[Qemu-devel] [PATCH] char-pty: fix glib assert

[Qemu-devel] [PATCH] char-pty: fix glib assert

[Juan Quintela]

I am having this assert over and over:

(process:30804): GLib-CRITICAL **: g_source_unref: assertion 'source != NULL' failed

gdb points to the line changed on this patch, and my reading of
commit:

2c716ba1506769c9be2caa02f0f6d6e7c00f4304

is that it should be timer_src what is unrefered there.

But I don't claim to fully understand this code, so ....

Any comment?

Thanks, Juan.

Signed-off-by: Juan Quintela <quintela@redhat.com>
CC: Peter Xu <peterx@redhat.com>
CC: Paolo Bonzini <pbonzini@redhat.com>


---
 chardev/char-pty.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/chardev/char-pty.c b/chardev/char-pty.c
index 89315e6807..c26e02bce8 100644
--- a/chardev/char-pty.c
+++ b/chardev/char-pty.c
@@ -57,8 +57,8 @@ static gboolean pty_chr_timer(gpointer opaque)
     PtyChardev *s = PTY_CHARDEV(opaque);
 
     qemu_mutex_lock(&chr->chr_write_lock);
+    g_source_unref(s->timer_src);
     s->timer_src = NULL;
-    g_source_unref(s->open_source);
     s->open_source = NULL;
     if (!s->connected) {
         /* Next poll ... */
-- 
2.14.3

Re: [Qemu-devel] [PATCH] char-pty: fix glib assert

[Laurent Vivier]

On 29/01/2018 17:17, Juan Quintela wrote:
> I am having this assert over and over:
> 
> (process:30804): GLib-CRITICAL **: g_source_unref: assertion 'source != NULL' failed
> 
> gdb points to the line changed on this patch, and my reading of
> commit:
> 
> 2c716ba1506769c9be2caa02f0f6d6e7c00f4304
> 
> is that it should be timer_src what is unrefered there.
> 
> But I don't claim to fully understand this code, so ....
> 
> Any comment?
> 
> Thanks, Juan.
> 
> Signed-off-by: Juan Quintela <quintela@redhat.com>
> CC: Peter Xu <peterx@redhat.com>
> CC: Paolo Bonzini <pbonzini@redhat.com>
> 
> 
> ---
>  chardev/char-pty.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/chardev/char-pty.c b/chardev/char-pty.c
> index 89315e6807..c26e02bce8 100644
> --- a/chardev/char-pty.c
> +++ b/chardev/char-pty.c
> @@ -57,8 +57,8 @@ static gboolean pty_chr_timer(gpointer opaque)
>      PtyChardev *s = PTY_CHARDEV(opaque);
>  
>      qemu_mutex_lock(&chr->chr_write_lock);
> +    g_source_unref(s->timer_src);
>      s->timer_src = NULL;
> -    g_source_unref(s->open_source);
>      s->open_source = NULL;
>      if (!s->connected) {
>          /* Next poll ... */
> 

If I read correctly the doc, we also need a
g_source_destroy(s->timer_src). Perhaps we can directly call
pty_chr_timer_cancel() here?

Thanks,
Laurent

Re: [Qemu-devel] [PATCH] char-pty: fix glib assert

[Peter Xu]

On Mon, Jan 29, 2018 at 05:40:11PM +0100, Laurent Vivier wrote:
> On 29/01/2018 17:17, Juan Quintela wrote:
> > I am having this assert over and over:
> > 
> > (process:30804): GLib-CRITICAL **: g_source_unref: assertion 'source != NULL' failed
> > 
> > gdb points to the line changed on this patch, and my reading of
> > commit:
> > 
> > 2c716ba1506769c9be2caa02f0f6d6e7c00f4304
> > 
> > is that it should be timer_src what is unrefered there.
> > 
> > But I don't claim to fully understand this code, so ....
> > 
> > Any comment?
> > 
> > Thanks, Juan.
> > 
> > Signed-off-by: Juan Quintela <quintela@redhat.com>
> > CC: Peter Xu <peterx@redhat.com>
> > CC: Paolo Bonzini <pbonzini@redhat.com>
> > 
> > 
> > ---
> >  chardev/char-pty.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/chardev/char-pty.c b/chardev/char-pty.c
> > index 89315e6807..c26e02bce8 100644
> > --- a/chardev/char-pty.c
> > +++ b/chardev/char-pty.c
> > @@ -57,8 +57,8 @@ static gboolean pty_chr_timer(gpointer opaque)
> >      PtyChardev *s = PTY_CHARDEV(opaque);
> >  
> >      qemu_mutex_lock(&chr->chr_write_lock);
> > +    g_source_unref(s->timer_src);
> >      s->timer_src = NULL;
> > -    g_source_unref(s->open_source);
> >      s->open_source = NULL;
> >      if (!s->connected) {
> >          /* Next poll ... */
> > 
> 
> If I read correctly the doc, we also need a
> g_source_destroy(s->timer_src). Perhaps we can directly call
> pty_chr_timer_cancel() here?

Yes.  Another fix is already queued by Paolo here:

  http://patchwork.ozlabs.org/patch/862763/

Sorry for that!

-- 
Peter Xu