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

Hello,

On Apr 17, 2007, at 19:16, Joachim Thiemann wrote:

> On 17/04/07, Michael Lorenz <macallan@netbsd.org> wrote:
>> Maybe - please run ofctl -p ( this will dump the OF tree including
>> properties ) and post the relevant part ( or the whole thing if you're
>> unsure )
>
> No prob, though ofctl is not in 3.1, I just compiled ofctl.c from
> -current.  I assume the result is not affected, though:

Nope, same thing. The OF interface hasn't changed in ages.

> ff87ab28: /pci@80000000/pci-bridge@d/pci1407,200@2
>
> vendor-id               00001407 ........ ........ ........   ....
> device-id               00000200 ........ ........ ........   ....
> revision-id             00000000 ........ ........ ........   ....
> class-code              00070003 ........ ........ ........   ....
> interrupts              00000001 ........ ........ ........   ....
> min-grant               00000000 ........ ........ ........   ....
> max-latency             00000000 ........ ........ ........   ....
> subsystem-vendor-id     00001407 ........ ........ ........   ....
> subsystem-id            00000200 ........ ........ ........   ....
> devsel-speed            00000002 ........ ........ ........   ....
> fast-back-to-back
> name                    70636931 3430372c 32303000 ........   
> "pci1407,200"
> compatible              70636931 3430372c 32303000 ........   
> "pci1407,200"
>            000c:       70636931 3430372c 32303000 ........   
> "pci1407,200"
>            0018:       70636963 6c617373 2c303730 30303300   
> "pciclass,070003"
> reg                     00011000 00000000 00000000 00000000   
> ................
>            0010:       00000000 21011010 00000000 00000000   
> ....!...........
>            0020:       00000000 00000010 21011014 00000000   
> ........!.......
>            0030:       00000000 00000000 00000010 ........   
> ............
> assigned-addresses

Hmm, so OF did try to set up something. The reg property looks weird 
though.

>> Maybe OF put something into device properties and for some reason
>> failed to program the card's config space - if that's the case I can
>> likely fix it.
>
> Your efforts are greatly appreciated :-) Do you need the changes to
> pucdata.c?  I already submitted a pr but can't seem to find it anymore
> (??? I am _sure_ I submitted a pr, but can't even see it the
> netbsd-bugs archive. Strange.)
>
> Anyways, it's simply
>        /* Lava Computers new serial */
>        {   "Lava Computers new serial",
>            {   0x1407, 0x0200, 0,      0       },
>            {   0xffff, 0xfffc, 0,      0       },
>            {
>                { PUC_PORT_TYPE_COM, 0x10, 0x00, COM_FREQ*4 },
>                { PUC_PORT_TYPE_COM, 0x14, 0x00, COM_FREQ*4 },
>            },
>        },
>
> The COM_FREQ multiplier was found "empirically" :-)

Looks like I'll need to look at the puc code - I have no idea what this 
code does ;)
One final thing to test though - could you please build a kernel 
without the puc driver and check if the card's config space looks any 
different?

have fun
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iQEVAwUBRiWj6MpnzkX8Yg2nAQJZKAf/W/nae2Tm10UF5ywFfwY9fCbdUYyK0Jxo
4vqHrDK8a+FHdApEocKI7D7Mr6iXX9CRtn1+xGOuSTD24kVQmAiL12/TmmQ5b98C
ZxVXeMp4GDQvDICEhQhNGOShNq4kUA7xdOmY091sgrHUzTriHR9CVaLjMD9B1fWS
Vnz/afkhJ+M3giIhkBD9A/DRyKSEe9Xp8+EioGfglRPaxvR0N/DNmIEdig2+2gWJ
dbgFpgXVad1Y42R1hG/lYPrViLRF8yHl4AcPHT3iKDxTONNRQh4COoY/7mUcBOku
mv7XrDiq+0i6KHewyJdAlJj4xSv3oIgX7VgJSQQxL7F2jio3U0xBCA==
=1efk
-----END PGP SIGNATURE-----