pmdapapi buglets from Coverity scan

[Nathan Scott <nathans@xxxxxxxxxx>]

Hi Lukas,

Coverity is reporting a couple of issues in pmdapapi that I
could use some help with if you could?  It's saying the sts
setting here is dead code...

                /*
                 * This is where we'd see if a requested counter was
                 * "one too many".  We must leave a note for the
                 * function to return an error, but must continue (so
                 * that reactivating other counters is still
                 * attempted).  
                 */
                sts = PM_ERR_VALUE;
                continue;

which at first glance does indeed seem to be the case (it'll
be overwritten on subsequent loop iterations by PAPI_add_event).

papi_store it then says has a loop control issue - apparently
it can only execute the "for (i = 0; i < result->numpmid; i++)"
loop once.  I think this is because the switch always returns a
result for the first PMID, when we should really be continuing
on through all metrics being stored to.

cheers.

--
Nathan