Andrew Morton wrote: > Ben Greear wrote: > >> SCSI device sda: 156301488 512-byte hdwr sectors (80026 MB) >> SCSI device sda: drive cache: write back >> SCSI device sda: 156301488 512-byte hdwr sectors (80026 MB) >> SCSI device sda: drive cache: write back >> sda: sda1 sda2 sda3 >> Attached scsi disk sda at scsi0, channel 0, id 0, lun 0 >> Unable to handle kernel paging request at virtual address e09a68e8 >> printing eip: >> c014976e >> *pde = 0151f067 >> Oops: 0000 [#1] >> PREEMPT SMP >> Modules linked in: ext3 jbd ata_piix libata sd_mod scsi_mod >> CPU: 1 >> EIP: 0060:[] Not tainted VLI >> EFLAGS: 00010006 (2.6.11-rc4) >> EIP is at cache_alloc_refill+0x20e/0x240 > > > I assume something has wrecked the slab caches. Please enable > CONFIG_DEBUG_SLAB and CONFIG_DEBUG_PAGEALLOC. Being in a rush, I had to return this system for a replacement system. Since this machine also had issues saving it's BIOS settings, the problem may be hardware related. (We did replace the RAM but that did NOT fix the BIOS problem, at least...) Continuing a bad weekend, the replacment system (Shuttle SB61G2) with same CPU, HD, CDROM, floppy and the new stick of RAM, also had issues. It did boot the various 2.6 kernels without problem, but when I put a 4-port tulip (DFE570tx) NIC in the system, and started traffic on all four ports, the system locked up within about 5 seconds. Interestingly, with HT enabled, I would get partial freeze for a bit, and then total freeze. I assume one of the virtual CPUs was locked and the other one managed to run a bit longer untill it also hit a deadlock. With HT disabled, when it locked, it locked completely at once. With HT, often the mouse worked much longer than the keyboard. I assume some problem with interrupts. Booting with noapic didn't help. I tried 3 different 4-port tulip NICs (two different manufacturers), same behaviour. In a final bit of desperation, I installed RedHat 9 with a custom 2.4.27 kernel. Wouldn't you know it: The system has been rock solid ever since. This system will probably be shipped out on Monday, so I can't offer much debugging info here either. I am attaching the lspci and dmesg output (from the 2.4.27 kernel, RedHat 9 install) in case someone else hits similar problems and a pattern emerges... Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com