gpu: nvgpu: update test_gk20a_runlist_wait_pending spec

Update description for test_gk20a_runlist_wait_pending.
Test is now using register IO callbacks instead of a thread to
clear fifo_eng_runlist_r pending bit.

Jira NVGPU-3793

Change-Id: I7ddae3f4574705eb57bd432693da91480c7f09dd
Signed-off-by: Thomas Fleury <tfleury@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2243621
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: Debarshi Dutta <ddutta@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This commit is contained in:
Thomas Fleury
2019-11-20 14:27:50 -05:00
committed by Alex Waterman
parent db62de9f00
commit a6c4eb4c14

View File

@@ -81,15 +81,17 @@ int test_gk20a_runlist_hw_submit(struct unit_module *m,
* Steps:
* - Check case where runlist is not pending (not wait).
* - Set register to indicate that runlist is NOT pending.
* - call gk20a_runlist_wait_pending.
* - Call gk20a_runlist_wait_pending.
* - Check case where some polling is needed until runlist is not pending:
* - Set register to indicate that runlist is pending.
* - Start a thread to clear this bit after some time.
* Time is chosed so that one ore more loops are required.
* - call gk20a_runlist_wait_pending.
* - Install register IO callbacks in order to control
* value read from fifo_eng_runlist_r register.
* - Configure callback to clear pending bit after one nvgpu_readl.
* - Call gk20a_runlist_wait_pending.
* - Configure callback to clear pending bit after two nvgpu_readl.
* - Call gk20a_runlist_wait_pending.
* - Check case where polling times out:
* - Set register to indicate that runlist is pending.
* - call gk20a_runlist_wait_pending.
* - Call gk20a_runlist_wait_pending.
*
* Output: Returns PASS if all branches gave expected results. FAIL otherwise.
*/