Some theoretically unchanged functions can have undesired changes if the
compiler decides to perform inlining in a different way (e.g. because of
newly added references). In such a case, it can be useful to discard
changes to functions that don't actually need modification.
Sadly, this currently doesn't work for functions missing the ftrace hook
(e.g. notrace code) as presence of the hook is checked before
identifying elements to ignore.
Look for functions/sections to ignore earlier.
Signed-off-by: Julien Thierry <jthierry@redhat.com>
In shadow-newpid.patch, a new structure member is created using the
klp_shadow_get_or_alloc() routine. The simplified and explained version
of it in the patch author guide has a typo that replaces
klp_shadow_get_or_alloc() with klp_shadow_get().
Signed-off-by: Yannick Cote <ycote@redhat.com>
Replace module-call-external.patch with a more fleshed-out version which
tests multiple things (including jump tables for newer kernels), and
also includes a runtime test.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Starting with Linux 5.8, vmlinux-specific KLP relas are applied early,
before all the special section initializations are done.
This means that jump labels can now be supported for cases where the
corresponding static keys live in the core kernel (vmlinux).
It also means that paravirt patching and alternatives can also now be
supported without the need for the .klp.arch sections.
This simplifies things greatly for newer kernels. We just have to make
sure that module-specific KLP relas aren't created for special sections.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
There were a few kernels (4.7 and 4.8) which didn't have support for
.klp.arch sections, but for which we still tried to use
CONFIG_LIVEPATCH. Those are inherently buggy, so just drop
CONFIG_LIVEPATCH support for them altogether.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
kpatch.ko has been quietly deprecated for a while, because there are
some known issues, including special section initialization ordering
issues. Starting with Linux 5.7, it will be completely broken because
kallsyms_lookup_name() will no longer be exported.
Add a warning to make its deprecation status more obvious.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Do some kpatch-build script cleanups to improve readability. This
is only a cleanup and shouldn't affect any functionality.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
With Linux commit 5190044c2965 ("modpost: move the namespace field in
Module.symvers last"), the format of Module.symvers has changed yet
again.
Use a completely different approach for figuring out the format. If a
column has "vmlinux", assume that's the "Module" column.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
We have recently encountered a situation when a patched function
had more than one jump label (static branches with the same static key
used to turn on/off some debugging feature). As it is often the case
with jump labels, their locations were far from obvious in the source
code, hidden in the chains of inline functions.
create-diff-object, however, exits after it has reported one jump label.
This is inconvenient, because, after one updates the patch to avoid
that jump label, the next build of the binary patch reveals another
one and fails again, and so on. It can be very time-consuming.
Let us report all jump labels first.
Before this commit the messages looked like this:
kpatch-build/create-diff-object: ERROR: dev.o:
kpatch_regenerate_special_section: 2084:
Found a jump label at ploop_req_state_process()+0x220, using key css_stacks_on.
Jump labels aren't currently supported. Use static_key_enabled() instead.
After:
dev.o: Found a jump label at ploop_req_state_process+0x220, key: css_stacks_on.
dev.o: Found a jump label at ploop_ioctl+0x2708, key: css_stacks_on.
kpatch-build/create-diff-object: ERROR: dev.o:
kpatch_regenerate_special_section: 2123:
Found 2 jump label(s) in the patched code.
Jump labels aren't currently supported. Use static_key_enabled() instead.
Signed-off-by: Evgenii Shatokhin <eshatokhin@virtuozzo.com>
The dynrela (aka .klp.rela) conversion logic is notoriously complex and
fragile. Simplify it and improve the comments.
This is mainly a cosmetic change. In theory it shouldn't change
functionality or break anything.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
On powerpc, "readelf -s" of vmlinux shows both .dynsym and .symtab.
.dynsym is just a subset of .symtab, so skip it to avoid duplicates.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
symtab_read() is quite fragile because it relies on the fact that the
first and second loops have the exact same conditions.
Instead just change the first loop to count all the lines in the file,
to get an upper bound for allocation. It's ok to over-allocate
slightly.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Reverse the if condition and use a 'continue' statement to reduce
indentation and improve readability.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
A symbol in the original object might get split in several sub-functions
in the patched object, which can themselves be bundled (and use a
separate rela section). References to local static variables from the
original function, might have been moved in one of the sub-functions
in the patched object.
Look for references to local static variables in the rela section
of child symbols in the patched object.
Signed-off-by: Julien Thierry <jthierry@redhat.com>
Consider symbols containing .part. in their names as sub-function
of the symbols they are derived from (if such symbol still exists in the
object file).
Signed-off-by: Julien Thierry <jthierry@redhat.com>
A symbol associated to a function can be split into multiple
sub-functions. Currently, kpatch only supports one child per function.
Extend this to support an arbitrary number of sub-function per function.
Signed-off-by: Julien Thierry <jthierry@redhat.com>
When a child symbol has changed, the parent symbol is only needed
in the output object if the child symbol is unpatchable on its own.
This is the case when the child symbol does not have its own profiling
call.
Only include unchanged parent symbols if their child has changed and
the child does not have a function profiling call.
Signed-off-by: Julien Thierry <jthierry@redhat.com>
Disabling the test case, until we find a workaround for functions
with the static key getting inlined. Read the comments from
@joe-lawrence in #1086 for detailed analysis.
Suggested-by: Joe Lawrence <joe.lawrence@redhat.com>
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Increment version to v0.9.1 and update v0.9.0's changelog section
following many fixes and improvements.
*Note* that the tree has been tagged to v0.9.0 earlier at
(commit: fd8209aa00). This is to update the tree accordingly. Full
list of changes:
v0.9.1:
- Handle ppc64le toc with only constants
- Don't strip callback section symbols
- Integration tests update
- Fix -Wconversion warnings
- Process debug sections last
v0.9.0:
- Many fixes in integration tests and adding rhel-8.0
- Updates to documentation
- Many updates and additions to the patch author guide
- Fix to relocations used for ZERO_PAGE(0)
- Simplify static local variables correlation
- Make symvers reading code more flexible
- Free sections in elf teardown
- Fix kpatch-test module unloading
- Disable the build of kpatch.ko module by default
- Simplify mangled function correlation
- Use whole word filename matching in find_parent_obj()
- Simplify relocation processing
Signed-off-by: Yannick Cote <ycote@redhat.com>
Update the rebase_patches hack^H^H^H tool to make it easier to define
environment variables to drive it, rather than sourcing source
/etc/os-release for everything. Fix a bunch of shellcheck warnings
along the way.
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
When a ppcle64 ".toc" section contains only constants, the compiler
might not (won't?) create a corresponding ".rela.toc" section.
In such cases, create-diff-object crashes, assuming ".rela.toc" exists
whenever .toc exists. Simply report that no rela are available when
looking up possible relocations in .toc.
Fixes#1078.
Signed-off-by: Julien Thierry <jthierry@redhat.com>
A binary patch may be used to fix network-related issues, so it is better to
apply it before networking services have started.
We encountered a situation in Virtuozzo 7, when the older kernels
conflicted with a new NetworkManager, ip utility and other system
components (https://www.mail-archive.com/devel@openvz.org/msg35123.html).
Binary patches were provided for these kernels to fix the issue but were
loaded after networking services in some cases. As a result, NetworkManager
and some other system components failed to work properly.
Let us make sure the patches are applied earlier during boot.
Signed-off-by: Evgenii Shatokhin <eshatokhin@virtuozzo.com>
Fetch pull-request data before resetting to a specified commit. This
will allow us to run integration tests on pull-request using master
repo, without cloning the original.
Signed-off-by: Artem Savkov <asavkov@redhat.com>
Julien Thierry
Yannick Cote
Yannick Cote
Josh Poimboeuf
Evgenii Shatokhin
Joe Lawrence
Yannick Cote
Kamalesh Babulal
Artem Savkov