diff options
| author | Glenn Kasten <gkasten@google.com> | 2012-02-02 14:16:03 -0800 |
|---|---|---|
| committer | Glenn Kasten <gkasten@google.com> | 2012-02-24 13:31:26 -0800 |
| commit | 8abf44d2f2bcd20a2835570efe89d89c19db426a (patch) | |
| tree | 29fc684b04d18621913145c807c4493f5b607d86 /drm | |
| parent | d6fd85a157ce2054b2304e6d171fa87ae09c363d (diff) | |
| download | frameworks_av-8abf44d2f2bcd20a2835570efe89d89c19db426a.zip frameworks_av-8abf44d2f2bcd20a2835570efe89d89c19db426a.tar.gz frameworks_av-8abf44d2f2bcd20a2835570efe89d89c19db426a.tar.bz2 | |
Fix tracking of hardware state for dump
At end of AudioFlinger::onFirstRef(), the hardware status was being left
in wrong state. It should be AUDIO_HW_IDLE but was AUDIO_HW_INIT.
mHardwareStatus was being set to AUDIO_HW_OUTPUT_OPEN too early, and so
a return would leave it in the wrong state until next hardware operation.
Take the hardware lock for dev->get_parameters, and update mHardwareStatus
before and after.
Keep hardware lock only for the duration of the dev->set_parameters.
Rename two constants in enum hardware_call_state to have the prefix
AUDIO_HW so they follow the naming conventions.
Add comments.
Change-Id: I6c7450b11f9b13adaeef9cec874333e478a58fc0
Diffstat (limited to 'drm')
0 files changed, 0 insertions, 0 deletions