@danmcd - If your memory goes back far enough to OS-5004, I would appreciate your review on illumos 15465 (fenix) which is similar/related.

OS-5004: SIOCGLIFADDR doesn't set sin6_scope_id for link-local IPv6 addresses (Resolved)

↳ smartos.org/bugview/OS-5004 | joyent/illumos-joyent a92bddf

fenix illumos 15465

BUG 15465: SIOCGLIFCONF doesn't set sin6_scope_id for link-local IPv6 addresses (In Progress)

↳ illumos.org/issues/15465 | code.illumos.org/c/illumos-gate/+/2700

Well, even if it does not go back that far :)

I am suspecting that the test-runner that (I think) MNX provides for libuv acquired a LL IPv6 address since the last libuv release, as the test should have always failed.

WTF didn't I upstream this back then?

It was upstreamed, but that was SIOCLIFADDR and this is SIOCLIFCONF

Oh shit.

ahh OS-5004 aka fenix illumos 6680

OS-5004: SIOCGLIFADDR doesn't set sin6_scope_id for link-local IPv6 addresses (Resolved)

↳ smartos.org/bugview/OS-5004 | joyent/illumos-joyent a92bddf

fenix illumos 6680

BUG 6680: SIOCGLIFADDR doesn't set sin6_scope_id for link-local IPv6 addresses (Closed)

↳ illumos.org/issues/6680

Okay, gotcha.

RTI it now if you haven't. I'm sure someone not-me will approve.

I'm checking with a few people here first, but will do, thanks!

pmooney: "For AVIC is dark and full of terrors"

big disappointment, IMO

APICv made a big perf difference on intel

AVIC seemed poised to deliver something similar, including a few facets of it which I thought were an improvement on the INTC design

but between the warning errata, and the fact that the inter-vCPU interrupt tables never seemed to grow support for x2APIC

the value prop definitely fell off some

maybe they'll get their act together

Zen-next (Zen5? Haven't followed roadmaps much since buying Zen2-powered HDC3 hardware.

)

This was mostly a problem in Zen 3, IIRC. I believe Zen 4 (Genoa) does have more there, but I'll have to dig a bit. Been a while.

[illumos-gate] 15456 Remove SPARC man pages from section 5 -- Peter Tribble <peter.tribble⊙gc>

[illumos-gate] 15458 Remove SPARC man pages from section 8 -- Peter Tribble <peter.tribble⊙gc>

illumos.org/opensolaris/bugdb/bug.html#!6425953 was apparantly filed by sommerfeld for the same issue as illumos.org/issues/15465 nearly 15 years ago

→ OpenSolaris issue 6425953: SIOCGLIFCONF doesn't fill in sin6_scope_id for link-local addresses (Accepted)

sorry, looked at wrong line, make that 17 years ago

Crikey, thanks alanc

It was also spotted as illumos 15319 (fenix), although that was a lot more recently!

BUG 15319: Should getifaddrs() return scope IDs for link local IPv6 addresses? (Duplicate)

↳ illumos.org/issues/15319

alanc: blast from the past!

likely filed when I was hip-deep in getting iked to deal with dynamically added addresses or some such.

I can't remember if you were working at all in in.ikev2d before <CENSORED> <CENSORED> us? Or if it was actual in.iked. (So long ago...)

I definitely did a bit of ikev2 implementation work

shortly before pulling the ejection lever

I haven't looked at the code but I bet illumos.org/opensolaris/bugdb/bug.html#!6425939 isn't fixed either.

→ OpenSolaris issue 6425939: PF_ROUTE doesn't fill in sin6_scope_id in rts_fill_msg_v6 (Accepted)