* [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device
@ 2023-03-10 16:05 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:05 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
Changes since v1:
* Don't dereference handle in tmc_etr_get_buffer() when not in perf mode.
* Fix some W=1 warnings
* Add a commit to rename child/output in terms of local/remote
-------------------
Currently there is a refcount leak in CTI when using system wide mode
or tracing multithreaded applications. See the last commit for a
reproducer. This prevents the module from being unloaded.
Historically there have been a few issues and fixes attempted around
here which have resulted in some extra logic and a member to keep
track of CTI being enabled 'struct coresight_device->ect_enabled'.
The fix in commit 665c157e0204 ("coresight: cti: Fix hang in
cti_disable_hw()") was also related to CTI having its own
enable/disable path which came later than other devices.
If we make CTI a helper device and enable helper devices adjacent to
the path we get very similar enable/disable behavior to now, but with
more reuse of the existing reference counting logic in the coresight
core code. This also affects CATU which can have a little bit of
its hard coded enable/disable code removed.
Enabling CATU on the generic path does require that input connections
are tracked so that it can get its associated ETR buffer.
Applies to coresight/next (669c4614236a7) but also requires the
realloc_array patch here [1].
Also available in full here [2].
[1]: https://lore.kernel.org/linux-arm-kernel/20230306152723.3090195-1-james.clark@arm.com/
[2]: https://gitlab.arm.com/linux-arm/linux-jc/-/tree/james-cs-cti-module-refcount-fix-v2
James Clark (9):
coresight: Use enum type for cs_mode wherever possible
coresight: Change name of pdata->conns
coresight: Rename nr_outports to nr_outconns
coresight: Rename connection members to allow for input connections
coresight: Dynamically add connections
coresight: Store in-connections as well as out-connections
coresight: Refactor out buffer allocation function for ETR
coresight: Enable and disable helper devices adjacent to the path
coresight: Fix CTI module refcount leak by making it a helper device
drivers/hwtracing/coresight/coresight-catu.c | 34 +-
drivers/hwtracing/coresight/coresight-core.c | 312 +++++++++++-------
.../hwtracing/coresight/coresight-cti-core.c | 56 ++--
.../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
drivers/hwtracing/coresight/coresight-cti.h | 4 +-
drivers/hwtracing/coresight/coresight-etb10.c | 3 +-
.../coresight/coresight-etm3x-core.c | 6 +-
.../coresight/coresight-etm4x-core.c | 6 +-
.../hwtracing/coresight/coresight-platform.c | 178 +++++++---
drivers/hwtracing/coresight/coresight-priv.h | 9 +-
drivers/hwtracing/coresight/coresight-stm.c | 6 +-
drivers/hwtracing/coresight/coresight-sysfs.c | 9 +-
.../hwtracing/coresight/coresight-tmc-etf.c | 2 +-
.../hwtracing/coresight/coresight-tmc-etr.c | 89 ++---
drivers/hwtracing/coresight/coresight-tmc.h | 2 +
drivers/hwtracing/coresight/coresight-tpdm.c | 4 +-
drivers/hwtracing/coresight/coresight-tpiu.c | 3 +-
drivers/hwtracing/coresight/coresight-trbe.c | 3 +-
drivers/hwtracing/coresight/ultrasoc-smb.c | 3 +-
drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
include/linux/coresight.h | 109 +++---
21 files changed, 530 insertions(+), 314 deletions(-)
--
2.34.1
^ permalink raw reply [flat|nested] 74+ messages in thread
* [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device
@ 2023-03-10 16:05 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:05 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
Changes since v1:
* Don't dereference handle in tmc_etr_get_buffer() when not in perf mode.
* Fix some W=1 warnings
* Add a commit to rename child/output in terms of local/remote
-------------------
Currently there is a refcount leak in CTI when using system wide mode
or tracing multithreaded applications. See the last commit for a
reproducer. This prevents the module from being unloaded.
Historically there have been a few issues and fixes attempted around
here which have resulted in some extra logic and a member to keep
track of CTI being enabled 'struct coresight_device->ect_enabled'.
The fix in commit 665c157e0204 ("coresight: cti: Fix hang in
cti_disable_hw()") was also related to CTI having its own
enable/disable path which came later than other devices.
If we make CTI a helper device and enable helper devices adjacent to
the path we get very similar enable/disable behavior to now, but with
more reuse of the existing reference counting logic in the coresight
core code. This also affects CATU which can have a little bit of
its hard coded enable/disable code removed.
Enabling CATU on the generic path does require that input connections
are tracked so that it can get its associated ETR buffer.
Applies to coresight/next (669c4614236a7) but also requires the
realloc_array patch here [1].
Also available in full here [2].
[1]: https://lore.kernel.org/linux-arm-kernel/20230306152723.3090195-1-james.clark@arm.com/
[2]: https://gitlab.arm.com/linux-arm/linux-jc/-/tree/james-cs-cti-module-refcount-fix-v2
James Clark (9):
coresight: Use enum type for cs_mode wherever possible
coresight: Change name of pdata->conns
coresight: Rename nr_outports to nr_outconns
coresight: Rename connection members to allow for input connections
coresight: Dynamically add connections
coresight: Store in-connections as well as out-connections
coresight: Refactor out buffer allocation function for ETR
coresight: Enable and disable helper devices adjacent to the path
coresight: Fix CTI module refcount leak by making it a helper device
drivers/hwtracing/coresight/coresight-catu.c | 34 +-
drivers/hwtracing/coresight/coresight-core.c | 312 +++++++++++-------
.../hwtracing/coresight/coresight-cti-core.c | 56 ++--
.../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
drivers/hwtracing/coresight/coresight-cti.h | 4 +-
drivers/hwtracing/coresight/coresight-etb10.c | 3 +-
.../coresight/coresight-etm3x-core.c | 6 +-
.../coresight/coresight-etm4x-core.c | 6 +-
.../hwtracing/coresight/coresight-platform.c | 178 +++++++---
drivers/hwtracing/coresight/coresight-priv.h | 9 +-
drivers/hwtracing/coresight/coresight-stm.c | 6 +-
drivers/hwtracing/coresight/coresight-sysfs.c | 9 +-
.../hwtracing/coresight/coresight-tmc-etf.c | 2 +-
.../hwtracing/coresight/coresight-tmc-etr.c | 89 ++---
drivers/hwtracing/coresight/coresight-tmc.h | 2 +
drivers/hwtracing/coresight/coresight-tpdm.c | 4 +-
drivers/hwtracing/coresight/coresight-tpiu.c | 3 +-
drivers/hwtracing/coresight/coresight-trbe.c | 3 +-
drivers/hwtracing/coresight/ultrasoc-smb.c | 3 +-
drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
include/linux/coresight.h | 109 +++---
21 files changed, 530 insertions(+), 314 deletions(-)
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* [PATCH v2 1/9] coresight: Use enum type for cs_mode wherever possible
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
mode is stored as a local_t, but it is also passed around a lot as a
plain u32, so use the correct type wherever local_t isn't currently
used. This helps a little bit with readability.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 7 ++++---
drivers/hwtracing/coresight/coresight-etb10.c | 3 ++-
drivers/hwtracing/coresight/coresight-etm3x-core.c | 6 +++---
drivers/hwtracing/coresight/coresight-etm4x-core.c | 6 +++---
drivers/hwtracing/coresight/coresight-priv.h | 9 ++-------
drivers/hwtracing/coresight/coresight-stm.c | 6 +++---
drivers/hwtracing/coresight/coresight-tmc-etf.c | 2 +-
drivers/hwtracing/coresight/coresight-tmc-etr.c | 7 +++----
drivers/hwtracing/coresight/coresight-tpdm.c | 4 ++--
drivers/hwtracing/coresight/coresight-tpiu.c | 3 ++-
drivers/hwtracing/coresight/coresight-trbe.c | 3 ++-
drivers/hwtracing/coresight/ultrasoc-smb.c | 3 ++-
drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
include/linux/coresight.h | 13 ++++++++++---
14 files changed, 40 insertions(+), 34 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index d3bf82c0de1d..e3b63fd52b9c 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -308,7 +308,7 @@ void coresight_set_assoc_ectdev_mutex(struct coresight_device *csdev,
EXPORT_SYMBOL_GPL(coresight_set_assoc_ectdev_mutex);
static int coresight_enable_sink(struct coresight_device *csdev,
- u32 mode, void *data)
+ enum cs_mode mode, void *data)
{
int ret;
@@ -416,7 +416,8 @@ static void coresight_disable_link(struct coresight_device *csdev,
csdev->enable = false;
}
-static int coresight_enable_source(struct coresight_device *csdev, u32 mode)
+static int coresight_enable_source(struct coresight_device *csdev,
+ enum cs_mode mode)
{
int ret;
@@ -516,7 +517,7 @@ void coresight_disable_path(struct list_head *path)
}
EXPORT_SYMBOL_GPL(coresight_disable_path);
-int coresight_enable_path(struct list_head *path, u32 mode, void *sink_data)
+int coresight_enable_path(struct list_head *path, enum cs_mode mode, void *sink_data)
{
int ret = 0;
diff --git a/drivers/hwtracing/coresight/coresight-etb10.c b/drivers/hwtracing/coresight/coresight-etb10.c
index 8aa6e4f83e42..eb99c445015a 100644
--- a/drivers/hwtracing/coresight/coresight-etb10.c
+++ b/drivers/hwtracing/coresight/coresight-etb10.c
@@ -225,7 +225,8 @@ static int etb_enable_perf(struct coresight_device *csdev, void *data)
return ret;
}
-static int etb_enable(struct coresight_device *csdev, u32 mode, void *data)
+static int etb_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *data)
{
int ret;
diff --git a/drivers/hwtracing/coresight/coresight-etm3x-core.c b/drivers/hwtracing/coresight/coresight-etm3x-core.c
index afc57195ee52..116a91d90ac2 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x-core.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x-core.c
@@ -552,8 +552,8 @@ static int etm_enable_sysfs(struct coresight_device *csdev)
return ret;
}
-static int etm_enable(struct coresight_device *csdev,
- struct perf_event *event, u32 mode)
+static int etm_enable(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode)
{
int ret;
u32 val;
@@ -671,7 +671,7 @@ static void etm_disable_sysfs(struct coresight_device *csdev)
static void etm_disable(struct coresight_device *csdev,
struct perf_event *event)
{
- u32 mode;
+ enum cs_mode mode;
struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
/*
diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c b/drivers/hwtracing/coresight/coresight-etm4x-core.c
index 1ea8f173cca0..2ff4c42ea688 100644
--- a/drivers/hwtracing/coresight/coresight-etm4x-core.c
+++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c
@@ -822,8 +822,8 @@ static int etm4_enable_sysfs(struct coresight_device *csdev)
return ret;
}
-static int etm4_enable(struct coresight_device *csdev,
- struct perf_event *event, u32 mode)
+static int etm4_enable(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode)
{
int ret;
u32 val;
@@ -989,7 +989,7 @@ static void etm4_disable_sysfs(struct coresight_device *csdev)
static void etm4_disable(struct coresight_device *csdev,
struct perf_event *event)
{
- u32 mode;
+ enum cs_mode mode;
struct etmv4_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
/*
diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h
index 595ce5862056..788ff19c60f6 100644
--- a/drivers/hwtracing/coresight/coresight-priv.h
+++ b/drivers/hwtracing/coresight/coresight-priv.h
@@ -82,12 +82,6 @@ enum etm_addr_type {
ETM_ADDR_TYPE_STOP,
};
-enum cs_mode {
- CS_MODE_DISABLED,
- CS_MODE_SYSFS,
- CS_MODE_PERF,
-};
-
/**
* struct cs_buffer - keep track of a recording session' specifics
* @cur: index of the current buffer
@@ -133,7 +127,8 @@ static inline void CS_UNLOCK(void __iomem *addr)
}
void coresight_disable_path(struct list_head *path);
-int coresight_enable_path(struct list_head *path, u32 mode, void *sink_data);
+int coresight_enable_path(struct list_head *path, enum cs_mode mode,
+ void *sink_data);
struct coresight_device *coresight_get_sink(struct list_head *path);
struct coresight_device *
coresight_get_enabled_sink(struct coresight_device *source);
diff --git a/drivers/hwtracing/coresight/coresight-stm.c b/drivers/hwtracing/coresight/coresight-stm.c
index 66a614c5492c..a1c27c901ad1 100644
--- a/drivers/hwtracing/coresight/coresight-stm.c
+++ b/drivers/hwtracing/coresight/coresight-stm.c
@@ -119,7 +119,7 @@ DEFINE_CORESIGHT_DEVLIST(stm_devs, "stm");
* @spinlock: only one at a time pls.
* @chs: the channels accociated to this STM.
* @stm: structure associated to the generic STM interface.
- * @mode: this tracer's mode, i.e sysFS, or disabled.
+ * @mode: this tracer's mode (enum cs_mode), i.e sysFS, or disabled.
* @traceid: value of the current ID for this component.
* @write_bytes: Maximus bytes this STM can write at a time.
* @stmsper: settings for register STMSPER.
@@ -192,8 +192,8 @@ static void stm_enable_hw(struct stm_drvdata *drvdata)
CS_LOCK(drvdata->base);
}
-static int stm_enable(struct coresight_device *csdev,
- struct perf_event *event, u32 mode)
+static int stm_enable(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode)
{
u32 val;
struct stm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etf.c b/drivers/hwtracing/coresight/coresight-tmc-etf.c
index 0ab1f73c2d06..14d3c1472455 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etf.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etf.c
@@ -302,7 +302,7 @@ static int tmc_enable_etf_sink_perf(struct coresight_device *csdev, void *data)
}
static int tmc_enable_etf_sink(struct coresight_device *csdev,
- u32 mode, void *data)
+ enum cs_mode mode, void *data)
{
int ret;
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 918d461fcf4a..ed589cfff1b5 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -1666,17 +1666,16 @@ static int tmc_enable_etr_sink_perf(struct coresight_device *csdev, void *data)
}
static int tmc_enable_etr_sink(struct coresight_device *csdev,
- u32 mode, void *data)
+ enum cs_mode mode, void *data)
{
switch (mode) {
case CS_MODE_SYSFS:
return tmc_enable_etr_sink_sysfs(csdev);
case CS_MODE_PERF:
return tmc_enable_etr_sink_perf(csdev, data);
+ default:
+ return -EINVAL;
}
-
- /* We shouldn't be here */
- return -EINVAL;
}
static int tmc_disable_etr_sink(struct coresight_device *csdev)
diff --git a/drivers/hwtracing/coresight/coresight-tpdm.c b/drivers/hwtracing/coresight/coresight-tpdm.c
index 9479a5e8c672..f4854af0431e 100644
--- a/drivers/hwtracing/coresight/coresight-tpdm.c
+++ b/drivers/hwtracing/coresight/coresight-tpdm.c
@@ -42,8 +42,8 @@ static void __tpdm_enable(struct tpdm_drvdata *drvdata)
CS_LOCK(drvdata->base);
}
-static int tpdm_enable(struct coresight_device *csdev,
- struct perf_event *event, u32 mode)
+static int tpdm_enable(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode)
{
struct tpdm_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 34d37abd2c8d..b0179f761c98 100644
--- a/drivers/hwtracing/coresight/coresight-tpiu.c
+++ b/drivers/hwtracing/coresight/coresight-tpiu.c
@@ -69,7 +69,8 @@ static void tpiu_enable_hw(struct csdev_access *csa)
CS_LOCK(csa->base);
}
-static int tpiu_enable(struct coresight_device *csdev, u32 mode, void *__unused)
+static int tpiu_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *__unused)
{
tpiu_enable_hw(&csdev->access);
atomic_inc(csdev->refcnt);
diff --git a/drivers/hwtracing/coresight/coresight-trbe.c b/drivers/hwtracing/coresight/coresight-trbe.c
index 1fc4fd79a1c6..5406e338c280 100644
--- a/drivers/hwtracing/coresight/coresight-trbe.c
+++ b/drivers/hwtracing/coresight/coresight-trbe.c
@@ -1005,7 +1005,8 @@ static int __arm_trbe_enable(struct trbe_buf *buf,
return ret;
}
-static int arm_trbe_enable(struct coresight_device *csdev, u32 mode, void *data)
+static int arm_trbe_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *data)
{
struct trbe_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
struct trbe_cpudata *cpudata = dev_get_drvdata(&csdev->dev);
diff --git a/drivers/hwtracing/coresight/ultrasoc-smb.c b/drivers/hwtracing/coresight/ultrasoc-smb.c
index b317342c7ce5..a2ced0b3c3cd 100644
--- a/drivers/hwtracing/coresight/ultrasoc-smb.c
+++ b/drivers/hwtracing/coresight/ultrasoc-smb.c
@@ -256,7 +256,8 @@ static int smb_enable_perf(struct coresight_device *csdev, void *data)
return 0;
}
-static int smb_enable(struct coresight_device *csdev, u32 mode, void *data)
+static int smb_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *data)
{
struct smb_drv_data *drvdata = dev_get_drvdata(csdev->dev.parent);
int ret = 0;
diff --git a/drivers/hwtracing/coresight/ultrasoc-smb.h b/drivers/hwtracing/coresight/ultrasoc-smb.h
index 7dfbe42e37a0..d2e14e8d2c8a 100644
--- a/drivers/hwtracing/coresight/ultrasoc-smb.h
+++ b/drivers/hwtracing/coresight/ultrasoc-smb.h
@@ -119,7 +119,7 @@ struct smb_drv_data {
struct mutex mutex;
bool reading;
pid_t pid;
- u32 mode;
+ enum cs_mode mode;
};
#endif
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index f19a47b9bb5a..1f878e8ed8c4 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -272,6 +272,12 @@ static struct coresight_dev_list (var) = { \
#define to_coresight_device(d) container_of(d, struct coresight_device, dev)
+enum cs_mode {
+ CS_MODE_DISABLED,
+ CS_MODE_SYSFS,
+ CS_MODE_PERF,
+};
+
#define source_ops(csdev) csdev->ops->source_ops
#define sink_ops(csdev) csdev->ops->sink_ops
#define link_ops(csdev) csdev->ops->link_ops
@@ -288,7 +294,8 @@ static struct coresight_dev_list (var) = { \
* @update_buffer: update buffer pointers after a trace session.
*/
struct coresight_ops_sink {
- int (*enable)(struct coresight_device *csdev, u32 mode, void *data);
+ int (*enable)(struct coresight_device *csdev, enum cs_mode mode,
+ void *data);
int (*disable)(struct coresight_device *csdev);
void *(*alloc_buffer)(struct coresight_device *csdev,
struct perf_event *event, void **pages,
@@ -320,8 +327,8 @@ struct coresight_ops_link {
*/
struct coresight_ops_source {
int (*cpu_id)(struct coresight_device *csdev);
- int (*enable)(struct coresight_device *csdev,
- struct perf_event *event, u32 mode);
+ int (*enable)(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode);
void (*disable)(struct coresight_device *csdev,
struct perf_event *event);
};
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 1/9] coresight: Use enum type for cs_mode wherever possible
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
mode is stored as a local_t, but it is also passed around a lot as a
plain u32, so use the correct type wherever local_t isn't currently
used. This helps a little bit with readability.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 7 ++++---
drivers/hwtracing/coresight/coresight-etb10.c | 3 ++-
drivers/hwtracing/coresight/coresight-etm3x-core.c | 6 +++---
drivers/hwtracing/coresight/coresight-etm4x-core.c | 6 +++---
drivers/hwtracing/coresight/coresight-priv.h | 9 ++-------
drivers/hwtracing/coresight/coresight-stm.c | 6 +++---
drivers/hwtracing/coresight/coresight-tmc-etf.c | 2 +-
drivers/hwtracing/coresight/coresight-tmc-etr.c | 7 +++----
drivers/hwtracing/coresight/coresight-tpdm.c | 4 ++--
drivers/hwtracing/coresight/coresight-tpiu.c | 3 ++-
drivers/hwtracing/coresight/coresight-trbe.c | 3 ++-
drivers/hwtracing/coresight/ultrasoc-smb.c | 3 ++-
drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
include/linux/coresight.h | 13 ++++++++++---
14 files changed, 40 insertions(+), 34 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index d3bf82c0de1d..e3b63fd52b9c 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -308,7 +308,7 @@ void coresight_set_assoc_ectdev_mutex(struct coresight_device *csdev,
EXPORT_SYMBOL_GPL(coresight_set_assoc_ectdev_mutex);
static int coresight_enable_sink(struct coresight_device *csdev,
- u32 mode, void *data)
+ enum cs_mode mode, void *data)
{
int ret;
@@ -416,7 +416,8 @@ static void coresight_disable_link(struct coresight_device *csdev,
csdev->enable = false;
}
-static int coresight_enable_source(struct coresight_device *csdev, u32 mode)
+static int coresight_enable_source(struct coresight_device *csdev,
+ enum cs_mode mode)
{
int ret;
@@ -516,7 +517,7 @@ void coresight_disable_path(struct list_head *path)
}
EXPORT_SYMBOL_GPL(coresight_disable_path);
-int coresight_enable_path(struct list_head *path, u32 mode, void *sink_data)
+int coresight_enable_path(struct list_head *path, enum cs_mode mode, void *sink_data)
{
int ret = 0;
diff --git a/drivers/hwtracing/coresight/coresight-etb10.c b/drivers/hwtracing/coresight/coresight-etb10.c
index 8aa6e4f83e42..eb99c445015a 100644
--- a/drivers/hwtracing/coresight/coresight-etb10.c
+++ b/drivers/hwtracing/coresight/coresight-etb10.c
@@ -225,7 +225,8 @@ static int etb_enable_perf(struct coresight_device *csdev, void *data)
return ret;
}
-static int etb_enable(struct coresight_device *csdev, u32 mode, void *data)
+static int etb_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *data)
{
int ret;
diff --git a/drivers/hwtracing/coresight/coresight-etm3x-core.c b/drivers/hwtracing/coresight/coresight-etm3x-core.c
index afc57195ee52..116a91d90ac2 100644
--- a/drivers/hwtracing/coresight/coresight-etm3x-core.c
+++ b/drivers/hwtracing/coresight/coresight-etm3x-core.c
@@ -552,8 +552,8 @@ static int etm_enable_sysfs(struct coresight_device *csdev)
return ret;
}
-static int etm_enable(struct coresight_device *csdev,
- struct perf_event *event, u32 mode)
+static int etm_enable(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode)
{
int ret;
u32 val;
@@ -671,7 +671,7 @@ static void etm_disable_sysfs(struct coresight_device *csdev)
static void etm_disable(struct coresight_device *csdev,
struct perf_event *event)
{
- u32 mode;
+ enum cs_mode mode;
struct etm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
/*
diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c b/drivers/hwtracing/coresight/coresight-etm4x-core.c
index 1ea8f173cca0..2ff4c42ea688 100644
--- a/drivers/hwtracing/coresight/coresight-etm4x-core.c
+++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c
@@ -822,8 +822,8 @@ static int etm4_enable_sysfs(struct coresight_device *csdev)
return ret;
}
-static int etm4_enable(struct coresight_device *csdev,
- struct perf_event *event, u32 mode)
+static int etm4_enable(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode)
{
int ret;
u32 val;
@@ -989,7 +989,7 @@ static void etm4_disable_sysfs(struct coresight_device *csdev)
static void etm4_disable(struct coresight_device *csdev,
struct perf_event *event)
{
- u32 mode;
+ enum cs_mode mode;
struct etmv4_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
/*
diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h
index 595ce5862056..788ff19c60f6 100644
--- a/drivers/hwtracing/coresight/coresight-priv.h
+++ b/drivers/hwtracing/coresight/coresight-priv.h
@@ -82,12 +82,6 @@ enum etm_addr_type {
ETM_ADDR_TYPE_STOP,
};
-enum cs_mode {
- CS_MODE_DISABLED,
- CS_MODE_SYSFS,
- CS_MODE_PERF,
-};
-
/**
* struct cs_buffer - keep track of a recording session' specifics
* @cur: index of the current buffer
@@ -133,7 +127,8 @@ static inline void CS_UNLOCK(void __iomem *addr)
}
void coresight_disable_path(struct list_head *path);
-int coresight_enable_path(struct list_head *path, u32 mode, void *sink_data);
+int coresight_enable_path(struct list_head *path, enum cs_mode mode,
+ void *sink_data);
struct coresight_device *coresight_get_sink(struct list_head *path);
struct coresight_device *
coresight_get_enabled_sink(struct coresight_device *source);
diff --git a/drivers/hwtracing/coresight/coresight-stm.c b/drivers/hwtracing/coresight/coresight-stm.c
index 66a614c5492c..a1c27c901ad1 100644
--- a/drivers/hwtracing/coresight/coresight-stm.c
+++ b/drivers/hwtracing/coresight/coresight-stm.c
@@ -119,7 +119,7 @@ DEFINE_CORESIGHT_DEVLIST(stm_devs, "stm");
* @spinlock: only one at a time pls.
* @chs: the channels accociated to this STM.
* @stm: structure associated to the generic STM interface.
- * @mode: this tracer's mode, i.e sysFS, or disabled.
+ * @mode: this tracer's mode (enum cs_mode), i.e sysFS, or disabled.
* @traceid: value of the current ID for this component.
* @write_bytes: Maximus bytes this STM can write at a time.
* @stmsper: settings for register STMSPER.
@@ -192,8 +192,8 @@ static void stm_enable_hw(struct stm_drvdata *drvdata)
CS_LOCK(drvdata->base);
}
-static int stm_enable(struct coresight_device *csdev,
- struct perf_event *event, u32 mode)
+static int stm_enable(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode)
{
u32 val;
struct stm_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etf.c b/drivers/hwtracing/coresight/coresight-tmc-etf.c
index 0ab1f73c2d06..14d3c1472455 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etf.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etf.c
@@ -302,7 +302,7 @@ static int tmc_enable_etf_sink_perf(struct coresight_device *csdev, void *data)
}
static int tmc_enable_etf_sink(struct coresight_device *csdev,
- u32 mode, void *data)
+ enum cs_mode mode, void *data)
{
int ret;
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 918d461fcf4a..ed589cfff1b5 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -1666,17 +1666,16 @@ static int tmc_enable_etr_sink_perf(struct coresight_device *csdev, void *data)
}
static int tmc_enable_etr_sink(struct coresight_device *csdev,
- u32 mode, void *data)
+ enum cs_mode mode, void *data)
{
switch (mode) {
case CS_MODE_SYSFS:
return tmc_enable_etr_sink_sysfs(csdev);
case CS_MODE_PERF:
return tmc_enable_etr_sink_perf(csdev, data);
+ default:
+ return -EINVAL;
}
-
- /* We shouldn't be here */
- return -EINVAL;
}
static int tmc_disable_etr_sink(struct coresight_device *csdev)
diff --git a/drivers/hwtracing/coresight/coresight-tpdm.c b/drivers/hwtracing/coresight/coresight-tpdm.c
index 9479a5e8c672..f4854af0431e 100644
--- a/drivers/hwtracing/coresight/coresight-tpdm.c
+++ b/drivers/hwtracing/coresight/coresight-tpdm.c
@@ -42,8 +42,8 @@ static void __tpdm_enable(struct tpdm_drvdata *drvdata)
CS_LOCK(drvdata->base);
}
-static int tpdm_enable(struct coresight_device *csdev,
- struct perf_event *event, u32 mode)
+static int tpdm_enable(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode)
{
struct tpdm_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 34d37abd2c8d..b0179f761c98 100644
--- a/drivers/hwtracing/coresight/coresight-tpiu.c
+++ b/drivers/hwtracing/coresight/coresight-tpiu.c
@@ -69,7 +69,8 @@ static void tpiu_enable_hw(struct csdev_access *csa)
CS_LOCK(csa->base);
}
-static int tpiu_enable(struct coresight_device *csdev, u32 mode, void *__unused)
+static int tpiu_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *__unused)
{
tpiu_enable_hw(&csdev->access);
atomic_inc(csdev->refcnt);
diff --git a/drivers/hwtracing/coresight/coresight-trbe.c b/drivers/hwtracing/coresight/coresight-trbe.c
index 1fc4fd79a1c6..5406e338c280 100644
--- a/drivers/hwtracing/coresight/coresight-trbe.c
+++ b/drivers/hwtracing/coresight/coresight-trbe.c
@@ -1005,7 +1005,8 @@ static int __arm_trbe_enable(struct trbe_buf *buf,
return ret;
}
-static int arm_trbe_enable(struct coresight_device *csdev, u32 mode, void *data)
+static int arm_trbe_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *data)
{
struct trbe_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
struct trbe_cpudata *cpudata = dev_get_drvdata(&csdev->dev);
diff --git a/drivers/hwtracing/coresight/ultrasoc-smb.c b/drivers/hwtracing/coresight/ultrasoc-smb.c
index b317342c7ce5..a2ced0b3c3cd 100644
--- a/drivers/hwtracing/coresight/ultrasoc-smb.c
+++ b/drivers/hwtracing/coresight/ultrasoc-smb.c
@@ -256,7 +256,8 @@ static int smb_enable_perf(struct coresight_device *csdev, void *data)
return 0;
}
-static int smb_enable(struct coresight_device *csdev, u32 mode, void *data)
+static int smb_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *data)
{
struct smb_drv_data *drvdata = dev_get_drvdata(csdev->dev.parent);
int ret = 0;
diff --git a/drivers/hwtracing/coresight/ultrasoc-smb.h b/drivers/hwtracing/coresight/ultrasoc-smb.h
index 7dfbe42e37a0..d2e14e8d2c8a 100644
--- a/drivers/hwtracing/coresight/ultrasoc-smb.h
+++ b/drivers/hwtracing/coresight/ultrasoc-smb.h
@@ -119,7 +119,7 @@ struct smb_drv_data {
struct mutex mutex;
bool reading;
pid_t pid;
- u32 mode;
+ enum cs_mode mode;
};
#endif
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index f19a47b9bb5a..1f878e8ed8c4 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -272,6 +272,12 @@ static struct coresight_dev_list (var) = { \
#define to_coresight_device(d) container_of(d, struct coresight_device, dev)
+enum cs_mode {
+ CS_MODE_DISABLED,
+ CS_MODE_SYSFS,
+ CS_MODE_PERF,
+};
+
#define source_ops(csdev) csdev->ops->source_ops
#define sink_ops(csdev) csdev->ops->sink_ops
#define link_ops(csdev) csdev->ops->link_ops
@@ -288,7 +294,8 @@ static struct coresight_dev_list (var) = { \
* @update_buffer: update buffer pointers after a trace session.
*/
struct coresight_ops_sink {
- int (*enable)(struct coresight_device *csdev, u32 mode, void *data);
+ int (*enable)(struct coresight_device *csdev, enum cs_mode mode,
+ void *data);
int (*disable)(struct coresight_device *csdev);
void *(*alloc_buffer)(struct coresight_device *csdev,
struct perf_event *event, void **pages,
@@ -320,8 +327,8 @@ struct coresight_ops_link {
*/
struct coresight_ops_source {
int (*cpu_id)(struct coresight_device *csdev);
- int (*enable)(struct coresight_device *csdev,
- struct perf_event *event, u32 mode);
+ int (*enable)(struct coresight_device *csdev, struct perf_event *event,
+ enum cs_mode mode);
void (*disable)(struct coresight_device *csdev,
struct perf_event *event);
};
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 2/9] coresight: Change name of pdata->conns
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
conns is actually for output connections. Change the name to make it
clearer and so that we can add input connections later.
No functional changes.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 26 +++++++++----------
.../hwtracing/coresight/coresight-platform.c | 12 ++++-----
.../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
include/linux/coresight.h | 2 +-
4 files changed, 21 insertions(+), 21 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index e3b63fd52b9c..c0eda7407fb8 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -119,7 +119,7 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
struct coresight_connection *conn;
for (i = 0; i < parent->pdata->nr_outport; i++) {
- conn = &parent->pdata->conns[i];
+ conn = &parent->pdata->out_conns[i];
if (conn->child_dev == csdev)
return conn->child_port;
}
@@ -137,7 +137,7 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
struct coresight_connection *conn;
for (i = 0; i < csdev->pdata->nr_outport; i++) {
- conn = &csdev->pdata->conns[i];
+ conn = &csdev->pdata->out_conns[i];
if (conn->child_dev == child)
return conn->outport;
}
@@ -606,7 +606,7 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outport; i++) {
struct coresight_device *child_dev;
- child_dev = csdev->pdata->conns[i].child_dev;
+ child_dev = csdev->pdata->out_conns[i].child_dev;
if (child_dev)
sink = coresight_find_enabled_sink(child_dev);
if (sink)
@@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outport; i++) {
struct coresight_device *child;
- child = csdev->pdata->conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].child_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
if (!coresight_get_ref(child))
goto err;
@@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
for (i--; i >= 0; i--) {
struct coresight_device *child;
- child = csdev->pdata->conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].child_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
coresight_put_ref(child);
}
@@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outport; i++) {
struct coresight_device *child;
- child = csdev->pdata->conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].child_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
coresight_put_ref(child);
}
@@ -794,7 +794,7 @@ static int _coresight_build_path(struct coresight_device *csdev,
for (i = 0; i < csdev->pdata->nr_outport; i++) {
struct coresight_device *child_dev;
- child_dev = csdev->pdata->conns[i].child_dev;
+ child_dev = csdev->pdata->out_conns[i].child_dev;
if (child_dev &&
_coresight_build_path(child_dev, sink, path) == 0) {
found = true;
@@ -964,7 +964,7 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
struct coresight_device *child_dev, *sink = NULL;
int child_depth = curr_depth;
- child_dev = csdev->pdata->conns[i].child_dev;
+ child_dev = csdev->pdata->out_conns[i].child_dev;
if (child_dev)
sink = coresight_find_sink(child_dev, &child_depth);
@@ -1334,7 +1334,7 @@ static int coresight_orphan_match(struct device *dev, void *data)
* an orphan connection whose name matches @csdev, link it.
*/
for (i = 0; i < i_csdev->pdata->nr_outport; i++) {
- conn = &i_csdev->pdata->conns[i];
+ conn = &i_csdev->pdata->out_conns[i];
/* Skip the port if FW doesn't describe it */
if (!conn->child_fwnode)
@@ -1375,7 +1375,7 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
int i, ret = 0;
for (i = 0; i < csdev->pdata->nr_outport; i++) {
- struct coresight_connection *conn = &csdev->pdata->conns[i];
+ struct coresight_connection *conn = &csdev->pdata->out_conns[i];
if (!conn->child_fwnode)
continue;
@@ -1412,7 +1412,7 @@ static int coresight_remove_match(struct device *dev, void *data)
* a connection whose name matches @csdev, remove it.
*/
for (i = 0; i < iterator->pdata->nr_outport; i++) {
- conn = &iterator->pdata->conns[i];
+ conn = &iterator->pdata->out_conns[i];
if (conn->child_dev == NULL || conn->child_fwnode == NULL)
continue;
@@ -1548,7 +1548,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
struct coresight_platform_data *pdata)
{
int i;
- struct coresight_connection *conns = pdata->conns;
+ struct coresight_connection *conns = pdata->out_conns;
for (i = 0; i < pdata->nr_outport; i++) {
/* If we have made the links, remove them now */
@@ -1560,7 +1560,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
*/
if (conns[i].child_fwnode) {
fwnode_handle_put(conns[i].child_fwnode);
- pdata->conns[i].child_fwnode = NULL;
+ pdata->out_conns[i].child_fwnode = NULL;
}
}
if (csdev)
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index 475899714104..5085525a32bb 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -27,9 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
struct coresight_platform_data *pdata)
{
if (pdata->nr_outport) {
- pdata->conns = devm_kcalloc(dev, pdata->nr_outport,
- sizeof(*pdata->conns), GFP_KERNEL);
- if (!pdata->conns)
+ pdata->out_conns = devm_kcalloc(dev, pdata->nr_outport,
+ sizeof(*pdata->out_conns), GFP_KERNEL);
+ if (!pdata->out_conns)
return -ENOMEM;
}
@@ -251,7 +251,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
break;
}
- conn = &pdata->conns[endpoint.port];
+ conn = &pdata->out_conns[endpoint.port];
if (conn->child_fwnode) {
dev_warn(dev, "Duplicate output port %d\n",
endpoint.port);
@@ -744,8 +744,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
int port = conns[i].outport;
/* Duplicate output port */
- WARN_ON(pdata->conns[port].child_fwnode);
- pdata->conns[port] = conns[i];
+ WARN_ON(pdata->out_conns[port].child_fwnode);
+ pdata->out_conns[port] = conns[i];
}
devm_kfree(&adev->dev, conns);
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index ed589cfff1b5..86d4a08aa833 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
return NULL;
for (i = 0; i < etr->pdata->nr_outport; i++) {
- tmp = etr->pdata->conns[i].child_dev;
+ tmp = etr->pdata->out_conns[i].child_dev;
if (tmp && coresight_is_catu_device(tmp))
return tmp;
}
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 1f878e8ed8c4..322d7273e122 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -109,7 +109,7 @@ union coresight_dev_subtype {
struct coresight_platform_data {
int nr_inport;
int nr_outport;
- struct coresight_connection *conns;
+ struct coresight_connection *out_conns;
};
/**
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 2/9] coresight: Change name of pdata->conns
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
conns is actually for output connections. Change the name to make it
clearer and so that we can add input connections later.
No functional changes.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 26 +++++++++----------
.../hwtracing/coresight/coresight-platform.c | 12 ++++-----
.../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
include/linux/coresight.h | 2 +-
4 files changed, 21 insertions(+), 21 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index e3b63fd52b9c..c0eda7407fb8 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -119,7 +119,7 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
struct coresight_connection *conn;
for (i = 0; i < parent->pdata->nr_outport; i++) {
- conn = &parent->pdata->conns[i];
+ conn = &parent->pdata->out_conns[i];
if (conn->child_dev == csdev)
return conn->child_port;
}
@@ -137,7 +137,7 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
struct coresight_connection *conn;
for (i = 0; i < csdev->pdata->nr_outport; i++) {
- conn = &csdev->pdata->conns[i];
+ conn = &csdev->pdata->out_conns[i];
if (conn->child_dev == child)
return conn->outport;
}
@@ -606,7 +606,7 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outport; i++) {
struct coresight_device *child_dev;
- child_dev = csdev->pdata->conns[i].child_dev;
+ child_dev = csdev->pdata->out_conns[i].child_dev;
if (child_dev)
sink = coresight_find_enabled_sink(child_dev);
if (sink)
@@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outport; i++) {
struct coresight_device *child;
- child = csdev->pdata->conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].child_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
if (!coresight_get_ref(child))
goto err;
@@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
for (i--; i >= 0; i--) {
struct coresight_device *child;
- child = csdev->pdata->conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].child_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
coresight_put_ref(child);
}
@@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outport; i++) {
struct coresight_device *child;
- child = csdev->pdata->conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].child_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
coresight_put_ref(child);
}
@@ -794,7 +794,7 @@ static int _coresight_build_path(struct coresight_device *csdev,
for (i = 0; i < csdev->pdata->nr_outport; i++) {
struct coresight_device *child_dev;
- child_dev = csdev->pdata->conns[i].child_dev;
+ child_dev = csdev->pdata->out_conns[i].child_dev;
if (child_dev &&
_coresight_build_path(child_dev, sink, path) == 0) {
found = true;
@@ -964,7 +964,7 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
struct coresight_device *child_dev, *sink = NULL;
int child_depth = curr_depth;
- child_dev = csdev->pdata->conns[i].child_dev;
+ child_dev = csdev->pdata->out_conns[i].child_dev;
if (child_dev)
sink = coresight_find_sink(child_dev, &child_depth);
@@ -1334,7 +1334,7 @@ static int coresight_orphan_match(struct device *dev, void *data)
* an orphan connection whose name matches @csdev, link it.
*/
for (i = 0; i < i_csdev->pdata->nr_outport; i++) {
- conn = &i_csdev->pdata->conns[i];
+ conn = &i_csdev->pdata->out_conns[i];
/* Skip the port if FW doesn't describe it */
if (!conn->child_fwnode)
@@ -1375,7 +1375,7 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
int i, ret = 0;
for (i = 0; i < csdev->pdata->nr_outport; i++) {
- struct coresight_connection *conn = &csdev->pdata->conns[i];
+ struct coresight_connection *conn = &csdev->pdata->out_conns[i];
if (!conn->child_fwnode)
continue;
@@ -1412,7 +1412,7 @@ static int coresight_remove_match(struct device *dev, void *data)
* a connection whose name matches @csdev, remove it.
*/
for (i = 0; i < iterator->pdata->nr_outport; i++) {
- conn = &iterator->pdata->conns[i];
+ conn = &iterator->pdata->out_conns[i];
if (conn->child_dev == NULL || conn->child_fwnode == NULL)
continue;
@@ -1548,7 +1548,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
struct coresight_platform_data *pdata)
{
int i;
- struct coresight_connection *conns = pdata->conns;
+ struct coresight_connection *conns = pdata->out_conns;
for (i = 0; i < pdata->nr_outport; i++) {
/* If we have made the links, remove them now */
@@ -1560,7 +1560,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
*/
if (conns[i].child_fwnode) {
fwnode_handle_put(conns[i].child_fwnode);
- pdata->conns[i].child_fwnode = NULL;
+ pdata->out_conns[i].child_fwnode = NULL;
}
}
if (csdev)
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index 475899714104..5085525a32bb 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -27,9 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
struct coresight_platform_data *pdata)
{
if (pdata->nr_outport) {
- pdata->conns = devm_kcalloc(dev, pdata->nr_outport,
- sizeof(*pdata->conns), GFP_KERNEL);
- if (!pdata->conns)
+ pdata->out_conns = devm_kcalloc(dev, pdata->nr_outport,
+ sizeof(*pdata->out_conns), GFP_KERNEL);
+ if (!pdata->out_conns)
return -ENOMEM;
}
@@ -251,7 +251,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
break;
}
- conn = &pdata->conns[endpoint.port];
+ conn = &pdata->out_conns[endpoint.port];
if (conn->child_fwnode) {
dev_warn(dev, "Duplicate output port %d\n",
endpoint.port);
@@ -744,8 +744,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
int port = conns[i].outport;
/* Duplicate output port */
- WARN_ON(pdata->conns[port].child_fwnode);
- pdata->conns[port] = conns[i];
+ WARN_ON(pdata->out_conns[port].child_fwnode);
+ pdata->out_conns[port] = conns[i];
}
devm_kfree(&adev->dev, conns);
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index ed589cfff1b5..86d4a08aa833 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
return NULL;
for (i = 0; i < etr->pdata->nr_outport; i++) {
- tmp = etr->pdata->conns[i].child_dev;
+ tmp = etr->pdata->out_conns[i].child_dev;
if (tmp && coresight_is_catu_device(tmp))
return tmp;
}
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 1f878e8ed8c4..322d7273e122 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -109,7 +109,7 @@ union coresight_dev_subtype {
struct coresight_platform_data {
int nr_inport;
int nr_outport;
- struct coresight_connection *conns;
+ struct coresight_connection *out_conns;
};
/**
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 3/9] coresight: Rename nr_outports to nr_outconns
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
Rename to avoid confusion between port number and the index in the
connection array. The port number is already stored in the connection,
and in a later commit the connection array will be appended to, so
the length of it will no longer reflect the number of ports.
No functional changes.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 32 ++++++++---------
.../hwtracing/coresight/coresight-platform.c | 34 +++++++++----------
.../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
include/linux/coresight.h | 10 +++---
4 files changed, 39 insertions(+), 39 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index c0eda7407fb8..617be08883da 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -118,7 +118,7 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
int i;
struct coresight_connection *conn;
- for (i = 0; i < parent->pdata->nr_outport; i++) {
+ for (i = 0; i < parent->pdata->nr_outconns; i++) {
conn = &parent->pdata->out_conns[i];
if (conn->child_dev == csdev)
return conn->child_port;
@@ -136,7 +136,7 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
int i;
struct coresight_connection *conn;
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
conn = &csdev->pdata->out_conns[i];
if (conn->child_dev == child)
return conn->outport;
@@ -397,9 +397,9 @@ static void coresight_disable_link(struct coresight_device *csdev,
link_subtype = csdev->subtype.link_subtype;
if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_MERG) {
- nr_conns = csdev->pdata->nr_inport;
+ nr_conns = csdev->pdata->nr_inconns;
} else if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_SPLIT) {
- nr_conns = csdev->pdata->nr_outport;
+ nr_conns = csdev->pdata->nr_outconns;
} else {
nr_conns = 1;
}
@@ -603,7 +603,7 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
/*
* Recursively explore each port found on this element.
*/
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child_dev;
child_dev = csdev->pdata->out_conns[i].child_dev;
@@ -719,7 +719,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
{
int i;
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child;
child = csdev->pdata->out_conns[i].child_dev;
@@ -749,7 +749,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
int i;
coresight_put_ref(csdev);
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child;
child = csdev->pdata->out_conns[i].child_dev;
@@ -791,7 +791,7 @@ static int _coresight_build_path(struct coresight_device *csdev,
}
/* Not a sink - recursively explore each port found on this element */
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child_dev;
child_dev = csdev->pdata->out_conns[i].child_dev;
@@ -960,7 +960,7 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
* Not a sink we want - or possible child sink may be better.
* recursively explore each port found on this element.
*/
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child_dev, *sink = NULL;
int child_depth = curr_depth;
@@ -1333,7 +1333,7 @@ static int coresight_orphan_match(struct device *dev, void *data)
* Circle throuch all the connection of that component. If we find
* an orphan connection whose name matches @csdev, link it.
*/
- for (i = 0; i < i_csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < i_csdev->pdata->nr_outconns; i++) {
conn = &i_csdev->pdata->out_conns[i];
/* Skip the port if FW doesn't describe it */
@@ -1374,7 +1374,7 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
{
int i, ret = 0;
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_connection *conn = &csdev->pdata->out_conns[i];
if (!conn->child_fwnode)
@@ -1411,7 +1411,7 @@ static int coresight_remove_match(struct device *dev, void *data)
* Circle throuch all the connection of that component. If we find
* a connection whose name matches @csdev, remove it.
*/
- for (i = 0; i < iterator->pdata->nr_outport; i++) {
+ for (i = 0; i < iterator->pdata->nr_outconns; i++) {
conn = &iterator->pdata->out_conns[i];
if (conn->child_dev == NULL || conn->child_fwnode == NULL)
@@ -1451,7 +1451,7 @@ static void coresight_remove_conns(struct coresight_device *csdev)
* doesn't have at least one input port, there is no point
* in searching all the devices.
*/
- if (csdev->pdata->nr_inport)
+ if (csdev->pdata->nr_inconns)
bus_for_each_dev(&coresight_bustype, NULL,
csdev, coresight_remove_match);
}
@@ -1550,7 +1550,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
int i;
struct coresight_connection *conns = pdata->out_conns;
- for (i = 0; i < pdata->nr_outport; i++) {
+ for (i = 0; i < pdata->nr_outconns; i++) {
/* If we have made the links, remove them now */
if (csdev && conns[i].child_dev)
coresight_remove_links(csdev, &conns[i]);
@@ -1587,9 +1587,9 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
link_subtype = desc->subtype.link_subtype;
if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_MERG)
- nr_refcnts = desc->pdata->nr_inport;
+ nr_refcnts = desc->pdata->nr_inconns;
else if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_SPLIT)
- nr_refcnts = desc->pdata->nr_outport;
+ nr_refcnts = desc->pdata->nr_outconns;
}
refcnts = kcalloc(nr_refcnts, sizeof(*refcnts), GFP_KERNEL);
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index 5085525a32bb..ed865e0621a9 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -26,8 +26,8 @@
static int coresight_alloc_conns(struct device *dev,
struct coresight_platform_data *pdata)
{
- if (pdata->nr_outport) {
- pdata->out_conns = devm_kcalloc(dev, pdata->nr_outport,
+ if (pdata->nr_outconns) {
+ pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
sizeof(*pdata->out_conns), GFP_KERNEL);
if (!pdata->out_conns)
return -ENOMEM;
@@ -84,7 +84,7 @@ static inline bool of_coresight_legacy_ep_is_input(struct device_node *ep)
}
static void of_coresight_get_ports_legacy(const struct device_node *node,
- int *nr_inport, int *nr_outport)
+ int *nr_inconns, int *nr_outconns)
{
struct device_node *ep = NULL;
struct of_endpoint endpoint;
@@ -114,8 +114,8 @@ static void of_coresight_get_ports_legacy(const struct device_node *node,
} while (ep);
- *nr_inport = in;
- *nr_outport = out;
+ *nr_inconns = in;
+ *nr_outconns = out;
}
static struct device_node *of_coresight_get_port_parent(struct device_node *ep)
@@ -164,7 +164,7 @@ of_coresight_count_ports(struct device_node *port_parent)
}
static void of_coresight_get_ports(const struct device_node *node,
- int *nr_inport, int *nr_outport)
+ int *nr_inconns, int *nr_outconns)
{
struct device_node *input_ports = NULL, *output_ports = NULL;
@@ -173,16 +173,16 @@ static void of_coresight_get_ports(const struct device_node *node,
if (input_ports || output_ports) {
if (input_ports) {
- *nr_inport = of_coresight_count_ports(input_ports);
+ *nr_inconns = of_coresight_count_ports(input_ports);
of_node_put(input_ports);
}
if (output_ports) {
- *nr_outport = of_coresight_count_ports(output_ports);
+ *nr_outconns = of_coresight_count_ports(output_ports);
of_node_put(output_ports);
}
} else {
/* Fall back to legacy DT bindings parsing */
- of_coresight_get_ports_legacy(node, nr_inport, nr_outport);
+ of_coresight_get_ports_legacy(node, nr_inconns, nr_outconns);
}
}
@@ -289,10 +289,10 @@ static int of_get_coresight_platform_data(struct device *dev,
struct device_node *node = dev->of_node;
/* Get the number of input and output port for this component */
- of_coresight_get_ports(node, &pdata->nr_inport, &pdata->nr_outport);
+ of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
/* If there are no output connections, we are done */
- if (!pdata->nr_outport)
+ if (!pdata->nr_outconns)
return 0;
ret = coresight_alloc_conns(dev, pdata);
@@ -690,7 +690,7 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
const union acpi_object *graph;
struct coresight_connection *conns, *ptr;
- pdata->nr_inport = pdata->nr_outport = 0;
+ pdata->nr_inconns = pdata->nr_outconns = 0;
graph = acpi_get_coresight_graph(adev);
if (!graph)
return -ENOENT;
@@ -718,11 +718,11 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
return dir;
if (dir == ACPI_CORESIGHT_LINK_MASTER) {
- if (ptr->outport >= pdata->nr_outport)
- pdata->nr_outport = ptr->outport + 1;
+ if (ptr->outport >= pdata->nr_outconns)
+ pdata->nr_outconns = ptr->outport + 1;
ptr++;
} else {
- WARN_ON(pdata->nr_inport == ptr->child_port + 1);
+ WARN_ON(pdata->nr_inconns == ptr->child_port + 1);
/*
* We do not track input port connections for a device.
* However we need the highest port number described,
@@ -730,8 +730,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
* record for an output connection. Hence, do not move
* the ptr for input connections
*/
- if (ptr->child_port >= pdata->nr_inport)
- pdata->nr_inport = ptr->child_port + 1;
+ if (ptr->child_port >= pdata->nr_inconns)
+ pdata->nr_inconns = ptr->child_port + 1;
}
}
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 86d4a08aa833..4711dfa7418c 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -781,7 +781,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
if (!IS_ENABLED(CONFIG_CORESIGHT_CATU))
return NULL;
- for (i = 0; i < etr->pdata->nr_outport; i++) {
+ for (i = 0; i < etr->pdata->nr_outconns; i++) {
tmp = etr->pdata->out_conns[i].child_dev;
if (tmp && coresight_is_catu_device(tmp))
return tmp;
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 322d7273e122..cdf0d1def778 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -102,13 +102,13 @@ union coresight_dev_subtype {
* struct coresight_platform_data - data harvested from the firmware
* specification.
*
- * @nr_inport: Number of elements for the input connections.
- * @nr_outport: Number of elements for the output connections.
- * @conns: Sparse array of nr_outport connections from this component.
+ * @nr_inconns: Number of elements for the input connections.
+ * @nr_outconns: Number of elements for the output connections.
+ * @conns: Sparse array of nr_outconns connections from this component.
*/
struct coresight_platform_data {
- int nr_inport;
- int nr_outport;
+ int nr_inconns;
+ int nr_outconns;
struct coresight_connection *out_conns;
};
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 3/9] coresight: Rename nr_outports to nr_outconns
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
Rename to avoid confusion between port number and the index in the
connection array. The port number is already stored in the connection,
and in a later commit the connection array will be appended to, so
the length of it will no longer reflect the number of ports.
No functional changes.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 32 ++++++++---------
.../hwtracing/coresight/coresight-platform.c | 34 +++++++++----------
.../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
include/linux/coresight.h | 10 +++---
4 files changed, 39 insertions(+), 39 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index c0eda7407fb8..617be08883da 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -118,7 +118,7 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
int i;
struct coresight_connection *conn;
- for (i = 0; i < parent->pdata->nr_outport; i++) {
+ for (i = 0; i < parent->pdata->nr_outconns; i++) {
conn = &parent->pdata->out_conns[i];
if (conn->child_dev == csdev)
return conn->child_port;
@@ -136,7 +136,7 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
int i;
struct coresight_connection *conn;
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
conn = &csdev->pdata->out_conns[i];
if (conn->child_dev == child)
return conn->outport;
@@ -397,9 +397,9 @@ static void coresight_disable_link(struct coresight_device *csdev,
link_subtype = csdev->subtype.link_subtype;
if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_MERG) {
- nr_conns = csdev->pdata->nr_inport;
+ nr_conns = csdev->pdata->nr_inconns;
} else if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_SPLIT) {
- nr_conns = csdev->pdata->nr_outport;
+ nr_conns = csdev->pdata->nr_outconns;
} else {
nr_conns = 1;
}
@@ -603,7 +603,7 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
/*
* Recursively explore each port found on this element.
*/
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child_dev;
child_dev = csdev->pdata->out_conns[i].child_dev;
@@ -719,7 +719,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
{
int i;
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child;
child = csdev->pdata->out_conns[i].child_dev;
@@ -749,7 +749,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
int i;
coresight_put_ref(csdev);
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child;
child = csdev->pdata->out_conns[i].child_dev;
@@ -791,7 +791,7 @@ static int _coresight_build_path(struct coresight_device *csdev,
}
/* Not a sink - recursively explore each port found on this element */
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child_dev;
child_dev = csdev->pdata->out_conns[i].child_dev;
@@ -960,7 +960,7 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
* Not a sink we want - or possible child sink may be better.
* recursively explore each port found on this element.
*/
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child_dev, *sink = NULL;
int child_depth = curr_depth;
@@ -1333,7 +1333,7 @@ static int coresight_orphan_match(struct device *dev, void *data)
* Circle throuch all the connection of that component. If we find
* an orphan connection whose name matches @csdev, link it.
*/
- for (i = 0; i < i_csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < i_csdev->pdata->nr_outconns; i++) {
conn = &i_csdev->pdata->out_conns[i];
/* Skip the port if FW doesn't describe it */
@@ -1374,7 +1374,7 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
{
int i, ret = 0;
- for (i = 0; i < csdev->pdata->nr_outport; i++) {
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_connection *conn = &csdev->pdata->out_conns[i];
if (!conn->child_fwnode)
@@ -1411,7 +1411,7 @@ static int coresight_remove_match(struct device *dev, void *data)
* Circle throuch all the connection of that component. If we find
* a connection whose name matches @csdev, remove it.
*/
- for (i = 0; i < iterator->pdata->nr_outport; i++) {
+ for (i = 0; i < iterator->pdata->nr_outconns; i++) {
conn = &iterator->pdata->out_conns[i];
if (conn->child_dev == NULL || conn->child_fwnode == NULL)
@@ -1451,7 +1451,7 @@ static void coresight_remove_conns(struct coresight_device *csdev)
* doesn't have at least one input port, there is no point
* in searching all the devices.
*/
- if (csdev->pdata->nr_inport)
+ if (csdev->pdata->nr_inconns)
bus_for_each_dev(&coresight_bustype, NULL,
csdev, coresight_remove_match);
}
@@ -1550,7 +1550,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
int i;
struct coresight_connection *conns = pdata->out_conns;
- for (i = 0; i < pdata->nr_outport; i++) {
+ for (i = 0; i < pdata->nr_outconns; i++) {
/* If we have made the links, remove them now */
if (csdev && conns[i].child_dev)
coresight_remove_links(csdev, &conns[i]);
@@ -1587,9 +1587,9 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
link_subtype = desc->subtype.link_subtype;
if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_MERG)
- nr_refcnts = desc->pdata->nr_inport;
+ nr_refcnts = desc->pdata->nr_inconns;
else if (link_subtype == CORESIGHT_DEV_SUBTYPE_LINK_SPLIT)
- nr_refcnts = desc->pdata->nr_outport;
+ nr_refcnts = desc->pdata->nr_outconns;
}
refcnts = kcalloc(nr_refcnts, sizeof(*refcnts), GFP_KERNEL);
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index 5085525a32bb..ed865e0621a9 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -26,8 +26,8 @@
static int coresight_alloc_conns(struct device *dev,
struct coresight_platform_data *pdata)
{
- if (pdata->nr_outport) {
- pdata->out_conns = devm_kcalloc(dev, pdata->nr_outport,
+ if (pdata->nr_outconns) {
+ pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
sizeof(*pdata->out_conns), GFP_KERNEL);
if (!pdata->out_conns)
return -ENOMEM;
@@ -84,7 +84,7 @@ static inline bool of_coresight_legacy_ep_is_input(struct device_node *ep)
}
static void of_coresight_get_ports_legacy(const struct device_node *node,
- int *nr_inport, int *nr_outport)
+ int *nr_inconns, int *nr_outconns)
{
struct device_node *ep = NULL;
struct of_endpoint endpoint;
@@ -114,8 +114,8 @@ static void of_coresight_get_ports_legacy(const struct device_node *node,
} while (ep);
- *nr_inport = in;
- *nr_outport = out;
+ *nr_inconns = in;
+ *nr_outconns = out;
}
static struct device_node *of_coresight_get_port_parent(struct device_node *ep)
@@ -164,7 +164,7 @@ of_coresight_count_ports(struct device_node *port_parent)
}
static void of_coresight_get_ports(const struct device_node *node,
- int *nr_inport, int *nr_outport)
+ int *nr_inconns, int *nr_outconns)
{
struct device_node *input_ports = NULL, *output_ports = NULL;
@@ -173,16 +173,16 @@ static void of_coresight_get_ports(const struct device_node *node,
if (input_ports || output_ports) {
if (input_ports) {
- *nr_inport = of_coresight_count_ports(input_ports);
+ *nr_inconns = of_coresight_count_ports(input_ports);
of_node_put(input_ports);
}
if (output_ports) {
- *nr_outport = of_coresight_count_ports(output_ports);
+ *nr_outconns = of_coresight_count_ports(output_ports);
of_node_put(output_ports);
}
} else {
/* Fall back to legacy DT bindings parsing */
- of_coresight_get_ports_legacy(node, nr_inport, nr_outport);
+ of_coresight_get_ports_legacy(node, nr_inconns, nr_outconns);
}
}
@@ -289,10 +289,10 @@ static int of_get_coresight_platform_data(struct device *dev,
struct device_node *node = dev->of_node;
/* Get the number of input and output port for this component */
- of_coresight_get_ports(node, &pdata->nr_inport, &pdata->nr_outport);
+ of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
/* If there are no output connections, we are done */
- if (!pdata->nr_outport)
+ if (!pdata->nr_outconns)
return 0;
ret = coresight_alloc_conns(dev, pdata);
@@ -690,7 +690,7 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
const union acpi_object *graph;
struct coresight_connection *conns, *ptr;
- pdata->nr_inport = pdata->nr_outport = 0;
+ pdata->nr_inconns = pdata->nr_outconns = 0;
graph = acpi_get_coresight_graph(adev);
if (!graph)
return -ENOENT;
@@ -718,11 +718,11 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
return dir;
if (dir == ACPI_CORESIGHT_LINK_MASTER) {
- if (ptr->outport >= pdata->nr_outport)
- pdata->nr_outport = ptr->outport + 1;
+ if (ptr->outport >= pdata->nr_outconns)
+ pdata->nr_outconns = ptr->outport + 1;
ptr++;
} else {
- WARN_ON(pdata->nr_inport == ptr->child_port + 1);
+ WARN_ON(pdata->nr_inconns == ptr->child_port + 1);
/*
* We do not track input port connections for a device.
* However we need the highest port number described,
@@ -730,8 +730,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
* record for an output connection. Hence, do not move
* the ptr for input connections
*/
- if (ptr->child_port >= pdata->nr_inport)
- pdata->nr_inport = ptr->child_port + 1;
+ if (ptr->child_port >= pdata->nr_inconns)
+ pdata->nr_inconns = ptr->child_port + 1;
}
}
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 86d4a08aa833..4711dfa7418c 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -781,7 +781,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
if (!IS_ENABLED(CONFIG_CORESIGHT_CATU))
return NULL;
- for (i = 0; i < etr->pdata->nr_outport; i++) {
+ for (i = 0; i < etr->pdata->nr_outconns; i++) {
tmp = etr->pdata->out_conns[i].child_dev;
if (tmp && coresight_is_catu_device(tmp))
return tmp;
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 322d7273e122..cdf0d1def778 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -102,13 +102,13 @@ union coresight_dev_subtype {
* struct coresight_platform_data - data harvested from the firmware
* specification.
*
- * @nr_inport: Number of elements for the input connections.
- * @nr_outport: Number of elements for the output connections.
- * @conns: Sparse array of nr_outport connections from this component.
+ * @nr_inconns: Number of elements for the input connections.
+ * @nr_outconns: Number of elements for the output connections.
+ * @conns: Sparse array of nr_outconns connections from this component.
*/
struct coresight_platform_data {
- int nr_inport;
- int nr_outport;
+ int nr_inconns;
+ int nr_outconns;
struct coresight_connection *out_conns;
};
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 4/9] coresight: Rename connection members to allow for input connections
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
References to outport and child_port will be confusing when we use the
same struct for inputs as well as outputs. Rename the members to refer
to the position in terms of local and remote instead.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 70 +++++++++----------
.../hwtracing/coresight/coresight-platform.c | 32 ++++-----
drivers/hwtracing/coresight/coresight-sysfs.c | 10 +--
.../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
include/linux/coresight.h | 23 +++---
5 files changed, 71 insertions(+), 66 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index 617be08883da..f457914e445e 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -120,8 +120,8 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
for (i = 0; i < parent->pdata->nr_outconns; i++) {
conn = &parent->pdata->out_conns[i];
- if (conn->child_dev == csdev)
- return conn->child_port;
+ if (conn->remote_dev == csdev)
+ return conn->remote_port;
}
dev_err(&csdev->dev, "couldn't find inport, parent: %s, child: %s\n",
@@ -138,8 +138,8 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
conn = &csdev->pdata->out_conns[i];
- if (conn->child_dev == child)
- return conn->outport;
+ if (conn->remote_dev == child)
+ return conn->port;
}
dev_err(&csdev->dev, "couldn't find outport, parent: %s, child: %s\n",
@@ -604,11 +604,11 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
* Recursively explore each port found on this element.
*/
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
- struct coresight_device *child_dev;
+ struct coresight_device *remote_dev;
- child_dev = csdev->pdata->out_conns[i].child_dev;
- if (child_dev)
- sink = coresight_find_enabled_sink(child_dev);
+ remote_dev = csdev->pdata->out_conns[i].remote_dev;
+ if (remote_dev)
+ sink = coresight_find_enabled_sink(remote_dev);
if (sink)
return sink;
}
@@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child;
- child = csdev->pdata->out_conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].remote_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
if (!coresight_get_ref(child))
goto err;
@@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
for (i--; i >= 0; i--) {
struct coresight_device *child;
- child = csdev->pdata->out_conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].remote_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
coresight_put_ref(child);
}
@@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child;
- child = csdev->pdata->out_conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].remote_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
coresight_put_ref(child);
}
@@ -792,11 +792,11 @@ static int _coresight_build_path(struct coresight_device *csdev,
/* Not a sink - recursively explore each port found on this element */
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
- struct coresight_device *child_dev;
+ struct coresight_device *remote_dev;
- child_dev = csdev->pdata->out_conns[i].child_dev;
- if (child_dev &&
- _coresight_build_path(child_dev, sink, path) == 0) {
+ remote_dev = csdev->pdata->out_conns[i].remote_dev;
+ if (remote_dev &&
+ _coresight_build_path(remote_dev, sink, path) == 0) {
found = true;
break;
}
@@ -961,12 +961,12 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
* recursively explore each port found on this element.
*/
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
- struct coresight_device *child_dev, *sink = NULL;
+ struct coresight_device *remote_dev, *sink = NULL;
int child_depth = curr_depth;
- child_dev = csdev->pdata->out_conns[i].child_dev;
- if (child_dev)
- sink = coresight_find_sink(child_dev, &child_depth);
+ remote_dev = csdev->pdata->out_conns[i].remote_dev;
+ if (remote_dev)
+ sink = coresight_find_sink(remote_dev, &child_depth);
if (sink)
found_sink = coresight_select_best_sink(found_sink,
@@ -1337,12 +1337,12 @@ static int coresight_orphan_match(struct device *dev, void *data)
conn = &i_csdev->pdata->out_conns[i];
/* Skip the port if FW doesn't describe it */
- if (!conn->child_fwnode)
+ if (!conn->remote_fwnode)
continue;
/* We have found at least one orphan connection */
- if (conn->child_dev == NULL) {
+ if (conn->remote_dev == NULL) {
/* Does it match this newly added device? */
- if (conn->child_fwnode == csdev->dev.fwnode) {
+ if (conn->remote_fwnode == csdev->dev.fwnode) {
ret = coresight_make_links(i_csdev,
conn, csdev);
if (ret)
@@ -1377,13 +1377,13 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_connection *conn = &csdev->pdata->out_conns[i];
- if (!conn->child_fwnode)
+ if (!conn->remote_fwnode)
continue;
- conn->child_dev =
- coresight_find_csdev_by_fwnode(conn->child_fwnode);
- if (conn->child_dev && conn->child_dev->has_conns_grp) {
+ conn->remote_dev =
+ coresight_find_csdev_by_fwnode(conn->remote_fwnode);
+ if (conn->remote_dev && conn->remote_dev->has_conns_grp) {
ret = coresight_make_links(csdev, conn,
- conn->child_dev);
+ conn->remote_dev);
if (ret)
break;
} else {
@@ -1414,10 +1414,10 @@ static int coresight_remove_match(struct device *dev, void *data)
for (i = 0; i < iterator->pdata->nr_outconns; i++) {
conn = &iterator->pdata->out_conns[i];
- if (conn->child_dev == NULL || conn->child_fwnode == NULL)
+ if (conn->remote_dev == NULL || conn->remote_fwnode == NULL)
continue;
- if (csdev->dev.fwnode == conn->child_fwnode) {
+ if (csdev->dev.fwnode == conn->remote_fwnode) {
iterator->orphan = true;
coresight_remove_links(iterator, conn);
/*
@@ -1425,8 +1425,8 @@ static int coresight_remove_match(struct device *dev, void *data)
* device acquired in parsing the connections from
* platform data.
*/
- fwnode_handle_put(conn->child_fwnode);
- conn->child_fwnode = NULL;
+ fwnode_handle_put(conn->remote_fwnode);
+ conn->remote_fwnode = NULL;
/* No need to continue */
break;
}
@@ -1552,15 +1552,15 @@ void coresight_release_platform_data(struct coresight_device *csdev,
for (i = 0; i < pdata->nr_outconns; i++) {
/* If we have made the links, remove them now */
- if (csdev && conns[i].child_dev)
+ if (csdev && conns[i].remote_dev)
coresight_remove_links(csdev, &conns[i]);
/*
* Drop the refcount and clear the handle as this device
* is going away
*/
- if (conns[i].child_fwnode) {
- fwnode_handle_put(conns[i].child_fwnode);
- pdata->out_conns[i].child_fwnode = NULL;
+ if (conns[i].remote_fwnode) {
+ fwnode_handle_put(conns[i].remote_fwnode);
+ pdata->out_conns[i].remote_fwnode = NULL;
}
}
if (csdev)
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index ed865e0621a9..c77238cdf448 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -252,13 +252,13 @@ static int of_coresight_parse_endpoint(struct device *dev,
}
conn = &pdata->out_conns[endpoint.port];
- if (conn->child_fwnode) {
+ if (conn->remote_fwnode) {
dev_warn(dev, "Duplicate output port %d\n",
endpoint.port);
ret = -EINVAL;
break;
}
- conn->outport = endpoint.port;
+ conn->port = endpoint.port;
/*
* Hold the refcount to the target device. This could be
* released via:
@@ -267,8 +267,8 @@ static int of_coresight_parse_endpoint(struct device *dev,
* 2) While removing the target device via
* coresight_remove_match()
*/
- conn->child_fwnode = fwnode_handle_get(rdev_fwnode);
- conn->child_port = rendpoint.port;
+ conn->remote_fwnode = fwnode_handle_get(rdev_fwnode);
+ conn->remote_port = rendpoint.port;
/* Connection record updated */
} while (0);
@@ -649,8 +649,8 @@ static int acpi_coresight_parse_link(struct acpi_device *adev,
dir = fields[3].integer.value;
if (dir == ACPI_CORESIGHT_LINK_MASTER) {
- conn->outport = fields[0].integer.value;
- conn->child_port = fields[1].integer.value;
+ conn->port = fields[0].integer.value;
+ conn->remote_port = fields[1].integer.value;
rdev = coresight_find_device_by_fwnode(&r_adev->fwnode);
if (!rdev)
return -EPROBE_DEFER;
@@ -662,14 +662,14 @@ static int acpi_coresight_parse_link(struct acpi_device *adev,
* 2) While removing the target device via
* coresight_remove_match().
*/
- conn->child_fwnode = fwnode_handle_get(&r_adev->fwnode);
+ conn->remote_fwnode = fwnode_handle_get(&r_adev->fwnode);
} else if (dir == ACPI_CORESIGHT_LINK_SLAVE) {
/*
* We are only interested in the port number
* for the input ports at this component.
- * Store the port number in child_port.
+ * Store the port number in remote_port.
*/
- conn->child_port = fields[0].integer.value;
+ conn->remote_port = fields[0].integer.value;
} else {
/* Invalid direction */
return -EINVAL;
@@ -718,11 +718,11 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
return dir;
if (dir == ACPI_CORESIGHT_LINK_MASTER) {
- if (ptr->outport >= pdata->nr_outconns)
- pdata->nr_outconns = ptr->outport + 1;
+ if (ptr->port >= pdata->nr_outconns)
+ pdata->nr_outconns = ptr->port + 1;
ptr++;
} else {
- WARN_ON(pdata->nr_inconns == ptr->child_port + 1);
+ WARN_ON(pdata->nr_inconns == ptr->remote_port + 1);
/*
* We do not track input port connections for a device.
* However we need the highest port number described,
@@ -730,8 +730,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
* record for an output connection. Hence, do not move
* the ptr for input connections
*/
- if (ptr->child_port >= pdata->nr_inconns)
- pdata->nr_inconns = ptr->child_port + 1;
+ if (ptr->remote_port >= pdata->nr_inconns)
+ pdata->nr_inconns = ptr->remote_port + 1;
}
}
@@ -741,10 +741,10 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
/* Copy the connection information to the final location */
for (i = 0; conns + i < ptr; i++) {
- int port = conns[i].outport;
+ int port = conns[i].port;
/* Duplicate output port */
- WARN_ON(pdata->out_conns[port].child_fwnode);
+ WARN_ON(pdata->out_conns[port].remote_fwnode);
pdata->out_conns[port] = conns[i];
}
diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
index 34d2a2d31d00..3da9868d9237 100644
--- a/drivers/hwtracing/coresight/coresight-sysfs.c
+++ b/drivers/hwtracing/coresight/coresight-sysfs.c
@@ -150,11 +150,11 @@ int coresight_make_links(struct coresight_device *orig,
do {
outs = devm_kasprintf(&orig->dev, GFP_KERNEL,
- "out:%d", conn->outport);
+ "out:%d", conn->port);
if (!outs)
break;
ins = devm_kasprintf(&target->dev, GFP_KERNEL,
- "in:%d", conn->child_port);
+ "in:%d", conn->remote_port);
if (!ins)
break;
link = devm_kzalloc(&orig->dev,
@@ -178,7 +178,7 @@ int coresight_make_links(struct coresight_device *orig,
* Install the device connection. This also indicates that
* the links are operational on both ends.
*/
- conn->child_dev = target;
+ conn->remote_dev = target;
return 0;
} while (0);
@@ -198,9 +198,9 @@ void coresight_remove_links(struct coresight_device *orig,
coresight_remove_sysfs_link(conn->link);
- devm_kfree(&conn->child_dev->dev, conn->link->target_name);
+ devm_kfree(&conn->remote_dev->dev, conn->link->target_name);
devm_kfree(&orig->dev, conn->link->orig_name);
devm_kfree(&orig->dev, conn->link);
conn->link = NULL;
- conn->child_dev = NULL;
+ conn->remote_dev = NULL;
}
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 4711dfa7418c..223081dc3ad4 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
return NULL;
for (i = 0; i < etr->pdata->nr_outconns; i++) {
- tmp = etr->pdata->out_conns[i].child_dev;
+ tmp = etr->pdata->out_conns[i].remote_dev;
if (tmp && coresight_is_catu_device(tmp))
return tmp;
}
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index cdf0d1def778..64bb5fc95afa 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -164,18 +164,23 @@ struct coresight_desc {
/**
* struct coresight_connection - representation of a single connection
- * @outport: a connection's output port number.
- * @child_port: remote component's port number @output is connected to.
- * @chid_fwnode: remote component's fwnode handle.
- * @child_dev: a @coresight_device representation of the component
- connected to @outport.
+ * @port: this connection's local port number. For input
+ * connections this is the in-port. For outputs it's the
+ * out-port.
+ * @remote_port: remote component's port number @port is connected to.
+ * For input connections this is the out-port number on
+ * remote device. For output connections it's the in-port
+ * on the remote device.
+ * @remote_fwnode: remote component's fwnode handle.
+ * @remote_dev: a @coresight_device representation of the component
+ * connected to @port.
* @link: Representation of the connection as a sysfs link.
*/
struct coresight_connection {
- int outport;
- int child_port;
- struct fwnode_handle *child_fwnode;
- struct coresight_device *child_dev;
+ int port;
+ int remote_port;
+ struct fwnode_handle *remote_fwnode;
+ struct coresight_device *remote_dev;
struct coresight_sysfs_link *link;
};
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 4/9] coresight: Rename connection members to allow for input connections
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
References to outport and child_port will be confusing when we use the
same struct for inputs as well as outputs. Rename the members to refer
to the position in terms of local and remote instead.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 70 +++++++++----------
.../hwtracing/coresight/coresight-platform.c | 32 ++++-----
drivers/hwtracing/coresight/coresight-sysfs.c | 10 +--
.../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
include/linux/coresight.h | 23 +++---
5 files changed, 71 insertions(+), 66 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index 617be08883da..f457914e445e 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -120,8 +120,8 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
for (i = 0; i < parent->pdata->nr_outconns; i++) {
conn = &parent->pdata->out_conns[i];
- if (conn->child_dev == csdev)
- return conn->child_port;
+ if (conn->remote_dev == csdev)
+ return conn->remote_port;
}
dev_err(&csdev->dev, "couldn't find inport, parent: %s, child: %s\n",
@@ -138,8 +138,8 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
conn = &csdev->pdata->out_conns[i];
- if (conn->child_dev == child)
- return conn->outport;
+ if (conn->remote_dev == child)
+ return conn->port;
}
dev_err(&csdev->dev, "couldn't find outport, parent: %s, child: %s\n",
@@ -604,11 +604,11 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
* Recursively explore each port found on this element.
*/
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
- struct coresight_device *child_dev;
+ struct coresight_device *remote_dev;
- child_dev = csdev->pdata->out_conns[i].child_dev;
- if (child_dev)
- sink = coresight_find_enabled_sink(child_dev);
+ remote_dev = csdev->pdata->out_conns[i].remote_dev;
+ if (remote_dev)
+ sink = coresight_find_enabled_sink(remote_dev);
if (sink)
return sink;
}
@@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child;
- child = csdev->pdata->out_conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].remote_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
if (!coresight_get_ref(child))
goto err;
@@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
for (i--; i >= 0; i--) {
struct coresight_device *child;
- child = csdev->pdata->out_conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].remote_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
coresight_put_ref(child);
}
@@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_device *child;
- child = csdev->pdata->out_conns[i].child_dev;
+ child = csdev->pdata->out_conns[i].remote_dev;
if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
coresight_put_ref(child);
}
@@ -792,11 +792,11 @@ static int _coresight_build_path(struct coresight_device *csdev,
/* Not a sink - recursively explore each port found on this element */
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
- struct coresight_device *child_dev;
+ struct coresight_device *remote_dev;
- child_dev = csdev->pdata->out_conns[i].child_dev;
- if (child_dev &&
- _coresight_build_path(child_dev, sink, path) == 0) {
+ remote_dev = csdev->pdata->out_conns[i].remote_dev;
+ if (remote_dev &&
+ _coresight_build_path(remote_dev, sink, path) == 0) {
found = true;
break;
}
@@ -961,12 +961,12 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
* recursively explore each port found on this element.
*/
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
- struct coresight_device *child_dev, *sink = NULL;
+ struct coresight_device *remote_dev, *sink = NULL;
int child_depth = curr_depth;
- child_dev = csdev->pdata->out_conns[i].child_dev;
- if (child_dev)
- sink = coresight_find_sink(child_dev, &child_depth);
+ remote_dev = csdev->pdata->out_conns[i].remote_dev;
+ if (remote_dev)
+ sink = coresight_find_sink(remote_dev, &child_depth);
if (sink)
found_sink = coresight_select_best_sink(found_sink,
@@ -1337,12 +1337,12 @@ static int coresight_orphan_match(struct device *dev, void *data)
conn = &i_csdev->pdata->out_conns[i];
/* Skip the port if FW doesn't describe it */
- if (!conn->child_fwnode)
+ if (!conn->remote_fwnode)
continue;
/* We have found at least one orphan connection */
- if (conn->child_dev == NULL) {
+ if (conn->remote_dev == NULL) {
/* Does it match this newly added device? */
- if (conn->child_fwnode == csdev->dev.fwnode) {
+ if (conn->remote_fwnode == csdev->dev.fwnode) {
ret = coresight_make_links(i_csdev,
conn, csdev);
if (ret)
@@ -1377,13 +1377,13 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
struct coresight_connection *conn = &csdev->pdata->out_conns[i];
- if (!conn->child_fwnode)
+ if (!conn->remote_fwnode)
continue;
- conn->child_dev =
- coresight_find_csdev_by_fwnode(conn->child_fwnode);
- if (conn->child_dev && conn->child_dev->has_conns_grp) {
+ conn->remote_dev =
+ coresight_find_csdev_by_fwnode(conn->remote_fwnode);
+ if (conn->remote_dev && conn->remote_dev->has_conns_grp) {
ret = coresight_make_links(csdev, conn,
- conn->child_dev);
+ conn->remote_dev);
if (ret)
break;
} else {
@@ -1414,10 +1414,10 @@ static int coresight_remove_match(struct device *dev, void *data)
for (i = 0; i < iterator->pdata->nr_outconns; i++) {
conn = &iterator->pdata->out_conns[i];
- if (conn->child_dev == NULL || conn->child_fwnode == NULL)
+ if (conn->remote_dev == NULL || conn->remote_fwnode == NULL)
continue;
- if (csdev->dev.fwnode == conn->child_fwnode) {
+ if (csdev->dev.fwnode == conn->remote_fwnode) {
iterator->orphan = true;
coresight_remove_links(iterator, conn);
/*
@@ -1425,8 +1425,8 @@ static int coresight_remove_match(struct device *dev, void *data)
* device acquired in parsing the connections from
* platform data.
*/
- fwnode_handle_put(conn->child_fwnode);
- conn->child_fwnode = NULL;
+ fwnode_handle_put(conn->remote_fwnode);
+ conn->remote_fwnode = NULL;
/* No need to continue */
break;
}
@@ -1552,15 +1552,15 @@ void coresight_release_platform_data(struct coresight_device *csdev,
for (i = 0; i < pdata->nr_outconns; i++) {
/* If we have made the links, remove them now */
- if (csdev && conns[i].child_dev)
+ if (csdev && conns[i].remote_dev)
coresight_remove_links(csdev, &conns[i]);
/*
* Drop the refcount and clear the handle as this device
* is going away
*/
- if (conns[i].child_fwnode) {
- fwnode_handle_put(conns[i].child_fwnode);
- pdata->out_conns[i].child_fwnode = NULL;
+ if (conns[i].remote_fwnode) {
+ fwnode_handle_put(conns[i].remote_fwnode);
+ pdata->out_conns[i].remote_fwnode = NULL;
}
}
if (csdev)
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index ed865e0621a9..c77238cdf448 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -252,13 +252,13 @@ static int of_coresight_parse_endpoint(struct device *dev,
}
conn = &pdata->out_conns[endpoint.port];
- if (conn->child_fwnode) {
+ if (conn->remote_fwnode) {
dev_warn(dev, "Duplicate output port %d\n",
endpoint.port);
ret = -EINVAL;
break;
}
- conn->outport = endpoint.port;
+ conn->port = endpoint.port;
/*
* Hold the refcount to the target device. This could be
* released via:
@@ -267,8 +267,8 @@ static int of_coresight_parse_endpoint(struct device *dev,
* 2) While removing the target device via
* coresight_remove_match()
*/
- conn->child_fwnode = fwnode_handle_get(rdev_fwnode);
- conn->child_port = rendpoint.port;
+ conn->remote_fwnode = fwnode_handle_get(rdev_fwnode);
+ conn->remote_port = rendpoint.port;
/* Connection record updated */
} while (0);
@@ -649,8 +649,8 @@ static int acpi_coresight_parse_link(struct acpi_device *adev,
dir = fields[3].integer.value;
if (dir == ACPI_CORESIGHT_LINK_MASTER) {
- conn->outport = fields[0].integer.value;
- conn->child_port = fields[1].integer.value;
+ conn->port = fields[0].integer.value;
+ conn->remote_port = fields[1].integer.value;
rdev = coresight_find_device_by_fwnode(&r_adev->fwnode);
if (!rdev)
return -EPROBE_DEFER;
@@ -662,14 +662,14 @@ static int acpi_coresight_parse_link(struct acpi_device *adev,
* 2) While removing the target device via
* coresight_remove_match().
*/
- conn->child_fwnode = fwnode_handle_get(&r_adev->fwnode);
+ conn->remote_fwnode = fwnode_handle_get(&r_adev->fwnode);
} else if (dir == ACPI_CORESIGHT_LINK_SLAVE) {
/*
* We are only interested in the port number
* for the input ports at this component.
- * Store the port number in child_port.
+ * Store the port number in remote_port.
*/
- conn->child_port = fields[0].integer.value;
+ conn->remote_port = fields[0].integer.value;
} else {
/* Invalid direction */
return -EINVAL;
@@ -718,11 +718,11 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
return dir;
if (dir == ACPI_CORESIGHT_LINK_MASTER) {
- if (ptr->outport >= pdata->nr_outconns)
- pdata->nr_outconns = ptr->outport + 1;
+ if (ptr->port >= pdata->nr_outconns)
+ pdata->nr_outconns = ptr->port + 1;
ptr++;
} else {
- WARN_ON(pdata->nr_inconns == ptr->child_port + 1);
+ WARN_ON(pdata->nr_inconns == ptr->remote_port + 1);
/*
* We do not track input port connections for a device.
* However we need the highest port number described,
@@ -730,8 +730,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
* record for an output connection. Hence, do not move
* the ptr for input connections
*/
- if (ptr->child_port >= pdata->nr_inconns)
- pdata->nr_inconns = ptr->child_port + 1;
+ if (ptr->remote_port >= pdata->nr_inconns)
+ pdata->nr_inconns = ptr->remote_port + 1;
}
}
@@ -741,10 +741,10 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
/* Copy the connection information to the final location */
for (i = 0; conns + i < ptr; i++) {
- int port = conns[i].outport;
+ int port = conns[i].port;
/* Duplicate output port */
- WARN_ON(pdata->out_conns[port].child_fwnode);
+ WARN_ON(pdata->out_conns[port].remote_fwnode);
pdata->out_conns[port] = conns[i];
}
diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
index 34d2a2d31d00..3da9868d9237 100644
--- a/drivers/hwtracing/coresight/coresight-sysfs.c
+++ b/drivers/hwtracing/coresight/coresight-sysfs.c
@@ -150,11 +150,11 @@ int coresight_make_links(struct coresight_device *orig,
do {
outs = devm_kasprintf(&orig->dev, GFP_KERNEL,
- "out:%d", conn->outport);
+ "out:%d", conn->port);
if (!outs)
break;
ins = devm_kasprintf(&target->dev, GFP_KERNEL,
- "in:%d", conn->child_port);
+ "in:%d", conn->remote_port);
if (!ins)
break;
link = devm_kzalloc(&orig->dev,
@@ -178,7 +178,7 @@ int coresight_make_links(struct coresight_device *orig,
* Install the device connection. This also indicates that
* the links are operational on both ends.
*/
- conn->child_dev = target;
+ conn->remote_dev = target;
return 0;
} while (0);
@@ -198,9 +198,9 @@ void coresight_remove_links(struct coresight_device *orig,
coresight_remove_sysfs_link(conn->link);
- devm_kfree(&conn->child_dev->dev, conn->link->target_name);
+ devm_kfree(&conn->remote_dev->dev, conn->link->target_name);
devm_kfree(&orig->dev, conn->link->orig_name);
devm_kfree(&orig->dev, conn->link);
conn->link = NULL;
- conn->child_dev = NULL;
+ conn->remote_dev = NULL;
}
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 4711dfa7418c..223081dc3ad4 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
return NULL;
for (i = 0; i < etr->pdata->nr_outconns; i++) {
- tmp = etr->pdata->out_conns[i].child_dev;
+ tmp = etr->pdata->out_conns[i].remote_dev;
if (tmp && coresight_is_catu_device(tmp))
return tmp;
}
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index cdf0d1def778..64bb5fc95afa 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -164,18 +164,23 @@ struct coresight_desc {
/**
* struct coresight_connection - representation of a single connection
- * @outport: a connection's output port number.
- * @child_port: remote component's port number @output is connected to.
- * @chid_fwnode: remote component's fwnode handle.
- * @child_dev: a @coresight_device representation of the component
- connected to @outport.
+ * @port: this connection's local port number. For input
+ * connections this is the in-port. For outputs it's the
+ * out-port.
+ * @remote_port: remote component's port number @port is connected to.
+ * For input connections this is the out-port number on
+ * remote device. For output connections it's the in-port
+ * on the remote device.
+ * @remote_fwnode: remote component's fwnode handle.
+ * @remote_dev: a @coresight_device representation of the component
+ * connected to @port.
* @link: Representation of the connection as a sysfs link.
*/
struct coresight_connection {
- int outport;
- int child_port;
- struct fwnode_handle *child_fwnode;
- struct coresight_device *child_dev;
+ int port;
+ int remote_port;
+ struct fwnode_handle *remote_fwnode;
+ struct coresight_device *remote_dev;
struct coresight_sysfs_link *link;
};
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 5/9] coresight: Dynamically add connections
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
Add a function for adding connections dynamically. This also removes
the 1:1 mapping between port number and the index into the connections
array. The only place this mapping was used was in the warning for
duplicate output ports, which has been replaced by a search. Other
uses of the port number already use the port member variable.
Being able to dynamically add connections will allow other devices like
CTI to re-use the connection mechanism despite not having explicit
connections described in the DT.
Signed-off-by: James Clark <james.clark@arm.com>
---
.../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
include/linux/coresight.h | 7 +-
2 files changed, 64 insertions(+), 20 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index c77238cdf448..16553f7dde12 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
struct coresight_platform_data *pdata)
{
if (pdata->nr_outconns) {
- pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
- sizeof(*pdata->out_conns), GFP_KERNEL);
+ pdata->out_conns = devm_krealloc_array(
+ dev, pdata->out_conns, pdata->nr_outconns,
+ sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
if (!pdata->out_conns)
return -ENOMEM;
}
@@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
return 0;
}
+/*
+ * Add a connection in the first free slot, or realloc
+ * if there is no space. @conn's contents is copied into the new slot.
+ *
+ * If the output port is already assigned on this device, return -EINVAL
+ */
+int coresight_add_conn(struct device *dev,
+ struct coresight_platform_data *pdata,
+ const struct coresight_connection *conn)
+{
+ int ret;
+ struct coresight_connection *free_conn = NULL;
+ struct coresight_connection *i;
+
+ /*
+ * Search for a free slot, and while looking for one, warn
+ * on any existing duplicate output port.
+ */
+ for (i = pdata->out_conns; i < pdata->out_conns + pdata->nr_outconns;
+ ++i) {
+ if (i->remote_fwnode && conn->port != -1 &&
+ i->port == conn->port) {
+ dev_warn(dev, "Duplicate output port %d\n", i->port);
+ return -EINVAL;
+ }
+ if (!i->remote_fwnode && !free_conn)
+ free_conn = i;
+ }
+
+ if (!free_conn) {
+ pdata->nr_outconns++;
+ ret = coresight_alloc_conns(dev, pdata);
+ if (ret)
+ return ret;
+ free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
+ }
+
+ *free_conn = *conn;
+ return 0;
+}
+EXPORT_SYMBOL_GPL(coresight_add_conn);
+
static struct device *
coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
{
@@ -224,7 +267,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
struct device_node *rep = NULL;
struct device *rdev = NULL;
struct fwnode_handle *rdev_fwnode;
- struct coresight_connection *conn;
+ struct coresight_connection conn;
do {
/* Parse the local port details */
@@ -251,14 +294,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
break;
}
- conn = &pdata->out_conns[endpoint.port];
- if (conn->remote_fwnode) {
- dev_warn(dev, "Duplicate output port %d\n",
- endpoint.port);
- ret = -EINVAL;
- break;
- }
- conn->port = endpoint.port;
+ conn.port = endpoint.port;
/*
* Hold the refcount to the target device. This could be
* released via:
@@ -267,8 +303,14 @@ static int of_coresight_parse_endpoint(struct device *dev,
* 2) While removing the target device via
* coresight_remove_match()
*/
- conn->remote_fwnode = fwnode_handle_get(rdev_fwnode);
- conn->remote_port = rendpoint.port;
+ conn.remote_fwnode = fwnode_handle_get(rdev_fwnode);
+ conn.remote_port = rendpoint.port;
+
+ ret = coresight_add_conn(dev, pdata, &conn);
+ if (ret) {
+ fwnode_handle_put(conn.remote_fwnode);
+ return ret;
+ }
/* Connection record updated */
} while (0);
@@ -741,13 +783,10 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
/* Copy the connection information to the final location */
for (i = 0; conns + i < ptr; i++) {
- int port = conns[i].port;
-
- /* Duplicate output port */
- WARN_ON(pdata->out_conns[port].remote_fwnode);
- pdata->out_conns[port] = conns[i];
+ rc = coresight_add_conn(&adev->dev, pdata, &conns[i]);
+ if (rc)
+ return rc;
}
-
devm_kfree(&adev->dev, conns);
return 0;
}
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 64bb5fc95afa..47fa58d6981d 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -173,7 +173,9 @@ struct coresight_desc {
* on the remote device.
* @remote_fwnode: remote component's fwnode handle.
* @remote_dev: a @coresight_device representation of the component
- * connected to @port.
+ * connected to @port. Will be looked up using
+ * @remote_fwnode once the remote's coresight_device has
+ * been created.
* @link: Representation of the connection as a sysfs link.
*/
struct coresight_connection {
@@ -614,5 +616,8 @@ static inline void coresight_write64(struct coresight_device *csdev, u64 val, u3
extern int coresight_get_cpu(struct device *dev);
struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
+int coresight_add_conn(struct device *dev,
+ struct coresight_platform_data *pdata,
+ const struct coresight_connection *conn);
#endif /* _LINUX_COREISGHT_H */
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 5/9] coresight: Dynamically add connections
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
Add a function for adding connections dynamically. This also removes
the 1:1 mapping between port number and the index into the connections
array. The only place this mapping was used was in the warning for
duplicate output ports, which has been replaced by a search. Other
uses of the port number already use the port member variable.
Being able to dynamically add connections will allow other devices like
CTI to re-use the connection mechanism despite not having explicit
connections described in the DT.
Signed-off-by: James Clark <james.clark@arm.com>
---
.../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
include/linux/coresight.h | 7 +-
2 files changed, 64 insertions(+), 20 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index c77238cdf448..16553f7dde12 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
struct coresight_platform_data *pdata)
{
if (pdata->nr_outconns) {
- pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
- sizeof(*pdata->out_conns), GFP_KERNEL);
+ pdata->out_conns = devm_krealloc_array(
+ dev, pdata->out_conns, pdata->nr_outconns,
+ sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
if (!pdata->out_conns)
return -ENOMEM;
}
@@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
return 0;
}
+/*
+ * Add a connection in the first free slot, or realloc
+ * if there is no space. @conn's contents is copied into the new slot.
+ *
+ * If the output port is already assigned on this device, return -EINVAL
+ */
+int coresight_add_conn(struct device *dev,
+ struct coresight_platform_data *pdata,
+ const struct coresight_connection *conn)
+{
+ int ret;
+ struct coresight_connection *free_conn = NULL;
+ struct coresight_connection *i;
+
+ /*
+ * Search for a free slot, and while looking for one, warn
+ * on any existing duplicate output port.
+ */
+ for (i = pdata->out_conns; i < pdata->out_conns + pdata->nr_outconns;
+ ++i) {
+ if (i->remote_fwnode && conn->port != -1 &&
+ i->port == conn->port) {
+ dev_warn(dev, "Duplicate output port %d\n", i->port);
+ return -EINVAL;
+ }
+ if (!i->remote_fwnode && !free_conn)
+ free_conn = i;
+ }
+
+ if (!free_conn) {
+ pdata->nr_outconns++;
+ ret = coresight_alloc_conns(dev, pdata);
+ if (ret)
+ return ret;
+ free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
+ }
+
+ *free_conn = *conn;
+ return 0;
+}
+EXPORT_SYMBOL_GPL(coresight_add_conn);
+
static struct device *
coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
{
@@ -224,7 +267,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
struct device_node *rep = NULL;
struct device *rdev = NULL;
struct fwnode_handle *rdev_fwnode;
- struct coresight_connection *conn;
+ struct coresight_connection conn;
do {
/* Parse the local port details */
@@ -251,14 +294,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
break;
}
- conn = &pdata->out_conns[endpoint.port];
- if (conn->remote_fwnode) {
- dev_warn(dev, "Duplicate output port %d\n",
- endpoint.port);
- ret = -EINVAL;
- break;
- }
- conn->port = endpoint.port;
+ conn.port = endpoint.port;
/*
* Hold the refcount to the target device. This could be
* released via:
@@ -267,8 +303,14 @@ static int of_coresight_parse_endpoint(struct device *dev,
* 2) While removing the target device via
* coresight_remove_match()
*/
- conn->remote_fwnode = fwnode_handle_get(rdev_fwnode);
- conn->remote_port = rendpoint.port;
+ conn.remote_fwnode = fwnode_handle_get(rdev_fwnode);
+ conn.remote_port = rendpoint.port;
+
+ ret = coresight_add_conn(dev, pdata, &conn);
+ if (ret) {
+ fwnode_handle_put(conn.remote_fwnode);
+ return ret;
+ }
/* Connection record updated */
} while (0);
@@ -741,13 +783,10 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
/* Copy the connection information to the final location */
for (i = 0; conns + i < ptr; i++) {
- int port = conns[i].port;
-
- /* Duplicate output port */
- WARN_ON(pdata->out_conns[port].remote_fwnode);
- pdata->out_conns[port] = conns[i];
+ rc = coresight_add_conn(&adev->dev, pdata, &conns[i]);
+ if (rc)
+ return rc;
}
-
devm_kfree(&adev->dev, conns);
return 0;
}
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 64bb5fc95afa..47fa58d6981d 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -173,7 +173,9 @@ struct coresight_desc {
* on the remote device.
* @remote_fwnode: remote component's fwnode handle.
* @remote_dev: a @coresight_device representation of the component
- * connected to @port.
+ * connected to @port. Will be looked up using
+ * @remote_fwnode once the remote's coresight_device has
+ * been created.
* @link: Representation of the connection as a sysfs link.
*/
struct coresight_connection {
@@ -614,5 +616,8 @@ static inline void coresight_write64(struct coresight_device *csdev, u64 val, u3
extern int coresight_get_cpu(struct device *dev);
struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
+int coresight_add_conn(struct device *dev,
+ struct coresight_platform_data *pdata,
+ const struct coresight_connection *conn);
#endif /* _LINUX_COREISGHT_H */
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
This will allow CATU to get its associated ETR in a generic way where
currently the enable path has some hard coded searches which avoid
the need to store input connections.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
.../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
include/linux/coresight.h | 25 ++++++++
4 files changed, 130 insertions(+), 13 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index f457914e445e..a8ba7493c09a 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fff
EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
static const struct cti_assoc_op *cti_assoc_ops;
+static int coresight_fixup_inputs(struct coresight_device *csdev);
ssize_t coresight_simple_show_pair(struct device *_dev,
struct device_attribute *attr, char *buf)
@@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct coresight_device *csdev)
csdev, coresight_orphan_match);
}
+/*
+ * Device connections are discovered before one/both devices have been created,
+ * so inputs must be added later.
+ */
+static int coresight_fixup_inputs(struct coresight_device *csdev)
+{
+ int i, ret = 0;
+ struct coresight_connection *out_conn;
+ struct coresight_connection in_conn;
+
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
+ out_conn = &csdev->pdata->out_conns[i];
+ if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
+ continue;
+
+ /* Reverse local/remote relationships for inputs */
+ in_conn.remote_dev = csdev;
+ in_conn.remote_port = out_conn->port;
+ in_conn.port = out_conn->remote_port;
+ in_conn.remote_fwnode = csdev->dev.fwnode;
+ ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
+ out_conn->remote_dev->pdata,
+ &in_conn);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
static int coresight_fixup_device_conns(struct coresight_device *csdev)
{
@@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct device *dev, void *data)
*/
fwnode_handle_put(conn->remote_fwnode);
conn->remote_fwnode = NULL;
+ conn->remote_dev = NULL;
+ /* No need to continue */
+ break;
+ }
+ }
+ for (i = 0; i < iterator->pdata->nr_inconns; i++) {
+ conn = &iterator->pdata->in_conns[i];
+ if (csdev == conn->remote_dev) {
+ conn->remote_fwnode = NULL;
+ conn->remote_dev = NULL;
/* No need to continue */
break;
}
}
-
/*
* Returning '0' ensures that all known component on the
* bus will be checked.
@@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct coresight_device *csdev,
for (i = 0; i < pdata->nr_outconns; i++) {
/* If we have made the links, remove them now */
- if (csdev && conns[i].remote_dev)
+ if (csdev && conns[i].remote_dev) {
coresight_remove_links(csdev, &conns[i]);
+ conns[i].remote_dev = NULL;
+ }
+
/*
* Drop the refcount and clear the handle as this device
* is going away
*/
if (conns[i].remote_fwnode) {
fwnode_handle_put(conns[i].remote_fwnode);
- pdata->out_conns[i].remote_fwnode = NULL;
+ conns[i].remote_fwnode = NULL;
}
}
+ for (i = 0; i < pdata->nr_inconns; i++) {
+ pdata->in_conns[i].remote_dev = NULL;
+ pdata->in_conns[i].remote_fwnode = NULL;
+ }
+
if (csdev)
coresight_remove_conns_sysfs_group(csdev);
}
-
struct coresight_device *coresight_register(struct coresight_desc *desc)
{
int ret;
@@ -1659,6 +1705,8 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
ret = coresight_create_conns_sysfs_group(csdev);
if (!ret)
ret = coresight_fixup_device_conns(csdev);
+ if (!ret)
+ ret = coresight_fixup_inputs(csdev);
if (!ret)
ret = coresight_fixup_orphan_conns(csdev);
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index 16553f7dde12..20e3351cbdc2 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -20,8 +20,7 @@
#include "coresight-priv.h"
/*
- * coresight_alloc_conns: Allocate connections record for each output
- * port from the device.
+ * coresight_alloc_conns: Allocate connections record for each input/output device.
*/
static int coresight_alloc_conns(struct device *dev,
struct coresight_platform_data *pdata)
@@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
if (!pdata->out_conns)
return -ENOMEM;
}
-
+ if (pdata->nr_inconns) {
+ pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
+ pdata->nr_inconns,
+ sizeof(*pdata->in_conns),
+ GFP_KERNEL | __GFP_ZERO);
+ if (!pdata->in_conns)
+ return -ENOMEM;
+ }
return 0;
}
@@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
}
EXPORT_SYMBOL_GPL(coresight_add_conn);
+/*
+ * Add a connection in the first free slot, or realloc
+ * if there is no space.
+ *
+ * Do nothing if the connection already exists because inputs are
+ * fixed up multiple times.
+ */
+int coresight_add_in_conn(struct device *dev,
+ struct coresight_platform_data *pdata,
+ struct coresight_connection *conn)
+{
+ int ret;
+ struct coresight_connection *free_conn = NULL;
+ int i;
+
+ /* Search for a free slot or exit if a duplicate is found */
+ if (pdata->in_conns) {
+ for (i = 0; i < pdata->nr_inconns; ++i) {
+ if (!free_conn && !pdata->in_conns[i].remote_fwnode)
+ free_conn = &pdata->in_conns[i];
+ if (pdata->in_conns[i].remote_fwnode ==
+ conn->remote_fwnode)
+ return 0;
+ }
+ }
+
+ if (!free_conn) {
+ pdata->nr_inconns++;
+ ret = coresight_alloc_conns(dev, pdata);
+ if (ret)
+ return ret;
+ free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
+ }
+
+ *free_conn = *conn;
+ return 0;
+}
+EXPORT_SYMBOL_GPL(coresight_add_in_conn);
+
static struct device *
coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
{
@@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
/*
* of_coresight_parse_endpoint : Parse the given output endpoint @ep
- * and fill the connection information in @conn
+ * and fill the connection information in @in_conn and @out_conn
*
* Parses the local port, remote device name and the remote port.
*
@@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct device *dev,
/* Get the number of input and output port for this component */
of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
- /* If there are no output connections, we are done */
- if (!pdata->nr_outconns)
- return 0;
-
ret = coresight_alloc_conns(dev, pdata);
if (ret)
return ret;
+ /* If there are no output connections, we are done */
+ if (!pdata->nr_outconns)
+ return 0;
+
parent = of_coresight_get_output_ports_node(node);
/*
* If the DT uses obsoleted bindings, the ports are listed
diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
index 3da9868d9237..2abf9639ac0f 100644
--- a/drivers/hwtracing/coresight/coresight-sysfs.c
+++ b/drivers/hwtracing/coresight/coresight-sysfs.c
@@ -202,5 +202,4 @@ void coresight_remove_links(struct coresight_device *orig,
devm_kfree(&orig->dev, conn->link->orig_name);
devm_kfree(&orig->dev, conn->link);
conn->link = NULL;
- conn->remote_dev = NULL;
}
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 47fa58d6981d..fd268b24c761 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -110,6 +110,7 @@ struct coresight_platform_data {
int nr_inconns;
int nr_outconns;
struct coresight_connection *out_conns;
+ struct coresight_connection *in_conns;
};
/**
@@ -177,6 +178,27 @@ struct coresight_desc {
* @remote_fwnode once the remote's coresight_device has
* been created.
* @link: Representation of the connection as a sysfs link.
+ *
+ * The full connection structure looks like this, where in_conns store references to
+ * the parent device in the same remote_dev member as output connections.
+ *
+ * +-----------------------------+ +-----------------------------+
+ * |coresight_device | |coresight_connection |
+ * |-----------------------------| |-----------------------------|
+ * ---->| | | |
+ * | | | | remote_dev*|------
+ * | |pdata->out_conns[nr_outconns]|----------->| | |
+ * | | | | | |
+ * | +-----------------------------+ +-----------------------------+ |
+ * | |
+ * | +-----------------------------+ +-----------------------------+ |
+ * | |coresight_connection | |coresight_device | |
+ * | |-----------------------------| |------------------------------ |
+ * | | | | |<-----
+ * -----|remote_dev* | | |
+ * | |<-----------|pdata->in_conns[nr_inconns] |
+ * | | | |
+ * +-----------------------------+ +-----------------------------+
*/
struct coresight_connection {
int port;
@@ -619,5 +641,8 @@ struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
int coresight_add_conn(struct device *dev,
struct coresight_platform_data *pdata,
const struct coresight_connection *conn);
+int coresight_add_in_conn(struct device *dev,
+ struct coresight_platform_data *pdata,
+ struct coresight_connection *conn);
#endif /* _LINUX_COREISGHT_H */
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
This will allow CATU to get its associated ETR in a generic way where
currently the enable path has some hard coded searches which avoid
the need to store input connections.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
.../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
include/linux/coresight.h | 25 ++++++++
4 files changed, 130 insertions(+), 13 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index f457914e445e..a8ba7493c09a 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fff
EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
static const struct cti_assoc_op *cti_assoc_ops;
+static int coresight_fixup_inputs(struct coresight_device *csdev);
ssize_t coresight_simple_show_pair(struct device *_dev,
struct device_attribute *attr, char *buf)
@@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct coresight_device *csdev)
csdev, coresight_orphan_match);
}
+/*
+ * Device connections are discovered before one/both devices have been created,
+ * so inputs must be added later.
+ */
+static int coresight_fixup_inputs(struct coresight_device *csdev)
+{
+ int i, ret = 0;
+ struct coresight_connection *out_conn;
+ struct coresight_connection in_conn;
+
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
+ out_conn = &csdev->pdata->out_conns[i];
+ if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
+ continue;
+
+ /* Reverse local/remote relationships for inputs */
+ in_conn.remote_dev = csdev;
+ in_conn.remote_port = out_conn->port;
+ in_conn.port = out_conn->remote_port;
+ in_conn.remote_fwnode = csdev->dev.fwnode;
+ ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
+ out_conn->remote_dev->pdata,
+ &in_conn);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
static int coresight_fixup_device_conns(struct coresight_device *csdev)
{
@@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct device *dev, void *data)
*/
fwnode_handle_put(conn->remote_fwnode);
conn->remote_fwnode = NULL;
+ conn->remote_dev = NULL;
+ /* No need to continue */
+ break;
+ }
+ }
+ for (i = 0; i < iterator->pdata->nr_inconns; i++) {
+ conn = &iterator->pdata->in_conns[i];
+ if (csdev == conn->remote_dev) {
+ conn->remote_fwnode = NULL;
+ conn->remote_dev = NULL;
/* No need to continue */
break;
}
}
-
/*
* Returning '0' ensures that all known component on the
* bus will be checked.
@@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct coresight_device *csdev,
for (i = 0; i < pdata->nr_outconns; i++) {
/* If we have made the links, remove them now */
- if (csdev && conns[i].remote_dev)
+ if (csdev && conns[i].remote_dev) {
coresight_remove_links(csdev, &conns[i]);
+ conns[i].remote_dev = NULL;
+ }
+
/*
* Drop the refcount and clear the handle as this device
* is going away
*/
if (conns[i].remote_fwnode) {
fwnode_handle_put(conns[i].remote_fwnode);
- pdata->out_conns[i].remote_fwnode = NULL;
+ conns[i].remote_fwnode = NULL;
}
}
+ for (i = 0; i < pdata->nr_inconns; i++) {
+ pdata->in_conns[i].remote_dev = NULL;
+ pdata->in_conns[i].remote_fwnode = NULL;
+ }
+
if (csdev)
coresight_remove_conns_sysfs_group(csdev);
}
-
struct coresight_device *coresight_register(struct coresight_desc *desc)
{
int ret;
@@ -1659,6 +1705,8 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
ret = coresight_create_conns_sysfs_group(csdev);
if (!ret)
ret = coresight_fixup_device_conns(csdev);
+ if (!ret)
+ ret = coresight_fixup_inputs(csdev);
if (!ret)
ret = coresight_fixup_orphan_conns(csdev);
diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
index 16553f7dde12..20e3351cbdc2 100644
--- a/drivers/hwtracing/coresight/coresight-platform.c
+++ b/drivers/hwtracing/coresight/coresight-platform.c
@@ -20,8 +20,7 @@
#include "coresight-priv.h"
/*
- * coresight_alloc_conns: Allocate connections record for each output
- * port from the device.
+ * coresight_alloc_conns: Allocate connections record for each input/output device.
*/
static int coresight_alloc_conns(struct device *dev,
struct coresight_platform_data *pdata)
@@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
if (!pdata->out_conns)
return -ENOMEM;
}
-
+ if (pdata->nr_inconns) {
+ pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
+ pdata->nr_inconns,
+ sizeof(*pdata->in_conns),
+ GFP_KERNEL | __GFP_ZERO);
+ if (!pdata->in_conns)
+ return -ENOMEM;
+ }
return 0;
}
@@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
}
EXPORT_SYMBOL_GPL(coresight_add_conn);
+/*
+ * Add a connection in the first free slot, or realloc
+ * if there is no space.
+ *
+ * Do nothing if the connection already exists because inputs are
+ * fixed up multiple times.
+ */
+int coresight_add_in_conn(struct device *dev,
+ struct coresight_platform_data *pdata,
+ struct coresight_connection *conn)
+{
+ int ret;
+ struct coresight_connection *free_conn = NULL;
+ int i;
+
+ /* Search for a free slot or exit if a duplicate is found */
+ if (pdata->in_conns) {
+ for (i = 0; i < pdata->nr_inconns; ++i) {
+ if (!free_conn && !pdata->in_conns[i].remote_fwnode)
+ free_conn = &pdata->in_conns[i];
+ if (pdata->in_conns[i].remote_fwnode ==
+ conn->remote_fwnode)
+ return 0;
+ }
+ }
+
+ if (!free_conn) {
+ pdata->nr_inconns++;
+ ret = coresight_alloc_conns(dev, pdata);
+ if (ret)
+ return ret;
+ free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
+ }
+
+ *free_conn = *conn;
+ return 0;
+}
+EXPORT_SYMBOL_GPL(coresight_add_in_conn);
+
static struct device *
coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
{
@@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
/*
* of_coresight_parse_endpoint : Parse the given output endpoint @ep
- * and fill the connection information in @conn
+ * and fill the connection information in @in_conn and @out_conn
*
* Parses the local port, remote device name and the remote port.
*
@@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct device *dev,
/* Get the number of input and output port for this component */
of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
- /* If there are no output connections, we are done */
- if (!pdata->nr_outconns)
- return 0;
-
ret = coresight_alloc_conns(dev, pdata);
if (ret)
return ret;
+ /* If there are no output connections, we are done */
+ if (!pdata->nr_outconns)
+ return 0;
+
parent = of_coresight_get_output_ports_node(node);
/*
* If the DT uses obsoleted bindings, the ports are listed
diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
index 3da9868d9237..2abf9639ac0f 100644
--- a/drivers/hwtracing/coresight/coresight-sysfs.c
+++ b/drivers/hwtracing/coresight/coresight-sysfs.c
@@ -202,5 +202,4 @@ void coresight_remove_links(struct coresight_device *orig,
devm_kfree(&orig->dev, conn->link->orig_name);
devm_kfree(&orig->dev, conn->link);
conn->link = NULL;
- conn->remote_dev = NULL;
}
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index 47fa58d6981d..fd268b24c761 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -110,6 +110,7 @@ struct coresight_platform_data {
int nr_inconns;
int nr_outconns;
struct coresight_connection *out_conns;
+ struct coresight_connection *in_conns;
};
/**
@@ -177,6 +178,27 @@ struct coresight_desc {
* @remote_fwnode once the remote's coresight_device has
* been created.
* @link: Representation of the connection as a sysfs link.
+ *
+ * The full connection structure looks like this, where in_conns store references to
+ * the parent device in the same remote_dev member as output connections.
+ *
+ * +-----------------------------+ +-----------------------------+
+ * |coresight_device | |coresight_connection |
+ * |-----------------------------| |-----------------------------|
+ * ---->| | | |
+ * | | | | remote_dev*|------
+ * | |pdata->out_conns[nr_outconns]|----------->| | |
+ * | | | | | |
+ * | +-----------------------------+ +-----------------------------+ |
+ * | |
+ * | +-----------------------------+ +-----------------------------+ |
+ * | |coresight_connection | |coresight_device | |
+ * | |-----------------------------| |------------------------------ |
+ * | | | | |<-----
+ * -----|remote_dev* | | |
+ * | |<-----------|pdata->in_conns[nr_inconns] |
+ * | | | |
+ * +-----------------------------+ +-----------------------------+
*/
struct coresight_connection {
int port;
@@ -619,5 +641,8 @@ struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
int coresight_add_conn(struct device *dev,
struct coresight_platform_data *pdata,
const struct coresight_connection *conn);
+int coresight_add_in_conn(struct device *dev,
+ struct coresight_platform_data *pdata,
+ struct coresight_connection *conn);
#endif /* _LINUX_COREISGHT_H */
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 7/9] coresight: Refactor out buffer allocation function for ETR
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
When CATU is moved to the generic enable/disable path system in the
next commit, it will need to call into ETR and get it to pre-allocate
its buffer so add a function for it.
No functional changes
Signed-off-by: James Clark <james.clark@arm.com>
---
.../hwtracing/coresight/coresight-tmc-etr.c | 50 ++++++++++++++++---
drivers/hwtracing/coresight/coresight-tmc.h | 2 +
2 files changed, 45 insertions(+), 7 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 223081dc3ad4..cb9621003aef 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -1169,7 +1169,7 @@ void tmc_etr_disable_hw(struct tmc_drvdata *drvdata)
drvdata->etr_buf = NULL;
}
-static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
+static struct etr_buf *tmc_etr_get_sysfs_buffer(struct coresight_device *csdev)
{
int ret = 0;
unsigned long flags;
@@ -1192,7 +1192,7 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
/* Allocate memory with the locks released */
free_buf = new_buf = tmc_etr_setup_sysfs_buf(drvdata);
if (IS_ERR(new_buf))
- return PTR_ERR(new_buf);
+ return new_buf;
/* Let's try again */
spin_lock_irqsave(&drvdata->spinlock, flags);
@@ -1223,17 +1223,33 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
drvdata->sysfs_buf = new_buf;
}
- ret = tmc_etr_enable_hw(drvdata, drvdata->sysfs_buf);
- if (!ret) {
- drvdata->mode = CS_MODE_SYSFS;
- atomic_inc(csdev->refcnt);
- }
out:
spin_unlock_irqrestore(&drvdata->spinlock, flags);
/* Free memory outside the spinlock if need be */
if (free_buf)
tmc_etr_free_sysfs_buf(free_buf);
+ return ret ? ERR_PTR(ret) : drvdata->sysfs_buf;
+}
+
+static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
+{
+ int ret;
+ unsigned long flags;
+ struct tmc_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+ struct etr_buf *sysfs_buf = tmc_etr_get_sysfs_buffer(csdev);
+
+ if (IS_ERR(sysfs_buf))
+ return PTR_ERR(sysfs_buf);
+
+ spin_lock_irqsave(&drvdata->spinlock, flags);
+ ret = tmc_etr_enable_hw(drvdata, sysfs_buf);
+ if (!ret) {
+ drvdata->mode = CS_MODE_SYSFS;
+ atomic_inc(csdev->refcnt);
+ }
+
+ spin_unlock_irqrestore(&drvdata->spinlock, flags);
if (!ret)
dev_dbg(&csdev->dev, "TMC-ETR enabled\n");
@@ -1241,6 +1257,26 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
return ret;
}
+struct etr_buf *tmc_etr_get_buffer(struct coresight_device *csdev,
+ enum cs_mode mode, void *data)
+{
+ struct perf_output_handle *handle = data;
+ struct etr_perf_buffer *etr_perf;
+
+ switch (mode) {
+ case CS_MODE_SYSFS:
+ return tmc_etr_get_sysfs_buffer(csdev);
+ case CS_MODE_PERF:
+ etr_perf = etm_perf_sink_config(handle);
+ if (WARN_ON(!etr_perf || !etr_perf->etr_buf))
+ return ERR_PTR(-EINVAL);
+ return etr_perf->etr_buf;
+ default:
+ return ERR_PTR(-EINVAL);
+ }
+}
+EXPORT_SYMBOL_GPL(tmc_etr_get_buffer);
+
/*
* alloc_etr_buf: Allocate ETR buffer for use by perf.
* The size of the hardware buffer is dependent on the size configured
diff --git a/drivers/hwtracing/coresight/coresight-tmc.h b/drivers/hwtracing/coresight/coresight-tmc.h
index 01c0382a29c0..b97da39652d2 100644
--- a/drivers/hwtracing/coresight/coresight-tmc.h
+++ b/drivers/hwtracing/coresight/coresight-tmc.h
@@ -332,5 +332,7 @@ struct coresight_device *tmc_etr_get_catu_device(struct tmc_drvdata *drvdata);
void tmc_etr_set_catu_ops(const struct etr_buf_operations *catu);
void tmc_etr_remove_catu_ops(void);
+struct etr_buf *tmc_etr_get_buffer(struct coresight_device *csdev,
+ enum cs_mode mode, void *data);
#endif
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 7/9] coresight: Refactor out buffer allocation function for ETR
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
When CATU is moved to the generic enable/disable path system in the
next commit, it will need to call into ETR and get it to pre-allocate
its buffer so add a function for it.
No functional changes
Signed-off-by: James Clark <james.clark@arm.com>
---
.../hwtracing/coresight/coresight-tmc-etr.c | 50 ++++++++++++++++---
drivers/hwtracing/coresight/coresight-tmc.h | 2 +
2 files changed, 45 insertions(+), 7 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 223081dc3ad4..cb9621003aef 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -1169,7 +1169,7 @@ void tmc_etr_disable_hw(struct tmc_drvdata *drvdata)
drvdata->etr_buf = NULL;
}
-static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
+static struct etr_buf *tmc_etr_get_sysfs_buffer(struct coresight_device *csdev)
{
int ret = 0;
unsigned long flags;
@@ -1192,7 +1192,7 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
/* Allocate memory with the locks released */
free_buf = new_buf = tmc_etr_setup_sysfs_buf(drvdata);
if (IS_ERR(new_buf))
- return PTR_ERR(new_buf);
+ return new_buf;
/* Let's try again */
spin_lock_irqsave(&drvdata->spinlock, flags);
@@ -1223,17 +1223,33 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
drvdata->sysfs_buf = new_buf;
}
- ret = tmc_etr_enable_hw(drvdata, drvdata->sysfs_buf);
- if (!ret) {
- drvdata->mode = CS_MODE_SYSFS;
- atomic_inc(csdev->refcnt);
- }
out:
spin_unlock_irqrestore(&drvdata->spinlock, flags);
/* Free memory outside the spinlock if need be */
if (free_buf)
tmc_etr_free_sysfs_buf(free_buf);
+ return ret ? ERR_PTR(ret) : drvdata->sysfs_buf;
+}
+
+static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
+{
+ int ret;
+ unsigned long flags;
+ struct tmc_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
+ struct etr_buf *sysfs_buf = tmc_etr_get_sysfs_buffer(csdev);
+
+ if (IS_ERR(sysfs_buf))
+ return PTR_ERR(sysfs_buf);
+
+ spin_lock_irqsave(&drvdata->spinlock, flags);
+ ret = tmc_etr_enable_hw(drvdata, sysfs_buf);
+ if (!ret) {
+ drvdata->mode = CS_MODE_SYSFS;
+ atomic_inc(csdev->refcnt);
+ }
+
+ spin_unlock_irqrestore(&drvdata->spinlock, flags);
if (!ret)
dev_dbg(&csdev->dev, "TMC-ETR enabled\n");
@@ -1241,6 +1257,26 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
return ret;
}
+struct etr_buf *tmc_etr_get_buffer(struct coresight_device *csdev,
+ enum cs_mode mode, void *data)
+{
+ struct perf_output_handle *handle = data;
+ struct etr_perf_buffer *etr_perf;
+
+ switch (mode) {
+ case CS_MODE_SYSFS:
+ return tmc_etr_get_sysfs_buffer(csdev);
+ case CS_MODE_PERF:
+ etr_perf = etm_perf_sink_config(handle);
+ if (WARN_ON(!etr_perf || !etr_perf->etr_buf))
+ return ERR_PTR(-EINVAL);
+ return etr_perf->etr_buf;
+ default:
+ return ERR_PTR(-EINVAL);
+ }
+}
+EXPORT_SYMBOL_GPL(tmc_etr_get_buffer);
+
/*
* alloc_etr_buf: Allocate ETR buffer for use by perf.
* The size of the hardware buffer is dependent on the size configured
diff --git a/drivers/hwtracing/coresight/coresight-tmc.h b/drivers/hwtracing/coresight/coresight-tmc.h
index 01c0382a29c0..b97da39652d2 100644
--- a/drivers/hwtracing/coresight/coresight-tmc.h
+++ b/drivers/hwtracing/coresight/coresight-tmc.h
@@ -332,5 +332,7 @@ struct coresight_device *tmc_etr_get_catu_device(struct tmc_drvdata *drvdata);
void tmc_etr_set_catu_ops(const struct etr_buf_operations *catu);
void tmc_etr_remove_catu_ops(void);
+struct etr_buf *tmc_etr_get_buffer(struct coresight_device *csdev,
+ enum cs_mode mode, void *data);
#endif
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
Currently CATU is the only helper device, and its enable and disable
calls are hard coded. To allow more helper devices to be added in a
generic way, remove these hard coded calls and just enable and disable
all helper devices.
This has to apply to helpers adjacent to the path, because they will
never be in the path. CATU was already discovered in this way, so
there is no change there.
One change that is needed is for CATU to call back into ETR to allocate
the buffer. Because the enable call was previously hard coded, it was
done at a point where the buffer was already allocated, but this is no
longer the case.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
.../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
include/linux/coresight.h | 3 +-
4 files changed, 99 insertions(+), 34 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-catu.c b/drivers/hwtracing/coresight/coresight-catu.c
index bc90a03f478f..24a08a2b96b1 100644
--- a/drivers/hwtracing/coresight/coresight-catu.c
+++ b/drivers/hwtracing/coresight/coresight-catu.c
@@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct catu_drvdata *drvdata)
return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
}
-static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
+static struct coresight_device *
+catu_get_etr_device(struct coresight_device *csdev)
+{
+ int i;
+ struct coresight_device *tmp;
+
+ for (i = 0; i < csdev->pdata->nr_inconns; i++) {
+ tmp = csdev->pdata->in_conns[i].remote_dev;
+ if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
+ tmp->subtype.sink_subtype ==
+ CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
+ return tmp;
+ }
+
+ return NULL;
+}
+
+static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode cs_mode,
+ void *data)
{
int rc;
u32 control, mode;
- struct etr_buf *etr_buf = data;
+ struct etr_buf *etr_buf = NULL;
struct device *dev = &drvdata->csdev->dev;
struct coresight_device *csdev = drvdata->csdev;
+ struct coresight_device *etrdev;
if (catu_wait_for_ready(drvdata))
dev_warn(dev, "Timeout while waiting for READY\n");
@@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
if (rc)
return rc;
+ etrdev = catu_get_etr_device(csdev);
+ if (etrdev) {
+ etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
+ if (IS_ERR(etr_buf))
+ return PTR_ERR(etr_buf);
+ }
control |= BIT(CATU_CONTROL_ENABLE);
if (etr_buf && etr_buf->mode == ETR_MODE_CATU) {
@@ -441,13 +466,14 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
return 0;
}
-static int catu_enable(struct coresight_device *csdev, void *data)
+static int catu_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *data)
{
int rc;
struct catu_drvdata *catu_drvdata = csdev_to_catu_drvdata(csdev);
CS_UNLOCK(catu_drvdata->base);
- rc = catu_enable_hw(catu_drvdata, data);
+ rc = catu_enable_hw(catu_drvdata, mode, data);
CS_LOCK(catu_drvdata->base);
return rc;
}
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index a8ba7493c09a..3e6ccd9e8d4e 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -441,6 +441,34 @@ static int coresight_enable_source(struct coresight_device *csdev,
return 0;
}
+static int coresight_enable_helper(struct coresight_device *csdev,
+ enum cs_mode mode, void *sink_data)
+{
+ int ret;
+
+ if (!helper_ops(csdev)->enable)
+ return 0;
+ ret = helper_ops(csdev)->enable(csdev, mode, sink_data);
+ if (ret)
+ return ret;
+
+ csdev->enable = true;
+ return 0;
+}
+
+static void coresight_disable_helper(struct coresight_device *csdev)
+{
+ int ret;
+
+ if (!helper_ops(csdev)->disable)
+ return;
+
+ ret = helper_ops(csdev)->disable(csdev, NULL);
+ if (ret)
+ return;
+ csdev->enable = false;
+}
+
/**
* coresight_disable_source - Drop the reference count by 1 and disable
* the device if there are no users left.
@@ -460,6 +488,18 @@ static bool coresight_disable_source(struct coresight_device *csdev)
return !csdev->enable;
}
+static void coresight_disable_helpers(struct coresight_device *csdev)
+{
+ int i;
+ struct coresight_device *helper;
+
+ for (i = 0; i < csdev->pdata->nr_outconns; ++i) {
+ helper = csdev->pdata->out_conns[i].remote_dev;
+ if (helper && helper->type == CORESIGHT_DEV_TYPE_HELPER)
+ coresight_disable_helper(helper);
+ }
+}
+
/*
* coresight_disable_path_from : Disable components in the given path beyond
* @nd in the list. If @nd is NULL, all the components, except the SOURCE are
@@ -509,6 +549,9 @@ static void coresight_disable_path_from(struct list_head *path,
default:
break;
}
+
+ /* Disable all helpers adjacent along the path last */
+ coresight_disable_helpers(csdev);
}
}
@@ -518,9 +561,28 @@ void coresight_disable_path(struct list_head *path)
}
EXPORT_SYMBOL_GPL(coresight_disable_path);
-int coresight_enable_path(struct list_head *path, enum cs_mode mode, void *sink_data)
+static int coresight_enable_helpers(struct coresight_device *csdev,
+ enum cs_mode mode, void *sink_data)
{
+ int i, ret = 0;
+ struct coresight_device *helper;
+
+ for (i = 0; i < csdev->pdata->nr_outconns; ++i) {
+ helper = csdev->pdata->out_conns[i].remote_dev;
+ if (!helper || helper->type != CORESIGHT_DEV_TYPE_HELPER)
+ continue;
+ ret = coresight_enable_helper(helper, mode, sink_data);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+int coresight_enable_path(struct list_head *path, enum cs_mode mode,
+ void *sink_data)
+{
int ret = 0;
u32 type;
struct coresight_node *nd;
@@ -530,6 +592,10 @@ int coresight_enable_path(struct list_head *path, enum cs_mode mode, void *sink_
csdev = nd->csdev;
type = csdev->type;
+ /* Enable all helpers adjacent to the path first */
+ ret = coresight_enable_helpers(csdev, mode, sink_data);
+ if (ret)
+ goto err;
/*
* ETF devices are tricky... They can be a link or a sink,
* depending on how they are configured. If an ETF has been
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index cb9621003aef..d6fd44fdf836 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -791,24 +791,6 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
}
EXPORT_SYMBOL_GPL(tmc_etr_get_catu_device);
-static inline int tmc_etr_enable_catu(struct tmc_drvdata *drvdata,
- struct etr_buf *etr_buf)
-{
- struct coresight_device *catu = tmc_etr_get_catu_device(drvdata);
-
- if (catu && helper_ops(catu)->enable)
- return helper_ops(catu)->enable(catu, etr_buf);
- return 0;
-}
-
-static inline void tmc_etr_disable_catu(struct tmc_drvdata *drvdata)
-{
- struct coresight_device *catu = tmc_etr_get_catu_device(drvdata);
-
- if (catu && helper_ops(catu)->disable)
- helper_ops(catu)->disable(catu, drvdata->etr_buf);
-}
-
static const struct etr_buf_operations *etr_buf_ops[] = {
[ETR_MODE_FLAT] = &etr_flat_buf_ops,
[ETR_MODE_ETR_SG] = &etr_sg_buf_ops,
@@ -1058,13 +1040,6 @@ static int tmc_etr_enable_hw(struct tmc_drvdata *drvdata,
if (WARN_ON(drvdata->etr_buf))
return -EBUSY;
- /*
- * If this ETR is connected to a CATU, enable it before we turn
- * this on.
- */
- rc = tmc_etr_enable_catu(drvdata, etr_buf);
- if (rc)
- return rc;
rc = coresight_claim_device(drvdata->csdev);
if (!rc) {
drvdata->etr_buf = etr_buf;
@@ -1072,7 +1047,6 @@ static int tmc_etr_enable_hw(struct tmc_drvdata *drvdata,
if (rc) {
drvdata->etr_buf = NULL;
coresight_disclaim_device(drvdata->csdev);
- tmc_etr_disable_catu(drvdata);
}
}
@@ -1162,8 +1136,6 @@ static void __tmc_etr_disable_hw(struct tmc_drvdata *drvdata)
void tmc_etr_disable_hw(struct tmc_drvdata *drvdata)
{
__tmc_etr_disable_hw(drvdata);
- /* Disable CATU device if this ETR is connected to one */
- tmc_etr_disable_catu(drvdata);
coresight_disclaim_device(drvdata->csdev);
/* Reset the ETR buf used by hardware */
drvdata->etr_buf = NULL;
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index fd268b24c761..c6ee1634d813 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -372,7 +372,8 @@ struct coresight_ops_source {
* @disable : Disable the device
*/
struct coresight_ops_helper {
- int (*enable)(struct coresight_device *csdev, void *data);
+ int (*enable)(struct coresight_device *csdev, enum cs_mode mode,
+ void *data);
int (*disable)(struct coresight_device *csdev, void *data);
};
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
Currently CATU is the only helper device, and its enable and disable
calls are hard coded. To allow more helper devices to be added in a
generic way, remove these hard coded calls and just enable and disable
all helper devices.
This has to apply to helpers adjacent to the path, because they will
never be in the path. CATU was already discovered in this way, so
there is no change there.
One change that is needed is for CATU to call back into ETR to allocate
the buffer. Because the enable call was previously hard coded, it was
done at a point where the buffer was already allocated, but this is no
longer the case.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
.../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
include/linux/coresight.h | 3 +-
4 files changed, 99 insertions(+), 34 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-catu.c b/drivers/hwtracing/coresight/coresight-catu.c
index bc90a03f478f..24a08a2b96b1 100644
--- a/drivers/hwtracing/coresight/coresight-catu.c
+++ b/drivers/hwtracing/coresight/coresight-catu.c
@@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct catu_drvdata *drvdata)
return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
}
-static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
+static struct coresight_device *
+catu_get_etr_device(struct coresight_device *csdev)
+{
+ int i;
+ struct coresight_device *tmp;
+
+ for (i = 0; i < csdev->pdata->nr_inconns; i++) {
+ tmp = csdev->pdata->in_conns[i].remote_dev;
+ if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
+ tmp->subtype.sink_subtype ==
+ CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
+ return tmp;
+ }
+
+ return NULL;
+}
+
+static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode cs_mode,
+ void *data)
{
int rc;
u32 control, mode;
- struct etr_buf *etr_buf = data;
+ struct etr_buf *etr_buf = NULL;
struct device *dev = &drvdata->csdev->dev;
struct coresight_device *csdev = drvdata->csdev;
+ struct coresight_device *etrdev;
if (catu_wait_for_ready(drvdata))
dev_warn(dev, "Timeout while waiting for READY\n");
@@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
if (rc)
return rc;
+ etrdev = catu_get_etr_device(csdev);
+ if (etrdev) {
+ etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
+ if (IS_ERR(etr_buf))
+ return PTR_ERR(etr_buf);
+ }
control |= BIT(CATU_CONTROL_ENABLE);
if (etr_buf && etr_buf->mode == ETR_MODE_CATU) {
@@ -441,13 +466,14 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
return 0;
}
-static int catu_enable(struct coresight_device *csdev, void *data)
+static int catu_enable(struct coresight_device *csdev, enum cs_mode mode,
+ void *data)
{
int rc;
struct catu_drvdata *catu_drvdata = csdev_to_catu_drvdata(csdev);
CS_UNLOCK(catu_drvdata->base);
- rc = catu_enable_hw(catu_drvdata, data);
+ rc = catu_enable_hw(catu_drvdata, mode, data);
CS_LOCK(catu_drvdata->base);
return rc;
}
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index a8ba7493c09a..3e6ccd9e8d4e 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -441,6 +441,34 @@ static int coresight_enable_source(struct coresight_device *csdev,
return 0;
}
+static int coresight_enable_helper(struct coresight_device *csdev,
+ enum cs_mode mode, void *sink_data)
+{
+ int ret;
+
+ if (!helper_ops(csdev)->enable)
+ return 0;
+ ret = helper_ops(csdev)->enable(csdev, mode, sink_data);
+ if (ret)
+ return ret;
+
+ csdev->enable = true;
+ return 0;
+}
+
+static void coresight_disable_helper(struct coresight_device *csdev)
+{
+ int ret;
+
+ if (!helper_ops(csdev)->disable)
+ return;
+
+ ret = helper_ops(csdev)->disable(csdev, NULL);
+ if (ret)
+ return;
+ csdev->enable = false;
+}
+
/**
* coresight_disable_source - Drop the reference count by 1 and disable
* the device if there are no users left.
@@ -460,6 +488,18 @@ static bool coresight_disable_source(struct coresight_device *csdev)
return !csdev->enable;
}
+static void coresight_disable_helpers(struct coresight_device *csdev)
+{
+ int i;
+ struct coresight_device *helper;
+
+ for (i = 0; i < csdev->pdata->nr_outconns; ++i) {
+ helper = csdev->pdata->out_conns[i].remote_dev;
+ if (helper && helper->type == CORESIGHT_DEV_TYPE_HELPER)
+ coresight_disable_helper(helper);
+ }
+}
+
/*
* coresight_disable_path_from : Disable components in the given path beyond
* @nd in the list. If @nd is NULL, all the components, except the SOURCE are
@@ -509,6 +549,9 @@ static void coresight_disable_path_from(struct list_head *path,
default:
break;
}
+
+ /* Disable all helpers adjacent along the path last */
+ coresight_disable_helpers(csdev);
}
}
@@ -518,9 +561,28 @@ void coresight_disable_path(struct list_head *path)
}
EXPORT_SYMBOL_GPL(coresight_disable_path);
-int coresight_enable_path(struct list_head *path, enum cs_mode mode, void *sink_data)
+static int coresight_enable_helpers(struct coresight_device *csdev,
+ enum cs_mode mode, void *sink_data)
{
+ int i, ret = 0;
+ struct coresight_device *helper;
+
+ for (i = 0; i < csdev->pdata->nr_outconns; ++i) {
+ helper = csdev->pdata->out_conns[i].remote_dev;
+ if (!helper || helper->type != CORESIGHT_DEV_TYPE_HELPER)
+ continue;
+ ret = coresight_enable_helper(helper, mode, sink_data);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+int coresight_enable_path(struct list_head *path, enum cs_mode mode,
+ void *sink_data)
+{
int ret = 0;
u32 type;
struct coresight_node *nd;
@@ -530,6 +592,10 @@ int coresight_enable_path(struct list_head *path, enum cs_mode mode, void *sink_
csdev = nd->csdev;
type = csdev->type;
+ /* Enable all helpers adjacent to the path first */
+ ret = coresight_enable_helpers(csdev, mode, sink_data);
+ if (ret)
+ goto err;
/*
* ETF devices are tricky... They can be a link or a sink,
* depending on how they are configured. If an ETF has been
diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index cb9621003aef..d6fd44fdf836 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -791,24 +791,6 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
}
EXPORT_SYMBOL_GPL(tmc_etr_get_catu_device);
-static inline int tmc_etr_enable_catu(struct tmc_drvdata *drvdata,
- struct etr_buf *etr_buf)
-{
- struct coresight_device *catu = tmc_etr_get_catu_device(drvdata);
-
- if (catu && helper_ops(catu)->enable)
- return helper_ops(catu)->enable(catu, etr_buf);
- return 0;
-}
-
-static inline void tmc_etr_disable_catu(struct tmc_drvdata *drvdata)
-{
- struct coresight_device *catu = tmc_etr_get_catu_device(drvdata);
-
- if (catu && helper_ops(catu)->disable)
- helper_ops(catu)->disable(catu, drvdata->etr_buf);
-}
-
static const struct etr_buf_operations *etr_buf_ops[] = {
[ETR_MODE_FLAT] = &etr_flat_buf_ops,
[ETR_MODE_ETR_SG] = &etr_sg_buf_ops,
@@ -1058,13 +1040,6 @@ static int tmc_etr_enable_hw(struct tmc_drvdata *drvdata,
if (WARN_ON(drvdata->etr_buf))
return -EBUSY;
- /*
- * If this ETR is connected to a CATU, enable it before we turn
- * this on.
- */
- rc = tmc_etr_enable_catu(drvdata, etr_buf);
- if (rc)
- return rc;
rc = coresight_claim_device(drvdata->csdev);
if (!rc) {
drvdata->etr_buf = etr_buf;
@@ -1072,7 +1047,6 @@ static int tmc_etr_enable_hw(struct tmc_drvdata *drvdata,
if (rc) {
drvdata->etr_buf = NULL;
coresight_disclaim_device(drvdata->csdev);
- tmc_etr_disable_catu(drvdata);
}
}
@@ -1162,8 +1136,6 @@ static void __tmc_etr_disable_hw(struct tmc_drvdata *drvdata)
void tmc_etr_disable_hw(struct tmc_drvdata *drvdata)
{
__tmc_etr_disable_hw(drvdata);
- /* Disable CATU device if this ETR is connected to one */
- tmc_etr_disable_catu(drvdata);
coresight_disclaim_device(drvdata->csdev);
/* Reset the ETR buf used by hardware */
drvdata->etr_buf = NULL;
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index fd268b24c761..c6ee1634d813 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -372,7 +372,8 @@ struct coresight_ops_source {
* @disable : Disable the device
*/
struct coresight_ops_helper {
- int (*enable)(struct coresight_device *csdev, void *data);
+ int (*enable)(struct coresight_device *csdev, enum cs_mode mode,
+ void *data);
int (*disable)(struct coresight_device *csdev, void *data);
};
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 9/9] coresight: Fix CTI module refcount leak by making it a helper device
2023-03-10 16:05 ` James Clark
@ 2023-03-10 16:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
The CTI module has some hard coded refcounting code that has a leak.
For example running perf and then trying to unload it fails:
perf record -e cs_etm// -a -- ls
rmmod coresight_cti
rmmod: ERROR: Module coresight_cti is in use
The coresight core already handles references of devices in use, so by
making CTI a normal helper device, we get working refcounting for free.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 73 +++----------------
.../hwtracing/coresight/coresight-cti-core.c | 56 +++++++++-----
.../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
drivers/hwtracing/coresight/coresight-cti.h | 4 +-
include/linux/coresight.h | 28 +------
5 files changed, 53 insertions(+), 112 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index 3e6ccd9e8d4e..94d84404fb29 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -253,48 +253,6 @@ void coresight_disclaim_device(struct coresight_device *csdev)
}
EXPORT_SYMBOL_GPL(coresight_disclaim_device);
-/* enable or disable an associated CTI device of the supplied CS device */
-static int
-coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
-{
- int ect_ret = 0;
- struct coresight_device *ect_csdev = csdev->ect_dev;
- struct module *mod;
-
- if (!ect_csdev)
- return 0;
- if ((!ect_ops(ect_csdev)->enable) || (!ect_ops(ect_csdev)->disable))
- return 0;
-
- mod = ect_csdev->dev.parent->driver->owner;
- if (enable) {
- if (try_module_get(mod)) {
- ect_ret = ect_ops(ect_csdev)->enable(ect_csdev);
- if (ect_ret) {
- module_put(mod);
- } else {
- get_device(ect_csdev->dev.parent);
- csdev->ect_enabled = true;
- }
- } else
- ect_ret = -ENODEV;
- } else {
- if (csdev->ect_enabled) {
- ect_ret = ect_ops(ect_csdev)->disable(ect_csdev);
- put_device(ect_csdev->dev.parent);
- module_put(mod);
- csdev->ect_enabled = false;
- }
- }
-
- /* output warning if ECT enable is preventing trace operation */
- if (ect_ret)
- dev_info(&csdev->dev, "Associated ECT device (%s) %s failed\n",
- dev_name(&ect_csdev->dev),
- enable ? "enable" : "disable");
- return ect_ret;
-}
-
/*
* Set the associated ect / cti device while holding the coresight_mutex
* to avoid a race with coresight_enable that may try to use this value.
@@ -302,8 +260,14 @@ coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
void coresight_set_assoc_ectdev_mutex(struct coresight_device *csdev,
struct coresight_device *ect_csdev)
{
+ struct coresight_connection conn = {};
+
mutex_lock(&coresight_mutex);
- csdev->ect_dev = ect_csdev;
+ conn.remote_fwnode = fwnode_handle_get(dev_fwnode(&ect_csdev->dev));
+ conn.remote_dev = ect_csdev;
+ conn.remote_port = conn.port = -1;
+ coresight_add_conn(csdev->dev.parent, csdev->pdata, &conn);
+ coresight_fixup_inputs(csdev);
mutex_unlock(&coresight_mutex);
}
EXPORT_SYMBOL_GPL(coresight_set_assoc_ectdev_mutex);
@@ -320,12 +284,8 @@ static int coresight_enable_sink(struct coresight_device *csdev,
if (!sink_ops(csdev)->enable)
return -EINVAL;
- ret = coresight_control_assoc_ectdev(csdev, true);
- if (ret)
- return ret;
ret = sink_ops(csdev)->enable(csdev, mode, data);
if (ret) {
- coresight_control_assoc_ectdev(csdev, false);
return ret;
}
csdev->enable = true;
@@ -343,7 +303,6 @@ static void coresight_disable_sink(struct coresight_device *csdev)
ret = sink_ops(csdev)->disable(csdev);
if (ret)
return;
- coresight_control_assoc_ectdev(csdev, false);
csdev->enable = false;
}
@@ -368,17 +327,11 @@ static int coresight_enable_link(struct coresight_device *csdev,
return outport;
if (link_ops(csdev)->enable) {
- ret = coresight_control_assoc_ectdev(csdev, true);
- if (!ret) {
- ret = link_ops(csdev)->enable(csdev, inport, outport);
- if (ret)
- coresight_control_assoc_ectdev(csdev, false);
- }
+ ret = link_ops(csdev)->enable(csdev, inport, outport);
+ if (!ret)
+ csdev->enable = true;
}
- if (!ret)
- csdev->enable = true;
-
return ret;
}
@@ -407,7 +360,6 @@ static void coresight_disable_link(struct coresight_device *csdev,
if (link_ops(csdev)->disable) {
link_ops(csdev)->disable(csdev, inport, outport);
- coresight_control_assoc_ectdev(csdev, false);
}
for (i = 0; i < nr_conns; i++)
@@ -424,12 +376,8 @@ static int coresight_enable_source(struct coresight_device *csdev,
if (!csdev->enable) {
if (source_ops(csdev)->enable) {
- ret = coresight_control_assoc_ectdev(csdev, true);
- if (ret)
- return ret;
ret = source_ops(csdev)->enable(csdev, NULL, mode);
if (ret) {
- coresight_control_assoc_ectdev(csdev, false);
return ret;
}
}
@@ -482,7 +430,6 @@ static bool coresight_disable_source(struct coresight_device *csdev)
if (atomic_dec_return(csdev->refcnt) == 0) {
if (source_ops(csdev)->disable)
source_ops(csdev)->disable(csdev, NULL);
- coresight_control_assoc_ectdev(csdev, false);
csdev->enable = false;
}
return !csdev->enable;
diff --git a/drivers/hwtracing/coresight/coresight-cti-core.c b/drivers/hwtracing/coresight/coresight-cti-core.c
index 277c890a1f1f..dbce6680759f 100644
--- a/drivers/hwtracing/coresight/coresight-cti-core.c
+++ b/drivers/hwtracing/coresight/coresight-cti-core.c
@@ -555,7 +555,10 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
mutex_lock(&ect_mutex);
/* exit if current is an ECT device.*/
- if ((csdev->type == CORESIGHT_DEV_TYPE_ECT) || list_empty(&ect_net))
+ if ((csdev->type == CORESIGHT_DEV_TYPE_HELPER &&
+ csdev->subtype.helper_subtype ==
+ CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI) ||
+ list_empty(&ect_net))
goto cti_add_done;
/* if we didn't find the csdev previously we used the fwnode name */
@@ -580,6 +583,22 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
mutex_unlock(&ect_mutex);
}
+static struct coresight_device *cti__get_cti_device(struct coresight_device *csdev)
+{
+ int i;
+ struct coresight_device *tmp;
+
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
+ tmp = csdev->pdata->out_conns[i].remote_dev;
+
+ if (tmp && tmp->type == CORESIGHT_DEV_TYPE_HELPER &&
+ tmp->subtype.helper_subtype ==
+ CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI)
+ return tmp;
+ }
+ return NULL;
+}
+
/*
* Removing the associated devices is easier.
* A CTI will not have a value for csdev->ect_dev.
@@ -588,20 +607,21 @@ static void cti_remove_assoc_from_csdev(struct coresight_device *csdev)
{
struct cti_drvdata *ctidrv;
struct cti_trig_con *tc;
+ struct coresight_device *cti_csdev = cti__get_cti_device(csdev);
struct cti_device *ctidev;
+ if (!cti_csdev)
+ return;
+
mutex_lock(&ect_mutex);
- if (csdev->ect_dev) {
- ctidrv = csdev_to_cti_drvdata(csdev->ect_dev);
- ctidev = &ctidrv->ctidev;
- list_for_each_entry(tc, &ctidev->trig_cons, node) {
- if (tc->con_dev == csdev) {
- cti_remove_sysfs_link(ctidrv, tc);
- tc->con_dev = NULL;
- break;
- }
+ ctidrv = csdev_to_cti_drvdata(cti_csdev);
+ ctidev = &ctidrv->ctidev;
+ list_for_each_entry(tc, &ctidev->trig_cons, node) {
+ if (tc->con_dev == csdev) {
+ cti_remove_sysfs_link(ctidrv, tc);
+ tc->con_dev = NULL;
+ break;
}
- csdev->ect_dev = NULL;
}
mutex_unlock(&ect_mutex);
}
@@ -646,8 +666,6 @@ static void cti_remove_conn_xrefs(struct cti_drvdata *drvdata)
list_for_each_entry(tc, &ctidev->trig_cons, node) {
if (tc->con_dev) {
- coresight_set_assoc_ectdev_mutex(tc->con_dev,
- NULL);
cti_remove_sysfs_link(drvdata, tc);
tc->con_dev = NULL;
}
@@ -795,27 +813,27 @@ static void cti_pm_release(struct cti_drvdata *drvdata)
}
/** cti ect operations **/
-int cti_enable(struct coresight_device *csdev)
+int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data)
{
struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
return cti_enable_hw(drvdata);
}
-int cti_disable(struct coresight_device *csdev)
+int cti_disable(struct coresight_device *csdev, void *data)
{
struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
return cti_disable_hw(drvdata);
}
-static const struct coresight_ops_ect cti_ops_ect = {
+static const struct coresight_ops_helper cti_ops_ect = {
.enable = cti_enable,
.disable = cti_disable,
};
static const struct coresight_ops cti_ops = {
- .ect_ops = &cti_ops_ect,
+ .helper_ops = &cti_ops_ect,
};
/*
@@ -922,8 +940,8 @@ static int cti_probe(struct amba_device *adev, const struct amba_id *id)
/* set up coresight component description */
cti_desc.pdata = pdata;
- cti_desc.type = CORESIGHT_DEV_TYPE_ECT;
- cti_desc.subtype.ect_subtype = CORESIGHT_DEV_SUBTYPE_ECT_CTI;
+ cti_desc.type = CORESIGHT_DEV_TYPE_HELPER;
+ cti_desc.subtype.helper_subtype = CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI;
cti_desc.ops = &cti_ops;
cti_desc.groups = drvdata->ctidev.con_groups;
cti_desc.dev = dev;
diff --git a/drivers/hwtracing/coresight/coresight-cti-sysfs.c b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
index e528cff9d4e2..d25dd2737b49 100644
--- a/drivers/hwtracing/coresight/coresight-cti-sysfs.c
+++ b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
@@ -112,11 +112,11 @@ static ssize_t enable_store(struct device *dev,
ret = pm_runtime_resume_and_get(dev->parent);
if (ret)
return ret;
- ret = cti_enable(drvdata->csdev);
+ ret = cti_enable(drvdata->csdev, CS_MODE_SYSFS, NULL);
if (ret)
pm_runtime_put(dev->parent);
} else {
- ret = cti_disable(drvdata->csdev);
+ ret = cti_disable(drvdata->csdev, NULL);
if (!ret)
pm_runtime_put(dev->parent);
}
diff --git a/drivers/hwtracing/coresight/coresight-cti.h b/drivers/hwtracing/coresight/coresight-cti.h
index 8b106b13a244..cb9ee616d01f 100644
--- a/drivers/hwtracing/coresight/coresight-cti.h
+++ b/drivers/hwtracing/coresight/coresight-cti.h
@@ -215,8 +215,8 @@ int cti_add_connection_entry(struct device *dev, struct cti_drvdata *drvdata,
const char *assoc_dev_name);
struct cti_trig_con *cti_allocate_trig_con(struct device *dev, int in_sigs,
int out_sigs);
-int cti_enable(struct coresight_device *csdev);
-int cti_disable(struct coresight_device *csdev);
+int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data);
+int cti_disable(struct coresight_device *csdev, void *data);
void cti_write_all_hw_regs(struct cti_drvdata *drvdata);
void cti_write_intack(struct device *dev, u32 ackval);
void cti_write_single_reg(struct cti_drvdata *drvdata, int offset, u32 value);
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index c6ee1634d813..cf7a8658ee88 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -40,8 +40,7 @@ enum coresight_dev_type {
CORESIGHT_DEV_TYPE_LINK,
CORESIGHT_DEV_TYPE_LINKSINK,
CORESIGHT_DEV_TYPE_SOURCE,
- CORESIGHT_DEV_TYPE_HELPER,
- CORESIGHT_DEV_TYPE_ECT,
+ CORESIGHT_DEV_TYPE_HELPER
};
enum coresight_dev_subtype_sink {
@@ -66,12 +65,7 @@ enum coresight_dev_subtype_source {
enum coresight_dev_subtype_helper {
CORESIGHT_DEV_SUBTYPE_HELPER_CATU,
-};
-
-/* Embedded Cross Trigger (ECT) sub-types */
-enum coresight_dev_subtype_ect {
- CORESIGHT_DEV_SUBTYPE_ECT_NONE,
- CORESIGHT_DEV_SUBTYPE_ECT_CTI,
+ CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI
};
/**
@@ -84,8 +78,6 @@ enum coresight_dev_subtype_ect {
* by @coresight_dev_subtype_source.
* @helper_subtype: type of helper this component is, as defined
* by @coresight_dev_subtype_helper.
- * @ect_subtype: type of cross trigger this component is, as
- * defined by @coresight_dev_subtype_ect
*/
union coresight_dev_subtype {
/* We have some devices which acts as LINK and SINK */
@@ -95,7 +87,6 @@ union coresight_dev_subtype {
};
enum coresight_dev_subtype_source source_subtype;
enum coresight_dev_subtype_helper helper_subtype;
- enum coresight_dev_subtype_ect ect_subtype;
};
/**
@@ -264,12 +255,9 @@ struct coresight_device {
bool activated; /* true only if a sink is part of a path */
struct dev_ext_attribute *ea;
struct coresight_device *def_sink;
- /* cross trigger handling */
- struct coresight_device *ect_dev;
/* sysfs links between components */
int nr_links;
bool has_conns_grp;
- bool ect_enabled; /* true only if associated ect device is enabled */
/* system configuration and feature lists */
struct list_head feature_csdev_list;
struct list_head config_csdev_list;
@@ -377,23 +365,11 @@ struct coresight_ops_helper {
int (*disable)(struct coresight_device *csdev, void *data);
};
-/**
- * struct coresight_ops_ect - Ops for an embedded cross trigger device
- *
- * @enable : Enable the device
- * @disable : Disable the device
- */
-struct coresight_ops_ect {
- int (*enable)(struct coresight_device *csdev);
- int (*disable)(struct coresight_device *csdev);
-};
-
struct coresight_ops {
const struct coresight_ops_sink *sink_ops;
const struct coresight_ops_link *link_ops;
const struct coresight_ops_source *source_ops;
const struct coresight_ops_helper *helper_ops;
- const struct coresight_ops_ect *ect_ops;
};
#if IS_ENABLED(CONFIG_CORESIGHT)
--
2.34.1
^ permalink raw reply related [flat|nested] 74+ messages in thread
* [PATCH v2 9/9] coresight: Fix CTI module refcount leak by making it a helper device
@ 2023-03-10 16:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-10 16:06 UTC (permalink / raw)
To: coresight
Cc: James Clark, Mathieu Poirier, Suzuki K Poulose, Mike Leach,
Leo Yan, Alexander Shishkin, linux-arm-kernel, linux-kernel
The CTI module has some hard coded refcounting code that has a leak.
For example running perf and then trying to unload it fails:
perf record -e cs_etm// -a -- ls
rmmod coresight_cti
rmmod: ERROR: Module coresight_cti is in use
The coresight core already handles references of devices in use, so by
making CTI a normal helper device, we get working refcounting for free.
Signed-off-by: James Clark <james.clark@arm.com>
---
drivers/hwtracing/coresight/coresight-core.c | 73 +++----------------
.../hwtracing/coresight/coresight-cti-core.c | 56 +++++++++-----
.../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
drivers/hwtracing/coresight/coresight-cti.h | 4 +-
include/linux/coresight.h | 28 +------
5 files changed, 53 insertions(+), 112 deletions(-)
diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
index 3e6ccd9e8d4e..94d84404fb29 100644
--- a/drivers/hwtracing/coresight/coresight-core.c
+++ b/drivers/hwtracing/coresight/coresight-core.c
@@ -253,48 +253,6 @@ void coresight_disclaim_device(struct coresight_device *csdev)
}
EXPORT_SYMBOL_GPL(coresight_disclaim_device);
-/* enable or disable an associated CTI device of the supplied CS device */
-static int
-coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
-{
- int ect_ret = 0;
- struct coresight_device *ect_csdev = csdev->ect_dev;
- struct module *mod;
-
- if (!ect_csdev)
- return 0;
- if ((!ect_ops(ect_csdev)->enable) || (!ect_ops(ect_csdev)->disable))
- return 0;
-
- mod = ect_csdev->dev.parent->driver->owner;
- if (enable) {
- if (try_module_get(mod)) {
- ect_ret = ect_ops(ect_csdev)->enable(ect_csdev);
- if (ect_ret) {
- module_put(mod);
- } else {
- get_device(ect_csdev->dev.parent);
- csdev->ect_enabled = true;
- }
- } else
- ect_ret = -ENODEV;
- } else {
- if (csdev->ect_enabled) {
- ect_ret = ect_ops(ect_csdev)->disable(ect_csdev);
- put_device(ect_csdev->dev.parent);
- module_put(mod);
- csdev->ect_enabled = false;
- }
- }
-
- /* output warning if ECT enable is preventing trace operation */
- if (ect_ret)
- dev_info(&csdev->dev, "Associated ECT device (%s) %s failed\n",
- dev_name(&ect_csdev->dev),
- enable ? "enable" : "disable");
- return ect_ret;
-}
-
/*
* Set the associated ect / cti device while holding the coresight_mutex
* to avoid a race with coresight_enable that may try to use this value.
@@ -302,8 +260,14 @@ coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
void coresight_set_assoc_ectdev_mutex(struct coresight_device *csdev,
struct coresight_device *ect_csdev)
{
+ struct coresight_connection conn = {};
+
mutex_lock(&coresight_mutex);
- csdev->ect_dev = ect_csdev;
+ conn.remote_fwnode = fwnode_handle_get(dev_fwnode(&ect_csdev->dev));
+ conn.remote_dev = ect_csdev;
+ conn.remote_port = conn.port = -1;
+ coresight_add_conn(csdev->dev.parent, csdev->pdata, &conn);
+ coresight_fixup_inputs(csdev);
mutex_unlock(&coresight_mutex);
}
EXPORT_SYMBOL_GPL(coresight_set_assoc_ectdev_mutex);
@@ -320,12 +284,8 @@ static int coresight_enable_sink(struct coresight_device *csdev,
if (!sink_ops(csdev)->enable)
return -EINVAL;
- ret = coresight_control_assoc_ectdev(csdev, true);
- if (ret)
- return ret;
ret = sink_ops(csdev)->enable(csdev, mode, data);
if (ret) {
- coresight_control_assoc_ectdev(csdev, false);
return ret;
}
csdev->enable = true;
@@ -343,7 +303,6 @@ static void coresight_disable_sink(struct coresight_device *csdev)
ret = sink_ops(csdev)->disable(csdev);
if (ret)
return;
- coresight_control_assoc_ectdev(csdev, false);
csdev->enable = false;
}
@@ -368,17 +327,11 @@ static int coresight_enable_link(struct coresight_device *csdev,
return outport;
if (link_ops(csdev)->enable) {
- ret = coresight_control_assoc_ectdev(csdev, true);
- if (!ret) {
- ret = link_ops(csdev)->enable(csdev, inport, outport);
- if (ret)
- coresight_control_assoc_ectdev(csdev, false);
- }
+ ret = link_ops(csdev)->enable(csdev, inport, outport);
+ if (!ret)
+ csdev->enable = true;
}
- if (!ret)
- csdev->enable = true;
-
return ret;
}
@@ -407,7 +360,6 @@ static void coresight_disable_link(struct coresight_device *csdev,
if (link_ops(csdev)->disable) {
link_ops(csdev)->disable(csdev, inport, outport);
- coresight_control_assoc_ectdev(csdev, false);
}
for (i = 0; i < nr_conns; i++)
@@ -424,12 +376,8 @@ static int coresight_enable_source(struct coresight_device *csdev,
if (!csdev->enable) {
if (source_ops(csdev)->enable) {
- ret = coresight_control_assoc_ectdev(csdev, true);
- if (ret)
- return ret;
ret = source_ops(csdev)->enable(csdev, NULL, mode);
if (ret) {
- coresight_control_assoc_ectdev(csdev, false);
return ret;
}
}
@@ -482,7 +430,6 @@ static bool coresight_disable_source(struct coresight_device *csdev)
if (atomic_dec_return(csdev->refcnt) == 0) {
if (source_ops(csdev)->disable)
source_ops(csdev)->disable(csdev, NULL);
- coresight_control_assoc_ectdev(csdev, false);
csdev->enable = false;
}
return !csdev->enable;
diff --git a/drivers/hwtracing/coresight/coresight-cti-core.c b/drivers/hwtracing/coresight/coresight-cti-core.c
index 277c890a1f1f..dbce6680759f 100644
--- a/drivers/hwtracing/coresight/coresight-cti-core.c
+++ b/drivers/hwtracing/coresight/coresight-cti-core.c
@@ -555,7 +555,10 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
mutex_lock(&ect_mutex);
/* exit if current is an ECT device.*/
- if ((csdev->type == CORESIGHT_DEV_TYPE_ECT) || list_empty(&ect_net))
+ if ((csdev->type == CORESIGHT_DEV_TYPE_HELPER &&
+ csdev->subtype.helper_subtype ==
+ CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI) ||
+ list_empty(&ect_net))
goto cti_add_done;
/* if we didn't find the csdev previously we used the fwnode name */
@@ -580,6 +583,22 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
mutex_unlock(&ect_mutex);
}
+static struct coresight_device *cti__get_cti_device(struct coresight_device *csdev)
+{
+ int i;
+ struct coresight_device *tmp;
+
+ for (i = 0; i < csdev->pdata->nr_outconns; i++) {
+ tmp = csdev->pdata->out_conns[i].remote_dev;
+
+ if (tmp && tmp->type == CORESIGHT_DEV_TYPE_HELPER &&
+ tmp->subtype.helper_subtype ==
+ CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI)
+ return tmp;
+ }
+ return NULL;
+}
+
/*
* Removing the associated devices is easier.
* A CTI will not have a value for csdev->ect_dev.
@@ -588,20 +607,21 @@ static void cti_remove_assoc_from_csdev(struct coresight_device *csdev)
{
struct cti_drvdata *ctidrv;
struct cti_trig_con *tc;
+ struct coresight_device *cti_csdev = cti__get_cti_device(csdev);
struct cti_device *ctidev;
+ if (!cti_csdev)
+ return;
+
mutex_lock(&ect_mutex);
- if (csdev->ect_dev) {
- ctidrv = csdev_to_cti_drvdata(csdev->ect_dev);
- ctidev = &ctidrv->ctidev;
- list_for_each_entry(tc, &ctidev->trig_cons, node) {
- if (tc->con_dev == csdev) {
- cti_remove_sysfs_link(ctidrv, tc);
- tc->con_dev = NULL;
- break;
- }
+ ctidrv = csdev_to_cti_drvdata(cti_csdev);
+ ctidev = &ctidrv->ctidev;
+ list_for_each_entry(tc, &ctidev->trig_cons, node) {
+ if (tc->con_dev == csdev) {
+ cti_remove_sysfs_link(ctidrv, tc);
+ tc->con_dev = NULL;
+ break;
}
- csdev->ect_dev = NULL;
}
mutex_unlock(&ect_mutex);
}
@@ -646,8 +666,6 @@ static void cti_remove_conn_xrefs(struct cti_drvdata *drvdata)
list_for_each_entry(tc, &ctidev->trig_cons, node) {
if (tc->con_dev) {
- coresight_set_assoc_ectdev_mutex(tc->con_dev,
- NULL);
cti_remove_sysfs_link(drvdata, tc);
tc->con_dev = NULL;
}
@@ -795,27 +813,27 @@ static void cti_pm_release(struct cti_drvdata *drvdata)
}
/** cti ect operations **/
-int cti_enable(struct coresight_device *csdev)
+int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data)
{
struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
return cti_enable_hw(drvdata);
}
-int cti_disable(struct coresight_device *csdev)
+int cti_disable(struct coresight_device *csdev, void *data)
{
struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
return cti_disable_hw(drvdata);
}
-static const struct coresight_ops_ect cti_ops_ect = {
+static const struct coresight_ops_helper cti_ops_ect = {
.enable = cti_enable,
.disable = cti_disable,
};
static const struct coresight_ops cti_ops = {
- .ect_ops = &cti_ops_ect,
+ .helper_ops = &cti_ops_ect,
};
/*
@@ -922,8 +940,8 @@ static int cti_probe(struct amba_device *adev, const struct amba_id *id)
/* set up coresight component description */
cti_desc.pdata = pdata;
- cti_desc.type = CORESIGHT_DEV_TYPE_ECT;
- cti_desc.subtype.ect_subtype = CORESIGHT_DEV_SUBTYPE_ECT_CTI;
+ cti_desc.type = CORESIGHT_DEV_TYPE_HELPER;
+ cti_desc.subtype.helper_subtype = CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI;
cti_desc.ops = &cti_ops;
cti_desc.groups = drvdata->ctidev.con_groups;
cti_desc.dev = dev;
diff --git a/drivers/hwtracing/coresight/coresight-cti-sysfs.c b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
index e528cff9d4e2..d25dd2737b49 100644
--- a/drivers/hwtracing/coresight/coresight-cti-sysfs.c
+++ b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
@@ -112,11 +112,11 @@ static ssize_t enable_store(struct device *dev,
ret = pm_runtime_resume_and_get(dev->parent);
if (ret)
return ret;
- ret = cti_enable(drvdata->csdev);
+ ret = cti_enable(drvdata->csdev, CS_MODE_SYSFS, NULL);
if (ret)
pm_runtime_put(dev->parent);
} else {
- ret = cti_disable(drvdata->csdev);
+ ret = cti_disable(drvdata->csdev, NULL);
if (!ret)
pm_runtime_put(dev->parent);
}
diff --git a/drivers/hwtracing/coresight/coresight-cti.h b/drivers/hwtracing/coresight/coresight-cti.h
index 8b106b13a244..cb9ee616d01f 100644
--- a/drivers/hwtracing/coresight/coresight-cti.h
+++ b/drivers/hwtracing/coresight/coresight-cti.h
@@ -215,8 +215,8 @@ int cti_add_connection_entry(struct device *dev, struct cti_drvdata *drvdata,
const char *assoc_dev_name);
struct cti_trig_con *cti_allocate_trig_con(struct device *dev, int in_sigs,
int out_sigs);
-int cti_enable(struct coresight_device *csdev);
-int cti_disable(struct coresight_device *csdev);
+int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data);
+int cti_disable(struct coresight_device *csdev, void *data);
void cti_write_all_hw_regs(struct cti_drvdata *drvdata);
void cti_write_intack(struct device *dev, u32 ackval);
void cti_write_single_reg(struct cti_drvdata *drvdata, int offset, u32 value);
diff --git a/include/linux/coresight.h b/include/linux/coresight.h
index c6ee1634d813..cf7a8658ee88 100644
--- a/include/linux/coresight.h
+++ b/include/linux/coresight.h
@@ -40,8 +40,7 @@ enum coresight_dev_type {
CORESIGHT_DEV_TYPE_LINK,
CORESIGHT_DEV_TYPE_LINKSINK,
CORESIGHT_DEV_TYPE_SOURCE,
- CORESIGHT_DEV_TYPE_HELPER,
- CORESIGHT_DEV_TYPE_ECT,
+ CORESIGHT_DEV_TYPE_HELPER
};
enum coresight_dev_subtype_sink {
@@ -66,12 +65,7 @@ enum coresight_dev_subtype_source {
enum coresight_dev_subtype_helper {
CORESIGHT_DEV_SUBTYPE_HELPER_CATU,
-};
-
-/* Embedded Cross Trigger (ECT) sub-types */
-enum coresight_dev_subtype_ect {
- CORESIGHT_DEV_SUBTYPE_ECT_NONE,
- CORESIGHT_DEV_SUBTYPE_ECT_CTI,
+ CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI
};
/**
@@ -84,8 +78,6 @@ enum coresight_dev_subtype_ect {
* by @coresight_dev_subtype_source.
* @helper_subtype: type of helper this component is, as defined
* by @coresight_dev_subtype_helper.
- * @ect_subtype: type of cross trigger this component is, as
- * defined by @coresight_dev_subtype_ect
*/
union coresight_dev_subtype {
/* We have some devices which acts as LINK and SINK */
@@ -95,7 +87,6 @@ union coresight_dev_subtype {
};
enum coresight_dev_subtype_source source_subtype;
enum coresight_dev_subtype_helper helper_subtype;
- enum coresight_dev_subtype_ect ect_subtype;
};
/**
@@ -264,12 +255,9 @@ struct coresight_device {
bool activated; /* true only if a sink is part of a path */
struct dev_ext_attribute *ea;
struct coresight_device *def_sink;
- /* cross trigger handling */
- struct coresight_device *ect_dev;
/* sysfs links between components */
int nr_links;
bool has_conns_grp;
- bool ect_enabled; /* true only if associated ect device is enabled */
/* system configuration and feature lists */
struct list_head feature_csdev_list;
struct list_head config_csdev_list;
@@ -377,23 +365,11 @@ struct coresight_ops_helper {
int (*disable)(struct coresight_device *csdev, void *data);
};
-/**
- * struct coresight_ops_ect - Ops for an embedded cross trigger device
- *
- * @enable : Enable the device
- * @disable : Disable the device
- */
-struct coresight_ops_ect {
- int (*enable)(struct coresight_device *csdev);
- int (*disable)(struct coresight_device *csdev);
-};
-
struct coresight_ops {
const struct coresight_ops_sink *sink_ops;
const struct coresight_ops_link *link_ops;
const struct coresight_ops_source *source_ops;
const struct coresight_ops_helper *helper_ops;
- const struct coresight_ops_ect *ect_ops;
};
#if IS_ENABLED(CONFIG_CORESIGHT)
--
2.34.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
2023-03-10 16:06 ` James Clark
@ 2023-03-14 5:35 ` Jinlong Mao
-1 siblings, 0 replies; 74+ messages in thread
From: Jinlong Mao @ 2023-03-14 5:35 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 3/11/2023 12:06 AM, James Clark wrote:
> This will allow CATU to get its associated ETR in a generic way where
> currently the enable path has some hard coded searches which avoid
> the need to store input connections.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
> .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
> drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
> include/linux/coresight.h | 25 ++++++++
> 4 files changed, 130 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index f457914e445e..a8ba7493c09a 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fff
> EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
>
> static const struct cti_assoc_op *cti_assoc_ops;
> +static int coresight_fixup_inputs(struct coresight_device *csdev);
>
> ssize_t coresight_simple_show_pair(struct device *_dev,
> struct device_attribute *attr, char *buf)
> @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct coresight_device *csdev)
> csdev, coresight_orphan_match);
> }
>
> +/*
> + * Device connections are discovered before one/both devices have been created,
> + * so inputs must be added later.
> + */
> +static int coresight_fixup_inputs(struct coresight_device *csdev)
> +{
> + int i, ret = 0;
> + struct coresight_connection *out_conn;
> + struct coresight_connection in_conn;
> +
> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> + out_conn = &csdev->pdata->out_conns[i];
> + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
> + continue;
Hi James,
If out_conn->remote_dev is null here, the in_conn of
out_conn->remote_dev->pdata will never be set.
For example, device A is connected to in_port 0 of device B. If device A
is probed first, the in_conn of device
B will not be set.
Do we need to add Defer probe return here ? I tested with defer probe
return, it works.
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
out_conn = &csdev->pdata->out_conns[i];
if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
- continue;
+ return -EPROBE_DEFER;
Thanks
Jinlong Mao
> +
> + /* Reverse local/remote relationships for inputs */
> + in_conn.remote_dev = csdev;
> + in_conn.remote_port = out_conn->port;
> + in_conn.port = out_conn->remote_port;
> + in_conn.remote_fwnode = csdev->dev.fwnode;
> + ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
> + out_conn->remote_dev->pdata,
> + &in_conn);
> + if (ret)
> + return ret;
> + }
> +
> + return 0;
> +}
>
> static int coresight_fixup_device_conns(struct coresight_device *csdev)
> {
> @@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct device *dev, void *data)
> */
> fwnode_handle_put(conn->remote_fwnode);
> conn->remote_fwnode = NULL;
> + conn->remote_dev = NULL;
> + /* No need to continue */
> + break;
> + }
> + }
> + for (i = 0; i < iterator->pdata->nr_inconns; i++) {
> + conn = &iterator->pdata->in_conns[i];
> + if (csdev == conn->remote_dev) {
> + conn->remote_fwnode = NULL;
> + conn->remote_dev = NULL;
> /* No need to continue */
> break;
> }
> }
> -
> /*
> * Returning '0' ensures that all known component on the
> * bus will be checked.
> @@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct coresight_device *csdev,
>
> for (i = 0; i < pdata->nr_outconns; i++) {
> /* If we have made the links, remove them now */
> - if (csdev && conns[i].remote_dev)
> + if (csdev && conns[i].remote_dev) {
> coresight_remove_links(csdev, &conns[i]);
> + conns[i].remote_dev = NULL;
> + }
> +
> /*
> * Drop the refcount and clear the handle as this device
> * is going away
> */
> if (conns[i].remote_fwnode) {
> fwnode_handle_put(conns[i].remote_fwnode);
> - pdata->out_conns[i].remote_fwnode = NULL;
> + conns[i].remote_fwnode = NULL;
> }
> }
> + for (i = 0; i < pdata->nr_inconns; i++) {
> + pdata->in_conns[i].remote_dev = NULL;
> + pdata->in_conns[i].remote_fwnode = NULL;
> + }
> +
> if (csdev)
> coresight_remove_conns_sysfs_group(csdev);
> }
> -
> struct coresight_device *coresight_register(struct coresight_desc *desc)
> {
> int ret;
> @@ -1659,6 +1705,8 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
> ret = coresight_create_conns_sysfs_group(csdev);
> if (!ret)
> ret = coresight_fixup_device_conns(csdev);
> + if (!ret)
> + ret = coresight_fixup_inputs(csdev);
> if (!ret)
> ret = coresight_fixup_orphan_conns(csdev);
>
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index 16553f7dde12..20e3351cbdc2 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -20,8 +20,7 @@
>
> #include "coresight-priv.h"
> /*
> - * coresight_alloc_conns: Allocate connections record for each output
> - * port from the device.
> + * coresight_alloc_conns: Allocate connections record for each input/output device.
> */
> static int coresight_alloc_conns(struct device *dev,
> struct coresight_platform_data *pdata)
> @@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
> if (!pdata->out_conns)
> return -ENOMEM;
> }
> -
> + if (pdata->nr_inconns) {
> + pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
> + pdata->nr_inconns,
> + sizeof(*pdata->in_conns),
> + GFP_KERNEL | __GFP_ZERO);
> + if (!pdata->in_conns)
> + return -ENOMEM;
> + }
> return 0;
> }
>
> @@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
> }
> EXPORT_SYMBOL_GPL(coresight_add_conn);
>
> +/*
> + * Add a connection in the first free slot, or realloc
> + * if there is no space.
> + *
> + * Do nothing if the connection already exists because inputs are
> + * fixed up multiple times.
> + */
> +int coresight_add_in_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
> + struct coresight_connection *conn)
> +{
> + int ret;
> + struct coresight_connection *free_conn = NULL;
> + int i;
> +
> + /* Search for a free slot or exit if a duplicate is found */
> + if (pdata->in_conns) {
> + for (i = 0; i < pdata->nr_inconns; ++i) {
> + if (!free_conn && !pdata->in_conns[i].remote_fwnode)
> + free_conn = &pdata->in_conns[i];
> + if (pdata->in_conns[i].remote_fwnode ==
> + conn->remote_fwnode)
> + return 0;
> + }
> + }
> +
> + if (!free_conn) {
> + pdata->nr_inconns++;
> + ret = coresight_alloc_conns(dev, pdata);
> + if (ret)
> + return ret;
> + free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
> + }
> +
> + *free_conn = *conn;
> + return 0;
> +}
> +EXPORT_SYMBOL_GPL(coresight_add_in_conn);
> +
> static struct device *
> coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
> {
> @@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
>
> /*
> * of_coresight_parse_endpoint : Parse the given output endpoint @ep
> - * and fill the connection information in @conn
> + * and fill the connection information in @in_conn and @out_conn
> *
> * Parses the local port, remote device name and the remote port.
> *
> @@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct device *dev,
> /* Get the number of input and output port for this component */
> of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
>
> - /* If there are no output connections, we are done */
> - if (!pdata->nr_outconns)
> - return 0;
> -
> ret = coresight_alloc_conns(dev, pdata);
> if (ret)
> return ret;
>
> + /* If there are no output connections, we are done */
> + if (!pdata->nr_outconns)
> + return 0;
> +
> parent = of_coresight_get_output_ports_node(node);
> /*
> * If the DT uses obsoleted bindings, the ports are listed
> diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
> index 3da9868d9237..2abf9639ac0f 100644
> --- a/drivers/hwtracing/coresight/coresight-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
> @@ -202,5 +202,4 @@ void coresight_remove_links(struct coresight_device *orig,
> devm_kfree(&orig->dev, conn->link->orig_name);
> devm_kfree(&orig->dev, conn->link);
> conn->link = NULL;
> - conn->remote_dev = NULL;
> }
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index 47fa58d6981d..fd268b24c761 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -110,6 +110,7 @@ struct coresight_platform_data {
> int nr_inconns;
> int nr_outconns;
> struct coresight_connection *out_conns;
> + struct coresight_connection *in_conns;
> };
>
> /**
> @@ -177,6 +178,27 @@ struct coresight_desc {
> * @remote_fwnode once the remote's coresight_device has
> * been created.
> * @link: Representation of the connection as a sysfs link.
> + *
> + * The full connection structure looks like this, where in_conns store references to
> + * the parent device in the same remote_dev member as output connections.
> + *
> + * +-----------------------------+ +-----------------------------+
> + * |coresight_device | |coresight_connection |
> + * |-----------------------------| |-----------------------------|
> + * ---->| | | |
> + * | | | | remote_dev*|------
> + * | |pdata->out_conns[nr_outconns]|----------->| | |
> + * | | | | | |
> + * | +-----------------------------+ +-----------------------------+ |
> + * | |
> + * | +-----------------------------+ +-----------------------------+ |
> + * | |coresight_connection | |coresight_device | |
> + * | |-----------------------------| |------------------------------ |
> + * | | | | |<-----
> + * -----|remote_dev* | | |
> + * | |<-----------|pdata->in_conns[nr_inconns] |
> + * | | | |
> + * +-----------------------------+ +-----------------------------+
> */
> struct coresight_connection {
> int port;
> @@ -619,5 +641,8 @@ struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
> int coresight_add_conn(struct device *dev,
> struct coresight_platform_data *pdata,
> const struct coresight_connection *conn);
> +int coresight_add_in_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
> + struct coresight_connection *conn);
>
> #endif /* _LINUX_COREISGHT_H */
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
@ 2023-03-14 5:35 ` Jinlong Mao
0 siblings, 0 replies; 74+ messages in thread
From: Jinlong Mao @ 2023-03-14 5:35 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 3/11/2023 12:06 AM, James Clark wrote:
> This will allow CATU to get its associated ETR in a generic way where
> currently the enable path has some hard coded searches which avoid
> the need to store input connections.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
> .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
> drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
> include/linux/coresight.h | 25 ++++++++
> 4 files changed, 130 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index f457914e445e..a8ba7493c09a 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fff
> EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
>
> static const struct cti_assoc_op *cti_assoc_ops;
> +static int coresight_fixup_inputs(struct coresight_device *csdev);
>
> ssize_t coresight_simple_show_pair(struct device *_dev,
> struct device_attribute *attr, char *buf)
> @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct coresight_device *csdev)
> csdev, coresight_orphan_match);
> }
>
> +/*
> + * Device connections are discovered before one/both devices have been created,
> + * so inputs must be added later.
> + */
> +static int coresight_fixup_inputs(struct coresight_device *csdev)
> +{
> + int i, ret = 0;
> + struct coresight_connection *out_conn;
> + struct coresight_connection in_conn;
> +
> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> + out_conn = &csdev->pdata->out_conns[i];
> + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
> + continue;
Hi James,
If out_conn->remote_dev is null here, the in_conn of
out_conn->remote_dev->pdata will never be set.
For example, device A is connected to in_port 0 of device B. If device A
is probed first, the in_conn of device
B will not be set.
Do we need to add Defer probe return here ? I tested with defer probe
return, it works.
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
out_conn = &csdev->pdata->out_conns[i];
if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
- continue;
+ return -EPROBE_DEFER;
Thanks
Jinlong Mao
> +
> + /* Reverse local/remote relationships for inputs */
> + in_conn.remote_dev = csdev;
> + in_conn.remote_port = out_conn->port;
> + in_conn.port = out_conn->remote_port;
> + in_conn.remote_fwnode = csdev->dev.fwnode;
> + ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
> + out_conn->remote_dev->pdata,
> + &in_conn);
> + if (ret)
> + return ret;
> + }
> +
> + return 0;
> +}
>
> static int coresight_fixup_device_conns(struct coresight_device *csdev)
> {
> @@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct device *dev, void *data)
> */
> fwnode_handle_put(conn->remote_fwnode);
> conn->remote_fwnode = NULL;
> + conn->remote_dev = NULL;
> + /* No need to continue */
> + break;
> + }
> + }
> + for (i = 0; i < iterator->pdata->nr_inconns; i++) {
> + conn = &iterator->pdata->in_conns[i];
> + if (csdev == conn->remote_dev) {
> + conn->remote_fwnode = NULL;
> + conn->remote_dev = NULL;
> /* No need to continue */
> break;
> }
> }
> -
> /*
> * Returning '0' ensures that all known component on the
> * bus will be checked.
> @@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct coresight_device *csdev,
>
> for (i = 0; i < pdata->nr_outconns; i++) {
> /* If we have made the links, remove them now */
> - if (csdev && conns[i].remote_dev)
> + if (csdev && conns[i].remote_dev) {
> coresight_remove_links(csdev, &conns[i]);
> + conns[i].remote_dev = NULL;
> + }
> +
> /*
> * Drop the refcount and clear the handle as this device
> * is going away
> */
> if (conns[i].remote_fwnode) {
> fwnode_handle_put(conns[i].remote_fwnode);
> - pdata->out_conns[i].remote_fwnode = NULL;
> + conns[i].remote_fwnode = NULL;
> }
> }
> + for (i = 0; i < pdata->nr_inconns; i++) {
> + pdata->in_conns[i].remote_dev = NULL;
> + pdata->in_conns[i].remote_fwnode = NULL;
> + }
> +
> if (csdev)
> coresight_remove_conns_sysfs_group(csdev);
> }
> -
> struct coresight_device *coresight_register(struct coresight_desc *desc)
> {
> int ret;
> @@ -1659,6 +1705,8 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
> ret = coresight_create_conns_sysfs_group(csdev);
> if (!ret)
> ret = coresight_fixup_device_conns(csdev);
> + if (!ret)
> + ret = coresight_fixup_inputs(csdev);
> if (!ret)
> ret = coresight_fixup_orphan_conns(csdev);
>
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index 16553f7dde12..20e3351cbdc2 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -20,8 +20,7 @@
>
> #include "coresight-priv.h"
> /*
> - * coresight_alloc_conns: Allocate connections record for each output
> - * port from the device.
> + * coresight_alloc_conns: Allocate connections record for each input/output device.
> */
> static int coresight_alloc_conns(struct device *dev,
> struct coresight_platform_data *pdata)
> @@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
> if (!pdata->out_conns)
> return -ENOMEM;
> }
> -
> + if (pdata->nr_inconns) {
> + pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
> + pdata->nr_inconns,
> + sizeof(*pdata->in_conns),
> + GFP_KERNEL | __GFP_ZERO);
> + if (!pdata->in_conns)
> + return -ENOMEM;
> + }
> return 0;
> }
>
> @@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
> }
> EXPORT_SYMBOL_GPL(coresight_add_conn);
>
> +/*
> + * Add a connection in the first free slot, or realloc
> + * if there is no space.
> + *
> + * Do nothing if the connection already exists because inputs are
> + * fixed up multiple times.
> + */
> +int coresight_add_in_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
> + struct coresight_connection *conn)
> +{
> + int ret;
> + struct coresight_connection *free_conn = NULL;
> + int i;
> +
> + /* Search for a free slot or exit if a duplicate is found */
> + if (pdata->in_conns) {
> + for (i = 0; i < pdata->nr_inconns; ++i) {
> + if (!free_conn && !pdata->in_conns[i].remote_fwnode)
> + free_conn = &pdata->in_conns[i];
> + if (pdata->in_conns[i].remote_fwnode ==
> + conn->remote_fwnode)
> + return 0;
> + }
> + }
> +
> + if (!free_conn) {
> + pdata->nr_inconns++;
> + ret = coresight_alloc_conns(dev, pdata);
> + if (ret)
> + return ret;
> + free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
> + }
> +
> + *free_conn = *conn;
> + return 0;
> +}
> +EXPORT_SYMBOL_GPL(coresight_add_in_conn);
> +
> static struct device *
> coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
> {
> @@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
>
> /*
> * of_coresight_parse_endpoint : Parse the given output endpoint @ep
> - * and fill the connection information in @conn
> + * and fill the connection information in @in_conn and @out_conn
> *
> * Parses the local port, remote device name and the remote port.
> *
> @@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct device *dev,
> /* Get the number of input and output port for this component */
> of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
>
> - /* If there are no output connections, we are done */
> - if (!pdata->nr_outconns)
> - return 0;
> -
> ret = coresight_alloc_conns(dev, pdata);
> if (ret)
> return ret;
>
> + /* If there are no output connections, we are done */
> + if (!pdata->nr_outconns)
> + return 0;
> +
> parent = of_coresight_get_output_ports_node(node);
> /*
> * If the DT uses obsoleted bindings, the ports are listed
> diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
> index 3da9868d9237..2abf9639ac0f 100644
> --- a/drivers/hwtracing/coresight/coresight-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
> @@ -202,5 +202,4 @@ void coresight_remove_links(struct coresight_device *orig,
> devm_kfree(&orig->dev, conn->link->orig_name);
> devm_kfree(&orig->dev, conn->link);
> conn->link = NULL;
> - conn->remote_dev = NULL;
> }
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index 47fa58d6981d..fd268b24c761 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -110,6 +110,7 @@ struct coresight_platform_data {
> int nr_inconns;
> int nr_outconns;
> struct coresight_connection *out_conns;
> + struct coresight_connection *in_conns;
> };
>
> /**
> @@ -177,6 +178,27 @@ struct coresight_desc {
> * @remote_fwnode once the remote's coresight_device has
> * been created.
> * @link: Representation of the connection as a sysfs link.
> + *
> + * The full connection structure looks like this, where in_conns store references to
> + * the parent device in the same remote_dev member as output connections.
> + *
> + * +-----------------------------+ +-----------------------------+
> + * |coresight_device | |coresight_connection |
> + * |-----------------------------| |-----------------------------|
> + * ---->| | | |
> + * | | | | remote_dev*|------
> + * | |pdata->out_conns[nr_outconns]|----------->| | |
> + * | | | | | |
> + * | +-----------------------------+ +-----------------------------+ |
> + * | |
> + * | +-----------------------------+ +-----------------------------+ |
> + * | |coresight_connection | |coresight_device | |
> + * | |-----------------------------| |------------------------------ |
> + * | | | | |<-----
> + * -----|remote_dev* | | |
> + * | |<-----------|pdata->in_conns[nr_inconns] |
> + * | | | |
> + * +-----------------------------+ +-----------------------------+
> */
> struct coresight_connection {
> int port;
> @@ -619,5 +641,8 @@ struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
> int coresight_add_conn(struct device *dev,
> struct coresight_platform_data *pdata,
> const struct coresight_connection *conn);
> +int coresight_add_in_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
> + struct coresight_connection *conn);
>
> #endif /* _LINUX_COREISGHT_H */
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 1/9] coresight: Use enum type for cs_mode wherever possible
2023-03-10 16:06 ` James Clark
@ 2023-03-16 16:47 ` Suzuki K Poulose
-1 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 16:47 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> mode is stored as a local_t, but it is also passed around a lot as a
> plain u32, so use the correct type wherever local_t isn't currently
> used. This helps a little bit with readability.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 7 ++++---
> drivers/hwtracing/coresight/coresight-etb10.c | 3 ++-
> drivers/hwtracing/coresight/coresight-etm3x-core.c | 6 +++---
> drivers/hwtracing/coresight/coresight-etm4x-core.c | 6 +++---
> drivers/hwtracing/coresight/coresight-priv.h | 9 ++-------
> drivers/hwtracing/coresight/coresight-stm.c | 6 +++---
> drivers/hwtracing/coresight/coresight-tmc-etf.c | 2 +-
> drivers/hwtracing/coresight/coresight-tmc-etr.c | 7 +++----
> drivers/hwtracing/coresight/coresight-tpdm.c | 4 ++--
> drivers/hwtracing/coresight/coresight-tpiu.c | 3 ++-
> drivers/hwtracing/coresight/coresight-trbe.c | 3 ++-
> drivers/hwtracing/coresight/ultrasoc-smb.c | 3 ++-
> drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
> include/linux/coresight.h | 13 ++++++++++---
> 14 files changed, 40 insertions(+), 34 deletions(-)
>
Looks good to me.
Suzuki
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 1/9] coresight: Use enum type for cs_mode wherever possible
@ 2023-03-16 16:47 ` Suzuki K Poulose
0 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 16:47 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> mode is stored as a local_t, but it is also passed around a lot as a
> plain u32, so use the correct type wherever local_t isn't currently
> used. This helps a little bit with readability.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 7 ++++---
> drivers/hwtracing/coresight/coresight-etb10.c | 3 ++-
> drivers/hwtracing/coresight/coresight-etm3x-core.c | 6 +++---
> drivers/hwtracing/coresight/coresight-etm4x-core.c | 6 +++---
> drivers/hwtracing/coresight/coresight-priv.h | 9 ++-------
> drivers/hwtracing/coresight/coresight-stm.c | 6 +++---
> drivers/hwtracing/coresight/coresight-tmc-etf.c | 2 +-
> drivers/hwtracing/coresight/coresight-tmc-etr.c | 7 +++----
> drivers/hwtracing/coresight/coresight-tpdm.c | 4 ++--
> drivers/hwtracing/coresight/coresight-tpiu.c | 3 ++-
> drivers/hwtracing/coresight/coresight-trbe.c | 3 ++-
> drivers/hwtracing/coresight/ultrasoc-smb.c | 3 ++-
> drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
> include/linux/coresight.h | 13 ++++++++++---
> 14 files changed, 40 insertions(+), 34 deletions(-)
>
Looks good to me.
Suzuki
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 5/9] coresight: Dynamically add connections
2023-03-10 16:06 ` James Clark
@ 2023-03-16 17:12 ` Suzuki K Poulose
-1 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 17:12 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> Add a function for adding connections dynamically. This also removes
> the 1:1 mapping between port number and the index into the connections
> array. The only place this mapping was used was in the warning for
> duplicate output ports, which has been replaced by a search. Other
> uses of the port number already use the port member variable.
>
> Being able to dynamically add connections will allow other devices like
> CTI to re-use the connection mechanism despite not having explicit
> connections described in the DT.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> .../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
> include/linux/coresight.h | 7 +-
> 2 files changed, 64 insertions(+), 20 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index c77238cdf448..16553f7dde12 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
> struct coresight_platform_data *pdata)
> {
> if (pdata->nr_outconns) {
> - pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
> - sizeof(*pdata->out_conns), GFP_KERNEL);
> + pdata->out_conns = devm_krealloc_array(
> + dev, pdata->out_conns, pdata->nr_outconns,
super minor nit:
pdata->out_conns = devm_krealloc_array(dev,
> + sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
> if (!pdata->out_conns)
> return -ENOMEM;
> }
> @@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
> return 0;
> }
>
> +/*
> + * Add a connection in the first free slot, or realloc
> + * if there is no space. @conn's contents is copied into the new slot.
> + *
> + * If the output port is already assigned on this device, return -EINVAL
> + */
> +int coresight_add_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
> + const struct coresight_connection *conn)
> +{
> + int ret;
> + struct coresight_connection *free_conn = NULL;
> + struct coresight_connection *i;
> +
> + /*
> + * Search for a free slot, and while looking for one, warn
> + * on any existing duplicate output port.
> + */
> + for (i = pdata->out_conns; i < pdata->out_conns + pdata->nr_outconns;
> + ++i) {
minor nit: I see why you have gone against using "i" as index into
the array. But I think having that as the index is still better
readable.
for (i = 0; i < pdata->nr_outconns; i++) {
struct coresight_connection *c = &pdata->out_conns[i];
> + if (i->remote_fwnode && conn->port != -1 &&
> + i->port == conn->port) {
> + dev_warn(dev, "Duplicate output port %d\n", i->port);
> + return -EINVAL;
> + }
> + if (!i->remote_fwnode && !free_conn)
> + free_conn = i;
> + }
> +
> + if (!free_conn) {
and:
/* No free slots */
if (i == pdata->nr_outconns) {
> + pdata->nr_outconns++;
> + ret = coresight_alloc_conns(dev, pdata);
> + if (ret)
> + return ret;
> + free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
> + }
> +
and:
pdata->out_conns[i] = *conn;
Otherwise looks good to me.
Suzuki
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 5/9] coresight: Dynamically add connections
@ 2023-03-16 17:12 ` Suzuki K Poulose
0 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 17:12 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> Add a function for adding connections dynamically. This also removes
> the 1:1 mapping between port number and the index into the connections
> array. The only place this mapping was used was in the warning for
> duplicate output ports, which has been replaced by a search. Other
> uses of the port number already use the port member variable.
>
> Being able to dynamically add connections will allow other devices like
> CTI to re-use the connection mechanism despite not having explicit
> connections described in the DT.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> .../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
> include/linux/coresight.h | 7 +-
> 2 files changed, 64 insertions(+), 20 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index c77238cdf448..16553f7dde12 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
> struct coresight_platform_data *pdata)
> {
> if (pdata->nr_outconns) {
> - pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
> - sizeof(*pdata->out_conns), GFP_KERNEL);
> + pdata->out_conns = devm_krealloc_array(
> + dev, pdata->out_conns, pdata->nr_outconns,
super minor nit:
pdata->out_conns = devm_krealloc_array(dev,
> + sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
> if (!pdata->out_conns)
> return -ENOMEM;
> }
> @@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
> return 0;
> }
>
> +/*
> + * Add a connection in the first free slot, or realloc
> + * if there is no space. @conn's contents is copied into the new slot.
> + *
> + * If the output port is already assigned on this device, return -EINVAL
> + */
> +int coresight_add_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
> + const struct coresight_connection *conn)
> +{
> + int ret;
> + struct coresight_connection *free_conn = NULL;
> + struct coresight_connection *i;
> +
> + /*
> + * Search for a free slot, and while looking for one, warn
> + * on any existing duplicate output port.
> + */
> + for (i = pdata->out_conns; i < pdata->out_conns + pdata->nr_outconns;
> + ++i) {
minor nit: I see why you have gone against using "i" as index into
the array. But I think having that as the index is still better
readable.
for (i = 0; i < pdata->nr_outconns; i++) {
struct coresight_connection *c = &pdata->out_conns[i];
> + if (i->remote_fwnode && conn->port != -1 &&
> + i->port == conn->port) {
> + dev_warn(dev, "Duplicate output port %d\n", i->port);
> + return -EINVAL;
> + }
> + if (!i->remote_fwnode && !free_conn)
> + free_conn = i;
> + }
> +
> + if (!free_conn) {
and:
/* No free slots */
if (i == pdata->nr_outconns) {
> + pdata->nr_outconns++;
> + ret = coresight_alloc_conns(dev, pdata);
> + if (ret)
> + return ret;
> + free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
> + }
> +
and:
pdata->out_conns[i] = *conn;
Otherwise looks good to me.
Suzuki
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 2/9] coresight: Change name of pdata->conns
2023-03-10 16:06 ` James Clark
@ 2023-03-16 17:17 ` Suzuki K Poulose
-1 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 17:17 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> conns is actually for output connections. Change the name to make it
> clearer and so that we can add input connections later.
>
> No functional changes.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 26 +++++++++----------
> .../hwtracing/coresight/coresight-platform.c | 12 ++++-----
> .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> include/linux/coresight.h | 2 +-
> 4 files changed, 21 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index e3b63fd52b9c..c0eda7407fb8 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -119,7 +119,7 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
> struct coresight_connection *conn;
>
> for (i = 0; i < parent->pdata->nr_outport; i++) {
> - conn = &parent->pdata->conns[i];
> + conn = &parent->pdata->out_conns[i];
> if (conn->child_dev == csdev)
> return conn->child_port;
> }
> @@ -137,7 +137,7 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
> struct coresight_connection *conn;
>
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> - conn = &csdev->pdata->conns[i];
> + conn = &csdev->pdata->out_conns[i];
> if (conn->child_dev == child)
> return conn->outport;
> }
> @@ -606,7 +606,7 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> struct coresight_device *child_dev;
>
> - child_dev = csdev->pdata->conns[i].child_dev;
> + child_dev = csdev->pdata->out_conns[i].child_dev;
> if (child_dev)
> sink = coresight_find_enabled_sink(child_dev);
> if (sink)
> @@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> struct coresight_device *child;
>
> - child = csdev->pdata->conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].child_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> if (!coresight_get_ref(child))
> goto err;
> @@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> for (i--; i >= 0; i--) {
> struct coresight_device *child;
>
> - child = csdev->pdata->conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].child_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> coresight_put_ref(child);
> }
> @@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> struct coresight_device *child;
>
> - child = csdev->pdata->conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].child_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> coresight_put_ref(child);
> }
> @@ -794,7 +794,7 @@ static int _coresight_build_path(struct coresight_device *csdev,
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> struct coresight_device *child_dev;
>
> - child_dev = csdev->pdata->conns[i].child_dev;
> + child_dev = csdev->pdata->out_conns[i].child_dev;
> if (child_dev &&
> _coresight_build_path(child_dev, sink, path) == 0) {
> found = true;
> @@ -964,7 +964,7 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
> struct coresight_device *child_dev, *sink = NULL;
> int child_depth = curr_depth;
>
> - child_dev = csdev->pdata->conns[i].child_dev;
> + child_dev = csdev->pdata->out_conns[i].child_dev;
> if (child_dev)
> sink = coresight_find_sink(child_dev, &child_depth);
>
> @@ -1334,7 +1334,7 @@ static int coresight_orphan_match(struct device *dev, void *data)
> * an orphan connection whose name matches @csdev, link it.
> */
> for (i = 0; i < i_csdev->pdata->nr_outport; i++) {
> - conn = &i_csdev->pdata->conns[i];
> + conn = &i_csdev->pdata->out_conns[i];
>
> /* Skip the port if FW doesn't describe it */
> if (!conn->child_fwnode)
> @@ -1375,7 +1375,7 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
> int i, ret = 0;
>
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> - struct coresight_connection *conn = &csdev->pdata->conns[i];
> + struct coresight_connection *conn = &csdev->pdata->out_conns[i];
>
> if (!conn->child_fwnode)
> continue;
> @@ -1412,7 +1412,7 @@ static int coresight_remove_match(struct device *dev, void *data)
> * a connection whose name matches @csdev, remove it.
> */
> for (i = 0; i < iterator->pdata->nr_outport; i++) {
> - conn = &iterator->pdata->conns[i];
> + conn = &iterator->pdata->out_conns[i];
>
> if (conn->child_dev == NULL || conn->child_fwnode == NULL)
> continue;
> @@ -1548,7 +1548,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> struct coresight_platform_data *pdata)
> {
> int i;
> - struct coresight_connection *conns = pdata->conns;
> + struct coresight_connection *conns = pdata->out_conns;
>
> for (i = 0; i < pdata->nr_outport; i++) {
> /* If we have made the links, remove them now */
> @@ -1560,7 +1560,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> */
> if (conns[i].child_fwnode) {
> fwnode_handle_put(conns[i].child_fwnode);
> - pdata->conns[i].child_fwnode = NULL;
> + pdata->out_conns[i].child_fwnode = NULL;
> }
> }
> if (csdev)
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index 475899714104..5085525a32bb 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -27,9 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
> struct coresight_platform_data *pdata)
> {
> if (pdata->nr_outport) {
> - pdata->conns = devm_kcalloc(dev, pdata->nr_outport,
> - sizeof(*pdata->conns), GFP_KERNEL);
> - if (!pdata->conns)
> + pdata->out_conns = devm_kcalloc(dev, pdata->nr_outport,
> + sizeof(*pdata->out_conns), GFP_KERNEL);
> + if (!pdata->out_conns)
> return -ENOMEM;
> }
>
> @@ -251,7 +251,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
> break;
> }
>
> - conn = &pdata->conns[endpoint.port];
> + conn = &pdata->out_conns[endpoint.port];
> if (conn->child_fwnode) {
> dev_warn(dev, "Duplicate output port %d\n",
> endpoint.port);
> @@ -744,8 +744,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
> int port = conns[i].outport;
>
> /* Duplicate output port */
> - WARN_ON(pdata->conns[port].child_fwnode);
> - pdata->conns[port] = conns[i];
> + WARN_ON(pdata->out_conns[port].child_fwnode);
> + pdata->out_conns[port] = conns[i];
> }
>
> devm_kfree(&adev->dev, conns);
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index ed589cfff1b5..86d4a08aa833 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> @@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
> return NULL;
>
> for (i = 0; i < etr->pdata->nr_outport; i++) {
> - tmp = etr->pdata->conns[i].child_dev;
> + tmp = etr->pdata->out_conns[i].child_dev;
> if (tmp && coresight_is_catu_device(tmp))
> return tmp;
> }
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index 1f878e8ed8c4..322d7273e122 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -109,7 +109,7 @@ union coresight_dev_subtype {
> struct coresight_platform_data {
> int nr_inport;
> int nr_outport;
> - struct coresight_connection *conns;
> + struct coresight_connection *out_conns;
The comment above this structure might need to reflect the name change.
Otherwise looks good to me.
Suzuki
> };
>
> /**
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 2/9] coresight: Change name of pdata->conns
@ 2023-03-16 17:17 ` Suzuki K Poulose
0 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 17:17 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> conns is actually for output connections. Change the name to make it
> clearer and so that we can add input connections later.
>
> No functional changes.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 26 +++++++++----------
> .../hwtracing/coresight/coresight-platform.c | 12 ++++-----
> .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> include/linux/coresight.h | 2 +-
> 4 files changed, 21 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index e3b63fd52b9c..c0eda7407fb8 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -119,7 +119,7 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
> struct coresight_connection *conn;
>
> for (i = 0; i < parent->pdata->nr_outport; i++) {
> - conn = &parent->pdata->conns[i];
> + conn = &parent->pdata->out_conns[i];
> if (conn->child_dev == csdev)
> return conn->child_port;
> }
> @@ -137,7 +137,7 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
> struct coresight_connection *conn;
>
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> - conn = &csdev->pdata->conns[i];
> + conn = &csdev->pdata->out_conns[i];
> if (conn->child_dev == child)
> return conn->outport;
> }
> @@ -606,7 +606,7 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> struct coresight_device *child_dev;
>
> - child_dev = csdev->pdata->conns[i].child_dev;
> + child_dev = csdev->pdata->out_conns[i].child_dev;
> if (child_dev)
> sink = coresight_find_enabled_sink(child_dev);
> if (sink)
> @@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> struct coresight_device *child;
>
> - child = csdev->pdata->conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].child_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> if (!coresight_get_ref(child))
> goto err;
> @@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> for (i--; i >= 0; i--) {
> struct coresight_device *child;
>
> - child = csdev->pdata->conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].child_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> coresight_put_ref(child);
> }
> @@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> struct coresight_device *child;
>
> - child = csdev->pdata->conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].child_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> coresight_put_ref(child);
> }
> @@ -794,7 +794,7 @@ static int _coresight_build_path(struct coresight_device *csdev,
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> struct coresight_device *child_dev;
>
> - child_dev = csdev->pdata->conns[i].child_dev;
> + child_dev = csdev->pdata->out_conns[i].child_dev;
> if (child_dev &&
> _coresight_build_path(child_dev, sink, path) == 0) {
> found = true;
> @@ -964,7 +964,7 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
> struct coresight_device *child_dev, *sink = NULL;
> int child_depth = curr_depth;
>
> - child_dev = csdev->pdata->conns[i].child_dev;
> + child_dev = csdev->pdata->out_conns[i].child_dev;
> if (child_dev)
> sink = coresight_find_sink(child_dev, &child_depth);
>
> @@ -1334,7 +1334,7 @@ static int coresight_orphan_match(struct device *dev, void *data)
> * an orphan connection whose name matches @csdev, link it.
> */
> for (i = 0; i < i_csdev->pdata->nr_outport; i++) {
> - conn = &i_csdev->pdata->conns[i];
> + conn = &i_csdev->pdata->out_conns[i];
>
> /* Skip the port if FW doesn't describe it */
> if (!conn->child_fwnode)
> @@ -1375,7 +1375,7 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
> int i, ret = 0;
>
> for (i = 0; i < csdev->pdata->nr_outport; i++) {
> - struct coresight_connection *conn = &csdev->pdata->conns[i];
> + struct coresight_connection *conn = &csdev->pdata->out_conns[i];
>
> if (!conn->child_fwnode)
> continue;
> @@ -1412,7 +1412,7 @@ static int coresight_remove_match(struct device *dev, void *data)
> * a connection whose name matches @csdev, remove it.
> */
> for (i = 0; i < iterator->pdata->nr_outport; i++) {
> - conn = &iterator->pdata->conns[i];
> + conn = &iterator->pdata->out_conns[i];
>
> if (conn->child_dev == NULL || conn->child_fwnode == NULL)
> continue;
> @@ -1548,7 +1548,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> struct coresight_platform_data *pdata)
> {
> int i;
> - struct coresight_connection *conns = pdata->conns;
> + struct coresight_connection *conns = pdata->out_conns;
>
> for (i = 0; i < pdata->nr_outport; i++) {
> /* If we have made the links, remove them now */
> @@ -1560,7 +1560,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> */
> if (conns[i].child_fwnode) {
> fwnode_handle_put(conns[i].child_fwnode);
> - pdata->conns[i].child_fwnode = NULL;
> + pdata->out_conns[i].child_fwnode = NULL;
> }
> }
> if (csdev)
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index 475899714104..5085525a32bb 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -27,9 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
> struct coresight_platform_data *pdata)
> {
> if (pdata->nr_outport) {
> - pdata->conns = devm_kcalloc(dev, pdata->nr_outport,
> - sizeof(*pdata->conns), GFP_KERNEL);
> - if (!pdata->conns)
> + pdata->out_conns = devm_kcalloc(dev, pdata->nr_outport,
> + sizeof(*pdata->out_conns), GFP_KERNEL);
> + if (!pdata->out_conns)
> return -ENOMEM;
> }
>
> @@ -251,7 +251,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
> break;
> }
>
> - conn = &pdata->conns[endpoint.port];
> + conn = &pdata->out_conns[endpoint.port];
> if (conn->child_fwnode) {
> dev_warn(dev, "Duplicate output port %d\n",
> endpoint.port);
> @@ -744,8 +744,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
> int port = conns[i].outport;
>
> /* Duplicate output port */
> - WARN_ON(pdata->conns[port].child_fwnode);
> - pdata->conns[port] = conns[i];
> + WARN_ON(pdata->out_conns[port].child_fwnode);
> + pdata->out_conns[port] = conns[i];
> }
>
> devm_kfree(&adev->dev, conns);
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index ed589cfff1b5..86d4a08aa833 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> @@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
> return NULL;
>
> for (i = 0; i < etr->pdata->nr_outport; i++) {
> - tmp = etr->pdata->conns[i].child_dev;
> + tmp = etr->pdata->out_conns[i].child_dev;
> if (tmp && coresight_is_catu_device(tmp))
> return tmp;
> }
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index 1f878e8ed8c4..322d7273e122 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -109,7 +109,7 @@ union coresight_dev_subtype {
> struct coresight_platform_data {
> int nr_inport;
> int nr_outport;
> - struct coresight_connection *conns;
> + struct coresight_connection *out_conns;
The comment above this structure might need to reflect the name change.
Otherwise looks good to me.
Suzuki
> };
>
> /**
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 3/9] coresight: Rename nr_outports to nr_outconns
2023-03-10 16:06 ` James Clark
@ 2023-03-16 17:18 ` Suzuki K Poulose
-1 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 17:18 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> Rename to avoid confusion between port number and the index in the
> connection array. The port number is already stored in the connection,
> and in a later commit the connection array will be appended to, so
> the length of it will no longer reflect the number of ports.
>
> No functional changes.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 32 ++++++++---------
> .../hwtracing/coresight/coresight-platform.c | 34 +++++++++----------
> .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> include/linux/coresight.h | 10 +++---
> 4 files changed, 39 insertions(+), 39 deletions(-)
>
...
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index 322d7273e122..cdf0d1def778 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -102,13 +102,13 @@ union coresight_dev_subtype {
> * struct coresight_platform_data - data harvested from the firmware
> * specification.
> *
> - * @nr_inport: Number of elements for the input connections.
> - * @nr_outport: Number of elements for the output connections.
> - * @conns: Sparse array of nr_outport connections from this component.
> + * @nr_inconns: Number of elements for the input connections.
> + * @nr_outconns: Number of elements for the output connections.
> + * @conns: Sparse array of nr_outconns connections from this component.
s/conns/out_conns/
Otherwise looks good to me.
Suzuki
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 3/9] coresight: Rename nr_outports to nr_outconns
@ 2023-03-16 17:18 ` Suzuki K Poulose
0 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 17:18 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> Rename to avoid confusion between port number and the index in the
> connection array. The port number is already stored in the connection,
> and in a later commit the connection array will be appended to, so
> the length of it will no longer reflect the number of ports.
>
> No functional changes.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 32 ++++++++---------
> .../hwtracing/coresight/coresight-platform.c | 34 +++++++++----------
> .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> include/linux/coresight.h | 10 +++---
> 4 files changed, 39 insertions(+), 39 deletions(-)
>
...
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index 322d7273e122..cdf0d1def778 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -102,13 +102,13 @@ union coresight_dev_subtype {
> * struct coresight_platform_data - data harvested from the firmware
> * specification.
> *
> - * @nr_inport: Number of elements for the input connections.
> - * @nr_outport: Number of elements for the output connections.
> - * @conns: Sparse array of nr_outport connections from this component.
> + * @nr_inconns: Number of elements for the input connections.
> + * @nr_outconns: Number of elements for the output connections.
> + * @conns: Sparse array of nr_outconns connections from this component.
s/conns/out_conns/
Otherwise looks good to me.
Suzuki
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
2023-03-10 16:06 ` James Clark
@ 2023-03-16 20:23 ` Suzuki K Poulose
-1 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 20:23 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> This will allow CATU to get its associated ETR in a generic way where
> currently the enable path has some hard coded searches which avoid
> the need to store input connections.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
> .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
> drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
> include/linux/coresight.h | 25 ++++++++
> 4 files changed, 130 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index f457914e445e..a8ba7493c09a 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fff
> EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
>
> static const struct cti_assoc_op *cti_assoc_ops;
> +static int coresight_fixup_inputs(struct coresight_device *csdev);
>
> ssize_t coresight_simple_show_pair(struct device *_dev,
> struct device_attribute *attr, char *buf)
> @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct coresight_device *csdev)
> csdev, coresight_orphan_match);
> }
>
> +/*
> + * Device connections are discovered before one/both devices have been created,
> + * so inputs must be added later.
> + */
> +static int coresight_fixup_inputs(struct coresight_device *csdev)
> +{
> + int i, ret = 0;
> + struct coresight_connection *out_conn;
> + struct coresight_connection in_conn;
> +
> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> + out_conn = &csdev->pdata->out_conns[i];
> + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
> + continue;
> +
> + /* Reverse local/remote relationships for inputs */
> + in_conn.remote_dev = csdev;
> + in_conn.remote_port = out_conn->port;
> + in_conn.port = out_conn->remote_port;
> + in_conn.remote_fwnode = csdev->dev.fwnode;
> + ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
> + out_conn->remote_dev->pdata,
> + &in_conn);
> + if (ret)
> + return ret;
> + }
> +
> + return 0;
> +}
>
> static int coresight_fixup_device_conns(struct coresight_device *csdev)
> {
> @@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct device *dev, void *data)
Removing the connection should be much simpler now with the inports
connection tracking.
i.e., coresight_remove_conns() need not iterate over all the devices on
the coresight bus. We had to do that to find all the devices connecting
to the removed device.
Instead with the in_conns we could readily get the csdev instance and
remove the connection from those devices.
Similarly for the out_conns can give you the remove device for removing
the in_conns on that device.
i.e,
static void coresight_remove_conns(csdev)
{
struct coresight_device *remote;
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
remote = csdev->data->out_conns[i].remote_dev;
if (!remote)
continue;
for (j = 0; j < remote->pdata->nr_inconns; j++) {
if (remote->pdata->in_conns[i].remote_dev == csdev)
/* Clear in_conns[i] on remote */
}
}
Similarly for in_conns
for (i = 0; i < csdev->pdata->nr_inconns; i++) {
remote = csdev->data->in_conns[i].remote_dev;
if (!remote)
continue;
for (j = 0; j < remote->pdata->nr_outconns; j++) {
if (remote->pdata->out_conns[i].remote_dev == csdev)
/* Clear out_conns[i] on remote */
}
> */
> fwnode_handle_put(conn->remote_fwnode);
> conn->remote_fwnode = NULL;
> + conn->remote_dev = NULL;
> + /* No need to continue */
> + break;
> + }
> + }
> + for (i = 0; i < iterator->pdata->nr_inconns; i++) {
> + conn = &iterator->pdata->in_conns[i];
> + if (csdev == conn->remote_dev) {
> + conn->remote_fwnode = NULL;
> + conn->remote_dev = NULL;
> /* No need to continue */
> break;
> }
> }
> -
> /*
> * Returning '0' ensures that all known component on the
> * bus will be checked.
> @@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct coresight_device *csdev,
>
> for (i = 0; i < pdata->nr_outconns; i++) {
> /* If we have made the links, remove them now */
> - if (csdev && conns[i].remote_dev)
> + if (csdev && conns[i].remote_dev) {
> coresight_remove_links(csdev, &conns[i]);
> + conns[i].remote_dev = NULL;
> + }
> +
> /*
> * Drop the refcount and clear the handle as this device
> * is going away
> */
> if (conns[i].remote_fwnode) {
> fwnode_handle_put(conns[i].remote_fwnode);
> - pdata->out_conns[i].remote_fwnode = NULL;
> + conns[i].remote_fwnode = NULL;
> }
> }
> + for (i = 0; i < pdata->nr_inconns; i++) {
> + pdata->in_conns[i].remote_dev = NULL;
> + pdata->in_conns[i].remote_fwnode = NULL;
> + }
> +
> if (csdev)
> coresight_remove_conns_sysfs_group(csdev);
> }
> -
> struct coresight_device *coresight_register(struct coresight_desc *desc)
> {
> int ret;
> @@ -1659,6 +1705,8 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
> ret = coresight_create_conns_sysfs_group(csdev);
> if (!ret)
> ret = coresight_fixup_device_conns(csdev);
> + if (!ret)
> + ret = coresight_fixup_inputs(csdev);
Don't we also need to fixup the "input" connections on this
device ? It would be good move all of this connection stuff
into a wrapper function.
> if (!ret)
> ret = coresight_fixup_orphan_conns(csdev);
>
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index 16553f7dde12..20e3351cbdc2 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -20,8 +20,7 @@
>
> #include "coresight-priv.h"
> /*
> - * coresight_alloc_conns: Allocate connections record for each output
> - * port from the device.
> + * coresight_alloc_conns: Allocate connections record for each input/output device.
> */
> static int coresight_alloc_conns(struct device *dev,
> struct coresight_platform_data *pdata)
> @@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
> if (!pdata->out_conns)
> return -ENOMEM;
> }
> -
> + if (pdata->nr_inconns) {
> + pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
> + pdata->nr_inconns,
> + sizeof(*pdata->in_conns),
> + GFP_KERNEL | __GFP_ZERO);
> + if (!pdata->in_conns)
> + return -ENOMEM;
> + }
devm_krealloc_array() skips any allocation if the requested new_size <=
origin_size. So this should be fine for the out_conns, to not realloc
the space again unnecessarily.
> return 0;
> }
>
> @@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
> }
> EXPORT_SYMBOL_GPL(coresight_add_conn);
Should this be named as add_out_conn in the previous patch
just to make it explicit and easier to read.
>
> +/*
> + * Add a connection in the first free slot, or realloc
> + * if there is no space.
> + *
> + * Do nothing if the connection already exists because inputs are
> + * fixed up multiple times.
> + */
> +int coresight_add_in_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
minor nit: Is there any reason why this can't simply be:
int coresight_add_in_conn(struct coresight_device *csdev,
struct coresight_connection *conn)
{
struct device *dev = csdev->dev.parent;
struct coresight_platform_data *pdata = csdev->pdata;
> + struct coresight_connection *conn)
Suzuki
> +{
> + int ret;
> + struct coresight_connection *free_conn = NULL;
> + int i;
> +
> + /* Search for a free slot or exit if a duplicate is found */
> + if (pdata->in_conns) {
> + for (i = 0; i < pdata->nr_inconns; ++i) {
> + if (!free_conn && !pdata->in_conns[i].remote_fwnode)
> + free_conn = &pdata->in_conns[i];
> + if (pdata->in_conns[i].remote_fwnode ==
> + conn->remote_fwnode)
> + return 0;
> + }
> + }
> +
> + if (!free_conn) {
> + pdata->nr_inconns++;
> + ret = coresight_alloc_conns(dev, pdata);
> + if (ret)
> + return ret;
> + free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
> + }
> +
> + *free_conn = *conn;
> + return 0;
> +}
> +EXPORT_SYMBOL_GPL(coresight_add_in_conn);
> +
> static struct device *
> coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
> {
> @@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
>
> /*
> * of_coresight_parse_endpoint : Parse the given output endpoint @ep
> - * and fill the connection information in @conn
> + * and fill the connection information in @in_conn and @out_conn
> *
> * Parses the local port, remote device name and the remote port.
> *
> @@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct device *dev,
> /* Get the number of input and output port for this component */
> of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
>
> - /* If there are no output connections, we are done */
> - if (!pdata->nr_outconns)
> - return 0;
> -
> ret = coresight_alloc_conns(dev, pdata);
> if (ret)
> return ret;
>
> + /* If there are no output connections, we are done */
> + if (!pdata->nr_outconns)
> + return 0;
> +
> parent = of_coresight_get_output_ports_node(node);
> /*
> * If the DT uses obsoleted bindings, the ports are listed
> diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
> index 3da9868d9237..2abf9639ac0f 100644
> --- a/drivers/hwtracing/coresight/coresight-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
> @@ -202,5 +202,4 @@ void coresight_remove_links(struct coresight_device *orig,
> devm_kfree(&orig->dev, conn->link->orig_name);
> devm_kfree(&orig->dev, conn->link);
> conn->link = NULL;
> - conn->remote_dev = NULL;
> }
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index 47fa58d6981d..fd268b24c761 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -110,6 +110,7 @@ struct coresight_platform_data {
> int nr_inconns;
> int nr_outconns;
> struct coresight_connection *out_conns;
> + struct coresight_connection *in_conns;
Please document the field above the structure.
Suzuki
> };
>
> /**
> @@ -177,6 +178,27 @@ struct coresight_desc {
> * @remote_fwnode once the remote's coresight_device has
> * been created.
> * @link: Representation of the connection as a sysfs link.
> + *
> + * The full connection structure looks like this, where in_conns store references to
> + * the parent device in the same remote_dev member as output connections.
> + *
> + * +-----------------------------+ +-----------------------------+
> + * |coresight_device | |coresight_connection |
> + * |-----------------------------| |-----------------------------|
> + * ---->| | | |
> + * | | | | remote_dev*|------
> + * | |pdata->out_conns[nr_outconns]|----------->| | |
> + * | | | | | |
> + * | +-----------------------------+ +-----------------------------+ |
> + * | |
> + * | +-----------------------------+ +-----------------------------+ |
> + * | |coresight_connection | |coresight_device | |
> + * | |-----------------------------| |------------------------------ |
> + * | | | | |<-----
> + * -----|remote_dev* | | |
> + * | |<-----------|pdata->in_conns[nr_inconns] |
> + * | | | |
> + * +-----------------------------+ +-----------------------------+
> */
> struct coresight_connection {
> int port;
> @@ -619,5 +641,8 @@ struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
> int coresight_add_conn(struct device *dev,
> struct coresight_platform_data *pdata,
> const struct coresight_connection *conn);
> +int coresight_add_in_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
> + struct coresight_connection *conn);
>
> #endif /* _LINUX_COREISGHT_H */
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
@ 2023-03-16 20:23 ` Suzuki K Poulose
0 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-16 20:23 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> This will allow CATU to get its associated ETR in a generic way where
> currently the enable path has some hard coded searches which avoid
> the need to store input connections.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
> .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
> drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
> include/linux/coresight.h | 25 ++++++++
> 4 files changed, 130 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index f457914e445e..a8ba7493c09a 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fff
> EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
>
> static const struct cti_assoc_op *cti_assoc_ops;
> +static int coresight_fixup_inputs(struct coresight_device *csdev);
>
> ssize_t coresight_simple_show_pair(struct device *_dev,
> struct device_attribute *attr, char *buf)
> @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct coresight_device *csdev)
> csdev, coresight_orphan_match);
> }
>
> +/*
> + * Device connections are discovered before one/both devices have been created,
> + * so inputs must be added later.
> + */
> +static int coresight_fixup_inputs(struct coresight_device *csdev)
> +{
> + int i, ret = 0;
> + struct coresight_connection *out_conn;
> + struct coresight_connection in_conn;
> +
> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> + out_conn = &csdev->pdata->out_conns[i];
> + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
> + continue;
> +
> + /* Reverse local/remote relationships for inputs */
> + in_conn.remote_dev = csdev;
> + in_conn.remote_port = out_conn->port;
> + in_conn.port = out_conn->remote_port;
> + in_conn.remote_fwnode = csdev->dev.fwnode;
> + ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
> + out_conn->remote_dev->pdata,
> + &in_conn);
> + if (ret)
> + return ret;
> + }
> +
> + return 0;
> +}
>
> static int coresight_fixup_device_conns(struct coresight_device *csdev)
> {
> @@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct device *dev, void *data)
Removing the connection should be much simpler now with the inports
connection tracking.
i.e., coresight_remove_conns() need not iterate over all the devices on
the coresight bus. We had to do that to find all the devices connecting
to the removed device.
Instead with the in_conns we could readily get the csdev instance and
remove the connection from those devices.
Similarly for the out_conns can give you the remove device for removing
the in_conns on that device.
i.e,
static void coresight_remove_conns(csdev)
{
struct coresight_device *remote;
for (i = 0; i < csdev->pdata->nr_outconns; i++) {
remote = csdev->data->out_conns[i].remote_dev;
if (!remote)
continue;
for (j = 0; j < remote->pdata->nr_inconns; j++) {
if (remote->pdata->in_conns[i].remote_dev == csdev)
/* Clear in_conns[i] on remote */
}
}
Similarly for in_conns
for (i = 0; i < csdev->pdata->nr_inconns; i++) {
remote = csdev->data->in_conns[i].remote_dev;
if (!remote)
continue;
for (j = 0; j < remote->pdata->nr_outconns; j++) {
if (remote->pdata->out_conns[i].remote_dev == csdev)
/* Clear out_conns[i] on remote */
}
> */
> fwnode_handle_put(conn->remote_fwnode);
> conn->remote_fwnode = NULL;
> + conn->remote_dev = NULL;
> + /* No need to continue */
> + break;
> + }
> + }
> + for (i = 0; i < iterator->pdata->nr_inconns; i++) {
> + conn = &iterator->pdata->in_conns[i];
> + if (csdev == conn->remote_dev) {
> + conn->remote_fwnode = NULL;
> + conn->remote_dev = NULL;
> /* No need to continue */
> break;
> }
> }
> -
> /*
> * Returning '0' ensures that all known component on the
> * bus will be checked.
> @@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct coresight_device *csdev,
>
> for (i = 0; i < pdata->nr_outconns; i++) {
> /* If we have made the links, remove them now */
> - if (csdev && conns[i].remote_dev)
> + if (csdev && conns[i].remote_dev) {
> coresight_remove_links(csdev, &conns[i]);
> + conns[i].remote_dev = NULL;
> + }
> +
> /*
> * Drop the refcount and clear the handle as this device
> * is going away
> */
> if (conns[i].remote_fwnode) {
> fwnode_handle_put(conns[i].remote_fwnode);
> - pdata->out_conns[i].remote_fwnode = NULL;
> + conns[i].remote_fwnode = NULL;
> }
> }
> + for (i = 0; i < pdata->nr_inconns; i++) {
> + pdata->in_conns[i].remote_dev = NULL;
> + pdata->in_conns[i].remote_fwnode = NULL;
> + }
> +
> if (csdev)
> coresight_remove_conns_sysfs_group(csdev);
> }
> -
> struct coresight_device *coresight_register(struct coresight_desc *desc)
> {
> int ret;
> @@ -1659,6 +1705,8 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
> ret = coresight_create_conns_sysfs_group(csdev);
> if (!ret)
> ret = coresight_fixup_device_conns(csdev);
> + if (!ret)
> + ret = coresight_fixup_inputs(csdev);
Don't we also need to fixup the "input" connections on this
device ? It would be good move all of this connection stuff
into a wrapper function.
> if (!ret)
> ret = coresight_fixup_orphan_conns(csdev);
>
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index 16553f7dde12..20e3351cbdc2 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -20,8 +20,7 @@
>
> #include "coresight-priv.h"
> /*
> - * coresight_alloc_conns: Allocate connections record for each output
> - * port from the device.
> + * coresight_alloc_conns: Allocate connections record for each input/output device.
> */
> static int coresight_alloc_conns(struct device *dev,
> struct coresight_platform_data *pdata)
> @@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
> if (!pdata->out_conns)
> return -ENOMEM;
> }
> -
> + if (pdata->nr_inconns) {
> + pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
> + pdata->nr_inconns,
> + sizeof(*pdata->in_conns),
> + GFP_KERNEL | __GFP_ZERO);
> + if (!pdata->in_conns)
> + return -ENOMEM;
> + }
devm_krealloc_array() skips any allocation if the requested new_size <=
origin_size. So this should be fine for the out_conns, to not realloc
the space again unnecessarily.
> return 0;
> }
>
> @@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
> }
> EXPORT_SYMBOL_GPL(coresight_add_conn);
Should this be named as add_out_conn in the previous patch
just to make it explicit and easier to read.
>
> +/*
> + * Add a connection in the first free slot, or realloc
> + * if there is no space.
> + *
> + * Do nothing if the connection already exists because inputs are
> + * fixed up multiple times.
> + */
> +int coresight_add_in_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
minor nit: Is there any reason why this can't simply be:
int coresight_add_in_conn(struct coresight_device *csdev,
struct coresight_connection *conn)
{
struct device *dev = csdev->dev.parent;
struct coresight_platform_data *pdata = csdev->pdata;
> + struct coresight_connection *conn)
Suzuki
> +{
> + int ret;
> + struct coresight_connection *free_conn = NULL;
> + int i;
> +
> + /* Search for a free slot or exit if a duplicate is found */
> + if (pdata->in_conns) {
> + for (i = 0; i < pdata->nr_inconns; ++i) {
> + if (!free_conn && !pdata->in_conns[i].remote_fwnode)
> + free_conn = &pdata->in_conns[i];
> + if (pdata->in_conns[i].remote_fwnode ==
> + conn->remote_fwnode)
> + return 0;
> + }
> + }
> +
> + if (!free_conn) {
> + pdata->nr_inconns++;
> + ret = coresight_alloc_conns(dev, pdata);
> + if (ret)
> + return ret;
> + free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
> + }
> +
> + *free_conn = *conn;
> + return 0;
> +}
> +EXPORT_SYMBOL_GPL(coresight_add_in_conn);
> +
> static struct device *
> coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
> {
> @@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
>
> /*
> * of_coresight_parse_endpoint : Parse the given output endpoint @ep
> - * and fill the connection information in @conn
> + * and fill the connection information in @in_conn and @out_conn
> *
> * Parses the local port, remote device name and the remote port.
> *
> @@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct device *dev,
> /* Get the number of input and output port for this component */
> of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
>
> - /* If there are no output connections, we are done */
> - if (!pdata->nr_outconns)
> - return 0;
> -
> ret = coresight_alloc_conns(dev, pdata);
> if (ret)
> return ret;
>
> + /* If there are no output connections, we are done */
> + if (!pdata->nr_outconns)
> + return 0;
> +
> parent = of_coresight_get_output_ports_node(node);
> /*
> * If the DT uses obsoleted bindings, the ports are listed
> diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
> index 3da9868d9237..2abf9639ac0f 100644
> --- a/drivers/hwtracing/coresight/coresight-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
> @@ -202,5 +202,4 @@ void coresight_remove_links(struct coresight_device *orig,
> devm_kfree(&orig->dev, conn->link->orig_name);
> devm_kfree(&orig->dev, conn->link);
> conn->link = NULL;
> - conn->remote_dev = NULL;
> }
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index 47fa58d6981d..fd268b24c761 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -110,6 +110,7 @@ struct coresight_platform_data {
> int nr_inconns;
> int nr_outconns;
> struct coresight_connection *out_conns;
> + struct coresight_connection *in_conns;
Please document the field above the structure.
Suzuki
> };
>
> /**
> @@ -177,6 +178,27 @@ struct coresight_desc {
> * @remote_fwnode once the remote's coresight_device has
> * been created.
> * @link: Representation of the connection as a sysfs link.
> + *
> + * The full connection structure looks like this, where in_conns store references to
> + * the parent device in the same remote_dev member as output connections.
> + *
> + * +-----------------------------+ +-----------------------------+
> + * |coresight_device | |coresight_connection |
> + * |-----------------------------| |-----------------------------|
> + * ---->| | | |
> + * | | | | remote_dev*|------
> + * | |pdata->out_conns[nr_outconns]|----------->| | |
> + * | | | | | |
> + * | +-----------------------------+ +-----------------------------+ |
> + * | |
> + * | +-----------------------------+ +-----------------------------+ |
> + * | |coresight_connection | |coresight_device | |
> + * | |-----------------------------| |------------------------------ |
> + * | | | | |<-----
> + * -----|remote_dev* | | |
> + * | |<-----------|pdata->in_conns[nr_inconns] |
> + * | | | |
> + * +-----------------------------+ +-----------------------------+
> */
> struct coresight_connection {
> int port;
> @@ -619,5 +641,8 @@ struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
> int coresight_add_conn(struct device *dev,
> struct coresight_platform_data *pdata,
> const struct coresight_connection *conn);
> +int coresight_add_in_conn(struct device *dev,
> + struct coresight_platform_data *pdata,
> + struct coresight_connection *conn);
>
> #endif /* _LINUX_COREISGHT_H */
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
2023-03-10 16:06 ` James Clark
@ 2023-03-17 11:04 ` Suzuki K Poulose
-1 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-17 11:04 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> Currently CATU is the only helper device, and its enable and disable
> calls are hard coded. To allow more helper devices to be added in a
> generic way, remove these hard coded calls and just enable and disable
> all helper devices.
>
> This has to apply to helpers adjacent to the path, because they will
> never be in the path. CATU was already discovered in this way, so
> there is no change there.
>
> One change that is needed is for CATU to call back into ETR to allocate
> the buffer. Because the enable call was previously hard coded, it was
> done at a point where the buffer was already allocated, but this is no
> longer the case.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
> drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
> .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
> include/linux/coresight.h | 3 +-
> 4 files changed, 99 insertions(+), 34 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-catu.c b/drivers/hwtracing/coresight/coresight-catu.c
> index bc90a03f478f..24a08a2b96b1 100644
> --- a/drivers/hwtracing/coresight/coresight-catu.c
> +++ b/drivers/hwtracing/coresight/coresight-catu.c
> @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct catu_drvdata *drvdata)
> return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
> }
>
> -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> +static struct coresight_device *
> +catu_get_etr_device(struct coresight_device *csdev)
> +{
> + int i;
> + struct coresight_device *tmp;
> +
> + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
> + tmp = csdev->pdata->in_conns[i].remote_dev;
> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
> + tmp->subtype.sink_subtype ==
> + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
> + return tmp;
> + }
> +
> + return NULL;
> +}
> +
> +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode cs_mode,
> + void *data)
> {
> int rc;
> u32 control, mode;
> - struct etr_buf *etr_buf = data;
> + struct etr_buf *etr_buf = NULL;
> struct device *dev = &drvdata->csdev->dev;
> struct coresight_device *csdev = drvdata->csdev;
> + struct coresight_device *etrdev;
>
> if (catu_wait_for_ready(drvdata))
> dev_warn(dev, "Timeout while waiting for READY\n");
> @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> if (rc)
> return rc;
>
> + etrdev = catu_get_etr_device(csdev);
> + if (etrdev) {
> + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
> + if (IS_ERR(etr_buf))
> + return PTR_ERR(etr_buf);
> + }
WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
return.
> control |= BIT(CATU_CONTROL_ENABLE);
>
> if (etr_buf && etr_buf->mode == ETR_MODE_CATU) {
> @@ -441,13 +466,14 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> return 0;
> }
>
> -static int catu_enable(struct coresight_device *csdev, void *data)
> +static int catu_enable(struct coresight_device *csdev, enum cs_mode mode,
> + void *data)
> {
> int rc;
> struct catu_drvdata *catu_drvdata = csdev_to_catu_drvdata(csdev);
>
> CS_UNLOCK(catu_drvdata->base);
> - rc = catu_enable_hw(catu_drvdata, data);
> + rc = catu_enable_hw(catu_drvdata, mode, data);
> CS_LOCK(catu_drvdata->base);
> return rc;
> }
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index a8ba7493c09a..3e6ccd9e8d4e 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -441,6 +441,34 @@ static int coresight_enable_source(struct coresight_device *csdev,
> return 0;
> }
>
> +static int coresight_enable_helper(struct coresight_device *csdev,
> + enum cs_mode mode, void *sink_data)
minor nit: s/sink_data/data/ ? Though it is always either sink_data
(perf mode) or NULL (sysfs mode), for the core code it is simply an
opaque data.
Rest looks fine to me.
Suzuki
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
@ 2023-03-17 11:04 ` Suzuki K Poulose
0 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-17 11:04 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> Currently CATU is the only helper device, and its enable and disable
> calls are hard coded. To allow more helper devices to be added in a
> generic way, remove these hard coded calls and just enable and disable
> all helper devices.
>
> This has to apply to helpers adjacent to the path, because they will
> never be in the path. CATU was already discovered in this way, so
> there is no change there.
>
> One change that is needed is for CATU to call back into ETR to allocate
> the buffer. Because the enable call was previously hard coded, it was
> done at a point where the buffer was already allocated, but this is no
> longer the case.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
> drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
> .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
> include/linux/coresight.h | 3 +-
> 4 files changed, 99 insertions(+), 34 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-catu.c b/drivers/hwtracing/coresight/coresight-catu.c
> index bc90a03f478f..24a08a2b96b1 100644
> --- a/drivers/hwtracing/coresight/coresight-catu.c
> +++ b/drivers/hwtracing/coresight/coresight-catu.c
> @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct catu_drvdata *drvdata)
> return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
> }
>
> -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> +static struct coresight_device *
> +catu_get_etr_device(struct coresight_device *csdev)
> +{
> + int i;
> + struct coresight_device *tmp;
> +
> + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
> + tmp = csdev->pdata->in_conns[i].remote_dev;
> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
> + tmp->subtype.sink_subtype ==
> + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
> + return tmp;
> + }
> +
> + return NULL;
> +}
> +
> +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode cs_mode,
> + void *data)
> {
> int rc;
> u32 control, mode;
> - struct etr_buf *etr_buf = data;
> + struct etr_buf *etr_buf = NULL;
> struct device *dev = &drvdata->csdev->dev;
> struct coresight_device *csdev = drvdata->csdev;
> + struct coresight_device *etrdev;
>
> if (catu_wait_for_ready(drvdata))
> dev_warn(dev, "Timeout while waiting for READY\n");
> @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> if (rc)
> return rc;
>
> + etrdev = catu_get_etr_device(csdev);
> + if (etrdev) {
> + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
> + if (IS_ERR(etr_buf))
> + return PTR_ERR(etr_buf);
> + }
WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
return.
> control |= BIT(CATU_CONTROL_ENABLE);
>
> if (etr_buf && etr_buf->mode == ETR_MODE_CATU) {
> @@ -441,13 +466,14 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> return 0;
> }
>
> -static int catu_enable(struct coresight_device *csdev, void *data)
> +static int catu_enable(struct coresight_device *csdev, enum cs_mode mode,
> + void *data)
> {
> int rc;
> struct catu_drvdata *catu_drvdata = csdev_to_catu_drvdata(csdev);
>
> CS_UNLOCK(catu_drvdata->base);
> - rc = catu_enable_hw(catu_drvdata, data);
> + rc = catu_enable_hw(catu_drvdata, mode, data);
> CS_LOCK(catu_drvdata->base);
> return rc;
> }
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index a8ba7493c09a..3e6ccd9e8d4e 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -441,6 +441,34 @@ static int coresight_enable_source(struct coresight_device *csdev,
> return 0;
> }
>
> +static int coresight_enable_helper(struct coresight_device *csdev,
> + enum cs_mode mode, void *sink_data)
minor nit: s/sink_data/data/ ? Though it is always either sink_data
(perf mode) or NULL (sysfs mode), for the core code it is simply an
opaque data.
Rest looks fine to me.
Suzuki
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 9/9] coresight: Fix CTI module refcount leak by making it a helper device
2023-03-10 16:06 ` James Clark
@ 2023-03-17 11:40 ` Suzuki K Poulose
-1 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-17 11:40 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> The CTI module has some hard coded refcounting code that has a leak.
> For example running perf and then trying to unload it fails:
>
> perf record -e cs_etm// -a -- ls
> rmmod coresight_cti
>
> rmmod: ERROR: Module coresight_cti is in use
>
> The coresight core already handles references of devices in use, so by
> making CTI a normal helper device, we get working refcounting for free.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 73 +++----------------
> .../hwtracing/coresight/coresight-cti-core.c | 56 +++++++++-----
> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
> include/linux/coresight.h | 28 +------
> 5 files changed, 53 insertions(+), 112 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index 3e6ccd9e8d4e..94d84404fb29 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -253,48 +253,6 @@ void coresight_disclaim_device(struct coresight_device *csdev)
> }
> EXPORT_SYMBOL_GPL(coresight_disclaim_device);
>
> -/* enable or disable an associated CTI device of the supplied CS device */
> -static int
> -coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
> -{
> - int ect_ret = 0;
> - struct coresight_device *ect_csdev = csdev->ect_dev;
> - struct module *mod;
> -
> - if (!ect_csdev)
> - return 0;
> - if ((!ect_ops(ect_csdev)->enable) || (!ect_ops(ect_csdev)->disable))
> - return 0;
> -
> - mod = ect_csdev->dev.parent->driver->owner;
> - if (enable) {
> - if (try_module_get(mod)) {
> - ect_ret = ect_ops(ect_csdev)->enable(ect_csdev);
> - if (ect_ret) {
> - module_put(mod);
> - } else {
> - get_device(ect_csdev->dev.parent);
> - csdev->ect_enabled = true;
> - }
> - } else
> - ect_ret = -ENODEV;
> - } else {
> - if (csdev->ect_enabled) {
> - ect_ret = ect_ops(ect_csdev)->disable(ect_csdev);
> - put_device(ect_csdev->dev.parent);
> - module_put(mod);
> - csdev->ect_enabled = false;
> - }
> - }
> -
> - /* output warning if ECT enable is preventing trace operation */
> - if (ect_ret)
> - dev_info(&csdev->dev, "Associated ECT device (%s) %s failed\n",
> - dev_name(&ect_csdev->dev),
> - enable ? "enable" : "disable");
> - return ect_ret;
> -}
> -
> /*
> * Set the associated ect / cti device while holding the coresight_mutex
> * to avoid a race with coresight_enable that may try to use this value.
> @@ -302,8 +260,14 @@ coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
> void coresight_set_assoc_ectdev_mutex(struct coresight_device *csdev,
> struct coresight_device *ect_csdev)
> {
> + struct coresight_connection conn = {};
> +
> mutex_lock(&coresight_mutex);
> - csdev->ect_dev = ect_csdev;
> + conn.remote_fwnode = fwnode_handle_get(dev_fwnode(&ect_csdev->dev));
> + conn.remote_dev = ect_csdev;
> + conn.remote_port = conn.port = -1;
> + coresight_add_conn(csdev->dev.parent, csdev->pdata, &conn);
> + coresight_fixup_inputs(csdev);
Can this cause duplicate entries to be created for the other "existing"
output connections of csdev ? IIUC, we just need to create the input
connection for the connection we are creating now and not all of the
out_conns. They should have been populated at coresight_register().
Now you may not hit this, because we don't have any actual "defined"
connections in the DT. But we need to be safe here.
> mutex_unlock(&coresight_mutex);
> }
> EXPORT_SYMBOL_GPL(coresight_set_assoc_ectdev_mutex);
> @@ -320,12 +284,8 @@ static int coresight_enable_sink(struct coresight_device *csdev,
> if (!sink_ops(csdev)->enable)
> return -EINVAL;
>
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (ret)
> - return ret;
> ret = sink_ops(csdev)->enable(csdev, mode, data);
> if (ret) {
> - coresight_control_assoc_ectdev(csdev, false);
> return ret;
> }
super minor nit: You could remove the { } too
> csdev->enable = true;
> @@ -343,7 +303,6 @@ static void coresight_disable_sink(struct coresight_device *csdev)
> ret = sink_ops(csdev)->disable(csdev);
> if (ret)
> return;
> - coresight_control_assoc_ectdev(csdev, false);
> csdev->enable = false;
> }
>
> @@ -368,17 +327,11 @@ static int coresight_enable_link(struct coresight_device *csdev,
> return outport;
>
> if (link_ops(csdev)->enable) {
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (!ret) {
> - ret = link_ops(csdev)->enable(csdev, inport, outport);
> - if (ret)
> - coresight_control_assoc_ectdev(csdev, false);
> - }
> + ret = link_ops(csdev)->enable(csdev, inport, outport);
> + if (!ret)
> + csdev->enable = true;
> }
>
> - if (!ret)
> - csdev->enable = true;
> -
> return ret;
> }
>
> @@ -407,7 +360,6 @@ static void coresight_disable_link(struct coresight_device *csdev,
>
> if (link_ops(csdev)->disable) {
> link_ops(csdev)->disable(csdev, inport, outport);
> - coresight_control_assoc_ectdev(csdev, false);
> }
>
super minor nit: You may remove the { } s.
> for (i = 0; i < nr_conns; i++)
> @@ -424,12 +376,8 @@ static int coresight_enable_source(struct coresight_device *csdev,
>
> if (!csdev->enable) {
> if (source_ops(csdev)->enable) {
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (ret)
> - return ret;
> ret = source_ops(csdev)->enable(csdev, NULL, mode);
> if (ret) {
> - coresight_control_assoc_ectdev(csdev, false);
> return ret;
> }
> }
> @@ -482,7 +430,6 @@ static bool coresight_disable_source(struct coresight_device *csdev)
> if (atomic_dec_return(csdev->refcnt) == 0) {
> if (source_ops(csdev)->disable)
> source_ops(csdev)->disable(csdev, NULL);
> - coresight_control_assoc_ectdev(csdev, false);
> csdev->enable = false;
> }
> return !csdev->enable;
> diff --git a/drivers/hwtracing/coresight/coresight-cti-core.c b/drivers/hwtracing/coresight/coresight-cti-core.c
> index 277c890a1f1f..dbce6680759f 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-core.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-core.c
> @@ -555,7 +555,10 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
> mutex_lock(&ect_mutex);
>
> /* exit if current is an ECT device.*/
> - if ((csdev->type == CORESIGHT_DEV_TYPE_ECT) || list_empty(&ect_net))
> + if ((csdev->type == CORESIGHT_DEV_TYPE_HELPER &&
> + csdev->subtype.helper_subtype ==
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI) ||
Should we add a helper to check this :
is_coresight_ctidev(csdev) ?
> + list_empty(&ect_net))
> goto cti_add_done;
>
> /* if we didn't find the csdev previously we used the fwnode name */
> @@ -580,6 +583,22 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
> mutex_unlock(&ect_mutex);
> }
>
> +static struct coresight_device *cti__get_cti_device(struct coresight_device *csdev)
minor nit: no perf tool naming style please ;-)
> +{
> + int i;
> + struct coresight_device *tmp;
> +
> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> + tmp = csdev->pdata->out_conns[i].remote_dev;
> +
> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_HELPER &&
> + tmp->subtype.helper_subtype ==
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI)
As mentioned above, a helper to check this would benefit reading.
> + return tmp;
> + }
> + return NULL;
> +}
> +
> /*
> * Removing the associated devices is easier.
> * A CTI will not have a value for csdev->ect_dev.
> @@ -588,20 +607,21 @@ static void cti_remove_assoc_from_csdev(struct coresight_device *csdev)
> {
> struct cti_drvdata *ctidrv;
> struct cti_trig_con *tc;
> + struct coresight_device *cti_csdev = cti__get_cti_device(csdev);
> struct cti_device *ctidev;
>
> + if (!cti_csdev)
> + return;
> +
> mutex_lock(&ect_mutex);
> - if (csdev->ect_dev) {
> - ctidrv = csdev_to_cti_drvdata(csdev->ect_dev);
> - ctidev = &ctidrv->ctidev;
> - list_for_each_entry(tc, &ctidev->trig_cons, node) {
> - if (tc->con_dev == csdev) {
> - cti_remove_sysfs_link(ctidrv, tc);
> - tc->con_dev = NULL;
> - break;
> - }
> + ctidrv = csdev_to_cti_drvdata(cti_csdev);
> + ctidev = &ctidrv->ctidev;
> + list_for_each_entry(tc, &ctidev->trig_cons, node) {
> + if (tc->con_dev == csdev) {
> + cti_remove_sysfs_link(ctidrv, tc);
> + tc->con_dev = NULL;
> + break;
> }
> - csdev->ect_dev = NULL;
> }
> mutex_unlock(&ect_mutex);
> }
> @@ -646,8 +666,6 @@ static void cti_remove_conn_xrefs(struct cti_drvdata *drvdata)
>
> list_for_each_entry(tc, &ctidev->trig_cons, node) {
> if (tc->con_dev) {
> - coresight_set_assoc_ectdev_mutex(tc->con_dev,
> - NULL);
> cti_remove_sysfs_link(drvdata, tc);
> tc->con_dev = NULL;
> }
> @@ -795,27 +813,27 @@ static void cti_pm_release(struct cti_drvdata *drvdata)
> }
>
> /** cti ect operations **/
> -int cti_enable(struct coresight_device *csdev)
> +int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data)
> {
> struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
>
> return cti_enable_hw(drvdata);
> }
>
> -int cti_disable(struct coresight_device *csdev)
> +int cti_disable(struct coresight_device *csdev, void *data)
> {
> struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
>
> return cti_disable_hw(drvdata);
> }
>
> -static const struct coresight_ops_ect cti_ops_ect = {
> +static const struct coresight_ops_helper cti_ops_ect = {
> .enable = cti_enable,
> .disable = cti_disable,
> };
>
> static const struct coresight_ops cti_ops = {
> - .ect_ops = &cti_ops_ect,
> + .helper_ops = &cti_ops_ect,
> };
>
> /*
> @@ -922,8 +940,8 @@ static int cti_probe(struct amba_device *adev, const struct amba_id *id)
>
> /* set up coresight component description */
> cti_desc.pdata = pdata;
> - cti_desc.type = CORESIGHT_DEV_TYPE_ECT;
> - cti_desc.subtype.ect_subtype = CORESIGHT_DEV_SUBTYPE_ECT_CTI;
> + cti_desc.type = CORESIGHT_DEV_TYPE_HELPER;
> + cti_desc.subtype.helper_subtype = CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI;
> cti_desc.ops = &cti_ops;
> cti_desc.groups = drvdata->ctidev.con_groups;
> cti_desc.dev = dev;
> diff --git a/drivers/hwtracing/coresight/coresight-cti-sysfs.c b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> index e528cff9d4e2..d25dd2737b49 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> @@ -112,11 +112,11 @@ static ssize_t enable_store(struct device *dev,
> ret = pm_runtime_resume_and_get(dev->parent);
> if (ret)
> return ret;
> - ret = cti_enable(drvdata->csdev);
> + ret = cti_enable(drvdata->csdev, CS_MODE_SYSFS, NULL);
> if (ret)
> pm_runtime_put(dev->parent);
> } else {
> - ret = cti_disable(drvdata->csdev);
> + ret = cti_disable(drvdata->csdev, NULL);
> if (!ret)
> pm_runtime_put(dev->parent);
> }
> diff --git a/drivers/hwtracing/coresight/coresight-cti.h b/drivers/hwtracing/coresight/coresight-cti.h
> index 8b106b13a244..cb9ee616d01f 100644
> --- a/drivers/hwtracing/coresight/coresight-cti.h
> +++ b/drivers/hwtracing/coresight/coresight-cti.h
> @@ -215,8 +215,8 @@ int cti_add_connection_entry(struct device *dev, struct cti_drvdata *drvdata,
> const char *assoc_dev_name);
> struct cti_trig_con *cti_allocate_trig_con(struct device *dev, int in_sigs,
> int out_sigs);
> -int cti_enable(struct coresight_device *csdev);
> -int cti_disable(struct coresight_device *csdev);
> +int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data);
> +int cti_disable(struct coresight_device *csdev, void *data);
> void cti_write_all_hw_regs(struct cti_drvdata *drvdata);
> void cti_write_intack(struct device *dev, u32 ackval);
> void cti_write_single_reg(struct cti_drvdata *drvdata, int offset, u32 value);
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index c6ee1634d813..cf7a8658ee88 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -40,8 +40,7 @@ enum coresight_dev_type {
> CORESIGHT_DEV_TYPE_LINK,
> CORESIGHT_DEV_TYPE_LINKSINK,
> CORESIGHT_DEV_TYPE_SOURCE,
> - CORESIGHT_DEV_TYPE_HELPER,
> - CORESIGHT_DEV_TYPE_ECT,
> + CORESIGHT_DEV_TYPE_HELPER
> };
>
> enum coresight_dev_subtype_sink {
> @@ -66,12 +65,7 @@ enum coresight_dev_subtype_source {
>
> enum coresight_dev_subtype_helper {
> CORESIGHT_DEV_SUBTYPE_HELPER_CATU,
> -};
> -
> -/* Embedded Cross Trigger (ECT) sub-types */
> -enum coresight_dev_subtype_ect {
> - CORESIGHT_DEV_SUBTYPE_ECT_NONE,
> - CORESIGHT_DEV_SUBTYPE_ECT_CTI,
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI
> };
>
> /**
> @@ -84,8 +78,6 @@ enum coresight_dev_subtype_ect {
> * by @coresight_dev_subtype_source.
> * @helper_subtype: type of helper this component is, as defined
> * by @coresight_dev_subtype_helper.
> - * @ect_subtype: type of cross trigger this component is, as
> - * defined by @coresight_dev_subtype_ect
> */
> union coresight_dev_subtype {
> /* We have some devices which acts as LINK and SINK */
> @@ -95,7 +87,6 @@ union coresight_dev_subtype {
> };
> enum coresight_dev_subtype_source source_subtype;
> enum coresight_dev_subtype_helper helper_subtype;
> - enum coresight_dev_subtype_ect ect_subtype;
> };
>
> /**
> @@ -264,12 +255,9 @@ struct coresight_device {
> bool activated; /* true only if a sink is part of a path */
> struct dev_ext_attribute *ea;
> struct coresight_device *def_sink;
> - /* cross trigger handling */
> - struct coresight_device *ect_dev;
> /* sysfs links between components */
> int nr_links;
> bool has_conns_grp;
> - bool ect_enabled; /* true only if associated ect device is enabled */
nit: Update comments above the struct too ?
Thanks
Suzuki
> /* system configuration and feature lists */
> struct list_head feature_csdev_list;
> struct list_head config_csdev_list;
> @@ -377,23 +365,11 @@ struct coresight_ops_helper {
> int (*disable)(struct coresight_device *csdev, void *data);
> };
>
> -/**
> - * struct coresight_ops_ect - Ops for an embedded cross trigger device
> - *
> - * @enable : Enable the device
> - * @disable : Disable the device
> - */
> -struct coresight_ops_ect {
> - int (*enable)(struct coresight_device *csdev);
> - int (*disable)(struct coresight_device *csdev);
> -};
> -
> struct coresight_ops {
> const struct coresight_ops_sink *sink_ops;
> const struct coresight_ops_link *link_ops;
> const struct coresight_ops_source *source_ops;
> const struct coresight_ops_helper *helper_ops;
> - const struct coresight_ops_ect *ect_ops;
> };
>
> #if IS_ENABLED(CONFIG_CORESIGHT)
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 9/9] coresight: Fix CTI module refcount leak by making it a helper device
@ 2023-03-17 11:40 ` Suzuki K Poulose
0 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-17 11:40 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 10/03/2023 16:06, James Clark wrote:
> The CTI module has some hard coded refcounting code that has a leak.
> For example running perf and then trying to unload it fails:
>
> perf record -e cs_etm// -a -- ls
> rmmod coresight_cti
>
> rmmod: ERROR: Module coresight_cti is in use
>
> The coresight core already handles references of devices in use, so by
> making CTI a normal helper device, we get working refcounting for free.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 73 +++----------------
> .../hwtracing/coresight/coresight-cti-core.c | 56 +++++++++-----
> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
> include/linux/coresight.h | 28 +------
> 5 files changed, 53 insertions(+), 112 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index 3e6ccd9e8d4e..94d84404fb29 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -253,48 +253,6 @@ void coresight_disclaim_device(struct coresight_device *csdev)
> }
> EXPORT_SYMBOL_GPL(coresight_disclaim_device);
>
> -/* enable or disable an associated CTI device of the supplied CS device */
> -static int
> -coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
> -{
> - int ect_ret = 0;
> - struct coresight_device *ect_csdev = csdev->ect_dev;
> - struct module *mod;
> -
> - if (!ect_csdev)
> - return 0;
> - if ((!ect_ops(ect_csdev)->enable) || (!ect_ops(ect_csdev)->disable))
> - return 0;
> -
> - mod = ect_csdev->dev.parent->driver->owner;
> - if (enable) {
> - if (try_module_get(mod)) {
> - ect_ret = ect_ops(ect_csdev)->enable(ect_csdev);
> - if (ect_ret) {
> - module_put(mod);
> - } else {
> - get_device(ect_csdev->dev.parent);
> - csdev->ect_enabled = true;
> - }
> - } else
> - ect_ret = -ENODEV;
> - } else {
> - if (csdev->ect_enabled) {
> - ect_ret = ect_ops(ect_csdev)->disable(ect_csdev);
> - put_device(ect_csdev->dev.parent);
> - module_put(mod);
> - csdev->ect_enabled = false;
> - }
> - }
> -
> - /* output warning if ECT enable is preventing trace operation */
> - if (ect_ret)
> - dev_info(&csdev->dev, "Associated ECT device (%s) %s failed\n",
> - dev_name(&ect_csdev->dev),
> - enable ? "enable" : "disable");
> - return ect_ret;
> -}
> -
> /*
> * Set the associated ect / cti device while holding the coresight_mutex
> * to avoid a race with coresight_enable that may try to use this value.
> @@ -302,8 +260,14 @@ coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
> void coresight_set_assoc_ectdev_mutex(struct coresight_device *csdev,
> struct coresight_device *ect_csdev)
> {
> + struct coresight_connection conn = {};
> +
> mutex_lock(&coresight_mutex);
> - csdev->ect_dev = ect_csdev;
> + conn.remote_fwnode = fwnode_handle_get(dev_fwnode(&ect_csdev->dev));
> + conn.remote_dev = ect_csdev;
> + conn.remote_port = conn.port = -1;
> + coresight_add_conn(csdev->dev.parent, csdev->pdata, &conn);
> + coresight_fixup_inputs(csdev);
Can this cause duplicate entries to be created for the other "existing"
output connections of csdev ? IIUC, we just need to create the input
connection for the connection we are creating now and not all of the
out_conns. They should have been populated at coresight_register().
Now you may not hit this, because we don't have any actual "defined"
connections in the DT. But we need to be safe here.
> mutex_unlock(&coresight_mutex);
> }
> EXPORT_SYMBOL_GPL(coresight_set_assoc_ectdev_mutex);
> @@ -320,12 +284,8 @@ static int coresight_enable_sink(struct coresight_device *csdev,
> if (!sink_ops(csdev)->enable)
> return -EINVAL;
>
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (ret)
> - return ret;
> ret = sink_ops(csdev)->enable(csdev, mode, data);
> if (ret) {
> - coresight_control_assoc_ectdev(csdev, false);
> return ret;
> }
super minor nit: You could remove the { } too
> csdev->enable = true;
> @@ -343,7 +303,6 @@ static void coresight_disable_sink(struct coresight_device *csdev)
> ret = sink_ops(csdev)->disable(csdev);
> if (ret)
> return;
> - coresight_control_assoc_ectdev(csdev, false);
> csdev->enable = false;
> }
>
> @@ -368,17 +327,11 @@ static int coresight_enable_link(struct coresight_device *csdev,
> return outport;
>
> if (link_ops(csdev)->enable) {
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (!ret) {
> - ret = link_ops(csdev)->enable(csdev, inport, outport);
> - if (ret)
> - coresight_control_assoc_ectdev(csdev, false);
> - }
> + ret = link_ops(csdev)->enable(csdev, inport, outport);
> + if (!ret)
> + csdev->enable = true;
> }
>
> - if (!ret)
> - csdev->enable = true;
> -
> return ret;
> }
>
> @@ -407,7 +360,6 @@ static void coresight_disable_link(struct coresight_device *csdev,
>
> if (link_ops(csdev)->disable) {
> link_ops(csdev)->disable(csdev, inport, outport);
> - coresight_control_assoc_ectdev(csdev, false);
> }
>
super minor nit: You may remove the { } s.
> for (i = 0; i < nr_conns; i++)
> @@ -424,12 +376,8 @@ static int coresight_enable_source(struct coresight_device *csdev,
>
> if (!csdev->enable) {
> if (source_ops(csdev)->enable) {
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (ret)
> - return ret;
> ret = source_ops(csdev)->enable(csdev, NULL, mode);
> if (ret) {
> - coresight_control_assoc_ectdev(csdev, false);
> return ret;
> }
> }
> @@ -482,7 +430,6 @@ static bool coresight_disable_source(struct coresight_device *csdev)
> if (atomic_dec_return(csdev->refcnt) == 0) {
> if (source_ops(csdev)->disable)
> source_ops(csdev)->disable(csdev, NULL);
> - coresight_control_assoc_ectdev(csdev, false);
> csdev->enable = false;
> }
> return !csdev->enable;
> diff --git a/drivers/hwtracing/coresight/coresight-cti-core.c b/drivers/hwtracing/coresight/coresight-cti-core.c
> index 277c890a1f1f..dbce6680759f 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-core.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-core.c
> @@ -555,7 +555,10 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
> mutex_lock(&ect_mutex);
>
> /* exit if current is an ECT device.*/
> - if ((csdev->type == CORESIGHT_DEV_TYPE_ECT) || list_empty(&ect_net))
> + if ((csdev->type == CORESIGHT_DEV_TYPE_HELPER &&
> + csdev->subtype.helper_subtype ==
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI) ||
Should we add a helper to check this :
is_coresight_ctidev(csdev) ?
> + list_empty(&ect_net))
> goto cti_add_done;
>
> /* if we didn't find the csdev previously we used the fwnode name */
> @@ -580,6 +583,22 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
> mutex_unlock(&ect_mutex);
> }
>
> +static struct coresight_device *cti__get_cti_device(struct coresight_device *csdev)
minor nit: no perf tool naming style please ;-)
> +{
> + int i;
> + struct coresight_device *tmp;
> +
> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> + tmp = csdev->pdata->out_conns[i].remote_dev;
> +
> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_HELPER &&
> + tmp->subtype.helper_subtype ==
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI)
As mentioned above, a helper to check this would benefit reading.
> + return tmp;
> + }
> + return NULL;
> +}
> +
> /*
> * Removing the associated devices is easier.
> * A CTI will not have a value for csdev->ect_dev.
> @@ -588,20 +607,21 @@ static void cti_remove_assoc_from_csdev(struct coresight_device *csdev)
> {
> struct cti_drvdata *ctidrv;
> struct cti_trig_con *tc;
> + struct coresight_device *cti_csdev = cti__get_cti_device(csdev);
> struct cti_device *ctidev;
>
> + if (!cti_csdev)
> + return;
> +
> mutex_lock(&ect_mutex);
> - if (csdev->ect_dev) {
> - ctidrv = csdev_to_cti_drvdata(csdev->ect_dev);
> - ctidev = &ctidrv->ctidev;
> - list_for_each_entry(tc, &ctidev->trig_cons, node) {
> - if (tc->con_dev == csdev) {
> - cti_remove_sysfs_link(ctidrv, tc);
> - tc->con_dev = NULL;
> - break;
> - }
> + ctidrv = csdev_to_cti_drvdata(cti_csdev);
> + ctidev = &ctidrv->ctidev;
> + list_for_each_entry(tc, &ctidev->trig_cons, node) {
> + if (tc->con_dev == csdev) {
> + cti_remove_sysfs_link(ctidrv, tc);
> + tc->con_dev = NULL;
> + break;
> }
> - csdev->ect_dev = NULL;
> }
> mutex_unlock(&ect_mutex);
> }
> @@ -646,8 +666,6 @@ static void cti_remove_conn_xrefs(struct cti_drvdata *drvdata)
>
> list_for_each_entry(tc, &ctidev->trig_cons, node) {
> if (tc->con_dev) {
> - coresight_set_assoc_ectdev_mutex(tc->con_dev,
> - NULL);
> cti_remove_sysfs_link(drvdata, tc);
> tc->con_dev = NULL;
> }
> @@ -795,27 +813,27 @@ static void cti_pm_release(struct cti_drvdata *drvdata)
> }
>
> /** cti ect operations **/
> -int cti_enable(struct coresight_device *csdev)
> +int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data)
> {
> struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
>
> return cti_enable_hw(drvdata);
> }
>
> -int cti_disable(struct coresight_device *csdev)
> +int cti_disable(struct coresight_device *csdev, void *data)
> {
> struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
>
> return cti_disable_hw(drvdata);
> }
>
> -static const struct coresight_ops_ect cti_ops_ect = {
> +static const struct coresight_ops_helper cti_ops_ect = {
> .enable = cti_enable,
> .disable = cti_disable,
> };
>
> static const struct coresight_ops cti_ops = {
> - .ect_ops = &cti_ops_ect,
> + .helper_ops = &cti_ops_ect,
> };
>
> /*
> @@ -922,8 +940,8 @@ static int cti_probe(struct amba_device *adev, const struct amba_id *id)
>
> /* set up coresight component description */
> cti_desc.pdata = pdata;
> - cti_desc.type = CORESIGHT_DEV_TYPE_ECT;
> - cti_desc.subtype.ect_subtype = CORESIGHT_DEV_SUBTYPE_ECT_CTI;
> + cti_desc.type = CORESIGHT_DEV_TYPE_HELPER;
> + cti_desc.subtype.helper_subtype = CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI;
> cti_desc.ops = &cti_ops;
> cti_desc.groups = drvdata->ctidev.con_groups;
> cti_desc.dev = dev;
> diff --git a/drivers/hwtracing/coresight/coresight-cti-sysfs.c b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> index e528cff9d4e2..d25dd2737b49 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> @@ -112,11 +112,11 @@ static ssize_t enable_store(struct device *dev,
> ret = pm_runtime_resume_and_get(dev->parent);
> if (ret)
> return ret;
> - ret = cti_enable(drvdata->csdev);
> + ret = cti_enable(drvdata->csdev, CS_MODE_SYSFS, NULL);
> if (ret)
> pm_runtime_put(dev->parent);
> } else {
> - ret = cti_disable(drvdata->csdev);
> + ret = cti_disable(drvdata->csdev, NULL);
> if (!ret)
> pm_runtime_put(dev->parent);
> }
> diff --git a/drivers/hwtracing/coresight/coresight-cti.h b/drivers/hwtracing/coresight/coresight-cti.h
> index 8b106b13a244..cb9ee616d01f 100644
> --- a/drivers/hwtracing/coresight/coresight-cti.h
> +++ b/drivers/hwtracing/coresight/coresight-cti.h
> @@ -215,8 +215,8 @@ int cti_add_connection_entry(struct device *dev, struct cti_drvdata *drvdata,
> const char *assoc_dev_name);
> struct cti_trig_con *cti_allocate_trig_con(struct device *dev, int in_sigs,
> int out_sigs);
> -int cti_enable(struct coresight_device *csdev);
> -int cti_disable(struct coresight_device *csdev);
> +int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data);
> +int cti_disable(struct coresight_device *csdev, void *data);
> void cti_write_all_hw_regs(struct cti_drvdata *drvdata);
> void cti_write_intack(struct device *dev, u32 ackval);
> void cti_write_single_reg(struct cti_drvdata *drvdata, int offset, u32 value);
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index c6ee1634d813..cf7a8658ee88 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -40,8 +40,7 @@ enum coresight_dev_type {
> CORESIGHT_DEV_TYPE_LINK,
> CORESIGHT_DEV_TYPE_LINKSINK,
> CORESIGHT_DEV_TYPE_SOURCE,
> - CORESIGHT_DEV_TYPE_HELPER,
> - CORESIGHT_DEV_TYPE_ECT,
> + CORESIGHT_DEV_TYPE_HELPER
> };
>
> enum coresight_dev_subtype_sink {
> @@ -66,12 +65,7 @@ enum coresight_dev_subtype_source {
>
> enum coresight_dev_subtype_helper {
> CORESIGHT_DEV_SUBTYPE_HELPER_CATU,
> -};
> -
> -/* Embedded Cross Trigger (ECT) sub-types */
> -enum coresight_dev_subtype_ect {
> - CORESIGHT_DEV_SUBTYPE_ECT_NONE,
> - CORESIGHT_DEV_SUBTYPE_ECT_CTI,
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI
> };
>
> /**
> @@ -84,8 +78,6 @@ enum coresight_dev_subtype_ect {
> * by @coresight_dev_subtype_source.
> * @helper_subtype: type of helper this component is, as defined
> * by @coresight_dev_subtype_helper.
> - * @ect_subtype: type of cross trigger this component is, as
> - * defined by @coresight_dev_subtype_ect
> */
> union coresight_dev_subtype {
> /* We have some devices which acts as LINK and SINK */
> @@ -95,7 +87,6 @@ union coresight_dev_subtype {
> };
> enum coresight_dev_subtype_source source_subtype;
> enum coresight_dev_subtype_helper helper_subtype;
> - enum coresight_dev_subtype_ect ect_subtype;
> };
>
> /**
> @@ -264,12 +255,9 @@ struct coresight_device {
> bool activated; /* true only if a sink is part of a path */
> struct dev_ext_attribute *ea;
> struct coresight_device *def_sink;
> - /* cross trigger handling */
> - struct coresight_device *ect_dev;
> /* sysfs links between components */
> int nr_links;
> bool has_conns_grp;
> - bool ect_enabled; /* true only if associated ect device is enabled */
nit: Update comments above the struct too ?
Thanks
Suzuki
> /* system configuration and feature lists */
> struct list_head feature_csdev_list;
> struct list_head config_csdev_list;
> @@ -377,23 +365,11 @@ struct coresight_ops_helper {
> int (*disable)(struct coresight_device *csdev, void *data);
> };
>
> -/**
> - * struct coresight_ops_ect - Ops for an embedded cross trigger device
> - *
> - * @enable : Enable the device
> - * @disable : Disable the device
> - */
> -struct coresight_ops_ect {
> - int (*enable)(struct coresight_device *csdev);
> - int (*disable)(struct coresight_device *csdev);
> -};
> -
> struct coresight_ops {
> const struct coresight_ops_sink *sink_ops;
> const struct coresight_ops_link *link_ops;
> const struct coresight_ops_source *source_ops;
> const struct coresight_ops_helper *helper_ops;
> - const struct coresight_ops_ect *ect_ops;
> };
>
> #if IS_ENABLED(CONFIG_CORESIGHT)
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device
2023-03-10 16:05 ` James Clark
@ 2023-03-21 15:12 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 15:12 UTC (permalink / raw)
To: James Clark
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James
On Fri, 10 Mar 2023 at 16:06, James Clark <james.clark@arm.com> wrote:
>
> Changes since v1:
>
> * Don't dereference handle in tmc_etr_get_buffer() when not in perf mode.
> * Fix some W=1 warnings
> * Add a commit to rename child/output in terms of local/remote
>
> -------------------
>
> Currently there is a refcount leak in CTI when using system wide mode
> or tracing multithreaded applications. See the last commit for a
> reproducer. This prevents the module from being unloaded.
>
> Historically there have been a few issues and fixes attempted around
> here which have resulted in some extra logic and a member to keep
> track of CTI being enabled 'struct coresight_device->ect_enabled'.
> The fix in commit 665c157e0204 ("coresight: cti: Fix hang in
> cti_disable_hw()") was also related to CTI having its own
> enable/disable path which came later than other devices.
>
> If we make CTI a helper device and enable helper devices adjacent to
> the path we get very similar enable/disable behavior to now, but with
> more reuse of the existing reference counting logic in the coresight
> core code. This also affects CATU which can have a little bit of
> its hard coded enable/disable code removed.
>
> Enabling CATU on the generic path does require that input connections
> are tracked so that it can get its associated ETR buffer.
>
> Applies to coresight/next (669c4614236a7) but also requires the
> realloc_array patch here [1].
>
> Also available in full here [2].
>
> [1]: https://lore.kernel.org/linux-arm-kernel/20230306152723.3090195-1-james.clark@arm.com/
> [2]: https://gitlab.arm.com/linux-arm/linux-jc/-/tree/james-cs-cti-module-refcount-fix-v2
>
> James Clark (9):
> coresight: Use enum type for cs_mode wherever possible
> coresight: Change name of pdata->conns
> coresight: Rename nr_outports to nr_outconns
> coresight: Rename connection members to allow for input connections
> coresight: Dynamically add connections
> coresight: Store in-connections as well as out-connections
> coresight: Refactor out buffer allocation function for ETR
> coresight: Enable and disable helper devices adjacent to the path
> coresight: Fix CTI module refcount leak by making it a helper device
>
> drivers/hwtracing/coresight/coresight-catu.c | 34 +-
> drivers/hwtracing/coresight/coresight-core.c | 312 +++++++++++-------
> .../hwtracing/coresight/coresight-cti-core.c | 56 ++--
> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
> drivers/hwtracing/coresight/coresight-etb10.c | 3 +-
> .../coresight/coresight-etm3x-core.c | 6 +-
> .../coresight/coresight-etm4x-core.c | 6 +-
> .../hwtracing/coresight/coresight-platform.c | 178 +++++++---
> drivers/hwtracing/coresight/coresight-priv.h | 9 +-
> drivers/hwtracing/coresight/coresight-stm.c | 6 +-
> drivers/hwtracing/coresight/coresight-sysfs.c | 9 +-
> .../hwtracing/coresight/coresight-tmc-etf.c | 2 +-
> .../hwtracing/coresight/coresight-tmc-etr.c | 89 ++---
> drivers/hwtracing/coresight/coresight-tmc.h | 2 +
> drivers/hwtracing/coresight/coresight-tpdm.c | 4 +-
> drivers/hwtracing/coresight/coresight-tpiu.c | 3 +-
> drivers/hwtracing/coresight/coresight-trbe.c | 3 +-
> drivers/hwtracing/coresight/ultrasoc-smb.c | 3 +-
> drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
> include/linux/coresight.h | 109 +++---
> 21 files changed, 530 insertions(+), 314 deletions(-)
>
> --
> 2.34.1
>
Looking at this overall - given that the only use of the in_conn is to
reference the connecting device from the helper,
i.e. coresight-catu.c:405: tmp = csdev->pdata->in_conns[i].remote_dev;
would it not be simpler to :
a)
in coresight_connection add a field:
struct coresight_device *origin_dev;
which mimics the origin / target model we already have in coresight_sysfs_link
then
b) the in_conns could simply be references to out_conn object from
origin_dev, rather than a complete coresight_connection with reversed
values, thus simplifying the in_conns handling code, and removing the
unused reversed feilds in the current in_conn object.
e.g. tmp = csdev->pdata->in_conns[i]->origin_dev
The remainder of the code would remain much the same, just adjusted
for in_conns as refs rather than independent conn objects
Regards
Mike
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device
@ 2023-03-21 15:12 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 15:12 UTC (permalink / raw)
To: James Clark
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James
On Fri, 10 Mar 2023 at 16:06, James Clark <james.clark@arm.com> wrote:
>
> Changes since v1:
>
> * Don't dereference handle in tmc_etr_get_buffer() when not in perf mode.
> * Fix some W=1 warnings
> * Add a commit to rename child/output in terms of local/remote
>
> -------------------
>
> Currently there is a refcount leak in CTI when using system wide mode
> or tracing multithreaded applications. See the last commit for a
> reproducer. This prevents the module from being unloaded.
>
> Historically there have been a few issues and fixes attempted around
> here which have resulted in some extra logic and a member to keep
> track of CTI being enabled 'struct coresight_device->ect_enabled'.
> The fix in commit 665c157e0204 ("coresight: cti: Fix hang in
> cti_disable_hw()") was also related to CTI having its own
> enable/disable path which came later than other devices.
>
> If we make CTI a helper device and enable helper devices adjacent to
> the path we get very similar enable/disable behavior to now, but with
> more reuse of the existing reference counting logic in the coresight
> core code. This also affects CATU which can have a little bit of
> its hard coded enable/disable code removed.
>
> Enabling CATU on the generic path does require that input connections
> are tracked so that it can get its associated ETR buffer.
>
> Applies to coresight/next (669c4614236a7) but also requires the
> realloc_array patch here [1].
>
> Also available in full here [2].
>
> [1]: https://lore.kernel.org/linux-arm-kernel/20230306152723.3090195-1-james.clark@arm.com/
> [2]: https://gitlab.arm.com/linux-arm/linux-jc/-/tree/james-cs-cti-module-refcount-fix-v2
>
> James Clark (9):
> coresight: Use enum type for cs_mode wherever possible
> coresight: Change name of pdata->conns
> coresight: Rename nr_outports to nr_outconns
> coresight: Rename connection members to allow for input connections
> coresight: Dynamically add connections
> coresight: Store in-connections as well as out-connections
> coresight: Refactor out buffer allocation function for ETR
> coresight: Enable and disable helper devices adjacent to the path
> coresight: Fix CTI module refcount leak by making it a helper device
>
> drivers/hwtracing/coresight/coresight-catu.c | 34 +-
> drivers/hwtracing/coresight/coresight-core.c | 312 +++++++++++-------
> .../hwtracing/coresight/coresight-cti-core.c | 56 ++--
> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
> drivers/hwtracing/coresight/coresight-etb10.c | 3 +-
> .../coresight/coresight-etm3x-core.c | 6 +-
> .../coresight/coresight-etm4x-core.c | 6 +-
> .../hwtracing/coresight/coresight-platform.c | 178 +++++++---
> drivers/hwtracing/coresight/coresight-priv.h | 9 +-
> drivers/hwtracing/coresight/coresight-stm.c | 6 +-
> drivers/hwtracing/coresight/coresight-sysfs.c | 9 +-
> .../hwtracing/coresight/coresight-tmc-etf.c | 2 +-
> .../hwtracing/coresight/coresight-tmc-etr.c | 89 ++---
> drivers/hwtracing/coresight/coresight-tmc.h | 2 +
> drivers/hwtracing/coresight/coresight-tpdm.c | 4 +-
> drivers/hwtracing/coresight/coresight-tpiu.c | 3 +-
> drivers/hwtracing/coresight/coresight-trbe.c | 3 +-
> drivers/hwtracing/coresight/ultrasoc-smb.c | 3 +-
> drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
> include/linux/coresight.h | 109 +++---
> 21 files changed, 530 insertions(+), 314 deletions(-)
>
> --
> 2.34.1
>
Looking at this overall - given that the only use of the in_conn is to
reference the connecting device from the helper,
i.e. coresight-catu.c:405: tmp = csdev->pdata->in_conns[i].remote_dev;
would it not be simpler to :
a)
in coresight_connection add a field:
struct coresight_device *origin_dev;
which mimics the origin / target model we already have in coresight_sysfs_link
then
b) the in_conns could simply be references to out_conn object from
origin_dev, rather than a complete coresight_connection with reversed
values, thus simplifying the in_conns handling code, and removing the
unused reversed feilds in the current in_conn object.
e.g. tmp = csdev->pdata->in_conns[i]->origin_dev
The remainder of the code would remain much the same, just adjusted
for in_conns as refs rather than independent conn objects
Regards
Mike
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 1/9] coresight: Use enum type for cs_mode wherever possible
2023-03-16 16:47 ` Suzuki K Poulose
@ 2023-03-21 15:12 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 15:12 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Thu, 16 Mar 2023 at 16:47, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > mode is stored as a local_t, but it is also passed around a lot as a
> > plain u32, so use the correct type wherever local_t isn't currently
> > used. This helps a little bit with readability.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-core.c | 7 ++++---
> > drivers/hwtracing/coresight/coresight-etb10.c | 3 ++-
> > drivers/hwtracing/coresight/coresight-etm3x-core.c | 6 +++---
> > drivers/hwtracing/coresight/coresight-etm4x-core.c | 6 +++---
> > drivers/hwtracing/coresight/coresight-priv.h | 9 ++-------
> > drivers/hwtracing/coresight/coresight-stm.c | 6 +++---
> > drivers/hwtracing/coresight/coresight-tmc-etf.c | 2 +-
> > drivers/hwtracing/coresight/coresight-tmc-etr.c | 7 +++----
> > drivers/hwtracing/coresight/coresight-tpdm.c | 4 ++--
> > drivers/hwtracing/coresight/coresight-tpiu.c | 3 ++-
> > drivers/hwtracing/coresight/coresight-trbe.c | 3 ++-
> > drivers/hwtracing/coresight/ultrasoc-smb.c | 3 ++-
> > drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
> > include/linux/coresight.h | 13 ++++++++++---
> > 14 files changed, 40 insertions(+), 34 deletions(-)
> >
>
> Looks good to me.
>
> Suzuki
>
>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 1/9] coresight: Use enum type for cs_mode wherever possible
@ 2023-03-21 15:12 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 15:12 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Thu, 16 Mar 2023 at 16:47, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > mode is stored as a local_t, but it is also passed around a lot as a
> > plain u32, so use the correct type wherever local_t isn't currently
> > used. This helps a little bit with readability.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-core.c | 7 ++++---
> > drivers/hwtracing/coresight/coresight-etb10.c | 3 ++-
> > drivers/hwtracing/coresight/coresight-etm3x-core.c | 6 +++---
> > drivers/hwtracing/coresight/coresight-etm4x-core.c | 6 +++---
> > drivers/hwtracing/coresight/coresight-priv.h | 9 ++-------
> > drivers/hwtracing/coresight/coresight-stm.c | 6 +++---
> > drivers/hwtracing/coresight/coresight-tmc-etf.c | 2 +-
> > drivers/hwtracing/coresight/coresight-tmc-etr.c | 7 +++----
> > drivers/hwtracing/coresight/coresight-tpdm.c | 4 ++--
> > drivers/hwtracing/coresight/coresight-tpiu.c | 3 ++-
> > drivers/hwtracing/coresight/coresight-trbe.c | 3 ++-
> > drivers/hwtracing/coresight/ultrasoc-smb.c | 3 ++-
> > drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
> > include/linux/coresight.h | 13 ++++++++++---
> > 14 files changed, 40 insertions(+), 34 deletions(-)
> >
>
> Looks good to me.
>
> Suzuki
>
>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 2/9] coresight: Change name of pdata->conns
2023-03-16 17:17 ` Suzuki K Poulose
@ 2023-03-21 15:12 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 15:12 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Thu, 16 Mar 2023 at 17:17, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > conns is actually for output connections. Change the name to make it
> > clearer and so that we can add input connections later.
> >
> > No functional changes.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-core.c | 26 +++++++++----------
> > .../hwtracing/coresight/coresight-platform.c | 12 ++++-----
> > .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> > include/linux/coresight.h | 2 +-
> > 4 files changed, 21 insertions(+), 21 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> > index e3b63fd52b9c..c0eda7407fb8 100644
> > --- a/drivers/hwtracing/coresight/coresight-core.c
> > +++ b/drivers/hwtracing/coresight/coresight-core.c
> > @@ -119,7 +119,7 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
> > struct coresight_connection *conn;
> >
> > for (i = 0; i < parent->pdata->nr_outport; i++) {
> > - conn = &parent->pdata->conns[i];
> > + conn = &parent->pdata->out_conns[i];
> > if (conn->child_dev == csdev)
> > return conn->child_port;
> > }
> > @@ -137,7 +137,7 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
> > struct coresight_connection *conn;
> >
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > - conn = &csdev->pdata->conns[i];
> > + conn = &csdev->pdata->out_conns[i];
> > if (conn->child_dev == child)
> > return conn->outport;
> > }
> > @@ -606,7 +606,7 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > struct coresight_device *child_dev;
> >
> > - child_dev = csdev->pdata->conns[i].child_dev;
> > + child_dev = csdev->pdata->out_conns[i].child_dev;
> > if (child_dev)
> > sink = coresight_find_enabled_sink(child_dev);
> > if (sink)
> > @@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > struct coresight_device *child;
> >
> > - child = csdev->pdata->conns[i].child_dev;
> > + child = csdev->pdata->out_conns[i].child_dev;
> > if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> > if (!coresight_get_ref(child))
> > goto err;
> > @@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> > for (i--; i >= 0; i--) {
> > struct coresight_device *child;
> >
> > - child = csdev->pdata->conns[i].child_dev;
> > + child = csdev->pdata->out_conns[i].child_dev;
> > if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> > coresight_put_ref(child);
> > }
> > @@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > struct coresight_device *child;
> >
> > - child = csdev->pdata->conns[i].child_dev;
> > + child = csdev->pdata->out_conns[i].child_dev;
> > if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> > coresight_put_ref(child);
> > }
> > @@ -794,7 +794,7 @@ static int _coresight_build_path(struct coresight_device *csdev,
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > struct coresight_device *child_dev;
> >
> > - child_dev = csdev->pdata->conns[i].child_dev;
> > + child_dev = csdev->pdata->out_conns[i].child_dev;
> > if (child_dev &&
> > _coresight_build_path(child_dev, sink, path) == 0) {
> > found = true;
> > @@ -964,7 +964,7 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
> > struct coresight_device *child_dev, *sink = NULL;
> > int child_depth = curr_depth;
> >
> > - child_dev = csdev->pdata->conns[i].child_dev;
> > + child_dev = csdev->pdata->out_conns[i].child_dev;
> > if (child_dev)
> > sink = coresight_find_sink(child_dev, &child_depth);
> >
> > @@ -1334,7 +1334,7 @@ static int coresight_orphan_match(struct device *dev, void *data)
> > * an orphan connection whose name matches @csdev, link it.
> > */
> > for (i = 0; i < i_csdev->pdata->nr_outport; i++) {
> > - conn = &i_csdev->pdata->conns[i];
> > + conn = &i_csdev->pdata->out_conns[i];
> >
> > /* Skip the port if FW doesn't describe it */
> > if (!conn->child_fwnode)
> > @@ -1375,7 +1375,7 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
> > int i, ret = 0;
> >
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > - struct coresight_connection *conn = &csdev->pdata->conns[i];
> > + struct coresight_connection *conn = &csdev->pdata->out_conns[i];
> >
> > if (!conn->child_fwnode)
> > continue;
> > @@ -1412,7 +1412,7 @@ static int coresight_remove_match(struct device *dev, void *data)
> > * a connection whose name matches @csdev, remove it.
> > */
> > for (i = 0; i < iterator->pdata->nr_outport; i++) {
> > - conn = &iterator->pdata->conns[i];
> > + conn = &iterator->pdata->out_conns[i];
> >
> > if (conn->child_dev == NULL || conn->child_fwnode == NULL)
> > continue;
> > @@ -1548,7 +1548,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> > struct coresight_platform_data *pdata)
> > {
> > int i;
> > - struct coresight_connection *conns = pdata->conns;
> > + struct coresight_connection *conns = pdata->out_conns;
> >
> > for (i = 0; i < pdata->nr_outport; i++) {
> > /* If we have made the links, remove them now */
> > @@ -1560,7 +1560,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> > */
> > if (conns[i].child_fwnode) {
> > fwnode_handle_put(conns[i].child_fwnode);
> > - pdata->conns[i].child_fwnode = NULL;
> > + pdata->out_conns[i].child_fwnode = NULL;
> > }
> > }
> > if (csdev)
> > diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> > index 475899714104..5085525a32bb 100644
> > --- a/drivers/hwtracing/coresight/coresight-platform.c
> > +++ b/drivers/hwtracing/coresight/coresight-platform.c
> > @@ -27,9 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
> > struct coresight_platform_data *pdata)
> > {
> > if (pdata->nr_outport) {
> > - pdata->conns = devm_kcalloc(dev, pdata->nr_outport,
> > - sizeof(*pdata->conns), GFP_KERNEL);
> > - if (!pdata->conns)
> > + pdata->out_conns = devm_kcalloc(dev, pdata->nr_outport,
> > + sizeof(*pdata->out_conns), GFP_KERNEL);
> > + if (!pdata->out_conns)
> > return -ENOMEM;
> > }
> >
> > @@ -251,7 +251,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
> > break;
> > }
> >
> > - conn = &pdata->conns[endpoint.port];
> > + conn = &pdata->out_conns[endpoint.port];
> > if (conn->child_fwnode) {
> > dev_warn(dev, "Duplicate output port %d\n",
> > endpoint.port);
> > @@ -744,8 +744,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
> > int port = conns[i].outport;
> >
> > /* Duplicate output port */
> > - WARN_ON(pdata->conns[port].child_fwnode);
> > - pdata->conns[port] = conns[i];
> > + WARN_ON(pdata->out_conns[port].child_fwnode);
> > + pdata->out_conns[port] = conns[i];
> > }
> >
> > devm_kfree(&adev->dev, conns);
> > diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> > index ed589cfff1b5..86d4a08aa833 100644
> > --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> > +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> > @@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
> > return NULL;
> >
> > for (i = 0; i < etr->pdata->nr_outport; i++) {
> > - tmp = etr->pdata->conns[i].child_dev;
> > + tmp = etr->pdata->out_conns[i].child_dev;
> > if (tmp && coresight_is_catu_device(tmp))
> > return tmp;
> > }
> > diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> > index 1f878e8ed8c4..322d7273e122 100644
> > --- a/include/linux/coresight.h
> > +++ b/include/linux/coresight.h
> > @@ -109,7 +109,7 @@ union coresight_dev_subtype {
> > struct coresight_platform_data {
> > int nr_inport;
> > int nr_outport;
> > - struct coresight_connection *conns;
> > + struct coresight_connection *out_conns;
>
> The comment above this structure might need to reflect the name change.
>
> Otherwise looks good to me.
>
> Suzuki
>
> > };
> >
> > /**
>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 2/9] coresight: Change name of pdata->conns
@ 2023-03-21 15:12 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 15:12 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Thu, 16 Mar 2023 at 17:17, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > conns is actually for output connections. Change the name to make it
> > clearer and so that we can add input connections later.
> >
> > No functional changes.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-core.c | 26 +++++++++----------
> > .../hwtracing/coresight/coresight-platform.c | 12 ++++-----
> > .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> > include/linux/coresight.h | 2 +-
> > 4 files changed, 21 insertions(+), 21 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> > index e3b63fd52b9c..c0eda7407fb8 100644
> > --- a/drivers/hwtracing/coresight/coresight-core.c
> > +++ b/drivers/hwtracing/coresight/coresight-core.c
> > @@ -119,7 +119,7 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
> > struct coresight_connection *conn;
> >
> > for (i = 0; i < parent->pdata->nr_outport; i++) {
> > - conn = &parent->pdata->conns[i];
> > + conn = &parent->pdata->out_conns[i];
> > if (conn->child_dev == csdev)
> > return conn->child_port;
> > }
> > @@ -137,7 +137,7 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
> > struct coresight_connection *conn;
> >
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > - conn = &csdev->pdata->conns[i];
> > + conn = &csdev->pdata->out_conns[i];
> > if (conn->child_dev == child)
> > return conn->outport;
> > }
> > @@ -606,7 +606,7 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > struct coresight_device *child_dev;
> >
> > - child_dev = csdev->pdata->conns[i].child_dev;
> > + child_dev = csdev->pdata->out_conns[i].child_dev;
> > if (child_dev)
> > sink = coresight_find_enabled_sink(child_dev);
> > if (sink)
> > @@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > struct coresight_device *child;
> >
> > - child = csdev->pdata->conns[i].child_dev;
> > + child = csdev->pdata->out_conns[i].child_dev;
> > if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> > if (!coresight_get_ref(child))
> > goto err;
> > @@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> > for (i--; i >= 0; i--) {
> > struct coresight_device *child;
> >
> > - child = csdev->pdata->conns[i].child_dev;
> > + child = csdev->pdata->out_conns[i].child_dev;
> > if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> > coresight_put_ref(child);
> > }
> > @@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > struct coresight_device *child;
> >
> > - child = csdev->pdata->conns[i].child_dev;
> > + child = csdev->pdata->out_conns[i].child_dev;
> > if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> > coresight_put_ref(child);
> > }
> > @@ -794,7 +794,7 @@ static int _coresight_build_path(struct coresight_device *csdev,
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > struct coresight_device *child_dev;
> >
> > - child_dev = csdev->pdata->conns[i].child_dev;
> > + child_dev = csdev->pdata->out_conns[i].child_dev;
> > if (child_dev &&
> > _coresight_build_path(child_dev, sink, path) == 0) {
> > found = true;
> > @@ -964,7 +964,7 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
> > struct coresight_device *child_dev, *sink = NULL;
> > int child_depth = curr_depth;
> >
> > - child_dev = csdev->pdata->conns[i].child_dev;
> > + child_dev = csdev->pdata->out_conns[i].child_dev;
> > if (child_dev)
> > sink = coresight_find_sink(child_dev, &child_depth);
> >
> > @@ -1334,7 +1334,7 @@ static int coresight_orphan_match(struct device *dev, void *data)
> > * an orphan connection whose name matches @csdev, link it.
> > */
> > for (i = 0; i < i_csdev->pdata->nr_outport; i++) {
> > - conn = &i_csdev->pdata->conns[i];
> > + conn = &i_csdev->pdata->out_conns[i];
> >
> > /* Skip the port if FW doesn't describe it */
> > if (!conn->child_fwnode)
> > @@ -1375,7 +1375,7 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
> > int i, ret = 0;
> >
> > for (i = 0; i < csdev->pdata->nr_outport; i++) {
> > - struct coresight_connection *conn = &csdev->pdata->conns[i];
> > + struct coresight_connection *conn = &csdev->pdata->out_conns[i];
> >
> > if (!conn->child_fwnode)
> > continue;
> > @@ -1412,7 +1412,7 @@ static int coresight_remove_match(struct device *dev, void *data)
> > * a connection whose name matches @csdev, remove it.
> > */
> > for (i = 0; i < iterator->pdata->nr_outport; i++) {
> > - conn = &iterator->pdata->conns[i];
> > + conn = &iterator->pdata->out_conns[i];
> >
> > if (conn->child_dev == NULL || conn->child_fwnode == NULL)
> > continue;
> > @@ -1548,7 +1548,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> > struct coresight_platform_data *pdata)
> > {
> > int i;
> > - struct coresight_connection *conns = pdata->conns;
> > + struct coresight_connection *conns = pdata->out_conns;
> >
> > for (i = 0; i < pdata->nr_outport; i++) {
> > /* If we have made the links, remove them now */
> > @@ -1560,7 +1560,7 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> > */
> > if (conns[i].child_fwnode) {
> > fwnode_handle_put(conns[i].child_fwnode);
> > - pdata->conns[i].child_fwnode = NULL;
> > + pdata->out_conns[i].child_fwnode = NULL;
> > }
> > }
> > if (csdev)
> > diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> > index 475899714104..5085525a32bb 100644
> > --- a/drivers/hwtracing/coresight/coresight-platform.c
> > +++ b/drivers/hwtracing/coresight/coresight-platform.c
> > @@ -27,9 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
> > struct coresight_platform_data *pdata)
> > {
> > if (pdata->nr_outport) {
> > - pdata->conns = devm_kcalloc(dev, pdata->nr_outport,
> > - sizeof(*pdata->conns), GFP_KERNEL);
> > - if (!pdata->conns)
> > + pdata->out_conns = devm_kcalloc(dev, pdata->nr_outport,
> > + sizeof(*pdata->out_conns), GFP_KERNEL);
> > + if (!pdata->out_conns)
> > return -ENOMEM;
> > }
> >
> > @@ -251,7 +251,7 @@ static int of_coresight_parse_endpoint(struct device *dev,
> > break;
> > }
> >
> > - conn = &pdata->conns[endpoint.port];
> > + conn = &pdata->out_conns[endpoint.port];
> > if (conn->child_fwnode) {
> > dev_warn(dev, "Duplicate output port %d\n",
> > endpoint.port);
> > @@ -744,8 +744,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
> > int port = conns[i].outport;
> >
> > /* Duplicate output port */
> > - WARN_ON(pdata->conns[port].child_fwnode);
> > - pdata->conns[port] = conns[i];
> > + WARN_ON(pdata->out_conns[port].child_fwnode);
> > + pdata->out_conns[port] = conns[i];
> > }
> >
> > devm_kfree(&adev->dev, conns);
> > diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> > index ed589cfff1b5..86d4a08aa833 100644
> > --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> > +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> > @@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
> > return NULL;
> >
> > for (i = 0; i < etr->pdata->nr_outport; i++) {
> > - tmp = etr->pdata->conns[i].child_dev;
> > + tmp = etr->pdata->out_conns[i].child_dev;
> > if (tmp && coresight_is_catu_device(tmp))
> > return tmp;
> > }
> > diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> > index 1f878e8ed8c4..322d7273e122 100644
> > --- a/include/linux/coresight.h
> > +++ b/include/linux/coresight.h
> > @@ -109,7 +109,7 @@ union coresight_dev_subtype {
> > struct coresight_platform_data {
> > int nr_inport;
> > int nr_outport;
> > - struct coresight_connection *conns;
> > + struct coresight_connection *out_conns;
>
> The comment above this structure might need to reflect the name change.
>
> Otherwise looks good to me.
>
> Suzuki
>
> > };
> >
> > /**
>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 3/9] coresight: Rename nr_outports to nr_outconns
2023-03-16 17:18 ` Suzuki K Poulose
@ 2023-03-21 17:33 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 17:33 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Thu, 16 Mar 2023 at 17:18, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > Rename to avoid confusion between port number and the index in the
> > connection array. The port number is already stored in the connection,
> > and in a later commit the connection array will be appended to, so
> > the length of it will no longer reflect the number of ports.
> >
> > No functional changes.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-core.c | 32 ++++++++---------
> > .../hwtracing/coresight/coresight-platform.c | 34 +++++++++----------
> > .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> > include/linux/coresight.h | 10 +++---
> > 4 files changed, 39 insertions(+), 39 deletions(-)
> >
>
> ...
>
> > diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> > index 322d7273e122..cdf0d1def778 100644
> > --- a/include/linux/coresight.h
> > +++ b/include/linux/coresight.h
> > @@ -102,13 +102,13 @@ union coresight_dev_subtype {
> > * struct coresight_platform_data - data harvested from the firmware
> > * specification.
> > *
> > - * @nr_inport: Number of elements for the input connections.
> > - * @nr_outport: Number of elements for the output connections.
> > - * @conns: Sparse array of nr_outport connections from this component.
> > + * @nr_inconns: Number of elements for the input connections.
> > + * @nr_outconns: Number of elements for the output connections.
> > + * @conns: Sparse array of nr_outconns connections from this component.
>
> s/conns/out_conns/
>
> Otherwise looks good to me.
>
> Suzuki
>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 3/9] coresight: Rename nr_outports to nr_outconns
@ 2023-03-21 17:33 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 17:33 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Thu, 16 Mar 2023 at 17:18, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > Rename to avoid confusion between port number and the index in the
> > connection array. The port number is already stored in the connection,
> > and in a later commit the connection array will be appended to, so
> > the length of it will no longer reflect the number of ports.
> >
> > No functional changes.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-core.c | 32 ++++++++---------
> > .../hwtracing/coresight/coresight-platform.c | 34 +++++++++----------
> > .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> > include/linux/coresight.h | 10 +++---
> > 4 files changed, 39 insertions(+), 39 deletions(-)
> >
>
> ...
>
> > diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> > index 322d7273e122..cdf0d1def778 100644
> > --- a/include/linux/coresight.h
> > +++ b/include/linux/coresight.h
> > @@ -102,13 +102,13 @@ union coresight_dev_subtype {
> > * struct coresight_platform_data - data harvested from the firmware
> > * specification.
> > *
> > - * @nr_inport: Number of elements for the input connections.
> > - * @nr_outport: Number of elements for the output connections.
> > - * @conns: Sparse array of nr_outport connections from this component.
> > + * @nr_inconns: Number of elements for the input connections.
> > + * @nr_outconns: Number of elements for the output connections.
> > + * @conns: Sparse array of nr_outconns connections from this component.
>
> s/conns/out_conns/
>
> Otherwise looks good to me.
>
> Suzuki
>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 4/9] coresight: Rename connection members to allow for input connections
2023-03-10 16:06 ` James Clark
@ 2023-03-21 17:33 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 17:33 UTC (permalink / raw)
To: James Clark
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James,
On Fri, 10 Mar 2023 at 16:06, James Clark <james.clark@arm.com> wrote:
>
> References to outport and child_port will be confusing when we use the
> same struct for inputs as well as outputs. Rename the members to refer
> to the position in terms of local and remote instead.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 70 +++++++++----------
> .../hwtracing/coresight/coresight-platform.c | 32 ++++-----
> drivers/hwtracing/coresight/coresight-sysfs.c | 10 +--
> .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> include/linux/coresight.h | 23 +++---
> 5 files changed, 71 insertions(+), 66 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index 617be08883da..f457914e445e 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -120,8 +120,8 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
>
> for (i = 0; i < parent->pdata->nr_outconns; i++) {
> conn = &parent->pdata->out_conns[i];
> - if (conn->child_dev == csdev)
> - return conn->child_port;
> + if (conn->remote_dev == csdev)
> + return conn->remote_port;
> }
>
> dev_err(&csdev->dev, "couldn't find inport, parent: %s, child: %s\n",
> @@ -138,8 +138,8 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
>
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> conn = &csdev->pdata->out_conns[i];
> - if (conn->child_dev == child)
> - return conn->outport;
> + if (conn->remote_dev == child)
> + return conn->port;
> }
>
> dev_err(&csdev->dev, "couldn't find outport, parent: %s, child: %s\n",
> @@ -604,11 +604,11 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
> * Recursively explore each port found on this element.
> */
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> - struct coresight_device *child_dev;
> + struct coresight_device *remote_dev;
>
> - child_dev = csdev->pdata->out_conns[i].child_dev;
> - if (child_dev)
> - sink = coresight_find_enabled_sink(child_dev);
> + remote_dev = csdev->pdata->out_conns[i].remote_dev;
> + if (remote_dev)
> + sink = coresight_find_enabled_sink(remote_dev);
> if (sink)
> return sink;
> }
> @@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> struct coresight_device *child;
>
> - child = csdev->pdata->out_conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].remote_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> if (!coresight_get_ref(child))
> goto err;
> @@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> for (i--; i >= 0; i--) {
> struct coresight_device *child;
>
> - child = csdev->pdata->out_conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].remote_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> coresight_put_ref(child);
> }
> @@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> struct coresight_device *child;
>
> - child = csdev->pdata->out_conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].remote_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> coresight_put_ref(child);
> }
> @@ -792,11 +792,11 @@ static int _coresight_build_path(struct coresight_device *csdev,
>
> /* Not a sink - recursively explore each port found on this element */
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> - struct coresight_device *child_dev;
> + struct coresight_device *remote_dev;
>
> - child_dev = csdev->pdata->out_conns[i].child_dev;
> - if (child_dev &&
> - _coresight_build_path(child_dev, sink, path) == 0) {
> + remote_dev = csdev->pdata->out_conns[i].remote_dev;
> + if (remote_dev &&
> + _coresight_build_path(remote_dev, sink, path) == 0) {
> found = true;
> break;
> }
> @@ -961,12 +961,12 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
> * recursively explore each port found on this element.
> */
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> - struct coresight_device *child_dev, *sink = NULL;
> + struct coresight_device *remote_dev, *sink = NULL;
> int child_depth = curr_depth;
>
> - child_dev = csdev->pdata->out_conns[i].child_dev;
> - if (child_dev)
> - sink = coresight_find_sink(child_dev, &child_depth);
> + remote_dev = csdev->pdata->out_conns[i].remote_dev;
> + if (remote_dev)
> + sink = coresight_find_sink(remote_dev, &child_depth);
>
> if (sink)
> found_sink = coresight_select_best_sink(found_sink,
> @@ -1337,12 +1337,12 @@ static int coresight_orphan_match(struct device *dev, void *data)
> conn = &i_csdev->pdata->out_conns[i];
>
> /* Skip the port if FW doesn't describe it */
> - if (!conn->child_fwnode)
> + if (!conn->remote_fwnode)
> continue;
> /* We have found at least one orphan connection */
> - if (conn->child_dev == NULL) {
> + if (conn->remote_dev == NULL) {
> /* Does it match this newly added device? */
> - if (conn->child_fwnode == csdev->dev.fwnode) {
> + if (conn->remote_fwnode == csdev->dev.fwnode) {
> ret = coresight_make_links(i_csdev,
> conn, csdev);
> if (ret)
> @@ -1377,13 +1377,13 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> struct coresight_connection *conn = &csdev->pdata->out_conns[i];
>
> - if (!conn->child_fwnode)
> + if (!conn->remote_fwnode)
> continue;
> - conn->child_dev =
> - coresight_find_csdev_by_fwnode(conn->child_fwnode);
> - if (conn->child_dev && conn->child_dev->has_conns_grp) {
> + conn->remote_dev =
> + coresight_find_csdev_by_fwnode(conn->remote_fwnode);
> + if (conn->remote_dev && conn->remote_dev->has_conns_grp) {
> ret = coresight_make_links(csdev, conn,
> - conn->child_dev);
> + conn->remote_dev);
> if (ret)
> break;
> } else {
> @@ -1414,10 +1414,10 @@ static int coresight_remove_match(struct device *dev, void *data)
> for (i = 0; i < iterator->pdata->nr_outconns; i++) {
> conn = &iterator->pdata->out_conns[i];
>
> - if (conn->child_dev == NULL || conn->child_fwnode == NULL)
> + if (conn->remote_dev == NULL || conn->remote_fwnode == NULL)
> continue;
>
> - if (csdev->dev.fwnode == conn->child_fwnode) {
> + if (csdev->dev.fwnode == conn->remote_fwnode) {
> iterator->orphan = true;
> coresight_remove_links(iterator, conn);
> /*
> @@ -1425,8 +1425,8 @@ static int coresight_remove_match(struct device *dev, void *data)
> * device acquired in parsing the connections from
> * platform data.
> */
> - fwnode_handle_put(conn->child_fwnode);
> - conn->child_fwnode = NULL;
> + fwnode_handle_put(conn->remote_fwnode);
> + conn->remote_fwnode = NULL;
> /* No need to continue */
> break;
> }
> @@ -1552,15 +1552,15 @@ void coresight_release_platform_data(struct coresight_device *csdev,
>
> for (i = 0; i < pdata->nr_outconns; i++) {
> /* If we have made the links, remove them now */
> - if (csdev && conns[i].child_dev)
> + if (csdev && conns[i].remote_dev)
> coresight_remove_links(csdev, &conns[i]);
> /*
> * Drop the refcount and clear the handle as this device
> * is going away
> */
> - if (conns[i].child_fwnode) {
> - fwnode_handle_put(conns[i].child_fwnode);
> - pdata->out_conns[i].child_fwnode = NULL;
> + if (conns[i].remote_fwnode) {
> + fwnode_handle_put(conns[i].remote_fwnode);
> + pdata->out_conns[i].remote_fwnode = NULL;
> }
> }
> if (csdev)
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index ed865e0621a9..c77238cdf448 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -252,13 +252,13 @@ static int of_coresight_parse_endpoint(struct device *dev,
> }
>
> conn = &pdata->out_conns[endpoint.port];
> - if (conn->child_fwnode) {
> + if (conn->remote_fwnode) {
> dev_warn(dev, "Duplicate output port %d\n",
> endpoint.port);
> ret = -EINVAL;
> break;
> }
> - conn->outport = endpoint.port;
> + conn->port = endpoint.port;
> /*
> * Hold the refcount to the target device. This could be
> * released via:
> @@ -267,8 +267,8 @@ static int of_coresight_parse_endpoint(struct device *dev,
> * 2) While removing the target device via
> * coresight_remove_match()
> */
> - conn->child_fwnode = fwnode_handle_get(rdev_fwnode);
> - conn->child_port = rendpoint.port;
> + conn->remote_fwnode = fwnode_handle_get(rdev_fwnode);
> + conn->remote_port = rendpoint.port;
> /* Connection record updated */
> } while (0);
>
> @@ -649,8 +649,8 @@ static int acpi_coresight_parse_link(struct acpi_device *adev,
>
> dir = fields[3].integer.value;
> if (dir == ACPI_CORESIGHT_LINK_MASTER) {
> - conn->outport = fields[0].integer.value;
> - conn->child_port = fields[1].integer.value;
> + conn->port = fields[0].integer.value;
> + conn->remote_port = fields[1].integer.value;
> rdev = coresight_find_device_by_fwnode(&r_adev->fwnode);
> if (!rdev)
> return -EPROBE_DEFER;
> @@ -662,14 +662,14 @@ static int acpi_coresight_parse_link(struct acpi_device *adev,
> * 2) While removing the target device via
> * coresight_remove_match().
> */
> - conn->child_fwnode = fwnode_handle_get(&r_adev->fwnode);
> + conn->remote_fwnode = fwnode_handle_get(&r_adev->fwnode);
> } else if (dir == ACPI_CORESIGHT_LINK_SLAVE) {
> /*
> * We are only interested in the port number
> * for the input ports at this component.
> - * Store the port number in child_port.
> + * Store the port number in remote_port.
> */
> - conn->child_port = fields[0].integer.value;
> + conn->remote_port = fields[0].integer.value;
> } else {
> /* Invalid direction */
> return -EINVAL;
> @@ -718,11 +718,11 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
> return dir;
>
> if (dir == ACPI_CORESIGHT_LINK_MASTER) {
> - if (ptr->outport >= pdata->nr_outconns)
> - pdata->nr_outconns = ptr->outport + 1;
> + if (ptr->port >= pdata->nr_outconns)
> + pdata->nr_outconns = ptr->port + 1;
> ptr++;
> } else {
> - WARN_ON(pdata->nr_inconns == ptr->child_port + 1);
> + WARN_ON(pdata->nr_inconns == ptr->remote_port + 1);
> /*
> * We do not track input port connections for a device.
> * However we need the highest port number described,
> @@ -730,8 +730,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
> * record for an output connection. Hence, do not move
> * the ptr for input connections
> */
> - if (ptr->child_port >= pdata->nr_inconns)
> - pdata->nr_inconns = ptr->child_port + 1;
> + if (ptr->remote_port >= pdata->nr_inconns)
> + pdata->nr_inconns = ptr->remote_port + 1;
> }
> }
>
> @@ -741,10 +741,10 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
>
> /* Copy the connection information to the final location */
> for (i = 0; conns + i < ptr; i++) {
> - int port = conns[i].outport;
> + int port = conns[i].port;
>
> /* Duplicate output port */
> - WARN_ON(pdata->out_conns[port].child_fwnode);
> + WARN_ON(pdata->out_conns[port].remote_fwnode);
> pdata->out_conns[port] = conns[i];
> }
>
> diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
> index 34d2a2d31d00..3da9868d9237 100644
> --- a/drivers/hwtracing/coresight/coresight-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
> @@ -150,11 +150,11 @@ int coresight_make_links(struct coresight_device *orig,
>
> do {
> outs = devm_kasprintf(&orig->dev, GFP_KERNEL,
> - "out:%d", conn->outport);
> + "out:%d", conn->port);
> if (!outs)
> break;
> ins = devm_kasprintf(&target->dev, GFP_KERNEL,
> - "in:%d", conn->child_port);
> + "in:%d", conn->remote_port);
> if (!ins)
> break;
> link = devm_kzalloc(&orig->dev,
> @@ -178,7 +178,7 @@ int coresight_make_links(struct coresight_device *orig,
> * Install the device connection. This also indicates that
> * the links are operational on both ends.
> */
> - conn->child_dev = target;
> + conn->remote_dev = target;
> return 0;
> } while (0);
>
> @@ -198,9 +198,9 @@ void coresight_remove_links(struct coresight_device *orig,
>
> coresight_remove_sysfs_link(conn->link);
>
> - devm_kfree(&conn->child_dev->dev, conn->link->target_name);
> + devm_kfree(&conn->remote_dev->dev, conn->link->target_name);
> devm_kfree(&orig->dev, conn->link->orig_name);
> devm_kfree(&orig->dev, conn->link);
> conn->link = NULL;
> - conn->child_dev = NULL;
> + conn->remote_dev = NULL;
> }
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index 4711dfa7418c..223081dc3ad4 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> @@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
> return NULL;
>
> for (i = 0; i < etr->pdata->nr_outconns; i++) {
> - tmp = etr->pdata->out_conns[i].child_dev;
> + tmp = etr->pdata->out_conns[i].remote_dev;
> if (tmp && coresight_is_catu_device(tmp))
> return tmp;
> }
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index cdf0d1def778..64bb5fc95afa 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -164,18 +164,23 @@ struct coresight_desc {
>
> /**
> * struct coresight_connection - representation of a single connection
> - * @outport: a connection's output port number.
> - * @child_port: remote component's port number @output is connected to.
> - * @chid_fwnode: remote component's fwnode handle.
> - * @child_dev: a @coresight_device representation of the component
> - connected to @outport.
> + * @port: this connection's local port number. For input
> + * connections this is the in-port. For outputs it's the
> + * out-port.
> + * @remote_port: remote component's port number @port is connected to.
> + * For input connections this is the out-port number on
> + * remote device. For output connections it's the in-port
> + * on the remote device.
> + * @remote_fwnode: remote component's fwnode handle.
> + * @remote_dev: a @coresight_device representation of the component
> + * connected to @port.
> * @link: Representation of the connection as a sysfs link.
> */
> struct coresight_connection {
> - int outport;
> - int child_port;
> - struct fwnode_handle *child_fwnode;
> - struct coresight_device *child_dev;
> + int port;
perhaps consider naming this explicitly local_port?
> + int remote_port;
> + struct fwnode_handle *remote_fwnode;
> + struct coresight_device *remote_dev;
> struct coresight_sysfs_link *link;
and while we are at it this could be sysfs_link?
> };
>
> --
> 2.34.1
>
Either way,
Reviewed-by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 4/9] coresight: Rename connection members to allow for input connections
@ 2023-03-21 17:33 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 17:33 UTC (permalink / raw)
To: James Clark
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James,
On Fri, 10 Mar 2023 at 16:06, James Clark <james.clark@arm.com> wrote:
>
> References to outport and child_port will be confusing when we use the
> same struct for inputs as well as outputs. Rename the members to refer
> to the position in terms of local and remote instead.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 70 +++++++++----------
> .../hwtracing/coresight/coresight-platform.c | 32 ++++-----
> drivers/hwtracing/coresight/coresight-sysfs.c | 10 +--
> .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
> include/linux/coresight.h | 23 +++---
> 5 files changed, 71 insertions(+), 66 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index 617be08883da..f457914e445e 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -120,8 +120,8 @@ static int coresight_find_link_inport(struct coresight_device *csdev,
>
> for (i = 0; i < parent->pdata->nr_outconns; i++) {
> conn = &parent->pdata->out_conns[i];
> - if (conn->child_dev == csdev)
> - return conn->child_port;
> + if (conn->remote_dev == csdev)
> + return conn->remote_port;
> }
>
> dev_err(&csdev->dev, "couldn't find inport, parent: %s, child: %s\n",
> @@ -138,8 +138,8 @@ static int coresight_find_link_outport(struct coresight_device *csdev,
>
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> conn = &csdev->pdata->out_conns[i];
> - if (conn->child_dev == child)
> - return conn->outport;
> + if (conn->remote_dev == child)
> + return conn->port;
> }
>
> dev_err(&csdev->dev, "couldn't find outport, parent: %s, child: %s\n",
> @@ -604,11 +604,11 @@ coresight_find_enabled_sink(struct coresight_device *csdev)
> * Recursively explore each port found on this element.
> */
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> - struct coresight_device *child_dev;
> + struct coresight_device *remote_dev;
>
> - child_dev = csdev->pdata->out_conns[i].child_dev;
> - if (child_dev)
> - sink = coresight_find_enabled_sink(child_dev);
> + remote_dev = csdev->pdata->out_conns[i].remote_dev;
> + if (remote_dev)
> + sink = coresight_find_enabled_sink(remote_dev);
> if (sink)
> return sink;
> }
> @@ -722,7 +722,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> struct coresight_device *child;
>
> - child = csdev->pdata->out_conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].remote_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> if (!coresight_get_ref(child))
> goto err;
> @@ -733,7 +733,7 @@ static int coresight_grab_device(struct coresight_device *csdev)
> for (i--; i >= 0; i--) {
> struct coresight_device *child;
>
> - child = csdev->pdata->out_conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].remote_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> coresight_put_ref(child);
> }
> @@ -752,7 +752,7 @@ static void coresight_drop_device(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> struct coresight_device *child;
>
> - child = csdev->pdata->out_conns[i].child_dev;
> + child = csdev->pdata->out_conns[i].remote_dev;
> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
> coresight_put_ref(child);
> }
> @@ -792,11 +792,11 @@ static int _coresight_build_path(struct coresight_device *csdev,
>
> /* Not a sink - recursively explore each port found on this element */
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> - struct coresight_device *child_dev;
> + struct coresight_device *remote_dev;
>
> - child_dev = csdev->pdata->out_conns[i].child_dev;
> - if (child_dev &&
> - _coresight_build_path(child_dev, sink, path) == 0) {
> + remote_dev = csdev->pdata->out_conns[i].remote_dev;
> + if (remote_dev &&
> + _coresight_build_path(remote_dev, sink, path) == 0) {
> found = true;
> break;
> }
> @@ -961,12 +961,12 @@ coresight_find_sink(struct coresight_device *csdev, int *depth)
> * recursively explore each port found on this element.
> */
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> - struct coresight_device *child_dev, *sink = NULL;
> + struct coresight_device *remote_dev, *sink = NULL;
> int child_depth = curr_depth;
>
> - child_dev = csdev->pdata->out_conns[i].child_dev;
> - if (child_dev)
> - sink = coresight_find_sink(child_dev, &child_depth);
> + remote_dev = csdev->pdata->out_conns[i].remote_dev;
> + if (remote_dev)
> + sink = coresight_find_sink(remote_dev, &child_depth);
>
> if (sink)
> found_sink = coresight_select_best_sink(found_sink,
> @@ -1337,12 +1337,12 @@ static int coresight_orphan_match(struct device *dev, void *data)
> conn = &i_csdev->pdata->out_conns[i];
>
> /* Skip the port if FW doesn't describe it */
> - if (!conn->child_fwnode)
> + if (!conn->remote_fwnode)
> continue;
> /* We have found at least one orphan connection */
> - if (conn->child_dev == NULL) {
> + if (conn->remote_dev == NULL) {
> /* Does it match this newly added device? */
> - if (conn->child_fwnode == csdev->dev.fwnode) {
> + if (conn->remote_fwnode == csdev->dev.fwnode) {
> ret = coresight_make_links(i_csdev,
> conn, csdev);
> if (ret)
> @@ -1377,13 +1377,13 @@ static int coresight_fixup_device_conns(struct coresight_device *csdev)
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> struct coresight_connection *conn = &csdev->pdata->out_conns[i];
>
> - if (!conn->child_fwnode)
> + if (!conn->remote_fwnode)
> continue;
> - conn->child_dev =
> - coresight_find_csdev_by_fwnode(conn->child_fwnode);
> - if (conn->child_dev && conn->child_dev->has_conns_grp) {
> + conn->remote_dev =
> + coresight_find_csdev_by_fwnode(conn->remote_fwnode);
> + if (conn->remote_dev && conn->remote_dev->has_conns_grp) {
> ret = coresight_make_links(csdev, conn,
> - conn->child_dev);
> + conn->remote_dev);
> if (ret)
> break;
> } else {
> @@ -1414,10 +1414,10 @@ static int coresight_remove_match(struct device *dev, void *data)
> for (i = 0; i < iterator->pdata->nr_outconns; i++) {
> conn = &iterator->pdata->out_conns[i];
>
> - if (conn->child_dev == NULL || conn->child_fwnode == NULL)
> + if (conn->remote_dev == NULL || conn->remote_fwnode == NULL)
> continue;
>
> - if (csdev->dev.fwnode == conn->child_fwnode) {
> + if (csdev->dev.fwnode == conn->remote_fwnode) {
> iterator->orphan = true;
> coresight_remove_links(iterator, conn);
> /*
> @@ -1425,8 +1425,8 @@ static int coresight_remove_match(struct device *dev, void *data)
> * device acquired in parsing the connections from
> * platform data.
> */
> - fwnode_handle_put(conn->child_fwnode);
> - conn->child_fwnode = NULL;
> + fwnode_handle_put(conn->remote_fwnode);
> + conn->remote_fwnode = NULL;
> /* No need to continue */
> break;
> }
> @@ -1552,15 +1552,15 @@ void coresight_release_platform_data(struct coresight_device *csdev,
>
> for (i = 0; i < pdata->nr_outconns; i++) {
> /* If we have made the links, remove them now */
> - if (csdev && conns[i].child_dev)
> + if (csdev && conns[i].remote_dev)
> coresight_remove_links(csdev, &conns[i]);
> /*
> * Drop the refcount and clear the handle as this device
> * is going away
> */
> - if (conns[i].child_fwnode) {
> - fwnode_handle_put(conns[i].child_fwnode);
> - pdata->out_conns[i].child_fwnode = NULL;
> + if (conns[i].remote_fwnode) {
> + fwnode_handle_put(conns[i].remote_fwnode);
> + pdata->out_conns[i].remote_fwnode = NULL;
> }
> }
> if (csdev)
> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> index ed865e0621a9..c77238cdf448 100644
> --- a/drivers/hwtracing/coresight/coresight-platform.c
> +++ b/drivers/hwtracing/coresight/coresight-platform.c
> @@ -252,13 +252,13 @@ static int of_coresight_parse_endpoint(struct device *dev,
> }
>
> conn = &pdata->out_conns[endpoint.port];
> - if (conn->child_fwnode) {
> + if (conn->remote_fwnode) {
> dev_warn(dev, "Duplicate output port %d\n",
> endpoint.port);
> ret = -EINVAL;
> break;
> }
> - conn->outport = endpoint.port;
> + conn->port = endpoint.port;
> /*
> * Hold the refcount to the target device. This could be
> * released via:
> @@ -267,8 +267,8 @@ static int of_coresight_parse_endpoint(struct device *dev,
> * 2) While removing the target device via
> * coresight_remove_match()
> */
> - conn->child_fwnode = fwnode_handle_get(rdev_fwnode);
> - conn->child_port = rendpoint.port;
> + conn->remote_fwnode = fwnode_handle_get(rdev_fwnode);
> + conn->remote_port = rendpoint.port;
> /* Connection record updated */
> } while (0);
>
> @@ -649,8 +649,8 @@ static int acpi_coresight_parse_link(struct acpi_device *adev,
>
> dir = fields[3].integer.value;
> if (dir == ACPI_CORESIGHT_LINK_MASTER) {
> - conn->outport = fields[0].integer.value;
> - conn->child_port = fields[1].integer.value;
> + conn->port = fields[0].integer.value;
> + conn->remote_port = fields[1].integer.value;
> rdev = coresight_find_device_by_fwnode(&r_adev->fwnode);
> if (!rdev)
> return -EPROBE_DEFER;
> @@ -662,14 +662,14 @@ static int acpi_coresight_parse_link(struct acpi_device *adev,
> * 2) While removing the target device via
> * coresight_remove_match().
> */
> - conn->child_fwnode = fwnode_handle_get(&r_adev->fwnode);
> + conn->remote_fwnode = fwnode_handle_get(&r_adev->fwnode);
> } else if (dir == ACPI_CORESIGHT_LINK_SLAVE) {
> /*
> * We are only interested in the port number
> * for the input ports at this component.
> - * Store the port number in child_port.
> + * Store the port number in remote_port.
> */
> - conn->child_port = fields[0].integer.value;
> + conn->remote_port = fields[0].integer.value;
> } else {
> /* Invalid direction */
> return -EINVAL;
> @@ -718,11 +718,11 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
> return dir;
>
> if (dir == ACPI_CORESIGHT_LINK_MASTER) {
> - if (ptr->outport >= pdata->nr_outconns)
> - pdata->nr_outconns = ptr->outport + 1;
> + if (ptr->port >= pdata->nr_outconns)
> + pdata->nr_outconns = ptr->port + 1;
> ptr++;
> } else {
> - WARN_ON(pdata->nr_inconns == ptr->child_port + 1);
> + WARN_ON(pdata->nr_inconns == ptr->remote_port + 1);
> /*
> * We do not track input port connections for a device.
> * However we need the highest port number described,
> @@ -730,8 +730,8 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
> * record for an output connection. Hence, do not move
> * the ptr for input connections
> */
> - if (ptr->child_port >= pdata->nr_inconns)
> - pdata->nr_inconns = ptr->child_port + 1;
> + if (ptr->remote_port >= pdata->nr_inconns)
> + pdata->nr_inconns = ptr->remote_port + 1;
> }
> }
>
> @@ -741,10 +741,10 @@ static int acpi_coresight_parse_graph(struct acpi_device *adev,
>
> /* Copy the connection information to the final location */
> for (i = 0; conns + i < ptr; i++) {
> - int port = conns[i].outport;
> + int port = conns[i].port;
>
> /* Duplicate output port */
> - WARN_ON(pdata->out_conns[port].child_fwnode);
> + WARN_ON(pdata->out_conns[port].remote_fwnode);
> pdata->out_conns[port] = conns[i];
> }
>
> diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
> index 34d2a2d31d00..3da9868d9237 100644
> --- a/drivers/hwtracing/coresight/coresight-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
> @@ -150,11 +150,11 @@ int coresight_make_links(struct coresight_device *orig,
>
> do {
> outs = devm_kasprintf(&orig->dev, GFP_KERNEL,
> - "out:%d", conn->outport);
> + "out:%d", conn->port);
> if (!outs)
> break;
> ins = devm_kasprintf(&target->dev, GFP_KERNEL,
> - "in:%d", conn->child_port);
> + "in:%d", conn->remote_port);
> if (!ins)
> break;
> link = devm_kzalloc(&orig->dev,
> @@ -178,7 +178,7 @@ int coresight_make_links(struct coresight_device *orig,
> * Install the device connection. This also indicates that
> * the links are operational on both ends.
> */
> - conn->child_dev = target;
> + conn->remote_dev = target;
> return 0;
> } while (0);
>
> @@ -198,9 +198,9 @@ void coresight_remove_links(struct coresight_device *orig,
>
> coresight_remove_sysfs_link(conn->link);
>
> - devm_kfree(&conn->child_dev->dev, conn->link->target_name);
> + devm_kfree(&conn->remote_dev->dev, conn->link->target_name);
> devm_kfree(&orig->dev, conn->link->orig_name);
> devm_kfree(&orig->dev, conn->link);
> conn->link = NULL;
> - conn->child_dev = NULL;
> + conn->remote_dev = NULL;
> }
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index 4711dfa7418c..223081dc3ad4 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> @@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
> return NULL;
>
> for (i = 0; i < etr->pdata->nr_outconns; i++) {
> - tmp = etr->pdata->out_conns[i].child_dev;
> + tmp = etr->pdata->out_conns[i].remote_dev;
> if (tmp && coresight_is_catu_device(tmp))
> return tmp;
> }
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index cdf0d1def778..64bb5fc95afa 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -164,18 +164,23 @@ struct coresight_desc {
>
> /**
> * struct coresight_connection - representation of a single connection
> - * @outport: a connection's output port number.
> - * @child_port: remote component's port number @output is connected to.
> - * @chid_fwnode: remote component's fwnode handle.
> - * @child_dev: a @coresight_device representation of the component
> - connected to @outport.
> + * @port: this connection's local port number. For input
> + * connections this is the in-port. For outputs it's the
> + * out-port.
> + * @remote_port: remote component's port number @port is connected to.
> + * For input connections this is the out-port number on
> + * remote device. For output connections it's the in-port
> + * on the remote device.
> + * @remote_fwnode: remote component's fwnode handle.
> + * @remote_dev: a @coresight_device representation of the component
> + * connected to @port.
> * @link: Representation of the connection as a sysfs link.
> */
> struct coresight_connection {
> - int outport;
> - int child_port;
> - struct fwnode_handle *child_fwnode;
> - struct coresight_device *child_dev;
> + int port;
perhaps consider naming this explicitly local_port?
> + int remote_port;
> + struct fwnode_handle *remote_fwnode;
> + struct coresight_device *remote_dev;
> struct coresight_sysfs_link *link;
and while we are at it this could be sysfs_link?
> };
>
> --
> 2.34.1
>
Either way,
Reviewed-by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
2023-03-16 20:23 ` Suzuki K Poulose
@ 2023-03-21 17:56 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 17:56 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James
On Thu, 16 Mar 2023 at 20:23, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > This will allow CATU to get its associated ETR in a generic way where
> > currently the enable path has some hard coded searches which avoid
> > the need to store input connections.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
> > .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
> > drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
> > include/linux/coresight.h | 25 ++++++++
> > 4 files changed, 130 insertions(+), 13 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> > index f457914e445e..a8ba7493c09a 100644
> > --- a/drivers/hwtracing/coresight/coresight-core.c
> > +++ b/drivers/hwtracing/coresight/coresight-core.c
> > @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fff
> > EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
> >
> > static const struct cti_assoc_op *cti_assoc_ops;
> > +static int coresight_fixup_inputs(struct coresight_device *csdev);
> >
> > ssize_t coresight_simple_show_pair(struct device *_dev,
> > struct device_attribute *attr, char *buf)
> > @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct coresight_device *csdev)
> > csdev, coresight_orphan_match);
> > }
> >
> > +/*
> > + * Device connections are discovered before one/both devices have been created,
> > + * so inputs must be added later.
> > + */
> > +static int coresight_fixup_inputs(struct coresight_device *csdev)
> > +{
> > + int i, ret = 0;
> > + struct coresight_connection *out_conn;
> > + struct coresight_connection in_conn;
> > +
> > + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> > + out_conn = &csdev->pdata->out_conns[i];
> > + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
> > + continue;
> > +
> > + /* Reverse local/remote relationships for inputs */
> > + in_conn.remote_dev = csdev;
> > + in_conn.remote_port = out_conn->port;
> > + in_conn.port = out_conn->remote_port;
> > + in_conn.remote_fwnode = csdev->dev.fwnode;
> > + ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
> > + out_conn->remote_dev->pdata,
> > + &in_conn);
> > + if (ret)
> > + return ret;
> > + }
> > +
> > + return 0;
> > +}
> >
> > static int coresight_fixup_device_conns(struct coresight_device *csdev)
> > {
> > @@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct device *dev, void *data)
>
> Removing the connection should be much simpler now with the inports
> connection tracking.
>
> i.e., coresight_remove_conns() need not iterate over all the devices on
> the coresight bus. We had to do that to find all the devices connecting
> to the removed device.
>
> Instead with the in_conns we could readily get the csdev instance and
> remove the connection from those devices.
>
> Similarly for the out_conns can give you the remove device for removing
> the in_conns on that device.
>
> i.e,
>
> static void coresight_remove_conns(csdev)
> {
> struct coresight_device *remote;
>
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> remote = csdev->data->out_conns[i].remote_dev;
>
> if (!remote)
> continue;
>
> for (j = 0; j < remote->pdata->nr_inconns; j++) {
> if (remote->pdata->in_conns[i].remote_dev == csdev)
> /* Clear in_conns[i] on remote */
>
> }
> }
>
> Similarly for in_conns
> for (i = 0; i < csdev->pdata->nr_inconns; i++) {
> remote = csdev->data->in_conns[i].remote_dev;
>
> if (!remote)
> continue;
>
> for (j = 0; j < remote->pdata->nr_outconns; j++) {
> if (remote->pdata->out_conns[i].remote_dev == csdev)
> /* Clear out_conns[i] on remote */
> }
>
>
> > */
> > fwnode_handle_put(conn->remote_fwnode);
> > conn->remote_fwnode = NULL;
> > + conn->remote_dev = NULL;
> > + /* No need to continue */
> > + break;
> > + }
> > + }
> > + for (i = 0; i < iterator->pdata->nr_inconns; i++) {
> > + conn = &iterator->pdata->in_conns[i];
> > + if (csdev == conn->remote_dev) {
> > + conn->remote_fwnode = NULL;
> > + conn->remote_dev = NULL;
> > /* No need to continue */
> > break;
> > }
> > }
> > -
> > /*
> > * Returning '0' ensures that all known component on the
> > * bus will be checked.
> > @@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> >
> > for (i = 0; i < pdata->nr_outconns; i++) {
> > /* If we have made the links, remove them now */
> > - if (csdev && conns[i].remote_dev)
> > + if (csdev && conns[i].remote_dev) {
> > coresight_remove_links(csdev, &conns[i]);
> > + conns[i].remote_dev = NULL;
> > + }
> > +
> > /*
> > * Drop the refcount and clear the handle as this device
> > * is going away
> > */
> > if (conns[i].remote_fwnode) {
> > fwnode_handle_put(conns[i].remote_fwnode);
> > - pdata->out_conns[i].remote_fwnode = NULL;
> > + conns[i].remote_fwnode = NULL;
> > }
> > }
> > + for (i = 0; i < pdata->nr_inconns; i++) {
> > + pdata->in_conns[i].remote_dev = NULL;
> > + pdata->in_conns[i].remote_fwnode = NULL;
> > + }
> > +
> > if (csdev)
> > coresight_remove_conns_sysfs_group(csdev);
> > }
> > -
> > struct coresight_device *coresight_register(struct coresight_desc *desc)
> > {
> > int ret;
> > @@ -1659,6 +1705,8 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
> > ret = coresight_create_conns_sysfs_group(csdev);
> > if (!ret)
> > ret = coresight_fixup_device_conns(csdev);
> > + if (!ret)
> > + ret = coresight_fixup_inputs(csdev);
> Don't we also need to fixup the "input" connections on this
> device ? It would be good move all of this connection stuff
> into a wrapper function.
>
> > if (!ret)
> > ret = coresight_fixup_orphan_conns(csdev);
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> > index 16553f7dde12..20e3351cbdc2 100644
> > --- a/drivers/hwtracing/coresight/coresight-platform.c
> > +++ b/drivers/hwtracing/coresight/coresight-platform.c
> > @@ -20,8 +20,7 @@
> >
> > #include "coresight-priv.h"
> > /*
> > - * coresight_alloc_conns: Allocate connections record for each output
> > - * port from the device.
> > + * coresight_alloc_conns: Allocate connections record for each input/output device.
> > */
> > static int coresight_alloc_conns(struct device *dev,
> > struct coresight_platform_data *pdata)
> > @@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
> > if (!pdata->out_conns)
> > return -ENOMEM;
> > }
> > -
> > + if (pdata->nr_inconns) {
> > + pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
> > + pdata->nr_inconns,
> > + sizeof(*pdata->in_conns),
> > + GFP_KERNEL | __GFP_ZERO);
> > + if (!pdata->in_conns)
> > + return -ENOMEM;
> > + }
>
> devm_krealloc_array() skips any allocation if the requested new_size <=
> origin_size. So this should be fine for the out_conns, to not realloc
> the space again unnecessarily.
>
> > return 0;
> > }
> >
> > @@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
> > }
> > EXPORT_SYMBOL_GPL(coresight_add_conn);
>
> Should this be named as add_out_conn in the previous patch
> just to make it explicit and easier to read.
>
> >
> > +/*
> > + * Add a connection in the first free slot, or realloc
> > + * if there is no space.
> > + *
> > + * Do nothing if the connection already exists because inputs are
> > + * fixed up multiple times.
> > + */
> > +int coresight_add_in_conn(struct device *dev,
> > + struct coresight_platform_data *pdata,
>
> minor nit: Is there any reason why this can't simply be:
>
> int coresight_add_in_conn(struct coresight_device *csdev,
> struct coresight_connection *conn)
> {
> struct device *dev = csdev->dev.parent;
> struct coresight_platform_data *pdata = csdev->pdata;
>
> > + struct coresight_connection *conn)
>
>
> Suzuki
>
>
> > +{
> > + int ret;
> > + struct coresight_connection *free_conn = NULL;
> > + int i;
> > +
> > + /* Search for a free slot or exit if a duplicate is found */
> > + if (pdata->in_conns) {
> > + for (i = 0; i < pdata->nr_inconns; ++i) {
> > + if (!free_conn && !pdata->in_conns[i].remote_fwnode)
> > + free_conn = &pdata->in_conns[i];
> > + if (pdata->in_conns[i].remote_fwnode ==
> > + conn->remote_fwnode)
> > + return 0;
> > + }
> > + }
> > +
As with the out conns this assumes there can never be a match after a free slot.
Mike
> > + if (!free_conn) {
> > + pdata->nr_inconns++;
> > + ret = coresight_alloc_conns(dev, pdata);
> > + if (ret)
> > + return ret;
> > + free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
> > + }
> > +
> > + *free_conn = *conn;
> > + return 0;
> > +}
> > +EXPORT_SYMBOL_GPL(coresight_add_in_conn);
> > +
> > static struct device *
> > coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
> > {
> > @@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
> >
> > /*
> > * of_coresight_parse_endpoint : Parse the given output endpoint @ep
> > - * and fill the connection information in @conn
> > + * and fill the connection information in @in_conn and @out_conn
> > *
> > * Parses the local port, remote device name and the remote port.
> > *
> > @@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct device *dev,
> > /* Get the number of input and output port for this component */
> > of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
> >
> > - /* If there are no output connections, we are done */
> > - if (!pdata->nr_outconns)
> > - return 0;
> > -
> > ret = coresight_alloc_conns(dev, pdata);
> > if (ret)
> > return ret;
> >
> > + /* If there are no output connections, we are done */
> > + if (!pdata->nr_outconns)
> > + return 0;
> > +
> > parent = of_coresight_get_output_ports_node(node);
> > /*
> > * If the DT uses obsoleted bindings, the ports are listed
> > diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
> > index 3da9868d9237..2abf9639ac0f 100644
> > --- a/drivers/hwtracing/coresight/coresight-sysfs.c
> > +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
> > @@ -202,5 +202,4 @@ void coresight_remove_links(struct coresight_device *orig,
> > devm_kfree(&orig->dev, conn->link->orig_name);
> > devm_kfree(&orig->dev, conn->link);
> > conn->link = NULL;
> > - conn->remote_dev = NULL;
> > }
> > diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> > index 47fa58d6981d..fd268b24c761 100644
> > --- a/include/linux/coresight.h
> > +++ b/include/linux/coresight.h
> > @@ -110,6 +110,7 @@ struct coresight_platform_data {
> > int nr_inconns;
> > int nr_outconns;
> > struct coresight_connection *out_conns;
> > + struct coresight_connection *in_conns;
>
> Please document the field above the structure.
>
> Suzuki
>
>
> > };
> >
> > /**
> > @@ -177,6 +178,27 @@ struct coresight_desc {
> > * @remote_fwnode once the remote's coresight_device has
> > * been created.
> > * @link: Representation of the connection as a sysfs link.
> > + *
> > + * The full connection structure looks like this, where in_conns store references to
> > + * the parent device in the same remote_dev member as output connections.
> > + *
> > + * +-----------------------------+ +-----------------------------+
> > + * |coresight_device | |coresight_connection |
> > + * |-----------------------------| |-----------------------------|
> > + * ---->| | | |
> > + * | | | | remote_dev*|------
> > + * | |pdata->out_conns[nr_outconns]|----------->| | |
> > + * | | | | | |
> > + * | +-----------------------------+ +-----------------------------+ |
> > + * | |
> > + * | +-----------------------------+ +-----------------------------+ |
> > + * | |coresight_connection | |coresight_device | |
> > + * | |-----------------------------| |------------------------------ |
> > + * | | | | |<-----
> > + * -----|remote_dev* | | |
> > + * | |<-----------|pdata->in_conns[nr_inconns] |
> > + * | | | |
> > + * +-----------------------------+ +-----------------------------+
> > */
> > struct coresight_connection {
> > int port;
> > @@ -619,5 +641,8 @@ struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
> > int coresight_add_conn(struct device *dev,
> > struct coresight_platform_data *pdata,
> > const struct coresight_connection *conn);
> > +int coresight_add_in_conn(struct device *dev,
> > + struct coresight_platform_data *pdata,
> > + struct coresight_connection *conn);
> >
> > #endif /* _LINUX_COREISGHT_H */
>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
@ 2023-03-21 17:56 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 17:56 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James
On Thu, 16 Mar 2023 at 20:23, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > This will allow CATU to get its associated ETR in a generic way where
> > currently the enable path has some hard coded searches which avoid
> > the need to store input connections.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
> > .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
> > drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
> > include/linux/coresight.h | 25 ++++++++
> > 4 files changed, 130 insertions(+), 13 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> > index f457914e445e..a8ba7493c09a 100644
> > --- a/drivers/hwtracing/coresight/coresight-core.c
> > +++ b/drivers/hwtracing/coresight/coresight-core.c
> > @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fff
> > EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
> >
> > static const struct cti_assoc_op *cti_assoc_ops;
> > +static int coresight_fixup_inputs(struct coresight_device *csdev);
> >
> > ssize_t coresight_simple_show_pair(struct device *_dev,
> > struct device_attribute *attr, char *buf)
> > @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct coresight_device *csdev)
> > csdev, coresight_orphan_match);
> > }
> >
> > +/*
> > + * Device connections are discovered before one/both devices have been created,
> > + * so inputs must be added later.
> > + */
> > +static int coresight_fixup_inputs(struct coresight_device *csdev)
> > +{
> > + int i, ret = 0;
> > + struct coresight_connection *out_conn;
> > + struct coresight_connection in_conn;
> > +
> > + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> > + out_conn = &csdev->pdata->out_conns[i];
> > + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
> > + continue;
> > +
> > + /* Reverse local/remote relationships for inputs */
> > + in_conn.remote_dev = csdev;
> > + in_conn.remote_port = out_conn->port;
> > + in_conn.port = out_conn->remote_port;
> > + in_conn.remote_fwnode = csdev->dev.fwnode;
> > + ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
> > + out_conn->remote_dev->pdata,
> > + &in_conn);
> > + if (ret)
> > + return ret;
> > + }
> > +
> > + return 0;
> > +}
> >
> > static int coresight_fixup_device_conns(struct coresight_device *csdev)
> > {
> > @@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct device *dev, void *data)
>
> Removing the connection should be much simpler now with the inports
> connection tracking.
>
> i.e., coresight_remove_conns() need not iterate over all the devices on
> the coresight bus. We had to do that to find all the devices connecting
> to the removed device.
>
> Instead with the in_conns we could readily get the csdev instance and
> remove the connection from those devices.
>
> Similarly for the out_conns can give you the remove device for removing
> the in_conns on that device.
>
> i.e,
>
> static void coresight_remove_conns(csdev)
> {
> struct coresight_device *remote;
>
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> remote = csdev->data->out_conns[i].remote_dev;
>
> if (!remote)
> continue;
>
> for (j = 0; j < remote->pdata->nr_inconns; j++) {
> if (remote->pdata->in_conns[i].remote_dev == csdev)
> /* Clear in_conns[i] on remote */
>
> }
> }
>
> Similarly for in_conns
> for (i = 0; i < csdev->pdata->nr_inconns; i++) {
> remote = csdev->data->in_conns[i].remote_dev;
>
> if (!remote)
> continue;
>
> for (j = 0; j < remote->pdata->nr_outconns; j++) {
> if (remote->pdata->out_conns[i].remote_dev == csdev)
> /* Clear out_conns[i] on remote */
> }
>
>
> > */
> > fwnode_handle_put(conn->remote_fwnode);
> > conn->remote_fwnode = NULL;
> > + conn->remote_dev = NULL;
> > + /* No need to continue */
> > + break;
> > + }
> > + }
> > + for (i = 0; i < iterator->pdata->nr_inconns; i++) {
> > + conn = &iterator->pdata->in_conns[i];
> > + if (csdev == conn->remote_dev) {
> > + conn->remote_fwnode = NULL;
> > + conn->remote_dev = NULL;
> > /* No need to continue */
> > break;
> > }
> > }
> > -
> > /*
> > * Returning '0' ensures that all known component on the
> > * bus will be checked.
> > @@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct coresight_device *csdev,
> >
> > for (i = 0; i < pdata->nr_outconns; i++) {
> > /* If we have made the links, remove them now */
> > - if (csdev && conns[i].remote_dev)
> > + if (csdev && conns[i].remote_dev) {
> > coresight_remove_links(csdev, &conns[i]);
> > + conns[i].remote_dev = NULL;
> > + }
> > +
> > /*
> > * Drop the refcount and clear the handle as this device
> > * is going away
> > */
> > if (conns[i].remote_fwnode) {
> > fwnode_handle_put(conns[i].remote_fwnode);
> > - pdata->out_conns[i].remote_fwnode = NULL;
> > + conns[i].remote_fwnode = NULL;
> > }
> > }
> > + for (i = 0; i < pdata->nr_inconns; i++) {
> > + pdata->in_conns[i].remote_dev = NULL;
> > + pdata->in_conns[i].remote_fwnode = NULL;
> > + }
> > +
> > if (csdev)
> > coresight_remove_conns_sysfs_group(csdev);
> > }
> > -
> > struct coresight_device *coresight_register(struct coresight_desc *desc)
> > {
> > int ret;
> > @@ -1659,6 +1705,8 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
> > ret = coresight_create_conns_sysfs_group(csdev);
> > if (!ret)
> > ret = coresight_fixup_device_conns(csdev);
> > + if (!ret)
> > + ret = coresight_fixup_inputs(csdev);
> Don't we also need to fixup the "input" connections on this
> device ? It would be good move all of this connection stuff
> into a wrapper function.
>
> > if (!ret)
> > ret = coresight_fixup_orphan_conns(csdev);
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> > index 16553f7dde12..20e3351cbdc2 100644
> > --- a/drivers/hwtracing/coresight/coresight-platform.c
> > +++ b/drivers/hwtracing/coresight/coresight-platform.c
> > @@ -20,8 +20,7 @@
> >
> > #include "coresight-priv.h"
> > /*
> > - * coresight_alloc_conns: Allocate connections record for each output
> > - * port from the device.
> > + * coresight_alloc_conns: Allocate connections record for each input/output device.
> > */
> > static int coresight_alloc_conns(struct device *dev,
> > struct coresight_platform_data *pdata)
> > @@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
> > if (!pdata->out_conns)
> > return -ENOMEM;
> > }
> > -
> > + if (pdata->nr_inconns) {
> > + pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
> > + pdata->nr_inconns,
> > + sizeof(*pdata->in_conns),
> > + GFP_KERNEL | __GFP_ZERO);
> > + if (!pdata->in_conns)
> > + return -ENOMEM;
> > + }
>
> devm_krealloc_array() skips any allocation if the requested new_size <=
> origin_size. So this should be fine for the out_conns, to not realloc
> the space again unnecessarily.
>
> > return 0;
> > }
> >
> > @@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
> > }
> > EXPORT_SYMBOL_GPL(coresight_add_conn);
>
> Should this be named as add_out_conn in the previous patch
> just to make it explicit and easier to read.
>
> >
> > +/*
> > + * Add a connection in the first free slot, or realloc
> > + * if there is no space.
> > + *
> > + * Do nothing if the connection already exists because inputs are
> > + * fixed up multiple times.
> > + */
> > +int coresight_add_in_conn(struct device *dev,
> > + struct coresight_platform_data *pdata,
>
> minor nit: Is there any reason why this can't simply be:
>
> int coresight_add_in_conn(struct coresight_device *csdev,
> struct coresight_connection *conn)
> {
> struct device *dev = csdev->dev.parent;
> struct coresight_platform_data *pdata = csdev->pdata;
>
> > + struct coresight_connection *conn)
>
>
> Suzuki
>
>
> > +{
> > + int ret;
> > + struct coresight_connection *free_conn = NULL;
> > + int i;
> > +
> > + /* Search for a free slot or exit if a duplicate is found */
> > + if (pdata->in_conns) {
> > + for (i = 0; i < pdata->nr_inconns; ++i) {
> > + if (!free_conn && !pdata->in_conns[i].remote_fwnode)
> > + free_conn = &pdata->in_conns[i];
> > + if (pdata->in_conns[i].remote_fwnode ==
> > + conn->remote_fwnode)
> > + return 0;
> > + }
> > + }
> > +
As with the out conns this assumes there can never be a match after a free slot.
Mike
> > + if (!free_conn) {
> > + pdata->nr_inconns++;
> > + ret = coresight_alloc_conns(dev, pdata);
> > + if (ret)
> > + return ret;
> > + free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
> > + }
> > +
> > + *free_conn = *conn;
> > + return 0;
> > +}
> > +EXPORT_SYMBOL_GPL(coresight_add_in_conn);
> > +
> > static struct device *
> > coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
> > {
> > @@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
> >
> > /*
> > * of_coresight_parse_endpoint : Parse the given output endpoint @ep
> > - * and fill the connection information in @conn
> > + * and fill the connection information in @in_conn and @out_conn
> > *
> > * Parses the local port, remote device name and the remote port.
> > *
> > @@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct device *dev,
> > /* Get the number of input and output port for this component */
> > of_coresight_get_ports(node, &pdata->nr_inconns, &pdata->nr_outconns);
> >
> > - /* If there are no output connections, we are done */
> > - if (!pdata->nr_outconns)
> > - return 0;
> > -
> > ret = coresight_alloc_conns(dev, pdata);
> > if (ret)
> > return ret;
> >
> > + /* If there are no output connections, we are done */
> > + if (!pdata->nr_outconns)
> > + return 0;
> > +
> > parent = of_coresight_get_output_ports_node(node);
> > /*
> > * If the DT uses obsoleted bindings, the ports are listed
> > diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c b/drivers/hwtracing/coresight/coresight-sysfs.c
> > index 3da9868d9237..2abf9639ac0f 100644
> > --- a/drivers/hwtracing/coresight/coresight-sysfs.c
> > +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
> > @@ -202,5 +202,4 @@ void coresight_remove_links(struct coresight_device *orig,
> > devm_kfree(&orig->dev, conn->link->orig_name);
> > devm_kfree(&orig->dev, conn->link);
> > conn->link = NULL;
> > - conn->remote_dev = NULL;
> > }
> > diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> > index 47fa58d6981d..fd268b24c761 100644
> > --- a/include/linux/coresight.h
> > +++ b/include/linux/coresight.h
> > @@ -110,6 +110,7 @@ struct coresight_platform_data {
> > int nr_inconns;
> > int nr_outconns;
> > struct coresight_connection *out_conns;
> > + struct coresight_connection *in_conns;
>
> Please document the field above the structure.
>
> Suzuki
>
>
> > };
> >
> > /**
> > @@ -177,6 +178,27 @@ struct coresight_desc {
> > * @remote_fwnode once the remote's coresight_device has
> > * been created.
> > * @link: Representation of the connection as a sysfs link.
> > + *
> > + * The full connection structure looks like this, where in_conns store references to
> > + * the parent device in the same remote_dev member as output connections.
> > + *
> > + * +-----------------------------+ +-----------------------------+
> > + * |coresight_device | |coresight_connection |
> > + * |-----------------------------| |-----------------------------|
> > + * ---->| | | |
> > + * | | | | remote_dev*|------
> > + * | |pdata->out_conns[nr_outconns]|----------->| | |
> > + * | | | | | |
> > + * | +-----------------------------+ +-----------------------------+ |
> > + * | |
> > + * | +-----------------------------+ +-----------------------------+ |
> > + * | |coresight_connection | |coresight_device | |
> > + * | |-----------------------------| |------------------------------ |
> > + * | | | | |<-----
> > + * -----|remote_dev* | | |
> > + * | |<-----------|pdata->in_conns[nr_inconns] |
> > + * | | | |
> > + * +-----------------------------+ +-----------------------------+
> > */
> > struct coresight_connection {
> > int port;
> > @@ -619,5 +641,8 @@ struct coresight_platform_data *coresight_get_platform_data(struct device *dev);
> > int coresight_add_conn(struct device *dev,
> > struct coresight_platform_data *pdata,
> > const struct coresight_connection *conn);
> > +int coresight_add_in_conn(struct device *dev,
> > + struct coresight_platform_data *pdata,
> > + struct coresight_connection *conn);
> >
> > #endif /* _LINUX_COREISGHT_H */
>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 5/9] coresight: Dynamically add connections
2023-03-16 17:12 ` Suzuki K Poulose
@ 2023-03-21 17:56 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 17:56 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James
On Thu, 16 Mar 2023 at 17:12, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > Add a function for adding connections dynamically. This also removes
> > the 1:1 mapping between port number and the index into the connections
> > array. The only place this mapping was used was in the warning for
> > duplicate output ports, which has been replaced by a search. Other
> > uses of the port number already use the port member variable.
> >
> > Being able to dynamically add connections will allow other devices like
> > CTI to re-use the connection mechanism despite not having explicit
> > connections described in the DT.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > .../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
> > include/linux/coresight.h | 7 +-
> > 2 files changed, 64 insertions(+), 20 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> > index c77238cdf448..16553f7dde12 100644
> > --- a/drivers/hwtracing/coresight/coresight-platform.c
> > +++ b/drivers/hwtracing/coresight/coresight-platform.c
> > @@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
> > struct coresight_platform_data *pdata)
> > {
> > if (pdata->nr_outconns) {
> > - pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
> > - sizeof(*pdata->out_conns), GFP_KERNEL);
> > + pdata->out_conns = devm_krealloc_array(
> > + dev, pdata->out_conns, pdata->nr_outconns,
>
> super minor nit:
> pdata->out_conns = devm_krealloc_array(dev,
>
>
> > + sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
> > if (!pdata->out_conns)
> > return -ENOMEM;
> > }
> > @@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
> > return 0;
> > }
> >
> > +/*
> > + * Add a connection in the first free slot, or realloc
> > + * if there is no space. @conn's contents is copied into the new slot.
> > + *
> > + * If the output port is already assigned on this device, return -EINVAL
> > + */
> > +int coresight_add_conn(struct device *dev,
> > + struct coresight_platform_data *pdata,
> > + const struct coresight_connection *conn)
> > +{
> > + int ret;
> > + struct coresight_connection *free_conn = NULL;
> > + struct coresight_connection *i;
> > +
> > + /*
> > + * Search for a free slot, and while looking for one, warn
> > + * on any existing duplicate output port.
> > + */
> > + for (i = pdata->out_conns; i < pdata->out_conns + pdata->nr_outconns;
> > + ++i) {
>
> minor nit: I see why you have gone against using "i" as index into
> the array. But I think having that as the index is still better
> readable.
>
> for (i = 0; i < pdata->nr_outconns; i++) {
> struct coresight_connection *c = &pdata->out_conns[i];
>
> > + if (i->remote_fwnode && conn->port != -1 &&
> > + i->port == conn->port) {
> > + dev_warn(dev, "Duplicate output port %d\n", i->port);
> > + return -EINVAL;
> > + }
This code assumes that slots are filled sequentially and that it is
not possible to release slots out of order - i.e. if we find a free
slot, there is not a match in a later slot.
I can't think how this could happen but a comment to confirm this
might be needed here.
When we had 1:1 port / array index then this check was guaranteed
Mike
> > + if (!i->remote_fwnode && !free_conn)
> > + free_conn = i;
> > + }
> > +
> > + if (!free_conn) {
>
> and:
> /* No free slots */
> if (i == pdata->nr_outconns) {
>
> > + pdata->nr_outconns++;
> > + ret = coresight_alloc_conns(dev, pdata);
> > + if (ret)
> > + return ret;
> > + free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
> > + }
> > +
>
> and:
> pdata->out_conns[i] = *conn;
>
>
> Otherwise looks good to me.
>
> Suzuki
>
>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 5/9] coresight: Dynamically add connections
@ 2023-03-21 17:56 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 17:56 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James
On Thu, 16 Mar 2023 at 17:12, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > Add a function for adding connections dynamically. This also removes
> > the 1:1 mapping between port number and the index into the connections
> > array. The only place this mapping was used was in the warning for
> > duplicate output ports, which has been replaced by a search. Other
> > uses of the port number already use the port member variable.
> >
> > Being able to dynamically add connections will allow other devices like
> > CTI to re-use the connection mechanism despite not having explicit
> > connections described in the DT.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > .../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
> > include/linux/coresight.h | 7 +-
> > 2 files changed, 64 insertions(+), 20 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
> > index c77238cdf448..16553f7dde12 100644
> > --- a/drivers/hwtracing/coresight/coresight-platform.c
> > +++ b/drivers/hwtracing/coresight/coresight-platform.c
> > @@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
> > struct coresight_platform_data *pdata)
> > {
> > if (pdata->nr_outconns) {
> > - pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
> > - sizeof(*pdata->out_conns), GFP_KERNEL);
> > + pdata->out_conns = devm_krealloc_array(
> > + dev, pdata->out_conns, pdata->nr_outconns,
>
> super minor nit:
> pdata->out_conns = devm_krealloc_array(dev,
>
>
> > + sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
> > if (!pdata->out_conns)
> > return -ENOMEM;
> > }
> > @@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
> > return 0;
> > }
> >
> > +/*
> > + * Add a connection in the first free slot, or realloc
> > + * if there is no space. @conn's contents is copied into the new slot.
> > + *
> > + * If the output port is already assigned on this device, return -EINVAL
> > + */
> > +int coresight_add_conn(struct device *dev,
> > + struct coresight_platform_data *pdata,
> > + const struct coresight_connection *conn)
> > +{
> > + int ret;
> > + struct coresight_connection *free_conn = NULL;
> > + struct coresight_connection *i;
> > +
> > + /*
> > + * Search for a free slot, and while looking for one, warn
> > + * on any existing duplicate output port.
> > + */
> > + for (i = pdata->out_conns; i < pdata->out_conns + pdata->nr_outconns;
> > + ++i) {
>
> minor nit: I see why you have gone against using "i" as index into
> the array. But I think having that as the index is still better
> readable.
>
> for (i = 0; i < pdata->nr_outconns; i++) {
> struct coresight_connection *c = &pdata->out_conns[i];
>
> > + if (i->remote_fwnode && conn->port != -1 &&
> > + i->port == conn->port) {
> > + dev_warn(dev, "Duplicate output port %d\n", i->port);
> > + return -EINVAL;
> > + }
This code assumes that slots are filled sequentially and that it is
not possible to release slots out of order - i.e. if we find a free
slot, there is not a match in a later slot.
I can't think how this could happen but a comment to confirm this
might be needed here.
When we had 1:1 port / array index then this check was guaranteed
Mike
> > + if (!i->remote_fwnode && !free_conn)
> > + free_conn = i;
> > + }
> > +
> > + if (!free_conn) {
>
> and:
> /* No free slots */
> if (i == pdata->nr_outconns) {
>
> > + pdata->nr_outconns++;
> > + ret = coresight_alloc_conns(dev, pdata);
> > + if (ret)
> > + return ret;
> > + free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
> > + }
> > +
>
> and:
> pdata->out_conns[i] = *conn;
>
>
> Otherwise looks good to me.
>
> Suzuki
>
>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 7/9] coresight: Refactor out buffer allocation function for ETR
2023-03-10 16:06 ` James Clark
@ 2023-03-21 18:01 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 18:01 UTC (permalink / raw)
To: James Clark
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Fri, 10 Mar 2023 at 16:07, James Clark <james.clark@arm.com> wrote:
>
> When CATU is moved to the generic enable/disable path system in the
> next commit, it will need to call into ETR and get it to pre-allocate
> its buffer so add a function for it.
>
> No functional changes
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> .../hwtracing/coresight/coresight-tmc-etr.c | 50 ++++++++++++++++---
> drivers/hwtracing/coresight/coresight-tmc.h | 2 +
> 2 files changed, 45 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index 223081dc3ad4..cb9621003aef 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> @@ -1169,7 +1169,7 @@ void tmc_etr_disable_hw(struct tmc_drvdata *drvdata)
> drvdata->etr_buf = NULL;
> }
>
> -static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> +static struct etr_buf *tmc_etr_get_sysfs_buffer(struct coresight_device *csdev)
> {
> int ret = 0;
> unsigned long flags;
> @@ -1192,7 +1192,7 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> /* Allocate memory with the locks released */
> free_buf = new_buf = tmc_etr_setup_sysfs_buf(drvdata);
> if (IS_ERR(new_buf))
> - return PTR_ERR(new_buf);
> + return new_buf;
>
> /* Let's try again */
> spin_lock_irqsave(&drvdata->spinlock, flags);
> @@ -1223,17 +1223,33 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> drvdata->sysfs_buf = new_buf;
> }
>
> - ret = tmc_etr_enable_hw(drvdata, drvdata->sysfs_buf);
> - if (!ret) {
> - drvdata->mode = CS_MODE_SYSFS;
> - atomic_inc(csdev->refcnt);
> - }
> out:
> spin_unlock_irqrestore(&drvdata->spinlock, flags);
>
> /* Free memory outside the spinlock if need be */
> if (free_buf)
> tmc_etr_free_sysfs_buf(free_buf);
> + return ret ? ERR_PTR(ret) : drvdata->sysfs_buf;
> +}
> +
> +static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> +{
> + int ret;
> + unsigned long flags;
> + struct tmc_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
> + struct etr_buf *sysfs_buf = tmc_etr_get_sysfs_buffer(csdev);
> +
> + if (IS_ERR(sysfs_buf))
> + return PTR_ERR(sysfs_buf);
> +
> + spin_lock_irqsave(&drvdata->spinlock, flags);
> + ret = tmc_etr_enable_hw(drvdata, sysfs_buf);
> + if (!ret) {
> + drvdata->mode = CS_MODE_SYSFS;
> + atomic_inc(csdev->refcnt);
> + }
> +
> + spin_unlock_irqrestore(&drvdata->spinlock, flags);
>
> if (!ret)
> dev_dbg(&csdev->dev, "TMC-ETR enabled\n");
> @@ -1241,6 +1257,26 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> return ret;
> }
>
> +struct etr_buf *tmc_etr_get_buffer(struct coresight_device *csdev,
> + enum cs_mode mode, void *data)
> +{
> + struct perf_output_handle *handle = data;
> + struct etr_perf_buffer *etr_perf;
> +
> + switch (mode) {
> + case CS_MODE_SYSFS:
> + return tmc_etr_get_sysfs_buffer(csdev);
> + case CS_MODE_PERF:
> + etr_perf = etm_perf_sink_config(handle);
> + if (WARN_ON(!etr_perf || !etr_perf->etr_buf))
> + return ERR_PTR(-EINVAL);
> + return etr_perf->etr_buf;
> + default:
> + return ERR_PTR(-EINVAL);
> + }
> +}
> +EXPORT_SYMBOL_GPL(tmc_etr_get_buffer);
> +
> /*
> * alloc_etr_buf: Allocate ETR buffer for use by perf.
> * The size of the hardware buffer is dependent on the size configured
> diff --git a/drivers/hwtracing/coresight/coresight-tmc.h b/drivers/hwtracing/coresight/coresight-tmc.h
> index 01c0382a29c0..b97da39652d2 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc.h
> +++ b/drivers/hwtracing/coresight/coresight-tmc.h
> @@ -332,5 +332,7 @@ struct coresight_device *tmc_etr_get_catu_device(struct tmc_drvdata *drvdata);
>
> void tmc_etr_set_catu_ops(const struct etr_buf_operations *catu);
> void tmc_etr_remove_catu_ops(void);
> +struct etr_buf *tmc_etr_get_buffer(struct coresight_device *csdev,
> + enum cs_mode mode, void *data);
>
> #endif
> --
> 2.34.1
>
Reviewed by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 7/9] coresight: Refactor out buffer allocation function for ETR
@ 2023-03-21 18:01 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-21 18:01 UTC (permalink / raw)
To: James Clark
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Fri, 10 Mar 2023 at 16:07, James Clark <james.clark@arm.com> wrote:
>
> When CATU is moved to the generic enable/disable path system in the
> next commit, it will need to call into ETR and get it to pre-allocate
> its buffer so add a function for it.
>
> No functional changes
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> .../hwtracing/coresight/coresight-tmc-etr.c | 50 ++++++++++++++++---
> drivers/hwtracing/coresight/coresight-tmc.h | 2 +
> 2 files changed, 45 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index 223081dc3ad4..cb9621003aef 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> @@ -1169,7 +1169,7 @@ void tmc_etr_disable_hw(struct tmc_drvdata *drvdata)
> drvdata->etr_buf = NULL;
> }
>
> -static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> +static struct etr_buf *tmc_etr_get_sysfs_buffer(struct coresight_device *csdev)
> {
> int ret = 0;
> unsigned long flags;
> @@ -1192,7 +1192,7 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> /* Allocate memory with the locks released */
> free_buf = new_buf = tmc_etr_setup_sysfs_buf(drvdata);
> if (IS_ERR(new_buf))
> - return PTR_ERR(new_buf);
> + return new_buf;
>
> /* Let's try again */
> spin_lock_irqsave(&drvdata->spinlock, flags);
> @@ -1223,17 +1223,33 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> drvdata->sysfs_buf = new_buf;
> }
>
> - ret = tmc_etr_enable_hw(drvdata, drvdata->sysfs_buf);
> - if (!ret) {
> - drvdata->mode = CS_MODE_SYSFS;
> - atomic_inc(csdev->refcnt);
> - }
> out:
> spin_unlock_irqrestore(&drvdata->spinlock, flags);
>
> /* Free memory outside the spinlock if need be */
> if (free_buf)
> tmc_etr_free_sysfs_buf(free_buf);
> + return ret ? ERR_PTR(ret) : drvdata->sysfs_buf;
> +}
> +
> +static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> +{
> + int ret;
> + unsigned long flags;
> + struct tmc_drvdata *drvdata = dev_get_drvdata(csdev->dev.parent);
> + struct etr_buf *sysfs_buf = tmc_etr_get_sysfs_buffer(csdev);
> +
> + if (IS_ERR(sysfs_buf))
> + return PTR_ERR(sysfs_buf);
> +
> + spin_lock_irqsave(&drvdata->spinlock, flags);
> + ret = tmc_etr_enable_hw(drvdata, sysfs_buf);
> + if (!ret) {
> + drvdata->mode = CS_MODE_SYSFS;
> + atomic_inc(csdev->refcnt);
> + }
> +
> + spin_unlock_irqrestore(&drvdata->spinlock, flags);
>
> if (!ret)
> dev_dbg(&csdev->dev, "TMC-ETR enabled\n");
> @@ -1241,6 +1257,26 @@ static int tmc_enable_etr_sink_sysfs(struct coresight_device *csdev)
> return ret;
> }
>
> +struct etr_buf *tmc_etr_get_buffer(struct coresight_device *csdev,
> + enum cs_mode mode, void *data)
> +{
> + struct perf_output_handle *handle = data;
> + struct etr_perf_buffer *etr_perf;
> +
> + switch (mode) {
> + case CS_MODE_SYSFS:
> + return tmc_etr_get_sysfs_buffer(csdev);
> + case CS_MODE_PERF:
> + etr_perf = etm_perf_sink_config(handle);
> + if (WARN_ON(!etr_perf || !etr_perf->etr_buf))
> + return ERR_PTR(-EINVAL);
> + return etr_perf->etr_buf;
> + default:
> + return ERR_PTR(-EINVAL);
> + }
> +}
> +EXPORT_SYMBOL_GPL(tmc_etr_get_buffer);
> +
> /*
> * alloc_etr_buf: Allocate ETR buffer for use by perf.
> * The size of the hardware buffer is dependent on the size configured
> diff --git a/drivers/hwtracing/coresight/coresight-tmc.h b/drivers/hwtracing/coresight/coresight-tmc.h
> index 01c0382a29c0..b97da39652d2 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc.h
> +++ b/drivers/hwtracing/coresight/coresight-tmc.h
> @@ -332,5 +332,7 @@ struct coresight_device *tmc_etr_get_catu_device(struct tmc_drvdata *drvdata);
>
> void tmc_etr_set_catu_ops(const struct etr_buf_operations *catu);
> void tmc_etr_remove_catu_ops(void);
> +struct etr_buf *tmc_etr_get_buffer(struct coresight_device *csdev,
> + enum cs_mode mode, void *data);
>
> #endif
> --
> 2.34.1
>
Reviewed by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device
2023-03-21 15:12 ` Mike Leach
@ 2023-03-22 9:26 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-22 9:26 UTC (permalink / raw)
To: Mike Leach
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On 21/03/2023 15:12, Mike Leach wrote:
> Hi James
>
> On Fri, 10 Mar 2023 at 16:06, James Clark <james.clark@arm.com> wrote:
>>
>> Changes since v1:
>>
>> * Don't dereference handle in tmc_etr_get_buffer() when not in perf mode.
>> * Fix some W=1 warnings
>> * Add a commit to rename child/output in terms of local/remote
>>
>> -------------------
>>
>> Currently there is a refcount leak in CTI when using system wide mode
>> or tracing multithreaded applications. See the last commit for a
>> reproducer. This prevents the module from being unloaded.
>>
>> Historically there have been a few issues and fixes attempted around
>> here which have resulted in some extra logic and a member to keep
>> track of CTI being enabled 'struct coresight_device->ect_enabled'.
>> The fix in commit 665c157e0204 ("coresight: cti: Fix hang in
>> cti_disable_hw()") was also related to CTI having its own
>> enable/disable path which came later than other devices.
>>
>> If we make CTI a helper device and enable helper devices adjacent to
>> the path we get very similar enable/disable behavior to now, but with
>> more reuse of the existing reference counting logic in the coresight
>> core code. This also affects CATU which can have a little bit of
>> its hard coded enable/disable code removed.
>>
>> Enabling CATU on the generic path does require that input connections
>> are tracked so that it can get its associated ETR buffer.
>>
>> Applies to coresight/next (669c4614236a7) but also requires the
>> realloc_array patch here [1].
>>
>> Also available in full here [2].
>>
>> [1]: https://lore.kernel.org/linux-arm-kernel/20230306152723.3090195-1-james.clark@arm.com/
>> [2]: https://gitlab.arm.com/linux-arm/linux-jc/-/tree/james-cs-cti-module-refcount-fix-v2
>>
>> James Clark (9):
>> coresight: Use enum type for cs_mode wherever possible
>> coresight: Change name of pdata->conns
>> coresight: Rename nr_outports to nr_outconns
>> coresight: Rename connection members to allow for input connections
>> coresight: Dynamically add connections
>> coresight: Store in-connections as well as out-connections
>> coresight: Refactor out buffer allocation function for ETR
>> coresight: Enable and disable helper devices adjacent to the path
>> coresight: Fix CTI module refcount leak by making it a helper device
>>
>> drivers/hwtracing/coresight/coresight-catu.c | 34 +-
>> drivers/hwtracing/coresight/coresight-core.c | 312 +++++++++++-------
>> .../hwtracing/coresight/coresight-cti-core.c | 56 ++--
>> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
>> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
>> drivers/hwtracing/coresight/coresight-etb10.c | 3 +-
>> .../coresight/coresight-etm3x-core.c | 6 +-
>> .../coresight/coresight-etm4x-core.c | 6 +-
>> .../hwtracing/coresight/coresight-platform.c | 178 +++++++---
>> drivers/hwtracing/coresight/coresight-priv.h | 9 +-
>> drivers/hwtracing/coresight/coresight-stm.c | 6 +-
>> drivers/hwtracing/coresight/coresight-sysfs.c | 9 +-
>> .../hwtracing/coresight/coresight-tmc-etf.c | 2 +-
>> .../hwtracing/coresight/coresight-tmc-etr.c | 89 ++---
>> drivers/hwtracing/coresight/coresight-tmc.h | 2 +
>> drivers/hwtracing/coresight/coresight-tpdm.c | 4 +-
>> drivers/hwtracing/coresight/coresight-tpiu.c | 3 +-
>> drivers/hwtracing/coresight/coresight-trbe.c | 3 +-
>> drivers/hwtracing/coresight/ultrasoc-smb.c | 3 +-
>> drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
>> include/linux/coresight.h | 109 +++---
>> 21 files changed, 530 insertions(+), 314 deletions(-)
>>
>> --
>> 2.34.1
>>
>
> Looking at this overall - given that the only use of the in_conn is to
> reference the connecting device from the helper,
>
> i.e. coresight-catu.c:405: tmp = csdev->pdata->in_conns[i].remote_dev;
>
> would it not be simpler to :
>
> a)
> in coresight_connection add a field:
>
> struct coresight_device *origin_dev;
>
> which mimics the origin / target model we already have in coresight_sysfs_link
> then
>
> b) the in_conns could simply be references to out_conn object from
> origin_dev, rather than a complete coresight_connection with reversed
> values, thus simplifying the in_conns handling code, and removing the
> unused reversed feilds in the current in_conn object.
>
> e.g. tmp = csdev->pdata->in_conns[i]->origin_dev
>
> The remainder of the code would remain much the same, just adjusted
> for in_conns as refs rather than independent conn objects
This was actually my first implementation because I also thought it
would be simpler. Unfortunately it didn't work because the realloc means
that storing references to the connections is impossible as they are
moved if the array needs to grow.
Based on Suzuki's and Jinlongs comments about missing some of the input
connections some of the time, now I've made a single place for fixing up
output connections (currently there are two) and put the new input code
there. So V3 will be a bit simpler in that any time a connection is made
the input is made at the same time so it's not conceptually any more
complicated than the current code.
>
> Regards
>
>
> Mike
>
> --
> Mike Leach
> Principal Engineer, ARM Ltd.
> Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device
@ 2023-03-22 9:26 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-22 9:26 UTC (permalink / raw)
To: Mike Leach
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On 21/03/2023 15:12, Mike Leach wrote:
> Hi James
>
> On Fri, 10 Mar 2023 at 16:06, James Clark <james.clark@arm.com> wrote:
>>
>> Changes since v1:
>>
>> * Don't dereference handle in tmc_etr_get_buffer() when not in perf mode.
>> * Fix some W=1 warnings
>> * Add a commit to rename child/output in terms of local/remote
>>
>> -------------------
>>
>> Currently there is a refcount leak in CTI when using system wide mode
>> or tracing multithreaded applications. See the last commit for a
>> reproducer. This prevents the module from being unloaded.
>>
>> Historically there have been a few issues and fixes attempted around
>> here which have resulted in some extra logic and a member to keep
>> track of CTI being enabled 'struct coresight_device->ect_enabled'.
>> The fix in commit 665c157e0204 ("coresight: cti: Fix hang in
>> cti_disable_hw()") was also related to CTI having its own
>> enable/disable path which came later than other devices.
>>
>> If we make CTI a helper device and enable helper devices adjacent to
>> the path we get very similar enable/disable behavior to now, but with
>> more reuse of the existing reference counting logic in the coresight
>> core code. This also affects CATU which can have a little bit of
>> its hard coded enable/disable code removed.
>>
>> Enabling CATU on the generic path does require that input connections
>> are tracked so that it can get its associated ETR buffer.
>>
>> Applies to coresight/next (669c4614236a7) but also requires the
>> realloc_array patch here [1].
>>
>> Also available in full here [2].
>>
>> [1]: https://lore.kernel.org/linux-arm-kernel/20230306152723.3090195-1-james.clark@arm.com/
>> [2]: https://gitlab.arm.com/linux-arm/linux-jc/-/tree/james-cs-cti-module-refcount-fix-v2
>>
>> James Clark (9):
>> coresight: Use enum type for cs_mode wherever possible
>> coresight: Change name of pdata->conns
>> coresight: Rename nr_outports to nr_outconns
>> coresight: Rename connection members to allow for input connections
>> coresight: Dynamically add connections
>> coresight: Store in-connections as well as out-connections
>> coresight: Refactor out buffer allocation function for ETR
>> coresight: Enable and disable helper devices adjacent to the path
>> coresight: Fix CTI module refcount leak by making it a helper device
>>
>> drivers/hwtracing/coresight/coresight-catu.c | 34 +-
>> drivers/hwtracing/coresight/coresight-core.c | 312 +++++++++++-------
>> .../hwtracing/coresight/coresight-cti-core.c | 56 ++--
>> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
>> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
>> drivers/hwtracing/coresight/coresight-etb10.c | 3 +-
>> .../coresight/coresight-etm3x-core.c | 6 +-
>> .../coresight/coresight-etm4x-core.c | 6 +-
>> .../hwtracing/coresight/coresight-platform.c | 178 +++++++---
>> drivers/hwtracing/coresight/coresight-priv.h | 9 +-
>> drivers/hwtracing/coresight/coresight-stm.c | 6 +-
>> drivers/hwtracing/coresight/coresight-sysfs.c | 9 +-
>> .../hwtracing/coresight/coresight-tmc-etf.c | 2 +-
>> .../hwtracing/coresight/coresight-tmc-etr.c | 89 ++---
>> drivers/hwtracing/coresight/coresight-tmc.h | 2 +
>> drivers/hwtracing/coresight/coresight-tpdm.c | 4 +-
>> drivers/hwtracing/coresight/coresight-tpiu.c | 3 +-
>> drivers/hwtracing/coresight/coresight-trbe.c | 3 +-
>> drivers/hwtracing/coresight/ultrasoc-smb.c | 3 +-
>> drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
>> include/linux/coresight.h | 109 +++---
>> 21 files changed, 530 insertions(+), 314 deletions(-)
>>
>> --
>> 2.34.1
>>
>
> Looking at this overall - given that the only use of the in_conn is to
> reference the connecting device from the helper,
>
> i.e. coresight-catu.c:405: tmp = csdev->pdata->in_conns[i].remote_dev;
>
> would it not be simpler to :
>
> a)
> in coresight_connection add a field:
>
> struct coresight_device *origin_dev;
>
> which mimics the origin / target model we already have in coresight_sysfs_link
> then
>
> b) the in_conns could simply be references to out_conn object from
> origin_dev, rather than a complete coresight_connection with reversed
> values, thus simplifying the in_conns handling code, and removing the
> unused reversed feilds in the current in_conn object.
>
> e.g. tmp = csdev->pdata->in_conns[i]->origin_dev
>
> The remainder of the code would remain much the same, just adjusted
> for in_conns as refs rather than independent conn objects
This was actually my first implementation because I also thought it
would be simpler. Unfortunately it didn't work because the realloc means
that storing references to the connections is impossible as they are
moved if the array needs to grow.
Based on Suzuki's and Jinlongs comments about missing some of the input
connections some of the time, now I've made a single place for fixing up
output connections (currently there are two) and put the new input code
there. So V3 will be a bit simpler in that any time a connection is made
the input is made at the same time so it's not conceptually any more
complicated than the current code.
>
> Regards
>
>
> Mike
>
> --
> Mike Leach
> Principal Engineer, ARM Ltd.
> Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 9/9] coresight: Fix CTI module refcount leak by making it a helper device
2023-03-10 16:06 ` James Clark
@ 2023-03-22 9:29 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-22 9:29 UTC (permalink / raw)
To: James Clark
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James
On Fri, 10 Mar 2023 at 16:07, James Clark <james.clark@arm.com> wrote:
>
> The CTI module has some hard coded refcounting code that has a leak.
> For example running perf and then trying to unload it fails:
>
> perf record -e cs_etm// -a -- ls
> rmmod coresight_cti
>
> rmmod: ERROR: Module coresight_cti is in use
>
> The coresight core already handles references of devices in use, so by
> making CTI a normal helper device, we get working refcounting for free.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 73 +++----------------
> .../hwtracing/coresight/coresight-cti-core.c | 56 +++++++++-----
> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
> include/linux/coresight.h | 28 +------
> 5 files changed, 53 insertions(+), 112 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index 3e6ccd9e8d4e..94d84404fb29 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -253,48 +253,6 @@ void coresight_disclaim_device(struct coresight_device *csdev)
> }
> EXPORT_SYMBOL_GPL(coresight_disclaim_device);
>
> -/* enable or disable an associated CTI device of the supplied CS device */
> -static int
> -coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
> -{
> - int ect_ret = 0;
> - struct coresight_device *ect_csdev = csdev->ect_dev;
> - struct module *mod;
> -
> - if (!ect_csdev)
> - return 0;
> - if ((!ect_ops(ect_csdev)->enable) || (!ect_ops(ect_csdev)->disable))
> - return 0;
> -
> - mod = ect_csdev->dev.parent->driver->owner;
> - if (enable) {
> - if (try_module_get(mod)) {
> - ect_ret = ect_ops(ect_csdev)->enable(ect_csdev);
> - if (ect_ret) {
> - module_put(mod);
> - } else {
> - get_device(ect_csdev->dev.parent);
> - csdev->ect_enabled = true;
> - }
> - } else
> - ect_ret = -ENODEV;
> - } else {
> - if (csdev->ect_enabled) {
> - ect_ret = ect_ops(ect_csdev)->disable(ect_csdev);
> - put_device(ect_csdev->dev.parent);
> - module_put(mod);
> - csdev->ect_enabled = false;
> - }
> - }
> -
> - /* output warning if ECT enable is preventing trace operation */
> - if (ect_ret)
> - dev_info(&csdev->dev, "Associated ECT device (%s) %s failed\n",
> - dev_name(&ect_csdev->dev),
> - enable ? "enable" : "disable");
> - return ect_ret;
> -}
> -
> /*
> * Set the associated ect / cti device while holding the coresight_mutex
> * to avoid a race with coresight_enable that may try to use this value.
> @@ -302,8 +260,14 @@ coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
> void coresight_set_assoc_ectdev_mutex(struct coresight_device *csdev,
> struct coresight_device *ect_csdev)
Now we have generic helpers this could become a generic "add helper"
routine - for helpers such as CTI that aren't picked up by the trace
data port graph that catches other things.
e.g.
coresight_set_assoc_helper_mutex(struct coresight_device *csdev,
struct coresight_device *helper_csdev)
Regards
Mike
> {
> + struct coresight_connection conn = {};
> +
> mutex_lock(&coresight_mutex);
> - csdev->ect_dev = ect_csdev;
> + conn.remote_fwnode = fwnode_handle_get(dev_fwnode(&ect_csdev->dev));
> + conn.remote_dev = ect_csdev;
> + conn.remote_port = conn.port = -1;
> + coresight_add_conn(csdev->dev.parent, csdev->pdata, &conn);
> + coresight_fixup_inputs(csdev);
> mutex_unlock(&coresight_mutex);
> }
> EXPORT_SYMBOL_GPL(coresight_set_assoc_ectdev_mutex);
> @@ -320,12 +284,8 @@ static int coresight_enable_sink(struct coresight_device *csdev,
> if (!sink_ops(csdev)->enable)
> return -EINVAL;
>
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (ret)
> - return ret;
> ret = sink_ops(csdev)->enable(csdev, mode, data);
> if (ret) {
> - coresight_control_assoc_ectdev(csdev, false);
> return ret;
> }
> csdev->enable = true;
> @@ -343,7 +303,6 @@ static void coresight_disable_sink(struct coresight_device *csdev)
> ret = sink_ops(csdev)->disable(csdev);
> if (ret)
> return;
> - coresight_control_assoc_ectdev(csdev, false);
> csdev->enable = false;
> }
>
> @@ -368,17 +327,11 @@ static int coresight_enable_link(struct coresight_device *csdev,
> return outport;
>
> if (link_ops(csdev)->enable) {
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (!ret) {
> - ret = link_ops(csdev)->enable(csdev, inport, outport);
> - if (ret)
> - coresight_control_assoc_ectdev(csdev, false);
> - }
> + ret = link_ops(csdev)->enable(csdev, inport, outport);
> + if (!ret)
> + csdev->enable = true;
> }
>
> - if (!ret)
> - csdev->enable = true;
> -
> return ret;
> }
>
> @@ -407,7 +360,6 @@ static void coresight_disable_link(struct coresight_device *csdev,
>
> if (link_ops(csdev)->disable) {
> link_ops(csdev)->disable(csdev, inport, outport);
> - coresight_control_assoc_ectdev(csdev, false);
> }
>
> for (i = 0; i < nr_conns; i++)
> @@ -424,12 +376,8 @@ static int coresight_enable_source(struct coresight_device *csdev,
>
> if (!csdev->enable) {
> if (source_ops(csdev)->enable) {
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (ret)
> - return ret;
> ret = source_ops(csdev)->enable(csdev, NULL, mode);
> if (ret) {
> - coresight_control_assoc_ectdev(csdev, false);
> return ret;
> }
> }
> @@ -482,7 +430,6 @@ static bool coresight_disable_source(struct coresight_device *csdev)
> if (atomic_dec_return(csdev->refcnt) == 0) {
> if (source_ops(csdev)->disable)
> source_ops(csdev)->disable(csdev, NULL);
> - coresight_control_assoc_ectdev(csdev, false);
> csdev->enable = false;
> }
> return !csdev->enable;
> diff --git a/drivers/hwtracing/coresight/coresight-cti-core.c b/drivers/hwtracing/coresight/coresight-cti-core.c
> index 277c890a1f1f..dbce6680759f 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-core.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-core.c
> @@ -555,7 +555,10 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
> mutex_lock(&ect_mutex);
>
> /* exit if current is an ECT device.*/
> - if ((csdev->type == CORESIGHT_DEV_TYPE_ECT) || list_empty(&ect_net))
> + if ((csdev->type == CORESIGHT_DEV_TYPE_HELPER &&
> + csdev->subtype.helper_subtype ==
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI) ||
> + list_empty(&ect_net))
> goto cti_add_done;
>
> /* if we didn't find the csdev previously we used the fwnode name */
> @@ -580,6 +583,22 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
> mutex_unlock(&ect_mutex);
> }
>
> +static struct coresight_device *cti__get_cti_device(struct coresight_device *csdev)
> +{
> + int i;
> + struct coresight_device *tmp;
> +
> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> + tmp = csdev->pdata->out_conns[i].remote_dev;
> +
> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_HELPER &&
> + tmp->subtype.helper_subtype ==
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI)
> + return tmp;
> + }
> + return NULL;
> +}
> +
> /*
> * Removing the associated devices is easier.
> * A CTI will not have a value for csdev->ect_dev.
> @@ -588,20 +607,21 @@ static void cti_remove_assoc_from_csdev(struct coresight_device *csdev)
> {
> struct cti_drvdata *ctidrv;
> struct cti_trig_con *tc;
> + struct coresight_device *cti_csdev = cti__get_cti_device(csdev);
> struct cti_device *ctidev;
>
> + if (!cti_csdev)
> + return;
> +
> mutex_lock(&ect_mutex);
> - if (csdev->ect_dev) {
> - ctidrv = csdev_to_cti_drvdata(csdev->ect_dev);
> - ctidev = &ctidrv->ctidev;
> - list_for_each_entry(tc, &ctidev->trig_cons, node) {
> - if (tc->con_dev == csdev) {
> - cti_remove_sysfs_link(ctidrv, tc);
> - tc->con_dev = NULL;
> - break;
> - }
> + ctidrv = csdev_to_cti_drvdata(cti_csdev);
> + ctidev = &ctidrv->ctidev;
> + list_for_each_entry(tc, &ctidev->trig_cons, node) {
> + if (tc->con_dev == csdev) {
> + cti_remove_sysfs_link(ctidrv, tc);
> + tc->con_dev = NULL;
> + break;
> }
> - csdev->ect_dev = NULL;
> }
> mutex_unlock(&ect_mutex);
> }
> @@ -646,8 +666,6 @@ static void cti_remove_conn_xrefs(struct cti_drvdata *drvdata)
>
> list_for_each_entry(tc, &ctidev->trig_cons, node) {
> if (tc->con_dev) {
> - coresight_set_assoc_ectdev_mutex(tc->con_dev,
> - NULL);
> cti_remove_sysfs_link(drvdata, tc);
> tc->con_dev = NULL;
> }
> @@ -795,27 +813,27 @@ static void cti_pm_release(struct cti_drvdata *drvdata)
> }
>
> /** cti ect operations **/
> -int cti_enable(struct coresight_device *csdev)
> +int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data)
> {
> struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
>
> return cti_enable_hw(drvdata);
> }
>
> -int cti_disable(struct coresight_device *csdev)
> +int cti_disable(struct coresight_device *csdev, void *data)
> {
> struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
>
> return cti_disable_hw(drvdata);
> }
>
> -static const struct coresight_ops_ect cti_ops_ect = {
> +static const struct coresight_ops_helper cti_ops_ect = {
> .enable = cti_enable,
> .disable = cti_disable,
> };
>
> static const struct coresight_ops cti_ops = {
> - .ect_ops = &cti_ops_ect,
> + .helper_ops = &cti_ops_ect,
> };
>
> /*
> @@ -922,8 +940,8 @@ static int cti_probe(struct amba_device *adev, const struct amba_id *id)
>
> /* set up coresight component description */
> cti_desc.pdata = pdata;
> - cti_desc.type = CORESIGHT_DEV_TYPE_ECT;
> - cti_desc.subtype.ect_subtype = CORESIGHT_DEV_SUBTYPE_ECT_CTI;
> + cti_desc.type = CORESIGHT_DEV_TYPE_HELPER;
> + cti_desc.subtype.helper_subtype = CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI;
> cti_desc.ops = &cti_ops;
> cti_desc.groups = drvdata->ctidev.con_groups;
> cti_desc.dev = dev;
> diff --git a/drivers/hwtracing/coresight/coresight-cti-sysfs.c b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> index e528cff9d4e2..d25dd2737b49 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> @@ -112,11 +112,11 @@ static ssize_t enable_store(struct device *dev,
> ret = pm_runtime_resume_and_get(dev->parent);
> if (ret)
> return ret;
> - ret = cti_enable(drvdata->csdev);
> + ret = cti_enable(drvdata->csdev, CS_MODE_SYSFS, NULL);
> if (ret)
> pm_runtime_put(dev->parent);
> } else {
> - ret = cti_disable(drvdata->csdev);
> + ret = cti_disable(drvdata->csdev, NULL);
> if (!ret)
> pm_runtime_put(dev->parent);
> }
> diff --git a/drivers/hwtracing/coresight/coresight-cti.h b/drivers/hwtracing/coresight/coresight-cti.h
> index 8b106b13a244..cb9ee616d01f 100644
> --- a/drivers/hwtracing/coresight/coresight-cti.h
> +++ b/drivers/hwtracing/coresight/coresight-cti.h
> @@ -215,8 +215,8 @@ int cti_add_connection_entry(struct device *dev, struct cti_drvdata *drvdata,
> const char *assoc_dev_name);
> struct cti_trig_con *cti_allocate_trig_con(struct device *dev, int in_sigs,
> int out_sigs);
> -int cti_enable(struct coresight_device *csdev);
> -int cti_disable(struct coresight_device *csdev);
> +int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data);
> +int cti_disable(struct coresight_device *csdev, void *data);
> void cti_write_all_hw_regs(struct cti_drvdata *drvdata);
> void cti_write_intack(struct device *dev, u32 ackval);
> void cti_write_single_reg(struct cti_drvdata *drvdata, int offset, u32 value);
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index c6ee1634d813..cf7a8658ee88 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -40,8 +40,7 @@ enum coresight_dev_type {
> CORESIGHT_DEV_TYPE_LINK,
> CORESIGHT_DEV_TYPE_LINKSINK,
> CORESIGHT_DEV_TYPE_SOURCE,
> - CORESIGHT_DEV_TYPE_HELPER,
> - CORESIGHT_DEV_TYPE_ECT,
> + CORESIGHT_DEV_TYPE_HELPER
> };
>
> enum coresight_dev_subtype_sink {
> @@ -66,12 +65,7 @@ enum coresight_dev_subtype_source {
>
> enum coresight_dev_subtype_helper {
> CORESIGHT_DEV_SUBTYPE_HELPER_CATU,
> -};
> -
> -/* Embedded Cross Trigger (ECT) sub-types */
> -enum coresight_dev_subtype_ect {
> - CORESIGHT_DEV_SUBTYPE_ECT_NONE,
> - CORESIGHT_DEV_SUBTYPE_ECT_CTI,
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI
> };
>
> /**
> @@ -84,8 +78,6 @@ enum coresight_dev_subtype_ect {
> * by @coresight_dev_subtype_source.
> * @helper_subtype: type of helper this component is, as defined
> * by @coresight_dev_subtype_helper.
> - * @ect_subtype: type of cross trigger this component is, as
> - * defined by @coresight_dev_subtype_ect
> */
> union coresight_dev_subtype {
> /* We have some devices which acts as LINK and SINK */
> @@ -95,7 +87,6 @@ union coresight_dev_subtype {
> };
> enum coresight_dev_subtype_source source_subtype;
> enum coresight_dev_subtype_helper helper_subtype;
> - enum coresight_dev_subtype_ect ect_subtype;
> };
>
> /**
> @@ -264,12 +255,9 @@ struct coresight_device {
> bool activated; /* true only if a sink is part of a path */
> struct dev_ext_attribute *ea;
> struct coresight_device *def_sink;
> - /* cross trigger handling */
> - struct coresight_device *ect_dev;
> /* sysfs links between components */
> int nr_links;
> bool has_conns_grp;
> - bool ect_enabled; /* true only if associated ect device is enabled */
> /* system configuration and feature lists */
> struct list_head feature_csdev_list;
> struct list_head config_csdev_list;
> @@ -377,23 +365,11 @@ struct coresight_ops_helper {
> int (*disable)(struct coresight_device *csdev, void *data);
> };
>
> -/**
> - * struct coresight_ops_ect - Ops for an embedded cross trigger device
> - *
> - * @enable : Enable the device
> - * @disable : Disable the device
> - */
> -struct coresight_ops_ect {
> - int (*enable)(struct coresight_device *csdev);
> - int (*disable)(struct coresight_device *csdev);
> -};
> -
> struct coresight_ops {
> const struct coresight_ops_sink *sink_ops;
> const struct coresight_ops_link *link_ops;
> const struct coresight_ops_source *source_ops;
> const struct coresight_ops_helper *helper_ops;
> - const struct coresight_ops_ect *ect_ops;
> };
>
> #if IS_ENABLED(CONFIG_CORESIGHT)
> --
> 2.34.1
>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 9/9] coresight: Fix CTI module refcount leak by making it a helper device
@ 2023-03-22 9:29 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-22 9:29 UTC (permalink / raw)
To: James Clark
Cc: coresight, Mathieu Poirier, Suzuki K Poulose, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
Hi James
On Fri, 10 Mar 2023 at 16:07, James Clark <james.clark@arm.com> wrote:
>
> The CTI module has some hard coded refcounting code that has a leak.
> For example running perf and then trying to unload it fails:
>
> perf record -e cs_etm// -a -- ls
> rmmod coresight_cti
>
> rmmod: ERROR: Module coresight_cti is in use
>
> The coresight core already handles references of devices in use, so by
> making CTI a normal helper device, we get working refcounting for free.
>
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 73 +++----------------
> .../hwtracing/coresight/coresight-cti-core.c | 56 +++++++++-----
> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
> include/linux/coresight.h | 28 +------
> 5 files changed, 53 insertions(+), 112 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index 3e6ccd9e8d4e..94d84404fb29 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -253,48 +253,6 @@ void coresight_disclaim_device(struct coresight_device *csdev)
> }
> EXPORT_SYMBOL_GPL(coresight_disclaim_device);
>
> -/* enable or disable an associated CTI device of the supplied CS device */
> -static int
> -coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
> -{
> - int ect_ret = 0;
> - struct coresight_device *ect_csdev = csdev->ect_dev;
> - struct module *mod;
> -
> - if (!ect_csdev)
> - return 0;
> - if ((!ect_ops(ect_csdev)->enable) || (!ect_ops(ect_csdev)->disable))
> - return 0;
> -
> - mod = ect_csdev->dev.parent->driver->owner;
> - if (enable) {
> - if (try_module_get(mod)) {
> - ect_ret = ect_ops(ect_csdev)->enable(ect_csdev);
> - if (ect_ret) {
> - module_put(mod);
> - } else {
> - get_device(ect_csdev->dev.parent);
> - csdev->ect_enabled = true;
> - }
> - } else
> - ect_ret = -ENODEV;
> - } else {
> - if (csdev->ect_enabled) {
> - ect_ret = ect_ops(ect_csdev)->disable(ect_csdev);
> - put_device(ect_csdev->dev.parent);
> - module_put(mod);
> - csdev->ect_enabled = false;
> - }
> - }
> -
> - /* output warning if ECT enable is preventing trace operation */
> - if (ect_ret)
> - dev_info(&csdev->dev, "Associated ECT device (%s) %s failed\n",
> - dev_name(&ect_csdev->dev),
> - enable ? "enable" : "disable");
> - return ect_ret;
> -}
> -
> /*
> * Set the associated ect / cti device while holding the coresight_mutex
> * to avoid a race with coresight_enable that may try to use this value.
> @@ -302,8 +260,14 @@ coresight_control_assoc_ectdev(struct coresight_device *csdev, bool enable)
> void coresight_set_assoc_ectdev_mutex(struct coresight_device *csdev,
> struct coresight_device *ect_csdev)
Now we have generic helpers this could become a generic "add helper"
routine - for helpers such as CTI that aren't picked up by the trace
data port graph that catches other things.
e.g.
coresight_set_assoc_helper_mutex(struct coresight_device *csdev,
struct coresight_device *helper_csdev)
Regards
Mike
> {
> + struct coresight_connection conn = {};
> +
> mutex_lock(&coresight_mutex);
> - csdev->ect_dev = ect_csdev;
> + conn.remote_fwnode = fwnode_handle_get(dev_fwnode(&ect_csdev->dev));
> + conn.remote_dev = ect_csdev;
> + conn.remote_port = conn.port = -1;
> + coresight_add_conn(csdev->dev.parent, csdev->pdata, &conn);
> + coresight_fixup_inputs(csdev);
> mutex_unlock(&coresight_mutex);
> }
> EXPORT_SYMBOL_GPL(coresight_set_assoc_ectdev_mutex);
> @@ -320,12 +284,8 @@ static int coresight_enable_sink(struct coresight_device *csdev,
> if (!sink_ops(csdev)->enable)
> return -EINVAL;
>
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (ret)
> - return ret;
> ret = sink_ops(csdev)->enable(csdev, mode, data);
> if (ret) {
> - coresight_control_assoc_ectdev(csdev, false);
> return ret;
> }
> csdev->enable = true;
> @@ -343,7 +303,6 @@ static void coresight_disable_sink(struct coresight_device *csdev)
> ret = sink_ops(csdev)->disable(csdev);
> if (ret)
> return;
> - coresight_control_assoc_ectdev(csdev, false);
> csdev->enable = false;
> }
>
> @@ -368,17 +327,11 @@ static int coresight_enable_link(struct coresight_device *csdev,
> return outport;
>
> if (link_ops(csdev)->enable) {
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (!ret) {
> - ret = link_ops(csdev)->enable(csdev, inport, outport);
> - if (ret)
> - coresight_control_assoc_ectdev(csdev, false);
> - }
> + ret = link_ops(csdev)->enable(csdev, inport, outport);
> + if (!ret)
> + csdev->enable = true;
> }
>
> - if (!ret)
> - csdev->enable = true;
> -
> return ret;
> }
>
> @@ -407,7 +360,6 @@ static void coresight_disable_link(struct coresight_device *csdev,
>
> if (link_ops(csdev)->disable) {
> link_ops(csdev)->disable(csdev, inport, outport);
> - coresight_control_assoc_ectdev(csdev, false);
> }
>
> for (i = 0; i < nr_conns; i++)
> @@ -424,12 +376,8 @@ static int coresight_enable_source(struct coresight_device *csdev,
>
> if (!csdev->enable) {
> if (source_ops(csdev)->enable) {
> - ret = coresight_control_assoc_ectdev(csdev, true);
> - if (ret)
> - return ret;
> ret = source_ops(csdev)->enable(csdev, NULL, mode);
> if (ret) {
> - coresight_control_assoc_ectdev(csdev, false);
> return ret;
> }
> }
> @@ -482,7 +430,6 @@ static bool coresight_disable_source(struct coresight_device *csdev)
> if (atomic_dec_return(csdev->refcnt) == 0) {
> if (source_ops(csdev)->disable)
> source_ops(csdev)->disable(csdev, NULL);
> - coresight_control_assoc_ectdev(csdev, false);
> csdev->enable = false;
> }
> return !csdev->enable;
> diff --git a/drivers/hwtracing/coresight/coresight-cti-core.c b/drivers/hwtracing/coresight/coresight-cti-core.c
> index 277c890a1f1f..dbce6680759f 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-core.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-core.c
> @@ -555,7 +555,10 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
> mutex_lock(&ect_mutex);
>
> /* exit if current is an ECT device.*/
> - if ((csdev->type == CORESIGHT_DEV_TYPE_ECT) || list_empty(&ect_net))
> + if ((csdev->type == CORESIGHT_DEV_TYPE_HELPER &&
> + csdev->subtype.helper_subtype ==
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI) ||
> + list_empty(&ect_net))
> goto cti_add_done;
>
> /* if we didn't find the csdev previously we used the fwnode name */
> @@ -580,6 +583,22 @@ static void cti_add_assoc_to_csdev(struct coresight_device *csdev)
> mutex_unlock(&ect_mutex);
> }
>
> +static struct coresight_device *cti__get_cti_device(struct coresight_device *csdev)
> +{
> + int i;
> + struct coresight_device *tmp;
> +
> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> + tmp = csdev->pdata->out_conns[i].remote_dev;
> +
> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_HELPER &&
> + tmp->subtype.helper_subtype ==
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI)
> + return tmp;
> + }
> + return NULL;
> +}
> +
> /*
> * Removing the associated devices is easier.
> * A CTI will not have a value for csdev->ect_dev.
> @@ -588,20 +607,21 @@ static void cti_remove_assoc_from_csdev(struct coresight_device *csdev)
> {
> struct cti_drvdata *ctidrv;
> struct cti_trig_con *tc;
> + struct coresight_device *cti_csdev = cti__get_cti_device(csdev);
> struct cti_device *ctidev;
>
> + if (!cti_csdev)
> + return;
> +
> mutex_lock(&ect_mutex);
> - if (csdev->ect_dev) {
> - ctidrv = csdev_to_cti_drvdata(csdev->ect_dev);
> - ctidev = &ctidrv->ctidev;
> - list_for_each_entry(tc, &ctidev->trig_cons, node) {
> - if (tc->con_dev == csdev) {
> - cti_remove_sysfs_link(ctidrv, tc);
> - tc->con_dev = NULL;
> - break;
> - }
> + ctidrv = csdev_to_cti_drvdata(cti_csdev);
> + ctidev = &ctidrv->ctidev;
> + list_for_each_entry(tc, &ctidev->trig_cons, node) {
> + if (tc->con_dev == csdev) {
> + cti_remove_sysfs_link(ctidrv, tc);
> + tc->con_dev = NULL;
> + break;
> }
> - csdev->ect_dev = NULL;
> }
> mutex_unlock(&ect_mutex);
> }
> @@ -646,8 +666,6 @@ static void cti_remove_conn_xrefs(struct cti_drvdata *drvdata)
>
> list_for_each_entry(tc, &ctidev->trig_cons, node) {
> if (tc->con_dev) {
> - coresight_set_assoc_ectdev_mutex(tc->con_dev,
> - NULL);
> cti_remove_sysfs_link(drvdata, tc);
> tc->con_dev = NULL;
> }
> @@ -795,27 +813,27 @@ static void cti_pm_release(struct cti_drvdata *drvdata)
> }
>
> /** cti ect operations **/
> -int cti_enable(struct coresight_device *csdev)
> +int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data)
> {
> struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
>
> return cti_enable_hw(drvdata);
> }
>
> -int cti_disable(struct coresight_device *csdev)
> +int cti_disable(struct coresight_device *csdev, void *data)
> {
> struct cti_drvdata *drvdata = csdev_to_cti_drvdata(csdev);
>
> return cti_disable_hw(drvdata);
> }
>
> -static const struct coresight_ops_ect cti_ops_ect = {
> +static const struct coresight_ops_helper cti_ops_ect = {
> .enable = cti_enable,
> .disable = cti_disable,
> };
>
> static const struct coresight_ops cti_ops = {
> - .ect_ops = &cti_ops_ect,
> + .helper_ops = &cti_ops_ect,
> };
>
> /*
> @@ -922,8 +940,8 @@ static int cti_probe(struct amba_device *adev, const struct amba_id *id)
>
> /* set up coresight component description */
> cti_desc.pdata = pdata;
> - cti_desc.type = CORESIGHT_DEV_TYPE_ECT;
> - cti_desc.subtype.ect_subtype = CORESIGHT_DEV_SUBTYPE_ECT_CTI;
> + cti_desc.type = CORESIGHT_DEV_TYPE_HELPER;
> + cti_desc.subtype.helper_subtype = CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI;
> cti_desc.ops = &cti_ops;
> cti_desc.groups = drvdata->ctidev.con_groups;
> cti_desc.dev = dev;
> diff --git a/drivers/hwtracing/coresight/coresight-cti-sysfs.c b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> index e528cff9d4e2..d25dd2737b49 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-sysfs.c
> @@ -112,11 +112,11 @@ static ssize_t enable_store(struct device *dev,
> ret = pm_runtime_resume_and_get(dev->parent);
> if (ret)
> return ret;
> - ret = cti_enable(drvdata->csdev);
> + ret = cti_enable(drvdata->csdev, CS_MODE_SYSFS, NULL);
> if (ret)
> pm_runtime_put(dev->parent);
> } else {
> - ret = cti_disable(drvdata->csdev);
> + ret = cti_disable(drvdata->csdev, NULL);
> if (!ret)
> pm_runtime_put(dev->parent);
> }
> diff --git a/drivers/hwtracing/coresight/coresight-cti.h b/drivers/hwtracing/coresight/coresight-cti.h
> index 8b106b13a244..cb9ee616d01f 100644
> --- a/drivers/hwtracing/coresight/coresight-cti.h
> +++ b/drivers/hwtracing/coresight/coresight-cti.h
> @@ -215,8 +215,8 @@ int cti_add_connection_entry(struct device *dev, struct cti_drvdata *drvdata,
> const char *assoc_dev_name);
> struct cti_trig_con *cti_allocate_trig_con(struct device *dev, int in_sigs,
> int out_sigs);
> -int cti_enable(struct coresight_device *csdev);
> -int cti_disable(struct coresight_device *csdev);
> +int cti_enable(struct coresight_device *csdev, enum cs_mode mode, void *data);
> +int cti_disable(struct coresight_device *csdev, void *data);
> void cti_write_all_hw_regs(struct cti_drvdata *drvdata);
> void cti_write_intack(struct device *dev, u32 ackval);
> void cti_write_single_reg(struct cti_drvdata *drvdata, int offset, u32 value);
> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> index c6ee1634d813..cf7a8658ee88 100644
> --- a/include/linux/coresight.h
> +++ b/include/linux/coresight.h
> @@ -40,8 +40,7 @@ enum coresight_dev_type {
> CORESIGHT_DEV_TYPE_LINK,
> CORESIGHT_DEV_TYPE_LINKSINK,
> CORESIGHT_DEV_TYPE_SOURCE,
> - CORESIGHT_DEV_TYPE_HELPER,
> - CORESIGHT_DEV_TYPE_ECT,
> + CORESIGHT_DEV_TYPE_HELPER
> };
>
> enum coresight_dev_subtype_sink {
> @@ -66,12 +65,7 @@ enum coresight_dev_subtype_source {
>
> enum coresight_dev_subtype_helper {
> CORESIGHT_DEV_SUBTYPE_HELPER_CATU,
> -};
> -
> -/* Embedded Cross Trigger (ECT) sub-types */
> -enum coresight_dev_subtype_ect {
> - CORESIGHT_DEV_SUBTYPE_ECT_NONE,
> - CORESIGHT_DEV_SUBTYPE_ECT_CTI,
> + CORESIGHT_DEV_SUBTYPE_HELPER_ECT_CTI
> };
>
> /**
> @@ -84,8 +78,6 @@ enum coresight_dev_subtype_ect {
> * by @coresight_dev_subtype_source.
> * @helper_subtype: type of helper this component is, as defined
> * by @coresight_dev_subtype_helper.
> - * @ect_subtype: type of cross trigger this component is, as
> - * defined by @coresight_dev_subtype_ect
> */
> union coresight_dev_subtype {
> /* We have some devices which acts as LINK and SINK */
> @@ -95,7 +87,6 @@ union coresight_dev_subtype {
> };
> enum coresight_dev_subtype_source source_subtype;
> enum coresight_dev_subtype_helper helper_subtype;
> - enum coresight_dev_subtype_ect ect_subtype;
> };
>
> /**
> @@ -264,12 +255,9 @@ struct coresight_device {
> bool activated; /* true only if a sink is part of a path */
> struct dev_ext_attribute *ea;
> struct coresight_device *def_sink;
> - /* cross trigger handling */
> - struct coresight_device *ect_dev;
> /* sysfs links between components */
> int nr_links;
> bool has_conns_grp;
> - bool ect_enabled; /* true only if associated ect device is enabled */
> /* system configuration and feature lists */
> struct list_head feature_csdev_list;
> struct list_head config_csdev_list;
> @@ -377,23 +365,11 @@ struct coresight_ops_helper {
> int (*disable)(struct coresight_device *csdev, void *data);
> };
>
> -/**
> - * struct coresight_ops_ect - Ops for an embedded cross trigger device
> - *
> - * @enable : Enable the device
> - * @disable : Disable the device
> - */
> -struct coresight_ops_ect {
> - int (*enable)(struct coresight_device *csdev);
> - int (*disable)(struct coresight_device *csdev);
> -};
> -
> struct coresight_ops {
> const struct coresight_ops_sink *sink_ops;
> const struct coresight_ops_link *link_ops;
> const struct coresight_ops_source *source_ops;
> const struct coresight_ops_helper *helper_ops;
> - const struct coresight_ops_ect *ect_ops;
> };
>
> #if IS_ENABLED(CONFIG_CORESIGHT)
> --
> 2.34.1
>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
2023-03-17 11:04 ` Suzuki K Poulose
@ 2023-03-22 9:29 ` Mike Leach
-1 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-22 9:29 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Fri, 17 Mar 2023 at 11:04, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > Currently CATU is the only helper device, and its enable and disable
> > calls are hard coded. To allow more helper devices to be added in a
> > generic way, remove these hard coded calls and just enable and disable
> > all helper devices.
> >
> > This has to apply to helpers adjacent to the path, because they will
> > never be in the path. CATU was already discovered in this way, so
> > there is no change there.
> >
> > One change that is needed is for CATU to call back into ETR to allocate
> > the buffer. Because the enable call was previously hard coded, it was
> > done at a point where the buffer was already allocated, but this is no
> > longer the case.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
> > drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
> > .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
> > include/linux/coresight.h | 3 +-
> > 4 files changed, 99 insertions(+), 34 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-catu.c b/drivers/hwtracing/coresight/coresight-catu.c
> > index bc90a03f478f..24a08a2b96b1 100644
> > --- a/drivers/hwtracing/coresight/coresight-catu.c
> > +++ b/drivers/hwtracing/coresight/coresight-catu.c
> > @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct catu_drvdata *drvdata)
> > return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
> > }
> >
> > -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> > +static struct coresight_device *
> > +catu_get_etr_device(struct coresight_device *csdev)
> > +{
> > + int i;
> > + struct coresight_device *tmp;
> > +
> > + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
> > + tmp = csdev->pdata->in_conns[i].remote_dev;
> > + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
> > + tmp->subtype.sink_subtype ==
> > + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
> > + return tmp;
> > + }
> > +
> > + return NULL;
> > +}
> > +
> > +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode cs_mode,
> > + void *data)
> > {
> > int rc;
> > u32 control, mode;
> > - struct etr_buf *etr_buf = data;
> > + struct etr_buf *etr_buf = NULL;
> > struct device *dev = &drvdata->csdev->dev;
> > struct coresight_device *csdev = drvdata->csdev;
> > + struct coresight_device *etrdev;
> >
> > if (catu_wait_for_ready(drvdata))
> > dev_warn(dev, "Timeout while waiting for READY\n");
> > @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> > if (rc)
> > return rc;
> >
> > + etrdev = catu_get_etr_device(csdev);
> > + if (etrdev) {
> > + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
> > + if (IS_ERR(etr_buf))
> > + return PTR_ERR(etr_buf);
> > + }
>
> WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
> return.
>
>
> > control |= BIT(CATU_CONTROL_ENABLE);
> >
> > if (etr_buf && etr_buf->mode == ETR_MODE_CATU) {
> > @@ -441,13 +466,14 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> > return 0;
> > }
> >
> > -static int catu_enable(struct coresight_device *csdev, void *data)
> > +static int catu_enable(struct coresight_device *csdev, enum cs_mode mode,
> > + void *data)
> > {
> > int rc;
> > struct catu_drvdata *catu_drvdata = csdev_to_catu_drvdata(csdev);
> >
> > CS_UNLOCK(catu_drvdata->base);
> > - rc = catu_enable_hw(catu_drvdata, data);
> > + rc = catu_enable_hw(catu_drvdata, mode, data);
> > CS_LOCK(catu_drvdata->base);
> > return rc;
> > }
> > diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> > index a8ba7493c09a..3e6ccd9e8d4e 100644
> > --- a/drivers/hwtracing/coresight/coresight-core.c
> > +++ b/drivers/hwtracing/coresight/coresight-core.c
> > @@ -441,6 +441,34 @@ static int coresight_enable_source(struct coresight_device *csdev,
> > return 0;
> > }
> >
> > +static int coresight_enable_helper(struct coresight_device *csdev,
> > + enum cs_mode mode, void *sink_data)
>
> minor nit: s/sink_data/data/ ? Though it is always either sink_data
> (perf mode) or NULL (sysfs mode), for the core code it is simply an
> opaque data.
>
> Rest looks fine to me.
>
> Suzuki
>
>
Reviewed by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
@ 2023-03-22 9:29 ` Mike Leach
0 siblings, 0 replies; 74+ messages in thread
From: Mike Leach @ 2023-03-22 9:29 UTC (permalink / raw)
To: Suzuki K Poulose
Cc: James Clark, coresight, Mathieu Poirier, Leo Yan,
Alexander Shishkin, linux-arm-kernel, linux-kernel
On Fri, 17 Mar 2023 at 11:04, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>
> On 10/03/2023 16:06, James Clark wrote:
> > Currently CATU is the only helper device, and its enable and disable
> > calls are hard coded. To allow more helper devices to be added in a
> > generic way, remove these hard coded calls and just enable and disable
> > all helper devices.
> >
> > This has to apply to helpers adjacent to the path, because they will
> > never be in the path. CATU was already discovered in this way, so
> > there is no change there.
> >
> > One change that is needed is for CATU to call back into ETR to allocate
> > the buffer. Because the enable call was previously hard coded, it was
> > done at a point where the buffer was already allocated, but this is no
> > longer the case.
> >
> > Signed-off-by: James Clark <james.clark@arm.com>
> > ---
> > drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
> > drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
> > .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
> > include/linux/coresight.h | 3 +-
> > 4 files changed, 99 insertions(+), 34 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-catu.c b/drivers/hwtracing/coresight/coresight-catu.c
> > index bc90a03f478f..24a08a2b96b1 100644
> > --- a/drivers/hwtracing/coresight/coresight-catu.c
> > +++ b/drivers/hwtracing/coresight/coresight-catu.c
> > @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct catu_drvdata *drvdata)
> > return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
> > }
> >
> > -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> > +static struct coresight_device *
> > +catu_get_etr_device(struct coresight_device *csdev)
> > +{
> > + int i;
> > + struct coresight_device *tmp;
> > +
> > + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
> > + tmp = csdev->pdata->in_conns[i].remote_dev;
> > + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
> > + tmp->subtype.sink_subtype ==
> > + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
> > + return tmp;
> > + }
> > +
> > + return NULL;
> > +}
> > +
> > +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode cs_mode,
> > + void *data)
> > {
> > int rc;
> > u32 control, mode;
> > - struct etr_buf *etr_buf = data;
> > + struct etr_buf *etr_buf = NULL;
> > struct device *dev = &drvdata->csdev->dev;
> > struct coresight_device *csdev = drvdata->csdev;
> > + struct coresight_device *etrdev;
> >
> > if (catu_wait_for_ready(drvdata))
> > dev_warn(dev, "Timeout while waiting for READY\n");
> > @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> > if (rc)
> > return rc;
> >
> > + etrdev = catu_get_etr_device(csdev);
> > + if (etrdev) {
> > + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
> > + if (IS_ERR(etr_buf))
> > + return PTR_ERR(etr_buf);
> > + }
>
> WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
> return.
>
>
> > control |= BIT(CATU_CONTROL_ENABLE);
> >
> > if (etr_buf && etr_buf->mode == ETR_MODE_CATU) {
> > @@ -441,13 +466,14 @@ static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
> > return 0;
> > }
> >
> > -static int catu_enable(struct coresight_device *csdev, void *data)
> > +static int catu_enable(struct coresight_device *csdev, enum cs_mode mode,
> > + void *data)
> > {
> > int rc;
> > struct catu_drvdata *catu_drvdata = csdev_to_catu_drvdata(csdev);
> >
> > CS_UNLOCK(catu_drvdata->base);
> > - rc = catu_enable_hw(catu_drvdata, data);
> > + rc = catu_enable_hw(catu_drvdata, mode, data);
> > CS_LOCK(catu_drvdata->base);
> > return rc;
> > }
> > diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> > index a8ba7493c09a..3e6ccd9e8d4e 100644
> > --- a/drivers/hwtracing/coresight/coresight-core.c
> > +++ b/drivers/hwtracing/coresight/coresight-core.c
> > @@ -441,6 +441,34 @@ static int coresight_enable_source(struct coresight_device *csdev,
> > return 0;
> > }
> >
> > +static int coresight_enable_helper(struct coresight_device *csdev,
> > + enum cs_mode mode, void *sink_data)
>
> minor nit: s/sink_data/data/ ? Though it is always either sink_data
> (perf mode) or NULL (sysfs mode), for the core code it is simply an
> opaque data.
>
> Rest looks fine to me.
>
> Suzuki
>
>
Reviewed by: Mike Leach <mike.leach@linaro.org>
--
Mike Leach
Principal Engineer, ARM Ltd.
Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device
[not found] ` <CAJ9a7VgbehyFVF1PUZSSGVzhdSY083AvN0bC2xKH=JEUp8oVbw@mail.gmail.com>
@ 2023-03-22 17:18 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-22 17:18 UTC (permalink / raw)
To: Mike Leach
Cc: Alexander Shishkin, Leo Yan, Mathieu Poirier, Suzuki K Poulose,
coresight, linux-arm-kernel, linux-kernel
On 22/03/2023 16:22, Mike Leach wrote:
> Hi
>
> On Wed, 22 Mar 2023 at 09:26, James Clark <james.clark@arm.com> wrote:
>
>>
>>
>> On 21/03/2023 15:12, Mike Leach wrote:
>>> Hi James
>>>
>>> On Fri, 10 Mar 2023 at 16:06, James Clark <james.clark@arm.com> wrote:
>>>>
>>>> Changes since v1:
>>>>
>>>> * Don't dereference handle in tmc_etr_get_buffer() when not in perf
>> mode.
>>>> * Fix some W=1 warnings
>>>> * Add a commit to rename child/output in terms of local/remote
>>>>
>>>> -------------------
>>>>
>>>> Currently there is a refcount leak in CTI when using system wide mode
>>>> or tracing multithreaded applications. See the last commit for a
>>>> reproducer. This prevents the module from being unloaded.
>>>>
>>>> Historically there have been a few issues and fixes attempted around
>>>> here which have resulted in some extra logic and a member to keep
>>>> track of CTI being enabled 'struct coresight_device->ect_enabled'.
>>>> The fix in commit 665c157e0204 ("coresight: cti: Fix hang in
>>>> cti_disable_hw()") was also related to CTI having its own
>>>> enable/disable path which came later than other devices.
>>>>
>>>> If we make CTI a helper device and enable helper devices adjacent to
>>>> the path we get very similar enable/disable behavior to now, but with
>>>> more reuse of the existing reference counting logic in the coresight
>>>> core code. This also affects CATU which can have a little bit of
>>>> its hard coded enable/disable code removed.
>>>>
>>>> Enabling CATU on the generic path does require that input connections
>>>> are tracked so that it can get its associated ETR buffer.
>>>>
>>>> Applies to coresight/next (669c4614236a7) but also requires the
>>>> realloc_array patch here [1].
>>>>
>>>> Also available in full here [2].
>>>>
>>>> [1]:
>> https://lore.kernel.org/linux-arm-kernel/20230306152723.3090195-1-james.clark@arm.com/
>>>> [2]:
>> https://gitlab.arm.com/linux-arm/linux-jc/-/tree/james-cs-cti-module-refcount-fix-v2
>>>>
>>>> James Clark (9):
>>>> coresight: Use enum type for cs_mode wherever possible
>>>> coresight: Change name of pdata->conns
>>>> coresight: Rename nr_outports to nr_outconns
>>>> coresight: Rename connection members to allow for input connections
>>>> coresight: Dynamically add connections
>>>> coresight: Store in-connections as well as out-connections
>>>> coresight: Refactor out buffer allocation function for ETR
>>>> coresight: Enable and disable helper devices adjacent to the path
>>>> coresight: Fix CTI module refcount leak by making it a helper device
>>>>
>>>> drivers/hwtracing/coresight/coresight-catu.c | 34 +-
>>>> drivers/hwtracing/coresight/coresight-core.c | 312 +++++++++++-------
>>>> .../hwtracing/coresight/coresight-cti-core.c | 56 ++--
>>>> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
>>>> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
>>>> drivers/hwtracing/coresight/coresight-etb10.c | 3 +-
>>>> .../coresight/coresight-etm3x-core.c | 6 +-
>>>> .../coresight/coresight-etm4x-core.c | 6 +-
>>>> .../hwtracing/coresight/coresight-platform.c | 178 +++++++---
>>>> drivers/hwtracing/coresight/coresight-priv.h | 9 +-
>>>> drivers/hwtracing/coresight/coresight-stm.c | 6 +-
>>>> drivers/hwtracing/coresight/coresight-sysfs.c | 9 +-
>>>> .../hwtracing/coresight/coresight-tmc-etf.c | 2 +-
>>>> .../hwtracing/coresight/coresight-tmc-etr.c | 89 ++---
>>>> drivers/hwtracing/coresight/coresight-tmc.h | 2 +
>>>> drivers/hwtracing/coresight/coresight-tpdm.c | 4 +-
>>>> drivers/hwtracing/coresight/coresight-tpiu.c | 3 +-
>>>> drivers/hwtracing/coresight/coresight-trbe.c | 3 +-
>>>> drivers/hwtracing/coresight/ultrasoc-smb.c | 3 +-
>>>> drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
>>>> include/linux/coresight.h | 109 +++---
>>>> 21 files changed, 530 insertions(+), 314 deletions(-)
>>>>
>>>> --
>>>> 2.34.1
>>>>
>>>
>>> Looking at this overall - given that the only use of the in_conn is to
>>> reference the connecting device from the helper,
>>>
>>> i.e. coresight-catu.c:405: tmp =
>> csdev->pdata->in_conns[i].remote_dev;
>>>
>>> would it not be simpler to :
>>>
>>> a)
>>> in coresight_connection add a field:
>>>
>>> struct coresight_device *origin_dev;
>>>
>>> which mimics the origin / target model we already have in
>> coresight_sysfs_link
>>> then
>>>
>>> b) the in_conns could simply be references to out_conn object from
>>> origin_dev, rather than a complete coresight_connection with reversed
>>> values, thus simplifying the in_conns handling code, and removing the
>>> unused reversed feilds in the current in_conn object.
>>>
>>> e.g. tmp = csdev->pdata->in_conns[i]->origin_dev
>>>
>>> The remainder of the code would remain much the same, just adjusted
>>> for in_conns as refs rather than independent conn objects
>>
>> This was actually my first implementation because I also thought it
>> would be simpler. Unfortunately it didn't work because the realloc means
>> that storing references to the connections is impossible as they are
>> moved if the array needs to grow.
>>
>
> Alternatively the out_conns could be references as well, creating a single
> connection object, avoiding the proliferation of additional objects that
> are just used to point back at the output element.
That would work, as long as there wouldn't be any objections based on
locality of elements in an array for performance?
IMO it's fine because starting and stopping a session are pretty
computationally expensive anyway. But it would require one extra change
to the output connections so it's a bit more churn.
>
> Mike
>
>
>
>
>> Based on Suzuki's and Jinlongs comments about missing some of the input
>> connections some of the time, now I've made a single place for fixing up
>> output connections (currently there are two) and put the new input code
>> there. So V3 will be a bit simpler in that any time a connection is made
>> the input is made at the same time so it's not conceptually any more
>> complicated than the current code.
>
>
>>
>>>
>>> Regards
>>>
>>>
>>> Mike
>>>
>>> --
>>> Mike Leach
>>> Principal Engineer, ARM Ltd.
>>> Manchester Design Centre. UK
>>
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device
@ 2023-03-22 17:18 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-22 17:18 UTC (permalink / raw)
To: Mike Leach
Cc: Alexander Shishkin, Leo Yan, Mathieu Poirier, Suzuki K Poulose,
coresight, linux-arm-kernel, linux-kernel
On 22/03/2023 16:22, Mike Leach wrote:
> Hi
>
> On Wed, 22 Mar 2023 at 09:26, James Clark <james.clark@arm.com> wrote:
>
>>
>>
>> On 21/03/2023 15:12, Mike Leach wrote:
>>> Hi James
>>>
>>> On Fri, 10 Mar 2023 at 16:06, James Clark <james.clark@arm.com> wrote:
>>>>
>>>> Changes since v1:
>>>>
>>>> * Don't dereference handle in tmc_etr_get_buffer() when not in perf
>> mode.
>>>> * Fix some W=1 warnings
>>>> * Add a commit to rename child/output in terms of local/remote
>>>>
>>>> -------------------
>>>>
>>>> Currently there is a refcount leak in CTI when using system wide mode
>>>> or tracing multithreaded applications. See the last commit for a
>>>> reproducer. This prevents the module from being unloaded.
>>>>
>>>> Historically there have been a few issues and fixes attempted around
>>>> here which have resulted in some extra logic and a member to keep
>>>> track of CTI being enabled 'struct coresight_device->ect_enabled'.
>>>> The fix in commit 665c157e0204 ("coresight: cti: Fix hang in
>>>> cti_disable_hw()") was also related to CTI having its own
>>>> enable/disable path which came later than other devices.
>>>>
>>>> If we make CTI a helper device and enable helper devices adjacent to
>>>> the path we get very similar enable/disable behavior to now, but with
>>>> more reuse of the existing reference counting logic in the coresight
>>>> core code. This also affects CATU which can have a little bit of
>>>> its hard coded enable/disable code removed.
>>>>
>>>> Enabling CATU on the generic path does require that input connections
>>>> are tracked so that it can get its associated ETR buffer.
>>>>
>>>> Applies to coresight/next (669c4614236a7) but also requires the
>>>> realloc_array patch here [1].
>>>>
>>>> Also available in full here [2].
>>>>
>>>> [1]:
>> https://lore.kernel.org/linux-arm-kernel/20230306152723.3090195-1-james.clark@arm.com/
>>>> [2]:
>> https://gitlab.arm.com/linux-arm/linux-jc/-/tree/james-cs-cti-module-refcount-fix-v2
>>>>
>>>> James Clark (9):
>>>> coresight: Use enum type for cs_mode wherever possible
>>>> coresight: Change name of pdata->conns
>>>> coresight: Rename nr_outports to nr_outconns
>>>> coresight: Rename connection members to allow for input connections
>>>> coresight: Dynamically add connections
>>>> coresight: Store in-connections as well as out-connections
>>>> coresight: Refactor out buffer allocation function for ETR
>>>> coresight: Enable and disable helper devices adjacent to the path
>>>> coresight: Fix CTI module refcount leak by making it a helper device
>>>>
>>>> drivers/hwtracing/coresight/coresight-catu.c | 34 +-
>>>> drivers/hwtracing/coresight/coresight-core.c | 312 +++++++++++-------
>>>> .../hwtracing/coresight/coresight-cti-core.c | 56 ++--
>>>> .../hwtracing/coresight/coresight-cti-sysfs.c | 4 +-
>>>> drivers/hwtracing/coresight/coresight-cti.h | 4 +-
>>>> drivers/hwtracing/coresight/coresight-etb10.c | 3 +-
>>>> .../coresight/coresight-etm3x-core.c | 6 +-
>>>> .../coresight/coresight-etm4x-core.c | 6 +-
>>>> .../hwtracing/coresight/coresight-platform.c | 178 +++++++---
>>>> drivers/hwtracing/coresight/coresight-priv.h | 9 +-
>>>> drivers/hwtracing/coresight/coresight-stm.c | 6 +-
>>>> drivers/hwtracing/coresight/coresight-sysfs.c | 9 +-
>>>> .../hwtracing/coresight/coresight-tmc-etf.c | 2 +-
>>>> .../hwtracing/coresight/coresight-tmc-etr.c | 89 ++---
>>>> drivers/hwtracing/coresight/coresight-tmc.h | 2 +
>>>> drivers/hwtracing/coresight/coresight-tpdm.c | 4 +-
>>>> drivers/hwtracing/coresight/coresight-tpiu.c | 3 +-
>>>> drivers/hwtracing/coresight/coresight-trbe.c | 3 +-
>>>> drivers/hwtracing/coresight/ultrasoc-smb.c | 3 +-
>>>> drivers/hwtracing/coresight/ultrasoc-smb.h | 2 +-
>>>> include/linux/coresight.h | 109 +++---
>>>> 21 files changed, 530 insertions(+), 314 deletions(-)
>>>>
>>>> --
>>>> 2.34.1
>>>>
>>>
>>> Looking at this overall - given that the only use of the in_conn is to
>>> reference the connecting device from the helper,
>>>
>>> i.e. coresight-catu.c:405: tmp =
>> csdev->pdata->in_conns[i].remote_dev;
>>>
>>> would it not be simpler to :
>>>
>>> a)
>>> in coresight_connection add a field:
>>>
>>> struct coresight_device *origin_dev;
>>>
>>> which mimics the origin / target model we already have in
>> coresight_sysfs_link
>>> then
>>>
>>> b) the in_conns could simply be references to out_conn object from
>>> origin_dev, rather than a complete coresight_connection with reversed
>>> values, thus simplifying the in_conns handling code, and removing the
>>> unused reversed feilds in the current in_conn object.
>>>
>>> e.g. tmp = csdev->pdata->in_conns[i]->origin_dev
>>>
>>> The remainder of the code would remain much the same, just adjusted
>>> for in_conns as refs rather than independent conn objects
>>
>> This was actually my first implementation because I also thought it
>> would be simpler. Unfortunately it didn't work because the realloc means
>> that storing references to the connections is impossible as they are
>> moved if the array needs to grow.
>>
>
> Alternatively the out_conns could be references as well, creating a single
> connection object, avoiding the proliferation of additional objects that
> are just used to point back at the output element.
That would work, as long as there wouldn't be any objections based on
locality of elements in an array for performance?
IMO it's fine because starting and stopping a session are pretty
computationally expensive anyway. But it would require one extra change
to the output connections so it's a bit more churn.
>
> Mike
>
>
>
>
>> Based on Suzuki's and Jinlongs comments about missing some of the input
>> connections some of the time, now I've made a single place for fixing up
>> output connections (currently there are two) and put the new input code
>> there. So V3 will be a bit simpler in that any time a connection is made
>> the input is made at the same time so it's not conceptually any more
>> complicated than the current code.
>
>
>>
>>>
>>> Regards
>>>
>>>
>>> Mike
>>>
>>> --
>>> Mike Leach
>>> Principal Engineer, ARM Ltd.
>>> Manchester Design Centre. UK
>>
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 5/9] coresight: Dynamically add connections
2023-03-21 17:56 ` Mike Leach
@ 2023-03-23 10:49 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-23 10:49 UTC (permalink / raw)
To: Mike Leach, Suzuki K Poulose
Cc: coresight, Mathieu Poirier, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 21/03/2023 17:56, Mike Leach wrote:
> Hi James
>
> On Thu, 16 Mar 2023 at 17:12, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>>
>> On 10/03/2023 16:06, James Clark wrote:
>>> Add a function for adding connections dynamically. This also removes
>>> the 1:1 mapping between port number and the index into the connections
>>> array. The only place this mapping was used was in the warning for
>>> duplicate output ports, which has been replaced by a search. Other
>>> uses of the port number already use the port member variable.
>>>
>>> Being able to dynamically add connections will allow other devices like
>>> CTI to re-use the connection mechanism despite not having explicit
>>> connections described in the DT.
>>>
>>> Signed-off-by: James Clark <james.clark@arm.com>
>>> ---
>>> .../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
>>> include/linux/coresight.h | 7 +-
>>> 2 files changed, 64 insertions(+), 20 deletions(-)
>>>
>>> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
>>> index c77238cdf448..16553f7dde12 100644
>>> --- a/drivers/hwtracing/coresight/coresight-platform.c
>>> +++ b/drivers/hwtracing/coresight/coresight-platform.c
>>> @@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
>>> struct coresight_platform_data *pdata)
>>> {
>>> if (pdata->nr_outconns) {
>>> - pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
>>> - sizeof(*pdata->out_conns), GFP_KERNEL);
>>> + pdata->out_conns = devm_krealloc_array(
>>> + dev, pdata->out_conns, pdata->nr_outconns,
>>
>> super minor nit:
>> pdata->out_conns = devm_krealloc_array(dev,
>>
>>
>>> + sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
>>> if (!pdata->out_conns)
>>> return -ENOMEM;
>>> }
>>> @@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
>>> return 0;
>>> }
>>>
>>> +/*
>>> + * Add a connection in the first free slot, or realloc
>>> + * if there is no space. @conn's contents is copied into the new slot.
>>> + *
>>> + * If the output port is already assigned on this device, return -EINVAL
>>> + */
>>> +int coresight_add_conn(struct device *dev,
>>> + struct coresight_platform_data *pdata,
>>> + const struct coresight_connection *conn)
>>> +{
>>> + int ret;
>>> + struct coresight_connection *free_conn = NULL;
>>> + struct coresight_connection *i;
>>> +
>>> + /*
>>> + * Search for a free slot, and while looking for one, warn
>>> + * on any existing duplicate output port.
>>> + */
>>> + for (i = pdata->out_conns; i < pdata->out_conns + pdata->nr_outconns;
>>> + ++i) {
>>
>> minor nit: I see why you have gone against using "i" as index into
>> the array. But I think having that as the index is still better
>> readable.
>>
>> for (i = 0; i < pdata->nr_outconns; i++) {
>> struct coresight_connection *c = &pdata->out_conns[i];
>>
>>> + if (i->remote_fwnode && conn->port != -1 &&
>>> + i->port == conn->port) {
>>> + dev_warn(dev, "Duplicate output port %d\n", i->port);
>>> + return -EINVAL;
>>> + }
>
> This code assumes that slots are filled sequentially and that it is
> not possible to release slots out of order - i.e. if we find a free
> slot, there is not a match in a later slot.
> I can't think how this could happen but a comment to confirm this
> might be needed here.
>
> When we had 1:1 port / array index then this check was guaranteed
>> Mike
I thought about this but I couldn't see an issue here. The loop always
runs to the end even if a free slot is found so it should find
duplicates in any order. Unless I'm missing some other edge case?
>
>
>
>>> + if (!i->remote_fwnode && !free_conn)
>>> + free_conn = i;
>>> + }
>>> +
>>> + if (!free_conn) {
>>
>> and:
>> /* No free slots */
>> if (i == pdata->nr_outconns) {
>>
>>> + pdata->nr_outconns++;
>>> + ret = coresight_alloc_conns(dev, pdata);
>>> + if (ret)
>>> + return ret;
>>> + free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
>>> + }
>>> +
>>
>> and:
>> pdata->out_conns[i] = *conn;
>>
>>
>> Otherwise looks good to me.
>>
>> Suzuki
>>
>>
>
>
> --
> Mike Leach
> Principal Engineer, ARM Ltd.
> Manchester Design Centre. UK
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 5/9] coresight: Dynamically add connections
@ 2023-03-23 10:49 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-23 10:49 UTC (permalink / raw)
To: Mike Leach, Suzuki K Poulose
Cc: coresight, Mathieu Poirier, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 21/03/2023 17:56, Mike Leach wrote:
> Hi James
>
> On Thu, 16 Mar 2023 at 17:12, Suzuki K Poulose <suzuki.poulose@arm.com> wrote:
>>
>> On 10/03/2023 16:06, James Clark wrote:
>>> Add a function for adding connections dynamically. This also removes
>>> the 1:1 mapping between port number and the index into the connections
>>> array. The only place this mapping was used was in the warning for
>>> duplicate output ports, which has been replaced by a search. Other
>>> uses of the port number already use the port member variable.
>>>
>>> Being able to dynamically add connections will allow other devices like
>>> CTI to re-use the connection mechanism despite not having explicit
>>> connections described in the DT.
>>>
>>> Signed-off-by: James Clark <james.clark@arm.com>
>>> ---
>>> .../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
>>> include/linux/coresight.h | 7 +-
>>> 2 files changed, 64 insertions(+), 20 deletions(-)
>>>
>>> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/drivers/hwtracing/coresight/coresight-platform.c
>>> index c77238cdf448..16553f7dde12 100644
>>> --- a/drivers/hwtracing/coresight/coresight-platform.c
>>> +++ b/drivers/hwtracing/coresight/coresight-platform.c
>>> @@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
>>> struct coresight_platform_data *pdata)
>>> {
>>> if (pdata->nr_outconns) {
>>> - pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
>>> - sizeof(*pdata->out_conns), GFP_KERNEL);
>>> + pdata->out_conns = devm_krealloc_array(
>>> + dev, pdata->out_conns, pdata->nr_outconns,
>>
>> super minor nit:
>> pdata->out_conns = devm_krealloc_array(dev,
>>
>>
>>> + sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
>>> if (!pdata->out_conns)
>>> return -ENOMEM;
>>> }
>>> @@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
>>> return 0;
>>> }
>>>
>>> +/*
>>> + * Add a connection in the first free slot, or realloc
>>> + * if there is no space. @conn's contents is copied into the new slot.
>>> + *
>>> + * If the output port is already assigned on this device, return -EINVAL
>>> + */
>>> +int coresight_add_conn(struct device *dev,
>>> + struct coresight_platform_data *pdata,
>>> + const struct coresight_connection *conn)
>>> +{
>>> + int ret;
>>> + struct coresight_connection *free_conn = NULL;
>>> + struct coresight_connection *i;
>>> +
>>> + /*
>>> + * Search for a free slot, and while looking for one, warn
>>> + * on any existing duplicate output port.
>>> + */
>>> + for (i = pdata->out_conns; i < pdata->out_conns + pdata->nr_outconns;
>>> + ++i) {
>>
>> minor nit: I see why you have gone against using "i" as index into
>> the array. But I think having that as the index is still better
>> readable.
>>
>> for (i = 0; i < pdata->nr_outconns; i++) {
>> struct coresight_connection *c = &pdata->out_conns[i];
>>
>>> + if (i->remote_fwnode && conn->port != -1 &&
>>> + i->port == conn->port) {
>>> + dev_warn(dev, "Duplicate output port %d\n", i->port);
>>> + return -EINVAL;
>>> + }
>
> This code assumes that slots are filled sequentially and that it is
> not possible to release slots out of order - i.e. if we find a free
> slot, there is not a match in a later slot.
> I can't think how this could happen but a comment to confirm this
> might be needed here.
>
> When we had 1:1 port / array index then this check was guaranteed
>> Mike
I thought about this but I couldn't see an issue here. The loop always
runs to the end even if a free slot is found so it should find
duplicates in any order. Unless I'm missing some other edge case?
>
>
>
>>> + if (!i->remote_fwnode && !free_conn)
>>> + free_conn = i;
>>> + }
>>> +
>>> + if (!free_conn) {
>>
>> and:
>> /* No free slots */
>> if (i == pdata->nr_outconns) {
>>
>>> + pdata->nr_outconns++;
>>> + ret = coresight_alloc_conns(dev, pdata);
>>> + if (ret)
>>> + return ret;
>>> + free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
>>> + }
>>> +
>>
>> and:
>> pdata->out_conns[i] = *conn;
>>
>>
>> Otherwise looks good to me.
>>
>> Suzuki
>>
>>
>
>
> --
> Mike Leach
> Principal Engineer, ARM Ltd.
> Manchester Design Centre. UK
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
2023-03-16 20:23 ` Suzuki K Poulose
@ 2023-03-29 11:59 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 11:59 UTC (permalink / raw)
To: Suzuki K Poulose, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 16/03/2023 20:23, Suzuki K Poulose wrote:
> On 10/03/2023 16:06, James Clark wrote:
>> This will allow CATU to get its associated ETR in a generic way where
>> currently the enable path has some hard coded searches which avoid
>> the need to store input connections.
>>
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>> drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
>> .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
>> drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
>> include/linux/coresight.h | 25 ++++++++
>> 4 files changed, 130 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-core.c
>> b/drivers/hwtracing/coresight/coresight-core.c
>> index f457914e445e..a8ba7493c09a 100644
>> --- a/drivers/hwtracing/coresight/coresight-core.c
>> +++ b/drivers/hwtracing/coresight/coresight-core.c
>> @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff,
>> 0x7fffffff, 0x7fffffff, 0x7fff
>> EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
>> static const struct cti_assoc_op *cti_assoc_ops;
>> +static int coresight_fixup_inputs(struct coresight_device *csdev);
>> ssize_t coresight_simple_show_pair(struct device *_dev,
>> struct device_attribute *attr, char *buf)
>> @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct
>> coresight_device *csdev)
>> csdev, coresight_orphan_match);
>> }
>> +/*
>> + * Device connections are discovered before one/both devices have
>> been created,
>> + * so inputs must be added later.
>> + */
>> +static int coresight_fixup_inputs(struct coresight_device *csdev)
>> +{
>> + int i, ret = 0;
>> + struct coresight_connection *out_conn;
>> + struct coresight_connection in_conn;
>> +
>> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>> + out_conn = &csdev->pdata->out_conns[i];
>> + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
>> + continue;
>> +
>> + /* Reverse local/remote relationships for inputs */
>> + in_conn.remote_dev = csdev;
>> + in_conn.remote_port = out_conn->port;
>> + in_conn.port = out_conn->remote_port;
>> + in_conn.remote_fwnode = csdev->dev.fwnode;
>> + ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
>> + out_conn->remote_dev->pdata,
>> + &in_conn);
>> + if (ret)
>> + return ret;
>> + }
>> +
>> + return 0;
>> +}
>> static int coresight_fixup_device_conns(struct coresight_device
>> *csdev)
>> {
>> @@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct
>> device *dev, void *data)
>
> Removing the connection should be much simpler now with the inports
> connection tracking.
>
> i.e., coresight_remove_conns() need not iterate over all the devices on
> the coresight bus. We had to do that to find all the devices connecting
> to the removed device.
Done in v3
>
> Instead with the in_conns we could readily get the csdev instance and
> remove the connection from those devices.
>
> Similarly for the out_conns can give you the remove device for removing
> the in_conns on that device.
>
> i.e,
>
> static void coresight_remove_conns(csdev)
> {
> struct coresight_device *remote;
>
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> remote = csdev->data->out_conns[i].remote_dev;
>
> if (!remote)
> continue;
>
> for (j = 0; j < remote->pdata->nr_inconns; j++) {
> if (remote->pdata->in_conns[i].remote_dev == csdev)
> /* Clear in_conns[i] on remote */
>
> }
> }
>
> Similarly for in_conns
> for (i = 0; i < csdev->pdata->nr_inconns; i++) {
> remote = csdev->data->in_conns[i].remote_dev;
>
> if (!remote)
> continue;
>
> for (j = 0; j < remote->pdata->nr_outconns; j++) {
> if (remote->pdata->out_conns[i].remote_dev == csdev)
> /* Clear out_conns[i] on remote */
> }
>
>
>> */
>> fwnode_handle_put(conn->remote_fwnode);
>> conn->remote_fwnode = NULL;
>> + conn->remote_dev = NULL;
>> + /* No need to continue */
>> + break;
>> + }
>> + }
>> + for (i = 0; i < iterator->pdata->nr_inconns; i++) {
>> + conn = &iterator->pdata->in_conns[i];
>> + if (csdev == conn->remote_dev) {
>> + conn->remote_fwnode = NULL;
>> + conn->remote_dev = NULL;
>> /* No need to continue */
>> break;
>> }
>> }
>> -
>> /*
>> * Returning '0' ensures that all known component on the
>> * bus will be checked.
>> @@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct
>> coresight_device *csdev,
>> for (i = 0; i < pdata->nr_outconns; i++) {
>> /* If we have made the links, remove them now */
>> - if (csdev && conns[i].remote_dev)
>> + if (csdev && conns[i].remote_dev) {
>> coresight_remove_links(csdev, &conns[i]);
>> + conns[i].remote_dev = NULL;
>> + }
>> +
>> /*
>> * Drop the refcount and clear the handle as this device
>> * is going away
>> */
>> if (conns[i].remote_fwnode) {
>> fwnode_handle_put(conns[i].remote_fwnode);
>> - pdata->out_conns[i].remote_fwnode = NULL;
>> + conns[i].remote_fwnode = NULL;
>> }
>> }
>> + for (i = 0; i < pdata->nr_inconns; i++) {
>> + pdata->in_conns[i].remote_dev = NULL;
>> + pdata->in_conns[i].remote_fwnode = NULL;
>> + }
>> +
>> if (csdev)
>> coresight_remove_conns_sysfs_group(csdev);
>> }
>> -
>> struct coresight_device *coresight_register(struct coresight_desc
>> *desc)
>> {
>> int ret;
>> @@ -1659,6 +1705,8 @@ struct coresight_device
>> *coresight_register(struct coresight_desc *desc)
>> ret = coresight_create_conns_sysfs_group(csdev);
>> if (!ret)
>> ret = coresight_fixup_device_conns(csdev);
>> + if (!ret)
>> + ret = coresight_fixup_inputs(csdev);
> Don't we also need to fixup the "input" connections on this
> device ?
> It would be good move all of this connection stuff
> into a wrapper function.
>
Should be simplified now with coresight_fixup_device_conns() dropped.
>> if (!ret)
>> ret = coresight_fixup_orphan_conns(csdev);
>> diff --git a/drivers/hwtracing/coresight/coresight-platform.c
>> b/drivers/hwtracing/coresight/coresight-platform.c
>> index 16553f7dde12..20e3351cbdc2 100644
>> --- a/drivers/hwtracing/coresight/coresight-platform.c
>> +++ b/drivers/hwtracing/coresight/coresight-platform.c
>> @@ -20,8 +20,7 @@
>> #include "coresight-priv.h"
>> /*
>> - * coresight_alloc_conns: Allocate connections record for each output
>> - * port from the device.
>> + * coresight_alloc_conns: Allocate connections record for each
>> input/output device.
>> */
>> static int coresight_alloc_conns(struct device *dev,
>> struct coresight_platform_data *pdata)
>> @@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
>> if (!pdata->out_conns)
>> return -ENOMEM;
>> }
>> -
>> + if (pdata->nr_inconns) {
>> + pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
>> + pdata->nr_inconns,
>> + sizeof(*pdata->in_conns),
>> + GFP_KERNEL | __GFP_ZERO);
>> + if (!pdata->in_conns)
>> + return -ENOMEM;
>> + }
>
> devm_krealloc_array() skips any allocation if the requested new_size <=
> origin_size. So this should be fine for the out_conns, to not realloc
> the space again unnecessarily.
Now out_conns is contiguous so adding always allocates.
>
>> return 0;
>> }
>> @@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
>> }
>> EXPORT_SYMBOL_GPL(coresight_add_conn);
>
> Should this be named as add_out_conn in the previous patch
> just to make it explicit and easier to read.
Done
>
>> +/*
>> + * Add a connection in the first free slot, or realloc
>> + * if there is no space.
>> + *
>> + * Do nothing if the connection already exists because inputs are
>> + * fixed up multiple times.
>> + */
>> +int coresight_add_in_conn(struct device *dev,
>> + struct coresight_platform_data *pdata,
>
> minor nit: Is there any reason why this can't simply be:
>
> int coresight_add_in_conn(struct coresight_device *csdev,
> struct coresight_connection *conn)
> {
> struct device *dev = csdev->dev.parent;
> struct coresight_platform_data *pdata = csdev->pdata;
>
No I can change it. I was going for consistency with
coresight_add_conn() which does need the device argument because it's
called both before and after the coresight_device is created. But yeah
it makes sense to simplify this one.
>> + struct coresight_connection *conn)
>
>
> Suzuki
>
>
>> +{
>> + int ret;
>> + struct coresight_connection *free_conn = NULL;
>> + int i;
>> +
>> + /* Search for a free slot or exit if a duplicate is found */
>> + if (pdata->in_conns) {
>> + for (i = 0; i < pdata->nr_inconns; ++i) {
>> + if (!free_conn && !pdata->in_conns[i].remote_fwnode)
>> + free_conn = &pdata->in_conns[i];
>> + if (pdata->in_conns[i].remote_fwnode ==
>> + conn->remote_fwnode)
>> + return 0;
>> + }
>> + }
>> +
>> + if (!free_conn) {
>> + pdata->nr_inconns++;
>> + ret = coresight_alloc_conns(dev, pdata);
>> + if (ret)
>> + return ret;
>> + free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
>> + }
>> +
>> + *free_conn = *conn;
>> + return 0;
>> +}
>> +EXPORT_SYMBOL_GPL(coresight_add_in_conn);
>> +
>> static struct device *
>> coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
>> {
>> @@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
>> /*
>> * of_coresight_parse_endpoint : Parse the given output endpoint @ep
>> - * and fill the connection information in @conn
>> + * and fill the connection information in @in_conn and @out_conn
>> *
>> * Parses the local port, remote device name and the remote port.
>> *
>> @@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct
>> device *dev,
>> /* Get the number of input and output port for this component */
>> of_coresight_get_ports(node, &pdata->nr_inconns,
>> &pdata->nr_outconns);
>> - /* If there are no output connections, we are done */
>> - if (!pdata->nr_outconns)
>> - return 0;
>> -
>> ret = coresight_alloc_conns(dev, pdata);
>> if (ret)
>> return ret;
>> + /* If there are no output connections, we are done */
>> + if (!pdata->nr_outconns)
>> + return 0;
>> +
>> parent = of_coresight_get_output_ports_node(node);
>> /*
>> * If the DT uses obsoleted bindings, the ports are listed
>> diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c
>> b/drivers/hwtracing/coresight/coresight-sysfs.c
>> index 3da9868d9237..2abf9639ac0f 100644
>> --- a/drivers/hwtracing/coresight/coresight-sysfs.c
>> +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
>> @@ -202,5 +202,4 @@ void coresight_remove_links(struct
>> coresight_device *orig,
>> devm_kfree(&orig->dev, conn->link->orig_name);
>> devm_kfree(&orig->dev, conn->link);
>> conn->link = NULL;
>> - conn->remote_dev = NULL;
>> }
>> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
>> index 47fa58d6981d..fd268b24c761 100644
>> --- a/include/linux/coresight.h
>> +++ b/include/linux/coresight.h
>> @@ -110,6 +110,7 @@ struct coresight_platform_data {
>> int nr_inconns;
>> int nr_outconns;
>> struct coresight_connection *out_conns;
>> + struct coresight_connection *in_conns;
>
> Please document the field above the structure.
>
> Suzuki
Done
>
>
>> };
>> /**
>> @@ -177,6 +178,27 @@ struct coresight_desc {
>> * @remote_fwnode once the remote's coresight_device has
>> * been created.
>> * @link: Representation of the connection as a sysfs link.
>> + *
>> + * The full connection structure looks like this, where in_conns
>> store references to
>> + * the parent device in the same remote_dev member as output
>> connections.
>> + *
>> + * +-----------------------------+
>> +-----------------------------+
>> + * |coresight_device |
>> |coresight_connection |
>> + * |-----------------------------|
>> |-----------------------------|
>> + * ---->| |
>> | |
>> + * | | |
>> | remote_dev*|------
>> + * |
>> |pdata->out_conns[nr_outconns]|----------->| | |
>> + * | | |
>> | | |
>> + * | +-----------------------------+
>> +-----------------------------+ |
>> + *
>> | |
>> + * | +-----------------------------+
>> +-----------------------------+ |
>> + * | |coresight_connection |
>> |coresight_device | |
>> + * | |-----------------------------|
>> |------------------------------ |
>> + * | | |
>> | |<-----
>> + * -----|remote_dev* |
>> | |
>> + * |
>> |<-----------|pdata->in_conns[nr_inconns] |
>> + * | |
>> | |
>> + * +-----------------------------+
>> +-----------------------------+
>> */
>> struct coresight_connection {
>> int port;
>> @@ -619,5 +641,8 @@ struct coresight_platform_data
>> *coresight_get_platform_data(struct device *dev);
>> int coresight_add_conn(struct device *dev,
>> struct coresight_platform_data *pdata,
>> const struct coresight_connection *conn);
>> +int coresight_add_in_conn(struct device *dev,
>> + struct coresight_platform_data *pdata,
>> + struct coresight_connection *conn);
>> #endif /* _LINUX_COREISGHT_H */
>
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
@ 2023-03-29 11:59 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 11:59 UTC (permalink / raw)
To: Suzuki K Poulose, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 16/03/2023 20:23, Suzuki K Poulose wrote:
> On 10/03/2023 16:06, James Clark wrote:
>> This will allow CATU to get its associated ETR in a generic way where
>> currently the enable path has some hard coded searches which avoid
>> the need to store input connections.
>>
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>> drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
>> .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
>> drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
>> include/linux/coresight.h | 25 ++++++++
>> 4 files changed, 130 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-core.c
>> b/drivers/hwtracing/coresight/coresight-core.c
>> index f457914e445e..a8ba7493c09a 100644
>> --- a/drivers/hwtracing/coresight/coresight-core.c
>> +++ b/drivers/hwtracing/coresight/coresight-core.c
>> @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff,
>> 0x7fffffff, 0x7fffffff, 0x7fff
>> EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
>> static const struct cti_assoc_op *cti_assoc_ops;
>> +static int coresight_fixup_inputs(struct coresight_device *csdev);
>> ssize_t coresight_simple_show_pair(struct device *_dev,
>> struct device_attribute *attr, char *buf)
>> @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct
>> coresight_device *csdev)
>> csdev, coresight_orphan_match);
>> }
>> +/*
>> + * Device connections are discovered before one/both devices have
>> been created,
>> + * so inputs must be added later.
>> + */
>> +static int coresight_fixup_inputs(struct coresight_device *csdev)
>> +{
>> + int i, ret = 0;
>> + struct coresight_connection *out_conn;
>> + struct coresight_connection in_conn;
>> +
>> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>> + out_conn = &csdev->pdata->out_conns[i];
>> + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
>> + continue;
>> +
>> + /* Reverse local/remote relationships for inputs */
>> + in_conn.remote_dev = csdev;
>> + in_conn.remote_port = out_conn->port;
>> + in_conn.port = out_conn->remote_port;
>> + in_conn.remote_fwnode = csdev->dev.fwnode;
>> + ret = coresight_add_in_conn(out_conn->remote_dev->dev.parent,
>> + out_conn->remote_dev->pdata,
>> + &in_conn);
>> + if (ret)
>> + return ret;
>> + }
>> +
>> + return 0;
>> +}
>> static int coresight_fixup_device_conns(struct coresight_device
>> *csdev)
>> {
>> @@ -1427,11 +1457,20 @@ static int coresight_remove_match(struct
>> device *dev, void *data)
>
> Removing the connection should be much simpler now with the inports
> connection tracking.
>
> i.e., coresight_remove_conns() need not iterate over all the devices on
> the coresight bus. We had to do that to find all the devices connecting
> to the removed device.
Done in v3
>
> Instead with the in_conns we could readily get the csdev instance and
> remove the connection from those devices.
>
> Similarly for the out_conns can give you the remove device for removing
> the in_conns on that device.
>
> i.e,
>
> static void coresight_remove_conns(csdev)
> {
> struct coresight_device *remote;
>
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> remote = csdev->data->out_conns[i].remote_dev;
>
> if (!remote)
> continue;
>
> for (j = 0; j < remote->pdata->nr_inconns; j++) {
> if (remote->pdata->in_conns[i].remote_dev == csdev)
> /* Clear in_conns[i] on remote */
>
> }
> }
>
> Similarly for in_conns
> for (i = 0; i < csdev->pdata->nr_inconns; i++) {
> remote = csdev->data->in_conns[i].remote_dev;
>
> if (!remote)
> continue;
>
> for (j = 0; j < remote->pdata->nr_outconns; j++) {
> if (remote->pdata->out_conns[i].remote_dev == csdev)
> /* Clear out_conns[i] on remote */
> }
>
>
>> */
>> fwnode_handle_put(conn->remote_fwnode);
>> conn->remote_fwnode = NULL;
>> + conn->remote_dev = NULL;
>> + /* No need to continue */
>> + break;
>> + }
>> + }
>> + for (i = 0; i < iterator->pdata->nr_inconns; i++) {
>> + conn = &iterator->pdata->in_conns[i];
>> + if (csdev == conn->remote_dev) {
>> + conn->remote_fwnode = NULL;
>> + conn->remote_dev = NULL;
>> /* No need to continue */
>> break;
>> }
>> }
>> -
>> /*
>> * Returning '0' ensures that all known component on the
>> * bus will be checked.
>> @@ -1552,21 +1591,28 @@ void coresight_release_platform_data(struct
>> coresight_device *csdev,
>> for (i = 0; i < pdata->nr_outconns; i++) {
>> /* If we have made the links, remove them now */
>> - if (csdev && conns[i].remote_dev)
>> + if (csdev && conns[i].remote_dev) {
>> coresight_remove_links(csdev, &conns[i]);
>> + conns[i].remote_dev = NULL;
>> + }
>> +
>> /*
>> * Drop the refcount and clear the handle as this device
>> * is going away
>> */
>> if (conns[i].remote_fwnode) {
>> fwnode_handle_put(conns[i].remote_fwnode);
>> - pdata->out_conns[i].remote_fwnode = NULL;
>> + conns[i].remote_fwnode = NULL;
>> }
>> }
>> + for (i = 0; i < pdata->nr_inconns; i++) {
>> + pdata->in_conns[i].remote_dev = NULL;
>> + pdata->in_conns[i].remote_fwnode = NULL;
>> + }
>> +
>> if (csdev)
>> coresight_remove_conns_sysfs_group(csdev);
>> }
>> -
>> struct coresight_device *coresight_register(struct coresight_desc
>> *desc)
>> {
>> int ret;
>> @@ -1659,6 +1705,8 @@ struct coresight_device
>> *coresight_register(struct coresight_desc *desc)
>> ret = coresight_create_conns_sysfs_group(csdev);
>> if (!ret)
>> ret = coresight_fixup_device_conns(csdev);
>> + if (!ret)
>> + ret = coresight_fixup_inputs(csdev);
> Don't we also need to fixup the "input" connections on this
> device ?
> It would be good move all of this connection stuff
> into a wrapper function.
>
Should be simplified now with coresight_fixup_device_conns() dropped.
>> if (!ret)
>> ret = coresight_fixup_orphan_conns(csdev);
>> diff --git a/drivers/hwtracing/coresight/coresight-platform.c
>> b/drivers/hwtracing/coresight/coresight-platform.c
>> index 16553f7dde12..20e3351cbdc2 100644
>> --- a/drivers/hwtracing/coresight/coresight-platform.c
>> +++ b/drivers/hwtracing/coresight/coresight-platform.c
>> @@ -20,8 +20,7 @@
>> #include "coresight-priv.h"
>> /*
>> - * coresight_alloc_conns: Allocate connections record for each output
>> - * port from the device.
>> + * coresight_alloc_conns: Allocate connections record for each
>> input/output device.
>> */
>> static int coresight_alloc_conns(struct device *dev,
>> struct coresight_platform_data *pdata)
>> @@ -33,7 +32,14 @@ static int coresight_alloc_conns(struct device *dev,
>> if (!pdata->out_conns)
>> return -ENOMEM;
>> }
>> -
>> + if (pdata->nr_inconns) {
>> + pdata->in_conns = devm_krealloc_array(dev, pdata->in_conns,
>> + pdata->nr_inconns,
>> + sizeof(*pdata->in_conns),
>> + GFP_KERNEL | __GFP_ZERO);
>> + if (!pdata->in_conns)
>> + return -ENOMEM;
>> + }
>
> devm_krealloc_array() skips any allocation if the requested new_size <=
> origin_size. So this should be fine for the out_conns, to not realloc
> the space again unnecessarily.
Now out_conns is contiguous so adding always allocates.
>
>> return 0;
>> }
>> @@ -79,6 +85,45 @@ int coresight_add_conn(struct device *dev,
>> }
>> EXPORT_SYMBOL_GPL(coresight_add_conn);
>
> Should this be named as add_out_conn in the previous patch
> just to make it explicit and easier to read.
Done
>
>> +/*
>> + * Add a connection in the first free slot, or realloc
>> + * if there is no space.
>> + *
>> + * Do nothing if the connection already exists because inputs are
>> + * fixed up multiple times.
>> + */
>> +int coresight_add_in_conn(struct device *dev,
>> + struct coresight_platform_data *pdata,
>
> minor nit: Is there any reason why this can't simply be:
>
> int coresight_add_in_conn(struct coresight_device *csdev,
> struct coresight_connection *conn)
> {
> struct device *dev = csdev->dev.parent;
> struct coresight_platform_data *pdata = csdev->pdata;
>
No I can change it. I was going for consistency with
coresight_add_conn() which does need the device argument because it's
called both before and after the coresight_device is created. But yeah
it makes sense to simplify this one.
>> + struct coresight_connection *conn)
>
>
> Suzuki
>
>
>> +{
>> + int ret;
>> + struct coresight_connection *free_conn = NULL;
>> + int i;
>> +
>> + /* Search for a free slot or exit if a duplicate is found */
>> + if (pdata->in_conns) {
>> + for (i = 0; i < pdata->nr_inconns; ++i) {
>> + if (!free_conn && !pdata->in_conns[i].remote_fwnode)
>> + free_conn = &pdata->in_conns[i];
>> + if (pdata->in_conns[i].remote_fwnode ==
>> + conn->remote_fwnode)
>> + return 0;
>> + }
>> + }
>> +
>> + if (!free_conn) {
>> + pdata->nr_inconns++;
>> + ret = coresight_alloc_conns(dev, pdata);
>> + if (ret)
>> + return ret;
>> + free_conn = &pdata->in_conns[pdata->nr_inconns - 1];
>> + }
>> +
>> + *free_conn = *conn;
>> + return 0;
>> +}
>> +EXPORT_SYMBOL_GPL(coresight_add_in_conn);
>> +
>> static struct device *
>> coresight_find_device_by_fwnode(struct fwnode_handle *fwnode)
>> {
>> @@ -249,7 +294,7 @@ static int of_coresight_get_cpu(struct device *dev)
>> /*
>> * of_coresight_parse_endpoint : Parse the given output endpoint @ep
>> - * and fill the connection information in @conn
>> + * and fill the connection information in @in_conn and @out_conn
>> *
>> * Parses the local port, remote device name and the remote port.
>> *
>> @@ -333,14 +378,14 @@ static int of_get_coresight_platform_data(struct
>> device *dev,
>> /* Get the number of input and output port for this component */
>> of_coresight_get_ports(node, &pdata->nr_inconns,
>> &pdata->nr_outconns);
>> - /* If there are no output connections, we are done */
>> - if (!pdata->nr_outconns)
>> - return 0;
>> -
>> ret = coresight_alloc_conns(dev, pdata);
>> if (ret)
>> return ret;
>> + /* If there are no output connections, we are done */
>> + if (!pdata->nr_outconns)
>> + return 0;
>> +
>> parent = of_coresight_get_output_ports_node(node);
>> /*
>> * If the DT uses obsoleted bindings, the ports are listed
>> diff --git a/drivers/hwtracing/coresight/coresight-sysfs.c
>> b/drivers/hwtracing/coresight/coresight-sysfs.c
>> index 3da9868d9237..2abf9639ac0f 100644
>> --- a/drivers/hwtracing/coresight/coresight-sysfs.c
>> +++ b/drivers/hwtracing/coresight/coresight-sysfs.c
>> @@ -202,5 +202,4 @@ void coresight_remove_links(struct
>> coresight_device *orig,
>> devm_kfree(&orig->dev, conn->link->orig_name);
>> devm_kfree(&orig->dev, conn->link);
>> conn->link = NULL;
>> - conn->remote_dev = NULL;
>> }
>> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
>> index 47fa58d6981d..fd268b24c761 100644
>> --- a/include/linux/coresight.h
>> +++ b/include/linux/coresight.h
>> @@ -110,6 +110,7 @@ struct coresight_platform_data {
>> int nr_inconns;
>> int nr_outconns;
>> struct coresight_connection *out_conns;
>> + struct coresight_connection *in_conns;
>
> Please document the field above the structure.
>
> Suzuki
Done
>
>
>> };
>> /**
>> @@ -177,6 +178,27 @@ struct coresight_desc {
>> * @remote_fwnode once the remote's coresight_device has
>> * been created.
>> * @link: Representation of the connection as a sysfs link.
>> + *
>> + * The full connection structure looks like this, where in_conns
>> store references to
>> + * the parent device in the same remote_dev member as output
>> connections.
>> + *
>> + * +-----------------------------+
>> +-----------------------------+
>> + * |coresight_device |
>> |coresight_connection |
>> + * |-----------------------------|
>> |-----------------------------|
>> + * ---->| |
>> | |
>> + * | | |
>> | remote_dev*|------
>> + * |
>> |pdata->out_conns[nr_outconns]|----------->| | |
>> + * | | |
>> | | |
>> + * | +-----------------------------+
>> +-----------------------------+ |
>> + *
>> | |
>> + * | +-----------------------------+
>> +-----------------------------+ |
>> + * | |coresight_connection |
>> |coresight_device | |
>> + * | |-----------------------------|
>> |------------------------------ |
>> + * | | |
>> | |<-----
>> + * -----|remote_dev* |
>> | |
>> + * |
>> |<-----------|pdata->in_conns[nr_inconns] |
>> + * | |
>> | |
>> + * +-----------------------------+
>> +-----------------------------+
>> */
>> struct coresight_connection {
>> int port;
>> @@ -619,5 +641,8 @@ struct coresight_platform_data
>> *coresight_get_platform_data(struct device *dev);
>> int coresight_add_conn(struct device *dev,
>> struct coresight_platform_data *pdata,
>> const struct coresight_connection *conn);
>> +int coresight_add_in_conn(struct device *dev,
>> + struct coresight_platform_data *pdata,
>> + struct coresight_connection *conn);
>> #endif /* _LINUX_COREISGHT_H */
>
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 5/9] coresight: Dynamically add connections
2023-03-16 17:12 ` Suzuki K Poulose
@ 2023-03-29 12:02 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 12:02 UTC (permalink / raw)
To: Suzuki K Poulose, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 16/03/2023 17:12, Suzuki K Poulose wrote:
> On 10/03/2023 16:06, James Clark wrote:
>> Add a function for adding connections dynamically. This also removes
>> the 1:1 mapping between port number and the index into the connections
>> array. The only place this mapping was used was in the warning for
>> duplicate output ports, which has been replaced by a search. Other
>> uses of the port number already use the port member variable.
>>
>> Being able to dynamically add connections will allow other devices like
>> CTI to re-use the connection mechanism despite not having explicit
>> connections described in the DT.
>>
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>> .../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
>> include/linux/coresight.h | 7 +-
>> 2 files changed, 64 insertions(+), 20 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-platform.c
>> b/drivers/hwtracing/coresight/coresight-platform.c
>> index c77238cdf448..16553f7dde12 100644
>> --- a/drivers/hwtracing/coresight/coresight-platform.c
>> +++ b/drivers/hwtracing/coresight/coresight-platform.c
>> @@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
>> struct coresight_platform_data *pdata)
>> {
>> if (pdata->nr_outconns) {
>> - pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
>> - sizeof(*pdata->out_conns), GFP_KERNEL);
>> + pdata->out_conns = devm_krealloc_array(
>> + dev, pdata->out_conns, pdata->nr_outconns,
>
> super minor nit:
> pdata->out_conns = devm_krealloc_array(dev,
>
This is actually clang-format's doing when using the kernel rules in the
root of the repo. I started using it because of some other style
comments I got before. Not sure if this time it's just done something
bad or it's technically ok.
I formatted everything in V3 with it so it should at least all be
consistent.
>
>> + sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
>> if (!pdata->out_conns)
>> return -ENOMEM;
>> }
>> @@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
>> return 0;
>> }
>> +/*
>> + * Add a connection in the first free slot, or realloc
>> + * if there is no space. @conn's contents is copied into the new slot.
>> + *
>> + * If the output port is already assigned on this device, return -EINVAL
>> + */
>> +int coresight_add_conn(struct device *dev,
>> + struct coresight_platform_data *pdata,
>> + const struct coresight_connection *conn)
>> +{
>> + int ret;
>> + struct coresight_connection *free_conn = NULL;
>> + struct coresight_connection *i;
>> +
>> + /*
>> + * Search for a free slot, and while looking for one, warn
>> + * on any existing duplicate output port.
>> + */
>> + for (i = pdata->out_conns; i < pdata->out_conns +
>> pdata->nr_outconns;
>> + ++i) {
>
> minor nit: I see why you have gone against using "i" as index into
> the array. But I think having that as the index is still better
> readable.
> for (i = 0; i < pdata->nr_outconns; i++) {
> struct coresight_connection *c = &pdata->out_conns[i];
> >> + if (i->remote_fwnode && conn->port != -1 &&
>> + i->port == conn->port) {
>> + dev_warn(dev, "Duplicate output port %d\n", i->port);
>> + return -EINVAL;
>> + }
>> + if (!i->remote_fwnode && !free_conn)
>> + free_conn = i;
>> + }
>> +
>> + if (!free_conn) {
>
> and:
> /* No free slots */
> if (i == pdata->nr_outconns) {
>
>> + pdata->nr_outconns++;
>> + ret = coresight_alloc_conns(dev, pdata);
>> + if (ret)
>> + return ret;
>> + free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
>> + }
>> +
>
> and:
> pdata->out_conns[i] = *conn;
>
> > Otherwise looks good to me.
>
> Suzuki
>
>
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 5/9] coresight: Dynamically add connections
@ 2023-03-29 12:02 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 12:02 UTC (permalink / raw)
To: Suzuki K Poulose, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 16/03/2023 17:12, Suzuki K Poulose wrote:
> On 10/03/2023 16:06, James Clark wrote:
>> Add a function for adding connections dynamically. This also removes
>> the 1:1 mapping between port number and the index into the connections
>> array. The only place this mapping was used was in the warning for
>> duplicate output ports, which has been replaced by a search. Other
>> uses of the port number already use the port member variable.
>>
>> Being able to dynamically add connections will allow other devices like
>> CTI to re-use the connection mechanism despite not having explicit
>> connections described in the DT.
>>
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>> .../hwtracing/coresight/coresight-platform.c | 77 ++++++++++++++-----
>> include/linux/coresight.h | 7 +-
>> 2 files changed, 64 insertions(+), 20 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-platform.c
>> b/drivers/hwtracing/coresight/coresight-platform.c
>> index c77238cdf448..16553f7dde12 100644
>> --- a/drivers/hwtracing/coresight/coresight-platform.c
>> +++ b/drivers/hwtracing/coresight/coresight-platform.c
>> @@ -27,8 +27,9 @@ static int coresight_alloc_conns(struct device *dev,
>> struct coresight_platform_data *pdata)
>> {
>> if (pdata->nr_outconns) {
>> - pdata->out_conns = devm_kcalloc(dev, pdata->nr_outconns,
>> - sizeof(*pdata->out_conns), GFP_KERNEL);
>> + pdata->out_conns = devm_krealloc_array(
>> + dev, pdata->out_conns, pdata->nr_outconns,
>
> super minor nit:
> pdata->out_conns = devm_krealloc_array(dev,
>
This is actually clang-format's doing when using the kernel rules in the
root of the repo. I started using it because of some other style
comments I got before. Not sure if this time it's just done something
bad or it's technically ok.
I formatted everything in V3 with it so it should at least all be
consistent.
>
>> + sizeof(*pdata->out_conns), GFP_KERNEL | __GFP_ZERO);
>> if (!pdata->out_conns)
>> return -ENOMEM;
>> }
>> @@ -36,6 +37,48 @@ static int coresight_alloc_conns(struct device *dev,
>> return 0;
>> }
>> +/*
>> + * Add a connection in the first free slot, or realloc
>> + * if there is no space. @conn's contents is copied into the new slot.
>> + *
>> + * If the output port is already assigned on this device, return -EINVAL
>> + */
>> +int coresight_add_conn(struct device *dev,
>> + struct coresight_platform_data *pdata,
>> + const struct coresight_connection *conn)
>> +{
>> + int ret;
>> + struct coresight_connection *free_conn = NULL;
>> + struct coresight_connection *i;
>> +
>> + /*
>> + * Search for a free slot, and while looking for one, warn
>> + * on any existing duplicate output port.
>> + */
>> + for (i = pdata->out_conns; i < pdata->out_conns +
>> pdata->nr_outconns;
>> + ++i) {
>
> minor nit: I see why you have gone against using "i" as index into
> the array. But I think having that as the index is still better
> readable.
> for (i = 0; i < pdata->nr_outconns; i++) {
> struct coresight_connection *c = &pdata->out_conns[i];
> >> + if (i->remote_fwnode && conn->port != -1 &&
>> + i->port == conn->port) {
>> + dev_warn(dev, "Duplicate output port %d\n", i->port);
>> + return -EINVAL;
>> + }
>> + if (!i->remote_fwnode && !free_conn)
>> + free_conn = i;
>> + }
>> +
>> + if (!free_conn) {
>
> and:
> /* No free slots */
> if (i == pdata->nr_outconns) {
>
>> + pdata->nr_outconns++;
>> + ret = coresight_alloc_conns(dev, pdata);
>> + if (ret)
>> + return ret;
>> + free_conn = &pdata->out_conns[pdata->nr_outconns - 1];
>> + }
>> +
>
> and:
> pdata->out_conns[i] = *conn;
>
> > Otherwise looks good to me.
>
> Suzuki
>
>
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
2023-03-17 11:04 ` Suzuki K Poulose
@ 2023-03-29 12:04 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 12:04 UTC (permalink / raw)
To: Suzuki K Poulose, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 17/03/2023 11:04, Suzuki K Poulose wrote:
> On 10/03/2023 16:06, James Clark wrote:
>> Currently CATU is the only helper device, and its enable and disable
>> calls are hard coded. To allow more helper devices to be added in a
>> generic way, remove these hard coded calls and just enable and disable
>> all helper devices.
>>
>> This has to apply to helpers adjacent to the path, because they will
>> never be in the path. CATU was already discovered in this way, so
>> there is no change there.
>>
>> One change that is needed is for CATU to call back into ETR to allocate
>> the buffer. Because the enable call was previously hard coded, it was
>> done at a point where the buffer was already allocated, but this is no
>> longer the case.
>>
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>> drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
>> drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
>> .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
>> include/linux/coresight.h | 3 +-
>> 4 files changed, 99 insertions(+), 34 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-catu.c
>> b/drivers/hwtracing/coresight/coresight-catu.c
>> index bc90a03f478f..24a08a2b96b1 100644
>> --- a/drivers/hwtracing/coresight/coresight-catu.c
>> +++ b/drivers/hwtracing/coresight/coresight-catu.c
>> @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct
>> catu_drvdata *drvdata)
>> return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
>> }
>> -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
>> +static struct coresight_device *
>> +catu_get_etr_device(struct coresight_device *csdev)
>> +{
>> + int i;
>> + struct coresight_device *tmp;
>> +
>> + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
>> + tmp = csdev->pdata->in_conns[i].remote_dev;
>> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
>> + tmp->subtype.sink_subtype ==
>> + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
>> + return tmp;
>> + }
>> +
>> + return NULL;
>> +}
>> +
>> +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode
>> cs_mode,
>> + void *data)
>> {
>> int rc;
>> u32 control, mode;
>> - struct etr_buf *etr_buf = data;
>> + struct etr_buf *etr_buf = NULL;
>> struct device *dev = &drvdata->csdev->dev;
>> struct coresight_device *csdev = drvdata->csdev;
>> + struct coresight_device *etrdev;
>> if (catu_wait_for_ready(drvdata))
>> dev_warn(dev, "Timeout while waiting for READY\n");
>> @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata
>> *drvdata, void *data)
>> if (rc)
>> return rc;
>> + etrdev = catu_get_etr_device(csdev);
>> + if (etrdev) {
>> + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
>> + if (IS_ERR(etr_buf))
>> + return PTR_ERR(etr_buf);
>> + }
>
> WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
> return.
>
I saw there was the pass-through mode below which I thought didn't need
an ETR device. I think I followed the code through and there was a way
for it to get there without an ETR in the existing version, but now I'm
not sure. Or does it still need the ETR device but it just doesn't
access the buffer?
>
>> control |= BIT(CATU_CONTROL_ENABLE);
>> if (etr_buf && etr_buf->mode == ETR_MODE_CATU) {
>> @@ -441,13 +466,14 @@ static int catu_enable_hw(struct catu_drvdata
>> *drvdata, void *data)
>> return 0;
>> }
>> -static int catu_enable(struct coresight_device *csdev, void *data)
>> +static int catu_enable(struct coresight_device *csdev, enum cs_mode
>> mode,
>> + void *data)
>> {
>> int rc;
>> struct catu_drvdata *catu_drvdata = csdev_to_catu_drvdata(csdev);
>> CS_UNLOCK(catu_drvdata->base);
>> - rc = catu_enable_hw(catu_drvdata, data);
>> + rc = catu_enable_hw(catu_drvdata, mode, data);
>> CS_LOCK(catu_drvdata->base);
>> return rc;
>> }
>> diff --git a/drivers/hwtracing/coresight/coresight-core.c
>> b/drivers/hwtracing/coresight/coresight-core.c
>> index a8ba7493c09a..3e6ccd9e8d4e 100644
>> --- a/drivers/hwtracing/coresight/coresight-core.c
>> +++ b/drivers/hwtracing/coresight/coresight-core.c
>> @@ -441,6 +441,34 @@ static int coresight_enable_source(struct
>> coresight_device *csdev,
>> return 0;
>> }
>> +static int coresight_enable_helper(struct coresight_device *csdev,
>> + enum cs_mode mode, void *sink_data)
>
> minor nit: s/sink_data/data/ ? Though it is always either sink_data
> (perf mode) or NULL (sysfs mode), for the core code it is simply an
> opaque data.
>
Done
> Rest looks fine to me.
>
> Suzuki
>
>
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
@ 2023-03-29 12:04 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 12:04 UTC (permalink / raw)
To: Suzuki K Poulose, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 17/03/2023 11:04, Suzuki K Poulose wrote:
> On 10/03/2023 16:06, James Clark wrote:
>> Currently CATU is the only helper device, and its enable and disable
>> calls are hard coded. To allow more helper devices to be added in a
>> generic way, remove these hard coded calls and just enable and disable
>> all helper devices.
>>
>> This has to apply to helpers adjacent to the path, because they will
>> never be in the path. CATU was already discovered in this way, so
>> there is no change there.
>>
>> One change that is needed is for CATU to call back into ETR to allocate
>> the buffer. Because the enable call was previously hard coded, it was
>> done at a point where the buffer was already allocated, but this is no
>> longer the case.
>>
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>> drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
>> drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
>> .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
>> include/linux/coresight.h | 3 +-
>> 4 files changed, 99 insertions(+), 34 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-catu.c
>> b/drivers/hwtracing/coresight/coresight-catu.c
>> index bc90a03f478f..24a08a2b96b1 100644
>> --- a/drivers/hwtracing/coresight/coresight-catu.c
>> +++ b/drivers/hwtracing/coresight/coresight-catu.c
>> @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct
>> catu_drvdata *drvdata)
>> return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
>> }
>> -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
>> +static struct coresight_device *
>> +catu_get_etr_device(struct coresight_device *csdev)
>> +{
>> + int i;
>> + struct coresight_device *tmp;
>> +
>> + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
>> + tmp = csdev->pdata->in_conns[i].remote_dev;
>> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
>> + tmp->subtype.sink_subtype ==
>> + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
>> + return tmp;
>> + }
>> +
>> + return NULL;
>> +}
>> +
>> +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode
>> cs_mode,
>> + void *data)
>> {
>> int rc;
>> u32 control, mode;
>> - struct etr_buf *etr_buf = data;
>> + struct etr_buf *etr_buf = NULL;
>> struct device *dev = &drvdata->csdev->dev;
>> struct coresight_device *csdev = drvdata->csdev;
>> + struct coresight_device *etrdev;
>> if (catu_wait_for_ready(drvdata))
>> dev_warn(dev, "Timeout while waiting for READY\n");
>> @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata
>> *drvdata, void *data)
>> if (rc)
>> return rc;
>> + etrdev = catu_get_etr_device(csdev);
>> + if (etrdev) {
>> + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
>> + if (IS_ERR(etr_buf))
>> + return PTR_ERR(etr_buf);
>> + }
>
> WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
> return.
>
I saw there was the pass-through mode below which I thought didn't need
an ETR device. I think I followed the code through and there was a way
for it to get there without an ETR in the existing version, but now I'm
not sure. Or does it still need the ETR device but it just doesn't
access the buffer?
>
>> control |= BIT(CATU_CONTROL_ENABLE);
>> if (etr_buf && etr_buf->mode == ETR_MODE_CATU) {
>> @@ -441,13 +466,14 @@ static int catu_enable_hw(struct catu_drvdata
>> *drvdata, void *data)
>> return 0;
>> }
>> -static int catu_enable(struct coresight_device *csdev, void *data)
>> +static int catu_enable(struct coresight_device *csdev, enum cs_mode
>> mode,
>> + void *data)
>> {
>> int rc;
>> struct catu_drvdata *catu_drvdata = csdev_to_catu_drvdata(csdev);
>> CS_UNLOCK(catu_drvdata->base);
>> - rc = catu_enable_hw(catu_drvdata, data);
>> + rc = catu_enable_hw(catu_drvdata, mode, data);
>> CS_LOCK(catu_drvdata->base);
>> return rc;
>> }
>> diff --git a/drivers/hwtracing/coresight/coresight-core.c
>> b/drivers/hwtracing/coresight/coresight-core.c
>> index a8ba7493c09a..3e6ccd9e8d4e 100644
>> --- a/drivers/hwtracing/coresight/coresight-core.c
>> +++ b/drivers/hwtracing/coresight/coresight-core.c
>> @@ -441,6 +441,34 @@ static int coresight_enable_source(struct
>> coresight_device *csdev,
>> return 0;
>> }
>> +static int coresight_enable_helper(struct coresight_device *csdev,
>> + enum cs_mode mode, void *sink_data)
>
> minor nit: s/sink_data/data/ ? Though it is always either sink_data
> (perf mode) or NULL (sysfs mode), for the core code it is simply an
> opaque data.
>
Done
> Rest looks fine to me.
>
> Suzuki
>
>
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
2023-03-14 5:35 ` Jinlong Mao
@ 2023-03-29 12:06 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 12:06 UTC (permalink / raw)
To: Jinlong Mao, coresight
Cc: Mathieu Poirier, Mike Leach, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 14/03/2023 05:35, Jinlong Mao wrote:
>
> On 3/11/2023 12:06 AM, James Clark wrote:
>> This will allow CATU to get its associated ETR in a generic way where
>> currently the enable path has some hard coded searches which avoid
>> the need to store input connections.
>>
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>> drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
>> .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
>> drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
>> include/linux/coresight.h | 25 ++++++++
>> 4 files changed, 130 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-core.c
>> b/drivers/hwtracing/coresight/coresight-core.c
>> index f457914e445e..a8ba7493c09a 100644
>> --- a/drivers/hwtracing/coresight/coresight-core.c
>> +++ b/drivers/hwtracing/coresight/coresight-core.c
>> @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff,
>> 0x7fffffff, 0x7fffffff, 0x7fff
>> EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
>> static const struct cti_assoc_op *cti_assoc_ops;
>> +static int coresight_fixup_inputs(struct coresight_device *csdev);
>> ssize_t coresight_simple_show_pair(struct device *_dev,
>> struct device_attribute *attr, char *buf)
>> @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct
>> coresight_device *csdev)
>> csdev, coresight_orphan_match);
>> }
>> +/*
>> + * Device connections are discovered before one/both devices have
>> been created,
>> + * so inputs must be added later.
>> + */
>> +static int coresight_fixup_inputs(struct coresight_device *csdev)
>> +{
>> + int i, ret = 0;
>> + struct coresight_connection *out_conn;
>> + struct coresight_connection in_conn;
>> +
>> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>> + out_conn = &csdev->pdata->out_conns[i];
>> + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
>> + continue;
>
> Hi James,
>
> If out_conn->remote_dev is null here, the in_conn of
> out_conn->remote_dev->pdata will never be set.
> For example, device A is connected to in_port 0 of device B. If device A
> is probed first, the in_conn of device
> B will not be set.
> Do we need to add Defer probe return here ? I tested with defer probe
> return, it works.
>
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> out_conn = &csdev->pdata->out_conns[i];
> if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
> - continue;
> + return -EPROBE_DEFER;
>
> Thanks
> Jinlong Mao
I think you are right but I thought that EPROBE_DEFER was too big of a
change and that it might break something in some unexpected way.
In V3 I used the orphan mechanism for inputs in the same way as outputs
so the problem should be gone now but without having to defer loading.
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 6/9] coresight: Store in-connections as well as out-connections
@ 2023-03-29 12:06 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 12:06 UTC (permalink / raw)
To: Jinlong Mao, coresight
Cc: Mathieu Poirier, Mike Leach, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 14/03/2023 05:35, Jinlong Mao wrote:
>
> On 3/11/2023 12:06 AM, James Clark wrote:
>> This will allow CATU to get its associated ETR in a generic way where
>> currently the enable path has some hard coded searches which avoid
>> the need to store input connections.
>>
>> Signed-off-by: James Clark <james.clark@arm.com>
>> ---
>> drivers/hwtracing/coresight/coresight-core.c | 56 +++++++++++++++--
>> .../hwtracing/coresight/coresight-platform.c | 61 ++++++++++++++++---
>> drivers/hwtracing/coresight/coresight-sysfs.c | 1 -
>> include/linux/coresight.h | 25 ++++++++
>> 4 files changed, 130 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-core.c
>> b/drivers/hwtracing/coresight/coresight-core.c
>> index f457914e445e..a8ba7493c09a 100644
>> --- a/drivers/hwtracing/coresight/coresight-core.c
>> +++ b/drivers/hwtracing/coresight/coresight-core.c
>> @@ -59,6 +59,7 @@ const u32 coresight_barrier_pkt[4] = {0x7fffffff,
>> 0x7fffffff, 0x7fffffff, 0x7fff
>> EXPORT_SYMBOL_GPL(coresight_barrier_pkt);
>> static const struct cti_assoc_op *cti_assoc_ops;
>> +static int coresight_fixup_inputs(struct coresight_device *csdev);
>> ssize_t coresight_simple_show_pair(struct device *_dev,
>> struct device_attribute *attr, char *buf)
>> @@ -1369,6 +1370,35 @@ static int coresight_fixup_orphan_conns(struct
>> coresight_device *csdev)
>> csdev, coresight_orphan_match);
>> }
>> +/*
>> + * Device connections are discovered before one/both devices have
>> been created,
>> + * so inputs must be added later.
>> + */
>> +static int coresight_fixup_inputs(struct coresight_device *csdev)
>> +{
>> + int i, ret = 0;
>> + struct coresight_connection *out_conn;
>> + struct coresight_connection in_conn;
>> +
>> + for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>> + out_conn = &csdev->pdata->out_conns[i];
>> + if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
>> + continue;
>
> Hi James,
>
> If out_conn->remote_dev is null here, the in_conn of
> out_conn->remote_dev->pdata will never be set.
> For example, device A is connected to in_port 0 of device B. If device A
> is probed first, the in_conn of device
> B will not be set.
> Do we need to add Defer probe return here ? I tested with defer probe
> return, it works.
>
> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
> out_conn = &csdev->pdata->out_conns[i];
> if (!out_conn->remote_dev || !out_conn->remote_dev->pdata)
> - continue;
> + return -EPROBE_DEFER;
>
> Thanks
> Jinlong Mao
I think you are right but I thought that EPROBE_DEFER was too big of a
change and that it might break something in some unexpected way.
In V3 I used the orphan mechanism for inputs in the same way as outputs
so the problem should be gone now but without having to defer loading.
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
2023-03-29 12:04 ` James Clark
@ 2023-03-29 13:23 ` Suzuki K Poulose
-1 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-29 13:23 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 29/03/2023 13:04, James Clark wrote:
>
>
> On 17/03/2023 11:04, Suzuki K Poulose wrote:
>> On 10/03/2023 16:06, James Clark wrote:
>>> Currently CATU is the only helper device, and its enable and disable
>>> calls are hard coded. To allow more helper devices to be added in a
>>> generic way, remove these hard coded calls and just enable and disable
>>> all helper devices.
>>>
>>> This has to apply to helpers adjacent to the path, because they will
>>> never be in the path. CATU was already discovered in this way, so
>>> there is no change there.
>>>
>>> One change that is needed is for CATU to call back into ETR to allocate
>>> the buffer. Because the enable call was previously hard coded, it was
>>> done at a point where the buffer was already allocated, but this is no
>>> longer the case.
>>>
>>> Signed-off-by: James Clark <james.clark@arm.com>
>>> ---
>>> drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
>>> drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
>>> .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
>>> include/linux/coresight.h | 3 +-
>>> 4 files changed, 99 insertions(+), 34 deletions(-)
>>>
>>> diff --git a/drivers/hwtracing/coresight/coresight-catu.c
>>> b/drivers/hwtracing/coresight/coresight-catu.c
>>> index bc90a03f478f..24a08a2b96b1 100644
>>> --- a/drivers/hwtracing/coresight/coresight-catu.c
>>> +++ b/drivers/hwtracing/coresight/coresight-catu.c
>>> @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct
>>> catu_drvdata *drvdata)
>>> return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
>>> }
>>> -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
>>> +static struct coresight_device *
>>> +catu_get_etr_device(struct coresight_device *csdev)
>>> +{
>>> + int i;
>>> + struct coresight_device *tmp;
>>> +
>>> + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
>>> + tmp = csdev->pdata->in_conns[i].remote_dev;
>>> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
>>> + tmp->subtype.sink_subtype ==
>>> + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
>>> + return tmp;
>>> + }
>>> +
>>> + return NULL;
>>> +}
>>> +
>>> +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode
>>> cs_mode,
>>> + void *data)
>>> {
>>> int rc;
>>> u32 control, mode;
>>> - struct etr_buf *etr_buf = data;
>>> + struct etr_buf *etr_buf = NULL;
>>> struct device *dev = &drvdata->csdev->dev;
>>> struct coresight_device *csdev = drvdata->csdev;
>>> + struct coresight_device *etrdev;
>>> if (catu_wait_for_ready(drvdata))
>>> dev_warn(dev, "Timeout while waiting for READY\n");
>>> @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata
>>> *drvdata, void *data)
>>> if (rc)
>>> return rc;
>>> + etrdev = catu_get_etr_device(csdev);
>>> + if (etrdev) {
>>> + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
>>> + if (IS_ERR(etr_buf))
>>> + return PTR_ERR(etr_buf);
>>> + }
>>
>> WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
>> return.
>>
>
> I saw there was the pass-through mode below which I thought didn't need
> an ETR device. I think I followed the code through and there was a way
> for it to get there without an ETR in the existing version, but now I'm
> not sure.
> Or does it still need the ETR device but it just doesn't
> access the buffer?
The first part is correct. Without an ETR, CATU wouldn't be a helper
device, and wouldn't get here in "enable CATU" via the helper route.
The CATU chooses the mode depending on the etr_buf mode.
Suzuki
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
@ 2023-03-29 13:23 ` Suzuki K Poulose
0 siblings, 0 replies; 74+ messages in thread
From: Suzuki K Poulose @ 2023-03-29 13:23 UTC (permalink / raw)
To: James Clark, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 29/03/2023 13:04, James Clark wrote:
>
>
> On 17/03/2023 11:04, Suzuki K Poulose wrote:
>> On 10/03/2023 16:06, James Clark wrote:
>>> Currently CATU is the only helper device, and its enable and disable
>>> calls are hard coded. To allow more helper devices to be added in a
>>> generic way, remove these hard coded calls and just enable and disable
>>> all helper devices.
>>>
>>> This has to apply to helpers adjacent to the path, because they will
>>> never be in the path. CATU was already discovered in this way, so
>>> there is no change there.
>>>
>>> One change that is needed is for CATU to call back into ETR to allocate
>>> the buffer. Because the enable call was previously hard coded, it was
>>> done at a point where the buffer was already allocated, but this is no
>>> longer the case.
>>>
>>> Signed-off-by: James Clark <james.clark@arm.com>
>>> ---
>>> drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
>>> drivers/hwtracing/coresight/coresight-core.c | 68 ++++++++++++++++++-
>>> .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
>>> include/linux/coresight.h | 3 +-
>>> 4 files changed, 99 insertions(+), 34 deletions(-)
>>>
>>> diff --git a/drivers/hwtracing/coresight/coresight-catu.c
>>> b/drivers/hwtracing/coresight/coresight-catu.c
>>> index bc90a03f478f..24a08a2b96b1 100644
>>> --- a/drivers/hwtracing/coresight/coresight-catu.c
>>> +++ b/drivers/hwtracing/coresight/coresight-catu.c
>>> @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct
>>> catu_drvdata *drvdata)
>>> return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY, 1);
>>> }
>>> -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
>>> +static struct coresight_device *
>>> +catu_get_etr_device(struct coresight_device *csdev)
>>> +{
>>> + int i;
>>> + struct coresight_device *tmp;
>>> +
>>> + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
>>> + tmp = csdev->pdata->in_conns[i].remote_dev;
>>> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
>>> + tmp->subtype.sink_subtype ==
>>> + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
>>> + return tmp;
>>> + }
>>> +
>>> + return NULL;
>>> +}
>>> +
>>> +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode
>>> cs_mode,
>>> + void *data)
>>> {
>>> int rc;
>>> u32 control, mode;
>>> - struct etr_buf *etr_buf = data;
>>> + struct etr_buf *etr_buf = NULL;
>>> struct device *dev = &drvdata->csdev->dev;
>>> struct coresight_device *csdev = drvdata->csdev;
>>> + struct coresight_device *etrdev;
>>> if (catu_wait_for_ready(drvdata))
>>> dev_warn(dev, "Timeout while waiting for READY\n");
>>> @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata
>>> *drvdata, void *data)
>>> if (rc)
>>> return rc;
>>> + etrdev = catu_get_etr_device(csdev);
>>> + if (etrdev) {
>>> + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
>>> + if (IS_ERR(etr_buf))
>>> + return PTR_ERR(etr_buf);
>>> + }
>>
>> WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
>> return.
>>
>
> I saw there was the pass-through mode below which I thought didn't need
> an ETR device. I think I followed the code through and there was a way
> for it to get there without an ETR in the existing version, but now I'm
> not sure.
> Or does it still need the ETR device but it just doesn't
> access the buffer?
The first part is correct. Without an ETR, CATU wouldn't be a helper
device, and wouldn't get here in "enable CATU" via the helper route.
The CATU chooses the mode depending on the etr_buf mode.
Suzuki
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
2023-03-29 13:23 ` Suzuki K Poulose
@ 2023-03-29 14:10 ` James Clark
-1 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 14:10 UTC (permalink / raw)
To: Suzuki K Poulose, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 29/03/2023 14:23, Suzuki K Poulose wrote:
> On 29/03/2023 13:04, James Clark wrote:
>>
>>
>> On 17/03/2023 11:04, Suzuki K Poulose wrote:
>>> On 10/03/2023 16:06, James Clark wrote:
>>>> Currently CATU is the only helper device, and its enable and disable
>>>> calls are hard coded. To allow more helper devices to be added in a
>>>> generic way, remove these hard coded calls and just enable and disable
>>>> all helper devices.
>>>>
>>>> This has to apply to helpers adjacent to the path, because they will
>>>> never be in the path. CATU was already discovered in this way, so
>>>> there is no change there.
>>>>
>>>> One change that is needed is for CATU to call back into ETR to allocate
>>>> the buffer. Because the enable call was previously hard coded, it was
>>>> done at a point where the buffer was already allocated, but this is no
>>>> longer the case.
>>>>
>>>> Signed-off-by: James Clark <james.clark@arm.com>
>>>> ---
>>>> drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
>>>> drivers/hwtracing/coresight/coresight-core.c | 68
>>>> ++++++++++++++++++-
>>>> .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
>>>> include/linux/coresight.h | 3 +-
>>>> 4 files changed, 99 insertions(+), 34 deletions(-)
>>>>
>>>> diff --git a/drivers/hwtracing/coresight/coresight-catu.c
>>>> b/drivers/hwtracing/coresight/coresight-catu.c
>>>> index bc90a03f478f..24a08a2b96b1 100644
>>>> --- a/drivers/hwtracing/coresight/coresight-catu.c
>>>> +++ b/drivers/hwtracing/coresight/coresight-catu.c
>>>> @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct
>>>> catu_drvdata *drvdata)
>>>> return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY,
>>>> 1);
>>>> }
>>>> -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
>>>> +static struct coresight_device *
>>>> +catu_get_etr_device(struct coresight_device *csdev)
>>>> +{
>>>> + int i;
>>>> + struct coresight_device *tmp;
>>>> +
>>>> + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
>>>> + tmp = csdev->pdata->in_conns[i].remote_dev;
>>>> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
>>>> + tmp->subtype.sink_subtype ==
>>>> + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
>>>> + return tmp;
>>>> + }
>>>> +
>>>> + return NULL;
>>>> +}
>>>> +
>>>> +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode
>>>> cs_mode,
>>>> + void *data)
>>>> {
>>>> int rc;
>>>> u32 control, mode;
>>>> - struct etr_buf *etr_buf = data;
>>>> + struct etr_buf *etr_buf = NULL;
>>>> struct device *dev = &drvdata->csdev->dev;
>>>> struct coresight_device *csdev = drvdata->csdev;
>>>> + struct coresight_device *etrdev;
>>>> if (catu_wait_for_ready(drvdata))
>>>> dev_warn(dev, "Timeout while waiting for READY\n");
>>>> @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata
>>>> *drvdata, void *data)
>>>> if (rc)
>>>> return rc;
>>>> + etrdev = catu_get_etr_device(csdev);
>>>> + if (etrdev) {
>>>> + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
>>>> + if (IS_ERR(etr_buf))
>>>> + return PTR_ERR(etr_buf);
>>>> + }
>>>
>>> WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
>>> return.
>>>
>>
>> I saw there was the pass-through mode below which I thought didn't need
>> an ETR device. I think I followed the code through and there was a way
>> for it to get there without an ETR in the existing version, but now I'm
>> not sure.
>
>
>> Or does it still need the ETR device but it just doesn't
>> access the buffer?
>
> The first part is correct. Without an ETR, CATU wouldn't be a helper
> device, and wouldn't get here in "enable CATU" via the helper route.
> The CATU chooses the mode depending on the etr_buf mode.
Ok thanks I will add that warning then
>
>
> Suzuki
>
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 74+ messages in thread
* Re: [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path
@ 2023-03-29 14:10 ` James Clark
0 siblings, 0 replies; 74+ messages in thread
From: James Clark @ 2023-03-29 14:10 UTC (permalink / raw)
To: Suzuki K Poulose, coresight
Cc: Mathieu Poirier, Mike Leach, Leo Yan, Alexander Shishkin,
linux-arm-kernel, linux-kernel
On 29/03/2023 14:23, Suzuki K Poulose wrote:
> On 29/03/2023 13:04, James Clark wrote:
>>
>>
>> On 17/03/2023 11:04, Suzuki K Poulose wrote:
>>> On 10/03/2023 16:06, James Clark wrote:
>>>> Currently CATU is the only helper device, and its enable and disable
>>>> calls are hard coded. To allow more helper devices to be added in a
>>>> generic way, remove these hard coded calls and just enable and disable
>>>> all helper devices.
>>>>
>>>> This has to apply to helpers adjacent to the path, because they will
>>>> never be in the path. CATU was already discovered in this way, so
>>>> there is no change there.
>>>>
>>>> One change that is needed is for CATU to call back into ETR to allocate
>>>> the buffer. Because the enable call was previously hard coded, it was
>>>> done at a point where the buffer was already allocated, but this is no
>>>> longer the case.
>>>>
>>>> Signed-off-by: James Clark <james.clark@arm.com>
>>>> ---
>>>> drivers/hwtracing/coresight/coresight-catu.c | 34 ++++++++--
>>>> drivers/hwtracing/coresight/coresight-core.c | 68
>>>> ++++++++++++++++++-
>>>> .../hwtracing/coresight/coresight-tmc-etr.c | 28 --------
>>>> include/linux/coresight.h | 3 +-
>>>> 4 files changed, 99 insertions(+), 34 deletions(-)
>>>>
>>>> diff --git a/drivers/hwtracing/coresight/coresight-catu.c
>>>> b/drivers/hwtracing/coresight/coresight-catu.c
>>>> index bc90a03f478f..24a08a2b96b1 100644
>>>> --- a/drivers/hwtracing/coresight/coresight-catu.c
>>>> +++ b/drivers/hwtracing/coresight/coresight-catu.c
>>>> @@ -395,13 +395,32 @@ static inline int catu_wait_for_ready(struct
>>>> catu_drvdata *drvdata)
>>>> return coresight_timeout(csa, CATU_STATUS, CATU_STATUS_READY,
>>>> 1);
>>>> }
>>>> -static int catu_enable_hw(struct catu_drvdata *drvdata, void *data)
>>>> +static struct coresight_device *
>>>> +catu_get_etr_device(struct coresight_device *csdev)
>>>> +{
>>>> + int i;
>>>> + struct coresight_device *tmp;
>>>> +
>>>> + for (i = 0; i < csdev->pdata->nr_inconns; i++) {
>>>> + tmp = csdev->pdata->in_conns[i].remote_dev;
>>>> + if (tmp && tmp->type == CORESIGHT_DEV_TYPE_SINK &&
>>>> + tmp->subtype.sink_subtype ==
>>>> + CORESIGHT_DEV_SUBTYPE_SINK_SYSMEM)
>>>> + return tmp;
>>>> + }
>>>> +
>>>> + return NULL;
>>>> +}
>>>> +
>>>> +static int catu_enable_hw(struct catu_drvdata *drvdata, enum cs_mode
>>>> cs_mode,
>>>> + void *data)
>>>> {
>>>> int rc;
>>>> u32 control, mode;
>>>> - struct etr_buf *etr_buf = data;
>>>> + struct etr_buf *etr_buf = NULL;
>>>> struct device *dev = &drvdata->csdev->dev;
>>>> struct coresight_device *csdev = drvdata->csdev;
>>>> + struct coresight_device *etrdev;
>>>> if (catu_wait_for_ready(drvdata))
>>>> dev_warn(dev, "Timeout while waiting for READY\n");
>>>> @@ -416,6 +435,12 @@ static int catu_enable_hw(struct catu_drvdata
>>>> *drvdata, void *data)
>>>> if (rc)
>>>> return rc;
>>>> + etrdev = catu_get_etr_device(csdev);
>>>> + if (etrdev) {
>>>> + etr_buf = tmc_etr_get_buffer(etrdev, cs_mode, data);
>>>> + if (IS_ERR(etr_buf))
>>>> + return PTR_ERR(etr_buf);
>>>> + }
>>>
>>> WARN_ON(!etrdev) ? We are not supposed to reach in the first place and
>>> return.
>>>
>>
>> I saw there was the pass-through mode below which I thought didn't need
>> an ETR device. I think I followed the code through and there was a way
>> for it to get there without an ETR in the existing version, but now I'm
>> not sure.
>
>
>> Or does it still need the ETR device but it just doesn't
>> access the buffer?
>
> The first part is correct. Without an ETR, CATU wouldn't be a helper
> device, and wouldn't get here in "enable CATU" via the helper route.
> The CATU chooses the mode depending on the etr_buf mode.
Ok thanks I will add that warning then
>
>
> Suzuki
>
^ permalink raw reply [flat|nested] 74+ messages in thread
end of thread, other threads:[~2023-03-29 14:25 UTC | newest]
Thread overview: 74+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-03-10 16:05 [PATCH v2 0/9] coresight: Fix CTI module refcount leak by making it a helper device James Clark
2023-03-10 16:05 ` James Clark
2023-03-10 16:06 ` [PATCH v2 1/9] coresight: Use enum type for cs_mode wherever possible James Clark
2023-03-10 16:06 ` James Clark
2023-03-16 16:47 ` Suzuki K Poulose
2023-03-16 16:47 ` Suzuki K Poulose
2023-03-21 15:12 ` Mike Leach
2023-03-21 15:12 ` Mike Leach
2023-03-10 16:06 ` [PATCH v2 2/9] coresight: Change name of pdata->conns James Clark
2023-03-10 16:06 ` James Clark
2023-03-16 17:17 ` Suzuki K Poulose
2023-03-16 17:17 ` Suzuki K Poulose
2023-03-21 15:12 ` Mike Leach
2023-03-21 15:12 ` Mike Leach
2023-03-10 16:06 ` [PATCH v2 3/9] coresight: Rename nr_outports to nr_outconns James Clark
2023-03-10 16:06 ` James Clark
2023-03-16 17:18 ` Suzuki K Poulose
2023-03-16 17:18 ` Suzuki K Poulose
2023-03-21 17:33 ` Mike Leach
2023-03-21 17:33 ` Mike Leach
2023-03-10 16:06 ` [PATCH v2 4/9] coresight: Rename connection members to allow for input connections James Clark
2023-03-10 16:06 ` James Clark
2023-03-21 17:33 ` Mike Leach
2023-03-21 17:33 ` Mike Leach
2023-03-10 16:06 ` [PATCH v2 5/9] coresight: Dynamically add connections James Clark
2023-03-10 16:06 ` James Clark
2023-03-16 17:12 ` Suzuki K Poulose
2023-03-16 17:12 ` Suzuki K Poulose
2023-03-21 17:56 ` Mike Leach
2023-03-21 17:56 ` Mike Leach
2023-03-23 10:49 ` James Clark
2023-03-23 10:49 ` James Clark
2023-03-29 12:02 ` James Clark
2023-03-29 12:02 ` James Clark
2023-03-10 16:06 ` [PATCH v2 6/9] coresight: Store in-connections as well as out-connections James Clark
2023-03-10 16:06 ` James Clark
2023-03-14 5:35 ` Jinlong Mao
2023-03-14 5:35 ` Jinlong Mao
2023-03-29 12:06 ` James Clark
2023-03-29 12:06 ` James Clark
2023-03-16 20:23 ` Suzuki K Poulose
2023-03-16 20:23 ` Suzuki K Poulose
2023-03-21 17:56 ` Mike Leach
2023-03-21 17:56 ` Mike Leach
2023-03-29 11:59 ` James Clark
2023-03-29 11:59 ` James Clark
2023-03-10 16:06 ` [PATCH v2 7/9] coresight: Refactor out buffer allocation function for ETR James Clark
2023-03-10 16:06 ` James Clark
2023-03-21 18:01 ` Mike Leach
2023-03-21 18:01 ` Mike Leach
2023-03-10 16:06 ` [PATCH v2 8/9] coresight: Enable and disable helper devices adjacent to the path James Clark
2023-03-10 16:06 ` James Clark
2023-03-17 11:04 ` Suzuki K Poulose
2023-03-17 11:04 ` Suzuki K Poulose
2023-03-22 9:29 ` Mike Leach
2023-03-22 9:29 ` Mike Leach
2023-03-29 12:04 ` James Clark
2023-03-29 12:04 ` James Clark
2023-03-29 13:23 ` Suzuki K Poulose
2023-03-29 13:23 ` Suzuki K Poulose
2023-03-29 14:10 ` James Clark
2023-03-29 14:10 ` James Clark
2023-03-10 16:06 ` [PATCH v2 9/9] coresight: Fix CTI module refcount leak by making it a helper device James Clark
2023-03-10 16:06 ` James Clark
2023-03-17 11:40 ` Suzuki K Poulose
2023-03-17 11:40 ` Suzuki K Poulose
2023-03-22 9:29 ` Mike Leach
2023-03-22 9:29 ` Mike Leach
2023-03-21 15:12 ` [PATCH v2 0/9] " Mike Leach
2023-03-21 15:12 ` Mike Leach
2023-03-22 9:26 ` James Clark
2023-03-22 9:26 ` James Clark
[not found] ` <CAJ9a7VgbehyFVF1PUZSSGVzhdSY083AvN0bC2xKH=JEUp8oVbw@mail.gmail.com>
2023-03-22 17:18 ` James Clark
2023-03-22 17:18 ` James Clark
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.