8278171: [vectorapi] Mask incorrectly computed for zero extending cast #6634

merykitty · 2021-12-01T12:03:15Z

When doing an unsigned upcast, we perform a signed cast followed by a bitwise and operation to clip the extended signed bit. The sign clip mask is currently calculated incorrectly, the correct mask should have the number of least significant bit set equal to the size of the source elements. This patch fixes this trivial issue and adds several tests for zero extension casts.

Thank you very much.

Progress

Change must not contain extraneous whitespace
Commit message must refer to an issue
Change must be properly reviewed

Issue

JDK-8278171: [vectorapi] Mask incorrectly computed for zero extending cast

Reviewers

Paul Sandoz (@PaulSandoz - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.java.net/jdk pull/6634/head:pull/6634
$ git checkout pull/6634

Update a local copy of the PR:
$ git checkout pull/6634
$ git pull https://git.openjdk.java.net/jdk pull/6634/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 6634

View PR using the GUI difftool:
$ git pr show -t 6634

Using diff file

Download this PR as a diff file:
https://git.openjdk.java.net/jdk/pull/6634.diff

bridgekeeper · 2021-12-01T12:04:33Z

👋 Welcome back merykitty! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

openjdk · 2021-12-01T12:07:02Z

@merykitty The following labels will be automatically applied to this pull request:

core-libs
hotspot

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing lists. If you would like to change these labels, use the /label pull request command.

merykitty · 2021-12-01T12:25:39Z

/label add hotspot-compiler

openjdk · 2021-12-01T12:26:38Z

@merykitty
The hotspot-compiler label was successfully added.

merykitty · 2021-12-02T08:59:20Z

@PaulSandoz Could you take a look at this PR? Also, could you create an issue for this PR, please.
Should this be split into 2, the first one fixes the bug and add tests while the second one implements the intrinsics.
Thank you very much.

PaulSandoz

I am inclined to separated out. Fix the bug, add the tests, and integrate for 18. Then enhance with the intrinsics for 19.

If you agree to that I will create two bugs.

PaulSandoz · 2021-12-02T19:46:39Z

src/hotspot/cpu/x86/x86.ad

+    case Op_VectorUCastB2X:
+      if (size_in_bits == 256 && UseAVX < 2) {
+        return false; // Implementation limitation
+      }
+      break;
+    case Op_VectorUCastS2X:
+      if (size_in_bits == 256 && UseAVX < 2) {
+        return false;
+      }
+      break;
+    case Op_VectorUCastI2X:
+      if (size_in_bits == 256 && UseAVX < 2) {
+        return false;
+      }
+      break;


Collapse cases, since each has the code code?

merykitty · 2021-12-02T22:04:36Z

@PaulSandoz Yes, I think that should be the case, thank you very much.

PaulSandoz · 2021-12-02T22:21:20Z

@merykitty here you go:

[vectorapi] Mask incorrectly computed for zero extending cast
https://bugs.openjdk.java.net/browse/JDK-8278171

[vectorapi] Add x64 intrinsics for unsigned (zero extended) casts
https://bugs.openjdk.java.net/browse/JDK-8278173

mlbridge · 2021-12-02T22:31:39Z

Webrevs

00: Full (4e913753)

merykitty · 2021-12-02T22:40:40Z

/label remove hotspot,hotspot-compiler

openjdk · 2021-12-02T22:41:04Z

@merykitty
The hotspot label was successfully removed.

The hotspot-compiler label was successfully removed.

PaulSandoz · 2021-12-02T23:22:05Z

Looks good. Running tests.

Something we should consider later: split this test out into smaller files and support the max vector. On supported platforms we could consider using the IR test framework to verify the correct IR node is generated.

PaulSandoz

Tests are good.

openjdk · 2021-12-03T02:31:48Z

⚠️ @merykitty the full name on your profile does not match the author name in this pull requests' HEAD commit. If this pull request gets integrated then the author name from this pull requests' HEAD commit will be used for the resulting commit. If you wish to push a new commit with a different author name, then please run the following commands in a local repository of your personal fork:

$ git checkout vectorUnsignedCastIntrinsics
$ git commit -c user.name='Preferred Full Name' --allow-empty -m 'Update full name'
$ git push

openjdk · 2021-12-03T02:31:50Z

@merykitty This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8278171: [vectorapi] Mask incorrectly computed for zero extending cast

Reviewed-by: psandoz

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 14 new commits pushed to the master branch:

19ce33d: 8258512: serviceability/sa/TestJmapCore.java timed out on macOS 10.13.6
c93552c: 8277069: [REDO] JDK-8276743 Make openjdk build Zip Archive generation "reproducible"
b8ac0d2: 8276681: Additional malformed Javadoc inline tags in JDK source
b310f30: 8277396: [TESTBUG] In DefaultButtonModelCrashTest.java, frame is accessed from main thread
15345e3: 8276661: (fs) UserDefinedFileAttributeView no longer works with long path (win)
652b5f8: 8276674: Malformed Javadoc inline tags in JDK source
e0f1fc7: 8277358: Accelerate CRC32-C
73a9654: 8276806: Use Objects.checkFromIndexSize where possible in java.base
30087cc: 8275342: Change nested classes in java.prefs to static nested classes
669c90e: 8278099: two sun/security/pkcs11/Signature tests failed with AssertionError
... and 4 more: https://git.openjdk.java.net/jdk/compare/7c4ef3abaff55b29dddc50f2aea8a04a3214f9fb...master

As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

As you do not have Committer status in this project an existing Committer must agree to sponsor your change. Possible candidates are the reviewers of this PR (@PaulSandoz) but any other Committer may sponsor as well.

➡️ To flag this PR as ready for integration with the above commit message, type /integrate in a new comment. (Afterwards, your sponsor types /sponsor in a new comment to perform the integration).

merykitty · 2021-12-03T03:41:59Z

/integrate

openjdk · 2021-12-03T03:42:58Z

@merykitty
Your change (at version 4e91375) is now ready to be sponsored by a Committer.

merykitty · 2021-12-03T03:44:09Z

Thanks a lot for your support and review. Do I need another review here?

PaulSandoz · 2021-12-03T17:05:02Z

/sponsor

PaulSandoz · 2021-12-03T17:06:03Z

Thanks a lot for your support and review. Do I need another review here?

No need for this one, for next PR for the intrinsics will likely need two HotSpot reviewers.

openjdk · 2021-12-03T17:07:20Z

Going to push as commit 2e30fa9.
Since your change was applied there have been 32 commits pushed to the master branch:

fbf096e: 8251400: Fix incorrect addition of library to test in JDK-8237858
0a09092: 8268288: jdk/jfr/api/consumer/streaming/TestOutOfProcessMigration.java fails with "Error: ShouldNotReachHere()"
0d938ce: 8278172: java/nio/channels/FileChannel/BlockDeviceSize.java should only run on Linux
0e7b6bc: 8278141: LIR_OpLoadKlass::_info shadows the field of the same name from LIR_Op
53a4342: 8278137: JFR: PrettyWriter uses incorrect year specifier
f723779: 8278079: C2: expand_dtrace_alloc_probe doesn't take effect in macro.cpp
3f28a21: 8278144: Javadoc for MemorySegment::set/MemorySegment::setAtIndex is missing throws tag
ba2a8e5: 8278205: jlink plugins should dump .class file in debug mode
18c54b4: 8274640: Cleanup unnecessary null comparison before instanceof check in java.desktop
dda8f26: 8278139: G1: Refactor G1BlockOffsetTablePart::block_at_or_preceding
... and 22 more: https://git.openjdk.java.net/jdk/compare/7c4ef3abaff55b29dddc50f2aea8a04a3214f9fb...master

Your commit was automatically rebased without conflicts.

openjdk · 2021-12-03T17:07:40Z

@PaulSandoz @merykitty Pushed as commit 2e30fa9.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

retain relevant changes

241ac80

openjdk bot added hotspot hotspot-dev@openjdk.org core-libs core-libs-dev@openjdk.org labels Dec 1, 2021

openjdk bot added the hotspot-compiler hotspot-compiler-dev@openjdk.org label Dec 1, 2021

PaulSandoz reviewed Dec 2, 2021

View reviewed changes

Merge branch 'master' into vectorUnsignedCastIntrinsics

105ac18

revert intrinsics

4e91375

merykitty changed the title ~~x86 - Vector unsigned casts intrinsics~~ 8278171: [vectorapi] Mask incorrectly computed for zero extending cast Dec 2, 2021

openjdk bot added the rfr Pull request is ready for review label Dec 2, 2021

openjdk bot removed hotspot hotspot-dev@openjdk.org hotspot-compiler hotspot-compiler-dev@openjdk.org labels Dec 2, 2021

PaulSandoz approved these changes Dec 3, 2021

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Dec 3, 2021

openjdk bot added the sponsor Pull request is ready to be sponsored label Dec 3, 2021

openjdk bot closed this Dec 3, 2021

openjdk bot added integrated Pull request has been integrated and removed ready Pull request is ready to be integrated rfr Pull request is ready for review sponsor Pull request is ready to be sponsored labels Dec 3, 2021

merykitty deleted the vectorUnsignedCastIntrinsics branch February 5, 2022 13:10

8278171: [vectorapi] Mask incorrectly computed for zero extending cast #6634

8278171: [vectorapi] Mask incorrectly computed for zero extending cast #6634

merykitty commented Dec 1, 2021 •

edited by openjdk bot

bridgekeeper bot commented Dec 1, 2021

openjdk bot commented Dec 1, 2021

merykitty commented Dec 1, 2021

openjdk bot commented Dec 1, 2021

merykitty commented Dec 2, 2021

PaulSandoz left a comment •

edited

PaulSandoz Dec 2, 2021

merykitty commented Dec 2, 2021

PaulSandoz commented Dec 2, 2021

mlbridge bot commented Dec 2, 2021

merykitty commented Dec 2, 2021

openjdk bot commented Dec 2, 2021

PaulSandoz commented Dec 2, 2021

PaulSandoz left a comment

openjdk bot commented Dec 3, 2021

openjdk bot commented Dec 3, 2021 •

edited

merykitty commented Dec 3, 2021

openjdk bot commented Dec 3, 2021

merykitty commented Dec 3, 2021

PaulSandoz commented Dec 3, 2021

PaulSandoz commented Dec 3, 2021

openjdk bot commented Dec 3, 2021

openjdk bot commented Dec 3, 2021

8278171: [vectorapi] Mask incorrectly computed for zero extending cast #6634

8278171: [vectorapi] Mask incorrectly computed for zero extending cast #6634

Conversation

merykitty commented Dec 1, 2021 • edited by openjdk bot

Progress

Issue

Reviewers

Reviewing

bridgekeeper bot commented Dec 1, 2021

openjdk bot commented Dec 1, 2021

merykitty commented Dec 1, 2021

openjdk bot commented Dec 1, 2021

merykitty commented Dec 2, 2021

PaulSandoz left a comment • edited

Choose a reason for hiding this comment

PaulSandoz Dec 2, 2021

Choose a reason for hiding this comment

merykitty commented Dec 2, 2021

PaulSandoz commented Dec 2, 2021

mlbridge bot commented Dec 2, 2021

Webrevs

merykitty commented Dec 2, 2021

openjdk bot commented Dec 2, 2021

PaulSandoz commented Dec 2, 2021

PaulSandoz left a comment

Choose a reason for hiding this comment

openjdk bot commented Dec 3, 2021

openjdk bot commented Dec 3, 2021 • edited

merykitty commented Dec 3, 2021

openjdk bot commented Dec 3, 2021

merykitty commented Dec 3, 2021

PaulSandoz commented Dec 3, 2021

PaulSandoz commented Dec 3, 2021

openjdk bot commented Dec 3, 2021

openjdk bot commented Dec 3, 2021

merykitty commented Dec 1, 2021 •

edited by openjdk bot

PaulSandoz left a comment •

edited

openjdk bot commented Dec 3, 2021 •

edited