Re: [PATCH v2] mm: fix possible OOB in numa_rebuild_large_mapping()

[Kefeng Wang <wangkefeng.wang@huawei.com>]

On 2024/6/12 23:27, David Hildenbrand wrote:
> On 12.06.24 14:28, Kefeng Wang wrote:
>> The large folio is mapped with folio size(not greater PMD_SIZE) aligned
>> virtual address during the pagefault, ie, 'addr = 
>> ALIGN_DOWN(vmf->address,
>> nr_pages * PAGE_SIZE)' in do_anonymous_page(). But after the mremap(),
>> the virtual address only requires PAGE_SIZE alignment. Also pte is moved
>> to new in move_page_tables(), then traversal of the new pte in the
>> numa_rebuild_large_mapping() could hit the following issue,
>>
>>     Unable to handle kernel paging request at virtual address 
>> 00000a80c021a788
>>     Mem abort info:
>>       ESR = 0x0000000096000004
>>       EC = 0x25: DABT (current EL), IL = 32 bits
>>       SET = 0, FnV = 0
>>       EA = 0, S1PTW = 0
>>       FSC = 0x04: level 0 translation fault
>>     Data abort info:
>>       ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000
>>       CM = 0, WnR = 0, TnD = 0, TagAccess = 0
>>       GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
>>     user pgtable: 4k pages, 48-bit VAs, pgdp=00002040341a6000
>>     [00000a80c021a788] pgd=0000000000000000, p4d=0000000000000000
>>     Internal error: Oops: 0000000096000004 [#1] SMP
>>     ...
>>     CPU: 76 PID: 15187 Comm: git Kdump: loaded Tainted: G        
>> W          6.10.0-rc2+ #209
>>     Hardware name: Huawei TaiShan 2280 V2/BC82AMDD, BIOS 1.79 08/21/2021
>>     pstate: 60400009 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
>>     pc : numa_rebuild_large_mapping+0x338/0x638
>>     lr : numa_rebuild_large_mapping+0x320/0x638
>>     sp : ffff8000b41c3b00
>>     x29: ffff8000b41c3b30 x28: ffff8000812a0000 x27: 00000000000a8000
>>     x26: 00000000000000a8 x25: 0010000000000001 x24: ffff20401c7170f0
>>     x23: 0000ffff33a1e000 x22: 0000ffff33a76000 x21: ffff20400869eca0
>>     x20: 0000ffff33976000 x19: 00000000000000a8 x18: ffffffffffffffff
>>     x17: 0000000000000000 x16: 0000000000000020 x15: ffff8000b41c36a8
>>     x14: 0000000000000000 x13: 205d373831353154 x12: 5b5d333331363732
>>     x11: 000000000011ff78 x10: 000000000011ff10 x9 : ffff800080273f30
>>     x8 : 000000320400869e x7 : c0000000ffffd87f x6 : 00000000001e6ba8
>>     x5 : ffff206f3fb5af88 x4 : 0000000000000000 x3 : 0000000000000000
>>     x2 : 0000000000000000 x1 : fffffdffc0000000 x0 : 00000a80c021a780
>>     Call trace:
>>      numa_rebuild_large_mapping+0x338/0x638
>>      do_numa_page+0x3e4/0x4e0
>>      handle_pte_fault+0x1bc/0x238
>>      __handle_mm_fault+0x20c/0x400
>>      handle_mm_fault+0xa8/0x288
>>      do_page_fault+0x124/0x498
>>      do_translation_fault+0x54/0x80
>>      do_mem_abort+0x4c/0xa8
>>      el0_da+0x40/0x110
>>      el0t_64_sync_handler+0xe4/0x158
>>      el0t_64_sync+0x188/0x190
>>
>> Fix it by making the start and end not only within the vma range, but
>> also within the page table range.
>>
>> Fixes: d2136d749d76 ("mm: support multi-size THP numa balancing")
>> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
>> ---
>> v2:
>> - don't pass nr_pages into numa_rebuild_large_mapping()
>> - address comment and suggestion from David
>>
>>   mm/memory.c | 14 ++++++++++----
>>   1 file changed, 10 insertions(+), 4 deletions(-)
>>
>> diff --git a/mm/memory.c b/mm/memory.c
>> index 0d309cfb703c..60f7a05ad0cd 100644
>> --- a/mm/memory.c
>> +++ b/mm/memory.c
>> @@ -5228,10 +5228,16 @@ static void numa_rebuild_large_mapping(struct 
>> vm_fault *vmf, struct vm_area_stru
>>                          bool ignore_writable, bool pte_write_upgrade)
>>   {
>>       int nr = pte_pfn(fault_pte) - folio_pfn(folio);
>> -    unsigned long start = max(vmf->address - nr * PAGE_SIZE, 
>> vma->vm_start);
>> -    unsigned long end = min(vmf->address + (folio_nr_pages(folio) - 
>> nr) * PAGE_SIZE, vma->vm_end);
>> -    pte_t *start_ptep = vmf->pte - (vmf->address - start) / PAGE_SIZE;
>> -    unsigned long addr;
>> +    unsigned long start, end, addr = vmf->address;
>> +    unsigned long addr_start = addr - (nr << PAGE_SHIFT);
>> +    unsigned long pt_start = ALIGN_DOWN(addr, PMD_SIZE);
>> +    pte_t *start_ptep;
>> +
>> +    /* Stay within the VMA and within the page table. */
>> +    start = max3(addr_start, pt_start, vma->vm_start);
>> +    end = min3(addr_start + folio_size(folio), pt_start + PMD_SIZE,
>> +           vma->vm_end);
>> +    start_ptep = vmf->pte - ((addr - start) >> PAGE_SHIFT);
>>       /* Restore all PTEs' mapping of the large folio */
>>       for (addr = start; addr != end; start_ptep++, addr += PAGE_SIZE) {
> 
> Should do the trick, hopefully ;)

At least passed our test, almost 100% occurred before :)

> 
> Acked-by: David Hildenbrand <david@redhat.com>
>

Thanks