Re: updated-fbmem-patch.patch

[Andrea Arcangeli <andrea@suse.de>]

On Wed, Apr 21, 2004 at 03:41:53PM -0700, Chris Wright wrote:
> * Andrea Arcangeli (andrea@suse.de) wrote:
> > http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.5/2.6.5-mm4/broken-out/updated-fbmem-patch.patch
> > 
> > this uses get_user for the set_cmap operation too.
> > 
> > --- 2.6.5-aa3/drivers/video/fbmem.c.~1~	2004-04-04 08:09:23.000000000 +0200
> > +++ 2.6.5-aa3/drivers/video/fbmem.c	2004-04-21 03:11:05.878951424 +0200
> > @@ -1034,11 +1034,11 @@ fb_ioctl(struct inode *inode, struct fil
> >  	case FBIOPUTCMAP:
> >  		if (copy_from_user(&cmap, (void *) arg, sizeof(cmap)))
> >  			return -EFAULT;
> > -		return (fb_set_cmap(&cmap, 0, info));
> > +		return (fb_set_cmap(&cmap, 1, info));
> 
> 0 is userspace, 1 is kernel space.  this change looks wrong.
> 
> Perhaps the change below so comment is in sync with code.

yes, I was mislead by the comment, that's why you will never ever see a
single comments describing parameters in my code (except if the stuff
really isn't obvious and it cannot be trivially deduced by the code and
in turn it _deserves_ a fat comment, definitely not in this case).
There's no way I can trust comments anyways, I perfectly know I cannot
stop after reading a comment, I went ahead and I checked if the comment
was right but it was too late and I was already biased by what I read in
the comment so I didn't notice the comment was wrong. This isn't a good
excuse, it's still my bad mistake, shame on me, but it certainly gives
more strenght to my no-comment-for-obvious-parameters policy. I don't
care if you cannot do a pdf of the whole kernel anymore, that pdf is
likely buggy anyways and you'd better not attempt to read it in the
first place.

Arjan and Andrew also notified me privately. thanks to all for noticing
and sorry for the stupid mistake of being influenced by a worthless
comment.