This is the mail archive of the systemtap@sourceware.org mailing list for the systemtap project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug runtime/20148] during module init, systemtap is calling cpufreq_register_notifier(), a function that can sleep, in atomic context


https://sourceware.org/bugzilla/show_bug.cgi?id=20148

David Smith <dsmith at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|kernel BUGs while doing     |during module init,
                   |parallel testing            |systemtap is calling
                   |                            |cpufreq_register_notifier()
                   |                            |, a function that can
                   |                            |sleep, in atomic context

--- Comment #1 from David Smith <dsmith at redhat dot com> ---
OK, I see at least one problem here. I've cut down the log to make it more
visible:

====
May 25 16:03:44 kernel: BUG: sleeping function called from invalid context at
kernel/mutex.c:576
May 25 16:03:44 kernel: in_atomic(): 1, irqs_disabled(): 0, pid: 27280, name:
stapio
May 25 16:03:44 kernel: Call Trace:
May 25 16:03:44 kernel: [c000000183ac7b50] [c0000000007c5240]
.cpufreq_register_notifier+0xa0/0xc0
May 25 16:03:44 kernel: [c000000183ac7bc0] [d0000000020a8c54]
.systemtap_module_init+0x394/0xad0 [s_27280]
====


systemtap_module_init() calls _stp_init_time(), which calls
cpufreq_register_notifier(). However, cpufreq_register_notifier() can sleep. At
the point where it was called, we were in atomic context.

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

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]