DragonFlyBSD

:Hi,
:
:I got curious about Alexander Shiryaev's USB harddisk bug and just
:upgraded my system on a USB harddisk to HEAD. It shows some different
:behavior here:
:
:All works unless I use both EHCI and NATA. Either one will work fine
:alone but used together will result in da0 (the USB harddisk) not being
:detected upon boot, ending on the mountroot> prompt.
:
:Sascha
:
:--
:http://yoyodyne.ath.cx

I've seen this too.  If I load EHCI as a module da0 doesn't get detected
    at boot time.

There's some sort of timing issue.  I'm certain it's not NATA's fault
    but instead it is an issue with assumptions the USB code makes when
    EHCI is present.  EHCI has to take over the OHCI driver and its a real
    mess (and badly designed).

-Matt
                    Matthew Dillon 
                    <dillon@backplane.com>

Hmm, strange that old ATA and EHCI work fine, though.

Sascha

I've spent all day trying to make EHCI work at boot time. There are
all sorts of ordering problems, too many to be able to fix it for this
release. Even making little changes I managed to cause my test box to
hang so this is going to have to wait until after the release.

I am going to try to figure out what vinum is doing now.

-Matt

I just tried this again, and with 2.5.1, it works for me. The laptop boots fine
from a USB drive using nata(4) and with ehci(4) loaded. da8 is detected and the
boot proceeds normally.

Sascha

Based on the fact that this is a 'works for me' from the original bug
reporter, I'm closing this.

Cheers,
Alex Hornung

Sorry, I'm re-opening this again.

I've tried it on another laptop and unfortunately the drive was not recognized
there.