HCRX-24 and missing consoles on my C200

This has never worked and I don't know why.
See:

	http://lists.debian.org/debian-hppa/2003/10/msg00073.html
	http://lists.debian.org/debian-hppa/2005/02/msg00120.html

I've previously asserted one does NOT need the sti= parameter if there is only gfx card in the system and it's being used as the firmware console. I don't know if that's still true.

STI firmware is failing to initialize itself, not the kernel initializing the card.

Double would have been nice. ;)
Next time, check the mailing list archives to determine if a particular email reached the list(s).

ciao,
grant
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

Thank You Grant, helpfully as usual. :)
Please, can you suggest to me any video card that is working on a C200 with linux (24 bpp)?

I'm bit confused about it, what do you mean with "the firmware is failing to initialize itself"?

I'm really sorry for this annoyng issue, next time I'll do as you suggested.

I've asked some help to the list owner.

Thank you again!

Ciao.

Fabio.

To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Someone should probably turn on debugging. Edit drivers/video/sticore.h and change the initial #if 0 to #if 1 and rebuild and boot.

As far as I know, this is failing the initial call to STI init_graph, I suspect either it doesn't like the default init values, or something is not quite right when setting up the glob_cfg. As I understand it, we set up the glob_cfg structures with nothing but the region pointers, and then call init_graph to set up the rest of the context which is passed between every further called STI function.

Or the STI rom is completely broken on this card and HP-UX uses software instead of the firmware routines.

I think I have a HCRX here (A4070 with A4072), I can try and fix this but not until at least January.

Cheers,
Kyle
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

STI is a seperate firmware from PDC and IODC. We have to invoke the initializaton functions from the OS to get it started. I'm not STI expert though. And we do have public documentation for STI that you are "encouraged" to read :^)

http://ftp.parisc-linux.org/docs/arch/sti.pdf

And sorry, I don't know offhand if any 24bpp graphics cards will work. Perhaps some are listed in the parisc graphics FAQ or
http://wiki.parisc-linux.org/.

It's not that bad...don't lose any sleep over it.

welcome.

ciao!
grant
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

Fabio wrote (repeatedly ;-)

First things first - this ain't really a Coral SGC at all. Its a Hyperbowl card (HCRX-24/A4071B). The lazy ***s in the HP graphics hardware group reused the card identifiers!

The "init_graph failed" is worrying - that same card with the same id is successfully detected and initialized in my B132L, and I'm pretty sure the same was true in my C180 (kernel 2.6.11-rc1-pa4).

That ought to be sti=10/8 according to your dmesg. But as Grant said, it shouldn't strictly be needed as long as there's just the one card in the box.

Ignore this - believe me, you don't have a Coral SGC.

I sent Helge Deller an A4071B a while back so he could work on the stifb driver when he had time. I guess he hasn't had time.

I'm gonna try a couple of tests on the B132L this evening - first thing is to make the Hyperbowl the console and see what happens.

-Andy
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

HPUX *never* supported this card on C-class boxes.
See page 157 (pdf page #) of C200 owners manual:
NOTICE: The A4077A Color Graphics Card, A4078A Dual Color Graphics Card,
A4079B HCRX-8Z graphics adapter, and the A4071B HCRX-24 graphics adapter with the A4072A Z Accelerator attached are not supported in the Model C160/C180.

getting it working might be quite "challenging". :)

grant

To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

I think I've already taken the torch on enough challenging HP graphics problems... mumble. I'll pass on this one.

:)
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

This looks like the same card that I have:

1. Coral SGC Graphics at 0xf4000000 [0] { 10, 0x0, 0x004, 0x00077 }
...
STI GSC/PCI core graphics driver Version 0.9a
id 2bcb015a-9a02587, conforms to spec rev. 8.04
graphics card name: HPA4071B_LZ

I tested this recently on my 715/100:

http://lists.parisc-linux.org/pipermail/parisc-linux/2005-November/027654.html

(Worked fine, except for a few small glitches.)

Stuart Brady
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

rebuild and boot.

suspect either it doesn't like the default init values, or something is not quite right when setting up the glob_cfg. As I understand it, we set up the glob_cfg structures with nothing but the region pointers, and then call init_graph to set up the rest of the context which is passed

instead of the firmware routines.

but not until at least January.

Hi folks,

Here's a datapoint with 2.6.15-rc5-pa1 on a B132L

Linux version 2.6.15-rc5-pa1 (root@weasel) (gcc version 3.3.5 (Gentoo Linux 3.3.5-r1, pie-8.7.7)) #1 Wed Dec 7 23:02:54 CET 2005

Kernel command line: root=/dev/sda2 sti=8/8 stifb=bpp:0 console=tty0 sti_font=VGA8x16 palo_kernel=2/boot/vmlinux

Found devices:
15. Coral SGC Graphics at 0xf4000000 [8/8] { 10, 0x0, 0x004, 0x00077 } 16. Gecko GSC Core Graphics at 0xf8000000 [8/24] { 10, 0x0, 0x016, 0x00085 }, additional addresses: 0xf0011000

STI GSC/PCI core graphics driver Version 0.9a
id 2bcb015a-9a02587, conforms to spec rev. 8.04
graphics card name: HPA4071B
id 2d08c0a7-9a02587, conforms to spec rev. 8.07
graphics card name: INTERNAL_EG_1280
sticon: Initializing STI text console.
Console: switching to colour STI console 160x64
Console: switching to colour frame buffer device 160x64
fb0: stifb 1280x1024-32 frame buffer device, HPA4071B, id: 2bcb015a, mmio:
0xf4100000
fb1: stifb 1280x1024-8 frame buffer device, INTERNAL_EG_1280, id:
2d08c0a7, mmio:
0xf8100000

Note that the A4071B/HCRX24 gets detected just fine. Its set in PDC as the console path, and also via the sti=8/8 parameter.

Nothing crashes, but the console is unusable. Most of the output is black on black, though I get bits of text in magenta, yellow and green depending on the intended colour. Tux has a bright green background and is otherwise pretty psychedelic. Back to Vis-EG for the time being.

As far as A4071B on C-Class goes, my company ran 50-60 C110's with that card and never had any trouble (HP-UX 10.20). I'm pretty sure I've had one in my C180 and it behaved exactly as it does in the B132L. If I get time I'll check tonight. C180 and C200 are basically the same Raven architecture, so if there's a general problem with initializing the card on Ravens it should show up.

Fabio: have you got the latest firmware on that C200? (apologies if you've already mentioned that before.)

ftp://ftp.itrc.hp.com/firmware_patches/hp/cpu/PF_CC2X0063.txt ftp://ftp.itrc.hp.com/firmware_patches/hp/cpu/PF_CC2X0063

-Andy

To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Have you tried running X, and then switching back to the console? That at least gets it working for me... I wonder which devices are affected by this? What about the A4071A?

I've tried reading the stifb source but I'm quite baffled by it. :-(

Stuart Brady


-- 
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Hi all,

Hi all,

I just committed a patch to stifb in kernel 2.6.15-rc6-pa1. Could you please try this one? It works for me... :-)

Stuart's hint with starting X and then switching back again pointed to a bug in the palette handling of stifb, which I hopefully fixed now.

X11 still only shows me 3/4 of the screen, but this seems to be another problem....

Regards,
Helge

To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Ooh! I'll test this... Unfortunately, I'm compiling on a 715/100 so bear with me for a moment. :)

Max Grabert found the fix for that... put "DefaultFbBpp 32" in the "Screen" section of your xorg.conf file. I'm not sure if that's a bug in stifb or not. If not, perhaps you could argue that there's a bug in xorg or the framebuffer code.

Cheers,

Stuart Brady


-- 
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

I'm sorry, but I didn't undestrand if you have fixed my problem with HCRX-24? :)

If yes the kernel with your patch is aivailable on cvs.parisc-linux.org or I've to patch the source and compile it?

Thank you.

Ciao.

Fabio.

Hello Fabio,

Probably not.
Your main problem is that you see this message:

STI GSC/PCI core graphics driver Version 0.9a

     id 2bcb015a-9a02587, conforms to spec rev. 8.04

STI init_graph failed (ret -1, errno 258)
sti: probe of 10:8 failed with error 1

"init_graph" is a graphic card firmware function, which seems to not work on your system. Some HP firmwares have support to "patch" the graphic card firmware before executing "init_graph". This is probably not the case for your c200.

So, please check if your machine runs the latest HP-C200 firmware.
For this please read the PARISC-Linux FAQ at http://www.parisc-linux.org/faq/index.html#PDCupdate ("How can I check if the PDC (firmware) revision is the latest?").

Additionally I will try write some patch-code during the next weeks, which might get your card working even on the C200 and other machines. Stay tuned.

2.6.15-rc6-pa1 is currently beeing built.
During the next hour I assume it should be finished and you may download it from here: http://cvs.parisc-linux.org/download/linux-2.6/
(I assume a B180 might be close to your c200 ?).

Helge
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

I'll give it a spin on the B132. The C180 is in storage right now, though with the cold weather we're having I could justify running it purely as a heater ;-) That thing really generates some heat when its chock of 64MB SIMMs.

-Andy

Unfortunately not. The B180 build has CCIO disabled, so it's much less useful than it might otherwise be. I keep threatening to rename the b180_defconfig to c110_defconfig and enable CCIO, HPPB, ZALON, LASI_82596, HIL, MUX and HARMONY.

Then I think I should just delete it and people with Raven-class machines should just use the 32 build and be happy about it.

So in the meantime we keep building a not-very-useful config.

To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Next Time! :)

Done, I've updated PDC Firmware to the latest release when I bought the C200, actually It runs 6.3

You have all my thanks! :)

Thank You!

Ciao.

Fabio.

Thank you, but I'm a bit cunfused, what is the kernel image that best suit my C200? (I've understand that 64Bit kernel is only useful if I have more than 3GB of RAM, so 32Bit kernel with working front-led is wonderfull for me! :) )

Thank You again.

Ciao.

Fabio.

To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Seems to work fine... except for a few black patches on Tux's feet.

I've tested using:

for n in 0 1; do

	for m in 0 4 1 5 2 6 3 7; do
		echo -e "\033[${n};3${m}mTesting\033[m"
	done

done

Works as expected.

Stuart Brady
_______________________________________________
parisc-linux mailing list
[email protected]
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

BTW, I'm doing all PA-Linux development on my Athlon desktop and cross-compile the parisc-linux kernel (e.g. with ftp://ftp.parisc-linux.org/x86-xcompilers/xc-20040917-hppa32.tgz). Then the Athlon PC runs a tftp-server and services the built kernel to the parisc servers. The kernel boots, but then has it's root directory on the local hard disk. It's way faster than native compilation, and as benefit I can boot all my parisc machines with this one single kernel (build once, boot everywhere ;-)).

Helge

To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]