Re: kernel panics on sio interrupt-level overflows

2005-04-04 Thread Kris Kennaway
On Mon, Apr 04, 2005 at 10:35:11AM +0200, Oleg Tarasov wrote: > Hello, > > Kris Kennaway <[EMAIL PROTECTED]> wrote: > > > > Does enabling WITNESS support (and possibly excluding > > WITNESS_SKIPSPIN) produce any diagnostics? > > > Kris > > I can enable WITNESS, but I don't know where to find

Re: kernel panics on sio interrupt-level overflows

2005-04-04 Thread Oleg Tarasov
Hello, Kris Kennaway <[EMAIL PROTECTED]> wrote: > Does enabling WITNESS support (and possibly excluding > WITNESS_SKIPSPIN) produce any diagnostics? > Kris I can enable WITNESS, but I don't know where to find that useful information I can provide. My server's uptime is 8 days without mpsafen

Re: kernel panics on sio interrupt-level overflows

2005-03-29 Thread Kris Kennaway
On Tue, Mar 29, 2005 at 12:09:19PM +0200, Oleg Tarasov wrote: > Hello, > > This problem had unusual solution. It seems the key to this problem is > that FreeBSD 5.3-STABLE-p5 (I have updated my system) has DEADLOCKS > when intensively using sio device on high speed (115200) with SMP support. > It

Re: kernel panics on sio interrupt-level overflows

2005-03-29 Thread Oleg Tarasov
Hello, This problem had unusual solution. It seems the key to this problem is that FreeBSD 5.3-STABLE-p5 (I have updated my system) has DEADLOCKS when intensively using sio device on high speed (115200) with SMP support. It seems that MPSAFE causes those deadlocks. Turning debug.mpsafenet="0" in