I could give you 2 extremes. The shortest lead time is a product we can build in our fabs and we can fully assemble and test it in our factory and ship it. And on a product where there are thousands of customers, hundreds and thousands of customers, a very broad, high-volume product with just lots and lots of customers buying it. Because on that product, we can build it inside plus we can build inventory because it has a broad usage. That will have the shortest lead time and most likely on the shelf. You go to the other extreme, and you take a product that runs at a foundry, gets probe, assembled and tested outside, hasn't been brought it yet either as SMSC product or could be one of our product that runs outside. And if you add on the top of that, especially a product that has a very narrow customer base, 1 or 2 customers buy it, and it doesn't really have a broad-based usage in consumer, industrial communication, PC, a lot of Microchip products, then we cannot build a lot of inventory because the inventory can go obsolete. And some of that product is largely distant -- built on specific customer demand, either on order or understanding with the customers. So that will have the longest lead time. Those 2 are the extremes. And you have -- and that extreme could be 18 weeks. The lowest that is built inside is off-the-shelf. And then, there are products all over the place. That are maybe fab-ed inside, tested outside or fab outside, tested inside, small customer base, large customer base, all over the place. We sell over 100,000 SKUs.
Christopher B. Danely - JP Morgan Chase & Co, Research Division: And then another, I guess, my follow-up question, is just on cash and cash management. So as you guys have talked about and as you're demonstrating, is that some of the best margins and some of the highest dividend yields in your space. Is it -- probably, you have your cash and growth continues to outpace dividend growth? And you also have a little bit of share count creep over the last several quarters. So I'm just wondering, do you think about having some sort of regular token buyback? Or do you think about maybe increasing the dividend a little more? And then if you could also address why you increased the revolver to, I mean to $2 billion up from $750 million?