On Thu, Jun 27, 2013 at 08:29:11AM +0200, Harald Hoyer wrote:
On 06/25/2013 06:22 PM, Dave Jones wrote:
> On Fri, Jun 14, 2013 at 04:02:22PM +0200, Harald Hoyer wrote:
> > On 06/13/2013 09:08 AM, harald(a)redhat.com wrote:
> > >
> > > This patch changes the version of kernel flavor/variants to end
> > > in "+$flavor" instead of ".$flavor". This makes it
easier to detect
> > > a flavor and a parser can separate it from the architecture.
> > >
> > > With that change we can correct kernel-install (of systemd) to call
> > > new-kernel-package with --package kernel-$flavor, because the $flavor
> > > can easily be extracted from the version string.
> > > ---
> > >
> > > After talking to the rpm team, "+" is the only char left as a
separator.
> > >
> > > Kernel compiles and runs:
> > > $ uname -r
> > > 3.10.0-0.rc5.git0.1.fc20.x86_64+debug
> > > $ echo /lib/modules/*debug
> > > /lib/modules/3.10.0-0.rc5.git0.1.fc20.x86_64+debug
> >
> >
> > Any comments?
>
> Let's just apply it and see what happens.
>
> Needs a refresh against current though..
>
> Applying: kernel.spec: let flavors/variants end with "+$flavor" in the
uname
> error: patch failed: kernel.spec:1871
> error: kernel.spec: patch does not apply
>
> Dave
>
ok, I have rebased the patch. One thing though with current git HEAD.
The recent changes with %{pae} show in the build log:
+ popd
~/data/git/fedora/kernel/kernel-3.9.fc20/linux-3.10.0-0.rc6.git0.2.fc20.x86_64
+ /usr/lib/rpm/find-debuginfo.sh --strict-build-id -r -p
'.*/usr/bin/perf(.debug)?|.*/usr/libexec/perf-core/.*|XXX' -o
perf-debuginfo.list -p
'.*/usr/lib64/python2.7/site-packages/perf.so(.debug)?|XXX' -o
python-perf-debuginfo.list -p
'.*/usr/bin/centrino-decode(.debug)?|.*/usr/bin/powernow-k8-decode(.debug)?|.*/usr/bin/cpupower(.debug)?|.*/usr/lib64/libcpupower.*|.*/usr/bin/turbostat(.debug)?|.*/usr/bin/x86_energy_perf_policy(.debug)?|XXX'
-o kernel-tools-debuginfo.list -p
'/.*/3.10.0-0.rc6.git0.2.fc20.x86_64/.*|/.*3.10.0-0.rc6.git0.2.fc20.x86_64(.debug)?'
-o debuginfo.list -p
'/.*/3.10.0-0.rc6.git0.2.fc20.x86_64.smp/.*|/.*3.10.0-0.rc6.git0.2.fc20.x86_64.smp(.debug)?'
-o debuginfosmp.list -p
'/.*/3.10.0-0.rc6.git0.2.fc20.x86_64.%{pae}/.*|/.*3.10.0-0.rc6.git0.2.fc20.x86_64.%{pae}(.debug)?'
-o 'debuginfo%{pae}.list' -p
'/.*/3.10.0-0.rc6.git0.2.fc20.x86_64.%{pae}debug/.*|/.*3.10.0-0.rc6.git0.2.fc20.x86_64.%{pae}debug(.debug)?'
-o 'debuginfo%{pae}debug.list' -p
'/.*/3.10.0-0.rc6.git0.2.fc20.x86_64.debug/.*|/.*3.10.0-0.rc6.git0.2.fc20.x86_64.debug(.debug)?'
-o debuginfodebug.list /home/harald/data/git/fedora/kernel/kernel-3.9.fc20
extracting debug info from
/home/harald/rpmbuild/BUILDROOT/kernel-3.10.0-0.rc6.git0.2.fc20.x86_64/usr/src/kernels/3.10.0-0.rc6.git0.2.fc20.x86_64/scripts/mod/mk_elfconfig
So, it _seems_ %{pae} was _not_ replaced in the %global. You might want to have
a look at this.
Huh. Seems it's expanded in the i686 build logs, but not x86_64.
josh