RE: [PATCH] printk: fixing the deadlock when calling printk in nmihandle

[Liu, Chuansheng]

> -----Original Message-----
> From: Peter Zijlstra [mailto:peterz@xxxxxxxxxxxxx]
> Sent: Wednesday, July 04, 2012 10:52 PM
> To: Liu, Chuansheng
> Cc: 'linux-kernel@xxxxxxxxxxxxxxx' (linux-kernel@xxxxxxxxxxxxxxx); kay@xxxxxxxx;
> gregkh@xxxxxxxxxxxxxxxxxxx; mingo@xxxxxxx
> Subject: RE: [PATCH] printk: fixing the deadlock when calling printk in nmi
> handle
> 
> On Wed, 2012-07-04 at 16:50 +0200, Peter Zijlstra wrote:
> > On Wed, 2012-07-04 at 14:46 +0000, Liu, Chuansheng wrote:
> > > > > I took many time to dig out this issue.
> > > >
> > > > Well, it was a known issue..
> > > Still feel unfair :)
> > > If this is a known issue that calling printk is not safe in nmi
> > > handler, Why not defense it like this patch before real solution coming out?
> >
> > Why replace one broken thing with another broken thing?
> 
> Also, I much prefer obviously broken over subtly broken.
Supporting your opinion totally, but from production view, it is not a good thing that system crash,
I prefer to lose one log instead:)
N‹§²æìr¸›yúèšØb²X¬¶ÇvØ^–)Þ{.nÇ+‰·¥Š{±‘êçzX§¶›¡Ü}©ž²ÆzÚ&j:+v‰¨¾«‘êçzZ+€Ê+zf£¢·hšˆ§~†­†Ûiÿûàz¹®w¥¢¸?™¨è­Ú&¢)ßf”ù^jÇy§m…á@A«a¶Úÿ0¶ìh®å’i