[fprint] uru4000, number of enroll stages

[Martin Hejnfelt]

Hi,

I'm using a Digital Persona U.are.U 4500 and using the current libfprint
from git. I can see in the git log that there were some changes in the
nr_enroll_stages, that was changed to 5 for all devices. This seems to
not make the U.are.U 4500 happy, as it goes into some sort of disco
mode, upon the second enroll stage. First enrollment stage goes fine,
second one it starts blinking with the red diodes and libfprint
continously hammers the log with this:

drv:debug [__ssm_call_handler] 0x2306670 entering state 4
fp:debug [fpi_img_new] length=111360
fp:debug [fpi_imgdev_image_captured] 
fp:debug [fpi_imgdev_image_captured] ignoring due to current state 2
drv:debug [__ssm_call_handler] 0x2306670 entering state 0
drv:debug [__ssm_call_handler] 0x2306670 entering state 1
uru4000:debug [imaging_run_state] hw header lines 289
drv:debug [__ssm_call_handler] 0x2306670 entering state 2
uru4000:debug [sm_read_regs] read 4 regs at 34
uru4000:debug [sm_read_reg_cb] reg value ec
drv:debug [__ssm_call_handler] 0x2306670 entering state 3
uru4000:debug [imaging_run_state] encryption id 39 -> key 7b0f322a
uru4000:debug [imaging_run_state] 0 02 77
uru4000:debug [imaging_run_state] decoding 77 lines
uru4000:debug [imaging_run_state] 1 00 1
uru4000:debug [imaging_run_state] skipping 1 lines
uru4000:debug [imaging_run_state] 2 02 49
uru4000:debug [imaging_run_state] decoding 49 lines
uru4000:debug [imaging_run_state] 3 01 1
uru4000:debug [imaging_run_state] skipping 1 lines
uru4000:debug [imaging_run_state] 4 00 1
uru4000:debug [imaging_run_state] skipping 1 lines
uru4000:debug [imaging_run_state] 5 02 49
uru4000:debug [imaging_run_state] decoding 49 lines
uru4000:debug [imaging_run_state] 6 00 1
uru4000:debug [imaging_run_state] skipping 1 lines
uru4000:debug [imaging_run_state] 7 02 111
uru4000:debug [imaging_run_state] decoding 111 lines
drv:debug [__ssm_call_handler] 0x2306670 entering state 4
fp:debug [fpi_img_new] length=111360
fp:debug [fpi_imgdev_image_captured] 
fp:debug [fpi_imgdev_image_captured] ignoring due to current state 2
drv:debug [__ssm_call_handler] 0x2306670 entering state 0
drv:debug [__ssm_call_handler] 0x2306670 entering state 1
uru4000:debug [imaging_run_state] hw header lines 289
drv:debug [__ssm_call_handler] 0x2306670 entering state 2
uru4000:debug [sm_read_regs] read 4 regs at 34
uru4000:debug [sm_read_reg_cb] reg value ff
drv:debug [__ssm_call_handler] 0x2306670 entering state 3
uru4000:debug [imaging_run_state] encryption id 3a -> key b7ac5a39
uru4000:debug [imaging_run_state] 0 02 77
uru4000:debug [imaging_run_state] decoding 77 lines
uru4000:debug [imaging_run_state] 1 00 1
uru4000:debug [imaging_run_state] skipping 1 lines
uru4000:debug [imaging_run_state] 2 02 49
uru4000:debug [imaging_run_state] decoding 49 lines
uru4000:debug [imaging_run_state] 3 01 1
uru4000:debug [imaging_run_state] skipping 1 lines
uru4000:debug [imaging_run_state] 4 00 1
uru4000:debug [imaging_run_state] skipping 1 lines
uru4000:debug [imaging_run_state] 5 02 49
uru4000:debug [imaging_run_state] decoding 49 lines
uru4000:debug [imaging_run_state] 6 00 1
uru4000:debug [imaging_run_state] skipping 1 lines
uru4000:debug [imaging_run_state] 7 02 111
uru4000:debug [imaging_run_state] decoding 111 lines
drv:debug [__ssm_call_handler] 0x2306670 entering state 4
fp:debug [fpi_img_new] length=111360
fp:debug [fpi_imgdev_image_captured] 

Now if I change the nr_enroll_stages in imgdev.c to 1, it all goes good
and the device works "properly" although I guess matching rate could
then be improved if this could be worked out?

Any clues here?

Thank you in advance!

Kind regards,
Martin Hejnfelt