Subject: Re: [ecasound] warning about system freezes
From: martijn (mlina_AT_dds.nl)
Date: Mon Oct 08 2001 - 22:58:48 EEST
Once upon a Mon, Oct 08 2001, Kai Vehmanen hit keys in the following order:
>
> Are you sure - 2.1.0 was released just yesterday? ;)
heh, since the CVS version showed me version 2.1.0, i figure i have ran into
that same bug. not sure though, i've been running without problems since i
clocked my CPU to it's normal speed.
> Have the crashes happened when running ecasound as a non-root
> user? If so, it's probably a driver issue. Bugs in drivers
> are probably the most common causes for kernel
> crashes.
No always record with root permissions. For playback i don't, but maybe i
should try and recreate the problem. my last CVS checkout was on Sept 26,
i'll take a look at that version.
> The ecasound 2.1.0 crash is not a real kernel crash.
> For instance, it's possible to ping the frozen machine.
> So the network stack is working, which again means
> that the kernel is running. It's just that ecasound
> threads are in a deadlock (or actually a livelock)
> and consume 100% of user-space resources (due to
> SCHED_FIFO scheduling, which is much like running
> apps under MS-DOS :)).
ah, yeah of course. realtime priority, i forgot. since linux 2.4.10 i can
finally let my power button shutdown my system cleanly, so i won't suffer too
much from this dead-or-alive lock. i've never done much debugging, only
looked at core dumps with gdb, to get some hints which function caused the
crash. i don't know if gdb is useful in debuging a realtime process. could
you give me a few hints? o, wait i'll first make sure i can reproduce the
'crash' ;)
martijn
-- To unsubscribe send message 'unsubscribe' in the body of the message to <ecasound-list-request_AT_wakkanet.fi>.
This archive was generated by hypermail 2b28 : Mon Oct 08 2001 - 22:44:23 EEST