X don't run correctly with new nvidia drivers (nvidia-glx-1.0-6629 and nvidia-kernel-1.0-6629). I get only a blank screen. The only workaround seems to be the 'Option "NvAGP" "0"' in /etc/X11/xorg.conf (but this obviously is a non-long term solution) Reproducible: Always Steps to Reproduce: 1. install sys-kernel/gentoo-dev-sources-2.6.9-r* 2. install x11-base/xorg-x11-6.8.0-r1 and configure xorg.conf to use "nvidia" X driver 3. install media-video/nvidia-kernel-1.0-6629 and media-video/nvidia-glx-1.0-6629 and load "nvidia.ko" module 4. run "startx" with default option Actual Results: I see only a blank screen. I don't see any graphics or text at all. Expected Results: X should work :) M/B: ASUS A7V133 (AGP 4X) Video Card: GeForce FX 5200 (AGP 8X)
Created attachment 43685 [details] Xorg log file when I've a blank screen
Created attachment 43686 [details] "emerge --info" list
Created attachment 43688 [details] nvidia-bug-report.log (without Xorg.0.log's chunk already submitted)
Can you try a different kernel (any 2.6 will do, but if the problem persists please try development-sources)?
I've tried the following combinations (with NvAGP "1" in xorg.conf, or NvAGP "3" & agpgart kernel module loaded): X doesn't work with 1.0.6629* and: gentoo-sources-2.4.27-r2, vanilla 2.6.7, vanilla 2.6.8, vanilla 2.6.9, development-sources-2.6.10-rc1 BTW with previous drivers 1.0.6111 I get the same behaviour with this video card. After running startx, after about 20-30 secs, I see some signs of hard disk activity (as if booted the window-manager). The keyboard seems to respond, but if I press CTRL+ALT+BACKSPACE and then CTRL+F* it stops responding.
Created attachment 43735 [details] complete and more recent log
Note: Before buying this GeForce FX 5200 I had a GeForce2 MX: X worked/booted without problems with nvidia-{kernel|glx} 1.0.6111 (in default NvAGP "3") and gentoo-sources-2.4.26-r9, openmosix-sources-2.4.27-r2, gentoo-dev-sources-2.6.9-r1 --- Now in /proc/driver/nvidia/agp/status I read a "disabled" status, but in BIOS AGP results working at 4X I've tried even some nvidia.ko module options (NVreg_EnableAGPSBA=1 NVreg_EnableAGPFW=1 NVreg_EnableVia4x=1 NVreg_ReqAGPRate=4) with no luck Should I suppose that the problem is related to AGP not enabled by nvidia{.ko|.o} kernel module in newer or almost newer (perhaps all 8X AGP) nvidia-based graphics cards, therefore putting NvAGP in xorg.conf to a value > 0 leads to blank screen problems?
Maybe have a look here (official forum for linux drivers): http://www.nvnews.net/vbulletin/showthread.php?t=40106 and post your problem there ... If zander posts any patches, we will gladly apply them.
I get the same behavior under both linux-2.6.8-gentoo-r4 and linux-2.6.9-gentoo-r1 on AMD64. Actually, it doesn't lock up all of the way. It will start to boot and if you wait a little while (< 1min) it will continue booting, but then repeat the lock up cycle. Then I grew impatient waiting for another period of stability and rebooted. It seems that the driver is locking up periodically, but I can't find any Xorg.log info different between 6111 and 6629 other than the following. (Note I have 3 NVidia cards in my system). < (WW) NVIDIA(1): WAIT (2, 1, 0x800004x, 0x08x, 0x08x) < (WW) NVIDIA(1): WAIT (1, 1, 0x800004x, 0x08x, 0x08x) < (WW) NVIDIA(1): WAIT (2, 6, 0x800004x, 0x08x, 0x08x) < (WW) NVIDIA(1): WAIT (1, 6, 0x800004x, 0x08x, 0x08x)
Maybe I got the point. Previously I had tried also with the stock NVIDIA-Linux-x86-1.0-6629-pkg1.run (installed without gentoo's patches): same blank screen (= surely the patches were not the problem). My problems (agp status disabled = agp card recognized as a PCI-like one, and forcing agp mode transfer with NvAGP > 0), probably, were due to modprobe nvidia.ko _before_ agpgart.ko or not to modprobe agpgart.ko at all. I've noticed that AGP is correctly enabled if I load first agpgart.ko and only then nvidia.ko (with NvAGP "3" in xorg.conf). BTW I've added also a '*' to alias char-major-195 nvidia ("alias char-major-195* nvidia") in /etc/modules.d/nvidia but I don't think I really need this one (I've only one nvidia card). Now I'm using development-sources-2.6.10-rc1, but IMHO it doesn't matter. cat /proc/driver/nvidia/agp/status Status: Enabled <-- Driver: NVIDIA <-- AGP Rate: 4x Fast Writes: Enabled SBA: Enabled Indeed, my agp problem in gentoo seems disappeared, for me this bug could be marked as "resolved", thanks to all :))
Jacob: What about you? Does comment #10 apply aswell
Still happening even with all of the fixes described in Comment 10. However, I still get (WW) NVIDIA(1): WAIT (2, 1, 0x800004x, 0x08x, 0x08x) In the Xorg log. I think I had the option set in the right place in XF86Config to not use the AGPGART module. But it seems to be loaded when I am running with 6111, and it may be compiled into the kernel #cat /proc/driver/nvidia/agp/status Status: Enabled Driver: AGPGART AGP Rate: 8x Fast Writes: Enabled SBA: Enabled
Jacob can you try configuring your cards so that one using the nvidia driver and the others use the "nv" driver.. Im wondering if there is some resource driver based conflict or if its something else.
Bingo. I did what you said in Comment 13 (made GeForce4 use nvidia and others use nv), and it booted up fine I guess the resources that are being contended for are listed in Comment 9. The following may be interesting because two cards use IRQ 16. mygentoo martin $ cat /proc/driver/nvidia/cards/0 Model: GeForce2 MX/MX 400 IRQ: 16 Video BIOS: ??.??.??.??.?? Card Type: PCI mygentoo martin $ cat /proc/driver/nvidia/cards/1 Model: GeForce FX 5200 IRQ: 18 Video BIOS: ??.??.??.??.?? Card Type: PCI mygentoo martin $ cat /proc/driver/nvidia/cards/2 Model: GeForce4 Ti 4200 with AGP8X IRQ: 16 Video BIOS: 04.28.40.10.00 Card Type: AGP
Several problems still: 1. Using the "nvidia" and "nv" drivers at the same time leads to an X memory leak. (need a new bug report?) 2. Removing one of the cards on IRQ 16 does not fix the problem. I suspect this indicates that the real problem is with using multiple "nvidia" drivers at the same time. I'm surprised no one else is having this problem. Note that 6111 drivers do NOT have this problem!!! It first appeared in 6629
Is there still a problem here? If so, explain.
I do not have anymore problems. This started working a while ago, and now definitely =works for me with: media-video/nvidia-kernel-1.0.8178-r3
Problem's fixed according to one so we'll mark it as such.