[jakobbotsch]

• Introduce GT_JTEST to replace GTF_JCMP_TEST
• Stop encoding JCMP conditions in GenTreeFlags by switching GT_JCMP to a GenTreeOpCC node. This removes GTF_JCMP_EQ and the LoongArch64/RISC-V specific mechanism to communicate condition codes to the backend via gtFlags.

[ghost]

Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch
See info in area-owners.md if you want to be subscribed.

Issue Details

---

• Introduce GT_JTEST to replace GTF_JCMP_TEST
• Stop encoding JCMP conditions in GenTreeFlags by switching GT_JCMP to a GenTreeOpCC node. This removes GTF_JCMP_EQ and the LoongArch64/RISC-V specific mechanism to communicate condition codes to the backend via gtFlags.

Author:	jakobbotsch
Assignees:	jakobbotsch
Labels:	area-CodeGen-coreclr
Milestone:	-

[jakobbotsch]

cc @shushanhf @clamp03. This is an alternative to #85217 that stops using GenTreeFlags to communicate the condition and cleans up the representation at the same time.

Also, unrelated to this PR, but I had to apply several patches to get the RISC-V/LoongArch cross-jits to build: https://gist.github.com/jakobbotsch/a11071cec499ef260fbe1b5c85c1b32f
You may want to look into some of these and why they're necessary.

[shushanhf]

cc @shushanhf @clamp03. This is an alternative to #85217 that stops using GenTreeFlags to communicate the condition and cleans up the representation at the same time.

Thansk.

Also, unrelated to this PR, but I had to apply several patches to get the RISC-V/LoongArch cross-jits to build: https://gist.github.com/jakobbotsch/a11071cec499ef260fbe1b5c85c1b32f You may want to look into some of these and why they're necessary.

I will add LA64's image for CI. dotnet/dotnet-buildtools-prereqs-docker#856

[jakobbotsch]

cc @dotnet/jit-contrib PTAL @TIHan -- related to what we have discussed a couple of times. No diffs.

@clamp03 @shushanhf If you can test the LA/RISC-V changes it would be great.

[shushanhf]

cc @dotnet/jit-contrib PTAL @TIHan -- related to what we have discussed a couple of times. No diffs.

@clamp03 @shushanhf If you can test the LA/RISC-V changes it would be great.

OK，I will test it later.

[shushanhf]

cc @dotnet/jit-contrib PTAL @TIHan -- related to what we have discussed a couple of times. No diffs.
@clamp03 @shushanhf If you can test the LA/RISC-V changes it would be great.

OK，I will test it later.

I just tested some cases. I think this is ok.
All the tests will be finished tomorrow morning.

[shushanhf]

cc @dotnet/jit-contrib PTAL @TIHan -- related to what we have discussed a couple of times. No diffs.
@clamp03 @shushanhf If you can test the LA/RISC-V changes it would be great.

OK，I will test it later.

All the tests will be finished tomorrow morning.

@jakobbotsch
For LoongArch64, this is ok.
Thanks

[clamp03]

Thank you. I will check.
CC #84834

[clamp03]

Checked on RISC-V. It works well. Thank you.