1.1.81 PCI probe hang

1.1.81 PCI probe hang

Post by Blake Jon » Fri, 20 Jan 1995 19:37:11



Well, I finally tried booting 1.1.81 tonight after running 1.1.66 quite well
for a month or so, and it hung very quickly.  I had already seen the first
four lines about bios32_init since 1.1.6x, and I did mail as suggested; but
then the PCI probing started, and I hung immediately.  Here's the information
(copied by hand):

-- snip --
bios32_init : BIOS32 Service Directory structure at 0x000e8130
bios32_init : BIOS32 Service Directory entry at 0xf482c
bios32_init : BIOS32 Service Directory structure at 0x000f0130

pcibios_init : PCI BIOS revision 2.00 entry at 0xf4800
Probing PCI Hardware.
general protection: 25a0
EIP: 0010:000f480e
EFLAGS: 00010046
eax: 0000000a   ebx: 0000004e   ecx: 00000000   edx: 0000004e
esi: 0019e964   edi: 00000008   ebp: 00008000   esp: 001a092c
Process swapper (pid: 0, process nr: 0, stackpage = 0019e9c0)
Stack:  001125a1 00000010 0000004e 00000000 01000000 0011320f 00000000 0000004e
        00000008 001a0960 00100000 001db400 0100004e 00000000 001132bd 00000000
        0011326c 001137a3 001db400 01000000 756e6547 001050a0 001c7538 0010dc33
Call Trace: 001125a1 01000000 0011320f 001132bd 0011326c 001137a3 01000000
        0010dc33 01000000 01000000 01000000
Code: eb 1b 90 90 90 80 fc b4 0f 85 0a 00 00 00 e8 4f 06 00 00 eb
kfree of non-kmalloced memory: 001a09b8, next=00000000, order=0
task[0] (swapper) killed: unable to recover
Kernel panic: Trying to free up swapper memory space
In swapper task - not syncing
-- snip --

The EIP wasn't very helpful, but here are some snippets of System.map
that seem relevant:

-- snip --
0011218c t _bios32_service
0011233c t _pcibios_init
0011242c T _pcibios_present
0011243c T _pcibios_find_class
0011249c T _pcibios_find_device
001124fc T _pcibios_read_config_byte
0011253c T _pcibios_read_config_word
0011257c T _pcibios_read_config_dword
001125bc T _pcibios_write_config_byte
001125fc T _pcibios_write_config_word
0011263c T _pcibios_write_config_dword
001126dc T _NCR53c810_test
0011282c T _pcibios_strerror
001128bc T _multi_function
001128ec T _interrupt_decode
0011292c T _bist_probe
0011295c T _revision_decode
00112c8c T _class_decode
00112f7c T _device_decode
0011314c T _vendor_decode
001131ec T _probe_devices
0011328c T _probe_pci
001132dc T _add_pci_resource
001133dc T _get_pci_list
001136cc T _bios32_init
001137ac t _get_long
0011382c t _put_long
001138dc t _find_vma

0019e948 d _bios32_entry
0019e94c d _bios32_indirect
0019e954 d _pci_list
0019e95c d _pci_index
0019e960 d _pcibios_entry
0019e964 d _pci_indirect
-- snip --

My system is a 486 DX2/66 with 256k cache and 16 MB RAM; in one PCI slot there
is a "DTC 2130S PCI Bus to IDE Single Port Controller" with 2 540MB IDE drives
attached, and an ATI Mach64 in the other one.  As far as I can gather, the
motherboard is an "AL1" of some sort ("Premio" perhaps?  That was the only
brand-name type word I saw on the board...); the manual is completely
unhelpful.  I hope this gets fixed before 1.2.0...

-Blake
--



      Get 2+ million digits of Pi from ftp://foo.caltech.edu/pub/pi/pi.gz


GM/CS d- H s g+ p2 !au a18 w+ v+(-) C++(++++) U[VIXH]++>+++S++$>+++L++++C---
P+>+++ L++>+++ 3 E-(---) N+>++ K++ W--- M-- V-- po- Y+>++ t !5 j- R- G' tv-(--)
b++ !D B-- e+(*)>---(++++) u++(*)(**) h+(-)(*) f? !r n+(-)(----) !y
--



      Get 2+ million digits of Pi from ftp://foo.caltech.edu/pub/pi/pi.gz

 
 
 

1. 1.1.81 kernel panic NCR 53c810

I just had a kernel panic on a 1.1.81 system (P90, NCR 53c810).
The message wasn't logged, because the panic killed the
swapper, but here's the error message was

scsi0: scan_scsi_buf too small (need 4607 bytes)

This happened during some heavy I/O activity, i.e. untaring
a 20 Meg file onto a MS-DOS partition.
--

The joy of engineering is to find a straight line on a double
logarithmic diagram.

2. FYI: Fips2.0 vs. Partition Manager on Win98

3. 1.1.81 and Mitsumi 4X IDE (ATAPI) CD-ROM problems

4. i2c-ali15x3 on hp omnibook xe3

5. /etc/utmp error - gcc 2.6.2 kernel 1.1.81 XF86 3.1

6. Apache Displaying Test Page On named based Virtual Hosts

7. iBCS broken with 1.1.81

8. where are Solaris 2.5 relnotes?

9. Kernel change summary 1.1.81 -> 1.1.82

10. Rebuild 1.1.81 & have vmlinuz &vmlinux

11. 1.1.81 Netstat Bug?

12. Multi-homed NFS servers was Re: 1.1.81 seems to break NFS

13. Kernel>1.1.81 Support EIDE???