From: FFADO <ffa...@ff...> - 2012-07-17 06:44:57
|
#242: ntp interferes with ffado ------------------------+--------------------------------------------------- Reporter: ppalmers | Owner: ppalmers Type: bug | Status: new Priority: major | Milestone: FFADO 2.x Component: | Version: FFADO 2.0-rc2 (1.999.42) Resolution: | Keywords: Device_name: all | ------------------------+--------------------------------------------------- Comment (by jwoithe): stefan: I've done some more thinking about all this. Regarding the difference between CLOCK_MONOTONIC and CLOCK_MONOTONIC_RAW, I think I've identified one possible way where the behaviour could be different on your "bad-RTC" PC. On this PC, CLOCK_MONOTONIC and CLOCK_MONOTONIC_RAW clearly do run at different rates, and the difference is significant. In FFADO's timing code, a request to sleep when CLOCK_MONOTONIC_RAW is in use will in fact use CLOCK_MONOTONIC (see [comment:20 comment 20]). For most systems the speed difference between the two clocks will be very small, so the distinction is largely irrelevant. However, on systems like your bad-RTC PC, I can imagine that the speed differences could cause some issues. I'm not convinced it's the sole reason for the negative time steps we've noted in the logs, but it could at least have something to do with the overall stability issues and explain the apparent differences in behaviour we see when the different clocks are used. In light of this I intend to run some more tests myself in the next few days. Hopefully I'll have a fresh revision to test in reasonably short order. -- Ticket URL: <http://subversion.ffado.org/ticket/242#comment:60> FFADO <http://subversion.ffado.org/index.fcgi> Free Firewire Audio Drivers |