From: Илья <il...@ma...> - 2013-08-27 20:25:26
|
Вторник, 27 августа 2013, 23:11 +09:30 от Jonathan Woithe <jw...@ju...>: >Hi all Hello again! >In the meantime, could you please check out r2384 or later and test with >your hardware? The kernel should no longer freeze, which will hopefully >allow us to continue working through the remaining issues. >Ok. I think I've found the problem (or at least the most significant >problem). As per my more detailed email in the "jackd+ffado causes kernel >lock ups" thread on ffado-devel, a bug was introduced in r2379 which appears >to overwhelm the kernel driver with async packets. I have committed r2384 >as an interim workaround; this disables the offending call so ffado should >now now longer freeze the kernel when run with an RME Fireface interface. I >will be committing a proper fix in due course. I have done some tests today. It was a tests with old ubuntu 12.04 with kenel ver. 3.2. I test a r2374 (revision before changes) and latest revision. I can report that it was 2384! Also I tested my new VIA controller in 12.04. I can report that there is no one kernel lock ups occurred. Whole logs here: rghost.net/download/48415386/38b636593a53498d7b7cf027962b65f5e2156683/rme%20800%20ubuntu%2012.04.zip for professional analysis. I can say that jackd do not started correctly, but it produce more interesting messages and do not locked up practically. >This interim fix will cause some inconsistent behaviour with the sample rate >if autosync clock mode is active. Until I implement the full fix it would >be best to restrict testing to master clock mode (or be prepared to >workaround whatever issues crop up). All this tests was done in master clock mode. >Regards > jonathan |