Em Mon, 30 Dec 2013 13:48:34 +0100
André Roth <[email protected]> escreveu:

description?

What bug are you trying to fix?

> Signed-off-by: André Roth <[email protected]>
> ---
>  lib/libdvbv5/dvb-scan.c | 36 +++++++++++++++++++++++-------------
>  1 file changed, 23 insertions(+), 13 deletions(-)
> 
> diff --git a/lib/libdvbv5/dvb-scan.c b/lib/libdvbv5/dvb-scan.c
> index e9ccc72..520bf9c 100644
> --- a/lib/libdvbv5/dvb-scan.c
> +++ b/lib/libdvbv5/dvb-scan.c
> @@ -96,9 +96,13 @@ int dvb_read_section_with_id(struct dvb_v5_fe_parms 
> *parms, int dmx_fd,
>       uint8_t *buf = NULL;
>       uint8_t *tbl = NULL;
>       ssize_t table_length = 0;
> +
> +     // handle sections

No C99 comments.

> +     int start_id = -1;
> +     int start_section = -1;
>       int first_section = -1;
>       int last_section = -1;
> -     int table_id = -1;
> +     /*int table_id = -1;*/

Why to comment? If it is buggy, just remove it.

>       int sections = 0;
>       struct dmx_sct_filter_params f;
>       struct dvb_table_header *h;
> @@ -108,7 +112,6 @@ int dvb_read_section_with_id(struct dvb_v5_fe_parms 
> *parms, int dmx_fd,
>       *table = NULL;
>  
>       // FIXME: verify known table
> -
>       memset(&f, 0, sizeof(f));
>       f.pid = pid;
>       f.filter.filter[0] = tid;
> @@ -185,24 +188,27 @@ int dvb_read_section_with_id(struct dvb_v5_fe_parms 
> *parms, int dmx_fd,
>  
>               h = (struct dvb_table_header *)buf;
>               dvb_table_header_init(h);
> +
> +             /* dvb_logdbg( "dvb_read_section: id %d, section %d/%d, 
> current: %d", h->id, h->section_id, h->last_section, h->current_next ); */
> +             if (start_id == h->id && start_section == h->section_id) {
> +                     dvb_logdbg( "dvb_read_section: section repeated, 
> reading done" );
> +                     break;
> +             }
> +             if (start_id == -1) start_id = h->id;
> +             if (start_section == -1) start_section = h->section_id;

One statement per line, please.

> +
>               if (id != -1 && h->id != id) { /* search for a specific table 
> id */
>                       continue;
> -             } else {
> -                     if (table_id == -1)
> -                             table_id = h->id;
> -                     else if (h->id != table_id) {
> -                             dvb_logwarn("dvb_read_section: table ID 
> mismatch reading multi section table: %d != %d", h->id, table_id);
> -                             continue;
> -                     }
>               }
>  
> +             /*dvb_logerr("dvb_read_section: got section %d, last %d, filter 
> %d", h->section_id, h->last_section, id );*/

Why are you adding a commented line?

>               /* handle the sections */
>               if (first_section == -1)
>                       first_section = h->section_id;
> -             else if (h->section_id == first_section)
> +             else if (start_id == h->id && h->section_id == first_section)
>                       break;
>  
> -             if (last_section == -1)
> +             if (last_section == -1 || h->last_section > last_section)
>                       last_section = h->last_section;
>  
>               if (!tbl) {
> @@ -228,10 +234,14 @@ int dvb_read_section_with_id(struct dvb_v5_fe_parms 
> *parms, int dmx_fd,
>               else
>                       dvb_logerr("dvb_read_section: no initializer for table 
> %d", tid);
>  
> -             if (++sections == last_section + 1)
> +             if (id != -1 && ++sections == last_section + 1) {
> +                     dvb_logerr("dvb_read_section: ++sections == 
> last_section + 1");
>                       break;
> +             }
>       }
> -     free(buf);
> +
> +     if (buf)
> +             free(buf);

Buffer is always allocated at this point. No need to test before free.

>  
>       dvb_dmx_stop(dmx_fd);
>  


-- 

Cheers,
Mauro
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to