On Mon, Feb 23, 2009 at 4:52 AM, Janne Grunau <j@xxxxxxxxxx> wrote: > Hi, > > I have some difficulties with the Hauppauge Nova-T 500 and 2.6.28 kernels. > > Following USB errors appear with 2.6.28.4 every 100ms: > > Feb 15 02:46:03 golem [ 7720.876132] usb 2-1: events/3 timed out on ep1in len=0/6 > Feb 15 02:46:03 golem [ 7720.976039] usb 2-1: events/3 timed out on ep1in len=0/6 > Feb 15 02:46:03 golem [ 7721.076068] usb 2-1: events/3 timed out on ep1in len=0/6 > > The device is still useable but occasionally I see the same error > with additional usb errors (iirc same kernel and config) > > Feb 17 20:33:17 golem [ 14.733031] ehci_hcd 0000:01:06.2: reused qh ffff8800bf80d140 schedule > Feb 17 20:33:17 golem [ 14.733040] usb 2-1: link qh64-0001/ffff8800bf80d140 start 63 [2/0 us] > Feb 17 20:33:17 golem [ 14.783035] usb 2-1: unlink qh64-0001/ffff8800bf80d140 start 63 [2/0 us] > Feb 17 20:33:17 golem [ 14.783107] usb 2-1: events/3 timed out on ep1in len=0/6 > ... > Feb 17 20:34:12 golem [ 128.130059] ehci_hcd 0000:01:06.2: reused qh ffff8800bf80d140 schedule > Feb 17 20:34:12 golem [ 128.130069] usb 2-1: link qh64-0001/ffff8800bf80d140 start 63 [2/0 us] > Feb 17 20:34:12 golem [ 128.131007] ehci_hcd 0000:01:06.2: force halt; handhake ffffc20000050014 00 > 004000 00000000 -> -110 > > ehci gives up and the device is unuseabe. > > This is still the case with 2.6.28.7. > > git bisect blames following change: > > | commit 99afb989b05b9fb1c7b3831ce4b7a000b214acdb > | Author: Devin Heitmueller <devin.heitmueller@xxxxxxxxx> > | Date: Sat Nov 15 07:13:07 2008 -0300 > | > | V4L/DVB (9639): Make dib0700 remote control support work with firmware v1.20 > > 2.6.28.x with DVB drivers from v4l-dvb hg works as expected bu I fail > to see which changeset fixed it. If you have an idea I'll test it. > Otherwise I'll bisect v4l-dvb hg. > > This should be fixed in 2.6.28-stable. > > Janne Hello Janneg, Thank you for reporting this issue. We actually rely on the bulk endpoint timing out to query the IR port if there is no keypress. I wonder if they introduced a new kernel warning in 2.6.28.4 when this occurs. Regarding the other issue, is there any correlation with the errors to plugging/unplugging the devices? Perhaps there is an issue with cancellation of the thread that does the polling when the device is disconnected. I actually don't have a Nova-T 500. I did all the development on the dib0700 based Pinnacle 801e. Devin -- Devin J. Heitmueller http://www.devinheitmueller.com AIM: devinheitmueller -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
