Re: [PATCH] printk: Remove separate printk_sched buffers and use printk buf instead

[Jan Kara <jack@suse.cz>]

On Tue 06-05-14 20:05:19, Steven Rostedt wrote:
> On Tue, 6 May 2014 16:37:14 -0700
> Andrew Morton <akpm@linux-foundation.org> wrote:
> 
> 
> > What I have in -next is very different from this version of the patch. 
> > What's happening?
> 
> Hmm, it looks massaged from what I originally sent in order to be
> placed after the other patches in the series. A quick review of the
> patch seems to be mostly the same logic, but it is a bit different.
> 
> Jan, did you update my patch?
  Yes, I was carrying it for quite long and with each rebase and rewrite of
the patch set it got massaged a bit. So in the end it's probably somewhat
different from what you originally submitted.

> > From: Steven Rostedt <rostedt@goodmis.org>
> > Subject: printk: remove separate printk_sched buffers and use printk buf instead
> > 
> > To prevent deadlocks with doing a printk inside the scheduler,
> > printk_sched() was created.  The issue is that printk has a console_sem
> > that it can grab and release.  The release does a wake up if there's a
> > task pending on the sem, and this wake up grabs the rq locks that is held
> > in the scheduler.  This leads to a possible deadlock if the wake up uses
> > the same rq as the one with the rq lock held already.
> > 
> > What printk_sched() does is to save the printk write in a per cpu buffer
> > and sets the PRINTK_PENDING_SCHED flag.  On a timer tick, if this flag is
> > set, the printk() is done against the buffer.
> > 
> > There's a couple of issues with this approach.
> > 
> > 1) If two printk_sched()s are called before the tick, the second one
> >    will overwrite the first one.
> > 
> > 2) The temporary buffer is 512 bytes and is per cpu.  This is a quite a
> >    bit of space wasted for something that is seldom used.
> > 
> > In order to remove this, the printk_sched() can use the printk buffer
> > instead, and delay the console_trylock()/console_unlock() to the queued
> > work.
> > 
> > Because printk_sched() would then be taking the logbuf_lock, the
> > logbuf_lock must not be held while doing anything that may call into the
> > scheduler functions, which includes wake ups.  Unfortunately, printk()
> > also has a console_sem that it uses, and on release, the up(&console_sem)
> > may do a wake up of any pending waiters.  This must be avoided while
> > holding the logbuf_lock.
> > 
> > Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
> 
> Jan, if you did massage my patch, you should add here what you did.
> Usually in [brackets]. Otherwise it puts the blame on me if something
> breaks, or adds confusion if I happen to send out another patch like I
> just did.
  Yeah, sorry for that. I should have been updating the changelog with what
I did...

								Honza
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR