Disassembler uses the same format for variables and registers

[ghost]

Consider:

        ;; v18 <- BinaryDoubleOp:20(+, v50, v72) T{_Double}
0xd8    1e69294c               faddd v12, v10, v9

The disassembler uses v[0-9]+ both for the temporary variables and for the FPU registers on Arm64.
This leads to some confusion as e.g. v18 might refer to both a variable and a register, holding different values.

We should probably come up with some separate naming scheme for the variables to distinguish them from the registers

[mraleph]

I am not sure where confusion can originate from. Disassembly can't refer to SSA names directly and IL instructions can't refer to register directly (only in location summaries) - so the namespaces are completely disjoint.

[ghost]

To me it's confusing that they get printed alongside each other.
Of course it's not the end of the world - with a little mental overhead I can easily tell which is which. But the less overhead the better.
It also makes it easier to for instance search back and forth to find the place where a given v is set or read if you don't get a number of false positives.

Are you reluctant to change these?

[mraleph]

Are you reluctant to change these?

No, I am just saying that I don't think it's really a problem. It's like reading two different languages that contain the same word, e.g. "bad" in English and Danish refer to different things but nobody is really confused because you know which language you are reading at any given time. I will acknowledge though that search clashes might be an issue (but again, you can easily filter things out with regexps - because IL comments start with \s*;;).

We could replace v with % or something similarly neutral (e.g. $).

[rmacnak-google]

For RISC-V, there's another similar overlap between registers and temps in unoptimized code called t0 etc.

        ;; t0 <- Constant(#Closure: (dynamic) => void from Function '_print@15225868': static.)
0x7f0e94700292    0207b283               ld t0, 32(pp)   Closure: (dynamic) => void from Function '_print@15225868': static.
0x7f0e94700296        1161               addi sp, sp, -8
0x7f0e94700298        e016               sd t0, 0(sp)

It's mildly annoying when searching through disassembly to get hits from the wrong kind.