You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After fixing the panic-on-load ( #163 ) I tried suspending. When I run acpiconf -s 3, the driver hangs in i915_wait_request -- for an infinite duration, because the function is being called with a timeout of NULL (by intel_engine_idle).
Changing i915_wait_request to treat a timeout of NULL as a timeout of 1 second prevents the hang but makes suspend fail (because the "are you ready for suspend" check is failing when the driver says "I failed to wait until the last request completed") but if I set error = 0 in i915_drm_suspend after the call to i915_gem_suspend then the suspend proceeds.
So... something seems to be broken in a way which makes the driver think that there's a request in progress when in fact there is no such request. It's possible that this is related to the problem in #163 since a nonexistent request not completing could easily be mistaken for a hung GPU...
After this, I can suspend and resume but get the "vertical stripes" syndrome mentioned in #149 .
The text was updated successfully, but these errors were encountered:
@lastewart has just posted a workaround in #149 that seems to work ... for at least two of us, anyway! I've just set the hw.acpi.lid_switch_state=S3 sysctl again after a long absence and life is just that little bit nicer. :)
After fixing the panic-on-load ( #163 ) I tried suspending. When I run
acpiconf -s 3
, the driver hangs ini915_wait_request
-- for an infinite duration, because the function is being called with a timeout of NULL (byintel_engine_idle
).Changing
i915_wait_request
to treat a timeout of NULL as a timeout of 1 second prevents the hang but makes suspend fail (because the "are you ready for suspend" check is failing when the driver says "I failed to wait until the last request completed") but if I seterror = 0
ini915_drm_suspend
after the call toi915_gem_suspend
then the suspend proceeds.So... something seems to be broken in a way which makes the driver think that there's a request in progress when in fact there is no such request. It's possible that this is related to the problem in #163 since a nonexistent request not completing could easily be mistaken for a hung GPU...
After this, I can suspend and resume but get the "vertical stripes" syndrome mentioned in #149 .
The text was updated successfully, but these errors were encountered: