Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 9 Nov 2020 21:29:13 -0600
From:   Bjorn Andersson <bjorn.andersson@linaro.org>
To:     Rishabh Bhatnagar <rishabhb@codeaurora.org>
Cc:     linux-remoteproc@vger.kernel.org, linux-kernel@vger.kernel.org,
        tsoni@codeaurora.org, psodagud@codeaurora.org,
        sidgup@codeaurora.org
Subject: Re: [PATCH 2/2] remoteproc: qcom: Add trace events for q6v5_pas
 driver
Message-ID: <20201110032913.GC332990@builder.lan>
References: <1604971241-29000-1-git-send-email-rishabhb@codeaurora.org>
 <1604971241-29000-3-git-send-email-rishabhb@codeaurora.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1604971241-29000-3-git-send-email-rishabhb@codeaurora.org>
Precedence: bulk

On Mon 09 Nov 19:20 CST 2020, Rishabh Bhatnagar wrote:

> Add tracepoints for q6v5_pas driver. These will help in
> analyzing the time taken by each step in remoteproc
> bootup/shutdown process and also serve as standard
> checkpoints in code.
> 

These tracepoints seems quite generic and useful to drivers other than
the Qualcomm PAS driver. Please move them into the framework instead.

> Signed-off-by: Rishabh Bhatnagar <rishabhb@codeaurora.org>
> ---
>  drivers/remoteproc/qcom_q6v5_pas.c | 11 +++++++++++
>  include/trace/events/q6v5_pas.h    | 34 ++++++++++++++++++++++++++++++++++
>  2 files changed, 45 insertions(+)
>  create mode 100644 include/trace/events/q6v5_pas.h
> 
> diff --git a/drivers/remoteproc/qcom_q6v5_pas.c b/drivers/remoteproc/qcom_q6v5_pas.c
> index 3837f23..b3c0a6a 100644
> --- a/drivers/remoteproc/qcom_q6v5_pas.c
> +++ b/drivers/remoteproc/qcom_q6v5_pas.c
> @@ -29,6 +29,9 @@
>  #include "qcom_q6v5.h"
>  #include "remoteproc_internal.h"
>  
> +#define CREATE_TRACE_POINTS
> +#include <trace/events/q6v5_pas.h>
> +
>  struct adsp_data {
>  	int crash_reason_smem;
>  	const char *firmware_name;
> @@ -121,12 +124,14 @@ static int adsp_load(struct rproc *rproc, const struct firmware *fw)
>  	struct qcom_adsp *adsp = (struct qcom_adsp *)rproc->priv;
>  	int ret;
>  
> +	trace_q6v5_pas("setting up memory and loading segments", rproc->name);
>  	ret = qcom_mdt_load(adsp->dev, fw, rproc->firmware, adsp->pas_id,
>  			    adsp->mem_region, adsp->mem_phys, adsp->mem_size,
>  			    &adsp->mem_reloc);
>  	if (ret)
>  		return ret;
>  
> +	trace_q6v5_pas("done loading segments", rproc->name);
>  	qcom_pil_info_store(adsp->info_name, adsp->mem_phys, adsp->mem_size);
>  
>  	return 0;
> @@ -137,6 +142,7 @@ static int adsp_start(struct rproc *rproc)
>  	struct qcom_adsp *adsp = (struct qcom_adsp *)rproc->priv;
>  	int ret;
>  
> +	trace_q6v5_pas("Voting for resources", rproc->name);
>  	qcom_q6v5_prepare(&adsp->q6v5);
>  
>  	ret = adsp_pds_enable(adsp, adsp->active_pds, adsp->active_pd_count);
> @@ -163,12 +169,14 @@ static int adsp_start(struct rproc *rproc)
>  	if (ret)
>  		goto disable_cx_supply;
>  
> +	trace_q6v5_pas("Before authenticate and reset", rproc->name);
>  	ret = qcom_scm_pas_auth_and_reset(adsp->pas_id);
>  	if (ret) {
>  		dev_err(adsp->dev,
>  			"failed to authenticate image and release reset\n");
>  		goto disable_px_supply;
>  	}
> +	trace_q6v5_pas("After authenticate and reset", rproc->name);
>  
>  	ret = qcom_q6v5_wait_for_start(&adsp->q6v5, msecs_to_jiffies(5000));
>  	if (ret == -ETIMEDOUT) {
> @@ -177,6 +185,7 @@ static int adsp_start(struct rproc *rproc)
>  		goto disable_px_supply;
>  	}
>  
> +	trace_q6v5_pas("Remoteproc is up", rproc->name);
>  	return 0;
>  
>  disable_px_supply:
> @@ -214,6 +223,7 @@ static int adsp_stop(struct rproc *rproc)
>  	int handover;
>  	int ret;
>  
> +	trace_q6v5_pas("Request stop", rproc->name);
>  	ret = qcom_q6v5_request_stop(&adsp->q6v5);
>  	if (ret == -ETIMEDOUT)
>  		dev_err(adsp->dev, "timed out on wait\n");
> @@ -227,6 +237,7 @@ static int adsp_stop(struct rproc *rproc)
>  	if (handover)
>  		qcom_pas_handover(&adsp->q6v5);
>  
> +	trace_q6v5_pas("Remoteproc is down", rproc->name);
>  	return ret;
>  }
>  
> diff --git a/include/trace/events/q6v5_pas.h b/include/trace/events/q6v5_pas.h
> new file mode 100644
> index 0000000..38ee5e2
> --- /dev/null
> +++ b/include/trace/events/q6v5_pas.h
> @@ -0,0 +1,34 @@
> +/* SPDX-License-Identifier: GPL-2.0-only */
> +/*
> + * Copyright (c) 2020, The Linux Foundation. All rights reserved.
> + */
> +
> +#undef TRACE_SYSTEM
> +#define TRACE_SYSTEM q6v5_pas
> +
> +#if !defined(_TRACE_Q6V5_PAS_H) || defined(TRACE_HEADER_MULTI_READ)
> +#define _TRACE_Q6V5_PAS_H
> +
> +#include <linux/tracepoint.h>
> +
> +TRACE_EVENT(q6v5_pas,
> +
> +	TP_PROTO(const char *event, const char *rproc_name),

Rather than distinguishing the trace events by the textual first
parameter, split it into individual trace events for each event.

Regards,
Bjorn

> +
> +	TP_ARGS(event, rproc_name),
> +
> +	TP_STRUCT__entry(
> +		__string(event, event)
> +		__string(rproc_name, rproc_name)
> +	),
> +
> +	TP_fast_assign(
> +		__assign_str(event, event);
> +		__assign_str(rproc_name, rproc_name);
> +	),
> +
> +	TP_printk("event=%s remoteproc:%s", __get_str(event), __get_str(rproc_name))
> +);
> +
> +#endif
> +#include <trace/define_trace.h>
> -- 
> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
> a Linux Foundation Collaborative Project
>