Mateusz Guzik <mjg_at_FreeBSD.org> wrote on
Date: Sat, 04 Feb 2023 17:51:27 UTC :

> The branch main has been updated by mjg:
> 
> URL: 
> https://cgit.FreeBSD.org/src/commit/?id=d9d5f2c042a51a9f0dd69eb1fc349efd81ffa483
> 
> commit d9d5f2c042a51a9f0dd69eb1fc349efd81ffa483
> Author: Mateusz Guzik <m...@freebsd.org>
> AuthorDate: 2023-02-04 17:47:41 +0000
> Commit: Mateusz Guzik <m...@freebsd.org>
> CommitDate: 2023-02-04 17:50:41 +0000
> 
> cpuset: add --count
> 
> Can be used to count the number of hardware threads in the cpu set.
> 
> For example:
> $ cpuset -g -p $$
> pid 2440 mask: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
> 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34,
> 35, 36, 37, 38, 39
> pid 2440 domain policy: first-touch mask: 0, 1
> $ cpuset -g --count -p $$
> 40
> 
> The intent is to replace calls to sysctl hw.ncpu and kern.smp.cpus which
> can be found in the tree, which are not adequate given existence of
> cpusets.
> 
> Right now only -g -p combination is supported to reduce complexity.
> As anything else errors out, this can be expanded later as needed.
> 
> Differential Revision: https://reviews.freebsd.org/D36351

This broke long standing commands sequences in some makefiles
that I use to build and run some benchmarks, the commands
at issue involve use of cpuset.

However I can show the oddity in a much simpler context:

# cpuset echo "text"
text

# cpuset echo "-text"
usage: cpuset [-l cpu-list] [-n policy:domain-list] [-s setid] cmd ...
       cpuset [-l cpu-list] [-n policy:domain-list] [-s setid] -p pid
       cpuset [-c] [-l cpu-list] [-n policy:domain-list] -C -p pid
       cpuset [-c] [-l cpu-list] [-n policy:domain-list]
              [-j jailid | -p pid | -t tid | -s setid | -x irq]
       cpuset -g [-cir]
              [-d domain | -j jailid | -p pid | -t tid | -s setid | -x irq]
       cpuset -g --count -p pid


By contrast, in my stable/13 context, so, showing the old behavior:

# cpuset echo "-text"
-text


===
Mark Millard
marklmi at yahoo.com


Reply via email to