This fixes tcflush(3), unbreaking cheribuild.py under arm64 Focal.
Details
Details
- Reviewers
imp - Group Reviewers
Linux Emulation - Commits
- rG1f86d04d02a6: linux: Partially implement TCSBRK
rG7e7859e7c2b9: linux: Partially implement TCSBRK
Diff Detail
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
| sys/compat/linux/linux_ioctl.c | ||
|---|---|---|
| 763 | arg == 0 means break, should this not be ioctl TCSBRK (arg=0) not implemented or something like that? | |
| sys/compat/linux/linux_ioctl.c | ||
|---|---|---|
| 763 | I think so too... I noticed that. The implementation, though, wouldn't be hard, since it's just a ioctl(TIOCSBRK), pause_sbt(0.4s), ioctl(TIOCCBRK) | |
| sys/compat/linux/linux_ioctl.c | ||
|---|---|---|
| 763 | My thinking was, this is additional information that doesn't help the user in any way. The point of this message is to spot the problem and to be able to grep for it, not to explain the actual situation. And yes, implementation is trivial; it's the real-world testcase I don't have. From my experience it's better to leave a (loud) stub than commit simple, but untested code. | |
| sys/compat/linux/linux_ioctl.c | ||
|---|---|---|
| 763 | Sigh, now I see what you mean - I thought it was about the additional arg, not the wrong ioctl name. | |
Comment Actions
Minor wording tweak, but now it's fine.
| sys/compat/linux/linux_ioctl.c | ||
|---|---|---|
| 763 | s/arg 0/arg == 0/ here for clarity. | |
| sys/compat/linux/linux_ioctl.c | ||
|---|---|---|
| 763 | I think I prefer the current version, due to consistency of how the rest of Linuxulator uses it. | |