Anandtech had a nice chat with Intel's VP and Director of Intel Client PC Operations and Enabling, Steve Smith about the yesterday Intel 6-series chipset SATA bug and it appeares that the bug is caused by a fauly transistor inside the 3Gbps PLL clocking tree.
The problem that plagues the 3Gbps ports however doesn't affect 6Gbps ports as these are connected to a different PLL clocking tree. It appears that the aforementioned transistor leaks current which ultimately leads to a failure of the ports. To make things worse (for Intel), the problem appeared on the B-stepping of the chipsets, which is the one that got shipped to customers.
Intel still keeps telling the world that this is not a full blown recall and calls it "a simple oversight". Quite a costly oversight if you ask us.
You can find more here