RE: [hyperv] BUG at drivers/hv/channel.c:462 while changing MTU

[KY Srinivasan]

> -----Original Message-----
> From: Greg Kroah-Hartman [mailto:gregkh@xxxxxxxxxxxxxxxxxxx]
> Sent: Monday, August 25, 2014 10:50 AM
> To: KY Srinivasan
> Cc: Dan Carpenter; Sitsofe Wheeler; Jason Wang; linux-
> kernel@xxxxxxxxxxxxxxx; David S. Miller; Daniel Borkmann;
> netdev@xxxxxxxxxxxxxxx; devel@xxxxxxxxxxxxxxxxxxxxxx; Haiyang Zhang
> Subject: Re: [hyperv] BUG at drivers/hv/channel.c:462 while changing MTU
> 
> On Mon, Aug 25, 2014 at 05:34:27PM +0000, KY Srinivasan wrote:
> >
> >
> > > -----Original Message-----
> > > From: Dan Carpenter [mailto:dan.carpenter@xxxxxxxxxx]
> > > Sent: Monday, August 25, 2014 2:37 AM
> > > To: Sitsofe Wheeler
> > > Cc: KY Srinivasan; Greg Kroah-Hartman; Jason Wang; linux-
> > > kernel@xxxxxxxxxxxxxxx; David S. Miller; Daniel Borkmann;
> > > netdev@xxxxxxxxxxxxxxx; devel@xxxxxxxxxxxxxxxxxxxxxx; Haiyang Zhang
> > > Subject: Re: [hyperv] BUG at drivers/hv/channel.c:462 while changing
> > > MTU
> > >
> > > The code here is:
> > >
> > > drivers/hv/channel.c
> > >    460          BUG_ON(ret != 0);
> > >    461          t = wait_for_completion_timeout(&info->waitevent, 5*HZ);
> > >    462          BUG_ON(t == 0);
> > >
> > > So it calls BUG_ON() if the teardown takes more than 5 seconds.
> > > It's most likely that there is a race condition somewhere.  It's
> > > also possible that it's just taking longer than 5 seconds for some
> > > odd reason and the bug would go away if we raised it to 60 seconds.
> > >
> > > BUG_ON() after 5 seconds seems like a very aggressive thing.
> >
> > Dan,
> >
> > I am going to audit all BUG_ON() instances.
> 
> Please remove them all, no kernel driver should ever crash the kernel and
> not give a user a chance to recover :(

I will clean this up.

K. Y
> 
> thanks,
> 
> greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/