| Message ID | 20241209122617.53341-1-luxu.kernel@bytedance.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [v4] riscv: mm: Fix the out of bound issue of vmemmap address | expand |
Hi Xu, On Mon, Dec 9, 2024 at 1:26 PM Xu Lu <luxu.kernel@bytedance.com> wrote: > > In sparse vmemmap model, the virtual address of vmemmap is calculated as: > ((struct page *)VMEMMAP_START - (phys_ram_base >> PAGE_SHIFT)). > And the struct page's va can be calculated with an offset: > (vmemmap + (pfn)). > > However, when initializing struct pages, kernel actually starts from the > first page from the same section that phys_ram_base belongs to. If the > first page's physical address is not (phys_ram_base >> PAGE_SHIFT), then > we get an va below VMEMMAP_START when calculating va for it's struct page. > > For example, if phys_ram_base starts from 0x82000000 with pfn 0x82000, the > first page in the same section is actually pfn 0x80000. During > init_unavailable_range(), we will initialize struct page for pfn 0x80000 > with virtual address ((struct page *)VMEMMAP_START - 0x2000), which is > below VMEMMAP_START as well as PCI_IO_END. > > This commit fixes this bug by introducing a new variable > 'vmemmap_start_pfn' which is aligned with memory section size and using > it to calculate vmemmap address instead of phys_ram_base. > > Fixes: a11dd49dcb93 ("riscv: Sparse-Memory/vmemmap out-of-bounds fix") > Signed-off-by: Xu Lu <luxu.kernel@bytedance.com> > --- > arch/riscv/include/asm/page.h | 1 + > arch/riscv/include/asm/pgtable.h | 2 +- > arch/riscv/mm/init.c | 17 ++++++++++++++++- > 3 files changed, 18 insertions(+), 2 deletions(-) > > diff --git a/arch/riscv/include/asm/page.h b/arch/riscv/include/asm/page.h > index 71aabc5c6713..125f5ecd9565 100644 > --- a/arch/riscv/include/asm/page.h > +++ b/arch/riscv/include/asm/page.h > @@ -122,6 +122,7 @@ struct kernel_mapping { > > extern struct kernel_mapping kernel_map; > extern phys_addr_t phys_ram_base; > +extern unsigned long vmemmap_start_pfn; > > #define is_kernel_mapping(x) \ > ((x) >= kernel_map.virt_addr && (x) < (kernel_map.virt_addr + kernel_map.size)) > diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h > index d4e99eef90ac..050fdc49b5ad 100644 > --- a/arch/riscv/include/asm/pgtable.h > +++ b/arch/riscv/include/asm/pgtable.h > @@ -87,7 +87,7 @@ > * Define vmemmap for pfn_to_page & page_to_pfn calls. Needed if kernel > * is configured with CONFIG_SPARSEMEM_VMEMMAP enabled. > */ > -#define vmemmap ((struct page *)VMEMMAP_START - (phys_ram_base >> PAGE_SHIFT)) > +#define vmemmap ((struct page *)VMEMMAP_START - vmemmap_start_pfn) > > #define PCI_IO_SIZE SZ_16M > #define PCI_IO_END VMEMMAP_START > diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c > index 0e8c20adcd98..d93271cb97b1 100644 > --- a/arch/riscv/mm/init.c > +++ b/arch/riscv/mm/init.c > @@ -33,6 +33,7 @@ > #include <asm/pgtable.h> > #include <asm/sections.h> > #include <asm/soc.h> > +#include <asm/sparsemem.h> > #include <asm/tlbflush.h> > > #include "../kernel/head.h" > @@ -62,6 +63,13 @@ EXPORT_SYMBOL(pgtable_l5_enabled); > phys_addr_t phys_ram_base __ro_after_init; > EXPORT_SYMBOL(phys_ram_base); > > +#ifdef CONFIG_SPARSEMEM_VMEMMAP > +#define VMEMMAP_ADDR_ALIGN (1ULL << SECTION_SIZE_BITS) > + > +unsigned long vmemmap_start_pfn __ro_after_init; > +EXPORT_SYMBOL(vmemmap_start_pfn); > +#endif > + > unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)] > __page_aligned_bss; > EXPORT_SYMBOL(empty_zero_page); > @@ -240,8 +248,12 @@ static void __init setup_bootmem(void) > * Make sure we align the start of the memory on a PMD boundary so that > * at worst, we map the linear mapping with PMD mappings. > */ > - if (!IS_ENABLED(CONFIG_XIP_KERNEL)) > + if (!IS_ENABLED(CONFIG_XIP_KERNEL)) { > phys_ram_base = memblock_start_of_DRAM() & PMD_MASK; > +#ifdef CONFIG_SPARSEMEM_VMEMMAP > + vmemmap_start_pfn = round_down(phys_ram_base, VMEMMAP_ADDR_ALIGN) >> PAGE_SHIFT; > +#endif > + } > > /* > * In 64-bit, any use of __va/__pa before this point is wrong as we > @@ -1101,6 +1113,9 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa) > kernel_map.xiprom_sz = (uintptr_t)(&_exiprom) - (uintptr_t)(&_xiprom); > > phys_ram_base = CONFIG_PHYS_RAM_BASE; > +#ifdef CONFIG_SPARSEMEM_VMEMMAP > + vmemmap_start_pfn = round_down(phys_ram_base, VMEMMAP_ADDR_ALIGN) >> PAGE_SHIFT; > +#endif > kernel_map.phys_addr = (uintptr_t)CONFIG_PHYS_RAM_BASE; > kernel_map.size = (uintptr_t)(&_end) - (uintptr_t)(&_start); > > -- > 2.20.1 > Thanks for the multiple revisions! It looks good to me, so: Reviewed-by: Alexandre Ghiti <alexghiti@rivosinc.com> Thanks again, Alex
Xu Lu <luxu.kernel@bytedance.com> writes: > In sparse vmemmap model, the virtual address of vmemmap is calculated as: > ((struct page *)VMEMMAP_START - (phys_ram_base >> PAGE_SHIFT)). > And the struct page's va can be calculated with an offset: > (vmemmap + (pfn)). > > However, when initializing struct pages, kernel actually starts from the > first page from the same section that phys_ram_base belongs to. If the > first page's physical address is not (phys_ram_base >> PAGE_SHIFT), then > we get an va below VMEMMAP_START when calculating va for it's struct page. > > For example, if phys_ram_base starts from 0x82000000 with pfn 0x82000, the > first page in the same section is actually pfn 0x80000. During > init_unavailable_range(), we will initialize struct page for pfn 0x80000 > with virtual address ((struct page *)VMEMMAP_START - 0x2000), which is > below VMEMMAP_START as well as PCI_IO_END. > > This commit fixes this bug by introducing a new variable > 'vmemmap_start_pfn' which is aligned with memory section size and using > it to calculate vmemmap address instead of phys_ram_base. > > Fixes: a11dd49dcb93 ("riscv: Sparse-Memory/vmemmap out-of-bounds fix") > Signed-off-by: Xu Lu <luxu.kernel@bytedance.com> Verified that the kernel is not trying to access vmemmap/struct pages below VMEMMAP_START. Tested-by: Björn Töpel <bjorn@rivosinc.com> Reviewed-by: Björn Töpel <bjorn@rivosinc.com>
diff --git a/arch/riscv/include/asm/page.h b/arch/riscv/include/asm/page.h index 71aabc5c6713..125f5ecd9565 100644 --- a/arch/riscv/include/asm/page.h +++ b/arch/riscv/include/asm/page.h @@ -122,6 +122,7 @@ struct kernel_mapping { extern struct kernel_mapping kernel_map; extern phys_addr_t phys_ram_base; +extern unsigned long vmemmap_start_pfn; #define is_kernel_mapping(x) \ ((x) >= kernel_map.virt_addr && (x) < (kernel_map.virt_addr + kernel_map.size)) diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h index d4e99eef90ac..050fdc49b5ad 100644 --- a/arch/riscv/include/asm/pgtable.h +++ b/arch/riscv/include/asm/pgtable.h @@ -87,7 +87,7 @@ * Define vmemmap for pfn_to_page & page_to_pfn calls. Needed if kernel * is configured with CONFIG_SPARSEMEM_VMEMMAP enabled. */ -#define vmemmap ((struct page *)VMEMMAP_START - (phys_ram_base >> PAGE_SHIFT)) +#define vmemmap ((struct page *)VMEMMAP_START - vmemmap_start_pfn) #define PCI_IO_SIZE SZ_16M #define PCI_IO_END VMEMMAP_START diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c index 0e8c20adcd98..d93271cb97b1 100644 --- a/arch/riscv/mm/init.c +++ b/arch/riscv/mm/init.c @@ -33,6 +33,7 @@ #include <asm/pgtable.h> #include <asm/sections.h> #include <asm/soc.h> +#include <asm/sparsemem.h> #include <asm/tlbflush.h> #include "../kernel/head.h" @@ -62,6 +63,13 @@ EXPORT_SYMBOL(pgtable_l5_enabled); phys_addr_t phys_ram_base __ro_after_init; EXPORT_SYMBOL(phys_ram_base); +#ifdef CONFIG_SPARSEMEM_VMEMMAP +#define VMEMMAP_ADDR_ALIGN (1ULL << SECTION_SIZE_BITS) + +unsigned long vmemmap_start_pfn __ro_after_init; +EXPORT_SYMBOL(vmemmap_start_pfn); +#endif + unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)] __page_aligned_bss; EXPORT_SYMBOL(empty_zero_page); @@ -240,8 +248,12 @@ static void __init setup_bootmem(void) * Make sure we align the start of the memory on a PMD boundary so that * at worst, we map the linear mapping with PMD mappings. */ - if (!IS_ENABLED(CONFIG_XIP_KERNEL)) + if (!IS_ENABLED(CONFIG_XIP_KERNEL)) { phys_ram_base = memblock_start_of_DRAM() & PMD_MASK; +#ifdef CONFIG_SPARSEMEM_VMEMMAP + vmemmap_start_pfn = round_down(phys_ram_base, VMEMMAP_ADDR_ALIGN) >> PAGE_SHIFT; +#endif + } /* * In 64-bit, any use of __va/__pa before this point is wrong as we @@ -1101,6 +1113,9 @@ asmlinkage void __init setup_vm(uintptr_t dtb_pa) kernel_map.xiprom_sz = (uintptr_t)(&_exiprom) - (uintptr_t)(&_xiprom); phys_ram_base = CONFIG_PHYS_RAM_BASE; +#ifdef CONFIG_SPARSEMEM_VMEMMAP + vmemmap_start_pfn = round_down(phys_ram_base, VMEMMAP_ADDR_ALIGN) >> PAGE_SHIFT; +#endif kernel_map.phys_addr = (uintptr_t)CONFIG_PHYS_RAM_BASE; kernel_map.size = (uintptr_t)(&_end) - (uintptr_t)(&_start);
In sparse vmemmap model, the virtual address of vmemmap is calculated as: ((struct page *)VMEMMAP_START - (phys_ram_base >> PAGE_SHIFT)). And the struct page's va can be calculated with an offset: (vmemmap + (pfn)). However, when initializing struct pages, kernel actually starts from the first page from the same section that phys_ram_base belongs to. If the first page's physical address is not (phys_ram_base >> PAGE_SHIFT), then we get an va below VMEMMAP_START when calculating va for it's struct page. For example, if phys_ram_base starts from 0x82000000 with pfn 0x82000, the first page in the same section is actually pfn 0x80000. During init_unavailable_range(), we will initialize struct page for pfn 0x80000 with virtual address ((struct page *)VMEMMAP_START - 0x2000), which is below VMEMMAP_START as well as PCI_IO_END. This commit fixes this bug by introducing a new variable 'vmemmap_start_pfn' which is aligned with memory section size and using it to calculate vmemmap address instead of phys_ram_base. Fixes: a11dd49dcb93 ("riscv: Sparse-Memory/vmemmap out-of-bounds fix") Signed-off-by: Xu Lu <luxu.kernel@bytedance.com> --- arch/riscv/include/asm/page.h | 1 + arch/riscv/include/asm/pgtable.h | 2 +- arch/riscv/mm/init.c | 17 ++++++++++++++++- 3 files changed, 18 insertions(+), 2 deletions(-)