[sane-devel] fixes for genesys backend

[Marc Cousin]

On Friday 07 August 2015 08:56:44 Stef wrote:
> On 06/08/2015 22:16, Michael Thayer wrote:
> > Hello Stef,
> > 
> > Stef <stef.dev <at> free.fr> writes:
> >> On 05/08/2015 10:27, Michael Thayer wrote:
> >>> Stef <stef.dev <at> free.fr> writes:
> >>>>        in genesys backend build 2508 (commit
> >>>> 
> >>>> 4eea901305c9cb66428023c93cbd1a63af8c5fba) I have just pushed fixes for
> > 
> >>>> several opened issues in the bug tracker:
> > [...]
> > 
> >>> Sorry to say this, but the issue which I reported previously with the
> >>> LIDE
> >>> 210 is not fixed when I apply 09daef4 locally.  When I move the usleep()
> >>> up
> >>> to the top of sanei_genesys_test_buffer_empty() however my scanner works
> > 
> > again.
> > [...]
> > 
> >>       and when you have an usleep at both places, how does it behave ?
> > 
> > Yes, that works too.  Seems slightly hacky.  I wonder whether a general
> > delay, e.g. after writing to a register on a non-super-speed device on a
> > super-speed port, might be what is needed?
> > 
> > Regards,
> > 
> > Michael
> 
>      Hello,
> 
>      thanks for the test report.
> 
>      Yes this this a hack : one usleep() for your GL847 device, and
> another for a GL124 (or USB3?). Since we don't have the timing
> requirements of the ASIC, the only thing we can do is trial an error.
> Both delays don't add up much when checking for data being available. We
> could even wait more between each check.


Problem solved for me too (LIDE110 and 210).

Thanks a lot !