I have an application that slaves UNIX time to GPS and has been working fine for years on various OS's such as Linux, Solaris, and IRIX. It is currenlty working fine with RedHat kernel release 2.6.18. However, when I ported it to RedHat kernel release 2.6.24, it exhibits an unstable behavior which manifests as a "ringing" about the intended set point. Extensive debugging tells me that the problem is that the kernel is slow to respond to changes commanded using the adjtimex call. It is slow to initiate the requested change and slow to stop it, resulting in both under and overshooting of the intended correction. The application has a cycle time of one second where it wakes up and checks the time delta and (optionally) the status of a previously commanded correction. As a result of this ringing, it is issuing a correction nearly every cycle and the UNIX/GPS time delta is ranging between +/- 100 microseconds rather than the intended +/- 20 microseconds. Does anyone know what could be causing the apparent slow kernel response, or what other possible cause it could be?
Thanks, Jesse --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Linux Users Group. To post a message, send email to [email protected] To unsubscribe, send email to [EMAIL PROTECTED] For more options, visit our group at http://groups.google.com/group/linuxusersgroup -~----------~----~----~----~------~----~------~--~---
