lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20201211213141.GB1921322@xps15>
Date:   Fri, 11 Dec 2020 14:31:41 -0700
From:   Mathieu Poirier <mathieu.poirier@...aro.org>
To:     Anshuman Khandual <anshuman.khandual@....com>
Cc:     linux-arm-kernel@...ts.infradead.org, coresight@...ts.linaro.org,
        linux-kernel@...r.kernel.org, suzuki.poulose@....com,
        mike.leach@...aro.org
Subject: Re: [RFC 10/11] coresgith: etm-perf: Connect TRBE sink with ETE
 source

On Tue, Nov 10, 2020 at 06:15:08PM +0530, Anshuman Khandual wrote:
> Unlike traditional sink devices, individual TRBE instances are not detected
> via DT or ACPI nodes. Instead TRBE instances are detected during CPU online
> process. Hence a path connecting ETE and TRBE on a given CPU would not have
> been established until then. This adds two coresight helpers that will help
> modify outward connections from a source device to establish and terminate
> path to a given sink device. But this method might not be optimal and would
> be reworked later.
> 
> Signed-off-by: Anshuman Khandual <anshuman.khandual@....com>
> ---
>  drivers/hwtracing/coresight/coresight-etm-perf.c | 30 ++++++++++++++++++++++++
>  drivers/hwtracing/coresight/coresight-etm-perf.h |  4 ++++
>  drivers/hwtracing/coresight/coresight-platform.c |  3 ++-
>  drivers/hwtracing/coresight/coresight-trbe.c     |  2 ++
>  include/linux/coresight.h                        |  2 ++
>  5 files changed, 40 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/hwtracing/coresight/coresight-etm-perf.c b/drivers/hwtracing/coresight/coresight-etm-perf.c
> index 1a37991..b4ab1d4 100644
> --- a/drivers/hwtracing/coresight/coresight-etm-perf.c
> +++ b/drivers/hwtracing/coresight/coresight-etm-perf.c
> @@ -664,3 +664,33 @@ void __exit etm_perf_exit(void)
>  {
>  	perf_pmu_unregister(&etm_pmu);
>  }
> +
> +#ifdef CONFIG_CORESIGHT_TRBE
> +void coresight_trbe_connect_ete(struct coresight_device *csdev_trbe, int cpu)
> +{
> +	struct coresight_device *csdev_ete = per_cpu(csdev_src, cpu);

As Suzuki pointed out that won't work if the TRBE gets probed before the
ETMv4-ETE.  I also agree with Suzuki this situation should be better handled
with a per csdev_trbe that should be declared in the coresight-core.c file.
That way both sysfs and perf have access to it.  

> +
> +	if (!csdev_ete) {
> +		pr_err("Corresponding ETE device not present on cpu %d\n", cpu);
> +		return;
> +	}
> +	csdev_ete->def_sink = csdev_trbe;

That should be done in function coresight_find_default_sink().  If
per_cpu(csdev_trbe, cpu) exists then that's the what we pick.  If not then move
along with coresight_find_sink().


> +	csdev_ete->pdata->nr_outport++;
> +	if (!csdev_ete->pdata->conns)
> +		coresight_alloc_conns(&csdev_ete->dev, csdev_ete->pdata);
> +	csdev_ete->pdata->conns[csdev_ete->pdata->nr_outport - 1].child_dev = csdev_trbe;

I don't think we have to go through all that dance since the TRBE is directly
connected to the ETE.  With the above about coresight_find_default_sink() in
mind, all we need to do is fix coresight_build_path() to check if the sink
parameter is the same as csdev->def_sink.  If so then just add the sink to the
patch, no need to follow ports as we do for other classic components.

Thanks,
Mathieu

> +}
> +
> +void coresight_trbe_remove_ete(struct coresight_device *csdev_trbe, int cpu)
> +{
> +	struct coresight_device *csdev_ete = per_cpu(csdev_src, cpu);
> +
> +	if (!csdev_ete) {
> +		pr_err("Corresponding ETE device not present on cpu %d\n", cpu);
> +		return;
> +	}
> +	csdev_ete->pdata->conns[csdev_ete->pdata->nr_outport - 1].child_dev = NULL;
> +	csdev_ete->def_sink = NULL;
> +	csdev_ete->pdata->nr_outport--;
> +}
> +#endif
> diff --git a/drivers/hwtracing/coresight/coresight-etm-perf.h b/drivers/hwtracing/coresight/coresight-etm-perf.h
> index 3e4f2ad..20386cf 100644
> --- a/drivers/hwtracing/coresight/coresight-etm-perf.h
> +++ b/drivers/hwtracing/coresight/coresight-etm-perf.h
> @@ -85,4 +85,8 @@ static inline void *etm_perf_sink_config(struct perf_output_handle *handle)
>  int __init etm_perf_init(void);
>  void __exit etm_perf_exit(void);
>  
> +#ifdef CONFIG_CORESIGHT_TRBE
> +void coresight_trbe_connect_ete(struct coresight_device *csdev, int cpu);
> +void coresight_trbe_remove_ete(struct coresight_device *csdev, int cpu);
> +#endif
>  #endif
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index c594f45..8fa7406 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -23,7 +23,7 @@
>   * coresight_alloc_conns: Allocate connections record for each output
>   * port from the device.
>   */
> -static int coresight_alloc_conns(struct device *dev,
> +int coresight_alloc_conns(struct device *dev,
>  				 struct coresight_platform_data *pdata)
>  {
>  	if (pdata->nr_outport) {
> @@ -35,6 +35,7 @@ static int coresight_alloc_conns(struct device *dev,
>  
>  	return 0;
>  }
> +EXPORT_SYMBOL_GPL(coresight_alloc_conns);
>  
>  static struct device *
>  coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
> diff --git a/drivers/hwtracing/coresight/coresight-trbe.c b/drivers/hwtracing/coresight/coresight-trbe.c
> index 48a8ec3..afd1a1c 100644
> --- a/drivers/hwtracing/coresight/coresight-trbe.c
> +++ b/drivers/hwtracing/coresight/coresight-trbe.c
> @@ -507,6 +507,7 @@ static void arm_trbe_probe_coresight_cpu(void *info)
>  	if (IS_ERR(cpudata->csdev))
>  		goto cpu_clear;
>  
> +	coresight_trbe_connect_ete(cpudata->csdev, cpudata->cpu);
>  	dev_set_drvdata(&cpudata->csdev->dev, cpudata);
>  	cpudata->trbe_dbm = get_trbe_flag_update();
>  	cpudata->trbe_align = 1ULL << get_trbe_address_align();
> @@ -586,6 +587,7 @@ static int arm_trbe_cpu_teardown(unsigned int cpu, struct hlist_node *node)
>  
>  	if (cpumask_test_cpu(cpu, &drvdata->supported_cpus)) {
>  		cpudata = per_cpu_ptr(drvdata->cpudata, cpu);
> +		coresight_trbe_remove_ete(cpudata->csdev, cpu);
>  		if (cpudata->csdev) {
>  			coresight_unregister(cpudata->csdev);
>  			cpudata->drvdata = NULL;
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index c2d0a2a..c657813 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -496,6 +496,8 @@ void coresight_relaxed_write64(struct coresight_device *csdev,
>  			       u64 val, u32 offset);
>  void coresight_write64(struct coresight_device *csdev, u64 val, u32 offset);
>  
> +int coresight_alloc_conns(struct device *dev,
> +			  struct coresight_platform_data *pdata);
>  
>  #else
>  static inline struct coresight_device *
> -- 
> 2.7.4
> 

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ