8341844: [lworld] Test crashing in genzgc in ZBarrier::mark_from_old_slow_path(zaddress)+0x4c #1321
Conversation
…slow_path(zaddress)+0x4c Generational zgc requires a store barrier in addition to the load barrier. load barrier followed by bulk copy isn't enough
|
👋 Welcome back dsimms! A progress list of the required criteria for merging this PR into |
|
@MrSimms 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: 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 2 new commits pushed to the
Please see this link for an up-to-date comparison between the source branch of this pull request and the ➡️ To integrate this PR with the above commit message to the |
Webrevs
|
|
/contributor add @fisk |
|
@MrSimms |
| // 1) possibly raw copy for any primitive payload before each map | ||
| // 2) load and store barrier for each oop | ||
| // 3) possibly raw copy for any primitive payload trailer | ||
| assert(lk == NON_ATOMIC_FLAT || lk == PAYLOAD, "Cannot support layout other than NON_ATOMIC_FLAT"); // Note: PAYLOAD is incorrect, resolve when transistioned to new flattening |
There was a problem hiding this comment.
Why is PAYLOAD incorrect here?
And what about ATOMIC_FLAT? Even with ZGC, it can contain an oop.
There was a problem hiding this comment.
You are correct there is a use single case under the md->contains_oops() condition with Z that is valid, a single oop (and only a single oop, no other data). Will adjust...
Given that we now have "LayoutKind", it would be useful to add an assert for the non-oops condition too, but that should probably go into RawAccessBarrier<decorators>::value_copy...I want fix that later (JDK-8346505) and get the merge work continued (could be baked into the "next gen" flattening work ?).
|
/integrate |
|
Going to push as commit 896cca5.
Your commit was automatically rebased without conflicts. |
Generational zgc requires a store barrier in addition to the load barrier.
load barrier followed by bulk copy isn't enough
Progress
Issue
Reviewers
Contributors
<eosterlund@openjdk.org>Reviewing
Using
gitCheckout this PR locally:
$ git fetch https://git.openjdk.org/valhalla.git pull/1321/head:pull/1321$ git checkout pull/1321Update a local copy of the PR:
$ git checkout pull/1321$ git pull https://git.openjdk.org/valhalla.git pull/1321/headUsing Skara CLI tools
Checkout this PR locally:
$ git pr checkout 1321View PR using the GUI difftool:
$ git pr show -t 1321Using diff file
Download this PR as a diff file:
https://git.openjdk.org/valhalla/pull/1321.diff
Using Webrev
Link to Webrev Comment