Thank you, Bastian! Your suggestion allowed me to move forward.

The error message appears to be generated only in the min_available_space
function of this file: gnuradio-runtime/lib/block_executor.cc .

I tried to derive an exact expression of the value I need to set for a
given payload size, but I couldn't in the time I had. Does anybody already
know it?

Another question I have is: will set_min_output_buffer have any effect on a
running flowgraph? Is that a good idea?

Thanks!


2016-03-14 21:26 GMT-03:00 Bastian Bloessl <[email protected]>:

> Hi,
>
>
> > On 14 Mar 2016, at 16:56, Francisco Albani <[email protected]>
> wrote:
> >
> > I wrote a custom message block that produces a PDU with complex samples.
> My plan is to connect it to a PDU to Tagged Stream and then to a UHD Sink
> (with len_tag_name = 'packet_len').
> >
> > The goal is to transmit a burst for each packet. This packets are made
> of approximately 14k samples and sometimes more.
> >
> > I'm unable to do this because of this error:
> >
> > thread[thread-per-block[1]: <block pdu_to_tagged_stream (2)>]: Buffer
> too small for min_noutput_items
> >
> > I tried rising the PDUtoTaggedStream block min/max output buffer size to
> its maximum values with no success.
> >
> > Attached is the minimum equivalent flowgraph that shows the same problem.
> >
> > Is this a Scheduler hardcoded limitation?
> >
> > Anybody knows a workaround?
> >
>
> the Tag Debug and Null Sink seem to not have an interface for adjusting
> the buffers, but if I add it (manually) to the python file, it works for me
>
> (self.blocks_pdu_to_tagged_stream_0).set_min_output_buffer(100000)
> (self.blocks_null_sink_0).set_min_output_buffer(100000)
> (self.blocks_tag_debug_0).set_min_output_buffer(100000)
>
> Best,
> Bastian
_______________________________________________
Discuss-gnuradio mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

Reply via email to