emerge x11-drivers/nvidia-drivers-100.14.19 fail [resolved]

Message

bjk2007 · Post by **bjk2007** » Tue Jan 08, 2008 9:35 pm

We just got a new video card for our gentoo server and had to recompile xorg-server, which was successful but x11-drivers/nvidia-drivers failed to compile. Here is the build log:

Code: Select all

 * Determining the location of the kernel source code
 * Found kernel source directory:
 *     /usr/src/linux
 * Found kernel object directory:
 *     /lib/modules/2.6.23.1/build
 * Found sources for kernel version:
 *     2.6.23.1
 * Checking for MTRR support ...
  [ ok ]
>>> Unpacking source...
Creating directory NVIDIA-Linux-x86_64-100.14.19-pkg2
Verifying archive integrity... OK
Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86_64 100.14.19.....
................................................................................
................................................
 * Applying NVIDIA_glx-defines.patch ...
  [ ok ]
 * Applying NVIDIA_glx-glheader.patch ...
  [ ok ]
 * Applying NVIDIA_i2c-hwmon.patch ...
  [ ok ]
 * Converting Makefile.kbuild to use M= instead of SUBDIRS= ...
  [ ok ]
>>> Source unpacked.
>>> Compiling source in /var/tmp/portage/x11-drivers/nvidia-drivers-100.14.19/wo
rk/NVIDIA-Linux-x86_64-100.14.19-pkg2/usr/src/nv ...
 * Preparing nvidia module
sh ./conftest.sh "x86_64-pc-linux-gnu-gcc" "x86_64-pc-linux-gnu-gcc" /usr/src/li
nux /lib/modules/2.6.23.1/build cc_sanity_check full_output
rm -f Makefile
sh ./conftest.sh "x86_64-pc-linux-gnu-gcc" "x86_64-pc-linux-gnu-gcc" /usr/src/li
nux /lib/modules/2.6.23.1/build select_makefile full_output
If you are using a Linux 2.4 kernel, please make sure
you either have configured kernel sources matching your
kernel or the correct set of kernel headers installed
on your system.

If you are using a Linux 2.6 kernel, please make sure
you have configured kernel sources matching your kernel
installed on your system. If you specified a separate
output directory using either the "KBUILD_OUTPUT" or
the "O" KBUILD parameter, make sure to specify this
directory with the SYSOUT environment variable or with
the equivalent nvidia-installer command line option.

Depending on where and how the kernel sources (or the
kernel headers) were installed, you may need to specify
their location with the SYSSRC environment variable or
the equivalent nvidia-installer command line option.

*** Unable to determine the target kernel version. ***

make: *** [select_makefile] Error 1
 *
 * ERROR: x11-drivers/nvidia-drivers-100.14.19 failed.
 * Call stack:
 *                         ebuild.sh, line 1701:  Called dyn_compile
 *                         ebuild.sh, line 1039:  Called qa_call 'src_compile'
 *                         ebuild.sh, line   44:  Called src_compile
 *   nvidia-drivers-100.14.19.ebuild, line  214:  Called linux-mod_src_compile
 *                  linux-mod.eclass, line  518:  Called die
 * The specific snippet of code:
 *                      emake HOSTCC="$(tc-getBUILD_CC)" CC="$(get-KERNEL_CC)" LDFLAGS="$(get_abi_LDFLAGS)" \
 *                                ${BUILD_FIXES} ${BUILD_PARAMS} ${BUILD_TARGETS} \
 *                              || die "Unable to make ${BUILD_FIXES} ${BUILD_PARAMS} ${BUILD_TARGETS}."
 *  The die message:
 *   Unable to make  IGNORE_CC_MISMATCH=yes V=1 SYSSRC=/usr/src/linux SYSOUT=/lib/modules/2.6.23.1/build clean module.

The linux symlink does point to the sources we used to compile our kernel. We used the traditional make && make modules_install method rather than genkernel to build our current kernel. Does anyone know what is going on? Thanks in advance.

PaulBredbury · Post by **PaulBredbury** » Wed Jan 09, 2008 9:57 am

"Traditional" != foolproof.

See wiki re "make".

bjk2007 · Post by **bjk2007** » Wed Jan 09, 2008 4:17 pm

I took a look at your link and got slightly confused at first. I'm assuming the part you were referencing was http://gentoo-wiki.com/HOWTO_Compile_Ke ... nd_modules . I rebuilt our kernel according to that procedure (up to http://gentoo-wiki.com/HOWTO_Compile_Ke ... igure_GRUB), and the system continues to run fine. However, nvidia-drivers is still failing to compile with an identical build error. I would appreciate a slightly better explanation of what exactly is going on and what we're trying to do if it's not a problem.

PaulBredbury · Post by **PaulBredbury** » Wed Jan 09, 2008 9:16 pm

In that case, IIRC this is the file checked, although I could well be wrong:

Code: Select all

$ cat /usr/src/linux/include/linux/utsrelease.h 
#define UTS_RELEASE "2.6.24-rc6-zen0-g4a1d51b0"

Anyway, your 2.6.23.1 kernel is out-of-date, so upgrade.

bjk2007 · Post by **bjk2007** » Wed Jan 09, 2008 9:35 pm

I'm not sure how much good upgrading a few bugfix versions is going to do, but I'll give it a try anyways. That'll give us a fully clean source directory too. Any other suggestions in the meantime would be appreciated

deathcon1 · Post by **deathcon1** » Wed Jan 09, 2008 9:39 pm

If you don't want to upgrade the kernel try this:

Code: Select all

#cd /usr/src/linux && make clean && make && make modules_install

bjk2007 · Post by **bjk2007** » Wed Jan 09, 2008 9:49 pm

The issue isn't really upgrading the kernel; it's whether doing so will make any difference. The newest stable kernel.org sources is 2.6.23.13; I'm running 2.6.23.1.

alienjon · Post by **alienjon** » Wed Jan 09, 2008 9:59 pm

Code: Select all

 * Found sources for kernel version: 
 *     2.6.23.1

{...}

If you are using a Linux 2.6 kernel, please make sure 
you have configured kernel sources matching your kernel 
installed on your system.

It looks like emerge finds your kernel just fine, I imagine it is having a problem reading from the configuration file, though. Do you keep your .config file in some non-standard location?

bjk2007 · Post by **bjk2007** » Wed Jan 09, 2008 10:34 pm

No, we keep the .config in /usr/src/linux/.config. I'm planning on double checking through the nvidia drivers wiki page to see if I can find anything in there. I should be posting my results if any in a few hours.

bjk2007 · Post by **bjk2007** » Thu Jan 10, 2008 12:29 am

I checked the nvidia drivers wiki page (http://gentoo-wiki.com/HOWTO_nVidia_Drivers), and the kernel settings are all correct as outlined. There's really nothing else in there before the drivers install that would be likely to cause the problem. Unless we're missing something here, I might try recompiling the kernel at work tomorrow with fresh sources just to see if that changes anything. If anyone else has any suggestions, they would be very much appreciated.

alienjon · Post by **alienjon** » Thu Jan 10, 2008 4:22 am

This isn't a solution, mind you, but as a work-around, you could always try installing a lower-version of the nvidia-drivers package.

bjk2007 · Post by **bjk2007** » Thu Jan 10, 2008 4:32 am

This isn't a solution, mind you, but as a work-around, you could always try installing a lower-version of the nvidia-drivers package.

Just tried that. Lower versions aren't any friendlier; same build error.

t-wata · Post by **t-wata** » Thu Jan 10, 2008 6:25 pm

I was facing the same issue and it has been solved after kernel update.
If you won't see any error during compiling test code like followings, then version of your kernel is not cause of this issue, otherwise, you need to update your kernel.

Code: Select all

$cat test.c
#define __KERNEL__ 1
#include <linux/utsname.h>
int main(){
}
$gcc -I/usr/src/linux/include test.c

bjk2007 · Post by **bjk2007** » Thu Jan 10, 2008 8:31 pm

An upgrade to the 2.6.23.9 vanilla kernel (latest unmasked sources in portage tree) has not fixed the issue.

I was facing the same issue and it has been solved after kernel update.
If you won't see any error during compiling test code like followings, then version of your kernel is not cause of this issue, otherwise, you need to update your kernel.

Would you mind elaborating on this a bit? I can't locate any relevant test.c files on the system.

Any suggestions from anyone on where to go from here?

alienjon · Post by **alienjon** » Thu Jan 10, 2008 9:15 pm

Sorry, I didn't catch which kernel you were using before. It may be the fact that you are using the vanilla sources maybe? I did a quick search online and found this in the ubuntu forums which indicates that something special needs to be done if using the vanilla kernel but as I'm at work right now and can't follow the link so I'm not sure what they are saying and if it is applicable to Gentoo.

Would you be able to try another kernel source? Maybe gentoo-sources or the like?

bjk2007 · Post by **bjk2007** » Thu Jan 10, 2008 9:47 pm

Unfortunately, it appears that the domain to which they are referring to is no longer registered. It looks like it isn't a compile error for them anyway. As far as different sources go, it's a possibility but last we checked the gentoo-sources are too old to support our sata controller. We actually had to use a sabayon livecd to install the base system because the 2007.0 one didn't support the hard drives. I'd have to check to figure out exactly what kernel version the split is at though. If it's worth finding adequate alternative sources I suppose we could try though.
UPDATE: Gentoo sources are up to the .23 level now so I guess I can give those a try.

Any further suggestions?

bjk2007 · Post by **bjk2007** » Thu Jan 10, 2008 11:09 pm

Just tried a gentoo sourced kernel. No go; same error. I'd really appreciate any more suggestions from anyone. Even being able to identify the exact cause of the compile failure would be a plus.

PaulBredbury · Post by **PaulBredbury** » Fri Jan 11, 2008 11:40 am

Look at your own output:

Code: Select all

sh ./conftest.sh

Go take a look at that file.

bjk2007 · Post by **bjk2007** » Fri Jan 11, 2008 2:15 pm

Look at your own output:
Code:
sh ./conftest.sh

Go take a look at that file.

I just did:

Code: Select all

#updatedb
#locate conftest.sh

as root, but nothing was found. It appears that this file does not exist.

PaulBredbury · Post by **PaulBredbury** » Fri Jan 11, 2008 7:04 pm

Well, look for it properly:

Code: Select all

find /var/tmp/portage -name conftest.sh

It will be there, unless you've deleted it - it's what shows the error message you're seeing

bjk2007 · Post by **bjk2007** » Sat Jan 12, 2008 2:57 am

I guess I've become too dependent on locate and never bothered to learn find. That'll be useful in the future.

Unfortunately, as much as I'd like to complete this problem, I'm just out of time since this server was supposed to be put into production before Christmas when we ran into freak hardware failures, not to mention the other projects I have backing up. We ended up being able to get the nv drivers to work after a new X -configure though, so we do at least have a GUI. Hopefully we've narrowed it down far enough to where someone who has this issue in the future can find this post and solve their problem. A big thank you to everyone who provided assistance as I tried to get this working!

For anyone who finds this thread in the future trying to solve the same problem, I did manage to match some of the output to a section of the file starting at line 285 ("select_makefile)"). If a person were to start from there they could fully identify the conditions that cause $RET to not be set to 0 in the nested if block starting on line 318. This is the cause of the long textual "Unable to determine the target kernel version" error. Hopefully you can figure out how to solve it from there.

darkless · Post by **darkless** » Thu Jan 17, 2008 9:48 am

I checked out conftest.sh and found out that it had a wrong asm folder reference. I needed to add the following symlink (I use x86 arch, modify as appropriate):

Code: Select all

# ln -s asm-x86 /lib/modules/`uname -r`/source/include/asm-i386

kernelOfTruth · Post by **kernelOfTruth** » Tue Jan 22, 2008 6:37 pm

*subscribes* unfortunately that didn't help

jcat · Post by **jcat** » Tue Jan 22, 2008 8:14 pm

I see you tried this in the end, but all I was thinking as I started reading this thread was, "just use the nv driver!".

Not much point in using the nvidia driver on a server, 2D performance for the nv driver is pretty good, and that's all you need.

Cheers,
jcat

spankmeister7 · Post by **spankmeister7** » Sun Feb 03, 2008 9:03 pm

I can't get this to work, either.

Before anyone asks, let me say that my /usr/src/linux symlink points directly to /usr/src/linux-2.6.24-gentoo.

Here's the build log output:

[32;01m*[0m Determining the location of the kernel source code
[32;01m*[0m Found kernel source directory:
[32;01m*[0m /usr/src/linux
[32;01m*[0m Found kernel object directory:
[32;01m*[0m /lib/modules/2.6.24-gentoo/build
[32;01m*[0m Found sources for kernel version:
[32;01m*[0m 2.6.24-gentoo
[32;01m*[0m Checking for MTRR support ...
[A[72C [34;01m[ [32;01mok[34;01m ][0m
>>> Unpacking source...
Creating directory NVIDIA-Linux-x86-100.14.19-pkg0
Verifying archive integrity... OK
Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86 100.14.19.............................................................................................................................
[32;01m*[0m Applying NVIDIA_glx-defines.patch ...
[A[72C [34;01m[ [32;01mok[34;01m ][0m
[32;01m*[0m Applying NVIDIA_glx-glheader.patch ...
[A[72C [34;01m[ [32;01mok[34;01m ][0m
[32;01m*[0m Applying NVIDIA_i2c-hwmon.patch ...
[A[72C [34;01m[ [32;01mok[34;01m ][0m
[32;01m*[0m Converting Makefile.kbuild to use M= instead of SUBDIRS= ...
[A[72C [34;01m[ [32;01mok[34;01m ][0m
>>> Source unpacked.
>>> Compiling source in /var/tmp/portage/x11-drivers/nvidia-drivers-100.14.19/work/NVIDIA-Linux-x86-100.14.19-pkg0/usr/src/nv ...
[32;01m*[0m Preparing nvidia module
rm -f Makefile
sh ./conftest.sh "i686-pc-linux-gnu-gcc" "i686-pc-linux-gnu-gcc" /usr/src/linux /lib/modules/2.6.24-gentoo/build cc_sanity_check full_output
sh ./conftest.sh "i686-pc-linux-gnu-gcc" "i686-pc-linux-gnu-gcc" /usr/src/linux /lib/modules/2.6.24-gentoo/build select_makefile full_output
If you are using a Linux 2.4 kernel, please make sure
you either have configured kernel sources matching your
kernel or the correct set of kernel headers installed
on your system.

If you are using a Linux 2.6 kernel, please make sure
you have configured kernel sources matching your kernel
installed on your system. If you specified a separate
output directory using either the "KBUILD_OUTPUT" or
the "O" KBUILD parameter, make sure to specify this
directory with the SYSOUT environment variable or with
the equivalent nvidia-installer command line option.

Depending on where and how the kernel sources (or the
kernel headers) were installed, you may need to specify
their location with the SYSSRC environment variable or
the equivalent nvidia-installer command line option.

*** Unable to determine the target kernel version. ***

make: *** [select_makefile] Error 1
[31;01m*[0m
[31;01m*[0m ERROR: x11-drivers/nvidia-drivers-100.14.19 failed.
[31;01m*[0m Call stack:
[31;01m*[0m ebuild.sh, line 1701: Called dyn_compile
[31;01m*[0m ebuild.sh, line 1039: Called qa_call 'src_compile'
[31;01m*[0m ebuild.sh, line 44: Called src_compile
[31;01m*[0m nvidia-drivers-100.14.19.ebuild, line 214: Called linux-mod_src_compile
[31;01m*[0m linux-mod.eclass, line 519: Called die
[31;01m*[0m The specific snippet of code:
[31;01m*[0m emake HOSTCC="$(tc-getBUILD_CC)" CC="$(get-KERNEL_CC)" LDFLAGS="$(get_abi_LDFLAGS)" \
[31;01m*[0m ${BUILD_FIXES} ${BUILD_PARAMS} ${BUILD_TARGETS} \
[31;01m*[0m || die "Unable to make ${BUILD_FIXES} ${BUILD_PARAMS} ${BUILD_TARGETS}."
[31;01m*[0m The die message:
[31;01m*[0m Unable to make IGNORE_CC_MISMATCH=yes V=1 SYSSRC=/usr/src/linux SYSOUT=/lib/modules/2.6.24-gentoo/build clean module.
[31;01m*[0m
[31;01m*[0m If you need support, post the topmost build error, and the call stack if relevant.
[31;01m*[0m A complete build log is located at '/var/tmp/portage/x11-drivers/nvidia-drivers-100.14.19/temp/build.log'.
[31;01m*[0m

I really do not want to use the nv driver as was the resolution in the listing above. Is there any help on getting the nvidia-drivers set up?

Thanks!

emerge x11-drivers/nvidia-drivers-100.14.19 fail [resolved]

emerge x11-drivers/nvidia-drivers-100.14.19 fail [resolved]

I still need help with this.