Re: Kernel mode prefetch abort with -current on J728

Alex Poylisher wrote:
>On Sat, May 20, 2006 at 02:58:15PM +0100, Robert Swindells wrote:
>> 
>> Alex Poylisher wrote:
>> >Have seen this on boot with every kernel since 3.99.19:
>> >
>> >wdc0 at pcmcia1 function 0: <CF Card, CFA 02GB CKS>
>> >wdc0: i/o mapped mode
>> >atabus0 at wdc0 channel 0
>> >wd0 at atabus0 drive0Fatal kernel prefetch abort at 0xc09bdb64
>> >trapframe: 0xc2960e74, spsr=a0000013
>> >r0=c0d96800, r1=000001ae, r2=00000000, r3=20000093
>> >r4=0000000b, r5=c20ecd14, r6=c0d95d00, r7=c0265258
>> >r8=00000002, r9=00000002, r10=c2055000, r11=c2960f58
>> >r12=00000000, ssp=c2960ec0, sir=c0d970f0, pc=c09bdb64
>> >
>> >Stopped at      0xcc09bdb64:    address 0xc09bdb64 is invalid
>> >andgt   r5, r5, #0x00000000
>> >db>
>> 
>> Can you get a stack trace at this point ? Type 'bt' at the prompt.

>Here it is:

>db> bt
>0xc01456b0
>	scp=0xc01456b0 rlv=0xc014513c (0xc014513c)
>	rsp=0xc2960f5c rfp=0xc2960f90
>	r10=0x00000000 r9=0xc02288bc
>	r8=0xc0da4c00 r7=0x00000000 r6=0xc20e5680 r5=0x00000000
>	r4=0xc20ecd14
>0xc0144db0
>	scp=0xc0144db0 rlv=0xc01d6ec8 (netbsd:apm_thread+0x78)
>	rsp=0xc2960f94 rfp=0xc2960fb0
>	r10=0xc02628a0 r9=0x00000000
>	r8=0xc0264980 r7=0x00000000 r6=0xc0236a14 r5=0xc0da4cc68
>	r4=0xc0da4c00
>netbsd:apm_thread+0xc
>	scp=0xc01d6e5c rlv=0xc01b7294 (netbsd:proc_trampoline+0x18)
>	rsp=0xc2960fb4 rfp=0xc0011e18
>	r6=0x0000000 r5=0xc0da4c00
>	r4=0xc01d6e50
>Bad frame pointer: 0xc0011e18
>db>

You could try taking out power management while we track down the
problem.

Comment out the following lines in your config and rebuild the
kernel:

# Power management
j720pwr*	at j720ssp?
hpcapm0		at j720pwr?		# power management
apmdev0		at hpcapm0		# APM