nls: utf8_wcstombs: use correct buffer size in error case

When utf8_wcstombs encounters a character that cannot be encoded, we
must not decrease the remaining output buffer size because nothing has
been written to the output buffer.

Signed-off-by: Clemens Ladisch <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This commit is contained in:
Clemens Ladisch 2009-04-24 10:11:40 +02:00 • committed by Greg Kroah-Hartman
commit e27ecdd94d

View file

@ -144,7 +144,6 @@ utf8_wcstombs(__u8 *s, const wchar_t *pwcs, int maxlen)
size = utf8_wctomb(op, *ip, maxlen);
if (size == -1) {
/* Ignore character and move on */
maxlen--;
} else {
op += size;
maxlen -= size;