On 3/7/2023 8:47 PM, Krzysztof Kozlowski wrote:
On 07/03/2023 05:41, Manikanta Mylavarapu wrote:
Add new binding document for multipd model remoteproc.
IPQ5018, IPQ9574 follows multipd model.
Signed-off-by: Manikanta Mylavarapu <quic_mmanikan@xxxxxxxxxxx>
---
.../bindings/remoteproc/qcom,multipd-pil.yaml | 282 ++++++++++++++++++
1 file changed, 282 insertions(+)
create mode 100644 Documentation/devicetree/bindings/remoteproc/qcom,multipd-pil.yaml
diff --git a/Documentation/devicetree/bindings/remoteproc/qcom,multipd-pil.yaml b/Documentation/devicetree/bindings/remoteproc/qcom,multipd-pil.yaml
new file mode 100644
index 000000000000..b788607f5abd
--- /dev/null
+++ b/Documentation/devicetree/bindings/remoteproc/qcom,multipd-pil.yaml
@@ -0,0 +1,282 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/remoteproc/qcom,multipd-pil.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Qualcomm Multipd Secure Peripheral Image Loader
+
+maintainers:
+ - Bjorn Andersson <andersson@xxxxxxxxxx>
+ - Mathieu Poirier <mathieu.poirier@xxxxxxxxxx>
+
+description:
+ Multipd Peripheral Image Loader loads firmware and boots Q6 pd, WCSS pd
+ remoteproc's on the Qualcomm IPQ5018, IPQ9574 SoC.
What is a "pd"?
Pd means protection domain.
It's similar to process in Linux. Here QDSP6 processor runs each wifi
radio functionality on a separate process. One process can't access
other process resources, so this is termed as PD i.e protection domain.
Here we have two pd's called root and user pd. We can correlate Root pd
as root and user pd as user in linux. Root pd has more privileges than
user pd.
From remoteproc driver perspective, root pd corresponds to QDSP6
processor bring up and user pd corresponds to Wifi radio (WCSS) bring up.
+
+properties:
+ compatible:
+ enum:
+ - qcom,ipq5018-q6-mpd
+ - qcom,ipq9574-q6-mpd
+
+ '#address-cells': true
+
+ '#size-cells': true
Why do you need both?
If really needed, these should be const. >
It's not required. I am going to remove it.
+
+ 'ranges': true
+
Same question - why do you need it?
It's not required. I am going to remove it.
+ reg:
+ maxItems: 1
+
+ interrupts-extended:
Instead interrupts
Sure. I will use 'interrupts'.
+ items:
+ - description: Watchdog interrupt
+ - description: Fatal interrupt
+ - description: Ready interrupt
+ - description: Handover interrupt
+ - description: Stop acknowledge interrupt
+
+ interrupt-names:
+ items:
+ - const: wdog
+ - const: fatal
+ - const: ready
+ - const: handover
+ - const: stop-ack
+
+ clocks:
+ minItems: 25
+ maxItems: 25
Drop both and instead describe the items. Anyway minItems are not needed
here.
Sure. I will drop min & max items and describe clocks.
+
+ clock-names:
+ minItems: 25
+ maxItems: 25
Drop both and instead list the names.
Sure. I will drop.
+
+ assigned-clocks:
+ minItems: 13
+ maxItems: 13
Drop, they do not have to be mentioned in the binding. If you think they
need to, then why?
+
+ assigned-clock-rates:
+ minItems: 13
+ maxItems: 13
Ditto
+
+ qcom,smem-states:
+ $ref: /schemas/types.yaml#/definitions/phandle-array
+ description: States used by the AP to signal the remoteprocessor
+ items:
+ - description: Shutdown Q6
+ - description: Stop Q6
+
+ qcom,smem-state-names:
+ description:
+ Names of the states used by the AP to signal the remoteprocessor
+ items:
+ - const: shutdown
+ - const: stop
+
+ memory-region:
+ items:
+ - description: Q6 pd reserved region
+
+ glink-edge:
+ $ref: /schemas/remoteproc/qcom,glink-edge.yaml#
unevaluatedProperties: false
Sure, will add.
+ description:
+ Qualcomm G-Link subnode which represents communication edge, channels
+ and devices related to the Modem.
+
+patternProperties:
+ "^remoteproc_pd1|remoteproc_pd2|remoteproc_pd3":
No, underscores are not allowed. Also, what is pd?
Sure, will remove underscores.
+ type: object
+ description:
+ In Multipd model, WCSS pd depends on Q6 pd i.e Q6 pd should be up before
+ WCSS. It can be achieved by keeping wcss pd node as subnode of Q6
+ device node.
+
+ properties:
+ compatible:
+ enum:
+ - "qcom,ipq5018-wcss-ahb-mpd"
+ - "qcom,ipq9574-wcss-ahb-mpd"
+ - "qcom,ipq5018-wcss-pcie-mpd"
Drop quotes
Sure, will remove it.
+
+ interrupts-extended:
Same as before
Sure, will use 'interrupts'.
+ items:
+ - description: Fatal interrupt
+ - description: Ready interrupt
+ - description: Spawn acknowledge interrupt
+ - description: Stop acknowledge interrupt
+
+ interrupt-names:
+ items:
+ - const: fatal
+ - const: ready
+ - const: spawn-ack
+ - const: stop-ack
+
+ qcom,smem-states:
+ $ref: /schemas/types.yaml#/definitions/phandle-array
+ description: States used by the AP to signal the remoteprocessor
+ items:
+ - description: Shutdown WCSS pd
+ - description: Stop WCSS pd
+ - description: Spawn WCSS pd
+
+ qcom,smem-state-names:
+ description:
+ Names of the states used by the AP to signal the remoteprocessor
remote processor
I will update.
+ items:
+ - const: shutdown
+ - const: stop
+ - const: spawn
This is confusing. Why your children have the same properties as parent?
Here both parent & child considered as remote processor. So once they
powered up/power down/crashed, they used to do some handshaking with
APPS processor. So interrupts are common between parent i.e root pd and
child i.e user pd
+
+ required:
+ - compatible
+
+ additionalProperties: false
+
+required:
+ - compatible
+ - reg
+ - interrupts-extended
interrupts
Sure. I will use 'interrupts' instead of interrupts-extended
+ - interrupt-names
+ - qcom,smem-states
+ - qcom,smem-state-names
+ - memory-region
+
+additionalProperties: false
+
+allOf:
+ - if:
+ properties:
+ compatible:
+ enum:
+ - qcom,ipq9574-q6-mpd
+ then:
+ properties:
+ assigned-clocks:
+ items:
+ - description: Phandle, clock specifier of GCC_ANOC_WCSS_AXI_M_CLK
+ - description: Phandle, clock specifier of GCC_WCSS_AHB_S_CLK
+ - description: Phandle, clock specifier of GCC_WCSS_ECAHB_CLK
+ - description: Phandle, clock specifier of GCC_WCSS_ACMT_CLK
+ - description: Phandle, clock specifier of GCC_WCSS_AXI_M_CLK
+ - description: Phandle, clock specifier of GCC_Q6_AXIM_CLK
+ - description: Phandle, clock specifier of GCC_Q6_AXIM2_CLK
+ - description: Phandle, clock specifier of GCC_Q6_AHB_CLK
+ - description: Phandle, clock specifier of GCC_Q6_AHB_S_CLK
+ - description: Phandle, clock specifier of GCC_Q6SS_BOOT_CLK
+ - description: Phandle, clock specifier of GCC_MEM_NOC_Q6_AXI_CLK
+ - description: Phandle, clock specifier of GCC_WCSS_Q6_TBU_CLK
+ - description: Phandle, clock specifier of GCC_SYS_NOC_WCSS_AHB_CLK
Eh, so here they are. But Why? Do you expect different clocks for
others? If so, where are they?
Anyway, drop useless "Phandle, clock specifier of". Clocks cannot be
anything else than phandle and a clock specifier. Instead of using some
cryptic ACRONYM_OR_SOME_CLK, describe them. Just like we do for other
bindings. You have plenty of good examples, so please start from them.
+ assigned-clock-rates:
+ items:
+ - description: Must be 266666667 HZ
+ - description: Must be 133333333 HZ
+ - description: Must be 133333333 HZ
+ - description: Must be 133333333 HZ
+ - description: Must be 266666667 HZ
+ - description: Must be 533000000 HZ
+ - description: Must be 342857143 HZ
+ - description: Must be 133333333 HZ
+ - description: Must be 133333333 HZ
+ - description: Must be 342857143 HZ
+ - description: Must be 533000000 HZ
+ - description: Must be 533000000 HZ
+ - description: Must be 133333333 HZ
???
If these are fixed, why this is in DT? DT is for variable and
non-discoverable pieces and you do not have here anything variable, but
fixed.
+
+examples:
+ - |
+ #include <dt-bindings/interrupt-controller/arm-gic.h>
+ #include <dt-bindings/clock/qcom,gcc-ipq5018.h>
+ #include <dt-bindings/reset/qcom,gcc-ipq5018.h>
Use 4 spaces for example indentation.
Sure, will use 4 spaces.
+
+ q6v5_wcss: remoteproc@cd00000 {
+ compatible = "qcom,ipq5018-q6-mpd";
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges;
+ reg = <0x0cd00000 0x4040>;
+ interrupts-extended = <&intc GIC_SPI 291 IRQ_TYPE_EDGE_RISING>,
+ <&wcss_smp2p_in 0 0>,
Wrong alignment of indentation
Sure, will update alignment.
+ <&wcss_smp2p_in 1 0>,
+ <&wcss_smp2p_in 2 0>,
+ <&wcss_smp2p_in 3 0>;
+ interrupt-names = "wdog",
+ "fatal",
+ "ready",
+ "handover",
+ "stop-ack";
+
+ qcom,smem-states = <&wcss_smp2p_out 0>,
+ <&wcss_smp2p_out 1>;
+ qcom,smem-state-names = "shutdown",
+ "stop";
+
+ memory-region = <&q6_region>;
+
+ glink-edge {
+ interrupts = <GIC_SPI 179 IRQ_TYPE_EDGE_RISING>;
+ label = "rtr";
+ qcom,remote-pid = <1>;
+ mboxes = <&apcs_glb 8>;
+ };
+
+ q6_wcss_pd1: remoteproc_pd1 {
+ compatible = "qcom,ipq5018-wcss-ahb-mpd";
+ interrupts-extended = <&wcss_smp2p_in 8 0>,
+ <&wcss_smp2p_in 9 0>,
+ <&wcss_smp2p_in 12 0>,
+ <&wcss_smp2p_in 11 0>;
+ interrupt-names = "fatal",
+ "ready",
+ "spawn-ack",
+ "stop-ack";
+ qcom,smem-states = <&wcss_smp2p_out 8>,
+ <&wcss_smp2p_out 9>,
+ <&wcss_smp2p_out 10>;
+ qcom,smem-state-names = "shutdown",
+ "stop",
+ "spawn";
+ };
+
+ q6_wcss_pd2: remoteproc_pd2 {
+ compatible = "qcom,ipq5018-wcss-pcie-mpd";
+ interrupts-extended = <&wcss_smp2p_in 16 0>,
+ <&wcss_smp2p_in 17 0>,
+ <&wcss_smp2p_in 20 0>,
+ <&wcss_smp2p_in 19 0>;
+ interrupt-names = "fatal",
+ "ready",
+ "spawn-ack",
+ "stop-ack";
+
+ qcom,smem-states = <&wcss_smp2p_out 16>,
+ <&wcss_smp2p_out 17>,
+ <&wcss_smp2p_out 18>;
+ qcom,smem-state-names = "shutdown",
+ "stop",
+ "spawn";
+ status = "okay";
Drop statuses from the example.
Sure, will drop status property.
Best regards,
Krzysztof
Thanks & Regards,
Manikanta.
