Interop: annotation processor to generate classes for message resolution #12

grimmerm · 2016-01-25T11:08:25Z

There is a need to simplify interop API to:

make it more type-safe
encourage exception handling
make its use less verbose

chumer · 2016-01-25T12:11:48Z

Please don't create additional branches in the main repo. (feature/instrumentation_api is an exception).
Instead create a fork and create a pull-request from your fork. Give people access to your fork for those you want to collaborate with.

chumer · 2016-01-25T12:12:08Z

Please delete this branch after you have closed the pull-request.

jtulach · 2016-01-25T12:21:58Z

I did an API signature check and following generated classes are public. I think they should be package private:

Added Classes
-------------

com.oracle.truffle.tck.ComplexNumberAEntryForeign
com.oracle.truffle.tck.ComplexNumberBEntryForeign
com.oracle.truffle.tck.ComplexNumberForeign
com.oracle.truffle.tck.ComplexNumbersAForeign
com.oracle.truffle.tck.ComplexNumbersBForeign

jtulach · 2016-01-25T12:30:13Z

truffle/com.oracle.truffle.api.interop.java/snapshot.sigtest

@@ -16,6 +16,8 @@ CLSS public abstract interface !annotation com.oracle.truffle.api.interop.java.M
 intf java.lang.annotation.Annotation
 meth public abstract java.lang.String message()

+CLSS abstract interface com.oracle.truffle.api.interop.java.package-info
+


Don't change the signature files, please.

package-info should never be part of a signature file, or should it?

chumer · 2016-01-25T12:32:55Z

@jtulach shouldn't be the API signature check be part of the gate? Or is it and it was just not run? We have the ability to spawn arbitrary stuff in parallel in travis, we should probably do that for the api signature check.

Re. (by @jtulach): yes, we can run it but how it shall report the "failure"?

jtulach · 2016-01-25T12:33:07Z

truffle/com.oracle.truffle.api.interop/src/com/oracle/truffle/api/interop/ForeignAccess.java

-     * @throws IllegalAccessError if the <code>receiver</code> does not support the
-     *             {@link Message#createNode() message represented} by <code>foreignNode</code>
+     * @throws IllegalStateException if any error occurred while accessing the <code>receiver</code>
+     *             object
     */


Didn't you want to deprecate this method?

woess · 2016-01-25T16:00:50Z

I would like to see some examples of the code generated by this annotation processor rather than having to infer it from the source code.

grimmerm · 2016-01-26T15:15:36Z

Unless there are objections, let this be merged on Wed.

jtulach · 2016-01-27T09:43:03Z

truffle/com.oracle.truffle.api.interop/src/com/oracle/truffle/api/interop/ForeignAccess.java

+     * @throws UnsupportedMessageException if the <code>receiver</code> does not support the
+     *             {@link Message#createNode() message represented} by <code>executeNode</code>
+     */
+    public static Object sendExecute(Node executeNode, VirtualFrame frame, TruffleObject receiver, Object[] arguments) throws UnsupportedTypeException, ArityException, UnsupportedMessageException {


Using varargs for the arguments argument might be more comfortable for the caller.

jtulach · 2016-01-27T10:38:20Z

I like the change and I am about to merge it once we solve problems with Javadoc.

…lasses-for-message-resolution

…enerate-classes-for-message-resolution AcceptMessage annotation. Its annotation processor. InteropException and its subclasses. Specialized sendXYZ methods in ForeignAccess class.

woess · 2016-01-27T13:22:17Z

For the record, I didn't review this change (not enough time).

@cached

* commit '3ca08c7bb8628665580317807776648d5504d6db': Add failing test for DSL processor bug with negated guard + @cached.

This patch adds the following match rules to generate bitfield move instruction on AArch64: * (RightShift (LeftShift value a) b) -> SBFX/SBFIZ * (UnsignedRightShift (LeftShift value a) b) -> UBFX/UBFIZ * (LeftShift (SignExtend value) a) -> SBFIZ Eg: lsl w0, w1, oracle#8 asr w0, w0, oracle#15 is optimized to: sbfx w0, w1, oracle#7, oracle#17 It also adds the rules to integrate the ZeroExtend with unsigned bitfield move operation. Eg: ubfiz w0, w1, oracle#5, oracle#12 and x0, x0, #0xffffffff is optimized to: ubfiz x0, x1, oracle#5, oracle#12 Change-Id: I2b635d4895db0d5f4c30630176f336e9a226ccaf

…inter. Trapping nullcheck might generate two uncompress instructions for the same compressed oop in Graal. One is inserted by the backend when it emits nullcheck. If the pointer is a compressed object, it should be uncompressed before the nullcheck is emitted. And another one is generated by the normal uncompressing operation. These two instructions are duplicated with each other. The generated codes on AArch64 like: ldr w0, [x0,oracle#112] lsl x2, x0, oracle#3 ; uncompressing (first) ldr xzr, [x2] ; implicit exception: deoptimizes ...... ; fixed operations lsl x0, x0, oracle#3 ; uncompressing (second) str w1, [x0,oracle#12] A simple way to avoid this is to apply the nullcheck to the uncompressed result if it exists instead of to the compressed pointer when generating the trapping nullcheck. With the modification, the codes above could be optimized to: ldr w0, [x0,oracle#112] lsl x0, x0, oracle#3 ; uncompressing ldr xzr, [x0] ; implicit exception: deoptimizes ...... ; fixed operations str w1, [x0,oracle#12] Change-Id: Iabfe47bbf984ed11c42555f84bdd0ccf2a5bdddb

Add support for killing compiler threads

Trapping nullcheck might generate two uncompress instructions for the same compressed oop on AArch64. One is inserted by the backend when it emits nullcheck. If the object is a compressed pointer, it is uncompressed before the nullcheck is emitted. And another one is generated by the uncompression node used for memory access. These two instructions are duplicated with each other. The generated codes on AArch64 like: ldr w0, [x0,oracle#112] lsl x2, x0, oracle#3 ; uncompressing (first) ldr xzr, [x2] ; implicit exception: deoptimizes ...... ; fixed operations lsl x0, x0, oracle#3 ; uncompressing (second) str w1, [x0,oracle#12] A simple way to avoid this is to creat a new uncompression node for the nullcheck, and let the value numbering remove the duplicated one if possible. Since the address lowering of AMD64 can handle the uncompressing computation for address, the created uncompression node is wrapped to an address node and the nullcheck is finally applied on the address. With the modification, the codes above could be optimized to: ldr w0, [x0,oracle#112] lsl x0, x0, oracle#3 ; uncompressing ldr xzr, [x0] ; implicit exception: deoptimizes ...... ; fixed operations str w1, [x0,oracle#12] Change-Id: Iabfe47bbf984ed11c42555f84bdd0ccf2a5bdddb

Trapping nullcheck might generate two uncompress instructions for the same compressed oop on AArch64. One is inserted by the backend when it emits nullcheck. If the object is a compressed pointer, it is uncompressed before the nullcheck is emitted. And another one is generated by the uncompression node used for memory access. These two instructions are duplicated with each other. The generated codes on AArch64 like: ldr w0, [x0,oracle#112] lsl x2, x0, oracle#3 ; uncompressing (first) ldr xzr, [x2] ; implicit exception: deoptimizes ...... ; fixed operations lsl x0, x0, oracle#3 ; uncompressing (second) str w1, [x0,oracle#12] A simple way to avoid this is to creat a new uncompression node for the nullcheck, and let the value numbering remove the duplicated one if possible. Since the address lowering of AMD64 can handle the uncompressing computation for address, the created uncompression node is wrapped to an address node and the nullcheck is finally applied on the address. With the modification, the codes above could be optimized to: ldr w0, [x0,oracle#112] lsl x0, x0, oracle#3 ; uncompressing ldr xzr, [x0] ; implicit exception: deoptimizes ...... ; fixed operations str w1, [x0,oracle#12]

Matthias Grimmer added 3 commits January 25, 2016 11:02

Add AcceptMessage annotation to generate classes for message resolution

022643e

Fix checkstyle errors

5a773f1

Eclipseformat

e465fe0

Generated ForeignAccess classes should be package private

67013f8

jtulach reviewed Jan 25, 2016
View reviewed changes

Matthias Grimmer added 3 commits January 25, 2016 15:13

Use InteropExceptions instead of IllegalArgumentExceptions

5862bc4

Revert signature file changes

463b9c0

Fix SLTckTests

c3905e8

Matthias Grimmer added 2 commits January 26, 2016 12:09

Add Javadoc to AcceptMessage and add a simple example

45d8a67

Fix checkstyle warnings and eclipseformat

06086c6

Fix line-length issues for codesnippets

a1450f4

jtulach reviewed Jan 27, 2016
View reviewed changes

Use varargs arguments for ForeignAccess#new/invoke/execute

682d35a

jtulach mentioned this pull request Jan 27, 2016

Switching to new version of the codesnippet doclet to prevent broken … graalvm/mx#12

Merged

Improve javadoc of ForeignAccess#send methods

33df340

jtulach self-assigned this Jan 27, 2016

Merge branch 'master' into interop/annotation-processor-to-generate-c…

69b2258

…lasses-for-message-resolution

jtulach merged commit bce21f2 into master Jan 27, 2016

jtulach deleted the interop/annotation-processor-to-generate-classes-for-message-resolution branch January 27, 2016 10:50

thomaswue added the oca-signed label Feb 1, 2016

dougxc pushed a commit that referenced this pull request Jul 11, 2016

Merge pull request #12 in G/truffle from eregon-dsl-bug to master

ed80f7f

* commit '3ca08c7bb8628665580317807776648d5504d6db': Add failing test for DSL processor bug with negated guard + @cached.

brunoborges unassigned jtulach Jan 8, 2018

armLeiJin mentioned this pull request Jul 17, 2019

Store store barrier at the end of initializer of an object without final fields #1504

Closed

kipply added a commit to Shopify/graal that referenced this pull request Apr 24, 2020

Merge pull request oracle#12 from Shopify/jit-pause-resume

9e6d306

Add support for killing compiler threads

jbiancot mentioned this pull request Dec 4, 2020

Pi4 2GB of ram, Ubuntu 20.04 - Image generator watchdog is aborting image generation. To configure the watchdog #3043

Closed

dragosroban mentioned this pull request Mar 10, 2021

No instances of sun.nio.ch.Iocp are allowed in the image heap as this class should be initialized at image runtime. #3248

Closed

JozGonzalez mentioned this pull request Aug 21, 2023

[Snyk] Upgrade minimist from 1.2.6 to 1.2.8 JozGonzalez/graal#4

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Interop: annotation processor to generate classes for message resolution #12

Interop: annotation processor to generate classes for message resolution #12

grimmerm commented Jan 25, 2016

chumer commented Jan 25, 2016

chumer commented Jan 25, 2016

jtulach commented Jan 25, 2016

jtulach Jan 25, 2016

woess Jan 25, 2016

chumer commented Jan 25, 2016

jtulach Jan 25, 2016

woess commented Jan 25, 2016

grimmerm commented Jan 26, 2016

jtulach Jan 27, 2016

jtulach commented Jan 27, 2016

woess commented Jan 27, 2016

Interop: annotation processor to generate classes for message resolution #12

Interop: annotation processor to generate classes for message resolution #12

Conversation

grimmerm commented Jan 25, 2016

chumer commented Jan 25, 2016

chumer commented Jan 25, 2016

jtulach commented Jan 25, 2016

jtulach Jan 25, 2016

Choose a reason for hiding this comment

woess Jan 25, 2016

Choose a reason for hiding this comment

chumer commented Jan 25, 2016

jtulach Jan 25, 2016

Choose a reason for hiding this comment

woess commented Jan 25, 2016

grimmerm commented Jan 26, 2016

jtulach Jan 27, 2016

Choose a reason for hiding this comment

jtulach commented Jan 27, 2016

woess commented Jan 27, 2016