RFR: 8310577: Fix -Wconversion warnings in interpreter and oops [v2]

[Matias Saavedra Silva]

On Thu, 22 Jun 2023 14:19:16 GMT, Coleen Phillimore <coleenp at openjdk.org> wrote:

>> Please review this change that's mostly casts and a few int type narrowing changes.  The casts are for size calculations between two addresses within the same area so can be int and are never going to be 64 bit.
>> I didn't change CompileThreshold to be 'int' because SA and JVMCI code expects it to be intx.  Maybe that should change though.
>> Tested with tier1 on Oracle platforms, and tier1-4 linux and windows x64.
>
> Coleen Phillimore has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Fred's suggestion.

Looks good! I have a couple of concerns but they might not be an issue.

src/hotspot/share/interpreter/bytecodeUtils.cpp line 356:

> 354: StackSlotAnalysisData::StackSlotAnalysisData(BasicType type) : _bci(INVALID), _type(type) {}
> 355: 
> 356: StackSlotAnalysisData::StackSlotAnalysisData(int bci, BasicType type) : _bci((u2)bci), _type(type) {

Looking at this more closely, it seems like bci is meant to be a u2 but it needs one extra bit for the "INVALID" value. Maybe it's possible to represent invalid bci's differently? That way we can just set _bci to be a u2.

src/hotspot/share/interpreter/bytecodes.cpp line 443:

> 441: #endif
> 442:   jchar bc_flags = 0;
> 443:   if (can_trap)           bc_flags |= _bc_can_trap;

I believe that bitwise-or will promote the operands to int. Will this cause any issues here?

-------------

Marked as reviewed by matsaave (Committer).

PR Review: https://git.openjdk.org/jdk/pull/14604#pullrequestreview-1493479959
PR Review Comment: https://git.openjdk.org/jdk/pull/14604#discussion_r1238698168
PR Review Comment: https://git.openjdk.org/jdk/pull/14604#discussion_r1238721770