linux-kernel - RE: [patch v3 3/3] mm, oom: do not panic for oom kills triggered from sysrq

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <alpine.DEB.2.10.1507141414410.16182@chino.kir.corp.google.com>
Date:	Tue, 14 Jul 2015 14:16:18 -0700 (PDT)
From:	David Rientjes <rientjes@...gle.com>
To:	Hillf Danton <hillf.zj@...baba-inc.com>
cc:	Andrew Morton <akpm@...ux-foundation.org>,
	Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
	Michal Hocko <mhocko@...e.cz>,
	linux-kernel <linux-kernel@...r.kernel.org>, linux-mm@...ck.org
Subject: RE: [patch v3 3/3] mm, oom: do not panic for oom kills triggered
 from sysrq

On Fri, 10 Jul 2015, Hillf Danton wrote:

> > I'm not sure I understand your point.
> > 
> > There are two oom killer panics: when panic_on_oom is enabled and when the
> > oom killer can't find an eligible process.
> > 
> > The change to the panic_on_oom panic is dealt with in check_panic_on_oom()
> > and the no eligible process panic is dealt with here.
> > 
> > If the sysctl is disabled, and there are no eligible processes to kill,
> > the change in behavior here is that we don't panic when triggered from
> > sysrq.  That's the change in the hunk above.
> > 
> When no eligible processes is selected to kill, we are sure that we skip one
> panic in check_panic_on_oom(), and we have no clear reason to panic again.
> 
> But we can simply answer the caller that there is no page, and let her
> decide what to do.
> 
> So I prefer to fold the two panic into one.
> 
> Hillf
> > > > -	if (p != (void *)-1UL) {
> > > > +	if (p && p != (void *)-1UL) {
> > > >  		oom_kill_process(oc, p, points, totalpages, NULL,
> > > >  				 "Out of memory");
> > > >  		killed = 1;
> 

I'm still not sure I understand your point, unfortunately.  The new check:

	if (!p && oc->order != -1) {
		dump_header(oc, NULL, NULL);
		panic("Out of memory and no killable processes...\n");
	}

ensures we never panic when called from sysrq.  This is done because 
userspace can easily race when there is a single eligible process to kill 
that exits or is otherwise killed and the sysrq+f ends up panicking the 
machine unexpectedly.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/