From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kay Sievers Date: Wed, 18 Feb 2004 08:27:19 +0000 Subject: Re: [PATCH] closing the loop -- retries and acks Message-Id: <20040218082719.GA9050@vrfy.org> List-Id: References: <403305A3.2070604@sympatico.ca> In-Reply-To: <403305A3.2070604@sympatico.ca> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-hotplug@vger.kernel.org On Wed, Feb 18, 2004 at 01:26:43AM -0500, Chris Friesen wrote: > This patch reduces the already low chances of lost messages. > > Once udevsend successfully sends a message, it waits with a timeout for > a response. If the response is good, then all is well, otherwise it > sends the message again. > > In udevd, when we receive a request we store the address of the > requestor, and then once the request has been handled we then send an > ack back so that they know they have been taken care of. > > This patch takes care of the basics, but there are still a few things left: > 1) tuning timeouts and retry numbers > 2) capabilities check on the ack > 3) pull out the udevd-specific stuff from struct hotplug_msg so that we > don't waste time/memory. udevd will have a different struct, one > element of which is hotplug_msg. udevsend still only knows about > hotplug_msg > 4) pull the messaging stuff into a set of headers or something to get > rid of code duplication > > Comments? It's sweet. But isn't the case of handling a lost a message purely theroretical? What problem we actually have is solved with this infrastructure? A possibly crashing or exiting udevd in the time window from accepting our message until its execution? Is the handling of this special case really needed? Oh, and we don't handle a crashing udev :) We have a remaining risk all over the place and it seems that we try to look better than the whole picture. We have failure paths with much much higher propability. So, it's nice, sure. But is it worth the extra complexity? thanks, Kay ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id56&alloc_id438&op=click _______________________________________________ Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net Linux-hotplug-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel