exec: restore EACCES of S_ISDIR execve()

Patch series "Fix S_ISDIR execve() errno".

Fix an errno change for execve() of directories, noticed by Marc Zyngier.
Along with the fix, include a regression test to avoid seeing this return
in the future.

This patch (of 2):

The return code for attempting to execute a directory has always been
EACCES.  Adjust the S_ISDIR exec test to reflect the old errno instead of
the general EISDIR for other kinds of "open" attempts on directories.

Fixes: 633fb6a ("exec: move S_ISREG() check earlier")
Reported-by: Marc Zyngier <[email protected]>
Signed-off-by: Kees Cook <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Tested-by: Greg Kroah-Hartman <[email protected]>
Reviewed-by: Greg Kroah-Hartman <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Link: https://lore.kernel.org/lkml/20200813151305.6191993b@why
Signed-off-by: Linus Torvalds <[email protected]>

Author: kees

fs/namei.c

@@ -2849,8 +2849,10 @@ static int may_open(const struct path *path, int acc_mode, int flag)
 	case S_IFLNK:
 		return -ELOOP;
 	case S_IFDIR:
-		if (acc_mode & (MAY_WRITE | MAY_EXEC))
+		if (acc_mode & MAY_WRITE)
 			return -EISDIR;
+		if (acc_mode & MAY_EXEC)
+			return -EACCES;
 		break;
 	case S_IFBLK:
 	case S_IFCHR: