From: Suzuki K Poulose <suzuki.poulose@arm.com>
To: Tao Zhang <taozha@codeaurora.org>,
Mathieu Poirier <mathieu.poirier@linaro.org>,
Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Mike Leach <mike.leach@linaro.org>, Leo Yan <leo.yan@linaro.org>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
coresight@lists.linaro.org, linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org,
Tingwei Zhang <tingwei@codeaurora.org>,
Mao Jinlong <jinlmao@codeaurora.org>,
Yuanfang Zhang <zhangyuanfang@codeaurora.org>
Subject: Re: [PATCH v1] coresight: add node to reset all coresight devices
Date: Fri, 23 Apr 2021 14:20:18 +0100 [thread overview]
Message-ID: <e753898c-0334-cf84-620e-d9b9f18d87e7@arm.com> (raw)
In-Reply-To: <1619166578-28690-1-git-send-email-taozha@codeaurora.org>
On 23/04/2021 09:29, Tao Zhang wrote:
> Add new reset_source_sink node to be able to disable all active
> coresight devices.
> In this way, we no longer need to manually disable all active
> coresight devices one by one. After enabling multiple coresight
> paths, users can reset their status more conveniently by this
> node.
>
What is the use case here ? Why would you trigger a reset for all the
sources/sink without gracefully completing any on-going sessions
(including the perf ones, which are driven by the kernel perf layer)
> This patch base on coresight-next repo
> http://git.linaro.org/kernel/coresight.git/log/?h=next
>
> And this patch depends on the following patch
> https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg2551216.html
Please post related patches as a series, possibly describing the overall
problem that you are trying to solve, in the cover letter.
>
> Signed-off-by: Tingwei Zhang <tingwei@codeaurora.org>
> Signed-off-by: Mao Jinlong <jinlmao@codeaurora.org>
> Signed-off-by: Tao Zhang <taozha@codeaurora.org>
> ---
> drivers/hwtracing/coresight/coresight-core.c | 72 ++++++++++++++++++++++++----
> 1 file changed, 64 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-core.c b/drivers/hwtracing/coresight/coresight-core.c
> index 7dfadb6..0001b6c 100644
> --- a/drivers/hwtracing/coresight/coresight-core.c
> +++ b/drivers/hwtracing/coresight/coresight-core.c
> @@ -107,6 +107,23 @@ static int coresight_source_is_unique(struct coresight_device *csdev)
> csdev, coresight_id_match);
> }
>
> +static int coresight_reset_sink(struct device *dev, void *data)
> +{
> + struct coresight_device *csdev = to_coresight_device(dev);
> +
> + if ((csdev->type == CORESIGHT_DEV_TYPE_SINK ||
> + csdev->type == CORESIGHT_DEV_TYPE_LINKSINK) &&
> + csdev->activated)
> + csdev->activated = false;
Why is this needed, when the disabl_path() should have taken care of this ?
> +
> + return 0;
> +}
> +
> +static void coresight_reset_all_sink(void)
> +{
> + bus_for_each_dev(&coresight_bustype, NULL, NULL, coresight_reset_sink);
> +}
> +
How can you disable all the active sinks when some of the sinks could be
driven by perf ?
> static int coresight_find_link_inport(struct coresight_device *csdev,
> struct coresight_device *parent)
> {
> @@ -1137,7 +1154,7 @@ int coresight_enable(struct coresight_device *csdev)
> }
> EXPORT_SYMBOL_GPL(coresight_enable);
>
> -void coresight_disable(struct coresight_device *csdev)
> +static void __coresight_disable(struct coresight_device *csdev)
> {
> int ret;
> struct list_head *path = NULL;
> @@ -1145,14 +1162,12 @@ void coresight_disable(struct coresight_device *csdev)
> struct coresight_path *cspath_next = NULL;
> struct coresight_device *src_csdev = NULL;
>
> - mutex_lock(&coresight_mutex);
> -
> ret = coresight_validate_source(csdev, __func__);
> if (ret)
> - goto out;
> + return;
>
> if (!csdev->enable || !coresight_disable_source(csdev))
> - goto out;
> + return;
>
> list_for_each_entry_safe(cspath, cspath_next, &cs_active_paths, link) {
> src_csdev = coresight_get_source(cspath->path);
> @@ -1165,12 +1180,16 @@ void coresight_disable(struct coresight_device *csdev)
> }
> }
> if (path == NULL)
> - goto out;
> + return;
>
> coresight_disable_path(path);
> coresight_release_path(path);
> +}
>
> -out:
> +void coresight_disable(struct coresight_device *csdev)
> +{
> + mutex_lock(&coresight_mutex);
> + __coresight_disable(csdev);
> mutex_unlock(&coresight_mutex);
> }
> EXPORT_SYMBOL_GPL(coresight_disable);
> @@ -1467,7 +1486,43 @@ int coresight_timeout(void __iomem *addr, u32 offset, int position, int value)
>
> return -EAGAIN;
> }
> -EXPORT_SYMBOL_GPL(coresight_timeout);
Why ?
> +
> +static ssize_t reset_source_sink_store(struct bus_type *bus,
> + const char *buf, size_t size)
> +{
> + int ret = 0;
> + unsigned long val;
> + struct coresight_path *cspath = NULL;
> + struct coresight_path *cspath_next = NULL;
> + struct coresight_device *csdev;
> +
> + ret = kstrtoul(buf, 10, &val);
> + if (ret)
> + return ret;
> +
> + mutex_lock(&coresight_mutex);
> +
> + list_for_each_entry_safe(cspath, cspath_next, &cs_active_paths, link) {
> + csdev = coresight_get_source(cspath->path);
> + if (!csdev)
> + continue;
> + atomic_set(csdev->refcnt, 1);
Is this safe ?
> + __coresight_disable(csdev);
> + }
> +
> + /* Reset all activated sinks */
> + coresight_reset_all_sink();
> +
> + mutex_unlock(&coresight_mutex);
> + return size;
> +}
> +static BUS_ATTR_WO(reset_source_sink);
> +
> +static struct attribute *coresight_reset_source_sink_attrs[] = {
> + &bus_attr_reset_source_sink.attr,
> + NULL,
> +};
> +ATTRIBUTE_GROUPS(coresight_reset_source_sink);
>
> /*
> * coresight_release_platform_data: Release references to the devices connected
> @@ -1680,6 +1735,7 @@ EXPORT_SYMBOL_GPL(coresight_alloc_device_name);
>
> struct bus_type coresight_bustype = {
> .name = "coresight",
> + .bus_groups = coresight_reset_source_sink_groups,
> };
>
> static int __init coresight_init(void)
>
next prev parent reply other threads:[~2021-04-23 13:20 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-23 8:29 [PATCH v1] coresight: add node to reset all coresight devices Tao Zhang
2021-04-23 10:02 ` Greg Kroah-Hartman
2021-04-28 9:58 ` taozha
2021-04-23 13:20 ` Suzuki K Poulose [this message]
2021-04-28 9:57 ` taozha
2021-04-28 10:09 ` Suzuki K Poulose
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e753898c-0334-cf84-620e-d9b9f18d87e7@arm.com \
--to=suzuki.poulose@arm.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=coresight@lists.linaro.org \
--cc=gregkh@linuxfoundation.org \
--cc=jinlmao@codeaurora.org \
--cc=leo.yan@linaro.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mathieu.poirier@linaro.org \
--cc=mike.leach@linaro.org \
--cc=taozha@codeaurora.org \
--cc=tingwei@codeaurora.org \
--cc=zhangyuanfang@codeaurora.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).