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

["Holger Hoffstätte" <holger@applied-asynchrony.com>]

On 2023-07-06 00:15, Suren Baghdasaryan wrote:
> On Mon, Jul 3, 2023 at 6:52 AM Holger Hoffstätte
> <holger@applied-asynchrony.com> wrote:
>>
>> On 2023-07-03 12:47, Jiri Slaby wrote:
>>> Cc Jacob Young (from kernel bugzilla)
>>>
>>> On 30. 06. 23, 19:40, Suren Baghdasaryan wrote:
>>>> On Fri, Jun 30, 2023 at 1:43 AM Jiri Slaby <jirislaby@kernel.org> wrote:
>>>>>
>>>>> On 30. 06. 23, 10:28, Jiri Slaby wrote:
>>>>>>    > 2348
>>>>>> clone3({flags=CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_THREAD|CLONE_SYSVSEM|CLONE_SETTLS|CLONE_PARENT_SETTID|CLONE_CHILD_CLEARTID, child_tid=0x7fcaa5882990, parent_tid=0x7fcaa5882990, exit_signal=0, stack=0x7fcaa5082000, stack_size=0x7ffe00, tls=0x7fcaa58826c0} => {parent_tid=[2351]}, 88) = 2351
>>>>>>    > 2350  <... clone3 resumed> => {parent_tid=[2372]}, 88) = 2372
>>>>>>    > 2351  <... clone3 resumed> => {parent_tid=[2354]}, 88) = 2354
>>>>>>    > 2351  <... clone3 resumed> => {parent_tid=[2357]}, 88) = 2357
>>>>>>    > 2354  <... clone3 resumed> => {parent_tid=[2355]}, 88) = 2355
>>>>>>    > 2355  <... clone3 resumed> => {parent_tid=[2370]}, 88) = 2370
>>>>>>    > 2370  mmap(NULL, 262144, PROT_READ|PROT_WRITE,
>>>>>> MAP_PRIVATE|MAP_ANONYMOUS, -1, 0 <unfinished ...>
>>>>>>    > 2370  <... mmap resumed>)               = 0x7fca68249000
>>>>>>    > 2372  <... clone3 resumed> => {parent_tid=[2384]}, 88) = 2384
>>>>>>    > 2384  <... clone3 resumed> => {parent_tid=[2388]}, 88) = 2388
>>>>>>    > 2388  <... clone3 resumed> => {parent_tid=[2392]}, 88) = 2392
>>>>>>    > 2392  <... clone3 resumed> => {parent_tid=[2395]}, 88) = 2395
>>>>>>    > 2395  write(2, "runtime: marked free object in s"..., 36 <unfinished
>>>>>> ...>
>>>>>>
>>>>>> I.e. IIUC, all are threads (CLONE_VM) and thread 2370 mapped ANON
>>>>>> 0x7fca68249000 - 0x7fca6827ffff and go in thread 2395 thinks for some
>>>>>> reason 0x7fca6824bec8 in that region is "bad".
>>>>
>>>> Thanks for the analysis Jiri.
>>>> Is it possible from these logs to identify whether 2370 finished the
>>>> mmap operation before 2395 tried to access 0x7fca6824bec8? That access
>>>> has to happen only after mmap finishes mapping the region.
>>>
>>> Hi,
>>>
>>> it's hard to tell, but I assume so.
>>>
>>> For now, forget about this go's overly complicated, hard to reproduce case and concentrate on the very nice reduced testcase in:
>>>    https://bugzilla.kernel.org/show_bug.cgi?id=217624
>>> ;)
>>>
>>> FWIW, I can reproduce using the test case too.
>>>
>>> thanks,
>>
>> As another (admittedly correlation-only) data point, I noticed at least hourly crashes
>> of Firefox-114 after upgrading to 6.4.1, which had never happened before with 6.3.x.
>> After reverting 0bff0aaea03e2a3ed6 - with a bit of context fixup due to follow-up
>> commits in 6.4.1 - it has been rock stable again, for several hours now.
> 
> Jiri, Holger, would you be able to try
> https://lore.kernel.org/all/20230705171213.2843068-2-surenb@google.com/
> and see if your issues still exist?

Just in time! Not 2 minutes ago I finished rebuilding 6.4.2 + the last version of
your patches on a second machine (old Intel Sandy Bridge workstation) to be my
crash test dummy. I removed the BROKEN dependency in mm/Kconfig, manually set
PER_VMA_LOCK=y and ... it seems to work?! Boots fine, Firefox seems to work
(but no exhaustive tests yet). I will also rerun a few reboot laps, just to
exercise this a bit harder and see if something comes up.

Tomorrow I'll also try again on my Zen2 Thinkpad and will report back.

Fingers crossed!

cheers
Holger