<AWT Dev> RFR: 8260931: Implement JEP 382: New macOS Rendering Pipeline

[Kevin Rushforth]

On Thu, 4 Feb 2021 10:35:02 GMT, Ajit Ghaisas <aghaisas at openjdk.org> wrote:

> **Description :**
> This is the implementation of [JEP 382 : New macOS Rendering Pipeline](https://bugs.openjdk.java.net/browse/JDK-8238361)
> It implements a Java 2D internal rendering pipeline for macOS using the Apple Metal API.
> The entire work on this was done under [OpenJDK Project - Lanai](http://openjdk.java.net/projects/lanai/)
> 
> We iterated through several Early Access (EA) builds and have reached a stage where it is ready to be integrated to openjdk/jdk. The latest EA build is available at - https://jdk.java.net/lanai/
> 
> A new option -Dsun.java2d.metal=true | True needs to be used to use this pipeline.
> 
> **Testing :**
> This implementation has been tested with the tests present at - [Test Plan for JEP 382: New macOS Rendering Pipeline](https://bugs.openjdk.java.net/browse/JDK-8251396)
> 
> **Note to reviewers :**
> 1) Default rendering pipeline on macOS has not been changed by this PR. OpenGL still stays as the default rendering pipeline and Metal rendering pipeline is optional to choose.
> 
> 2) To apply and test this PR - 
> To enable the metal pipeline you must specify on command line -Dsun.java2d.metal=true (No message will be printed in this case) or -Dsun.java2d.metal=True (A message indicating Metal rendering pipeline is enabled gets printed)

I left a few additional comments. Overall this looks good to me.

src/java.desktop/macosx/native/libawt_lwawt/java2d/metal/MTLGraphicsConfig.m line 82:

> 80:     (JNIEnv *env, jclass mtlgc)
> 81: {
> 82:     FILE *f = popen("/usr/sbin/system_profiler SPDisplaysDataType", "r");

How robust is this? It seems like the contents of this could be an implementation detail and subject to change. Is it documented by Apple?

src/java.desktop/macosx/native/libawt_lwawt/java2d/metal/MTLGraphicsConfig.m line 83:

> 81: {
> 82:     FILE *f = popen("/usr/sbin/system_profiler SPDisplaysDataType", "r");
> 83:     bool metalSupported = JNI_FALSE;

This code is mixing types; it should be `jboolean metalSupported`

src/java.desktop/macosx/native/libawt_lwawt/java2d/metal/MTLMaskFill.m line 26:

> 24:  */
> 25: 
> 26: #ifndef HEADLESS

I see a few occurrences of `#ifndef HEADLESS` in the metal pipeline. Is this needed? I don't see any of the other native macos files in Java2D (e.g., the OpenGL pipeline) doing the same. Will this file ever be compiled with HEADLESS being undefined?

src/java.desktop/macosx/native/libawt_lwawt/java2d/metal/MTLPaints.h line 41:

> 39: /**
> 40:  * The MTLPaint class represents paint mode (color, gradient, e.t.c.)
> 41:  * */

Minor: `* */` --> `*/`; also typo: `e.t.c.` should be `etc.`

src/java.desktop/macosx/native/libawt_lwawt/java2d/metal/MTLSurfaceDataBase.h line 37:

> 35: 
> 36: /**
> 37:  * The MTLSDOps structure describes a native OpenGL surface and contains all

Should that be "Metal" and not "OpenGL" ?

src/java.desktop/macosx/native/libawt_lwawt/java2d/metal/MTLTextRenderer.m line 77:

> 75:  * be safe to use this one glyph cache for all screens in a multimon
> 76:  * environment, since the glyph cache texture is shared between all contexts,
> 77:  * and (in theory) OpenGL drivers should be smart enough to manage that

`Metal` drivers?

src/java.desktop/macosx/native/libawt_lwawt/java2d/metal/MTLTexurePool.m line 29:

> 27: #import "Trace.h"
> 28: 
> 29: #define SCREEN_MEMORY_SIZE_4K (4096*2160*4) //~33,7 mb

This means that a 4k display with a narrower aspect ratio wouldn't fit (assuming there ever were to be such a thing). What would happen if you encountered a screen that was, say, 4k * 2.5K?

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

PR: https://git.openjdk.java.net/jdk/pull/2403