On Fri, 2006-04-14 at 14:06 +0800, Jaya Kumar wrote:
> On 3/26/06, Jim Gettys <[EMAIL PROTECTED]> wrote:
> > On Sun, 2006-03-26 at 20:24 +0800, Jaya Kumar wrote:
> > >
> > > AD1888 support in ALSA seems complete to me. Are you having a specific
> > > problem with it?
> >
> > It works as far as it goes, as far as we know.
> >
> > However, it could go further ensuring parts of the chips not in use are
> > off, and more importantly, entirely powering the chip off when not in
> > use, which the 1888 is capable of, and many other chips are not able to
> > do.
> 
> I have a couple of questions here. Are you planning to use the 2.6.16
> cs5535audio driver for your first rev boards? The PM support I posted
> goes as far as turning off the AC Link. The upper layer of ALSA
> suspend has the responsibility for disabling the codec on the other
> side. Currently what ALSA does is to shutdown the following: EAPD, HP
> Amp, ADC & DAC, Analog Mixer with Vref On. I suppose it could go
> further but that would involve testing whether it resumes properly. 

On a human power machine, we really care about milliamps.

And yes, full off of the chip needs testing, and we gather it is a bit
interesting getting it to wake up again.

> I
> could add a ad1888 specific build_ops->suspend in the codec layer to
> shutdown completely. I'd need someone to test this for me since
> nothing here seems to use the AD1888. Unless of course you're willing
> to ship a test board to India.

Well, I will probably be able to do so this summer.  We are having a
large number (500) boards built for developers to hack on.

Though I also may have a couple sample AD1888's that can plug into old
Pentium 4 mother boards that have a funky slot that often existed on
that era motherboard.  I forget the name of this slot.

When we have developer boards, we expect to be able to ship worldwide;
one of OLPC's sponsors is Brightstar, who have worldwide logistics
capability (including India), and has volunteered to help us distribute
and support the developer boards.


> 
> > As you might imagine, we care a whole lot about what our minimum power
> > consumption is, since the CPU will be able to suspend to RAM while
> 
> Just to be certain, are you using the GX2 or the GX3 cpu on your first
> rev boards? Are you using the CS5535 or the CS5536 bridge?

GX2.  CS5536.

> 
> > keeping the display running (using our DCON chip).  So our bottom end
> > power consumption when, say, reading a book and not doing anything,
> > should be in the neighborhood of .5 watts; even if we continue to
> 
> Impressive goals.

Yup.

> 
> > forward packets in the mesh, (which we can also do with the CPU
> > suspended, due to the choice of the Marvel wirless chip), we'll be in
> > the neighborhood of 1 watt.
> 
> Do you already have drivers for all this stuff? I'd be happy to help
> in that area.

Nope.  A simple matter of programming as they say ;-).

I'm sure we could use help.

> 
> >
> > So power consumption that would be lost in the noise on most systems
> > will be very significant on our system.  We'll be in the values normally
> > only seen on PDA's.
> >                               Regards,
> >                                   - Jim
> >
> >
> > >
> > > There are other items on your kernel work list that look attractive to
> > > me. Let me know if you actually need help on them or already have
> > > people to work on them.
> > >
> > > Best regards,
> > > jayakumar
> > > --
> > > olpc-software mailing list
> > > [email protected]
> > > https://www.redhat.com/mailman/listinfo/olpc-software
> > --
> > Jim Gettys
> > One Laptop Per Child
> >
> >
> >
-- 
Jim Gettys
One Laptop Per Child


--
olpc-software mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/olpc-software

Reply via email to