On Friday 29 Apr 2011 2:33:42 pm [email protected] wrote:
> Well, I downloaded version 0.4.0 and the latest fprint-demo. fprint-demo
> doesn't see the fingerprint scanner.
>
> However, running the example codes I get this when trying to enroll or
> verify-live.
>
> $sudo ./enroll
> Found device claimed by Validity VFS101 driver
> Opened device. It's now time to enroll your finger.
>
> You will need to successfully scan your finger 3 times to complete the
> process.
>
> Scan your finger now.
> vfs101:error [async_recv_cb] transfer not completed, status = 6
> async:error [fpi_drvcb_enroll_stage_completed] BUG at async.c:161
> sync:error [fp_enroll_finger_img] unrecognised return code -5
> Enroll failed with error -22
>
Most likely, the protocol is different than vfs101/201.
Maybe, it has different registers? maybe takes different values for dev init?
who knows?
So, just adding the dev ids in vfs101.c won't help.
We'll need to understand the protocol from usb-sniff logs.
It's a daunting task, to be honest. VFS101 was made possible due to
Ray Lehtiniemi's efforts to reverse-engineer the protocol from sniff logs.
Kudos to him. And of course, vfs101.c primary author - Sergio Cerlesi. :)
Even the current vfs101.c doesn't cover up all the states for my VFS201 sensor
- but, at least it can get the image out.
> Now it says VFS101, but I left the code alone and added the VFS301 to the
> list of supported devices for the driver.
>
> $ diff vfs101.c vfs101.c.new
> 1542a1543
> > { .vendor = 0x138a, .product = 0x0005 },
>
That is because you _explicitly_ added the devid in the driver - so, while
scanning the devices, libfprint sees that vfs101 driver *claims* to support
the device that it found using usb ids.
> ------------------
> The other examples gave other messages.
>
> $sudo ./img_capture
> Found device claimed by Validity VFS101 driver
> Opened device. It's now time to scan your finger.
>
> image capture failed, code -95
>
>
> Verify failed because the was no file to verify.
>
> cpp-test seems to do nothing, but no error.
>
cpp-test is a simple C++ program that just tries to link to the
fprint library - to make sure that we can still call the libfprint
functions from C++ code.
It doesn't serve any other purpose.
> That's all for tonight (morning). I'll have to dig into the code to see
> what the messages mean.
>
> Brian
IOW, there is a lot more work to be done for Validity sensors.
We're just starting off with some usable code.
But, it needs a lot of patience - which, I hope, we'll have ;)
I'd rather prefer to have an open source driver than anything
closed-source from Validity Inc. ;)
Kunal
_______________________________________________
fprint mailing list
[email protected]
http://lists.reactivated.net/mailman/listinfo/fprint
