Repository navigation
Enable runsc core tagging when supported by the kernel - #2887
Benjamin Elder (BenTheElder) merged 1 commit into
Conversation
7f51b14 to
472a374
Compare
Benjamin Elder (BenTheElder)
left a comment
There was a problem hiding this comment.
thanks, including some vetted review findings ... this crashes on my local dev environment
472a374 to
92de95a
Compare
|
This is consistently causing e2e failures, looks like the little CI runners are having a timing issue with this in |
|
Tagging networking folks to see if someone has bandwidth to help out with that. |
|
Fabricio Voznika (@fvoznika) Can you please make sure that the PR pass the DCO check? I can't move it back to the merge queue |
|
DCO check needs fixing (sorry, we just completed that transition, |
Head branch was pushed to by a user without write access
92de95a to
cb200af
Compare
|
DCO still failing, not sure why :/ |
|
The test PR merged so this should be otherwise fine |
|
I think you might need to rebase on upstream/main. It's confusing because github shows this PR as one commit, but the DCO check is seeing many commits |
cb200af to
6118881
Compare
|
Benjamin Elder (@BenTheElder) it should be good to go now. |
6118881 to
f158354
Compare
|
Benjamin Elder (@BenTheElder) had to resolve a few conflicts...it's ready to go. |
Probe prctl(PR_SCHED_CORE_GET) once and pass --enable-core-tags to runsc when core scheduling is available on the host kernel to isolate gVisor sandboxes across SMT siblings. Signed-off-by: Fabricio Voznika <fvoznika@google.com>
f158354 to
3399ba7
Compare
|
This is the top of the merge queue and should go in finally. |
Probe prctl(PR_SCHED_CORE_GET) once and pass --enable-core-tags to runsc when core scheduling is available on the host kernel to isolate gVisor sandboxes across SMT siblings.
Fixes #2136
AI disclosure: AI tooling assisted in writing and testing parts of this change.
Signed-off-by: Fabricio Voznika fvoznika@google.com