----- Am 15. Jul 2019 um 18:32 schrieb Ravindra Kumar Meena rmeena...@gmail.com:
>>
>> This is also a global variable:
>>
>> + static bool ts_recevied[ RTEMS_RECORD_CLIENT_MAXIMUM_CPU_COUNT ] = {
>> false };
>>
>> Please don't modify flush_items(), add everything to print_item().
>>
>> You don't need this extra boolean flag, just use
>>
>> if (timestamp_begin == 0) {
>> timestamp_begin = item->ns;
>> }
>>
>> timestamp_end = item->ns;
>>
>
> Have made changes.
>
> https://github.com/rmeena840/rtems-tools/commit/10c17a29a3810cfd2efc96636032d76a0241321a
>
>
> I have also added packet_size and content_size in both client and metadata.
>
> https://github.com/rmeena840/rtems-tools/commit/88a53f93865fb54f8bcf2afc323baba3feac98c3
>
> https://github.com/rmeena840/rtems-tools/commit/2e4c130f42352df3333427ebafd7c36b1443f16c
>
> Since our binary stream files do not have struct padding so both should be
> equal in values.
>
> Have a look.
I will have a look at this tomorrow. Please add all data used by LTTNG to the
streams.
The next step is to figure out which events are used for the context switch
view in Trace Compass and convert the RTEMS record data so that a corresponding
LTTNG event is output. We can use the Trace Compass to test this.
_______________________________________________
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel
