My previous RFC[1] tried to add recordmcount as the mcount subcommand
of objtool. As a necessary first step that required enabling building
of objtool for more than the x86 architecture.
Some folks have been working on enabling objtool checking functionality
for arm64. Rather than repeat that work here I aim to show a minimal
set which ensures that objtool builds for any architecture. This
will allow for not only building the check and ORC subcommands
but also incorporating more subcommands -- such as recordmcount.
I changed the Makefile to use the SRCARCH to determine which subcommands
should be available and set the appropriate SUBCMD_ variables. Those
variables then get used in the Build file to conditionally build them into
objtool. When the files are missing suitable empty definitions are located
in arch/missing so that the compilation will succeed while also allowing
objtool to report that the command is unavailable on the architecture.
Since the series does not add support for stack validation or checking
to any new architectures there's no reason to make KConfig or Makefile
changes which would normally be used to test this. So I've been forcing
builds of objtool with:
make O=build-ARCH ARCH=foo CROSS_COMPILE=foo-linux-gnu- defconfig
make O=build-ARCH ARCH=foo CROSS_COMPILE=foo-linux-gnu- tools/objtool
And running the resulting binary in qemu-static to verify that it
shows all objtool subcommands are supported on x86 and unsupported on
another arch.
[1] https://lore.kernel.org/lkml/[email protected]/
Matt Helsley (5):
objtool: Exit successfully when requesting help
objtool: Move struct objtool_file into arch-independent header
objtool: Add support for relocations without addends
objtool: Enable compilation of objtool for all architectures
objtool: Report missing support for subcommands
tools/objtool/Build | 10 +++--
tools/objtool/Makefile | 11 +++++-
tools/objtool/arch.h | 40 +++++++++++++++++++
tools/objtool/arch/missing/Build | 3 ++
tools/objtool/arch/missing/check.c | 16 ++++++++
tools/objtool/arch/missing/orc_dump.c | 13 ++++++
tools/objtool/arch/missing/orc_gen.c | 16 ++++++++
tools/objtool/arch/x86/Build | 1 +
tools/objtool/{ => arch/x86}/special.c | 4 +-
tools/objtool/{ => arch/x86}/special.h | 2 +-
tools/objtool/builtin-check.c | 5 +++
tools/objtool/builtin-orc.c | 9 ++++-
tools/objtool/builtin.h | 2 +
tools/objtool/check.c | 5 ++-
tools/objtool/check.h | 48 +---------------------
tools/objtool/elf.c | 55 +++++++++++++++++++++-----
tools/objtool/elf.h | 5 ++-
tools/objtool/objtool.c | 39 +++++++++++++++---
tools/objtool/objtool.h | 20 ++++++++++
tools/objtool/orc.h | 3 +-
tools/objtool/orc_dump.c | 1 +
tools/objtool/orc_gen.c | 3 ++
22 files changed, 235 insertions(+), 76 deletions(-)
create mode 100644 tools/objtool/arch/missing/Build
create mode 100644 tools/objtool/arch/missing/check.c
create mode 100644 tools/objtool/arch/missing/orc_dump.c
create mode 100644 tools/objtool/arch/missing/orc_gen.c
rename tools/objtool/{ => arch/x86}/special.c (98%)
rename tools/objtool/{ => arch/x86}/special.h (95%)
create mode 100644 tools/objtool/objtool.h
base-commit: 6e7f2eacf09811d092c1b41263108ac7fe0d089d
--
2.20.1
Hi Matt,
On 5/11/20 6:35 PM, Matt Helsley wrote:
> My previous RFC[1] tried to add recordmcount as the mcount subcommand
> of objtool. As a necessary first step that required enabling building
> of objtool for more than the x86 architecture.
>
> Some folks have been working on enabling objtool checking functionality
> for arm64. Rather than repeat that work here I aim to show a minimal
> set which ensures that objtool builds for any architecture. This
> will allow for not only building the check and ORC subcommands
> but also incorporating more subcommands -- such as recordmcount.
>
Thanks for putting the effort to simplify adding support of the
different command for new architectures.
I have a few some comments on some commits.
Cheers,
--
Julien Thierry