kernel headers - Now a BIGGER Issue

Christopher A Williams chriswfedora at cawllc.com
Mon Jan 14 15:04:56 UTC 2013


On Sun, 2013-01-13 at 07:49 -0800, Tom London wrote:

<snip...>
-- 
Christopher A Williams <chriswfedora at cawllc.com>

> > found in:
> >
> > /usr/src/kernels/3.6.11-3.fc18.x86_64/include/linux
> >
> > Put that all together and I suspect that the actual issue isn't so much
> > the different tree structure - it's consistent on F18 between the 3.6
> > and 3.7 series kernels. Rather, something has changed in the
> > updates-testing kernel package which is causing the Workstation scripts
> > to not be able to properly traverse the tree to find what they need.
> >
> > Hope that makes sense! As mentioned, I believe an appropriate symlink
> > will provide a work-around, but we should look at both the development
> > packages and the workstation scripts to evaluate where the most
> > appropriate fix should go. I'm sure this won't be the only package that
> > has this issue.
> >
> > Unfortunately, since I'm getting on another airplane today, I won't have
> > much more opportunity to look at this for a while. Hopefully someone
> > else will be able to pick things up from here...
> >
> > Chris
> >
> > --
> > Christopher A. Williams <chriswfedora at cawllc.com>
> >
> 
> Not sure this helps, but:
> 
> [root at tlondon ~]# locate /usr/\*/linux/version.h
> /usr/include/linux/version.h
> /usr/src/kernels/3.8.0-0.rc2.git3.1.fc19.x86_64/include/generated/uapi/linux/version.h
> /usr/src/kernels/3.8.0-0.rc2.git4.2.local.fc19.x86_64/include/generated/uapi/linux/version.h
> /usr/src/kernels/3.8.0-0.rc2.git4.2.local2.fc19.x86_64/include/generated/uapi/linux/version.h
> /usr/src/kernels/3.8.0-0.rc3.git0.3.local.fc19.x86_64/include/generated/uapi/linux/version.h
> [root at tlondon ~]#
> [root at tlondon ~]# ls -l $(locate /usr/\*/linux/version.h)
> -rw-r--r--. 1 root root 97 Jan 11 10:40 /usr/include/linux/version.h
> -rw-r--r--. 4 root root 97 Jan  7 15:25
> /usr/src/kernels/3.8.0-0.rc2.git3.1.fc19.x86_64/include/generated/uapi/linux/version.h
> -rw-r--r--. 4 root root 97 Jan  7 15:25
> /usr/src/kernels/3.8.0-0.rc2.git4.2.local2.fc19.x86_64/include/generated/uapi/linux/version.h
> -rw-r--r--. 4 root root 97 Jan  7 15:25
> /usr/src/kernels/3.8.0-0.rc2.git4.2.local.fc19.x86_64/include/generated/uapi/linux/version.h
> -rw-r--r--. 4 root root 97 Jan  7 15:25
> /usr/src/kernels/3.8.0-0.rc3.git0.3.local.fc19.x86_64/include/generated/uapi/linux/version.h
> [root at tlondon ~]#

OK - Thanks for extra info. I'm now in a situation where I must do
something. The latest kernel update came through this morning and we now
have the kernel 3.7 headers problem with Workstation hitting the
branched repo as opposed to the updates-testing repo.

...And because of the way Workstation 9.0.1 is trying to deal with it, I
believe this now falls squarely back on the Fedora team to look at and
resolve. You can't blame Workstation or its licensing model - don't even
start.

The update to kernel 3.7.2-201.fc18.x86_64 has the same kernel headers
problem. Except Workstation comes back with a specific error message
saying that it cannot find the header files. It then specifically asks
for the location of the header files. it even gives you a nic file
picker that's defaulting to the directory /usr/src/kernels for you to
start looking. Giving it the
location /usr/src/kernels/3.7.2-201.fc18.x86_64 doesn't work.

I'm assuming this is the correct location for the kernel header files.

Since Workstation is clearly allowing you to specify the location of the
header files and still can't find them, something with the new kernel
3.7 is, in my opinion, broken.

Can we please have a look at this? I will BZ as necessary...

Chris




More information about the test mailing list