Re: [PATCH v3] mm/huge_memory: don't unpoison huge_zero_folio

[Miaohe Lin]

On 2024/5/17 15:02, Anshuman Khandual wrote:
> On 5/16/24 17:56, Miaohe Lin wrote:
>> When I did memory failure tests recently, below panic occurs:
>>
>>  kernel BUG at include/linux/mm.h:1135!
>>  invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
>>  CPU: 9 PID: 137 Comm: kswapd1 Not tainted 6.9.0-rc4-00491-gd5ce28f156fe-dirty #14
>>  RIP: 0010:shrink_huge_zero_page_scan+0x168/0x1a0
>>  RSP: 0018:ffff9933c6c57bd0 EFLAGS: 00000246
>>  RAX: 000000000000003e RBX: 0000000000000000 RCX: ffff88f61fc5c9c8
>>  RDX: 0000000000000000 RSI: 0000000000000027 RDI: ffff88f61fc5c9c0
>>  RBP: ffffcd7c446b0000 R08: ffffffff9a9405f0 R09: 0000000000005492
>>  R10: 00000000000030ea R11: ffffffff9a9405f0 R12: 0000000000000000
>>  R13: 0000000000000000 R14: 0000000000000000 R15: ffff88e703c4ac00
>>  FS:  0000000000000000(0000) GS:ffff88f61fc40000(0000) knlGS:0000000000000000
>>  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>  CR2: 000055f4da6e9878 CR3: 0000000c71048000 CR4: 00000000000006f0
>>  Call Trace:
>>   <TASK>
>>   do_shrink_slab+0x14f/0x6a0
>>   shrink_slab+0xca/0x8c0
>>   shrink_node+0x2d0/0x7d0
>>   balance_pgdat+0x33a/0x720
>>   kswapd+0x1f3/0x410
>>   kthread+0xd5/0x100
>>   ret_from_fork+0x2f/0x50
>>   ret_from_fork_asm+0x1a/0x30
>>   </TASK>
>>  Modules linked in: mce_inject hwpoison_inject
>>  ---[ end trace 0000000000000000 ]---
>>  RIP: 0010:shrink_huge_zero_page_scan+0x168/0x1a0
>>  RSP: 0018:ffff9933c6c57bd0 EFLAGS: 00000246
>>  RAX: 000000000000003e RBX: 0000000000000000 RCX: ffff88f61fc5c9c8
>>  RDX: 0000000000000000 RSI: 0000000000000027 RDI: ffff88f61fc5c9c0
>>  RBP: ffffcd7c446b0000 R08: ffffffff9a9405f0 R09: 0000000000005492
>>  R10: 00000000000030ea R11: ffffffff9a9405f0 R12: 0000000000000000
>>  R13: 0000000000000000 R14: 0000000000000000 R15: ffff88e703c4ac00
>>  FS:  0000000000000000(0000) GS:ffff88f61fc40000(0000) knlGS:0000000000000000
>>  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>  CR2: 000055f4da6e9878 CR3: 0000000c71048000 CR4: 00000000000006f0
>>
>> The root cause is that HWPoison flag will be set for huge_zero_folio
>> without increasing the folio refcnt. But then unpoison_memory() will
>> decrease the folio refcnt unexpectly as it appears like a successfully
> 
> Small nit, a typo in here   ^^^^^ s/unexpectly/unexpectedly/.

Thanks for finding out this typo. And thanks Andrew for fixing the typo.

> 
>> hwpoisoned folio leading to VM_BUG_ON_PAGE(page_ref_count(page) == 0)
>> when releasing huge_zero_folio.
>>
>> Skip unpoisoning huge_zero_folio in unpoison_memory() to fix this issue.
>> We're not prepared to unpoison huge_zero_folio yet.
>>
>> Fixes: 478d134e9506 ("mm/huge_memory: do not overkill when splitting huge_zero_page")
> 
> The target commit looks right.
> 
>> Signed-off-by: Miaohe Lin <linmiaohe@xxxxxxxxxx>
>> Acked-by: David Hildenbrand <david@xxxxxxxxxx>
>> Reviewed-by: Yang Shi <shy828301@xxxxxxxxx>
>> Reviewed-by: Oscar Salvador <osalvador@xxxxxxx>
>> Cc: <stable@xxxxxxxxxxxxxxx>
>> ---
>> v3:
>>  Move up is_huge_zero_folio() check and change return value to
>> -EOPNOTSUPP per Oscar.
>>  Collect Reviewed-by and Acked-by tag. Thanks.
>> v2:
>>  Change to simply check for the huge zero page per David. Thanks.
>> ---
>>  mm/memory-failure.c | 7 +++++++
>>  1 file changed, 7 insertions(+)
>>
>> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
>> index 16ada4fb02b7..a9fe9eda593f 100644
>> --- a/mm/memory-failure.c
>> +++ b/mm/memory-failure.c
>> @@ -2546,6 +2546,13 @@ int unpoison_memory(unsigned long pfn)
>>  		goto unlock_mutex;
>>  	}
>>  
>> +	if (is_huge_zero_folio(folio)) {
>> +		unpoison_pr_info("Unpoison: huge zero page is not supported %#lx\n",
>> +				 pfn, &unpoison_rs);
>> +		ret = -EOPNOTSUPP;
>> +		goto unlock_mutex;
>> +	}
>> +
>>  	if (!PageHWPoison(p)) {
>>  		unpoison_pr_info("Unpoison: Page was already unpoisoned %#lx\n",
>>  				 pfn, &unpoison_rs);
> 
> This patch applies on latest linux-next but not on latest mainline as
> is_huge_zero_folio() is absent there.

It could be simply replaced with is_huge_zero_page().

> 
> Reviewed-by: Anshuman Khandual <anshuman.khandual@xxxxxxx>

Thanks.
.