Re: [PATCH V2] firmware: arm_scmi: bus: bypass set fwnode for scmi cpufreq

On Tue, Aug 13, 2024 at 11:51:24PM -0700, Saravana Kannan wrote:

On Tue, Aug 13, 2024 at 8:36 PM Peng Fan [off-list ref] wrote:

quoted

quoted

Subject: RE: [PATCH V2] firmware: arm_scmi: bus: bypass set fwnode
for scmi cpufreq

Hi,

quoted

quoted

quoted

Subject: Re: [PATCH V2] firmware: arm_scmi: bus: bypass set fwnode

for

quoted

scmi cpufreq

On Tue, Aug 13, 2024 at 01:52:30PM +0000, Peng Fan wrote:

quoted

quoted

Subject: Re: [PATCH V2] firmware: arm_scmi: bus: bypass set

fwnode

quoted

quoted

for scmi cpufreq

On Tue, Aug 13, 2024 at 10:25:31AM +0000, Peng Fan wrote:

quoted

quoted

Subject: Re: [PATCH V2] firmware: arm_scmi: bus: bypass set

fwnode

quoted

quoted

for scmi cpufreq

On Jul 29, 2024 at 15:03:25 +0800, Peng Fan (OSS) wrote:

quoted

From: Peng Fan <peng.fan@nxp.com>

Two drivers scmi_cpufreq.c and scmi_perf_domain.c both

use

quoted

quoted

quoted

quoted

quoted

quoted

SCMI_PROTCOL_PERF protocol, but with different name, so

two

quoted

quoted

scmi

quoted

quoted

quoted

devices will be created. But the fwnode->dev could only
point to one

device.

quoted

If scmi cpufreq device created earlier, the fwnode->dev will
point to the scmi cpufreq device. Then the fw_devlink will
link performance domain user device(consumer) to the scmi
cpufreq

device(supplier).

quoted

quoted

quoted

But actually the performance domain user device, such as
GPU,

should

quoted

use the scmi perf device as supplier. Also if 'cpufreq.off=1'
in bootargs, the GPU driver will defer probe always, because
of the scmi cpufreq

The commit message itself seems very specific to some

platform

quoted

quoted

quoted

quoted

quoted

to

me.

quoted

quoted

What about platforms that don't atall have a GPU? Why

would

quoted

quoted

quoted

they

quoted

quoted

care about this?

It is a generic issue if a platform has performance domain to
serve scmi cpufreq and device performance level.

quoted

quoted

device not ready.

Because for cpufreq, no need use fw_devlink. So bypass

setting

quoted

quoted

quoted

quoted

fwnode

quoted

for scmi cpufreq device.

Fixes: 96da4a99ce50 ("firmware: arm_scmi: Set fwnode for

the

quoted

quoted

quoted

quoted

quoted

scmi_device")
Signed-off-by: Peng Fan <peng.fan@nxp.com>
---

V2:
 Use A!=B to replace !(A == B)  Add fixes tag  This might be
a workaround, but since this is a fix, it is simple for
backporting.

More than a workaround, it feels like a HACK to me.

quoted

V1:



 drivers/firmware/arm_scmi/bus.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/firmware/arm_scmi/bus.c

b/drivers/firmware/arm_scmi/bus.c index

96b2e5f9a8ef..be91a82e0cda

quoted

100644

--- a/drivers/firmware/arm_scmi/bus.c
+++ b/drivers/firmware/arm_scmi/bus.c

@@ -395,7 +395,8 @@ __scmi_device_create(struct

device_node

quoted

quoted

*np, struct device *parent,

quoted

  scmi_dev->id = id;
  scmi_dev->protocol_id = protocol;
  scmi_dev->dev.parent = parent;
- device_set_node(&scmi_dev->dev,

of_fwnode_handle(np));

quoted

quoted

quoted

quoted

quoted

+ if ((protocol != SCMI_PROTOCOL_PERF) ||

strcmp(name,

quoted

quoted

quoted

quoted

"cpufreq"))

quoted

+         device_set_node(&scmi_dev->dev,

of_fwnode_handle(np));

I kind of disagree with the idea here to be specific about the
PROTOCOL_PERF or cpufreq. This is a generic arm scmi bus

driver

quoted

quoted

right?

quoted

quoted

Why bring in specific code into a bus driver? We will never
fix the actual root cause of the issue this way.

The root cause is fwnode devlink only supports one fwnode, one

device.

quoted

1:1 match. But current arm scmi driver use one fwnode for two

devices.

quoted

If your platform has scmi cpufreq and scmi device performance

domain,

quoted

you might see that some devices are consumer of scmi cpufreq,

but

quoted

quoted

quoted

actually they should be consumer of scmi device performance

domain.

quoted

I not have a good idea that this is fw devlink design that only
allows
1 fwnode has 1 device or not. If yes, that arm scmi should be

fixed.

quoted

quoted

quoted

quoted

If not, fw devlink should be updated.

The current patch is the simplest method for stable tree fixes
as I could work out.

So this is the same root cause at the end of the issues you had
with IMX pinctrl coexistence...i.e. the SCMI stack creates scmi
devices that embeds the protocol node, BUT since you can have
multiple device/drivers doing different things on different
resources within the same protocol you can end with 2 devices
having the same embedded device_node, since we dont really

have

quoted

quoted

quoted

anything else to

use

quoted

quoted

as device_node, I rigth ?

I think, yes. And you remind me that with PINCTRL_SCMI and
CONFIG_PINCTRL_IMX_SCMI both enabled, the scmi pinctrl node

will

quoted

only

quoted

take one to set the fwnode device pointer depends on the order to

run

quoted

__scmi_device_create.

So not only perf, pinctrl also has issue here, fwnode devlink will
not work properly for pinctrl/perf.

...mmm ... the standard generic Pinctrl driver and the IMX Pintrcl are
mutually exclusive in the code (@probe time) itself as far as I can
remember about what you did, so why devlink should have that issue
there ?
Have you seen any issue in this regards while having loaded
pinctrl_scmi and pinctrl_imx_scmi ?

No. it works well in my setup. I am just worried that there might be
issues because fwnode only has one dev pointer, see device_add.

quoted

I want to have a better look next days about this devlink issue that
you reported...it still not clear to me...from device_link_add() docs,
it seems indeed that it will return the old existing link if a link
exist already between that same supplier/consumer devices

pair....but

quoted

from the code (at first sight) it seems that the check is made agains
the devices not their embeded device_nodes, but here (and in
pinctrl/imx case) you will have 2 distinct consumer devices (with

same

quoted

device_node)...I may have missed something in your exaplanation....

It might be false alarm for pinctrl, but it is true issue for perf.

Just give a recheck. Pinctrl has same issue. With PINCTRL_SCMI
and PINCTRL_IMX_SCMI both enabled, two scmi devices
will be created. So it depends on device creation order, 1st
created device will be used as supplier.

On i.MX, there is no issue with both enabled, it is because
the imx scmi device is created first. If the generic pinctrl scmi
device created first, imx will have issue.

In the end, this is generic fw_devlink limitation or arm scmi driver
issue.

+1 to what Cristian and Dhruva said in other parts of this thread.

This patch itself is definitely a hack.

The problem isn't so much that fw_devlink doesn't want to support
multiple devices getting instantiated from one DT node. The problem is
that there's no way to know which of the multiple devices is the real
supplier just by looking at the information in devicetree/firmware
(the fw in fw_devlink). And keep in mind that one of the main
requirements of fw_devlink is to work before any driver is loaded and
not depend on drivers for correctness of the dependency information
because it needs to work on a fully modular kernel too. So, fw_devlink
just picks the first device that's instantiated from a DT node.

I really hate folks creating multiple devices from one DT node. One IP
block can support multiple things, there's no need to instantiate
multiple devices for it. The same driver could have just as easily
registered with multiple frameworks. So, ideally I'd want us to fix
this issue in the SCMI framework code. In the case where the same SCMI
node is creating two devices, can they both probe successfully? If
yes, why are we not using a child node or a separate node for this
second device? If it's always one or the other, why are we creating
two devices? Can you please point to specific upstream DT examples for
me to get a better handle on what's going on?

Btw, there is the deferred_probe_timeout command line option that can
be used so that fw_devlink stops enforcing dependencies where there
are no supplier drivers for a device after a timeout. It's not ideal,
but it's something to unblock you.

The best fw_devlink could do is just not enforce any dependencies if
there is more than one device instantiated for a given supplier DT
node.

So I'll try to describe in the following how the SCMI stack is
designed and behaves as it stands now....probably not exactly in a
concise way ... :P

The protocol-nodes in the DT are used to describe the presence and
sometime additional properties of an SCMI Protocol like Clock,
Peformance, Pinctrl etc, beside being used in phandles to refer to
specific per-protocol resources from other drivers.
(arch/arm64/boot/dts/arm/juno-scmi.dtsi to see how protocol nodes
 are referred )

Note also that the SCNI protocol is extensible and growing, we have 10
stanard protocols currently and potentially a number of additional vendor
protocols numbers that can be defined in the future.

The core SCMI stack knows how to build,send and track the proper SCMI
messages for all of these SCMI protocols and their well defined SCMI
commands, and exposes a common API (include/linux/scmi_protocol.h) in the
classic "file-operations" style for the driver-users that wants to use
such SCMI facilities.

As an example, clk-scmi is one of this SCMI drivers: it plugs on one side
on the standard CLK frameowrk and then registers also as an SCMI driver
with the SCMI stack, so that it can access and use the related specific
scmi_clk_ops to map common operations like clocks enable/disable into an
SCMI message exchange.

When the SCMI core stack comes up it queries the SCMI/fw upfront via the
dedicated Base protocol to discover, amongst other stuff, the list of
supported protocols by the current platorm.

An SCMI driver like clk-scmi needs some sort of device to drive: the
driver itself declares the protocol(CLK) that he wants to use and a name,
and the SCMI bus will create a device accordingly, using the related
protocol device_node and the name choosen (so that multiple devices
can exist with different names for different drivers but using the same
protocol)

Then, the clk-scmi driver at probe time, after a successufl match on
protocol/name, will obtain the related clk_scmi ops from the core
stack together with an opaque handle for that instance of the SCMI
stack: this first SCMI/clk user will trigger the specific core protocol
initialization code to be run.

The clk-scmi driver can then issue any CLK related ops he wants
during its lifetime.

Protocols are initialized on first use: when a SCMI driver appears that
wants to use protocol-X ops, protocol-X will be initialized...any
subsequent driver appearing that wants to use protocol-X will use the
same initialized protocol stack.

...BUT you can have possibly multiple drivers willing to use the same
protocol, probably on a disjoint set of resources in that protocol:
an example of this is the Sensor protocol which is used by scmi-hwmon
for a few classic sensors, BUT also by scmi-iio for another class of
distinct sensors (and plugging into a different subsystem): such drivers
will both use Sensor protocol and both will have a device created an
"hwmon"/sensor vs "iio"/sensor...

...the problem is, as it start to become apparent, I have ONLY one single
DT node to represent the protocol which I can use to embed in the
created devices...

In this scenario I would like to have multiple distinct DT nodes describing
the same protocol so that I can instantiate different devices using the same
protocol but using a distinct device_node...just I am not sure how to create
such a duplicate node out of thin-air in a non-dirty way...moreover I am not
even sure this is doable, since I fear that such a solution will even more
break the stack, since how the phandle references across the DT are even
supposed to work if I magically create a duplicate of a device_node
urelated to any real DT...

...sorry for my ignorance in this area, any advice will be very much welcome.

Thanks,
Cristian