On 1/18/24 01:58, Konrad Dybcio wrote:

On 1/17/24 18:34, Sibi Sankar wrote:

quoted

From: Shivnandan Kumar <redacted>

This patch introduces a client driver that interacts with the SCMI QCOM
vendor protocol and passes on the required tuneables to start various
features running on the SCMI controller.

Signed-off-by: Shivnandan Kumar <redacted>
Co-developed-by: Ramakrishna Gottimukkula <redacted>
Signed-off-by: Ramakrishna Gottimukkula <redacted>
Co-developed-by: Amir Vajid <redacted>
Signed-off-by: Amir Vajid <redacted>
Co-developed-by: Sibi Sankar <redacted>
Signed-off-by: Sibi Sankar <redacted>
---

[...]

quoted

+
+struct cpufreq_memfreq_map {
+    unsigned int            cpufreq_mhz;
+    unsigned int            memfreq_khz;
+};

Weird use of tabs

will fix it in the next re-spin.

[...]

quoted

+static int get_mask(struct device_node *np, u32 *mask)
+{
+    struct device_node *dev_phandle;
+    struct device *cpu_dev;
+    int cpu, i = 0;
+    int ret = -ENODEV;

Don't initialize ret here, return 0 instead of breaking and return
enodev otherwise.

ack

quoted

+
+    dev_phandle = of_parse_phandle(np, "qcom,cpulist", i++);
+    while (dev_phandle) {
+        for_each_possible_cpu(cpu) {
+            cpu_dev = get_cpu_device(cpu);
+            if (cpu_dev && cpu_dev->of_node == dev_phandle) {
+                *mask |= BIT(cpu);
+                ret = 0;
+                break;
+            }
+        }

of_cpu_node_to_id()

ack

quoted

+        dev_phandle = of_parse_phandle(np, "qcom,cpulist", i++);
+    }
+
+    return ret;
+}

quoted

+
+static struct cpufreq_memfreq_map *init_cpufreq_memfreq_map(struct 
device *dev,
+                                struct device_node *of_node,
+                                u32 *cnt)

I really feel like this is trying to reinvent OPP..

if you structure your entries like so:

opp-0 {
     opp-hz = /bits/ 64 <12341234 43214321>;
};

you'll be able to use all the fantastic APIs that have been
created over the years!

I didn't know listing multiple frequencies in a opp was allowed. We can
probably get away with it here since we just parse the data here and not
populate data in the opp core.

[...]

quoted

+            monitor->mon_type = (of_property_read_bool(monitor_np, 
"qcom,compute-mon")) ? 1 : 0;
+            monitor->ipm_ceil = (of_property_read_bool(monitor_np, 
"qcom,compute-mon")) ? 0 : 20000000;

What does it even mean for a monitor to be a compute mon?

When a monitor is marked compute-mon it means that the table is
followed religiously irrespective whether the instruction per miss
count threshold (ipm) is exceeded or not. Equivalent to having
a cpufreq map -> l3/DDR bw mapping upstream.

There seem to be no dt-bindings for properties referenced in this
driver, neither in the series nor in the dependencies. This is
strictly required.

Ack

Thanks again for reviewing the series. :)

-Sibi

Konrad