Re: snapper does not work on PowerBook5,6

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

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

Hello,

On Oct 17, 2010, at 3:44 PM, Frank Wille wrote:

> > Umm, what does snapper's interrupts property look like on your
> > machine? The DMA and the status change interrupt are different types
> > on mine.
>
> Yes, I think that's the same on all snapper machines:
> 0000001e 00000001 00000001 00000000 00000002 00000000
>
> I wonder for what this irq 30 is needed? Plugging in headphones is  
> detected
> by an additional interrupt (irq 61 in my case).

Argh, I forgot that snapper uses a separate gpio interrupt for  
headphone detection. So that interrupt is only for error conditions.

> > > Maybe there is no interrupt problem, but a DMA problem? I found the
> > > location
> > > where the freeze happens: It is an infinite loop in dbdma_stop(),
> > > while
> > > waiting for two status flags to indicate DMA is flushed and stopped.
> > >
> > > But since dbdma is also used for wdc0 I'm not sure what can be wrong
> > > here.
> >
> > This is getting weirder and weirder.
>
> Are there any docs about using dbdma for audio? I admit that I don't  
> even
> understand yet how dbdma knows what to do with the data it is  
> reading. :)

It's hard wired IIRC, the dbdma registers associated with the audio  
chip are for audio DMA ;)

> I just compared the source with OpenBSD. Seems they have already  
> improved
> dbdma.c to use bus_dma(9). And FreeBSD has a very clean  
> reimplementation of
> everything, but I saw no special handling for newer PowerBooks either.

Yeah, macppc predates bus_space. Some drivers have been bus_spacified  
during ppcoea_renovation but not all of them.

> BTW, there was already a posting more than 4 years ago, which  
> describes
> similar snapper problems on a PowerBook6,8:
> http://mail-index.netbsd.org/port-macppc/2006/02/10/0001.html
>
> He didn't hear sound either, but surprisingly he received a single  
> irq1.

Hmm, I must have missed that.

> > Try to force the DMA interrupts to be level triggered.
>
> This will give me infinite interrupts. More than I ever wanted. :)
> No, I think level-triggered is correct.

At least that verifies that the interrupt is enabled and working.  
Leaves the question why they don't fire on your hardware - I need to  
stare at the code for a bit, maybe it's something trivial like having  
to explicitly clear them before use or something.

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

iQEVAwUBTLxAAcpnzkX8Yg2nAQLRAggAjMs295n635yC2jTA4skouFHh9eDTjz3N
1fMOi7BndyB3VmQSUuxSlfglWJrJTyEYC7U+UZLaEJZ5EjU4DAsrBxQxRuuo+B8v
MDjYxbV0uxuq7+IVOruH6iG8+6ouCPgmL0fDvY7ZcJVvBrsuv9xObJlmMfdstmia
h0tlr4jChnzQB3EGWpe6z9rCBkjh3mlrgqY+VHJ78JIxftW8y88Sr/j79pSwT5Rw
jkDxNJd+2BtOanaBON+g8I8f/MbZG6CpUTdkdrsqccOs6Ia1spZBEzlxBoo45UDH
lwFQ+pS8vkK+yQtibm1iTTRCrpnKlPXxMhF9F4rA80OVu+1hBaWu5g==
=IqCS
-----END PGP SIGNATURE-----