Re: [Qemu-devel] [PATCH] coverity-model: replay data is considered trusted

[Paolo Bonzini]

On 15/05/2018 14:00, Markus Armbruster wrote:
> Paolo Bonzini <address@hidden> writes:
> 
>> Replay data is not considered a possible attack vector; add a model that
>> does not use getc so that "tainted data" warnings are suppressed.
>>
>> Signed-off-by: Paolo Bonzini <address@hidden>
>> ---
>>  scripts/coverity-model.c | 13 +++++++++++++
>>  1 file changed, 13 insertions(+)
>>
>> diff --git a/scripts/coverity-model.c b/scripts/coverity-model.c
>> index c702804f41..576f48de33 100644
>> --- a/scripts/coverity-model.c
>> +++ b/scripts/coverity-model.c
>> @@ -103,6 +103,19 @@ static int get_keysym(const name2keysym_t *table,
>    /* Tainting */
> 
>    typedef struct {} name2keysym_t;
>    static int get_keysym(const name2keysym_t *table,
>                          const char *name)
>    {
>        int result;
>        if (result > 0) {
>            __coverity_tainted_string_sanitize_content__(name);
>            return result;
>        } else {
>            return 0;
>>      }
>>  }
>>  
>> +
> 
> Does the new model go under /* Tainting */ ?

Yes, it does.  Any chance you can do the change yourself?...

Paolo

> If yes, I'd like to have just one blank line here.
> 
>> +/* Replay data is considered trusted.  */
> 
> If no, I'd like to insert one here.
> 
>> +uint8_t replay_get_byte(void)
>> +{
>> +     uint8_t byte = 0;
>> +     if (replay_file) {
>> +         uint8_t c;
>> +         byte = c;
>> +     }
>> +     return byte;
>> +}
>> +
>> +
>>  /*
>>   * GLib memory allocation functions.
>>   *