On Mon, 2009-02-23 at 17:41 -0500, Theodore Tso wrote: > On Mon, Feb 23, 2009 at 11:16:59PM +0100, Peter Zijlstra wrote: > > > > We're working on adding arguments to the function/graph tracer, it would > > fit all your above requirements and doesn't need any source modification > > to boot. > > *Excellent*. I would love to have that funtionality. How do you plan > to make available complex data structures (i.e., suppose I want > inode->i_ino printed out)? I assume this will require writing some > "easy to generate" glue code that would presumably be some kind of > kernel module? That doesn't bother me (after all that's what > SystemTap does), as long as generation of the glue code can be largely > automated ---- so that you can take something approximately like a > DTrace or SystemTap script, and with some perl or python helper, > translate it into glue code that gets compiled into a kernel module. > Is something like that what you have in mind? Yeah, we were also looking at using sparse and term rewrite systems on top of the regular C parse tree to generate stuff. -- To unsubscribe from this list: send the line "unsubscribe linux-kbuild" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
