[Bug 222094] New: sched_setattr: bug in `ERRORS` section w.r.t. `attr.sched_flags contains a flag other than SCHED_FLAG_RESET_ON_FORK`

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=222094

            Bug ID: 222094
           Summary: sched_setattr: bug in `ERRORS` section w.r.t.
                    `attr.sched_flags contains a flag other than
                    SCHED_FLAG_RESET_ON_FORK`
           Product: Documentation
           Version: unspecified
          Hardware: All
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: man-pages
          Assignee: documentation_man-pages@kernel-bugs.osdl.org
          Reporter: yaneurabeya@gmail.com
        Regression: No

Per https://www.man7.org/linux/man-pages/man2/sched_getattr.2.html:

This scenario documented in the manpage implies that `attr.sched_flags` must be
set to `SCHED_FLAG_RESET_ON_FORK`:

>       EINVAL attr.sched_flags contains a flag other than
>       SCHED_FLAG_RESET_ON_FORK.

The second scenario implies that `attr.sched_flags` is a bitmask which can
contain `SCHED_FLAG_UTIL_CLAMP_MIN` or `SCHED_FLAG_UTIL_CLAMP_MAX`:

>       EINVAL attr.sched_flags contains SCHED_FLAG_UTIL_CLAMP_MIN or
>       SCHED_FLAG_UTIL_CLAMP_MAX, and attr.sched_util_min or
>       attr.sched_util_max are out of bounds.

The first case contradicts the second case.

Doing some digging in the kernel source code [1], I couldn't find an obvious
spot where the first scenario was actually implemented -> **I think the
`SCHED_FLAG_RESET_ON_FORK ` scenario is a documentation bug.**

1.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/kernel/sched/syscalls.c?id=786244f70322e41c937e69f0f935bfd11a9611bf

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.