From: Pavel Begunkov <[email protected]>
To: Oleg Nesterov <[email protected]>, Tejun Heo <[email protected]>
Cc: [email protected], Jens Axboe <[email protected]>,
Andrew Morton <[email protected]>,
Christian Brauner <[email protected]>,
Tycho Andersen <[email protected]>,
Thomas Gleixner <[email protected]>,
[email protected], Julian Orth <[email protected]>,
Peter Zijlstra <[email protected]>
Subject: Re: [PATCH 2/2] kernel: rerun task_work while freezing in get_signal()
Date: Tue, 9 Jul 2024 20:26:56 +0100 [thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>
On 7/9/24 20:07, Oleg Nesterov wrote:
> Hi Tejun,
>
> Thanks for looking at this, can you review this V2 patch from Pavel?
> To me it makes sense even without 1/2 which I didn't even bother to
> read. At least as a simple workaround for now.
They are kind of separate but without 1/2 this patch creates
another infinite loop, even though it's harder to hit and
is io_uring specific.
> On 07/09, Tejun Heo wrote:
>>
>> Hello,
>>
>> On Tue, Jul 09, 2024 at 03:05:21PM +0100, Pavel Begunkov wrote:
>>>> -----------------------------------------------------------------------
>>>> Either way I have no idea whether a cgroup_task_frozen() task should
>>>> react to task_work_add(TWA_SIGNAL) or not.
>>>>
>>>> Documentation/admin-guide/cgroup-v2.rst says
>>>>
>>>> Writing "1" to the file causes freezing of the cgroup and all
>>>> descendant cgroups. This means that all belonging processes will
>>>> be stopped and will not run until the cgroup will be explicitly
>>>> unfrozen.
>>>>
>>>> AFAICS this is not accurate, they can run but can't return to user-mode.
>>>> So I guess task_work_run() is fine.
>>>
>>> IIUC it's a user facing doc, so maybe it's accurate enough from that
>>> perspective. But I do agree that the semantics around task_work is
>>> not exactly clear.
>>
>> A good correctness test for cgroup freezer is whether it'd be safe to
>> snapshot and restore the tasks in the cgroup while frozen.
>
> Well, I don't really understand what can snapshot/restore actually mean...
CRIU, I assume. I'll try it ...
> I forgot everything about cgroup freezer and I am already sleeping, but even
> if we forget about task_work_add/TIF_NOTIFY_SIGNAL/etc, afaics ptrace can
> change the state of cgroup_task_frozen() task between snapshot and restore ?
... but I'm inclined to think the patch makes sense regardless,
we're replacing an infinite loop with wait-wake-execute-wait.
--
Pavel Begunkov
next prev parent reply other threads:[~2024-07-09 19:26 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-07-07 16:32 [PATCH 0/2] fix task_work interation with freezing Pavel Begunkov
2024-07-07 16:32 ` [PATCH 1/2] io_uring/io-wq: limit retrying worker initialisation Pavel Begunkov
2024-07-07 16:32 ` [PATCH 2/2] kernel: rerun task_work while freezing in get_signal() Pavel Begunkov
2024-07-08 10:42 ` Oleg Nesterov
2024-07-08 15:40 ` Pavel Begunkov
2024-07-08 18:48 ` Peter Zijlstra
2024-07-09 10:36 ` Oleg Nesterov
2024-07-09 14:05 ` Pavel Begunkov
2024-07-09 16:39 ` Tejun Heo
2024-07-09 19:07 ` Oleg Nesterov
2024-07-09 19:26 ` Pavel Begunkov [this message]
2024-07-09 19:38 ` Oleg Nesterov
2024-07-09 19:55 ` Pavel Begunkov
2024-07-10 0:54 ` Tejun Heo
2024-07-10 17:53 ` Pavel Begunkov
2024-07-10 19:10 ` Oleg Nesterov
2024-07-10 19:20 ` Tejun Heo
2024-07-10 21:34 ` Oleg Nesterov
2024-07-10 22:01 ` Tejun Heo
2024-07-10 22:17 ` Oleg Nesterov
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox