On Tue, 23 Sep 2025 18:59:07 +0200
Michal Koutný <[email protected]> wrote:

> If ever PID_MAX_DEFAULT changes, it must be compatible with tracing
> hashmaps assumptions.
> 
> Link: https://lore.kernel.org/r/[email protected]/
> Signed-off-by: Michal Koutný <[email protected]>
> ---
>  kernel/trace/trace_sched_switch.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> Drive-by flushing an old idea. Take it or leave it.
> 
> diff --git a/kernel/trace/trace_sched_switch.c 
> b/kernel/trace/trace_sched_switch.c
> index cb49f7279dc80..aabae7daaac91 100644
> --- a/kernel/trace/trace_sched_switch.c
> +++ b/kernel/trace/trace_sched_switch.c
> @@ -243,6 +243,8 @@ int trace_create_savedcmd(void)
>  int trace_save_cmdline(struct task_struct *tsk)
>  {
>       unsigned tpid, idx;
> +     /* so that map_pid_to_cmdline indexing is efficient */
> +     BUILD_BUG_ON(!is_power_of_2(PID_MAX_DEFAULT));

Please move this to just above the test. And it doesn't need a comment,
as it's obvious for why it needs to be a power of two when it's above
the test.

>  
>       /* treat recording of idle task as a success */
>       if (!tsk->pid)

Like this:

        /* treat recording of idle task as a success */
        if (!tsk->pid)
                return 1;

+       BUILD_BUG_ON(!is_power_of_2(PID_MAX_DEFAULT));
+
        tpid = tsk->pid & (PID_MAX_DEFAULT - 1);

-- Steve

Reply via email to