[PATCH v5] mm: huge-vmap: fail gracefully on unexpected huge vmap mappings

[ard.biesheuvel@linaro.org (Ard Biesheuvel)]

On 9 June 2017 at 09:22, Mark Rutland <mark.rutland@arm.com> wrote:
> On Fri, Jun 09, 2017 at 08:22:26AM +0000, Ard Biesheuvel wrote:
>> Existing code that uses vmalloc_to_page() may assume that any
>> address for which is_vmalloc_addr() returns true may be passed
>> into vmalloc_to_page() to retrieve the associated struct page.
>>
>> This is not un unreasonable assumption to make, but on architectures
>> that have CONFIG_HAVE_ARCH_HUGE_VMAP=y, it no longer holds, and we
>> need to ensure that vmalloc_to_page() does not go off into the weeds
>> trying to dereference huge PUDs or PMDs as table entries.
>>
>> Given that vmalloc() and vmap() themselves never create huge
>> mappings or deal with compound pages at all, there is no correct
>> answer in this case, so return NULL instead, and issue a warning.
>>
>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>> ---
>> v5: - fix typo
>>
>> v4: - use pud_bad/pmd_bad instead of pud_huge/pmd_huge, which don't require
>>       changes to hugetlb.h, and give us what we need on all architectures
>>     - move WARN_ON_ONCE() calls out of conditionals

^^^

>>     - add explanatory comment
>>
>>  mm/vmalloc.c | 15 +++++++++++++--
>>  1 file changed, 13 insertions(+), 2 deletions(-)
>>
>> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
>> index 34a1c3e46ed7..0fcd371266a4 100644
>> --- a/mm/vmalloc.c
>> +++ b/mm/vmalloc.c
>> @@ -287,10 +287,21 @@ struct page *vmalloc_to_page(const void *vmalloc_addr)
>>       if (p4d_none(*p4d))
>>               return NULL;
>>       pud = pud_offset(p4d, addr);
>> -     if (pud_none(*pud))
>> +
>> +     /*
>> +      * Don't dereference bad PUD or PMD (below) entries. This will also
>> +      * identify huge mappings, which we may encounter on architectures
>> +      * that define CONFIG_HAVE_ARCH_HUGE_VMAP=y. Such regions will be
>> +      * identified as vmalloc addresses by is_vmalloc_addr(), but are
>> +      * not [unambiguously] associated with a struct page, so there is
>> +      * no correct value to return for them.
>> +      */
>> +     WARN_ON_ONCE(pud_bad(*pud));
>> +     if (pud_none(*pud) || pud_bad(*pud))
>>               return NULL;
>
> Nit: the WARN_ON_ONCE() can be folded into the conditional:
>
>         if (pud_none(*pud) || WARN_ON_ONCE(pud_bad(*pud)))
>                 reutrn NULL;
>
>>       pmd = pmd_offset(pud, addr);
>> -     if (pmd_none(*pmd))
>> +     WARN_ON_ONCE(pmd_bad(*pmd));
>> +     if (pmd_none(*pmd) || pmd_bad(*pmd))
>>               return NULL;
>
> Likewise here.
>

Actually, it was Dave who requested them to be taken out of the conditional.

> Otherwise, looks good to me. FWIW:
>
> Acked-by: Mark Rutland <mark.rutland@arm.com>
>

Thanks,
Ard.