[igt-dev] [PATCH i-g-t 05/23] i915/perf: Bump timestamp tolerance for DG1

[Umesh Nerlige Ramappa]

On Tue, Sep 06, 2022 at 04:06:37PM +0300, Lionel Landwerlin wrote:
>On 23/08/2022 21:30, Umesh Nerlige Ramappa wrote:
>>Timestamp deltas obtained from the batch vs the delta obtained from oa
>>reports have a delta that's between 4 to 5 us in DG1. Adjust the test
>>accordingly.
>>
>>Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa at intel.com>
>
>It's odd...
>
>4/5us between 2 commands that follow each other exactly in a command stream?
>
>Even if the PIPE_CONTROL stalls, it's supposed to write the timestamp 
>after it's done stalling.
>
>And the MI_REPORT_PERF_COUNT command is just after.
>
>
>If you're sure there is no other explanation : Reviewed-by: Lionel 
>Landwerlin <lionel.g.landwerlin at intel.com>

Hmm, I debugged this a while ago. I will give it a quick retry and see 
if it's still the same issue.

>
>
>>---
>>  tests/i915/perf.c | 5 +++--
>>  1 file changed, 3 insertions(+), 2 deletions(-)
>>
>>diff --git a/tests/i915/perf.c b/tests/i915/perf.c
>>index ca3ccc17..95600562 100644
>>--- a/tests/i915/perf.c
>>+++ b/tests/i915/perf.c
>>@@ -4084,8 +4084,9 @@ static void gen12_single_ctx_helper(void)
>>  	delta_delta = delta_ts64_ns > delta_oa32_ns ?
>>  		      (delta_ts64_ns - delta_oa32_ns) :
>>  		      (delta_oa32_ns - delta_ts64_ns);
>>-	if (delta_delta > 500) {
>>-		igt_debug("delta_delta exceeds margin, skipping..\n");
>>+	if (delta_delta > 5000) {
>>+		igt_debug("delta_delta = %d. exceeds margin, skipping..\n",
>>+			  delta_delta);
>>  		exit(EAGAIN);
>>  	}
>
>