From mboxrd@z Thu Jan 1 00:00:00 1970 From: Greg KH Date: Wed, 03 Dec 2008 14:59:42 +0000 Subject: Re: usb coldplug problem Message-Id: <20081203145942.GA12203@kroah.com> List-Id: References: <403A9FAC.2010904@free.fr> In-Reply-To: <403A9FAC.2010904@free.fr> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-hotplug@vger.kernel.org On Wed, Dec 03, 2008 at 11:13:32AM +0000, Bernd Porr wrote: > I totally take Greg's point that the firmware load should be via a firmware > request from the driver which then in turn triggers the firmware helper in > userspace. When I checked that last time (about a year ago) this was too > buggy and I stuck with the comedi firmware upload which has been running > for about 4 years without probs. Might be a good time to use the firmware > helper. Yes, the comedi firmware upload will be changing sometime in the future to use the proper interfaces. > But in comedi it is so that the command "comedi_config" associates a driver > with a device. This works fine with hotplug. I plug in my usbdux, this > triggers a udev event and then a udev rule can decide which /dev/comedi? is > actaully taking care of the box. If you want to question that mechanism > then comedi itself has to be rewritten. And I'm sure that nobody here in > comedi-land would like to do that. It just works too good. I'm going to be questioning that when we get to looking closer at the comedi/userspace interface to make sure it is sane and matches the rest of the kernel. But that's not going to happen until I get all of the comedi drivers merged into the staging tree so that we can clean everything up properly. thanks, greg k-h