[linux-dvb] [PATCH] Fix Hauppauge Nova-T 500 USB disconnects
Soeren.Moch at stud.uni-hannover.de
Fri Jul 27 16:18:00 CEST 2007
>>> Respect the patch, it works for me. I have not any disconnect in some
>>> days, althought it is dificult to say that the problem is 100% solved,
>>> because disconnects are random.
Some days ago I wrote a patch for the Nova-T-Stick (see thread
"dib0700 problems (nova-T usb stick)" ). As I could see there,
disconnects are not random, they can only occur directly after a
streamon firmware call. The dib0700 bridge seems to write unintentionally
to the usb bus (or at least driving current into D+/D- or switching
pull-up/down resistors on/off or doing similar strange things) after a
streamon firmware call, usually causing protocol errors (not a real
problem, maybe only for running isochronous transfers to other devices
on the bus), but causing a usb disconnect when accidentally hitting a SOF
packet. So avoiding streamon firmware calls will decrease the disconnect
rate. This is the same idea as in your patches.
>> Relying on device name seems somewhat hackish, however, and doesn't fix
>> anything for nova-t stick ;)
> There are other form to do the same, more easily, for all the dibcom
> but there is a problem: streaming state don't go off when closing the
> althought perphaps that is not important.
> See the atached patch.
This new patch will not help for a Nova-T stick due to the "i2c gating
problem" I have seen there (see thread mentioned above). When simply
starting usb streaming and leaving it on all the time, you are not able
set new tuning parameters, which of course makes this stick useless.
My hope was that Patrick (or some guy at dibcom) will come along with
a real solution for the dib0700 misbehavior after a streamon call.
A new aspect may be, that the disconnect and i2c gating problems might not
be caused by the same source, since the Nova-T 500 seems to have the
disconnect problem, but no i2c gating problem.
More information about the linux-dvb