[ibm-acpi-devel] [ltp] Re: thinkpad-acpi release 0.17-20071002 uploaded to ibm-acpi.sf.net

Henrique de Moraes Holschuh linux-thinkpad@linux-thinkpad.org
Thu, 4 Oct 2007 00:33:57 -0300


On Wed, 03 Oct 2007, Chris Hanson wrote:
> Henrique de Moraes Holschuh wrote:
> > Try booting with ec_intr=0 in the kernel command line.  Who knows, it might
> > even help. I have seen weirder stuff...
> 
> No difference.

It was a shot in the dark, anyway. Well, let's keep trying.

> > Does the "printk on EINVAL" debug patch I sent ever triggers?  What values
> > it reports?
> 
> It never triggers.

I simply don't understand what is happening.  brightness_set can return
EINVAL only from that place, or if brightness_get is returning EINVAL.  And
brightness_get can't return EINVAL, just EIO.  The mutex_* primitives also
don't do EINVAL, just EINTR.

The sysfs handlers are all in the backlight class, so the bug might even be
there... but I can't reproduce it here, and we should be using very similar
kernels.  The procfs ones can return EINVAL by themselves, if you echo a
command it doesn't recognize, but that's it, and it wouldn't explain
problems in the sysfs handlers.

Now for the last resort questions: are you sure your thinkpad­acpi is free
from bad merges from patch, etc? what gcc are you using, and are you
optimizing the kernel for size (mine is optimized for size)?

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh