On Sat, Nov 19, 2022 at 3:01 PM Ian Rogers <irogers@google.com> wrote:
>
> On Fri, Nov 18, 2022 at 9:47 AM Nick Desaulniers
> <ndesaulniers@google.com> wrote:
> >
> > On Thu, Nov 17, 2022 at 2:03 PM Ian Rogers <irogers@google.com> wrote:
> > >
> > > Install libsubcmd and then get headers from there, this avoids
> > > inadvertent dependencies on things in tools/lib. Fix V=1
> > > support. Clean up how HOSTCC is used to override CC to avoid CFLAGS
> > > being set for say gcc, and then CC being overridden to clang. Support
> > > HOSTCFLAGS as a make option.
> >
> > Hi Ian,
> > Thanks for the patch set! I'm particularly interested in fixing the
> > plumbing of HOSTCFLAGS into objtool properly.
> >
> > Does this patch series depend on recent work you were doing to perf
> > and libsubcmd? When I apply it to mainline (84368d882b96) or
> > linux-next (20221118), I observe build failures with it.
> >
> > ```
> > /bin/sh: line 1: cd:
> > /android0/kernel-all/tools/objtool/libsubcmd/tools/objtool/: No such
> > file or directory
> > ../../scripts/Makefile.include:17: *** output directory
> > "/android0/kernel-all/tools/objtool/libsubcmd/tools/objtool/" does not
> > exist.  Stop.
> > make[3]: *** [Makefile:82:
> > /android0/kernel-all/tools/objtool/libsubcmd/libsubcmd.a] Error 2
> > make[3]: *** Waiting for unfinished jobs....
> > builtin-check.c:6:10: fatal error: 'subcmd/parse-options.h' file not found
> > #include <subcmd/parse-options.h>
> >          ^~~~~~~~~~~~~~~~~~~~~~~~
> > ```
> > Which I also observe when just doing a regular kernel build and not
> > just tools/objtool/.
> >
> > I tried to apply other recent series first, like
> > https://lore.kernel.org/lkml/20221109184914.1357295-1-irogers@google.com/
> > https://lore.kernel.org/linux-perf-users/20221116072211.2837834-2-irogers@google.com/
> > but wasn't quite able to get the result to build.
> >
> > My plan to validate HOSTCFLAGS is:
> > $ make LLVM=1 -j128 mrproper defconfig tools/objtool HOSTCFLAGS=-Os
> > V=1 1> before.txt
> > $ <apply patch series>
> > $ make LLVM=1 -j128 mrproper defconfig tools/objtool HOSTCFLAGS=-Os
> > V=1 1> after.txt
> > $ meld before.txt after.txt
>
> Thanks Nick! I think:
> https://lore.kernel.org/lkml/20221109184914.1357295-3-irogers@google.com/
> is necessary and already in Arnaldo's tree:
> https://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git/log/?h=perf%2Fcore
> However, with that the output directory in libsubcmd is calculated
> incorrectly and results in the error you see. I'll work on a v2 patch
> set.
>
> Thanks,
> Ian

So clearing subdir when invoking the make for libsubcmd rectified the
issue. v2 patch set is here:
https://lore.kernel.org/lkml/20221122001125.765003-1-irogers@google.com/
I've attached the before.txt and after.txt, they show the -Os is being
passed correctly in the after.txt case.

Thanks,
Ian

> > >
> > > Ian Rogers (3):
> > >   objtool: Install libsubcmd in build
> > >   objtool: Properly support make V=1
> > >   objtool: Alter how HOSTCC is forced
> > >
> > >  tools/objtool/Build    |  2 --
> > >  tools/objtool/Makefile | 68 +++++++++++++++++++++++++++++-------------
> > >  2 files changed, 48 insertions(+), 22 deletions(-)
> > >
> > > --
> > > 2.38.1.584.g0f3c55d4c2-goog
> > >
> >
> >
> > --
> > Thanks,
> > ~Nick Desaulniers