Applied Thanks for your work on this massive rework.
On Fri, Aug 7, 2009 at 11:42 PM, Corey Ashford<cjash...@linux.vnet.ibm.com> wrote: > Hello, > > Looking at the code in pfmlib_gen_powerpc.c again, I realized that during > the change to consolidate the Power arch chips code to use common > structures, and get rid of switch statements, I made a mistake with the > group_vector and event_count arrays. > > I had made the incorrect assumption that the #defines for PFMLIB_POWER4_PMU > (and similar) had started at 0 and went to 7, but in reality, these numbers > start at 110 and go to 117. The result is that arrays are created which > have entries 0..109 unused, wasting space. The attached patch offsets the > indexes by FIRST_POWER_PMU which is set to the value of PFMLIB_PPC970_PMU > (110). > > Also, as per a recent discussion about the use of malloc() in the dispatch > call, I decided to just eliminate the call and use C99's ability to do > dynamic memory allocation on the stack. So this patch undoes the recent > change that checks the return value of malloc(), and the single exit point > which was used to call free(). Since free is no longer needed, I went back > to using multiple exit points, because I think the code is more clear that > way. > > Also fixed one place where the indentation inside a loop was wrong. > > None of these changes are necessary. The code works as is, so this is just > a clean-up and optimization. > > Tested on Power5. > > Signed-off-by: Corey Ashford <cjash...@us.ibm.com> > > ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ perfmon2-devel mailing list perfmon2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/perfmon2-devel