drm/i915/psr: Fix race in intel_psr_work()
authorDhinakaran Pandiyan <dhinakaran.pandiyan@gmail.com>
Mon, 25 Jun 2018 05:47:40 +0000 (22:47 -0700)
committerDhinakaran Pandiyan <dhinakaran.pandiyan@gmail.com>
Tue, 26 Jun 2018 18:44:55 +0000 (11:44 -0700)
commitc12e0643a05d978657877630d4da1ace06ea3720
tree461efd4f52fe98a6c5c99df846cab9fff52f8181
parentcf5d862db2e301b8e487f42f99c5cf6f5228ddae
drm/i915/psr: Fix race in intel_psr_work()

Commit 5422b37c907e ("drm/i915/psr: Kill delays when activating psr
back.") switched from delayed work to the plain variant and while doing so
removed the check for work_busy() before scheduling a PSR activation.
This appears to cause consecutive executions of psr_activate() in this
scenario - after a worker picks up the PSR work item for execution and
before the work function can acquire the PSR mutex, a psr_flush() can
get hold of the mutex and schedule another PSR work. Without a psr_exit()
between the two psr_activate() calls, warning messages get printed.
Further, since we drop the mutex in the midst of psr_work() to wait for
PSR to idle, another work item can also get scheduled. Fix this by
returning if PSR was already active.

Fixes: 5422b37c907e ("drm/i915/psr: Kill delays when activating psr back.")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106948
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180625054741.3919-1-dhinakaran.pandiyan@intel.com
drivers/gpu/drm/i915/intel_psr.c