Re: RFR: 8303134: JFR: Missing stack trace during chunk rotation stress [v8]
Greetings,
please help review this fix for some problematic situations in JFR where data can be lost. Most problems originate from writing event data in the wrong epoch due to safepointing. Detailed information about the changes is in the JIRA issue.
Testing: jdk_jfr, stress testing.
Thanks Markus
Markus Grönlund has updated the pull request incrementally with one additional commit since the last revision: build x86_32 ------------- Changes: - all: https://git.openjdk.org/jdk/pull/14761/files - new: https://git.openjdk.org/jdk/pull/14761/files/08210ac9..97854c2e Webrevs: - full: https://webrevs.openjdk.org/?repo=jdk&pr=14761&range=07 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=14761&range=06-07 Stats: 1 line in 1 file changed: 0 ins; 0 del; 1 mod Patch: https://git.openjdk.org/jdk/pull/14761.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/14761/head:pull/14761 PR: https://git.openjdk.org/jdk/pull/14761
On Tue, 11 Jul 2023 19:20:32 GMT, Markus Grönlund <mgronlun@openjdk.org> wrote:
Greetings,
please help review this fix for some problematic situations in JFR where data can be lost. Most problems originate from writing event data in the wrong epoch due to safepointing. Detailed information about the changes is in the JIRA issue.
Testing: jdk_jfr, stress testing.
Thanks Markus
Markus Grönlund has updated the pull request incrementally with one additional commit since the last revision:
build x86_32
The C2 changes look good to me. I guess this needs some testing on platforms not supported by Oracle. src/hotspot/share/opto/library_call.cpp line 3024:
3022: 3023: // Load the current value of the notified field in the JfrThreadLocal. 3024: Node* notified_offset = basic_plus_adr(top(), tls_ptr, in_bytes(NOTIFY_OFFSET_JFR));
Suggestion: Node* notified_offset = basic_plus_adr(top(), tls_ptr, in_bytes(NOTIFY_OFFSET_JFR)); ------------- Marked as reviewed by thartmann (Reviewer). PR Review: https://git.openjdk.org/jdk/pull/14761#pullrequestreview-1527596166 PR Review Comment: https://git.openjdk.org/jdk/pull/14761#discussion_r1262003063
On Tue, 11 Jul 2023 19:20:32 GMT, Markus Grönlund <mgronlun@openjdk.org> wrote:
Greetings,
please help review this fix for some problematic situations in JFR where data can be lost. Most problems originate from writing event data in the wrong epoch due to safepointing. Detailed information about the changes is in the JIRA issue.
Testing: jdk_jfr, stress testing.
Thanks Markus
Markus Grönlund has updated the pull request incrementally with one additional commit since the last revision:
build x86_32
FYI: Performed `jdk_jfr` test on my 64-core linux-riscv64 server, result is clean. Hope that helps. ------------- PR Comment: https://git.openjdk.org/jdk/pull/14761#issuecomment-1633670208
On Thu, 13 Jul 2023 06:56:06 GMT, Fei Yang <fyang@openjdk.org> wrote:
FYI: Performed `jdk_jfr` test on my 64-core linux-riscv64 server, result is clean. Hope that helps.
Thanks Fei, it helps! ------------- PR Comment: https://git.openjdk.org/jdk/pull/14761#issuecomment-1634654935
participants (3)
-
Fei Yang
-
Markus Grönlund
-
Tobias Hartmann