Re: [PATCH] Fix performance problems with mprotect()

["Bruce Rogers" <BROGERS@novell.com>]

I'm not sure how this BUG_ON would get triggered with current x86 paging semantics, and probably isn't needed I think.
Am I missing something?
The code is almost exactly the same as the existing update_intpte, which has the same BUG_ON, and so I just left it in.
(BTW:I considered merging the new functions into the non "sync" versions but opted not to.)

- Bruce

>>> On 1/7/2008 at 8:18 AM, in message <p73d4sdsly4.fsf@bingen.suse.de>, Andi Kleen
<andi@firstfloor.org> wrote:
> "Bruce Rogers" <brogers@novell.com> writes:
>> +
>> +        /* Allowed to change in Accessed/Dirty flags only. */
>> +        BUG_ON((t ^ old) & ~(intpte_t)(_PAGE_ACCESSED|_PAGE_DIRTY));
> 
> 
> Are you sure that BUG_ON can't be triggered from the hypercall?
> It should error out in this case I think.
> 
> -Andi