[OE-core] [PATCH V2] perf: make a copy of kernel source to perf workdir

[Richard Purdie]

On Mon, 2018-04-23 at 13:51 -0400, Bruce Ashfield wrote:
> On 2018-04-23 4:13 AM, Hongxu Jia wrote:
> > 
> > Since perf modify kernel source which shared by multiple
> > recipes, it probably caused kernel-devsrc compile failure
> > at world build.
> > ...
> > > 
> > > 0 blocks
> > > cpio: ./tools/perf/arch/arm/util/sedr7ORqk: Cannot stat:
> > No such file or directory
> > > 
> > > 0 blocks
> > ...
> > cpio tried to find a file at ${S}/tools/perf and failed
> > if the input list is not valid.
> > 
> > Make a copy of kernel shared source directory into a perf workdir
> > could fix the issue.
> > 
> > Drop `Fix for rebuilding' which is obsolete
> > 
> > Signed-off-by: Hongxu Jia <hongxu.jia at windriver.com>
> > ---
> >   meta/recipes-kernel/perf/perf.bb | 11 ++++++-----
> >   1 file changed, 6 insertions(+), 5 deletions(-)
> > 
> > diff --git a/meta/recipes-kernel/perf/perf.bb b/meta/recipes-
> > kernel/perf/perf.bb
> > index 872b5f0..8b40806 100644
> > --- a/meta/recipes-kernel/perf/perf.bb
> > +++ b/meta/recipes-kernel/perf/perf.bb
> > @@ -61,7 +61,7 @@ export PERL_ARCHLIB = "${STAGING_LIBDIR}${PERL_OW
> > N_DIR}/perl/${@get_perl_version
> >   
> >   inherit kernelsrc
> >   
> > -B = "${WORKDIR}/${BPN}-${PV}"
> > +S = "${WORKDIR}/${BP}"
> >   SPDX_S = "${S}/tools/perf"
> >   
> >   # The LDFLAGS is required or some old kernels fails due missing
> > @@ -119,11 +119,12 @@ do_install() {
> >   	fi
> >   }
> >   
> > -do_configure_prepend () {
> > -    # Fix for rebuilding
> > -    rm -rf ${B}/
> > -    mkdir -p ${B}/
> > +do_configure[prefuncs] += "git_clone_kernel_source"
> > +git_clone_kernel_source () {
> > +    git clone --single-branch ${STAGING_KERNEL_DIR} ${S}
> I probably would have used "cp -a -r", or similar. Since it
> is possible that a kernel may not actually be a git repo (linux-yocto
> always is, but that is just one data point).
> 
> Also, did you try just copying the tools SRC dir + whatever
> infrastructure is needed to build ? That will save us a LOT
> of i/o when moving the source around for perf to build.
> 
> But outside of those details/mechanics, the approach is sound.

Our "copy hardlink tree" function, whatever its called is sometimes of
use in these cases...

I think I'm the current owner of a bug in this area which I've sadly
not gotten to yet :(.

Also, have we considered the alternative - not copying tools/perf for
the kernel devsrc?

Cheers,

Richard