802.3 LLC driver completely hangs up system at boot time

Byeong-ryeol Kim (jinbo21@soback.kornet.nm.kr)
Sun, 4 Jan 1998 20:58:54 +0900 (KST)

Of course, I know experimental drivers sometimes does not work at all,
and brings harmful effects to the behavior of system or system itself.
Today, I experienced such driver myself, and I think it is the
last driver that does not work and behaves as a killer of kernel.

I do not know this was reported before. I submit a report
in order to warn for those like me, but not to complain. :-)
It took 2 days to find this fact, sigh...

I usually compile kernel with almost modules modules as possible
as I can.
Because (I know) 802.3 LLC driver is experimental and I could not
compile it before, I have been disabling it in kernel configuration
until the day before yesterday..

I enabled CONFIG_LLC again as a test when compiling kernel-2.1.76(7).
This time It was compiled without any problem.
But I could not load kernel at boot time, and the system was locked
up completely.
So I should have push reset button to reboot.
There was no logging in the /var/log/messages, etc.
Here are some messages about it which I wrote manually.
It does not seem to be related with versions or sort of compilers.
I used gcc-2.7.2.3-8 of redhat 5.0, egcs-1.0.1-prerelease,
egcs-971225, but results were the same)

2.1.76, 2.1.77 kernels boots without any problem, if I disable the
'CONFIG_LLC' option when configuring kernel. Compilers used in this
case(without LLC) are gcc-2.7.2.3-8 of redhat 5.0 and egcs-1.0.1-
prerelease.


I have been using modutils-2.1.71, kernel-2.1.76 compiled with
almost all modules except some modules(including LLC) on redhat
5.0 system. The hardware components are as follows:

MainBoard: ASUS P/I-P55T2P4 rev. 3.0
BIOS: 0205(Award 4.51PG)
CPU: intel MMX 166 (stepping 4)
RAM: SamSung 60ns EDO RAM 64M without parity
HDD: Quantum F/B 1280A & Conner CFA 850A
NIC: 3c509b combo
Serial: ZyXEL ISDN E2864I external TA

BTW, I found no symbols corresponding with EIP number below.
...
POSIX conformance testing by UNIFIX
Unable to handle kernel paging request at virtual address 01007f43
current->tss.cr3=00101000, %cr3=00101000
*pde=00000000
Ooops: 0000
CPU: 0
EIP: 0010:[<c0119080>]
EFLAGS: 00010246
eax:00000000 ebx:00000000 ecx:ffffffff edx:04040403
....
Process: swapper (pid:0, process nr:0, stackpage=c0107000)
Stack: c02ee800 c01e1ed68 c01c70ae 0000000a 81a470ae
...
Call Trace: [<c01c70ae>][<c0119115>]...
Code: f2 ae f7 d1 49 66 8b 46 10 89 cd 66 89 44 24 12 ...
Kernel Panic: Attend to kill the idle task!
In swapper task - not syncing

BTW, can we use LLC driver before kernel 2.2.0?

"Where there is a will, there is a way." jinbo21@soback.kornet.nm.kr
kbeyl@kids.kotel.co.kr
For the future of you and me! hitel: jinbo21