Re: [PATCH 1/3] sched: add sched_task_call()
[Posted February 24, 2015 by corbet]
| From: |
| Peter Zijlstra <peterz-AT-infradead.org> |
| To: |
| Josh Poimboeuf <jpoimboe-AT-redhat.com> |
| Subject: |
| Re: [PATCH 1/3] sched: add sched_task_call() |
| Date: |
| Tue, 17 Feb 2015 10:24:50 +0100 |
| Message-ID: |
| <20150217092450.GI5029@twins.programming.kicks-ass.net> |
| Cc: |
| Andrew Morton <akpm-AT-linux-foundation.org>, Ingo Molnar <mingo-AT-redhat.com>, Jiri Kosina <jkosina-AT-suse.cz>, Seth Jennings <sjenning-AT-redhat.com>, linux-kernel-AT-vger.kernel.org |
| Archive‑link: | |
Article |
On Mon, Feb 16, 2015 at 04:05:05PM -0600, Josh Poimboeuf wrote:
> Yeah, I can understand that. I definitely want to avoid touching the
> scheduler code. Basically I'm trying to find a way to atomically do the
> following:
>
> if (task is sleeping) {
> walk the stack
> if (certain set of functions isn't on the stack)
> set (or clear) a thread flag for the task
> }
>
> Any ideas on how I can achieve that? So far my ideas are:
So far stack unwinding has basically been a best effort debug output
kind of thing, you're wanting to make the integrity of the kernel depend
on it.
You require an absolute 100% correctness of the stack unwinder -- where
today it is; as stated above; a best effort debug output thing.
That is a _big_ change.
Has this been properly considered; has all the asm of the relevant
architectures been audited? Are you planning on maintaining that level
of audit for all new patches?
Because the way you propose to do things, we'll end up with silent but
deadly fail if the unwinder is less than 100% correct. No way to easily
debug that, no warns, just silent corruption.
Are you really really sure you want to go do this?
