[RISC-V] Support displaying GTF_ADDRMODE_NO_CSE flag on GT_CAST

[namu-lee]

Follow-up to #122088.
Add support to display the GTF_ADDRMODE_NO_CSE flag on GT_CAST.

Before

N024 (  2,  5) [000389] -----+---U-                               |           |     +--*  CAST      long <- uint $386

After

N024 (  2,  5) [000389] -----+-N-U-                               |           |     +--*  CAST      long <- uint $386

Part of #84834, cc @dotnet/samsung

[dotnet-policy-service]

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

[Copilot]

Pull request overview

Updates CoreCLR JIT tree dumping so nodes marked with GTF_ADDRMODE_NO_CSE display as GTF_DONT_CSE (intended to help diagnose complex addressing-mode patterns, including those involving casts).

Changes:

• Extends the gtDispNode flag printing logic to include GT_CAST when deciding to display GTF_ADDRMODE_NO_CSE as GTF_DONT_CSE.

---

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

Updates CoreCLR JIT tree dump formatting to better reflect address-mode-related “no CSE” markings, including when they occur on cast nodes.

Changes:

• Extend gtDispNode’s flag-printing logic to consider IsPartOfAddressMode() nodes.
• Treat GT_CAST similarly to GT_ADD/GT_MUL/GT_LSH for displaying the GTF_ADDRMODE_NO_CSE intent (printed as GTF_DONT_CSE).

---

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

This PR refactors JIT tree debug flag printing in Compiler::gtDispNode to use the existing GenTree::IsPartOfAddressMode() helper when deciding to display GTF_ADDRMODE_NO_CSE-marked nodes as GTF_DONT_CSE. This keeps address-mode-related behavior consistent with other parts of the JIT that already use IsPartOfAddressMode().

Changes:

• Replace inline operator checks (GT_ADD/GT_MUL/GT_LSH + GTF_ADDRMODE_NO_CSE) with tree->IsPartOfAddressMode().
• Preserve existing behavior for clearing GTF_REVERSE_OPS on non-binary/non-multiop nodes, while making the control flow slightly more direct.

---

You can also share your feedback on Copilot code review. Take the survey.

[JulieLeeMSFT]

@namu-lee, please address test failures and convert this PR back when it is ready for review.

[jakobbotsch]

Test failures were presumably CI issues, going to rerun CI