[pvrusb2] Re: [linux-usb-devel] USB EHCI VIA disconnect problem

[Helmut Toplitzer]

So. After a looong time I decided to drop you a note
whats going on with my "small but funny" usb/ide problem:
(Maybe others are interested into this too)

First: I didn't bought a new USB controller.

Alan, your patch went quite well and it looks like the device
isn't disconnecting any longer. So I did some
usb-storage transfer tests and it looks like it had
no unwanted side-effects.

However the IDE-transfer problem still exists, and I
posted it to the linux-ide list. (It's not directly a
USB problem because also the CPU utilization has
influence on it)

The hang after some time using the ehci-hcd driver still
happens but I was able to isolate it. It looks like the
also onboard audio-module (snd_via82xx,snd_ac97_codec)
is getting wrong with either the ehci-hcd driver or
the pvrusb2 driver using ehci-hcd. Till using pvrusb2 with
ehci-hcd I had no problems using the audio output. Unloading
ehci-hcd switching back to uhci-hcd solves the problem again.
Directing audio output to /dev/null also works without further 
interruptions.
(Maybe this is also not a perfect solution :-) but
someone gets an idea whats going comletely wrong here)

Beside this the msp3400 module for my pvrusb2 device 
included in the mainline kernel is not usable for me, because
it hangs too after some channel-switches. So not funny enough,
the msp3400 module bundled with the pvrusb2-driver (an older
version of this module) solves at least this problem. (Looks
like v4l-people screwed something up with inclusion into the
mainline kernel).

Ok. That's it for now. So I'm now a bit glueless to which
of the problems I will stick next ...
(Just send me some ideas :-)

Howere I'm willing to test patches if someone has an idea
whats happening here.

Cheers
helmut

PS: You're right if you think that I already reached
a high level of insanity.


-- 
My GNUpg fingerprint http://www.gnupg.org
4563 F4FB 0B7E 8698 53CD  00E9 E319 35BD 6A91 1656