8.0-RC2 on Rackmac3,1 - fatal kernel trap

[Nathan Whitehorn]

I just took a look at the data layout in the kernel, and translations 
immediately precedes ofwcall, so there is no question that the buffer 
size has been exceeded and that the zeroing the translations buffer has 
killed ofwcall. I'll send you a new CD with a larger buffer (and bounds 
checking) sometime at the beginning of next week. Hopefully, the change 
will also be merged in 8.0-RELEASE.
-Nathan

Nathan Whitehorn wrote:
> This is the middle of reading out the current MMU translations from 
> Open Firmware (the OF_getprop() call at line 902 of mmu_oea64.c). It's 
> possible the size of the translations buffer has been exceeded, and 
> memory has been trashed, which would also explain why there is a 
> difference between CD and netbooting. Peter, any ideas?
> -Nathan
>
> Jacob Lambert wrote:
>> Here's a transcript with 'show registers' and 'bt' after trying to
>> boot your 9.0-20091106-SNAP-powerpc-bootonly.iso.
>>
>> ---
>> 0 > boot cd:,\boot\loader cd:0 load-size=3994c adler32=a40c7b0c
>>
>> Loading ELF
>> Consoles: Open Firmware console
>>
>> FreeBSD/powerpc Open Firmware loader, Revision 0.1
>> (root at elpis.tachypleus.net, Fri Nov  6 04:12:54 UTC 2009)
>> Memory: 2097152KB
>> Booted from: /ht/pci at 5/ata-6/disk at 0
>>
>> Loading /boot/defaults/loader.conf
>> /boot/kernel/kernel data=0x72f8b0+0x199f98 
>> syms=[0x4+0x5fad0+0x4+0x83f05]
>> \
>> Hit [Enter] to boot immediately, or any other key for command prompt.
>> Booting [/boot/kernel/kernel] in 9 seconds...
>>
>> Type '?' for a list of commands, 'help' for more detailed help.
>> OK boot
>> Kernel entry at 0x100100 ...
>> KDB: debugger backends: ddb
>> KDB: current backend: ddb
>>
>> fatal kernel trap:
>>
>>    exception       = 0x7 (program)
>>    srr0            = 0x0
>>    srr1            = 0x83030
>>    lr              = 0x582038
>>    curthread       = 0
>>
>> [thread pid -1 tid 0 ]
>> Stopped at      0:      illegal instruction 0
>> db> show registers
>> r0                   0
>> r1            0x7e9e20  tmpstk+0x1e20
>> r2                   0
>> r3            0x829998  args.5817
>> r4          0xff88ed18
>> r5            0x829998  args.5817
>> r6            0x9b60ac  translations
>> r7               0x960  dsisize+0x8bc
>> r8                 0x1
>> r9                   0
>> r10          0x1c32ce4
>> r11                  0
>> r12         0x24000028
>> r13                  0
>> r14                  0
>> r15                  0
>> r16                  0
>> r17                  0
>> r18                  0
>> r19           0x100100  __start
>> r20           0xeaf000
>> r21                0x4
>> r22          0x1c2e330
>> r23           0x9c9848  end
>> r24           0x100100  __start
>> r25           0x9c5f54  mmu_kernel_obj
>> r26              0x960  dsisize+0x8bc
>> r27           0x9b60ac  translations
>> r28           0x7e00d4
>> r29         0xff88ed18
>> r30             0x3030  dsisize+0x2f8c
>> r31           0x7e9e20  tmpstk+0x1e20
>> srr0                 0
>> srr1           0x83030
>> lr            0x582038  openfirmware+0x120
>> ctr                  0
>> cr          0x24000022
>> xer                  0
>> dar                  0
>> dsisr                0
>> 0:      illegal instruction 0
>> db> bt
>> Tracing pid -1 tid 0 td 0
>> 0x007e9e80: at ofw_std_getprop+0x3c
>> 0x007e9ea0: at OF_getprop+0x8c
>> 0x007e9ed0: at moea64_bridge_bootstrap+0x5c8
>> 0x007e9f90: at kernbase+0xfc
>> db>
>> ---
>>
>> On Fri, Nov 6, 2009 at 11:57 AM, Nathan Whitehorn
>> <nwhitehorn at freebsd.org> wrote:
>>  
>>> Jacob Lambert wrote:
>>>    
>>>> All-righty, I've got three boot transcripts for you're viewing 
>>>> pleasure.
>>>>
>>>> First, I snagged a copy of 8.0-CURRENT-200906-powerpc-disc1.iso and
>>>> burned it. I've removed the ATI video card so this is all done from
>>>> the serial console with Open Firmware. Here's the transcript:
>>>>
>>>>       
>>> [snip]
>>>    
>>>> Next, being a bored and curious sysadmin, I managed to find that you
>>>> had just uploaded 9.0-20091106-SNAP-powerpc-bootonly.iso to your home
>>>> folder on people.freebsd.org. I suppose it's premature for me to be
>>>> fiddling with it but like I said, I'm bored. :) Here's the transcript:
>>>>
>>>>       
>>> That is remarkably bored and curious: you beat my email to tell you 
>>> about
>>> it! I'd be very interested in the output of 'show registers' and 
>>> also 'bt',
>>> if the kernel is on the real stack yet.
>>> -Nathan
>>>
>>>     
>>
>>
>>
>>   
>
> _______________________________________________
> freebsd-ppc at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-ppc
> To unsubscribe, send any mail to "freebsd-ppc-unsubscribe at freebsd.org"