Recent changes
I have two questions about recent changes made to the kernel driver:
1. Looking at the intel_setup_overlay() after the commit ecbec53b1d00ba582f71b210ed96cafc05ebd189, it seems there is potential double unlock for the struct_mutex in the (unlikely) case when intel_overlay_map_regs() failed.
2. In i915_wait_request(), after the commit 4697995b98417c6da9ab2708a36f5e2bc926c8ac, shouldn't the calls
i915_driver_irq_preinstall(ring->dev); i915_driver_irq_postinstall(ring->dev); be replaced with ring->dev->driver->irq_preinstall(ring->dev); ring->dev->driver->irq_postinstall(ring->dev);
?
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 6 lines · Konstantin Belousov
I have two questions about recent changes made to the kernel driver:
1. Looking at the
intel_setup_overlay()after the commitecbec53b1d00ba582f71b210ed96cafc05ebd189, it seems there is potential double unlock for thestruct_mutexin the (unlikely) case whenintel_overlay_map_regs()failed.
Dan Carpenter pointed out a further mistake in that we weren't handling the pin count correctly for 945/965/g33.
Quoted 8 lines
2. In
i915_wait_request(), after the commit4697995b98417c6da9ab2708a36f5e2bc926c8ac, shouldn't the callsi915_driver_irq_preinstall(ring->dev); i915_driver_irq_postinstall(ring->dev); be replaced with ring->dev->driver->irq_preinstall(ring->dev); ring->dev->driver->irq_postinstall(ring->dev);?
Yes.
-Chris
Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre