On Sun, May 15, 2022 at 01:24:13PM -0700, Richard Cochran wrote:
> On Tue, Nov 23, 2021 at 02:14:13AM +0200, Vladimir Oltean wrote:
>
> > diff --git a/ts2phc.h b/ts2phc.h
> > new file mode 100644
> > index 000000000000..14cb2b0c21a3
> > --- /dev/null
> > +++ b/ts2phc.h
> > @@ -0,0 +1,23 @@
> > +/**
> > + * @file ts2phc.h
> > + * @brief Structure definitions for ts2phc
> > + * @note Copyright 2020 Vladimir Oltean <olte...@gmail.com>
> > + * @note SPDX-License-Identifier: GPL-2.0+
> > + */
> > +#ifndef HAVE_TS2PHC_H
> > +#define HAVE_TS2PHC_H
> > +
> > +struct ts2phc_sink_array;
> > +
> > +struct ts2phc_private {
> > + struct ts2phc_pps_source *src;
> > + STAILQ_HEAD(sink_ifaces_head, ts2phc_pps_sink) sinks;
> > + unsigned int n_sinks;
> > + struct ts2phc_sink_array *polling_array;
> > + struct config *cfg;
> > +};
> > +
> > +#include "ts2phc_pps_source.h"
> > +#include "ts2phc_pps_sink.h"
>
> It is a bit strange to put the include directives after the struct
> definitions. After all, ts2phc_private has a pointer to a
> ts2phc_pps_source. Wonder why compiler doesn't complain about missing
> forward declaration.
>
> Anyhow, please put the include directives at the top.
Ok, I will add a forward declaration for struct ts2phc_private in
ts2phc_pps_source and ts2phc_pps_sink.
>
> > +#endif
>
> > @@ -65,49 +64,55 @@ ts2phc_pps_sink_offset(struct ts2phc_pps_sink *sink,
> > struct ts2phc_source_timestamp ts,
> > int64_t *offset, uint64_t *local_ts);
> >
> > -static STAILQ_HEAD(pps_sink_ifaces_head, ts2phc_pps_sink) ts2phc_sinks =
> > - STAILQ_HEAD_INITIALIZER(ts2phc_sinks);
> > -
> > -static unsigned int ts2phc_n_sinks;
> > -
> > -static int ts2phc_pps_sink_array_create(void)
> > +static int ts2phc_pps_sink_array_create(struct ts2phc_private *priv)
> > {
> > + struct ts2phc_sink_array *polling_array;
> > struct ts2phc_pps_sink *sink;
> > unsigned int i;
> >
> > - if (polling_array.sink) {
> > - return 0;
> > + polling_array = malloc(sizeof(*polling_array));
>
> This polling_array is never freed. Please fix it, so that there is no
> memory leak on normal program termination.
>
> (I use valgrind in pre-release testing to make sure no new memory
> issues have appeared.)
>
> Thanks,
> Richard
Bad patch splitting, sorry. I have free(polling_array) in the next patch
"ts2phc: instantiate a full clock structure for every PPS sink".
Anyway I've moved it here, will resend after retesting if there are no
other comments.
_______________________________________________
Linuxptp-devel mailing list
Linuxptp-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
