Hello, Hallo I want to setup accelerated X on a PowerMac 9600. I'm using LinuxPPC, but had no luck so far and someone from the linuxppc-user mailing list gave me the hint that the chip in the machine I'm using is from a german company (Formac), so I ask here if Suse-linux has solved this (is Suse-linux's X accelerated on european PowerMac's 9600?) What I know so far: Xpmac (at least Xpmac-Rage128 rev.9) doesn't support acceleration on this 9600. While booting, my selfcompiled 2.2.15pre9 kernel says '/bandit/formacGA5' and 'unsupported', and XF86_FBDev (Version 3.3.3) doesn't seem to understand the hardware, the lower half of the screen remains black (or whatever was on the console) and the upper half shows the content twice and totally wrong (some bit depth problems). The kernel (rsync from linuxcare.com.au) has all mac video drivers compiled in except for the matrox acceleration. There's no option in xconfig like 'formac' or 'GA5'. Thanks! Danke, Merci. Christian.
On Mon, May 01, Christian Jaeger wrote:
What I know so far: Xpmac (at least Xpmac-Rage128 rev.9) doesn't support acceleration on this 9600. While booting, my selfcompiled 2.2.15pre9 kernel says '/bandit/formacGA5' and 'unsupported', and XF86_FBDev (Version 3.3.3) doesn't seem to understand the hardware, the lower half of the screen remains black (or whatever was on the console) and the upper half shows the content twice and totally wrong (some bit depth problems). The kernel (rsync from linuxcare.com.au) has all mac video drivers compiled in except for the matrox acceleration. There's no option in xconfig like 'formac' or 'GA5'.
The latest Xpmac has better support for the known cards. What is exactly in your machine? Try the command /sbin/lspci /sbin/lspci -n Gruss Olaf -- $ man clone BUGS Main feature not yet implemented...
At 11:23 Uhr +0200 1.5.2000, Olaf Hering wrote:
The latest Xpmac has better support for the known cards.
I now have tried Xpmac.rev10 (the version from Kevin Hendricks with USB and rage128 support), without options, and also with -mach64, -ct6555, -tt128. Without options, it's unaccelerated, and with each of the 3 options it hangs or crashes (tt128 works best, it runs quite a while before crashing, while booting the window manager).
What is exactly in your machine? Try the command /sbin/lspci
lspci 00:0b.0 Host bridge: Apple Computer Inc. Bandit PowerPC host bridge (rev 03) 00:10.0 Class ff00: Apple Computer Inc. Grand Central I/O (rev 02) lspci -n 00:0b.0 Class 0600: 106b:0001 (rev 03) 00:10.0 Class ff00: 106b:0002 (rev 02) That's all what's shown from lspci on this machine. Startup messages: MacOS display is /bandit/formacGA5 Using unsupported 1024x768 formacGA5 at 92000800, depth=16, pitch=2048 Console: switching to colour frame buffer device 128x48 fb0: Open Firmware frame buffer device on /bandit/formacGA5 Christian. (Hope someone knows a solution? !...)
On Tue, May 02, Christian Jaeger wrote:
lspci 00:0b.0 Host bridge: Apple Computer Inc. Bandit PowerPC host bridge (rev 03) 00:10.0 Class ff00: Apple Computer Inc. Grand Central I/O (rev 02) lspci -n 00:0b.0 Class 0600: 106b:0001 (rev 03) 00:10.0 Class ff00: 106b:0002 (rev 02)
That's all what's shown from lspci on this machine.
It is an Onboard chip.
Startup messages:
MacOS display is /bandit/formacGA5 Using unsupported 1024x768 formacGA5 at 92000800, depth=16, pitch=2048 Console: switching to colour frame buffer device 128x48 fb0: Open Firmware frame buffer device on /bandit/formacGA5
You should uncheck the "No video driver" option in BootX. Gruss Olaf -- $ man clone BUGS Main feature not yet implemented...
At 23:16 Uhr +0200 2.5.2000, Olaf Hering wrote:
MacOS display is /bandit/formacGA5 Using unsupported 1024x768 formacGA5 at 92000800, depth=16, pitch=2048 Console: switching to colour frame buffer device 128x48 fb0: Open Firmware frame buffer device on /bandit/formacGA5
You should uncheck the "No video driver" option in BootX.
I HAVE this unchecked! (I suppose it uses open firmware because no native driver gets activated?) I don't give any kernel arguments, however. Which should I give? Christian.
On Wed, May 03, Christian Jaeger wrote:
At 23:16 Uhr +0200 2.5.2000, Olaf Hering wrote:
MacOS display is /bandit/formacGA5 Using unsupported 1024x768 formacGA5 at 92000800, depth=16, pitch=2048 Console: switching to colour frame buffer device 128x48 fb0: Open Firmware frame buffer device on /bandit/formacGA5
You should uncheck the "No video driver" option in BootX.
I HAVE this unchecked! (I suppose it uses open firmware because no native driver gets activated?)
I don't give any kernel arguments, however. Which should I give?
Ok, if it is really unchecked then no Framebuffer Driver can use the card and the kernel switch back to the OpenFirmware driver. Seems to be a unknown card. Can you check the type of the chip on the card? Maybe you find some entries in /proc/device-tree/bandit/formacGA5 with the name vendor_id. Just do a hexdump vendor_id. Or look at the chip itself. Gruss Olaf -- $ man clone BUGS Main feature not yet implemented...
At 0:04 Uhr +0200 4.5.2000, Olaf Hering wrote:
Can you check the type of the chip on the card? Maybe you find some entries in /proc/device-tree/bandit/formacGA5 with the name vendor_id. Just do a hexdump vendor_id.
There's NO '/proc/device-tree/bandit/formacGA5', and also no similar entry. I've done a find -name '*ormac*' and a grep -r ormac on /proc/device-tree/ and nothing was found. There is, however, /proc/fb which contains '0 OFfb /bandit/formacGA5'. You find a detailed listing at the end of this mail. (BTW: there are really two 'bandit' directories listed in /proc/device-tree/ ?!!)
Or look at the chip itself.
OK, I've looked into the machine. Could have done it earlier :-/: It's not an onboard chip, it's a PCI card. There's no other video connector on the back of this Mac than the connector at the PCI slot. ((Mal von einer nicht-verwendeten vorgestanzten Aussparung abgesehen, welche immer noch zu ist.)) So it seems that Apple delivered the machine with this card. The card has written 'Formac Pro Formace' and 'Part number PNGA 50-2' on it. I can't see on the chip itself, and I don't have a screwdriver here at the moment so I can't pull out the card. I'll look at it tomorrow. Christian. [root@nebelhorn chris]# cd /proc/device-tree/ [root@nebelhorn device-tree]# ll total 0 -r--r--r-- 1 root root 4 May 4 01:15 #address-cells -r--r--r-- 1 root root 4 May 4 01:15 #size-cells dr-xr-xr-x 13 root root 0 May 4 01:05 ./ dr-xr-xr-x 108 root root 0 May 4 2000 ../ dr-xr-xr-x 2 root root 0 May 4 01:15 AAPL,ROM/ -r--r--r-- 1 root root 4 May 4 01:15 AAPL,cpu-id dr-xr-xr-x 3 root root 0 May 4 01:15 PowerPC,60?/ dr-xr-xr-x 2 root root 0 May 4 01:15 aliases/ dr-xr-xr-x 4 root root 0 May 4 01:08 bandit/ dr-xr-xr-x 4 root root 0 May 4 01:08 bandit/ dr-xr-xr-x 2 root root 0 May 4 01:15 chosen/ -r--r--r-- 1 root root 4 May 4 01:15 clock-frequency -r--r--r-- 1 root root 18 May 4 01:15 compatible dr-xr-xr-x 2 root root 0 May 4 01:15 hammerhead/ dr-xr-xr-x 2 root root 0 May 4 01:15 memory/ -r--r--r-- 1 root root 16 May 4 01:15 model -r--r--r-- 1 root root 12 May 4 01:15 name dr-xr-xr-x 2 root root 0 May 4 01:15 openprom/ dr-xr-xr-x 2 root root 0 May 4 01:15 options/ dr-xr-xr-x 12 root root 0 May 4 01:15 packages/ [root@nebelhorn device-tree]# cd bandit/ [root@nebelhorn bandit]# ll total 0 -r--r--r-- 1 root root 4 May 4 01:16 #address-cells -r--r--r-- 1 root root 4 May 4 01:16 #size-cells dr-xr-xr-x 4 root root 0 May 4 01:08 ./ dr-xr-xr-x 13 root root 0 May 4 01:05 ../ -r--r--r-- 1 root root 4 May 4 01:16 AAPL,address -r--r--r-- 1 root root 4 May 4 01:16 AAPL,interrupts -r--r--r-- 1 root root 8 May 4 01:16 bus-range -r--r--r-- 1 root root 4 May 4 01:16 clock-frequency -r--r--r-- 1 root root 4 May 4 01:16 device_type -r--r--r-- 1 root root 24 May 4 01:16 driver-ist dr-xr-xr-x 10 root root 0 May 4 01:16 gc/ -r--r--r-- 1 root root 14 May 4 01:16 model -r--r--r-- 1 root root 7 May 4 01:16 name dr-xr-xr-x 2 root root 0 May 4 01:10 pci106b,1/ -r--r--r-- 1 root root 72 May 4 01:16 ranges -r--r--r-- 1 root root 8 May 4 01:16 reg -r--r--r-- 1 root root 13 May 4 01:16 slot-names [root@nebelhorn bandit]# cd pci106b,1/ [root@nebelhorn pci106b,1]# ll total 0 dr-xr-xr-x 2 root root 0 May 4 01:10 ./ dr-xr-xr-x 4 root root 0 May 4 01:08 ../ -r--r--r-- 1 root root 4 May 4 01:16 class-code -r--r--r-- 1 root root 4 May 4 01:16 device-id -r--r--r-- 1 root root 4 May 4 01:16 devsel-speed -r--r--r-- 1 root root 0 May 4 01:16 fast-back-to-back -r--r--r-- 1 root root 4 May 4 01:16 max-latency -r--r--r-- 1 root root 4 May 4 01:16 min-grant -r--r--r-- 1 root root 10 May 4 01:16 name -r--r--r-- 1 root root 20 May 4 01:16 reg -r--r--r-- 1 root root 4 May 4 01:16 revision-id -r--r--r-- 1 root root 4 May 4 01:16 vendor-id [root@nebelhorn pci106b,1]# cat class-code |xxd 0000000: 0006 0000 .... [root@nebelhorn pci106b,1]# cat device-id |xxd 0000000: 0000 0001 .... [root@nebelhorn pci106b,1]# cat devsel-speed |xxd 0000000: 0000 0001 .... [root@nebelhorn pci106b,1]# cat fast-back-to-back |xxd [root@nebelhorn pci106b,1]# cat max-latency |xxd 0000000: 0000 0000 .... [root@nebelhorn pci106b,1]# cat min-grant |xxd 0000000: 0000 0000 .... [root@nebelhorn pci106b,1]# cat name |xxd 0000000: 7063 6931 3036 622c 3100 pci106b,1. [root@nebelhorn pci106b,1]# cat reg|xxd 0000000: 0000 5800 0000 0000 0000 0000 0000 0000 ..X............. 0000010: 0000 0000 .... [root@nebelhorn pci106b,1]# cat revision-id |xxd 0000000: 0000 0003 .... [root@nebelhorn pci106b,1]# cat vendor-id |xxd 0000000: 0000 106b ...k [root@nebelhorn pci106b,1]# -- | __ . _ |_ | /| / /_\ | \ Christian Jaeger |/ |/ \_. |_/ -Programmierung Rektorat, ETH Zentrum, HG F 69.2 Burgstr. 26 8092 Zürich, 01/63 2 2062 8037 Zürich, 01/273 65 46 www.rektorat.ethz.ch
At 0:04 Uhr +0200 4.5.2000, Olaf Hering wrote:
Can you check the type of the chip on the card? Maybe you find some entries in /proc/device-tree/bandit/formacGA5 with the name vendor_id. Just do a hexdump vendor_id.
Or look at the chip itself.
OK, I've finally looked at the chip(s): I'm not completely sure which one is the important one, there are two big chips (probably it's the former..) - Imagine 128 Series II Number Nine C 9646XF044 - (Texas Instruments) VIDEO TVP3030 - 220PPA 9718706 76A769W BTW: X is not only slow here because of slow drawing, there is also a delay of around 1 second for each window to be opened (i.e. opening a new window from within NEdit (which normally is very fast), or file open/save dialogs all take one extra second of time). Is this normal for unaccelerated X or is there also another (configuration) problem? (The same system works very well on my BlueG3 at home) Vielen Dank .Christian. *-----------------------------------------------------------------------* Web Office ETH Zuerich Christian Jaeger Abteilung Oeffentlichkeitsarbeit (AOA) ETH Zentrum CH-8092 Zurich office: HG J46 e-mail: weboffice@ethz.ch phone: +41 (0)1 63 2 5780 fax: +41 (0)1 63 2 3525 home: (0)1 273 65 46 *-----------------------------------------------------------------------*
participants (3)
-
Christian Jaeger -
Christian Jaeger -
Olaf Hering