The following commit has been merged into the x86/core branch of tip:

Commit-ID:     f666c92090a41ac5524dade63ff96b3adcf8c2ab
Gitweb:        https://git.kernel.org/tip/f666c92090a41ac5524dade63ff96b3adcf8c2ab
Author:        Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
AuthorDate:    Wed, 16 Oct 2024 14:14:55 +03:00
Committer:     Ingo Molnar <mingo@kernel.org>
CommitterDate: Wed, 19 Mar 2025 11:12:29 +01:00

x86/mm/ident_map: Fix theoretical virtual address overflow to zero

The current calculation of the 'next' virtual address in the
page table initialization functions in arch/x86/mm/ident_map.c
doesn't protect against wrapping to zero.

This is a theoretical issue that cannot happen currently,
the problematic case is possible only if the user sets a
high enough x86_mapping_info::offset value - which no
current code in the upstream kernel does.

( The wrapping to zero only occurs if the top PGD entry is accessed.
  There are no such users upstream. Only hibernate_64.c uses
  x86_mapping_info::offset, and it operates on the direct mapping
  range, which is not the top PGD entry. )

Should such an overflow happen, it can result in page table
corruption and a hang.

To future-proof this code, replace the manual 'next' calculation
with p?d_addr_end() which handles wrapping correctly.

[ Backporter's note: there's no need to backport this patch. ]

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kai Huang <kai.huang@intel.com>
Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/20241016111458.846228-2-kirill.shutemov@linux.intel.com
---
 arch/x86/mm/ident_map.c | 14 +++-----------
 1 file changed, 3 insertions(+), 11 deletions(-)

diff --git a/arch/x86/mm/ident_map.c b/arch/x86/mm/ident_map.c
index 5ab7bd2..bd5d101 100644
--- a/arch/x86/mm/ident_map.c
+++ b/arch/x86/mm/ident_map.c
@@ -101,9 +101,7 @@ static int ident_pud_init(struct x86_mapping_info *info, pud_t *pud_page,
 		pmd_t *pmd;
 		bool use_gbpage;
 
-		next = (addr & PUD_MASK) + PUD_SIZE;
-		if (next > end)
-			next = end;
+		next = pud_addr_end(addr, end);
 
 		/* if this is already a gbpage, this portion is already mapped */
 		if (pud_leaf(*pud))
@@ -154,10 +152,7 @@ static int ident_p4d_init(struct x86_mapping_info *info, p4d_t *p4d_page,
 		p4d_t *p4d = p4d_page + p4d_index(addr);
 		pud_t *pud;
 
-		next = (addr & P4D_MASK) + P4D_SIZE;
-		if (next > end)
-			next = end;
-
+		next = p4d_addr_end(addr, end);
 		if (p4d_present(*p4d)) {
 			pud = pud_offset(p4d, 0);
 			result = ident_pud_init(info, pud, addr, next);
@@ -199,10 +194,7 @@ int kernel_ident_mapping_init(struct x86_mapping_info *info, pgd_t *pgd_page,
 		pgd_t *pgd = pgd_page + pgd_index(addr);
 		p4d_t *p4d;
 
-		next = (addr & PGDIR_MASK) + PGDIR_SIZE;
-		if (next > end)
-			next = end;
-
+		next = pgd_addr_end(addr, end);
 		if (pgd_present(*pgd)) {
 			p4d = p4d_offset(pgd, 0);
 			result = ident_p4d_init(info, p4d, addr, next);