[2/2] igt/kms_frontbuffer_tracking: Bump the wait time for FBC

Submitted by Chris Wilson on Jan. 25, 2018, 9:28 p.m.

Details

Message ID 20180125212849.2261-2-chris@chris-wilson.co.uk
State New
Series "Series without cover letter"
Headers show

Commit Message

Chris Wilson Jan. 25, 2018, 9:28 p.m.
It is taking longer than a couple of seconds for the FBC worker to be
executed after scheduling; and then will take a minimum of a vblank
interval for it activate. So wait longer to reduce the flip flops.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 tests/kms_frontbuffer_tracking.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Patch hide | download patch | download mbox

diff --git a/tests/kms_frontbuffer_tracking.c b/tests/kms_frontbuffer_tracking.c
index 79e4f5862..fb5627535 100644
--- a/tests/kms_frontbuffer_tracking.c
+++ b/tests/kms_frontbuffer_tracking.c
@@ -920,7 +920,7 @@  static bool fbc_stride_not_supported(void)
 
 static bool fbc_wait_until_enabled(void)
 {
-	return igt_wait(fbc_is_enabled(IGT_LOG_DEBUG), 2000, 1);
+	return igt_wait(fbc_is_enabled(IGT_LOG_DEBUG), 5000, 8);
 }
 
 static bool psr_wait_until_enabled(void)

Comments

Chris Wilson Jan. 26, 2018, 9:39 p.m.
Quoting Chris Wilson (2018-01-25 21:28:49)
> It is taking longer than a couple of seconds for the FBC worker to be
> executed after scheduling; and then will take a minimum of a vblank
> interval for it activate. So wait longer to reduce the flip flops.
> 
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Any acks?

> ---
>  tests/kms_frontbuffer_tracking.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tests/kms_frontbuffer_tracking.c b/tests/kms_frontbuffer_tracking.c
> index 79e4f5862..fb5627535 100644
> --- a/tests/kms_frontbuffer_tracking.c
> +++ b/tests/kms_frontbuffer_tracking.c
> @@ -920,7 +920,7 @@ static bool fbc_stride_not_supported(void)
>  
>  static bool fbc_wait_until_enabled(void)
>  {
> -       return igt_wait(fbc_is_enabled(IGT_LOG_DEBUG), 2000, 1);
> +       return igt_wait(fbc_is_enabled(IGT_LOG_DEBUG), 5000, 8);
>  }
>  
>  static bool psr_wait_until_enabled(void)
> -- 
> 2.15.1
>
Chris Wilson Feb. 3, 2018, 10:13 a.m.
Quoting Chris Wilson (2018-01-26 21:39:28)
> Quoting Chris Wilson (2018-01-25 21:28:49)
> > It is taking longer than a couple of seconds for the FBC worker to be
> > executed after scheduling; and then will take a minimum of a vblank
> > interval for it activate. So wait longer to reduce the flip flops.
> > 
> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> 
> Any acks?

Still looking for a response here. CI says that 2s isn't enough for FBC
to kick in (it spends most of that time waiting for an invalidate, and
then at the every end about 5ms waiting for the next vblank).

In terms of human response, we don't care so much how long it takes to
enable FBC, but how long it takes to disable; similarly for PSR. The
downside to it taking longer than expected to enable FBC is power
consumption. Is the display block hooked up to rapl?
-Chris
Rodrigo Vivi Feb. 5, 2018, 9:54 p.m.
On Sat, Feb 03, 2018 at 10:13:36AM +0000, Chris Wilson wrote:
> Quoting Chris Wilson (2018-01-26 21:39:28)
> > Quoting Chris Wilson (2018-01-25 21:28:49)
> > > It is taking longer than a couple of seconds for the FBC worker to be
> > > executed after scheduling; and then will take a minimum of a vblank
> > > interval for it activate. So wait longer to reduce the flip flops.
> > > 
> > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > 
> > Any acks?
> 
> Still looking for a response here. CI says that 2s isn't enough for FBC
> to kick in (it spends most of that time waiting for an invalidate, and
> then at the every end about 5ms waiting for the next vblank).
> 
> In terms of human response, we don't care so much how long it takes to
> enable FBC, but how long it takes to disable; similarly for PSR. The
> downside to it taking longer than expected to enable FBC is power
> consumption. Is the display block hooked up to rapl?

Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

> -Chris
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx