Re: i915 suspend crash: BUG: unable to handle kernel NULL pointer deferrence

[Keith Packard <keithp@keithp.com>]

[-- Attachment #1: Type: text/plain, Size: 1797 bytes --]

On Fri, 12 Aug 2011 08:57:41 -0700, "Justin P. Mattock" <justinmattock@gmail.com> wrote:

> [ 96.384971] [drm:i915_get_vblank_timestamp] *ERROR* Invalid crtc 0

You're not using KMS are you...

UMS mode disables a ton of features in the driver; you've hit a couple
that weren't getting skipped in all cases.

Here's some patches that make sure two of the disabled features are
correctly bypassed. If these work for you, I'll create a couple of
patches, one per bug. Afterwards, you should really consider switching
to KMS...

(meanwhile, thanks for finding bugs in ancient code!)

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 02f96fd..9cbb0cd 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -2058,8 +2058,10 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->get_vblank_counter = gm45_get_vblank_counter;
 	}
 
-
-	dev->driver->get_vblank_timestamp = i915_get_vblank_timestamp;
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
+		dev->driver->get_vblank_timestamp = i915_get_vblank_timestamp;
+	else
+		dev->driver->get_vblank_timestamp = NULL;
 	dev->driver->get_scanout_position = i915_get_crtc_scanoutpos;
 
 	if (IS_IVYBRIDGE(dev)) {
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 87677d6..f107423 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -871,7 +871,8 @@ int i915_restore_state(struct drm_device *dev)
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	intel_init_clock_gating(dev);
+	if (drm_core_check_feature(dev, DRIVER_MODESET))
+		intel_init_clock_gating(dev);
 
 	if (IS_IRONLAKE_M(dev)) {
 		ironlake_enable_drps(dev);


-- 
keith.packard@intel.com

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]