8257228: G1: SIGFPE in G1ConcurrentRefine::create(int*) due to buffers_to_cards overflow #1489

DamonFool · 2020-11-28T08:26:57Z

Hi all,

SIGFPE was observed by running:

java -XX:G1ConcRefinementThresholdStep=16G -XX:G1UpdateBufferSize=1G -version

The reason is that buffers_to_cards [1] returns 0 for 'step' due to overflow.
It would be better to add overflow check logic is it.

Testing:

tier1 on Linux/x64

Thanks.
Best regards,
Jie

[1] https://github.com/openjdk/jdk/blob/master/src/hotspot/share/gc/g1/g1ConcurrentRefine.cpp#L235

Progress

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

Issue

JDK-8257228: G1: SIGFPE in G1ConcurrentRefine::create(int*) due to buffers_to_cards overflow

Reviewers

Kim Barrett (@kimbarrett - Reviewer)
Thomas Schatzl (@tschatzl - Reviewer) ⚠️ Review applies to 8bdeb20

Download

$ git fetch https://git.openjdk.java.net/jdk pull/1489/head:pull/1489
$ git checkout pull/1489

…s_to_cards overflow

DamonFool · 2020-11-28T08:27:25Z

/issue add JDK-8257228
/test
/label add hotspot-gc
/cc hotspot-gc

bridgekeeper · 2020-11-28T08:27:38Z

👋 Welcome back jiefu! 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 · 2020-11-28T08:28:46Z

@DamonFool This issue is referenced in the PR title - it will now be updated.

openjdk · 2020-11-28T08:28:54Z

@DamonFool
The hotspot-gc label was successfully added.

openjdk · 2020-11-28T08:29:25Z

@DamonFool The hotspot-gc label was already applied.

mlbridge · 2020-11-28T08:32:05Z

Webrevs

kimbarrett

This is only a problem when using values for some command line options that
are far out of the "normal" range. That's just a general problem; we have
far too many options, and some of them interact in interesting ways, so that
it's pretty much impossible to fully test or check for problem cases. And we
don't want to set artificially low limit values for individual options
because it's hard to know what some application might find useful. In this
case, it does look like we can reasonably do more checking though.

kimbarrett · 2020-11-29T07:43:29Z

src/hotspot/share/gc/g1/g1ConcurrentRefine.cpp

+  if (value == 0) return 0;
+  size_t res = value * G1UpdateBufferSize;
+
+  if (res < value || res < G1UpdateBufferSize) { // Check overflow


That's not a correct multiplication overflow check.

kimbarrett · 2020-11-29T07:54:20Z

src/hotspot/share/gc/g1/g1ConcurrentRefine.cpp

+  if (res < value || res < G1UpdateBufferSize) { // Check overflow
+    vm_exit_during_initialization("buffers_to_cards overflow!");
+  }
+  return res;


buffers_to_cards is only used with command line options as arguments. So
it's not inappropriate to use vm_exit_during_initialization, even though
this function isn't obviously about initialization. Maybe it should be
changed to

// Convert configuration values in units of buffers to number of cards.
static size_t configuration_buffers_to_cards(size_t value, const char* value_name) {
... use value_name in the error message
}

An alternative would be to move the checking to constraint functions for
the command line options. That puts the checking far from the code where
it matters though, but maybe a comment in the conversion function referring
to the option constraints would be sufficient. This might provide better
reporting?

DamonFool · 2020-11-30T06:56:53Z

/test

openjdk · 2020-11-30T06:57:18Z

@DamonFool you need to get approval to run the tests in tier1 for commits up until 76dec20

DamonFool · 2020-11-30T07:03:15Z

Hi @kimbarrett ,

Thanks for your review and comments.

The mul-overflow checking had been fixed.
The error msg had been refined.

For the sake of maintenance, I didn't use constraint function.

There are quite a few flags involved in it such as:

G1UpdateBufferSize
ParallelGCThreads
G1ConcRefinementThresholdStep
G1ConcRefinementGreenZone
G1ConcRefinementYellowZone
G1ConcRefinementRedZone

Maybe, you can add the constraint function for each of them this time.
But, when a new flag is used in the future, the constraint function may be missed.
Also, it seems a little strange to add this kind of constraints to flags like ParallelGCThreads.
So I prefer to fixing it in buffers_to_cards.

What do you think?

Thanks.
Best regards,
Jie

DamonFool · 2020-11-30T09:03:22Z

/test

openjdk · 2020-11-30T09:03:51Z

@DamonFool you need to get approval to run the tests in tier1 for commits up until e3ea5de

kimbarrett · 2020-11-30T09:42:49Z

test/hotspot/jtreg/gc/g1/TestBuffersToCardsOverflow.java

+
+public class TestBuffersToCardsOverflow {
+    public static void main(String... args) throws Exception {
+        ProcessTools.executeTestJava("-XX:G1ConcRefinementThresholdStep=16G",


It seems like this will fail to start on a 32bit platform, because the threshold step exceeds the possible range.

Good catch!
Fixed. Thanks.

kimbarrett · 2020-11-30T09:46:41Z

The mul-overflow checking had been fixed.
The error msg had been refined.

These updates look okay.

For the sake of maintenance, I didn't use constraint function.

There are quite a few flags involved in it such as:

[...]

Maybe, you can add the constraint function for each of them this time.
But, when a new flag is used in the future, the constraint function may be missed.
Also, it seems a little strange to add this kind of constraints to flags like ParallelGCThreads.
So I prefer to fixing it in buffers_to_cards.

The constraint function could be on G1UpdateBufferSize, checking each of
those other values for being not too large for its given value. But I agree
that a constraint function is far away, so the improved buffers_to_cards
does seem better.

tschatzl

Lgtm

openjdk · 2020-11-30T11:31:13Z

@DamonFool 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:

8257228: G1: SIGFPE in G1ConcurrentRefine::create(int*) due to buffers_to_cards overflow

Reviewed-by: kbarrett, tschatzl

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 1 new commit pushed to the master branch:

fe5cccc: 8254631: Better support ALPN byte wire values in SunJSSE

Please see this link for an up-to-date comparison between the source branch of this pull request and the master branch.
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.

tschatzl · 2020-11-30T13:25:27Z

(Just trying out what happens if I try to issue a test request)

/test

DamonFool · 2020-12-02T00:06:10Z

Lgtm

Thanks @tschatzl for your review.

@kimbarrett , are you OK with this change?
Thanks.

kimbarrett · 2020-12-02T03:31:53Z

src/hotspot/share/gc/g1/g1ConcurrentRefine.cpp

@@ -243,15 +252,17 @@ static size_t calc_min_yellow_zone_size() {

 static size_t calc_init_green_zone() {
  size_t green = G1ConcRefinementGreenZone;
+  char* name = (char*) "G1ConcRefinementGreenZone";


Change the type of name to const char* and eliminate the cast here and on line 258.

Amazing!
It gets compiled without the casts by just adding 'const'.

Updated. Thanks.

DamonFool · 2020-12-02T04:13:15Z

/test

openjdk · 2020-12-02T04:13:33Z

@DamonFool you need to get approval to run the tests in tier1 for commits up until efa0946

DamonFool · 2020-12-02T06:49:19Z

/integrate

openjdk · 2020-12-02T06:50:25Z

@DamonFool Since your change was applied there has been 1 commit pushed to the master branch:

fe5cccc: 8254631: Better support ALPN byte wire values in SunJSSE

Your commit was automatically rebased without conflicts.

Pushed as commit f2a0988.

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

8257228: G1: SIGFPE in G1ConcurrentRefine::create(int*) due to buffer…

6b37c51

…s_to_cards overflow

openjdk bot added the rfr Pull request is ready for review label Nov 28, 2020

openjdk bot added the hotspot-gc hotspot-gc-dev@openjdk.org label Nov 28, 2020

kimbarrett suggested changes Nov 29, 2020

View reviewed changes

DamonFool added 3 commits November 30, 2020 14:13

Fix mul-overflow-check and error reporting

cb7328b

Refine the erro msg

5a6730c

Merge branch 'master' into JDK-8257228

76dec20

openjdk bot added the test-request label Nov 30, 2020

Fix build error without PCH

e3ea5de

kimbarrett suggested changes Nov 30, 2020

View reviewed changes

Only run the test on 64-bit machines

8bdeb20

tschatzl approved these changes Nov 30, 2020

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Nov 30, 2020

kimbarrett suggested changes Dec 2, 2020

View reviewed changes

Merge branch 'master' into JDK-8257228

7ab8830

Eliminate the casts

efa0946

kimbarrett approved these changes Dec 2, 2020

View reviewed changes

openjdk bot closed this Dec 2, 2020

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 Dec 2, 2020

DamonFool deleted the JDK-8257228 branch December 2, 2020 06:52

8257228: G1: SIGFPE in G1ConcurrentRefine::create(int*) due to buffers_to_cards overflow #1489

8257228: G1: SIGFPE in G1ConcurrentRefine::create(int*) due to buffers_to_cards overflow #1489

DamonFool commented Nov 28, 2020 •

edited by openjdk bot

DamonFool commented Nov 28, 2020

bridgekeeper bot commented Nov 28, 2020

openjdk bot commented Nov 28, 2020

openjdk bot commented Nov 28, 2020

openjdk bot commented Nov 28, 2020

mlbridge bot commented Nov 28, 2020 •

edited

kimbarrett left a comment

kimbarrett Nov 29, 2020

kimbarrett Nov 29, 2020

DamonFool commented Nov 30, 2020

openjdk bot commented Nov 30, 2020

DamonFool commented Nov 30, 2020

DamonFool commented Nov 30, 2020

openjdk bot commented Nov 30, 2020

kimbarrett Nov 30, 2020

DamonFool Nov 30, 2020

kimbarrett commented Nov 30, 2020

tschatzl left a comment

openjdk bot commented Nov 30, 2020 •

edited

tschatzl commented Nov 30, 2020

DamonFool commented Dec 2, 2020

kimbarrett Dec 2, 2020

DamonFool Dec 2, 2020

DamonFool commented Dec 2, 2020

openjdk bot commented Dec 2, 2020

DamonFool commented Dec 2, 2020

openjdk bot commented Dec 2, 2020

8257228: G1: SIGFPE in G1ConcurrentRefine::create(int*) due to buffers_to_cards overflow #1489

8257228: G1: SIGFPE in G1ConcurrentRefine::create(int*) due to buffers_to_cards overflow #1489

Conversation

DamonFool commented Nov 28, 2020 • edited by openjdk bot

Progress

Issue

Reviewers

Download

DamonFool commented Nov 28, 2020

bridgekeeper bot commented Nov 28, 2020

openjdk bot commented Nov 28, 2020

openjdk bot commented Nov 28, 2020

openjdk bot commented Nov 28, 2020

mlbridge bot commented Nov 28, 2020 • edited

Webrevs

kimbarrett left a comment

Choose a reason for hiding this comment

kimbarrett Nov 29, 2020

Choose a reason for hiding this comment

kimbarrett Nov 29, 2020

Choose a reason for hiding this comment

DamonFool commented Nov 30, 2020

openjdk bot commented Nov 30, 2020

DamonFool commented Nov 30, 2020

DamonFool commented Nov 30, 2020

openjdk bot commented Nov 30, 2020

kimbarrett Nov 30, 2020

Choose a reason for hiding this comment

DamonFool Nov 30, 2020

Choose a reason for hiding this comment

kimbarrett commented Nov 30, 2020

tschatzl left a comment

Choose a reason for hiding this comment

openjdk bot commented Nov 30, 2020 • edited

tschatzl commented Nov 30, 2020

DamonFool commented Dec 2, 2020

kimbarrett Dec 2, 2020

Choose a reason for hiding this comment

DamonFool Dec 2, 2020

Choose a reason for hiding this comment

DamonFool commented Dec 2, 2020

openjdk bot commented Dec 2, 2020

DamonFool commented Dec 2, 2020

openjdk bot commented Dec 2, 2020

DamonFool commented Nov 28, 2020 •

edited by openjdk bot

mlbridge bot commented Nov 28, 2020 •

edited

openjdk bot commented Nov 30, 2020 •

edited