Re: snapper does not work on PowerBook5,6

[Michael <macallan%netbsd.org@localhost>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

On Oct 13, 2010, at 3:52 PM, Frank Wille wrote:

> Michael wrote:
>
> > > Here the whole system freezes (interrupt storm?) when I send a
> > > bigger file to /dev/audio. With very short
> > > files nothing will happen at all.
> >
> > This is weird.
>
> I correct myself: the freeze happens always, also with small files.
>
>
> > Got any dmesg output?

> codec at ki2c0 address 0x6a not configured

This is the culprit. Apparently Apple renamed the node for the audio  
codec again or did something else to it that keeps the deq 'driver'  
from attaching, which leads to this:

> snapper0: software codec

I distinctly remember seeing a reference to a TAS3004 in your OF tree.
Getting deq to attach to it will likely fix the audio problem.

> > Apple seems to reorganize its device tree with
> > every model no matter if the actual hardware changed or not. Snapper
> > needs the deq driver to attach in order to find the codec which  
> > should
> > live on iic at ki2c at obio.
>
> Seems this codec was not found. The OF structure on my PowerBook  
> looks like
> this:
>
> ff98be28:       /i2c@18000
> ff98ca00:         /i2c-bus@0
> ff98cbc8:           /codec@6a

Exactly.

> So there are two problem in deq.c:
> 1. The node is called "codec" instead of "deq".
> 2. There is no "i2c-address" property.

Then it should use the reg value instead - 0x6a looks right to me.

> I hacked that in to make deq attach:
> [...]
> snapper0 at obio0 offset 0x0: irq 30,1,2
> pmu0 at obio0 offset 0x16000 irq 47:  rev. 12
> pmu0: initializing IIC bus
> pmu0: initializing RTC
> pmu0: pmu not configured
> pmu0: power-mgt not configured
> smartbat0 at pmu0 addr 0: smart battery
> flags: 5 AC present
> ki2c0 at obio0
> iic0 at ki2c0: I2C bus
> deq0 at ki2c0 Apple Digital Equalizer, addr 0x6a
> wdc0 at obio0 offset 0x20000 irq 24, level triggered: DMA transfer
> atabus0 at wdc0 channel 0
> snapper0: codec: TAS3004
> audio0 at snapper0: full duplex, playback, capture
> [...]
>
> But still no difference. The system freezes when writing to /dev/ 
> audio. :(

Damn, there goes that theory.
The weird thing is that the DMA machinery didn't really change at all  
from the first PCI Macs all the way up to the last G5s - I have no  
idea what could cause this. Now it's probably time to sprinkle printfs  
and see where exactly the hangup happens.
Umm, reminds me - do you get interrupts from snapper? It should list  
three - one of input, one for output and one for status changes ( like  
(un)plugging headphones, errors etc. )

have fun
Michael

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iQEVAwUBTLYfo8pnzkX8Yg2nAQLdzwf7BNcdCvbDiw0+MzSar1E4TtQ5vUu+yX3e
9pWFkJRXo5l9hSPo7D9RQxelEOqLFaFrEsv4fu6DqXYFV/1xte98hHtmHgDKVPUO
wgefJGp3bXiQcvb8udsDByIVoUCmvwmyZ4XmsrjulFTfPrLS4UfCJ8dxHEFDbemn
MASKa7zWccuiWpDT3yZ/SU6Qk2qfALo9JmpLlsP8yIDHtmBdAR5IqKU68MQSI1xx
ZN/67ttCcxeyJhOo8xXsXKWS2D/oX6alQFMTmI0Vn2oRJGm27e/+pDL2K4pLiJ8h
qxyL7DeTZqAVzPFT3lLU0tQO2LZSoRndt4koN5s8R+Cwkgz8gMDMBw==
=D6pU
-----END PGP SIGNATURE-----