cifs: fix cifsConvertToUCS() for the mapchars case
authorJeff Layton <jlayton@redhat.com>
Tue, 17 May 2011 19:28:21 +0000 (15:28 -0400)
committerSteve Conklin <sconklin@canonical.com>
Thu, 2 Jun 2011 19:23:31 +0000 (14:23 -0500)
BugLink: http://bugs.launchpad.net/bugs/788691

commit 11379b5e33950048ad66825da7f462b0d0da9d73 upstream.

As Metze pointed out, commit 84cdf74e broke mapchars option:

    Commit "cifs: fix unaligned accesses in cifsConvertToUCS"
    (84cdf74e8096a10dd6acbb870dd404b92f07a756) does multiple steps
    in just one commit (moving the function and changing it without
    testing).

    put_unaligned_le16(temp, &target[j]); is never called for any
    codepoint the goes via the 'default' switch statement. As a result
    we put just zero (or maybe uninitialized) bytes into the target
    buffer.

His proposed patch looks correct, but doesn't apply to the current head
of the tree. This patch should also fix it.

Reported-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Steve Conklin <sconklin@canonical.com>

fs/cifs/cifs_unicode.c

index 23d43cd..1b2e180 100644 (file)
@@ -277,6 +277,7 @@ cifsConvertToUCS(__le16 *target, const char *source, int srclen,
 
        for (i = 0, j = 0; i < srclen; j++) {
                src_char = source[i];
+               charlen = 1;
                switch (src_char) {
                case 0:
                        put_unaligned(0, &target[j]);
@@ -316,16 +317,13 @@ cifsConvertToUCS(__le16 *target, const char *source, int srclen,
                                dst_char = cpu_to_le16(0x003f);
                                charlen = 1;
                        }
-                       /*
-                        * character may take more than one byte in the source
-                        * string, but will take exactly two bytes in the
-                        * target string
-                        */
-                       i += charlen;
-                       continue;
                }
+               /*
+                * character may take more than one byte in the source string,
+                * but will take exactly two bytes in the target string
+                */
+               i += charlen;
                put_unaligned(dst_char, &target[j]);
-               i++; /* move to next char in source string */
        }
 
 ctoUCS_out: