Revert "x86/mm/ident_map: Use gbpages only where full GB page should be mapped."

This reverts commit d794734.

While the original change tries to fix a bug, it also unintentionally broke
existing systems, see the regressions reported at:

  https://lore.kernel.org/all/[email protected]/

Since d794734 was also marked for -stable, let's back it out before
causing more damage.

Note that due to another upstream change the revert was not 100% automatic:

  0a845e0 mm/treewide: replace pud_large() with pud_leaf()

Signed-off-by: Ingo Molnar <[email protected]>
Cc: <[email protected]>
Cc: Russ Anderson <[email protected]>
Cc: Steve Wahl <[email protected]>
Cc: Dave Hansen <[email protected]>
Link: https://lore.kernel.org/all/[email protected]/
Fixes: d794734 ("x86/mm/ident_map: Use gbpages only where full GB page should be mapped.")

Author: Ingo Molnar

arch/x86/mm/ident_map.c

@@ -26,31 +26,18 @@ static int ident_pud_init(struct x86_mapping_info *info, pud_t *pud_page,
 	for (; addr < end; addr = next) {
 		pud_t *pud = pud_page + pud_index(addr);
 		pmd_t *pmd;
-		bool use_gbpage;
 
 		next = (addr & PUD_MASK) + PUD_SIZE;
 		if (next > end)
 			next = end;
 
-		/* if this is already a gbpage, this portion is already mapped */
-		if (pud_leaf(*pud))
-			continue;
-
-		/* Is using a gbpage allowed? */
-		use_gbpage = info->direct_gbpages;
-
-		/* Don't use gbpage if it maps more than the requested region. */
-		/* at the begining: */
-		use_gbpage &= ((addr & ~PUD_MASK) == 0);
-		/* ... or at the end: */
-		use_gbpage &= ((next & ~PUD_MASK) == 0);
-
-		/* Never overwrite existing mappings */
-		use_gbpage &= !pud_present(*pud);
-
-		if (use_gbpage) {
+		if (info->direct_gbpages) {
 			pud_t pudval;
 
+			if (pud_present(*pud))
+				continue;
+
+			addr &= PUD_MASK;
 			pudval = __pud((addr - info->offset) | info->page_flag);
 			set_pud(pud, pudval);
 			continue;