[RFC] status error patch

[castet.matthieu at free.fr]

Selon Duncan Sands <baldrick at free.fr>:

> > > If you immediately resubmit the urb, won't it just come back with
> > > -EILSEQ again?
> > >
> > no if there is enough data to fill all the packets of the iso urb.
>
> I've been reading about iso urbs, and it looks like EILSEQ is not
> unreasonable when there is a timeout.  My plan is to add iso support
> to the speedtouch driver.  That will (1) help me learn about iso, (2)
> help me to adapt usbatm to make life easier for people with iso capable
> devices, and (3) make it clearer whether this is a problem with the
> iso implementation in usbatm or a problem with your modem.

Ok, but if it is the the modem that is bogous, the subdriver will need a way to
mask some errors or it will need to always provide a patched usbatm lib( I find
 this ugly) : if you look in other linux drivers, you well see that no all
device are perfect and some need some quirks...

I don't think it this due to the iso implementation in usbatm, as there is the
same error in original adi driver.


thanks

Matthieu