| Message ID | 20190410105923.18546-7-mika.kuoppala@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [1/7] drm/i915: Use dedicated rc6 enabling sequence for gen11 | expand |
Quoting Mika Kuoppala (2019-04-10 11:59:23) > There is a chance we can see spurious interrupts in live > now. We have more engines enabled and that with more elaborate > access patterns with pm and display, increases the chances > hardware just makes a social call, without anything to work on. > > Remove the error as we have tests to actually probe if > we really miss interrupt, instead of getting spurious ones. > > Note that now we do write to intr_dw even with a zero > value. This is considered advantegous as the write > is an ack that sw is done. Writing 0 back to 0 handled interrupts seems fair. > Cc: Chris Wilson <chris@chris-wilson.co.uk> > Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> I think both interrupt handling patches should go in first, as they are bug fixes that may be required even before rps is fixed. -Chris
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 3531ef79ae21..d934545445e1 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -3041,14 +3041,8 @@ gen11_gt_bank_handler(struct drm_i915_private * const i915, intr_dw = raw_reg_read(regs, GEN11_GT_INTR_DW(bank)); - if (unlikely(!intr_dw)) { - DRM_ERROR("GT_INTR_DW%u blank!\n", bank); - return; - } - for_each_set_bit(bit, &intr_dw, 32) { - const u32 ident = gen11_gt_engine_identity(i915, - bank, bit); + const u32 ident = gen11_gt_engine_identity(i915, bank, bit); gen11_gt_identity_handler(i915, ident); }
There is a chance we can see spurious interrupts in live now. We have more engines enabled and that with more elaborate access patterns with pm and display, increases the chances hardware just makes a social call, without anything to work on. Remove the error as we have tests to actually probe if we really miss interrupt, instead of getting spurious ones. Note that now we do write to intr_dw even with a zero value. This is considered advantegous as the write is an ack that sw is done. Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> --- drivers/gpu/drm/i915/i915_irq.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-)