From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sasha Khapyorsky Subject: Re: [PATCHv3] opensm: Better handling of non responsive SMAs Date: Wed, 9 Jun 2010 22:37:07 +0300 Message-ID: <20100609193707.GK20172@me> References: <20100609182911.GA26769@comcast.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20100609182911.GA26769-Wuw85uim5zDR7s880joybQ@public.gmane.org> Sender: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Hal Rosenstock Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-rdma@vger.kernel.org On 14:29 Wed 09 Jun , Hal Rosenstock wrote: > > In order to better handle non responsive SMAs (when link is physically up > but the SMA does not respond), a timeout based mechanism for SMPs is added > to better enable forward progress in a more timely fashion. So rather than > waiting for timeouts and outstanding wire SMPs to drop below some configured > value, there is also an additional limit introduced for transaction based SMPs. > These timeout based SMPs are capped at a configured maximum value. > > Two new options are added for this: > max_wire_smps2 indicates the second (higher) limit of timeout based > SMPs supported. When this limit is reached, timeout based SMPs > are not longer sent (until the number of outstanding ones drops below > this limit). > max_smps_timeout indicates the number of microseconds used for the > timeout in between sending SMPs when the oustanding wire count is above > max_wire_smps and below max_wire_smps2. It defaults to the transaction > timeout times the transaction retries. > > The timeout based SMP mechanism can be disabled by setting max_wire_smps2 > to the same value as max_wire_smps. This is equivalent to the (current) > algorithm prior to this change. By default, this mechanism is disabled. > > Signed-off-by: Hal Rosenstock Applied. Thanks. Sasha -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html