On Sun, Jul 24, 2011 at 8:57 AM, Mauro Carvalho Chehab <mchehab@xxxxxxxxxx> wrote: > I proposed the same fix sometime ago, when Gerd reported this issue > for me. His feedback was that this partially fixed the issue, but > he reported that he also needed to increase the set_power_mode delay > from 5 to 50 ms in order to fully initialize the cx231xx hardware, > as on the enclosed patch. It seems he tested with vanilla Fedora kernel. > > So, I suspect that HZ may be affecting this driver as well. As you know, > if HZ is configured with 100, the minimum msleep() delay will be 10. > so, instead of waiting for 5ms, it will wait for 10ms for the device > to powerup. > > It would be great to configure HZ with 1000 and see the differences with > and without Gerd's patch. > > Cheers, > Mauro. I don't dispute the possibility that there is some *other* bug that effects users who have some other value for HZ, but neither I nor the other use saw it. Without this patch though, the device is broken for *everybody*. I would suggest checking in this patch, and separately the HZ issue can be investigated. I'll see if I can find some cycles today to reconfigure my kernel with a different HZ. Will also check the datasheets and see if Conexant documented any sort of time for power ramping. It's not uncommon for such documentation to include a diagram showing timing for power up. Devin -- Devin J. Heitmueller - Kernel Labs http://www.kernellabs.com -- 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
