Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1267482 > unrolled thread
| Started by | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| First post | 2015-11-11 23:20 +0100 |
| Last post | 2015-11-12 12:00 +0100 |
| Articles | 20 on this page of 27 — 3 participants |
Back to article view | Back to linux.kernel
[PATCH V3 00/26] Coresight integration with perf Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:20 +0100
[PATCH V3 07/26] coresight: moving PM runtime operations to core framework Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 17/26] coresight: etb10: moving to local atomic operations Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 24/26] perf tools: adding perf_session to *info_prive_size() Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 05/26] coresight: etm3x: implementing 'cpu_id()' API Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 08/26] coresight: etm3x: adding operation mode for etm_enable() Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 23/26] perf tools: making function set_max_cpu_num() non static Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 20/26] coresight: etm-perf: new PMU driver for ETM tracers Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 06/26] coresight: associating path with session rather than tracer Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 11/26] coresight: etm3x: changing default trace configuration Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 25/26] perf tools: making coresight PMU listable Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 12/26] coresight: etm3x: consolidating initial config Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 19/26] coresight: etb10: implementing AUX space API Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 10/26] coresight: etm3x: set progbit to stop trace collection Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 03/26] coresight: etm3x: unlocking tracers in default arch init Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 14/26] coresight: etm3x: adding perf_get/set_config() API Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 09/26] coresight: add API to get sink from path Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 16/26] perf: changing pmu::setup_aux() parameter to include event Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 18/26] coresight: adding operation mode for sink->enable() Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 01/26] coresight: etm3x: moving etm_readl/writel to header file Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 26/26] perf tools: adding coresight etm PMU record capabilities Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 21/26] coresight: introducing a global trace ID function Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
[PATCH V3 22/26] coresight: updating documentation to reflect integration with perf Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-11 23:30 +0100
Re: [PATCH V3 02/26] coresight: etm3x: moving sysFS entries to dedicated file Arnd Bergmann <arnd@arndb.de> - 2015-11-11 23:40 +0100
Re: [PATCH V3 02/26] coresight: etm3x: moving sysFS entries to dedicated file Mathieu Poirier <mathieu.poirier@linaro.org> - 2015-11-13 17:20 +0100
Re: [PATCH V3 02/26] coresight: etm3x: moving sysFS entries to dedicated file Arnd Bergmann <arnd@arndb.de> - 2015-11-13 17:30 +0100
Re: [PATCH V3 00/26] Coresight integration with perf Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-11-12 12:00 +0100
Page 1 of 2 [1] 2 Next page →
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:20 +0100 |
| Subject | [PATCH V3 00/26] Coresight integration with perf |
| Message-ID | <qtLUd-hJ-3@gated-at.bofh.it> |
This patchset aims to integrate configuration and control of the Coresight tracers with the perf sub-system. The goal is to use PMUs to represent tracers and the auxiliary buffer enhancement to collect processor traces. As such a lot of work is done to move the current Coresight sysFS oriented configuration and control commands to perf's AUX API. For the time being the work concentrates on ETMv3 and ETB1.0 sink buffer. Work on ETMv4 and other type of sink buffers will follow once a foundation has been established. Best regards, Mathieu Changes since V2: * Rebased to v4.3. * Used the -M option to "git format-patch". * Decoupled tracers and path - both are now completely independent. * Moved path enablement/release to Perf's add/del functions. * Moved PM runtime operations to the coresight core rather than individual drivers. * Enhanced mutual exclusion scheme between perf and sysFS. * Removed architecture specific define in perf cmd line tool. * Got rid of unused fields in structure 'cs_buffers'. * Enhanced changelog for patch 16/26. Changes since V1: * Fixed typos in typographical error in documentation. * Moved to a multi session support scheme. * Split static and dynamic tracer configuration. * Fixed configuration for user and kernel space tracing. * Using WARN_ON_ONCE() rather than WARN_ON(). * Implemented strategy to prevent tracers from being used simultaneously. * Changed sink_ops::unset_buffer() to sink_ops::reset_buffer(). * Moves ETM's sysFS interface from driver core to dedicated file. * Removed spinlock in "etm_cpu_id()". * Aggregated PMU driver pieces in a single patch. * Added user space changes and rebased everything to v4.3-rc5. Mathieu Poirier (26): coresight: etm3x: moving etm_readl/writel to header file coresight: etm3x: moving sysFS entries to dedicated file coresight: etm3x: unlocking tracers in default arch init coresight: etm3x: splitting struct etm_drvdata coresight: etm3x: implementing 'cpu_id()' API coresight: associating path with session rather than tracer coresight: moving PM runtime operations to core framework coresight: etm3x: adding operation mode for etm_enable() coresight: add API to get sink from path coresight: etm3x: set progbit to stop trace collection coresight: etm3x: changing default trace configuration coresight: etm3x: consolidating initial config coresight: etm3x: implementing user/kernel mode tracing coresight: etm3x: adding perf_get/set_config() API coresight: etm3x: implementing perf_enable/disable() API perf: changing pmu::setup_aux() parameter to include event coresight: etb10: moving to local atomic operations coresight: adding operation mode for sink->enable() coresight: etb10: implementing AUX space API coresight: etm-perf: new PMU driver for ETM tracers coresight: introducing a global trace ID function coresight: updating documentation to reflect integration with perf perf tools: making function set_max_cpu_num() non static perf tools: adding perf_session to *info_prive_size() perf tools: making coresight PMU listable perf tools: adding coresight etm PMU record capabilities Documentation/trace/coresight.txt | 138 +- MAINTAINERS | 4 + arch/x86/kernel/cpu/perf_event_intel_bts.c | 4 +- arch/x86/kernel/cpu/perf_event_intel_pt.c | 5 +- drivers/hwtracing/coresight/Kconfig | 1 + drivers/hwtracing/coresight/Makefile | 4 +- drivers/hwtracing/coresight/coresight-etb10.c | 268 ++- drivers/hwtracing/coresight/coresight-etm-perf.c | 438 +++++ drivers/hwtracing/coresight/coresight-etm-perf.h | 32 + drivers/hwtracing/coresight/coresight-etm.h | 143 +- .../hwtracing/coresight/coresight-etm3x-sysfs.c | 1442 +++++++++++++++ drivers/hwtracing/coresight/coresight-etm3x.c | 1875 +++++--------------- drivers/hwtracing/coresight/coresight-etm4x.c | 8 +- drivers/hwtracing/coresight/coresight-funnel.c | 2 - drivers/hwtracing/coresight/coresight-priv.h | 15 + .../coresight/coresight-replicator-qcom.c | 4 - drivers/hwtracing/coresight/coresight-replicator.c | 2 - drivers/hwtracing/coresight/coresight-tmc.c | 7 +- drivers/hwtracing/coresight/coresight-tpiu.c | 4 +- drivers/hwtracing/coresight/coresight.c | 330 ++-- include/linux/coresight-pmu.h | 39 + include/linux/coresight.h | 36 +- include/linux/perf_event.h | 2 +- kernel/events/ring_buffer.c | 2 +- tools/perf/arch/arm/util/Build | 2 + tools/perf/arch/arm/util/auxtrace.c | 48 + tools/perf/arch/arm/util/cs_etm.c | 466 +++++ tools/perf/arch/arm/util/cs_etm.h | 44 + tools/perf/arch/arm/util/pmu.c | 34 + tools/perf/arch/x86/util/intel-bts.c | 4 +- tools/perf/arch/x86/util/intel-pt.c | 4 +- tools/perf/config/Makefile | 17 +- tools/perf/util/auxtrace.c | 8 +- tools/perf/util/auxtrace.h | 7 +- tools/perf/util/cpumap.c | 2 +- tools/perf/util/cpumap.h | 1 + 36 files changed, 3787 insertions(+), 1655 deletions(-) create mode 100644 drivers/hwtracing/coresight/coresight-etm-perf.c create mode 100644 drivers/hwtracing/coresight/coresight-etm-perf.h create mode 100644 drivers/hwtracing/coresight/coresight-etm3x-sysfs.c create mode 100644 include/linux/coresight-pmu.h create mode 100644 tools/perf/arch/arm/util/auxtrace.c create mode 100644 tools/perf/arch/arm/util/cs_etm.c create mode 100644 tools/perf/arch/arm/util/cs_etm.h create mode 100644 tools/perf/arch/arm/util/pmu.c -- 1.9.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[toc] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 07/26] coresight: moving PM runtime operations to core framework |
| Message-ID | <qtM3T-lk-1@gated-at.bofh.it> |
| In reply to | #1267482 |
Moving PM runtime operations in Coresight devices enable() and
disable() API to the framework core when a path is setup. That
way the runtime core doesn't have to be involved everytime a
path is enabled. It also avoids calling runtime PM operations
in IRQ context.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etb10.c | 4 ----
drivers/hwtracing/coresight/coresight-etm3x.c | 3 ---
drivers/hwtracing/coresight/coresight-etm4x.c | 6 ------
drivers/hwtracing/coresight/coresight-funnel.c | 2 --
drivers/hwtracing/coresight/coresight-replicator-qcom.c | 4 ----
drivers/hwtracing/coresight/coresight-replicator.c | 2 --
drivers/hwtracing/coresight/coresight-tmc.c | 5 -----
drivers/hwtracing/coresight/coresight-tpiu.c | 2 --
drivers/hwtracing/coresight/coresight.c | 9 ++++++++-
9 files changed, 8 insertions(+), 29 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etb10.c b/drivers/hwtracing/coresight/coresight-etb10.c
index 77d0f9c1118d..1301edc44629 100644
--- a/drivers/hwtracing/coresight/coresight-etb10.c
+++ b/drivers/hwtracing/coresight/coresight-etb10.c
@@ -137,8 +137,6 @@ static int etb_enable(struct coresight_device *csdev)
struct etb_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
unsigned long flags;
- pm_runtime_get_sync(drvdata->dev);
-
spin_lock_irqsave(&drvdata->spinlock, flags);
etb_enable_hw(drvdata);
drvdata->enable = true;
@@ -247,8 +245,6 @@ static void etb_disable(struct coresight_device *csdev)
drvdata->enable = false;
spin_unlock_irqrestore(&drvdata->spinlock, flags);
- pm_runtime_put(drvdata->dev);
-
dev_info(drvdata->dev, "ETB disabled\n");
}
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index 74526f044a06..660c1e07601b 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -358,7 +358,6 @@ static int etm_enable(struct coresight_device *csdev)
struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
int ret;
- pm_runtime_get_sync(csdev->dev.parent);
spin_lock(&drvdata->spinlock);
/*
@@ -382,7 +381,6 @@ static int etm_enable(struct coresight_device *csdev)
return 0;
err:
spin_unlock(&drvdata->spinlock);
- pm_runtime_put(csdev->dev.parent);
return ret;
}
@@ -439,7 +437,6 @@ static void etm_disable(struct coresight_device *csdev)
spin_unlock(&drvdata->spinlock);
put_online_cpus();
- pm_runtime_put(csdev->dev.parent);
dev_info(drvdata->dev, "ETM tracing disabled\n");
}
diff --git a/drivers/hwtracing/coresight/coresight-etm4x.c b/drivers/hwtracing/coresight/coresight-etm4x.c
index 254a81a4e6f4..a62e94814c37 100644
--- a/drivers/hwtracing/coresight/coresight-etm4x.c
+++ b/drivers/hwtracing/coresight/coresight-etm4x.c
@@ -72,7 +72,6 @@ static int etm4_trace_id(struct coresight_device *csdev)
if (!drvdata->enable)
return drvdata->trcid;
- pm_runtime_get_sync(drvdata->dev);
spin_lock_irqsave(&drvdata->spinlock, flags);
CS_UNLOCK(drvdata->base);
@@ -81,7 +80,6 @@ static int etm4_trace_id(struct coresight_device *csdev)
CS_LOCK(drvdata->base);
spin_unlock_irqrestore(&drvdata->spinlock, flags);
- pm_runtime_put(drvdata->dev);
return trace_id;
}
@@ -185,7 +183,6 @@ static int etm4_enable(struct coresight_device *csdev)
struct etmv4_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
int ret;
- pm_runtime_get_sync(drvdata->dev);
spin_lock(&drvdata->spinlock);
/*
@@ -205,7 +202,6 @@ static int etm4_enable(struct coresight_device *csdev)
return 0;
err:
spin_unlock(&drvdata->spinlock);
- pm_runtime_put(drvdata->dev);
return ret;
}
@@ -254,8 +250,6 @@ static void etm4_disable(struct coresight_device *csdev)
spin_unlock(&drvdata->spinlock);
put_online_cpus();
- pm_runtime_put(drvdata->dev);
-
dev_info(drvdata->dev, "ETM tracing disabled\n");
}
diff --git a/drivers/hwtracing/coresight/coresight-funnel.c b/drivers/hwtracing/coresight/coresight-funnel.c
index 2e36bde7fcb4..a47bba361833 100644
--- a/drivers/hwtracing/coresight/coresight-funnel.c
+++ b/drivers/hwtracing/coresight/coresight-funnel.c
@@ -69,7 +69,6 @@ static int funnel_enable(struct coresight_device *csdev, int inport,
{
struct funnel_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
- pm_runtime_get_sync(drvdata->dev);
funnel_enable_hw(drvdata, inport);
dev_info(drvdata->dev, "FUNNEL inport %d enabled\n", inport);
@@ -95,7 +94,6 @@ static void funnel_disable(struct coresight_device *csdev, int inport,
struct funnel_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
funnel_disable_hw(drvdata, inport);
- pm_runtime_put(drvdata->dev);
dev_info(drvdata->dev, "FUNNEL inport %d disabled\n", inport);
}
diff --git a/drivers/hwtracing/coresight/coresight-replicator-qcom.c b/drivers/hwtracing/coresight/coresight-replicator-qcom.c
index 584059e9e866..8149087e8966 100644
--- a/drivers/hwtracing/coresight/coresight-replicator-qcom.c
+++ b/drivers/hwtracing/coresight/coresight-replicator-qcom.c
@@ -48,8 +48,6 @@ static int replicator_enable(struct coresight_device *csdev, int inport,
{
struct replicator_state *drvdata = dev_get_drvdata(csdev->dev.parent);
- pm_runtime_get_sync(drvdata->dev);
-
CS_UNLOCK(drvdata->base);
/*
@@ -86,8 +84,6 @@ static void replicator_disable(struct coresight_device *csdev, int inport,
CS_LOCK(drvdata->base);
- pm_runtime_put(drvdata->dev);
-
dev_info(drvdata->dev, "REPLICATOR disabled\n");
}
diff --git a/drivers/hwtracing/coresight/coresight-replicator.c b/drivers/hwtracing/coresight/coresight-replicator.c
index 963ac197c253..a0fbb2e05389 100644
--- a/drivers/hwtracing/coresight/coresight-replicator.c
+++ b/drivers/hwtracing/coresight/coresight-replicator.c
@@ -41,7 +41,6 @@ static int replicator_enable(struct coresight_device *csdev, int inport,
{
struct replicator_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
- pm_runtime_get_sync(drvdata->dev);
dev_info(drvdata->dev, "REPLICATOR enabled\n");
return 0;
}
@@ -51,7 +50,6 @@ static void replicator_disable(struct coresight_device *csdev, int inport,
{
struct replicator_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
- pm_runtime_put(drvdata->dev);
dev_info(drvdata->dev, "REPLICATOR disabled\n");
}
diff --git a/drivers/hwtracing/coresight/coresight-tmc.c b/drivers/hwtracing/coresight/coresight-tmc.c
index a57c7ec1661f..5e2a71767870 100644
--- a/drivers/hwtracing/coresight/coresight-tmc.c
+++ b/drivers/hwtracing/coresight/coresight-tmc.c
@@ -242,12 +242,9 @@ static int tmc_enable(struct tmc_drvdata *drvdata, enum tmc_mode mode)
{
unsigned long flags;
- pm_runtime_get_sync(drvdata->dev);
-
spin_lock_irqsave(&drvdata->spinlock, flags);
if (drvdata->reading) {
spin_unlock_irqrestore(&drvdata->spinlock, flags);
- pm_runtime_put(drvdata->dev);
return -EBUSY;
}
@@ -381,8 +378,6 @@ out:
drvdata->enable = false;
spin_unlock_irqrestore(&drvdata->spinlock, flags);
- pm_runtime_put(drvdata->dev);
-
dev_info(drvdata->dev, "TMC disabled\n");
}
diff --git a/drivers/hwtracing/coresight/coresight-tpiu.c b/drivers/hwtracing/coresight/coresight-tpiu.c
index 7214efd10db5..e19b86e61c38 100644
--- a/drivers/hwtracing/coresight/coresight-tpiu.c
+++ b/drivers/hwtracing/coresight/coresight-tpiu.c
@@ -74,7 +74,6 @@ static int tpiu_enable(struct coresight_device *csdev)
{
struct tpiu_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
- pm_runtime_get_sync(csdev->dev.parent);
tpiu_enable_hw(drvdata);
dev_info(drvdata->dev, "TPIU enabled\n");
@@ -98,7 +97,6 @@ static void tpiu_disable(struct coresight_device *csdev)
struct tpiu_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
tpiu_disable_hw(drvdata);
- pm_runtime_put(csdev->dev.parent);
dev_info(drvdata->dev, "TPIU disabled\n");
}
diff --git a/drivers/hwtracing/coresight/coresight.c b/drivers/hwtracing/coresight/coresight.c
index 4beea67fd5a1..a4103a216e33 100644
--- a/drivers/hwtracing/coresight/coresight.c
+++ b/drivers/hwtracing/coresight/coresight.c
@@ -24,6 +24,7 @@
#include <linux/coresight.h>
#include <linux/of_platform.h>
#include <linux/delay.h>
+#include <linux/pm_runtime.h>
#include "coresight-priv.h"
@@ -361,7 +362,8 @@ out:
/*
* A path from this element to a sink has been found. The elements
* leading to the sink are already enqueued, all that is left to do
- * is add a node for this element.
+ * is tell the PM runtime core we need this element and add a node
+ * for it.
*/
node = kzalloc(sizeof(struct coresight_node), GFP_KERNEL);
if (!node)
@@ -369,6 +371,7 @@ out:
node->csdev = csdev;
list_add(&node->link, path);
+ pm_runtime_get_sync(csdev->dev.parent);
return 0;
}
@@ -400,9 +403,13 @@ struct list_head *coresight_build_path(struct coresight_device *csdev)
*/
void coresight_release_path(struct list_head *path)
{
+ struct coresight_device *csdev;
struct coresight_node *nd, *next;
list_for_each_entry_safe(nd, next, path, link) {
+ csdev = nd->csdev;
+
+ pm_runtime_put_sync(csdev->dev.parent);
list_del(&nd->link);
kfree(nd);
}
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 17/26] coresight: etb10: moving to local atomic operations |
| Message-ID | <qtM3T-lk-3@gated-at.bofh.it> |
| In reply to | #1267482 |
Moving to use local atomic operations to take advantage of the
lockless implementation, something that will come handy when
the ETB is accessed from the Perf subsystem. Also changing the
name of the variable to something more meaningful.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etb10.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etb10.c b/drivers/hwtracing/coresight/coresight-etb10.c
index 1301edc44629..09e306361e4f 100644
--- a/drivers/hwtracing/coresight/coresight-etb10.c
+++ b/drivers/hwtracing/coresight/coresight-etb10.c
@@ -10,6 +10,7 @@
* GNU General Public License for more details.
*/
+#include <asm/local.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
@@ -71,7 +72,7 @@
* @csdev: component vitals needed by the framework.
* @miscdev: specifics to handle "/dev/xyz.etb" entry.
* @spinlock: only one at a time pls.
- * @in_use: synchronise user space access to etb buffer.
+ * @reading: synchronise user space access to etb buffer.
* @buf: area of memory where ETB buffer content gets sent.
* @buffer_depth: size of @buf.
* @enable: this ETB is being used.
@@ -84,7 +85,7 @@ struct etb_drvdata {
struct coresight_device *csdev;
struct miscdevice miscdev;
spinlock_t spinlock;
- atomic_t in_use;
+ local_t reading;
u8 *buf;
u32 buffer_depth;
bool enable;
@@ -277,7 +278,7 @@ static int etb_open(struct inode *inode, struct file *file)
struct etb_drvdata *drvdata = container_of(file->private_data,
struct etb_drvdata, miscdev);
- if (atomic_cmpxchg(&drvdata->in_use, 0, 1))
+ if (local_cmpxchg(&drvdata->reading, 0, 1))
return -EBUSY;
dev_dbg(drvdata->dev, "%s: successfully opened\n", __func__);
@@ -313,7 +314,7 @@ static int etb_release(struct inode *inode, struct file *file)
{
struct etb_drvdata *drvdata = container_of(file->private_data,
struct etb_drvdata, miscdev);
- atomic_set(&drvdata->in_use, 0);
+ local_set(&drvdata->reading, 0);
dev_dbg(drvdata->dev, "%s: released\n", __func__);
return 0;
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 24/26] perf tools: adding perf_session to *info_prive_size() |
| Message-ID | <qtM3T-lk-7@gated-at.bofh.it> |
| In reply to | #1267482 |
On some architecture the size of the private header may
be dependent on the number of tracers used in the session. As
such adding a "struct perf_session" parameter, which should
contain all the required information.
Also adjusting the existing client of the interface to take
the new parameter into account.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
tools/perf/arch/x86/util/intel-bts.c | 4 +++-
tools/perf/arch/x86/util/intel-pt.c | 4 +++-
tools/perf/util/auxtrace.c | 7 ++++---
tools/perf/util/auxtrace.h | 6 ++++--
4 files changed, 14 insertions(+), 7 deletions(-)
diff --git a/tools/perf/arch/x86/util/intel-bts.c b/tools/perf/arch/x86/util/intel-bts.c
index 9b94ce520917..6efdc5e86d97 100644
--- a/tools/perf/arch/x86/util/intel-bts.c
+++ b/tools/perf/arch/x86/util/intel-bts.c
@@ -60,7 +60,9 @@ struct branch {
u64 misc;
};
-static size_t intel_bts_info_priv_size(struct auxtrace_record *itr __maybe_unused)
+static size_t
+intel_bts_info_priv_size(struct auxtrace_record *itr __maybe_unused,
+ struct perf_session *session __maybe_unused)
{
return INTEL_BTS_AUXTRACE_PRIV_SIZE;
}
diff --git a/tools/perf/arch/x86/util/intel-pt.c b/tools/perf/arch/x86/util/intel-pt.c
index 2ca10d796c0b..1fd9a16f668c 100644
--- a/tools/perf/arch/x86/util/intel-pt.c
+++ b/tools/perf/arch/x86/util/intel-pt.c
@@ -273,7 +273,9 @@ intel_pt_pmu_default_config(struct perf_pmu *intel_pt_pmu)
return attr;
}
-static size_t intel_pt_info_priv_size(struct auxtrace_record *itr __maybe_unused)
+static size_t
+intel_pt_info_priv_size(struct auxtrace_record *itr __maybe_unused,
+ struct perf_session *session __maybe_unused)
{
return INTEL_PT_AUXTRACE_PRIV_SIZE;
}
diff --git a/tools/perf/util/auxtrace.c b/tools/perf/util/auxtrace.c
index a980e7c50ee0..430df9575dbe 100644
--- a/tools/perf/util/auxtrace.c
+++ b/tools/perf/util/auxtrace.c
@@ -478,10 +478,11 @@ void auxtrace_heap__pop(struct auxtrace_heap *heap)
heap_array[last].ordinal);
}
-size_t auxtrace_record__info_priv_size(struct auxtrace_record *itr)
+size_t auxtrace_record__info_priv_size(struct auxtrace_record *itr,
+ struct perf_session *session)
{
if (itr)
- return itr->info_priv_size(itr);
+ return itr->info_priv_size(itr, session);
return 0;
}
@@ -852,7 +853,7 @@ int perf_event__synthesize_auxtrace_info(struct auxtrace_record *itr,
int err;
pr_debug2("Synthesizing auxtrace information\n");
- priv_size = auxtrace_record__info_priv_size(itr);
+ priv_size = auxtrace_record__info_priv_size(itr, session);
ev = zalloc(sizeof(struct auxtrace_info_event) + priv_size);
if (!ev)
return -ENOMEM;
diff --git a/tools/perf/util/auxtrace.h b/tools/perf/util/auxtrace.h
index bf72b77a588a..d76177169484 100644
--- a/tools/perf/util/auxtrace.h
+++ b/tools/perf/util/auxtrace.h
@@ -289,7 +289,8 @@ struct auxtrace_record {
int (*recording_options)(struct auxtrace_record *itr,
struct perf_evlist *evlist,
struct record_opts *opts);
- size_t (*info_priv_size)(struct auxtrace_record *itr);
+ size_t (*info_priv_size)(struct auxtrace_record *itr,
+ struct perf_session *session);
int (*info_fill)(struct auxtrace_record *itr,
struct perf_session *session,
struct auxtrace_info_event *auxtrace_info,
@@ -425,7 +426,8 @@ int auxtrace_parse_snapshot_options(struct auxtrace_record *itr,
int auxtrace_record__options(struct auxtrace_record *itr,
struct perf_evlist *evlist,
struct record_opts *opts);
-size_t auxtrace_record__info_priv_size(struct auxtrace_record *itr);
+size_t auxtrace_record__info_priv_size(struct auxtrace_record *itr,
+ struct perf_session *session);
int auxtrace_record__info_fill(struct auxtrace_record *itr,
struct perf_session *session,
struct auxtrace_info_event *auxtrace_info,
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 05/26] coresight: etm3x: implementing 'cpu_id()' API |
| Message-ID | <qtM3T-lk-17@gated-at.bofh.it> |
| In reply to | #1267482 |
Other than plainly parsing the device tree there is no way to
know which CPU a tracer is affined to. As such adding an
interface to lookup the CPU field enclosed in the etm_drvdata
structure that was initialised at boot time.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etm3x.c | 8 ++++++++
include/linux/coresight.h | 3 +++
2 files changed, 11 insertions(+)
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index 67af548a874d..74526f044a06 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -312,6 +312,13 @@ static void etm_enable_hw(void *info)
dev_dbg(drvdata->dev, "cpu: %d enable smp call done\n", drvdata->cpu);
}
+static int etm_cpu_id(struct coresight_device *csdev)
+{
+ struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+
+ return drvdata->cpu;
+}
+
int etm_get_trace_id(struct etm_drvdata *drvdata)
{
unsigned long flags;
@@ -438,6 +445,7 @@ static void etm_disable(struct coresight_device *csdev)
}
static const struct coresight_ops_source etm_source_ops = {
+ .cpu_id = etm_cpu_id,
.trace_id = etm_trace_id,
.enable = etm_enable,
.disable = etm_disable,
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index a7cabfa23b55..bf62b265bf52 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -205,12 +205,15 @@ struct coresight_ops_link {
/**
* struct coresight_ops_source - basic operations for a source
* Operations available for sources.
+ * @cpu_id: returns the value of the CPU number this component
+ * is associated to.
* @trace_id: returns the value of the component's trace ID as known
to the HW.
* @enable: enables tracing for a source.
* @disable: disables tracing for a source.
*/
struct coresight_ops_source {
+ int (*cpu_id)(struct coresight_device *csdev);
int (*trace_id)(struct coresight_device *csdev);
int (*enable)(struct coresight_device *csdev);
void (*disable)(struct coresight_device *csdev);
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 08/26] coresight: etm3x: adding operation mode for etm_enable() |
| Message-ID | <qtM3U-lk-19@gated-at.bofh.it> |
| In reply to | #1267482 |
Adding a new mode to source API enable() in order to
distinguish where the request comes from. That way it is
possible to perform different operations based on where
the request was issued from.
The ETM4x driver is also modified to keep in sync with the
new interface.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etm.h | 5 +-
.../hwtracing/coresight/coresight-etm3x-sysfs.c | 4 +-
drivers/hwtracing/coresight/coresight-etm3x.c | 68 +++++++++++++++++++---
drivers/hwtracing/coresight/coresight-etm4x.c | 2 +-
drivers/hwtracing/coresight/coresight-priv.h | 6 ++
drivers/hwtracing/coresight/coresight.c | 6 +-
include/linux/coresight.h | 2 +-
7 files changed, 76 insertions(+), 17 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etm.h b/drivers/hwtracing/coresight/coresight-etm.h
index 9317510d555f..a8663780a908 100644
--- a/drivers/hwtracing/coresight/coresight-etm.h
+++ b/drivers/hwtracing/coresight/coresight-etm.h
@@ -13,6 +13,7 @@
#ifndef _CORESIGHT_CORESIGHT_ETM_H
#define _CORESIGHT_CORESIGHT_ETM_H
+#include <asm/local.h>
#include <linux/spinlock.h>
#include "coresight-priv.h"
@@ -214,7 +215,7 @@ struct etm_config {
* @port_size: port size as reported by ETMCR bit 4-6 and 21.
* @arch: ETM/PTM version number.
* @use_cpu14: true if management registers need to be accessed via CP14.
- * @enable: is this ETM/PTM currently tracing.
+ * @mode: this tracer's mode, i.e sysFS, Perf or disabled.
* @sticky_enable: true if ETM base configuration has been done.
* @boot_enable:true if we should start tracing at boot time.
* @os_unlock: true if access to management registers is allowed.
@@ -238,7 +239,7 @@ struct etm_drvdata {
int port_size;
u8 arch;
bool use_cp14;
- bool enable;
+ local_t mode;
bool sticky_enable;
bool boot_enable;
bool os_unlock;
diff --git a/drivers/hwtracing/coresight/coresight-etm3x-sysfs.c b/drivers/hwtracing/coresight/coresight-etm3x-sysfs.c
index 410d29e2b90c..6da50d94ec6f 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x-sysfs.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x-sysfs.c
@@ -811,7 +811,7 @@ static ssize_t cntr_val_show(struct device *dev,
if (WARN_ON_ONCE(!config))
return -EINVAL;
- if (!drvdata->enable) {
+ if (!local_read(&drvdata->mode)) {
spin_lock(&drvdata->spinlock);
for (i = 0; i < drvdata->nr_cntr; i++)
ret += sprintf(buf, "counter %d: %x\n",
@@ -1072,7 +1072,7 @@ static ssize_t seq_curr_state_show(struct device *dev,
if (WARN_ON_ONCE(!config))
return -EINVAL;
- if (!drvdata->enable) {
+ if (!local_read(&drvdata->mode)) {
val = config->seq_curr_state;
goto out;
}
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index 660c1e07601b..5a2be0590754 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -327,7 +327,7 @@ int etm_get_trace_id(struct etm_drvdata *drvdata)
if (!drvdata)
goto out;
- if (!drvdata->enable)
+ if (!local_read(&drvdata->mode))
return drvdata->traceid;
pm_runtime_get_sync(drvdata->dev);
@@ -353,7 +353,7 @@ static int etm_trace_id(struct coresight_device *csdev)
return etm_get_trace_id(drvdata);
}
-static int etm_enable(struct coresight_device *csdev)
+static int etm_enable_sysfs(struct coresight_device *csdev)
{
struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
int ret;
@@ -372,18 +372,44 @@ static int etm_enable(struct coresight_device *csdev)
goto err;
}
- drvdata->enable = true;
drvdata->sticky_enable = true;
-
spin_unlock(&drvdata->spinlock);
dev_info(drvdata->dev, "ETM tracing enabled\n");
return 0;
+
err:
spin_unlock(&drvdata->spinlock);
return ret;
}
+static int etm_enable(struct coresight_device *csdev, u32 mode)
+{
+ int ret;
+ u32 val;
+ struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+
+ val = local_cmpxchg(&drvdata->mode, CS_MODE_DISABLED, mode);
+
+ /* Someone is already using the tracer */
+ if (val)
+ return -EBUSY;
+
+ switch (mode) {
+ case CS_MODE_SYSFS:
+ ret = etm_enable_sysfs(csdev);
+ break;
+ default:
+ ret = -EINVAL;
+ }
+
+ /* The tracer didn't start */
+ if (ret)
+ local_set(&drvdata->mode, CS_MODE_DISABLED);
+
+ return ret;
+}
+
static void etm_disable_hw(void *info)
{
int i;
@@ -415,7 +441,7 @@ static void etm_disable_hw(void *info)
dev_dbg(drvdata->dev, "cpu: %d disable smp call done\n", drvdata->cpu);
}
-static void etm_disable(struct coresight_device *csdev)
+static void etm_disable_sysfs(struct coresight_device *csdev)
{
struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
@@ -433,7 +459,6 @@ static void etm_disable(struct coresight_device *csdev)
* ensures that register writes occur when cpu is powered.
*/
smp_call_function_single(drvdata->cpu, etm_disable_hw, drvdata, 1);
- drvdata->enable = false;
spin_unlock(&drvdata->spinlock);
put_online_cpus();
@@ -441,6 +466,33 @@ static void etm_disable(struct coresight_device *csdev)
dev_info(drvdata->dev, "ETM tracing disabled\n");
}
+static void etm_disable(struct coresight_device *csdev)
+{
+ u32 mode;
+ struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+
+ /*
+ * For as long as the tracer isn't disabled another entity can't
+ * change its status. As such we can read the status here without
+ * fearing it will change under us.
+ */
+ mode = local_read(&drvdata->mode);
+
+ switch (mode) {
+ case CS_MODE_DISABLED:
+ break;
+ case CS_MODE_SYSFS:
+ etm_disable_sysfs(csdev);
+ break;
+ default:
+ WARN_ON_ONCE(mode);
+ return;
+ }
+
+ if (mode)
+ local_set(&drvdata->mode, CS_MODE_DISABLED);
+}
+
static const struct coresight_ops_source etm_source_ops = {
.cpu_id = etm_cpu_id,
.trace_id = etm_trace_id,
@@ -468,7 +520,7 @@ static int etm_cpu_callback(struct notifier_block *nfb, unsigned long action,
etmdrvdata[cpu]->os_unlock = true;
}
- if (etmdrvdata[cpu]->enable)
+ if (local_read(&etmdrvdata[cpu]->mode))
etm_enable_hw(etmdrvdata[cpu]);
spin_unlock(&etmdrvdata[cpu]->spinlock);
break;
@@ -481,7 +533,7 @@ static int etm_cpu_callback(struct notifier_block *nfb, unsigned long action,
case CPU_DYING:
spin_lock(&etmdrvdata[cpu]->spinlock);
- if (etmdrvdata[cpu]->enable)
+ if (local_read(&etmdrvdata[cpu]->mode))
etm_disable_hw(etmdrvdata[cpu]);
spin_unlock(&etmdrvdata[cpu]->spinlock);
break;
diff --git a/drivers/hwtracing/coresight/coresight-etm4x.c b/drivers/hwtracing/coresight/coresight-etm4x.c
index a62e94814c37..6290231c365b 100644
--- a/drivers/hwtracing/coresight/coresight-etm4x.c
+++ b/drivers/hwtracing/coresight/coresight-etm4x.c
@@ -178,7 +178,7 @@ static void etm4_enable_hw(void *info)
dev_dbg(drvdata->dev, "cpu: %d enable smp call done\n", drvdata->cpu);
}
-static int etm4_enable(struct coresight_device *csdev)
+static int etm4_enable(struct coresight_device *csdev, u32 mode)
{
struct etmv4_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
int ret;
diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h
index 7b193a34d709..2dc59a5c5b1e 100644
--- a/drivers/hwtracing/coresight/coresight-priv.h
+++ b/drivers/hwtracing/coresight/coresight-priv.h
@@ -34,6 +34,12 @@
#define TIMEOUT_US 100
#define BMVAL(val, lsb, msb) ((val & GENMASK(msb, lsb)) >> lsb)
+enum cs_mode {
+ CS_MODE_DISABLED,
+ CS_MODE_SYSFS,
+ CS_MODE_PERF,
+};
+
static inline void CS_LOCK(void __iomem *addr)
{
do {
diff --git a/drivers/hwtracing/coresight/coresight.c b/drivers/hwtracing/coresight/coresight.c
index a4103a216e33..7e4064b31c99 100644
--- a/drivers/hwtracing/coresight/coresight.c
+++ b/drivers/hwtracing/coresight/coresight.c
@@ -222,7 +222,7 @@ static void coresight_disable_link(struct coresight_device *csdev,
csdev->enable = false;
}
-static int coresight_enable_source(struct coresight_device *csdev)
+static int coresight_enable_source(struct coresight_device *csdev, u32 mode)
{
int ret;
@@ -234,7 +234,7 @@ static int coresight_enable_source(struct coresight_device *csdev)
if (!csdev->enable) {
if (source_ops(csdev)->enable) {
- ret = source_ops(csdev)->enable(csdev);
+ ret = source_ops(csdev)->enable(csdev, mode);
if (ret)
return ret;
}
@@ -443,7 +443,7 @@ int coresight_enable(struct coresight_device *csdev)
if (ret)
goto err_path;
- ret = coresight_enable_source(csdev);
+ ret = coresight_enable_source(csdev, CS_MODE_SYSFS);
if (ret)
goto err_source;
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 851ecb22397e..61dfb8d511ea 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -213,7 +213,7 @@ struct coresight_ops_link {
struct coresight_ops_source {
int (*cpu_id)(struct coresight_device *csdev);
int (*trace_id)(struct coresight_device *csdev);
- int (*enable)(struct coresight_device *csdev);
+ int (*enable)(struct coresight_device *csdev, u32 mode);
void (*disable)(struct coresight_device *csdev);
};
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 23/26] perf tools: making function set_max_cpu_num() non static |
| Message-ID | <qtM3U-lk-23@gated-at.bofh.it> |
| In reply to | #1267482 |
For memory allocation purposes, code located in other places
then util/cpumap.c may want to know how many CPUs the system has.
This patch is making function set_max_cpu_num() available to
other parts of the perf tool so that global variable
@max_cpu_num gets the right value when referenced by cpu__max_cpu().
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
tools/perf/util/cpumap.c | 2 +-
tools/perf/util/cpumap.h | 1 +
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/tools/perf/util/cpumap.c b/tools/perf/util/cpumap.c
index 3667e2123e5b..97ef46e4a0f6 100644
--- a/tools/perf/util/cpumap.c
+++ b/tools/perf/util/cpumap.c
@@ -373,7 +373,7 @@ out:
}
/* Determine highest possible cpu in the system for sparse allocation */
-static void set_max_cpu_num(void)
+void set_max_cpu_num(void)
{
const char *mnt;
char path[PATH_MAX];
diff --git a/tools/perf/util/cpumap.h b/tools/perf/util/cpumap.h
index 0af9cecb4c51..6f7dce7dcca6 100644
--- a/tools/perf/util/cpumap.h
+++ b/tools/perf/util/cpumap.h
@@ -14,6 +14,7 @@ struct cpu_map {
int map[];
};
+void set_max_cpu_num(void);
struct cpu_map *cpu_map__new(const char *cpu_list);
struct cpu_map *cpu_map__dummy_new(void);
struct cpu_map *cpu_map__read(FILE *file);
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 20/26] coresight: etm-perf: new PMU driver for ETM tracers |
| Message-ID | <qtM3U-lk-29@gated-at.bofh.it> |
| In reply to | #1267482 |
Perf is a well known and used tool for performance monitoring
and much more. A such it is an ideal candidate for integration
with coresight based HW tracing.
This patch introduces a PMU that represent a coresight tracer to
the Perf core.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/Makefile | 3 +-
drivers/hwtracing/coresight/coresight-etm-perf.c | 438 +++++++++++++++++++++++
drivers/hwtracing/coresight/coresight-etm-perf.h | 32 ++
drivers/hwtracing/coresight/coresight-etm3x.c | 7 +
include/linux/coresight-pmu.h | 27 ++
5 files changed, 506 insertions(+), 1 deletion(-)
create mode 100644 drivers/hwtracing/coresight/coresight-etm-perf.c
create mode 100644 drivers/hwtracing/coresight/coresight-etm-perf.h
create mode 100644 include/linux/coresight-pmu.h
diff --git a/drivers/hwtracing/coresight/Makefile b/drivers/hwtracing/coresight/Makefile
index 233d66cf22d3..cf8c6d689747 100644
--- a/drivers/hwtracing/coresight/Makefile
+++ b/drivers/hwtracing/coresight/Makefile
@@ -9,6 +9,7 @@ obj-$(CONFIG_CORESIGHT_SINK_ETBV10) += coresight-etb10.o
obj-$(CONFIG_CORESIGHT_LINKS_AND_SINKS) += coresight-funnel.o \
coresight-replicator.o
obj-$(CONFIG_CORESIGHT_SOURCE_ETM3X) += coresight-etm3x.o coresight-etm-cp14.o \
- coresight-etm3x-sysfs.o
+ coresight-etm3x-sysfs.o \
+ coresight-etm-perf.o
obj-$(CONFIG_CORESIGHT_SOURCE_ETM4X) += coresight-etm4x.o
obj-$(CONFIG_CORESIGHT_QCOM_REPLICATOR) += coresight-replicator-qcom.o
diff --git a/drivers/hwtracing/coresight/coresight-etm-perf.c b/drivers/hwtracing/coresight/coresight-etm-perf.c
new file mode 100644
index 000000000000..590d9b132e6c
--- /dev/null
+++ b/drivers/hwtracing/coresight/coresight-etm-perf.c
@@ -0,0 +1,438 @@
+/*
+ * Copyright(C) 2015 Linaro Limited. All rights reserved.
+ * Author: Mathieu Poirier <mathieu.poirier@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/coresight.h>
+#include <linux/coresight-pmu.h>
+#include <linux/cpumask.h>
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/perf_event.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include "coresight-priv.h"
+
+static struct pmu etm_pmu;
+static bool etm_perf_up;
+
+/**
+ * struct etm_event_data - Coresight specifics associated to an event
+ * @mask: hold the CPU(s) this event was set for.
+ * @source_config: per CPU tracer configuration associated to a
+ * trace session.
+ * @sink_config: per CPU AUX configuration associated to a
+ * trace session.
+ * @path: the path, from source (first element) to
+ * sink (last element).
+ */
+struct etm_event_data {
+ cpumask_t mask;
+ void **source_config;
+ void **sink_config;
+ struct list_head **path;
+};
+
+static DEFINE_PER_CPU(struct perf_output_handle, ctx_handle);
+static DEFINE_PER_CPU(struct coresight_device *, csdev_src);
+
+/* ETMv3.5/PTM's ETMCR is 'config' */
+PMU_FORMAT_ATTR(cycacc, "config:" __stringify(ETM_OPT_CYCACC));
+PMU_FORMAT_ATTR(timestamp, "config:" __stringify(ETM_OPT_TS));
+
+static struct attribute *etm_config_formats_attr[] = {
+ &format_attr_cycacc.attr,
+ &format_attr_timestamp.attr,
+ NULL,
+};
+
+static struct attribute_group etm_pmu_format_group = {
+ .name = "format",
+ .attrs = etm_config_formats_attr,
+};
+
+static const struct attribute_group *etm_pmu_attr_groups[] = {
+ &etm_pmu_format_group,
+ NULL,
+};
+
+static void etm_event_read(struct perf_event *event) {}
+static void etm_event_destroy(struct perf_event *event) {}
+
+static int etm_event_init(struct perf_event *event)
+{
+ if (event->attr.type != etm_pmu.type)
+ return -ENOENT;
+
+ if (event->cpu >= nr_cpu_ids)
+ return -EINVAL;
+
+ event->destroy = etm_event_destroy;
+
+ return 0;
+}
+
+static void *alloc_event_data(int cpu)
+{
+ int size;
+ struct etm_event_data *event_data;
+ void *source_config, *sink_config;
+ struct list_head *path;
+
+ event_data = kzalloc(sizeof(struct etm_event_data), GFP_KERNEL);
+ if (!event_data)
+ return NULL;
+
+ if (cpu != -1)
+ size = 1;
+ else
+ size = num_online_cpus();
+
+ source_config = kcalloc(size, sizeof(void *), GFP_KERNEL);
+ if (!source_config)
+ goto source_config_err;
+
+ sink_config = kcalloc(size, sizeof(void *), GFP_KERNEL);
+ if (!sink_config)
+ goto sink_config_err;
+
+ path = kcalloc(size, sizeof(struct list_head *), GFP_KERNEL);
+ if (!path)
+ goto path_err;
+
+ cpumask_clear(&event_data->mask);
+ event_data->source_config = source_config;
+ event_data->sink_config = sink_config;
+ event_data->path = (struct list_head **)path;
+
+out:
+ return event_data;
+
+path_err:
+ kfree(sink_config);
+sink_config_err:
+ kfree(source_config);
+source_config_err:
+ kfree(event_data);
+ event_data = NULL;
+ goto out;
+}
+
+static void free_event_data(struct etm_event_data *event_data)
+{
+ int cpu;
+ cpumask_t *mask = &event_data->mask;
+
+ for_each_cpu(cpu, mask) {
+ kfree(event_data->source_config[cpu]);
+ kfree(event_data->sink_config[cpu]);
+ kfree(event_data->path[cpu]);
+ }
+
+ kfree(event_data->source_config);
+ kfree(event_data->sink_config);
+ kfree(event_data->path);
+ kfree(event_data);
+}
+
+static void *etm_setup_aux(struct perf_event *event, void **pages,
+ int nr_pages, bool overwrite)
+{
+ int cpu;
+ cpumask_t *mask, initialised;
+ struct etm_event_data *event_data = NULL;
+ struct coresight_device *csdev;
+
+ event_data = alloc_event_data(event->cpu);
+ if (!event_data)
+ return NULL;
+
+ mask = &event_data->mask;
+
+ if (event->cpu != -1)
+ cpumask_set_cpu(event->cpu, mask);
+ else
+ cpumask_copy(mask, cpu_online_mask);
+
+ cpumask_clear(&initialised);
+ for_each_cpu(cpu, mask) {
+ struct coresight_device *sink;
+
+ csdev = per_cpu(csdev_src, cpu);
+ if (!csdev)
+ goto err;
+
+ /* Get the tracer's config from perf */
+ if (!source_ops(csdev)->get_config)
+ goto err;
+
+ event_data->source_config[cpu] =
+ source_ops(csdev)->get_config(csdev, event);
+
+ if (!event_data->source_config[cpu])
+ goto err;
+
+ /*
+ * Building a path doesn't enable it, it simply builds a
+ * list of devices from source to sink that can be
+ * referenced later when the path is actually needed.
+ */
+ event_data->path[cpu] = coresight_build_path(csdev);
+ if (!event_data->path[cpu])
+ goto err;
+
+ /*
+ * Remember which paths have been initialised so that they
+ * can be undone in case of errors.
+ */
+ cpumask_set_cpu(cpu, &initialised);
+
+ /* Grab the sink at the end of the path */
+ sink = coresight_get_sink(event_data->path[cpu]);
+ if (!sink)
+ goto err;
+
+ if (!sink_ops(sink)->setup_aux)
+ goto err;
+
+ /* Finally get the AUX specific data from the sink buffer */
+ event_data->sink_config[cpu] =
+ sink_ops(sink)->setup_aux(sink, cpu, pages,
+ nr_pages, overwrite);
+ if (!event_data->sink_config[cpu])
+ goto err;
+ }
+
+out:
+ return event_data;
+
+err:
+ for_each_cpu(cpu, &initialised)
+ coresight_release_path(event_data->path[cpu]);
+
+ free_event_data(event_data);
+ event_data = NULL;
+ goto out;
+}
+
+static void etm_free_aux(void *data)
+{
+ free_event_data(data);
+}
+
+static void etm_event_stop(struct perf_event *event, int mode)
+{
+ int cpu = smp_processor_id();
+ struct coresight_device *csdev = per_cpu(csdev_src, cpu);
+
+ if (event->hw.state == PERF_HES_STOPPED)
+ return;
+
+ if (!csdev)
+ return;
+
+ /* stop tracer */
+ source_ops(csdev)->disable(csdev);
+
+ /* tell the core */
+ event->hw.state = PERF_HES_STOPPED;
+
+
+ if (mode & PERF_EF_UPDATE) {
+ struct coresight_device *sink;
+ struct perf_output_handle *handle = this_cpu_ptr(&ctx_handle);
+ struct etm_event_data *event_data = perf_get_aux(handle);
+
+ if (WARN_ON_ONCE(handle->event != event))
+ return;
+
+ if (WARN_ON_ONCE(!event_data))
+ return;
+
+ sink = coresight_get_sink(event_data->path[cpu]);
+ if (WARN_ON_ONCE(!sink))
+ return;
+
+ /* update trace information */
+ if (!sink_ops(sink)->update_buffer)
+ return;
+
+ sink_ops(sink)->update_buffer(sink, handle,
+ event_data->sink_config[cpu]);
+ }
+}
+
+static void etm_event_start(struct perf_event *event, int flags)
+{
+ int cpu = smp_processor_id();
+ struct coresight_device *csdev = per_cpu(csdev_src, cpu);
+
+ if (!csdev)
+ goto fail;
+
+ /* tell the perf core the event is alive */
+ event->hw.state = 0;
+
+ if (source_ops(csdev)->enable(csdev, CS_MODE_PERF))
+ goto fail;
+
+ return;
+
+fail:
+ event->hw.state = PERF_HES_STOPPED;
+}
+
+static void etm_event_del(struct perf_event *event, int mode)
+{
+ bool lost = true;
+ int cpu = smp_processor_id();
+ unsigned long size;
+ struct coresight_device *sink;
+ struct perf_output_handle *handle = this_cpu_ptr(&ctx_handle);
+ struct etm_event_data *event_data = perf_get_aux(handle);
+
+ if (WARN_ON_ONCE(!event_data))
+ return;
+
+ etm_event_stop(event, PERF_EF_UPDATE);
+
+ sink = coresight_get_sink(event_data->path[cpu]);
+ if (!sink)
+ return;
+
+ if (!sink_ops(sink)->reset_buffer)
+ return;
+
+ size = sink_ops(sink)->reset_buffer(sink, handle,
+ event_data->sink_config[cpu],
+ &lost);
+
+ perf_aux_output_end(handle, size, lost);
+
+ coresight_disable_path(event_data->path[cpu]);
+}
+
+static int etm_event_add(struct perf_event *event, int mode)
+{
+
+ int ret = -EBUSY, cpu = smp_processor_id();
+ struct etm_event_data *event_data;
+ struct perf_output_handle *handle = this_cpu_ptr(&ctx_handle);
+ struct hw_perf_event *hwc = &event->hw;
+ struct coresight_device *csdev = per_cpu(csdev_src, cpu);
+ struct coresight_device *sink;
+
+ if (handle->event)
+ goto out;
+
+ event_data = perf_aux_output_begin(handle, event);
+ ret = -EINVAL;
+ if (WARN_ON_ONCE(!event_data))
+ goto fail_stop;
+
+ sink = coresight_get_sink(event_data->path[cpu]);
+ if (!sink)
+ goto fail_end_stop;
+
+ if (!sink_ops(sink)->set_buffer)
+ goto fail_end_stop;
+
+ ret = sink_ops(sink)->set_buffer(sink, handle,
+ event_data->sink_config[cpu]);
+
+ if (ret)
+ goto fail_end_stop;
+
+ if (!source_ops(csdev)->set_config) {
+ ret = -EINVAL;
+ goto fail_end_stop;
+ }
+
+ source_ops(csdev)->set_config(csdev, event_data->source_config[cpu]);
+
+ ret = coresight_enable_path(event_data->path[cpu], CS_MODE_PERF);
+ if (ret)
+ goto fail_end_stop;
+
+ if (mode & PERF_EF_START) {
+ etm_event_start(event, 0);
+ if (hwc->state & PERF_HES_STOPPED) {
+ etm_event_del(event, 0);
+ return -EBUSY;
+ }
+ }
+
+out:
+ return ret;
+
+fail_end_stop:
+ perf_aux_output_end(handle, 0, true);
+fail_stop:
+ hwc->state = PERF_HES_STOPPED;
+ goto out;
+}
+
+int etm_perf_symlink(struct coresight_device *csdev, bool link)
+{
+ char entry[sizeof("cpu9999999")];
+ int ret = 0, cpu = source_ops(csdev)->cpu_id(csdev);
+ struct device *pmu_dev = etm_pmu.dev;
+ struct device *cs_dev = &csdev->dev;
+
+ sprintf(entry, "cpu%d", cpu);
+
+ if (!etm_perf_up)
+ return -EPROBE_DEFER;
+
+ if (link) {
+ ret = sysfs_create_link(&pmu_dev->kobj, &cs_dev->kobj, entry);
+ if (ret)
+ return ret;
+ per_cpu(csdev_src, cpu) = csdev;
+ } else {
+ sysfs_remove_link(&pmu_dev->kobj, entry);
+ per_cpu(csdev_src, cpu) = NULL;
+ }
+
+ return 0;
+}
+
+static int __init etm_perf_init(void)
+{
+ int ret;
+
+ etm_pmu.capabilities = PERF_PMU_CAP_EXCLUSIVE;
+
+ etm_pmu.attr_groups = etm_pmu_attr_groups;
+ etm_pmu.task_ctx_nr = perf_sw_context;
+ etm_pmu.read = etm_event_read;
+ etm_pmu.event_init = etm_event_init;
+ etm_pmu.setup_aux = etm_setup_aux;
+ etm_pmu.free_aux = etm_free_aux;
+ etm_pmu.stop = etm_event_stop;
+ etm_pmu.start = etm_event_start;
+ etm_pmu.del = etm_event_del;
+ etm_pmu.add = etm_event_add;
+
+ ret = perf_pmu_register(&etm_pmu, CORESIGHT_ETM_PMU_NAME, -1);
+ if (ret == 0)
+ etm_perf_up = true;
+
+ return ret;
+}
+module_init(etm_perf_init);
diff --git a/drivers/hwtracing/coresight/coresight-etm-perf.h b/drivers/hwtracing/coresight/coresight-etm-perf.h
new file mode 100644
index 000000000000..87f5a134eb6f
--- /dev/null
+++ b/drivers/hwtracing/coresight/coresight-etm-perf.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright(C) 2015 Linaro Limited. All rights reserved.
+ * Author: Mathieu Poirier <mathieu.poirier@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef _CORESIGHT_ETM_PERF_H
+#define _CORESIGHT_ETM_PERF_H
+
+struct coresight_device;
+
+#ifdef CONFIG_CORESIGHT
+int etm_perf_symlink(struct coresight_device *csdev, bool link);
+
+#else
+static inline int etm_perf_symlink(struct coresight_device *csdev, bool link)
+{ return -EINVAL; }
+
+#endif /* CONFIG_CORESIGHT */
+
+#endif
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index 4ff3e1419fde..c07b7d3eee86 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -35,6 +35,7 @@
#include <asm/sections.h>
#include "coresight-etm.h"
+#include "coresight-etm-perf.h"
static int boot_enable;
module_param_named(boot_enable, boot_enable, int, S_IRUGO);
@@ -870,6 +871,12 @@ static int etm_probe(struct amba_device *adev, const struct amba_id *id)
goto err_arch_supported;
}
+ ret = etm_perf_symlink(drvdata->csdev, true);
+ if (ret) {
+ coresight_unregister(drvdata->csdev);
+ goto err_arch_supported;
+ }
+
pm_runtime_put(&adev->dev);
dev_info(dev, "%s initialized\n", (char *)id->data);
diff --git a/include/linux/coresight-pmu.h b/include/linux/coresight-pmu.h
new file mode 100644
index 000000000000..6c5386b23b10
--- /dev/null
+++ b/include/linux/coresight-pmu.h
@@ -0,0 +1,27 @@
+/*
+ * Copyright(C) 2015 Linaro Limited. All rights reserved.
+ * Author: Mathieu Poirier <mathieu.poirier@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef _LINUX_CORESIGHT_PMU_H
+#define _LINUX_CORESIGHT_PMU_H
+
+#define CORESIGHT_ETM_PMU_NAME "cs_etm"
+
+/* ETMv3.5/PTM's ETMCR config bit */
+#define ETM_OPT_CYCACC 12
+#define ETM_OPT_TS 28
+
+#endif
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 06/26] coresight: associating path with session rather than tracer |
| Message-ID | <qtM3U-lk-39@gated-at.bofh.it> |
| In reply to | #1267482 |
When using the Coresight framework from the sysFS interface a
tracer is always handling a single session and as such, a path
can be associated with a tracer. But when supporting multiple
session per tracer there is no guarantee that sessions will always
have the same path from source to sink.
This patch is removing the automatic association between path and
tracers. The building of a path and enablement of the components
in the path are decoupled, allowing for the association of a path
with a session rather than a tracer.
To keep backward functionality with the current sysFS access methods
a per-cpu place holder is used to keep a handle on the path built when
tracers are enabled. Lastly APIs to build paths and enable tracers are
made public so that other subsystem can interact with the Coresight
framework.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-priv.h | 5 +
drivers/hwtracing/coresight/coresight.c | 300 ++++++++++++++++++---------
include/linux/coresight.h | 2 -
3 files changed, 208 insertions(+), 99 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h
index 62fcd98cc7cf..7b193a34d709 100644
--- a/drivers/hwtracing/coresight/coresight-priv.h
+++ b/drivers/hwtracing/coresight/coresight-priv.h
@@ -52,6 +52,11 @@ static inline void CS_UNLOCK(void __iomem *addr)
} while (0);
}
+void coresight_disable_path(struct list_head *path);
+int coresight_enable_path(struct list_head *path);
+struct list_head *coresight_build_path(struct coresight_device *csdev);
+void coresight_release_path(struct list_head *path);
+
#ifdef CONFIG_CORESIGHT_SOURCE_ETM3X
extern int etm_readl_cp14(u32 off, unsigned int *val);
extern int etm_writel_cp14(u32 off, u32 val);
diff --git a/drivers/hwtracing/coresight/coresight.c b/drivers/hwtracing/coresight/coresight.c
index a3dcafb81700..4beea67fd5a1 100644
--- a/drivers/hwtracing/coresight/coresight.c
+++ b/drivers/hwtracing/coresight/coresight.c
@@ -29,6 +29,22 @@
static DEFINE_MUTEX(coresight_mutex);
+/**
+ * struct coresight_node - elements of a path, from source to sink
+ * @csdev: Address of an element.
+ * @link: hook to the list.
+ */
+struct coresight_node {
+ struct coresight_device *csdev;
+ struct list_head link;
+};
+
+/*
+ * When operating Coresight drivers from the sysFS interface, only a single
+ * path can exist from a tracer (associated to a CPU) to a sink.
+ */
+static DEFINE_PER_CPU(struct list_head *, sysfs_path);
+
static int coresight_id_match(struct device *dev, void *data)
{
int trace_id, i_trace_id;
@@ -68,15 +84,12 @@ static int coresight_source_is_unique(struct coresight_device *csdev)
csdev, coresight_id_match);
}
-static int coresight_find_link_inport(struct coresight_device *csdev)
+static int coresight_find_link_inport(struct coresight_device *csdev,
+ struct coresight_device *parent)
{
int i;
- struct coresight_device *parent;
struct coresight_connection *conn;
- parent = container_of(csdev->path_link.next,
- struct coresight_device, path_link);
-
for (i = 0; i < parent->nr_outport; i++) {
conn = &parent->conns[i];
if (conn->child_dev == csdev)
@@ -89,15 +102,12 @@ static int coresight_find_link_inport(struct coresight_device *csdev)
return 0;
}
-static int coresight_find_link_outport(struct coresight_device *csdev)
+static int coresight_find_link_outport(struct coresight_device *csdev,
+ struct coresight_device *child)
{
int i;
- struct coresight_device *child;
struct coresight_connection *conn;
- child = container_of(csdev->path_link.prev,
- struct coresight_device, path_link);
-
for (i = 0; i < csdev->nr_outport; i++) {
conn = &csdev->conns[i];
if (conn->child_dev == child)
@@ -138,14 +148,19 @@ static void coresight_disable_sink(struct coresight_device *csdev)
}
}
-static int coresight_enable_link(struct coresight_device *csdev)
+static int coresight_enable_link(struct coresight_device *csdev,
+ struct coresight_device *parent,
+ struct coresight_device *child)
{
int ret;
int link_subtype;
int refport, inport, outport;
- inport = coresight_find_link_inport(csdev);
- outport = coresight_find_link_outport(csdev);
+ if (!parent || !child)
+ return -EINVAL;
+
+ inport = coresight_find_link_inport(csdev, parent);
+ outport = coresight_find_link_outport(csdev, child);
link_subtype = csdev->subtype.link_subtype;
if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_MERG)
@@ -168,14 +183,19 @@ static int coresight_enable_link(struct coresight_device *csdev)
return 0;
}
-static void coresight_disable_link(struct coresight_device *csdev)
+static void coresight_disable_link(struct coresight_device *csdev,
+ struct coresight_device *parent,
+ struct coresight_device *child)
{
int i, nr_conns;
int link_subtype;
int refport, inport, outport;
- inport = coresight_find_link_inport(csdev);
- outport = coresight_find_link_outport(csdev);
+ if (!parent || !child)
+ return;
+
+ inport = coresight_find_link_inport(csdev, parent);
+ outport = coresight_find_link_outport(csdev, child);
link_subtype = csdev->subtype.link_subtype;
if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_MERG) {
@@ -235,109 +255,167 @@ static void coresight_disable_source(struct coresight_device *csdev)
}
}
-static int coresight_enable_path(struct list_head *path)
+void coresight_disable_path(struct list_head *path)
{
- int ret = 0;
- struct coresight_device *cd;
-
- /*
- * At this point we have a full @path, from source to sink. The
- * sink is the first entry and the source the last one. Go through
- * all the components and enabled them one by one.
- */
- list_for_each_entry(cd, path, path_link) {
- if (cd == list_first_entry(path, struct coresight_device,
- path_link)) {
- ret = coresight_enable_sink(cd);
- } else if (list_is_last(&cd->path_link, path)) {
- /*
- * Don't enable the source just yet - this needs to
- * happen at the very end when all links and sink
- * along the path have been configured properly.
- */
- ;
- } else {
- ret = coresight_enable_link(cd);
+ struct coresight_node *nd;
+ struct coresight_device *csdev, *parent, *child;
+
+ list_for_each_entry(nd, path, link) {
+ csdev = nd->csdev;
+
+ switch (csdev->type) {
+ case CORESIGHT_DEV_TYPE_SINK:
+ case CORESIGHT_DEV_TYPE_LINKSINK:
+ coresight_disable_sink(csdev);
+ break;
+ case CORESIGHT_DEV_TYPE_SOURCE:
+ /* sources are disabled from either sysFS or Perf */
+ break;
+ case CORESIGHT_DEV_TYPE_LINK:
+ parent = list_prev_entry(nd, link)->csdev;
+ child = list_next_entry(nd, link)->csdev;
+ coresight_disable_link(csdev, parent, child);
+ break;
+ default:
+ break;
}
- if (ret)
- goto err;
}
+}
- return 0;
-err:
- list_for_each_entry_continue_reverse(cd, path, path_link) {
- if (cd == list_first_entry(path, struct coresight_device,
- path_link)) {
- coresight_disable_sink(cd);
- } else if (list_is_last(&cd->path_link, path)) {
- ;
- } else {
- coresight_disable_link(cd);
+int coresight_enable_path(struct list_head *path)
+{
+
+ int ret = 0;
+ struct coresight_node *nd;
+ struct coresight_device *csdev, *parent, *child;
+
+ list_for_each_entry_reverse(nd, path, link) {
+ csdev = nd->csdev;
+
+ switch (csdev->type) {
+ case CORESIGHT_DEV_TYPE_SINK:
+ case CORESIGHT_DEV_TYPE_LINKSINK:
+ ret = coresight_enable_sink(csdev);
+ if (ret)
+ goto err;
+ break;
+ case CORESIGHT_DEV_TYPE_SOURCE:
+ /* sources are enabled from either sysFS or Perf */
+ break;
+ case CORESIGHT_DEV_TYPE_LINK:
+ parent = list_prev_entry(nd, link)->csdev;
+ child = list_next_entry(nd, link)->csdev;
+ ret = coresight_enable_link(csdev, parent, child);
+ if (ret)
+ goto err;
+ break;
+ default:
+ goto err;
}
}
+out:
return ret;
+err:
+ coresight_disable_path(path);
+ goto out;
}
-static int coresight_disable_path(struct list_head *path)
+/**
+ * _coresight_build_path - recursively build a path from a @csdev to a sink.
+ * @csdev: The device to start from.
+ * @path: The list to add devices to.
+ *
+ * The tree of Coresight device is traversed until an activated sink is
+ * found. From there the sink is added to the list along with all the
+ * devices that led to that point - the end result is a list from source
+ * to sink. In that list the source is the first device and the sink the
+ * last one.
+ */
+static int _coresight_build_path(struct coresight_device *csdev,
+ struct list_head *path)
{
- struct coresight_device *cd;
-
- list_for_each_entry_reverse(cd, path, path_link) {
- if (cd == list_first_entry(path, struct coresight_device,
- path_link)) {
- coresight_disable_sink(cd);
- } else if (list_is_last(&cd->path_link, path)) {
- /*
- * The source has already been stopped, no need
- * to do it again here.
- */
- ;
- } else {
- coresight_disable_link(cd);
+ int i;
+ bool found = false;
+ struct coresight_node *node;
+ struct coresight_connection *conn;
+
+ /* An activated sink has been found. Enqueue the element */
+ if ((csdev->type == CORESIGHT_DEV_TYPE_SINK ||
+ csdev->type == CORESIGHT_DEV_TYPE_LINKSINK) && csdev->activated)
+ goto out;
+
+ /* Not a sink - recursively explore each port found on this element */
+ for (i = 0; i < csdev->nr_outport; i++) {
+ conn = &csdev->conns[i];
+ if (_coresight_build_path(conn->child_dev, path) == 0) {
+ found = true;
+ break;
}
}
+ if (!found)
+ return -ENODEV;
+
+out:
+ /*
+ * A path from this element to a sink has been found. The elements
+ * leading to the sink are already enqueued, all that is left to do
+ * is add a node for this element.
+ */
+ node = kzalloc(sizeof(struct coresight_node), GFP_KERNEL);
+ if (!node)
+ return -ENOMEM;
+
+ node->csdev = csdev;
+ list_add(&node->link, path);
+
return 0;
}
-static int coresight_build_paths(struct coresight_device *csdev,
- struct list_head *path,
- bool enable)
+struct list_head *coresight_build_path(struct coresight_device *csdev)
{
- int i, ret = -EINVAL;
- struct coresight_connection *conn;
+ struct list_head *path;
- list_add(&csdev->path_link, path);
+ path = kzalloc(sizeof(struct list_head), GFP_KERNEL);
+ if (!path)
+ return NULL;
- if ((csdev->type == CORESIGHT_DEV_TYPE_SINK ||
- csdev->type == CORESIGHT_DEV_TYPE_LINKSINK) &&
- csdev->activated) {
- if (enable)
- ret = coresight_enable_path(path);
- else
- ret = coresight_disable_path(path);
- } else {
- for (i = 0; i < csdev->nr_outport; i++) {
- conn = &csdev->conns[i];
- if (coresight_build_paths(conn->child_dev,
- path, enable) == 0)
- ret = 0;
- }
+ INIT_LIST_HEAD(path);
+
+ if (_coresight_build_path(csdev, path)) {
+ kfree(path);
+ path = NULL;
}
- if (list_first_entry(path, struct coresight_device, path_link) != csdev)
- dev_err(&csdev->dev, "wrong device in %s\n", __func__);
+ return path;
+}
- list_del(&csdev->path_link);
+/**
+ * coresight_release_path - release a previously built path.
+ * @path: the path to release.
+ *
+ * Go through all the elements of a path and 1) removed it from the list and
+ * 2) free the memory allocated for each node.
+ */
+void coresight_release_path(struct list_head *path)
+{
+ struct coresight_node *nd, *next;
- return ret;
+ list_for_each_entry_safe(nd, next, path, link) {
+ list_del(&nd->link);
+ kfree(nd);
+ }
+
+ kfree(path);
+ path = NULL;
}
int coresight_enable(struct coresight_device *csdev)
{
int ret = 0;
- LIST_HEAD(path);
+ int cpu;
+ struct list_head *path;
mutex_lock(&coresight_mutex);
if (csdev->type != CORESIGHT_DEV_TYPE_SOURCE) {
@@ -348,22 +426,47 @@ int coresight_enable(struct coresight_device *csdev)
if (csdev->enable)
goto out;
- if (coresight_build_paths(csdev, &path, true)) {
- dev_err(&csdev->dev, "building path(s) failed\n");
+ path = coresight_build_path(csdev);
+ if (!path) {
+ pr_err("building path(s) failed\n");
goto out;
}
- if (coresight_enable_source(csdev))
- dev_err(&csdev->dev, "source enable failed\n");
+ ret = coresight_enable_path(path);
+ if (ret)
+ goto err_path;
+
+ ret = coresight_enable_source(csdev);
+ if (ret)
+ goto err_source;
+
+ /*
+ * When working from sysFS it is important to keep track
+ * of the paths that were created so that they can be
+ * undone in 'coresight_disable()'. Since there can only
+ * be a single session per tracer (when working from sysFS)
+ * a per-cpu variable will do just fine.
+ */
+ cpu = source_ops(csdev)->cpu_id(csdev);
+ per_cpu(sysfs_path, cpu) = path;
+
out:
mutex_unlock(&coresight_mutex);
return ret;
+
+err_source:
+ coresight_disable_path(path);
+
+err_path:
+ coresight_release_path(path);
+ goto out;
}
EXPORT_SYMBOL_GPL(coresight_enable);
void coresight_disable(struct coresight_device *csdev)
{
- LIST_HEAD(path);
+ int cpu;
+ struct list_head *path;
mutex_lock(&coresight_mutex);
if (csdev->type != CORESIGHT_DEV_TYPE_SOURCE) {
@@ -373,9 +476,12 @@ void coresight_disable(struct coresight_device *csdev)
if (!csdev->enable)
goto out;
+ cpu = source_ops(csdev)->cpu_id(csdev);
+ path = per_cpu(sysfs_path, cpu);
coresight_disable_source(csdev);
- if (coresight_build_paths(csdev, &path, false))
- dev_err(&csdev->dev, "releasing path(s) failed\n");
+ coresight_disable_path(path);
+ coresight_release_path(path);
+ per_cpu(sysfs_path, cpu) = NULL;
out:
mutex_unlock(&coresight_mutex);
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index bf62b265bf52..851ecb22397e 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -152,7 +152,6 @@ struct coresight_connection {
by @coresight_ops.
* @dev: The device entity associated to this component.
* @refcnt: keep track of what is in use.
- * @path_link: link of current component into the path being enabled.
* @orphan: true if the component has connections that haven't been linked.
* @enable: 'true' if component is currently part of an active path.
* @activated: 'true' only if a _sink_ has been activated. A sink can be
@@ -168,7 +167,6 @@ struct coresight_device {
const struct coresight_ops *ops;
struct device dev;
atomic_t *refcnt;
- struct list_head path_link;
bool orphan;
bool enable; /* true only if configured as part of a path */
bool activated; /* true only if a sink is part of a path */
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 11/26] coresight: etm3x: changing default trace configuration |
| Message-ID | <qtM3U-lk-31@gated-at.bofh.it> |
| In reply to | #1267482 |
Changing default configuration to include the entire address
range rather than just the kernel. That way traces are more
inclusive and it is easier to narrow down if needed.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etm.h | 2 ++
drivers/hwtracing/coresight/coresight-etm3x.c | 29 ++++++++++++---------------
2 files changed, 15 insertions(+), 16 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etm.h b/drivers/hwtracing/coresight/coresight-etm.h
index a8663780a908..f01862fc98dd 100644
--- a/drivers/hwtracing/coresight/coresight-etm.h
+++ b/drivers/hwtracing/coresight/coresight-etm.h
@@ -146,6 +146,7 @@
* @startstop_ctrl: setting for register ETMTSSCR.
* @enable_event: setting for register ETMTEEVR.
* @enable_ctrl1: setting for register ETMTECR1.
+ * @enable_ctrl2: setting for register ETMTECR2.
* @fifofull_level: setting for register ETMFFLR.
* @addr_idx: index for the address comparator selection.
* @addr_val: value for address comparator register.
@@ -179,6 +180,7 @@ struct etm_config {
u32 startstop_ctrl;
u32 enable_event;
u32 enable_ctrl1;
+ u32 enable_ctrl2;
u32 fifofull_level;
u8 addr_idx;
u32 addr_val[ETM_MAX_ADDR_CMP];
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index d1235711b576..4c54e104b610 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -607,26 +607,23 @@ static void etm_init_arch_data(void *info)
static void etm_init_default_data(struct etm_config *config)
{
- u32 flags = (1 << 0 | /* instruction execute*/
- 3 << 3 | /* ARM instruction */
- 0 << 5 | /* No data value comparison */
- 0 << 7 | /* No exact mach */
- 0 << 8 | /* Ignore context ID */
- 0 << 10); /* Security ignored */
-
if (WARN_ON_ONCE(!config))
return;
- config->ctrl = (ETMCR_CYC_ACC | ETMCR_TIMESTAMP_EN);
- config->enable_ctrl1 = ETMTECR1_ADDR_COMP_1;
- config->addr_val[0] = (u32) _stext;
- config->addr_val[1] = (u32) _etext;
- config->addr_acctype[0] = flags;
- config->addr_acctype[1] = flags;
- config->addr_type[0] = ETM_ADDR_TYPE_RANGE;
- config->addr_type[1] = ETM_ADDR_TYPE_RANGE;
-
etm_set_default(config);
+
+ /*
+ * Taken verbatim from the TRM:
+ *
+ * To trace all memory:
+ * set bit [24] in register 0x009, the ETMTECR1, to 1
+ * set all other bits in register 0x009, the ETMTECR1, to 0
+ * set all bits in register 0x007, the ETMTECR2, to 0
+ * set register 0x008, the ETMTEEVR, to 0x6F (TRUE).
+ */
+ config->enable_ctrl1 = BIT(24);
+ config->enable_ctrl2 = 0x0;
+ config->enable_event = ETM_HARD_WIRE_RES_A;
}
static void etm_init_trace_id(struct etm_drvdata *drvdata)
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 25/26] perf tools: making coresight PMU listable |
| Message-ID | <qtM3U-lk-37@gated-at.bofh.it> |
| In reply to | #1267482 |
Adding the required mechanic allowing 'perf list pmu' to
discover coresight ETM/PTM tracers.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
MAINTAINERS | 1 +
tools/perf/arch/arm/util/Build | 2 ++
tools/perf/arch/arm/util/pmu.c | 34 ++++++++++++++++++++++++++++++++++
tools/perf/config/Makefile | 17 ++++++++++++-----
4 files changed, 49 insertions(+), 5 deletions(-)
create mode 100644 tools/perf/arch/arm/util/pmu.c
diff --git a/MAINTAINERS b/MAINTAINERS
index 747c65316167..c31e68c998d2 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1000,6 +1000,7 @@ F: drivers/hwtracing/coresight/*
F: Documentation/trace/coresight.txt
F: Documentation/devicetree/bindings/arm/coresight.txt
F: Documentation/ABI/testing/sysfs-bus-coresight-devices-*
+F: tools/perf/arch/arm/util/pmu.c
ARM/CORGI MACHINE SUPPORT
M: Richard Purdie <rpurdie@rpsys.net>
diff --git a/tools/perf/arch/arm/util/Build b/tools/perf/arch/arm/util/Build
index d22e3d07de3d..66ab0b05549c 100644
--- a/tools/perf/arch/arm/util/Build
+++ b/tools/perf/arch/arm/util/Build
@@ -2,3 +2,5 @@ libperf-$(CONFIG_DWARF) += dwarf-regs.o
libperf-$(CONFIG_LIBUNWIND) += unwind-libunwind.o
libperf-$(CONFIG_LIBDW_DWARF_UNWIND) += unwind-libdw.o
+
+libperf-$(CONFIG_AUXTRACE) += pmu.o
diff --git a/tools/perf/arch/arm/util/pmu.c b/tools/perf/arch/arm/util/pmu.c
new file mode 100644
index 000000000000..af9fb666b44f
--- /dev/null
+++ b/tools/perf/arch/arm/util/pmu.c
@@ -0,0 +1,34 @@
+/*
+ * Copyright(C) 2015 Linaro Limited. All rights reserved.
+ * Author: Mathieu Poirier <mathieu.poirier@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <string.h>
+#include <linux/coresight-pmu.h>
+#include <linux/perf_event.h>
+
+#include "../../util/pmu.h"
+
+struct perf_event_attr
+*perf_pmu__get_default_config(struct perf_pmu *pmu __maybe_unused)
+{
+#ifdef HAVE_AUXTRACE_SUPPORT
+ if (!strcmp(pmu->name, CORESIGHT_ETM_PMU_NAME)) {
+ /* add ETM default config here */
+ pmu->selectable = true;
+ }
+#endif
+ return NULL;
+}
diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile
index 38a08539f4bf..64bdb82b3337 100644
--- a/tools/perf/config/Makefile
+++ b/tools/perf/config/Makefile
@@ -626,12 +626,19 @@ ifdef LIBBABELTRACE
endif
ifndef NO_AUXTRACE
- ifeq ($(feature-get_cpuid), 0)
- msg := $(warning Your gcc lacks the __get_cpuid() builtin, disables support for auxtrace/Intel PT, please install a newer gcc);
- NO_AUXTRACE := 1
+ ifeq ($(ARCH),x86)
+ ifeq ($(feature-get_cpuid), 0)
+ msg := $(warning Your gcc lacks the __get_cpuid() builtin, disables support for auxtrace/Intel PT, please install a newer gcc);
+ NO_AUXTRACE := 1
+ else
+ $(call detected,CONFIG_AUXTRACE)
+ CFLAGS += -DHAVE_AUXTRACE_SUPPORT
+ endif
else
- $(call detected,CONFIG_AUXTRACE)
- CFLAGS += -DHAVE_AUXTRACE_SUPPORT
+ ifeq ($(ARCH),$(filter $(ARCH), arm arm64))
+ $(call detected,CONFIG_AUXTRACE)
+ CFLAGS += -DHAVE_AUXTRACE_SUPPORT
+ endif
endif
endif
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 12/26] coresight: etm3x: consolidating initial config |
| Message-ID | <qtM3U-lk-41@gated-at.bofh.it> |
| In reply to | #1267482 |
There is really no point in having two functions to take care
of doing the initials tracer configuration. As such moving
everything to 'etm_set_default()'.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etm3x.c | 37 ++++++++++-----------------
1 file changed, 14 insertions(+), 23 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index 4c54e104b610..e973e4d13bcb 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -41,7 +41,6 @@ module_param_named(boot_enable, boot_enable, int, S_IRUGO);
/* The number of ETM/PTM currently registered */
static int etm_count;
static struct etm_drvdata *etmdrvdata[NR_CPUS];
-static void etm_init_default_data(struct etm_config *config);
/*
* Memory mapped writes to clear os lock are not supported on some processors
@@ -199,7 +198,7 @@ struct etm_config *get_etm_config(struct etm_drvdata *drvdata)
return NULL;
/* Set default config */
- etm_init_default_data(config);
+ etm_set_default(config);
drvdata->config = config;
out:
return drvdata->config;
@@ -212,6 +211,19 @@ void etm_set_default(struct etm_config *config)
if (WARN_ON_ONCE(!config))
return;
+ /*
+ * Taken verbatim from the TRM:
+ *
+ * To trace all memory:
+ * set bit [24] in register 0x009, the ETMTECR1, to 1
+ * set all other bits in register 0x009, the ETMTECR1, to 0
+ * set all bits in register 0x007, the ETMTECR2, to 0
+ * set register 0x008, the ETMTEEVR, to 0x6F (TRUE).
+ */
+ config->enable_ctrl1 = BIT(24);
+ config->enable_ctrl2 = 0x0;
+ config->enable_event = ETM_HARD_WIRE_RES_A;
+
config->trigger_event = ETM_DEFAULT_EVENT_VAL;
config->enable_event = ETM_HARD_WIRE_RES_A;
@@ -605,27 +617,6 @@ static void etm_init_arch_data(void *info)
CS_LOCK(drvdata->base);
}
-static void etm_init_default_data(struct etm_config *config)
-{
- if (WARN_ON_ONCE(!config))
- return;
-
- etm_set_default(config);
-
- /*
- * Taken verbatim from the TRM:
- *
- * To trace all memory:
- * set bit [24] in register 0x009, the ETMTECR1, to 1
- * set all other bits in register 0x009, the ETMTECR1, to 0
- * set all bits in register 0x007, the ETMTECR2, to 0
- * set register 0x008, the ETMTEEVR, to 0x6F (TRUE).
- */
- config->enable_ctrl1 = BIT(24);
- config->enable_ctrl2 = 0x0;
- config->enable_event = ETM_HARD_WIRE_RES_A;
-}
-
static void etm_init_trace_id(struct etm_drvdata *drvdata)
{
/*
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 19/26] coresight: etb10: implementing AUX space API |
| Message-ID | <qtM3U-lk-35@gated-at.bofh.it> |
| In reply to | #1267482 |
Adding an ETB10 specific AUX area operations to be used
by the perf framework when events are initialised.
Part of this operation involves modeling the mmap'ed area
based on the specific ways a sink buffer gathers information.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etb10.c | 228 ++++++++++++++++++++++++++
include/linux/coresight.h | 20 ++-
2 files changed, 246 insertions(+), 2 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etb10.c b/drivers/hwtracing/coresight/coresight-etb10.c
index dc4d707b28aa..27ddf9ecc7cc 100644
--- a/drivers/hwtracing/coresight/coresight-etb10.c
+++ b/drivers/hwtracing/coresight/coresight-etb10.c
@@ -28,6 +28,11 @@
#include <linux/coresight.h>
#include <linux/amba/bus.h>
#include <linux/clk.h>
+#include <linux/circ_buf.h>
+#include <linux/mm.h>
+#include <linux/perf_event.h>
+
+#include <asm/local.h>
#include "coresight-priv.h"
@@ -65,6 +70,26 @@
#define ETB_FRAME_SIZE_WORDS 4
/**
+ * struct cs_buffer - keep track of a recording session' specifics
+ * @cur: index of the current buffer
+ * @nr_pages: max number of pages granted to us
+ * @offset: offset within the current buffer
+ * @data_size: how much we collected in this run
+ * @lost: other than zero if we had a HW buffer wrap around
+ * @snapshot: is this run in snapshot mode
+ * @data_pages: a handle the ring buffer
+ */
+struct cs_buffers {
+ unsigned int cur;
+ unsigned int nr_pages;
+ unsigned long offset;
+ local_t data_size;
+ local_t lost;
+ bool snapshot;
+ void **data_pages;
+};
+
+/**
* struct etb_drvdata - specifics associated to an ETB component
* @base: memory mapped base address for this component.
* @dev: the device entity associated to this component.
@@ -266,9 +291,212 @@ static void etb_disable(struct coresight_device *csdev)
dev_info(drvdata->dev, "ETB disabled\n");
}
+static void *etb_setup_aux(struct coresight_device *csdev, int cpu,
+ void **pages, int nr_pages, bool overwrite)
+{
+ int node;
+ struct cs_buffers *buf;
+
+ if (cpu == -1)
+ cpu = smp_processor_id();
+ node = cpu_to_node(cpu);
+
+ buf = kzalloc_node(sizeof(struct cs_buffers), GFP_KERNEL, node);
+ if (!buf)
+ return NULL;
+
+ buf->snapshot = overwrite;
+ buf->nr_pages = nr_pages;
+ buf->data_pages = pages;
+
+ return buf;
+}
+
+static int etb_set_buffer(struct coresight_device *csdev,
+ struct perf_output_handle *handle,
+ void *sink_config)
+{
+ int ret = 0;
+ unsigned long head;
+ struct cs_buffers *buf = sink_config;
+
+ /* wrap head around to the amount of space we have */
+ head = handle->head & ((buf->nr_pages << PAGE_SHIFT) - 1);
+
+ /* find the page to write to */
+ buf->cur = head / PAGE_SIZE;
+
+ /* and offset within that page */
+ buf->offset = head % PAGE_SIZE;
+
+ local_set(&buf->data_size, 0);
+
+ return ret;
+}
+
+static unsigned long etb_reset_buffer(struct coresight_device *csdev,
+ struct perf_output_handle *handle,
+ void *sink_config, bool *lost)
+{
+ unsigned long size = 0;
+ struct cs_buffers *buf = sink_config;
+
+ if (buf) {
+ /*
+ * In snapshot mode ->data_size holds the new address of the
+ * ring buffer's head. The size itself is the whole address
+ * range since we want the latest information.
+ */
+ if (buf->snapshot)
+ handle->head = local_xchg(&buf->data_size,
+ buf->nr_pages << PAGE_SHIFT);
+
+ /*
+ * Tell the tracer PMU how much we got in this run and if
+ * something went wrong along the way. Nobody else can use
+ * this cs_buffers instance until we are done. As such
+ * resetting parameters here and squaring off with the ring
+ * buffer API in the tracer PMU is fine.
+ */
+ *lost = local_xchg(&buf->lost, 0);
+ size = local_xchg(&buf->data_size, 0);
+ }
+
+ return size;
+}
+
+static void etb_update_buffer(struct coresight_device *csdev,
+ struct perf_output_handle *handle,
+ void *sink_config)
+{
+ int i, cur;
+ u8 *buf_ptr;
+ u32 read_ptr, write_ptr, start;
+ u32 status, read_data, to_read;
+ unsigned long flags, offset;
+ struct cs_buffers *buf = sink_config;
+ struct etb_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+
+ if (!buf)
+ return;
+
+ spin_lock_irqsave(&drvdata->spinlock, flags);
+
+ CS_UNLOCK(drvdata->base);
+ etb_disable_hw(drvdata);
+
+ /* unit is in words, not bytes */
+ read_ptr = readl_relaxed(drvdata->base + ETB_RAM_READ_POINTER);
+ write_ptr = readl_relaxed(drvdata->base + ETB_RAM_WRITE_POINTER);
+
+ /*
+ * Entries should be aligned to the frame size. If they are not
+ * go back to the last alignement point to give decoding tools a
+ * chance to fix things.
+ */
+ if (write_ptr % ETB_FRAME_SIZE_WORDS) {
+ dev_err(drvdata->dev,
+ "write_ptr: %lu not aligned to formatter frame size\n",
+ (unsigned long)write_ptr);
+
+ write_ptr &= ~ETB_FRAME_SIZE_WORDS;
+ local_inc(&buf->lost);
+ }
+
+ /*
+ * Get a hold of the status register and see if a wrap around
+ * has occurred. If so adjust things accordingly. Otherwise
+ * start at the beginning and go until the write pointer has
+ * been reached.
+ */
+ status = readl_relaxed(drvdata->base + ETB_STATUS_REG);
+ if (status & ETB_STATUS_RAM_FULL) {
+ local_inc(&buf->lost);
+ to_read = drvdata->buffer_depth * ETB_FRAME_SIZE_WORDS;
+ start = write_ptr;
+ } else {
+ to_read = CIRC_CNT(write_ptr, read_ptr,
+ drvdata->buffer_depth * ETB_FRAME_SIZE_WORDS);
+ start = read_ptr;
+ }
+
+ /*
+ * Make sure we don't overwrite data that hasn't been consumed yet.
+ * It is entirely possible that the HW buffer has more data than the
+ * ring buffer can currently handle. If so adjust the start address
+ * to take only the last traces.
+ *
+ * In snapshot mode we are looking to get the latest traces only and as
+ * such, we don't care about not overwriting data that hasn't been
+ * processed by user space.
+ *
+ * Since metrics related to ETBs is in words, multiply by the
+ * amount of byte per word to have the right units.
+ */
+ if (!buf->snapshot && to_read > handle->size) {
+ unsigned int capacity = drvdata->buffer_depth *
+ ETB_FRAME_SIZE_WORDS;
+ to_read = handle->size;
+ /* advance the start pointer to get the latest trace data */
+ start += capacity - handle->size;
+ /* make sure we are still aligned */
+ start &= ~ETB_FRAME_SIZE_WORDS;
+ /* wrap around if we've reach the end of the HW buffer */
+ start &= capacity - 1;
+ /* let the decoder know we've skipped ahead */
+ local_inc(&buf->lost);
+ }
+
+ /* finally tell HW where we want to start reading from */
+ writel_relaxed(start, drvdata->base + ETB_RAM_READ_POINTER);
+
+ cur = buf->cur;
+ offset = buf->offset;
+ for (i = 0; i < to_read; i += 4) {
+ buf_ptr = buf->data_pages[cur] + offset;
+ read_data = readl_relaxed(drvdata->base +
+ ETB_RAM_READ_DATA_REG);
+ *buf_ptr++ = read_data >> 0;
+ *buf_ptr++ = read_data >> 8;
+ *buf_ptr++ = read_data >> 16;
+ *buf_ptr++ = read_data >> 24;
+
+ offset += 4;
+ if (offset >= PAGE_SIZE) {
+ offset = 0;
+ cur++;
+ /* wrap around at the end of the buffer */
+ cur &= buf->nr_pages - 1;
+ }
+ }
+
+ /* reset ETB buffer for next run */
+ writel_relaxed(0x0, drvdata->base + ETB_RAM_READ_POINTER);
+ writel_relaxed(0x0, drvdata->base + ETB_RAM_WRITE_POINTER);
+
+ /*
+ * In snapshot mode all we have to do is communicate to
+ * perf_aux_output_end() the address of the current head. In full
+ * trace mode the same function expects a size to move rb->aux_head
+ * forward.
+ */
+ if (buf->snapshot)
+ local_set(&buf->data_size, (cur * PAGE_SIZE) + offset);
+ else
+ local_add(to_read, &buf->data_size);
+
+ etb_enable_hw(drvdata);
+ CS_LOCK(drvdata->base);
+ spin_unlock_irqrestore(&drvdata->spinlock, flags);
+}
+
static const struct coresight_ops_sink etb_sink_ops = {
.enable = etb_enable,
.disable = etb_disable,
+ .setup_aux = etb_setup_aux,
+ .set_buffer = etb_set_buffer,
+ .reset_buffer = etb_reset_buffer,
+ .update_buffer = etb_update_buffer,
};
static const struct coresight_ops etb_cs_ops = {
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 80b3465bd459..810a8b2511c4 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -14,6 +14,7 @@
#define _LINUX_CORESIGHT_H
#include <linux/device.h>
+#include <linux/perf_event.h>
#include <linux/sched.h>
/* Peripheral id registers (0xFD0-0xFEC) */
@@ -181,12 +182,27 @@ struct coresight_device {
/**
* struct coresight_ops_sink - basic operations for a sink
* Operations available for sinks
- * @enable: enables the sink.
- * @disable: disables the sink.
+ * @enable: enables the sink.
+ * @disable: disables the sink.
+ * @setup_aux: initialises perf's ring buffer for trace collection.
+ * @set_buffer: initialises buffer mechanic before a trace session.
+ * @reset_buffer: finalises buffer mechanic after a trace session.
+ * @update_buffer: update buffer pointers after a trace session.
*/
struct coresight_ops_sink {
int (*enable)(struct coresight_device *csdev, u32 mode);
void (*disable)(struct coresight_device *csdev);
+ void *(*setup_aux)(struct coresight_device *csdev, int cpu,
+ void **pages, int nr_pages, bool overwrite);
+ int (*set_buffer)(struct coresight_device *csdev,
+ struct perf_output_handle *handle,
+ void *sink_config);
+ unsigned long (*reset_buffer)(struct coresight_device *csdev,
+ struct perf_output_handle *handle,
+ void *sink_config, bool *lost);
+ void (*update_buffer)(struct coresight_device *csdev,
+ struct perf_output_handle *handle,
+ void *sink_config);
};
/**
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 10/26] coresight: etm3x: set progbit to stop trace collection |
| Message-ID | <qtM3U-lk-45@gated-at.bofh.it> |
| In reply to | #1267482 |
There is no need to use the event enable's "always false" event to stop trace collection. For that purpose setting the programming bit (ETMCR:10) is enough. Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> --- drivers/hwtracing/coresight/coresight-etm3x.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c index 5a2be0590754..d1235711b576 100644 --- a/drivers/hwtracing/coresight/coresight-etm3x.c +++ b/drivers/hwtracing/coresight/coresight-etm3x.c @@ -422,9 +422,6 @@ static void etm_disable_hw(void *info) CS_UNLOCK(drvdata->base); etm_set_prog(drvdata); - /* Program trace enable to low by using always false event */ - etm_writel(drvdata, ETM_HARD_WIRE_RES_A | ETM_EVENT_NOT_A, ETMTEEVR); - /* Read back sequencer and counters for post trace analysis */ config->seq_curr_state = (etm_readl(drvdata, ETMSQR) & ETM_SQR_MASK); -- 1.9.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 03/26] coresight: etm3x: unlocking tracers in default arch init |
| Message-ID | <qtM3V-lk-51@gated-at.bofh.it> |
| In reply to | #1267482 |
Calling function 'smp_call_function_single()' to unlock a
tracer and calling it again right after to perform the
default initialisation doesn't make sense.
Moving 'etm_os_unlock()' just before making the default
initialisation results in the same outcome while saving
one call to 'smp_call_function_single()'.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etm3x.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index 2a1950d0753d..737b6164f4a5 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -47,11 +47,11 @@ static struct etm_drvdata *etmdrvdata[NR_CPUS];
* and OS lock must be unlocked before any memory mapped access on such
* processors, otherwise memory mapped reads/writes will be invalid.
*/
-static void etm_os_unlock(void *info)
+static void etm_os_unlock(struct etm_drvdata *drvdata)
{
- struct etm_drvdata *drvdata = (struct etm_drvdata *)info;
/* Writing any value to ETMOSLAR unlocks the trace registers */
etm_writel(drvdata, 0x0, ETMOSLAR);
+ drvdata->os_unlock = true;
isb();
}
@@ -478,6 +478,9 @@ static void etm_init_arch_data(void *info)
u32 etmccr;
struct etm_drvdata *drvdata = info;
+ /* Make sure all registers are accessible */
+ etm_os_unlock(drvdata);
+
CS_UNLOCK(drvdata->base);
/* First dummy read */
@@ -602,9 +605,6 @@ static int etm_probe(struct amba_device *adev, const struct amba_id *id)
get_online_cpus();
etmdrvdata[drvdata->cpu] = drvdata;
- if (!smp_call_function_single(drvdata->cpu, etm_os_unlock, drvdata, 1))
- drvdata->os_unlock = true;
-
if (smp_call_function_single(drvdata->cpu,
etm_init_arch_data, drvdata, 1))
dev_err(dev, "ETM arch init failed\n");
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 14/26] coresight: etm3x: adding perf_get/set_config() API |
| Message-ID | <qtM3V-lk-59@gated-at.bofh.it> |
| In reply to | #1267482 |
Adding a source operation to build a tracer's configuration from
a perf_event. That way possibly complex parsing of the information
conveyed by the event doesn't have to be carried out every time
the configuration is needed.
Since event configuration can change between concurrent sessions,
the possibility of associating a tracer with a configuration is
also provided. As such Perf can assign session configuration to
tracers as it see fit.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/Kconfig | 1 +
drivers/hwtracing/coresight/coresight-etm3x.c | 62 +++++++++++++++++++++++++++
include/linux/coresight.h | 7 ++-
3 files changed, 69 insertions(+), 1 deletion(-)
diff --git a/drivers/hwtracing/coresight/Kconfig b/drivers/hwtracing/coresight/Kconfig
index 6c8921140f02..e252dd1522e5 100644
--- a/drivers/hwtracing/coresight/Kconfig
+++ b/drivers/hwtracing/coresight/Kconfig
@@ -4,6 +4,7 @@
menuconfig CORESIGHT
bool "CoreSight Tracing Support"
select ARM_AMBA
+ select PERF_EVENTS
help
This framework provides a kernel interface for the CoreSight debug
and trace drivers to register themselves with. It's intended to build
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index a83c67d13b21..dd319ef0f1ac 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -31,6 +31,7 @@
#include <linux/seq_file.h>
#include <linux/uaccess.h>
#include <linux/clk.h>
+#include <linux/perf_event.h>
#include <asm/sections.h>
#include "coresight-etm.h"
@@ -315,6 +316,40 @@ void etm_config_trace_mode(struct etm_drvdata *drvdata,
config->addr_type[1] = ETM_ADDR_TYPE_RANGE;
}
+#define ETM3X_SUPPORTED_OPTIONS (ETMCR_CYC_ACC | ETMCR_TIMESTAMP_EN)
+
+static int etm_parse_event_config(struct etm_drvdata *drvdata,
+ struct etm_config *config,
+ struct perf_event *event)
+{
+ u32 mode = 0;
+ u64 event_config = event->attr.config;
+
+ if (event->attr.exclude_kernel)
+ mode = ETM_MODE_EXCL_KERN;
+
+ if (event->attr.exclude_user)
+ mode = ETM_MODE_EXCL_USER;
+
+ /*
+ * By default the tracers are configured to trace the whole address
+ * range. Narrow the field only if requested by user space.
+ */
+ if (mode)
+ etm_config_trace_mode(drvdata, config, mode);
+
+ /*
+ * At this time only cycle accurate and timestamp options are
+ * available.
+ */
+ if (event_config & ~ETM3X_SUPPORTED_OPTIONS)
+ return -EINVAL;
+
+ config->ctrl = event_config;
+
+ return 0;
+}
+
static void etm_enable_hw(void *info)
{
int i;
@@ -428,6 +463,31 @@ static int etm_trace_id(struct coresight_device *csdev)
return etm_get_trace_id(drvdata);
}
+static void *etm_get_config(struct coresight_device *csdev,
+ struct perf_event *event)
+{
+ struct etm_config *config = NULL;
+ struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+
+ config = kzalloc(sizeof(struct etm_config), GFP_KERNEL);
+ if (!config)
+ return config;
+
+ etm_set_default(config);
+
+ if (etm_parse_event_config(drvdata, config, event))
+ return ERR_PTR(-EINVAL);
+
+ return config;
+}
+
+static void etm_set_config(struct coresight_device *csdev, void *config)
+{
+ struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+
+ drvdata->config = config;
+}
+
static int etm_enable_sysfs(struct coresight_device *csdev)
{
struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
@@ -568,6 +628,8 @@ static void etm_disable(struct coresight_device *csdev)
static const struct coresight_ops_source etm_source_ops = {
.cpu_id = etm_cpu_id,
.trace_id = etm_trace_id,
+ .get_config = etm_get_config,
+ .set_config = etm_set_config,
.enable = etm_enable,
.disable = etm_disable,
};
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 61dfb8d511ea..f2148bc2bcfe 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -206,13 +206,18 @@ struct coresight_ops_link {
* @cpu_id: returns the value of the CPU number this component
* is associated to.
* @trace_id: returns the value of the component's trace ID as known
- to the HW.
+ * to the HW.
+ * @get_config: builds the ETM configuration after events' specifics.
+ * @set_config: associate a tracer with a configuration.
* @enable: enables tracing for a source.
* @disable: disables tracing for a source.
*/
struct coresight_ops_source {
int (*cpu_id)(struct coresight_device *csdev);
int (*trace_id)(struct coresight_device *csdev);
+ void *(*get_config)(struct coresight_device *csdev,
+ struct perf_event *event);
+ void (*set_config)(struct coresight_device *csdev, void *config);
int (*enable)(struct coresight_device *csdev, u32 mode);
void (*disable)(struct coresight_device *csdev);
};
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 09/26] coresight: add API to get sink from path |
| Message-ID | <qtM3T-lk-15@gated-at.bofh.it> |
| In reply to | #1267482 |
Add an API allowing external code to quickly get a handle on the
sink within a path. The sink is always last, but adding an API allows
to keep the path's node structure private and remove redundant checks.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-priv.h | 1 +
drivers/hwtracing/coresight/coresight.c | 15 +++++++++++++++
2 files changed, 16 insertions(+)
diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h
index 2dc59a5c5b1e..ed116b303e87 100644
--- a/drivers/hwtracing/coresight/coresight-priv.h
+++ b/drivers/hwtracing/coresight/coresight-priv.h
@@ -60,6 +60,7 @@ static inline void CS_UNLOCK(void __iomem *addr)
void coresight_disable_path(struct list_head *path);
int coresight_enable_path(struct list_head *path);
+struct coresight_device *coresight_get_sink(struct list_head *path);
struct list_head *coresight_build_path(struct coresight_device *csdev);
void coresight_release_path(struct list_head *path);
diff --git a/drivers/hwtracing/coresight/coresight.c b/drivers/hwtracing/coresight/coresight.c
index 7e4064b31c99..47a4e109f700 100644
--- a/drivers/hwtracing/coresight/coresight.c
+++ b/drivers/hwtracing/coresight/coresight.c
@@ -322,6 +322,21 @@ err:
goto out;
}
+struct coresight_device *coresight_get_sink(struct list_head *path)
+{
+ struct coresight_device *csdev;
+
+ if (!path)
+ return NULL;
+
+ csdev = list_last_entry(path, struct coresight_node, link)->csdev;
+ if (csdev->type != CORESIGHT_DEV_TYPE_SINK &&
+ csdev->type != CORESIGHT_DEV_TYPE_LINKSINK)
+ return NULL;
+
+ return csdev;
+}
+
/**
* _coresight_build_path - recursively build a path from a @csdev to a sink.
* @csdev: The device to start from.
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 16/26] perf: changing pmu::setup_aux() parameter to include event |
| Message-ID | <qtM3V-lk-65@gated-at.bofh.it> |
| In reply to | #1267482 |
For Coresight ETMv3/4 tracers the event carries information
about trace modes such as user or kernel space tracing and
whether tracing is allowed when operating in secure mode.
Since this information is to be embedded in the private
structure returned by setup_aux(), changing the first
parameter to be of type struct perf_event * so that all
the necessary information can be conveyed. Also changing
current customer of the API to reflect the modification.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
arch/x86/kernel/cpu/perf_event_intel_bts.c | 4 +++-
arch/x86/kernel/cpu/perf_event_intel_pt.c | 5 +++--
include/linux/perf_event.h | 2 +-
kernel/events/ring_buffer.c | 2 +-
4 files changed, 8 insertions(+), 5 deletions(-)
diff --git a/arch/x86/kernel/cpu/perf_event_intel_bts.c b/arch/x86/kernel/cpu/perf_event_intel_bts.c
index d1c0f254afbe..773600b2f313 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_bts.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_bts.c
@@ -68,10 +68,12 @@ static size_t buf_size(struct page *page)
}
static void *
-bts_buffer_setup_aux(int cpu, void **pages, int nr_pages, bool overwrite)
+bts_buffer_setup_aux(struct perf_event *event, void **pages,
+ int nr_pages, bool overwrite)
{
struct bts_buffer *buf;
struct page *page;
+ int cpu = event->cpu;
int node = (cpu == -1) ? cpu : cpu_to_node(cpu);
unsigned long offset;
size_t size = nr_pages << PAGE_SHIFT;
diff --git a/arch/x86/kernel/cpu/perf_event_intel_pt.c b/arch/x86/kernel/cpu/perf_event_intel_pt.c
index 42169283448b..f8b881210648 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_pt.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_pt.c
@@ -863,10 +863,11 @@ static int pt_buffer_init_topa(struct pt_buffer *buf, unsigned long nr_pages,
* Return: Our private PT buffer structure.
*/
static void *
-pt_buffer_setup_aux(int cpu, void **pages, int nr_pages, bool snapshot)
+pt_buffer_setup_aux(struct perf_event *event, void **pages,
+ int nr_pages, bool snapshot)
{
struct pt_buffer *buf;
- int node, ret;
+ int node, ret, cpu = event->cpu;
if (!nr_pages)
return NULL;
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 092a0e8a479a..0d9964b5ed4d 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -292,7 +292,7 @@ struct pmu {
/*
* Set up pmu-private data structures for an AUX area
*/
- void *(*setup_aux) (int cpu, void **pages,
+ void *(*setup_aux) (struct perf_event *event, void **pages,
int nr_pages, bool overwrite);
/* optional */
diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
index 182bc30899d5..87171b700c77 100644
--- a/kernel/events/ring_buffer.c
+++ b/kernel/events/ring_buffer.c
@@ -522,7 +522,7 @@ int rb_alloc_aux(struct ring_buffer *rb, struct perf_event *event,
goto out;
}
- rb->aux_priv = event->pmu->setup_aux(event->cpu, rb->aux_pages, nr_pages,
+ rb->aux_priv = event->pmu->setup_aux(event, rb->aux_pages, nr_pages,
overwrite);
if (!rb->aux_priv)
goto out;
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 18/26] coresight: adding operation mode for sink->enable() |
| Message-ID | <qtM3V-lk-63@gated-at.bofh.it> |
| In reply to | #1267482 |
Adding an operation mode to sink->enable() API in order
to prevent simultaneous access from different callers.
TPIU and TMC won't be supplemented with the AUX area
API immediately and as such ignore the new mode.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etb10.c | 31 +++++++++++++++++++++------
drivers/hwtracing/coresight/coresight-priv.h | 2 +-
drivers/hwtracing/coresight/coresight-tmc.c | 2 +-
drivers/hwtracing/coresight/coresight-tpiu.c | 2 +-
drivers/hwtracing/coresight/coresight.c | 10 ++++-----
include/linux/coresight.h | 2 +-
6 files changed, 33 insertions(+), 16 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etb10.c b/drivers/hwtracing/coresight/coresight-etb10.c
index 09e306361e4f..dc4d707b28aa 100644
--- a/drivers/hwtracing/coresight/coresight-etb10.c
+++ b/drivers/hwtracing/coresight/coresight-etb10.c
@@ -73,9 +73,9 @@
* @miscdev: specifics to handle "/dev/xyz.etb" entry.
* @spinlock: only one at a time pls.
* @reading: synchronise user space access to etb buffer.
+ * @mode: this ETB is being used.
* @buf: area of memory where ETB buffer content gets sent.
* @buffer_depth: size of @buf.
- * @enable: this ETB is being used.
* @trigger_cntr: amount of words to store after a trigger.
*/
struct etb_drvdata {
@@ -86,9 +86,9 @@ struct etb_drvdata {
struct miscdevice miscdev;
spinlock_t spinlock;
local_t reading;
+ local_t mode;
u8 *buf;
u32 buffer_depth;
- bool enable;
u32 trigger_cntr;
};
@@ -133,18 +133,34 @@ static void etb_enable_hw(struct etb_drvdata *drvdata)
CS_LOCK(drvdata->base);
}
-static int etb_enable(struct coresight_device *csdev)
+static int etb_enable(struct coresight_device *csdev, u32 mode)
{
- struct etb_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+ u32 val;
unsigned long flags;
+ struct etb_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+
+ val = local_cmpxchg(&drvdata->mode,
+ CS_MODE_DISABLED, mode);
+ /*
+ * When accessing from Perf, a HW buffer can be handled
+ * by a single trace entity. In sysFS mode many tracers
+ * can be logging to the same HW buffer.
+ */
+ if (val == CS_MODE_PERF)
+ return -EBUSY;
+
+ /* Nothing to do, the tracer is already enabled. */
+ if (val == CS_MODE_SYSFS)
+ goto out;
spin_lock_irqsave(&drvdata->spinlock, flags);
etb_enable_hw(drvdata);
- drvdata->enable = true;
spin_unlock_irqrestore(&drvdata->spinlock, flags);
+out:
dev_info(drvdata->dev, "ETB enabled\n");
return 0;
+
}
static void etb_disable_hw(struct etb_drvdata *drvdata)
@@ -243,9 +259,10 @@ static void etb_disable(struct coresight_device *csdev)
spin_lock_irqsave(&drvdata->spinlock, flags);
etb_disable_hw(drvdata);
etb_dump_hw(drvdata);
- drvdata->enable = false;
spin_unlock_irqrestore(&drvdata->spinlock, flags);
+ local_set(&drvdata->mode, CS_MODE_DISABLED);
+
dev_info(drvdata->dev, "ETB disabled\n");
}
@@ -263,7 +280,7 @@ static void etb_dump(struct etb_drvdata *drvdata)
unsigned long flags;
spin_lock_irqsave(&drvdata->spinlock, flags);
- if (drvdata->enable) {
+ if (local_read(&drvdata->mode) == CS_MODE_SYSFS) {
etb_disable_hw(drvdata);
etb_dump_hw(drvdata);
etb_enable_hw(drvdata);
diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h
index 932f34a84d96..333eddaed339 100644
--- a/drivers/hwtracing/coresight/coresight-priv.h
+++ b/drivers/hwtracing/coresight/coresight-priv.h
@@ -62,7 +62,7 @@ static inline void CS_UNLOCK(void __iomem *addr)
}
void coresight_disable_path(struct list_head *path);
-int coresight_enable_path(struct list_head *path);
+int coresight_enable_path(struct list_head *path, u32 mode);
struct coresight_device *coresight_get_sink(struct list_head *path);
struct list_head *coresight_build_path(struct coresight_device *csdev);
void coresight_release_path(struct list_head *path);
diff --git a/drivers/hwtracing/coresight/coresight-tmc.c b/drivers/hwtracing/coresight/coresight-tmc.c
index 5e2a71767870..d7e83da7a621 100644
--- a/drivers/hwtracing/coresight/coresight-tmc.c
+++ b/drivers/hwtracing/coresight/coresight-tmc.c
@@ -265,7 +265,7 @@ static int tmc_enable(struct tmc_drvdata *drvdata, enum tmc_mode mode)
return 0;
}
-static int tmc_enable_sink(struct coresight_device *csdev)
+static int tmc_enable_sink(struct coresight_device *csdev, u32 mode)
{
struct tmc_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
diff --git a/drivers/hwtracing/coresight/coresight-tpiu.c b/drivers/hwtracing/coresight/coresight-tpiu.c
index e19b86e61c38..0a09135382f6 100644
--- a/drivers/hwtracing/coresight/coresight-tpiu.c
+++ b/drivers/hwtracing/coresight/coresight-tpiu.c
@@ -70,7 +70,7 @@ static void tpiu_enable_hw(struct tpiu_drvdata *drvdata)
CS_LOCK(drvdata->base);
}
-static int tpiu_enable(struct coresight_device *csdev)
+static int tpiu_enable(struct coresight_device *csdev, u32 mode)
{
struct tpiu_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
diff --git a/drivers/hwtracing/coresight/coresight.c b/drivers/hwtracing/coresight/coresight.c
index 47a4e109f700..52311e77da9d 100644
--- a/drivers/hwtracing/coresight/coresight.c
+++ b/drivers/hwtracing/coresight/coresight.c
@@ -121,13 +121,13 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
return 0;
}
-static int coresight_enable_sink(struct coresight_device *csdev)
+static int coresight_enable_sink(struct coresight_device *csdev, u32 mode)
{
int ret;
if (!csdev->enable) {
if (sink_ops(csdev)->enable) {
- ret = sink_ops(csdev)->enable(csdev);
+ ret = sink_ops(csdev)->enable(csdev, mode);
if (ret)
return ret;
}
@@ -283,7 +283,7 @@ void coresight_disable_path(struct list_head *path)
}
}
-int coresight_enable_path(struct list_head *path)
+int coresight_enable_path(struct list_head *path, u32 mode)
{
int ret = 0;
@@ -296,7 +296,7 @@ int coresight_enable_path(struct list_head *path)
switch (csdev->type) {
case CORESIGHT_DEV_TYPE_SINK:
case CORESIGHT_DEV_TYPE_LINKSINK:
- ret = coresight_enable_sink(csdev);
+ ret = coresight_enable_sink(csdev, mode);
if (ret)
goto err;
break;
@@ -454,7 +454,7 @@ int coresight_enable(struct coresight_device *csdev)
goto out;
}
- ret = coresight_enable_path(path);
+ ret = coresight_enable_path(path, CS_MODE_SYSFS);
if (ret)
goto err_path;
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index f2148bc2bcfe..80b3465bd459 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -185,7 +185,7 @@ struct coresight_device {
* @disable: disables the sink.
*/
struct coresight_ops_sink {
- int (*enable)(struct coresight_device *csdev);
+ int (*enable)(struct coresight_device *csdev, u32 mode);
void (*disable)(struct coresight_device *csdev);
};
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Mathieu Poirier <mathieu.poirier@linaro.org> |
|---|---|
| Date | 2015-11-11 23:30 +0100 |
| Subject | [PATCH V3 01/26] coresight: etm3x: moving etm_readl/writel to header file |
| Message-ID | <qtM3V-lk-73@gated-at.bofh.it> |
| In reply to | #1267482 |
Moving functions etm_readl/writel to file "coresight-etm.h"
so that the main ETM3x driver can be split in more than one
file.
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
---
drivers/hwtracing/coresight/coresight-etm.h | 29 +++++++++++++++++++++++++++
drivers/hwtracing/coresight/coresight-etm3x.c | 29 ---------------------------
2 files changed, 29 insertions(+), 29 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-etm.h b/drivers/hwtracing/coresight/coresight-etm.h
index b4481eb29304..34f7db881fa7 100644
--- a/drivers/hwtracing/coresight/coresight-etm.h
+++ b/drivers/hwtracing/coresight/coresight-etm.h
@@ -251,4 +251,33 @@ enum etm_addr_type {
ETM_ADDR_TYPE_START,
ETM_ADDR_TYPE_STOP,
};
+
+static inline void etm_writel(struct etm_drvdata *drvdata,
+ u32 val, u32 off)
+{
+ if (drvdata->use_cp14) {
+ if (etm_writel_cp14(off, val)) {
+ dev_err(drvdata->dev,
+ "invalid CP14 access to ETM reg: %#x", off);
+ }
+ } else {
+ writel_relaxed(val, drvdata->base + off);
+ }
+}
+
+static inline unsigned int etm_readl(struct etm_drvdata *drvdata, u32 off)
+{
+ u32 val;
+
+ if (drvdata->use_cp14) {
+ if (etm_readl_cp14(off, &val)) {
+ dev_err(drvdata->dev,
+ "invalid CP14 access to ETM reg: %#x", off);
+ }
+ } else {
+ val = readl_relaxed(drvdata->base + off);
+ }
+
+ return val;
+}
#endif
diff --git a/drivers/hwtracing/coresight/coresight-etm3x.c b/drivers/hwtracing/coresight/coresight-etm3x.c
index d630b7ece735..c1dc095c3fb0 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x.c
@@ -42,35 +42,6 @@ module_param_named(boot_enable, boot_enable, int, S_IRUGO);
static int etm_count;
static struct etm_drvdata *etmdrvdata[NR_CPUS];
-static inline void etm_writel(struct etm_drvdata *drvdata,
- u32 val, u32 off)
-{
- if (drvdata->use_cp14) {
- if (etm_writel_cp14(off, val)) {
- dev_err(drvdata->dev,
- "invalid CP14 access to ETM reg: %#x", off);
- }
- } else {
- writel_relaxed(val, drvdata->base + off);
- }
-}
-
-static inline unsigned int etm_readl(struct etm_drvdata *drvdata, u32 off)
-{
- u32 val;
-
- if (drvdata->use_cp14) {
- if (etm_readl_cp14(off, &val)) {
- dev_err(drvdata->dev,
- "invalid CP14 access to ETM reg: %#x", off);
- }
- } else {
- val = readl_relaxed(drvdata->base + off);
- }
-
- return val;
-}
-
/*
* Memory mapped writes to clear os lock are not supported on some processors
* and OS lock must be unlocked before any memory mapped access on such
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
Page 1 of 2 [1] 2 Next page →
Back to top | Article view | linux.kernel
csiph-web