[linux-dvb] Problems with ca_zap

[Manu Abraham]

Allan Stirling wrote:
>  From pat.c:
> 
> static uint16_t parse_descriptor(struct program_descriptor 
> *p_program_descriptor,
>                                                                uint8_t 
> *buf, uint16_t pos)
> {
>        p_program_descriptor->program_number =
>                ((p_program_descriptor->program_number | buf[pos + 0]) << 
> 8) | buf[pos + 1];
> 
> Doesn't look quite right. Why do we care what the program_number is 
> before it's been parsed?
> 

Hmm, that is required because buf is 8 bits. All what i do is, get buf 
into program_number and shift program_number to the left by 8.

> If I change this to
> 
>        p_program_descriptor->program_number =
>                ( buf[pos + 0] << 8) | buf[pos + 1];
> 

This would be wrong however, since if you left shift an 8 bit buffer by 
8 all you will get is 0.

The problem would be solved, if program_number is initialized to 0 just 
before reading in from the buffer.


Manu