Messages in this thread | | | Date | Tue, 15 Apr 2008 10:21:29 -0400 | From | "Frank Ch. Eigler" <> | Subject | Re: [RFC PATCH 1/2] Marker probes in futex.c |
| |
Hi -
On Tue, Apr 15, 2008 at 04:04:30PM +0200, Ingo Molnar wrote: > [...] > > I have not seen any counter argument for the in-depth analysis of the > > instruction cache impact of the optimized markers I've done. Arguing > > that the markers are "bloated" based only on "size kernel/sched.o" > > output is a bit misleading. > > uhm, i'm not sure what you mean - how else would you quantify bloat than > to look at the size of the affected subsystem?
For example, by measuring how much of that "bloat" is actually experienced (e.g., in the form of presence in cache or additional time taken.) when that instrumentation is turned off vs. on. That should sound familiar, as this is the sort of data that you outlined (but I don't recall whether you actually provided numbers/recipes) when describing the "dyn-ftrace" nop-padding work.
- FChE
| |