chore: pin celeris main (ea3d528) in the refapps, debugvars and the bench server - #312
Merged
Merged
Conversation
…ench server
Bumps the ten modules that pin celeris from the v1.5.11 release tag to a
pseudo-version of celeris main, so the nightly and soak validate the code
that is actually a release candidate rather than the last shipped tag.
celeris main is green at ea3d528 and carries four io_uring fixes landed
today that the v1.5.11 pin does not have:
#521 the process-global H2 stream worker pool starved once
GOMAXPROCS*4 streaming handlers were live
#522 completeSend deadlocked the worker against itself on cs.detachMu,
and SEND_ZC ENOMEM was treated as fatal, killing healthy conns
#523 a dropped H2 eventfd POLL_ADD was swallowed, leaving the eventfd
permanently deaf
#525 the SEND_ZC first-CQE branch was keyed on the mutable w.sendZC
instead of the CQE's own F_MORE flag, stranding connections at
cs.sending == true where the dirty-list flush skips them; plus
r.pending is now reconciled against the kernel's submitted count
A pseudo-version rather than a tag on purpose: nothing is published, and
the pin reverts to a release tag when one is cut.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Bumps the ten modules that pin celeris from the v1.5.11 release tag to a pseudo-version of celeris main, so the nightly and the 24h soak validate the release candidate rather than the last shipped tag.
celeris main is green at
ea3d528and carries four io_uring fixes landed today thatv1.5.11does not have:GOMAXPROCS*4streaming handlers were live (2/8 handlers started on a 2-worker pool, 8/8 after)completeSenddeadlocked the worker against itself oncs.detachMu; SEND_ZCENOMEMwas treated as fatal, killing 16 healthy conns per runPOLL_ADDwas swallowed whileh2PollArmedis never cleared, leaving the eventfd permanently deafw.sendZCinstead of the CQE's ownF_MORE, stranding conns atcs.sending == truewhere the dirty-list flush skips them; plusr.pendingreconciled against the kernel's submitted countAll ten modules build. A pseudo-version rather than a tag on purpose: nothing is published, and the pin goes back to a release tag when one is cut.
Why now: celeris main went green with #525, so this is the first point where a nightly would exercise the fixed engine. Known-still-open against this pin: goceleris/celeris#484 (inbound truncation) and the goceleris/celeris#519 residue.