So what can someone do to compile the binary NVIDIA drivers? Does anyone
have a list of files that I need to symlink? I have also tried
downloading the linux-source-2.6.29 package, but that fails to provide
some files the driver needs as well.
--
To UNSUBSCRIBE, email to
On Mon, Mar 30, 2009 at 08:19:32PM -0400, Lennart Sorensen wrote:
Out of couriosity (and to try and fix the nvidia driver build system)...
Documented by who? Debian or the linux kernel?
Documentation/kbuild/modules.txt in the Linux source.
Bastian
--
Fascinating, a totally parochial
On Tue, Mar 31, 2009 at 10:16:39AM +0200, Bastian Blank wrote:
On Mon, Mar 30, 2009 at 08:19:32PM -0400, Lennart Sorensen wrote:
Out of couriosity (and to try and fix the nvidia driver build system)...
Documented by who? Debian or the linux kernel?
Documentation/kbuild/modules.txt in the
On Sat, Mar 28, 2009 at 11:28:07PM +0100, Bastian Blank wrote:
severity 521515 wishlist
tags 521515 wontfix
thanks
On Fri, Mar 27, 2009 at 09:02:04PM -0400, Aaron M. Ucko wrote:
As of 2.6.29-1, that no longer
holds, causing trouble for packages such as
Russ Allbery r...@debian.org writes:
Could you open a bug on the openafs package when you get a chance
(tomorrow or whenever) and we'll continue this there? It looks like
Filed as #521745. Kernel maintainers, apologies for the topic drift here.
--
Aaron M. Ucko, KB1CJC (amu at
Processing commands for cont...@bugs.debian.org:
severity 521515 wishlist
Bug#521515: linux-headers-2.6.29-1-amd64: please restore symlinks into -common
Severity set to `wishlist' from `normal'
tags 521515 wontfix
Bug#521515: linux-headers-2.6.29-1-amd64: please restore symlinks into -common
severity 521515 wishlist
tags 521515 wontfix
thanks
On Fri, Mar 27, 2009 at 09:02:04PM -0400, Aaron M. Ucko wrote:
As of 2.6.29-1, that no longer
holds, causing trouble for packages such as openafs-modules-source that
don't entirely defer to the kbuild
u...@debian.org (Aaron M. Ucko) writes:
Forwarding to Russ, who maintains a package that this change (which came
without warning TTBOMK) seriously inconveniences; please keep him Cc:ed.
[...]
From: Bastian Blank wa...@debian.org
There is no reliable way to detect the include paths of the
Forwarding to Russ, who maintains a package that this change (which
came without warning TTBOMK) seriously inconveniences; please keep him
Cc:ed.
---BeginMessage---
severity 521515 wishlist
tags 521515 wontfix
thanks
On Fri, Mar 27, 2009 at 09:02:04PM -0400, Aaron M. Ucko wrote:
Russ Allbery r...@debian.org writes:
OpenAFS *does* use kbuild. Aaron, what exactly breaks? Example error
messages? Is it just the symlinking to standardize the names of the
header files across platforms that doesn't work?
Yes:
| CC [M]
u...@debian.org (Aaron M. Ucko) writes:
OpenAFS could probably adapt by changing h, netinet, and sys under
MODLOAD-* from symlinks to .../include/linux to directories containing
forwarding headers; I'm not sure which specific headers would need such
treatment, but I suspect there are quite a
u...@debian.org (Aaron M. Ucko) writes:
Not quite; I believe they'd become linux/in.h, linux/types.h, etc.
Oh, good point.
Moreover, a lot of the inclusions were indirect; for instance, the
#include directive for sys/types.h ultimately came from afs/stds.h
(copied from .../src/config/). As
Russ Allbery r...@debian.org writes:
egrep -r '#include +[](sys|netinet|h)/' *.[ch] LINUX/ | awk '{ print $2 }'
| sed -e 's,\.\./,,' -e 's/[]//g' | sort -u
and as near as I can tell, on Linux, all those headers have the same name
but just have no directory prefix. So netinet/in.h becomes
Aaron M. Ucko u...@debian.org writes:
Historically, /usr/src/linux-headers-VERSION-FLAVOR used to contain
symlinks into /usr/src/linux-headers-VERSION-common for anything not at
least potentially flavor-specific. As of 2.6.29-1, that no longer
holds, causing trouble for packages such as
Package: linux-headers-2.6.29-1-amd64
Version: 2.6.29-1
Severity: normal
Historically, /usr/src/linux-headers-VERSION-FLAVOR used to contain
symlinks into /usr/src/linux-headers-VERSION-common for anything not at
least potentially flavor-specific. As of 2.6.29-1, that no longer
holds, causing
15 matches
Mail list logo
