Re: [PATCH tip/core/urgent] fix several lockdep splats, allow multiple splats

[Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>]

* Ingo Molnar (mingo@elte.hu) wrote:
> 
> * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> 
> > Hello!
> > 
> > This patchset contains four RCU lockdep splat fixes, courtesy of David 
> > Howells, Peter Zijlstra, and Trond Myklebust, [...]
> 
> I've applied #1 and #2 - but shouldnt #4 and #5 go via the NFS tree?
> 
> > [...] as well as an enhancement by Lai Jiangshan that permits collecting 
> > more than one RCU lockdep splat per boot.
> 
> Hm, this #3 patch i disagree with quite fundamentally: one of the big virtues 
> of lockdep is that it complains only once and then shuts up and lets the 
> system work. It allows distro debug kernels to have lockdep enabled, etc.
> 
> One bugreport per bootup per user is the most we can expect really. Not 
> disabling it risks getting a stream of repeat messages, annoyed testers and 
> gives us _less_ bugreports in the end.
> 
> Also, often the _first_ warning is the most reliable one - sometimes there's 
> interactions, and the first bug causing a second warning as well, etc. So 
> reporting just the highest-quality (i.e. first) issue we detect is the best 
> approach.

I recommend creating a kernel command line parameter that would tweak
the number of messages printed by lockdep. The default would indeed by 1
message, but people in a debugging marathon can specify a larger value
so they won't have to reboot between each individual lockdep error.

Thanks,

Mathieu

> 
> Thanks,
> 
> 	Ingo

-- 
Mathieu Desnoyers
Operating System Efficiency R&D Consultant
EfficiOS Inc.
http://www.efficios.com