Re: [PATCH 10/10] trace: Replace printk and WARN_ON with WARN

From: Steven Rostedt
Date: Fri Mar 27 2020 - 18:51:57 EST


On Fri, 27 Mar 2020 21:23:57 +0000
Jules Irenge <jbi.octave@xxxxxxxxx> wrote:

> Coccinelle suggests replacing printk and WARN_ON with WARN
>
> SUGGESTION: printk + WARN_ON can be just WARN.
> Signed-off-by: Jules Irenge <jbi.octave@xxxxxxxxx>
> ---
> kernel/trace/trace.c | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
> index 6b11e4e2150c..1fe31272ea73 100644
> --- a/kernel/trace/trace.c
> +++ b/kernel/trace/trace.c
> @@ -1799,9 +1799,7 @@ static int run_tracer_selftest(struct tracer *type)
> /* the test is responsible for resetting too */
> tr->current_trace = saved_tracer;
> if (ret) {
> - printk(KERN_CONT "FAILED!\n");
> - /* Add the warning after printing 'FAILED' */

NACK! Did you not read the above comment. The FAILED goes with another
print and should NOT be part of the WARN_ON.

-- Steve

> - WARN_ON(1);
> + WARN(1, "FAILED!\n");
> return -1;
> }
> /* Only reset on passing, to avoid touching corrupted buffers */