On Mon, Nov 21, 2011 at 4:48 PM, David Daney <ddaney@xxxxxxxxxxxxxxxxxx> wrote: > > Really Linus, did you read the other half of the message you quoted? Sorry no. It was just an automatic reaction along the lines of "oh christ, is this argument still going on"? I'm ok with magic gcc function attributes, although I still don't see why you need that symbol so badly. Quite frankly, some MIPS-only patch is *not* important-enough to worry about this. I'd rather have the simpler "that symbol doesn't exist, and MIPS just needs an #ifdef". I realize that we *used* to have code that did that if (something_that_evaluates_to_zero) { .. use HPAGE_MASK .. but considering that we don't have that any more and people are happy, I'm not all that convinced that we need to try to re-introduce it. Linus