Re: [PATCH 2/2] perf inject: Emit instruction records on ETM trace discontinuity

Em Tue, Feb 13, 2018 at 03:22:33PM -0700, Mathieu Poirier escreveu:

On Wed, Feb 07, 2018 at 01:57:25PM +0000, Robert Walker wrote:

...

@@ -946,20 +982,19 @@ static int cs_etm__run_decoder(struct cs_etm_queue *etmq) /* Run trace decoder until buffer consumed or end of trace */ do { processed = 0;

this...

...

	err = cs_etm_decoder__process_data_block(
		etmq->decoder,
		etmq->offset,
		&buffer.buf[buffer_used],
		buffer.len - buffer_used,
		&processed);

and this should have gone in the first patch.

...

	if (err)
		return err;

etmq->offset += processed; buffer_used += processed;

• 	/* Process each packet in this chunk */
  

And probably this too.

With the above changes:

Acked-by: Mathieu Poirier mathieu.poirier@linaro.org

Hi Robert,

Can you please address Mathieu's comments, and if you agree with them, resubmit?

Thanks!

- Arnaldo