Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 328681 - x11-drivers/nvidia-drivers-173.14.27 - install: cannot stat `usr/include/cuda/*.h': No such file or directory
Summary: x11-drivers/nvidia-drivers-173.14.27 - install: cannot stat `usr/include/cuda...
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Library (show other bugs)
Hardware: x86 Linux
: High normal (vote)
Assignee: Michal Januszewski (RETIRED)
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-07-16 23:07 UTC by Alex Weiss
Modified: 2010-07-27 00:10 UTC (History)
6 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Alex Weiss 2010-07-16 23:07:32 UTC
Hi all,

Not much to say here save that, when portage tries to emerge the latest stable nvidia-drivers (173.14.27), I get the following before the program exits.

 * Installing nvidia module
install: cannot stat `usr/include/cuda/*.h': No such file or directory
!!! doins: usr/include/cuda/*.h does not exist
 * ERROR: x11-drivers/nvidia-drivers-173.14.27 failed:
 *   failed to install cuda headers
 * 
 * Call stack:
 *     ebuild.sh, line  54:  Called src_install
 *   environment, line 4124:  Called die
 * The specific snippet of code:
 *           doins usr/include/cuda/*.h || die "failed to install cuda headers";
 * 
 * If you need support, post the output of 'emerge --info =x11-drivers/nvidia-drivers-173.14.27',
 * the complete build log and the output of 'emerge -pqv =x11-drivers/nvidia-drivers-173.14.27'.
 * The complete build log is located at '/var/tmp/portage/x11-drivers/nvidia-drivers-173.14.27/temp/build.log'.
 * The ebuild environment file is located at '/var/tmp/portage/x11-drivers/nvidia-drivers-173.14.27/temp/environment'.
 * S: '/var/tmp/portage/x11-drivers/nvidia-drivers-173.14.27/work/NVIDIA-Linux-x86-173.14.27-pkg0'
 * QA Notice: file does not exist:
 * 
 *      doins: usr/include/cuda/*.h does not exist
 * QA Notice: make jobserver unavailable:
 * 
 *      make[2]: warning: jobserver unavailable: using -j1.  Add `+' to parent make rule.

>>> Failed to emerge x11-drivers/nvidia-drivers-173.14.27, Log file:

>>>  '/var/tmp/portage/x11-drivers/nvidia-drivers-173.14.27/temp/build.log'

 * Messages for package x11-drivers/nvidia-drivers-173.14.27:

 * ERROR: x11-drivers/nvidia-drivers-173.14.27 failed:
 *   failed to install cuda headers
 * 
 * Call stack:
 *     ebuild.sh, line  54:  Called src_install
 *   environment, line 4124:  Called die
 * The specific snippet of code:
 *           doins usr/include/cuda/*.h || die "failed to install cuda headers";
 * 
 * If you need support, post the output of 'emerge --info =x11-drivers/nvidia-drivers-173.14.27',
 * the complete build log and the output of 'emerge -pqv =x11-drivers/nvidia-drivers-173.14.27'.
 * The complete build log is located at '/var/tmp/portage/x11-drivers/nvidia-drivers-173.14.27/temp/build.log'.
 * The ebuild environment file is located at '/var/tmp/portage/x11-drivers/nvidia-drivers-173.14.27/temp/environment'.
 * S: '/var/tmp/portage/x11-drivers/nvidia-drivers-173.14.27/work/NVIDIA-Linux-x86-173.14.27-pkg0'


Reproducible: Always

Steps to Reproduce:
1.emerge -uD world
2.
3.

Actual Results:  
The emerge crashed.

Expected Results:  
For nvidia-drivers to emerge
Comment 1 Markos Chandras (RETIRED) gentoo-dev 2010-07-17 17:18:17 UTC
full build.log please
Comment 2 Alex Weiss 2010-07-17 17:25:36 UTC
Hi,

I just re-emerged it and something has apparently changed such that it works again. I suppose it's okay to close the bug.

Alex
Comment 3 Martin Väth 2010-07-17 20:53:53 UTC
I reopen the bug, because I can reproduce it:
The problem is that after the unpack phase there are no CUDA-related files in
the working directory for 173.14.27 (no such problems with 173.14.25)

More precisely, after the unpack phase of both 173.14.25 and 174.14.27 then
# find /var/tmp/portage/x11-drivers -name "*cuda*"
produces only the output

/var/tmp/portage/[...]/nvidia-drivers-173.14.25/[...]/usr/include/cuda
/var/tmp/portage/[...]/nvidia-drivers-173.14.25/[...]/usr/include/cuda/cudaGL.h
/var/tmp/portage/[...]/nvidia-drivers-173.14.25/[...]/usr/include/cuda/cuda.h
/var/tmp/portage/[...]/nvidia-drivers-173.14.25/[...]/usr/libcuda.so.173.14.25

but there is no such file (not even the directory /usr/include/cuda)
in the [...]/nvidia-drivers-173.14.27/[...] directory.

It looks as if these files are not contained in the archive or the archive
is broken somehow. (Since lots of other files _are_ unpackad, I suppose that
the unpacking function itself is OK.)
Comment 4 Matthew Schultz 2010-07-18 23:52:00 UTC
I'm able to reproduce this problem as well.
Comment 5 Scott Marks 2010-07-19 09:06:29 UTC
I just tried this after a fresh emerge --sync, and it failed in the same way.
Comment 6 Joel Parthemore 2010-07-20 11:13:41 UTC
Yep, same exact symptoms.  I am wondering whether to roll back to the earlier X server release or hope for a fix to this in the next few days.  Be happy to include my "emerge --info" if that helps, but it sounds like everyone is hitting the same thing.
Comment 7 Michal Januszewski (RETIRED) gentoo-dev 2010-07-27 00:10:12 UTC
This seems to have already been fixed by Doug (a check for the presence of the CUDA directory was added to the ebuild). Closing.