llvm dev - Jan 2014 - [LLVMdev] How to force a MachineFunctionPass to be the last one ?

Hi,

I would like to execute a MachineFunctionPass after all other passes
which modify the machine code.
In other words, if we call llc to generate assembly file, that pass
should run right before the "Assembly Printer" pass.
Is there any official way to enforce this ?

Best regards,
Sebastien

On Jan 21, 2014, at 2:20 PM, sebastien riou <matic at nimp.co.uk> wrote:
> Hi,
> 
> I would like to execute a MachineFunctionPass after all other passes
> which modify the machine code.
> In other words, if we call llc to generate assembly file, that pass
> should run right before the "Assembly Printer" pass.
> Is there any official way to enforce this ?
This makes sense and has come up before in discussions. I can envision a set of
analyses that need to run after all MI modification but before asm printing. The
StackMapLiveness pass is one such example. No enforcement mechanism exists yet,
but it’s safe to say that nothing with modify MI after the addPreEmitPasses
hook. For now you can just schedule your pass at the end with comments.

We could introduce a fake MI analysis that should be invalidated by any pass
that mutates the MI. Then check that it’s valid during code emission. That part
is trivial.

I think the real issue is that MC lowering can potentially change opcodes, so
you can never really do this reliably in general :(
I’m not sure how fixable that problem is. Maybe others can provide more
suggestions.

-Andy

On Jan 21, 2014, at 3:20 PM, Andrew Trick <atrick at apple.com> wrote:
> 
> On Jan 21, 2014, at 2:20 PM, sebastien riou <matic at nimp.co.uk>
wrote:
> 
>> Hi,
>> 
>> I would like to execute a MachineFunctionPass after all other passes
>> which modify the machine code.
>> In other words, if we call llc to generate assembly file, that pass
>> should run right before the "Assembly Printer" pass.
>> Is there any official way to enforce this ?
> 
> This makes sense and has come up before in discussions. I can envision a
set of analyses that need to run after all MI modification but before asm
printing. The StackMapLiveness pass is one such example. No enforcement
mechanism exists yet, but it’s safe to say that nothing with modify MI after the
addPreEmitPasses hook. For now you can just schedule your pass at the end with
comments.
> 
> We could introduce a fake MI analysis that should be invalidated by any
pass that mutates the MI. Then check that it’s valid during code emission. That
part is trivial.
> 
> I think the real issue is that MC lowering can potentially change opcodes,
so you can never really do this reliably in general :(
> I’m not sure how fixable that problem is. Maybe others can provide more
suggestions.
You can add your pass to the end of

X86PassConfig::addPreEmitPass()

and it will currently be the last thing to run before the asm printer, but I
don’t know if thats a guarantee or not.

In theory anyone could add something after the call to preEmitPass and before
this code

FunctionPass *Printer = getTarget().createAsmPrinter(*this, *AsmStreamer);
PM.add(Printer);

but if they do you can shout loudly :)

Thanks,
Pete
> 
> -Andy
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev