Message ID | 1487866978-25863-1-git-send-email-mark.rutland@arm.com |
---|---|
State | Accepted |
Commit | d81bbe6d882461dec4b71dbe2aa85565fcca4187 |
Headers | show |
On Thu, Feb 23, 2017 at 04:25:26PM +0000, Ard Biesheuvel wrote: > On 23 February 2017 at 16:22, Mark Rutland <mark.rutland@arm.com> wrote: > > This reverts commit 0bfc445dec9dd8130d22c9f4476eed7598524129. > > > > When we change the permissions of regions mapped using contiguous > > entries, the architecture requires us to follow a Break-Before-Make > > strategy, breaking *all* associated entries before we can change any of > > the following properties from the entries: > > > > - presence of the contiguous bit > > - output address > > - attributes > > - permissiones > > > > Failure to do so can result in a number of problems (e.g. TLB conflict > > aborts and/or erroneous results from TLB lookups). > > > > See ARM DDI 0487A.k_iss10775, "Misprogramming of the Contiguous bit", > > page D4-1762. > > > > We do not take this into account when altering the permissions of kernel > > segments in mark_rodata_ro(), where we change the permissions of live > > contiguous entires one-by-one, leaving them transiently inconsistent. > > This has been observed to result in failures on some fast model > > configurations. > > > > Unfortunately, we cannot follow Break-Before-Make here as we'd have to > > unmap kernel text and data used to perform the sequence. > > > > For the timebeing, revert commit 0bfc445dec9dd813 so as to avoid issues > > resulting from this misuse of the contiguous bit. > > > > Signed-off-by: Mark Rutland <mark.rutland@arm.com> > > Reported-by: Jean-Philippe Brucker <jean-philippe.brucker@arm.com> > > Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org> > > Cc: Catalin Marinas <catalin.marinas@arm.com> > > Cc: Will Deacon <Will.Deacon@arm.com> > > Cc: stable@vger.kernel.org # v4.10 > > Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Cheers. > > --- > > arch/arm64/mm/mmu.c | 34 ++++------------------------------ > > 1 file changed, 4 insertions(+), 30 deletions(-) > > > > I'm aware that our hugtlbpage code has a similar issue. I'm looking into that > > now, and will address that with separate patches. It should be possible to use > > BBM there as it's a userspace mapping. > > Are you looking into this issue as well? I'm looking into it now. Thanks, Mark.
diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c index 17243e4..c391b1f 100644 --- a/arch/arm64/mm/mmu.c +++ b/arch/arm64/mm/mmu.c @@ -108,10 +108,8 @@ static bool pgattr_change_is_safe(u64 old, u64 new) static void alloc_init_pte(pmd_t *pmd, unsigned long addr, unsigned long end, unsigned long pfn, pgprot_t prot, - phys_addr_t (*pgtable_alloc)(void), - bool page_mappings_only) + phys_addr_t (*pgtable_alloc)(void)) { - pgprot_t __prot = prot; pte_t *pte; BUG_ON(pmd_sect(*pmd)); @@ -129,18 +127,7 @@ static void alloc_init_pte(pmd_t *pmd, unsigned long addr, do { pte_t old_pte = *pte; - /* - * Set the contiguous bit for the subsequent group of PTEs if - * its size and alignment are appropriate. - */ - if (((addr | PFN_PHYS(pfn)) & ~CONT_PTE_MASK) == 0) { - if (end - addr >= CONT_PTE_SIZE && !page_mappings_only) - __prot = __pgprot(pgprot_val(prot) | PTE_CONT); - else - __prot = prot; - } - - set_pte(pte, pfn_pte(pfn, __prot)); + set_pte(pte, pfn_pte(pfn, prot)); pfn++; /* @@ -159,7 +146,6 @@ static void alloc_init_pmd(pud_t *pud, unsigned long addr, unsigned long end, phys_addr_t (*pgtable_alloc)(void), bool page_mappings_only) { - pgprot_t __prot = prot; pmd_t *pmd; unsigned long next; @@ -186,18 +172,7 @@ static void alloc_init_pmd(pud_t *pud, unsigned long addr, unsigned long end, /* try section mapping first */ if (((addr | next | phys) & ~SECTION_MASK) == 0 && !page_mappings_only) { - /* - * Set the contiguous bit for the subsequent group of - * PMDs if its size and alignment are appropriate. - */ - if (((addr | phys) & ~CONT_PMD_MASK) == 0) { - if (end - addr >= CONT_PMD_SIZE) - __prot = __pgprot(pgprot_val(prot) | - PTE_CONT); - else - __prot = prot; - } - pmd_set_huge(pmd, phys, __prot); + pmd_set_huge(pmd, phys, prot); /* * After the PMD entry has been populated once, we @@ -207,8 +182,7 @@ static void alloc_init_pmd(pud_t *pud, unsigned long addr, unsigned long end, pmd_val(*pmd))); } else { alloc_init_pte(pmd, addr, next, __phys_to_pfn(phys), - prot, pgtable_alloc, - page_mappings_only); + prot, pgtable_alloc); BUG_ON(pmd_val(old_pmd) != 0 && pmd_val(old_pmd) != pmd_val(*pmd));
This reverts commit 0bfc445dec9dd8130d22c9f4476eed7598524129. When we change the permissions of regions mapped using contiguous entries, the architecture requires us to follow a Break-Before-Make strategy, breaking *all* associated entries before we can change any of the following properties from the entries: - presence of the contiguous bit - output address - attributes - permissiones Failure to do so can result in a number of problems (e.g. TLB conflict aborts and/or erroneous results from TLB lookups). See ARM DDI 0487A.k_iss10775, "Misprogramming of the Contiguous bit", page D4-1762. We do not take this into account when altering the permissions of kernel segments in mark_rodata_ro(), where we change the permissions of live contiguous entires one-by-one, leaving them transiently inconsistent. This has been observed to result in failures on some fast model configurations. Unfortunately, we cannot follow Break-Before-Make here as we'd have to unmap kernel text and data used to perform the sequence. For the timebeing, revert commit 0bfc445dec9dd813 so as to avoid issues resulting from this misuse of the contiguous bit. Signed-off-by: Mark Rutland <mark.rutland@arm.com> Reported-by: Jean-Philippe Brucker <jean-philippe.brucker@arm.com> Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Will Deacon <Will.Deacon@arm.com> Cc: stable@vger.kernel.org # v4.10 --- arch/arm64/mm/mmu.c | 34 ++++------------------------------ 1 file changed, 4 insertions(+), 30 deletions(-) I'm aware that our hugtlbpage code has a similar issue. I'm looking into that now, and will address that with separate patches. It should be possible to use BBM there as it's a userspace mapping. Mark. -- 1.9.1