Re: [PATCH 2/2] perf: Drop the skip argument fromperf_arch_fetch_regs_caller

From: Frederic Weisbecker
Date: Fri May 21 2010 - 06:29:27 EST


On Fri, May 21, 2010 at 12:24:37PM +0200, Peter Zijlstra wrote:
> On Fri, 2010-05-21 at 10:11 +0200, Frederic Weisbecker wrote:
> > +static inline void perf_fetch_caller_regs(struct pt_regs *regs)
> > {
> > - unsigned long ip;
> > -
> > memset(regs, 0, sizeof(*regs));
>
> btw, do we really need that memset?
>
> As long as we don't actually copy out to userspace, it really doesn't
> matter what is in there, and we should be setting all relevant registers
> anyway.


We probably need to decouple that yeah.

- If we don't do callchains, we don't need to deref bp
- If we don't copy to userspace, we don't need to zeroe

We should do the filling much later for tracepoints in fact,
but that doesn't concern syscalls or kprobes.

But well, I'll iterate that step by step.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/