Date: Tue, 23 Mar 2004 20:15:25 -0500 From: jason <jason@ec.rr.com> To: Eirik Oeverby <ltning@anduin.net> Cc: freebsd-current@freebsd.org Subject: Re: sound issues in -CURRENT Message-ID: <4060E12D.7070402@ec.rr.com> In-Reply-To: <40601BDF.8040708@anduin.net> References: <20040309070524.40ad96f5.ryan@slipgate.org> <20040312232251.48c0c57d.ryan@slipgate.org> <405FED9B.3050106@anduin.net> <405FF4B9.30905@DeepCore.dk> <1080034989.10553.12.camel@pav.hide.vol.cz> <40601BDF.8040708@anduin.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Eirik Oeverby wrote: > Hoi, > > I have all debugging and witness stuff disabled, though I am using > SCHED_ULE. Apart from that, the problem seems to be somewhat less > evident when playing from the network (no playback-related disk i/o), > but it only reduces the problem - it does not eliminate it. > > As I have stated earlier, the problem gets worse over time. After a > clean reboot (even though all the apps I use are started and running), > the problem shows up once every few minutes. Now, after a couple of > days uptime, it happens a few times per minute - enough to annoy me > ever so slightly when trying to tap the rhytm with my foot ;) > > My HDD is running in ATA-100 mode (no PIO) and is reasonably fast. > FreeBSD 4.9 works perfectly on the same machine, with all the same > software. > > /Eirik > > Pav Lucistnik wrote: > >> V út, 23. 03. 2004 v 09:26, Sřren Schmidt píše: >> >> >>>> Heavy disk I/O seems to be a trigger for the problem. I always hear >>>> the >>>> screwups, but i.e. when building mozilla or openoffice (during the >>>> extract phase with heavy I/O), sound almost slows down to a complete >>>> stop, stutters and hicks and cracks all the time - to a point where >>>> the >>>> music is almost impossible to even identify anymore. >>> >> >> >>> I think you are suffering from the high interrupt latency in -current, >>> this combined with a soundcard with a small buffer would easily lead >>> to your problems. A slow disk or PIO mode could also be troublesome >>> in some cases. You do have witness etc turned off right ? >> >> >> >> I hear this too, on both my cards (Soundblaster 128 and onboard AC97 >> codec). I have all possible debugging turned off. Heavy disc activity >> like untarring triggers the symptoms, which sounds like buffer underruns >> on the soundcard. Almost sounds like a new ATA stack is really effective >> (leaving little CPU cycles for feeding data to soundcard). >> > _______________________________________________ > freebsd-current@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to > "freebsd-current-unsubscribe@freebsd.org" > I have had this problem before myself. It would happen to me under load, but not that often. Usaully I'd be cvsuping, streaming or playing music, and I do a lot of compiling because I like to keep things up to date. I think I may have first noticied it on current before 5.2.1 with the switch to sched_ule. I have since added the patch to ule that had been cirrculating on the current mailing list before it was commited. It may have speed some things up, but I have yet to notice any skipping or sound problems in xmms. I will but some load on the system while streaming, but a week or less without problems is too soon to tell. At least I think it might be. Jason
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4060E12D.7070402>