On Thu, Oct 9, 2008 at 12:57 PM, Ira Weiny <[EMAIL PROTECTED]> wrote: > On Thu, 09 Oct 2008 09:17:39 -0700 > Al Chu <[EMAIL PROTECTED]> wrote: > >> Hey Hal, >> >> On Wed, 2008-10-08 at 21:54 -0400, Hal Rosenstock wrote: >> > Hi Al, >> > >> > On Wed, Oct 8, 2008 at 8:40 PM, Al Chu <[EMAIL PROTECTED]> wrote: >> > > >> > > Note that these scripts specify -R (reset only) for perfquery, so there >> > > aren't any issues with multiple perfquery outputs that may need to be >> > > parsed multiple times/differently. >> > >> > But this takes many more resets to work, right ? Why not have loop >> > ports (for reset) only do this is all ports is not supported ? >> >> Do you mean remove the --loop_ports option and loop the ports by default >> even if AllPortSelect isn't supported? >> >> I suppose that'd be fine. The reason I added an option was I just >> didn't want to change default behavior. But if everyone is happy just >> making it happen by default, I'm ok with that as well. >> > > I vote yes. Have the loop ports happen "under the covers" if the user > specifies '-a' this includes summing the values just like AllPortSelect would > in the switch. If the user specifies "0xff" I think that would be a good way > to request specific functionality and (as per Al's other email) return an > error > if AllPortSelect is not supported.
I think this neglects whether a single response to the user is intended, This would cause one response per port on a read. -- Hal > Ira > > _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
