x86/mm: Fix no-change case in try_preserve_large_page()

try_preserve_large_page() checks if new_prot is the same as
old_prot.  If so, it simply sets do_split to 0, and returns
with no-operation.  However, old_prot is set as a 4KB pgprot
value while new_prot is a large page pgprot value.

Now that old_prot is initially set from p?d_pgprot() as a
large page pgprot value, fix it by not overwriting old_prot
with a 4KB pgprot value.

Signed-off-by: Toshi Kani <toshi.kani@hpe.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Juergen Gross <jgross@suse.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Konrad Wilk <konrad.wilk@oracle.com>
Cc: Robert Elliot <elliott@hpe.com>
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/1442514264-12475-12-git-send-email-toshi.kani@hpe.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

arch/x86/mm/pageattr.c

@@ -518,7 +518,7 @@ try_preserve_large_page(pte_t *kpte, unsigned long address,
 	 * up accordingly.
 	 */
 	old_pte = *kpte;
-	old_prot = req_prot = pgprot_large_2_4k(old_prot);
+	req_prot = pgprot_large_2_4k(old_prot);
 
 	pgprot_val(req_prot) &= ~pgprot_val(cpa->mask_clr);
 	pgprot_val(req_prot) |= pgprot_val(cpa->mask_set);