Hello, I have a question about the indirectbr instruction. I attempted to use it according to the example in the Assembly Language Reference manual, but got an "expected instruction opcode" error. Poking about on the web I found this document: http://nondot.org/sabre/LLVMNotes/IndirectGoto.txt which appears to be a Nov 2, 2009 proposal to add indirectbr and blockaddress() to the IR language. So I suspect indirectbr is not supported in the llvm-as 2.5 packaged in Fedora 11, is that correct? Dustin
On Jan 2, 2010, at 11:03 AM, Dustin Laurence wrote:> Hello, > > I have a question about the indirectbr instruction. I attempted to > use > it according to the example in the Assembly Language Reference manual, > but got an "expected instruction opcode" error. Poking about on the > web > I found this document: > > http://nondot.org/sabre/LLVMNotes/IndirectGoto.txt > > which appears to be a Nov 2, 2009 proposal to add indirectbr and > blockaddress() to the IR language. So I suspect indirectbr is not > supported in the llvm-as 2.5 packaged in Fedora 11, is that correct?Yes, that is correct. It is supported in the trunk sources, but it has not yet been released.
On 01/02/2010 11:24 AM, Bob Wilson wrote:> Yes, that is correct.That *would* explain why I couldn't figure out how to make llvm-as understand it. :-)> ...It is supported in the trunk sources, but it has > not yet been released.OK. I'll stick with my workaround of using an integer state code and a switch statement for the moment just in the interests of minimizing the extra complexity beyond what I already have learning the LLVM IR, but at some point should think about just installing the trunk version (I did a trial build and had no problems, so I don't anticipate that would be a huge headache). Dustin
On 01/02/2010 11:24 AM, Bob Wilson wrote:> Yes, that is correct. It is supported in the trunk sources, but it has > not yet been released.Hmm. Would the same also be true of the "inbounds" keyword for GEP? It doesn't seem to be recognized ("expected type"). Dustin