[llvm] r190773 - Implement function prefix data as an IR feature.
peter at pcc.me.uk
Mon Sep 23 13:23:48 PDT 2013
On Mon, Sep 23, 2013 at 01:14:05PM -0700, Richard Mitton wrote:
> I have a small question about this -
> Why does the prefix data have to begin with a jump?
> Couldn't you just make the entry point label appear after the prefix
> data? And then use negative offsets to access it given the function's
> address? It seems like the jump adds a performance hit which isn't
> .word 0x1234 // prefix data
> push rbp
> // etc
I answered this question here:
In general I hope to include an FAQ in the blog post which will answer
these sorts of questions.
More information about the llvm-commits