[sane-devel] Adding Scanner to Backend Configuration

[Stéphane VOLTZ]

Le mercredi 23 mai 2007, gmurray at cloudnet.com a écrit :
> Hugh wrote:
> > I wanted to know if anyone would like to look at them.
>
> I would like to see the filtered output.  Please post/send me
> a url.
>
> In the initial URBS there seemed to be a lot of unnecessary data
> sent.  In the first 1393 URB numbers, there is a block of
> 0x0200 bytes that is resent to the device 312 times.  The
> block is constant, except for the first byte.  Sending the
> block to the device returns an URB with just 1 byte.
> That byte becomes the first byte of the 0x0200 byte block
> in the next resend.  Mysterious, but interesting.
>
> regards,
> Gerald

	For me it is a bulk read on endpoint 0x83. Which is consistent with the lsusb 
output. The windows driver sends out a useless buffer of 512 bytes, while the 
actual data is the one byte read back.
	With a modified 'parse1.awk' that doesn't skip this, the USB log look like:

URB     8  bulk_out len     4  wrote 0x01 0x46 0x01 0x00
URB     9  bulk_in  len     1  read  0x08
URB    10  bulk_out len     5  wrote 0x00 0x02 0x01 0x00 0x01
URB    11  bulk_out len     5  wrote 0x00 0x02 0x01 0x00 0x00
URB    12  bulk_out len     5  wrote 0x00 0x01 0x01 0x00 0x00
URB    13  bulk_out len     5  wrote 0x00 0x01 0x01 0x00 0x28
URB    14  bulk_out len     5  wrote 0x00 0x90 0x01 0x00 0x4f
URB    15  bulk_out len     5  wrote 0x00 0x92 0x01 0x00 0xff
URB    16  bulk_out len     4  wrote 0x01 0x91 0x01 0x00
URB    17  bulk_in  len     1  read  0x00
URB    18  bulk_out len     5  wrote 0x00 0x93 0x01 0x00 0x00
URB    19  bulk_out len     5  wrote 0x00 0x91 0x01 0x00 0x1f
URB    20  bulk_out len     5  wrote 0x00 0x95 0x01 0x00 0x1f
URB    21  bulk_out len     5  wrote 0x00 0x97 0x01 0x00 0x1f
URB    22  bulk_out len     5  wrote 0x00 0x9b 0x01 0x00 0x00
URB    23  bulk_out len     5  wrote 0x00 0x9c 0x01 0x00 0x07
.....
URB    37  bulk_out len     5  wrote 0x00 0x99 0x01 0x00 0x28
URB    38  bulk_out len     5  wrote 0x00 0x9a 0x01 0x00 0x03
URB    39  bulk_out len     5  wrote 0x00 0x80 0x01 0x00 0x10
URB    40  bulk_out len     5  wrote 0x00 0x8d 0x01 0x00 0x00
URB    41  bulk_out len     5  wrote 0x00 0x8d 0x01 0x00 0x04
URB    42  bulk_out len     4  wrote 0x01 0x8b 0x01 0x00
URB    43  bulk_in  len     1  read  0x00
URB    44  bulk_out len     4  wrote 0x01 0x8b 0x01 0x00
URB    45  bulk_in  len     1  read  0x00

	These look single register read/write for me:
	first byte read/write flag, second is register index, then a LSB register 
count. Then there is the value written, or a bulk in for the value read (on 
endpoint 83). There read/write flag of value 0x02, 0x04 and 0x05. 0x05 is for 
data reading.

When processed with custom scripts it gives a pseudo code like this:
write(0x8d,1)=0x00 
write(0x8d,1)=0x04 
read_reg(0x8b,1)=0x00 
read_reg(0x8b,1)=0x00 
write(0x85,1)=0x00 
write(0x87,1)=0x00 
write(0x88,1)=0x70 
read_reg(0x8b,1)=0x00 
read_reg(0x8b,1)=0x00 
write(0x85,1)=0x03 
write(0x87,1)=0x00 
write(0x88,1)=0x28 
read_reg(0x8b,1)=0x00 
read_reg(0x8b,1)=0x00 
write(0x85,1)=0x06 
write(0x87,1)=0x00 
write(0x88,1)=0x28 
read_reg(0x00,1)=0x40 
read_reg(0x01,1)=0x28 
read_reg(0x02,1)=0x00 
read_reg(0x03,1)=0x00 
read_reg(0x04,1)=0x0c 
read_reg(0x05,1)=0x00 
.....
..... all values from 0x00 to 0xFE, look like a 255 registers ASIC 
.....
read_reg(0xfd,1)=0x00 
read_reg(0xfe,1)=0x00 
write(0x60,1)=0x15 
write(0x10,1)=0x05 
write(0x9b,1)=0x01 
read_reg(0x46,1)=0x08 
write(0x80,1)=0x12 
write(0x11,1)=0x25 
write(0x9b,1)=0x01 
read_reg(0x46,1)=0x08 
read_reg(0x46,1)=0x08 
read_reg(0x46,1)=0x08 
write(0x80,1)=0x12 
write(0x11,1)=0x25 
write(0x01,1)=0x28 
write(0x04,1)=0x0c 
write(0x05,1)=0x00 
write(0x06,1)=0x00 
write(0x01,1)=0x29 
write(0x71,1)=0x01 
URB   595  bulk_out len     5  wrote 0x02 0x30 0x01 0x00 0x11 
write(0x71,1)=0x16 
write(0x72,1)=0x01

	Does it look familiar to anyone ? What ASIC could it be ?

Regards,
	Stef