Re: [PATCH RFC V2 3/5] jump_label: if a key has already beeninitialized, don't nop it out
From: Steven Rostedt
Date: Thu Oct 06 2011 - 18:20:53 EST
On Thu, 2011-10-06 at 15:10 -0700, Jeremy Fitzhardinge wrote:
> On 10/06/2011 03:06 PM, Steven Rostedt wrote:
> But bear in mind, I'm not using jump-label for tracing. While its
> important for the "disabled" state to be quick, performance of the
> "enabled" state is also important.
Sorry, I'm still thinking jump-label for tracing over.
But that said, having the nop match is the best of both worlds.
I think having a update_jumplabel.c that is just like the
record_mcount.c which modifies the code right after it was compiled is
the best thing to do. That is, have the assembler determine what size
jumps to use and update them to nops right in the object file before
linking. This should be rather trivial to do.
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/