[sane-devel] CanoScan3000F init/button scan reg/val recognizable?

[amth@suomi24.fi]

Quick cleanup of the usbsnoop-3000f-plugin.log (URBs 6-151+) turned into
registers and values:

Action		Reg	Val	Note
0x83 Write	0x92	0x00
0x83 Write	0xcf	0x00
0x83 Write	0xa3	0x10
0x83 Write	0xa1	0x3f
0x83 Write	0x97	0x0a
0x84 Read	0x8b	0x5f
0x84 Read	0x8b	0x5f
0x83 Write	0x97	0x0f
0x84 Read	0x8b	0xc3	0x8e reports twice, second with val=3D0x00?!?
0x8c BlindWrite	0x04
0x8c BlindWrite	0x00
0x8c BlindWrite	0x04
0x8c BlindWrite	0x00
0x8c BlindWrite	0x04
0x8c BlindWrite	0x00
0x8c BlindWrite	0x04
0x8c BlindWrite	0x00
0x8c BlindWrite	0x04
0x8c BlindWrite	0x00
0x83 Write	0x97	0x00
0x83 Write	0x0f	0x92 0x60
0x84 Read	0x8b	0xc3
0x83 BlindWrite	0x92 0x60
0x84 Read	0x8b	0xc3
0x83 BlindWrite	0x92 0x00
0x83 BlindWrite	0x97 0x00
0x83 BlindWrite 0xa8 0x80 0x83 0x20 0x85 0x50
0x83 BlindWrite	0x83 0x00 0x85 0xc3 0x93 0x80
0x83 BlindWrite	0xa8 0x18
0x8c BlindWrite	0x02
0x8c BlindWrite	0x00		0x8e reports twice, second with val=3D0x00?!?
0x83 Write	0x92	0x00
0x83 Write	0xcf	0x00
0x83 Write	0xa3	0x10
0x83 Write	0xa1	0x3f
0x83 Write	0x97	0x0a
0x84 Read	0x8b	0x5f
0x83 Write	0x97	0x0f
0x84 Read	0x8b	0xc3	0x8e reports twice, second with val=3D0x00?!?
0x83 Write	0x92	0x00
0x83 Write	0xcf	0x00
0x83 Write	0xa3	0x10
0x83 Write	0xa1	0x3f
0x83 Write	0x97	0x0a
0x84 Read	0x8b	0x5f
0x83 Write	0x97	0x0f
0x84 Read	0x8b	0xc3	0x8e reports twice, second with val=3D0x00?!?
...

After each line the is a USB like this:

0x8e Read		0x01

That is obviously USB's common acknowledge signal?

After that there's a continuous loop that is obviously front button scan/=
check?

So the 0x92 ... 0x8b+0x8e=3D0x00 is the loop for the init and standby rou=
tine?

And until those "BlindWrite"s end, that's the init of the scanner?

So as there are so much questionmarks (and names like those "BlindWrite"
steps that are just my own wrong name for them I came up with), maybe I
should leech a USB specification docs for more close info, or could someo=
ne
tell me what to do next (except for clearing up actual scanned image logs=
)...

So what to do next, as I haven't yet found what backend is the closest on=
e
to massage for a quick-fix testing?


--
amth