On 27 June 2015 at 14:49, Jiong Wang <jiong.w...@arm.com> wrote: > > Andreas Schwab writes: > >> Jiong Wang <jiong.w...@arm.com> writes: >> >>> Andreas Schwab writes: >>> >>>> spawn -ignore SIGHUP /opt/gcc/gcc-20150627/Build/gcc/xgcc >>>> -B/opt/gcc/gcc-20150627/Build/gcc/ -fno-diagnostics-show-caret >>>> -fdiagnostics-color=never -O0 -w -fpie -c -o pr65593.o >>>> /opt/gcc/gcc-20150627/gcc/testsuite/gcc.c-torture/compile/pr65593.c >>>> /tmp/cc0Pymaf.s: Assembler messages: >>>> /tmp/cc0Pymaf.s:11: Error: unknown relocation modifier at operand 2 -- >>>> `ldr x0,[x0,#:gotpage_lo15:a]' >>>> /tmp/cc0Pymaf.s:19: Error: unknown relocation modifier at operand 2 -- >>>> `ldr x2,[x0,#:gotpage_lo15:bar]' >>> >>> Andreas, >>> >>> The binutils patch has been upstreamed already. Please checkout the >>> latest binutils code. >> >> That must work with the current binutils. > > I see, agree. > > Will adopt current ILP32 feature detection mechanism into -fpic. > > And some new TLS patches may need similiar check also. > > Thanks. > Hi Jiong,
I'm not sure I fully understand your answer. I believe you are going to change the testcase, right? I'm seeing whole toolchain build failures (when building glibc) since you committed this, because of the new relocations. I understand your suggestion is to move to trunk binutils to have support for these new relocations. Any other things I could do to avoid changing the binutils version I use? Thanks Christophe. > -- > Regards, > Jiong >
