Bug 3278 - hrtimers cause kernel crash
Summary: hrtimers cause kernel crash
Status: RESOLVED FIXED
Alias: None
Product: systemtap
Classification: Unclassified
Component: runtime (show other bugs)
Version: unspecified
: P1 critical
Target Milestone: ---
Assignee: Josh Stone
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-09-28 16:13 UTC by Martin Hunt
Modified: 2006-09-28 17:51 UTC (History)
0 users

See Also:
Host:
Target:
Build:
Last reconfirmed:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Hunt 2006-09-28 16:13:27 UTC
This is new.  I run tests daily and didn't see this until today.

in testsuite/systemtap.samples,
> stap poll_map.stp

It will run OK, then after 30 seconds attempt to exit and crash the kernel.

OOPS on x86_64 xen guest:
BUG: soft lockup detected on CPU#0!

Call Trace:
 <IRQ> [<ffffffff802ad0fe>] softlockup_tick+0xdb/0xf6
 [<ffffffff8026d3a7>] timer_interrupt+0x38a/0x3d8
 [<ffffffff80210e82>] handle_IRQ_event+0x2d/0x60
 [<ffffffff802ad447>] __do_IRQ+0xa5/0x107
 [<ffffffff8026b2df>] do_IRQ+0xe7/0xf5
 [<ffffffff80394454>] evtchn_do_upcall+0x86/0xe0
 [<ffffffff8822eb28>] :stap_1825:enter_probe_1446+0x0/0x5f5
 [<ffffffff8025e1a6>] do_hypervisor_callback+0x1e/0x2c
 [<ffffffff8822eb28>] :stap_1825:enter_probe_1446+0x0/0x5f5
 [<ffffffff8020622a>] hypercall_page+0x22a/0x1000
 [<ffffffff8020622a>] hypercall_page+0x22a/0x1000
 [<ffffffff80393bb7>] force_evtchn_callback+0xa/0xb
 [<ffffffff8024ee85>] hrtimer_run_queues+0x109/0x176
 [<ffffffff80290645>] run_timer_softirq+0x23/0x1be
 [<ffffffff802122de>] __do_softirq+0x62/0xdd
 [<ffffffff8025e674>] call_softirq+0x1c/0x28
 [<ffffffff8026b469>] do_softirq+0x31/0x98
 [<ffffffff8026b2e4>] do_IRQ+0xec/0xf5
 [<ffffffff80394454>] evtchn_do_upcall+0x86/0xe0
 [<ffffffff8025e1a6>] do_hypervisor_callback+0x1e/0x2c

OOP on i686 smp:

Sep 28 11:40:16 dragon kernel: BUG: soft lockup detected on CPU#1!
Sep 28 11:40:16 dragon kernel:  <c044a94a> softlockup_tick+0xad/0xc4  <c042d858>
update_process_times+0x39/0x5c
Sep 28 11:40:16 dragon kernel:  <c0418af3> smp_apic_timer_interrupt+0x5a/0x63 
<f8e69fbf> enter_probe_1483+0x0/0x66a [stap_16376]
Sep 28 11:40:16 dragon kernel:  <c040490f> apic_timer_interrupt+0x1f/0x24 
<f8e69fbf> enter_probe_1483+0x0/0x66a [stap_16376]
Sep 28 11:40:16 dragon kernel:  <c04390f2> hrtimer_run_queues+0xed/0x143 
<c042d6bd> run_timer_softirq+0x25/0x187
Sep 28 11:40:16 dragon kernel:  <c0434225> __rcu_process_callbacks+0xd9/0x152 
<f8e6bcaf> __stp_time_timer_callback+0x0/0x95 [stap_16376]
Sep 28 11:40:16 dragon kernel:  <c0429dd9> __do_softirq+0x58/0xc2  <c04064e9>
do_softirq+0x46/0x51
Sep 28 11:40:16 dragon kernel:  =======================
Sep 28 11:40:16 dragon kernel:  <c040490f> apic_timer_interrupt+0x1f/0x24 
<c0402fe3> default_idle+0x0/0x55Sep 28 11:40:16 dragon kernel:  <c040300f>
default_idle+0x2c/0x55  <c04030c7> cpu_idle+0x8f/0xa8
Comment 1 Josh Stone 2006-09-28 16:16:31 UTC
I just made some changes with hrtimers, so I will investigate this...
Comment 2 Josh Stone 2006-09-28 17:51:12 UTC
Fixed in tapsets.cxx r1.146