Re: [Qemu-devel] [PING 0.14] fix crash when a read handler calls qemu_set_fd_handler2()

[Anthony Liguori]

On 01/27/2011 01:50 PM, Stefan Hajnoczi wrote:
> On Thu, Jan 27, 2011 at 5:14 PM, Anthony Liguori<address@hidden>  wrote:
>    
> > On 01/27/2011 10:44 AM, Rick Vernam wrote:
> >      
> > > > > Please reply to this message with a subject starting with "[PING 0.14]"
> > > > > and a link to patchwork, gmane, or whatnot; or "[PULL 0.14]" for tree
> > > > > pulls.  Also, I believe all patches meant for stable should also start
> > > > > having "[PATCH 0.14]", too.
> > > > >
> > > > >            
> > > I am not the author of this patch, but it fixed a crash that was bugging
> > > me
> > > consistently.
> > >
> > > http://patchwork.ozlabs.org/patch/62420/
> > >
> > >        
> > Stefan, I thought we're carryign a patch from you that fixes this problem?
> >   Am I mistaken?
> >      
> Yes, that crash was fixed in qemu.git by
> 0290b57bdfec83ca78b6d119ea9847bb17943328 "Delete IOHandlers after
> potentially running them".
>
> To be fair Corentin and Yoshiaki seemed in favor of applying this one
> regardless:
> http://patchwork.ozlabs.org/patch/80335/
>    

Yeah, the logic is bad there though.  The ioh gets freed as soon as that 
condition is over so the assignments to NULL are pointless.

But at any rate, let's just get rid of the whole dang thing and get a 
glib based main loop in as soon as we can.

Regards,

Anthony Liguori

> Stefan
>