[LLVMdev] How should LLVM interpreter handle llvm.objectsize.i64

Eric Christopher echristo at gmail.com
Thu May 9 13:14:58 PDT 2013

My suggestion would be to always lower to "unknown". See the code in,
IIRC, codegenprepare (or fast-isel) that does this.


On Thu, May 9, 2013 at 7:03 AM, Issac Torvalds <lyh.kernel at gmail.com> wrote:
> Hello LLVMer,
> I use dragonegg to generate LLVM bitcode. Then I use LLVM interpreter to
> execute what I get from compilation. However an error occurred and the error
> message is:
> "LLVM ERROR: Code generator does not support intrinsic function
> 'llvm.objectsize.i64'!."
> As far as I know, objectsize intrinsic is equivalent to gcc built-in
> function
> __builtin_object_size. But I don't call __builtin_object_size explicitly in
> my
> source code. Maybe headers what I include call the built-in somewhere. How
> do
> I eliminate such built-in function call in LLVM bitcode, or what code should
> I
> add into IntrinsicLowering.cpp to handle this problem.
> Any suggestion or comment is welcomed.
> Thanks
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

More information about the llvm-dev mailing list