[Intel-gfx] BUG: sleeping function called from invalid context on 3.10.10-rt7

Steven Rostedt rostedt at goodmis.org
Fri Oct 11 16:38:14 CEST 2013


On Fri, 11 Oct 2013 15:30:22 +0200
Sebastian Andrzej Siewior <bigeasy at linutronix.de> wrote:

> On 10/11/2013 02:37 PM, Steven Rostedt wrote:
> > On Fri, 11 Oct 2013 12:18:00 +0200
> > Sebastian Andrzej Siewior <bigeasy at linutronix.de> wrote:
> > 
> >> * Mario Kleiner | 2013-09-26 18:16:47 [+0200]:
> >>
> >>> Good! I will do that. Thanks for clarifying the irq and constraints
> >>> on raw locks in the other thread.
> >>
> >> Are there any suggestions for "now"?  preempt_disable_nort() like Luis
> >> suggesed?
> >>
> > 
> > The preempt_disable_nort() is rather pointless, because the
> > preempt_disable() was added specifically for -rt. When PREEMPT_RT is
> > not enabled, preemption is disabled there already by the previous calls
> > to spin_lock().
> 
> Either way. Then I remove the preempt_enable/disable call. Any
> objections?
> 

I have no issues with it, but it may cause issues with timings for the
device. But I see Mario is looking into that :-)

-- Steve



More information about the Intel-gfx mailing list