Skip to contentSkip to conversation

trivial fixes for git-am and fix for !RT_FULL

RT-3.6 Paul Gortmaker 2 participants 2 messages Oct 10, 2012 – Oct 11, 2012

lore
Share & tools
lore
#0 lore raw

I've used the 3.6 RT patches as a baseline for a git repo of patches (identical to how stable kernels are managed) since I like how it gives visibility into how each patch evolves and adds tracking to the series.

On that baseline I've layered on some trivial fixes (bad ordering in the From/Subject line, missing subjects etc) that trip up git-am and would need fixing anyway before considering them for mainlining. Aside from two whitespace things, these changes don't alter the final source at all.

There is one build fix which I've added for the !RT_FULL case; trivial missing ifdef usage.

I'll keep the master branch in this repo to track directly the exact content released by Thomas in his tarballs. The fixes I've done are on the v3.6.1-rt1-fixes branch that can be browsed here:

http://git.kernel.org/?p=linux/kernel/git/paulg/3.6-rt-patches.git;a=shortlog;h=refs/heads/v3.6.1-rt1-fixes

Q: Looks interesting. How would I use this?

A: If you have been using tglx's tarballs of patches, then use of this
is essentially identical. You'd simply do:

    git clone git://git.kernel.org/pub/scm/linux/kernel/git/paulg/3.6-rt-patches.git
    cd 3.6-rt-patches
    git checkout v3.6.1-rt1-fixes

and point quilt (or git-quilt-import etc) at your new dir of patches, just as you would have when using tglx's tarball of patches. Or checkout the master branch if you want to use tglx's latest.

Paul
---

The following changes since commit 19aa1730ceb5605bf3cf6e24aeb990643a925cfc:

RT: 3.6.1 baseline patch set (2012-10-09 19:23:14 -0400)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/paulg/3.6-rt-patches.git v3.6.1-rt1-fixes

for you to fetch changes up to 9d13f3023b994b00d649eb70a219bd57cc4ec41e:

timer.c: integrate build fix for !RT_FULL (2012-10-09 20:26:30 -0400)

----------------------------------------------------------------
Paul Gortmaker (10):
rename softirq-preempt-fix-3-re.txt to have a .patch suffix
signal-rt-fix: reorder Subject and From fix-rt-int3-x86_32-3.2-rt: reorder Subject and From
latency-hist: strip trailing newline panic-disable-random-on-rt: add Subject to the patch hrtimer-add-missing-debug_activate-aid: divorce from ancient announce rename hrtimer-add-missing-debug_activate-aid patch
futex-requeue-pi-fix: reorder Subject and From
ping-sysrq: fix space-before-tab warning
timer.c: integrate build fix for !RT_FULL

 fix-rt-int3-x86_32-3.2-rt.patch                    |  2 +-
 futex-requeue-pi-fix.patch                         |  2 +-
 ...=> hrtimer-add-missing-debug_activate-aid.patch | 23 +----------
 latency-hist.patch                                 |  3 +-
 oleg-signal-rt-fix.patch                           |  2 +-
 panic-disable-random-on-rt.patch                   |  1 +
 ping-sysrq.patch                                   |  2 +-
 series                                             |  5 ++-
 ...-fix-3-re.txt => softirq-preempt-fix-3-re.patch |  0
 timer.c-fix-build-fail-for-RT_FULL.patch           | 45 ++++++++++++++++++++++
 10 files changed, 55 insertions(+), 30 deletions(-)
 rename hrtimer-add-missing-debug_activate-aid-was-re-announce-3-0-6-rt17.patch => hrtimer-add-missing-debug_activate-aid.patch (77%)
 rename softirq-preempt-fix-3-re.txt => softirq-preempt-fix-3-re.patch (100%)
 create mode 100644 timer.c-fix-build-fail-for-RT_FULL.patch
#1 lore raw

Paul,

Quoted 9 lines · Paul Gortmaker

I've used the 3.6 RT patches as a baseline for a git repo of patches (identical to how stable kernels are managed) since I like how it gives visibility into how each patch evolves and adds tracking to the series.

On that baseline I've layered on some trivial fixes (bad ordering in the From/Subject line, missing subjects etc) that trip up git-am and would need fixing anyway before considering them for mainlining. Aside from two whitespace things, these changes don't alter the final source at all.

Quoted 6 lines

There is one build fix which I've added for the !RT_FULL case; trivial missing ifdef usage.

I'll keep the master branch in this repo to track directly the exact content released by Thomas in his tarballs. The fixes I've done are on the v3.6.1-rt1-fixes branch that can be browsed here:

Nice service. I grabbed it and replaced my own series with it, so we all can enjoy the fixed patch files in the next release.

Thanks,

tglx