From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.6 required=3.0 tests=DKIM_SIGNED,DKIM_VALID, DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI, SIGNED_OFF_BY,SPF_PASS,URIBL_BLOCKED,USER_AGENT_MUTT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 80331C04EB8 for ; Wed, 5 Dec 2018 03:31:58 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 2828820851 for ; Wed, 5 Dec 2018 03:31:58 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (1024-bit key) header.d=linaro.org header.i=@linaro.org header.b="VAA+kmeK" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 2828820851 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linaro.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726810AbeLEDb4 (ORCPT ); Tue, 4 Dec 2018 22:31:56 -0500 Received: from mail-wr1-f52.google.com ([209.85.221.52]:46742 "EHLO mail-wr1-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1725864AbeLEDb4 (ORCPT ); Tue, 4 Dec 2018 22:31:56 -0500 Received: by mail-wr1-f52.google.com with SMTP id l9so18080682wrt.13 for ; Tue, 04 Dec 2018 19:31:54 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; h=date:from:to:cc:subject:message-id:references:mime-version :content-disposition:in-reply-to:user-agent; bh=wlB088EpjbYek5NUXwNcHjHRuylhxCk0ZK1/tQY/sk0=; b=VAA+kmeKItk/BPU5YyORYApHXpEKcLqBlZ2zrXSWIYyciaPnf9bBiQg17/3Nh5bZou pro8CuOxcGIdlfPY9V028c+10gPduIHKtmlHgo0oMlYtsO1+83hOiRXagQcVoY6+DF2t SfOMDOktR3NFuwJYM3JwHdQUWwSsuKRUYAAYI= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:date:from:to:cc:subject:message-id:references :mime-version:content-disposition:in-reply-to:user-agent; bh=wlB088EpjbYek5NUXwNcHjHRuylhxCk0ZK1/tQY/sk0=; b=XU/9p6ScioXG604rYYUySXcTmvqPFa6sO5HcV5xDnjB4cFZ+/Qy4l+8nF9WvtslJiM kXqvx7dznAlYqvk4SizH/2xhFjmwK3wHP65pDD7acaSqRM9UZgZYufnNTL5QTqthWJZA DsCD2c79XXur5xmXgv4iDoO0rYHFL4sPglq/j1sHpxvDBsXENv5Lqyv6xKW5pBMBrphZ U05tO2/FF5R1hPeqc4t1P9F3uqUy1T2Kx+krSsmO6yv8DUyCOCCuZDudPzmjujA/FQdO kZezE76X5fZ03UzJ3DG6M7ZVCl1XdfbVtJgeaojKwdjy+hDMxde4z/L3bEf12hZRoJ2c MAjg== X-Gm-Message-State: AA+aEWbjZ+ucJnrto+xdifsfUThKFgFklIB3RncLZrO0LUS3Naz1hhoZ 9RwBCLxvqX+2rL2lXEu6gLTWow== X-Google-Smtp-Source: AFSGD/Vl8s3XYOM+zzdpiCJXzy0gKzWjIhjss1fXYwqw6EPg8PHIChEVrGdnCb8kPjgTouv3jyHZsw== X-Received: by 2002:adf:92c7:: with SMTP id 65mr6836282wrn.228.1543980713789; Tue, 04 Dec 2018 19:31:53 -0800 (PST) Received: from leoy-ThinkPad-X240s ([209.250.228.18]) by smtp.gmail.com with ESMTPSA id w12sm15911882wrr.23.2018.12.04.19.31.48 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Tue, 04 Dec 2018 19:31:52 -0800 (PST) Date: Wed, 5 Dec 2018 11:31:45 +0800 From: leo.yan@linaro.org To: Mathieu Poirier Cc: Arnaldo Carvalho de Melo , Alexander Shishkin , Jiri Olsa , Namhyung Kim , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, Mike Leach , Robert Walker , Al Grant , Coresight ML Subject: Re: [PATCH v1 3/5] perf cs-etm: Support for NO_SYNC packet Message-ID: <20181205033145.GA15964@leoy-ThinkPad-X240s> References: <1541912383-19915-1-git-send-email-leo.yan@linaro.org> <1541912383-19915-4-git-send-email-leo.yan@linaro.org> <20181119182759.GA608@xps15> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20181119182759.GA608@xps15> User-Agent: Mutt/1.10+31 (9cdd884) (2018-06-19) Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Nov 19, 2018 at 11:27:59AM -0700, Mathieu Poirier wrote: > On Sun, Nov 11, 2018 at 12:59:41PM +0800, Leo Yan wrote: > > As described in OpenCSD (CoreSight decoder lib), in the decoding stream > > it includes one trace element with type OCSD_GEN_TRC_ELEM_NO_SYNC; the > > element indicates 'either at start of decode, or after overflow / bad > > packet', we should take it as a signal for the tracing off and this will > > cause tracing discontinuity. From the trace dump with 'perf script', > > sometimes the element OCSD_GEN_TRC_ELEM_NO_SYNC collaborates with > > element OCSD_GEN_TRC_ELEM_TRACE_ON to show the tracing flow have been > > turned off and on, in this case the cs-etm code has handled TRACE_ON > > packet well so we observe the tracing discontinuity; but in another case > > it only inserts the element OCSD_GEN_TRC_ELEM_NO_SYNC into instructions > > packets, we miss to handle the case if has only standalone NO_SYNC > > element and users cannot receive the info for tracing discontinuity. > > > > This patch introduces new type CS_ETM_TRACE_OFF to generate packet for > > receiving element OCSD_GEN_TRC_ELEM_NO_SYNC from decoder; when generate > > sample, CS_ETM_TRACE_OFF packet has almost the same behaviour with > > Here you have used the word "almost", leading me to beleive there are cases > where the handling of TRACE_ON and NO_SYNC packets differ, but the > implementation enacts the same behavior for both. > > Mike or Robert, can you confirm that TRACE_ON and NO_SYNC packets should be > treated the same way? I also would like to get suggestions/comments from Mike and Rob. For NO_SYNC packet, Mike before has given some explination for it: "looking at the decoder flow, when the decoder is reset, then it is returned to an unsynchronised state and a NO_SYNC will be output. With perf, it can concatenate multiple trace buffers into a single section of the perf.data file. To ensure that the decode process can see the start of a new buffer, the drivers will insert a barrier packet between different buffers so that the decoder can spot the boundaries. When the decoder hits a barrier packet it will automatically reset so that it correctly decodes the next trace buffer. This could be what you are seeing here." So I think there still have some difference between TRACE_ON and NO_SYNC packets, TRACE_ON packet indicates the start of trace and it's also possible caused by tracing discontinuity; NO_SYNC packets usually caused by an unsynchronised state. But both of them presents discontinuity in perf trace data. I prefer to use two different packet types to present them, this can let the code to be more readable and easier to be maintained later. If you agree with this, I also will rephrase the commit log to avoid confusion that these two packets are the same thing. > Leo, if handling of the TRACE_ON and NO_SYNC packets is the same then we should > treat them the same way in cs_etm_decoder__gen_trace_elem_printer(), i.e simply > call cs_etm_decoder__buffer_trace_on(). That way packet handling in cs-etm.c > doesn't change. Otherwise see my comments below. > > > CS_ETM_TRACE_ON packet: both of them invokes cs_etm__flush() to generate > > samples for the previous instructions packet, and in cs_etm__sample() it > > also needs to generate samples if TRACE_OFF packet is followed by one > > sequential instructions packet. This patch also converts the address to > > 0 for TRACE_OFF packet, this is same with TRACE_ON packet as well. > > > > Signed-off-by: Leo Yan > > --- > > tools/perf/util/cs-etm-decoder/cs-etm-decoder.c | 10 ++++++++++ > > tools/perf/util/cs-etm-decoder/cs-etm-decoder.h | 7 ++++--- > > tools/perf/util/cs-etm.c | 15 +++++++++++---- > > 3 files changed, 25 insertions(+), 7 deletions(-) > > > > diff --git a/tools/perf/util/cs-etm-decoder/cs-etm-decoder.c b/tools/perf/util/cs-etm-decoder/cs-etm-decoder.c > > index 5efb616..9d52727 100644 > > --- a/tools/perf/util/cs-etm-decoder/cs-etm-decoder.c > > +++ b/tools/perf/util/cs-etm-decoder/cs-etm-decoder.c > > @@ -369,6 +369,14 @@ cs_etm_decoder__buffer_range(struct cs_etm_decoder *decoder, > > } > > > > static ocsd_datapath_resp_t > > +cs_etm_decoder__buffer_trace_off(struct cs_etm_decoder *decoder, > > + const uint8_t trace_chan_id) > > +{ > > + return cs_etm_decoder__buffer_packet(decoder, trace_chan_id, > > + CS_ETM_TRACE_OFF); > > +} > > + > > +static ocsd_datapath_resp_t > > cs_etm_decoder__buffer_trace_on(struct cs_etm_decoder *decoder, > > const uint8_t trace_chan_id) > > { > > @@ -389,6 +397,8 @@ static ocsd_datapath_resp_t cs_etm_decoder__gen_trace_elem_printer( > > case OCSD_GEN_TRC_ELEM_UNKNOWN: > > break; > > case OCSD_GEN_TRC_ELEM_NO_SYNC: > > + resp = cs_etm_decoder__buffer_trace_off(decoder, > > + trace_chan_id); > > If we want to handle NO_SYNC element types, why introduce a "trace_off" > function? Wouldn't it make more sense to call it > cs_etm_decoder__buffer_no_sync() ? Will change to cs_etm_decoder__buffer_no_sync(). > > decoder->trace_on = false; > > break; > > case OCSD_GEN_TRC_ELEM_TRACE_ON: > > diff --git a/tools/perf/util/cs-etm-decoder/cs-etm-decoder.h b/tools/perf/util/cs-etm-decoder/cs-etm-decoder.h > > index 9351bd1..a38c97c 100644 > > --- a/tools/perf/util/cs-etm-decoder/cs-etm-decoder.h > > +++ b/tools/perf/util/cs-etm-decoder/cs-etm-decoder.h > > @@ -23,9 +23,10 @@ struct cs_etm_buffer { > > }; > > > > enum cs_etm_sample_type { > > - CS_ETM_EMPTY = 0, > > - CS_ETM_RANGE = 1 << 0, > > - CS_ETM_TRACE_ON = 1 << 1, > > + CS_ETM_EMPTY = 0, > > + CS_ETM_RANGE = 1 << 0, > > + CS_ETM_TRACE_ON = 1 << 1, > > + CS_ETM_TRACE_OFF = 1 << 2, > > CS_ETM_NO_SYNC, see above comment. And please don't use indentation. Will do this. > > }; > > > > enum cs_etm_isa { > > diff --git a/tools/perf/util/cs-etm.c b/tools/perf/util/cs-etm.c > > index f4fa877..2a0cef9 100644 > > --- a/tools/perf/util/cs-etm.c > > +++ b/tools/perf/util/cs-etm.c > > @@ -517,8 +517,9 @@ static inline int cs_etm__t32_instr_size(struct cs_etm_queue *etmq, > > > > static inline u64 cs_etm__first_executed_instr(struct cs_etm_packet *packet) > > { > > - /* Returns 0 for the CS_ETM_TRACE_ON packet */ > > - if (packet->sample_type == CS_ETM_TRACE_ON) > > + /* Returns 0 for TRACE_ON and TRACE_OFF packets */ > > + if (packet->sample_type == CS_ETM_TRACE_ON || > > + packet->sample_type == CS_ETM_TRACE_OFF) > > return 0; > > > > return packet->start_addr; > > @@ -527,8 +528,9 @@ static inline u64 cs_etm__first_executed_instr(struct cs_etm_packet *packet) > > static inline > > u64 cs_etm__last_executed_instr(const struct cs_etm_packet *packet) > > { > > - /* Returns 0 for the CS_ETM_TRACE_ON packet */ > > - if (packet->sample_type == CS_ETM_TRACE_ON) > > + /* Returns 0 for TRACE_ON and TRACE_OFF packets */ > > + if (packet->sample_type == CS_ETM_TRACE_ON || > > + packet->sample_type == CS_ETM_TRACE_OFF) > > return 0; > > > > return packet->end_addr - packet->last_instr_size; > > @@ -930,6 +932,10 @@ static int cs_etm__sample(struct cs_etm_queue *etmq) > > if (etmq->prev_packet->sample_type == CS_ETM_TRACE_ON) > > generate_sample = true; > > > > + /* Generate sample for tracing off packet */ > > + if (etmq->prev_packet->sample_type == CS_ETM_TRACE_OFF) > > + generate_sample = true; > > + > > /* Generate sample for branch taken packet */ > > if (etmq->prev_packet->sample_type == CS_ETM_RANGE && > > etmq->prev_packet->last_instr_taken_branch) > > @@ -1085,6 +1091,7 @@ static int cs_etm__run_decoder(struct cs_etm_queue *etmq) > > cs_etm__sample(etmq); > > break; > > case CS_ETM_TRACE_ON: > > + case CS_ETM_TRACE_OFF: > > /* > > * Discontinuity in trace, flush > > * previous branch stack > > -- > > 2.7.4 > >