Re: [PATCH 1/1] mmc: add MODALIAS linkage for MMC/SD devices

[Andy Whitcroft <apw@canonical.com>]

On Tue, Jan 27, 2009 at 08:26:26PM +0100, Kay Sievers wrote:
> On Tue, Jan 27, 2009 at 20:14, Pierre Ossman <drzeus-mmc@drzeus.cx> wrote:
> > On Sun, 25 Jan 2009 17:00:11 +0100
> > Kay Sievers <kay.sievers@vrfy.org> wrote:
> >
> >> On Sun, Jan 25, 2009 at 16:48, Pierre Ossman <drzeus-mmc@drzeus.cx> wrote:
> >> >
> >> > My point was to have the kernel explicitly ask for the module it wants
> >> > as there is no decent device to driver mapping scheme.
> >>
> >> Yep, which is what we do not want. Aliases are "aliases", and not
> >> "module names". We need to add a matching alias to the module then.
> >> Direct module names can not properly defined/blacklisted in userspace,
> >> and we would need to work around that.
> >> Every modalias should be
> >> "<subsystem>:<whatever-name-fits-for-the-subsystem>" to plug properly
> >> into the autoloading infrastructure. We rather have no modalias at
> >> all, then a kernel module name there.
> >>
> >
> > The thing is that asking for a module is the only thing we can do here.
> > We can dress it up and give it some special coding to not cause
> > problems, but the code will always be "ask userspace to load
> > mmc_block", even if we replace "mmc_block" with "mmc:foobargazonk".
> >
> > Given that, do you have any preferences for a solution? If we cannot
> > simply have "mmc_block", then I'm leaning to "mmc:block" for now. The
> > contents of the aliases is just an opaque string as far as userspace is
> > concerned, right?
> 
> Right, it does not matter, we just need to be able to fnmatch()
> modalias and the string in the module.
> 
> We prefer every modalias to be prefixed with the subsystem, to be able
> to hook userspace configs into the alias processing. The string itself
> is completely up to the subsystem to decide whatever fits. If there
> could ever be several types of strings per susbsystem, it is usual to
> prefix the string with some character, but some subsystems know that
> this is not needed, and just put a plain single string there.

As the contents as opaque we can put whatever we want after the mmc:,
the last patch I did simply used the mmc type string as a key in case it
were useful.  Right now it is as we do only load the block for the SD
and MMC cases.  But later it would be for all once the SDIO card
support, supports memory.

So asre we happy with the last textual patch or should I refactor back
to mmc:block.

-apw