I believe this is occurring before the nvidia driver gets involved. Have you tried downgrading nvidia? On Oct 27, 2010 11:15 PM, "Heiko Baums" wrote:

Am Wed, 27 Oct 2010 21:36:52 -0500 schrieb "David C. Rankin" :

...

Guys,

To add to the nvidia issues, after update to the latest driver (nvidia 260.19.12-1) the machine stops during boot due to (Invalid video mode, press enter to see list....)

Pressing enter then lists the available modes. However, when I enter one of the listed modes, it is rejected and I get prompted again with the (Invalid video mode, press enter to see list....).

I don't know whether this is a bug, a KMS thing, or what, but I have never had any problems passing vga=0x31a on the kernel line with the nvidia driver before. (I know with ATI, KMS early is recommended, and no vga= on the kernel line)

Is anybody else seeing this? Should I just remove the vga= line?

After letting the (Invalid video mode, press enter to see list....) prompt time-out, it all continues fine and the nvidia driver loads without issue. What say the experts?

There's an option to scanning your video card for valid values. If I recall correctly you just can enter "scan" at that prompt. Then you can set vga= to one of the possible values.

But if you're using KMS you need to remove the vga kernel parameter. If you don't want to use KMS, add the kernel parameters vga and nomodeset.

Heiko

Am Thu, 28 Oct 2010 00:06:53 -0500 schrieb "David C. Rankin" :

Not yet, but I don't know what else could have caused this. I have used scan (same thing as just pressing enter) and I get about 50 valid modes to enter. Everything thing from vga to 1920x1200. I have tried entering the listed modes (e.g. 31A) and it just gives me the same 'invalid mode' prompt.

I've been through the KMS page that just refers you to https://wiki.archlinux.org/index.php/NVIDIA for nvidia cards. This card is an 8800GT so I have always used:

# pacman -S nvidia nvidia-utils

to install the driver and I've never had any issues. I haven't specifically set anything for KMS with nvidia boxes. So I'm at a loss at what changed.

The problem I have is that I don't know what process does the mode check and then throws the invalid mode error? Is it a kernel process, nvidia driver process. (I suspect a driver process since it is setting the framebuffer, but I could be totally wrong)

I'm open to all suggestions. I'll give the downgrade a try tomorrow or Friday when I get a chance. Thanks for the suggestions.

This problem has nothing to do with the nvidia driver. The nvidia driver is only for X. And entering "scan" at this prompt is not the same as just pressing Enter. Enter shows you all possible vesa modes. Entering "scan" shows you only the modes which are valid for your video card. And you need to decide whether you want to use KMS or not. If you want to use KMS you need to remove the vga kernel parameter. Otherwise you need to add the two kernel parameters vga and nomodeset (both of them) to the kernel line of your bootloader. Heiko

I do have the /etc/modprobe.d/nouveau_blacklist.conf file on my machine, too, but I remember that the nouveau kernel driver did get loaded until I added !nouveau to the MODULES list in my rc.conf. So 1) Can you do an "lsmod | grep nouveau" to see whether nouveau maybe still gets loaded. (If so, it might be interesting to learn why even though the above file is present.) 2) If the answer to (1) is yes, try adding !nouveau to the MODULES list in rc.conf and see whether this fixes your problem. Of course, I'm just guessing here. Cheers, Norbert David C. Rankin [2010.10.28 1803 -0500]:

On 10/28/2010 04:38 AM, Dan Vrátil wrote:

...

Hello David,

this cannot be related to nvidia drivers because of two things: 1) the drivers are loaded after the kernel boots and 2) the proprietary nvidia drivers do not support KMS (and according to nvidia they are not going too in near future).

Make sure, you have nouveau blacklisted (/etc/modprobe.d/nouveau_blacklist.conf) because you don't need it when you use the proprietary drivers and last time I tested it, it had problems with the vga boot option.

Dan

Thank you Heika, Attila, Dan,

The mystery still remains. I didn't do it, but noveau was already blacklisted on my box:

[17:32 archangel:/etc/modprobe.d] # l total 32 drwxr-xr-x 2 root root 4096 Oct 19 19:34 . drwxr-xr-x 96 root root 12288 Oct 28 00:49 .. -rw-r--r-- 1 root root 704 Jun 28 08:45 framebuffer_blacklist.conf -rw-r--r-- 1 root root 55 Jun 7 2009 modprobe.conf -rw-r--r-- 1 root root 18 Oct 18 09:28 nouveau_blacklist.conf -rw-r--r-- 1 root root 196 Sep 18 17:58 usb-load-ehci-first.conf [17:32 archangel:/etc/modprobe.d] # cat nouveau_blacklist.conf blacklist nouveau

Also, on the (press enter or scan issue), I'm sure you are 100% correct Heiko, but no matter whether I press enter or issue scan with this 8800GT, the list of modes returned is the same. Also, both enter or scan take the same amount of time to complete -- at least on this box.

If the proprietary driver doesn't support KMS and isn't going to for a while, then I wonder what changed to make my normal kernel line (that I've used for years) suddenly start causing the invalid mode error? The kernel line is:

kernel /vmlinuz26 root=/dev/mapper/nvidia_baaccajap5 ro vga=0x31a

As a test, I booted LTS with the vga=0x31a and it booted fine and into the correct framebuffer. So this is looking more and more kernel related.

I have also booted Arch without vga= in the kernel line and I do NOT get the invalid mode prompt, but then I'm left with a 640x480 framebuffer. That looks terrible on a 23" display.

Oddly, when I boot with vga= and get the error, somehow after the timeout, I get the right framebuffer.

I'm stumped. Any idea on "what could have changed?" that could be causing the mode error?

-- David C. Rankin, J.D.,P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com

-- () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments

On 10/28/2010 06:34 PM, C Anthony Risinger wrote:

On Oct 28, 2010, at 6:23 PM, Norbert Zeh wrote:

...

I do have the /etc/modprobe.d/nouveau_blacklist.conf file on my machine, too, but I remember that the nouveau kernel driver did get loaded until I added !nouveau to the MODULES list in my rc.conf.

I believe the proprietary driver package puts that blacklist file in place.

See what package owns it with (IIRC):

# pacman -Qm /etc/modprobe.d/nouveau_blacklist.conf

C Anthony [mobile]

[02:23 archangel:/etc/clamav] # pacman -Qo /etc/modprobe.d/nouveau_blacklist.conf /etc/modprobe.d/nouveau_blacklist.conf is owned by nvidia 260.19.12-1 You are correct, Sir! -- David C. Rankin, J.D.,P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com

Am Thu, 28 Oct 2010 18:03:09 -0500 schrieb "David C. Rankin" :

Also, on the (press enter or scan issue), I'm sure you are 100% correct Heiko, but no matter whether I press enter or issue scan with this 8800GT, the list of modes returned is the same. Also, both enter or scan take the same amount of time to complete -- at least on this box.

If the proprietary driver doesn't support KMS and isn't going to for a while, then I wonder what changed to make my normal kernel line (that I've used for years) suddenly start causing the invalid mode error? The kernel line is:

kernel /vmlinuz26 root=/dev/mapper/nvidia_baaccajap5 ro vga=0x31a

As a test, I booted LTS with the vga=0x31a and it booted fine and into the correct framebuffer. So this is looking more and more kernel related.

Have you tried to change vga=0x31a to a value which is shown when you enter scan on the prompt? If not, please, do so. And, please, enter "scan" and not only press Enter. Have you added nomodeset to the kernel line in /boot/grub/menu.lst? If not, please, do so. This has nothing to do with the nvidia and nouveau driver, because at this time there are no modules loaded. Please, add both kernel parameters vga= with a valid value and nomodeset to the kernel line in /boot/grub/menu.lst. This line has to look like this: kernel /vmlinuz26 ... nomodeset ... vga=...

I have also booted Arch without vga= in the kernel line and I do NOT get the invalid mode prompt, but then I'm left with a 640x480 framebuffer. That looks terrible on a 23" display.

If you don't add the vga parameter to the kernel line you don't get a framebuffer probably except if you are using KMS. Heiko

On 29.10.2010 02:35, Nicolas Bigaouette wrote:

It might have to do something with a module not being included in the initrd? Just guessing...

or maybe nouveau included in initrd? another wild guess.

On Fri, Oct 29, 2010 at 07:47:29AM -0500, David C. Rankin wrote:

On 10/28/2010 08:57 PM, Dave Reisner wrote:

...

Specify the mode as decimal instead of hex -- in this case it'd be 794.

Dave,

That worked! Thanks. But, Grrrr., what changed to make whatever checks now not like the hex designation? I have used the hex designation for at least a *decade*.

I have no idea why this happens, but I came across this myself with my htpc. d

At Freitag, 29. Oktober 2010 22:36 David C. Rankin wrote:

Do you (all) think this needs a bug opened? There must be a reason

why, but

given the lack of chatter I have seen about it, it must be fairly rare. Dunno if it is worth chasing. Arch-dev thoughts?

More for upstream than for achlinux and perhaps it will very hard to find it. Some time ago i switched from 791 to 0x317 because of the same reason as you now. It seems for me normal user that one of this 2 ways works even better than the other. I would say that you can save your time and enjoy that it works now again. See you, Attila