Re: [RFC] New markers description without format strings

* Frank Ch. Eigler ([email protected]) wrote:
> Hi -
> 
> On Mon, May 07, 2007 at 10:02:24AM -0400, Mathieu Desnoyers wrote:
> > I'll skip the implementation details, but I am considering changing the
> > current marker implementation, which uses format strings, for something
> > that would look like :
> > 
> > trace_mark(subsysname, eventname, mark_types(MARK_CHAR, MARK_ULLONG),
> >            mychar, myull);
> 
> Clever, but it is starting to look too wordy, and if I squint just
> enough, it looks close to the old "decorate the function name with the
> parameter types" technique.
> 

The difference is that I don't have to generate every combination of
function names, which would result in a huge header file.

> > [...] Moreover, this should be more efficient to do a switch() on
> > the enum members than to parse the format string.
> 
> FWIW, systemtap would not need to parse either the string or the type
> tag array at run time (but rather at translate time), so the
> performance difference does not affect it.
> 

As long as systemtap does not use the arguments passed as parameter, I
can understand this, but how does it expect to "know" what parameters a
specific marker receives ?


> - FChE
> 

-- 
Mathieu Desnoyers
Computer Engineering Ph.D. Student, Ecole Polytechnique de Montreal
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68