SeaBIOS and USB?

[Kevin O'Connor]

Hi,

The 2009 GSOC is winding down.  I'd like to get some feedback on what
people's plans are for USB support.

The GSOC experience demonstrated both some successes and challanges.
I believe Jason was able to launch some executables with the code
which demonstrates that it can work.  However, there were some issues
with memory management.

Besides the direct issue of not reserving the rom's memory, the memory
needed is also an issue.  The usbrom itself is just squeezing into
128K and it looks like it needs an additional 1M of working memory.
In practice, I don't think the USB bootup code should use more than
30K of option rom space or more than 20K of reserved memory.

Anyway, it would be helpful to hear from those that are planning to
work in this area and the work that they are planning.

Thanks,
-Kevin

[Leandro Dorileo]

Hi Kevin

On Thu, Aug 27, 2009 at 10:55 PM, Kevin O'Connor wrote:
> [...]
> Anyway, it would be helpful to hear from those that are planning to
> work in this area and the work that they are planning.

I`m still working on supporting OHCI devices. I haven`t prepared
patches for review but a preliminary work can be found here[1]. It`s
most a hack and you will find for example an endless loop after the
usb initialization function.

[1] - http://vps.dorilex.net/~dorileo/libpayload/ohci_dev/

Regards....

[Jason Wang]

hi kevin,
   since this week is the first week of new semester, i should go though
each department for checking in schedule.
about the usbrom i think the usbrom memory still should be analyzed in the
following days. it should be basic memory
of libpayload and usb stack. PMM code will be completed in a few days. I
will test it later.
anyway as you said it would be helpful to hear from anyone who is interested
in usbrom project.

On Fri, Aug 28, 2009 at 10:55 AM, Kevin O'Connor  wrote:
> Hi,
>
> The 2009 GSOC is winding down.  I'd like to get some feedback on what
> people's plans are for USB support.
>
> The GSOC experience demonstrated both some successes and challanges.
> I believe Jason was able to launch some executables with the code
> which demonstrates that it can work.  However, there were some issues
> with memory management.
>
> Besides the direct issue of not reserving the rom's memory, the memory
> needed is also an issue.  The usbrom itself is just squeezing into
> 128K and it looks like it needs an additional 1M of working memory.
> In practice, I don't think the USB bootup code should use more than
> 30K of option rom space or more than 20K of reserved memory.
>
> Anyway, it would be helpful to hear from those that are planning to
> work in this area and the work that they are planning.
>
> Thanks,
> -Kevin

[Kevin O'Connor]

On Thu, Aug 27, 2009 at 10:55:31PM -0400, Kevin O'Connor wrote:
> Anyway, it would be helpful to hear from those that are planning to
> work in this area and the work that they are planning.

To include myself on the discussion of plans - I've been reading up on
the USB specs and am looking at what it would take to add UHCI
keyboard support to SeaBIOS.

-Kevin