Repository navigation
image: udev's coldplug keeps cpu, so a core attached before systemd comes online - #258
Merged
Merged
Conversation
…omes online The coldplug filtered in 630fa2c replays block, net, virtio, misc, input, rtc, ptp and ttyS0, and not cpu. A core attached while a consumer's init is PID 1 has no udev to see it arrive: what onlines it is that replay reaching 40-spin-hotadd.rules. With v20261113.03 every spin workspace came up with one core whatever it was sized to (spin's kvm run 43527839765). task boot:hotadd attaches a core before systemd starts and reads the guest's online cores after: 0 on v20261113.03's image, 0-1 with this. image/build.sh refuses an image whose filter lost cpu, as it does input. Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
udev's coldplug (
10-coldplug.conf, filtered in 630fa2c) replayscpuagain.Why
spin boots every machine with one core and attaches the workspace's others while its own init is PID 1, before systemd and udev start. What onlines them is udev's coldplug replaying their
addto40-spin-hotadd.rules. The filter leftcpuout, so with v20261113.03 and v20261114.01 every workspace comes up with one core whatever it was sized to: spin's kvm run 43527839765 failed four tests with "still waiting for the guest to have its floor's two cores".How it is held
task boot:hotadd(TestACoreAttachedBeforeSystemdStartsComesOnline): an init that waits for a second core, the host attaches one over QMP, systemd starts, and a unit reads/sys/devices/system/cpu/online. On the image main builds today:present: 0-1 online: 0, failed. On this branch's image:present: 0-1 online: 0-1, passed. Both on a laptop under KVM.image/build.shrefuses an image whose filter has no--subsystem-match=cpu, as it does forinput.Not done
cpuin the replay costs the first login was not measured; it is one device per core. Worth oneboot:ssh-firstrun on the bench fleet against 630fa2c's 436 -> 402 ms.boot:hotaddis not in a workflow:compat.ymlboots the previous release's image, which would fail it until a release carries this.spin needs a release with this and a new pin.
🤖 Generated with Claude Code