Prev: sound/pci: Eliminate use after free
Next: fsck more often when powerfail is detected (was Re: wishful thinking about atomic, multi-sector or full MD stripe width, writes in storage)
From: Frederic Weisbecker on 4 Apr 2010 08:20
Hi,

On tip:master, while turning on lock events with perf through

perf lock record

I get random kernel crashes, sometimes about weird unaligned
accesses, sometimes about scheduler that complains.

I hope someone has an idea about this.

In three different attempts (got to force reboot each time), I got:

First (task_tgid_nr_ns() is called from perf_event_pid()):

[ 565.464201] Kernel unaligned access at TPC[486b74] task_tgid_nr_ns+0x8/0x54
[ 565.475801] sun4v_data_access_exception: ADDR[000060f8b13a2004] CTX[0000] TYPE[0009], going.
[ 565.488610] \|/ ____ \|/
1>[ 565.492705] Unable to handle kernel NULL pointer dereference
[ 565.492719] Unable to handle kernel NULL pointer dereference
[ 565.492733] Unable to handle kernel NULL pointer dereference
[ 565.492747] Unable to handle kernel NULL pointer dereference
[ 565.492761] Unable to handle kernel NULL pointer dereference
[ 565.492776] Unable to handle kernel NULL pointer dereference
[ 565.492790] Unable to handle kernel NULL pointer dereference
[ 565.492804] Unable to handle kernel NULL pointer dereference
[ 565.492818] Unable to handle kernel NULL pointer dereference
[ 565.492832] Unable to handle kernel NULL pointer dereference
[ 565.492847] Unable to handle kernel NULL pointer dereference

Second:

[ 250.508047] Kernel unaligned access at TPC[4d1a0c] perf_swevent_ctx_event+0x16c/0x1b0

(this one happened in asm/processor_64.h: prefetch(), probably while
walking to the context's event list)

Third:

[ 60.147895] e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX/TX
[ 130.637924] sun4v_data_access_exception: ADDR[ffffa010933a6060] CTX[0000] TYPE[0009], going.
[ 130.639329] kernel BUG at kernel/sched.c:1166!
[ 130.639364] \|/ ____ \|/
[ 130.639370] "@'/ .. \`@"
[ 130.639377] /_| \__/ |_\
[ 130.639382] \__U_/
[ 130.639394] swapper(0): Kernel bad sw trap 5 [#1]
[ 130.639430] TSTATE: 0000000080e01605 TPC: 000000000045a02c TNPC: 000000000045a030 Y: 00000000 Tainted: G
W
[ 130.639462] TPC: <resched_task+0x44/0xa8>
[ 130.639475] g0: fffff803f685d980 g1: 0000000000000000 g2: 0000000000000027 g3: 0000000000007d0d
[ 130.639493] g4: fffff803f685d980 g5: fffff800160a0000 g6: fffff803f6864000 g7: 0000000000b54c00
[ 130.639511] o0: 0000000000828580 o1: 000000000000048e o2: 0000000000000000 o3: 0000000000000000
[ 130.639528] o4: 0000000000000002 o5: 0000000000000001 sp: fffff803ff932ff1 ret_pc: 000000000045a024
[ 130.639548] RPC: <resched_task+0x3c/0xa8>
[ 130.639561] l0: 0000000000000000 l1: 000000000000000e l2: ffffffffffffffff l3: 0000000000000104
[ 130.639580] l4: fffff803f685d980 l5: 0006000000000000 l6: 000000000000000e l7: 00000000008a4180
[ 130.639597] i0: fffff803f62c6960 i1: fffff803f7fa18e0 i2: 0000000000000001 i3: 0000000000000000
[ 130.639616] i4: 0000000000b28340 i5: 0000000000b25cc0 i6: fffff803ff9330b1 i7: 0000000000461bc8
[ 130.639644] I7: <check_preempt_wakeup+0x148/0x1f8>
[ 130.639655] Instruction DUMP: 9210248e 7fff3f4d 90122180 <91d02005> c25a6008 82086008 0ac84014 92026008
4005c092
[ 130.639703] Kernel panic - not syncing: Aiee, killing interrupt handler!
[ 130.639715] Call Trace:
[ 130.639742] [000000000073795c] panic+0x58/0x124
[ 130.639763] [000000000046fb80] do_exit+0x64/0x770
[ 130.639782] [0000000000427d4c] die_if_kernel+0x264/0x290
[ 130.639801] [000000000042a138] bad_trap+0x78/0xe8
[ 130.639824] [00000000004220b0] tl0_resv104+0x30/0xa0
[ 130.639841] [000000000045a02c] resched_task+0x44/0xa8
[ 130.639861] [0000000000461bc8] check_preempt_wakeup+0x148/0x1f8
[ 130.639883] [0000000000466554] try_to_wake_up+0x484/0x570
[ 130.639902] [000000000046668c] wake_up_process+0xc/0x20
[ 130.639920] [00000000004678b4] load_balance+0xfb4/0x10f0
[ 130.639940] [0000000000467b60] rebalance_domains+0x170/0x204
[ 130.639960] [0000000000467c30] run_rebalance_domains+0x3c/0x100
[ 130.639985] [00000000004734a4] __do_softirq+0x1b8/0x378
[ 130.640004] [000000000042a354] do_softirq+0x8c/0xcc
[ 130.640021] [0000000000472ec0] irq_exit+0x68/0xd0
[ 130.640044] [000000000042f0f8] timer_interrupt+0xb8/0xec

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 | 
Pages: 1
Prev: sound/pci: Eliminate use after free
Next: fsck more often when powerfail is detected (was Re: wishful thinking about atomic, multi-sector or full MD stripe width, writes in storage)