fs/namespace: untag user pointers in copy_mount_options

This patch is a part of a series that extends kernel ABI to allow to pass
tagged user pointers (with the top byte set to something else other than
0x00) as syscall arguments.

In copy_mount_options a user address is being subtracted from TASK_SIZE.
If the address is lower than TASK_SIZE, the size is calculated to not
allow the exact_copy_from_user() call to cross TASK_SIZE boundary.
However if the address is tagged, then the size will be calculated
incorrectly.

Untag the address before subtracting.

Link: http://lkml.kernel.org/r/1de225e4a54204bfd7f25dac2635e31aa4aa1d90.1563904656.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <[email protected]>
Reviewed-by: Khalid Aziz <[email protected]>
Reviewed-by: Vincenzo Frascino <[email protected]>
Reviewed-by: Kees Cook <[email protected]>
Reviewed-by: Catalin Marinas <[email protected]>
Cc: Al Viro <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: Eric Auger <[email protected]>
Cc: Felix Kuehling <[email protected]>
Cc: Jens Wiklander <[email protected]>
Cc: Mauro Carvalho Chehab <[email protected]>
Cc: Mike Rapoport <[email protected]>
Cc: Will Deacon <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

Author: xairy

fs/namespace.c

@@ -3028,7 +3028,7 @@ void *copy_mount_options(const void __user * data)
 	 * the remainder of the page.
 	 */
 	/* copy_from_user cannot cross TASK_SIZE ! */
-	size = TASK_SIZE - (unsigned long)data;
+	size = TASK_SIZE - (unsigned long)untagged_addr(data);
 	if (size > PAGE_SIZE)
 		size = PAGE_SIZE;