On Sun, May 22, 2005 at 04:55:48PM -0400, Tyler Trafford wrote: > On 5/22/05, Axel Thimm <[EMAIL PROTECTED]> wrote: > > On Sun, May 22, 2005 at 08:02:37PM +0200, Axel Thimm wrote: > > > could some of you share your timings on loading the driver? I had > > > contradictory input on this (there was a claim on ~10 seconds!), so > > > I'd like to see whether there is some truth on that in certain > > > hardware configs. Please include kernel & ivtv version. > > > > Hm, first feedback gives number 5-8secs, far too much still :/ > > > > As Bryan Mayland pointed out most of this time goes into loading the > > audio firmware. > > > > An (untested) suggestion by Hauppauge engineering is to upload at > > 400Khz. The Mako chip can cope with it, as most probably the tuner. > > > > What will most probably not work well at 400KHz is the IR blaster, > > which will require speeds <= 100KHz. > > > > Optimizing the driver load time is interesting for suspending and > > resuming. > > I also notice that in addition to what "time" gives you, it takes > several more seconds for /dev/video* actually show up if you are using > udev.
I assume that a suspend/resume cycle will not remove/recreate device nodes. Or better said s/assume/hope/ ... :) -- Axel.Thimm at ATrpms.net
pgpRvcVH1LxDA.pgp
Description: PGP signature
