On Tue, 2 Oct 2012 17:43:43 -0500 Jesse Barnes <jbar...@virtuousgeek.org> wrote:
> To match IVB. > > Signed-off-by: Jesse Barnes <jbar...@virtuousgeek.org> > --- > drivers/gpu/drm/i915/i915_irq.c | 18 +++++------------- > 1 file changed, 5 insertions(+), 13 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c > index d915126..096b387 100644 > --- a/drivers/gpu/drm/i915/i915_irq.c > +++ b/drivers/gpu/drm/i915/i915_irq.c > @@ -1957,6 +1957,7 @@ static int valleyview_irq_postinstall(struct drm_device > *dev) > u32 enable_mask; > u32 hotplug_en = I915_READ(PORT_HOTPLUG_EN); > u32 pipestat_enable = PLANE_FLIP_DONE_INT_EN_VLV; > + u32 render_irqs; > u16 msid; > > enable_mask = I915_DISPLAY_PORT_INTERRUPT; > @@ -1996,21 +1997,12 @@ static int valleyview_irq_postinstall(struct > drm_device *dev) > I915_WRITE(VLV_IIR, 0xffffffff); > I915_WRITE(VLV_IIR, 0xffffffff); > > - dev_priv->gt_irq_mask = ~0; > - > - I915_WRITE(GTIIR, I915_READ(GTIIR)); > I915_WRITE(GTIIR, I915_READ(GTIIR)); > I915_WRITE(GTIMR, dev_priv->gt_irq_mask); > - I915_WRITE(GTIER, GT_GEN6_BLT_FLUSHDW_NOTIFY_INTERRUPT | > - GT_GEN6_BLT_CS_ERROR_INTERRUPT | > - GT_GEN6_BLT_USER_INTERRUPT | > - GT_GEN6_BSD_USER_INTERRUPT | > - GT_GEN6_BSD_CS_ERROR_INTERRUPT | > - GT_GEN7_L3_PARITY_ERROR_INTERRUPT | > - GT_PIPE_NOTIFY | > - GT_RENDER_CS_ERROR_INTERRUPT | > - GT_SYNC_STATUS | > - GT_USER_INTERRUPT); > + > + render_irqs = GT_USER_INTERRUPT | GEN6_BSD_USER_INTERRUPT | > + GEN6_BLITTER_USER_INTERRUPT; > + I915_WRITE(GTIER, render_irqs); > POSTING_READ(GTIER); > > /* ack & enable invalid PTE error interrupts */ Were parity errors intentionally dropped? I think it's fine to drop it, but maybe add it to the commit message. -- Ben Widawsky, Intel Open Source Technology Center _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
