On Sat, Sep 20, 2003 at 05:07:54PM -0400, Mark M. Hoffman wrote:
> * Greg KH <greg at kroah.com> [2003-09-15 11:30:26 -0700]:
> > Ugh, in looking over the bus drivers that are currently in the cvs tree,
> > but not in the kernel tree, I ran accross this horrible bit of code in
> > the i2c-i810.c driver:
> >
> > static unsigned char *mem;
> >
> > static inline void outlong(unsigned int dat, int off)
> > {
> > *((unsigned int *) (mem + off)) = dat;
> > }
> >
> > static inline unsigned int readlong(int off)
> > {
> > return *((unsigned int *) (mem + off));
> > }
> >
> > Will someone please fix this up properly to not directly access memory!
> > That is not portable at all, and will do bad things on non-i386 based
> > hardware platforms.
> >
> > After that's done, I'll worry about porting it to 2.6, but not before...
>
> Is this what you want? If so I'll update it (and the other two).
> Er, I'll get rid of the outlong/readlong indirection too...
Yes, this is the proper way to do things.
Thanks for looking into this and making the changes.
greg k-h
