During CD install of Florence betas 1 & 2 on my Dell PowerEdge 2400, the following oops was printed prior to starting the GUI. Installation continued successfully, though one module (shown in /proc/modules) was shown to be stuck in an initializing state (I think it was raid1, but don't hold me to that). Attached dmesg and decoded oops are from beta 2. raid0 personality registered raid1 personality registered raid5: measuring checksumming speed 8regs : 1537.716 MB/sec 32regs : 867.537 MB/sec invalid operand: 0000 CPU: 0 EIP: 0010:[<f889ab27>] EFLAGS: 00010206 eax: f6eddeac ebx: f6ec2f40 ecx: 8005003b edx: 0000000f esi: f6ec0000 edi: 00000004 ebp: 00000000 esp: f6eddeac ds: 0018 es: 0018 ss: 0018 Process loader (pid: 51, stackpage=f6edd000) Stack: 00000001 00001e3d 00001eaa c023b4e3 00000001 c0115d12 00000282 00000001 c023b4e3 00000020 000003e8 00000f1e ffffffff 000008e5 f889bcb8 f889c16b 00000000 00000f1f f889bc52 00000f40 f6ec0000 f6ec2f40 f6ec2f40 f6ec0000 Call Trace: [<c0115d12>] [<f889bcb8>] [<f889c16b>] [<f889bc52>] [<f889bd32>] [<f889c35c>] [<f889a000>] [<c01168b1>] [<f8895000>] [<f889a060>] [<c0108ed3>] Code: 0f 11 00 0f 11 48 10 0f 11 50 20 0f 11 58 30 0f 18 4e 00 0f raid5 personality registered
Created attachment 7525 [details] dmesg.txt
Created attachment 7526 [details] oops-decoded.txt
AFAIK it's a known problem and fixed in kernel 2.4.0 final.
Re-assigning to Michael Johnson
This defect is considered MUST-FIX for Florence Gold release
It's not fixed in 2.4.0-final, but we do have a patch to fix it in our current source tree and it will be in our next kernel.
Yes, confirmed fixed in our internal kernels.