On Tue, Mar 8, 2022 at 4:26 AM Sumit Garg <[email protected]> wrote: > > Hi, > > We have recently observed Linux kernel build failures while using SDK > toolchain (built using gcc-cross-canadian_<version>.bb) especially > after upstream changed the way to detect presence of GCC plugins here > [1]. Logs below: > > HOSTCXX scripts/gcc-plugins/arm_ssp_per_task_plugin.so > HOSTCC scripts/kallsyms > HOSTCC scripts/sorttable > HOSTCC scripts/asn1_compiler > SYSNR arch/arm/include/generated/asm/unistd-nr.h > GEN arch/arm/include/generated/asm/mach-types.h > SYSTBL arch/arm/include/generated/calls-oabi.S > SYSTBL arch/arm/include/generated/calls-eabi.S > HOSTCC scripts/mod/mk_elfconfig > CC scripts/mod/empty.o > cc1: error: cannot load plugin > ./scripts/gcc-plugins/arm_ssp_per_task_plugin.so > ./scripts/gcc-plugins/arm_ssp_per_task_plugin.so: undefined symbol: > _ZN8opt_pass14set_pass_paramEjb > make[1]: *** [scripts/Makefile.build:280: scripts/mod/empty.o] Error 1 > make: *** [Makefile:1214: prepare0] Error 2 > > This is reproducible with current OE core dunfell (gcc9.3) and > honister (gcc 11). But strangely this issue is not seen while building > kernel as part of OE build using normal cross toolchain (built using > gcc-cross_<version>.bb). > > I have tried to compare the two recipes: gcc-cross_<version>.bb and > gcc-cross-canadian_<version>.bb but can't find anything explicit which > can cause this behaviour. Also, following hack does disable GCC > plugins detection during kernel build leading to a successful build:
It is likely a variant of a bug that I'm working on now, feel free to add your steps to reproduce the issue there. https://bugzilla.yoctoproject.org/show_bug.cgi?id=14681 Cheers, Bruce > > diff --git a/meta/recipes-devtools/gcc/gcc-cross-canadian.inc > b/meta/recipes-devtools/gcc/gcc-cross-canadian.inc > index 553ef7fe62..5529932472 100644 > --- a/meta/recipes-devtools/gcc/gcc-cross-canadian.inc > +++ b/meta/recipes-devtools/gcc/gcc-cross-canadian.inc > @@ -73,8 +73,6 @@ FILES_${PN} = "\ > ${gcclibdir}/${TARGET_SYS}/${BINV}/lib* \ > ${gcclibdir}/${TARGET_SYS}/${BINV}/include \ > ${gcclibdir}/${TARGET_SYS}/${BINV}/include-fixed \ > - ${gcclibdir}/${TARGET_SYS}/${BINV}/plugin/include/ \ > - ${gcclibdir}/${TARGET_SYS}/${BINV}/plugin/gtype.* \ > ${includedir}/c++/${BINV} \ > ${prefix}/${TARGET_SYS}/bin/* \ > ${prefix}/${TARGET_SYS}/lib/* \ > > Thoughts? > > [1] > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=1e860048c53ee77ee9870dcce94847a28544b753 > > Regards, > Sumit > > > -- - Thou shalt not follow the NULL pointer, for chaos and madness await thee at its end - "Use the force Harry" - Gandalf, Star Trek II
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#162902): https://lists.openembedded.org/g/openembedded-core/message/162902 Mute This Topic: https://lists.openembedded.org/mt/89633477/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
