On Wed, 25 Oct 2017 14:22:58 +0200 Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote: > On Tue, Oct 24, 2017 at 07:39:02PM +0200, Christian Gromm wrote: > > On Fri, 20 Oct 2017 16:56:10 +0200 > > Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote: > > > > > On Fri, Oct 20, 2017 at 04:20:33PM +0200, Christian Gromm wrote: > > > > On 18.10.2017 16:29, Greg KH wrote: > > > > > On Wed, Oct 18, 2017 at 04:02:33PM +0200, Christian Gromm > > > > > wrote: > > > > > > On 18.10.2017 14:12, Greg KH wrote: > > > > > > > On Mon, Oct 16, 2017 at 10:46:09AM +0200, Christian Gromm > > > > > > > wrote: > > > > > > > > The following patch adapts the driver to use the device > > > > > > > > model by: > > > > > > > > > > > > > > > > - adopting the MOST bus_type > > > > > > > > - registering the core as a busdriver > > > > > > > > - removing private kobject/kset usage > > > > > > > > - removing private lists and structures to track > > > > > > > > registered modules and making use of the device model > > > > > > > > API > > > > > > > > - removing prefix of modules > > > > > > > > - allowing adapter drivers (a.k.a. HDM) to > > > > > > > > register MOST devices > > > > > > > > - registering AIM modules as components with the > > > > > > > > core to build the user space experience of the driver > > > > > > > > stack > > > > > > > > - using attribute groups to create the sysfs > > > > > > > > files > > > > > > > > - renaming variables to prevent collision with > > > > > > > > the introduced device structs. > > > > > > > > > > > > > > Hint, when you have to enumerate a list of different > > > > > > > things you do in a single patch, that is a _huge_ sign > > > > > > > that the patch needs to be broken up into smaller pieces, > > > > > > > each piece only doing one logical thing. > > > > > > > > > > > > > > > > > > > This is a huge reconstruction of the driver architecture > > > > > > that requires more than one thing be done at a time. > > > > > > Breaking this up into tiny patches might render the sources > > > > > > broken, if you don't apply the complete series. (Which is > > > > > > also a no-go, right?) > > > > > > > > > > > > Since staging is meant to be the place to fix things up, > > > > > > I thought I'd get away with this. Anyways, I'll try... > > > > > > > > > > Try to make it into something that you would want to be able > > > > > to review :) > > > > > > > > > > > > > Well... ok then. > > > > > > > > Do think you can spend an extra minute in Prague next week, so > > > > we can talk about what's next? > > > > > > Sure, but to start with, getting this patch series into a mergable > > > state is "what is next" :) > > > > > > > What about Wednesday, the slot right before (or right after) lunch > > time? Does this happen to fit in your schedule? > > Ugh, I missed this email earlier, sorry. How about during the coffee > break time today? Want to meet at the registration desk at 3:45? > Too bad, this time I missed it. Seems like email isn't the best choice when trying to arrange a spontaneous meeting, is it? Anyways, I'll try to break down the patches and resend them. This will take me some time, though. I just finished to split #1, which was ok using git's interactive rebase, since the changes were limited to files. But #2 is killing me. regards Chris _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
