Sargun Dhillon

In regards to sampling cycles / instructions across the same interval, wouldn't putting them in the same perf event group solve that problem so the kernel schedules them together?

Dumb question, why don't we (KS) just store our save / state in a parallel part of local storage, rather than injecting it into the saved state from kittens?

So, it seems to be difficult to reproduce. I relaunched Chrome, and it stopped happening for a while. Then it started happening again...another relaunch later, it fixed itself. Looking at...

This addresses https://github.com/cloudflare/pingora/issues/230

I do not think the failure is related.

I feel like this is a common-ish rust pattern, where you have into_inner(), that then returns the original thing the object was built from. We'd probably want to flush, but...

Bump?

@miminar and @dmage can you make the assertion that the cross-mount functionality is no longer in use?

@dmage Specifically, the "stat" based mount middleware enablement. Not cross-mounting in general. Specifically, the functionality introduced in: https://github.com/distribution/distribution/pull/1857 It looks like that was removed from the openshift registry.

Some log traces: ``` 2024-10-05T01:05:15.792912Z DEBUG runtime.spawn{kind=task task.name= task.id=175 loc.file="third-party/rust/vendor/pingora-core-0.3.0/src/services/listening.rs" loc.line=169 loc.col=38}: pingora_proxy::proxy_h1: upstream event: Some(Body(Some(b"123\n"), false)) 2024-10-05T01:05:15.792962Z TRACE runtime.spawn{kind=task task.name= task.id=175 loc.file="third-party/rust/vendor/pingora-core-0.3.0/src/services/listening.rs" loc.line=169 loc.col=38}: pingora_core::protocols::http::v1::body: Writing Body, size: 4...