-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
rhel-10.0
-
None
-
No
-
None
-
rhel-kernel-rts-time
-
0
-
False
-
False
-
-
None
-
None
-
None
-
None
-
Unspecified
-
Unspecified
-
Unspecified
-
-
x86_64
-
None
When running stalld with the queue_track (bpf) backend in logging mode and turning off idle detection, stalld incorrectly reports stalled threads.
Removing --no_idle_detection from the command line prevents this behavior. Also it seems to only occurr with the queue_track backend and not the sched_debug backend
$ sudo stalld --verbose --foreground --backend=queue_track --no_idle_detection
single_threaded_main: checking cpu 23 - rt: 0 - starving: 1
stalld: cpu 0: pid: 8522 starving for 575
stalld: cpu 1: pid: 9113 starving for 480
stalld: cpu 2: pid: 8864 starving for 580
stalld: cpu 3: pid: 8832 starving for 470
stalld: cpu 4: pid: 8386 starving for 570
stalld: cpu 5: pid: 8502 starving for 380
stalld: cpu 6: pid: 8434 starving for 450
stalld: cpu 7: pid: 7018 starving for 470
stalld: cpu 8: pid: 7087 starving for 420
stalld: cpu 9: pid: 9820 starving for 510
stalld: cpu 10: pid: 8422 starving for 510
stalld: cpu 11: pid: 9420 starving for 490