From: Tejun Heo <tj@kernel.org>
To: Jan Kratochvil <jan.kratochvil@redhat.com>
Cc: Denys Vlasenko <vda.linux@googlemail.com>,
Oleg Nesterov <oleg@redhat.com>,
linux-kernel@vger.kernel.org, torvalds@linux-foundation.org,
akpm@linux-foundation.org, indan@nul.nu
Subject: Re: PTRACE_DETACH without stop [Re: [PATCH 04/11] ptrace: implement PTRACE_INTERRUPT]
Date: Sun, 15 May 2011 18:35:23 +0200 [thread overview]
Message-ID: <20110515163523.GH23665@htj.dyndns.org> (raw)
In-Reply-To: <20110515161000.GE31855@host1.jankratochvil.net>
Hey, again.
On Sun, May 15, 2011 at 06:10:00PM +0200, Jan Kratochvil wrote:
> I do not think this change is much related to this patchset.
>
> But having to PTRACE_INTERRUPT the tracee before PTRACE_DETACH has no
> advantage, it is just a performance (see transparent tracking of 10000+ thread
> https://lkml.org/lkml/2011/5/15/115
> ) problem and also getting it correct. As when one wait()s and gets
> WIFSTOPPED one needs to respawn to signal otherwise the signal gets lost on
> PTRACE_ATTACH. How to respawn it? By PTRACE_INTERRUPT with DATA==signal?
> Or PTRACE_CONT with DATA==signal? With rapid signalling of the tracee the
> debugger may never have a chance to correctly quit. Handling other cases
> transparently for the original parent also may not be fully clear.
>
> It would be nice to write documentation already while discussing this patch,
> I do not know if PTRACE_INTERRUPT respects DATA etc., it may show ptrace is
> still tricky.
Hmmm... you seem to be a bit confused about different ptrace traps
(again, a lot of this probably stems from incorrect manpage). They
don't interact with each other that way and there is no correctness
issue (if there is, it's a but and should be fixed) of
INTERRUPT+DETACH compared to DETACH without interrupt. The only
difference is performance, so let's concentrate on that.
The reason why I'm reluctant to drop trapped requirement from both
SEIZE and DETACH is that removing those synchronization points opens
up a lot of corner cases. I'm they can all be dealt with but it's
gonna be measurably more complex and possibly fragile. Given that the
current implementation mandates traps on both ATTACH and DETACH, I'd
like to leave that part alone unless there are pretty strong
arguments, and, sure, performance might as well be strong enough.
But, I'd really like to know how much it actually costs to attach and
detach a lot of tasks before deciding anything.
Thank you.
--
tejun
next prev parent reply other threads:[~2011-05-15 16:35 UTC|newest]
Thread overview: 115+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-05-08 15:48 [PATCHSET ptrace] ptrace: implement PTRACE_SEIZE/INTERRUPT and group stop notification Tejun Heo
2011-05-08 15:48 ` [PATCH 01/11] job control: rename signal->group_stop and flags to jobctl and rearrange flags Tejun Heo
2011-05-08 15:48 ` [PATCH 02/11] ptrace: implement PTRACE_SEIZE Tejun Heo
2011-05-09 16:18 ` Oleg Nesterov
2011-05-10 9:46 ` Tejun Heo
2011-05-10 13:20 ` Oleg Nesterov
2011-05-10 13:47 ` Tejun Heo
2011-05-10 18:19 ` Oleg Nesterov
2011-05-15 15:56 ` PTRACE_SEIZE should not stop [Re: [PATCH 02/11] ptrace: implement PTRACE_SEIZE] Jan Kratochvil
2011-05-15 16:26 ` Tejun Heo
2011-05-15 17:15 ` Jan Kratochvil
2011-05-15 17:25 ` Tejun Heo
2011-05-15 19:48 ` Jan Kratochvil
2011-05-16 8:31 ` Tejun Heo
2011-05-16 12:26 ` Jan Kratochvil
2011-05-16 12:42 ` Tejun Heo
2011-05-16 13:03 ` Jan Kratochvil
2011-05-16 13:51 ` Tejun Heo
2011-05-16 13:21 ` Jan Kratochvil
2011-05-16 13:45 ` Tejun Heo
2011-05-16 13:48 ` Jan Kratochvil
2011-05-16 13:54 ` Tejun Heo
2011-05-08 15:48 ` [PATCH 03/11] ptrace: ptrace_check_attach(): rename @kill to @ignore_state and add comments Tejun Heo
2011-05-08 15:48 ` [PATCH 04/11] ptrace: implement PTRACE_INTERRUPT Tejun Heo
2011-05-08 21:58 ` Denys Vlasenko
2011-05-09 10:09 ` Tejun Heo
2011-05-09 10:55 ` Denys Vlasenko
2011-05-09 16:58 ` Oleg Nesterov
2011-05-10 9:50 ` Tejun Heo
2011-05-10 14:06 ` Oleg Nesterov
2011-05-10 14:20 ` Tejun Heo
2011-05-10 18:08 ` Oleg Nesterov
2011-05-11 8:29 ` Tejun Heo
2011-05-12 17:06 ` Oleg Nesterov
2011-05-12 17:21 ` Tejun Heo
2011-05-10 21:59 ` Denys Vlasenko
2011-05-11 9:19 ` Tejun Heo
2011-05-11 12:23 ` Denys Vlasenko
2011-05-11 13:22 ` Tejun Heo
2011-05-11 16:20 ` Bryan Donlan
2011-05-11 19:24 ` Tejun Heo
2011-05-15 16:10 ` PTRACE_DETACH without stop [Re: [PATCH 04/11] ptrace: implement PTRACE_INTERRUPT] Jan Kratochvil
2011-05-15 16:35 ` Tejun Heo [this message]
2011-05-15 17:39 ` Jan Kratochvil
2011-05-16 9:01 ` Tejun Heo
2011-05-16 12:08 ` Jan Kratochvil
2011-05-16 12:24 ` Tejun Heo
2011-05-08 15:48 ` [PATCH 05/11] ptrace: restructure ptrace_getsiginfo() Tejun Heo
2011-05-08 15:49 ` [PATCH 06/11] ptrace: make group stop state visible via PTRACE_GETSIGINFO Tejun Heo
2011-05-10 16:55 ` Oleg Nesterov
2011-05-10 17:11 ` Oleg Nesterov
2011-05-11 8:08 ` Tejun Heo
2011-05-12 16:47 ` Oleg Nesterov
2011-05-12 17:15 ` Tejun Heo
2011-05-08 15:49 ` [PATCH 07/11] ptrace: add JOBCTL_TRAPPED Tejun Heo
2011-05-08 15:49 ` [PATCH 08/11] ptrace: move fallback JOBCTL_TRAPPING clearing to get_signal_to_deliver() Tejun Heo
2011-05-11 15:48 ` Oleg Nesterov
2011-05-11 19:17 ` Tejun Heo
2011-05-12 15:40 ` Oleg Nesterov
2011-05-08 15:49 ` [PATCH 09/11] job control: reorganize wait_task_stopped() Tejun Heo
2011-05-11 15:48 ` Oleg Nesterov
2011-05-11 19:29 ` Tejun Heo
2011-05-12 15:42 ` Oleg Nesterov
2011-05-12 16:02 ` Tejun Heo
2011-05-12 17:25 ` Oleg Nesterov
2011-05-12 17:32 ` Tejun Heo
2011-05-12 17:33 ` Tejun Heo
2011-05-12 18:33 ` Oleg Nesterov
2011-05-13 8:46 ` Tejun Heo
2011-05-13 17:21 ` Oleg Nesterov
2011-05-14 10:56 ` Tejun Heo
2011-05-15 14:40 ` waitpid(WNOHANG) should report SIGCHLD-notified signals [Re: [PATCH 09/11] job control: reorganize wait_task_stopped()] Jan Kratochvil
2011-05-15 16:47 ` Tejun Heo
2011-05-15 17:01 ` Tejun Heo
2011-05-15 17:47 ` Jan Kratochvil
2011-05-16 9:13 ` Tejun Heo
2011-05-16 12:11 ` Jan Kratochvil
2011-05-16 12:27 ` Tejun Heo
2011-05-16 12:39 ` Jan Kratochvil
2011-05-16 12:46 ` Tejun Heo
2011-05-08 15:49 ` [PATCH 10/11] ptrace: move JOBCTL_TRAPPING wait to wait(2) and ptrace_check_attach() Tejun Heo
2011-05-11 16:49 ` Oleg Nesterov
2011-05-11 17:00 ` Oleg Nesterov
2011-05-11 19:45 ` Tejun Heo
2011-05-11 19:53 ` Tejun Heo
2011-05-12 10:23 ` Tejun Heo
2011-05-12 16:06 ` Oleg Nesterov
2011-05-12 15:59 ` Oleg Nesterov
2011-05-12 16:07 ` Tejun Heo
2011-05-12 18:20 ` Oleg Nesterov
2011-05-13 9:13 ` Tejun Heo
2011-05-13 18:34 ` Oleg Nesterov
2011-05-08 15:49 ` [PATCH 11/11] ptrace: implement group stop notification for ptracer Tejun Heo
2011-05-08 22:42 ` Denys Vlasenko
2011-05-09 10:10 ` Tejun Heo
2011-05-10 22:37 ` Denys Vlasenko
2011-05-11 9:05 ` Tejun Heo
2011-05-11 12:01 ` Denys Vlasenko
2011-05-11 13:13 ` Tejun Heo
2011-05-11 19:58 ` Oleg Nesterov
2011-05-11 20:18 ` Tejun Heo
2011-05-11 20:21 ` Tejun Heo
2011-05-12 10:24 ` Tejun Heo
2011-05-15 14:02 ` getter PTRACE_GETSIGINFO should not modify anything [Re: [PATCH 11/11] ptrace: implement group stop notification for ptracer] Jan Kratochvil
2011-05-15 14:28 ` Tejun Heo
2011-05-15 17:17 ` Jan Kratochvil
2011-05-15 17:28 ` Tejun Heo
2011-05-15 20:06 ` Jan Kratochvil
2011-05-16 8:43 ` Tejun Heo
2011-05-16 12:17 ` Jan Kratochvil
2011-05-16 12:56 ` Tejun Heo
2011-05-16 13:00 ` Ingo Molnar
2011-05-08 22:27 ` [PATCHSET ptrace] ptrace: implement PTRACE_SEIZE/INTERRUPT and group stop notification Denys Vlasenko
2011-05-09 9:48 ` Tejun Heo
2011-05-15 13:55 ` ptrace-testsuite status [Re: [PATCHSET ptrace] ptrace: implement PTRACE_SEIZE/INTERRUPT and group stop notification] Jan Kratochvil
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 \
--in-reply-to=20110515163523.GH23665@htj.dyndns.org \
--to=tj@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=indan@nul.nu \
--cc=jan.kratochvil@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=oleg@redhat.com \
--cc=torvalds@linux-foundation.org \
--cc=vda.linux@googlemail.com \
/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;
as well as URLs for NNTP newsgroup(s).