8262188: Add test to verify trace page sizes logging on Linux #2696

kstefanj · 2021-02-23T21:23:18Z

Please review this new test that checks the output from -Xlog:pagesize and compares it to what's really used according to /proc/self/smaps.

We already have a few tests that check the output from -Xlog:pagesize, the difference is that those tests expect some page size given the parameters used. This new test instead check if the values printed out are correct with regards to what is recorded in /proc/self/maps.

I wrote the test to see if all our logging was correct, and found those two issues that are now fixed:
JDK-8261029: Code heap page sizes not traced correctly using os::trace_page_sizes
JDK-8261230: GC tracing of page sizes are wrong in a few places

The test now passes and it will be a good way to prevent incorrect page size logging to be added going forward.

Testing
The new test has been run through mach5 without any problems and also manually on configurations with explicit large pages enabled.

Progress

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

Issue

JDK-8262188: Add test to verify trace page sizes logging on Linux

Reviewers

Thomas Stuefe (@tstuefe - Reviewer) ⚠️ Review applies to b55bb7f
Thomas Schatzl (@tschatzl - Reviewer) ⚠️ Review applies to d869aed

Download

$ git fetch https://git.openjdk.java.net/jdk pull/2696/head:pull/2696
$ git checkout pull/2696

bridgekeeper · 2021-02-23T21:23:35Z

👋 Welcome back sjohanss! 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 · 2021-02-23T21:24:47Z

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

hotspot-runtime

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 · 2021-02-23T21:46:56Z

Webrevs

tstuefe

Hi Stefan,

looks good. Minor points below.

Cheers, Thomas

tstuefe · 2021-02-24T10:16:19Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+ * @requires os.family == "linux"
+ * @requires os.arch=="amd64" | os.arch=="x86_64"
+ * @requires vm.gc != "Z"
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:+UseLargePages -XX:LargePageSizeInBytes=1g TestTracePageSizes


Should we also test with -XX:LargePageSizeInBytes=2M?

Good point, 2m might not always be the default so I guess it makes sense to explicitly test both.

tstuefe · 2021-02-24T10:17:26Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:-SegmentedCodeCache TestTracePageSizes
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:-SegmentedCodeCache -XX:+UseLargePages TestTracePageSizes
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:-SegmentedCodeCache -XX:+UseTransparentHugePages TestTracePageSizes
+ */


When do you use multiple @run, when multiple @test?

I might be using this wrong, but the way I use it is that for each @test I can have a set of @requires and then I do a set of runs under these requirements.

Sometimes specifying explicit tests for the different GCs are considered to expensive but in this case I think it is fine.

tstuefe · 2021-02-24T10:20:07Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:+UseSerialGC TestTracePageSizes
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:+UseSerialGC -XX:+UseLargePages TestTracePageSizes
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:+UseSerialGC -XX:+UseTransparentHugePages TestTracePageSizes
+*/


I see you don't explicitly test UseSHM. Because the number of tests would be too large?

I don't explicitly test HugeTLBFS either, but yes that is what's being tested when just setting +UseLargePages =) Would you prefer if I had one @run for -XX:+UseHugeTLBFS and one for -XX:+UseSHM.

I think this is fine for now; hopefully we'll get rid of UseSHM.

tstuefe · 2021-02-24T10:26:23Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+    // will generate a match.
+    private static void parseSmaps() throws Exception {
+        Pattern smapsPattern = Pattern.compile(RangeWithPageSize.smapsPatternString, Pattern.DOTALL);
+        Scanner smapsScanner = new Scanner(new File("/proc/self/smaps"));


Would that work if mappings are concurrently modified? (I guess it has to otherwise one could not read reliably from /proc at all)

I can't say for sure, not sure exactly how the file is consumed by the Scanner. The segments and the fields in those segments that we really care about should not change and therefore I think we should be fine.

I looked at https://www.kernel.org/doc/Documentation/filesystems/proc.txt

Note: reading /proc/PID/maps or /proc/PID/smaps is inherently racy (consistent output can be achieved only in the single read call). This typically manifests when doing partial reads of these files while the memory map is being modified. Despite the races, we do provide the following guarantees: 1) The mapped addresses never go backwards, which implies no two regions will ever overlap. 2) If there is something at a given vaddr during the entirety of the life of the smaps/maps walk, there will be some output for it.

So, if Scanner reads the file in one go its okay.

I'd say lets see how this goes then. If this test fails intermittently, we know where to look.

I did a very basic test in jshell and created a Scanner on a file that I then updated between calls to nextLine(). The output from the scanner did not include my additions, so at least for a small file is seems like the contents is read in one go. But that doesn't say that is always the case.

Bu I agree, we know where to look if we see intermittent failures.

tstuefe · 2021-02-24T10:27:12Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+    // transparent huge pages. The pattern is not greedy and will
+    // match as little as possible so each "segment" in the file
+    // will generate a match.
+    private static void parseSmaps() throws Exception {


Code may be easier to read if the pattern were defined up here.

I agree, I moved it around and ended up leaving it below, but now with the comment it certainly make sense to have it here.

tstuefe · 2021-02-24T10:45:44Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+        Pattern traceLinePattern = Pattern.compile(traceLinePatternString);
+
+        // The test needs to be run with page size logging printed to ps-$pid.log.
+        Scanner fileScanner = new Scanner(new File("./ps-" + ProcessHandle.current().pid() + ".log"));


So we scan the log file while it is being written by ourselves? This seems slightly unsafe. Are we sure it has been completely flushed at this point? I'm mainly concerned about things like running this test in a current dir on NFS or somesuch.

An alternative would be to let the testee write the log file and print out a copy of /proc/self/smaps to stdout, then to scan and compare in a separate process.

But if you think this is safe enough, I'm fine with this too.

Not really while it is being written, since all the output from -Xlog:pagesize happens before the java test is executed and from what I can tell the log-file should be properly flushed.

As you say doing the scan and compare in a separate process afterwards would make any concurrent modification issue and flush issue go away. But it would also slightly complicate the test and make it depend on the test library. For debugging it is very nice to be able to run the test standalone and that is a bit harder when depending on the test lib.

Until proven unsafe I would prefer to leave it as is.

I'd say so too. If that pattern works, I may use it in other places too.

kstefanj

Thanks Thomas for looking at this and highlighting some potential problems.

I think we should be fine, but I will see if I can trigger a failure caused by any of your concerns :)

kstefanj · 2021-02-24T11:25:44Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:-SegmentedCodeCache TestTracePageSizes
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:-SegmentedCodeCache -XX:+UseLargePages TestTracePageSizes
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:-SegmentedCodeCache -XX:+UseTransparentHugePages TestTracePageSizes
+ */


I might be using this wrong, but the way I use it is that for each @test I can have a set of @requires and then I do a set of runs under these requirements.

Sometimes specifying explicit tests for the different GCs are considered to expensive but in this case I think it is fine.

kstefanj · 2021-02-24T11:41:49Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:+UseSerialGC TestTracePageSizes
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:+UseSerialGC -XX:+UseLargePages TestTracePageSizes
+ * @run main/othervm -XX:+AlwaysPreTouch -Xlog:pagesize:ps-%p.log -XX:+UseSerialGC -XX:+UseTransparentHugePages TestTracePageSizes
+*/


I don't explicitly test HugeTLBFS either, but yes that is what's being tested when just setting +UseLargePages =) Would you prefer if I had one @run for -XX:+UseHugeTLBFS and one for -XX:+UseSHM.

kstefanj · 2021-02-24T11:43:26Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+    // transparent huge pages. The pattern is not greedy and will
+    // match as little as possible so each "segment" in the file
+    // will generate a match.
+    private static void parseSmaps() throws Exception {


I agree, I moved it around and ended up leaving it below, but now with the comment it certainly make sense to have it here.

kstefanj · 2021-02-24T12:08:36Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+    // will generate a match.
+    private static void parseSmaps() throws Exception {
+        Pattern smapsPattern = Pattern.compile(RangeWithPageSize.smapsPatternString, Pattern.DOTALL);
+        Scanner smapsScanner = new Scanner(new File("/proc/self/smaps"));


I can't say for sure, not sure exactly how the file is consumed by the Scanner. The segments and the fields in those segments that we really care about should not change and therefore I think we should be fine.

kstefanj · 2021-02-24T12:29:02Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+        Pattern traceLinePattern = Pattern.compile(traceLinePatternString);
+
+        // The test needs to be run with page size logging printed to ps-$pid.log.
+        Scanner fileScanner = new Scanner(new File("./ps-" + ProcessHandle.current().pid() + ".log"));


Not really while it is being written, since all the output from -Xlog:pagesize happens before the java test is executed and from what I can tell the log-file should be properly flushed.

As you say doing the scan and compare in a separate process afterwards would make any concurrent modification issue and flush issue go away. But it would also slightly complicate the test and make it depend on the test library. For debugging it is very nice to be able to run the test standalone and that is a bit harder when depending on the test lib.

Until proven unsafe I would prefer to leave it as is.

tstuefe

I'm fine with this test. It will come in useful.

I leave it up to you to check the reading of proc/self/smaps more robust. It would be okay for me to use this test as it is and if it starts failing intermittently to modify it.

openjdk · 2021-02-24T12:44:00Z

@kstefanj 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:

8262188: Add test to verify trace page sizes logging on Linux

Reviewed-by: stuefe, 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 64 new commits pushed to the master branch:

702ca62: 8262185: G1: Prune collection set candidates early
8bc8542: 8262195: Harden tests that use the HostsFileNameService (jdk.net.hosts.file property)
20c93b3: 8261914: IfNode::fold_compares_helper faces non-canonicalized bool when running JRuby JSON workload
ddd550a: 8261308: C2: assert(inner->is_valid_counted_loop(T_INT) && inner->is_strip_mined()) failed: OuterStripMinedLoop should have been removed
03d888f: 8261804: Remove field _processing_is_mt, calculate it instead
6800ba4: 8257500: Drawing MultiResolutionImage with ImageObserver "leaks" memory
65a245e: 8262329: Fix JFR parser exception messages
a4c2496: 8259535: ECDSA SignatureValue do not always have the specified length
2515c42: 8262332: serviceability/sa/ClhsdbJhisto.java fails with Test ERROR java.lang.RuntimeException: 'ParselTongue' missing from stdout/stderr
07061fc: 8256417: Exclude TestJFRWithJMX test from running with PodMan
... and 54 more: https://git.openjdk.java.net/jdk/compare/29c72631ee3f4f1f88a36f25b933642ea9659681...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.

Updated with an additional run with setting page size = 2m and moved the regexp-string to where it is used.

tschatzl

I think this is good. There are two typos in there, please fix before pushing.

tschatzl · 2021-03-01T10:44:55Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+        Scanner fileScanner = new Scanner(new File("./ps-" + ProcessHandle.current().pid() + ".log"));
+        while (fileScanner.hasNextLine()) {
+            String line = fileScanner.nextLine();
+            if (line.matches(traceLinePatternString)){


Missing space before the brace

tschatzl · 2021-03-01T10:45:44Z

test/hotspot/jtreg/runtime/os/TestTracePageSizes.java

+
+    public RangeWithPageSize(String start, String end, String pageSize, String vmFlags) {
+        this.start = Long.parseUnsignedLong(start, 16);
+        this.end = Long.parseUnsignedLong(end, 16);;


two ";" at the end

kstefanj · 2021-03-01T11:09:44Z

Thanks for the reviews Thomas and Thomas =)

kstefanj · 2021-03-01T12:03:07Z

/integrate

openjdk · 2021-03-01T12:04:09Z

@kstefanj Since your change was applied there have been 65 commits pushed to the master branch:

0a7fff4: 8261636: The test mapping in hugetlbfs_sanity_check should consider LargePageSizeInBytes
702ca62: 8262185: G1: Prune collection set candidates early
8bc8542: 8262195: Harden tests that use the HostsFileNameService (jdk.net.hosts.file property)
20c93b3: 8261914: IfNode::fold_compares_helper faces non-canonicalized bool when running JRuby JSON workload
ddd550a: 8261308: C2: assert(inner->is_valid_counted_loop(T_INT) && inner->is_strip_mined()) failed: OuterStripMinedLoop should have been removed
03d888f: 8261804: Remove field _processing_is_mt, calculate it instead
6800ba4: 8257500: Drawing MultiResolutionImage with ImageObserver "leaks" memory
65a245e: 8262329: Fix JFR parser exception messages
a4c2496: 8259535: ECDSA SignatureValue do not always have the specified length
2515c42: 8262332: serviceability/sa/ClhsdbJhisto.java fails with Test ERROR java.lang.RuntimeException: 'ParselTongue' missing from stdout/stderr
... and 55 more: https://git.openjdk.java.net/jdk/compare/29c72631ee3f4f1f88a36f25b933642ea9659681...master

Your commit was automatically rebased without conflicts.

Pushed as commit fe8e370.

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

8262188: Add test to verify trace page sizes logging on Linux

544ea5e

openjdk bot added the hotspot-runtime hotspot-runtime-dev@openjdk.org label Feb 23, 2021

8262188: Self review

b55bb7f

kstefanj marked this pull request as ready for review February 23, 2021 21:43

openjdk bot added the rfr Pull request is ready for review label Feb 23, 2021

tstuefe reviewed Feb 24, 2021

View reviewed changes

kstefanj commented Feb 24, 2021

View reviewed changes

tstuefe approved these changes Feb 24, 2021

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Feb 24, 2021

8262188: Thomas review

d869aed

Updated with an additional run with setting page size = 2m and moved the regexp-string to where it is used.

tschatzl approved these changes Mar 1, 2021

View reviewed changes

8262188: Schatzl review

c45bbcb

openjdk bot closed this Mar 1, 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 1, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

8262188: Add test to verify trace page sizes logging on Linux #2696

8262188: Add test to verify trace page sizes logging on Linux #2696

kstefanj commented Feb 23, 2021 •

edited by openjdk bot

Loading

bridgekeeper bot commented Feb 23, 2021

openjdk bot commented Feb 23, 2021

mlbridge bot commented Feb 23, 2021 •

edited

Loading

tstuefe left a comment

tstuefe Feb 24, 2021

kstefanj Feb 24, 2021

tstuefe Feb 24, 2021

kstefanj Feb 24, 2021

tstuefe Feb 24, 2021

kstefanj Feb 24, 2021

tstuefe Feb 24, 2021

tstuefe Feb 24, 2021

kstefanj Feb 24, 2021

tstuefe Feb 24, 2021

kstefanj Feb 24, 2021

tstuefe Feb 24, 2021

kstefanj Feb 24, 2021

tstuefe Feb 24, 2021

kstefanj Feb 24, 2021

tstuefe Feb 24, 2021

kstefanj left a comment

kstefanj Feb 24, 2021

kstefanj Feb 24, 2021

kstefanj Feb 24, 2021

kstefanj Feb 24, 2021

kstefanj Feb 24, 2021

tstuefe left a comment

openjdk bot commented Feb 24, 2021 •

edited

Loading

tschatzl left a comment

tschatzl Mar 1, 2021

tschatzl Mar 1, 2021

kstefanj commented Mar 1, 2021

kstefanj commented Mar 1, 2021

openjdk bot commented Mar 1, 2021

8262188: Add test to verify trace page sizes logging on Linux #2696

8262188: Add test to verify trace page sizes logging on Linux #2696

Conversation

kstefanj commented Feb 23, 2021 • edited by openjdk bot Loading

Progress

Issue

Reviewers

Download

bridgekeeper bot commented Feb 23, 2021

openjdk bot commented Feb 23, 2021

mlbridge bot commented Feb 23, 2021 • edited Loading

Webrevs

tstuefe 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

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

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

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kstefanj 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

Choose a reason for hiding this comment

tstuefe left a comment

Choose a reason for hiding this comment

openjdk bot commented Feb 24, 2021 • edited Loading

tschatzl left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kstefanj commented Mar 1, 2021

kstefanj commented Mar 1, 2021

openjdk bot commented Mar 1, 2021

kstefanj commented Feb 23, 2021 •

edited by openjdk bot

Loading

mlbridge bot commented Feb 23, 2021 •

edited

Loading

openjdk bot commented Feb 24, 2021 •

edited

Loading