RFR: 8252237: C2: Call to compute_separating_interferences has wrong argument order

[Aleksey Shipilev]

On Tue, 1 Dec 2020 07:35:33 GMT, Rahul Raghavan <rraghavan at openjdk.org> wrote:

> **# [.../src/hotspot/share/opto/coalesce.hpp] -**
> **112** uint compute_separating_interferences(Node *dst_copy, Node *src_copy, Block *b, uint bindex, RegMask &rm, **uint rm_size, uint reg_degree,** uint lr1, uint lr2);
> 
> 
> **# [.../src/hotspot/share/opto/coalesce.cpp] -**
> ..........
> **538** uint PhaseConservativeCoalesce::compute_separating_interferences(Node *dst_copy, Node *src_copy, Block *b, uint bindex, RegMask &rm, **uint reg_degree, uint rm_size,** uint lr1, uint lr2 ) {
> .................
> **747** reg_degree = compute_separating_interferences(dst_copy, src_copy, b, bindex, rm, **rm_size, reg_degree,** lr1, lr2 );
> 
> 
> **# So fixing the argument order - [.../src/hotspot/share/opto/coalesce.cpp]**
> -uint PhaseConservativeCoalesce::compute_separating_interferences(Node *dst_copy, Node *src_copy, Block *b, uint bindex, RegMask &rm, **uint reg_degree, uint rm_size,** uint lr1, uint lr2 ) {
> +uint PhaseConservativeCoalesce::compute_separating_interferences(Node *dst_copy, Node *src_copy, Block *b, uint bindex, RegMask &rm, **uint rm_size, uint reg_degree,** uint lr1, uint lr2 ) {

This change looks good, but my question is more wide: does it actually break/fixes any compiler bug? I would have thought changing RA would have some impact. In other words, are there test failures on higher tiers?

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

PR: https://git.openjdk.java.net/jdk/pull/1533