[LTP] [PATCH] numa01.sh: unset default cpuset awareness before verify physcpubind

Shizhao Chen shichen@redhat.com
Mon Sep 11 13:22:27 CEST 2023


On systems with cpu isolation configured (a very common technique used in the
real time kernel), $run_on_cpu may happen to be an isolated core and hence
unaccessible to numactl.

Unset default cpuset awareness with --all, so numactl can use all possible
CPUs/nodes for following physcpubind settings.

Signed-off-by: Shizhao Chen <shichen@redhat.com>
---
 testcases/kernel/numa/numa01.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/testcases/kernel/numa/numa01.sh b/testcases/kernel/numa/numa01.sh
index 367813ce9..0181cd8f5 100755
--- a/testcases/kernel/numa/numa01.sh
+++ b/testcases/kernel/numa/numa01.sh
@@ -183,7 +183,7 @@ test4()
     no_of_cpus=$(tst_ncpus)
     # not sure whether cpu's can't be in odd number
     run_on_cpu=$(($((no_of_cpus+1))/2))
-    numactl --physcpubind=$run_on_cpu support_numa pause & #just
waits for sigint
+    numactl --all --physcpubind=$run_on_cpu support_numa pause &
#just waits for sigint
     pid=$!
     var=`awk '{ print $2 }' /proc/$pid/stat`
     while [ $var = '(numactl)' ]; do
-- 
2.41.0



More information about the ltp mailing list