Re: [PATCH] ARM: add get_user() support for 8 byte types

[Russell King - ARM Linux <linux@arm.linux.org.uk>]

On Mon, Nov 12, 2012 at 06:31:50PM -0600, Rob Clark wrote:
> right, that is what I was worried about..  but what about something
> along the lines of:
> 
> 		case 8: {						\
> 			if (sizeof(x) < 8)				\
> 				__get_user_x(__r2, __p, __e, __l, 4);	\
> 			else						\
> 				__get_user_x(__r2, __p, __e, __l, 8);	\
> 			break;						\
> 		}							\
> 
> maybe we need a special variant of __get_user_8() instead to get the
> right 32bits on big vs little endian systems, but I think something
> roughly along these lines could work.

The problem with that is... big endian systems - where the 32-bit word we
want is the second one, so you can't just reduce that down to a 32-bit
access like that.  You also need to add an offset to the pointer in the BE
case (which can be done.)

I'd suggest calling the 4-byte version in there __get_user_xb() and doing
the 4-byte offset for BE inside that (or aliasing it to __get_user_x for
LE systems).