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

[Helmut Toplitzer]

Hi!


Just in case ....

Maybe Ingo Flaschberger's idea should be the default (tolerating errors).



Bye 
Helmut


----------  Forwarded Message  ----------

Subject: Re: [linux-usb-devel] Re: [pvrusb2] [Linux-usb-users] USB EHCI VIA 
disconnect problem
Date: Friday 03 February 2006 20:35
From: Alan Stern <stern at rowland.harvard.edu>
To: Helmut Toplitzer <pvrusb2 at toplitzer.net>
Cc: pvrusb2 at isely.net, USB development list 
<linux-usb-devel at lists.sourceforge.net>

On Fri, 3 Feb 2006, Helmut Toplitzer wrote:
> HI!
>
> As you suggested I bought a new USB controller (ALi, ohci/ehci) and
> disabled the old one.
>
> OK. It now works great. Well, until it stops. (Without hang
> this time but it stops. I need too restart mythtv.)
>
>
> Feb  3 18:13:45 kernel: pvrusb2 /*---TRACE_CTL----*/ pvr2_encoder_start
> Feb  3 18:13:45 kernel: pvrusb2 /*---TRACE_READ---*/ pvr2_ioread_start
> id=cf8af380
> Feb  3 18:18:44 kernel: ehci_hcd 0000:00:07.3: devpath 5 ep4in 3strikes
> Feb  3 18:18:44 kernel: pvrusb2 /*---TRACE_READ---*/ pvr2_ioread_stop
> id=cf8af380
>
> Stopping myth:
>
> Feb  3 18:20:53 lircd 0.7.1pre2[10482]: removed client
> Feb  3 18:21:06 kernel: pvrusb2 /*--TRACE_STREAM--*/ disable
> Feb  3 18:21:07 kernel: pvrusb2 subsys mask changing 0x1e:0x0 from 0x1f to
> 0x1 Feb  3 18:21:07 kernel: pvrusb2 /*---TRACE_CTL----*/ pvr2_encoder_stop
> Feb  3 18:21:07 kernel: pvrusb2 /*---TRACE_CTL----*/
> pvr2_hdw_cmd_usbstream(0) Feb  3 18:21:07 kernel: pvrusb2
> /*---TRACE_CTL----*/ decoder disable
>
> Funny. Isn't it?

At this point I think you have to take it up with the pvrusb2 or mythtv
developers.  It looks like the program stopped as soon as it got a single
error.  It should be smart enough to retry the operation, perhaps after a
short delay.  Or reset the device if that doesn't work.  And if nothing
succeeds it should fail gracefully instead of just stopping.

The log message from ehci-hcd says that there was some kind of error in
the USB communication.  The information isn't very specific, and it could
be caused by something as simple as a poor-quality USB cable.  Or it could
be the device crashes.

Alan Stern

-------------------------------------------------------

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