[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CANLsYkxRVvWUxEAmRQ7nCuS-NaOogN4sYOipxBW5zsozyu+y2g@mail.gmail.com>
Date: Thu, 15 Aug 2019 11:37:46 -0600
From: Mathieu Poirier <mathieu.poirier@...aro.org>
To: Yabin Cui <yabinc@...gle.com>
Cc: Suzuki K Poulose <suzuki.poulose@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] coresight: tmc-etr: Fix perf_data check.
On Mon, 12 Aug 2019 at 13:03, Yabin Cui <yabinc@...gle.com> 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.
>
> Fixes: 3147da92a8a8 ("coresight: tmc-etr: Allocate and free ETR memory buffers for CPU-wide scenarios")
> Signed-off-by: Yabin Cui <yabinc@...gle.com>
> ---
>
> v1 -> v2: rename perf_data to perf_buf. Add fixes tag.
>
> ---
> drivers/hwtracing/coresight/coresight-tmc-etr.c | 6 +++---
> drivers/hwtracing/coresight/coresight-tmc.h | 6 +++---
> 2 files changed, 6 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c b/drivers/hwtracing/coresight/coresight-tmc-etr.c
> index 17006705287a..90d1548ad268 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_buf != etr_buf)) {
> lost = true;
> spin_unlock_irqrestore(&drvdata->spinlock, flags);
> goto out;
> @@ -1497,7 +1497,7 @@ tmc_update_etr_buffer(struct coresight_device *csdev,
>
> CS_LOCK(drvdata->base);
> /* Reset perf specific data */
> - drvdata->perf_data = NULL;
> + drvdata->perf_buf = NULL;
> spin_unlock_irqrestore(&drvdata->spinlock, flags);
>
> size = etr_buf->len;
> @@ -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_buf = etr_perf->etr_buf;
Ok for the fix. Looking a things again I don't see a need to do the
assignment for each event - this needs to be done only when the device
is assocated with a monitored process. Please move it here [1].
Thanks,
Mathieu
[1]. https://elixir.bootlin.com/linux/v5.3-rc4/source/drivers/hwtracing/coresight/coresight-tmc-etr.c#L1572
>
> /*
> * No HW configuration is needed if the sink is already in
> diff --git a/drivers/hwtracing/coresight/coresight-tmc.h b/drivers/hwtracing/coresight/coresight-tmc.h
> index 1ed50411cc3c..f9a0c95e9ba2 100644
> --- a/drivers/hwtracing/coresight/coresight-tmc.h
> +++ b/drivers/hwtracing/coresight/coresight-tmc.h
> @@ -178,8 +178,8 @@ struct etr_buf {
> * device configuration register (DEVID)
> * @idr: Holds etr_bufs allocated for this ETR.
> * @idr_mutex: Access serialisation for idr.
> - * @perf_data: PERF buffer for ETR.
> - * @sysfs_data: SYSFS buffer for ETR.
> + * @sysfs_buf: SYSFS buffer for ETR.
> + * @perf_buf: PERF buffer for ETR.
> */
> struct tmc_drvdata {
> void __iomem *base;
> @@ -202,7 +202,7 @@ struct tmc_drvdata {
> struct idr idr;
> struct mutex idr_mutex;
> struct etr_buf *sysfs_buf;
> - void *perf_data;
> + struct etr_buf *perf_buf;
> };
>
> struct etr_buf_operations {
> --
> 2.23.0.rc1.153.gdeed80330f-goog
>
Powered by blists - more mailing lists