> > It tries to use USB resume mode, and looks at the controller to
> > figure out if it's got to recover from powerdown (my laptop does
> > this even without save-to-disk). There were also some issues
> > with cleanup after controller reset, and with reinitializing.
>
> What power management support are you using in the kernel / user space?
> APM? ACPI?
For me, under Linux: APM works, ACPI doesn't (yet).
YMMV, some folk have working ACPI.
> Are you saying that your laptop does save-to-RAM if you leave it alone
> too long? Is this the case you are fixing? But this patch also works on
> save-to-disk?
No, this is a pretty standard mode for use with Linux:
open laptop is in some higher powered mode, but shut the
lid and only a few chips get any power -- RAM, CPU, and
a few more. In my case, the OHCI controller doesn't get
power ... or perhaps something else resets it on powerup,
not that it really matters in terms of what to do.
Some people expected that USB would only be powered down
for save-to-disk; that's not universally true.
> I also note that the USB version in the root hub emulation is hardwired
> (and you updated from 1.0 to 1.1). I note that the OHCI spec doesn't
> have something like the UHCI Serial Bus Release Number register that
> specifies a particular USB revision. However it does contain a
> potentially applicable register that tells us what version of the OHCI
> spec is being used (HcRevision - see 7.1.1 of the OHCI spec). Any
> thoughts about using this?
Avoid. I just toggled that bit since I was tired of seeing
the dumps of that descriptor lie, when I know there's code
that does the "1.1 thing" not the "old 1.0 thing".
> Also, I think this needs to be folded into the UHCI drivers.
Well, it'll be different code, but the same sort of "smarts"
will be needed. It didn't look like the UHCI drivers have any
support there yet.
- Dave
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
