VME and the staging tree

Greg KH gregkh at linuxfoundation.org
Thu Apr 26 15:17:33 UTC 2012


On Thu, Apr 26, 2012 at 09:40:24AM +0100, Martyn Welch wrote:
> On 25/04/12 21:56, Greg KH wrote:
> > On Wed, Apr 25, 2012 at 04:03:22PM -0400, Emilio G. Cota wrote:
> >> On Wed, Apr 25, 2012 at 11:25:32 -0700, Greg KH wrote:
> >>> I have reviewed the code, and only have one major comment at this point
> >>> in time:
> >>>
> >>> What's up with the device ioctl in
> >>> drivers/staging/vme/devices/vme_user.c?  In the code, it says this is
> >>> something that should not be used, yet there is lots of code in the
> >>> ioctl itself, implying that it is used?  There's also lots of comments
> >>> in the function saying things should be fixed up.
> >>>
> >>> Is this ioctl still needed?  Can I just remove it and all will be fine?
> >>> If so, I'll be glad to do so.
> >>
> >> IIRC vme_user is just there to provide backwards compatibility with
> >> the old VME driver hosted at vmelinux.org--the IOCTL thus comes imposed
> >> by that old interface.
> >>
> 
> That's partially true - the old driver provided an extensive set of ioctls to
> control the driver, it had no kernel API. I had to cut it down quite a bit to
> make it at all sensible. It appears that vmelinux.org has gone (it's been down
> for a while now), so the information about that API will be hard to come by.
> 
> >> I don't think there are any users of it (please Martyn confirm this),
> >> so I'd advocate for getting rid of vme_user.[ch] altogether.
> > 
> > Ok, that seems reasonable to me.  If Martyn agrees, I'll go delete those
> > files.
> > 
> 
> I'm fairly sure that there are people using the vme_user module to access the
> VME bus to do simple things, however I also agree that it isn't really in a
> mergeable state at this point in time. Would it be possible to merge the VME
> core and bridge drivers and leave the VME device drivers (the contents of
> devices/) in staging for a bit longer? The pio2 driver is basically a GPIO
> device driver, so that should probably be merged into drivers/gpio/ and it
> will give us a bit more time to consider how to proceed with vme_user.

Ok, that sounds good to me, I'll go do that right now unless Emilio
objects?

greg k-h



More information about the devel mailing list