| Message ID | 1440595802-20359-1-git-send-email-ard.biesheuvel@linaro.org |
|---|---|
| State | New |
| Headers | show |
> On 27 aug. 2015, at 11:32, Will Deacon <will.deacon@arm.com> wrote: > > Hi Ard, > >> On Wed, Aug 26, 2015 at 02:30:02PM +0100, Ard Biesheuvel wrote: >> Currently, we infer the UEFI memory region mapping permissions >> from the memory region type (i.e., runtime services code are >> mapped RWX and runtime services data mapped RW-). This appears to >> work fine but is not entirely UEFI spec compliant. So instead, use >> the designated permission attributes to decide how these regions >> should be mapped. >> >> Since UEFIv2.5 introduces a new EFI_MEMORY_RO permission attribute, >> and redefines EFI_MEMORY_WP as a cacheability attribute, use only >> the former as a read-only attribute. For setting the PXN bit, the >> corresponding EFI_MEMORY_XP attribute is used. >> >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> >> --- >> Changes since v1: >> - rewrote page size and alignment check to be more legible >> - use code that is STRICT_MM_TYPECHECKS compliant >> >> Example output of a recent Tianocore build on FVP Foundation model >> is attached below. >> >> arch/arm64/kernel/efi.c | 37 +++++++++++++------- >> 1 file changed, 24 insertions(+), 13 deletions(-) >> >> diff --git a/arch/arm64/kernel/efi.c b/arch/arm64/kernel/efi.c >> index ab21e0d58278..c8d587f46f3e 100644 >> --- a/arch/arm64/kernel/efi.c >> +++ b/arch/arm64/kernel/efi.c >> @@ -235,7 +235,7 @@ static bool __init efi_virtmap_init(void) >> >> for_each_efi_memory_desc(&memmap, md) { >> u64 paddr, npages, size; >> - pgprot_t prot; >> + pteval_t prot_val; >> >> if (!(md->attribute & EFI_MEMORY_RUNTIME)) >> continue; >> @@ -247,22 +247,33 @@ static bool __init efi_virtmap_init(void) >> memrange_efi_to_native(&paddr, &npages); >> size = npages << PAGE_SHIFT; >> >> - pr_info(" EFI remap 0x%016llx => %p\n", >> - md->phys_addr, (void *)md->virt_addr); >> + if (!is_normal_ram(md)) >> + prot_val = PROT_DEVICE_nGnRE; >> + else >> + prot_val = pgprot_val(PAGE_KERNEL_EXEC); >> >> /* >> - * Only regions of type EFI_RUNTIME_SERVICES_CODE need to be >> - * executable, everything else can be mapped with the XN bits >> - * set. >> + * On 64 KB granule kernels, only use strict permissions when >> + * the region does not share a 64 KB page frame with another >> + * region at either end. >> */ >> - if (!is_normal_ram(md)) >> - prot = __pgprot(PROT_DEVICE_nGnRE); >> - else if (md->type == EFI_RUNTIME_SERVICES_CODE) >> - prot = PAGE_KERNEL_EXEC; >> - else >> - prot = PAGE_KERNEL; >> + if (PAGE_SIZE == EFI_PAGE_SIZE || >> + (PAGE_ALIGNED(md->virt_addr) && >> + PAGE_ALIGNED(md->phys_addr + md->num_pages * EFI_PAGE_SIZE))) { > > Why do you use virt_addr instead of phys_addr for the base check? No reason in particular, as far as I remember, so i should probably change that
diff --git a/arch/arm64/kernel/efi.c b/arch/arm64/kernel/efi.c index ab21e0d58278..c8d587f46f3e 100644 --- a/arch/arm64/kernel/efi.c +++ b/arch/arm64/kernel/efi.c @@ -235,7 +235,7 @@ static bool __init efi_virtmap_init(void) for_each_efi_memory_desc(&memmap, md) { u64 paddr, npages, size; - pgprot_t prot; + pteval_t prot_val; if (!(md->attribute & EFI_MEMORY_RUNTIME)) continue; @@ -247,22 +247,33 @@ static bool __init efi_virtmap_init(void) memrange_efi_to_native(&paddr, &npages); size = npages << PAGE_SHIFT; - pr_info(" EFI remap 0x%016llx => %p\n", - md->phys_addr, (void *)md->virt_addr); + if (!is_normal_ram(md)) + prot_val = PROT_DEVICE_nGnRE; + else + prot_val = pgprot_val(PAGE_KERNEL_EXEC); /* - * Only regions of type EFI_RUNTIME_SERVICES_CODE need to be - * executable, everything else can be mapped with the XN bits - * set. + * On 64 KB granule kernels, only use strict permissions when + * the region does not share a 64 KB page frame with another + * region at either end. */ - if (!is_normal_ram(md)) - prot = __pgprot(PROT_DEVICE_nGnRE); - else if (md->type == EFI_RUNTIME_SERVICES_CODE) - prot = PAGE_KERNEL_EXEC; - else - prot = PAGE_KERNEL; + if (PAGE_SIZE == EFI_PAGE_SIZE || + (PAGE_ALIGNED(md->virt_addr) && + PAGE_ALIGNED(md->phys_addr + md->num_pages * EFI_PAGE_SIZE))) { + + if (md->attribute & EFI_MEMORY_RO) + prot_val |= PTE_RDONLY; + if (md->attribute & EFI_MEMORY_XP) + prot_val |= PTE_PXN; + } + + pr_info(" EFI remap 0x%016llx => %p (R%c%c)\n", + md->phys_addr, (void *)md->virt_addr, + prot_val & PTE_RDONLY ? '-' : 'W', + prot_val & PTE_PXN ? '-' : 'X'); - create_pgd_mapping(&efi_mm, paddr, md->virt_addr, size, prot); + create_pgd_mapping(&efi_mm, paddr, md->virt_addr, size, + __pgprot(prot_val)); } return true; }
Currently, we infer the UEFI memory region mapping permissions from the memory region type (i.e., runtime services code are mapped RWX and runtime services data mapped RW-). This appears to work fine but is not entirely UEFI spec compliant. So instead, use the designated permission attributes to decide how these regions should be mapped. Since UEFIv2.5 introduces a new EFI_MEMORY_RO permission attribute, and redefines EFI_MEMORY_WP as a cacheability attribute, use only the former as a read-only attribute. For setting the PXN bit, the corresponding EFI_MEMORY_XP attribute is used. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> --- Changes since v1: - rewrote page size and alignment check to be more legible - use code that is STRICT_MM_TYPECHECKS compliant Example output of a recent Tianocore build on FVP Foundation model is attached below. arch/arm64/kernel/efi.c | 37 +++++++++++++------- 1 file changed, 24 insertions(+), 13 deletions(-)