--- On Fri, 26/8/11, Andreas Oberritter <obi@xxxxxxxxxxx> wrote: > I first thought that you were talking about a > regression in Linux 3.0.x. Heh, yes and no. I am talking about a regression that I am definitely seeing in 3.0.x. However, I cannot say which kernel the problem first appeared in. > This initial event with status=0 exists since 2002. It's > used to notify a new tuning operation to the event listener. > > http://www.linuxtv.org/cgi-bin/viewvc.cgi/DVB/driver/dvb_frontend.c?revision=1.6.2.30&view=markup OK, that's different. I've only noticed this regression because xine has started having trouble using a brand new DVB adapter. Debugging the problem has shown that the first event received after a FE_SET_FRONTEND ioctl() has frequency == 0, which is considered an error. Reading the documentation for FE_SET_FRONTEND lead me to believe that it would send only a single event once tuning had completed, which is not what the code does. > It's not my code and my patch doesn't create any new event. Those patches don't, no. I was assuming that you were patching code that you had patched earlier. My bad, it seems. > Your example code can't work. You need to call FE_GET_EVENT > or FE_READ_STATUS. And that's why I only called it "pseudocode" :-). > > So I'm going to say "No", your patches don't restore the old behaviour. > > Yes. The patch is restoring a different old behaviour. The > behaviour you're referring to has never been in the kernel. ;-) Yikes! Documentation bug, anyone? Cheers, Chris -- 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
