nufan_wfk at yahoo
Feb 7, 2003, 1:34 PM
Post #3 of 22
> You'll notice that these come in pairs. This is how
> reads work. The master
> sends the address, the slave responds.
ok, so this is the driver reading the value of reg
0x9, and the card responding with "it's 0xc6":
845uS: 21 W 09
497uS: 21 r C6
> > 21,W,0A,8E - write 0x83 to reg 0x0a
> Yes. Hmm, that's kinda wonky. Typically reading
> and writing have two separate register spaces.
err that should be write 0x8e... this (is the only)
one i understand... i think :)
> I highly suggest you concentrate on one chip at a
> time otherwise you're going to get confused.
heh yeah. Don't worry.. Right now i'm only interested
in the saa7115. I've made some progress so far.
> > 50,r,84,12,00,00 - read multiple data? explain?
> Register width is dependant on the chip, and which
> register or register set it's writing to.
if i understand correctly, this is a response to a
read request, so it makes sense now.
> > 40,W,12,00,13,0C,00 - write 4bytes to reg 0x12?
> Write 0x0c00 to register 0x0013 in set 12. -or-
> Write 0x130c00 to register 0x00 in set 12.
What about write 0x00130c00 to register 0x12? that's
my hunch (who writes 24-bit values? :)
Anyhow, obersvations so far (using the saa7114 as a
0x87 (enable digitizer) seems to be left on (0x1)
except when making changes to 0x88, and possibly 0x08,
0x30, 0x31, 0x32, 0x34, 0x35, 0x36, 0x80 (not always
though) .. This is different than our driver currently
works... we leave it off and turn it on when we want
and for register 0x02 (input select: tuner, composite,
svideo are options? maybe just tuner/composite and
Here's my interpretation (this is where he switches
from tuner to composite (i'm assuming his card is like
mine and has a svideo->composite adapter)
(previous value of 0x02 was 0x84)
160.621mS: 21 W 02 88 //set input to ?
873uS: 21 W 09 //request value of 0x09
497uS: 21 r 56 // value is 0x56
585uS: 21 W 09 C6 //set 0x09 to C6
1.010mS: 21 W 02 80 //switch input to direct-in
845uS: 21 W 09 //request value of 0x09
497uS: 21 r C6 // value is c6
562uS: 21 W 09 56 //set 0x09 to 56
2.220mS: 21 W 09 //request value of 0x09
499uS: 21 r 56 // value is 56
this sequence happens twice with about a second in
between.. Windows driver is wierd i suppose. Based on
the values in the saa7114 driver, 0x02 (low nibble) is
video source, and 0x09 is the chrominance trap bypass
(hi nibble).. These values are different than those of
the 7114 though.. appears 0x56 is the correct setting
for line-in and 0xc6 is correct for tuner.
Oh and john, thanks for the quick response. you be
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
ivtv-devel mailing list
ivtv-devel [at] lists