Hello,
Here's a collection of patches improving robustness to failures in
the device resume/suspend path. Those failures are pretty hard to
reproduce (happens once in a while on a deqp-vk run), so I used a
mechanism to fake them.
Faking a FW boot failure is kinda tricky though, which means the
last patch has only been partially tested:
- the fast reset path is well tested because that's the default on
a device suspend
- the slow reset has been tested with a hack replacing fast resets
by slow resets
- the fast -> slow reset fallback has been tested by faking boot
failures after a fast reset, but these are not real, which means
we can't really validate if the MCU recovers fine after a slow
reset
On the other hand, this implementation doesn't look like it could
do more harm than the current one (the only difference is the
extra GPU soft-reset that happens between the fast and slow FW
boot).
Regards,
Boris
Boris Brezillon (5):
drm/panthor: Preserve the result returned by panthor_fw_resume()
drm/panthor: Be robust against runtime PM resume failures in the
suspend path
drm/panthor: Ignore devfreq_{suspend,resume}_device() failures
drm/panthor: Be robust against resume failures
drm/panthor: Fix the fast-reset logic
drivers/gpu/drm/panthor/panthor_devfreq.c | 12 ++--
drivers/gpu/drm/panthor/panthor_devfreq.h | 4 +-
drivers/gpu/drm/panthor/panthor_device.c | 78 +++++++++++------------
drivers/gpu/drm/panthor/panthor_device.h | 28 ++++++++
drivers/gpu/drm/panthor/panthor_drv.c | 2 +-
drivers/gpu/drm/panthor/panthor_fw.c | 62 +++++-------------
drivers/gpu/drm/panthor/panthor_gpu.c | 14 ++--
drivers/gpu/drm/panthor/panthor_mmu.c | 3 +-
drivers/gpu/drm/panthor/panthor_sched.c | 4 +-
9 files changed, 102 insertions(+), 105 deletions(-)
--
2.46.2
