On Mon, Jan 20, 2014 at 03:53:58PM -0800, Greg Kroah-Hartman wrote:

On Tue, Jan 21, 2014 at 12:26:52AM +0100, Jens Taprogge wrote:

quoted

Hello,

I have bisected the boot failure of a Lenovo Thinkpad X1 Carbon to the
merge ea4ebd1cb093c8ec5d7f4bf9070dc444184c3971.  What I call "boot
failure" refers to the fact that no console is shown, the machine only
reads from the disc for a very short time, and the machine does not
respond to Ctrl-Alt-Del, etc.

As it turns out reverting e696c68363740d3a8fe495e353de3ff26b86257a
(Input: serio - fix sysfs layout) on top of v3.13 makes the machine boot
again.  So far I have now clue how this commit can affect early boot
/ showing of the console.

Does it ring any bells for you?

No, not at all, I don't see how that patch could prevent booting,
especially as it is fixing a regression (i.e. putting things back the
way they originally were.)

very odd.

greg k-h

I agree.  This machine had some very weird boot issues a while ago.
I was starting to suspect it had some alignment issue of some vector
that the EFI firmware uses when handing over control to the kernel.
(Note that I know very little about EFI.)  Does that make any sense?

See this thread http://www.spinics.net/lists/stable/msg09886.html for
the old discussion.  Back then, the problem disappeared with later
version of the kernel / compiler.

Best Regards,
Jens