K.R. (Randy) Lewis, We were able to confirm (today) that it is the /MP kernel that has the issue defined below. The single CPU build ran the huge file copy just fine. R Re-post in misc@ (sorry for posting wrongly before) ----------------- Encountered the following: 4.3 i386 MP kernel w/Raidframe - throws a: splassert: uvm_map_p: want 1 have 1 Here's what was going

We took a USB 250GB drive and wiped it (unallocated partition) using partition magic on one of those, what do you call it....oh yeah...WindowsXP thingys... (expedient, ok?) Then plugged it into a 4.3 OpenBSD MacMini, did a disklabel and established a partition sd0i of type MSDOS. Then did a newfs t msdos sd0i. All of this works great, each time we do it. Then mounted it on the MacMini

Matthias Scheler <tron@NetBSD.org> writes: On Fri, Jul 27, 2007 at 06:29:30PM +0200, Jarle Greipsland wrote: There seems to be a problem with these changes. With an older kernel (4.99.20 from mid-June) the dmesg part related to the mpt device is: [...] mpt0 at pci2 dev 1 function 0: Symbios Logic FC909A mpt0: interrupting at ioapic0 pin 19 (irq 5) scsibus3 at mpt0