Re: [PATCH 2/5] ftrace: Comment __ftrace_hash_rec_update() and make filter_hash bool

From: Mark Rutland
Date: Wed Jun 05 2024 - 06:15:52 EST


On Tue, Jun 04, 2024 at 05:28:19PM -0400, Steven Rostedt wrote:
> From: "Steven Rostedt (Google)" <rostedt@xxxxxxxxxxx>
>
> The function __ftrace_hash_rec_update() parameter "filter_hash" is only
> used for true or false (boolean), but is of type int. It already has an
> "inc" parameter that is boolean. This is confusing, make "filter_hash"
> boolean as well.
>
> While at it, add some documentation to that function especially since it
> holds the guts of the filtering logic of ftrace.
>
> Signed-off-by: Steven Rostedt (Google) <rostedt@xxxxxxxxxxx>
> ---
> kernel/trace/ftrace.c | 14 +++++++++++++-
> 1 file changed, 13 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> index 9dcdefe9d1aa..93c7c5fd4249 100644
> --- a/kernel/trace/ftrace.c
> +++ b/kernel/trace/ftrace.c
> @@ -1701,8 +1701,20 @@ static bool skip_record(struct dyn_ftrace *rec)
> !(rec->flags & FTRACE_FL_ENABLED);
> }
>
> +/*
> + * This is the main engine to the ftrace updates.
> + *
> + * It will iterate through all the available ftrace functions
> + * (the ones that ftrace can have callbacks to) and set the flags
> + * to the associated dyn_ftrace records.

I beleive s/to/in/ here, to make this one of:

set the flags in the associated dyn_ftrace records.

... rather than:

set the flags to the associated dyn_ftrace records.

> + *
> + * @filter_hash: True if for the filter hash is udpated, false for the
> + * notrace hash

Typo: s/udpated/updated/

... though I couldn't parse this regardless; maybe:

@filter_hash: true to update the filter hash, false to update
the notrace hash

Mark.

> + * @inc: True to add this hash, false to remove it (increment the
> + * recorder counters or decrement them).
> + */
> static bool __ftrace_hash_rec_update(struct ftrace_ops *ops,
> - int filter_hash,
> + bool filter_hash,
> bool inc)
> {
> struct ftrace_hash *hash;
> --
> 2.43.0
>
>