Hi,
Ingo, Steven, I get this patch from 3.4 preempt-rt patch set, It seems that this patch
fix relayfs bug not only for rt kernel, but also for mainline.
When I'm using below ktap script to tracing all event tracepoints, without this patch,
the system will hang in few seconds, the patch indeed fix the problem as the changelog pointed.
function eventfun (e) {
printf("%d %d\t%s\t%s", cpu(), pid(), execname(), e.annotate)
}
kdebug.probe("tp:", eventfun)
kdebug.probe_end(function () {
printf("probe end\n")
})
This patch is old, I can 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)
I hope you can remember this :)
so why we didn't commit this patch into mainline? any concern?
Thanks.
------------------------------------->
Subject: relay: fix timer madness
From: Ingo Molnar <[email protected]>
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 <[email protected]>
---
kernel/relay.c | 14 +++++---------
1 file changed, 5 insertions(+), 9 deletions(-)
Index: linux-rt-rebase.q/kernel/relay.c
===================================================================
--- linux-rt-rebase.q.orig/kernel/relay.c
+++ linux-rt-rebase.q/kernel/relay.c
@@ -319,6 +319,10 @@ static void wakeup_readers(unsigned long
{
struct rchan_buf *buf = (struct rchan_buf *)data;
wake_up_interruptible(&buf->read_wait);
+ /*
+ * Stupid polling for now:
+ */
+ mod_timer(&buf->timer, jiffies + 1);
}
/**
@@ -336,6 +340,7 @@ static void __relay_reset(struct rchan_b
init_waitqueue_head(&buf->read_wait);
kref_init(&buf->kref);
setup_timer(&buf->timer, wakeup_readers, (unsigned long)buf);
+ mod_timer(&buf->timer, jiffies + 1);
} else
del_timer_sync(&buf->timer);
@@ -604,15 +609,6 @@ size_t relay_switch_subbuf(struct rchan_
buf->subbufs_produced++;
buf->dentry->d_inode->i_size += 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;
On Sat, 20 Apr 2013 15:37:08 +0800 "zhangwei(Jovi)" <[email protected]> wrote:
> Hi,
>
> Ingo, Steven, I get this patch from 3.4 preempt-rt patch set, It seems that this patch
> fix relayfs bug not only for rt kernel, but also for mainline.
>
> When I'm using below ktap script to tracing all event tracepoints, without this patch,
> the system will hang in few seconds, the patch indeed fix the problem as the changelog pointed.
>
> function eventfun (e) {
> printf("%d %d\t%s\t%s", cpu(), pid(), execname(), e.annotate)
> }
>
> kdebug.probe("tp:", eventfun)
>
> kdebug.probe_end(function () {
> printf("probe end\n")
> })
>
>
> This patch is old, I can 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)
>
> I hope you can remember this :)
>
> so why we didn't commit this patch into mainline? any concern?
>
> Thanks.
>
> ------------------------------------->
> Subject: relay: fix timer madness
> From: Ingo Molnar <[email protected]>
>
> 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 <[email protected]>
(This version of the patch should have your signed-off-by)
> @@ -604,15 +609,6 @@ size_t relay_switch_subbuf(struct rchan_
> buf->subbufs_produced++;
> buf->dentry->d_inode->i_size += 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);
> }
We've "fixed" the printk-inside-runqueue-lock deadlocks via icky
hackery in wake_up_klogd(). I guess we could do it the same way here.
But the two approaches are conceptually very similar and this version
in relay.c is much simpler.