[llvm-bugs] [Bug 27679] New: Gold 2.24 on ARMv7 Thumb2 fails to self-host Clang/LLVM

[via llvm-bugs]

https://llvm.org/bugs/show_bug.cgi?id=27679

            Bug ID: 27679
           Summary: Gold 2.24 on ARMv7 Thumb2 fails to self-host
                    Clang/LLVM
           Product: new-bugs
           Version: unspecified
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: renato.golin at linaro.org
                CC: llvm-bugs at lists.llvm.org
    Classification: Unclassified

When using Gold 2.24 on Thumb2, llvm-tlbgen fails. BFD ld of the same version
works fine.

Stack dump:
0.    Program arguments: /home/linaro/devel/llvm/bisect/build2/bin/llvm-tblgen
-gen-intrinsic -I /home/linaro/devel/llvm/bisect/llvm/include/llvm/IR -I
/home/linaro/devel/llvm/bisect/llvm/lib/Target -I
/home/linaro/devel/llvm/bisect/llvm/include
/home/linaro/devel/llvm/bisect/llvm/include/llvm/IR/Intrinsics.td -o
/home/linaro/devel/llvm/bisect/build2/include/llvm/IR/Intrinsics.gen.tmp 
Bus error (core dumped)

Luckily, gold is failing to link stage2 llvm-tblgen, instead of stage1 Clang,
which would be a lot harder to pick up.

The main difference seems to be on branch literals, so likely some relocation
problem:

 * GOOD:
bhi.n d4dae
<_ZN12_GLOBAL__N_119RegisterInfoEmitter13runTargetDescERN4llvm11raw_ostreamERNS1_13CodeGenT
...
bl  10315c <_ZN4llvm11raw_ostream5writeEPKcj>
...
b.n d4dc6
<_ZN12_GLOBAL__N_119RegisterInfoEmitter13runTargetDescERN4llvm11raw_ostreamERNS1_13CodeGenTar

 * BAD:

bhi.n d29be
<_ZN12_GLOBAL__N_119RegisterInfoEmitter13runTargetDescERN4llvm11raw_ostreamERNS1_13CodeGenT
...
bl  100d6c <_ZN4llvm11raw_ostream5writeEPKcj>                                   
...
b.n d29d6
<_ZN12_GLOBAL__N_119RegisterInfoEmitter13runTargetDescERN4llvm11raw_ostreamERNS1_13CodeGenTar

Both binaries attached for reference.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20160508/99e1cbc3/attachment.html>