2019-08-09 20:24:52

by Yabin Cui

Subject: [PATCH] coresight: tmc-etr: Fix perf_data check.

When tracing etm data of multiple threads on multiple cpus through
perf interface, each cpu has a unique etr_perf_buffer while sharing
the same etr device. There is no guarantee that the last cpu starts
etm tracing also stops last. This makes perf_data check fail.

Fix it by checking etr_buf instead of etr_perf_buffer.

Signed-off-by: Yabin Cui <[email protected]>
---
drivers/hwtracing/coresight/coresight-tmc-etr.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
index 17006705287a..f466f05afe08 100644
--- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
+++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
@@ -1484,7 +1484,7 @@ tmc_update_etr_buffer(struct coresight_device *csdev,
goto out;
}

- if (WARN_ON(drvdata->perf_data != etr_perf)) {
+ if (WARN_ON(drvdata->perf_data != etr_buf)) {
lost = true;
spin_unlock_irqrestore(&drvdata->spinlock, flags);
goto out;
@@ -1556,7 +1556,7 @@ static int tmc_enable_etr_sink_perf(struct coresight_device *csdev, void *data)
}

etr_perf->head = PERF_IDX2OFF(handle->head, etr_perf);
- drvdata->perf_data = etr_perf;
+ drvdata->perf_data = etr_perf->etr_buf;

/*
* No HW configuration is needed if the sink is already in
--
2.23.0.rc1.153.gdeed80330f-goog

2019-08-12 10:46:38

by Suzuki K Poulose

Subject: Re: [PATCH] coresight: tmc-etr: Fix perf_data check.

On 09/08/2019 21:23, Yabin Cui wrote:
> When tracing etm data of multiple threads on multiple cpus through
> perf interface, each cpu has a unique etr_perf_buffer while sharing
> the same etr device. There is no guarantee that the last cpu starts
> etm tracing also stops last. This makes perf_data check fail.
>
> Fix it by checking etr_buf instead of etr_perf_buffer.

Please could you add a Fixes tag for this:

Fixes: 3147da92a8a81fc3 ("coresight: tmc-etr: Allocate and free ETR memory
buffers for CPU-wide scenarios")

as the problem was introduced as a side effect of the above patch ?

>
> Signed-off-by: Yabin Cui <[email protected]>
> ---
> drivers/hwtracing/coresight/coresight-tmc-etr.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index 17006705287a..f466f05afe08 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> @@ -1484,7 +1484,7 @@ tmc_update_etr_buffer(struct coresight_device *csdev,
> goto out;
> }
>
> - if (WARN_ON(drvdata->perf_data != etr_perf)) {
> + if (WARN_ON(drvdata->perf_data != etr_buf)) {
> lost = true;
> spin_unlock_irqrestore(&drvdata->spinlock, flags);
> goto out;
> @@ -1556,7 +1556,7 @@ static int tmc_enable_etr_sink_perf(struct coresight_device *csdev, void *data)
> }
>
> etr_perf->head = PERF_IDX2OFF(handle->head, etr_perf);
> - drvdata->perf_data = etr_perf;
> + drvdata->perf_data = etr_perf->etr_buf;

minor nit: Now that we are storing the etr_buf instead of the etr_perf_buf in
perf_data, we could make the "perf_data" => "perf_buf" inline with the
sysfs_buf.

Either ways:

Reviewed-by: Suzuki K Poulose <[email protected]>