On Sat, 9 Feb 2019 at 18:08, Heinrich Schuchardt xypron.glpk@gmx.de wrote:

We have to ensure while printing UTF16 strings that we do not exceed the end of the print buffer.

Signed-off-by: Heinrich Schuchardt xypron.glpk@gmx.de

lib/vsprintf.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)

Reviewed-by: Simon Glass sjg@chromium.org

Please see below.

diff --git a/lib/vsprintf.c b/lib/vsprintf.c index 4213441fbf..48304607b7 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -279,13 +279,17 @@ static char *string(char *buf, char *end, char *s, int field_width, static char *string16(char *buf, char *end, u16 *s, int field_width, int precision, int flags) {

•   u16 *str = s ? s : L"<NULL>";
  

•   ssize_t len = utf16_strnlen(str, precision);
  

•   const u16 *str = s ? s : L"<NULL>";
  

•   ssize_t i, len = utf16_strnlen(str, precision);
  
    if (!(flags & LEFT))
            for (; len < field_width; --field_width)
                    ADDCH(buf, ' ');
  

•   utf16_utf8_strncpy(&buf, str, len);
  

•   for (i = 0; i < len && buf <= end - MAX_UTF8_PER_UTF16; ++i) {
  

I'm a bit worried that this is overly conservative. Could utf16_get() perhaps return the actual length?

•           s32 s = utf16_get(&str);
  

•           utf8_put(s, &buf);
  

•   }
    for (; len < field_width; --field_width)
            ADDCH(buf, ' ');
    return buf;
  

-- 2.20.1

Regards, Simon