From mboxrd@z Thu Jan 1 00:00:00 1970 From: Greg KH Date: Tue, 13 Jan 2004 20:36:30 +0000 Subject: Re: about split the udev Message-Id: <20040113203630.GB10464@kroah.com> List-Id: References: <3ACA40606221794F80A5670F0AF15F8402D4EE96@PDSMSX403.ccr.corp.intel.com> In-Reply-To: <3ACA40606221794F80A5670F0AF15F8402D4EE96@PDSMSX403.ccr.corp.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-hotplug@vger.kernel.org On Tue, Jan 13, 2004 at 07:05:26PM +0000, Tristan Wibberley wrote: > Greg KH wrote: > > | I am talking about udev splitting into 2 pieces: > | udev - called the same way udev is called today. > | creates a tiny message queue message based on the hotplug > | event and then exits. This can probably be done in under > | 6Kb of binary code. > | udevd - receives the udev message queue messages. Orders them > | in the proper order (based on the SEQ_NUM from the > | hotplug event), and then acts on them. Size should be > | about what we have today for the udev binary 50Kb or so. > | > | We need to do this in order to properly handle out of order hotplug > | events. > > Will that solve the problem? If udevd keeps up with the messages, and > the udev that supplies the second message pre-empts the one that > provides the first, will udevd process the second message because it > doesn't know that there is an earlier one about to come in? > > If udevd waits for contiguous SEQ_NUMs to ensure this doesn't happen, > then a dead/locked-up udev will freeze everything (udevd won't know > which of the events it has should wait until after the unknown one). I > think a different solution is required. No, udev will have to peek at future messages to see if we got them out of order, and re-order them. And a time-out will have to be in place in case we somehow lose one entirely. thanks, greg k-h ------------------------------------------------------- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html _______________________________________________ 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