On Mon, Dec 1, 2008 at 3:13 PM, petercarm <linuxtv@xxxxxxxxxxxxxxxxxxxxx> wrote: > The LIRC project provides drivers for otherwise unsupported RC hardware. > For dvb-usb receivers it just piggybacks off the /dev/input device > file. No driver from LIRC is used. The LIRC involvement is purely in > userspace. This is not a LIRC problem. The problem arises whether or > not lircd is started. I recognize that lirc is purely in userspace, but it would be desirable to setup a comparable environment that polls the /dev/input device so I can attempt to reproduce the behavior locally. That's interesting that it occurs even if lircd is not running. > The problem is suppressed if the disable_rc_polling parameter in dvb_usb > is set to 1 but obviously this disables all RC support. This whole situation is suspicious since the IR polling has no interaction with the i2c subsystem. In fact, it no longer even uses the same request endpoint. Can you provide any quantitative data as to the frequency of failure and under what circumstances this occurs? Does it prevent tuning from succeeding, or do you lose lock after the device is capturing? Do you see i2c errors when not capturing? How often do the errors occur? Are you unable to get tuning lock in 100% of attempts? Are the i2c errors always against the mt2060, or are you seeing other i2c errors as well? You mentioned that "The Nova-T 500 and the Nova TD USB stick are the two dibcom devices". I realize this, but my question was, are you seeing failure with both of these two devices, and are these the only two devices you see errors with? Devin -- Devin J. Heitmueller http://www.devinheitmueller.com AIM: devinheitmueller _______________________________________________ linux-dvb mailing list linux-dvb@xxxxxxxxxxx http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
