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