On Tue, 2 Sep 2014, Dylan Alex Simon wrote:
> Some combination of kernel debugging options and killing processes let it
> survive long enough to write the backtrace to disk. A simple modprobe/rmmod
> wasn't enough, though, it required a few tries removing the device and then
> rmmod (though has definitely happend on just one removal before). Let me know
> if there's anything else I can try.
>
> <insmod>
> [ 28.855960] thingm 0003:27B8:01ED.0004: hidraw3: USB HID v1.01 Device [ThingM blink(1) mk2] on usb-0000:00:12.2-3.1.4/input0
> <rmmod;insmod>
> [ 147.037008] thingm 0003:27B8:01ED.0004: hidraw3: USB HID v1.01 Device [ThingM blink(1) mk2] on usb-0000:00:12.2-3.1.4/input0
> <unplug>
> [ 218.496688] usb 1-3.1.4: USB disconnect, device number 7
> [ 218.502278] hid : failed to write color
> [ 218.506131] hid : failed to write color
> <plug>
> [ 233.557300] usb 1-3.1.4: new full-speed USB device number 8 using ehci-pci
> [ 233.657195] usb 1-3.1.4: config 1 interface 0 altsetting 0 has 2 endpoint descriptors, different from the interface descriptor's value: 1
> [ 233.660402] thingm 0003:27B8:01ED.0005: hidraw3: USB HID v1.01 Device [ThingM blink(1) mk2] on usb-0000:00:12.2-3.1.4/input0
> <rmmod>
> [ 253.682724] BUG: unable to handle kernel paging request at ffffffffa00af0cf
> [ 253.682807] IP:
> [ 253.682812] [<ffffffffa00af0cf>] 0xffffffffa00af0cf
> [ 253.682817] PGD 1814067 PUD 1815063 PMD 42cace067 PTE 0
> [ 253.682820] Oops: 0010 [#1] SMP
> [ 253.682830] Modules linked in: led_class cuse fuse snd_emu10k1 snd_hwdep snd_util_mem snd_ac97_codec ac97_bus snd_rawmidi snd_seq_device snd_pcm snd_timer ipt_ULOG [last unloaded: hid_thingm]
> [ 253.682833] CPU: 0 PID: 849 Comm: kworker/0:2 Not tainted 3.16.1-00001-g98fed6d #145
> [ 253.682835] Hardware name: empty empty/S8010-LE, BIOS 'V2.03B ' 03/15/2012
> [ 253.682838] Workqueue: events 0xffffffffa00af040
> [ 253.682840] task: ffff88042e330050 ti: ffff880429d8c000 task.ti: ffff880429d8c000
> [ 253.682844] RIP: 0010:[<ffffffffa00af0cf>] [<ffffffffa00af0cf>] 0xffffffffa00af0cf
> [ 253.682846] RSP: 0018:ffff880429d8fdd0 EFLAGS: 00010286
> [ 253.682847] RAX: 0000000000000009 RBX: ffff88042ca83af0 RCX: 0000000000000302
> [ 253.682849] RDX: 0000000000000078 RSI: 0000000000000286 RDI: ffff88042caaade0
> [ 253.682850] RBP: ffff880429d8fdf0 R08: ffff8804acaaade0 R09: 0000000000000282
> [ 253.682852] R10: ffff88042c93dbc0 R11: 000000000000001f R12: ffff88042c885e80
> [ 253.682853] R13: 0000000000000000 R14: ffff88043ec14e00 R15: ffff88043ec113c0
> [ 253.682856] FS: 00007f58ebcec700(0000) GS:ffff88043ec00000(0000) knlGS:0000000000000000
> [ 253.682857] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
> [ 253.682859] CR2: ffffffffa00af0cf CR3: 0000000001813000 CR4: 00000000000407f0
> [ 253.682859] Stack:
> [ 253.682863] 010000000000a000 0001000000000063 000000008867f918 ffff88042ca83af0
> [ 253.682866] ffff880429d8fe38 ffffffff81052c2f ffff88043ec113c0 000000003ec113c0
> [ 253.682869] ffff88043ec113c0 ffff88043ec113e8 ffff88042e330050 ffff88042c885eb0
> [ 253.682870] Call Trace:
> [ 253.682878] [<ffffffff81052c2f>] process_one_work+0x14f/0x400
> [ 253.682882] [<ffffffff81053423>] worker_thread+0x63/0x540
> [ 253.682886] [<ffffffff810533c0>] ? create_and_start_worker+0x60/0x60
> [ 253.682889] [<ffffffff81059038>] kthread+0xe8/0x100
> [ 253.682893] [<ffffffff81058f50>] ? kthread_create_on_node+0x1b0/0x1b0
> [ 253.682897] [<ffffffff815323ec>] ret_from_fork+0x7c/0xb0
> [ 253.682900] [<ffffffff81058f50>] ? kthread_create_on_node+0x1b0/0x1b0
Alright, this supports my original hunch. I think I see the race. Could
you please try the patch below? Thanks.
diff --git a/drivers/hid/hid-thingm.c b/drivers/hid/hid-thingm.c
index 134be89..743a517 100644
--- a/drivers/hid/hid-thingm.c
+++ b/drivers/hid/hid-thingm.c
@@ -208,7 +208,7 @@ unregister_red:
static void thingm_remove_rgb(struct thingm_rgb *rgb)
{
- flush_work(&rgb->work);
+ cancel_work_sync(&rgb->work);
led_classdev_unregister(&rgb->red.ldev);
led_classdev_unregister(&rgb->green.ldev);
led_classdev_unregister(&rgb->blue.ldev);
--
Jiri Kosina
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
