Created attachment 96339 [details]
dmesg

System Environment:
--------------------------
Platform: Haswell
Kernel: drm-intel-fixes/0f4706d2740f2a221cd502922b22e522009041d9

Bug detailed description:
-----------------------------
It fails on all platforms with igt master branch.

Bisect shows: a388f3b98f039403a4583748db980d4a44dd6748 is the first bad commit.
commit a388f3b98f039403a4583748db980d4a44dd6748
Author:     Daniel Vetter <daniel.vetter@ffwll.ch>
AuthorDate: Sat Mar 22 13:22:11 2014 +0100
Commit:     Daniel Vetter <daniel.vetter@ffwll.ch>
CommitDate: Sat Mar 22 14:28:03 2014 +0100

    lib: allow igt_skip_on_simulation outside of fixtures.

    Thomas noticed that in simulation mode a lot of the tests fall over
    instead of skipping properly. This is due to recently added
    self-checks which ensure that any call to igt_skip happens either
    within a fixture or subtest block (or it's a simple test without
    subtests). This is to catch bugs since pretty much always not wrapping
    up hardware setup and checks into these blocks is a bug.

    Bug simulation skipping is a bit different, so allow that exception.
    Otherwise we'd need to fix up piles of tests (and likely need to play
    a game of whack-a-mole).

    Also add a library testcase for all the different variants to make
    sure it really works.

    Cc: Thomas Wood <thomas.wood@intel.com>
    Cc: Ben Widawsky <benjamin.widawsky@intel.com>
    Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>


run ./drv_suspend --run-subtest debugfs-reader
output:
IGT-Version: 1.6-g7a81094 (x86_64) (Linux: 3.14.0-rc5_drm-intel-fixes_0f4706_20140323+ x86_64)
rtcwake: wakeup from "mem" using /dev/rtc0 at Tue Mar 25 18:05:04 2014
Subtest debugfs-reader: SUCCESS
drv_suspend: igt_core.c:624: igt_fail: Assertion `!test_with_subtests || in_fixture' failed.
Aborted (core dumped)