Re: Kernel-Messages translation

Peter Barker (pbarker@brucehall19.anu.edu.au)
Thu, 19 Jun 1997 16:16:55 -0500 (EST)

All,
Seem to have a problem with PCI compiled in the kernel. It's a
long story.
My machine does not (AFAIK ;-) have PCI. According to a 2.0.18
kernel (hey; I needed to have it around for a course...):

Jun 19 11:31:09 brucehall19 kernel: Console: colour VGA+ 80x25, 1 virtual
console (max 63)
Jun 19 11:31:09 brucehall19 kernel: pcibios_init : BIOS32 Service
Directory structure at 0x000ffe80
Jun 19 11:31:09 brucehall19 kernel: pcibios_init : BIOS32 Service
Directory entry at 0xffe7b
Jun 19 11:31:09 brucehall19 kernel: bios32_service(1229148196) : not
present
Jun 19 11:31:09 brucehall19 kernel: pci_init: no BIOS32 detected
Jun 19 11:31:09 brucehall19 kernel: Calibrating delay loop.. ok - 49.77
BogoMIPS

What? It appears to have _found_ a "Service Directory Structure", but then
fails to access it?

The same happens on 2.1.43 kernels, but I believe the Service Directory
structure "moves around"

This is all well and good, and I wouldn't care terribly much since my
hardware is oft described as dodgy. Unfortunately, 2.0.30 reliably
comes up with:

-----
pcibios_init : BIOS32 Service Directory entry at 0xffe7b
bios32_service(1229148196) : not present
Probing PCI hardware.
gen. prot: 0000
CPU: 0
EIP: 0010: [<0000001d>]
EFLAGS: 00010046
eax: 0000b1eb ebx: 00000000 ecx: 00000000 eax: 01800000
esi: 0016bca1 edi: 0f000000 ebp: 00190690 esp: 0016df24
ds: 0018 es: 0018 fs: 0018 gs: 0018 ss: 0018
Process swapper (pid: 0, process nr: 0, stackpage=0016bfe0
Stack: 0016df8f 001900ac 00000012 00000046 0010ed8f 00000000 00000000 0000000e
0016df8f 01800000 0009e000 001618a0 00000000 00000000 0000000e 0016df8f
01800000 0009e000 001900ac 0016dfa4 00000016 00000000 00000000 01800000
Call Trace: [<0010ed84>] [<001618a0>] [<00161c2a>] [<001091a5>]
Code: e8 00 f0 a5 fe 00 f0 87 e9 00 f0 5a ef 00 f0 5a ef 00 f0 5a
kfree of non-kmalloced memory: 0016e028, next= 00000000 order=0
kfree of non-kmalloced memory: 00160181, next= 00000000 order=0
kfree of non-kmalloced memory: 0016e52c, next= 00000000 order=0
idle task may not sleep
idle task may not sleep
idle task may not sleep

---
(You may not be surprised this was copied by hand :-)


A feeble attempt at resolving those numbers:

---
Moriarty:/usr/src/linux/scripts# ./ksymoops /usr/src/linux/System.map <
~pbarker/oops
Using `/usr/src/linux/System.map' to map addresses to symbols.


Trace: 10ed84 <pcibios_read_config_byte+34/50>
Trace: 1618a0 <scan_bus+40/360>
Trace: 161c2a <pci_init+6a/90>
Trace: 1091a5 <start_kernel+d5/1b0>


Code:

---


OK, so you may say "well... don't do that then." I normally don't. The
problem is that vendors such as RedHat (sorry to single you out...) have
pci support in their install/recovery disks. Made moving from 4.1->4.2
annoying.


On a (hopefully unrelated) topic, has anyone else had problems with things
like kernel-compiles just dying at random spots with SIGIOTs and similar?
I know, I know, check my memory (see above for comments on hardware), but
I've gotten a few hints in here at other people having the same problem...


OK. Enough already. If you need more info, ask.


later,


Peter Barker           <AKA: Talon>   |   N    ,-_| \ /---- Moree, NSW 
Year: 3     Course: Engineering/Sci   | W + E /    ./}                
p.barker@bohm.anu.edu.au              |   S   \_,-._*<-- Canberra      
http://brucehall19.anu.edu.au         |            v     [35S, 149E]