Re: [PATCH 1/2] mm/compaction: skip folios containing hwpoisoned pages

[Kaitao Cheng <kaitao.cheng@linux.dev>]

在 2026/10/20 21:58, David Hildenbrand (Arm) 写道:
> On 10/20/26 15:31, Kaitao Cheng wrote:
>>
>>
>> 在 2026/10/20 21:07, David Hildenbrand (Arm) 写道:
>>> On 10/20/26 13:27, Kaitao Cheng wrote:
>>>>
>>>>
>>>> 在 2026/10/11 21:14, Huang, Ying 写道:
>>>>
>>>> On architectures with hardware error correction and exception recovery
>>>> support, accessing faulty memory automatically triggers an exception,
>>>> which itself guarantees atomicity.
>>>>
>>>> On architectures without such support, even if we implement this logic
>>>> in copy_mc_user_highpage() and copy_mc_highpage(), checking the hwpoison
>>>> flag and copying the data remain two separate operations. They cannot
>>>> be performed atomically, so there is still a race window between them,
>>>> however small.
>>>
>>> Yes. And you need proper architecture support to fix it?
>>
>> It does require architecture support, but that doesn't seem practical, haha.
> 
> Well, but that's required when an arch wants to rule out all races.
> 
>>
>>>> Moving the hwpoison check after the copy might address this issue.
>>>> Is there anything wrong with my understanding?
>>>
>>> Will architectures that are not prepared for copy_mc_user_highpage() not simply
>>> crash if they trigger an MCE during an access?
>>>
>>
>> Do we need to check both before and after the copy? This feels a bit
>> redundant, and I'm not sure whether such a small race window warrants
>> an additional check.
>>
> 
> I don't see how checking after the copy makes any sense? :)
> 
> You already crashed if there was a problem, why check? Or what am I missing?
In cases such as transient errors, software poison injection, repeated
bit flips in memory, or bit flips in the ECC check bits themselves, a
page may be faulty while memcpy() still completes normally. If the copied
data is correct, only the PG_hwpoison flag is lost during migration. But
if the copied data is corrupted and the kernel or a process subsequently
treats it as valid, I think the consequences could be more serious than
an immediate crash. However, I cannot estimate how likely this is to happen.

-- 
Thanks
Kaitao Cheng