Date: Tue, 12 Apr 2022 22:10:57 +0530

The perf benchmark for collections: numa hits failure =
in system

configuration with CPU's more than 1024. These benchmar=
ks uses

"sched_getaffinity" and "sched_setaffinity" in the code t=
o

work with affinity.

Example snippet fr=
om numa benchmark:

<<>>

perf: bench/numa.c:=
302: bind_to_node: Assertion `!(ret)' failed.

Aborted (core dumpe=
d)

<<>>

bind_to_node functio=
n uses "sched_getaffinity" to save the cpumask.

This fails with E=
INVAL because the default mask size in glibc is 1024

To overcome this 1024 CPUs mask size limitation of cpu_set_t,

=
change the mask size using the CPU_*_S macros ie, use CPU_ALLOC to

allocate cpumask, CPU_ALLOC_SIZE for size, CPU_SET_S to set mask bit.

Fix all the relevant places in the code to use mask s=
ize which is large

enough to represent number of possible CPU's i=
n the system.

This patchset also address a fix for=
 parse_setup_cpu_list function in

numa bench to check if input CP=
U is online before binding task to

that CPU. This is to fix failu=
res where, though CPU number is within

max CPU, it could happen t=
hat CPU is offline. Here, sched_setaffinity

will result in failur=
e when using cpumask having that cpu bit set

in the mask.

Patch 1 address fix for parse_setup_cpu_list to check if =
CPU used to bind

task is online. Patch 2 has fix for bench numa t=
o work with machines

having #CPUs > 1K

Athira Rajeev (2):

  tools/perf: Fix perf bench numa testcase t=
o check if CPU used to bind

    task is online

  perf b=
ench: Fix numa bench to fix usage of affinity for machines with

 =
   #CPUs > 1K

Changelog:

v2 -> v3<=
/pre>Link to the v2 version :
https://lore.=
kernel.org/all/20220406175113.87881-1-atrajeev@linux.vnet.ibm.com/

 - From the v2 version, patch 1 and patch 2 are now pa=
rt of upstream.
 - This v3 version separates patch 3 and patch 4 =
to address review
   comments from arnaldo which includes using s=
ysfs__read_str for reading
   sysfs file and fixing the compilati=
on issues observed in debian

Tesed the patches on powerpc with CPU =
> 1K and other configurations as well, verified the perf bench numa with=
 the patch set.Tested-by: Disha Goel <disgoel@linux.vnet.ibm.com>

 tools/perf/bench/numa.c  | 136 +++++++++++=
++++++++++++++++++----------
 tools/perf/util/header.c |  51 ++++=
+++++++++++
 tools/perf/util/header.h |   1 +
 3 files =
changed, 153 insertions(+), 35 deletions(-)



--=-L2vVa6joYbbKvsUwutrU--