On Thu, 21 Mar 2019 20:31:52 +0100
Peter Zijlstra <pet...@infradead.org> wrote:

> > 
> > No I didn't. Some users only care about performance, but find memory
> > cheap.  
> 
> Because cache-misses are free?

If I ever did implement this, I would try to get all the data out of
line as much as possible, where only a nop would be inserted:

        jmp lockdep_code
        raw_locking
1:

[..]

lockdep_code:
        do all that lockdep needs
        raw_locking
        jmp 1b

and have that converted to:


        nop // jmp lockdep_code
        raw_locking
1:

when disabled. So the only extra pressure on the icache is the nop.

-- Steve

        

Reply via email to