<div>Dale,</div>
<div> </div>
<div>I've looked a bit at ChooseConstraintTuple in llvm-gcc, as well as the code generator in Clang and the inline asm handling in LLVM (ala visitInlineAsm).</div>
<div> </div>
<div>While I see how Clang could do it to some extent by looking at the expression types of the operands, it seems that LLVM would be in a better position for choosing multiple alternative constraints as it would know what all the operands are, and where they currently are, particularly if part of an earlier transformation resulted in operands Clang thought were in memory were already in registers. And LLVm might need to step in anyway if registers can be allocated.</div>
<div> </div>
<div>Or is the constraint handling always performed before such changes as these would be made, such that Clang could know this?</div>
<div> </div>
<div>Or for a more off-the wall question, could asm constraints just be ignored, and let the target code in LLVM choose the appropriate instruction opcodes, and provide any additional piping needed to get the operands in and out of the given assembly instructions?</div>
<div> </div>
<div>But I guess that would be way more complicated than just selecting the best-matching constraint alternatives.</div>
<div> </div>
<div>What I learned from ChooseConstraintTuple in llvm-gcc is that the constraints need to be considered in parallel. That is, you collect the first alternative of each constraint string, give it a weight value, including a value meaning the combination won't work, collect the second alternative of each constraint string, give it a weigh, and if better than the first, save it, and so on for each constraint string, finally using the one with the best weight, or erroring if none match at all.</div>
<div> </div>
<div>This would require a few things:</div>
<div> </div>
<div>1. Clang needs to properly output the constraints. I recently changed Clang to only output the first alternative, as it was outputting the comma, which confused the code generator. Since apparently multiple constraint letters can be used (not the true multiple-alternative constraints), we probably would need to put in some separator character, such as a vertical bar character ("|") to delimit the constraints. I think I even saw a mention of "|" at one point in looking at the LLVM code, as if maybe it was considered at one time, but I couldn't tell what it applied to.</div>
<div> </div>
<div>2. Clang might need some change to the constraint representation with respeced to the logic it uses to determine if an operand is to be indirect (using "*"), because seeing "r" and "m" in alternatives results in both the "allowsRegister" and "allowsMemory" flags being set, and the allowsMemory flag is used to add the indirect flag. However, I saw code in LLVM that will insert code for the indirection even if the indirect flag is not present, such that I think we might be able to eliminate that flag and let LLVM deal with it.</div>
<div> </div>
<div>3. LLVM needs code to select the alterinative constraint to use. To do this, the inline asm handler probably needs to be changed to store the alternatives separately to make it easier to walk them for the selection, perhaps using a SubConstraintInfo class in an array owned by ConstraintInfo. After the selection, which determines the one subconstraint index needed, the data for the selected alternatives could then be copied to the ConstraintInfo class to reduce impact on the code using it.</div>
<div> </div>
<div>I'm willing to have a go at doing this myself, but I think it might be beyond me, particularly since I have no experience modifying LLVM yet, and this is a pretty hairy section. So it would be great if you or Chris or whoever really knows this code could take it on.</div>
<div> </div>
<div>Of course, this is just my initial brainstorming proposal, to see if you guys would agree that the back-end is more appropriate for the alternative constraint selection.</div>
<div> </div>
<div>-John</div>
<div><br> </div>
<div class="gmail_quote">On Thu, Jul 29, 2010 at 11:41 AM, Dale Johannesen <span dir="ltr"><<a href="mailto:dalej@apple.com">dalej@apple.com</a>></span> wrote:<br>
<blockquote style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class="gmail_quote">
<div style="WORD-WRAP: break-word">
<div>The LLVM asm parser doesn't support multiple alternative constraints; you need to have Clang pick one tuple in the front end. See ChooseConstraintTuple in llvm-gcc for prior art.</div><br>
<div>
<div>
<div></div>
<div class="h5">
<div>On Jul 29, 2010, at 11:36 AMPDT, John Thompson wrote:</div><br></div></div>
<blockquote type="cite">
<div>
<div></div>
<div class="h5">
<div>I'm trying to fix the handling of multiple alternate constraints in Clang (see <a href="http://gcc.gnu.org/onlinedocs/gcc/Multi_002dAlternative.html#Multi_002dAlternative" target="_blank">http://gcc.gnu.org/onlinedocs/gcc/Multi_002dAlternative.html#Multi_002dAlternative</a>).</div>
<div> </div>
<div>How should the constraints be represented in the .ll file?</div>
<div> </div>
<div>Clang currently will assert because the code generator sees a constraints string with the wrong number of comma-separated items.</div>
<div><br></div>
<div>Basically, after some editing, it just concatenates them together. The commas used in the string for a multiple alternative constraint gets confused with the commas separating the constraints.</div>
<div> </div>
<div>Briefly looking at the LLVM asm parser, I can't really see anything that tells me multiple alternative constraints are supported.</div>
<div> </div>
<div>But if they are, please tell me the string format, and perhaps update the docs to show this.</div>
<div> </div>
<div>If they are not yet supported in LLVM, what should I do?</div>
<div> </div>
<div>-John<br clear="all"><br>-- <br>John Thompson<br><a href="mailto:John.Thompson.JTSoftware@gmail.com" target="_blank">John.Thompson.JTSoftware@gmail.com</a><br><br></div></div></div>_______________________________________________<br>
LLVM Developers mailing list<br><a href="mailto:LLVMdev@cs.uiuc.edu" target="_blank">LLVMdev@cs.uiuc.edu</a> <a href="http://llvm.cs.uiuc.edu/" target="_blank">http://llvm.cs.uiuc.edu</a><br><a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev</a><br>
</blockquote></div><br></div></blockquote></div><br><br clear="all"><br>-- <br>John Thompson<br><a href="mailto:John.Thompson.JTSoftware@gmail.com">John.Thompson.JTSoftware@gmail.com</a><br><br>