On Sat, May 10, 2008 at 10:22:10PM +0200, Fabian Off wrote:
> Am Samstag, 10. Mai 2008 19:19:34 schrieb Kevin O'Connor:
> > Am Dienstag, 6. Mai 2008 19:47:01 schrieb Marc Poulhiès:
> > > Jumping to Kernel...
> >
> > This generally indicates a problem with the target kernel.  The
> > messages indicate that haret successfully loaded the kernel and jumped
> > to it.
> >
> > -Kevin
> Well, could you tell us how to debug this? We (or better: I) know, the kernel 
> I used works very well w/ my iPaq when flashed directly. Thus, there 
> shouldn't be a problem with the kernel. Is there anything we can do to debug 
> this further?

Hi Fabian,

After the "Jumping to Kernel...", the Linux kernel has control.  So,
one needs to debug the kernel.

The easiest way to do this is if you can hook up a serial console.
(Does your device have serial?)  With a serial enabled kernel, you
should be able to find where in the kernel it is halting.

You should also double check that the right mtype is in use.  Haret is
autodetecting h5000 and is therefore using mtype "H5400" (220).  The
kernel must also be compiled to use this mtype.

You can also try enabling "set kernelcrc 1" before booting in haret.
This will cause haret to do a crc check on the copied files just prior
to jumping to the kernel - any corruption that haret may have
introduced should be detected by this.

Also, it is possible that wince is putting the hardware in a state
that your kernel just can't cope with.  If this is the case, the only
way to work through it is to debug the kernel until you find where it
is failing.

Hope this helps,
Haret mailing list

Reply via email to