[Arm-netbook] EOMA-68 passthrough implementation

[lkcl luke]

On Tue, Jul 10, 2012 at 9:22 PM, Derek LaHousse <dlahouss at mtu.edu> wrote:

>>  achh, i'd forgotten about that.  hmmm, in a commercial pass-through
>> card that might involve dropping an 8-bit micro onto the PCB.  darn.
>
> Luke, Ryan and I were talking on IRC, we're both confused by your last
> statement.  I'm going to lay out assumptions, maybe we're not using the
> same terms.
>
> EOMA card = cpu-card or passthrough card, the Master side of most busses
> motherboard = where the card plugs in, the Slave or device side

 we stopped using the term "motherboard" and used the term "I/O Board"
because a "motherboard" has the main CPU.  but other than that, yes.

> Since there is no audio passed across the EOMA header, a passthrough
> card would need to go to great lengths to handle audio.

 assuming that you want audio handled at all: yes.

>  A completely passive passthrough would have no audio ports at all.

 correct.

>  However, a USB
> audio device on the motherboard

 %s / motherboard / "I/O Board" /g

> would be passed across the USB bus to
> whatever controls the passthrough card.

 *click*... you're right! :)  you're absolutely right.  yes.  of
course.  the USB audio IC would be on the I/O Board, thus it would be
USB data, that would just be "passed through".  yes.

 i know where my confusion stemmed from: i had had conversations
[off-list, some months ago] with someone about extracting HDMI Audio.

> There shouldn't be an EEPROM or uC on the passthrough card, as the spec
> doesn't call for an EEPROM on the EOMA card.

 no, that's correct, and if you review the thread again you'll see
that no mention of placing an I2C EEPROM on the passthrough card was
made.  there is however a requirement to place an I2C EEPROM *on the
I/O Board*, and for it to be readable by some device.

>  The EEPROM is on the motherboard,

 %s / motherboard / "I/O Board" /g

> and whatever controls the passthrough would need to read
> it.

 yes.  meaning that the I2C signals need to be "passed through" - i.e.
be "made available somehow", just as henrik said.

>  Or are you saying the passthrough card needs to scale the HDMI
> signal to the proper output, according to the EEPROM?
>
> To put a finer point on it:
> I have purchased an EOMA68-compliant tablet (1366x768 screen).  It has
> onboard storage and a USB audio card driving some pretty crappy
> speakers.  I don't yet own an EOMA68 CPUCard, so I will test my
> equipment with my laptop and an EOMA68-compliant passthrough card.
> This passthrough card looks pretty ugly.  It's shaped like a PCMCIA
> card, with a 2mm key which blocks it from going in my laptop (as it
> shouldn't).  Then, it expands out past flush to provide me a USB port,
> an HDMI, and an eSATA.  I hook all three up to my laptop, where I have
> an HDMI-out, a USB, and an eSATA.  I don't connect the I2C to my laptop,
> because I don't have a port.
> My laptop sees external storage, an audio card, and an external monitor.
> I enable the monitor, and my HDMI puts out a 1080p image.  Because the
> screen is only 768... What happens?  Who scales the image?

 yep.  wark, wark, ooops.   the information about the size of the
screen is stored in the I2C EEPROM.  but just as you say, there is no
way for this nice lovely laptop (or desktop PC) to read the I2C port.
this is why i said that it may be necessary to have a small 8-bit
processor on-board commercial-grade passthrough cards, in order to
convert the I2C signal into say... USB.  or something.  and/or to
handle and instruct the HDMI IC to perform scaling etc. etc.

 but for now i would not worry about all this: just make the I2C
signals available.  if you really really _really_ want to deal with
it, consider sourcing an I2C to USB converter IC and a 2-port USB hub,
chain them together and add the I2C signals (converted to USB) into
the USB stream.  but i would not recommend bothering: it increases the
number of ICs on the board by a factor of 3!

> An additional question I hid in there is: According to spec, is it okay
> for the EOMA68 card to extend beyond flush?

 yes it is.

 l.