-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
rhel-9.8
-
None
-
No
-
Moderate
-
1
-
rhel-kernel-rts-time
-
ssg_core_kernel
-
13
-
False
-
False
-
-
No
-
CK Parent Issues In Progress
-
None
-
None
-
Unspecified
-
Unspecified
-
Unspecified
-
-
All
-
None
The queue_track backend has been observed to experience stalling during a startup.
How to reproduce:
- Start an RT, high priority, CPU bound task on CPU 1: taskset -c 1 chrt -f 90 bash -c 'while :; do :; done' &
- Start a normal CPU bound task on CPU 1: taskset -c 1 bash -c 'while :; do :; done' &
- Start stalld on CPU 0: taskset -c 0 stalld -v -b queue_track
- Observe the stalld process hangs.
- Kill the RT process and stalld starts.
This is the stalld process kernel stack trace while it is hanging:
[root@dell-per450-03 test]# cat /proc/$(pgrep stalld)/stack [<0>] __wait_rcu_gp+0x116/0x120 [<0>] synchronize_rcu_tasks+0x4e/0xc0 [<0>] register_ftrace_direct_multi+0x183/0x260 [<0>] bpf_trampoline_update+0x33b/0x4e0 [<0>] __bpf_trampoline_link_prog+0xcc/0x1b0 [<0>] bpf_trampoline_link_prog+0x27/0x40 [<0>] bpf_tracing_prog_attach+0x312/0x500 [<0>] link_create+0x19f/0x280 [<0>] __sys_bpf+0x8d2/0xea0 [<0>] __x64_sys_bpf+0x1a/0x30 [<0>] do_syscall_64+0x5c/0xe0 [<0>] entry_SYSCALL_64_after_hwframe+0x76/0x7e
- clones
-
-
- In Progress
-