Howto install old driver (3123) on Red Hat 9?

[Gemini]

I have a GeForce 3 card made by Creative. New NVidia drivers hang Linux on this system, so I have been using 1.0-3123 until now. I upgraded the machine to Red Hat Linux 9 and noticed that I cannot build from the 3123 source RPMs any more because of the new threading implementation. Is there any way around this? I have tried every release after 3123, including the latest RH 9 compatible one, and they all hang the system when X is started. It will respond to ping over the network but nothing else. I have never found out if this is a known problem with the card, which seems to me to be of a common make (Creative 3d Blaster GeForce3 Titanium 500). The motherboard is Abit TH7II. Lspci -vv says

01:00.0 VGA compatible controller: nVidia Corporation NV20 [GeForce3 Ti 500] (rev a3) (prog-if 00 [VGA])
Subsystem: Micro-Star International Co., Ltd.: Unknown device 5305
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66Mhz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 32 (1250ns min, 250ns max)
Interrupt: pin A routed to IRQ 10
Region 0: Memory at dc000000 (32-bit, non-prefetchable) [size=16M]
Region 1: Memory at d4000000 (32-bit, prefetchable) [size=64M]
Region 2: Memory at d8000000 (32-bit, prefetchable) [size=512K]
Expansion ROM at <unassigned> [disabled] [size=64K]
Capabilities: [60] Power Management version 2
Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Capabilities: [44] AGP version 2.0
Status: RQ=31 SBA- 64bit- FW- Rate=x1,x2,x4
Command: RQ=0 SBA- AGP- 64bit- FW- Rate=<none>


The version 3123 driver did pretty much everything I wanted it to, so I do not necessarily need the latest and greatest driver. If anyone has an idea how to build the old driver on RH9, it would be greatly appreciated.


... Forgot to mention the kernel version, which is Red Hat's RPM-packaged 2.4.20-9. I have had the hang problem on Mandrake 8.2, 9.0 and 9.1 as well as Red Hat 8.0, with various vendor and custom kernels, so I think the hang is caused by a hardware problem.

[WaxyLemon]

The short answer is 3123 wont work in RedHat 9. If you compile a custom kernel from kernel.org, sans redhat patches, then you can build 3123. But then GLX will be broken anyway.

So I have RedHat 9, Linux 2.4.21-pre7, and NVIDIA 1.0-3123 with no OpenGL support. The only reason why I don't just use the open source nv driver is because I need DVI and TwinView support.

Waiting patiently for the day that I can have openGL, twinview, and an uptime of > 1 hour.

[Gemini]

The reason I do not just use the open source nv driver is that OpenGL 3d performance matters a great deal to me (3d molecular graphics)... So is the system hang a known problem with GeForce3 (NV20) and drivers newer than 3123? On my box, it is not some difficult-to-reproduce random thing, but rather just starting X will immediately hang the system every time, with every driver/kernel combination. Do I just happen to have a rare bad card or is this something so common that NVidia might actually fix it in a new driver release?

I still have Red Hat 8.0 root partition intact on the hard drive, so I guess I'll go back to that for the moment.

[dkosovic]

You could try building NVIDIA_kernel-1.0-3123.phoebe.src.rpm available from:

http://rebus.webz.cz/

RedHat Phoebe-3 uses the same new kernel threads as RedHat 9.0. Follow the instructions on how to disable TLS.


Doug.

[Gemini]

Originally posted by dkosovic
You could try building NVIDIA_kernel-1.0-3123.phoebe.src.rpm available from:

http://rebus.webz.cz/

RedHat Phoebe-3 uses the same new kernel threads as RedHat 9.0. Follow the instructions on how to disable TLS.

That worked for me. Thanks a lot! I had to disable TLS from /etc/profile for convenience, which is unfortunate, but there you are.

Does NVidia care at all about making new Linux drivers work with these slightly older cards?

Cheers & thanks again.