[parisc-linux] 2.4.16-pa24 crashes on c3k

[parisc-linux] 2.4.16-pa24 crashes on c3k

[Grant Grundler]

Hi all,
I can't boot 2.4.16-pa24 on my c3k.
2.4.16-pa20 was working fine.
I'll put the defective .config, vmlinux, System.map on
ftp://gsyprf10.external.hp.com/pub/c3000/2.4.16-pa24-crash.tgz

I'm using serial console. But I can't capture output on my OB500 laptop.
It doesn't want to talk to the serial port - no input or output. 
I'm not understanding why since the same settings/cable work with
the OB600 (much older machine).  I'll try to get a full dump available
in the next couple of days one way or another. Maybe try yet another laptop.

Last couple of lines before stack trace are:

...
SuperIO: Found NS87560 Legacy I/O device at 00:0e.1 (IRQ 64)
SuperIO: Serial port 1 at 0x3f8
SuperIO: Serial port 2 at 0x2f8
SuperIO: Parallel port at 0x378
SuperIO: Floppy controller at 0x3f0
WARNING! Stack pointer and cr30 do not correspond

Dumping Stack from 12298000 to 1229c740:
...
[ does not look like a valid stack trace - data patterns don't look
like what I've seen before and too long ]

Kernel Fault: Code=26 regs=1229c740 (Addr=00000074)

GR02: 1010d51c
GR30: 1229c740
IOAQ: 1010e798
IIR 4b5300e8
CR30: 1229c000 CR31: 10318000
ORIG_R28: 10326dbf

grant

Re: [parisc-linux] 2.4.16-pa24 crashes on c3k

[Tom]

On Tue, 25 Dec 2001, Grant Grundler wrote:
> Hi all,
> I can't boot 2.4.16-pa24 on my c3k.
> 2.4.16-pa20 was working fine.
> I'll put the defective .config, vmlinux, System.map on
> ftp://gsyprf10.external.hp.com/pub/c3000/2.4.16-pa24-crash.tgz

I just reloaded my C180 with the 0.93 ISO and updated to -pa24 and it's
working fine. (Note to all - the 0.93 ISO will cheerfully allow you to
partition a drive as 35MB type F0, 250MB swap and 4GB root, load all the
software, and boot with the stock kernel and map. Then, when you compile
your new -pa24 kernel, it barfs because you forgot that /boot must end in
the first 2GB... Fortunately I rebooted off the CD, deleted the swap
partition, created a new swap and /boot in that same 250MB area and all is
well.)

My only kernel changes were to the SCSI stuff, I believe, and to remove
all the NICs except the onboard.

Re: [parisc-linux] 2.4.16-pa24 crashes on c3k

[Grant Grundler]

Tom wrote:
> I just reloaded my C180 with the 0.93 ISO and updated to -pa24 and it's
> working fine.

Ok - good to know. I wasn't using the default config file.
But I tried the default config and that dies with similar
console output. I need to lookup the IOAQ and GR02 and try
to figure out how it got there. Must be something related
to suckyio since I'm not using any graphics console and that's
about when console switches over.

BTW, with -pa20, I did try eepro100 driver with both an i82559
and i82550 NIC. I didn't stress either hard (55Mb/s to my laptop
via crossover cable) but they basically work with ftp. I know one
small patch is needed for the 2.4.16 eepro100 driver and hopefully
we'll pick that up with the next merge.

grant

Re: [parisc-linux] 2.4.16-pa24 crashes on c3k

[Grant Grundler]

Grant Grundler wrote:
> Must be something related to suckyio since I'm not using any graphics
> console and that's about when console switches over.

Good guess. Just committed 2.4.16-pa25 which fixes.
suckyio requires the first PCI HBA be numbered 0 (and not 1).
Was broken by -pa24 commit.
diff will be posted to parisc-linux-cvs as usual.

grant

Re: [parisc-linux] 2.4.16-pa24 crashes on c3k

[Thomas Bogendoerfer]

On Tue, Dec 25, 2001 at 11:02:47PM -0700, Grant Grundler wrote:
> Grant Grundler wrote:
> > Must be something related to suckyio since I'm not using any graphics
> > console and that's about when console switches over.
> 
> Good guess. Just committed 2.4.16-pa25 which fixes.
> suckyio requires the first PCI HBA be numbered 0 (and not 1).

Linux Tulip driver version 0.9.15-pre9 (Nov 6, 2001)

PCI or EISA Host Bus Adapter 3 not registered. in32(0x30030) returning -1
WARNING! Stack pointer and cr30 do not correspond!

Dumping Stack from 10858000 to 108587c0:

that's -pa25 on my B132.

> Was broken by -pa24 commit.

I had the problem as above, when playing with willy's dino patch (hand applied,
so I missed a chunk). I tried to understand the pci resource assigning code
and failed so far. If my headache settles down, I'm might look into that 
again today.

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessary a
good idea.                                 [ Alexander Viro on linux-kernel ]

Re: [parisc-linux] 2.4.16-pa24 crashes on c3k

[Grant Grundler]

"Craig D. Lansing" wrote:
> Did you ever get this resolved? 

Well, for the systems I can test (A500/C3000), yes.
IIRC, Thomas Bogendoerf said he fixed it for B132 and C200.

> I saw Tom's reply that it could be related to /boot extending beyond 2GB,

In general yes. But given recent kernel changes around PCI/EISA HBA numbering
I'd say it's a kernel problem.

> J2240, I have tried 32-2.4.16-pa24 and pa29 and am getting stack dumps
> similar to yours.  I have tried the pre-compiled kernels and creating from
> source to no avail.  The general steps are:
> 
> 1. Download kernel or kernel source
> 2. Compile if from source
> 3. Copy vmlinux***, System.map.***, and Config.*** to /boot
> 4. Link vmlinux to vmlinux***
> 5. Edit /etc/palo.conf
> 6. run palo (no args)

You don't need to run palo again. It doesn't hurt either.
Just point /boot/vmlinux (or whatever palo.conf says) to the new kernel.

Just a few more nits interesting to other kernel developers.
I hope Ryan or Thomas can sort this out for you.

> Linux version 2.4.9-32 (root@paer) (gcc version 3.0.2 (Debian)) #1 Fri Nov
> 30 19:36:30 MST 2001
...
> model   9000/782/J2240
...
> Found devices:
> 1. U2-IOA BC Runway Port (12) at 0xfff88000 [8], versions 0x580, 0xf, 0xb
> 2. unknown device (13) at 0xfc000000 [8/0], versions 0x5d, 0x3, 0xa
> 3. SummitHawk Dino RS-232 (10) at 0xfc003000 [8/0/63], versions 0x5d, 0x0,
> 0x8c
> 4. Dino PCI Bridge (13) at 0xfc004000 [8/4], versions 0x680, 0x3, 0xa
> 5. SummitHawk 230 Core BA (11) at 0xffd00000 [8/12], versions 0x5c, 0x0,
> 0x81,
> additional addresses: 0xffc00000
> 6. SummitHawk 230 Core Centronics (10) at 0xffd02000 [8/12/0], versions
> 0x5c, 0x
> 0, 0x74,  additional addresses: 0xffd03000
> 7. SummitHawk 230 Core Audio (10) at 0xffd04000 [8/12/1], versions 0x5c,
> 0x4, 0x
> 7b
> 8. SummitHawk 230 Core SCSI (10) at 0xffd06000 [8/12/5], versions 0x5c, 0x0,
> 0x8
> 2
> 9. SummitHawk 230 Core PS/2 Port (10) at 0xffd08000 [8/12/7], versions 0x5c,
> 0x0
> , 0x84
> 10. SummitHawk 230 Core PS/2 Port (10) at 0xffd08100 [8/12/8], versions
> 0x5c, 0x
> 0, 0x84
> 11. U2-IOA BC GSC+ Port (7) at 0xfc03f000 [8/63], versions 0x501, 0x1, 0xc
> 12. U2-IOA BC Runway Port (12) at 0xfff8a000 [10], versions 0x580, 0xf, 0xb
> 13. Cujo PCI Bridge (13) at 0xf1604000 [10/4], versions 0x682, 0x1, 0xa,
> additi
> onal addresses: 0xf6000000
> 14. SummitHawk Dino RS-232 (10) at 0xf1607000 [10/4/63], versions 0x5d, 0x0,
> 0x8
> c
> 15. U2-IOA BC GSC+ Port (7) at 0xf163f000 [10/63], versions 0x501, 0x1, 0xc
> 16. SummitHawk U+ (0) at 0xfffa0000 [32], versions 0x5b5, 0x0, 0x4
> 17. SummitHawk U+ (0) at 0xfffa2000 [34], versions 0x5b5, 0x0, 0x4
> 18. SummitHawk Memory (1) at 0xfffb1000 [49], versions 0x83, 0x0, 0x9
> CONFIG_SMP=n  ignoring additional CPUs
> Warning : device (0, 0x5b5, 0x0, 0x4) NOT claimed by CPU
> CPU(s): 1 x PA8200 (PCX-U+) at 236.000000 MHz
> ccio found U2 at 0xfff88000
> io_io_low: 0xf8000000
> 19. unknown device (13) at 0xfc000000 [8/0], versions 0x5d, 0x3, 0xa
> 20. SummitHawk Dino RS-232 (10) at 0xfc003000 [8/0/63], versions 0x5d, 0x0,
> 0x8c
> 21. Dino PCI Bridge (13) at 0xfc004000 [8/4], versions 0x680, 0x3, 0xa

I think the devices showing up twice has been fixed in 2.4.16 kernels.

> 22. SummitHawk 230 Core BA (11) at 0xffd00000 [8/12], versions 0x5c, 0x0,
> 0x81,
>  additional addresses: 0xffc00000
> 23. SummitHawk 230 Core Centronics (10) at 0xffd02000 [8/12/0], versions
> 0x5c, 0
> x0, 0x74,  additional addresses: 0xffd03000
> 24. SummitHawk 230 Core Audio (10) at 0xffd04000 [8/12/1], versions 0x5c,
> 0x4, 0
> x7b
> 25. SummitHawk 230 Core SCSI (10) at 0xffd06000 [8/12/5], versions 0x5c,
> 0x0, 0x
> 82
> 26. SummitHawk 230 Core PS/2 Port (10) at 0xffd08000 [8/12/7], versions
> 0x5c, 0x
> 0, 0x84
> 27. SummitHawk 230 Core PS/2 Port (10) at 0xffd08100 [8/12/8], versions
> 0x5c, 0x
> 0, 0x84
> 28. U2-IOA BC GSC+ Port (7) at 0xfc03f000 [8/63], versions 0x501, 0x1, 0xc
> ccio found U2 at 0xfff8a000
> io_io_low: 0xf1600000
> 29. Cujo PCI Bridge (13) at 0xf1604000 [10/4], versions 0x682, 0x1, 0xa,
> additi
> onal addresses: 0xf6000000
> 30. SummitHawk Dino RS-232 (10) at 0xf1607000 [10/4/63], versions 0x5d, 0x0,
> 0x8
> c
> 31. U2-IOA BC GSC+ Port (7) at 0xf163f000 [10/63], versions 0x501, 0x1, 0xc
> Lasi version 0 at 0xffd00000 found.
> Cujo version unknown found at 0xfc000000

hmm...another nit. We should know the version by now.

> Dino version 3.1 found at 0xfc004000
> Cujo version 2.0 found at 0xf1604000
> Enabling Cujo 2.0 bug workaround
> POSIX conformance testing by UNIFIX
> Linux NET4.0 for Linux 2.4
> Based upon Swansea University Computer Society NET3.039
> Starting kswapd v1.8
> parport_init_chip: initialize bidirectional-mode.
> parport0: PC-style at 0xffd02800, irq 88 [PCSPP,TRISTATE]
> STI PCI ROM enabled at 0xf1ff0000
> PCI graphic ROM found at f1ff0000 (64 kB), fb @ f6000000 (32 MB)
> PCI region 0: f6000000 with 32768 kB
> PCI region 1: 00000000 with 0 kB
> PCI region 2: 00000000 with 0 kB
> PCI region 3: 00000000 with 0 kB
> PCI region 4: 00000000 with 0 kB
> PCI region 5: 00000000 with 0 kB
> PCI region 6: f1ff0000 with 64 kB
> PCI region 7: 00000000 with 0 kB
> PCI region 8: 00000000 with 0 kB
> PCI region 9: 00000000 with 0 kB
> PCI region 10: 00000000 with 0 kB
> PCI region 11: 00000000 with 0 kB
> Trying PCI STI ROM at f1ff0000, PCI hpa at f6000000
> I'm not trying to set up the PCI graphic card now... It would freeze your
> machin
> e
> Unable to handle STI device 'Hewlett-Packard Company A4977A Visualize EG'

This should work now too.
Or at least it's only know to NOT work with the C3000.

grant

Re: [parisc-linux] 2.4.16-pa24 crashes on c3k

[Craig D. Lansing]

Does this mean that the answer is to just keep downloading kernels until it
works?  BTW, somewhere in the process my problems with Eth0/1 were solved.
All I can think of was the one install that I tried from source.  I guess
the `make modules` or `make module_install` fixed something without my
knowing it.  Anyway, I can now telnet, ftp, ping, etc. my J2240 from my
WinDoze PCs so I'm a little further along.

----- Original Message -----
From: Grant Grundler <grundler@dsl2.external.hp.com>
To: Craig D. Lansing <lansing@beldar.com>
Cc: <parisc-linux@lists.parisc-linux.org>
Sent: Monday, December 31, 2001 12:22 PM
Subject: Re: [parisc-linux] 2.4.16-pa24 crashes on c3k


> "Craig D. Lansing" wrote:
> > Did you ever get this resolved?
>
> Well, for the systems I can test (A500/C3000), yes.
> IIRC, Thomas Bogendoerf said he fixed it for B132 and C200.
>
> > I saw Tom's reply that it could be related to /boot extending beyond
2GB,
>
> In general yes. But given recent kernel changes around PCI/EISA HBA
numbering
> I'd say it's a kernel problem.
>
> > J2240, I have tried 32-2.4.16-pa24 and pa29 and am getting stack dumps
> > similar to yours.  I have tried the pre-compiled kernels and creating
from
> > source to no avail.  The general steps are:
> >
> > 1. Download kernel or kernel source
> > 2. Compile if from source
> > 3. Copy vmlinux***, System.map.***, and Config.*** to /boot
> > 4. Link vmlinux to vmlinux***
> > 5. Edit /etc/palo.conf
> > 6. run palo (no args)
>
> You don't need to run palo again. It doesn't hurt either.
> Just point /boot/vmlinux (or whatever palo.conf says) to the new kernel.
>
> Just a few more nits interesting to other kernel developers.
> I hope Ryan or Thomas can sort this out for you.
>
> > Linux version 2.4.9-32 (root@paer) (gcc version 3.0.2 (Debian)) #1 Fri
Nov
> > 30 19:36:30 MST 2001
> ...
> > model   9000/782/J2240
> ...
> > Found devices:
> > 1. U2-IOA BC Runway Port (12) at 0xfff88000 [8], versions 0x580, 0xf,
0xb
> > 2. unknown device (13) at 0xfc000000 [8/0], versions 0x5d, 0x3, 0xa
> > 3. SummitHawk Dino RS-232 (10) at 0xfc003000 [8/0/63], versions 0x5d,
0x0,
> > 0x8c
> > 4. Dino PCI Bridge (13) at 0xfc004000 [8/4], versions 0x680, 0x3, 0xa
> > 5. SummitHawk 230 Core BA (11) at 0xffd00000 [8/12], versions 0x5c, 0x0,
> > 0x81,
> > additional addresses: 0xffc00000
> > 6. SummitHawk 230 Core Centronics (10) at 0xffd02000 [8/12/0], versions
> > 0x5c, 0x
> > 0, 0x74,  additional addresses: 0xffd03000
> > 7. SummitHawk 230 Core Audio (10) at 0xffd04000 [8/12/1], versions 0x5c,
> > 0x4, 0x
> > 7b
> > 8. SummitHawk 230 Core SCSI (10) at 0xffd06000 [8/12/5], versions 0x5c,
0x0,
> > 0x8
> > 2
> > 9. SummitHawk 230 Core PS/2 Port (10) at 0xffd08000 [8/12/7], versions
0x5c,
> > 0x0
> > , 0x84
> > 10. SummitHawk 230 Core PS/2 Port (10) at 0xffd08100 [8/12/8], versions
> > 0x5c, 0x
> > 0, 0x84
> > 11. U2-IOA BC GSC+ Port (7) at 0xfc03f000 [8/63], versions 0x501, 0x1,
0xc
> > 12. U2-IOA BC Runway Port (12) at 0xfff8a000 [10], versions 0x580, 0xf,
0xb
> > 13. Cujo PCI Bridge (13) at 0xf1604000 [10/4], versions 0x682, 0x1, 0xa,
> > additi
> > onal addresses: 0xf6000000
> > 14. SummitHawk Dino RS-232 (10) at 0xf1607000 [10/4/63], versions 0x5d,
0x0,
> > 0x8
> > c
> > 15. U2-IOA BC GSC+ Port (7) at 0xf163f000 [10/63], versions 0x501, 0x1,
0xc
> > 16. SummitHawk U+ (0) at 0xfffa0000 [32], versions 0x5b5, 0x0, 0x4
> > 17. SummitHawk U+ (0) at 0xfffa2000 [34], versions 0x5b5, 0x0, 0x4
> > 18. SummitHawk Memory (1) at 0xfffb1000 [49], versions 0x83, 0x0, 0x9
> > CONFIG_SMP=n  ignoring additional CPUs
> > Warning : device (0, 0x5b5, 0x0, 0x4) NOT claimed by CPU
> > CPU(s): 1 x PA8200 (PCX-U+) at 236.000000 MHz
> > ccio found U2 at 0xfff88000
> > io_io_low: 0xf8000000
> > 19. unknown device (13) at 0xfc000000 [8/0], versions 0x5d, 0x3, 0xa
> > 20. SummitHawk Dino RS-232 (10) at 0xfc003000 [8/0/63], versions 0x5d,
0x0,
> > 0x8c
> > 21. Dino PCI Bridge (13) at 0xfc004000 [8/4], versions 0x680, 0x3, 0xa
>
> I think the devices showing up twice has been fixed in 2.4.16 kernels.
>
> > 22. SummitHawk 230 Core BA (11) at 0xffd00000 [8/12], versions 0x5c,
0x0,
> > 0x81,
> >  additional addresses: 0xffc00000
> > 23. SummitHawk 230 Core Centronics (10) at 0xffd02000 [8/12/0], versions
> > 0x5c, 0
> > x0, 0x74,  additional addresses: 0xffd03000
> > 24. SummitHawk 230 Core Audio (10) at 0xffd04000 [8/12/1], versions
0x5c,
> > 0x4, 0
> > x7b
> > 25. SummitHawk 230 Core SCSI (10) at 0xffd06000 [8/12/5], versions 0x5c,
> > 0x0, 0x
> > 82
> > 26. SummitHawk 230 Core PS/2 Port (10) at 0xffd08000 [8/12/7], versions
> > 0x5c, 0x
> > 0, 0x84
> > 27. SummitHawk 230 Core PS/2 Port (10) at 0xffd08100 [8/12/8], versions
> > 0x5c, 0x
> > 0, 0x84
> > 28. U2-IOA BC GSC+ Port (7) at 0xfc03f000 [8/63], versions 0x501, 0x1,
0xc
> > ccio found U2 at 0xfff8a000
> > io_io_low: 0xf1600000
> > 29. Cujo PCI Bridge (13) at 0xf1604000 [10/4], versions 0x682, 0x1, 0xa,
> > additi
> > onal addresses: 0xf6000000
> > 30. SummitHawk Dino RS-232 (10) at 0xf1607000 [10/4/63], versions 0x5d,
0x0,
> > 0x8
> > c
> > 31. U2-IOA BC GSC+ Port (7) at 0xf163f000 [10/63], versions 0x501, 0x1,
0xc
> > Lasi version 0 at 0xffd00000 found.
> > Cujo version unknown found at 0xfc000000
>
> hmm...another nit. We should know the version by now.
>
> > Dino version 3.1 found at 0xfc004000
> > Cujo version 2.0 found at 0xf1604000
> > Enabling Cujo 2.0 bug workaround
> > POSIX conformance testing by UNIFIX
> > Linux NET4.0 for Linux 2.4
> > Based upon Swansea University Computer Society NET3.039
> > Starting kswapd v1.8
> > parport_init_chip: initialize bidirectional-mode.
> > parport0: PC-style at 0xffd02800, irq 88 [PCSPP,TRISTATE]
> > STI PCI ROM enabled at 0xf1ff0000
> > PCI graphic ROM found at f1ff0000 (64 kB), fb @ f6000000 (32 MB)
> > PCI region 0: f6000000 with 32768 kB
> > PCI region 1: 00000000 with 0 kB
> > PCI region 2: 00000000 with 0 kB
> > PCI region 3: 00000000 with 0 kB
> > PCI region 4: 00000000 with 0 kB
> > PCI region 5: 00000000 with 0 kB
> > PCI region 6: f1ff0000 with 64 kB
> > PCI region 7: 00000000 with 0 kB
> > PCI region 8: 00000000 with 0 kB
> > PCI region 9: 00000000 with 0 kB
> > PCI region 10: 00000000 with 0 kB
> > PCI region 11: 00000000 with 0 kB
> > Trying PCI STI ROM at f1ff0000, PCI hpa at f6000000
> > I'm not trying to set up the PCI graphic card now... It would freeze
your
> > machin
> > e
> > Unable to handle STI device 'Hewlett-Packard Company A4977A Visualize
EG'
>
> This should work now too.
> Or at least it's only know to NOT work with the C3000.
>
> grant
>
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux
>

Re: [parisc-linux] 2.4.16-pa24 crashes on c3k

[Grant Grundler]

"Craig D. Lansing" wrote:
> Does this mean that the answer is to just keep downloading kernels until it
> works?

I meant either look at the host bus controller enumeration code and
figure out the problem with EISA/PCI or wait until someone does.

grant