8349772: C2 hits MemLimit during incremental inlining#30555
Draft
dafedafe wants to merge 1 commit intoopenjdk:masterfrom
Draft
8349772: C2 hits MemLimit during incremental inlining#30555dafedafe wants to merge 1 commit intoopenjdk:masterfrom
dafedafe wants to merge 1 commit intoopenjdk:masterfrom
Conversation
|
👋 Welcome back dfenacci! A progress list of the required criteria for merging this PR into |
|
❗ This change is not yet ready to be integrated. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The issue was very intermittent but an old run that reproduced (for the crash with
-XX:-TieredCompilation -XX:+AlwaysIncrementalInline) suggested the following was happening:The test builds a very long MH adapter chain (in some runs up to ~10,000 transformations). Many small MH adapters are late-inlined, and C2 performs repeated incremental-inline/prune/IGVN rounds.
In each round, parser/IR/support data is allocated from compile arenas. Even when pruning removes dead nodes (live node count seems to remain moderate, e.g. ~13k), arena usage appears to be reclaimed only when the compilation ends.
So, for runs with long/complex chains, peak arena usage seems to cross MemLimit, which triggers a crash in debug (or a compilation bailout in product/non-crash mode).
Disabling/stopping late inlining at that point would probably only postpone the resource failure to a later phase.
JDK-8349820 disabled the explicit test-level MemLimit override. This change keeps a safeguard by using a larger explicit limit (2G) in debug builds.
Progress
Issue
Reviewing
Using
gitCheckout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/30555/head:pull/30555$ git checkout pull/30555Update a local copy of the PR:
$ git checkout pull/30555$ git pull https://git.openjdk.org/jdk.git pull/30555/headUsing Skara CLI tools
Checkout this PR locally:
$ git pr checkout 30555View PR using the GUI difftool:
$ git pr show -t 30555Using diff file
Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/30555.diff