RE: More detail on Deskpro XL6200 NIC (was: Legacy ethernet cards inFreeBSD)

2000-11-17 Thread Koster, K.J. 


 *ahem* How about compiling a kernel *without* the pcn driver? 
 I want to know why you didn't consider this combination. (Well, I
 know why: because giving me the answer straight away rather than
 making me drag it out of you would have been too easy. I would never
 be that lucky.)

I did not consider a kernel without pcn compiled in is because you asked me
to use the pcn driver. If you want straight answers, ask straight questions.
:)

For your information, I did hack the kernel for the lnc driver to *not*
probe for the card, because I (wrongly) guessed that perhaps the lnc probe
was screwing things up for the pnc probe. I figured that disabling the lnc
driver in the visual config would do the trick, but as I type this it occurs
to me that that might only disable the ISA card probe.

I've tried a kernel without the pnc driver compiled in. Clean build
directory and cvsupped to RELENG_4 as of 30 minutes ago. (typed from the
screen, names may be misspelled, number are corrent, dmesg output attached)

Fatal trap 12: page fault in kernel mode
fault virtaul address:  0x616776
fault code: supervisor read, page not present
instruction ptr:0x8:0x616776
stackprt:   0x10:0xc5d1bc4c
frameptr:   0x10:0xc5d1bd0c
code segment:   base 0x0, limit 0xf, type 0x1b
DPL 0, pres 1, def32 1 gran 1
processor eflagsinterrupt enabled, resume, IOPL 0
current proc.   1 (init)
interrupt mask  none
trap nr   12

It just keeps getting better and better. *sigh* This is the first out of
three machines that are to move up to 4.2. I planned a day for the first and
an hour for each next one.

I'm pretty sure the hardware is good. It survived a couple of buildworlds.
It works in 4.0.

I will see if I have time later today to rebuild the beast with kernel
debugging stuff enabled to give more details. Right now I have other work to
do.

Are there any specific things that I should try? I'm a little lost.

Kees Jan


 You are only young once,
   but you can stay immature all your life.




This is my own custom kernel (4.2)

Nov 17 12:38:52 dl6101 /kernel: Copyright (c) 1992-2000 The FreeBSD Project.
Nov 17 12:38:52 dl6101 /kernel: Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 
1991, 1992, 1993, 1994
Nov 17 12:38:52 dl6101 /kernel: The Regents of the University of California. All 
rights reserved.
Nov 17 12:38:52 dl6101 /kernel: FreeBSD 4.2-BETA #0: Fri Nov 17 12:24:14 CET 2000
Nov 17 12:38:52 dl6101 /kernel: [EMAIL PROTECTED]:/usr/src/sys/compile/QUAM
Nov 17 12:38:52 dl6101 /kernel: Timecounter "i8254"  frequency 1193182 Hz
Nov 17 12:38:52 dl6101 /kernel: Timecounter "TSC"  frequency 25039 Hz
Nov 17 12:38:52 dl6101 /kernel: CPU: Pentium Pro (200.01-MHz 686-class CPU)
Nov 17 12:38:52 dl6101 /kernel: Origin = "GenuineIntel"  Id = 0x616  Stepping = 6
Nov 17 12:38:52 dl6101 /kernel: 
Features=0xf9ffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR,PGE,MCA,CMOV
Nov 17 12:38:52 dl6101 /kernel: real memory  = 67108864 (65536K bytes)
Nov 17 12:38:52 dl6101 /kernel: avail memory = 62468096 (61004K bytes)
Nov 17 12:38:52 dl6101 /kernel: bios32: Bad BIOS32 Service Directory
Nov 17 12:38:52 dl6101 /kernel: Preloaded elf kernel "kernel" at 0xc02e6000.
Nov 17 12:38:52 dl6101 /kernel: Pentium Pro MTRR support enabled
Nov 17 12:38:52 dl6101 /kernel: npx0: math processor on motherboard
Nov 17 12:38:52 dl6101 /kernel: npx0: INT 16 interface
Nov 17 12:38:52 dl6101 /kernel: pcib0: Intel 82454KX/GX (Orion) host to PCI bridge 
on motherboard
Nov 17 12:38:52 dl6101 /kernel: pci0: PCI bus on pcib0
Nov 17 12:38:53 dl6101 /kernel: lnc0: PCNet/PCI Ethernet adapter port 0x7000-0x701f 
irq 5 at device 11.0 on pci0
Nov 17 12:38:53 dl6101 /kernel: lnc0: PCnet-32 VL-Bus address 00:80:5f:fa:a0:82
Nov 17 12:38:53 dl6101 /kernel: lnc0: driver is using old-style compatability shims
Nov 17 12:38:53 dl6101 /kernel: amd0: Tekram DC390(T)/AMD53c974 SCSI Host Adapter 
port 0x7100-0x717f irq 10 at device 12.0 on pci0
Nov 17 12:38:53 dl6101 /kernel: pci0: Matrox MGA Millennium 2064W graphics 
accelerator at 13.0
Nov 17 12:38:53 dl6101 /kernel: rl0: RealTek 8139 10/100BaseTX port 0x7200-0x72ff 
mem 0x430-0x43000ff irq 9 at device 14.0 on pci0
Nov 17 12:38:53 dl6101 /kernel: rl0: Ethernet address: 00:e0:4c:68:5c:0c
Nov 17 12:38:53 dl6101 /kernel: miibus0: MII bus on rl0
Nov 17 12:38:53 dl6101 /kernel: rlphy0: RealTek internal media interface on miibus0
Nov 17 12:38:53 dl6101 /kernel: rlphy0:  10baseT, 10baseT-FDX, 100baseTX, 
100baseTX-FDX, auto
Nov 17 12:38:53 dl6101 /kernel: isab0: PCI to EISA bridge (vendor=0e11 device=0001) 
at device 15.0 on pci0
Nov 17 12:38:53 dl6101 /kernel: eisa0: EISA bus on isab0
Nov 17 12:38:53 dl6101 /kernel: mainboard0: CPQ0521 (System Board) on eisa0 slot 0
Nov 17 12:38:53 dl6101 /kernel: isa0: ISA bus on isab0
Nov 17 12:38:53 dl6101 /kernel: chip0: Intel 82453KX/GX

Re: More detail on Deskpro XL6200 NIC (was: Legacy ethernet cards inFreeBSD)

2000-11-17 Thread Robert Swindells 


There used to be some very strange code in the lnc driver that relied
on the ISA probe for the original PCnet-PCI device failing, then probed
it in the if_lnc_pci stub.

Your card shouldn't really be identified as a VL-Bus one when it is a
PCI device, so maybe the lnc probe has been broken by the removal of
it's PCI stub that used to be in pci/if_lnc_pci.c

The datasheet for the 79C970 doesn't seem to be available on AMD's
website anymore, so I couldn't check on the device ID, but I think
that the one specified in the header was deliberately incorrect.

Robert



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

RE: More detail on Deskpro XL6200 NIC (was: Legacy ethernet cards inFreeBSD)

2000-11-17 Thread Koster, K.J. 

Dear Bill,

I've cvsupped, which gave me the patched version of the pcn driver. Thanks
for the prompt response.

I've build various kernels.

Kernel with pcn and lnc compiled in: trap 12
Kernel with lnc only:trap 12
Kernel with pcn only:works, but no network
Kernel with neihter: works, but no network

A backtrace reveals that the box dies with trap 12 in the function:

  __set_videodriver_set_sym_vga_driver()
  alloc_scp()
  scopen()
  cnopen()
  spec_open()
  spec_vnoperate()
  ufs_vnoperatespec()
  vn_open()
  open()
  syscall2()
  Xint0x80_syscall()

The last thing that the box prints is "Addtional TCP options: ." and then
the date. *beng* trap 12 etc

Yes, it does print the "." in the end, so it completes the additional TCP
options successfully.

I'm lost here. If anyone has any suggestions I will happily try them on
Monday. Otherwise I'm moving down a version or two. Right now I'm going home
for the weekend (to a system that actually *works*).

Kees Jan


 You are only young once,
   but you can stay immature all your life.


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message