[Bug runtime/31472] systemtap.base/tracepoints.stp causes reboot with RHEL9 linux 5.14.0-428.el9.x86_64

["wcohen at redhat dot com" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=31472

--- Comment #3 from William Cohen <wcohen at redhat dot com> ---
Created attachment 15400
  --> https://sourceware.org/bugzilla/attachment.cgi?id=15400&action=edit
smaller reproducer using just mmap_lock:* tracepoints for the reboot

Took a look at the difference between the RHEL9 427 and 428 kernel tracepoints.
 Suspected there was a set of tracepoints that were triggering the reboot in
the RHEL9 428 kernel.  With a pared down tracepoints.stp check which wildcard
tracepoints were causing the problem.  The following three tracepoints are
causing the reboot on RHEL9 428 kernel:

kernel.trace("mmap_lock:mmap_lock_acquire_returned")
kernel.trace("mmap_lock:mmap_lock_released")
kernel.trace("mmap_lock:mmap_lock_start_locking")

Check f39 and fedora rawhide and they both have those tracepoints. Attempted to
fedora 39 kernel-6.7.9-200.fc39.x86_64 and the reproducer ran with out issue.
The reproduce also ran without issue on rawhide with
kernel-6.8.0-0.rc7.20240308git3aaa8ce7a335.60.fc41.x86_64

Looks like the next step is to compare the difference between RHEL9 and
f39/rawhide.

-- 
You are receiving this mail because:
You are the assignee for the bug.