From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Williams Date: Fri, 18 Apr 2003 02:08:44 +0000 Subject: Re: [RFC] /sbin/hotplug multiplexor Message-Id: List-Id: References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-hotplug@vger.kernel.org Stephen Williams wrote: > david-b@pacbell.net said: > >>How often do you connect a new USB or PCI device? If it takes a full >>second to react, that's OK; and Linux is usually faster than that. > > > When I was running on a 66MHz system w/ 128Meg RAM, it took on > the order of 30 seconds to get around to invoking fxload. The > initial execute of /sbin/hotplug may be quick, but that script > sure takes a while to run. I think performance is an issue. david-b@pacbell.net said: > Systems that are "under-powered" have other tools available, like > having /sbin/hotplug be a custom C program. Hmm, where have I heard > of such a thing before, Greg? :) OK, temper-tantrum follows... A CPU is pegged in my K-WizzyHz workstation doing a Verilog simulation or some image processing, and I hotplug a camera to get more images. Now /sbin/hotplug starts running a bunch of processes, causing a lot of kernel activity to dispatch the hotplug event to umpteen processes, fiddling with a zillion minor configuration files and subscripts until it gets figured out and a kernel module for a USB mass storage device is loaded, all the while thrashing caches and invoking page allocations- deallocations, and disrupting the background task, not to mention the desktop. A ridiculous burden is a ridiculous burden no matter how fast the host system. I want my super-system doing *my* work, not spending barrels of mips and I/O cycles picking its nose. I used my earlier mentioned 66MHz system to receive high resolution 8.5x11" color images, several images per second; yet plugging in a minor USB device was a major disruption for half a minute. Something is very wrong with this picture. Incidentally, with REMOVER support, unplugging that minor device was nearly instantaneous on this "under-powered" machine. So getting the event to a process and dispatching it should be fast. It's been a while since I've been in the guts of the hotplug scripts, but my impression remains that they are in desperate need of diet, even for the normal case. -- Steve Williams "The woods are lovely, dark and deep. steve at icarus.com But I have promises to keep, steve at picturel.com and lines to code before I sleep, http://www.picturel.com And lines to code before I sleep." ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ 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