Support Loongarch64 by Panxuefeng-loongson · Pull Request #1440 · java-native-access/jna

Panxuefeng-loongson · 2022-05-25T03:17:58Z

Co-Authored-By: Ao Qi

Panxuefeng-loongson · 2022-05-25T03:24:36Z

libffi code have been merged into upstream, see: libffi/libffi@f259a6f. I have successfully built JNA based on it, tests pass all.

Note: We have Loongarch machines in the GCC compile farm, see: https://cfarm.tetaneutral.net/machines/list/. I think you can build the Loongarch port code with it.

Panxuefeng-loongson · 2022-05-25T03:26:25Z

@dbwiddis @theaoqi Hi all, please review the code for this pr.

dbwiddis · 2022-05-25T03:26:33Z

Changes look sane -- there is one TODO in the code. Can you elaborate on if this is still uncertain?

Panxuefeng-loongson · 2022-05-25T03:33:32Z

Changes look sane -- there is one TODO in the code. Can you elaborate on if this is still uncertain?

TODO where I am not very sure, I need to check with my colleagues. Thank you for your valuable comments!

Panxuefeng-loongson · 2022-05-25T06:58:44Z

On my Loongarch machine，multiArchPath is loongarch64-linux-gnu. So, I delete Todo code and please review again. @dbwiddis @theaoqi

theaoqi

The change LGTM. I am not sure if we should wait for the libffi update which contains LoongArch.

theaoqi · 2022-05-25T03:44:26Z

        }
+        else if (Platform.ARCH.equals("loongarch64")) {
+            libc = "-gnuabi64";   // TODO: not sure
+        }


Maybe this could be removed, see https://github.com/loongson/jna/blob/ed826d277fedabbe44d54d392c1219c60c73f0ea/src/com/sun/jna/NativeLibrary.java#L987 . However checking again is suggested by me.

The comment is outdated, you already did it.

Panxuefeng-loongson · 2022-05-26T01:54:36Z

Hi, @dbwiddis. The change looks good to @theaoqi and I have revised your suggestion. Please review again. Thanks

matthiasblaesing

Thank you - generally looks sane to me. The build currently fails as no initial linux-loongarch64.jar is provided. Please initiate it from out-of-date.jar.

In the long run the embedded copy of libffi needs also be updated.

Panxuefeng-loongson · 2022-05-28T00:34:00Z

Thank you - generally looks sane to me. The build currently fails as no initial linux-loongarch64.jar is provided. Please initiate it from out-of-date.jar.

In the long run the embedded copy of libffi needs also be updated.

libffi just merged into upstream a few days ago, see: libffi/libffi@f259a6f and jna needs to sync with it.

By the way, I think linux-loongarch64.jar has been initialized in build.xml,see:
<copy file="${lib.native}/out-of-date.jar" tofile="${lib.native}/linux-loongarch64.jar" overwrite="true"/>

matthiasblaesing · 2022-05-28T19:37:29Z

Thank you - generally looks sane to me. The build currently fails as no initial linux-loongarch64.jar is provided. Please initiate it from out-of-date.jar.

By the way, I think linux-loongarch64.jar has been initialized in build.xml,see: <copy file="${lib.native}/out-of-date.jar" tofile="${lib.native}/linux-loongarch64.jar" overwrite="true"/>

It is, if a breaking change of the native interface would have been detected. This is not the case. You need to add the empty file by hand. Have a look at the unittests attached to this PR, which all fail because of this.

Panxuefeng-loongson · 2022-05-29T12:25:29Z

Hi, @matthiasblaesing . Thank you for your advice, I realized my mistake and fixed it.

matthiasblaesing · 2022-05-29T18:10:07Z

Thanks for the update. I have one (maybe two) last request(s) before this good to go in:

please add an entry to CHANGES.md regarding "Support for Loongarch64" to the features section of the next version
I had a look at the raw git commit. This looks in general good. I'm not sure regarding your name. I don't know chinese customs about how to record your name, but the current author entry looks more like a user name and less like a real name. Could please check that? (running git log -n 1 in your branch show you the recorded author information)

theaoqi · 2022-05-30T01:39:30Z

 --------
 * [#1433](https://github.com/java-native-access/jna/pull/1433): Add `CFEqual`, `CFDictionaryRef.ByReference`, `CFStringRef.ByReference` to `c.s.j.p.mac.CoreFoundation` - [@shalupov](https://github.com/shalupov)
 * [#978](https://github.com/java-native-access/jna/issues/978): Remove use of finalizers in JNA and improve concurrency for `Memory`, `CallbackReference` and `NativeLibrary` - [@matthiasblaesing](https://github.com/matthiasblaesing).
+* [#1440](https://github.com/java-native-access/jna/pull/1440): Support Loongarch64 - [@Panxuefeng-loongson](https://github.com/Panxuefeng-loongson).


“Support for LoongArch64” may be better.

“Support for LoongArch64” may be better.

Done.

Co-Authored-By: Ao Qi <[email protected]>

matthiasblaesing · 2022-05-30T19:00:32Z

Looks good to me. Thank you.

matthiasblaesing · 2022-05-30T19:00:32Z

Looks good to me. Thank you.

theaoqi mentioned this pull request May 25, 2022

Loongarch: Add Loongarch port #1406

Closed

Panxuefeng-loongson force-pushed the support-Loongarch64 branch from 350c0c1 to e5646c5 Compare May 25, 2022 06:54

theaoqi approved these changes May 25, 2022

View reviewed changes

matthiasblaesing reviewed May 27, 2022

View reviewed changes

Panxuefeng-loongson force-pushed the support-Loongarch64 branch from e5646c5 to 44210ec Compare May 29, 2022 12:20

Panxuefeng-loongson force-pushed the support-Loongarch64 branch 3 times, most recently from 91ebb08 to ac553b8 Compare May 30, 2022 01:25

theaoqi reviewed May 30, 2022

View reviewed changes

Panxuefeng-loongson force-pushed the support-Loongarch64 branch 2 times, most recently from 2d84162 to b0d9853 Compare May 30, 2022 01:47

Support for LoongArch64

0066b08

Co-Authored-By: Ao Qi <[email protected]>

Panxuefeng-loongson force-pushed the support-Loongarch64 branch from b0d9853 to 0066b08 Compare May 30, 2022 01:54

Panxuefeng-loongson requested a review from matthiasblaesing May 30, 2022 02:08

matthiasblaesing merged commit bf60e51 into java-native-access:master May 30, 2022

Panxuefeng-loongson deleted the support-Loongarch64 branch May 31, 2022 00:14

Panxuefeng-loongson mentioned this pull request Nov 15, 2022

Can we sync with the jna community to support the LoongArch architecture? renaissance-benchmarks/renaissance#369

Closed

matthiasblaesing mentioned this pull request Aug 17, 2023

add loongarch support #1547

Closed

Conversation

Panxuefeng-loongson commented May 25, 2022

Uh oh!

Panxuefeng-loongson commented May 25, 2022

Uh oh!

Panxuefeng-loongson commented May 25, 2022

Uh oh!

dbwiddis commented May 25, 2022

Uh oh!

Panxuefeng-loongson commented May 25, 2022

Uh oh!

Panxuefeng-loongson commented May 25, 2022

Uh oh!

theaoqi left a comment

Choose a reason for hiding this comment

Uh oh!

theaoqi May 25, 2022

Choose a reason for hiding this comment

Uh oh!

theaoqi May 25, 2022

Choose a reason for hiding this comment

Uh oh!

Panxuefeng-loongson commented May 26, 2022

Uh oh!

matthiasblaesing left a comment

Choose a reason for hiding this comment

Uh oh!

Panxuefeng-loongson commented May 28, 2022

Uh oh!

matthiasblaesing commented May 28, 2022

Uh oh!

Panxuefeng-loongson commented May 29, 2022

Uh oh!

matthiasblaesing commented May 29, 2022

Uh oh!

theaoqi May 30, 2022

Choose a reason for hiding this comment

Uh oh!

Panxuefeng-loongson May 30, 2022

Choose a reason for hiding this comment

Uh oh!

matthiasblaesing commented May 30, 2022

Uh oh!

matthiasblaesing commented May 30, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants