Re: [PATCH 0/2] Debug bits for the tty ldisc problem

Re: [PATCH 0/2] Debug bits for the tty ldisc problem

[Peter Zijlstra]

On Mon, 2009-06-15 at 16:32 +0100, Alan Cox wrote:
> I'm not 100% sure the close one is
> safe for locking but as the lock detector has been busted since 2.6.30 its a
> bit hard to be sure.

Would you have some specifics?

Re: [PATCH 0/2] Debug bits for the tty ldisc problem

[Alan Cox]

On Mon, 15 Jun 2009 16:39:56 +0200
Peter Zijlstra <peterz@infradead.org> wrote:

> On Mon, 2009-06-15 at 16:32 +0100, Alan Cox wrote:
> > I'm not 100% sure the close one is
> > safe for locking but as the lock detector has been busted since 2.6.30 its a
> > bit hard to be sure.
> 
> Would you have some specifics?

I stuck it in bugzilla basically it blows up with a 

BUG: MAX_STACK_TRACE_ENTRIES too low! turning off the locking correctness
validator.

and has done for ages.

Re: [PATCH 0/2] Debug bits for the tty ldisc problem

[Ingo Molnar]

* Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> On Mon, 15 Jun 2009 16:39:56 +0200
> Peter Zijlstra <peterz@infradead.org> wrote:
> 
> > On Mon, 2009-06-15 at 16:32 +0100, Alan Cox wrote:
> > > I'm not 100% sure the close one is
> > > safe for locking but as the lock detector has been busted since 2.6.30 its a
> > > bit hard to be sure.
> > 
> > Would you have some specifics?
> 
> I stuck it in bugzilla basically it blows up with a 
> 
> BUG: MAX_STACK_TRACE_ENTRIES too low! turning off the locking 
> correctness validator.
> 
> and has done for ages.

It's hitting a simple limit. Please try the patch below - it 
quadruples the value printed in the warning (the number of entries) 
- does that get your system going?

Lock validation is getting more and more popular, covering now a lot 
more than just locks (workqueues, timers, etc.), so the max number 
of entries is steadily increasing. Some configs will run into this 
faster.

	Ingo

diff --git a/kernel/lockdep_internals.h b/kernel/lockdep_internals.h
index 699a2ac..031f4c6 100644
--- a/kernel/lockdep_internals.h
+++ b/kernel/lockdep_internals.h
@@ -65,7 +65,7 @@ enum {
  * Stack-trace: tightly packed array of stack backtrace
  * addresses. Protected by the hash_lock.
  */
-#define MAX_STACK_TRACE_ENTRIES	262144UL
+#define MAX_STACK_TRACE_ENTRIES	1048576UL
 
 extern struct list_head all_lock_classes;
 extern struct lock_chain lock_chains[];

[PATCH 0/2] Debug bits for the tty ldisc problem

[Alan Cox]

See if this turns anything up or fixes it. I'm not 100% sure the close one is
safe for locking but as the lock detector has been busted since 2.6.30 its a
bit hard to be sure.
---

Alan Cox (2):
      ldisc: debug aids
      ldisc: Make sure the ldisc isn't active when we close it


 drivers/char/tty_ldisc.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

-- 
	"In flight refueling scares me. It's like two elephants 
			mating at mach one"
				-- Arjan van de Ven

[PATCH 1/2] ldisc: Make sure the ldisc isn't active when we close it

[Alan Cox]

From: Alan Cox <alan@linux.intel.com>

Signed-off-by: Alan Cox <alan@linux.intel.com>
---

 drivers/char/tty_ldisc.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)


diff --git a/drivers/char/tty_ldisc.c b/drivers/char/tty_ldisc.c
index 94b3e06..874c248 100644
--- a/drivers/char/tty_ldisc.c
+++ b/drivers/char/tty_ldisc.c
@@ -793,6 +793,8 @@ void tty_ldisc_hangup(struct tty_struct *tty)
 		/* Avoid racing set_ldisc */
 		mutex_lock(&tty->ldisc_mutex);
 		/* Switch back to N_TTY */
+		tty_ldisc_halt(tty);
+		tty_ldisc_wait_idle(tty);
 		tty_ldisc_reinit(tty);
 		/* At this point we have a closed ldisc and we want to
 		   reopen it. We could defer this to the next open but

[PATCH 2/2] ldisc: debug aids

[Alan Cox]

From: Alan Cox <alan@linux.intel.com>


---

 drivers/char/tty_ldisc.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)


diff --git a/drivers/char/tty_ldisc.c b/drivers/char/tty_ldisc.c
index 874c248..a19e935 100644
--- a/drivers/char/tty_ldisc.c
+++ b/drivers/char/tty_ldisc.c
@@ -207,6 +207,7 @@ static void tty_ldisc_put(struct tty_ldisc *ld)
 	ldo->refcount--;
 	module_put(ldo->owner);
 	spin_unlock_irqrestore(&tty_ldisc_lock, flags);
+	WARN_ON(ld->refcount);
 	kfree(ld);
 }