From: Cristian Marussi <cristian.marussi@arm.com>
To: Sudeep Holla <sudeep.holla@arm.com>
Cc: linux-kernel@vger.kernel.org,
linux-arm-kernel@lists.infradead.org, lukasz.luba@arm.com,
james.quinlan@broadcom.com, Jonathan.Cameron@Huawei.com,
f.fainelli@gmail.com, etienne.carriere@linaro.org,
thara.gopinath@linaro.org, vincent.guittot@linaro.org,
souvik.chakravarty@arm.com
Subject: Re: [PATCH v6 37/37] firmware: arm_scmi: add dynamic scmi devices creation
Date: Mon, 15 Mar 2021 09:03:43 +0000 [thread overview]
Message-ID: <20210315090343.GH30179@e120937-lin> (raw)
In-Reply-To: <20210315083327.p6gculrdl3vlavhp@bogus>
Hi
On Mon, Mar 15, 2021 at 08:33:27AM +0000, Sudeep Holla wrote:
> Hi Cristian,
>
> Sorry for the delay.
>
No worries.
> On Tue, Feb 02, 2021 at 10:15:55PM +0000, Cristian Marussi wrote:
> > Having added the support for SCMI protocols as modules in order to let
> > vendors extend the SCMI core with their own additions it seems odd to
> > then force SCMI drivers built on top to use a static device table to
> > declare their devices since this way any new SCMI drivers addition
> > would need the core SCMI device table to be updated too.
> >
> > Remove the static core device table and let SCMI drivers to simply declare
> > which device/protocol pair they need at initialization time: the core will
> > then take care to generate such devices dynamically during platform
> > initialization or at module loading time, as long as the requested
> > underlying protocol is defined in the DT.
> >
> > Signed-off-by: Cristian Marussi <cristian.marussi@arm.com>
> > ---
> > v4 --> v5
> > - using klist instead of custom lists
> > v3 --> v4
> > - add a few comments
> > ---
> > drivers/firmware/arm_scmi/bus.c | 30 +++
> > drivers/firmware/arm_scmi/common.h | 5 +
> > drivers/firmware/arm_scmi/driver.c | 309 +++++++++++++++++++++++++----
> > 3 files changed, 310 insertions(+), 34 deletions(-)
> >
> > diff --git a/drivers/firmware/arm_scmi/bus.c b/drivers/firmware/arm_scmi/bus.c
> > index 88e5057f4e85..88149a46e6d9 100644
> > --- a/drivers/firmware/arm_scmi/bus.c
> > +++ b/drivers/firmware/arm_scmi/bus.c
> > @@ -51,6 +51,31 @@ static int scmi_dev_match(struct device *dev, struct device_driver *drv)
> > return 0;
> > }
> >
> > +static int scmi_match_by_id_table(struct device *dev, void *data)
> > +{
> > + struct scmi_device *sdev = to_scmi_dev(dev);
> > + struct scmi_device_id *id_table = data;
> > +
> > + return sdev->protocol_id == id_table->protocol_id &&
> > + !strcmp(sdev->name, id_table->name);
> > +}
> > +
> > +struct scmi_device *scmi_find_child_dev(struct device *parent,
> > + int prot_id, const char *name)
> > +{
> > + struct scmi_device_id id_table;
> > + struct device *dev;
> > +
> > + id_table.protocol_id = prot_id;
> > + id_table.name = name;
> > +
> > + dev = device_find_child(parent, &id_table, scmi_match_by_id_table);
> > + if (!dev)
> > + return NULL;
> > +
> > + return to_scmi_dev(dev);
> > +}
> > +
> > const struct scmi_protocol *scmi_get_protocol(int protocol_id)
> > {
> > const struct scmi_protocol *proto;
> > @@ -114,6 +139,10 @@ int scmi_driver_register(struct scmi_driver *driver, struct module *owner,
> > {
> > int retval;
> >
> > + retval = scmi_request_protocol_device(driver->id_table);
> > + if (retval)
> > + return retval;
> > +
> > driver->driver.bus = &scmi_bus_type;
> > driver->driver.name = driver->name;
> > driver->driver.owner = owner;
> > @@ -130,6 +159,7 @@ EXPORT_SYMBOL_GPL(scmi_driver_register);
> > void scmi_driver_unregister(struct scmi_driver *driver)
> > {
> > driver_unregister(&driver->driver);
> > + scmi_unrequest_protocol_device(driver->id_table);
> > }
> > EXPORT_SYMBOL_GPL(scmi_driver_unregister);
> >
> > diff --git a/drivers/firmware/arm_scmi/common.h b/drivers/firmware/arm_scmi/common.h
> > index 1e2046c61d43..9a0519db4865 100644
> > --- a/drivers/firmware/arm_scmi/common.h
> > +++ b/drivers/firmware/arm_scmi/common.h
> > @@ -307,6 +307,11 @@ struct scmi_transport_ops {
> > bool (*poll_done)(struct scmi_chan_info *cinfo, struct scmi_xfer *xfer);
> > };
> >
> > +int scmi_request_protocol_device(const struct scmi_device_id *id_table);
> > +void scmi_unrequest_protocol_device(const struct scmi_device_id *id_table);
>
> Sorry for being pedantic, I don't like these names. I would prefer
> something like scmi_protocol_device_{create,destroy/delete}_request.
> The action the function does needs to be at the end of the function name.
> Atleast that is something I follow. I haven't checked all the previous
> patches, just this function made to look at both the name style and the
> name itself.
>
Ok.
>
> > +struct scmi_device *scmi_find_child_dev(struct device *parent,
> > + int prot_id, const char *name);
> > +
>
> scmi_child_dev_find based on what I mentioned above. Please change all
> other non-static functions even if I have not mentioned. Try to cover
> all the new functions introduced in this series, existing ones we can
> take up later.
>
Ok I'll do.
> > /**
> > * struct scmi_desc - Description of SoC integration
> > *
> > diff --git a/drivers/firmware/arm_scmi/driver.c b/drivers/firmware/arm_scmi/driver.c
> > index dcdfd94b47f7..9fc979e3b16f 100644
> > --- a/drivers/firmware/arm_scmi/driver.c
> > +++ b/drivers/firmware/arm_scmi/driver.c
> > @@ -21,6 +21,7 @@
> > #include <linux/io.h>
> > #include <linux/kernel.h>
> > #include <linux/ktime.h>
> > +#include <linux/list.h>
> > #include <linux/module.h>
> > #include <linux/of_address.h>
> > #include <linux/of_device.h>
> > @@ -56,6 +57,14 @@ static DEFINE_MUTEX(scmi_list_mutex);
> > /* Track the unique id for the transfers for debug & profiling purpose */
> > static atomic_t transfer_last_id;
> >
> > +static DEFINE_IDR(scmi_requested_devices);
> > +static DEFINE_MUTEX(scmi_requested_devices_mtx);
> > +
> > +struct scmi_requested_dev {
> > + const struct scmi_device_id *id_table;
> > + struct list_head node;
> > +};
> > +
> > /**
> > * struct scmi_xfers_info - Structure to manage transfer information
> > *
> > @@ -113,6 +122,8 @@ struct scmi_protocol_instance {
> > * @protocols_mtx: A mutex to protect protocols instances initialization.
> > * @protocols_imp: List of protocols implemented, currently maximum of
> > * MAX_PROTOCOLS_IMP elements allocated by the base protocol
> > + * @active_protocols: IDR storing device_nodes for protocols actually defined
> > + * in the DT and confirmed as implemented by fw.
> > * @notify_priv: Pointer to private data structure specific to notifications.
> > * @node: List head
> > * @users: Number of users of this instance
> > @@ -130,6 +141,7 @@ struct scmi_info {
> > /* Ensure mutual exclusive access to protocols instance array */
> > struct mutex protocols_mtx;
> > u8 *protocols_imp;
> > + struct idr active_protocols;
> > void *notify_priv;
> > struct list_head node;
> > int users;
> > @@ -936,6 +948,13 @@ static void scmi_devm_put_protocol(struct scmi_device *sdev, u8 protocol_id)
> > WARN_ON(ret);
> > }
> >
> > +static inline
> > +struct scmi_handle *scmi_handle_get_from_info(struct scmi_info *info)
> > +{
> > + info->users++;
>
> Doesn't it race with anything ? I have already forgotten how this is used
> and in what context this gets called.
>
> > + return &info->handle;
> > +}
> > +
> > /**
> > * scmi_handle_get() - Get the SCMI handle for a device
> > *
> > @@ -957,8 +976,7 @@ struct scmi_handle *scmi_handle_get(struct device *dev)
> > list_for_each(p, &scmi_list) {
> > info = list_entry(p, struct scmi_info, node);
> > if (dev->parent == info->dev) {
> > - handle = &info->handle;
> > - info->users++;
> > + handle = scmi_handle_get_from_info(info);
>
> Ah here it is. Any particular reasons for moving it to separate function ?
>
Answering both of the above, users++ is protected by scmi_list_mutex as it
was already and the reason for moving the get in a separate function is to
able to call it with mutex alrady acquired, so both here from scmi_handle_get(),
which acquires itself that mutex, BUT also from scmi_request_protocol_device()
which acquires already the mutex while scanning the list of all the existing
scmi instances; it is a sort of get_handle_unlocked but since I changed
also the param from dev to info (since I have it already) I named it get_from_info
instead. I could add a get_from_info_unlocked(0 to clarify the intent.
I'll rework all of the above and post the whole series on top of for-next-scmi +
Jonathan immutable branch locally merged.
Thanks
Cristian
> --
> Regards,
> Sudeep
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
prev parent reply other threads:[~2021-03-15 9:07 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-02 22:15 [PATCH v6 0/37] SCMI vendor protocols and modularization Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 01/37] firmware: arm_scmi: review protocol registration interface Cristian Marussi
2021-03-08 4:38 ` Sudeep Holla
2021-03-08 11:24 ` Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 02/37] firmware: arm_scmi: introduce protocol handle definitions Cristian Marussi
2021-03-08 5:50 ` Sudeep Holla
2021-03-08 11:37 ` Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 03/37] firmware: arm_scmi: introduce devres get/put protocols operations Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 04/37] firmware: arm_scmi: add devm_acquire_protocol helper Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 05/37] firmware: arm_scmi: make notifications aware of protocols users Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 06/37] firmware: arm_scmi: introduce new devres notification ops Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 07/37] firmware: arm_scmi: refactor events registration Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 08/37] firmware: arm_scmi: convert events registration to protocol handles Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 09/37] firmware: arm_scmi: add new protocol handle core xfer ops Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 10/37] firmware: arm_scmi: add helper to access revision area memory Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 11/37] firmware: arm_scmi: port Base protocol to new interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 12/37] firmware: arm_scmi: port Perf protocol to new protocols interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 13/37] cpufreq: scmi: port driver to the new scmi_perf_proto_ops interface Cristian Marussi
2021-02-03 3:03 ` Viresh Kumar
2021-02-03 12:10 ` Cristian Marussi
2021-02-04 11:43 ` Sudeep Holla
2021-02-04 2:14 ` Viresh Kumar
2021-02-02 22:15 ` [PATCH v6 14/37] firmware: arm_scmi: remove legacy scmi_perf_ops protocol interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 15/37] firmware: arm_scmi: port Power protocol to new protocols interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 16/37] firmware: arm_scmi: port GenPD driver to the new scmi_power_proto_ops interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 17/37] firmware: arm_scmi: remove legacy scmi_power_ops protocol interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 18/37] firmware: arm_scmi: port Clock protocol to new protocols interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 19/37] clk: scmi: port driver to the new scmi_clk_proto_ops interface Cristian Marussi
2021-03-10 14:19 ` Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 20/37] firmware: arm_scmi: remove legacy scmi_clk_ops protocol interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 21/37] firmware: arm_scmi: port Reset protocol to new protocols interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 22/37] reset: reset-scmi: port driver to the new scmi_reset_proto_ops interface Cristian Marussi
2021-03-03 14:46 ` Philipp Zabel
2021-02-02 22:15 ` [PATCH v6 23/37] firmware: arm_scmi: remove legacy scmi_reset_ops protocol interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 24/37] firmware: arm_scmi: port Sensor protocol to new protocols interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 25/37] hwmon: (scmi) port driver to the new scmi_sensor_proto_ops interface Cristian Marussi
2021-02-03 1:30 ` Guenter Roeck
2021-02-02 22:15 ` [PATCH v6 26/37] firmware: arm_scmi: remove legacy scmi_sensor_ops protocol interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 27/37] firmware: arm_scmi: port SystemPower protocol to new protocols interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 28/37] firmware: arm_scmi: port Voltage " Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 29/37] regulator: scmi: port driver to the new scmi_voltage_proto_ops interface Cristian Marussi
2021-02-03 13:24 ` Mark Brown
2021-02-03 14:39 ` Cristian Marussi
2021-02-03 16:18 ` Mark Brown
2021-02-02 22:15 ` [PATCH v6 30/37] firmware: arm_scmi: remove legacy scmi_voltage_ops protocol interface Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 31/37] firmware: arm_scmi: make references to handle const Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 32/37] firmware: arm_scmi: cleanup legacy protocol init code Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 33/37] firmware: arm_scmi: cleanup unused core xfer wrappers Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 34/37] firmware: arm_scmi: cleanup events registration transient code Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 35/37] firmware: arm_scmi: make notify_priv really private Cristian Marussi
2021-03-08 6:12 ` Sudeep Holla
2021-03-08 11:39 ` Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 36/37] firmware: arm_scmi: add protocol modularization support Cristian Marussi
2021-03-08 7:34 ` Sudeep Holla
2021-03-09 8:04 ` Cristian Marussi
2021-02-02 22:15 ` [PATCH v6 37/37] firmware: arm_scmi: add dynamic scmi devices creation Cristian Marussi
2021-03-15 8:33 ` Sudeep Holla
2021-03-15 9:03 ` Cristian Marussi [this message]
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=20210315090343.GH30179@e120937-lin \
--to=cristian.marussi@arm.com \
--cc=Jonathan.Cameron@Huawei.com \
--cc=etienne.carriere@linaro.org \
--cc=f.fainelli@gmail.com \
--cc=james.quinlan@broadcom.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=lukasz.luba@arm.com \
--cc=souvik.chakravarty@arm.com \
--cc=sudeep.holla@arm.com \
--cc=thara.gopinath@linaro.org \
--cc=vincent.guittot@linaro.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).