Re: [PATCH v4] oom: add trace points for debugging.

[KAMEZAWA Hiroyuki]

On Tue, 13 Dec 2011 14:58:51 -0800
Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:

> On Tue, 13 Dec 2011 18:12:25 +0900
> KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> wrote:
> 
> > Changelog:
> >  - devided into oom tracepoint and task tracepoint.
> >  - task tracepoint traces fork/rename
> >  - oom tracepoint traces modification to oom_score_adj.
> > 
> > dropped acks because of total design changes.
> > 
> > From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
> > Subject: [PATCH] tracepoint: add tracepoints for debugging oom_score_adj.
> > 
> > oom_score_adj is used for guarding processes from OOM-Killer. One of problem
> > is that it's inherited at fork(). When a daemon set oom_score_adj and
> > make children, it's hard to know where the value is set.
> 
> This sounds like a really thin justification for patching the kernel. 
> "Help! I don't know what my code is doing!".
> 
> Alternatives would include grepping your source code for
> "oom_score_adj", or running "strace -f"!
> 
> I suspect you did have a good reason for making this change, but it
> wasn't explained very well?
> 

Ah, one example is this thread.
- http://marc.info/?t=132273239800008&r=1&w=4
oom_score_adj was inherited by sshd by bug?

And IIUC, I saw other issues with other task launchers as Gnome.

If the problem is caused by some daemon, strace -f is not very good.
And the user will not have the source codes. So, I thought it's better
to have tracepoints to see what is wrong in application chains.


Thanks,
-Kame


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/