Bug 20394

In about 50% of all cases, the kernel will fail to recognize the second
IDE Master correctly. Instead of the expected message 

ad0: 4134MB <IBM-DHEA-34331> [8400/16/63] at ata0-master using UDMA33
ad1: 9671MB <IBM-DTTA-351010> [19650/16/63] at ata0-slave using UDMA33
ad2: 14664MB <IBM-DJNA-351520> [29795/16/63] atata1-master using UDMA33
acd0: CDROM <Pioneer CD-ROM ATAPI Model DR-A04S0105> at ata1-slave using WDMA2

it reports as follows:

ad0: 4134MB <IBM-DHEA-34331> [8400/16/63] at ata0-master using UDMA33
ad1: 9671MB <IBM-DTTA-351010> [19650/16/63] at ata0-slave using UDMA33
ad2: 127MB <IBM\^A\M-D\^BNA-\^A5!5 0> [16257/16/1] at ata1-master using PIO4
acd0: CDROM <Pioneer CD-ROM ATAPI Model DR-A04S 0105> at ata1-slave using WDMA2

The ad2 model string can look different from the one above. At other times,
the kernel panics at exactly the ad2 position with a divide-by-zero error.

How-To-Repeat: 
Problem appear at boottime. It seems to happen a lot more often after a reboot
when the OS that ran before the reboot was Windows (no joke).