GCC 9 introduces a new option, -flive-patching. It disables certain
optimizations which could make a compilation unsafe for later live
patching of the running kernel.
The option is used only if CONFIG_LIVEPATCH is enabled and $(CC)
supports it.
Performance impact of the option was measured on three different
Intel machines - two bigger NUMA boxes and one smaller UMA box. Kernel
intensive (IO, scheduling, networking) benchmarks were selected, plus a
set of HPC workloads from NAS Parallel Benchmark. The tests were done on
upstream kernel 5.0-rc8 with openSUSE Leap 15.0 userspace.
The majority of the tests is unaffected. The only significant exception
is the scheduler section which suffers 1-3% degradation.
Evaluated-by: Giovanni Gherdovich <[email protected]>
Signed-off-by: Miroslav Benes <[email protected]>
---
The patch is based on master branch of the livepatching tree on
git.kernel.org.
Many thanks to Giovanni who ran the whole testing campaign and analyzed
the results. I archived the dashboard and detailed data. If anyone is
interested, we could probably set up a public website somewhere.
Makefile | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/Makefile b/Makefile
index f070e0d65186..464a693a22e9 100644
--- a/Makefile
+++ b/Makefile
@@ -789,6 +789,10 @@ KBUILD_CFLAGS_KERNEL += -ffunction-sections -fdata-sections
LDFLAGS_vmlinux += --gc-sections
endif
+ifdef CONFIG_LIVEPATCH
+KBUILD_CFLAGS += $(call cc-option, -flive-patching=inline-clone)
+endif
+
# arch Makefile may override CC so keep this after arch Makefile is included
NOSTDINC_FLAGS += -nostdinc -isystem $(shell $(CC) -print-file-name=include)
--
2.21.0
On Thu, Apr 04, 2019 at 08:44:11PM +0200, Miroslav Benes wrote:
> GCC 9 introduces a new option, -flive-patching. It disables certain
> optimizations which could make a compilation unsafe for later live
> patching of the running kernel.
>
> The option is used only if CONFIG_LIVEPATCH is enabled and $(CC)
> supports it.
>
> Performance impact of the option was measured on three different
> Intel machines - two bigger NUMA boxes and one smaller UMA box. Kernel
> intensive (IO, scheduling, networking) benchmarks were selected, plus a
> set of HPC workloads from NAS Parallel Benchmark. The tests were done on
> upstream kernel 5.0-rc8 with openSUSE Leap 15.0 userspace.
>
> The majority of the tests is unaffected. The only significant exception
> is the scheduler section which suffers 1-3% degradation.
>
> Evaluated-by: Giovanni Gherdovich <[email protected]>
> Signed-off-by: Miroslav Benes <[email protected]>
> ---
> The patch is based on master branch of the livepatching tree on
> git.kernel.org.
>
> Many thanks to Giovanni who ran the whole testing campaign and analyzed
> the results. I archived the dashboard and detailed data. If anyone is
> interested, we could probably set up a public website somewhere.
Acked-by: Josh Poimboeuf <[email protected]>
While we're on the subject, I'm glad to see ongoing activity on
klp-convert (though I haven't had a chance to follow the discussions).
What's the status of the rest of the needed bits?
- kgraft-analysis-tool - will this also be submitted for inclusion in
the kernel tree?
- Documentation about how to create a patch
- The other bits I think Nicolai has been working on to eliminate all
the manual steps
--
Josh
> While we're on the subject, I'm glad to see ongoing activity on
> klp-convert (though I haven't had a chance to follow the discussions).
> What's the status of the rest of the needed bits?
>
> - kgraft-analysis-tool - will this also be submitted for inclusion in
> the kernel tree?
Yes, it is the next step.
> - Documentation about how to create a patch
I know we talked about it, but I don't remember what we came up with. We
should probably tie to Nicolai's work. It will change the process a lot.
> - The other bits I think Nicolai has been working on to eliminate all
> the manual steps
Miroslav
On Thu 2019-04-04 15:18:55, Josh Poimboeuf wrote:
> On Thu, Apr 04, 2019 at 08:44:11PM +0200, Miroslav Benes wrote:
> > GCC 9 introduces a new option, -flive-patching. It disables certain
> > optimizations which could make a compilation unsafe for later live
> > patching of the running kernel.
> >
> > The option is used only if CONFIG_LIVEPATCH is enabled and $(CC)
> > supports it.
> >
> > Performance impact of the option was measured on three different
> > Intel machines - two bigger NUMA boxes and one smaller UMA box. Kernel
> > intensive (IO, scheduling, networking) benchmarks were selected, plus a
> > set of HPC workloads from NAS Parallel Benchmark. The tests were done on
> > upstream kernel 5.0-rc8 with openSUSE Leap 15.0 userspace.
> >
> > The majority of the tests is unaffected. The only significant exception
> > is the scheduler section which suffers 1-3% degradation.
> >
> > Evaluated-by: Giovanni Gherdovich <[email protected]>
> > Signed-off-by: Miroslav Benes <[email protected]>
> > ---
> > The patch is based on master branch of the livepatching tree on
> > git.kernel.org.
> >
> > Many thanks to Giovanni who ran the whole testing campaign and analyzed
> > the results. I archived the dashboard and detailed data. If anyone is
> > interested, we could probably set up a public website somewhere.
>
> Acked-by: Josh Poimboeuf <[email protected]>
I have committed the patch into the branch for-5.2/core.
Best Regards,
Petr