Re: [PATCH v4 29/33] x86/mm: try VMA lock-based page fault handling first

["Linux regression tracking #adding (Thorsten Leemhuis)" <regressions@leemhuis.info>]

[CCing the regression list, as it should be in the loop for regressions:
https://docs.kernel.org/admin-guide/reporting-regressions.html]

On 29.06.23 16:40, Jiri Slaby wrote:
> 
> On 27. 02. 23, 18:36, Suren Baghdasaryan wrote:
>> Attempt VMA lock-based page fault handling first, and fall back to the
>> existing mmap_lock-based handling if that fails.
> [...]
>> +    fault = handle_mm_fault(vma, address, flags |
>> FAULT_FLAG_VMA_LOCK, regs);
>> +    vma_end_read(vma);
>> +
>> +    if (!(fault & VM_FAULT_RETRY)) {
>> +        count_vm_vma_lock_event(VMA_LOCK_SUCCESS);
>> +        goto done;
>> +    }
>> +    count_vm_vma_lock_event(VMA_LOCK_RETRY);
> 
> This is apparently not strong enough as it causes go build failures like:
> 
> [  409s] strconv
> [  409s] releasep: m=0x579e2000 m->p=0x5781c600 p->m=0x0 p->status=2
> [  409s] fatal error: releasep: invalid p state
> [  409s]
> 
> [  325s] hash/adler32
> [  325s] hash/crc32
> [  325s] cmd/internal/codesign
> [  336s] fatal error: runtime: out of memory
> 
> There are many kinds of similar errors. It happens in 1-3 out of 20
> builds only.
> 
> If I revert the commit on top of 6.4, they all dismiss. Any idea?
> 
> The downstream report:
> https://bugzilla.suse.com/show_bug.cgi?id=1212775
> [...]

Thanks for the report. To be sure the issue doesn't fall through the
cracks unnoticed, I'm adding it to regzbot, the Linux kernel regression
tracking bot:

#regzbot ^introduced 0bff0aaea03e2a3ed6bfa3021
https://bugzilla.suse.com/show_bug.cgi?id=1212775
#regzbot title mm: failures when building go in 1-3 out of 20 builds
#regzbot ignore-activity

Developers: When fixing the issue, remember to add 'Link:' tags pointing
to the report (the parent of this mail). See page linked in footer for
details.

Ciao, Thorsten (wearing his 'the Linux kernel's regression tracker' hat)
--
Everything you wanna know about Linux kernel regression tracking:
https://linux-regtracking.leemhuis.info/about/#tldr
That page also explains what to do if mails like this annoy you.