Messages in this thread | | | From | Roland McGrath <> | Subject | Re: [PATCH 3/4] x86_64 ia32 syscall audit fast-path | Date | Mon, 21 Jul 2008 16:54:19 -0700 (PDT) |
| |
> do you have a delta patch against tip/x86/audit-speedup by any chance? > That is a topic branch of your previous drop, which got tested as well > to a certain degree. Would make it easier to see what changed, would > make the merge have a more nuanced history, etc. > > or, alternatively, if you have a -git based branch that i could pull, > that would be nice as well. (i can compare old-x86/audit-speedup to > new-x86/audit-speedup.)
Here is a GIT branch with freshly-rebased patches from today's upstream. There was some merging fixup to be done after the x86/step changes went in.
I don't think it's useful to have a history merged upstream that includes the original broken 3/4 patch. The newer version both has the straight line paths Linus wanted, and works right on AMD. If we had the old version in GIT, then there would be a point in the history that builds a kernel that's broken for all 32-bit processes on 64-bit AMD hardware. Anyone doing a bisect in the future could get unlucky hit that along the way, etc. Anyway, you'd have to redo this merging work with x86/step just to get one that builds now.
The following changes since commit 93ded9b8fd42abe2c3607097963d8de6ad9117eb: Linus Torvalds (1): Merge git://git.kernel.org/.../gregkh/usb-2.6
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/frob/linux-2.6-roland.git x86/auditsc
Roland McGrath (4): x86_64: remove bogus optimization in sysret_signal x86_64 syscall audit fast-path x86_64 ia32 syscall audit fast-path i386 syscall audit fast-path
arch/x86/ia32/ia32entry.S | 78 +++++++++++++++++++++++++++++++++++++++++-- arch/x86/kernel/entry_32.S | 48 ++++++++++++++++++++++++++- arch/x86/kernel/entry_64.S | 49 +++++++++++++++++++++++++--- kernel/auditsc.c | 3 +- 4 files changed, 166 insertions(+), 12 deletions(-)
Thanks, Roland
| |