[PATCH V2] relay: fix timer madness

[PATCH V2] relay: fix timer madness

[zhangwei(Jovi)]

When I'm using ktap script to tracing all event tracepoints by relay
transport, without this patch, the system will hang in few seconds.

I found the original patch discussion in 2007.
http://marc.info/?l=linux-kernel&m=118544794717162&w=2
(In that mail thread, the patch didn't fix that problem, but it fix
the problem I encountered now)

Changed from v1:
mod timer interval changed from jiffies+1 to HZ/10, as Ingo suggested.

Original patch changelog from Ingo in 2007:

Remove timer calls (!!!) from deep within the tracing infrastructure.
This was totally bogus code that can cause lockups and worse.
Poll the buffer every 2 jiffies for now.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: "zhangwei(Jovi)" <jovi.zhangwei@huawei.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
 kernel/relay.c |   14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/kernel/relay.c b/kernel/relay.c
index b91488b..87af4ce 100644
--- a/kernel/relay.c
+++ b/kernel/relay.c
@@ -339,6 +339,10 @@ static void wakeup_readers(unsigned long data)
 {
 	struct rchan_buf *buf = (struct rchan_buf *)data;
 	wake_up_interruptible(&buf->read_wait);
+	/*
+	 * Stupid polling for now:
+	 */
+	mod_timer(&buf->timer, HZ / 10);
 }

 /**
@@ -356,6 +360,7 @@ static void __relay_reset(struct rchan_buf *buf, unsigned int init)
 		init_waitqueue_head(&buf->read_wait);
 		kref_init(&buf->kref);
 		setup_timer(&buf->timer, wakeup_readers, (unsigned long)buf);
+		mod_timer(&buf->timer, HZ / 10);
 	} else
 		del_timer_sync(&buf->timer);

@@ -739,15 +744,6 @@ size_t relay_switch_subbuf(struct rchan_buf *buf, size_t length)
 		else
 			buf->early_bytes += buf->chan->subbuf_size -
 					    buf->padding[old_subbuf];
-		smp_mb();
-		if (waitqueue_active(&buf->read_wait))
-			/*
-			 * Calling wake_up_interruptible() from here
-			 * will deadlock if we happen to be logging
-			 * from the scheduler (trying to re-grab
-			 * rq->lock), so defer it.
-			 */
-			mod_timer(&buf->timer, jiffies + 1);
 	}

 	old = buf->data;
-- 
1.7.9.7

Re: [PATCH V2] relay: fix timer madness

[Li Zefan]

On 2013/7/10 10:18, zhangwei(Jovi) wrote:
> When I'm using ktap script to tracing all event tracepoints by relay
> transport, without this patch, the system will hang in few seconds.
> 
> I found the original patch discussion in 2007.
> http://marc.info/?l=linux-kernel&m=118544794717162&w=2
> (In that mail thread, the patch didn't fix that problem, but it fix
> the problem I encountered now)
> 
> Changed from v1:
> mod timer interval changed from jiffies+1 to HZ/10, as Ingo suggested.
> 
> Original patch changelog from Ingo in 2007:
> 
> Remove timer calls (!!!) from deep within the tracing infrastructure.
> This was totally bogus code that can cause lockups and worse.
> Poll the buffer every 2 jiffies for now.
> 
> Signed-off-by: Ingo Molnar <mingo@elte.hu>
> Signed-off-by: "zhangwei(Jovi)" <jovi.zhangwei@huawei.com>
> Cc: Steven Rostedt <rostedt@goodmis.org>
> Cc: Jens Axboe <axboe@kernel.dk>
> Cc: Al Viro <viro@zeniv.linux.org.uk>
> Cc: Eric Dumazet <edumazet@google.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

I don't think this patch should have Andrew's signed-off-by?

Re: [PATCH V2] relay: fix timer madness

[Andrew Morton]

On Wed, 10 Jul 2013 11:37:26 +0800 Li Zefan <lizefan@huawei.com> wrote:

> On 2013/7/10 10:18, zhangwei(Jovi) wrote:
> > When I'm using ktap script to tracing all event tracepoints by relay
> > transport, without this patch, the system will hang in few seconds.
> > 
> > I found the original patch discussion in 2007.
> > http://marc.info/?l=linux-kernel&m=118544794717162&w=2
> > (In that mail thread, the patch didn't fix that problem, but it fix
> > the problem I encountered now)
> > 
> > Changed from v1:
> > mod timer interval changed from jiffies+1 to HZ/10, as Ingo suggested.
> > 
> > Original patch changelog from Ingo in 2007:
> > 
> > Remove timer calls (!!!) from deep within the tracing infrastructure.
> > This was totally bogus code that can cause lockups and worse.
> > Poll the buffer every 2 jiffies for now.
> > 
> > Signed-off-by: Ingo Molnar <mingo@elte.hu>
> > Signed-off-by: "zhangwei(Jovi)" <jovi.zhangwei@huawei.com>
> > Cc: Steven Rostedt <rostedt@goodmis.org>
> > Cc: Jens Axboe <axboe@kernel.dk>
> > Cc: Al Viro <viro@zeniv.linux.org.uk>
> > Cc: Eric Dumazet <edumazet@google.com>
> > Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> 
> I don't think this patch should have Andrew's signed-off-by?

I guess not, unless it was taken from -mm, which would be odd, as I
have the old version.

v1 has been in my tree for a few months - Ingo requested some updates
but nothing happened and I have not checked whether v2 addresses his
requests.

Re: [PATCH V2] relay: fix timer madness

[Dan Carpenter]

On Wed, Jul 10, 2013 at 10:18:54AM +0800, zhangwei(Jovi) wrote:
> @@ -339,6 +339,10 @@ static void wakeup_readers(unsigned long data)
>  {
>  	struct rchan_buf *buf = (struct rchan_buf *)data;
>  	wake_up_interruptible(&buf->read_wait);
> +	/*
> +	 * Stupid polling for now:
> +	 */
> +	mod_timer(&buf->timer, HZ / 10);

mod_timer() takes an offset so probably "jiffies + HZ / 10" was
intended here and also below.  Certainly passing "HZ / 10" doesn't
make any kind of sense.

regards,
dan carpenter