Messages in this thread | | | Date | Sat, 26 Jan 2008 08:03:23 -0500 (EST) | From | Steven Rostedt <> | Subject | Re: [PATCH 01/23 -v6] printk - dont wakeup klogd with interrupts disabled |
| |
On Sat, 26 Jan 2008, Peter Zijlstra wrote: > > Right, while this might appear to be sound, I hate to point out that we > only want to exclude it when current holds the rq->lock. This could lead > to spuriously missing klogd wakeups because other cpus hold our > rq->lock.
Hmm, good point. Not hard to solve ;-)
> > Now, is this klogd thing only relevant for dumping stuff into the log, > or also for writing stuff out to the serial line?
It's only responsible for the log (userland stuff).
-- Steve
| |