On SIGINFO, print a space-separated list or remaining processes to
standard error.
MFC after: 1 week
Sponsored by: Klara, Inc.
Sponsored by: NetApp, Inc.
Details
Details
- Reviewers
kib kevans markj - Group Reviewers
Klara - Commits
- rGeddd8aa99ca8: pwait: Add a SIGINFO handler
Diff Detail
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Skipped - Unit
Tests Skipped - Build Status
Buildable 74990 Build 71873: arc lint + arc unit
Event Timeline
| bin/pwait/pwait.c | ||
|---|---|---|
| 252 | This is unreliable. We activate the signal knotes before even checking for the fact that the signal could be ignored. This allows more than one signal events to be lumped together. I think that the existing mechanism is better. | |
| bin/pwait/pwait.c | ||
|---|---|---|
| 252 | So what does the event look like if two signals are delivered simultaneously? The exiting mechanism was unfixably unreliable because there is no pkevent() so it was possible for SIGINFO to be delivered outside the kevent() call and be missed until after the next one returned. | |
| bin/pwait/pwait.c | ||
|---|---|---|
| 252 | No, I was wrong. there is per-signal knotes. The knote list is shared with the EVFILT_PROC filter, but individual signals are counted separately. So for curproc, using the EVFILT_SIGNAL should be fine. | |