Comment 51 for bug 578506

Revision history for this message
rloth (rloth) wrote : Re: [Kernel] ACPI: EC: input buffer is not empty, aborting transaction

@Pedro Carrasco : perhaps your directory structure is not exactly what the patch expects.

You might use the alternate patch syntax:
    patch /pathto/originalfile /pathto/patchfile

--------------------------------------------------------------------
In my case i didn't apply the omnibook patch, because I have another laptop model : samsung r719

NB : Amazingly, when I boot in a (non patched) natty live cd the laptop works fine, and it also works in windows. But when I install natty the problem reappears *on the second boot*

So I used the ec.c patch from comment 66 in https://bugzilla.kernel.org/show_bug.cgi?id=14733
(the recommended one for 2.6.38) Unfortunately it didn't really work!

Sure, I don't get flooded with the EC input buffer message any more and at last I can boot, but many other things are broke:
  - unexpected remounts
  - xorg problems
  - keyboard malfunction
  - all similar issues to those that had been described for maverick in https://bugs.launchpad.net/ubuntu/+source/linux/+bug/610055

I'm wondering if these are separate problems that just suddenly appear because the system actually boots, or if they are connected to the original thing. I suspect the latter because i get similar behaviour with noacpi option at boot and non patched kernel.

Btw, I tried the patch successively on both the "generic-pae" and "generic" config, it's the same problem in both.