LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Slackware (https://www.linuxquestions.org/questions/slackware-14/)
-   -   Nvidia 7174 Woes (https://www.linuxquestions.org/questions/slackware-14/nvidia-7174-woes-320038/)

odin123 05-04-2005 09:24 AM

Nvidia 7174 Woes
 
Hi all,

I'm an old user of Slack (though still a newbie), and after using SuSE 9.1 for a while and finding it too slow on my Toshiba S3000-400 laptop, I installed Slack 10.1 again.

The 2.4.29 kernel was fine but gave me too many problems with cd writing (if hdc as scsi in lilo, then I can't read cds burned by windows). So I took a deep breath and installed and compiled the 2.6.10 kernel using Patrick's sample config file.

All was fine until I installed Nvidia's 7174 driver. It installed alright, and I changed xorg.conf's file to have "nvidia" instead of "vesa" but when I startx, the screen shuts down and I have to reset, and wait for the ext3 recovery, etc. etc. Now I can't even go back to "vesa" to have X.

If I modprobe nvidia, I get the message "Fatal: Inserting nvidia (/lib/modules/2.6.10/kernel/drivers/video/nvidia.ko): invalid module format, though I'm pretty sure I included it in the compile.

Any ideas? No more compiles, please (it takes forever)!

Desperately wanting to keep Slackware!

Thanks.

piete 05-04-2005 10:24 AM

Hey Odin,

You're clearly not the only one finding these problems (http://www.nvnews.net/vbulletin/showthread.php?t=49417). You'll notice right at the bottom, things work fine =)

The basic steps you need to take, therefore, are:

1. download fresh kernel source (2.6.11 seems to be the favourite with 7174) from kernel.org (not Slackware.com!!)
2. copy .config file from your existing 2.6 kernel into the fresh downloaded one and run "make oldconfig"
3. make all && make install
4. then reinstall the nvidia driver

There are gazillions of 2.6-kernel-compile posts, but if you need more personalised help, just gimme a yell =)

- Piete.

[Edit]
The other option (marginally faster than kernel recompile) is to try an older nvidia driver, maybe 6629 or similar. From reading the posts, it seems the problem lies more in the 2.6 in Testing/ than in the nvidia driver, tho'.

odin123 05-04-2005 11:48 AM

Hey, Piete,

Thanks a lot, will do and get back to you once it's done!

Cheers!

jong357 05-04-2005 12:59 PM

I'm really, really curious. Is this a Geforce Go card?

odin123 05-04-2005 01:12 PM

It's a GeForce2 Go with 16Mb of memory.

Piete, could you help with some step-by-step help on how to add the new kernel on top of the 2.6.10 I have? Thanks a million!

jong357 05-04-2005 01:35 PM

Yea, I knew it.... I hate to burst your bubble but I wouldn't expect too much. I'm currently building 2.6.11.8 as we speak just to see if some of the 2.6.11rc5 changes actually make a difference as that link from above may suggest. I actually have read that thread before but chalked it up to being irrelevant. I must have missed his card identifier in the xorg.cong.txt....

I've been fighting this problem for years, literally, years. A modified twice over v4620 from what was once minion.de is the only thing that will work on older geforce go cards... You can see my rants all over the Linux nvidia forums dating back atleast 2 years... Actually, if you spend an entire day searching thru their forums, you'll find hundreds of pissed of laptop owners... I'm afraid Nvidia could care less about it's laptop users. Seriously. They don't care. They refuse to even acknowledge laptops under windows so I suppose it's nice to atleast see that they "claim" to support laptops under Linux. Anyway....

There should be a nice walk through on kernel compile/instalations at the top of this forum... It's been beaten to death so it's a sticky.... ;)

If and when you get fed up with trying to get nvidia drivers working, I can give you a link to ones that work. I'm crossing my fingers but I've been burned too much to get my hopes up too much with a new kernel....

slackb0t 05-04-2005 02:10 PM

I will take the link to the drivers that work :) I am running the nv driver still..

piete 05-04-2005 02:54 PM

Try following the steps in the sticky posts at the top of this forum: the 2.6 build mechanism is principly the same regardless of release number.

If that doesn't work I wrote one for someone else a while back: http://www.kaear.co.uk/default.asp?pageID=36

Also googling is a great resource.

If you *still* have problems, give me an email and we'll make some custom instructions just for you! I would say that by applying a modicum of intellect, the challenge of compiling & installing a new kernel becomes very very simple. The only real trick that I didn't find out about until very recently is copying your old .config file into a new kernel source tree and running "make oldconfig" on the new tree to give you an updated kernel config without loosing all the settings on your current kernel.

Of course, if you're still using a stock 9.1 kernel, "make oldconfig" isn't going to be much use!

A note or two (from experience, not necessarily totally correct):

* You can only have one full kernel installed at any one time. Although you can have multiple entries in lilo for multiple versions of kernel, the installed modules are only for the last kernel you ran "make modules_install" on.
* Nothing on your system is changed until you run "make install" and "make modules_install". I advise running them both (assuming a successful build) or NEITHER. Don't do any half measures.

Anyway, do some reading & research, and see if you can't pick up the process =) If you don't understand much of what I've said, then read some more and re-read what I've mentioned above - it should all come about quickly!

- Piete.

jong357 05-04-2005 04:38 PM

wierd... Angelfire yanked my web page for using it as a "free file server" but now it's mysteriously up again... Grab it while you can....

http://www.angelfire.com/linux/madpenguin1/nvidia.htm

I also had no joy on the new kernel. I wasn't getting the "invalid module format" anymore or the "/usr/lib/libGL.so is not an ELF file - it has the wrong magic bytes at the start" stuff when I'd run ldconfig... Seems the newer kernel did something but it didn't fix my problem. I get a hard lock whenever I start X. All black screen. The drivers at that link are the only ones that will fix that issue..... Oh well... Good luck tho. I'm burned out on it.

slackb0t 05-04-2005 07:23 PM

ok.. i installed the drivers from your site... and it works.. I get the nvidia splash screen etc.. however, occasionally everything locks up and I have to hard reset.. I can still move the mouse but the keyboard and comp are unresponsive..

It never happened before.. and there doesn't seem to be any one thing that I do to make it happen.. it is fairly random.. sometimes takes an hour.. other times a few min.. I checked syslog to see if I could find what happened but there is nothing there.. I guess it locks up and can't write the error.. any suggestions??

btw.. what I tried so far
1. lowering the refresh rates for the monitor..
2. Commenting out the 'amount of ram'
3. reinstalling the drivers
4. Double and triple checked the xorg.conf.. got everthing right in there.. loading glx but not dri or glcore etc.

any further help would be appreciated.. i am writing this using the nv driver again :(

jong357 05-04-2005 08:34 PM

Sorry to highjack the thread here. Still let us know if the new kernel worked out for you.

Interesting... Never heard of that. I ran that driver on 2.6.10 for a few weeks with no problem. Of course, this is on LFS not Slack. Still, it seems unlikely that its an OS specific problem. I've run those drivers on Slack for years with no problems (not kernel 2.6.x tho).

Now, I didn't patch that "2.6.9 and above" driver for the VMALLOC issue. Just changed the pci_find_class function to pci_get_class.... The only thing I can think of off hand is to patch it further. I ran without the vmalloc addition for a week with no probs and decided it just didn't affect me. I really don't know... You might try this:



Code:

--- NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/nv.c        2004-07-28 00:33:20.000000000 +0900
+++ NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/nv.c        2004-10-21 22:47:03.000000000 +0900
@@ -19,6 +19,8 @@
 MODULE_ALIAS_CHARDEV_MAJOR(NV_MAJOR_DEVICE_NUMBER);
 #endif
 
+#define __VMALLOC_RESERVE (128 << 20)
+
 /*
  * our global state; one per device
  */

I just found a patch that may address my problem while hunting for you... I'm off to try it. Thats a patch by the way. Save it in a text file as blah.diff, drop it in your source and "patch -Np1 -i blah.diff" just incase you didnt know. Then reinstall the driver. I don't see how thats going to fix anything but you never know..

EDIT - That patch above has already been applied to the driver. I also ditched the .tar.bz2 in favor of the native .run format. Just run it like you would any other nvidia driver.

slackb0t 05-04-2005 09:09 PM

ok.. I patched it and reinstalled.. I am writing this post using the nvidia driver.. so far so good :P I will post back if I get any lock ups..

thanks for all your help.. hopefully the problem is solved *crosses fingers*

jong357 05-04-2005 09:59 PM

Well, my 5000th atempt yielded nothing. par for the course. I'm rooting for you tho...

slackb0t 05-04-2005 11:39 PM

well.. hasn't locked up all night ... and i have been trying to make it happen... opening up lots of programs etc.. all good for now.. thx again for the help

jong357 05-05-2005 12:20 AM

Nice... Make sure you let me know if it freezes again. If I don't hear back from you, I'll apply that patch and upload the new tarball. I suppose it wouldn't hurt anyway...

odin123 05-05-2005 01:29 AM

By the way, it MIGHT be a distro-related issue: I was running 2.6 on Suse 9.1 (nvidia 6111) and it was just fine. Of course, it might also be the new kernel.

Just for a try, while waiting to download the 2.6.11, I installed the original 2.6.10, not the one in Slack's "Testing" directory, but had the same lockup.

And thanks for the steps, Piete!

piete 05-05-2005 07:06 AM

I'd be suprised if it was a distro related error, but I understand SuSE do patch their kernels.

Although, thinking about it, there's no reason you couldn't use the SuSE 9.1 kernel in Slack ... if it's the kernel that's the problem. I have no justification for this other than a niggling feeling (I was once told this is Bolshevism, and don't care to have that discussion again ... ) that it might be the compilation of kernel driver that is at fault - which is based on your kernel headers and toolchain. If that is the case, then try upgrading a few bits & pieces (namely the headers - DO NOT touch glibc, gcc or binutils). You should just be able to run upgradepkg on the kernel headers package and use a new set of headers from Slack 10 or 10.1 without affecting the rest of your system (if you're worried, do more research and back up your data before you try it ;) )

To jong & slackb0t: nvidia drivers causing hard lockups but leaving your mouse working is nothing new.

(http://ubuntuforums.org/archive/index.php/t-24703.html, http://www.nvnews.net/vbulletin/show...ht=6629+100%25)

The latter thread has me posting under "Austerity", but the symptoms are basically the same. Screen locks, becomes all but useless. SSH into the box and tun "top", and you'll see X is using 100% of the processor cycles - you may not have the same issue, of course. Unfortunatly, they talk a lot without providing actual fixes for the problem.

My fix? I discovered my graphics card was knackered (the fan driver on the PCB had been blown!), so got a new one. No more problem.

With a laptop, however, I can appreciate that buying a new one is not an option =)

Good luck all, I wish I could help more!

jong357 05-05-2005 04:23 PM

Yea, I've seen those threads... My X server won't even start tho. Just a pitch black screen right after I type 'startx'.... Switching VC's, 3 finger salute, nothing does the trick besides a hard reboot. I assume I'm geting an oops... Nvidia has known about this for almost 2 years. They just don't see it as a worth while thing to expend resources on since it only affects older cards which are not in distribution anymore.... I did try the patches for 6629 that were supposed to fix a similar issue but no joy. Other laptop owners on that same thread reported that it didn't work for them either.

Sucks to be me... I paid 1700 dollars for this thing 3 years ago. Buying a new one is out of the question. These modified drivers I'm using are gold.... ;) I get nervous with each new kernel release now... :cry: Only a matter of time before I won't know how to patch them...

Let us know how the battle goes Odin...

slackb0t 05-05-2005 04:55 PM

jong.. I haven't had a lock up since that patch.. I have been on this thing all last night and today.. I think it is fixed.

thank you again.

odin123 05-05-2005 06:06 PM

Ok guys,

Got the 2.6.11.8 kernel and compiled it, tried the 7174 driver, nothing.
Tried ALL the nvidia drivers, same thing; tried Jong's drivers, same.

I hate to ask the question, and I know it has been answered a zillion times:

WHY?

Why did I burn three days of my precious life in front of a stupid screen trying to get a stupid driver to work?

WHY, since (gasp) Windows (XP) installed in a jiffy, got ALL my hardware, NEVER locked up on me, is FREE (pirated, sorry), and I can run EVERYTHING on it?

I'm not trying to raise hell here, it's not the right place I know, but I'm pretty disappointed.

I think that if Linux is to win over Windows, they have to get their act together, and seriously so!

odin123 05-05-2005 06:09 PM

That said, thanks to all of you guys for trying to help me, you're a great bunch!

slackb0t 05-05-2005 06:37 PM

Not sure it's a linux problem.. more of an nvidia problem.. In fact the nvidia driver for the gforce2go card under XP has not been supported by nvidia in a while.. the only drivers that work under XP are a few years old as well..

It's funny how experiences differ though.. i agree XP was somewhat easier to install and start using.. but what I hated was that if something did go wrong you were screwed.. like the blue screen of death.. i mean wtf was that.. it would display a bunch of code then reboot and the only solution usually was reinstall or live with it... At least with linux you 'should' be able to solve most problems or at least know the cause.. btw I am not knocking windows.. I used ever version from 3.11 all the way up and enjoyed most of my experiences along the way.. no operating system is perfect but at least with slack I know good people are working on it and others will do their best to help you..

btw.. I still think you should be able to get jong's drivers working.. they seem to work for a lot of people.. did you make sure to uninstall the other drivers first?? maybe it is something simple for those ones.. not sure.. anyway good luck.. I hope to see you post here again with some success..

gbonvehi 05-05-2005 07:14 PM

Quote:

Originally posted by odin123
Ok guys,

Got the 2.6.11.8 kernel and compiled it, tried the 7174 driver, nothing.
Tried ALL the nvidia drivers, same thing; tried Jong's drivers, same.

I hate to ask the question, and I know it has been answered a zillion times:

WHY?

Why did I burn three days of my precious life in front of a stupid screen trying to get a stupid driver to work?

WHY, since (gasp) Windows (XP) installed in a jiffy, got ALL my hardware, NEVER locked up on me, is FREE (pirated, sorry), and I can run EVERYTHING on it?

I'm not trying to raise hell here, it's not the right place I know, but I'm pretty disappointed.

I think that if Linux is to win over Windows, they have to get their act together, and seriously so!

Well, I'm pretty disapointed with you too. If you use something that you like and never crash, why you're not paying for it? And you're complaining about something that's free, why complain? Why not telling, hey, i've this software and hardware, can anybody help me fixing it? Post a freaking bug request in nvidia forum to see if somebody can help you.
I know, I'm pretty sure you're "student and don't have money", well, we use free software and don't complain about it, but we're also trying to help to make it better...

jong357 05-05-2005 08:03 PM

Yea, It's most defenately an Nvidia issue. They have been going down hill for a little while now and I think ATI might be pulling ahead. Nvidia has NEVER supported laptop cards as far as I'm aware of. If your running Windows on a laptop with Nvidia, you have to get your drivers from Dell, HP, et. all.... I don't know... Sorry it didn't work out. If your determined enough, go hit the Nvidia forums and lurk around... Personally, I'll never buy Nvidia again. I'm plagued with problems on both puters...

Slackbot, glad all is well with you. I repackaged those drivers. Might want to grab all of them if your interested. I'm playing hide-and-go-seek with Angelfire right now. They have a bot that goes thru all their web sites and will tally up the ratio of .htm's to media files... :-) Thats what they told me when they replied to my inquiry. They re-instated it for 72 hours until I "develop my web site more"... So I just uploaded a bunch of old htm files and didn't link any of them.... :-) We'll se if they catch on.... I really need some web space apparently...

jong357 05-05-2005 08:09 PM

Odin, did you try my driver (2.6.9 and above) with kernel 2.6.10 or kernel 2.6.11.x? I haven't tested it with 2.6.11.x yet... I know it works for 2.6.10. Food for thought.

I'm off to try it on 2.6.11.8 right now....

Well, the building of it bombed out on me on 2.6.11.8. Said it couldn't complete the instalation or driver build failed or whatever it said. The day I've been dreading is here at last it would seem. I have many depreciated functions in my log along with other stuff... Sigh.....

Google, here I come.... :study: If you havent tried it yet on 2.6.10, I suggest you do so...

odin123 05-06-2005 03:27 AM

gbonvehi,

The money is not an issue (in fact, had I not mentioned the money, you would have said that I should be happy with linux-for-free better than windows-for money.

And "free" doesn't mean that it "can" be broken and that we should "shut up" and meekly thank the people who are giving to us just because it's "free." (For more on the concept of "free," look at the selling prices of RHE, SuSE, and even Slack. For "free," look at the amount of time--and time is money--one has to spend on fixing trivial things like sound or video or modem stuff in Linux; for "free," look at the money spent on dial-up connections--yes, they DO exist--in order to get answers and stuff. This is not complaining, these are hard facts.)

I don't like it when criticizing linux has become a taboo subject. As grown up people, we have the right, no, in fact, the duty, to cast a critical eye (and critical doesn't mean negative, by the way) on the things we use, even more so when these things, in this case linux, takes up so much of our daily (and nightly!) time.

When I spend three days setting up Slackware in order to have most things running fine and not getting results, I have the right to be disappointed, with all due thanks to Patrick and to all other good and earnest people who are working day and night against difficult odds to make us happy.

Without constructive criticism, linux would NOT be where it is today, gbonvehi, and the attitude of "don't say a bad word about Linux or else..." is not appreciated.

I have scoured the nvidia and the linuxquestions forums for an answer, and probably paid more money on my dial-up line than you can imagine in order to get answers and kernels, etc.

People like Piete, Jong 357, Slackb0t, and many others here and in the Linux community have earnestly tried to help me and help each other out and I warmly thank them as they and I will continue our search for the Holy Grail of Linux.

It is interesting to note that your only contribution to this thread was to be negative and aggressive. I wish you had given some suggestion about my problem, though.

As to continuing to use Linux, don't worry, I will.

(and, by the way, I'm no student, and I have money, thank you)

gbonvehi 05-06-2005 03:43 AM

Sorry, I usually try to help people, but I hate when people doesn't pay for something that's good (pirate) + a bad night = bad answer..
I didn't want you to say that Linux is like God (I'm a person that actually likes some stuff Microsoft did/does and admit it), I just wanted you to know that people is doing what they can to make it better, it's not stable or supported like Windows because a lot of people helps in their spare time, so it takes more time..
Most developers knows what Windows can do, and want to get there, but you've to be patience, there was a big improvement in Linux these years so I'm pretty sure the time to support hardware and software almost equally to Windows is getting closer..

Cheers,
Guillermo Bonvehí

keefaz 05-06-2005 04:29 AM

For video problem, try to disable framebuffer to see if that change something
Also for nvidia, do NOT compile in the Kernel :
CONFIG_FB_RIVA (lot of trouble) and CONFIG_FB_VGA16 (use CONFIG_FB_VESA instead)

odin123 05-06-2005 05:53 AM

jong357, I tried your drivers but got some error message half-way through install (on 2.6.11.8, if I remember correctly; I did so many installs and compiles and reformats the last three days that I lost track somehow...)

I guess I'll be back to 2.4 and wait until Slack 11 comes out, maybe they and the nvidia people will have sorted things out!

Thanks again to all those who helped, I appreciate the effort!

jong357 05-06-2005 01:00 PM

O.k.. Thats what I was saying. 2.6.11.8 does not work with the 4xxx/5xxx/6xxx versions of nvidia. From my expierence and numerous posts on the web, they made changes to the kernel (yet again) that break all but the latest nvidia drivers. Try booting up with 2.6.10 like you were doing and then install my "2.6.9 and above" driver. That SHOULD work....

Also, If your moving back to 2.4.2x, why not down load my 2.4 version of 4620. That works great as well..... Atleast concerning the black screen of death after startx...

I'm currently trying to get my hands on a patch that will REALLY help me out in the way of fixing my 4620 driver for the latest kernel. If anyone has "nvidia_display_driver-1.0-6111.diff" from here:

http://unixforge.org/~matthias-chris...-1.0-6111.diff

I beg of you... Please get it to me.... It's looking like I have to abandon my hope of getting newer drivers to work on my laptop and try to maintain my working 4620 as long as I can.... That guy has some really agressive perms on his server. I don't know why you would post a link to the patch in a public forum and then chmod the hell out of your pub directory... Makes no sense....

odin123 05-06-2005 03:45 PM

Thanks, jong357!

Slack on 2.4.29 with nvidia 6111 works like a charm.

I'm happy enough and I can't see why, in retrospective, I made such a big fuss about wanting to go to 2.6 and above, since 2.4 is doing almost everything I want (except for some annoyances with the cd as scsi in lilo but that's ok).

As I said, I'll wait for Slack 11 and hope that Patrick will make the jump to 2.6 with a kernel that can handle nvidia!

Cheers!

jong357 05-06-2005 09:47 PM

Yea... I'm not happy with 2.6.x either. With each kernel release, more things break. Working drivers no longer work and supported hardware no longer seems to be supported. I've gotten lucky with my laptop as far as 2.6 goes but it breaks my new desktop in multiple areas... They don't call 2.4 the stable branch for nothing...

slackb0t 05-15-2005 08:04 AM

Quote:

Originally posted by jong357
Odin, did you try my driver (2.6.9 and above) with kernel 2.6.10 or kernel 2.6.11.x? I haven't tested it with 2.6.11.x yet... I know it works for 2.6.10. Food for thought.

I'm off to try it on 2.6.11.8 right now....

Well, the building of it bombed out on me on 2.6.11.8. Said it couldn't complete the instalation or driver build failed or whatever it said. The day I've been dreading is here at last it would seem. I have many depreciated functions in my log along with other stuff... Sigh.....

Google, here I come.... :study: If you havent tried it yet on 2.6.10, I suggest you do so...

I just upgraded to the 2.6.11.9 kernel from current/testing ... I can no longer use your driver.. as you stated above i get the same messeages in /var/log/nvid... depreciated functions etc... then it just fails..

Do you have any suggestions for this kernel.. or if I want an nvidia driver do I need to bust back to 2.6.10?

thanks for any help

(just when I thought I had this all worked out.. I am writting from the nv driver again :( )

jong357 05-15-2005 09:44 AM

It's 2.6.10 or below.... I know, it sucks. I looked into what it would take to get them running on 2.6.11 and it's way over my head. I have a patch for 6629 that lets it (v6629) run on 2.6.11 but there were many changes made from 4620 to 6629 in and of itself...

I'm completely put out by nvidia and I'm just not going to give them anymore of my money. That simple. They don't state it anywhere but it's obvious they drop support for older cards after only a couple years. If your lucky, you have a 3 year old card that still works with the latest driver. I ussually don't have the money to buy a new graphics card every 2-3 years and I CERTAINLY don't have the money to buy a new laptop every 2-3 years....

If it's feasable, you might consider sticking with the stable kernel branch. It's still actively being developed and I've never had a problem with it.. ide-cdrom support is about the only good thing I see in 2.6.xx...

mziab 05-22-2005 10:03 AM

I'm also experiencing problems with drivers newer than 1.0-4620. On my Geforce 4 Ti4200 (NV28) TwinView doesn't work on any newer drivers, so I'm stuck with 1.0-4620. Anyway, I've recently managed to get 4620 working with 2.6.11.9. It's the result of a lot of luck and experimentation.
Quite recently a person called moocha released a modified ebuild for version 6111 with patches to make it work on kernels 2.6.11+. One of the patches is, in fact, extremely useful - moocha_agp_backend_h_fix.patch. It helped to get most of the work done.
I have some some additional patches of my own. nv-pci_find_class.patch just changes all uses of pci_find_class to pci_get_class. I've used it before with kernel 2.6.9 and 2.6.10.
I had to replace all occurrences of page_remap_range to io_page_remap_range. remap3.patch does this task.
I also whipped up a quick and dirty patch (agp-fix.patch), so AGP actually works. Without it, everything compiled nicely and xorg worked, but I got "cannot verify AGP usage" in the logs and glxgears was really slow.
One last problem I have, and don't know how to resolve, is a kernel panic when I reboot my system. However, there's a really simple solution to it - setting "NvAGP" to "1".
I'd also like to note I applied all the patches on a Nvidia installer already patched for 2.6. I've won my war with kernel 2.6.11 and hope this'll help you win yours :)

jong357 05-22-2005 11:54 AM

Interesting.... Thanks much.. If you could, leave those patches/links up for just a bit longer...

Nice find on the last link. Thats the same 4620 driver that I'm using. Ever since minion.de removed their drivers, I assumed they were lost to all...

jong357 05-22-2005 02:34 PM

Wow.... You rock man... Seriously. Nice find. Good footwork too. They work perfectly so far on 2.6.11.10... That was probably the most usefull 1st post ever in the history of any linux forum (in my book anyway) ;) Much appreciated...

So, for anyone stumbling upon this thread in the future who could benefit from it, I made a unified diff against the 4620 version that can be found in mziab's post above. That Nvidia driver was made by Zander (linux nvidia dev) awhile back so it would run on 2.6 kernels.
Code:

diff -urN NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/agp_backend.h NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/agp_backend.h
--- NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/agp_backend.h        1969-12-31 19:00:00.000000000 -0500
+++ NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/agp_backend.h        2005-05-22 13:23:16.000000000 -0400
@@ -0,0 +1,118 @@
+/*
+ * AGPGART backend specific includes. Not for userspace consumption.
+ *
+ * Copyright (C) 2002-2003 Dave Jones
+ * Copyright (C) 1999 Jeff Hartmann
+ * Copyright (C) 1999 Precision Insight, Inc.
+ * Copyright (C) 1999 Xi Graphics, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * JEFF HARTMANN, OR ANY OTHER CONTRIBUTORS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
+ * OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _AGP_BACKEND_H
+#define _AGP_BACKEND_H 1
+
+#ifdef __KERNEL__
+
+#ifndef TRUE
+#define TRUE 1
+#endif
+
+#ifndef FALSE
+#define FALSE 0
+#endif
+
+enum chipset_type {
+        NOT_SUPPORTED,
+        SUPPORTED,
+};
+
+struct agp_version {
+        u16 major;
+        u16 minor;
+};
+
+struct agp_kern_info {
+        struct agp_version version;
+        struct pci_dev *device;
+        enum chipset_type chipset;
+        unsigned long mode;
+        off_t aper_base;
+        size_t aper_size;
+        int max_memory;                /* In pages */
+        int current_memory;
+        int cant_use_aperture;
+        unsigned long page_mask;
+        struct vm_operations_struct *vm_ops;
+};
+
+/*
+ * The agp_memory structure has information about the block of agp memory
+ * allocated.  A caller may manipulate the next and prev pointers to link
+ * each allocated item into a list.  These pointers are ignored by the backend.
+ * Everything else should never be written to, but the caller may read any of
+ * the items to detrimine the status of this block of agp memory.
+ */
+
+struct agp_memory {
+        int key;
+        struct agp_memory *next;
+        struct agp_memory *prev;
+        size_t page_count;
+        int num_scratch_pages;
+        unsigned long *memory;
+        off_t pg_start;
+        u32 type;
+        u32 physical;
+        u8 is_bound;
+        u8 is_flushed;
+};
+
+#define AGP_NORMAL_MEMORY 0
+
+extern void agp_free_memory(struct agp_memory *);
+extern struct agp_memory *agp_allocate_memory(size_t, u32);
+extern int agp_copy_info(struct agp_kern_info *);
+extern int agp_bind_memory(struct agp_memory *, off_t);
+extern int agp_unbind_memory(struct agp_memory *);
+extern void agp_enable(u32);
+extern int agp_backend_acquire(void);
+extern void agp_backend_release(void);
+
+/*
+ * Interface between drm and agp code.  When agp initializes, it makes
+ * the below structure available via inter_module_register(), drm might
+ * use it.  Keith Owens <kaos@ocs.com.au> 28 Oct 2000.
+ */
+typedef struct {
+        void                        (* free_memory) (struct agp_memory *);
+        struct agp_memory *        (* allocate_memory) (size_t, u32);
+        int                        (* bind_memory) (struct agp_memory *, off_t);
+        int                        (* unbind_memory) (struct agp_memory *);
+        void                        (* enable) (u32);
+        int                        (* acquire) (void);
+        void                        (* release) (void);
+        int                        (* copy_info) (struct agp_kern_info *);
+} drm_agp_t;
+
+extern const drm_agp_t *drm_agp_p;
+
+#endif                                /* __KERNEL__ */
+#endif                                /* _AGP_BACKEND_H */
diff -urN NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/nv.c NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/nv.c
--- NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/nv.c        2004-01-13 08:34:57.000000000 -0500
+++ NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/nv.c        2005-05-22 13:25:15.000000000 -0400
@@ -238,7 +238,7 @@
    count = 0;
    dev = (struct pci_dev *) 0;
 
-    dev = pci_find_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
+    dev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
    while (dev)
    {
        if ((dev->vendor != 0x10de) || (dev->device < 0x20))
@@ -292,7 +292,7 @@
        }
 
    next:
-        dev = pci_find_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
+        dev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
    }
 
    nv_printf(NV_DBG_INFO, "NVRM: found %d device%s\n", count, count ? "" : "s");
@@ -535,7 +535,7 @@
    u8    cap_ptr;
    int    func, slot;
 
-    dev = pci_find_class(class << 8, NULL);
+    dev = pci_get_class(class << 8, NULL);
    do {
        for (func = 0; func < 8; func++) {
            slot = PCI_SLOT(dev->devfn);
@@ -546,7 +546,7 @@
            if (cap_ptr)
                return fn;
        }
-        dev = pci_find_class(class << 8, dev);
+        dev = pci_get_class(class << 8, dev);
    } while (dev);
 
    return NULL;
diff -urN NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/nv-linux.h NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/nv-linux.h
--- NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/nv-linux.h        2004-01-13 08:34:31.000000000 -0500
+++ NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/nv-linux.h        2005-05-22 13:25:31.000000000 -0400
@@ -115,7 +115,7 @@
 #if defined (CONFIG_AGP) || defined (CONFIG_AGP_MODULE)
 #ifndef NOAGPGART
 #  define AGPGART
-#  include <linux/agp_backend.h>
+#  include "agp_backend.h"
 #  include <linux/agpgart.h>
 #endif
 #endif
@@ -428,16 +428,16 @@
  * RedHat without means to identify the change, hence this hack.
  */
 #ifdef KERNEL_2_6
-#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(vma, a, ## b)
+#define NV_REMAP_PAGE_RANGE(a, b...) io_remap_page_range(vma, a, ## b)
 #else
 #if defined(REMAP_PAGE_RANGE_5)
-#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(vma, a, ## b)
+#define NV_REMAP_PAGE_RANGE(a, b...) io_remap_page_range(vma, a, ## b)
 #elif defined(REMAP_PAGE_RANGE_4)
-#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(a, ## b)
+#define NV_REMAP_PAGE_RANGE(a, b...) io_remap_page_range(a, ## b)
 #else
-#warning "conftest.sh failed, assuming old Linux 2.4 remap_page_range(4)!"
+#warning "conftest.sh failed, assuming old Linux 2.4 io_remap_page_range(4)!"
 
-#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(a, ## b)
+#define NV_REMAP_PAGE_RANGE(a, b...) io_remap_page_range(a, ## b)
 #endif
 #endif /* KERNEL_2_6 */
 
diff -urN NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/os-agp.c NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/os-agp.c
--- NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/os-agp.c        2003-12-04 13:24:15.000000000 -0500
+++ NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/os-agp.c        2005-05-22 13:23:08.000000000 -0400
@@ -47,7 +47,6 @@
 
 agp_kern_info        agpinfo;
 agp_gart              gart;
-const drm_agp_t      *drm_agp_p;
 
 #if defined(CONFIG_MTRR)
 #define MTRR_DEL(gart) if ((gart).mtrr > 0) mtrr_del((gart).mtrr, 0, 0);
@@ -81,12 +80,7 @@
      * Also, the Linux 2.5 AGP GART driver is modularized, agpgart.o does not
      * include backend drivers.
      */
-    if (!(drm_agp_p = inter_module_get_request("drm_agp", "agpgart")))
-    {
-        nv_printf(NV_DBG_ERRORS,
-            "NVRM: AGPGART: unable to retrieve symbol table\n");
-        return 1;
-    }
+    request_module("%s", "agpgart");
   
    /* NOTE: from here down, return an error code of '-1'
      * that indicates that agpgart is loaded, but we failed to use it
@@ -94,7 +88,7 @@
      * the memory controller.
      */
 
-    if (drm_agp_p->acquire())
+    if (agp_backend_acquire())
    {
        nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: backend in use\n");
        return -1;
@@ -119,10 +113,10 @@
      * chipsets via this function. If this Linux 2.4 kernels behaves the same
      * way, we have no way to know.
      */
-    drm_agp_p->copy_info(&agpinfo);
+    agp_copy_info(&agpinfo);
 #else
-    if (drm_agp_p->copy_info(&agpinfo)) {
-        drm_agp_p->release();
+    if (agp_copy_info(&agpinfo)) {
+        agp_backend_release();
        inter_module_put("drm_agp");
        return -1;
    }
@@ -140,7 +134,7 @@
          */
        nv_printf(NV_DBG_ERRORS,
            "NVRM: AGPGART: unable to set MTRR write-combining\n");
-        drm_agp_p->release();
+        agp_backend_release();
        inter_module_put("drm_agp");
        return -1;
    }
@@ -157,7 +151,7 @@
    {
        nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: unable to remap aperture\n");
        MTRR_DEL(gart);
-        drm_agp_p->release();
+        agp_backend_release();
        inter_module_put("drm_agp");
        return -1;
    }
@@ -169,7 +163,7 @@
        nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: unable to allocate bitmap\n");
        NV_IOUNMAP(gart.aperture, RM_PAGE_SIZE);
        MTRR_DEL(gart);
-        drm_agp_p->release();
+        agp_backend_release();
        inter_module_put("drm_agp");
        return -1;
    }
@@ -181,7 +175,7 @@
        os_free_mem(bitmap);
        NV_IOUNMAP(gart.aperture, RM_PAGE_SIZE);
        MTRR_DEL(gart);
-        drm_agp_p->release();
+        agp_backend_release();
        inter_module_put("drm_agp");
        return -1;
    }
@@ -192,7 +186,7 @@
    if (!(agp_rate & 0x00000004)) agpinfo.mode &= ~0x00000004;
    if (!(agp_rate & 0x00000002)) agpinfo.mode &= ~0x00000002;
   
-    drm_agp_p->enable(agpinfo.mode);
+    agp_enable(agpinfo.mode);
   
    *ap_phys_base  = (void*) agpinfo.aper_base;
    *ap_mapped_base = (void*) gart.aperture;
@@ -227,7 +221,7 @@
        NV_IOUNMAP(gart.aperture, RM_PAGE_SIZE);
    }
 
-    drm_agp_p->release();
+    agp_backend_release();
 
    /*
      * XXX Same as above; the inter_module_* mechanism will go away at some
@@ -280,7 +274,7 @@
        return RM_ERROR;
    }
 
-    ptr = drm_agp_p->allocate_memory(PageCount, AGP_NORMAL_MEMORY);
+    ptr = agp_allocate_memory(PageCount, AGP_NORMAL_MEMORY);
    if (ptr == NULL)
    {
        *pAddress = (void*) 0;
@@ -288,7 +282,7 @@
        return RM_ERR_NO_FREE_MEM;
    }
   
-    err = drm_agp_p->bind_memory(ptr, *Offset);
+    err = agp_bind_memory(ptr, *Offset);
    if (err)
    {
        // this happens a lot when the aperture itself fills up..
@@ -305,7 +299,7 @@
    if (status != RM_OK)
    {
        nv_printf(NV_DBG_ERRORS, "NVRM: memory allocation failed\n");
-        drm_agp_p->unbind_memory(ptr);
+        agp_unbind_memory(ptr);
        goto fail;
    }
 
@@ -320,7 +314,7 @@
    return RM_OK;
 
 fail:
-    drm_agp_p->free_memory(ptr);
+    agp_free_memory(ptr);
    *pAddress = (void*) 0;
 
    return RM_ERROR;
@@ -359,7 +353,7 @@
    {
        nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: unable to remap %lu pages\n",
            (unsigned long)agp_data->num_pages);
-        drm_agp_p->unbind_memory(agp_data->ptr);
+        agp_unbind_memory(agp_data->ptr);
        goto fail;
    }
   
@@ -448,8 +442,8 @@
    {
        size_t pages = ptr->page_count;
 
-        drm_agp_p->unbind_memory(ptr);
-        drm_agp_p->free_memory(ptr);
+        agp_unbind_memory(ptr);
+        agp_free_memory(ptr);
 
        nv_printf(NV_DBG_INFO, "NVRM: AGPGART: freed %ld pages\n",
            (unsigned long)pages);

Now, I didn't include the VMALLOC_RESERVE define in that patch. I've been running without it but others, like slackb0t, may have to additionally patch for that.
Code:

--- NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/nv.c        2004-07-28 00:33:20.000000000 +0900
+++ NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/nv.c        2004-10-21 22:47:03.000000000 +0900
@@ -19,6 +19,8 @@
 MODULE_ALIAS_CHARDEV_MAJOR(NV_MAJOR_DEVICE_NUMBER);
 #endif
 
+#define __VMALLOC_RESERVE (128 << 20)
+
 /*
  * our global state; one per device
  */

I think that one may give some fuzz because the first "MODULE_ALIAS" line isn't actually present in 4620's nv.c... Doesn't matter, it should still patch and work. I also noticed from the gentoo tarball I downloaded, from mziab's link, has a slightly different patch for the same thing. Either one should work.
Code:

--- NVIDIA-Linux-x86-1.0-4620-pkg0.orig/usr/src/nv/nv.c        2004-08-31 22:52:17.000000000 +0100
+++ NVIDIA-Linux-x86-1.0-4620-pkg0/usr/src/nv/nv.c        2004-09-20 02:49:42.399538232 +0100
@@ -19,6 +19,10 @@
 MODULE_ALIAS_CHARDEV_MAJOR(NV_MAJOR_DEVICE_NUMBER);
 #endif
 
+#ifndef __VMALLOC_RESERVE
+unsigned int __VMALLOC_RESERVE;
+#endif
+
 /*
  * our global state; one per device
  */

I haven't used that one, altho I know the first VMALLOC patch seems to get the job done according to slackb0t... If your having problems with X completely locking up after awhile, you might want to patch the nvidia driver further...

I also took your advice on not using AGPGART by setting "NvAGP" to "1" in my xorg.conf file. I'll try switching it back to 3 and see if I expierence the same problem.

I've also updated my Angelfire Site with the appropriate drivers and patches...

Again, thanks alot for tossing that one in my lap and welcome to LinuxQuestions.org!!! :D

BenneJezzerette 05-22-2005 05:25 PM

I downloaded that very same Nvidia Driver for my card, GeForce 2 MX 400, and it messed things up totally. Biggest change on the systme is, well, the system. I recently upgraded the Motherboard to a PCChips M825G Rev 9.2a with a VIA KM 266Pro/8235, AMD Sempron 2500+ and 512MB DDR PC2700 RAM, and the new kernals don't even have support for this chipset yet. Although the Linux 2.4.29 seems to like it ok, just the agpart does not load, because it can't see the VIA Chipset correctly. The new board has an Onboard Video at 8X AGP but shares the RAM and I want at least 1GB before I do that, so my 4X AGP nVidia GeForce 2 MX 400 is going to do.

The problem I had is this, without the nVidia Driver I get a 1400x1050 24 display, with it, it reverts back to 1024 x 768 24, this was unacceptable. Had to remove the driver so that I could get my display back to what I like.

I posted the new board and chipset on the kernaltrap.org site and am hoping they see it so that I cna start to consider the new Kernels again as the latest Linux 2.6.11.10 is out and the Linux 2.6.11.12 mm rc4 is on the horizon. Once they get the new VIA chipset in, I think they will be getting more adept at making the updates to the newer hardware that is out now.

slackb0t 05-22-2005 08:59 PM

Ok.. I am using this new patch/driver.. all seems to be working so far.. I did have to add the NvAGP 1 option... I was locking on exit.. seems to be ok now...

btw: Jong.. I did use both patchs just to be on the safe side...

one more thing.. does anyone know how to change the terrible slash screen.. not so much to turn it off but to replace it with something a little darker.. that thing is blinding at night.

jong357 05-22-2005 11:25 PM

I'm not sure... I think that may be hard coded somewhere. It is blinding tho.. Sometimes I get up before the sun comes up and yea, once you hit that white screen, it's almost painfull... Well, it is painfull.... If you want to turn it completely off, thats easy..

Option "NoLogo" "true"

put that under your card section in xorg.conf

ramdog 05-23-2005 03:14 PM

extracting .run file
 
Hi all,

I'm trying to install Slack 10.1 on my laptop (Compaq R3210CA). I've upgraded the kernel from the default 2.4.29 to 2.6.11.9 .

I'm trying to do the same extract-patch-install procedure outlined in mziab's post. However, I can't seem to exract the files from the .run file. I've tried using differen .run files, all with the same problem. The extractor quits after saying something about UID 1000, GID 1000.

Sorry I can't have a more detailed description, but the only way I can access the net now is by going into Windows.

Any help will be much appreciated.

Rameses

jong357 05-23-2005 03:20 PM

Try logging in as root and then do your thing.. That or maybe try to:

chmod 777 NVIDIA-BlahBlahBlah.run

ramdog 05-23-2005 03:20 PM

I was in root...

EDIT:

I tried chmod 777 but there was nothing different. same error

jong357 05-23-2005 03:23 PM

Oh.. Well, I suppose an exact error message would be helpfull then.

ramdog 05-23-2005 03:34 PM

Here we go, I've copied the entire thing, from when I run sh NVIDIA-blah-blah.run --extract-only[

-start-
Creating directory NVIDIA-Linux-x86-1.0-4620-pkg0
Verifying archive integrity... OK
Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86 1.0-4620...tar: ./usr/lib/libGLcore.so.1.0.4620: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/lib/libGL.so.1.0.4620: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/lib/libGL.la: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
...tar: ./usr/lib/tls/libGLcore.so.1.0.4620: Cannot change ownership to uid 1000, gid 1000: Operationnot permitted
.tar: ./usr/lib/tls/libGL.so.1.0.4620: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/lib/tls/libGL.la: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/lib/tls: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/lib: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
......tar: ./usr/X11R6/lib/modules/extensions/libglx.so.1.0.4620: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
..tar: ./usr/X11R6/lib/modules/extensions/tls/libglx.so.1.0.4620: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/X11R6/lib/modules/extensions/tls: Cannot change ownership to uid 1000, gid 1000: Operationnot permitted
tar: ./usr/X11R6/lib/modules/extensions: Cannot change ownership to uid 1000, gid 1000: Operation notpermitted
..tar: ./usr/X11R6/lib/modules/drivers/nvidia_drv.o: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/X11R6/lib/modules/drivers: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/X11R6/lib/modules: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/X11R6/lib/libXvMCNVIDIA.a: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/X11R6/lib/libXvMCNVIDIA.so.1.0.4620: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/X11R6/lib: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/X11R6: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
...tar: ./usr/include/GL/gl.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/include/GL/glx.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/include/GL/glxtokens.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/include/GL: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/include: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
..tar: ./usr/share/doc/NVIDIA_Changelog: Cannot change ownership to uid 1000, gid 1000: Operation notpermitted
.tar: ./usr/share/doc/XF86Config.sample: Cannot change ownership to uid 1000, gid 1000: Operation notpermitted
.tar: ./usr/share/doc/README: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/share/doc/README.DE: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/share/doc: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/share: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
..tar: ./usr/src/nv/makedevices.sh: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/nv.c: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/os-agp.c: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/os-interface.c: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/os-registry.c: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/os-agp.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/os-interface.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/nv-linux.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/nv-misc.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/nv-memdbg.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/nv.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/rmretval.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/nvtypes.h: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/README: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/conftest.sh: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/nv-kernel.o: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
..tar: ./usr/src/nv/precompiled: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/src/nv/Makefile.kbuild: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/Makefile.nvidia: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv/Makefile: Cannot create symlink to `Makefile.kbuild': Operation not permitted
.tar: ./usr/src/nv/gcc-version-check.c: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/src/nv: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr/src: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/bin/mkprecompiled: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/bin/tls_test: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/bin/makeself.sh: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/bin/nvidia-settings: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./usr/bin: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./usr: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: ./LICENSE: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./pkg-history.txt: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./Makefile: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./.manifest: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
.tar: ./nvidia-installer: Cannot change ownership to uid 1000, gid 1000: Operation not permitted
tar: Error exit delayed from previous errors
Extraction failed.
-end-

BenneJezzerette 05-23-2005 03:39 PM

Re: extracting .run file
 
Quote:

Originally posted by ramdog
Hi all,

I'm trying to install Slack 10.1 on my laptop (Compaq R3210CA). I've upgraded the kernel from the default 2.4.29 to 2.6.11.9 .

I'm trying to do the same extract-patch-install procedure outlined in mziab's post. However, I can't seem to exract the files from the .run file. I've tried using differen .run files, all with the same problem. The extractor quits after saying something about UID 1000, GID 1000.

Sorry I can't have a more detailed description, but the only way I can access the net now is by going into Windows.

Any help will be much appreciated.

Rameses

Question, did you actually get it installed first, then do the Kernel update? Did you go through the instructions here for your kernel compile too?? It should work. I put Slackware on an E-Machiine M2105 and it was a bear at most, and had to use the May, 2 2005 RC I downloaded to even get it to work. I recently updated to Linux 2.4.30 off the Slackware site so I am pretty much up to date with this. now, the display driver

sh NVIDIA*.run and yes it works fine, that is how I do it. Then let it go and you will be set. As for extracting it, did you get the read me off the nvidia site, it has some nice tips on how to make it work.


from the READ ME I have

The .run file accepts many commandline options. Here are a few of the
more common options:

--info
Print embedded info about the .run file and exit.

--check
Check integrity of the archive and exit.

--extract-only
Extract the contents of ./NVIDIA-Linux-x86-1.0-7174.run,
but do not run 'nvidia-installer'.

--help
Print usage information for the common commandline options
and exit.

--advanced-options
Print usage information for the common commandline options as
well as the advanced options, and then exit.

Installation will also install the utility `nvidia-installer`, which may
be later used to uninstall drivers, auto-download updated drivers, etc.


jong357 05-23-2005 04:36 PM

Usually when you get those errors, it still extracts fine. It just couldn't change the permissions for the files. Sounds pretty basic, but, did you check to see if you have an NVIDIA-BlahBlahBlah folder in the same directory?

Also, which drivers are these? Where did you get them from. What version are you trying to extract? Somebody assigned -g 1000 -o 1000 to the drivers your trying to use, which is pretty bogus. Thats probably why your getting that error because you don't have a Group "1000" or Owner "1000"... They should be root:root....

If they did extract and the perms weren't changed like I suspect, just do a:

chown -R root.root NVIDIA-NameOfFolder

Then patch and make a new .run file if you wish...

BenneJezzerette 05-23-2005 05:51 PM

Quote:

Originally posted by jong357
Usually when you get those errors, it still extracts fine. It just couldn't change the permissions for the files. Sounds pretty basic, but, did you check to see if you have an NVIDIA-BlahBlahBlah folder in the same directory?

Also, which drivers are these? Where did you get them from. What version are you trying to extract? Somebody assigned -g 1000 -o 1000 to the drivers your trying to use, which is pretty bogus. Thats probably why your getting that error because you don't have a Group "1000" or Owner "1000"... They should be root:root....

If they did extract and the perms weren't changed like I suspect, just do a:

chown -R root.root NVIDIA-NameOfFolder

Then patch and make a new .run file if you wish...

you can also not even make a new run file, just type make install, from the read me

ramdog 05-23-2005 06:56 PM

Benne: I downloaded the 2.6.11.9 kernel source from kernel.org, and compiled my own kernel after going thru the xconfig. Then I put it into my /boot partition, edited lilo.conf so I can select which kernel I want to boot with at startup.
I managed to install the 7174 drivers using the sh NVIDIA-blah.run command. But 7174 doesn't work. So I was trying to use the 2.6 modifed 4620 and patch it for 2.6.11, as per the instructions somewhere on page 3 of this thread.

jong: I'm using the .run file from your site! Anyways, I checked carefully this time and I found that the files did indeed extract into the folder. Sorry for any confusion.

I'll give the patch and install a try right now. I'll let you know how it goes.

Again, sorry for any confusion,
R

BenneJezzerette 05-23-2005 07:11 PM

Quote:

Originally posted by ramdog
Benne: I downloaded the 2.6.11.9 kernel source from kernel.org, and compiled my own kernel after going thru the xconfig. Then I put it into my /boot partition, edited lilo.conf so I can select which kernel I want to boot with at startup.
I managed to install the 7174 drivers using the sh NVIDIA-blah.run command. But 7174 doesn't work. So I was trying to use the 2.6 modifed 4620 and patch it for 2.6.11, as per the instructions somewhere on page 3 of this thread.

jong: I'm using the .run file from your site! Anyways, I checked carefully this time and I found that the files did indeed extract into the folder. Sorry for any confusion.

I'll give the patch and install a try right now. I'll let you know how it goes.

Again, sorry for any confusion,
R

god Luck, I am going to attemt the Linux 2.6.11.10 and see what mess I can make of my system.


All times are GMT -5. The time now is 11:59 PM.