Re: [PATCH 2/2] tracing: Use strncpy instead of memcpy for string keys in hist triggers

[Tom Zanussi]

Hi Steve,

On Mon, 2019-03-04 at 16:56 -0500, Steven Rostedt wrote:
> On Mon, 4 Mar 2019 16:50:00 -0500
> Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> 
> > > +++ b/kernel/trace/trace_events_hist.c
> > > @@ -4695,9 +4695,10 @@ static inline void add_to_key(char
> > > *compound_key, void *key,
> > >  		/* ensure NULL-termination */
> > >  		if (size > key_field->size - 1)
> > >  			size = key_field->size - 1;
> > > -	}
> > >  
> > > -	memcpy(compound_key + key_field->offset, key, size);
> > > +		strncpy(compound_key + key_field->offset, (char *)key,
> > > size);
> > > +	} else
> > > +		memcpy(compound_key + key_field->offset, key, size);
> > >  }
> > >    
> > 
> > Shouldn't we use strncpy() in save_comm() too. Feels safer.
> 
> Note, if that is changed, it can be another patch. This one is fine
> as
> is. I just was looking at other use cases of memcpy() in that file.
> 

Hmm, I don't think it's really necessary - it's not used in a key so
don't care about anything after the null, and TASK_COMM_LEN is used in
the memcpy.

Tom

> -- Steve