On Mon, Sep 14, 2009 at 02:31:16PM -0400, Steven Rostedt wrote: > On Mon, 2009-09-14 at 20:23 +0200, Peter Zijlstra wrote: > > On Mon, 2009-09-14 at 14:17 -0400, Steven Rostedt wrote: > > > Frederic, how big can one of those events get. The ring buffer (and > > > TRACE_EVENT) allow up to almost a page size, which is very hefty for the > > > stack. This code needs to either be rewritten or we need to set a limit > > > to the size of a profile entry. > > > > Yeah, that needs to get a re-write.. I've complained about this when it > > went in. > > One answer is to create a per cpu buffer that is big enough to hold the > data needed. Then you can disable interrupts an use it without worry. > > If you need to also handle NMIs, then create a per_cpu NMI buffer too, > and use that if "in_nmi()" is true. > > -- Steve Looks like a nice idea. Peter, does that sound acceptable to you to disable interrupts during a profiled tracepoint event? -- To unsubscribe from this list: send the line "unsubscribe linux-next" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
