linux-kernel - RE: [PATCH] Avoid dead lock of console related locks in panic case

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: <A5ED84D3BB3A384992CBB9C77DEDA4D414A0351F@USINDEM103.corp.hds.com>
Date:	Sat, 1 Dec 2012 01:04:22 +0000
From:	Seiji Aguchi <seiji.aguchi@....com>
To:	Andrew Morton <akpm@...ux-foundation.org>
CC:	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"joe@...ches.com" <joe@...ches.com>,
	"gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
	"kay@...y.org" <kay@...y.org>,
	"jim.cromie@...il.com" <jim.cromie@...il.com>,
	"mingo@...e.hu" <mingo@...e.hu>,
	"sboyd@...eaurora.org" <sboyd@...eaurora.org>,
	"jason.wessel@...driver.com" <jason.wessel@...driver.com>,
	"a.p.zijlstra@...llo.nl" <a.p.zijlstra@...llo.nl>,
	"rostedt@...dmis.org" <rostedt@...dmis.org>,
	"dle-develop@...ts.sourceforge.net" 
	<dle-develop@...ts.sourceforge.net>,
	Satoru Moriya <satoru.moriya@....com>
Subject: RE: [PATCH] Avoid dead lock of console related locks in panic case

> What I actually meant was: can "this" CPU avoid stopping other CPUs so early?  If we stop the other CPUs when this CPU is ready to
> stop itself then there will never be such deadlocks.

Let me explain my opinion.
When we focus on the deadlock only,   the code will be simple by moving smp_send_stop() at the end of panic().

But, panic situation is not normal.
I don't think that keeping running multiple cpus is safe, because they may touch corrupted data/variables and unnecessary 
panic/BUG() may happen.

IMO, cpus should be stopped "as early as" possible when panic happens.
And then panic() has to take minimal steps with a single cpu.
 - output messages
 - kicking troubleshooting features like kdump/kmsg_dump

Seiji


--
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/