Created attachment 275869 [details]
See attached files
Created attachment 275871 [details]
Comment on attachment 275871 [details]
this log is useless. never post something built with MAKEFLAGS=-s in your bug reports. please re-emerge with a proper log and post that as an attachment.
would be useful to see if binutils-2.21 fails as well
This is weird.. I tried removing "-s" from the MAKEOPTS and now suddently it compiles just fine.
Is it ok? I don't think so.. now I will try to recompile it re-adding the "-s" to the make options and let see what happens..
Ok, confirmed. Readding '-s' to MAKEOPTS breaks the compiling phase.
Created attachment 276023 [details]
I've switched to i686-pc-linux-gnu-4.4.5 (from i686-pc-linux-gnu-4.4.5-hardenednopie) and I removed the "-s" flag from MAKEOPTS and it crashes again.
See attached build.log (it's a tarball that contains ~8mb plain text file) for more information
and have you tried binutils-2.21 ?
/usr/lib/gcc/i686-pc-linux-gnu/4.4.5/../../../../i686-pc-linux-gnu/bin/ld: BFD (GNU Binutils) 2.21 internal error, aborting at /var/tmp/portage/sys-devel/binutils-2.21/work/binutils-2.21/bfd/elf32-i386.c line 3042 in elf_i386_relocate_section
This is the snippet of code:
" if ((input_section->flags & SEC_ALLOC) == 0
|| h->plt.offset == (bfd_vma) -1)
And here is a comment from the previous if:
/* Since STT_GNU_IFUNC symbol must go through PLT, we handle
it here if it is defined in a non-shared object. */
So maybe a non shared object is in that command and shouldn't be? Not sure how to check that though :(
i can't reproduce with recent versions:
specifically i tried:
MAKEOPTS='-j2 -s' emerge glibc
(In reply to SpanKY from comment #11)
and that's on an hardened/linux/amd64/no-multilib system