[Forgot to include llvmdev in my reply, sending a second time]
Duncan Sands wrote:
22:40|melis@juggle2:~/projects/llvm_gc> llc root.ll
22:40|melis@juggle2:~/projects/llvm_gc> gcc -o r root.s
error in /tmp/ccGhPZDb.o(.eh_frame); no .eh_frame_hdr table will be
the assembler you pasted assembles for me on x86-32 linux.
Ah, sorry for not being clear. The piece of assembler shown was for the
case where the gcroot call was removed and no exception handling code was
generated. I should have included the assembly for root.ll as shown.
I'm actually puzzled that exception handling code is generated, and more
so that it is caused by the llvm.gcroot call. If I comment that one out
all no exception handling stuff is created:
The EscapeEnumerator code in ShadowStackGC.cpp adds cleanup code using
invoke and unwind - might be related to this?
Okay, so that's probably where the EH stuff comes from. Does the assembler
error indicate a bug or is there something else I need to do (like using a
gcc flag, or perhaps using ld instead of gcc) to generate an executable?