Hi Marcel,
On 28 February 2014 08:00, Marcel Holtmann <marcel@xxxxxxxxxxxx> wrote:
> Hi Marcin,
>
>> If hfp_gw received line with few commands, it called command
>> handler one time, for first command. Next comamnd was processed
>> only if next line was received.
>> Now, after every response from gw, we call proces_input to be
>> sure that all data has been be processed.
>> It will process next command only if response for previous was sent.
>> ---
>> src/shared/hfp.c | 16 +++++++++-------
>> 1 file changed, 9 insertions(+), 7 deletions(-)
>>
>> diff --git a/src/shared/hfp.c b/src/shared/hfp.c
>> index 854cf46..0681b19 100644
>> --- a/src/shared/hfp.c
>> +++ b/src/shared/hfp.c
>> @@ -121,23 +121,21 @@ static void process_input(struct hfp_gw *hfp)
>>
>> *ptr = '\0';
>> count = asprintf(&ptr, "%s%s", str, str2);
>> - str = ptr;
>> } else {
>> - count = ptr - str;
>> *ptr = '\0';
>> + count = asprintf(&ptr, "%s", str);
>> }
>
> the whole point here is to not allocate a string if the command is not wrapped in the ring buffer. No idea what you are fixing here. And it is also not described in the commit message.
>
>> hfp->result_pending = true;
>>
>> + len = ringbuf_drain(hfp->read_buf, count + 1);
>> +
>> if (hfp->command_callback)
>> - hfp->command_callback(str, hfp->command_data);
>> + hfp->command_callback(ptr, hfp->command_data);
>> else
>> hfp_gw_send_result(hfp, HFP_RESULT_ERROR);
>>
>> - len = ringbuf_drain(hfp->read_buf, count + 1);
>> -
>> - if (str == ptr)
>> - free(ptr);
>> + free(ptr);
>> }
>>
>> static void read_watch_destroy(void *user_data)
>> @@ -341,6 +339,8 @@ bool hfp_gw_send_result(struct hfp_gw *hfp, enum hfp_result result)
>>
>> hfp->result_pending = false;
>>
>> + process_input(hfp);
>> +
>> return true;
>> }
>>
>> @@ -356,6 +356,8 @@ bool hfp_gw_send_error(struct hfp_gw *hfp, enum hfp_error error)
>>
>> hfp->result_pending = false;
>>
>> + process_input(hfp);
>> +
>> return true;
>> }
>
> We can certainly do this, but normally AT commands work that you get one, send a response and then get the next one. Except for unsolicited notification, but in HFP gateway case, we are the ones sending these and not the headset.
>
> What is this all fixing?
>
> Regards
>
> Marcel
>
If we can receive one command in time, lets just ignore this patch
BR
Marcin
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
