8321278: C2: Partial peeling fails with assert "last_peel <- first_not_peeled" #18353

rwestrel · 2024-03-18T17:15:38Z

The assert fails because peeling happens at a single entry
Region. That Region only has a single input because other inputs
were found unreachable and removed by
PhaseIdealLoop::Dominators(). The fix I propose is to have
PhaseIdealLoop::Dominators() remove the Region and its Phis
entirely in this case.

Progress

Change must be properly reviewed (1 review required, with at least 1 Reviewer)
Change must not contain extraneous whitespace
Commit message must refer to an issue

Issue

JDK-8321278: C2: Partial peeling fails with assert "last_peel <- first_not_peeled" (Bug - P3)

Reviewers

Christian Hagedorn (@chhagedorn - Reviewer)
Tobias Hartmann (@TobiHartmann - Reviewer) ⚠️ Review applies to d201aee2

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/18353/head:pull/18353
$ git checkout pull/18353

Update a local copy of the PR:
$ git checkout pull/18353
$ git pull https://git.openjdk.org/jdk.git pull/18353/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 18353

View PR using the GUI difftool:
$ git pr show -t 18353

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/18353.diff

Webrev

Link to Webrev Comment

bridgekeeper · 2024-03-18T17:16:08Z

👋 Welcome back roland! 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 · 2024-03-18T17:16:26Z

@rwestrel 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:

8321278: C2: Partial peeling fails with assert "last_peel <- first_not_peeled"

Reviewed-by: chagedorn, thartmann

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 281 new commits pushed to the master branch:

142c311: 8328744: Parallel: Parallel GC throws OOM before heap is fully expanded
cb2a671: 8324121: SIGFPE in PhaseIdealLoop::extract_long_range_checks
0c1b254: 8326438: C2: assert(ld->in(1)->Opcode() == Op_LoadN) failed: Assumption invalid: input to DecodeN is not LoadN
29ba4b7: 8328705: GHA: Cross-compilation jobs do not require build JDK
f67ec19: 8079786: [macosx] Test java/awt/Frame/DisposeParentGC/DisposeParentGC.java fails for Mac only
93579c2: 8323552: AbstractMemorySegmentImpl#mismatch returns -1 when comparing distinct areas of the same instance of MemorySegment
b235682: 8328709: AIX os::get_summary_cpu_info support Power 10
19a0151: 8328700: Unused import and variable should be deleted in regex package
cf9b5a7: 8328776: [AIX] remove checked_vmgetinfo, use vmgetinfo directly
acc4a82: 8328862: Remove unused GrowableArrayFilterIterator
... and 271 more: https://git.openjdk.org/jdk/compare/784f11c35d7f8646c7354c756ac8801a3d685874...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.

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

openjdk · 2024-03-18T17:16:55Z

@rwestrel 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.

mlbridge · 2024-03-18T17:19:36Z

Webrevs

test/hotspot/jtreg/compiler/loopopts/TestPartialPeelingAtSingleInputRegion.java

chhagedorn · 2024-03-22T16:01:46Z

test/hotspot/jtreg/compiler/loopopts/TestPartialPeelingAtSingleInputRegion.java

+    }
+
+    public static void main(String[] args) {
+        for (int i = 0; i < 50_000; ++i) {


Just a minor thing: Do you really need 50000 iterations or would fewer be sufficient to trigger the bug?

Good catch. It reproduces with 10000. I pushed a comment that makes that change.

chhagedorn · 2024-03-22T16:06:00Z

src/hotspot/share/opto/domgraph.cpp

+      Node* dom = tdom->_control;
+      if (dom != C->root() && dom->is_Region() && dom->req() == 2) {
+        remove_single_entry_region(t, tdom, dom, _igvn);
+      }
+      _idom[t->_control->_idx] = dom; // Set immediate dominator


Removing the regions during Dominators() seems reasonable. I guess doing a pass of IGVN after Dominators() is probably too much to get these regions removed?

Could you also remove the region and the phi where the unreachable loops are cleaned up and the region and phis become single-entry nodes? I.e. here:

jdk/src/hotspot/share/opto/domgraph.cpp

Lines 453 to 463 in ce7ebaa

// Kill dead input path

assert( !visited.test(whead->in(i)->_idx),

"input with no loop must be dead" );

_igvn.delete_input_of(whead, i);

for (DUIterator_Fast jmax, j = whead->fast_outs(jmax); j < jmax; j++) {

Node* p = whead->fast_out(j);

if( p->is_Phi() ) {

_igvn.delete_input_of(p, i);

}

}

i--; // Rerun same iteration

Thanks for reviewing this.

Removing the regions during Dominators() seems reasonable. I guess doing a pass of IGVN after Dominators() is probably too much to get these regions removed?

It does feel like a lot of overhead for such a simple corner case.

Could you also remove the region and the phi where the unreachable loops are cleaned up and the region and phis become single-entry nodes? I.e. here:

I considered it but it felt harder. The algorithm collects cfg nodes in dfs and then iterate over them several times. If we remove a region at the point you mention, it would need to be removed from the dfs node list. Or we would need to make later iterations over the dfs node list handle dead region nodes. It's not a problem when it's done later as I propose here.

Thanks for reviewing this.

Removing the regions during Dominators() seems reasonable. I guess doing a pass of IGVN after Dominators() is probably too much to get these regions removed?

It does feel like a lot of overhead for such a simple corner case.

Indeed, I don't think it's worth.

Could you also remove the region and the phi where the unreachable loops are cleaned up and the region and phis become single-entry nodes? I.e. here:

I considered it but it felt harder. The algorithm collects cfg nodes in dfs and then iterate over them several times. If we remove a region at the point you mention, it would need to be removed from the dfs node list. Or we would need to make later iterations over the dfs node list handle dead region nodes. It's not a problem when it's done later as I propose here.

I see, thanks for the explanation. Then it makes sense to handle this edge-case like you proposed to keep things simple. Maybe you can add a comment accordingly why we remove the region at this point.

I see, thanks for the explanation. Then it makes sense to handle this edge-case like you proposed to keep things simple. Maybe you can add a comment accordingly why we remove the region at this point.

Sure. I added a comment. Let me know if it looks ok to you.

…eInputRegion.java Co-authored-by: Christian Hagedorn <christian.hagedorn@oracle.com>

TobiHartmann

Looks good to me too.

chhagedorn

Looks good, thanks for the update!

rwestrel · 2024-03-25T13:44:15Z

@chhagedorn @TobiHartmann thanks for the reviews!

rwestrel · 2024-03-25T13:44:20Z

/integrate

openjdk · 2024-03-25T13:45:18Z

Going to push as commit af15c68.
Since your change was applied there have been 281 commits pushed to the master branch:

142c311: 8328744: Parallel: Parallel GC throws OOM before heap is fully expanded
cb2a671: 8324121: SIGFPE in PhaseIdealLoop::extract_long_range_checks
0c1b254: 8326438: C2: assert(ld->in(1)->Opcode() == Op_LoadN) failed: Assumption invalid: input to DecodeN is not LoadN
29ba4b7: 8328705: GHA: Cross-compilation jobs do not require build JDK
f67ec19: 8079786: [macosx] Test java/awt/Frame/DisposeParentGC/DisposeParentGC.java fails for Mac only
93579c2: 8323552: AbstractMemorySegmentImpl#mismatch returns -1 when comparing distinct areas of the same instance of MemorySegment
b235682: 8328709: AIX os::get_summary_cpu_info support Power 10
19a0151: 8328700: Unused import and variable should be deleted in regex package
cf9b5a7: 8328776: [AIX] remove checked_vmgetinfo, use vmgetinfo directly
acc4a82: 8328862: Remove unused GrowableArrayFilterIterator
... and 271 more: https://git.openjdk.org/jdk/compare/784f11c35d7f8646c7354c756ac8801a3d685874...master

Your commit was automatically rebased without conflicts.

openjdk · 2024-03-25T13:45:24Z

@rwestrel Pushed as commit af15c68.

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

fix & test

5cbab30

openjdk bot added the rfr Pull request is ready for review label Mar 18, 2024

openjdk bot added the hotspot-compiler hotspot-compiler-dev@openjdk.org label Mar 18, 2024

chhagedorn reviewed Mar 22, 2024

View reviewed changes

rwestrel and others added 2 commits March 22, 2024 17:35

Update test/hotspot/jtreg/compiler/loopopts/TestPartialPeelingAtSingl…

bb2cb9e

…eInputRegion.java Co-authored-by: Christian Hagedorn <christian.hagedorn@oracle.com>

review

d201aee

chhagedorn approved these changes Mar 25, 2024

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Mar 25, 2024

TobiHartmann approved these changes Mar 25, 2024

View reviewed changes

comment

8fbb6f9

chhagedorn approved these changes Mar 25, 2024

View reviewed changes

openjdk bot added the integrated Pull request has been integrated label Mar 25, 2024

openjdk bot closed this Mar 25, 2024

openjdk bot removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Mar 25, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

8321278: C2: Partial peeling fails with assert "last_peel <- first_not_peeled" #18353

8321278: C2: Partial peeling fails with assert "last_peel <- first_not_peeled" #18353

rwestrel commented Mar 18, 2024 •

edited by openjdk bot

Loading

bridgekeeper bot commented Mar 18, 2024

openjdk bot commented Mar 18, 2024 •

edited

Loading

openjdk bot commented Mar 18, 2024

mlbridge bot commented Mar 18, 2024 •

edited

Loading

chhagedorn Mar 22, 2024

rwestrel Mar 22, 2024

chhagedorn Mar 22, 2024

rwestrel Mar 22, 2024

chhagedorn Mar 25, 2024 •

edited

Loading

rwestrel Mar 25, 2024

TobiHartmann left a comment

chhagedorn left a comment

rwestrel commented Mar 25, 2024

rwestrel commented Mar 25, 2024

openjdk bot commented Mar 25, 2024

openjdk bot commented Mar 25, 2024

	// Kill dead input path
	assert( !visited.test(whead->in(i)->_idx),
	"input with no loop must be dead" );
	_igvn.delete_input_of(whead, i);
	for (DUIterator_Fast jmax, j = whead->fast_outs(jmax); j < jmax; j++) {
	Node* p = whead->fast_out(j);
	if( p->is_Phi() ) {
	_igvn.delete_input_of(p, i);
	}
	}
	i--; // Rerun same iteration

8321278: C2: Partial peeling fails with assert "last_peel <- first_not_peeled" #18353

8321278: C2: Partial peeling fails with assert "last_peel <- first_not_peeled" #18353

Conversation

rwestrel commented Mar 18, 2024 • edited by openjdk bot Loading

Progress

Issue

Reviewers

Reviewing

Webrev

bridgekeeper bot commented Mar 18, 2024

openjdk bot commented Mar 18, 2024 • edited Loading

openjdk bot commented Mar 18, 2024

mlbridge bot commented Mar 18, 2024 • edited Loading

Webrevs

chhagedorn Mar 22, 2024

Choose a reason for hiding this comment

rwestrel Mar 22, 2024

Choose a reason for hiding this comment

chhagedorn Mar 22, 2024

Choose a reason for hiding this comment

rwestrel Mar 22, 2024

Choose a reason for hiding this comment

chhagedorn Mar 25, 2024 • edited Loading

Choose a reason for hiding this comment

rwestrel Mar 25, 2024

Choose a reason for hiding this comment

TobiHartmann left a comment

Choose a reason for hiding this comment

chhagedorn left a comment

Choose a reason for hiding this comment

rwestrel commented Mar 25, 2024

rwestrel commented Mar 25, 2024

openjdk bot commented Mar 25, 2024

openjdk bot commented Mar 25, 2024

rwestrel commented Mar 18, 2024 •

edited by openjdk bot

Loading

openjdk bot commented Mar 18, 2024 •

edited

Loading

mlbridge bot commented Mar 18, 2024 •

edited

Loading

chhagedorn Mar 25, 2024 •

edited

Loading