as a side note for 15091 -- has anyone tried it on recent versions of go?

at least on 1.26.3, strip -x seems to work

tests/.../v4mapped might be another gitignore slip-through

andyf: I think that's one of yours

but apparenly log -S doesn't do what I want mid-merge

[illumos-gate] 18295 Add header tests to cover 18293 18294 -- Gordon Ross <gordon.w.ross⊙gc>

richlowe - thanks, I will add a check to my RTI flow, and follow up on that.

andyf: the best I can think of is having nightly do `git status -su | awk '$1 == "?"'` when it completes

probably annoying as you're working, though.

with the very obvious acknowledgement of the danger involved, is there any tool that'd let you map/peek/poke a device's registers?

from userland

richlowe, yes, but at least I can make my "prepare for RTI" thing do it.

jbk: pcitool?

our version seems to only deal with interrupts

though maybe there's other PCI ioctls?

jbk: github.com/illumos/illumos-gate/blo…src/cmd/pcitool/pcitool_usage.c#L35 - your man page doesn't document them, but the source code does

that's what i get for relying on the man page :)

wouldn't want people using it who didn't know what they were doing 8-)

fair enough :) I mean, it could cause havoc...

/usr/lib/pci/pcieadm has grown "bar write" and others recently too.

the manual page was added hastily because pcitool underpinned some interrupt remapping stuff on sun4v

before that it wasn't shipped

if I remember it wasn't even really available unless support gave it to you to work around something. I think other than that you either worked there or asked dan mick for it

:o