On 2024/7/10 8:14, Hugh Dickins wrote:
> On Tue, 9 Jul 2024, Miaohe Lin wrote:
>
>> A kernel crash was observed when migrating hugetlb folio:
>>
>> BUG: kernel NULL pointer dereference, address: 0000000000000008
>> PGD 0 P4D 0
>> Oops: Oops: 0002 [#1] PREEMPT SMP NOPTI
>> CPU: 0 PID: 3435 Comm: bash Not tainted 6.10.0-rc6-00450-g8578ca01f21f #66
>> RIP: 0010:__folio_undo_large_rmappable+0x70/0xb0
>> RSP: 0018:ffffb165c98a7b38 EFLAGS: 00000097
>> RAX: fffffbbc44528090 RBX: 0000000000000000 RCX: 0000000000000000
>> RDX: ffffa30e000a2800 RSI: 0000000000000246 RDI: ffffa3153ffffcc0
>> RBP: fffffbbc44528000 R08: 0000000000002371 R09: ffffffffbe4e5868
>> R10: 0000000000000001 R11: 0000000000000001 R12: ffffa3153ffffcc0
>> R13: fffffbbc44468000 R14: 0000000000000001 R15: 0000000000000001
>> FS: 00007f5b3a716740(0000) GS:ffffa3151fc00000(0000) knlGS:0000000000000000
>> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> CR2: 0000000000000008 CR3: 000000010959a000 CR4: 00000000000006f0
>> Call Trace:
>> <TASK>
>> __folio_migrate_mapping+0x59e/0x950
>> __migrate_folio.constprop.0+0x5f/0x120
>> move_to_new_folio+0xfd/0x250
>> migrate_pages+0x383/0xd70
>> soft_offline_page+0x2ab/0x7f0
>> soft_offline_page_store+0x52/0x90
>> kernfs_fop_write_iter+0x12c/0x1d0
>> vfs_write+0x380/0x540
>> ksys_write+0x64/0xe0
>> do_syscall_64+0xb9/0x1d0
>> entry_SYSCALL_64_after_hwframe+0x77/0x7f
>> RIP: 0033:0x7f5b3a514887
>> RSP: 002b:00007ffe138fce68 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
>> RAX: ffffffffffffffda RBX: 000000000000000c RCX: 00007f5b3a514887
>> RDX: 000000000000000c RSI: 0000556ab809ee10 RDI: 0000000000000001
>> RBP: 0000556ab809ee10 R08: 00007f5b3a5d1460 R09: 000000007fffffff
>> R10: 0000000000000000 R11: 0000000000000246 R12: 000000000000000c
>> R13: 00007f5b3a61b780 R14: 00007f5b3a617600 R15: 00007f5b3a616a00
>>
>> It's because hugetlb folio is passed to __folio_undo_large_rmappable()
>> unexpectedly. large_rmappable flag is imperceptibly set to hugetlb folio
>> since commit f6a8dd98a2ce ("hugetlb: convert alloc_buddy_hugetlb_folio to
>> use a folio"). Then commit be9581ea8c05 ("mm: fix crashes from deferred
>> split racing folio migration") makes folio_migrate_mapping() call
>> folio_undo_large_rmappable() triggering the bug. Fix this issue by
>> clearing large_rmappable flag for hugetlb folios. They don't need that
>> flag set anyway.
>
> Gosh, thanks a lot for catching this: it had not crossed my mind that
> a folio which passes (folio_test_large and) folio_test_large_rmappable
> might not be suitable for folio_undo_large_rmappable.
>
>>
>> Fixes: f6a8dd98a2ce ("hugetlb: convert alloc_buddy_hugetlb_folio to use a folio")
>
> That's in 6.10-rc, isn't it?
>
>> Fixes: be9581ea8c05 ("mm: fix crashes from deferred split racing folio migration")
>
> And that's in mm-hotfixes-stable intended for 6.10 final.
>
>> Signed-off-by: Miaohe Lin <linmiaohe@xxxxxxxxxx>
>> Cc: <stable@xxxxxxxxxxxxxxx>
>
> So if all goes to plan, this shouldn't need the Cc stable.
I think you are right. Cc stable should be removed. But this patch has been merged into
mm-hotfixes-stable branch, so might Andrew can kindly help modify this?
>
> I certainly deserve blame for not thinking of this possibility: but how
> was it working before my commit, when the folio_undo_large_rmappable()
> was being called from mem_cgroup_migrate()? I think that was just as
> liable to crash too.
I reproduced the crash today with commit be9581ea8c05 ("mm: fix crashes from deferred split racing folio migration") reverted:
BUG: kernel NULL pointer dereference, address: 0000000000000008
#PF: supervisor write access in kernel mode
#PF: error_code(0x0002) - not-present page
PGD 0 P4D 0
Oops: 0002 [#1] PREEMPT SMP NOPTI
CPU: 11 PID: 1481 Comm: bash Not tainted 6.9.0-rc4-00265-gf6a8dd98a2ce-dirty #76
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.14.0-0-g155821a1990b-prebuilt.qemu.org 04/01/2014
RIP: 0010:folio_undo_large_rmappable+0xa0/0xe0
RSP: 0018:ffffa4104950fbd0 EFLAGS: 00000097
RAX: ffff9e06001c8800 RBX: ffffc8f6614f0090 RCX: 0000000000000001
RDX: 0000000000000000 RSI: 0000000000000286 RDI: ffff9e060b23db98
RBP: ffffc8f6614f0000 R08: 0000000000002453 R09: ffffffffaead2a48
R10: 0000000000000000 R11: 0000000000000001 R12: ffff9e060b23db98
R13: 0000000000000000 R14: 0000000000000000 R15: ffffa4104950fcc8
FS: 00007f6821eb0740(0000) GS:ffff9e0ddfcc0000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000008 CR3: 0000000884502000 CR4: 00000000000006f0
Call Trace:
<TASK>
mem_cgroup_migrate+0x186/0x1d0
migrate_folio_extra+0x5c/0x90
move_to_new_folio+0xff/0x250
migrate_pages+0x702/0xd20
soft_offline_page+0x29b/0x7a0
soft_offline_page_store+0x52/0x90
kernfs_fop_write_iter+0x12c/0x1d0
vfs_write+0x387/0x550
ksys_write+0x64/0xe0
do_syscall_64+0xc2/0x1d0
entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7f6821d14887
RSP: 002b:00007ffeecdb19a8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
RAX: ffffffffffffffda RBX: 000000000000000c RCX: 00007f6821d14887
RDX: 000000000000000c RSI: 0000563cfd799e10 RDI: 0000000000000001
RBP: 0000563cfd799e10 R08: 00007f6821dd1460 R09: 000000007fffffff
R10: 0000000000000000 R11: 0000000000000246 R12: 000000000000000c
R13: 00007f6821e1b780 R14: 00007f6821e17600 R15: 00007f6821e16a00
This requires memory_hugetlb_accounting is enabled on cgroup2. Or folio_memcg of hugetlb folio will be NULL
and thus simply return before calling folio_undo_large_rmappable() in mem_cgroup_migrate().
memory_hugetlb_accounting isnot enable in my test env, so I didn't trigger this bug earlier.
So Fixes: be9581ea8c05 ("mm: fix crashes from deferred split racing folio migration") tag might also be removed?
>
> I would like to hear definitively from Matthew, whether a hugetlb page
> should or should not be reported as large_rmappable - is your patch here
> just fixing a surprise, or in danger of adding another surprise somewhere?
IIUC, large_rmappable is only used for thp. See below code:
static inline bool is_transparent_hugepage(const struct folio *folio)
{
if (!folio_test_large(folio))
return false;
return is_huge_zero_folio(folio) ||
folio_test_large_rmappable(folio);
}
But I might be miss something.
Thanks.
.
