Re: [PATCH 2/7] tracing: Change strlen to sizeof for hist trigger static strings

[Joe Perches <joe@perches.com>]

On Wed, 2018-12-19 at 13:46 -0600, Tom Zanussi wrote:
> Hi Steve,
> 
> On Wed, 2018-12-19 at 14:40 -0500, Steven Rostedt wrote:
> > On Tue, 18 Dec 2018 14:33:21 -0600
> > Tom Zanussi <zanussi@kernel.org> wrote:
> > 
> > > From: Tom Zanussi <tom.zanussi@linux.intel.com>
> > > 
> > > There's no need to use strlen() for static strings when the length
> > > is
> > > already known, so update trace_events_hist.c with sizeof() for
> > > those
> > > cases.
> > > 
> > > Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>
> > > ---
> > >  kernel/trace/trace_events_hist.c | 38 +++++++++++++++++++---------
> > > ----------
> > >  1 file changed, 19 insertions(+), 19 deletions(-)
> > > 
> > > diff --git a/kernel/trace/trace_events_hist.c
> > > b/kernel/trace/trace_events_hist.c
> > > index d29bf8a8e1dd..25d06b3ae1f6 100644
> > > --- a/kernel/trace/trace_events_hist.c
> > > +++ b/kernel/trace/trace_events_hist.c
> > > @@ -507,7 +507,7 @@ static int synth_field_string_size(char *type)
> > >  	start = strstr(type, "char[");
> > >  	if (start == NULL)
> > >  		return -EINVAL;
> > > -	start += strlen("char[");
> > > +	start += sizeof("char[") - 1;
> > >  
> > >  	end = strchr(type, ']');
> > >  	if (!end || end < start)
> > > @@ -1843,8 +1843,8 @@ static int parse_action(char *str, struct
> > > hist_trigger_attrs *attrs)
> > >  	if (attrs->n_actions >= HIST_ACTIONS_MAX)
> > >  		return ret;
> > >  
> > > -	if ((strncmp(str, "onmatch(", strlen("onmatch(")) == 0) ||
> > > -	    (strncmp(str, "onmax(", strlen("onmax(")) == 0)) {
> > > +	if ((strncmp(str, "onmatch(", sizeof("onmatch(") - 1) ==
> > > 0) ||
> > > +	    (strncmp(str, "onmax(", sizeof("onmax(") - 1) == 0)) {
> > >  		attrs->action_str[attrs->n_actions] = kstrdup(str,
> > > GFP_KERNEL);
> > >  		if (!attrs->action_str[attrs->n_actions]) {
> > >  			ret = -ENOMEM;
> > > @@ -1861,34 +1861,34 @@ static int parse_assignment(char *str,
> > > struct hist_trigger_attrs *attrs)
> > >  {
> > >  	int ret = 0;
> > >  
> > > -	if ((strncmp(str, "key=", strlen("key=")) == 0) ||
> > > -	    (strncmp(str, "keys=", strlen("keys=")) == 0)) {
> > > +	if ((strncmp(str, "key=", sizeof("key=") - 1) == 0) ||
> > > +	    (strncmp(str, "keys=", sizeof("keys=") - 1) == 0)) {
> > >  		attrs->keys_str = kstrdup(str, GFP_KERNEL);
> > 
> > I'll apply this as is, but since there's a lot of these, I wonder if
> > we
> > should make a marcro for this:
> > 
> > #define strcmp_const(str, str_const) strncmp(str, str_const,
> > sizeof(str_const) - 1)
> > 
> > ?
> > 
> > This would help prevent bugs due to typos and bad cut and paste.
> > 
> 
> Yeah, I had considered it but wasn't sure it was worth it.  Since
> you're suggesting it is, I can send another patch on top of these, or
> feel free if you want to too.  ;-)

I believe the 'strlen("foo") -> sizeof("foo") - 1'
conversions do not change objects at all.

strlen("constant") is already optimized by gcc to a
constant value when fed a constant string.

the strcmp_const macro does seem to make sense as
the copy/paste/typo possibility is real.