Re: [PATCH] remove LOCK_SECTION from x86_64 spin_lock asm

[Andi Kleen <ak@suse.de>]

On Thu, Sep 16, 2004 at 09:44:31AM +0200, Ingo Molnar wrote:
> it only works on pointer less kernels because the spinlock profile 
> unwinding is _conditional_ on an FP kernel right now:
> 
>  #if defined(CONFIG_SMP) && defined(CONFIG_FRAME_POINTER)
>  unsigned long profile_pc(struct pt_regs *regs)
>  {
>  ...
> 
> on non-FP kernels you'll see all the overhead in the single spin_lock()
> function, agreed?

Agreed.  Ok, should be easy to fix with the same scheme, except
that it has to use offset 0 instead of 1 for the !FP case.
I will cook up a patch.

BTW the SMP check is also wrong because preemptive kernels also
need this unwinding, although they should not spend too much
time in locking.

> > > in this respect - it might work if you can detect for sure at build time
> > > whether there's any local variable. Tricks like this really tend to
> > > haunt us later.
> > 
> > There are already lots of such assumptions in the kernel (e.g. WCHAN
> > and others).  I don't think adding one more is a big issue.
> 
> wchan has only one assumption: that that all __sched section functions

It's more than WCHAN, e.g. sysrq-t assumes it too. 

-Andi