Messages in this thread | | | Date | Tue, 23 Feb 2010 23:53:38 +0530 | From | Rabin Vincent <> | Subject | Re: [PATCH 07/10] ftrace: pass KBUILD_CFLAGS to record_mcount.pl |
| |
On Tue, Feb 23, 2010 at 08:30:46AM -0500, Steven Rostedt wrote: > On Sun, 2010-02-14 at 01:18 +0530, Rabin Vincent wrote: > > On ARM, we have two ABIs, and the ABI used is controlled via a config > > option. Object files built with one ABI can't be merged with object > > files built with the other ABI. So, record_mcount.pl needs to use the > > same compiler flags as the kernel when generating the object file with > > the mcount locations. Ensure this by passing CFLAGS to the script. > > I'm fine with this for now, but I'm wondering if we should pass the > autoconf.h to the script and parse that instead. This would give us all > set config options that we can look at.
The correct ABI options are constructed using cc-option and friends, and that logic would need to be duplicated in recordmcount.pl if we had to parse autoconf.h.
I originally tried to add a new RECORDMCOUNT_CFLAGS variable in the ARM Makefile and have that used here instead of KBUILD_CFLAGS (since KBUILD_CFLAGS includes a whole bunch of possibly irrelevant options), but coudn't figure out why my new variable wasn't accessible in Makefile.build.
> > But I need to go back and start looking at converting recordmcount.pl to > recordmcount.c again. > > > > > Signed-off-by: Rabin Vincent <rabin@rab.in> > > --- > > scripts/Makefile.build | 3 ++- > > 1 files changed, 2 insertions(+), 1 deletions(-) > > > > diff --git a/scripts/Makefile.build b/scripts/Makefile.build > > index 0b94d2f..2535c11 100644 > > --- a/scripts/Makefile.build > > +++ b/scripts/Makefile.build > > @@ -209,7 +209,8 @@ ifdef CONFIG_FTRACE_MCOUNT_RECORD > > cmd_record_mcount = set -e ; perl $(srctree)/scripts/recordmcount.pl "$(ARCH)" \ > > "$(if $(CONFIG_CPU_BIG_ENDIAN),big,little)" \ > > "$(if $(CONFIG_64BIT),64,32)" \ > > - "$(OBJDUMP)" "$(OBJCOPY)" "$(CC)" "$(LD)" "$(NM)" "$(RM)" "$(MV)" \ > > + "$(OBJDUMP)" "$(OBJCOPY)" "$(CC) $(KBUILD_CFLAGS)" \ > > + "$(LD)" "$(NM)" "$(RM)" "$(MV)" \ > > Doesn't this need to be with the change of recordmcount.pl too? > Otherwise applying this patch alone will break it. We need every step of > the patches to work otherwise we risk breaking a kernel bisect.
Dynamic ftrace for ARM is only enabled in the last patch of the series, so this woudn't be breaking it. No modification should be required for record_mcount.pl for other archs because this doesn't add/remove any parameters.
Rabin
| |