vfs: Fix EOVERFLOW testing in put_compat_statfs64
authorEric Sandeen <sandeen@redhat.com>
Wed, 2 Oct 2019 21:17:54 +0000 (16:17 -0500)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 3 Oct 2019 21:21:35 +0000 (14:21 -0700)
Today, put_compat_statfs64() disallows nearly any field value over
2^32 if f_bsize is only 32 bits, but that makes no sense.
compat_statfs64 is there for the explicit purpose of providing 64-bit
fields for f_files, f_ffree, etc.  And f_bsize is always only 32 bits.

As a result, 32-bit userspace gets -EOVERFLOW for i.e.  large file
counts even with -D_FILE_OFFSET_BITS=64 set.

In reality, only f_bsize and f_frsize can legitimately overflow
(fields like f_type and f_namelen should never be large), so test
only those fields.

This bug was discussed at length some time ago, and this is the proposal
Al suggested at https://lkml.org/lkml/2018/8/6/640.  It seemed to get
dropped amid the discussion of other related changes, but this
part seems obviously correct on its own, so I've picked it up and
sent it, for expediency.

Fixes: 64d2ab32efe3 ("vfs: fix put_compat_statfs64() does not handle errors")
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/statfs.c

index eea7af6f2f229e85f71e3ff3ac5d541e88956856..2616424012ea7bf4bc7f87a935466c74bb267397 100644 (file)
@@ -318,19 +318,10 @@ COMPAT_SYSCALL_DEFINE2(fstatfs, unsigned int, fd, struct compat_statfs __user *,
 static int put_compat_statfs64(struct compat_statfs64 __user *ubuf, struct kstatfs *kbuf)
 {
        struct compat_statfs64 buf;
-       if (sizeof(ubuf->f_bsize) == 4) {
-               if ((kbuf->f_type | kbuf->f_bsize | kbuf->f_namelen |
-                    kbuf->f_frsize | kbuf->f_flags) & 0xffffffff00000000ULL)
-                       return -EOVERFLOW;
-               /* f_files and f_ffree may be -1; it's okay
-                * to stuff that into 32 bits */
-               if (kbuf->f_files != 0xffffffffffffffffULL
-                && (kbuf->f_files & 0xffffffff00000000ULL))
-                       return -EOVERFLOW;
-               if (kbuf->f_ffree != 0xffffffffffffffffULL
-                && (kbuf->f_ffree & 0xffffffff00000000ULL))
-                       return -EOVERFLOW;
-       }
+
+       if ((kbuf->f_bsize | kbuf->f_frsize) & 0xffffffff00000000ULL)
+               return -EOVERFLOW;
+
        memset(&buf, 0, sizeof(struct compat_statfs64));
        buf.f_type = kbuf->f_type;
        buf.f_bsize = kbuf->f_bsize;