New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

8262837: handle split_USE correctly #2791

Closed

kuaiwei wants to merge 3 commits into openjdk:master from kuaiwei:reg_split

Contributor

kuaiwei commented Mar 2, 2021 •

edited by openjdk bot

Loading

I look into reg_split.cpp and found a potential issue.
Function split_USE usually creates a new spill copy node, but in some cases it just connect def to use and no new node created. But in caller side, they will assume it will create a new node. The code is like:

maxlrg = split_USE(...)
if (!maxlrg) {
   return 0;
}
insidx++;  // Reset iterator to skip USE side split

So if no node is created, the iterator index is updated and the next instruction will be skipped.

The change is let split_USE return the new node count, so the caller can update its index and maxlrg.

Progress

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

Issue

JDK-8262837: handle split_USE correctly

Reviewers

Vladimir Kozlov (@vnkozlov - Reviewer)

Download

$ git fetch https://git.openjdk.java.net/jdk pull/2791/head:pull/2791
$ git checkout pull/2791


          8262837: handle split_USE correctly

4d2e315

bridgekeeper bot commented Mar 2, 2021

👋 Welcome back kuaiwei! 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 bot added the rfr label

openjdk bot commented Mar 2, 2021

@kuaiwei The following label will be automatically applied to this pull request:

hotspot-compiler

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

Contributor Author

kuaiwei commented Mar 2, 2021

/label add hotspot-compiler

openjdk bot added the hotspot-compiler label

openjdk bot commented Mar 2, 2021

@kuaiwei The hotspot-compiler label was already applied.

mlbridge bot commented Mar 2, 2021 •

edited

Loading

Webrevs

vnkozlov suggested changes

View reviewed changes

src/hotspot/share/opto/reg_split.cpp Outdated Show resolved Hide resolved

src/hotspot/share/opto/reg_split.cpp Outdated

                 }
                 Node *spill = get_spillcopy_wide(spill_type, def, use, useidx );
-                if( !spill ) return 0;        // Bailed out
+                if( !spill ) return -1;        // Bailed out
                 // Insert SpillCopy before the USE, which uses the reaching DEF as
                 // its input, and defs a new live range, which is used by this node.
                 insert_proj( b, bindex, spill, maxlrg++ );

Contributor

vnkozlov Mar 3, 2021

maxlrg++ again

Contributor Author

kuaiwei Mar 3, 2021

Remove post-increment

src/hotspot/share/opto/reg_split.cpp

		@@ -1099,12 +1105,13 @@ uint PhaseChaitin::Split(uint maxlrg, ResourceArea* split_arena) {
		Node spill = new MachSpillCopyNode(MachSpillCopyNode::MemToReg, def, dmask, tmp_rm);
		insert_proj( b, insidx, spill, maxlrg );

Contributor

vnkozlov Mar 3, 2021

I think it missed maxlrg increment. May be also use insidx++ here so in the following code you need to increment only by delta.

Contributor Author

kuaiwei Mar 3, 2021

added maxlrg++ and insidx++, the following insidx is added with delta

src/hotspot/share/opto/reg_split.cpp Outdated

                 uint                 bidx, pidx, slidx, insidx, inpidx, twoidx;
                 uint                 non_phi = 1, spill_cnt = 0;
+                int                  delta;

Contributor

vnkozlov Mar 3, 2021

I think it should be local variable where it is used.

Contributor Author

kuaiwei Mar 3, 2021 •

edited

Loading

changed to local variable.

src/hotspot/share/opto/reg_split.cpp Outdated

                 // Insert SpillCopy before the USE, which uses the reaching DEF as
                 // its input, and defs a new live range, which is used by this node.
                 insert_proj( b, bindex, spill, maxlrg++ );
                 // Use the spill/clone
                 use->set_req(useidx,spill);
-                // return updated live range count
-                return maxlrg;
+                // return generated node count

Contributor

vnkozlov Mar 3, 2021

The comment is confusing. I would suggest to add comment at the header of this method to describe all returned values [-1, 0, 1].

Contributor Author

kuaiwei Mar 3, 2021

Add comment in method header of split_USE and remove this comment.

kuaiwei added 2 commits

March 3, 2021 16:04


          8262837: handle split_USE correctly

6ef827c


          8262837: handle split_USE correctly

1fa77b7

vnkozlov approved these changes

View reviewed changes

Contributor

vnkozlov left a comment

Good.
Passed tier1-4 testing.

openjdk bot commented Mar 3, 2021 •

edited

Loading

@kuaiwei 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:

8262837: handle split_USE correctly

Reviewed-by: kvn

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 33 new commits pushed to the master branch:

84c93d5: 8257137: Revise smov and umov in aarch64 assembler
d93fa0d: 8253940: com/sun/jdi/JdwpAttachTest.java failed with "RuntimeException: ERROR: LingeredApp.startApp was able to attach"
104a262: 8224775: test/jdk/com/sun/jdi/JdwpListenTest.java failed to attach
2848938: 8262927: Explicitly state fields examined for BigDecimal.hashCode
b397472: 8262915: java.awt.color.ColorSpace.getName() is not thread-safe
268d9b7: 8261447: MethodInvocationCounters frequently run into overflow
75aa154: 8259267: Refactor LoaderLeak shell test as java test.
a118185: 8261862: Expand discussion of rationale for BigDecimal equals/compareTo semantics
2d2ef08: 8262885: Shenandoah: FullGC prologue does not need to save/restore heap has_forwarded_object flag
1d2c1e6: 8248314: Parallel: Parallelize parallel full gc Adjust Roots phase
... and 23 more: https://git.openjdk.java.net/jdk/compare/dd33a8eaaab1e0a9bc816d08daedbfbb7f0d6c6a...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 (@vnkozlov) 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).

openjdk bot added the ready label

Contributor Author

kuaiwei commented Mar 4, 2021

Good.
Passed tier1-4 testing.

Thanks for your help.

Contributor Author

kuaiwei commented Mar 4, 2021

/integrate

openjdk bot added the sponsor label

openjdk bot commented Mar 4, 2021

@kuaiwei
Your change (at version 1fa77b7) is now ready to be sponsored by a Committer.

Contributor

vnkozlov commented Mar 4, 2021

/sponsor

openjdk bot closed this

openjdk bot added integrated and removed sponsor ready rfr labels

openjdk bot commented Mar 4, 2021

@vnkozlov @kuaiwei Since your change was applied there have been 44 commits pushed to the master branch:

bd1a806: 8263040: fix for JDK-8262122 fails validate-source
a6427c8: 8259709: Disable SHA-1 XML Signatures
ef5e13d: 8263030: Remove Shenandoah leftovers from ReferenceProcessor
222a17e: 8262122: [TESTBUG] Shenandoah-specific variant of TestReferenceRefersTo
a777e82: 8254717: isAssignableFrom checks in KeyFactorySpi.engineGetKeySpec appear to be backwards
d2c4ed0: 8262000: jdk/jfr/event/gc/detailed/TestPromotionFailedEventWithParallelScavenge.java failed with "OutOfMemoryError: Java heap space"
94f26e4: 8261450: JShell crashes with SIOOBE in tab completion
b817855: 8262935: Add missing logging to sun.net.httpserver.ServerImpl
2251319: 8262828: Format of OS information is different on macOS
4cfecce: 8261730: C2 compilation fails with assert(store->find_edge(load) != -1) failed: missing precedence edge
... and 34 more: https://git.openjdk.java.net/jdk/compare/dd33a8eaaab1e0a9bc816d08daedbfbb7f0d6c6a...master

Your commit was automatically rebased without conflicts.

Pushed as commit f56c918.

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

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

hotspot-compiler integrated