[OE-core] [PATCH] gcc: Add recipes for 4.6.0

Tom Zanussi tom.zanussi at intel.com
Sun Apr 24 20:59:15 UTC 2011


On Sun, 2011-04-24 at 11:41 -0700, Khem Raj wrote:
> On Sat, Apr 23, 2011 at 6:01 PM, Tom Zanussi <tom.zanussi at intel.com> wrote:
> > On Wed, 2011-04-20 at 17:18 -0700, Kamble, Nitin A wrote:
> >> Met 1st build failure with gcc4.6.6. I think it is not a gcc issue, but the recipe code issue; And Tom/Darren can take care of it.
> >>
> >
> > Hi,
> >
> > The patchset applied fine for me, but I'm getting do_patch errors when building:
> >
> > *** Manual patch resolution mode ***
> > Dropping to a shell, so patch rejects can be fixed manually.
> > Run "quilt refresh" when patch is corrected, press CTRL+D to exit.
> >
> > Applying patch gcc-uclibc-locale-ctype_touplow_t.patch
> > patching file libstdc++-v3/config/locale/generic/c_locale.h
> > Hunk #1 FAILED at 41.
> > 1 out of 1 hunk FAILED -- saving rejects to file libstdc++-v3/config/locale/generic/c_locale.h.rej
> > patching file libstdc++-v3/config/os/gnu-linux/ctype_base.h
> > Hunk #2 FAILED at 47.
> > 1 out of 2 hunks FAILED -- saving rejects to file libstdc++-v3/config/os/gnu-linux/ctype_base.h.rej
> > patching file libstdc++-v3/config/locale/generic/c_locale.cc
> > Applied patch gcc-uclibc-locale-ctype_touplow_t.patch (forced; needs refresh)
> > trz at elmorro:/usr/local/src/yocto/test/build/tmp/work/x86_64-poky-linux/gcc-cross-intermediate-4.6.0-r0/gcc-4.6.0$
> >
> > I actually fixed a couple of these up, but they keep coming...
> >
> > So is there anything special I need to do to get this building cleanly?
> > I'm using master, and added the following to my local.conf:
> >
> > GCCVERSION = "4.6.0"
> > SDKGCCVERSION = "4.6.0"
> >
> > and did bitbake linux-yocto.
> >
> >
> > Once I have it working and can fix compile errors, I can send a patch to
> > fix it up if necessary, but note that this particular gcc warning
> > (unused-but-set-variable) is on its way to being disabled in the
> > upstream kernel.  See the lkml thread, "Re: annoying new gcc 4.6.0
> > warnings.":
> 
> I looked at the errors in kernel's do_compile_perf task
> 
> | bench/sched-pipe.c: In function 'bench_sched_pipe':
> | bench/sched-pipe.c:58:6: error: variable 'ret' set but not used
> [-Werror=unused-but-set-variable]
> | cc1: all warnings being treated as errors
> |
> | make: *** [bench/sched-pipe.o] Error 1
> 
> Looking at sourcecode of tools/perf/bench/sched-pipe.c there is a
> dummy local variable ret which is never used
> and seemingly is used to workaround another warning where return
> values of read()
> and write() were being ignored. I think warning is legitimate here
> 

Hmm, the comment in the code pretty much says that this use of 'ret' was
what the author intended:

        /*
         * why does "ret" exist?
         * discarding returned value of read(), write()
         * causes error in building environment for perf
         */

Anyway, looks like it's already been fixed upstream.  See commit
fb7d0b3cefb80a105f7fd26bbc62e0cbf9192822

which adds __used to the ret value.

By the way, I just added that change to just the ret value, and many
more come after it, so it's just the first of many of the same warning.
~3000, if you believe the thread...

Tom

> >
> > https://lkml.org/lkml/2011/4/21/325
> >
> > Tom
> >
> >> Thanks,
> >> Nitin
> >>
> >>
> >> | PERF_VERSION = 2.6.37.5.308.g69cfbd^M
> >> | make: Leaving directory `/build_disk/poky_build/build_gcc_4.6.0/tmp/work/qemux86_64-poky-linux/linux-yocto-2.6.37+git1+727fc4769aa920fc5bb3eb9a81bf92d0e6340903_1+69cfbdf9f1ff461a75e5b77d6d7ba35e97db4cc3-r16/linux/tools/perf'^M
> >> | make: Entering directory `/build_disk/poky_build/build_gcc_4.6.0/tmp/work/qemux86_64-poky-linux/linux-yocto-2.6.37+git1+727fc4769aa920fc5bb3eb9a81bf92d0e6340903_1+69cfbdf9f1ff461a75e5b77d6d7ba35e97db4cc3-r16/linux/tools/perf'^M
> >> | Makefile:512: No libdw.h found or old libdw.h found or elfutils is older than 0.138, disables dwarf support. Please install new elfutils-devel/libdw-dev^M
> >> | Makefile:568: newt not found, disables TUI support. Please install newt-devel or libnewt-dev^M
> >> |     GEN common-cmds.h^M
> >> |     * new build flags or prefix^M
> >> |     CC perf.o^M
> >> |     CC builtin-annotate.o^M
> >> |     CC builtin-bench.o^M
> >> |     CC bench/sched-messaging.o^M
> >> |     CC bench/sched-pipe.o^M
> >> | bench/sched-pipe.c: In function 'bench_sched_pipe':^M
> >> | bench/sched-pipe.c:58:6: error: variable 'ret' set but not used [-Werror=unused-but-set-variable]^M
> >> | cc1: all warnings being treated as errors^M
> >> | ^M
> >> | make: *** [bench/sched-pipe.o] Error 1^M
> >> | ERROR: Function 'do_compile_perf' failed (see /build_disk/poky_build/build_gcc_4.6.0/tmp/work/qemux86_64-poky-linux/linux-yocto-2.6.37+git1+727fc4769aa920fc5bb3eb9a81bf92d0e6340903_1+69cfbdf9f1ff461a75e5b77d6d7ba35e97db4cc3-r16/temp/log.do_compile_perf.24035 for further information)^M
> >> | make:^M
> >> NOTE: package linux-yocto-2.6.37+git1+727fc4769aa920fc5bb3eb9a81bf92d0e6340903_1+69cfbdf9f1ff461a75e5b77d6d7ba35e97db4cc3-r16: task do_compile_perf: Failed
> >>
> >>
> >> > -----Original Message-----
> >> > From: openembedded-core-bounces at lists.openembedded.org
> >> > [mailto:openembedded-core-bounces at lists.openembedded.org] On Behalf Of
> >> > Kamble, Nitin A
> >> > Sent: Wednesday, April 20, 2011 2:58 PM
> >> > To: Patches and discussions about the oe-core layer
> >> > Subject: Re: [OE-core] [PATCH] gcc: Add recipes for 4.6.0
> >> >
> >> > Khem,
> >> >   Good news. All the gcc recipes built without any errors for arm, ppc,
> >> > mips, x86, x86-64. Now I am trying to build the poky-image-sato-sdk
> >> > image for each architecture. Will keep you posted.
> >> >
> >> > Thanks & Regards,
> >> > Nitin
> >> >
> >> > _______________________________________________
> >> > Openembedded-core mailing list
> >> > Openembedded-core at lists.openembedded.org
> >> > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
> >
> >
> >






More information about the Openembedded-core mailing list