On Thu, 18 Jun 2026 17:14:06 GMT, Ashutosh Mehra <[email protected]> wrote:
>> This patch is reorganizing the code to split `get_processor_features()` into >> multiple helper functions. Details are present in the issue >> https://bugs.openjdk.org/browse/JDK-8380448. >> It should not change the behavior of any VM flag, except >> `UseVectorizedHashCodeIntrinsic`. The earlier code always set >> `UseVectorizedHashCodeIntrinsic` to `true` if UseAVX >= 2. That means >> specifying `-XX:-UseVectorizedHashCodeIntrinsic` on command line wouldn't >> disable use of this intrinsic. This looks like a bug and has been fixed in >> this patch. >> >> --------- >> - [x] I confirm that I make this contribution in accordance with the >> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai). > > Ashutosh Mehra has updated the pull request with a new target base due to a > merge or a rebase. The pull request now contains 11 commits: > > - Add code missed out when merging master > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Merge branch 'master' into vm_version_x86-reorg > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Address review comment > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Simplify setting of IntelJccErratumMitigation flag > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Fix setting of CopyAVX3Threshold > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Fix another mistake > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Fix build failure > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Replace COMPILER2_OR_JVMCI with COMPILER2 > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Merge branch 'master' into vm_version_x86-reorg > > Signed-off-by: Ashutosh Mehra <[email protected]> > - Minor updates > > Signed-off-by: Ashutosh Mehra <[email protected]> > - ... and 1 more: https://git.openjdk.org/jdk/compare/39d2d165...8821f232 src/hotspot/cpu/x86/globals_x86.hpp line 189: > 187: constraint(CopyAVX3ThresholdConstraintFunc,AfterErgo) > \ > 188: > \ > 189: product(bool, IntelJccErratumMitigation, false, DIAGNOSTIC, > \ Drive-by comment. So don't expect me to give a full review ;) Is this flag adequately tested? Because if you now make changes, I'm afraid we might not catch if it breaks. The only test I see that sets the flag is this one: `test/hotspot/jtreg/compiler/c2/irTests/TestPadding.java` And it only sets the flag to `on`. We should have at least a smoke test that both sets the flag to on and off. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/31301#discussion_r3440626605
