regulator: add QCOM RPMh regulator driver

This patch series adds a driver and device tree binding documentation for PMIC regulator control via Resource Power Manager-hardened (RPMh) on some Qualcomm Technologies, Inc. SoCs such as SDM845. RPMh is a hardware block which contains several accelerators which are used to manage various hardware resources that are shared between the processors of the SoC. The final hardware state of a regulator is determined within RPMh by performing max aggregation of the requests made by all of the processors.

The RPMh regulator driver depends upon the RPMh driver [1] and command DB driver [2] which are both still undergoing review. It also depends upon three recent regulator changes: [3], [4], and [5].

Changes since v3 [6]:
- Removed support for DT properties qcom,regulator-initial-microvolt and qcom,headroom-microvolt
- Renamed DT property qcom,allowed-drms-modes to be qcom,regulator-drms-modes
- Updated DT binding documentation to mention which common regulator bindings can be used for qcom-rpmh-regulator devices
- Added voltage caching so that voltage requests are only sent to RPMh after the regulator has been enabled at least once
- Changed 'voltage_selector' default value to be -ENOTRECOVERABLE to interact with [5]
- Initialized 'enabled' to -EINVAL so that unused regulators are disabled by regulator_late_cleanup()
- Removed rpmh_regulator_load_default_parameters() as it is no longer needed
- Updated the mode usage description in qcom,rpmh-regulator.h

Changes since v2 [7]:
- Replaced '_' with '-' in device tree supply property names
- Renamed qcom_rpmh-regulator.c to be qcom-rpmh-regulator.c
- Updated various DT property names to use "microvolt" and "microamp"
- Moved allowed modes constraint specification out of the driver [4]
- Replaced rpmh_client with device pointer to match new RPMh API [1]
- Corrected drms mode threshold checking
- Initialized voltage_selector to -EINVAL when not specified in DT
- Added constants for PMIC regulator hardware modes
- Corrected type sign of mode mapping tables
- Made variable names for mode arrays plural
- Simplified Kconfig depends on
- Removed unnecessary constants and struct fields
- Added some descriptive comments

Changes since v1 [8]:
- Addressed review feedback from Doug, Mark, and Stephen
- Replaced set_voltage()/get_voltage() callbacks with set_voltage_sel()/ get_voltage_sel()
- Added set_bypass()/get_bypass() callbacks for BOB pass-through mode control
- Removed top-level PMIC data structures
- Removed initialization variables from structs and passed them as function parameters
- Removed various comments and error messages
- Simplified mode handling
- Refactored per-PMIC rpmh-regulator data specification
- Simplified probe function
- Moved header into DT patch
- Removed redundant property listings from DT binding documentation

[1]: https://lkml.org/lkml/2018/5/9/729
[2]: https://lkml.org/lkml/2018/4/10/714
[3]: https://lkml.org/lkml/2018/4/18/556
[4]: https://lkml.org/lkml/2018/5/11/696
[5]: https://lkml.org/lkml/2018/5/15/1005
[6]: https://lkml.org/lkml/2018/5/11/701
[7]: https://lkml.org/lkml/2018/4/13/687
[8]: https://lkml.org/lkml/2018/3/16/1431

David Collins (2):
regulator: dt-bindings: add QCOM RPMh regulator bindings
regulator: add QCOM RPMh regulator driver

 .../bindings/regulator/qcom,rpmh-regulator.txt     | 185 +++++
 drivers/regulator/Kconfig                          |   9 +
 drivers/regulator/Makefile                         |   1 +
 drivers/regulator/qcom-rpmh-regulator.c            | 865 +++++++++++++++++++++
 .../dt-bindings/regulator/qcom,rpmh-regulator.h    |  36 +
 5 files changed, 1096 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/regulator/qcom,rpmh-regulator.txt
 create mode 100644 drivers/regulator/qcom-rpmh-regulator.c
 create mode 100644 include/dt-bindings/regulator/qcom,rpmh-regulator.h

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

Introduce bindings for RPMh regulator devices found on some Qualcomm Technlogies, Inc. SoCs. These devices allow a given processor within the SoC to make PMIC regulator requests which are aggregated within the RPMh hardware block along with requests from other processors in the SoC to determine the final PMIC regulator hardware state.

---
 .../bindings/regulator/qcom,rpmh-regulator.txt     | 185 +++++++++++++++++++++
 .../dt-bindings/regulator/qcom,rpmh-regulator.h    |  36 ++++
 2 files changed, 221 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/regulator/qcom,rpmh-regulator.txt
 create mode 100644 include/dt-bindings/regulator/qcom,rpmh-regulator.h

new file mode 100644
index 0000000..aaac975
@@ -0,0 +1,185 @@
+Qualcomm Technologies, Inc. RPMh Regulators
+
+rpmh-regulator devices support PMIC regulator management via the Voltage
+Regulator Manager (VRM) and Oscillator Buffer (XOB) RPMh accelerators.  The APPS
+processor communicates with these hardware blocks via a Resource State
+Coordinator (RSC) using command packets.  The VRM allows changing three
+parameters for a given regulator: enable state, output voltage, and operating
+mode.  The XOB allows changing only a single parameter for a given regulator:
+its enable state.  Despite its name, the XOB is capable of controlling the
+enable state of any PMIC peripheral.  It is used for clock buffers, low-voltage
+switches, and LDO/SMPS regulators which have a fixed voltage and mode.
+
+=======================
+Required Node Structure
+=======================
+
+RPMh regulators must be described in two levels of device nodes.  The first
+level describes the PMIC containing the regulators and must reside within an
+RPMh device node.  The second level describes each regulator within the PMIC
+which is to be used on the board.  Each of these regulators maps to a single
+RPMh resource.
+
+The names used for regulator nodes must match those supported by a given PMIC.
+Supported regulator node names:
+	PM8998:		smps1 - smps13, ldo1 - ldo28, lvs1 - lvs2
+	PMI8998:	bob
+	PM8005:		smps1 - smps4
+
+========================
+First Level Nodes - PMIC
+========================
+
+- compatible
+	Usage:      required
+	Value type: <string>
+	Definition: Must be one of: "qcom,pm8998-rpmh-regulators",
+		    "qcom,pmi8998-rpmh-regulators" or
+		    "qcom,pm8005-rpmh-regulators".
+
+- qcom,pmic-id
+	Usage:      required
+	Value type: <string>
+	Definition: RPMh resource name suffix used for the regulators found on
+		    this PMIC.  Typical values: "a", "b", "c", "d", "e", "f".
+
+- vdd-s1-supply
+- vdd-s2-supply
+- vdd-s3-supply
+- vdd-s4-supply
+	Usage:      optional (PM8998 and PM8005 only)
+	Value type: <phandle>
+	Definition: phandle of the parent supply regulator of one or more of the
+		    regulators for this PMIC.
+
+- vdd-s5-supply
+- vdd-s6-supply
+- vdd-s7-supply
+- vdd-s8-supply
+- vdd-s9-supply
+- vdd-s10-supply
+- vdd-s11-supply
+- vdd-s12-supply
+- vdd-s13-supply
+- vdd-l1-l27-supply
+- vdd-l2-l8-l17-supply
+- vdd-l3-l11-supply
+- vdd-l4-l5-supply
+- vdd-l6-supply
+- vdd-l7-l12-l14-l15-supply
+- vdd-l9-supply
+- vdd-l10-l23-l25-supply
+- vdd-l13-l19-l21-supply
+- vdd-l16-l28-supply
+- vdd-l18-l22-supply
+- vdd-l20-l24-supply
+- vdd-l26-supply
+- vin-lvs-1-2-supply
+	Usage:      optional (PM8998 only)
+	Value type: <phandle>
+	Definition: phandle of the parent supply regulator of one or more of the
+		    regulators for this PMIC.
+
+- vdd-bob-supply
+	Usage:      optional (PMI8998 only)
+	Value type: <phandle>
+	Definition: BOB regulator parent supply phandle
+
+===============================
+Second Level Nodes - Regulators
+===============================
+
+- qcom,regulator-drms-modes
+	Usage:      required if regulator-allow-set-load is specified;
+		    VRM regulators only
+	Value type: <prop-encoded-array>
+	Definition: A list of integers specifying the PMIC regulator modes which
+		    can be configured at runtime based upon consumer load needs.
+		    Supported values are RPMH_REGULATOR_MODE_* which are defined
+		    in [1] (i.e. 0 to 3).
+
+- qcom,drms-mode-max-microamps
+	Usage:      required if regulator-allow-set-load is specified;
+		    VRM regulators only
+	Value type: <prop-encoded-array>
+	Definition: A list of integers specifying the maximum allowed load
+		    current in microamps for each of the modes listed in
+		    qcom,regulator-drms-modes (matched 1-to-1 in order).
+		    Elements must be specified in order from lowest to highest
+		    value.
+
+- qcom,always-wait-for-ack
+	Usage:      optional
+	Value type: <empty>
+	Definition: Boolean flag which indicates that the application processor
+		    must wait for an ACK or a NACK from RPMh for every request
+		    sent for this regulator including those which are for a
+		    strictly lower power state.
+
+Other properties defined in Documentation/devicetree/bindings/regulator.txt
+may also be used.  regulator-initial-mode and regulator-allowed-modes may be
+specified for VRM regulators using mode values from [1].  regulator-allow-bypass
+may be specified for BOB type regulators managed via VRM.
+regulator-allow-set-load may be specified for VRM regulators.  It is used in
+conjunction with qcom,regulator-drms-modes and qcom,drms-mode-max-microamps to
+define a dynamic total load current to mode mapping.
+
+[1] include/dt-bindings/regulator/qcom,rpmh-regulator.h
+
+========
+Examples
+========
+
+#include <dt-bindings/regulator/qcom,rpmh-regulator.h>
+
+&apps_rsc {
+	pm8998-rpmh-regulators {
+		compatible = "qcom,pm8998-rpmh-regulators";
+		qcom,pmic-id = "a";
+
+		vdd-l7-l12-l14-l15-supply = <&pm8998_s5>;
+
+		smps2 {
+			regulator-min-microvolt = <1100000>;
+			regulator-max-microvolt = <1100000>;
+		};
+
+		pm8998_s5: smps5 {
+			regulator-min-microvolt = <1904000>;
+			regulator-max-microvolt = <2040000>;
+		};
+
+		ldo7 {
+			regulator-min-microvolt = <1800000>;
+			regulator-max-microvolt = <1800000>;
+			regulator-initial-mode = <RPMH_REGULATOR_MODE_LPM>;
+			regulator-allowed-modes =
+				<RPMH_REGULATOR_MODE_LPM
+				 RPMH_REGULATOR_MODE_HPM>;
+			regulator-allow-set-load;
+			qcom,regulator-drms-modes =
+				<RPMH_REGULATOR_MODE_LPM
+				 RPMH_REGULATOR_MODE_HPM>;
+			qcom,drms-mode-max-microamps = <10000 1000000>;
+		};
+
+		lvs1 {
+			regulator-min-microvolt = <1800000>;
+			regulator-max-microvolt = <1800000>;
+		};
+	};
+
+	pmi8998-rpmh-regulators {
+		compatible = "qcom,pmi8998-rpmh-regulators";
+		qcom,pmic-id = "b";
+
+		bob {
+			regulator-min-microvolt = <3312000>;
+			regulator-max-microvolt = <3600000>;
+			regulator-allowed-modes =
+				<RPMH_REGULATOR_MODE_AUTO
+				 RPMH_REGULATOR_MODE_HPM>;
+			regulator-initial-mode = <RPMH_REGULATOR_MODE_AUTO>;
+		};
+	};
+};

new file mode 100644
index 0000000..86713dc
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2018, The Linux Foundation. All rights reserved. */
+
+#ifndef __QCOM_RPMH_REGULATOR_H
+#define __QCOM_RPMH_REGULATOR_H
+
+/*
+ * These mode constants may be used to specify modes for various RPMh regulator
+ * device tree properties (e.g. regulator-initial-mode).  Each type of regulator
+ * supports a subset of the possible modes.
+ *
+ * %RPMH_REGULATOR_MODE_RET:	Retention mode in which only an extremely small
+ *				load current is allowed.  This mode is supported
+ *				by LDO and SMPS type regulators.
+ * %RPMH_REGULATOR_MODE_LPM:	Low power mode in which a small load current is
+ *				allowed.  This mode corresponds to PFM for SMPS
+ *				and BOB type regulators.  This mode is supported
+ *				by LDO, HFSMPS, BOB, and PMIC4 FTSMPS type
+ *				regulators.
+ * %RPMH_REGULATOR_MODE_AUTO:	Auto mode in which the regulator hardware
+ *				automatically switches between LPM and HPM based
+ *				upon the real-time load current.  This mode is
+ *				supported by HFSMPS, BOB, and PMIC4 FTSMPS type
+ *				regulators.
+ * %RPMH_REGULATOR_MODE_HPM:	High power mode in which the full rated current
+ *				of the regulator is allowed.  This mode
+ *				corresponds to PWM for SMPS and BOB type
+ *				regulators.  This mode is supported by all types
+ *				of regulators.
+ */
+#define RPMH_REGULATOR_MODE_RET		0
+#define RPMH_REGULATOR_MODE_LPM		1
+#define RPMH_REGULATOR_MODE_AUTO	2
+#define RPMH_REGULATOR_MODE_HPM		3
+
+#endif

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

Hi,

I'm confused as to why we are specifying the maximum current the device
can deliver in a given mode in the DT - surely that's a fixed property
of the hardware?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180530/a7eec198/attachment-0001.sig>

Hi,

Said another way: you're saying that you'd expect the "max-microamps" table to have one fewer element than the listed modes? So in the above example you'd have:

drms-modes: LPM, HPM
max-microamps: 10000

...or in a more complicated case, you could have:

drms-modes: RET, LPM, AUTO, HPM
max-microamps: 1, 100, 10000

Did I understand you correctly?

I'm personally OK with that color for the shed. I kinda like the symmetry of having the same number of elements in both lists (and being able to print an error message if someone ends up asking for too much current), but it's not a big deal for me either way.

-Doug

No, I'm saying that I don't know why that property exists at all. This
sounds like it's intended to be the amount of current the regulator can
deliver in each mode which is normally a design property of the silicon.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180530/ae2343e4/attachment.sig>

Hi,

Ah, got it. So the whole point here is to be able to implement either the function "set_load" or the function "get_optimum_mode". We need some sort of table to convert from current to mode. That's what this table does.

My argument to David was that since set_load / get_optimum_mode were features of the regulator core these should actually be standard properties and not Qualcomm-specific ones.

-Doug

We do need that table, my expectation would be that this table would be
in the driver as it's not something I'd expect to vary between different
systems but rather be a property of the silicon design. No sense in
every single board having to copy it in.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180530/e99e0527/attachment-0001.sig>

Hi,

Ah, got it! I'd be OK with it being hardcoded in the driver.

At one point I think David was making the argument that some boards have different noise needs for the rails and thus you might want to change modes at different currents. I don't know if this is realistic but I believe it was part of his original argument for why this needed to be specified. If we can hardcode this in the driver I'm fine with it... That would actually solve many of my objections too...

-Doug

Hello Mark and Doug,

The DRMS modes to use and max allowed current per mode need to be specified at the board level in device tree instead of hard-coded per regulator type in the driver. There are at least two use cases driving this need: LDOs shared between RPMh client processors and SMPSes requiring PWM mode in certain circumstances.

For LDOs the maximum low power mode (LPM) current is typically 10 mA or 30 mA (depending upon subtype) per hardware documentation. Unfortunately, sharing control of regulators with other processors adds some subtlety to the LPM current limit that should actually be applied at runtime.

Consider the case of a regulator with physical 10 mA LPM max current. Say that modem and application processors each have a load on the regulator that draws 9 mA. If they each respect the 10 mA limit, then they'd each vote for LPM. The VRM block in RPMh hardware will aggregate these requests together using a max function which will result in the regulator being set to LPM, even though the total load is 18 mA (which would require high power mode (HPM)). To get around this corner case, a LPM max current of 1 uA can be used for all LDO supplies that have non-application processor consumers. Thus, any non-zero regulator_set_load() current request will result in setting the regulator to HPM (which is always safe).

The second situation that needs board-level DRMS mode and current limit specification is SMPS regulator AUTO mode to PWM (HPM) mode switching. SMPS regulators should theoretically always be able to operate in AUTO mode as it switches automatically between PWM mode (which can provide the maximum current) and PFM mode (which supports lower current but has higher efficiency). However, there may be board/system issues that require switching to PWM mode for certain use cases as it has better load regulation (i.e. no PFM ripple for lower loads) and supports more aggressive load current steps (i.e. greater A/ns).

If a Linux consumer requires the ability to force a given SMPS regulator from AUTO mode into PWM mode and that SMPS is shared by other Linux consumers (which may be the case, but at least must be guaranteed to work architecturally), then regulator_set_load() is the only option since it provides aggregation, where as regulator_set_mode() does not. regulator_set_load() can be utilized in this case by specifying AUTO mode and PWM mode as drms modes and specifying some particular AUTO mode maximum current (that is known by the consumer) in device tree. The consumer can then call regulator_set_load() with the imposed AUTO mode limit + delta when PWM mode is required and a lower value when AUTO mode is sufficient.

Note that I previously mentioned the need for board-level drms mode and current limit specification in [1].

Take care,
David

[1]: https://lkml.org/lkml/2018/3/22/802

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

Hi,

Is there any plan to change the way this works for future versions of RPMh?

Mark: I'm leaving this firmly in your hands. I can see David's points here. I could even believe that some of this stuff could be board specific where one board might have slightly different capacitors or they might be placed differently and might need a higher power mode to keep the signal clean.

-Doug

Hi Doug,

As far as I know, no. Adding VRM logic to sum RPMh client load current requests, compare the sums to thresholds, select corresponding modes, and override with explicit modes was determined to be too complex for RPMh. Instead, much simpler max aggregation was implemented along with design guidelines for PMICs to ensure that mode values are ordered from strictly lower to higher power.

Take care,
David

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

Is there really no way to improve the RPM firmware?

This is, of course, why the regulator API aggregates this stuff based on the current not based on having every individual user select a mode.

That's obviously just never going to work well, the best you can do here is just pretend that the other components are always operating at full power (which I assume all the other components are doing too...) or not try to use load based mode switching in the first place.

That's obviously broken though, what you're describing is just clearly nothing to do with load so trying to configure it using load is just going to lead to problems later on. Honestly it sounds like you just want to force the regulator into forced PWM mode all the time, otherwise you need to look into implementing support for describing the thing you're actually trying to do and add a mechanism for per consumer mode configuration.

This has been a frequent pattern with these RPM drivers, trying to find
some way to shoehorn something that happens to work right now into the
code. That's going to make things fragile and hard to maintain, we need
code that does the thing it's saying it does so that it's easier to
understand and work with - getting things running isn't enough, they
need to be clear.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180531/53042fa7/attachment.sig>

Hello Mark,

This aggregation takes place in a discrete hardware block, not in a general purpose processor. Theoretically, future chips could have redesigned VRM hardware; however, there is no plan to make such a change.

I will remove the DT-based mode and max load current mapping support. In its place, I'll implement hard coded LPM current limits for LDOs of 10 mA or 30 mA (depending upon subtype) like is done in other regulator drivers.

If we actually encounter an issue caused by the APPS processor and another RPMh client both voting for LPM when HPM is needed for the combination, then we can disable load-based mode control for the affected regulator in DT and configure its initial mode as HPM.

I agree that using regulator_set_load() to handle SMPS AUTO mode to PWM mode switching is hacky. Since there is no natural way to pick SMPS modes based on load current, I will remove the functionality completely. In its place, we can configure the SMPSes to have an initial mode of AUTO. If a use case requiring PWM mode arises, then the consumer driver responsible for it can call regulator_set_mode() to switch between AUTO and PWM modes explicitly.

Since regulator_set_mode() does not support aggregation currently, this technique would work only if there is exactly one consumer per regulator that needs explicit mode control. Should we hit a situation that needs multiple consumers to have such mode control, then we could simply default the SMPS to PWM mode.

I'll also start looking into per-consumer mode configuration and regulator_set_mode() aggregation within the regulator framework. I think that we should be able to function without it for now; however, it would be good to have going forward.

Take care,
David

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

Add the QCOM RPMh regulator driver to manage PMIC regulators which are controlled via RPMh on some Qualcomm Technologies, Inc. SoCs. RPMh is a hardware block which contains several accelerators which are used to manage various hardware resources that are shared between the processors of the SoC. The final hardware state of a regulator is determined within RPMh by performing max aggregation of the requests made by all of the processors.

Add support for PMIC regulator control via the voltage regulator manager (VRM) and oscillator buffer (XOB) RPMh accelerators. VRM supports manipulation of enable state, voltage, and mode. XOB supports manipulation of enable state.

---
 drivers/regulator/Kconfig               |   9 +
 drivers/regulator/Makefile              |   1 +
 drivers/regulator/qcom-rpmh-regulator.c | 865 ++++++++++++++++++++++++++++++++
 3 files changed, 875 insertions(+)
 create mode 100644 drivers/regulator/qcom-rpmh-regulator.c

index 4efae3b..1a69bdc 100644
@@ -671,6 +671,15 @@ config REGULATOR_QCOM_RPM
 	  Qualcomm RPM as a module. The module will be named
 	  "qcom_rpm-regulator".
 
+config REGULATOR_QCOM_RPMH
+	tristate "Qualcomm Technologies, Inc. RPMh regulator driver"
+	depends on QCOM_RPMH || COMPILE_TEST
+	help
+	  This driver supports control of PMIC regulators via the RPMh hardware
+	  block found on Qualcomm Technologies Inc. SoCs.  RPMh regulator
+	  control allows for voting on regulator state between multiple
+	  processors within the SoC.
+
 config REGULATOR_QCOM_SMD_RPM
 	tristate "Qualcomm SMD based RPM regulator driver"
 	depends on QCOM_SMD_RPM

index d81fb02..906f048 100644
@@ -77,6 +77,7 @@ obj-$(CONFIG_REGULATOR_MT6323)	+= mt6323-regulator.o
 obj-$(CONFIG_REGULATOR_MT6380)	+= mt6380-regulator.o
 obj-$(CONFIG_REGULATOR_MT6397)	+= mt6397-regulator.o
 obj-$(CONFIG_REGULATOR_QCOM_RPM) += qcom_rpm-regulator.o
+obj-$(CONFIG_REGULATOR_QCOM_RPMH) += qcom-rpmh-regulator.o
 obj-$(CONFIG_REGULATOR_QCOM_SMD_RPM) += qcom_smd-regulator.o
 obj-$(CONFIG_REGULATOR_QCOM_SPMI) += qcom_spmi-regulator.o
 obj-$(CONFIG_REGULATOR_PALMAS) += palmas-regulator.o

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

Hi,

Do you really need this extra boolean? Can't you just check if "enabled" is still "-EINVAL"? If it is then you don't pass the voltage along.

...this would mean that you'd also need to send the voltage vote when the regulator core tries to disable unused regulators at the end of bootup, but that should be OK right? If we never touched a regulator anywhere at probe time and we're about to vote to disable it, we know there's nobody requiring it to still be on. We can vote for the voltage now without fear of messing up a vote that the BIOS left in place.

In theory this should also allow you to assert your vote about the voltage of a regulator that has never been enabled, which (if I understand correctly) you consider to be a feature.

---

Other than that comment, everything else here looks good to go if Mark is good with it. As per the previous threads there are some things that I don't like a ton, but I feel it is between you and Mark to decide if you're good with it. A summary of those issues are:

1. I still believe that when we disable a regulator in Linux we should tell RPMh that we vote for the lowest voltage. ...but if Mark is happy with the way the driver works now I won't push it.

2. As per my comments in the bindings patch, I still believe that "qcom,drms-mode-max-microamps" belongs in the core. Again, up to Mark.

3. As per my comments in the bindings patch, I still believe that we're just adding lots of noise to the DT most of the time by specifying both "qcom,regulator-drms-modes" and "regulator-allowed-modes". Again, up to Mark.

-Doug

Hello Doug,

Removing 'ever_enabled' and caching the voltage when 'enabled == -EINVAL' seems workable. I'm a little concerned about this resulting in voltage = regulator-min-microvolt requests being sent for all regulators that are not explicitly enabled by Linux consumers before late_initcall_sync(). Theoretically all of the boot loader hand-off cases should be taken care of by this point so it should be safe.

I'll make this change.

Take care,
David

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

So, this is a very big driver and obviously it being RPM based it doesn't look like other regulators which is causing problems, especially when coupled with the desire to implement a bunch of more exotic features like the mode setting. I think this review is going to go a lot more smoothly if you split this up into a base driver with just normal, standard stuff that doesn't add too many custom properties or unusual ways of working and then a series of patches on top of that adding things like the mode adjustment and interaction with other RPM clients.

We've got other RPM based regulators in tree already so the baseline bit
shouldn't be too hard, that'll make the rest of the patches much smaller
and easier to review and mean that the bits that are simpler and easier
to cope with don't need to be reposted.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180530/fd8d3adf/attachment.sig>

Hello Mark,

I'll split up the patches so that reviewing is easier. For the base patch, would you prefer that I remove *all* mode support (handled by generic regulator framework DT properties) or only remove the special purpose drms mode handling support (i.e. qcom,regulator-drms-modes and qcom,drms-mode-max-microamps)?

Thanks,
David

The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

Standard operations should be fine, just weird custom stuff that's been
causing issues.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180531/cc95ff7c/attachment.sig>