linux-kernel - Re: [PATCH] bugfix oom kill init lead panic

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

Message-ID: <alpine.DEB.2.10.1511301407080.10460@chino.kir.corp.google.com>
Date:	Mon, 30 Nov 2015 14:08:23 -0800 (PST)
From:	David Rientjes <rientjes@...gle.com>
To:	"Chenjie (K)" <chenjie6@...wei.com>
cc:	Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
	linux-kernel@...r.kernel.org, David.Woodhouse@...el.com,
	zhihui.gao@...wei.com, lizefan@...wei.com, stable@...r.kernel.org
Subject: Re: [PATCH] bugfix oom kill init lead panic

On Mon, 30 Nov 2015, Chenjie (K) wrote:

> My kernel version is 3.10 ,but the 4.3 is the same
> and the newest code is
> 
> 	for_each_process(p) {
> 		if (!process_shares_mm(p, mm))
> 			continue;
> 		if (same_thread_group(p, victim))
> 			continue;
> 		if (unlikely(p->flags & PF_KTHREAD))
> 			continue;
> 		if (p->signal->oom_score_adj == OOM_SCORE_ADJ_MIN)
> 			continue;
> 
> so this not add the ！is_global_init also.
> 
> when we vfork (CLONE_VM) a process,the copy_mm
> 	if (clone_flags & CLONE_VM) {
> 		atomic_inc(&oldmm->mm_users);
> 		mm = oldmm;
> 		goto good_mm;
> 	}
> use the parent mm.
> 

I think it might be a legitimate fix, but if the oom killer is killing pid 
9134 in your log then I assume the next call to the oom killer will panic 
the system anyway unless there is actually a process using less memory 
that can be killed.  Would you mind enabling vm.oom_dump_tasks (it should 
default to enabled) and post the entire oom killer log?