Nick,<br> <br> Is there a clean way to tell the difference between dst and src operands in operations without assignment "=" (ie, store)?<br> <br><br><div class="gmail_quote">On Mon, Oct 24, 2011 at 9:52 AM, Ryan Taylor <span dir="ltr"><<a href="mailto:ryta1203@gmail.com" target="_blank">ryta1203@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Nick,<br><br> I forgot to thank you, thanks!<div><div></div><div><br><br><div class="gmail_quote">On Sat, Oct 22, 2011 at 2:25 PM, Nick Lewycky <span dir="ltr"><<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div>Ryan Taylor wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Nick,<br>
<br>
Ah, forgot the -o, thanks, silly mistake.<br>
<br>
So how would you extract "add" from the instruction "%A"?<br>
</blockquote>
<br></div>
I->getOpcodeName().<div><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Yes, this is sort of what I am trying to do. The instnamer works fine<br>
for the local variables and I already had the constants sorted out.<br>
</blockquote>
<br></div>
Great!<br><font color="#888888">
<br>
Nick<br>
<br>
</font><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div>
<br>
On Sat, Oct 22, 2011 at 1:21 PM, Nick Lewycky <<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a><br></div><div><div></div><div>
<mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a>>> wrote:<br>
<br>
Ryan Taylor wrote:<br>
<br>
Nick,<br>
<br>
Also, I forgot to mention I had no luck with instnamer, it<br>
still left<br>
the local variables as "%slotNum", it didn't name them, unless I<br>
used<br>
-instnamer wrong:<br>
<br>
opt -instnamer <file.bc> file2.bc<br>
<br>
<br>
Almost, use -o to specify output: opt -instnamer file.bc -o file2.bc<br>
<br>
<br>
Those are the ones I am refering to. The description for<br>
instnamer says that it names unnamed instructions (not<br>
operands), or am I confused on the terminology here?<br>
<br>
<br>
Ah, I see. The operand of an instruction is some other Value, but<br>
it's not a subclass of Value like Instruction is. Allow me to elaborate.<br>
<br>
Here's some example IR:<br>
<br>
declare i32 @test(i32 %arg) {<br>
%A = add i32 %arg, 1<br>
%B = mul i32 %A, 2<br>
ret i32 %B<br>
}<br>
<br>
The llvm::Value hierarchy contains Instruction, Argument and<br>
Constant (among others). The operands of %A are "i32 %arg" and "i32<br>
1" where %arg is an Argument and 1 is a Constant.<br>
<br>
So, saying that "but it doesn't name operands" is moot, because it<br>
goes through and names all the arguments and instructions, which<br>
means that it's going to name all the operands -- except for constants.<br>
<br>
Firstly, constants (like "i32 1") aren't allowed to have names.<br>
Secondly, some Constants (GlobalValues which includes functions and<br>
global variables) are allowed to have names, but the instnamer won't<br>
touch them.<br>
<br>
<br>
For example, if I print out I->getName I get "add" not "x" or<br>
"y", but when I do Value *V = I->getOperands(loop) and then do<br>
V->getName, then it prints out the name of the operand. Am I<br>
going about this backwards? It sounds like it from the<br>
terminology you are using (calling an operand an instruction).<br>
<br>
<br>
If you have the Instruction* for "%A", then getName() will return<br>
"A", not "add". It may be the case that you have "%add = add i32<br>
%arg, 1" in which case it would return "add". :)<br>
<br>
If you call %A->getOperand(0) then you'll get the Value* whose<br>
getName() returns "arg", and also you can cast pointer to Argument*.<br>
<br>
<br>
I don't mean to be contentious (as I really appreciate your time<br>
and help) but apparently someone does use it, me. When going<br>
from source to source it's needed to keep track of the<br>
variables. Or am I missing something here too?<br>
<br>
<br>
Sure, no problem! I'm happy to explain how LLVM works.<br>
<br>
I'm not sure what you mean when you say you're going<br>
source-to-source through LLVM. Are you taking a language (say C++)<br>
compiling it to LLVM IR, then trying to produce another language<br>
back out (say Javascript)? I would give up on trying to map the<br>
output variable names back to the input ones. Think of LLVM IR like<br>
you would x86 assembly, that information is long gone.<br>
<br>
If you mean that you're doing LLVM IR -> LLVM IR, then instead of<br>
names use the Value pointers directly. Like names, they refer to the<br>
values.<br>
<br>
Nick<br>
<br>
<br>
?<br>
<br>
<br>
On Sat, Oct 22, 2011 at 12:51 PM, Nick Lewycky <<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a><br>
<mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a>><br></div></div><div><div></div><div>
<mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a> <mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a>>>> wrote:<br>
<br>
Ryan Taylor wrote:<br>
<br>
Nick,<br>
<br>
Unfortunately this doesn't answer my question I<br>
don't think. It<br>
seems that -instnamer, as you mention, names the<br>
instructions<br>
but still<br>
does not name the local variables.<br>
<br>
<br>
What other local variables are you referring to? When AsmWriter<br>
prints "%y = add i32 %x, 1", the name of that add<br>
instruction is "y"<br>
and "x" is the name of another instruction or argument. If<br>
it has no<br>
name, the AsmWriter emits a number ("%0"), by counting from<br>
the top.<br>
The only other locals could be function arguments, and instnamer<br>
names those too.<br>
<br>
<br>
So there really is no way to do this shy of creating (or<br>
basically<br>
copying) the API from AsmWriter (seems very dedundant to<br>
me)?<br>
This seems<br>
like a large failing?<br>
<br>
<br>
Correct, you'd have to copy that logic.<br>
<br>
It's not a large failing because nobody uses names of<br>
non-globals<br>
for anything. When we want to refer to a value, we use the<br>
Value*.<br>
<br>
Nick<br>
<br>
<br>
On Fri, Oct 21, 2011 at 7:03 PM, Nick Lewycky<br>
<<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a> <mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a>><br>
<mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a> <mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a>>><br>
<mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a> <mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a>><br>
<mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a> <mailto:<a href="mailto:nicholas@mxc.ca" target="_blank">nicholas@mxc.ca</a>>>>> wrote:<br>
<br>
Ryan Taylor wrote:<br>
<br>
It looks like the AsmWriter is generating the local<br>
variables<br>
(SlotNum)s<br>
on the fly in that file (AsmWriter.cpp), so is<br>
there any<br>
way at<br>
all to<br>
get this information from the operation itself, via<br>
Instruction,<br>
Value<br>
or Type?<br>
<br>
<br>
Nope! As you noticed, they're created on the fly...<br>
<br>
...when the Value or Type is anonymous. If you want<br>
them to be<br>
persistent, values can have names via. the setName()<br>
call. "opt<br>
-instnamer" will name all your instructions, for<br>
example.<br>
<br>
Nick<br>
<br>
<br>
<br>
<br>
<br>
<br>
</div></div></blockquote>
<br>
</blockquote></div><br>
</div></div></blockquote></div><br>