On Thu, 19 Feb 2004, Simon Schuetz wrote: > > b) use a real SMP host? > I'm running the test on AMD K6 with 300 and 350 MHz. > Might be that they have no SMP support. Just to clarify, the 300 MHz and 350 MHz are two separate hosts? Or do you have multiple CPUs in one host (=SMP machine)? Btw, did the crash produce any syslog data (see /var/log/*log)? Can you try to copy-paste the crash message to ksymoops like this: on_the_host_that_crashed % ksymoops<enter> unable to handle kernel paging request at address C0000010 Oops: 0002 EIP: 0010:XXXXXXXX eax: xxxxxxxx ebx: xxxxxxxx ecx: xxxxxxxx edx: xxxxxxxx esi: xxxxxxxx edi: xxxxxxxx ebp: xxxxxxxx ds: xxxx es: xxxx fs: xxxx gs: xxxx Pid: xx, process nr: xx xx xx xx xx xx xx xx xx xx xx <ctrl+d> > Is SMP necessary or do I have to recompile without SMP support. I recommend compiling without SMP support because it has not been tested. Try without SMP support and tell us how it went. If it works without SMP support, we definitely have a bug and know where to trace it. We don't have two real HIP enabled hosts at my hands right at the moment, so we cannot test with SMP support. The problem is not guaranteed to reproduce on virtual hosts, but we can try test it with virtual hosts and later on with real hosts. -- Miika Komu miika@xxxxxx http://www.iki.fi/miika/