The kernel permissions are fixed after the kernel page table is created:
avoid that by mapping the kernel 'correctly' the first time.
Patch 2 is a cleanup patch on which the next patches are based on, not
necessary for this patchset though and relies on patch 1.
Patch 3 introduces a new helper to set kernel mapping permissions while
avoiding all the casts when using set_memory_* API.
Patch 4 is the bulk of this work and deals with mapping the kernel with
the right permissions.
Changes in v4:
* Add patch 1 as noted by Jisheng
* Changes patch 2 title as suggested by Anup
* Add Reviewed-by from Anup
Changes in v3:
* Add a patch that factorizes kernel address conversions
* Add a helper called set_kernel_memory in its own patch, as suggested by
Christoph
* Prefer IS_ENABLED over #ifdef, as suggested by Christoph
* Split overly long lines, as suggested by Christoph
* Simplify kernel mapping by mapping ALL text as readonly and taking advantage
of already present code that enables write for init text before
free_initmem_default.
Changes in v2:
* Rebased on top of for-next (and "riscv: mm: fix build errors caused by
mk_pmd()")
* Get rid of protect_kernel_linear_mapping_text_rodata as suggested by
Jisheng
* Improve code in general compared to previous RFC
Alexandre Ghiti (4):
riscv: Remove CONFIG_PHYS_RAM_BASE_FIXED
riscv: Simplify xip and !xip kernel address conversion macros
riscv: Introduce set_kernel_memory helper
riscv: Map the kernel with correct permissions the first time
arch/riscv/Kconfig | 6 --
arch/riscv/include/asm/page.h | 27 ++++----
arch/riscv/include/asm/pgtable.h | 2 +
arch/riscv/include/asm/sections.h | 17 +++++
arch/riscv/include/asm/set_memory.h | 13 ++--
arch/riscv/kernel/setup.c | 11 +--
arch/riscv/mm/init.c | 102 ++++++++++++----------------
arch/riscv/mm/pageattr.c | 10 +++
8 files changed, 95 insertions(+), 93 deletions(-)
--
2.30.2
Make the physical RAM base address available for all kernels, not only
XIP kernels as it will allow to simplify address conversions macros.
Signed-off-by: Alexandre Ghiti <[email protected]>
---
arch/riscv/Kconfig | 6 ------
1 file changed, 6 deletions(-)
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
index b58596b141fc..3d8e7e4bb45c 100644
--- a/arch/riscv/Kconfig
+++ b/arch/riscv/Kconfig
@@ -493,13 +493,8 @@ config STACKPROTECTOR_PER_TASK
def_bool y
depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_TLS
-config PHYS_RAM_BASE_FIXED
- bool "Explicitly specified physical RAM address"
- default n
-
config PHYS_RAM_BASE
hex "Platform Physical RAM address"
- depends on PHYS_RAM_BASE_FIXED
default "0x80000000"
help
This is the physical address of RAM in the system. It has to be
@@ -512,7 +507,6 @@ config XIP_KERNEL
# This prevents XIP from being enabled by all{yes,mod}config, which
# fail to build since XIP doesn't support large kernels.
depends on !COMPILE_TEST
- select PHYS_RAM_BASE_FIXED
help
Execute-In-Place allows the kernel to run from non-volatile storage
directly addressable by the CPU, such as NOR flash. This saves RAM
--
2.30.2
This helper should be used for setting permissions to the kernel
mapping as it takes pointers as arguments and then avoids explicit cast
to unsigned long needed for the set_memory_* API.
Suggested-by: Christoph Hellwig <[email protected]>
Signed-off-by: Alexandre Ghiti <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
---
arch/riscv/include/asm/set_memory.h | 5 +++++
arch/riscv/mm/pageattr.c | 10 ++++++++++
2 files changed, 15 insertions(+)
diff --git a/arch/riscv/include/asm/set_memory.h b/arch/riscv/include/asm/set_memory.h
index 086f757e8ba3..7a411fed9e0e 100644
--- a/arch/riscv/include/asm/set_memory.h
+++ b/arch/riscv/include/asm/set_memory.h
@@ -16,6 +16,7 @@ int set_memory_rw(unsigned long addr, int numpages);
int set_memory_x(unsigned long addr, int numpages);
int set_memory_nx(unsigned long addr, int numpages);
int set_memory_rw_nx(unsigned long addr, int numpages);
+int set_kernel_memory(char *start, char *end, int (*set_memory)(unsigned long, int));
void protect_kernel_text_data(void);
#else
static inline int set_memory_ro(unsigned long addr, int numpages) { return 0; }
@@ -24,6 +25,10 @@ static inline int set_memory_x(unsigned long addr, int numpages) { return 0; }
static inline int set_memory_nx(unsigned long addr, int numpages) { return 0; }
static inline void protect_kernel_text_data(void) {}
static inline int set_memory_rw_nx(unsigned long addr, int numpages) { return 0; }
+static inline int set_kernel_memory(char *start, char *end, int (*set_memory)(unsigned long, int))
+{
+ return 0;
+}
#endif
#if defined(CONFIG_64BIT) && defined(CONFIG_STRICT_KERNEL_RWX)
diff --git a/arch/riscv/mm/pageattr.c b/arch/riscv/mm/pageattr.c
index 5e49e4b4a4cc..c47ac6a432ac 100644
--- a/arch/riscv/mm/pageattr.c
+++ b/arch/riscv/mm/pageattr.c
@@ -156,6 +156,16 @@ int set_memory_nx(unsigned long addr, int numpages)
return __set_memory(addr, numpages, __pgprot(0), __pgprot(_PAGE_EXEC));
}
+int set_kernel_memory(char *startp, char *endp,
+ int (*set_memory)(unsigned long start, int num_pages))
+{
+ unsigned long start = (unsigned long)startp;
+ unsigned long end = (unsigned long)endp;
+ int num_pages = PAGE_ALIGN(end - start) >> PAGE_SHIFT;
+
+ return set_memory(start, num_pages);
+}
+
int set_direct_map_invalid_noflush(struct page *page)
{
int ret;
--
2.30.2
For 64b kernels, we map all the kernel with write and execute permissions
and afterwards remove writability from text and executability from data.
For 32b kernels, the kernel mapping resides in the linear mapping, so we
map all the linear mapping as writable and executable and afterwards we
remove those properties for unused memory and kernel mapping as
described above.
Change this behavior to directly map the kernel with correct permissions
and avoid going through the whole mapping to fix the permissions.
At the same time, this fixes an issue introduced by commit 2bfc6cd81bd1
("riscv: Move kernel mapping outside of linear mapping") as reported
here https://github.com/starfive-tech/linux/issues/17.
Signed-off-by: Alexandre Ghiti <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
---
arch/riscv/include/asm/page.h | 13 +++-
arch/riscv/include/asm/sections.h | 17 +++++
arch/riscv/include/asm/set_memory.h | 8 ---
arch/riscv/kernel/setup.c | 11 +--
arch/riscv/mm/init.c | 102 ++++++++++++----------------
5 files changed, 75 insertions(+), 76 deletions(-)
diff --git a/arch/riscv/include/asm/page.h b/arch/riscv/include/asm/page.h
index 6e004d8fda4d..349e4f9874cc 100644
--- a/arch/riscv/include/asm/page.h
+++ b/arch/riscv/include/asm/page.h
@@ -95,6 +95,7 @@ extern unsigned long va_kernel_pa_offset;
#endif
extern unsigned long va_kernel_xip_pa_offset;
extern unsigned long pfn_base;
+extern uintptr_t load_sz;
#define ARCH_PFN_OFFSET (pfn_base)
#else
#define va_pa_offset 0
@@ -108,6 +109,11 @@ extern unsigned long pfn_base;
extern unsigned long kernel_virt_addr;
#ifdef CONFIG_64BIT
+#define is_kernel_mapping(x) \
+ ((x) >= kernel_virt_addr && (x) < (kernel_virt_addr + load_sz))
+#define is_linear_mapping(x) \
+ ((x) >= PAGE_OFFSET && (x) < kernel_virt_addr)
+
#define linear_mapping_pa_to_va(x) ((void *)((unsigned long)(x) + va_pa_offset))
#define kernel_mapping_pa_to_va(y) ({ \
unsigned long _y = y; \
@@ -127,10 +133,15 @@ extern unsigned long kernel_virt_addr;
#define __va_to_pa_nodebug(x) ({ \
unsigned long _x = x; \
- (_x < kernel_virt_addr) ? \
+ is_linear_mapping(_x) ? \
linear_mapping_va_to_pa(_x) : kernel_mapping_va_to_pa(_x); \
})
#else
+#define is_kernel_mapping(x) \
+ ((x) >= kernel_virt_addr && (x) < (kernel_virt_addr + load_sz))
+#define is_linear_mapping(x) \
+ ((x) >= PAGE_OFFSET)
+
#define __pa_to_va_nodebug(x) ((void *)((unsigned long) (x) + va_pa_offset))
#define __va_to_pa_nodebug(x) ((unsigned long)(x) - va_pa_offset)
#endif /* CONFIG_64BIT */
diff --git a/arch/riscv/include/asm/sections.h b/arch/riscv/include/asm/sections.h
index 8a303fb1ee3b..32336e8a17cb 100644
--- a/arch/riscv/include/asm/sections.h
+++ b/arch/riscv/include/asm/sections.h
@@ -6,6 +6,7 @@
#define __ASM_SECTIONS_H
#include <asm-generic/sections.h>
+#include <linux/mm.h>
extern char _start[];
extern char _start_kernel[];
@@ -13,4 +14,20 @@ extern char __init_data_begin[], __init_data_end[];
extern char __init_text_begin[], __init_text_end[];
extern char __alt_start[], __alt_end[];
+static inline bool is_va_kernel_text(uintptr_t va)
+{
+ uintptr_t start = (uintptr_t)_start;
+ uintptr_t end = (uintptr_t)__init_data_begin;
+
+ return va >= start && va < end;
+}
+
+static inline bool is_va_kernel_lm_alias_text(uintptr_t va)
+{
+ uintptr_t start = (uintptr_t)lm_alias(_start);
+ uintptr_t end = (uintptr_t)lm_alias(__init_data_begin);
+
+ return va >= start && va < end;
+}
+
#endif /* __ASM_SECTIONS_H */
diff --git a/arch/riscv/include/asm/set_memory.h b/arch/riscv/include/asm/set_memory.h
index 7a411fed9e0e..c0b41ed218e1 100644
--- a/arch/riscv/include/asm/set_memory.h
+++ b/arch/riscv/include/asm/set_memory.h
@@ -17,13 +17,11 @@ int set_memory_x(unsigned long addr, int numpages);
int set_memory_nx(unsigned long addr, int numpages);
int set_memory_rw_nx(unsigned long addr, int numpages);
int set_kernel_memory(char *start, char *end, int (*set_memory)(unsigned long, int));
-void protect_kernel_text_data(void);
#else
static inline int set_memory_ro(unsigned long addr, int numpages) { return 0; }
static inline int set_memory_rw(unsigned long addr, int numpages) { return 0; }
static inline int set_memory_x(unsigned long addr, int numpages) { return 0; }
static inline int set_memory_nx(unsigned long addr, int numpages) { return 0; }
-static inline void protect_kernel_text_data(void) {}
static inline int set_memory_rw_nx(unsigned long addr, int numpages) { return 0; }
static inline int set_kernel_memory(char *start, char *end, int (*set_memory)(unsigned long, int))
{
@@ -31,12 +29,6 @@ static inline int set_kernel_memory(char *start, char *end, int (*set_memory)(un
}
#endif
-#if defined(CONFIG_64BIT) && defined(CONFIG_STRICT_KERNEL_RWX)
-void protect_kernel_linear_mapping_text_rodata(void);
-#else
-static inline void protect_kernel_linear_mapping_text_rodata(void) {}
-#endif
-
int set_direct_map_invalid_noflush(struct page *page);
int set_direct_map_default_noflush(struct page *page);
bool kernel_page_present(struct page *page);
diff --git a/arch/riscv/kernel/setup.c b/arch/riscv/kernel/setup.c
index 4db4d0b5911f..b3d0895ce5f7 100644
--- a/arch/riscv/kernel/setup.c
+++ b/arch/riscv/kernel/setup.c
@@ -290,11 +290,6 @@ void __init setup_arch(char **cmdline_p)
init_resources();
sbi_init();
- if (IS_ENABLED(CONFIG_STRICT_KERNEL_RWX)) {
- protect_kernel_text_data();
- protect_kernel_linear_mapping_text_rodata();
- }
-
#ifdef CONFIG_SWIOTLB
swiotlb_init(1);
#endif
@@ -333,11 +328,9 @@ subsys_initcall(topology_init);
void free_initmem(void)
{
- unsigned long init_begin = (unsigned long)__init_begin;
- unsigned long init_end = (unsigned long)__init_end;
-
if (IS_ENABLED(CONFIG_STRICT_KERNEL_RWX))
- set_memory_rw_nx(init_begin, (init_end - init_begin) >> PAGE_SHIFT);
+ set_kernel_memory(lm_alias(__init_begin), lm_alias(__init_end),
+ IS_ENABLED(CONFIG_64BIT) ? set_memory_rw : set_memory_rw_nx);
free_initmem_default(POISON_FREE_INITMEM);
}
diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
index 2d80088f33d5..6b70c345cfc4 100644
--- a/arch/riscv/mm/init.c
+++ b/arch/riscv/mm/init.c
@@ -425,6 +425,42 @@ asmlinkage void __init __copy_data(void)
}
#endif
+#ifdef CONFIG_STRICT_KERNEL_RWX
+static __init pgprot_t pgprot_from_va(uintptr_t va)
+{
+ if (is_va_kernel_text(va))
+ return PAGE_KERNEL_READ_EXEC;
+
+ /*
+ * In 64b kernel, the kernel mapping is outside the linear mapping so we
+ * must protect its linear mapping alias from being executed and written.
+ * And rodata section is marked readonly in mark_rodata_ro.
+ */
+ if (IS_ENABLED(CONFIG_64BIT) && is_va_kernel_lm_alias_text(va))
+ return PAGE_KERNEL_READ;
+
+ return PAGE_KERNEL;
+}
+
+void mark_rodata_ro(void)
+{
+ set_kernel_memory(__start_rodata, _data, set_memory_ro);
+ if (IS_ENABLED(CONFIG_64BIT))
+ set_kernel_memory(lm_alias(__start_rodata), lm_alias(_data),
+ set_memory_ro);
+
+ debug_checkwx();
+}
+#else
+static __init pgprot_t pgprot_from_va(uintptr_t va)
+{
+ if (IS_ENABLED(CONFIG_64BIT) && !is_kernel_mapping(va))
+ return PAGE_KERNEL;
+
+ return PAGE_KERNEL_EXEC;
+}
+#endif /* CONFIG_STRICT_KERNEL_RWX */
+
/*
* setup_vm() is called from head.S with MMU-off.
*
@@ -454,7 +490,8 @@ uintptr_t xiprom, xiprom_sz;
#define xiprom_sz (*((uintptr_t *)XIP_FIXUP(&xiprom_sz)))
#define xiprom (*((uintptr_t *)XIP_FIXUP(&xiprom)))
-static void __init create_kernel_page_table(pgd_t *pgdir, uintptr_t map_size)
+static void __init create_kernel_page_table(pgd_t *pgdir, uintptr_t map_size,
+ __always_unused bool early)
{
uintptr_t va, end_va;
@@ -473,7 +510,7 @@ static void __init create_kernel_page_table(pgd_t *pgdir, uintptr_t map_size)
map_size, PAGE_KERNEL);
}
#else
-static void __init create_kernel_page_table(pgd_t *pgdir, uintptr_t map_size)
+static void __init create_kernel_page_table(pgd_t *pgdir, uintptr_t map_size, bool early)
{
uintptr_t va, end_va;
@@ -481,7 +518,7 @@ static void __init create_kernel_page_table(pgd_t *pgdir, uintptr_t map_size)
for (va = kernel_virt_addr; va < end_va; va += map_size)
create_pgd_mapping(pgdir, va,
load_pa + (va - kernel_virt_addr),
- map_size, PAGE_KERNEL_EXEC);
+ map_size, early ? PAGE_KERNEL_EXEC : pgprot_from_va(va));
}
#endif
@@ -558,7 +595,7 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa)
* us to reach paging_init(). We map all memory banks later
* in setup_vm_final() below.
*/
- create_kernel_page_table(early_pg_dir, map_size);
+ create_kernel_page_table(early_pg_dir, map_size, true);
#ifndef __PAGETABLE_PMD_FOLDED
/* Setup early PMD for DTB */
@@ -634,22 +671,6 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa)
#endif
}
-#if defined(CONFIG_64BIT) && defined(CONFIG_STRICT_KERNEL_RWX)
-void protect_kernel_linear_mapping_text_rodata(void)
-{
- unsigned long text_start = (unsigned long)lm_alias(_start);
- unsigned long init_text_start = (unsigned long)lm_alias(__init_text_begin);
- unsigned long rodata_start = (unsigned long)lm_alias(__start_rodata);
- unsigned long data_start = (unsigned long)lm_alias(_data);
-
- set_memory_ro(text_start, (init_text_start - text_start) >> PAGE_SHIFT);
- set_memory_nx(text_start, (init_text_start - text_start) >> PAGE_SHIFT);
-
- set_memory_ro(rodata_start, (data_start - rodata_start) >> PAGE_SHIFT);
- set_memory_nx(rodata_start, (data_start - rodata_start) >> PAGE_SHIFT);
-}
-#endif
-
static void __init setup_vm_final(void)
{
uintptr_t va, map_size;
@@ -682,21 +703,15 @@ static void __init setup_vm_final(void)
map_size = best_map_size(start, end - start);
for (pa = start; pa < end; pa += map_size) {
va = (uintptr_t)__va(pa);
- create_pgd_mapping(swapper_pg_dir, va, pa,
- map_size,
-#ifdef CONFIG_64BIT
- PAGE_KERNEL
-#else
- PAGE_KERNEL_EXEC
-#endif
- );
+ create_pgd_mapping(swapper_pg_dir, va, pa, map_size,
+ pgprot_from_va(va));
}
}
#ifdef CONFIG_64BIT
/* Map the kernel */
- create_kernel_page_table(swapper_pg_dir, PMD_SIZE);
+ create_kernel_page_table(swapper_pg_dir, PMD_SIZE, false);
#endif
/* Clear fixmap PTE and PMD mappings */
@@ -727,35 +742,6 @@ static inline void setup_vm_final(void)
}
#endif /* CONFIG_MMU */
-#ifdef CONFIG_STRICT_KERNEL_RWX
-void __init protect_kernel_text_data(void)
-{
- unsigned long text_start = (unsigned long)_start;
- unsigned long init_text_start = (unsigned long)__init_text_begin;
- unsigned long init_data_start = (unsigned long)__init_data_begin;
- unsigned long rodata_start = (unsigned long)__start_rodata;
- unsigned long data_start = (unsigned long)_data;
- unsigned long max_low = (unsigned long)(__va(PFN_PHYS(max_low_pfn)));
-
- set_memory_ro(text_start, (init_text_start - text_start) >> PAGE_SHIFT);
- set_memory_ro(init_text_start, (init_data_start - init_text_start) >> PAGE_SHIFT);
- set_memory_nx(init_data_start, (rodata_start - init_data_start) >> PAGE_SHIFT);
- /* rodata section is marked readonly in mark_rodata_ro */
- set_memory_nx(rodata_start, (data_start - rodata_start) >> PAGE_SHIFT);
- set_memory_nx(data_start, (max_low - data_start) >> PAGE_SHIFT);
-}
-
-void mark_rodata_ro(void)
-{
- unsigned long rodata_start = (unsigned long)__start_rodata;
- unsigned long data_start = (unsigned long)_data;
-
- set_memory_ro(rodata_start, (data_start - rodata_start) >> PAGE_SHIFT);
-
- debug_checkwx();
-}
-#endif
-
#ifdef CONFIG_KEXEC_CORE
/*
* reserve_crashkernel() - reserves memory for crash kernel
--
2.30.2
To simplify the kernel address conversion code, make the same definition of
kernel_mapping_pa_to_va and kernel_mapping_va_to_pa compatible for both xip
and !xip kernel by defining XIP_OFFSET to 0 in !xip kernel.
Signed-off-by: Alexandre Ghiti <[email protected]>
Reviewed-by: Anup Patel <[email protected]>
---
arch/riscv/include/asm/page.h | 14 +++-----------
arch/riscv/include/asm/pgtable.h | 2 ++
2 files changed, 5 insertions(+), 11 deletions(-)
diff --git a/arch/riscv/include/asm/page.h b/arch/riscv/include/asm/page.h
index 6a7761c86ec2..6e004d8fda4d 100644
--- a/arch/riscv/include/asm/page.h
+++ b/arch/riscv/include/asm/page.h
@@ -93,9 +93,7 @@ extern unsigned long va_pa_offset;
#ifdef CONFIG_64BIT
extern unsigned long va_kernel_pa_offset;
#endif
-#ifdef CONFIG_XIP_KERNEL
extern unsigned long va_kernel_xip_pa_offset;
-#endif
extern unsigned long pfn_base;
#define ARCH_PFN_OFFSET (pfn_base)
#else
@@ -103,6 +101,7 @@ extern unsigned long pfn_base;
#ifdef CONFIG_64BIT
#define va_kernel_pa_offset 0
#endif
+#define va_kernel_xip_pa_offset 0
#define ARCH_PFN_OFFSET (PAGE_OFFSET >> PAGE_SHIFT)
#endif /* CONFIG_MMU */
@@ -110,29 +109,22 @@ extern unsigned long kernel_virt_addr;
#ifdef CONFIG_64BIT
#define linear_mapping_pa_to_va(x) ((void *)((unsigned long)(x) + va_pa_offset))
-#ifdef CONFIG_XIP_KERNEL
#define kernel_mapping_pa_to_va(y) ({ \
unsigned long _y = y; \
(_y >= CONFIG_PHYS_RAM_BASE) ? \
(void *)((unsigned long)(_y) + va_kernel_pa_offset + XIP_OFFSET) : \
(void *)((unsigned long)(_y) + va_kernel_xip_pa_offset); \
})
-#else
-#define kernel_mapping_pa_to_va(x) ((void *)((unsigned long)(x) + va_kernel_pa_offset))
-#endif
#define __pa_to_va_nodebug(x) linear_mapping_pa_to_va(x)
#define linear_mapping_va_to_pa(x) ((unsigned long)(x) - va_pa_offset)
-#ifdef CONFIG_XIP_KERNEL
#define kernel_mapping_va_to_pa(y) ({ \
unsigned long _y = y; \
(_y < kernel_virt_addr + XIP_OFFSET) ? \
((unsigned long)(_y) - va_kernel_xip_pa_offset) : \
((unsigned long)(_y) - va_kernel_pa_offset - XIP_OFFSET); \
})
-#else
-#define kernel_mapping_va_to_pa(x) ((unsigned long)(x) - va_kernel_pa_offset)
-#endif
+
#define __va_to_pa_nodebug(x) ({ \
unsigned long _x = x; \
(_x < kernel_virt_addr) ? \
@@ -141,7 +133,7 @@ extern unsigned long kernel_virt_addr;
#else
#define __pa_to_va_nodebug(x) ((void *)((unsigned long) (x) + va_pa_offset))
#define __va_to_pa_nodebug(x) ((unsigned long)(x) - va_pa_offset)
-#endif
+#endif /* CONFIG_64BIT */
#ifdef CONFIG_DEBUG_VIRTUAL
extern phys_addr_t __virt_to_phys(unsigned long x);
diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
index bde8ce3bfe7c..d98e931a31e5 100644
--- a/arch/riscv/include/asm/pgtable.h
+++ b/arch/riscv/include/asm/pgtable.h
@@ -77,6 +77,8 @@
#ifdef CONFIG_XIP_KERNEL
#define XIP_OFFSET SZ_8M
+#else
+#define XIP_OFFSET 0
#endif
#ifndef __ASSEMBLY__
--
2.30.2
Hi Alexandre,
On Fri, 4 Jun 2021 13:49:48 +0200
Alexandre Ghiti <[email protected]> wrote:
> To simplify the kernel address conversion code, make the same definition of
> kernel_mapping_pa_to_va and kernel_mapping_va_to_pa compatible for both xip
> and !xip kernel by defining XIP_OFFSET to 0 in !xip kernel.
>
> Signed-off-by: Alexandre Ghiti <[email protected]>
> Reviewed-by: Anup Patel <[email protected]>
> ---
> arch/riscv/include/asm/page.h | 14 +++-----------
> arch/riscv/include/asm/pgtable.h | 2 ++
> 2 files changed, 5 insertions(+), 11 deletions(-)
>
> diff --git a/arch/riscv/include/asm/page.h b/arch/riscv/include/asm/page.h
> index 6a7761c86ec2..6e004d8fda4d 100644
> --- a/arch/riscv/include/asm/page.h
> +++ b/arch/riscv/include/asm/page.h
> @@ -93,9 +93,7 @@ extern unsigned long va_pa_offset;
> #ifdef CONFIG_64BIT
> extern unsigned long va_kernel_pa_offset;
> #endif
> -#ifdef CONFIG_XIP_KERNEL
> extern unsigned long va_kernel_xip_pa_offset;
> -#endif
> extern unsigned long pfn_base;
> #define ARCH_PFN_OFFSET (pfn_base)
> #else
> @@ -103,6 +101,7 @@ extern unsigned long pfn_base;
> #ifdef CONFIG_64BIT
> #define va_kernel_pa_offset 0
> #endif
> +#define va_kernel_xip_pa_offset 0
> #define ARCH_PFN_OFFSET (PAGE_OFFSET >> PAGE_SHIFT)
> #endif /* CONFIG_MMU */
>
> @@ -110,29 +109,22 @@ extern unsigned long kernel_virt_addr;
>
> #ifdef CONFIG_64BIT
> #define linear_mapping_pa_to_va(x) ((void *)((unsigned long)(x) + va_pa_offset))
> -#ifdef CONFIG_XIP_KERNEL
> #define kernel_mapping_pa_to_va(y) ({ \
> unsigned long _y = y; \
> (_y >= CONFIG_PHYS_RAM_BASE) ? \
For !XIP, IIRC, one rule for riscv kernel is one unified Image for all RV64GC
or RV32GC platforms. When CONFIG_PHYS_RAM_BASE is available for all
riscv platforms after patch1, I'm not sure the common unified Image is still
possible or not.
One possible solution would be forcing CONFIG_PHYS_RAM_BASE = 0 if !XIP in patch1
or remove patch1 instead define CONFIG_PHY_RAM_BASE 0 for !XIP
Thanks
> (void *)((unsigned long)(_y) + va_kernel_pa_offset + XIP_OFFSET) : \
> (void *)((unsigned long)(_y) + va_kernel_xip_pa_offset); \
> })
> -#else
> -#define kernel_mapping_pa_to_va(x) ((void *)((unsigned long)(x) + va_kernel_pa_offset))
> -#endif
> #define __pa_to_va_nodebug(x) linear_mapping_pa_to_va(x)
>
> #define linear_mapping_va_to_pa(x) ((unsigned long)(x) - va_pa_offset)
> -#ifdef CONFIG_XIP_KERNEL
> #define kernel_mapping_va_to_pa(y) ({ \
> unsigned long _y = y; \
> (_y < kernel_virt_addr + XIP_OFFSET) ? \
> ((unsigned long)(_y) - va_kernel_xip_pa_offset) : \
> ((unsigned long)(_y) - va_kernel_pa_offset - XIP_OFFSET); \
> })
> -#else
> -#define kernel_mapping_va_to_pa(x) ((unsigned long)(x) - va_kernel_pa_offset)
> -#endif
> +
> #define __va_to_pa_nodebug(x) ({ \
> unsigned long _x = x; \
> (_x < kernel_virt_addr) ? \
> @@ -141,7 +133,7 @@ extern unsigned long kernel_virt_addr;
> #else
> #define __pa_to_va_nodebug(x) ((void *)((unsigned long) (x) + va_pa_offset))
> #define __va_to_pa_nodebug(x) ((unsigned long)(x) - va_pa_offset)
> -#endif
> +#endif /* CONFIG_64BIT */
>
> #ifdef CONFIG_DEBUG_VIRTUAL
> extern phys_addr_t __virt_to_phys(unsigned long x);
> diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
> index bde8ce3bfe7c..d98e931a31e5 100644
> --- a/arch/riscv/include/asm/pgtable.h
> +++ b/arch/riscv/include/asm/pgtable.h
> @@ -77,6 +77,8 @@
>
> #ifdef CONFIG_XIP_KERNEL
> #define XIP_OFFSET SZ_8M
> +#else
> +#define XIP_OFFSET 0
> #endif
>
> #ifndef __ASSEMBLY__
Le 4/06/2021 ? 14:47, Jisheng Zhang a ?crit?:
> Hi Alexandre,
>
> On Fri, 4 Jun 2021 13:49:48 +0200
> Alexandre Ghiti <[email protected]> wrote:
>
>> To simplify the kernel address conversion code, make the same definition of
>> kernel_mapping_pa_to_va and kernel_mapping_va_to_pa compatible for both xip
>> and !xip kernel by defining XIP_OFFSET to 0 in !xip kernel.
>>
>> Signed-off-by: Alexandre Ghiti <[email protected]>
>> Reviewed-by: Anup Patel <[email protected]>
>> ---
>> arch/riscv/include/asm/page.h | 14 +++-----------
>> arch/riscv/include/asm/pgtable.h | 2 ++
>> 2 files changed, 5 insertions(+), 11 deletions(-)
>>
>> diff --git a/arch/riscv/include/asm/page.h b/arch/riscv/include/asm/page.h
>> index 6a7761c86ec2..6e004d8fda4d 100644
>> --- a/arch/riscv/include/asm/page.h
>> +++ b/arch/riscv/include/asm/page.h
>> @@ -93,9 +93,7 @@ extern unsigned long va_pa_offset;
>> #ifdef CONFIG_64BIT
>> extern unsigned long va_kernel_pa_offset;
>> #endif
>> -#ifdef CONFIG_XIP_KERNEL
>> extern unsigned long va_kernel_xip_pa_offset;
>> -#endif
>> extern unsigned long pfn_base;
>> #define ARCH_PFN_OFFSET (pfn_base)
>> #else
>> @@ -103,6 +101,7 @@ extern unsigned long pfn_base;
>> #ifdef CONFIG_64BIT
>> #define va_kernel_pa_offset 0
>> #endif
>> +#define va_kernel_xip_pa_offset 0
>> #define ARCH_PFN_OFFSET (PAGE_OFFSET >> PAGE_SHIFT)
>> #endif /* CONFIG_MMU */
>>
>> @@ -110,29 +109,22 @@ extern unsigned long kernel_virt_addr;
>>
>> #ifdef CONFIG_64BIT
>> #define linear_mapping_pa_to_va(x) ((void *)((unsigned long)(x) + va_pa_offset))
>> -#ifdef CONFIG_XIP_KERNEL
>> #define kernel_mapping_pa_to_va(y) ({ \
>> unsigned long _y = y; \
>> (_y >= CONFIG_PHYS_RAM_BASE) ? \
>
> For !XIP, IIRC, one rule for riscv kernel is one unified Image for all RV64GC
> or RV32GC platforms. When CONFIG_PHYS_RAM_BASE is available for all
> riscv platforms after patch1, I'm not sure the common unified Image is still
> possible or not.
>
I completely did not handle the 32b case and was convinced that
0x8000_0000 was indeed the beginning of the RAM for all riscv64 plaforms.
> One possible solution would be forcing CONFIG_PHYS_RAM_BASE = 0 if !XIP in patch1
> or remove patch1 instead define CONFIG_PHY_RAM_BASE 0 for !XIP
Setting CONFIG_PHYS_RAM_BASE to 0 for !xip kernels is IMO not that great
since it represents something that is real but wrong (unlike XIP_OFFSET
which is 0 for !xip kernel and in that case, this is true). I will think
to a better solution and am open to suggestions.
Thanks Jisheng again for your reviews, really appreciated!
Alex
>
> Thanks
>
>> (void *)((unsigned long)(_y) + va_kernel_pa_offset + XIP_OFFSET) : \
>> (void *)((unsigned long)(_y) + va_kernel_xip_pa_offset); \
>> })
>> -#else
>> -#define kernel_mapping_pa_to_va(x) ((void *)((unsigned long)(x) + va_kernel_pa_offset))
>> -#endif
>> #define __pa_to_va_nodebug(x) linear_mapping_pa_to_va(x)
>>
>> #define linear_mapping_va_to_pa(x) ((unsigned long)(x) - va_pa_offset)
>> -#ifdef CONFIG_XIP_KERNEL
>> #define kernel_mapping_va_to_pa(y) ({ \
>> unsigned long _y = y; \
>> (_y < kernel_virt_addr + XIP_OFFSET) ? \
>> ((unsigned long)(_y) - va_kernel_xip_pa_offset) : \
>> ((unsigned long)(_y) - va_kernel_pa_offset - XIP_OFFSET); \
>> })
>> -#else
>> -#define kernel_mapping_va_to_pa(x) ((unsigned long)(x) - va_kernel_pa_offset)
>> -#endif
>> +
>> #define __va_to_pa_nodebug(x) ({ \
>> unsigned long _x = x; \
>> (_x < kernel_virt_addr) ? \
>> @@ -141,7 +133,7 @@ extern unsigned long kernel_virt_addr;
>> #else
>> #define __pa_to_va_nodebug(x) ((void *)((unsigned long) (x) + va_pa_offset))
>> #define __va_to_pa_nodebug(x) ((unsigned long)(x) - va_pa_offset)
>> -#endif
>> +#endif /* CONFIG_64BIT */
>>
>> #ifdef CONFIG_DEBUG_VIRTUAL
>> extern phys_addr_t __virt_to_phys(unsigned long x);
>> diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
>> index bde8ce3bfe7c..d98e931a31e5 100644
>> --- a/arch/riscv/include/asm/pgtable.h
>> +++ b/arch/riscv/include/asm/pgtable.h
>> @@ -77,6 +77,8 @@
>>
>> #ifdef CONFIG_XIP_KERNEL
>> #define XIP_OFFSET SZ_8M
>> +#else
>> +#define XIP_OFFSET 0
>> #endif
>>
>> #ifndef __ASSEMBLY__
>
>
>
> _______________________________________________
> linux-riscv mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-riscv
>
On Fri, 04 Jun 2021 04:49:46 PDT (-0700), [email protected] wrote:
> The kernel permissions are fixed after the kernel page table is created:
> avoid that by mapping the kernel 'correctly' the first time.
>
> Patch 2 is a cleanup patch on which the next patches are based on, not
> necessary for this patchset though and relies on patch 1.
>
> Patch 3 introduces a new helper to set kernel mapping permissions while
> avoiding all the casts when using set_memory_* API.
>
> Patch 4 is the bulk of this work and deals with mapping the kernel with
> the right permissions.
>
> Changes in v4:
> * Add patch 1 as noted by Jisheng
> * Changes patch 2 title as suggested by Anup
> * Add Reviewed-by from Anup
>
> Changes in v3:
> * Add a patch that factorizes kernel address conversions
> * Add a helper called set_kernel_memory in its own patch, as suggested by
> Christoph
> * Prefer IS_ENABLED over #ifdef, as suggested by Christoph
> * Split overly long lines, as suggested by Christoph
> * Simplify kernel mapping by mapping ALL text as readonly and taking advantage
> of already present code that enables write for init text before
> free_initmem_default.
>
> Changes in v2:
> * Rebased on top of for-next (and "riscv: mm: fix build errors caused by
> mk_pmd()")
> * Get rid of protect_kernel_linear_mapping_text_rodata as suggested by
> Jisheng
> * Improve code in general compared to previous RFC
>
> Alexandre Ghiti (4):
> riscv: Remove CONFIG_PHYS_RAM_BASE_FIXED
> riscv: Simplify xip and !xip kernel address conversion macros
> riscv: Introduce set_kernel_memory helper
> riscv: Map the kernel with correct permissions the first time
>
> arch/riscv/Kconfig | 6 --
> arch/riscv/include/asm/page.h | 27 ++++----
> arch/riscv/include/asm/pgtable.h | 2 +
> arch/riscv/include/asm/sections.h | 17 +++++
> arch/riscv/include/asm/set_memory.h | 13 ++--
> arch/riscv/kernel/setup.c | 11 +--
> arch/riscv/mm/init.c | 102 ++++++++++++----------------
> arch/riscv/mm/pageattr.c | 10 +++
> 8 files changed, 95 insertions(+), 93 deletions(-)
Thanks, these are on for-next. I had a few conflicts, let me know if
something went wrong.
On Fri, 4 Jun 2021 at 13:51, Alexandre Ghiti <[email protected]> wrote:
>
> Make the physical RAM base address available for all kernels, not only
> XIP kernels as it will allow to simplify address conversions macros.
Am I just reading it wrong or won't this patch make it so that the same kernel
can't run on two chips with physical ram starting at different addresses?
/Emil
> ---
> arch/riscv/Kconfig | 6 ------
> 1 file changed, 6 deletions(-)
>
> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> index b58596b141fc..3d8e7e4bb45c 100644
> --- a/arch/riscv/Kconfig
> +++ b/arch/riscv/Kconfig
> @@ -493,13 +493,8 @@ config STACKPROTECTOR_PER_TASK
> def_bool y
> depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_TLS
>
> -config PHYS_RAM_BASE_FIXED
> - bool "Explicitly specified physical RAM address"
> - default n
> -
> config PHYS_RAM_BASE
> hex "Platform Physical RAM address"
> - depends on PHYS_RAM_BASE_FIXED
> default "0x80000000"
> help
> This is the physical address of RAM in the system. It has to be
> @@ -512,7 +507,6 @@ config XIP_KERNEL
> # This prevents XIP from being enabled by all{yes,mod}config, which
> # fail to build since XIP doesn't support large kernels.
> depends on !COMPILE_TEST
> - select PHYS_RAM_BASE_FIXED
> help
> Execute-In-Place allows the kernel to run from non-volatile storage
> directly addressable by the CPU, such as NOR flash. This saves RAM
> --
> 2.30.2
>
>
> _______________________________________________
> linux-riscv mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-riscv
On Sat, 12 Jun 2021 16:23:03 PDT (-0700), [email protected] wrote:
> On Fri, 4 Jun 2021 at 13:51, Alexandre Ghiti <[email protected]> wrote:
>>
>> Make the physical RAM base address available for all kernels, not only
>> XIP kernels as it will allow to simplify address conversions macros.
>
> Am I just reading it wrong or won't this patch make it so that the same kernel
> can't run on two chips with physical ram starting at different addresses?
IIUC we were in that position, at least without relocatable kernels.
Maybe I'm misunderstanding this, though?
>
> /Emil
>
>> ---
>> arch/riscv/Kconfig | 6 ------
>> 1 file changed, 6 deletions(-)
>>
>> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
>> index b58596b141fc..3d8e7e4bb45c 100644
>> --- a/arch/riscv/Kconfig
>> +++ b/arch/riscv/Kconfig
>> @@ -493,13 +493,8 @@ config STACKPROTECTOR_PER_TASK
>> def_bool y
>> depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_TLS
>>
>> -config PHYS_RAM_BASE_FIXED
>> - bool "Explicitly specified physical RAM address"
>> - default n
>> -
>> config PHYS_RAM_BASE
>> hex "Platform Physical RAM address"
>> - depends on PHYS_RAM_BASE_FIXED
>> default "0x80000000"
>> help
>> This is the physical address of RAM in the system. It has to be
>> @@ -512,7 +507,6 @@ config XIP_KERNEL
>> # This prevents XIP from being enabled by all{yes,mod}config, which
>> # fail to build since XIP doesn't support large kernels.
>> depends on !COMPILE_TEST
>> - select PHYS_RAM_BASE_FIXED
>> help
>> Execute-In-Place allows the kernel to run from non-volatile storage
>> directly addressable by the CPU, such as NOR flash. This saves RAM
>> --
>> 2.30.2
>>
>>
>> _______________________________________________
>> linux-riscv mailing list
>> [email protected]
>> http://lists.infradead.org/mailman/listinfo/linux-riscv
On Sat, 12 Jun 2021 17:23:51 -0700 (PDT)
Palmer Dabbelt <[email protected]> wrote:
> On Sat, 12 Jun 2021 16:23:03 PDT (-0700), [email protected] wrote:
> > On Fri, 4 Jun 2021 at 13:51, Alexandre Ghiti <[email protected]> wrote:
> >>
> >> Make the physical RAM base address available for all kernels, not only
> >> XIP kernels as it will allow to simplify address conversions macros.
> >
> > Am I just reading it wrong or won't this patch make it so that the same kernel
> > can't run on two chips with physical ram starting at different addresses?
I mentioned this point in http://lists.infradead.org/pipermail/linux-riscv/2021-June/006840.html
>
> IIUC we were in that position, at least without relocatable kernels.
> Maybe I'm misunderstanding this, though?
Just my humble opinion, before this series patch, at least geneirc Image
for RV64 + MMU + !XIP is doable.
Thanks
>
> >
> > /Emil
> >
> >> ---
> >> arch/riscv/Kconfig | 6 ------
> >> 1 file changed, 6 deletions(-)
> >>
> >> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> >> index b58596b141fc..3d8e7e4bb45c 100644
> >> --- a/arch/riscv/Kconfig
> >> +++ b/arch/riscv/Kconfig
> >> @@ -493,13 +493,8 @@ config STACKPROTECTOR_PER_TASK
> >> def_bool y
> >> depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_TLS
> >>
> >> -config PHYS_RAM_BASE_FIXED
> >> - bool "Explicitly specified physical RAM address"
> >> - default n
> >> -
> >> config PHYS_RAM_BASE
> >> hex "Platform Physical RAM address"
> >> - depends on PHYS_RAM_BASE_FIXED
> >> default "0x80000000"
> >> help
> >> This is the physical address of RAM in the system. It has to be
> >> @@ -512,7 +507,6 @@ config XIP_KERNEL
> >> # This prevents XIP from being enabled by all{yes,mod}config, which
> >> # fail to build since XIP doesn't support large kernels.
> >> depends on !COMPILE_TEST
> >> - select PHYS_RAM_BASE_FIXED
> >> help
> >> Execute-In-Place allows the kernel to run from non-volatile storage
> >> directly addressable by the CPU, such as NOR flash. This saves RAM
> >> --
> >> 2.30.2
> >>
> >>
> >> _______________________________________________
> >> linux-riscv mailing list
> >> [email protected]
> >> http://lists.infradead.org/mailman/listinfo/linux-riscv
>
> _______________________________________________
> linux-riscv mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-riscv
Hi,
Le 13/06/2021 ? 02:44, Jisheng Zhang a ?crit?:
> On Sat, 12 Jun 2021 17:23:51 -0700 (PDT)
> Palmer Dabbelt <[email protected]> wrote:
>
>> On Sat, 12 Jun 2021 16:23:03 PDT (-0700), [email protected] wrote:
>>> On Fri, 4 Jun 2021 at 13:51, Alexandre Ghiti <[email protected]> wrote:
>>>>
>>>> Make the physical RAM base address available for all kernels, not only
>>>> XIP kernels as it will allow to simplify address conversions macros.
>>>
>>> Am I just reading it wrong or won't this patch make it so that the same kernel
>>> can't run on two chips with physical ram starting at different addresses?
>
> I mentioned this point in http://lists.infradead.org/pipermail/linux-riscv/2021-June/006840.html
>
>>
>> IIUC we were in that position, at least without relocatable kernels.
>> Maybe I'm misunderstanding this, though?
>
> Just my humble opinion, before this series patch, at least geneirc Image
> for RV64 + MMU + !XIP is doable.
>
This patch declares that the physical ram address is at 0x8000_0000,
whatever the chip, which may not be the case in practice. I did not
expect Palmer would take this one and had planned to simply push a v5
without the first 2 patches, but things happened this week that
prevented me to do that. IMO, we should just wait for a v5 that I'll
push when possible (probably today or in the coming days).
Thanks,
Alex
> Thanks
>
>>
>>>
>>> /Emil
>>>
>>>> ---
>>>> arch/riscv/Kconfig | 6 ------
>>>> 1 file changed, 6 deletions(-)
>>>>
>>>> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
>>>> index b58596b141fc..3d8e7e4bb45c 100644
>>>> --- a/arch/riscv/Kconfig
>>>> +++ b/arch/riscv/Kconfig
>>>> @@ -493,13 +493,8 @@ config STACKPROTECTOR_PER_TASK
>>>> def_bool y
>>>> depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_TLS
>>>>
>>>> -config PHYS_RAM_BASE_FIXED
>>>> - bool "Explicitly specified physical RAM address"
>>>> - default n
>>>> -
>>>> config PHYS_RAM_BASE
>>>> hex "Platform Physical RAM address"
>>>> - depends on PHYS_RAM_BASE_FIXED
>>>> default "0x80000000"
>>>> help
>>>> This is the physical address of RAM in the system. It has to be
>>>> @@ -512,7 +507,6 @@ config XIP_KERNEL
>>>> # This prevents XIP from being enabled by all{yes,mod}config, which
>>>> # fail to build since XIP doesn't support large kernels.
>>>> depends on !COMPILE_TEST
>>>> - select PHYS_RAM_BASE_FIXED
>>>> help
>>>> Execute-In-Place allows the kernel to run from non-volatile storage
>>>> directly addressable by the CPU, such as NOR flash. This saves RAM
>>>> --
>>>> 2.30.2
>>>>
>>>>
>>>> _______________________________________________
>>>> linux-riscv mailing list
>>>> [email protected]
>>>> http://lists.infradead.org/mailman/listinfo/linux-riscv
>>
>> _______________________________________________
>> linux-riscv mailing list
>> [email protected]
>> http://lists.infradead.org/mailman/listinfo/linux-riscv
>
>
>
> _______________________________________________
> linux-riscv mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-riscv
>