linux-kernel - Re: LED (leds_hp_disk) causes BUG: scheduling while atomic

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <1231401855.8422.15.camel@maya.local>
Date:	Thu, 08 Jan 2009 09:04:15 +0100
From:	Daniel Troeder <daniel@...in-box.com>
To:	Pavel Machek <pavel@...e.cz>
Cc:	deller@....de, kernel list <linux-kernel@...r.kernel.org>
Subject: Re: LED (leds_hp_disk) causes BUG: scheduling while atomic

Am Dienstag, den 06.01.2009, 11:14 +0100 schrieb Pavel Machek:
> > I have a HP Compaq nc2400 notebook, and having recently installed 2.6.28
> > (gentoo patchset, self configured), I have observed the following:
> > 
> > When {I manually, udev automatically} load the leds_hp_disk module, my
> > disk-led flashes regularly _without_ disk activity. When I look at my
> > dmesg I see the following:
> 
> Yes, that's a (mis-feature).
> 
> > [  724.376881] BUG: scheduling while atomic: swapper/0/0x10000100
> > [  724.376886] Modules linked in: leds_hp_disk i915 drm snd_pcm_oss
> 
> And that's a bug. I know what causes it, but i don't have right fix
> yet.
> 
> If you are interested, take latest -mm (mmotm from userweb) and apply
> this manually:
> 
> --- linux-mm/drivers/hwmon/hp_accel.c   2009-01-06 10:54:40.000000000
> +0100
> +++ linux/drivers/hwmon/hp_accel.c      2008-12-10 09:36:21.000000000
> +0100
> @@ -174,11 +174,35 @@
>  }
>  
>  static struct led_classdev hpled_led = {
> -       .name                   = "hp:red:hddprotection",
> -       .default_trigger        = "heartbeat",
> +       .name                   = "hp:red:hddprotect",
> +       .default_trigger        = "none",
>         .brightness_set         = hpled_set,
>  };
> 
> 
> ...led will stop blinking, and messages will be gone. (But note that
> if you turn on heartbeat or similar trigger, you'll still get the
> "scheduling in interrupt" nastyness...)
> 							Pavel
Hello :)

Sorry it took me so long to reply - I built a kernel based on
vanilla-2.6.28 and applied patches with quilt to .DATE=2009-01-05-12-50.
That worked well, except that when booted, the xfs driver corrupts each
time my MUAs (evolution) configuration :(

But not your problem... the LED-scheduling-problem went away (no more
blinking, on more kernel-BUG and I get:

[   31.822317] lis3lv02d: laptop model unknown, using default axes
configuration
[   31.822367] Registered led device: hp:red:hddprotect
[   32.059496] input: ST LIS3LV02DL Accelerometer
as /devices/platform/lis3lv02d/input/input8
[   32.081954] lis3lv02d driver loaded.

$ lsmod | egrep 'lis3lv02d|hp_accel'
hp_accel                6656  0 
led_class               3908  2 iwl3945,hp_accel
lis3lv02d               5956  1 hp_accel

/sys/devices/platform/lis3lv02d
/sys/bus/platform/devices/lis3lv02d
/sys/bus/acpi/drivers/lis3lv02d
/sys/module/lis3lv02d
/sys/module/hp_accel

I was shortly confused, because CONFIG_LEDS_HP_DISK went away and got
integrated into CONFIG_SENSORS_LIS3LV02D with
lis3lv02d-merge-with-leds-hp-disk.patch :)

Bye,
Daniel

Download attachment "signature.asc" of type "application/pgp-signature" (198 bytes)