On Wed, Jun 28, 2023 at 5:34 PM Andy Shevchenko
<andriy.shevchenko@xxxxxxxxxxxxxxx> wrote:
>
> We don't need to traverse over the entire string and replace
> occurrences of a character with '\0'. The first match will
> suffice. Hence, replace strreplace() with strchrnul().
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> ---
> mm/kasan/report_generic.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/mm/kasan/report_generic.c b/mm/kasan/report_generic.c
> index 51a1e8a8877f..63a34eac4a8c 100644
> --- a/mm/kasan/report_generic.c
> +++ b/mm/kasan/report_generic.c
> @@ -264,6 +264,7 @@ static void print_decoded_frame_descr(const char *frame_descr)
> while (num_objects--) {
> unsigned long offset;
> unsigned long size;
> + char *p;
>
> /* access offset */
> if (!tokenize_frame_descr(&frame_descr, token, sizeof(token),
> @@ -282,7 +283,7 @@ static void print_decoded_frame_descr(const char *frame_descr)
> return;
>
> /* Strip line number; without filename it's not very helpful. */
> - strreplace(token, ':', '\0');
> + p[strchrnul(token, ':') - token] = '\0';
Why not just
*(strchrnul(token, ':')) = '\0';
?
