Bug 60002

Created attachment 78651 [details] [review]
fix for initial console blank state

So, after some discussions on IRC with one part of the solution is to force a full modeset to make sure we unblank if we started the test with a blanked state.

A patch for this is attached could you try it on your box with the command sequence I posted in my previous comment?

A secondary issue that can happen is blanking _during_ the test, for that we need to prevent blanking from within the test app, we'll follow-up with a patch for that too.

(In reply to comment #1)
> Created attachment 78651 [details] [review] [review]
> fix for initial console blank state
> 
> So, after some discussions on IRC with one part of the solution is to force
> a full modeset to make sure we unblank if we started the test with a blanked
> state.
> 
> A patch for this is attached could you try it on your box with the command
> sequence I posted in my previous comment?
> 
> A secondary issue that can happen is blanking _during_ the test, for that we
> need to prevent blanking from within the test app, we'll follow-up with a
> patch for that too.

I tried the patch you appended on -nightly top, the result is just like yanbing reported: "The first time is pass,but after that always fail". I list the phenomenon below, if anything you needed I haven't described clearly, please comment.
-------------------------------------------------------------
[root@x-pnv1 tests]# ./kms_flip --run-subtest blocking-absolute-wf_vblank
Using monotonic timestamps
running testcase: blocking-absolute-wf_vblank
could not find requested crtc 3
Beginning blocking-absolute-wf_vblank on crtc 4, connector 5
  1024x600 60 1024 1048 1184 1438 600 603 604 628 0xa 0x48 54200
.............................................................
blocking-absolute-wf_vblank on crtc 4, connector 5: PASSED

[root@x-pnv1 tests]# echo $?
0
[root@x-pnv1 tests]# ./kms_flip --run-subtest blocking-absolute-wf_vblank
Using monotonic timestamps
running testcase: blocking-absolute-wf_vblank
could not find requested crtc 3
Beginning blocking-absolute-wf_vblank on crtc 4, connector 5
  1024x600 60 1024 1048 1184 1438 600 603 604 628 0xa 0x48 54200
.run_test_step:732 failed, ret=-16, errno=16
Aborted (core dumped)
[root@x-pnv1 tests]# echo $?
134

(In reply to comment #2)
> (In reply to comment #1)
> 
> I tried the patch you appended on -nightly top, the result is just like
> yanbing reported: "The first time is pass,but after that always fail". I
> list the phenomenon below, if anything you needed I haven't described
> clearly, please comment.
> -------------------------------------------------------------
> [root@x-pnv1 tests]# ./kms_flip --run-subtest blocking-absolute-wf_vblank
> Using monotonic timestamps
> running testcase: blocking-absolute-wf_vblank
> could not find requested crtc 3
> Beginning blocking-absolute-wf_vblank on crtc 4, connector 5
>   1024x600 60 1024 1048 1184 1438 600 603 604 628 0xa 0x48 54200
> .............................................................
> blocking-absolute-wf_vblank on crtc 4, connector 5: PASSED
> 
> [root@x-pnv1 tests]# echo $?
> 0
> [root@x-pnv1 tests]# ./kms_flip --run-subtest blocking-absolute-wf_vblank
> Using monotonic timestamps
> running testcase: blocking-absolute-wf_vblank
> could not find requested crtc 3
> Beginning blocking-absolute-wf_vblank on crtc 4, connector 5
>   1024x600 60 1024 1048 1184 1438 600 603 604 628 0xa 0x48 54200
> .run_test_step:732 failed, ret=-16, errno=16
> Aborted (core dumped)
> [root@x-pnv1 tests]# echo $?
> 134

Thanks, so this is a separate issue. Could you try the same patch for bug #59834?

Can you please retest with the following patch applied to latest -nightly?

https://patchwork.kernel.org/patch/2709081/

(In reply to comment #4)
> Can you please retest with the following patch applied to latest -nightly?
> 
> https://patchwork.kernel.org/patch/2709081/

I tried this patch on -nightly top, this issue also exist as well as comment 2.

Could you rerun this test? It's not clear where exactly in the source the failure happens, the new igt_assert() should help in this.

(In reply to comment #6)
> Could you rerun this test? It's not clear where exactly in the source the
> failure happens, the new igt_assert() should help in this.

I use latest nightly kernel to run this case again and find this issue always exist. 

------------------------------------------------------

[root@x-pnv1 tests]# ./kms_flip --run-subtest blocking-absolute-wf_vblank
Using monotonic timestamps
no crtc with a compatible encoder (crtc_idx_mask 00000001)
Beginning blocking-absolute-wf_vblank on crtc 4, connector 5
  1024x600 60 1024 1048 1184 1438 600 603 604 628 0xa 0x48 54200
.............................................................
blocking-absolute-wf_vblank on crtc 4, connector 5: PASSED

Subtest blocking-absolute-wf_vblank: SUCCESS
[root@x-pnv1 tests]# ./kms_flip --run-subtest blocking-absolute-wf_vblank
Using monotonic timestamps
no crtc with a compatible encoder (crtc_idx_mask 00000001)
Beginning blocking-absolute-wf_vblank on crtc 4, connector 5
  1024x600 60 1024 1048 1184 1438 600 603 604 628 0xa 0x48 54200
.Test assertion failure function run_test_step, file kms_flip.c:831:
Failed assertion: (do_wait_for_vblank(o, o->pipe, target_seq, &vbl_reply)) == 0
Subtest blocking-absolute-wf_vblank: FAIL
DPMS property not found on 5

This bug seems to have been neglected a bit, please retest on current -nightly.

Closing due to inactivity. Please reopen if this is still an issue.

Fixed.
# ./kms_flip --run-subtest blocking-absolute-wf_vblank
IGT-Version: 1.7-g67e29a3 (i686) (Linux: 3.16.0-rc2_drm-intel-nightly_9bfcb9_20140702+ i686)
Using monotonic timestamps
Beginning blocking-absolute-wf_vblank on crtc 9, connector 10
  1024x600 60 1024 1072 1104 1200 600 603 609 625 0xa 0x48 45000
....................................................................................................................................................................................
blocking-absolute-wf_vblank on crtc 9, connector 10: PASSED

Subtest blocking-absolute-wf_vblank: SUCCESS
# ./kms_flip --run-subtest blocking-absolute-wf_vblank
IGT-Version: 1.7-g67e29a3 (i686) (Linux: 3.16.0-rc2_drm-intel-nightly_9bfcb9_20140702+ i686)
Using monotonic timestamps
Beginning blocking-absolute-wf_vblank on crtc 9, connector 10
  1024x600 60 1024 1072 1104 1200 600 603 609 625 0xa 0x48 45000
....................................................................................................................................................................................
blocking-absolute-wf_vblank on crtc 9, connector 10: PASSED

Subtest blocking-absolute-wf_vblank: SUCCESS

Closing old verified.