Michael Schubert

iproute2 retries up to 10 times with a larger buffer: https://git.kernel.org/pub/scm/linux/kernel/git/shemminger/iproute2.git/tree/lib/bpf.c?h=v4.12.0#n1319

I pushed a first tag `v0.1.0` earlier today (bcc release `v0.18.0`) and plan to push a new tag for every new bcc release. Hopefully that will make tracking a bit...

> Currently when we log in, we pick up a single coreroller team (based on the github organization and team) ... Does this break with multiple teams? (re https://github.com/kinvolk/coreroller/pull/16#issuecomment-429826148)

We now explain that in the stage1-builder README: https://github.com/kinvolk/stage1-builder#using-custom-built-stage1-kvm-images-on-semaphore-ci

It requires another clang pragma; see https://github.com/weaveworks/tcptracer-bpf/pull/43/commits/c3113a625af78739f8e117b3b1198b937cace795 which is part of #43

> so #43 is yet to be merged? Yes.

> Every time I start tcptracer-bpf, it misses 28 kretprobes on tcp_v4_connect. How did you tests this? I don't see "every time I start tcptracer-bpf". I do see miss-hits, but...

> as far as I see kprobe_profile doesn't get reset I does get reset. The high number of hits then stems from the offset guess phase.

Sorry, let me try to rephrase: accessing **response** header values (such as `Content-Range`) is currently not possible through a go-swagger generated client. The **request** header parameter definitions are in the...

I see this as well. macOS Big Sur `11.4` Anki `2.1.44`.