yeah, and if people want to run propolis, a brand for omnios for that would make more sense

yeah, there's nothing in propolis which depends on oxide-specific bits

how can I inject NMI from UTM (qemu)?

`nmi 0` from the qemu query monitor, if you can get to it

hm. I see...

with serial port -> target, one can map monitor.

C-a c gets you to the monitor

whether the target supports an NMI varies

[illumos-gate] 15921 ptools should not be isaexec'd -- Richard Lowe <richlowe⊙rn>

nice.

I have a branch doing more, but let's let that settle first...

do you have ipf in that branch too?

yes

I have literally every consumer of isaexec except in.iked which we can't fix

am, ok

there is "14128 ipf: only build 64-bit commands" - and branch I apparently have forgot to bind with gerrit:D

well, feel free to go before me :)

or look at the thing I just sent you and take the ipf bits if they seem better

ok

I remember it was unpleasant, so it's likely they don't seem better

I was wondering what the plan for in.iked was, as we don't do anything with the closed bits other than unpack them

in.iked can be repackaged to be 64-bit only.

But it's closed so no love/hope for ARM there.

danmcd: we would have to find a way to move it

which isn't worth the effort, really

because as you say, it doesn't actually solve any real problem.

the easiest we could do is hardlink it

(to the real binary, rather than isaexec)

Yeah, hardlink would suffice if the goal is, "kill isaexec".

the goal is "deliver things in a way that makes sense going forward"

so either way

I have a branch in flight that rationalizes some other closed-bin things too, but I'm not sure if that's for illumos-gate yet.

The IKE certutils are stuck at 32-bit, but they honestly shouldn't need to be 32-bit NOR closed (i.e. they could be rewritten to use openssl and/or pktool).