in general this looks fine! thanks!

ok, if you have tested it between macosx and linux and freebsd then that's good enough for me.

ok, this is the next thing to dig into and figure out how to land.

wait, did this change at some point?

Ah, you haven't gone through and marked them as "done".

please address lwhsu's comments!

this is fine; we can update this after it lands to configure it as a bitmask and per-softc.

I think its fine to get it up and going, but in theory being able to transmit on all of the rings shouldn't stall traffic. If it is then chances are your RX thread is keeping your TX queue moving forward.

please include a reference in the commit message to where these are defined (eg if they're in a spec or just local to freebsd)

let's figure this out

hi! this doesn't have a commit message in the description; would you please update that? thanks!

I can't find where this is documented - all of the SDMs I've looked at so far describe IA32_PERF_STATUS but do not cover the format of the register above bit 31.

get rid of the bootverbose change and putit in a different diff, and then i'll approve/land

interesting, is there no way to force a firmware reload without just power cycling the device?
eg if i wanted to test two different firmware images by unloading and reloading the driver?

oh interesting, what did you observe this on?

landed, i fixed it locally

Bounce that to -ppc mailing list too. Justin would likely know better for the non-POWER chips.

update!