Messages in this thread |  | | | Date | Fri, 02 Feb 2007 16:40:07 -0500 | | From | Jeff Garzik <> | | Subject | Re: Should io(read|write)(8|16|32)_rep take (const|) volatile u(8|16|32) __iomem *addr? |
| |
Michael K. Edwards wrote: > Doubtless this is on purpose, but it's not clear to me why that should > be true for these particular functions/macros. I shouldn't have to > cast away the volatile on a pointer to hardware registers in order to > pass it into writel(), should I? And it shouldn't be forbidden for > the caller to declare the pointer volatile so that the compiler > catches attempts to pass it into non-volatile interfaces, should it?
Therein lies your bug: you should not be annotating your __iomem pointers with volatile.
That's why volatile is found in the implementation and not the prototypes of the accessor functions.
Jeff
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
|  |