Re: ioctls

[Pete Zaitcev <zaitcev@redhat.com>]

> (i)
> First one has the definitions:
> 
> #define _IOR(type,nr,size)      _IOC(_IOC_READ,(type),(nr),sizeof(size))
> #define _IOW(type,nr,size)      _IOC(_IOC_WRITE,(type),(nr),sizeof(size))
> 
> These are really unfortunate. I suppose I'll submit a patch
> to change the definition into
> 
> #define _IOR(group,nr,argtype)  _IOC(_IOC_READ,(group),(nr),sizeof(argtype))
> 
> Really a lot of people have been fooled into believing that
> the parameter "size" is a size.  But it is a data type.

Great idea! I actualy wanted it changed too, but could not find
two suitable English words. Perhaps that's why Linus blew it.
Wait... you aren't a native speaker either? :-)  Anyway, I like it.
Just about anything is better than what we have now.

> (ii)
> In all cases where the size is wrong (a largish number of cases),
> do we want to define the "correct" ioctls, and leave the old
> ones with _OLD suffix as deprecated?

I would not touch anything, but add a comment near every
instance (/* THIS IS BROKEN DOUBLE sizeof, DO NOT COPY */).

> (iii)
> For userspace it is difficult to get ioctl definitions.
> All the obscure ioctls live in <linux/foo.h> and including
> lots of such headers is a sure way to get a source that
> doesnt compile. Typedef clashes, things outside #ifdef __KERNEL__
> that use things inside, etc etc.
> Would anyone object against creating a directory with a
> name like kernelapi and slowly moving manifest constants,
> ioctl definitions, and definitions of the argument structs
> to files there?

There was a more comprehensive idea floated by Chrishtoff
Hellwig last week. We ought to ask Arjan (or find someone
else) to maintain glibc-kernelheaders as a community tarball
at kernel.org somewhere. Please, don't look at me. OK, only
if Arjan _REALLY_ refuses... Or perhaps you want to take
it yourself? This is basically your idea, only halfway
done: everything is copied already. Also, we do not need
a buy-in from Linus (though I suspect he'd support it).

-- Pete