8271567: AArch64: AES Galois CounterMode (GCM) interleaved implementation using vector instructions

[theRealAph]

[New PR because of the 17u updates SNAFU.]

This is a backport of the big AES/GCM patch from JDK head. It's a
major change and it's had very little time (almost a day) to mature in
head, so I've set things up so that there is no change at all unless
you explicitly enable -XX:+UseAESGCMIntrinsics, which defaults
to false. When we've had more time we can perhaps turn it on by default.

---

Progress

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

Issue

• JDK-8271567: AArch64: AES Galois CounterMode (GCM) interleaved implementation using vector instructions

Reviewers

• Andrew Dinn (@adinn - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.java.net/jdk17u pull/216/head:pull/216
$ git checkout pull/216

Update a local copy of the PR:
$ git checkout pull/216
$ git pull https://git.openjdk.java.net/jdk17u pull/216/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 216

View PR using the GUI difftool:
$ git pr show -t 216

Using diff file

Download this PR as a diff file:
https://git.openjdk.java.net/jdk17u/pull/216.diff

[bridgekeeper]

👋 Welcome back aph! 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]

This backport pull request has now been updated with issue from the original commit.

[mlbridge]

Webrevs

• 02: Full - Incremental (02ebfa3a)
• 01: Full - Incremental (4adf21b3)
• 00: Full (20eb98fe)

[adinn]

Looks good.

[openjdk]

@theRealAph 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:

8271567: AArch64: AES Galois CounterMode (GCM) interleaved implementation using vector instructions

Reviewed-by: adinn

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 5 new commits pushed to the master branch:

• 25bfea6: 8276130: Fix Github Actions of JDK17u to account for update version scheme
• 22f1513: 8273575: memory leak in appendBootClassPath(), paths must be deallocated
• 38bfd62: 8273965: some testlibrary_tests/ir_framework tests fail when c1 disabled
• f0fb814: 8174819: java/nio/file/WatchService/LotsOfEvents.java fails intermittently
• 08cee6f: 8275104: IR framework does not handle client VM builds correctly

Please see this link for an up-to-date comparison between the source branch of this pull request and the master branch.
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.

➡️ To integrate this PR with the above commit message to the master branch, type /integrate in a new comment.

[theRealAph]

A small change: use the existing -XX:+UseAESCTRIntrinsicsflag rather than invent yet another one.

[adinn]

I believe that change was in when I reviewed this. So all is good.

[theRealAph]

Can I get some maintainer attention for this one? Adds AES/GCM acceleration, very low risk because change because it's disabled by default.

[DamonFool]

Hi @theRealAph ,

I also had a no risk backport to jdk11u openjdk/jdk11u-dev#599 .
But I noticed that you had just rejected it in the JBS without any comments.

So may I ask the reason for the rejection?
Thanks.

[theRealAph]

Hi @theRealAph ,

I also had a no risk backport to jdk11u openjdk/jdk11u-dev#599 . But I noticed that you had just rejected it in the JBS without any comments.

So may I ask the reason for the rejection? Thanks.

Absolutely you may, but you should ask in the right place, which is in the appropriate bug report.

[theRealAph]

/integrate

[openjdk]

Going to push as commit 076941f.
Since your change was applied there have been 62 commits pushed to the master branch:

• f1a8a22: 8276205: Shenandoah: CodeCache_lock should always be held for initializing code cache iteration
• 95573d4: 8268882: C2: assert(n->outcnt() != 0 || C->top() == n || n->is_Proj()) failed: No dead instructions after post-alloc
• f9eed14: 8275766: (tz) Update Timezone Data to 2021e
• 816b5cd: 8263059: security/infra/java/security/cert/CertPathValidator/certification/ComodoCA.java fails due to revoked cert
• e5c70af: 8276854: Windows GHA builds fail due to broken Cygwin
• e8eec1f: 8273595: tools/jpackage tests do not work on apt-based Linux distros like Debian
• 9ff9e62: 8271315: Redo: Nimbus JTree renderer properties persist across L&F changes
• e7e83d5: 8272369: java/io/File/GetXSpace.java failed with "RuntimeException: java.nio.file.NoSuchFileException: /run/user/0"
• 1d11bcd: 8273486: Zero: Handle DiagnoseSyncOnValueBasedClasses VM option
• 95726bf: 8276105: C2: Conv(D|F)2(I|L)Nodes::Ideal should handle rounding correctly
• ... and 52 more: https://git.openjdk.java.net/jdk17u/compare/ccfe697f2330cd385e6b9842412fc6cc86494ed8...master

Your commit was automatically rebased without conflicts.

[openjdk]

@theRealAph Pushed as commit 076941f.

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

[ijuma]

It's been nearly a year, can we consider enabling this by default now?

[theRealAph]

It's been nearly a year, can we consider enabling this by default now?

I think so. Perhaps for 11 too.