Quoting Billy Biggs ([EMAIL PROTECTED]): > I'd really like to be able to get an accurate timestamp on when the > last VBI occurred. As well, it would be nice to know conclusively when > I've missed a VBI, rather than have to infer it from timestamps. > > Michel Danzer has done a DRM patch for the Radeon cards that gives a > VBI interrupt plus a sequence number. I'm also getting him to put a > timestamp in there. We were trying to determine how best to do the > timestamps: ideally I'd like to use some monotonic clock that we can > have common between video, ALSA, and V4L. This would be best for > PVR-type apps. > > Thoughts on what's best to do? I'm thinking I'll just have some > struct in kernel space with a 64bit timestamp plus a 64bit VBI interrupt > counter, and have it so you can call an FB ioctl to get it. My initial > implementation would be rdtsc timestamps.
Timestamps and vsync counter would be nice. I would like to have an official FB ioctl for this. -- Best regards, Denis Oliver Kropp .------------------------------------------. | DirectFB - Hardware accelerated graphics | | http://www.directfb.org/ | "------------------------------------------" Convergence GmbH -- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe directfb-dev" as subject.
