Linux-ARM-Kernel Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Mike Leach <mike.leach@arm.com>
To: Yeoreum Yun <yeoreum.yun@arm.com>,
	James Clark <james.clark@linaro.org>, Leo Yan <leo.yan@arm.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Mathieu Poirier <mathieu.poirier@linaro.org>,
	coresight@lists.linaro.org, linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-rt-devel@lists.linux.dev
Cc: Suzuki K Poulose <suzuki.poulose@arm.com>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
	Clark Williams <clrkwllms@kernel.org>,
	Steven Rostedt <rostedt@goodmis.org>
Subject: Re: [PATCH v11 4/9] coresight: etm3x: fix inconsistencies with sysfs configuration
Date: Fri, 18 Sep 2026 14:57:30 +0100	[thread overview]
Message-ID: <9b7086de-df87-4d5c-8a7a-fc2f7c724c0b@arm.com> (raw)
In-Reply-To: <20260915-separate_etm_cfg_v2-v11-4-d2b258d51747@arm.com>

Hi,

On 9/15/26 12:34, Yeoreum Yun wrote:
> The current ETM3x configuration via sysfs can lead to the following
> inconsistencies:
> 
>    - If a configuration is modified via sysfs while a perf session is
>      active, the running configuration may differ between before
>      a sched-out and after a subsequent sched-in.
> 
> To resolve these issues, separate the configuration into:
> 
>    - active_config: the configuration applied to the current session
>    - config: the configuration set via sysfs
> 

Same comment as for the etm4 config naming


> Additionally:
> 
>    - Since active_config and related fields are accessed only by the local CPU
>      in etm_enable/disable_sysfs_smp_call() (similar to perf enable/disable),
>      remove the lock/unlock from the sysfs enable/disable path and
>      starting/dying_cpu path except when to access config fields only.
> 
>    - Some of sysfs interface read etm register directly. To reduce lock
>      scope while the etm_enable_hw()/etm_disable_hw(), handle it via
>      IPI so that registers could be read from on proper CPU.
> 
> Fixes: 1925a470ce69 ("coresight: etm3x: splitting struct etm_drvdata")
> Signed-off-by: Yeoreum Yun <yeoreum.yun@arm.com>
> ---
>   drivers/hwtracing/coresight/coresight-etm.h        |  4 +-
>   drivers/hwtracing/coresight/coresight-etm3x-core.c | 68 ++++++++++--------
>   .../hwtracing/coresight/coresight-etm3x-sysfs.c    | 80 +++++++++++++++-------
>   3 files changed, 100 insertions(+), 52 deletions(-)
> 
> diff --git a/drivers/hwtracing/coresight/coresight-etm.h b/drivers/hwtracing/coresight/coresight-etm.h
> index 1d753cca29439..f3796162168d4 100644
> --- a/drivers/hwtracing/coresight/coresight-etm.h
> +++ b/drivers/hwtracing/coresight/coresight-etm.h
> @@ -226,7 +226,8 @@ struct etm_config {
>    * @etmccr:	value of register ETMCCR.
>    * @etmccer:	value of register ETMCCER.
>    * @traceid:	value of the current ID for this component.
> - * @config:	structure holding configuration parameters.
> + * @active_config:	structure holding current running configuration.
> + * @config:	structure holding sysfs mode configuration.
>    */
>   struct etm_drvdata {
>   	struct csdev_access		csa;
> @@ -248,6 +249,7 @@ struct etm_drvdata {
>   	u32				etmccr;
>   	u32				etmccer;
>   	u32				traceid;
> +	struct etm_config		active_config;
>   	struct etm_config		config;
>   };
>   
> diff --git a/drivers/hwtracing/coresight/coresight-etm3x-core.c b/drivers/hwtracing/coresight/coresight-etm3x-core.c
> index 862ad0786699c..fd76a57e5f861 100644
> --- a/drivers/hwtracing/coresight/coresight-etm3x-core.c
> +++ b/drivers/hwtracing/coresight/coresight-etm3x-core.c
> @@ -308,7 +308,7 @@ void etm_config_trace_mode(struct etm_config *config)
>   static int etm_parse_event_config(struct etm_drvdata *drvdata,
>   				  struct perf_event *event)
>   {
> -	struct etm_config *config = &drvdata->config;
> +	struct etm_config *config = &drvdata->active_config;
>   	struct perf_event_attr *attr = &event->attr;
>   	u8 ts_level;
>   
> @@ -367,7 +367,7 @@ static int etm_enable_hw(struct etm_drvdata *drvdata)
>   {
>   	int i, rc;
>   	u32 etmcr;
> -	struct etm_config *config = &drvdata->config;
> +	struct etm_config *config = &drvdata->active_config;
>   	struct coresight_device *csdev = drvdata->csdev;
>   
>   	CS_UNLOCK(drvdata->csa.base);
> @@ -442,32 +442,30 @@ static int etm_enable_hw(struct etm_drvdata *drvdata)
>   struct etm_enable_arg {
>   	struct etm_drvdata *drvdata;
>   	struct coresight_path *path;
> +	struct etm_config config;
>   	int rc;
>   };
>   
>   static void etm_enable_sysfs_smp_call(void *info)
>   {
>   	struct etm_enable_arg *arg = info;
> +	struct etm_drvdata *drvdata;
>   	struct coresight_device *csdev;
>   
>   	if (WARN_ON(!arg))
>   		return;
>   
> -	csdev = arg->drvdata->csdev;
> -	if (!coresight_take_mode(csdev, CS_MODE_SYSFS)) {
> -		/* Someone is already using the tracer */
> -		arg->rc = -EBUSY;
> -		return;
> -	}
> +	drvdata = arg->drvdata;
> +	csdev = drvdata->csdev;
>   
> -	arg->rc = etm_enable_hw(arg->drvdata);
> +	drvdata->active_config = arg->config;
> +	drvdata->traceid = arg->path->trace_id;
>   
> -	/* The tracer didn't start */
> -	if (arg->rc) {
> -		coresight_set_mode(csdev, CS_MODE_DISABLED);
> +	arg->rc = etm_enable_hw(arg->drvdata);
> +	if (arg->rc)
>   		return;
> -	}
>   
> +	drvdata->sticky_enable = true;
>   	csdev->path = arg->path;
>   }
>   
> @@ -512,9 +510,10 @@ static int etm_enable_sysfs(struct coresight_device *csdev, struct coresight_pat
>   	struct etm_enable_arg arg = { };
>   	int ret;
>   
> -	spin_lock(&drvdata->spinlock);
> -
> -	drvdata->traceid = path->trace_id;
> +	if (!coresight_take_mode(csdev, CS_MODE_SYSFS)) {
> +		/* Someone is already using the tracer */
> +		return -EBUSY;
> +	}
>   
>   	/*
>   	 * Configure the ETM only if the CPU is online.  If it isn't online
> @@ -523,23 +522,27 @@ static int etm_enable_sysfs(struct coresight_device *csdev, struct coresight_pat
>   	if (cpu_online(drvdata->cpu)) {
>   		arg.drvdata = drvdata;
>   		arg.path = path;
> +
> +		scoped_guard(spinlock, &drvdata->spinlock) {
> +			arg.config = drvdata->config;
> +		}
> +

Again I think the arg.config is unnecessary

Regards

Mike

>   		ret = smp_call_function_single(drvdata->cpu,
>   					       etm_enable_sysfs_smp_call, &arg, 1);
>   		if (!ret)
>   			ret = arg.rc;
> -		if (!ret)
> -			drvdata->sticky_enable = true;
>   	} else {
>   		ret = -ENODEV;
>   	}
>   
> -	if (ret)
> +	if (!ret) {
> +		dev_dbg(&csdev->dev, "ETM tracing enabled\n");
> +	} else {
>   		etm_release_trace_id(drvdata);
> +		/* The tracer didn't start */
> +		coresight_set_mode(csdev, CS_MODE_DISABLED);
> +	}
>   
> -	spin_unlock(&drvdata->spinlock);
> -
> -	if (!ret)
> -		dev_dbg(&csdev->dev, "ETM tracing enabled\n");
>   	return ret;
>   }
>   
> @@ -565,7 +568,7 @@ static int etm_enable(struct coresight_device *csdev, struct perf_event *event,
>   static void etm_disable_hw(struct etm_drvdata *drvdata)
>   {
>   	int i;
> -	struct etm_config *config = &drvdata->config;
> +	struct etm_config *config = &drvdata->active_config;
>   	struct coresight_device *csdev = drvdata->csdev;
>   
>   	CS_UNLOCK(drvdata->csa.base);
> @@ -593,7 +596,6 @@ static void etm_disable_sysfs_smp_call(void *info)
>   	etm_disable_hw(drvdata);
>   
>   	drvdata->csdev->path = NULL;
> -	coresight_set_mode(drvdata->csdev, CS_MODE_DISABLED);
>   }
>   
>   static void etm_disable_perf(struct coresight_device *csdev)
> @@ -630,8 +632,9 @@ static void etm_disable_perf(struct coresight_device *csdev)
>   static void etm_disable_sysfs(struct coresight_device *csdev)
>   {
>   	struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
> -
> -	spin_lock(&drvdata->spinlock);
> +	struct etm_config *config = &drvdata->config;
> +	const struct etm_config *active_config = &drvdata->active_config;
> +	int i;
>   
>   	/*
>   	 * Executing etm_disable_hw on the cpu whose ETM is being disabled
> @@ -640,7 +643,14 @@ static void etm_disable_sysfs(struct coresight_device *csdev)
>   	smp_call_function_single(drvdata->cpu, etm_disable_sysfs_smp_call,
>   				 drvdata, 1);
>   
> -	spin_unlock(&drvdata->spinlock);
> +	/*
> +	 * Userspace may read seq_curr_state and cntr_val through sysfs
> +	 * after the sysfs-session has been disabled.
> +	 */
> +	for (i = 0; i < drvdata->nr_cntr; i++)
> +		config->cntr_val[i] = active_config->cntr_val[i];
> +
> +	config->seq_curr_state = active_config->seq_curr_state;
>   
>   	/*
>   	 * we only release trace IDs when resetting sysfs.
> @@ -649,6 +659,8 @@ static void etm_disable_sysfs(struct coresight_device *csdev)
>   	 * prior trace id allocation method
>   	 */
>   
> +	coresight_set_mode(drvdata->csdev, CS_MODE_DISABLED);
> +
>   	dev_dbg(&csdev->dev, "ETM tracing disabled\n");
>   }
>   
> diff --git a/drivers/hwtracing/coresight/coresight-etm3x-sysfs.c b/drivers/hwtracing/coresight/coresight-etm3x-sysfs.c
> index d4846d6364945..2c8c129228589 100644
> --- a/drivers/hwtracing/coresight/coresight-etm3x-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-etm3x-sysfs.c
> @@ -6,10 +6,48 @@
>   
>   #include <linux/pid_namespace.h>
>   #include <linux/pm_runtime.h>
> +#include <linux/smp.h>
>   #include <linux/sysfs.h>
>   #include "coresight-etm.h"
>   #include "coresight-priv.h"
>   
> +struct etm_readl_cslocked_arg {
> +	struct etm_drvdata *drvdata;
> +	u32 off;
> +	unsigned long *val;
> +};
> +
> +static void etm_readl_cslocked_smp_call(void *info)
> +{
> +	struct etm_readl_cslocked_arg *arg = info;
> +
> +	CS_UNLOCK(arg->drvdata->csa.base);
> +	*arg->val = etm_readl(arg->drvdata, arg->off);
> +	CS_LOCK(arg->drvdata->csa.base);
> +}
> +
> +static int etm_readl_cslocked(struct etm_drvdata *drvdata, u32 off,
> +			      unsigned long *val)
> +{
> +	int ret;
> +	struct etm_readl_cslocked_arg arg;
> +
> +	ret = pm_runtime_get_sync(drvdata->csdev->dev.parent);
> +	if (ret < 0)
> +		goto out;
> +
> +	arg.drvdata = drvdata;
> +	arg.off = off;
> +	arg.val = val;
> +
> +	ret = smp_call_function_single(drvdata->cpu,
> +				       etm_readl_cslocked_smp_call,
> +				       (void *)&arg, 1);
> +out:
> +	pm_runtime_put(drvdata->csdev->dev.parent);
> +	return ret;
> +}
> +
>   static ssize_t nr_addr_cmp_show(struct device *dev,
>   				struct device_attribute *attr, char *buf)
>   {
> @@ -45,18 +83,13 @@ static DEVICE_ATTR_RO(nr_ctxid_cmp);
>   static ssize_t etmsr_show(struct device *dev,
>   			  struct device_attribute *attr, char *buf)
>   {
> -	unsigned long flags, val;
> +	int ret;
> +	unsigned long val;
>   	struct etm_drvdata *drvdata = dev_get_drvdata(dev->parent);
>   
> -	pm_runtime_get_sync(dev->parent);
> -	spin_lock_irqsave(&drvdata->spinlock, flags);
> -	CS_UNLOCK(drvdata->csa.base);
> -
> -	val = etm_readl(drvdata, ETMSR);
> -
> -	CS_LOCK(drvdata->csa.base);
> -	spin_unlock_irqrestore(&drvdata->spinlock, flags);
> -	pm_runtime_put(dev->parent);
> +	ret = etm_readl_cslocked(drvdata, ETMSR, &val);
> +	if (ret)
> +		return ret;
>   
>   	return sprintf(buf, "%#lx\n", val);
>   }
> @@ -74,7 +107,7 @@ static ssize_t reset_store(struct device *dev,
>   	ret = kstrtoul(buf, 16, &val);
>   	if (ret)
>   		return ret;
> -	if (coresight_get_mode(drvdata->csdev))
> +	if (coresight_get_mode(drvdata->csdev) == CS_MODE_SYSFS)
>   		return -EBUSY;
>   
>   	if (val) {
> @@ -719,7 +752,9 @@ static DEVICE_ATTR_RW(cntr_rld_event);
>   static ssize_t cntr_val_show(struct device *dev,
>   			     struct device_attribute *attr, char *buf)
>   {
> +	int ret;
>   	u32 val;
> +	unsigned long val2;
>   	struct etm_drvdata *drvdata = dev_get_drvdata(dev->parent);
>   	struct etm_config *config = &drvdata->config;
>   
> @@ -728,7 +763,10 @@ static ssize_t cntr_val_show(struct device *dev,
>   		val = config->cntr_val[config->cntr_idx];
>   		spin_unlock(&drvdata->spinlock);
>   	} else {
> -		val = etm_readl(drvdata, ETMCNTVRn(config->cntr_idx));
> +		ret = etm_readl_cslocked(drvdata, ETMCNTVRn(config->cntr_idx), &val2);
> +		if (ret)
> +			return ret;
> +		val = val2;
>   	}
>   
>   	return sysfs_emit(buf, "%#x\n", val);
> @@ -746,7 +784,7 @@ static ssize_t cntr_val_store(struct device *dev,
>   	ret = kstrtoul(buf, 16, &val);
>   	if (ret)
>   		return ret;
> -	if (coresight_get_mode(drvdata->csdev))
> +	if (coresight_get_mode(drvdata->csdev) == CS_MODE_SYSFS)
>   		return -EBUSY;
>   
>   	spin_lock(&drvdata->spinlock);
> @@ -934,7 +972,8 @@ static DEVICE_ATTR_RW(seq_13_event);
>   static ssize_t seq_curr_state_show(struct device *dev,
>   				   struct device_attribute *attr, char *buf)
>   {
> -	unsigned long val, flags;
> +	int ret;
> +	unsigned long val;
>   	struct etm_drvdata *drvdata = dev_get_drvdata(dev->parent);
>   	struct etm_config *config = &drvdata->config;
>   
> @@ -943,15 +982,10 @@ static ssize_t seq_curr_state_show(struct device *dev,
>   		goto out;
>   	}
>   
> -	pm_runtime_get_sync(dev->parent);
> -	spin_lock_irqsave(&drvdata->spinlock, flags);
> -
> -	CS_UNLOCK(drvdata->csa.base);
> -	val = (etm_readl(drvdata, ETMSQR) & ETM_SQR_MASK);
> -	CS_LOCK(drvdata->csa.base);
> -
> -	spin_unlock_irqrestore(&drvdata->spinlock, flags);
> -	pm_runtime_put(dev->parent);
> +	ret = etm_readl_cslocked(drvdata, ETMSQR, &val);
> +	if (ret)
> +		return ret;
> +	val &= ETM_SQR_MASK;
>   out:
>   	return sprintf(buf, "%#lx\n", val);
>   }
> 



  reply	other threads:[~2026-09-18 13:58 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-15 11:34 [PATCH v11 0/9] fix several inconsistencies with sysfs configuration in etmX Yeoreum Yun
2026-09-15 11:34 ` [PATCH v11 1/9] coresight: etm4x: prohibit modifying ss_status and cntr_val while session is enabled Yeoreum Yun
2026-09-18 11:14   ` Mike Leach
2026-09-18 17:08     ` Yeoreum Yun
2026-09-15 11:34 ` [PATCH v11 2/9] coresight: etm3x: prohibit modifying cntr_val and reset " Yeoreum Yun
2026-09-18 11:15   ` Mike Leach
2026-09-15 11:34 ` [PATCH v11 3/9] coresight: etm4x: fix inconsistencies with sysfs configuration Yeoreum Yun
2026-09-18 13:49   ` Mike Leach
2026-09-18 17:00     ` Yeoreum Yun
2026-09-15 11:34 ` [PATCH v11 4/9] coresight: etm3x: " Yeoreum Yun
2026-09-18 13:57   ` Mike Leach [this message]
2026-09-18 17:09     ` Yeoreum Yun
2026-09-15 11:34 ` [PATCH v11 5/9] coresight: etm3x: remove redundant cpu online check on etm_enable_sysfs() Yeoreum Yun
2026-09-18 13:58   ` Mike Leach
2026-09-15 11:34 ` [PATCH v11 6/9] coresight: etm4x: introduce struct etm4_caps Yeoreum Yun
2026-09-18 14:01   ` Mike Leach
2026-09-15 11:34 ` [PATCH v11 7/9] coresight: etm4x: exclude ss_status from drvdata->config Yeoreum Yun
2026-09-18 14:04   ` Mike Leach
2026-09-18 17:13     ` Yeoreum Yun
2026-09-15 11:34 ` [PATCH v11 8/9] coresight: etm4x: remove s_ex_level from config Yeoreum Yun
2026-09-18 14:05   ` Mike Leach
2026-09-15 11:34 ` [PATCH v11 9/9] coresight: etm3x: introduce struct etm_caps Yeoreum Yun
2026-09-18 14:57   ` Mike Leach
2026-09-24 15:25 ` [PATCH v11 0/9] fix several inconsistencies with sysfs configuration in etmX Leo Yan
2026-09-24 17:28   ` Yeoreum Yun

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=9b7086de-df87-4d5c-8a7a-fc2f7c724c0b@arm.com \
    --to=mike.leach@arm.com \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=bigeasy@linutronix.de \
    --cc=clrkwllms@kernel.org \
    --cc=coresight@lists.linaro.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=james.clark@linaro.org \
    --cc=leo.yan@arm.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-rt-devel@lists.linux.dev \
    --cc=mathieu.poirier@linaro.org \
    --cc=rostedt@goodmis.org \
    --cc=suzuki.poulose@arm.com \
    --cc=yeoreum.yun@arm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox