Bug #10708
closedEMG misses deregistration of pm_ops in remove
0%
Description
After successful probe of the platform_device, pm_ops are also registered. But after removing it, pm_ops are just suspended (in environment model). For races it means, that pm_ops thread is still active and the next probe operation will work in parallel with the first pm_ops. So, a normal deregistration of pm_ops (with pthread_join inside for races) is required.
Updated by Pavel Andrianov over 3 years ago
- Status changed from New to Resolved
After Ilya's fixes the problem disappeared. Mark as resolved. If Ilya says, that the fix is final, it should be closed.
Updated by Ilja Zakharov over 3 years ago
I will add more pydoc annotations to the code in the next few days without changing any functional code. So you can begin testing the branch before considering its merge to the master branch.
Updated by Evgeny Novikov over 3 years ago
- Status changed from Resolved to Open
- Priority changed from Normal to Urgent
- Target version set to 3.1
There is neither information on a branch where this bug was fixed nor information on a large improvement in progress. There is no much sense to test it prior to the end.
Updated by Ilja Zakharov over 3 years ago
- Status changed from Open to Resolved
Prepared the branch finally: emg-new-abstraction
Updated by Evgeny Novikov over 3 years ago
- Status changed from Resolved to Open
There are degradations shown by CI.
Updated by Evgeny Novikov over 3 years ago
- Status changed from Open to Resolved
Ilja seems to fix all regressions. Let's see.
Updated by Evgeny Novikov over 3 years ago
- Status changed from Resolved to Closed
After some more fixes, tests passed, so, I merged the branch to master in 0975f23d1.