8263971: C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN #3190

rwestrel · 2021-03-25T09:28:15Z

For the inner for loop of the test case:

1- the control input of the DivI node in the loop body is cleared
because the range of values for i23, the loop's iv, is known to
never be null.

2- pre/main/post loops are added

3- the main loop is unrolled enough that it's fully unrolled but
actually never entered

4- because the main loop is no longer a looop, DiVI nodes from the
main loop can float and are scheduled in the pre loop.

5- the main loop is never executed (the values of the iv for the main
loop after unrolling fall outside the (3, 68] range of values of the
initial loop), one of the main loop DivI nodes divides by 0 and
because it doesn't stay in the main loop, the crash occurs.

The calls to PhaseIdealLoop::cast_incr_before_loop() in
PhaseIdealLoop::insert_pre_post_loops() should guarantee that nodes
that uses the loop iv stay control dependent on the test that guards
the main loop. For that it inserts a CastII node with the
_carry_dependency flag set. With 8256730, that CastII for the main
loop is pushed through the iv AddI of the pre loop. In the process, it
looses the _carry_dependency flag which causes the CastII node to be
optimized out and the DivI nodes to float.

With the proposed fix, CastII nodes created by the logic from 8256730
have the _carry_dependency flag set if the CastII node that's
transform has it.

Progress

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

Issue

JDK-8263971: C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN

Reviewers

Christian Hagedorn (@chhagedorn - Reviewer) ⚠️ Review applies to 1fef7af
Tobias Hartmann (@TobiHartmann - Reviewer) ⚠️ Review applies to 1fef7af

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.java.net/jdk pull/3190/head:pull/3190
$ git checkout pull/3190

Update a local copy of the PR:
$ git checkout pull/3190
$ git pull https://git.openjdk.java.net/jdk pull/3190/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 3190

View PR using the GUI difftool:
$ git pr show -t 3190

Using diff file

Download this PR as a diff file:
https://git.openjdk.java.net/jdk/pull/3190.diff

bridgekeeper · 2021-03-25T09:29:06Z

👋 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 · 2021-03-25T09:30:36Z

@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 · 2021-03-25T09:34:01Z

Webrevs

DamonFool · 2021-03-25T09:44:50Z

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

+ * @bug 8263971
+ * @summary C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN
+ *
+ * @run main/othervm -Xcomp -XX:-TieredCompilation -XX:CompileOnly=TestLostDependencyOnZeroTripGuard -XX:+StressGCM -XX:StressSeed=886771365 TestLostDependencyOnZeroTripGuard


-XX:+UnlockDiagnosticVMOptions -XX:+IgnoreUnrecognizedVMOptions should be added for the test, right?

-XX:+UnlockDiagnosticVMOptions should be enough.

Thanks for looking at this

-XX:+UnlockDiagnosticVMOptions -XX:+IgnoreUnrecognizedVMOptions should be added for the test, right?

I added -XX:+UnlockDiagnosticVMOptions so the test can be run with a release build.

I don't think -XX:+UnlockDiagnosticVMOptions is enough since StressGCM is a C2-only flag.

What will happen if there is no c2 compiler?
Thanks.

I don't think -XX:+UnlockDiagnosticVMOptions is enough since StressGCM is a C2-only flag.

What will happen if there is no c2 compiler?

Right. Command line option added.

I don't think -XX:+UnlockDiagnosticVMOptions is enough since StressGCM is a C2-only flag.
What will happen if there is no c2 compiler?

Right. Command line option added.

Thanks for your update.

What will happen if there is no c2 compiler?

Ah, yes, good point! I was only thinking about product builds.

There are actually more compiler tests that have -XX:+IgnoreUnrecognizedVMOptions missing as for example: Vec_mulAddS2I that uses C2-specific flags. I'll file a bug to clean that up separately.

chhagedorn

Otherwise, looks good to me!

chhagedorn · 2021-03-25T12:15:29Z

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

+ * @bug 8263971
+ * @summary C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN
+ *
+ * @run main/othervm -Xcomp -XX:-TieredCompilation -XX:CompileOnly=TestLostDependencyOnZeroTripGuard -XX:+StressGCM -XX:StressSeed=886771365 TestLostDependencyOnZeroTripGuard


-XX:+UnlockDiagnosticVMOptions should be enough.

openjdk · 2021-03-25T12:42:40Z

@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:

8263971: C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN

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 77 new commits pushed to the master branch:

bcdf469: 8264337: VM crashed when -XX:+VerifySharedSpaces
1a681fa: 8263560: Remove needless wrapping with BufferedInputStream
182b11c: 8264016: [JVMCI] add some thread local fields for use by JVMCI
8cf1c62: 8263754: HexFormat 'fromHex' methods should be static
a5d7de2: 8263404: RsaPrivateKeySpec is always recognized as RSAPrivateCrtKeySpec in RSAKeyFactory.engineGetKeySpec
128c0c9: 8248418: jpackage fails to extract main class and version from app module linked in external runtime
fd45694: 8264344: Outdated links in JavaComponentAccessibility.m
f17ea9e: 8262899: TestRedirectLinks fails
963f1fc: 8264309: JFR: Improve .jfc parser
364cce1: 8264332: Use the blessed modifier order in jdk.charsets
... and 67 more: https://git.openjdk.java.net/jdk/compare/bc91596ca108ed366c35027cb01417d93543a1e0...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 · 2021-03-25T12:43:22Z

@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:

8263971: C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN

Reviewed-by: chagedorn

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 25 new commits pushed to the master branch:

a1e717f: 8264146: Make Mutex point to rather than embed _name
f69afba: 8263300: add HtmlId for the block containing a class's description.
d82464f: 8263528: Make static page ids safe from collision with language elements
d602ae0: 8263884: Clean up os::is_allocatable() across Posix platforms
a9d287a: 8260388: Listing (sub)packages at package level of API documentation
8120064: 8263781: C2: Cannot hoist independent load above arraycopy
9689863: 8262295: C2: Out-of-Bounds Array Load from Clone Source
a678a38: 8263743: redundant lock in SSLSocketImpl
3fcb499: 8263768: JFormattedTextField.AbstractFormatter.getDocumentFilter()/getNavigationFilter() spec doesn't mention what the default impls return and what does it mean
4155533: 8258753: StartTlsResponse.close() hangs due to synchronization issues
... and 15 more: https://git.openjdk.java.net/jdk/compare/bc91596ca108ed366c35027cb01417d93543a1e0...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.

rwestrel · 2021-03-25T12:50:57Z

Otherwise, looks good to me!

Thanks for the review.

TobiHartmann

Looks good to me.

TobiHartmann · 2021-03-29T06:55:06Z

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

+ * @summary C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN
+ *
+ * @run main/othervm -Xcomp -XX:-TieredCompilation -XX:CompileOnly=TestLostDependencyOnZeroTripGuard -XX:+UnlockDiagnosticVMOptions
+ *                   -XX:+IgnoreUnrecognizedVMOptions -XX:+StressGCM -XX:StressSeed=886771365 TestLostDependencyOnZeroTripGuard


Is it worth adding another @run with no fixed seed? Just to give this a chance to still trigger the bug if the implementation changes in the future such that the fixed seed does not work anymore.

Thanks for the review. Would you drop -XX:+StressGCM as well or keep it?

I would keep it.

Done in updated change.

rwestrel · 2021-03-30T07:38:47Z

/integrate

openjdk · 2021-03-30T07:41:31Z

@rwestrel Since your change was applied there have been 77 commits pushed to the master branch:

bcdf469: 8264337: VM crashed when -XX:+VerifySharedSpaces
1a681fa: 8263560: Remove needless wrapping with BufferedInputStream
182b11c: 8264016: [JVMCI] add some thread local fields for use by JVMCI
8cf1c62: 8263754: HexFormat 'fromHex' methods should be static
a5d7de2: 8263404: RsaPrivateKeySpec is always recognized as RSAPrivateCrtKeySpec in RSAKeyFactory.engineGetKeySpec
128c0c9: 8248418: jpackage fails to extract main class and version from app module linked in external runtime
fd45694: 8264344: Outdated links in JavaComponentAccessibility.m
f17ea9e: 8262899: TestRedirectLinks fails
963f1fc: 8264309: JFR: Improve .jfc parser
364cce1: 8264332: Use the blessed modifier order in jdk.charsets
... and 67 more: https://git.openjdk.java.net/jdk/compare/bc91596ca108ed366c35027cb01417d93543a1e0...master

Your commit was automatically rebased without conflicts.

Pushed as commit 8100a20.

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

rwestrel added 2 commits March 24, 2021 17:03

fix

0760a28

test

9666e06

openjdk bot added the rfr Pull request is ready for review label Mar 25, 2021

openjdk bot added the hotspot-compiler hotspot-compiler-dev@openjdk.org label Mar 25, 2021

DamonFool reviewed Mar 25, 2021

View reviewed changes

chhagedorn approved these changes Mar 25, 2021

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Mar 25, 2021

test with -XX:+UnlockDiagnosticVMOptions

2697f04

chhagedorn approved these changes Mar 25, 2021

View reviewed changes

test with -XX:+IgnoreUnrecognizedVMOptions

1fef7af

chhagedorn approved these changes Mar 25, 2021

View reviewed changes

TobiHartmann approved these changes Mar 29, 2021

View reviewed changes

2nd run with no seed

f6e736b

openjdk bot closed this Mar 30, 2021

openjdk bot added integrated Pull request has been integrated and removed ready Pull request is ready to be integrated rfr Pull request is ready for review labels Mar 30, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

8263971: C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN #3190

8263971: C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN #3190

rwestrel commented Mar 25, 2021 •

edited by openjdk bot

Loading

bridgekeeper bot commented Mar 25, 2021

openjdk bot commented Mar 25, 2021

mlbridge bot commented Mar 25, 2021 •

edited

Loading

DamonFool Mar 25, 2021

chhagedorn Mar 25, 2021

rwestrel Mar 25, 2021 •

edited

Loading

DamonFool Mar 25, 2021

rwestrel Mar 25, 2021

DamonFool Mar 25, 2021

chhagedorn Mar 25, 2021

chhagedorn left a comment

chhagedorn Mar 25, 2021

openjdk bot commented Mar 25, 2021 •

edited

Loading

openjdk bot commented Mar 25, 2021

rwestrel commented Mar 25, 2021

TobiHartmann left a comment

TobiHartmann Mar 29, 2021

rwestrel Mar 29, 2021

TobiHartmann Mar 30, 2021

rwestrel Mar 30, 2021

rwestrel commented Mar 30, 2021

openjdk bot commented Mar 30, 2021

8263971: C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN #3190

8263971: C2 crashes with SIGFPE with -XX:+StressGCM and -XX:+StressIGVN #3190

Conversation

rwestrel commented Mar 25, 2021 • edited by openjdk bot Loading

Progress

Issue

Reviewers

Reviewing

bridgekeeper bot commented Mar 25, 2021

openjdk bot commented Mar 25, 2021

mlbridge bot commented Mar 25, 2021 • edited Loading

Webrevs

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rwestrel Mar 25, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

chhagedorn left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

openjdk bot commented Mar 25, 2021 • edited Loading

openjdk bot commented Mar 25, 2021

rwestrel commented Mar 25, 2021

TobiHartmann left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rwestrel commented Mar 30, 2021

openjdk bot commented Mar 30, 2021

rwestrel commented Mar 25, 2021 •

edited by openjdk bot

Loading

mlbridge bot commented Mar 25, 2021 •

edited

Loading

rwestrel Mar 25, 2021 •

edited

Loading

openjdk bot commented Mar 25, 2021 •

edited

Loading