[PATCH] D114581: [RISCV] Fix a crash in decoding LMUL in VTYPE

Mon Nov 29 19:53:35 PST 2021

craig.topper added a comment.

In D114581#3160295 <https://reviews.llvm.org/D114581#3160295>, @benshi001 wrote:

> In D114581#3160270 <https://reviews.llvm.org/D114581#3160270>, @craig.topper wrote:
>
>> I've notified some of SiFive's binutils developers about the issues. Including that vlmul=0b100 causes objdump to print a regular vsetvli with `(null)` where the lmul string should go. They will be printing the immediate form.
>>
>> So I think the LLVM disassembler should decode any immediate value. The RISCVInstPrinter should print the immediate form if any extra bits or reserved encodings are used. The assembly parser should be updated to accept the immediate form if it doesn't already.
>
> So how about I do "RISCVInstPrinter should print the immediate form" in current patch? And I will make another patch for "assembly parser should be updated to accept the immediate form" ?

Sure.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D114581/new/

https://reviews.llvm.org/D114581