Yes, at least use the existing environment variable that others already use. That keeps it more uniform. uart_cpu_getdev() looks a bit clunky in its implementation, I'd love it if we could eventually make it support the OPAL console without needing even more cases.

Overall looks fine.

@tuexen this needs revisited to clean up a bit. As @nwhitehorn mentioned, it looks a little strange in how it attaches, so I need to rethink it. Right now the patch exists to help those who have a Talos or Blackbird and want to boot from serial, but output to the onboard VGA/HDMI.

Address comments. Update diff.

I just confirmed it still works on powerpc, nothing broke.

You found the bugs with the kyua tests, did you retest post-changes?

Fix stack space comment.

My understanding is you need both mentor approval and src approval. Consider it src approved by me.

In D21449#468106, @markj wrote:

Please take a look at D21491, which moves the page array into KVA on amd64. It looks like you're doing that already, but I don't see what's preventing the vm_page address range from being allocated from the kernel map's vmem arena(s).

I like the concept overall. A couple nits. Also, can you document this somewhere, maybe in of_real.c?

Thanks for testing the alternative.

That's the goal, yes. There are still bugs with powerpc, that we hope to iron out beforehand. However, if consensus says, we can punt on that for a little bit, but that would also delay a lot of other things I want to get in once we have lld (like ifuncs for pmap, etc).

Committed in r354525.

Looks good.

The trampoline for ppc64 is at 0x3ffffffffffe000 + offset, in the shared page.

Rebase on top of D22170

Address @andrew's feedback. Use the right sub-mask to get the DFSC bits.

Address @bdragon's feedback.