Cancel a work not waiting for it to finish.
Sponsored by: Serenity Cyber Security, LLC
MFC after: 1 week
Details
Details
Diff Detail
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Passed - Unit
No Test Coverage - Build Status
Buildable 54667 Build 51556: arc lint + arc unit
Event Timeline
| sys/compat/linuxkpi/common/src/linux_work.c | ||
|---|---|---|
| 374 | Should this case be asserted? | |
| sys/compat/linuxkpi/common/src/linux_work.c | ||
|---|---|---|
| 374 | Yes please. I see this is copied from other functions, perhaps the to-state should be WORK_ST_INVALID, and linux_update_state() can assert that pstate[c] is never invalid. | |
| 384 | We do this non-atomically after the state has transitioned to idle. What if some other thread tries to queue work, so the state goes back from ST_IDLE to ST_TASK, but then queues work before the taskqueue_cancel() call? The state will be left at ST_TASK even though all work is cancelled. | |