Re: gpt, wedges and raidframe on -current

[Hauke Fath <hf%spg.tu-darmstadt.de@localhost>]

At 12:23 Uhr -0800 25.02.2011, Brian Buhrow wrote:
>My guess is that the
>new code which boots off of the gpt partitions doesn't also do the skip
>ahead 64 blocks trick to see if there's a bootable raid partition.   If it
>doesn't, that could explain why the non-raid gpt partition boots and the one
>containing a raid set doesn't.

Hmmm... Just to be sure - does the primary bootstrap (the bootxx_*) go to
the start of the outer gpt (the dk0 raid partition), or to the start of the
inner gpt (the dk1 root partition)?

From browsing <sys/arch/i386/stand/mbr/gpt.S>, it looks as if the sector 0
boot code knows how to walk the gpt table. But does it descend recursively,
i.e. look at gpt partitions within a (special, e.g. raid) gpt partition?

And does a primary bootstrap look at other partitions but the one it was
written to?

Puzzled,
        hauke

-- 
     The ASCII Ribbon Campaign                    Hauke Fath
()     No HTML/RTF in email            Institut für Nachrichtentechnik
/\     No Word docs in email                     TU Darmstadt
     Respect for open standards              Ruf +49-6151-16-3281