GeForce4 Ti 4200 Redhat 8.0 fresh install question - Page 2

bwkaz · 01-13-03, 07:54 AM

Quote:

Originally posted by Tom1976

Code:

CPU0
  0:    1259321          XT-PIC  timer
  1:       1984          XT-PIC  keyboard
  2:          0          XT-PIC  cascade
  5:     117573          XT-PIC  usb-uhci, serial
  6:          0          XT-PIC  usb-uhci, usb-uhci
  8:          1          XT-PIC  rtc
 10:          0          XT-PIC  ehci-hcd, cmpci
 11:     207302          XT-PIC  usb-uhci, nvidia
 12:      82134          XT-PIC  PS/2 Mouse
 14:      11672          XT-PIC  ide0
 15:      47917          XT-PIC  ide1
NMI:          0
ERR:          0

Have you tried disabling extra USB ports? It seems you have four controllers installed (

); does disabling two of them that you don't use change the IRQ sharing?

Tom1976 · 01-13-03, 08:26 AM

How do I disable them? Is it a BIOS setting?

bwkaz · 01-13-03, 11:52 AM

Yes, in the BIOS. I don't know where -- that depends on your chipset and BIOS writer (AMI, Award, Phoenix, or whatever else).

Tom1976 · 01-13-03, 05:53 PM

My ASUS board uses the Award BIOS.
Browsing through the menu, I only found one field related to USB, the "USB Legacy Support" field. By default it was enabled, so I set it to disable.

During boot, when it got to "detecting new hardware", I was taken to a drop USB screen. I selected drop, and the boot continued. Here is my /proc/interrupt now:

CPU0
0: 545770 XT-PIC timer
1: 2418 XT-PIC keyboard
2: 0 XT-PIC cascade
5: 69338 XT-PIC usb-uhci, serial
8: 1 XT-PIC rtc
10: 0 XT-PIC ehci-hcd, cmpci
11: 87521 XT-PIC usb-uhci, nvidia
12: 62140 XT-PIC PS/2 Mouse
14: 10862 XT-PIC ide0
15: 19877 XT-PIC ide1
NMI: 0
ERR: 0

yes, a couple usb's are gone, but that one still sharing on 11 is still there......

Ideas?

bwkaz · 01-13-03, 09:16 PM

:-/

Rats, I was hoping that the USB controller that it got rid of would be the right one.

Any way to force the other controller to IRQ 6?

No, not really any ideas, unfortunately, unless modinfo usb-uhci tells you of some way to force a certain IRQ...

junkieclown · 01-14-03, 03:56 PM

yesterday I was able to check my ti4200 on another system.
The system was based on a SIS 735 chipset. interestingly, Quake3 Linux version played on this other system for nearly a minute before crashing. The crashing screen looked exactly like it does on my Asus KT333 based system. Blue-green chunks.

Then I installed the card's drivers on the Windows98 system that shares that sis735 hardware. At first I thought aha, bad hardware, since NeedForSpeed / Porsche Unleashed crashed soon after starting. Couple of minutes of play. However this was right after completing the card installation. And I had accidentally left windows at 256 colors. And this is Windows98 after all: it's not really supposed to be stable. After a reboot or two to adjust colors in Windows to 32bit, the card stayed stable while playing that game for multiple races. I couldn't play it hour after hour to check stability under stressful conditions, as there wasn't time, but it was clearly not exhibiting the "crash in 20 seconds or less" behavior it has under Linux.
Unfortunately that was about the only 3d game installed on that Windose system.

I guess my options are to hope Nvidia comes across with support for their product in a timely fashion (oops too late for that! ) or just to chalk this up to experience and buy some Intel hardware they feel like supporting.

junkieclown · 01-14-03, 04:07 PM

Quote:

Originally posted by Tom1976
My ASUS board uses the Award BIOS.
Browsing through the menu, I only found one field related to USB, the "USB Legacy Support" field. By default it was enabled, so I set it to disable.

During boot, when it got to "detecting new hardware", I was taken to a drop USB screen. I selected drop, and the boot continued. Here is my /proc/interrupt now:

CPU0
0: 545770 XT-PIC timer
1: 2418 XT-PIC keyboard
2: 0 XT-PIC cascade
5: 69338 XT-PIC usb-uhci, serial
8: 1 XT-PIC rtc
10: 0 XT-PIC ehci-hcd, cmpci
11: 87521 XT-PIC usb-uhci, nvidia
12: 62140 XT-PIC PS/2 Mouse
14: 10862 XT-PIC ide0
15: 19877 XT-PIC ide1
NMI: 0
ERR: 0

yes, a couple usb's are gone, but that one still sharing on 11 is still there......

Ideas?

Do you have peripherals that absolutely need USB2 ?
Your usb2 controller (ehci) is still in the mix. I got the Nvidia to its own interrupt not by turning off usb this or that in the BIOS, but by disabling the USB2 hardware. There's a jumper on the bottom edge of the motherboard that you change (page 24 of my asus a7v333 manual).
Changing that jumper and enabling "reset configuration data" (under the "boot" screen of the bIOS setup utility) caused all the PCI hardware to be re-recognized and with USB2 disabled at the electrical level, that left Nvida all alone on irq11.
You still have USB 1 type support and you can leave it enabled it in the BIOS.

However, I am not at all sure that this even matters. The Nvidia drivers just do not support our hardware and we're SOL until they do.

bwkaz · 01-14-03, 04:57 PM

Quote:

Originally posted by junkieclown
The Nvidia drivers just do not support our hardware and we're SOL until they do.

That's funny, because my KT333 works just fine with the nVidia drivers. Of course, it's not an Asus board, either, it's a cheap-a** Biostar, so that might have something to do with it...

MSI GF4 Ti 4200 128MB as well...

junkieclown · 01-14-03, 05:59 PM

do you use agpgart or nvidia's builtin agp support?
did you mess with drive strength settings in any way?
do anything unusual to get the drivers working?

what kernel are you using?

are you passing any options to the kernel via lilo/grub?

what i find "funny" is that tom and i have perfectly good hardware, we're not idiots who're incapable of reading documents, and the drivers do NOT work for us in 3d.
With my GF2 MX board, sure it works great. With my GF4 ti board, no it's not working.

I will continue to regard that as non-support for my Nvidia and related standard hardware until they fix it. After all, it works fine in Windows for Tom on the same hardware.
(PS-It seems that nvidia develops gart drivers for Windows. But you can see in the release notes for Linux no mention of Kt333 support unless it's lumped in with a generic support of Kt266...kt333 is hardly new, so i assume that's what they're doing.
Thing is, the gf4 crashes under agpgart just the same.)

nicoe · 01-14-03, 06:24 PM

Hello,

I've got the same problems as you have but I cannot test if mine is hardware related or not
(I don't have windows and don't have the opportunity to install it).

My config :

ASUS A7V33
Chaintech GeForce 4 Ti 4200 64M
Athlon XP 1800
768M DDR SDRAM
Debian Sarge

I compiled two kernel

2.4.19 :
no mtrr, no agp, no rtc

2.4.20 :
mtrr, no agp, rtc

My bios settings are the following :
Everything default except :
Video Memory Cache Mode : UC

XFree settings :
Default + Use of Nvidia AGP

For booting I append this line to lilo : pci=biosirq but doesn't have any effect. As you can see :

nicoe@smarties ~ % grep nvidia /proc/interrupts
11: 222046 XT-PIC usb-uhci, nvidia

The strange thing is that with the kernel 2.4.19, AGP is disabled but I can play (bzflag) for hours without problems.

With the 2.4.20 kernel, AGP is enabled but as soon as I play a little (20-30s or so) -> Freeze.
But I can launch glxgears without any problems.

bwkaz · 01-14-03, 06:43 PM

Sorry, should have added stuff like that.

I use agpgart, and (at least on this box) kernel 2.4.18 and .19. I've used the drivers all the way back to 2.4.4 or so, though, just not with the KT333 chipset (obviously -- it's newer than that).

No messing with drive strength. It's set to I believe E9 (possibly set to auto-try-to-detect and E9 is the current setting), but don't quote me on that. I have no idea what drive strength even changes, but I do remember someone saying that setting either nibble to F causes problems... in any case, I don't know what it does, so I don't know for sure what it's set at.

Anything unusual? No, just compiled from tarballs, exactly like every nVidia driver I've used since way back around 0.9-769 or so. I did set fast writes and SBA to enabled in os-registry.c (but only because I didn't want to pass the option on every autoprobe).

Kernel, again, currently 2.4.19 plus -preempt (I applied -preempt myself, but it worked with vanilla 2.4.19 also). I should try 2.4.20 one day, but won't soon.

It has MTRR support, agpgart a module, rtc a module (both loaded); if you want, I can post/attach the full .config file.

/proc/cmdline shows:

BOOT_IMAGE=LFS-gcc3 ro root=305 hdc=ide-scsi

Using lilo.

And my /proc/interrupts looks like:

Code:

           CPU0       
  0:   17447979          XT-PIC  timer
  1:     170380          XT-PIC  keyboard
  2:          0          XT-PIC  cascade
  8:          1          XT-PIC  rtc
 10:     959909          XT-PIC  usb-uhci, usb-uhci, eth0
 11:    6987900          XT-PIC  nvidia
 12:    1133045          XT-PIC  YMF724F
 14:     418556          XT-PIC  ide0
 15:      11510          XT-PIC  ide1
NMI:          0 
ERR:          0

so nvidia isn't sharing anything with anybody.

Oh yeah, now that I think about it, this might be relevant -- I'm still using 3123. I tried 4191 a couple days after release, and got the same error that I had been getting with 2960, "__divdi3 not found in libc6.so with link time reference" messages when I try to load the libglx.so X module (it's a problem with gcc3-compiled glibc/libstdc++/something like that). It took a driver upgrade last time to fix it, and I'm waiting for word from nVidia on when it'll get fixed again -- probably in the next release.

If there's anything else you want to know, ask away, of course.

junkieclown · 01-15-03, 02:20 PM

thanks for the additional information.

I have tried the 1.0-3xxx driver (using it now) with my Ti-4200 as a result of your post. Unfortunately this also crashes with my system: setting BIOS to use normal "PIC" interrupts, accepting setup defaults in the settings having to do with AGP, setting NvAGP to 0 in the XF86Config....
tuxracer crashes. RTCW crashes.
glxgears doesn't crash --of course I don't run it for very long, but inspired by that I changed tuxracer to run in a window and when run that way, Tuxracer won't crash.

Fullscreen OpenGL anything crashes with my Ti4200, regardless of agp setting. GForce2 is fine as always.

This IS a driver issue.

01-13-03, 11:52 AM	#15
bwkaz Registered User Join Date: Sep 2002 Posts: 2,262	Yes, in the BIOS. I don't know where -- that depends on your chipset and BIOS writer (AMI, Award, Phoenix, or whatever else). __________________ Registered Linux User #219692

01-13-03, 09:16 PM	#17
bwkaz Registered User Join Date: Sep 2002 Posts: 2,262	:-/ Rats, I was hoping that the USB controller that it got rid of would be the right one. Any way to force the other controller to IRQ 6? No, not really any ideas, unfortunately, unless modinfo usb-uhci tells you of some way to force a certain IRQ... __________________ Registered Linux User #219692

01-14-03, 03:56 PM	#18
junkieclown Helium Abuser Join Date: Nov 2002 Location: Summers in Rangoon, but mostly in my evil underground lair and Volcano laboratory. though Posts: 45	yesterday I was able to check my ti4200 on another system. The system was based on a SIS 735 chipset. interestingly, Quake3 Linux version played on this other system for nearly a minute before crashing. The crashing screen looked exactly like it does on my Asus KT333 based system. Blue-green chunks. Then I installed the card's drivers on the Windows98 system that shares that sis735 hardware. At first I thought aha, bad hardware, since NeedForSpeed / Porsche Unleashed crashed soon after starting. Couple of minutes of play. However this was right after completing the card installation. And I had accidentally left windows at 256 colors. And this is Windows98 after all: it's not really supposed to be stable. After a reboot or two to adjust colors in Windows to 32bit, the card stayed stable while playing that game for multiple races. I couldn't play it hour after hour to check stability under stressful conditions, as there wasn't time, but it was clearly not exhibiting the "crash in 20 seconds or less" behavior it has under Linux. Unfortunately that was about the only 3d game installed on that Windose system. I guess my options are to hope Nvidia comes across with support for their product in a timely fashion (oops too late for that! ) or just to chalk this up to experience and buy some Intel hardware they feel like supporting. Last edited by junkieclown; 01-14-03 at 04:30 PM.

01-14-03, 05:59 PM	#21
junkieclown Helium Abuser Join Date: Nov 2002 Location: Summers in Rangoon, but mostly in my evil underground lair and Volcano laboratory. though Posts: 45	do you use agpgart or nvidia's builtin agp support? did you mess with drive strength settings in any way? do anything unusual to get the drivers working? what kernel are you using? are you passing any options to the kernel via lilo/grub? what i find "funny" is that tom and i have perfectly good hardware, we're not idiots who're incapable of reading documents, and the drivers do NOT work for us in 3d. With my GF2 MX board, sure it works great. With my GF4 ti board, no it's not working. I will continue to regard that as non-support for my Nvidia and related standard hardware until they fix it. After all, it works fine in Windows for Tom on the same hardware. (PS-It seems that nvidia develops gart drivers for Windows. But you can see in the release notes for Linux no mention of Kt333 support unless it's lumped in with a generic support of Kt266...kt333 is hardly new, so i assume that's what they're doing. Thing is, the gf4 crashes under agpgart just the same.) Last edited by junkieclown; 01-14-03 at 06:56 PM.

01-14-03, 06:43 PM	#23
bwkaz Registered User Join Date: Sep 2002 Posts: 2,262	Sorry, should have added stuff like that. I use agpgart, and (at least on this box) kernel 2.4.18 and .19. I've used the drivers all the way back to 2.4.4 or so, though, just not with the KT333 chipset (obviously -- it's newer than that). No messing with drive strength. It's set to I believe E9 (possibly set to auto-try-to-detect and E9 is the current setting), but don't quote me on that. I have no idea what drive strength even changes, but I do remember someone saying that setting either nibble to F causes problems... in any case, I don't know what it does, so I don't know for sure what it's set at. Anything unusual? No, just compiled from tarballs, exactly like every nVidia driver I've used since way back around 0.9-769 or so. I did set fast writes and SBA to enabled in os-registry.c (but only because I didn't want to pass the option on every autoprobe). Kernel, again, currently 2.4.19 plus -preempt (I applied -preempt myself, but it worked with vanilla 2.4.19 also). I should try 2.4.20 one day, but won't soon. It has MTRR support, agpgart a module, rtc a module (both loaded); if you want, I can post/attach the full .config file. /proc/cmdline shows: BOOT_IMAGE=LFS-gcc3 ro root=305 hdc=ide-scsi Using lilo. And my /proc/interrupts looks like: Code: CPU0 0: 17447979 XT-PIC timer 1: 170380 XT-PIC keyboard 2: 0 XT-PIC cascade 8: 1 XT-PIC rtc 10: 959909 XT-PIC usb-uhci, usb-uhci, eth0 11: 6987900 XT-PIC nvidia 12: 1133045 XT-PIC YMF724F 14: 418556 XT-PIC ide0 15: 11510 XT-PIC ide1 NMI: 0 ERR: 0 so nvidia isn't sharing anything with anybody. Oh yeah, now that I think about it, this might be relevant -- I'm still using 3123. I tried 4191 a couple days after release, and got the same error that I had been getting with 2960, "__divdi3 not found in libc6.so with link time reference" messages when I try to load the libglx.so X module (it's a problem with gcc3-compiled glibc/libstdc++/something like that). It took a driver upgrade last time to fix it, and I'm waiting for word from nVidia on when it'll get fixed again -- probably in the next release. If there's anything else you want to know, ask away, of course. __________________ Registered Linux User #219692

01-13-03, 08:26 AM	#14
Tom1976 Registered User Join Date: Jan 2003 Posts: 13	How do I disable them? Is it a BIOS setting?

01-13-03, 05:53 PM	#16
Tom1976 Registered User Join Date: Jan 2003 Posts: 13	My ASUS board uses the Award BIOS. Browsing through the menu, I only found one field related to USB, the "USB Legacy Support" field. By default it was enabled, so I set it to disable. During boot, when it got to "detecting new hardware", I was taken to a drop USB screen. I selected drop, and the boot continued. Here is my /proc/interrupt now: CPU0 0: 545770 XT-PIC timer 1: 2418 XT-PIC keyboard 2: 0 XT-PIC cascade 5: 69338 XT-PIC usb-uhci, serial 8: 1 XT-PIC rtc 10: 0 XT-PIC ehci-hcd, cmpci 11: 87521 XT-PIC usb-uhci, nvidia 12: 62140 XT-PIC PS/2 Mouse 14: 10862 XT-PIC ide0 15: 19877 XT-PIC ide1 NMI: 0 ERR: 0 yes, a couple usb's are gone, but that one still sharing on 11 is still there...... Ideas?

01-14-03, 06:24 PM	#22
nicoe Registered User Join Date: Jan 2003 Posts: 3	Hello, I've got the same problems as you have but I cannot test if mine is hardware related or not (I don't have windows and don't have the opportunity to install it). My config : ASUS A7V33 Chaintech GeForce 4 Ti 4200 64M Athlon XP 1800 768M DDR SDRAM Debian Sarge I compiled two kernel 2.4.19 : no mtrr, no agp, no rtc 2.4.20 : mtrr, no agp, rtc My bios settings are the following : Everything default except : Video Memory Cache Mode : UC XFree settings : Default + Use of Nvidia AGP For booting I append this line to lilo : pci=biosirq but doesn't have any effect. As you can see : nicoe@smarties ~ % grep nvidia /proc/interrupts 11: 222046 XT-PIC usb-uhci, nvidia The strange thing is that with the kernel 2.4.19, AGP is disabled but I can play (bzflag) for hours without problems. With the 2.4.20 kernel, AGP is enabled but as soon as I play a little (20-30s or so) -> Freeze. But I can launch glxgears without any problems.

01-15-03, 02:20 PM	#24
junkieclown Helium Abuser Join Date: Nov 2002 Location: Summers in Rangoon, but mostly in my evil underground lair and Volcano laboratory. though Posts: 45	thanks for the additional information. I have tried the 1.0-3xxx driver (using it now) with my Ti-4200 as a result of your post. Unfortunately this also crashes with my system: setting BIOS to use normal "PIC" interrupts, accepting setup defaults in the settings having to do with AGP, setting NvAGP to 0 in the XF86Config.... tuxracer crashes. RTCW crashes. glxgears doesn't crash --of course I don't run it for very long, but inspired by that I changed tuxracer to run in a window and when run that way, Tuxracer won't crash. Fullscreen OpenGL anything crashes with my Ti4200, regardless of agp setting. GForce2 is fine as always. This IS a driver issue.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Redhat 8.0 NVIDIA works - INSTRUCTIONS	STEEL1	NVIDIA Linux	267	04-15-03 06:48 PM
nForce + RedHat 7.3 Success Story	rtz	NVIDIA Linux	47	02-15-03 02:20 AM
nforce on RedHat 8.0	shaun680	NVIDIA Linux	8	10-17-02 12:36 PM
For Sale: Gainward GeForce4 Ti 4200	MikeC	For Sale/Trade	9	10-09-02 03:56 PM
Redhat 8.0, no rpm, can't compile.	Ironphil	NVIDIA Linux	6	10-02-02 03:13 PM