Re: Heads up: TI OMAP3 and AM335x switched to FDT

["John D. Baker" <jdbaker%consolidated.net@localhost>]

On Tue, 3 Dec 2019, John D. Baker wrote:

> On Tue, 3 Dec 2019, Jared McNeill wrote:
> 
> > In the EFI bootloader, you can get a list of devices with the 'dev'
> > command:
> 
> Mine shows:
> 
>   > dev
>   hd0 (7600 MB): VenHw(E61D73B9-A384-4ACC-AEAB-82E828F3628B)/Usb(0x6,0x0)
>   hd1 (3688 MB): VenHw(E61D73B9-A384-4ACC-AEAB-82E828F3628B)/Usb(0x6,0x1)
>   net0 xx:xx:xx:xx:xx:xx
> 
>   default: net0
> 
> even though both the SD card (hd0) and the on-board eMMC have MBRs,
> disklabels and a number of partitions.

I think I see what's going on.  evbarm seems to be one of those ports
where partition "c" is the whole disk if the MBR partition table doesn't
define a "NetBSD" partition and "c" is the NetBSD partition and "d" is
the whole disk if it does.

When I partitioned my SD/eMMC devices, I operated on the "c is whole disk"
model (MBR in sector 0, disklabel in sector 1).

Apparently, "bootarm.efi" assumes the "d is whole disk, c is NetBSD part"
model, so expects the disklabel at the start of the NetBSD part.

I can probably retcon the PC-style partitioning, but it looks like I'll
have to dig up one of my uSD-USB adapters and do the fixup in an x86
machine as even the old BEAGLEBONE kernel is suffering "sdhc0: cmd timeout"
errors trying to operate on the SD card (fsck -f).

-- 
|/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
|\ / jdbaker[snail]consolidated[flyspeck]net  OpenBSD            FreeBSD
| X  No HTML/proprietary data in email.   BSD just sits there and works!
|/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645