + first - 1 : NO_IRQ;
+ return irq_find_mapping(root_domain, hwirq);
}

-static void __init or1k_irq_init(void)
+
+static int or1k_map(struct irq_domain *d, unsigned int irq, irq_hw_number_t hw)
{
- struct irq_chip_generic *gc;
- struct irq_chip_type *ct;
+ irq_set_chip_and_handler_name(irq, &or1k_dev,
+ handle_level_irq, "level");
+ irq_set_status_flags(irq, IRQ_LEVEL | IRQ_NOPROBE);

- /* Disable all interrupts until explicitly requested */
- mtspr(SPR_PICMR, (0UL));
+ return 0;
+}

- gc = irq_alloc_generic_chip("or1k-PIC", 1, 0, 0, handle_level_irq);
- ct = gc->chip_types;
+static const struct irq_domain_ops or1k_irq_domain_ops = {
+ .xlate = irq_domain_xlate_onecell,
+ .map = or1k_map,
+};

- ct->chip.irq_unmask = or1k_pic_unmask;
- ct->chip.irq_mask = or1k_pic_mask;
- ct->chip.irq_ack = or1k_pic_ack;
- ct->chip.irq_mask_ack = or1k_pic_mask_ack;
- ct->chip.irq_set_type = or1k_pic_set_type;
+/*
+ * This sets up the IRQ domain for the PIC built in to the OpenRISC
+ * 1000 CPU. This is the "root" domain as these are the interrupts
+ * that directly trigger an exception in the CPU.
+ */
+static void __init or1k_irq_init(void)
+{
+ struct device_node *intc = NULL;

- /* The OR1K PIC can handle both level and edge trigged
- * interrupts in roughly the same manner
- */
-#if 0
- /* FIXME: chip.type??? */
- ct->chip.type = IRQ_TYPE_EDGE_BOTH | IRQ_TYPE_LEVEL_MASK;
-#endif
+ /* The interrupt controller device node is mandatory */
+ intc = of_find_compatible_node(NULL, NULL, "opencores,or1k-pic");
+ BUG_ON(!intc);

- irq_setup_generic_chip(gc, IRQ_MSK(NR_IRQS), 0,
- IRQ_NOREQUEST, IRQ_LEVEL | IRQ_NOPROBE);
+ /* Disable all interrupts until explicitly requested */
+ mtspr(SPR_PICMR, (0UL));
+
+ root_domain = irq_domain_add_linear(intc, 32,
+ &or1k_irq_domain_ops, NULL);
}

void __init init_IRQ(void)
@@ -164,10 +186,3 @@ void __irq_entry do_IRQ(struct pt_regs *regs)
irq_exit();
set_irq_regs(old_regs);
}
-
-unsigned int irq_create_of_mapping(struct device_node *controller,
- const u32 *intspec, unsigned int intsize)
-{
- return intspec[0];
-}
-EXPORT_SYMBOL_GPL(irq_create_of_mapping);
--
1.7.0.4

2012-05-08 11:25:05

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 0/6] OpenRISC patches for review

On Tuesday 08 May 2012, Jonas Bonn wrote:
> Here's a couple of patches that I'm planning on queuing up for 3.5. These
> have been through the openrisc mailing list, but there's not much response
> on that channel, so I thought I'd run these once by the "big" list.
>
> Comments welcome, otherwise these will go into 3.5.

They all look good to me.

Acked-by: Arnd Bergmann <[email protected]>