[Intel-gfx] [SNB: Regression] Re: [PATCH 02/19] drm/i915: split irq handling into per-chipset functions

[Jesse Barnes]

On Wed, 18 May 2011 21:51:23 +1000
Ted Phelps <phelps at gnusto.com> wrote:

> 
> Jesse Barnes writes:
> > Set the IRQ handling functions in driver load so they'll just be used
> > directly, rather than branching over most of the code in the chipset
> > functions.
> 
> I'm seeing a kernel panic on my SNB i7-2600K with keithp/drm-intel-next
> (61e499b).  I've bisected to 4697995 -- the commit mentioned in the
> subject.
> 
> My photo of the stack trace is a bit blurry, but below is my attempted
> transcription of the the stack.  I can't read the addresses at all;
> please let me know if you'd like me to obtain a better dump.
> 
> Call Trace:
>  <IRQ>
>  queue_work_op+0x1d/0x30
>  queue_work+0x35/0x50
>  ironlake_irq_handler+0xec9/0xfa0 [i915]
>  handle_irq_event_percpu+0x75/0x260
>  handle_edge_irq+0x6f/0x110
>  handle_irq+0x1d/0x30
>  do_IRQ+0x58/0xe9
>  common_interrupt+0x13/0x13
>  ... more stuff that I assume is irrelevent ...
> 
> Please let me know if I can provide additional useful information.

Please try this patch.

-- 
Jesse Barnes, Intel Open Source Technology Center


diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 349a03e..e0aed1e 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -1736,6 +1736,7 @@ void ironlake_irq_preinstall(struct drm_device *dev)
 
 	INIT_WORK(&dev_priv->hotplug_work, i915_hotplug_work_func);
 	INIT_WORK(&dev_priv->error_work, i915_error_work_func);
+	INIT_WORK(&dev_priv->rps_work, gen6_pm_rps_work);
 
 	I915_WRITE(HWSTAM, 0xeffe);