[patch v6 21/21] record-btrace: add (reverse-)stepping support

Metzger, Markus T markus.t.metzger@intel.com
Wed Nov 6 15:06:00 GMT 2013


> -----Original Message-----
> From: gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] On Behalf Of Jan Kratochvil
> Sent: Sunday, October 06, 2013 9:52 PM


> >  @item btrace
> > -Hardware-supported instruction recording.  This method does not allow
> > -replaying and reverse execution.
> > +Hardware-supported instruction recording.  This method does not record
> > +data.  It allows limited replay and reverse execution.
> 
> As discussed in
> 	Message-ID:
> <A78C989F6D9628469189715575E55B230A9D4934@IRSMSX104.ger.corp.intel.
> com>
> 
> the documentation should mention the full history is usually not kept as the
> btrace buffer commonly overflows.  Otherwise user may be curious why the
> btrace history is always relatively short.

I added a sentence to describe this in the documentation.


> I would see it worth a comment even in
> 	(gdb) help record btrace
> 	Start branch trace recording.

I don't think that this is necessary or helpful.  If we start documenting the
trace size here, we should also say that trace is collected per thread with
no correlation between threads and that tracing induces a run-time overhead
and that it may also affect thread interleaving which may cause bug symptoms
to disappear, and ....


Regards,
Markus.

Intel GmbH
Dornacher Strasse 1
85622 Feldkirchen/Muenchen, Deutschland
Sitz der Gesellschaft: Feldkirchen bei Muenchen
Geschaeftsfuehrer: Christian Lamprechter, Hannes Schwaderer, Douglas Lusk
Registergericht: Muenchen HRB 47456
Ust.-IdNr./VAT Registration No.: DE129385895
Citibank Frankfurt a.M. (BLZ 502 109 00) 600119052



More information about the Gdb-patches mailing list