Re: [PATCH 0/2] block: Allow concurrent BB context changes

[Hanna Czenczek <hreitz@redhat.com>]

On 10.02.24 09:46, Michael Tokarev wrote:
> 09.02.2024 19:51, Hanna Czenczek :
>> On 09.02.24 15:08, Michael Tokarev wrote:
>>> 02.02.2024 17:47, Hanna Czenczek :
>>>> Hi,
>>>>
>>>> Without the AioContext lock, a BB's context may kind of change at any
>>>> time (unless it has a root node, and I/O requests are pending). That
>>>> also means that its own context (BlockBackend.ctx) and that of its 
>>>> root
>>>> node can differ sometimes (while the context is being changed).
>>>
>>> How relevant this is for -stable (8.2 at least) which does not have
>>> "scsi: eliminate AioContext lock" patchset, and in particular,:
>>> v8.2.0-124-geaad0fe260 "scsi: only access SCSIDevice->requests from
>>> one thread"?
>>>
>>> The issue first patch "block-backend: Allow concurrent context changes"
>>> fixes (RHEL-19381) seems to be for 8.1.something, so it exists in 8.2
>>> too, and this particular fix applies to 8.2.
>>>
>>> But with other changes around all this, I'm a bit lost as of what 
>>> should
>>> be done on stable.  Not even thinking about 7.2 here :)
>>
>> Ah, sorry, yes.  Since we do still have the AioContext lock, this 
>> series won’t be necessary in -stable.  Sorry for the noise!
>
> Hm. Now I'm confused even more.. :)
>
> ad89367202 "block-backend: Allow concurrent context changes" - the first
> one in this series - apparently is needed, as it fixes an issue reported
> for qemu 8.1 (https://issues.redhat.com/browse/RHEL-19381).  Or is it not
> the case?

Ah, yes, I got confused there.  There are two (unfortunately? 
fortunately? Red-Hat-internal) comments, one of which describes the 
crash that’s fixed here, so I thought that bug described this crash.  
But the actual description in the report describes something different 
(more like what’s fixed by 
https://lists.nongnu.org/archive/html/qemu-devel/2024-01/msg03649.html, 
but I’m not entirely sure yet).

So basically I got the bug link wrong.  We now have 
https://issues.redhat.com/browse/RHEL-24593, which has been reported 
only against 8.2.

Hanna

> FWIW, truth is born in the noise, not in silence ;)
>
> Thanks,
>
> /mjt
>