Re: [PATCH] printk: Print cpu number along with time

From: "Jörn Engel" <joern@logfs.org>
To: Jiri Kosina <jkosina@suse.cz>
Cc: David Rientjes <rientjes@google.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Rik van Riel <riel@redhat.com>,
	linux-kernel@vger.kernel.org, peterz@infradead.org,
	akpm@linux-foundation.org, cxie@redhat.com,
	Jiri Slaby <jslaby@suse.cz>
Subject: Re: [PATCH] printk: Print cpu number along with time
Date: Thu, 24 Apr 2014 19:29:16 -0400	[thread overview]
Message-ID: <20140424232916.GD25446@logfs.org> (raw)
In-Reply-To: <alpine.LNX.2.00.1404250020370.8903@pobox.suse.cz>

On Fri, 25 April 2014 00:21:20 +0200, Jiri Kosina wrote:
> On Thu, 24 Apr 2014, David Rientjes wrote:
> 
> > > Well, if you have dmesg dump from panic that happens every other year, and 
> > > you have to do post-mortem analysis on it, I am pretty sure you would love 
> > > to be able to figure out how the stack traces would look like without 
> > > inter-CPU interleaving. And I am pretty sure you wouldn't want to 
> > > insert/enable a tracepoint and wait another two years for the bug to 
> > > trigger again.
> > 
> > Sounds like the appropriate fix would be to serialize stack dumping to the 
> > kernel log.
> 
> That's rather difficult, as it can be done both from NMI and non-NMI 
> contexts, and those are very hard to synchronize.

It is also easy to get wrong and cause a panic() that doesn't reboot
but instead gets stuck somewhere.

Plus, you might want similar serialization for other multi-line debug
messages.  Scsi error messages come to mind, but we have many more
examples.

Jörn

--
Any headline which ends in a question mark can be answered by the word 'no'.
-- Betteridge's Law