Re: [PATCH] man/man2/sched_setattr.2: update `EINVAL` case with `.sched_flags`

[Alejandro Colomar <alx+linux-man@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2291 bytes --]

Hi Enji,

> Date: 2026-10-23 16:56:51-0700
> From: "Enji Cooper (yaneurabeya)" <yaneurabeya@gmail.com>
>
> Hi Alex!
> 
> > On Oct 23, 2026, at 12:37 PM, Alejandro Colomar <alx+linux-man@kernel.org> wrote:
> > 
> > Hi Enji,
> > 
> >> Date: 2026-10-23 11:58:57-0700
> >> From: "Enji Cooper (yaneurabeya)" <yaneurabeya@gmail.com>
> >> 
> >> In more recent versions of the kernel, `.sched_flags` is a treated as a
> >> bitmask that supports values like `SCHED_FLAG_UTIL_CLAMP_MIN`, etc.
> >> 
> >> Account for the additional flags so the documentation is up to date with
> >> the kernel implementation of `sched_setattr`.
> >> 
> >> Ref: linux.git 480a6ca2dc6e ("sched/uclamp: Allow to reset a task uclamp constraint value")
> >> 
> >> Signed-off-by: Enji Cooper <yaneurabeya@gmail.com>
> >> ---
> >> man/man2/sched_setattr.2 | 5 ++++-
> >> 1 file changed, 4 insertions(+), 1 deletion(-)
> >> 
> >> diff --git a/man/man2/sched_setattr.2 b/man/man2/sched_setattr.2
> >> index 80a0ac726dcf..64acdd88884f 100644
> >> --- a/man/man2/sched_setattr.2
> >> +++ b/man/man2/sched_setattr.2
> >> @@ -417,7 +417,10 @@ .SH ERRORS
> >> .B EINVAL
> >> .I attr.sched_flags
> >> contains a flag other than
> >> -.BR SCHED_FLAG_RESET_ON_FORK .
> >> +.BR SCHED_FLAG_RESET_ON_FORK ,
> >> +.BR SCHED_FLAG_UTIL_CLAMP_MIN ,
> >> +or
> >> +.BR SCHED_FLAG_UTIL_CLAMP_MAX .
> > 
> > Why does this list not include flags such as SCHED_FLAG_RESET_ON_FORK?
> > It is documented in the same manual page, so I expect it's valid here,
> > right?
> 
> I actually added the `SCHED_FLAG_UTIL_CLAMP_MIN` and
> `SCHED_FLAG_UTIL_CLAMP_MAX` flags to the list of valid flags, in
> addition to `SCHED_FLAG_RESET_ON_FORK`.

D'oh!  I was really confused yesterday!  :D
I meant to write SCHED_FLAG_RECLAIM, but then I pasted the wrong one.

> I’m kind of on the the fence with how I worded it though — it’s kind
> of wordy and not really futureproof. I’ll take a look at some of the
> other manpages to see if there’s a better way to word it.

Yes, this is what I meant in the end.  We could reword it to say:

	Invalid flags specified in
	.IR attr.sched_flags .


Have a lovely day!
Alex

> 
> Cheers!
> -Enji



-- 
<https://www.alejandro-colomar.es>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]