diff options
| author | Russell King <rmk+kernel@arm.linux.org.uk> | 2015-06-06 21:41:09 +0100 |
|---|---|---|
| committer | Russell King <rmk+kernel@arm.linux.org.uk> | 2015-09-15 16:19:48 +0100 |
| commit | 0fc6f44d9683c61678da4b0eebc89e8fa624de39 (patch) | |
| tree | bf0e5d5e975a34c36e352d48609e0a1df7e871c8 /drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c | |
| parent | f84a97d4804a09240372dc7b195f9d6162152228 (diff) | |
drm/i2c: tda998x: re-implement "Fix EDID read timeout on HDMI connect"
Commit 6833d26ef823 ("drm: tda998x: Fix EDID read timeout on HDMI
connect") used a weak scheme to try and delay reading EDID on a HDMI
connect event. It is weak because delaying the notification of a
hotplug event does not stop userspace from trying to read the EDID
within the 100ms delay.
The solution provided here solves this issue:
* When a HDMI connection event is detected, mark a blocking flag for
EDID reads, and start a timer for the delay.
* If an EDID read is attempted, and the blocking flag is set, wait
for the blocking flag to clear.
* When the timer expires, clear the blocking flag and wake any thread
waiting for the EDID read.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c')
0 files changed, 0 insertions, 0 deletions