lkml.org 
[lkml]   [2009]   [Jul]   [8]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 2/3] ring-buffer: make lockless

On Wed, 8 Jul 2009, Mathieu Desnoyers wrote:

> * Steven Rostedt (rostedt@goodmis.org) wrote:
> > From: Steven Rostedt <srostedt@redhat.com>
> >
> > This patch converts the ring buffers into a completely lockless
> > buffer recording system. The read side still takes locks since
> > we still serialize readers. But the writers are the ones that
> > must be lockless (those can happen in NMIs).
>
> Hi Steven,
>
> Have you started modeling this scheme with Promela as I previously
> discussed with you ? Paul McKenney did a lot of nice papers on the
> topic, and given the complexity level of this lockless buffering
> algorithm, I'd be much more comfortable if it was formally verified. The
> idea here is to make sure we don't start an endless bug-chasing hunt as
> trace buffers gets more exposure to various workloads.

I would when I get time. Too many things on my plate at the moment. But
I do constantly run my ring-buffer-benchmark module, which highly stresses
the buffer. I have not had to fix a bug in it for a long time, and nothing
that needed a design change.

I'll try it out (or better yet, maybe someone else will) when I get more
time.

-- Steve



\
 
 \ /
  Last update: 2009-07-08 23:19    [W:0.064 / U:4.480 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site