RE: [SLE] No X after Kernel Update
X-Mailinglist: suse-linux-e To: "'kranked@canada.com'"
Delivered-To: canada.com%kranked@canada.com Delivered-To: mailing list suse-linux-e@suse.com Cc: "Suse-Linux-E List (E-mail)" Received: (cpmta 9790 invoked from network); 11 Oct 2001 07:59:03 -0700 Received: from unknown (HELO yucntsys1.yucom.be) (212.8.180.1) by smtp.c009.snv.cp.net (209.228.34.109) with SMTP; 11 Oct 2001 07:59:03 -0700 Received: from mail pickup service by yucntsys1.yucom.be with Microsoft SMTPSVC; Thu, 11 Oct 2001 16:59:02 +0200 Received: from yucntsys1.yucom.be ([212.8.180.1]) by yucntsys1.yucom.be with Microsoft SMTPSVC(5.5.1877.687.68); Thu, 11 Oct 2001 10:05:42 +0200 Received: from pigeon.net7.be ([195.207.34.251]) by yucntsys1.yucom.be with Microsoft SMTPSVC(5.5.1877.687.68); Thu, 11 Oct 2001 10:05:42 +0200 Received: from pigeon.advalvas.be (pigeon.advalvas.be [195.207.34.251]) by
Received: from [202.58.118.7] by pigeon.advalvas.be (NTMail 5.06.0014/KF7358.44.3e3845f8) with ESMTP id agfpidaa for michelangelo@advalvas.be; Thu, 11 Oct 2001 09:42:22 +0200 Received: (qmail 19044 invoked by alias); 11 Oct 2001 07:37:14 -0000 Received: (qmail 19035 invoked from network); 11 Oct 2001 07:37:14 -0000 Content-Length: 6277 Content-Type: text/plain; charset="iso-8859-1" Mailing-List: contact suse-linux-e-help@suse.com; run by ezmlm X-Received: 11 Oct 2001 14:59:03 GMT Precedence: bulk List-Unsubscribe: mailto:suse-linux-e-unsubscribe@suse.com List-Post: mailto:suse-linux-e@suse.com Subject: RE: [SLE] No X after Kernel Update X-No-Archive: yes List-Help: mailto:suse-linux-e-help@suse.com From: "Heaton, Simon"
Return-Path: Return-Path: suse-linux-e-return-75261-michelangelo=advalvas.be@suse.com Mime-Version: 1.0 Date: Thu, 11 Oct 2001 08:39:11 +0100 Message-Id: X-Mailer: Internet Mail Service (5.5.2653.19) Hi Mark,
I suspect that your problem was related to the following:
If you download the binary nvidia driver from their site, it is built to only load if the kernel version is "2.4.4-4GB" as this is the default kernel used in the SuSE 7.2 release. Now you can check to see what kernel you are using by "uname -r". So after you upgraded to to v2.4.10 kernel, the nvdriver will simply not work. So what you have to do next is to download the driver source and rebuild it. It is easiest if you perform this rebuild while running the 2.4.10 kernel. Incidentally I suspect that the reason
modprobe did not find the NVdriver binary you downloaded will have been
it will have installed the binary in /lib/modules/2.4.4-4GB. Once you are running the v2.4.10 kernel, it will expect to find modules in /lib/modules/2.4.10.
So maybe try the following:
Download the v2.4.10 kernel source and build it as you usually do. Also download the nvidia driver source. Before you boot the v2.4.10 kernel, modify the /etc/inittab and set the default run level to 2 (ie. do not try and start X). Once you have rebooted as v2.4.10 and verified that everything is working, build and install the nvidia driver as per the instructions included in the nvidia kit. After you have installed the nvidia driver, check that it has indeed been installed correctly: Say uname -r returns "2.4.10", then look in /lib/modules/2.4.10 for a subdirectory called "video". If it is NOT there then we need to think again. If it is there, just try switching to run level 5 using "init 5" and hopefully you should be up and running. If not then we think again!
Good luck.
Kind regards,
Simon
-----Original Message----- From: kranked@canada.com [mailto:kranked@canada.com] Sent: Wednesday, October 10, 2001 10:16 PM To: Simon.Heaton@softwareag.co.uk Subject: RE: [SLE] No X after Kernel Update
On Wed, 10 October 2001, "Heaton, Simon" wrote:
Hi Simon, Hope you don't mind the direct e-mail as I'm sure the list has seen enough of my problem :) I am back to using kernel 2.4.4-4GB, but am still anxious to give it another shot. I was actually right in the middle of downloading the kernel source rpm and rebuild it that way, but low and behold mantel switched to 2.4.4-11 on my at 75% of my d'load haha. When I got dumped back to the console I checked lsmod for NVdriver or nvidia and nothing was there. Also doing a modprobe NVdriver produced a "error:cannot locate module NVdriver". Can I asked how it is that you updated to the 2.4.10 kernel as I did it
easy easy way and just downloaded the rpm and did a Uvh then ran mk_initrd
and then lilo. The kernel ran fine, but I wonder if not haveing the source led the the nvidia driver not being able to find the proper kernel header files or something. Also which way did you install the nvida drivers? Once again I took the easy route and just d'loaded the SuSE 7.2 rpm's but I'm almost sure now I should have used the source rpm's becuase in the Nvidia readme it said to use the source for systems that had upgraded kernels from the default that shipped with the distro. Thanks for the help .
Cheers Mark Doucette
Hi Mark,
I have the new nvidia drivers working fine with 2.4.10. As a start check
the
following:
When you are dumped back at the text terminal when X fails to start, log in as root and check what kernel modules are loaded with lsmod. What is the output from the following command: uname -r ? Now try the following: modprobe NVdriver. What message do you get if any?
Kind regards,
Simon
-----Original Message----- From: kranked@canada.com [mailto:kranked@canada.com] Sent: Monday, October 08, 2001 8:58 PM To: suse-linux-e@suse.com Subject: Re: [SLE] No X after Kernel Update
Hi again,
Well since I already have the new kernel installed I'm not sure how I could rebuild and install the src one. I didn't think it would allow me to install the same kernel version over itself. I tried removing the kernel the simple -e way but that didn't work.
So far I've d'loaded the SusE specific newest drivers from nvidia and X still isn't working. Here's a snip from /var/log/SaX.log from when I try to set up X through sax2 -m 0=nvidia <snip> Chipset RIVA TNT2 Ultra found (II) Loading /usr/X11R6/lib/modules/libvgahw.a (II) Module vgahw: vendor="The XFree86 Project" compiled for 4.0.3, module version = 0.1.0 (**) NVIDIA(0): Depth 16, (--) framebuffer bpp 16 (==) NVIDIA(0): RGB weight 565 (==) NVIDIA(0): Default visual is TrueColor (==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0) (--) NVIDIA(0): Linear framebuffer at 0xE6000000 (--) NVIDIA(0): MMIO registers at 0xE4000000 (EE) NVIDIA(0): Failed to initialize the NVdriver kernel module! (EE) NVIDIA(0): *** Aborting *** (II) Unloading /usr/X11R6/lib/modules/libvgahw.a (EE) Screen(s) found, but none have a usable configuration.
Fatal server error: no screens found <snip>
My next guess would be to upgrade to the newest XFree, but it just seems odd that I can't run it after just upgrading the kernel as many others have.
There were no errors when installing the new rpm's from nvidia. I made sure to remove the original NVIDIA_GLX prior to installing the new version as well.
If there are any other logs that would help to solve this then please let me know as I can reach them from this comp now.
Thanks in advance for all the help
Mark Doucette.
__________________________________________________________ Get your FREE personalized e-mail at http://www.canada.com
-- The content of this e-mail is confidential, may contain privileged material and is intended solely for the recipient(s) named above. If you receive
Hi Mark, First of all let me say that I have NOT yet downloaded the v2.4.12 source, however it will be very unlikely that there is any change with regards to compilation etc. Now my first suggestion on how to have successful implementations of newly built kernels is to switch to using grub as the boot loader insteed of lilo, if you are not already doing so. Having switched to grub a couple of months ago I have to say that I have found it easy to configure and faultless in performance. (I should also say that lilo too is faultless in performance, but for me somewhat more complex in configuration.) My second suggestion is to ensure that you have a recovery path should the system fail to boot. This is VERY important. Anyway, here are some crib notes for you. I have assumed that you are logged on as root and that you have a working X environment and that your are using some xterm. Also that you are not already running a 2.4.12 kernel: Unpack the NVidia driver source into /usr/src/kernel-modules. Now unpack the kernel source in /usr/src, it should create a directory with a name like "2.4.12". Then cd into this directory. If you wish to modify the default configuration of the supplied kernel type: make xconfig. Configure to your hearts content! Once you have the desired configuration save and exit. At the command line type: make dep and once this has completed, make clean and once this has completed, make bzImage. This last make is the bit thats compiles your new kernel. Once this has completed you should have a new kernel (bzImage) in the subdirectory: ./arch/i386/boot. Now from the same directory in which you executed the previous makes, type: make modules. This will build any modules etc. It is at this point that you need to exercise a bit of care as what happens next depends on your current configuration. Essentially, what we want to do now is install the newly built modules. In order to do this we would type: make modules_install. What this would do is create the directory structure /lib/modules/2.4.12, so if you already had the directory /lib/modules/2.4.12 it's contents could get trashed. Furthermore, I have assumed that the newly built kernel has the version of "2.4.12", but this may not be the case. It would be for a stock kernel but your SuSE kernel may be different. To check this, from the directory in which you performed the makes, more the following file: ./include/linux/version.h. The string you see here will be the version that the kernel would return from uname -r. If this string is different from the simple "2.4.12" (eg was 2.4.12-4GB), then the module directory should be /lib/modules/2.4.12-4GB. Install the new kernel for booting. This is really specific to your configuration, whether you are using lilo or grub or something else?? Read your documentation. You may have to perform several additional operations here if you are to successfully boot your new kernel. If still puzzled then post a query. Modify the /etc/inittab and set the default run level to 2 (from 5). This means X will not start when you reboot. Reboot. Once the system has rebooted, log on as root and cd to the directory within which the NVdriver source is located. There is (I recall) a readme which will tell you how to configure the /etc/X11/X86config file. To build the driver for your new kernel follow the documentation for building, and then for installation of the driver. If your driver has installed OK, you should see this directory: /lib/modules/2.4.12/video within which should be the new driver called NVdriver. Now type "init 5", hopefully this will start X together with your new driver. I hope the above is correct; I am typing this at work and cannot remember exact details. If you have any questions/further probs, just ask. Kind regards, Simon -----Original Message----- From: kranked@canada.com [mailto:kranked@canada.com] Sent: 13 October 2001 07:13 To: Simon.Heaton@softwareag.co.uk Subject: RE: [SLE] No X after Kernel Update Hello again Simon. Once again..apologies for the direct e-mail. I have gone through a miriad of things to get this to work but so far nothing. I think the biggest problem lies in just d'loading the kernel rpm from mantel and taking the 'easy route'. I think something is missing that is needed to --rebuild the nvidia source rpm because I get a boatload of errors when I try it, and no newly built kernel rpm to be found :). Now this is what I'm going to do next. I'm downloading the latest kernel from mantel linux-2.4.12.SuSE-0.tar.bz2 in source form and am going to try that route. The problem is is that i've never done a kernel from anything other than rpm format so I'm apprehensive for sure. I'm reading the 7.2 Configuration manual on how to build the kernel, but if you wouldn't mind could you lay it out for me as to the steps you took to get a working kernel from source with NVIDIA drivers as such. This will be my first real kernel compile I suppose and I don't want to really screw things up. Any and all help/pointers you could provide w! ! ould be very very much appreciated. Thanks again, and hope to hear back from you. Mark Doucette. On Thu, 11 October 2001, "Heaton, Simon" wrote: pigeon.advalvas.be (NTMail 5.06.0014/KF7358.44.3e3845f8) with ESMTP id agfpidaa for dragonbe@dma.be; Thu, 11 Oct 2001 09:42:22 +0200 that that the this
in error, please notify Software AG immediately and delete this e-mail.
Software AG (UK) Limited Registered in England & Wales 1310740 Registered Office: Hudson House, Hudson Way, Pride Park, Derby DE24 8HS
-- To unsubscribe send e-mail to suse-linux-e-unsubscribe@suse.com For additional commands send e-mail to suse-linux-e-help@suse.com
Also check the FAQ at http://www.suse.com/support/faq and the archives at http://lists.suse.com
__________________________________________________________ Get your FREE personalized e-mail at http://www.canada.com -- The content of this e-mail is confidential, may contain privileged material and is intended solely for the recipient(s) named above. If you receive this in error, please notify Software AG immediately and delete this e-mail. Software AG (UK) Limited Registered in England & Wales 1310740 Registered Office: Hudson House, Hudson Way, Pride Park, Derby DE24 8HS
participants (1)
-
Heaton, Simon