[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: HDD on Etrax100LX IDE developer board

On Wed, 12 Jun 2002, Sangmoon Kim wrote:
> I attatched the vmlinux and System.map again.
> I also attatched the Oops message for mke2fs /dev/hda1.
> I recompile the kernel and reproduced the message.

Great. This time they seem to match!

> It's weird that the SRP doesn't seem to point to some useful code.
> Thanks in advance.

Oh, SRP points to useful code now (the line after a jump subroutine call) 
but the instruction pointer, IRP, does not point to a location that could 
have caused the Oops :(. c0016410 is not preceded by any memory access at 
all actually... 

c001640a:       41c2                    moveq 1,$r12
c001640c:       649a                    move.d [$r4],$r9
c001640e:       e9c4                    lsl.d $r9,$r12
c0016410:       2f2e 0000 0040          add.d 40000000,$r2
c0016416:       ed23                    lsrq 13,$r2
c0016418:       c623                    lslq 6,$r2

Explanations could be that the code memory is overwritten or that the oops 
somehow is corrupted. What exact kernel version are you using, are you using 
the "official" one distributed with the developer board ? If this really 
is a bug somewhere it should have been noticed long ago in our own harddisk 
products, that is what is troubling me.