On Thu, 26 Oct 2017, Andy Lutomirski wrote: > > Here's the first "real" version of this series. It's based on Linus' > tree from a couple days ago.
Not seeing any problems in this series. I was hoping it would remove more lines of code than it adds, and it's the reverse, but that seems to be mostly the new push/pop stuff. I'm taking it we couldn't just replace the 'mov' instructions in the save/restore macros.. I *do* ask that when we have code like this: POP_EXTRA_REGS POP_C_REGS addq $8, %rsp that "addq $8" would have a comment about what stack entry we're skipping. I *think* it's 'orig_eax', but wouldn't it be nice to say so? Linus