Skip to contentSkip to conversation

initialize (multiple) PHYs for a HCD

PATCH Superseded Martin Blumenstingl 5 participants 29 messages Feb 18, 2018 – Mar 24, 2018

Subsystems (11)

v10 of 11 Superseded Latest is v11 →

lore
Share & tools
lore

Martin Blumenstingl

author reviewer
#0 lore raw

v10 cover initialize (multiple) PHYs for a HCD

The goal of this series is to initialize multiple PHYs on a USB host controller, which is needed on Amlogic Meson GXL and GXM SoCs.
- Amlogic Meson GXL and GXM SoCs come with a dwc3 IP block which has two or three USB2 ports enabled on the internal root-hub. The SoCs also provide separate USB2 PHYs, one per port. All USB2 PHYs (which are internally "connected" to the dwc3 roothub) need to be powered on, otherwise USB devices cannot be enumerated (even if just one PHY is disabled and if the device is plugged into another, enabled port)
- dwc3 already takes one USB2 and one USB3 PHY and initializes these correctly
- some other HCI platform drivers (like ehci-platform.c, xhci-mtk.c and ohci-platform.c) do not have a limitation on the number of PHYs - they support any number of USB PHYs (typically one PHY per actual port)

This series adds a new "PHY wrapper". This can be configured through devicetree by passing a "phys" property (with any number of PHY handles) to the USB controller.
Additionally it this integrates this new PHY wrapper into hcd.c which automatically enables it for all USB controller drivers (tested on an Amlogic Meson GXL SoC which uses a dwc3 controller). Since this leaves some duplicate code (which now becomes obsolete!) in various drivers this is being cleaned up as well.

Changes since v9 at [12]:
- changed "skip_phy_initialization" from PATCH #2 into a bitfield and moved it below the "remove_phy" bitfield as suggested by Alan Stern (thanks!)
- rebased against usb-next (which is currently identical to v4.16-rc1)

Changes since v8 at [10]:
- add new patch #2 which allows disabling "USB PHY management" in the USB HCD core driver. this also affects patch #4 which now checks the skip_phy_initialization flag before calling usb_phy_roothub_{init,power_on} (other functions don't need to be guarded since they are null-safe - see patch #3)
- integrated my other series 'remove driver-specific "multiple PHY" handling' from [11] into this one so the result of this "generic" code (= removing duplicate code in various drivers) can be seen directly. affects patches #5-8. (note: patch #1 "usb: mtu3: remove custom USB PHY handling" from that series was dropped, because it broke device mode support - Chunfeng Yun will come up with a separate patch later. patch #5 "usb: chipidea: do not set the "phy" field in struct usb_hcd" from that series was rewritten and is now integrated into the new patch #2 from this series)
- collected all {Reviewed,Tested,Acked}-by
- dropped RfC prefix since this v8 was out for review for two weeks and it only got Reviewed-by's and Tested-by's since then

Changes since v7 at [8]:
- updated the description in the cover-letter
- wrapped the only call usb_phy_roothub_power_{on,off} in hcd_bus_{resume,suspend} during system-suspend (excluding it from being called during runtime/auto suspend). this also fixes a problem on Meson GXL which does not detect that a new device has been connected if the PHYs are powered off. Thanks to Manu Gautam for suggesting this! (affects patch #3)
- after a discussion about the dt-bindings of this patch I went back to simply specifying a "phys" property (without a corresponding phy-names property) to pass all PHYs directly in the node of the USB controller. From my understanding the root hub node was "artificial" and we should describe it's properties as part of the controller, rather than introducing a sub-node. see also the discussion with Arnd in [9] (this affects patch #1 and #2)
- dropped patch #4 ("dt-bindings: usb: xhci: include the roothub and a device in the example") because we're back to just specifying a "phys" property directly within the controller.
NOTE: some HCI platform drivers are using the "phys" property as well. With this series phy_{init,power_on,power_off,exit} will be called twice for each PHY when using a ehci-platform.c or xhci-mtk.c USB controller (because the controller driver and our new, generic code now manage the PHYs). the PHY framework handles this gracefully (by ref-counting all operations). a follow-up series will be sent which removes the custom handling from the affected drivers (so only the new PHY wrapper will manage the PHYs state after that follow-up series).
- fixed an unnecessary whitespace change spotted by Alan Stern (thanks!) in patch #3
- removed all Tested-by's since we're changing code again

Changes since v6 at [6]:
- fixed unnecessary whitespace change (noticed by Alan Stern - thanks)
- added PATCH #4 to clarify (with an example) how I understood how Rob wanted the dt-binding to look like (see [0]). please wait for an ACK from Rob on this one to make sure that we're not introducing some broken binding (as discussed with Arnd: [7])
- added Alan Stern's ACK on PATCH #3
- added Neil's Tested-by on PATCH #2 and #3

Changes since RfC v5 at [5]:
- dropped RfC prefix
- removed noisy dev_err if no roothub node was found (spotted by Xiaolong Ye's kbuild test robot - thank you for that!)
- moved the call to usb_phy_roothub_power_off() within hcd_bus_suspend() to make sure that the PHYs are turned off if the "race with a root-hub wakeup event" condition is met (in this case the PHYs are turned on again, with the old code we did break the PHYs internal ref-counting because we never turned the PHYs off before turning them on again in case of that special "race with a root-hub wakeup event").
additionally we're not handling the status returned by usb_phy_roothub_power_off() anymore (the bus is already turned off and we tried to turn off all PHYs as well - only the PHYs which failed to power off will stay in the current state). thanks to Alan Stern for the suggestion
- removed return value from usb_phy_roothub_power_off() because none of my code uses it anymore. thanks to Alan Stern for the suggestion

Changes since v4 at [4]:
- renamed the subject of the cover-letter (old name was:
"initialize (multiple) PHYs in xhci-plat")
- back into RFC status (see below for the reasons)
- dropped Tested-by from Chunfeng Yun (same reasons as RFC status)
- reworded cover-letter and commit messages from "platform-roothub" to "roothub PHY wrapper"
- moved code from drivers/usb/host/platform-roothub.* to drivers/usb/core/phy.* and the changes to drivers/usb/host/xhci-plat.c to drivers/usb/core/hcd.c as suggested by Mathias Nyman (as a benefit this will enable the new logic for non-xHCI controllers as well - however this was not tested yet)
- rename the structs, function names, etc from platform_roothub_* to usb_phy_roothub*

Changes since RFCv3 at [3]:
- moved the DT binding change from patch #3 to patch #1 as suggested by Rob Herring (and slightly adjusted the commit message to account for that)
- added Tested-by from Chunfeng Yun (who confirmed that the whole concept and implementation works fine on Mediatek SoCs - many thanks again!) to patch #2
- added Rob Herring's ACK to patches 1 and 3
- dropped RFC status (RFCv3 -> PATCH v4)

Changes since RFCv2 at [2]:
- split phy_{init,exit} and phy_power_{on,off} handling. up until RFCv2 I called phy_init plus phy_power_on in platform_roothub_power_on and phy_power_off plus phy_exit in platform_roothub_power_off. However, Chunfeng Yun (a Mediatek SoC developer - many thanks for testing my series and providing great feedback) reported that only using phy_power_off (and omitting phy_exit) during system suspend fixes an issue where USB devices would be re-enumerated when resuming. His original problem description: "In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device.". This fix affects patch #2 and #3 as we now have platform_roothub_init (which calls phy_init internally), platform_roothub_power_on (which calls phy_power_on internally), platform_roothub_power_off (which calls phy_power_off internally) and platform_roothub_exit (which calls phy_exit internally). suspend and resume only call platform_roothub_power_{on,off} to prevent the issue described by Chunfeng Yun (unfortunately I cannot test this because the Amlogic platform currently does not support system suspend).
- dropped two struct forward declarations from platform-roothub.h which are not used in the header file (thanks to Chunfeng Yun for spotting this)

Changes since RFCv1 at [1]:
- split the usb-xhci dt-binding documentation into a separate patch
- fixed a typo ("usb-phy" -> "phys" in the dt-binding example)
- rebased to apply against latest usb-next

[0] http://lists.infradead.org/pipermail/linux-amlogic/2016-November/001818.html
[1] http://marc.info/?l=linux-usb&m=148414866303604&w=2
[2] https://www.spinics.net/lists/linux-usb/msg158967.html
[3] https://www.spinics.net/lists/devicetree/msg190426.html
[4] http://lists.infradead.org/pipermail/linux-amlogic/2017-September/004685.html
[5] http://lists.infradead.org/pipermail/linux-amlogic/2017-October/004924.html
[6] http://lists.infradead.org/pipermail/linux-amlogic/2017-October/005049.html
[7] http://lists.infradead.org/pipermail/linux-amlogic/2017-October/005121.html
[8] http://lists.infradead.org/pipermail/linux-amlogic/2017-October/005124.html
[9] http://lists.infradead.org/pipermail/linux-amlogic/2017-October/005192.html
[10] http://lists.infradead.org/pipermail/linux-amlogic/2018-January/006274.html
[11] https://marc.info/?l=linux-usb&m=151683941207898&w=2
[12] http://lists.infradead.org/pipermail/linux-amlogic/2018-February/006412.html

Martin Blumenstingl (8):
dt-bindings: usb: add the documentation for USB HCDs
usb: add a flag to skip PHY initialization to struct usb_hcd
usb: core: add a wrapper for the USB PHYs on the HCD
usb: core: hcd: integrate the PHY wrapper into the HCD core
usb: host: xhci-mtk: remove custom USB PHY handling
usb: host: ehci-platform: remove custom USB PHY handling
usb: host: ohci-platform: remove custom USB PHY handling
usb: core: hcd: remove support for initializing a single PHY

 .../devicetree/bindings/usb/mediatek,mtk-xhci.txt  |   5 +-
 .../devicetree/bindings/usb/mediatek,mtu3.txt      |   5 +-
 Documentation/devicetree/bindings/usb/usb-ehci.txt |   6 +-
 Documentation/devicetree/bindings/usb/usb-hcd.txt  |   9 ++
 Documentation/devicetree/bindings/usb/usb-ohci.txt |   6 +-
 Documentation/devicetree/bindings/usb/usb-uhci.txt |   3 +
 Documentation/devicetree/bindings/usb/usb-xhci.txt |   5 +
 drivers/usb/chipidea/host.c                        |   6 +-
 drivers/usb/core/Makefile                          |   2 +-
 drivers/usb/core/hcd.c                             |  64 ++++-----
 drivers/usb/core/phy.c                             | 158 +++++++++++++++++++++
 drivers/usb/core/phy.h                             |   7 +
 drivers/usb/host/ehci-fsl.c                        |   2 +
 drivers/usb/host/ehci-platform.c                   |  55 +------
 drivers/usb/host/ehci-tegra.c                      |   1 +
 drivers/usb/host/ohci-omap.c                       |   1 +
 drivers/usb/host/ohci-platform.c                   |  56 +-------
 drivers/usb/host/xhci-mtk.c                        |  98 +------------
 drivers/usb/host/xhci-plat.c                       |   1 +
 include/linux/usb/hcd.h                            |   8 +-
 20 files changed, 253 insertions(+), 245 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/usb/usb-hcd.txt
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
Signature · 1 line
2.16.1
author reviewer
#1 lore raw

PATCH v10 1/8 dt-bindings: usb: add the documentation for USB HCDs

A USB HCD may have several PHYs which need to be configured before the the HCD starts working.
This adds the documentation for such a USB HCD as well as a reference to the new "usb-hcd.txt" from all bindings that implement a USB HCD which support one USB PHY per port.

---
 Documentation/devicetree/bindings/usb/mediatek,mtk-xhci.txt | 5 ++++-
 Documentation/devicetree/bindings/usb/mediatek,mtu3.txt     | 5 ++++-
 Documentation/devicetree/bindings/usb/usb-ehci.txt          | 6 ++++--
 Documentation/devicetree/bindings/usb/usb-hcd.txt           | 9 +++++++++
 Documentation/devicetree/bindings/usb/usb-ohci.txt          | 6 ++++--
 Documentation/devicetree/bindings/usb/usb-uhci.txt          | 3 +++
 Documentation/devicetree/bindings/usb/usb-xhci.txt          | 5 +++++
 7 files changed, 33 insertions(+), 6 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/usb/usb-hcd.txt
Documentation/devicetree/bindings/usb/mediatek,mtk-xhci.txt +4 -1
index 88d9f4a4b280..266c2d917a28 100644
@@ -32,7 +32,7 @@ Required properties:
 	"mcu_ck": mcu_bus clock for register access,
 	"dma_ck": dma_bus clock for data transfer by DMA
 
- - phys : a list of phandle + phy specifier pairs
+ - phys : see usb-hcd.txt in the current directory
 
 Optional properties:
  - wakeup-source : enable USB remote wakeup;
@@ -52,6 +52,9 @@ Optional properties:
 	See: Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt
  - imod-interval-ns: default interrupt moderation interval is 5000ns
 
+additionally the properties from usb-hcd.txt (in the current directory) are
+supported.
+
 Example:
 usb30: usb at 11270000 {
 	compatible = "mediatek,mt8173-xhci";
Documentation/devicetree/bindings/usb/mediatek,mtu3.txt +4 -1
index d589a1ef96a1..3382b5cb471d 100644
@@ -17,7 +17,7 @@ Required properties:
  - clock-names : must contain "sys_ck" for clock of controller,
 	the following clocks are optional:
 	"ref_ck", "mcu_ck" and "dam_ck";
- - phys : a list of phandle + phy specifier pairs
+ - phys : see usb-hcd.txt in the current directory
  - dr_mode : should be one of "host", "peripheral" or "otg",
 	refer to usb/generic.txt
 
@@ -53,6 +53,9 @@ Optional properties:
  - mediatek,u3p-dis-msk : mask to disable u3ports, bit0 for u3port0,
 	bit1 for u3port1, ... etc;
 
+additionally the properties from usb-hcd.txt (in the current directory) are
+supported.
+
 Sub-nodes:
 The xhci should be added as subnode to mtu3 as shown in the following example
 if host mode is enabled. The DT binding details of xhci can be found in:
Documentation/devicetree/bindings/usb/usb-ehci.txt +4 -2
index 3efde12b5d68..0f1b75386207 100644
@@ -16,10 +16,12 @@ Optional properties:
  - has-transaction-translator : boolean, set this if EHCI have a Transaction
 				Translator built into the root hub.
  - clocks : a list of phandle + clock specifier pairs
- - phys : phandle + phy specifier pair
- - phy-names : "usb"
+ - phys : see usb-hcd.txt in the current directory
  - resets : phandle + reset specifier pair
 
+additionally the properties from usb-hcd.txt (in the current directory) are
+supported.
+
 Example (Sequoia 440EPx):
     ehci at e0000300 {
 	   compatible = "ibm,usb-ehci-440epx", "usb-ehci";
Documentation/devicetree/bindings/usb/usb-hcd.txt new file +9 -0
new file mode 100644
index 000000000000..50529b838c9c
@@ -0,0 +1,9 @@
+Generic USB HCD (Host Controller Device) Properties
+
+Optional properties:
+- phys: a list of all USB PHYs on this HCD
+
+Example:
+	&usb1 {
+		phys = <&usb2_phy1>, <&usb3_phy1>;
+	};
Documentation/devicetree/bindings/usb/usb-ohci.txt +4 -2
index 09e70c875bc6..a8d2103d1f3d 100644
@@ -13,10 +13,12 @@ Optional properties:
 - remote-wakeup-connected: remote wakeup is wired on the platform
 - num-ports : u32, to override the detected port count
 - clocks : a list of phandle + clock specifier pairs
-- phys : phandle + phy specifier pair
-- phy-names : "usb"
+- phys : see usb-hcd.txt in the current directory
 - resets : a list of phandle + reset specifier pairs
 
+additionally the properties from usb-hcd.txt (in the current directory) are
+supported.
+
 Example:
 
 	ohci0: usb at 1c14400 {
Documentation/devicetree/bindings/usb/usb-uhci.txt +3 -0
index 298133416c97..cc2e6f7d602e 100644
@@ -6,6 +6,9 @@ Required properties:
 - reg : Should contain 1 register ranges(address and length)
 - interrupts : UHCI controller interrupt
 
+additionally the properties from usb-hcd.txt (in the current directory) are
+supported.
+
 Example:
 
 	uhci at d8007b00 {
Documentation/devicetree/bindings/usb/usb-xhci.txt +5 -0
index e2ea59bbca93..2f7663bb69ff 100644
@@ -32,6 +32,11 @@ Optional properties:
   - usb3-lpm-capable: determines if platform is USB3 LPM capable
   - quirk-broken-port-ped: set if the controller has broken port disable mechanism
   - imod-interval-ns: default interrupt moderation interval is 5000ns
+  - phys : see usb-hcd.txt in the current directory
+
+additionally the properties from usb-hcd.txt (in the current directory) are
+supported.
+
 
 Example:
 	usb at f0931000 {
Signature · 1 line
2.16.1
author reviewer
#2 lore raw

PATCH v10 2/8 usb: add a flag to skip PHY initialization to struct usb_hcd

In reply to Martin Blumenstingl

The USB HCD core driver parses the device-tree node for "phys" and "usb-phys" properties. It also manages the power state of these PHYs automatically.
However, drivers may opt-out of this behavior by setting "phy" or "usb_phy" in struct usb_hcd to a non-null value. An example where this is required is the "Qualcomm USB2 controller", implemented by the chipidea driver. The hardware requires that the PHY is only powered on after the "reset completed" event from the controller is received.

A follow-up patch will allow the USB HCD core driver to manage more than one PHY. Add a new "skip_phy_initialization" bitflag to struct usb_hcd so drivers can opt-out of any PHY management provided by the USB HCD core driver.

This also updates the existing drivers so they use the new flag if they want to opt out of the PHY management provided by the USB HCD core driver. This means that for these drivers the new "multiple PHY" handling (which will be added in a follow-up patch) will be disabled as well.

---
 drivers/usb/chipidea/host.c      | 6 ++----
 drivers/usb/core/hcd.c           | 4 ++--
 drivers/usb/host/ehci-fsl.c      | 2 ++
 drivers/usb/host/ehci-platform.c | 4 ++--
 drivers/usb/host/ehci-tegra.c    | 1 +
 drivers/usb/host/ohci-omap.c     | 1 +
 drivers/usb/host/ohci-platform.c | 4 ++--
 drivers/usb/host/xhci-plat.c     | 1 +
 include/linux/usb/hcd.h          | 6 ++++++
 9 files changed, 19 insertions(+), 10 deletions(-)
drivers/usb/chipidea/host.c +2 -4
index 19d60ed7e41f..af45aa3222b5 100644
@@ -124,10 +124,8 @@ static int host_start(struct ci_hdrc *ci)
 
 	hcd->power_budget = ci->platdata->power_budget;
 	hcd->tpl_support = ci->platdata->tpl_support;
-	if (ci->phy)
-		hcd->phy = ci->phy;
-	else
-		hcd->usb_phy = ci->usb_phy;
+	if (ci->phy || ci->usb_phy)
+		hcd->skip_phy_initialization = 1;
 
 	ehci = hcd_to_ehci(hcd);
 	ehci->caps = ci->hw_bank.cap;
drivers/usb/core/hcd.c +2 -2
index fc32391a34d5..f2307470a31e 100644
@@ -2727,7 +2727,7 @@ int usb_add_hcd(struct usb_hcd *hcd,
 	int retval;
 	struct usb_device *rhdev;
 
-	if (IS_ENABLED(CONFIG_USB_PHY) && !hcd->usb_phy) {
+	if (IS_ENABLED(CONFIG_USB_PHY) && !hcd->skip_phy_initialization) {
 		struct usb_phy *phy = usb_get_phy_dev(hcd->self.sysdev, 0);
 
 		if (IS_ERR(phy)) {
@@ -2745,7 +2745,7 @@ int usb_add_hcd(struct usb_hcd *hcd,
 		}
 	}
 
-	if (IS_ENABLED(CONFIG_GENERIC_PHY) && !hcd->phy) {
+	if (IS_ENABLED(CONFIG_GENERIC_PHY) && !hcd->skip_phy_initialization) {
 		struct phy *phy = phy_get(hcd->self.sysdev, "usb");
 
 		if (IS_ERR(phy)) {
drivers/usb/host/ehci-fsl.c +2 -0
index c5094cb88cd5..0a9fd2022acf 100644
@@ -155,6 +155,8 @@ static int fsl_ehci_drv_probe(struct platform_device *pdev)
 			retval = -ENODEV;
 			goto err2;
 		}
+
+		hcd->skip_phy_initialization = 1;
 	}
 #endif
 	return retval;
drivers/usb/host/ehci-platform.c +2 -2
index b065a960adc2..b91eea8c73ae 100644
@@ -219,9 +219,9 @@ static int ehci_platform_probe(struct platform_device *dev)
 			if (IS_ERR(priv->phys[phy_num])) {
 				err = PTR_ERR(priv->phys[phy_num]);
 					goto err_put_hcd;
-			} else if (!hcd->phy) {
+			} else {
 				/* Avoiding phy_get() in usb_add_hcd() */
-				hcd->phy = priv->phys[phy_num];
+				hcd->skip_phy_initialization = 1;
 			}
 		}
 
drivers/usb/host/ehci-tegra.c +1 -0
index c809f7d2f08f..a6f4389f7e88 100644
@@ -461,6 +461,7 @@ static int tegra_ehci_probe(struct platform_device *pdev)
 		goto cleanup_clk_en;
 	}
 	hcd->usb_phy = u_phy;
+	hcd->skip_phy_initialization = 1;
 
 	tegra->needs_double_reset = of_property_read_bool(pdev->dev.of_node,
 		"nvidia,needs-double-reset");
drivers/usb/host/ohci-omap.c +1 -0
index 0201c49bc4fc..d8d35d456456 100644
@@ -230,6 +230,7 @@ static int ohci_omap_reset(struct usb_hcd *hcd)
 		} else {
 			return -EPROBE_DEFER;
 		}
+		hcd->skip_phy_initialization = 1;
 		ohci->start_hnp = start_hnp;
 	}
 #endif
drivers/usb/host/ohci-platform.c +2 -2
index 1e6c954f4b3f..62ef36a9333f 100644
@@ -186,9 +186,9 @@ static int ohci_platform_probe(struct platform_device *dev)
 			if (IS_ERR(priv->phys[phy_num])) {
 				err = PTR_ERR(priv->phys[phy_num]);
 				goto err_put_hcd;
-			} else if (!hcd->phy) {
+			} else {
 				/* Avoiding phy_get() in usb_add_hcd() */
-				hcd->phy = priv->phys[phy_num];
+				hcd->skip_phy_initialization = 1;
 			}
 		}
 
drivers/usb/host/xhci-plat.c +1 -0
index 6f038306c14d..6700e5ee82ad 100644
@@ -284,6 +284,7 @@ static int xhci_plat_probe(struct platform_device *pdev)
 		ret = usb_phy_init(hcd->usb_phy);
 		if (ret)
 			goto put_usb3_hcd;
+		hcd->skip_phy_initialization = 1;
 	}
 
 	ret = usb_add_hcd(hcd, irq, IRQF_SHARED);
include/linux/usb/hcd.h +6 -0
index 176900528822..693502c84c04 100644
@@ -151,6 +151,12 @@ struct usb_hcd {
 	unsigned		msix_enabled:1;	/* driver has MSI-X enabled? */
 	unsigned		msi_enabled:1;	/* driver has MSI enabled? */
 	unsigned		remove_phy:1;	/* auto-remove USB phy */
+	/*
+	 * do not manage the PHY state in the HCD core, instead let the driver
+	 * handle this (for example if the PHY can only be turned on after a
+	 * specific event)
+	 */
+	unsigned		skip_phy_initialization:1;
 
 	/* The next flag is a stopgap, to be removed when all the HCDs
 	 * support the new root-hub polling mechanism. */
Signature · 1 line
2.16.1
#3 lore raw

PATCH v10 2/8 usb: add a flag to skip PHY initialization to struct usb_hcd

Quoted 161 lines · Martin Blumenstingl

The USB HCD core driver parses the device-tree node for "phys" and "usb-phys" properties. It also manages the power state of these PHYs automatically.
However, drivers may opt-out of this behavior by setting "phy" or "usb_phy" in struct usb_hcd to a non-null value. An example where this is required is the "Qualcomm USB2 controller", implemented by the chipidea driver. The hardware requires that the PHY is only powered on after the "reset completed" event from the controller is received.

A follow-up patch will allow the USB HCD core driver to manage more than one PHY. Add a new "skip_phy_initialization" bitflag to struct usb_hcd so drivers can opt-out of any PHY management provided by the USB HCD core driver.

This also updates the existing drivers so they use the new flag if they want to opt out of the PHY management provided by the USB HCD core driver. This means that for these drivers the new "multiple PHY" handling (which will be added in a follow-up patch) will be disabled as well.

---
 drivers/usb/chipidea/host.c      | 6 ++----
 drivers/usb/core/hcd.c           | 4 ++--
 drivers/usb/host/ehci-fsl.c      | 2 ++
 drivers/usb/host/ehci-platform.c | 4 ++--
 drivers/usb/host/ehci-tegra.c    | 1 +
 drivers/usb/host/ohci-omap.c     | 1 +
 drivers/usb/host/ohci-platform.c | 4 ++--
 drivers/usb/host/xhci-plat.c     | 1 +
 include/linux/usb/hcd.h          | 6 ++++++
 9 files changed, 19 insertions(+), 10 deletions(-)
drivers/usb/chipidea/host.c +2 -4
index 19d60ed7e41f..af45aa3222b5 100644
@@ -124,10 +124,8 @@ static int host_start(struct ci_hdrc *ci)
 
 	hcd->power_budget = ci->platdata->power_budget;
 	hcd->tpl_support = ci->platdata->tpl_support;
-	if (ci->phy)
-		hcd->phy = ci->phy;
-	else
-		hcd->usb_phy = ci->usb_phy;
+	if (ci->phy || ci->usb_phy)
+		hcd->skip_phy_initialization = 1;
 
 	ehci = hcd_to_ehci(hcd);
 	ehci->caps = ci->hw_bank.cap;
drivers/usb/core/hcd.c +2 -2
index fc32391a34d5..f2307470a31e 100644
@@ -2727,7 +2727,7 @@ int usb_add_hcd(struct usb_hcd *hcd,
 	int retval;
 	struct usb_device *rhdev;
 
-	if (IS_ENABLED(CONFIG_USB_PHY) && !hcd->usb_phy) {
+	if (IS_ENABLED(CONFIG_USB_PHY) && !hcd->skip_phy_initialization) {
 		struct usb_phy *phy = usb_get_phy_dev(hcd->self.sysdev, 0);
 
 		if (IS_ERR(phy)) {
@@ -2745,7 +2745,7 @@ int usb_add_hcd(struct usb_hcd *hcd,
 		}
 	}
 
-	if (IS_ENABLED(CONFIG_GENERIC_PHY) && !hcd->phy) {
+	if (IS_ENABLED(CONFIG_GENERIC_PHY) && !hcd->skip_phy_initialization) {
 		struct phy *phy = phy_get(hcd->self.sysdev, "usb");
 
 		if (IS_ERR(phy)) {
drivers/usb/host/ehci-fsl.c +2 -0
index c5094cb88cd5..0a9fd2022acf 100644
@@ -155,6 +155,8 @@ static int fsl_ehci_drv_probe(struct platform_device *pdev)
 			retval = -ENODEV;
 			goto err2;
 		}
+
+		hcd->skip_phy_initialization = 1;
 	}
 #endif
 	return retval;
drivers/usb/host/ehci-platform.c +2 -2
index b065a960adc2..b91eea8c73ae 100644
@@ -219,9 +219,9 @@ static int ehci_platform_probe(struct platform_device *dev)
 			if (IS_ERR(priv->phys[phy_num])) {
 				err = PTR_ERR(priv->phys[phy_num]);
 					goto err_put_hcd;
-			} else if (!hcd->phy) {
+			} else {
 				/* Avoiding phy_get() in usb_add_hcd() */
-				hcd->phy = priv->phys[phy_num];
+				hcd->skip_phy_initialization = 1;
 			}
 		}
 
drivers/usb/host/ehci-tegra.c +1 -0
index c809f7d2f08f..a6f4389f7e88 100644
@@ -461,6 +461,7 @@ static int tegra_ehci_probe(struct platform_device *pdev)
 		goto cleanup_clk_en;
 	}
 	hcd->usb_phy = u_phy;
+	hcd->skip_phy_initialization = 1;
 
 	tegra->needs_double_reset = of_property_read_bool(pdev->dev.of_node,
 		"nvidia,needs-double-reset");
drivers/usb/host/ohci-omap.c +1 -0
index 0201c49bc4fc..d8d35d456456 100644
@@ -230,6 +230,7 @@ static int ohci_omap_reset(struct usb_hcd *hcd)
 		} else {
 			return -EPROBE_DEFER;
 		}
+		hcd->skip_phy_initialization = 1;
 		ohci->start_hnp = start_hnp;
 	}
 #endif
drivers/usb/host/ohci-platform.c +2 -2
index 1e6c954f4b3f..62ef36a9333f 100644
@@ -186,9 +186,9 @@ static int ohci_platform_probe(struct platform_device *dev)
 			if (IS_ERR(priv->phys[phy_num])) {
 				err = PTR_ERR(priv->phys[phy_num]);
 				goto err_put_hcd;
-			} else if (!hcd->phy) {
+			} else {
 				/* Avoiding phy_get() in usb_add_hcd() */
-				hcd->phy = priv->phys[phy_num];
+				hcd->skip_phy_initialization = 1;
 			}
 		}
 
drivers/usb/host/xhci-plat.c +1 -0
index 6f038306c14d..6700e5ee82ad 100644
@@ -284,6 +284,7 @@ static int xhci_plat_probe(struct platform_device *pdev)
 		ret = usb_phy_init(hcd->usb_phy);
 		if (ret)
 			goto put_usb3_hcd;
+		hcd->skip_phy_initialization = 1;
 	}
 
 	ret = usb_add_hcd(hcd, irq, IRQF_SHARED);
include/linux/usb/hcd.h +0 -0
index 176900528822..693502c84c04 100644
@@ -151,6 +151,12 @@ struct usb_hcd {
 	unsigned		msix_enabled:1;	/* driver has MSI-X enabled? */
 	unsigned		msi_enabled:1;	/* driver has MSI enabled? */

include/linux/usb/hcd.h struct usb_hcd { · line 162

Quoted 10 lines
include/linux/usb/hcd.h +6 -0
 	unsigned		remove_phy:1;	/* auto-remove USB phy */
+	/*
+	 * do not manage the PHY state in the HCD core, instead let the driver
+	 * handle this (for example if the PHY can only be turned on after a
+	 * specific event)
+	 */
+	unsigned		skip_phy_initialization:1;
 
 	/* The next flag is a stopgap, to be removed when all the HCDs
 	 * support the new root-hub polling mechanism. */

For chipidea part:

Signature · 2 lines
Best Regards,
Peter Chen
author reviewer
#4 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

In reply to Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)		+= of.o
 usbcore-$(CONFIG_USB_PCI)		+= hcd-pci.o
drivers/usb/core/phy.c new file +158 -0

158 lines not on this page · raw message

drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);
Signature · 1 line
2.16.1
#5 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

+some TI folks

Hi Martin,

Quoted 19 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 4 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

Quoted 101 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)		+= of.o
 usbcore-$(CONFIG_USB_PCI)		+= hcd-pci.o
drivers/usb/core/phy.c new file +74 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+	struct phy		*phy;
+	struct list_head	list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+	struct usb_phy_roothub *roothub_entry;
+
+	roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+	if (!roothub_entry)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&roothub_entry->list);
+
+	return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+				   struct list_head *list)
+{
+	struct usb_phy_roothub *roothub_entry;
+	struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+	if (IS_ERR_OR_NULL(phy)) {
+		if (!phy || PTR_ERR(phy) == -ENODEV)
+			return 0;
+		else
+			return PTR_ERR(phy);
+	}
+
+	roothub_entry = usb_phy_roothub_alloc(dev);
+	if (IS_ERR(roothub_entry))
+		return PTR_ERR(roothub_entry);
+
+	roothub_entry->phy = phy;
+
+	list_add_tail(&roothub_entry->list, list);
+
+	return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+	struct usb_phy_roothub *phy_roothub;
+	struct usb_phy_roothub *roothub_entry;
+	struct list_head *head;
+	int i, num_phys, err;
+
+	num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+					      "#phy-cells");
+	if (num_phys <= 0)
+		return NULL;
+
+	phy_roothub = usb_phy_roothub_alloc(dev);
+	if (IS_ERR(phy_roothub))
+		return phy_roothub;
+

drivers/usb/core/phy.c line 84

Quoted 10 lines
drivers/usb/core/phy.c new file +10 -0
+	for (i = 0; i < num_phys; i++) {
+		err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+		if (err)
+			goto err_out;
+	}
+
+	head = &phy_roothub->list;
+
+	list_for_each_entry(roothub_entry, head, list) {
+		err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 21 lines
drivers/usb/core/phy.c new file +21 -0
+		if (err)
+			goto err_exit_phys;
+	}
+
+	return phy_roothub;
+
+err_exit_phys:
+	list_for_each_entry_continue_reverse(roothub_entry, head, list)
+		phy_exit(roothub_entry->phy);
+
+err_out:
+	return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+	struct usb_phy_roothub *roothub_entry;
+	struct list_head *head;
+	int err, ret = 0;
+

drivers/usb/core/phy.c line 115

Quoted 10 lines
drivers/usb/core/phy.c new file +10 -0
+	if (!phy_roothub)
+		return 0;
+
+	head = &phy_roothub->list;
+
+	list_for_each_entry(roothub_entry, head, list) {
+		err = phy_exit(roothub_entry->phy);
+		if (err)
+			ret = ret;
+	}

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 31 lines
drivers/usb/core/phy.c new file +31 -0
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+	struct usb_phy_roothub *roothub_entry;
+	struct list_head *head;
+	int err;
+
+	if (!phy_roothub)
+		return 0;
+
+	head = &phy_roothub->list;
+
+	list_for_each_entry(roothub_entry, head, list) {
+		err = phy_power_on(roothub_entry->phy);
+		if (err)
+			goto err_out;
+	}
+
+	return 0;
+
+err_out:
+	list_for_each_entry_continue_reverse(roothub_entry, head, list)
+		phy_power_off(roothub_entry->phy);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);

drivers/usb/core/phy.c line 156

Quoted 10 lines
drivers/usb/core/phy.c new file +10 -0
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+	struct usb_phy_roothub *roothub_entry;
+
+	if (!phy_roothub)
+		return;
+
+	list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+		phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 9 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +0 -0
new file mode 100644
index 000000000000..6fde59bfbff8

drivers/usb/core/phy.h line 7

Quoted 7 lines
drivers/usb/core/phy.h new file +7 -0
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);

The following patch fixes the issue for me. Let me know what you think and I can post it officially.

drivers/usb/core/phy.c +15 -28
index 09b7c43..23232d3 100644
@@ -59,8 +59,6 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
 {
 	struct usb_phy_roothub *phy_roothub;
-	struct usb_phy_roothub *roothub_entry;
-	struct list_head *head;
 	int i, num_phys, err;
 
 	num_phys = of_count_phandle_with_args(dev->of_node, "phys",
@@ -75,25 +73,10 @@ struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
 	for (i = 0; i < num_phys; i++) {
 		err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
 		if (err)
-			goto err_out;
-	}
-
-	head = &phy_roothub->list;
-
-	list_for_each_entry(roothub_entry, head, list) {
-		err = phy_init(roothub_entry->phy);
-		if (err)
-			goto err_exit_phys;
+			return ERR_PTR(err);
 	}
 
 	return phy_roothub;
-
-err_exit_phys:
-	list_for_each_entry_continue_reverse(roothub_entry, head, list)
-		phy_exit(roothub_entry->phy);
-
-err_out:
-	return ERR_PTR(err);
 }
 EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
 
@@ -106,13 +89,8 @@ int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
 	if (!phy_roothub)
 		return 0;
 
-	head = &phy_roothub->list;
-
-	list_for_each_entry(roothub_entry, head, list) {
-		err = phy_exit(roothub_entry->phy);
-		if (err)
-			ret = ret;
-	}
+	/* TODO: usb_phy_roothub_del_phy */
+	/* TODO: usb_phy_roothub_free */
 
 	return ret;
 }
@@ -130,16 +108,23 @@ int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
 	head = &phy_roothub->list;
 
 	list_for_each_entry(roothub_entry, head, list) {
-		err = phy_power_on(roothub_entry->phy);
+		err = phy_init(roothub_entry->phy);
 		if (err)
 			goto err_out;
+		err = phy_power_on(roothub_entry->phy);
+		if (err) {
+			phy_exit(roothub_entry->phy);
+			goto err_out;
+		}
 	}
 
 	return 0;
 
 err_out:
-	list_for_each_entry_continue_reverse(roothub_entry, head, list)
+	list_for_each_entry_continue_reverse(roothub_entry, head, list) {
 		phy_power_off(roothub_entry->phy);
+		phy_exit(roothub_entry->phy);
+	}
 
 	return err;
 }
@@ -152,7 +137,9 @@ void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
 	if (!phy_roothub)
 		return;
 
-	list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+	list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list) {
 		phy_power_off(roothub_entry->phy);
+		phy_exit(roothub_entry->phy);
+	}
 }
 EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
Signature · 4 lines
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
author reviewer
#6 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Hi Roger,

Quoted 24 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 19 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 7 lines
Quoted 4 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms
- fix this in the usb_phy_roothub code
- fix this in the PHY driver
- somewhere else

Quoted 10 lines
Quoted 10 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +2 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
Quoted 8 lines

drivers/usb/core/phy.h line 7

Quoted 6 lines
drivers/usb/core/phy.h new file +5 -0
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);

The following patch fixes the issue for me. Let me know what you think and I can post it officially.

Chunfeng: could you please test whether this patch breaks the suspend/resume cycle on your Mediatek platforms? see [0] where I used a similar approach with a much older version of the "initialize multiple PHYs per HCD" patch

Quoted 99 lines
drivers/usb/core/phy.c +15 -29
index 09b7c43..23232d3 100644
@@ -59,8 +59,6 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
 {
        struct usb_phy_roothub *phy_roothub;
-       struct usb_phy_roothub *roothub_entry;
-       struct list_head *head;
        int i, num_phys, err;
 
        num_phys = of_count_phandle_with_args(dev->of_node, "phys",
@@ -75,25 +73,10 @@ struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
        for (i = 0; i < num_phys; i++) {
                err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
                if (err)
-                       goto err_out;
-       }
-
-       head = &phy_roothub->list;
-
-       list_for_each_entry(roothub_entry, head, list) {
-               err = phy_init(roothub_entry->phy);
-               if (err)
-                       goto err_exit_phys;
+                       return ERR_PTR(err);
        }
 
        return phy_roothub;
-
-err_exit_phys:
-       list_for_each_entry_continue_reverse(roothub_entry, head, list)
-               phy_exit(roothub_entry->phy);
-
-err_out:
-       return ERR_PTR(err);
 }
 EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
 
@@ -106,13 +89,8 @@ int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
        if (!phy_roothub)
                return 0;
 
-       head = &phy_roothub->list;
-
-       list_for_each_entry(roothub_entry, head, list) {
-               err = phy_exit(roothub_entry->phy);
-               if (err)
-                       ret = ret;
-       }
+       /* TODO: usb_phy_roothub_del_phy */
+       /* TODO: usb_phy_roothub_free */
 
        return ret;
 }
@@ -130,16 +108,23 @@ int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
        head = &phy_roothub->list;
 
        list_for_each_entry(roothub_entry, head, list) {
-               err = phy_power_on(roothub_entry->phy);
+               err = phy_init(roothub_entry->phy);
                if (err)
                        goto err_out;
+               err = phy_power_on(roothub_entry->phy);
+               if (err) {
+                       phy_exit(roothub_entry->phy);
+                       goto err_out;
+               }
        }
 
        return 0;
 
 err_out:
-       list_for_each_entry_continue_reverse(roothub_entry, head, list)
+       list_for_each_entry_continue_reverse(roothub_entry, head, list) {
                phy_power_off(roothub_entry->phy);
+               phy_exit(roothub_entry->phy);
+       }
 
        return err;
 }
@@ -152,7 +137,9 @@ void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
        if (!phy_roothub)
                return;
 
-       list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+       list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list) {
                phy_power_off(roothub_entry->phy);
+               phy_exit(roothub_entry->phy);
+       }
 }
 EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
 
 
--

cheers,
-roger

Quoted 1 line

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

Regards
Martin

[0] http://lists.infradead.org/pipermail/linux-amlogic/2017-July/004374.html

maintainer
#7 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Hi,

Quoted 37 lines · Martin Blumenstingl

Hi Roger,

Quoted 24 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 19 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 7 lines
Quoted 4 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

[...]

Quoted 75 lines
Quoted 22 lines
Quoted 22 lines
(quoted diff) +22 -0
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

Review

Quoted 5 lines
(quoted diff) +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
(quoted diff) +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

Review

Quoted 6 lines
(quoted diff) +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 3 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

You are right

Quoted 73 lines
Quoted 40 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
(quoted diff) +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

Review

Quoted 5 lines
(quoted diff) +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms
- fix this in the usb_phy_roothub code
- fix this in the PHY driver
- somewhere else

Quoted 10 lines
Quoted 10 lines
(quoted diff) +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +2 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
Quoted 8 lines

drivers/usb/core/phy.h line 7

Quoted 6 lines
drivers/usb/core/phy.h new file +5 -0
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);

The following patch fixes the issue for me. Let me know what you think and I can post it officially.

Quoted 4 lines

Chunfeng: could you please test whether this patch breaks the suspend/resume cycle on your Mediatek platforms? see [0] where I used a similar approach with a much older version of the "initialize multiple PHYs per HCD" patch

If add Roger's one, it do disconnect plugged devices on MTK platforms, due to re-initialize u2 phys when resume,as you mentioned above

Quoted 108 lines
Quoted 100 lines
drivers/usb/core/phy.c +15 -29
index 09b7c43..23232d3 100644
@@ -59,8 +59,6 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
 {
        struct usb_phy_roothub *phy_roothub;
-       struct usb_phy_roothub *roothub_entry;
-       struct list_head *head;
        int i, num_phys, err;
 
        num_phys = of_count_phandle_with_args(dev->of_node, "phys",
@@ -75,25 +73,10 @@ struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
        for (i = 0; i < num_phys; i++) {
                err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
                if (err)
-                       goto err_out;
-       }
-
-       head = &phy_roothub->list;
-
-       list_for_each_entry(roothub_entry, head, list) {
-               err = phy_init(roothub_entry->phy);
-               if (err)
-                       goto err_exit_phys;
+                       return ERR_PTR(err);
        }
 
        return phy_roothub;
-
-err_exit_phys:
-       list_for_each_entry_continue_reverse(roothub_entry, head, list)
-               phy_exit(roothub_entry->phy);
-
-err_out:
-       return ERR_PTR(err);
 }
 EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
 
@@ -106,13 +89,8 @@ int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
        if (!phy_roothub)
                return 0;
 
-       head = &phy_roothub->list;
-
-       list_for_each_entry(roothub_entry, head, list) {
-               err = phy_exit(roothub_entry->phy);
-               if (err)
-                       ret = ret;
-       }
+       /* TODO: usb_phy_roothub_del_phy */
+       /* TODO: usb_phy_roothub_free */
 
        return ret;
 }
@@ -130,16 +108,23 @@ int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
        head = &phy_roothub->list;
 
        list_for_each_entry(roothub_entry, head, list) {
-               err = phy_power_on(roothub_entry->phy);
+               err = phy_init(roothub_entry->phy);
                if (err)
                        goto err_out;
+               err = phy_power_on(roothub_entry->phy);
+               if (err) {
+                       phy_exit(roothub_entry->phy);
+                       goto err_out;
+               }
        }
 
        return 0;
 
 err_out:
-       list_for_each_entry_continue_reverse(roothub_entry, head, list)
+       list_for_each_entry_continue_reverse(roothub_entry, head, list) {
                phy_power_off(roothub_entry->phy);
+               phy_exit(roothub_entry->phy);
+       }
 
        return err;
 }
@@ -152,7 +137,9 @@ void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
        if (!phy_roothub)
                return;
 
-       list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+       list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list) {
                phy_power_off(roothub_entry->phy);
+               phy_exit(roothub_entry->phy);
+       }
 }
 EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
 
 
--

cheers,
-roger

Quoted 1 line

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

Regards
Martin

[0] http://lists.infradead.org/pipermail/linux-amlogic/2017-July/004374.html

#8 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

In reply to Martin Blumenstingl

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

Quoted 1 line

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

Quoted 19 lines

- somewhere else

Quoted 17 lines
Quoted 16 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);

<snip>

Signature · 4 lines
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
author reviewer
#9 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 51 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Quoted 7 lines

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 3 lines
Quoted 1 line

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

@Roger: what do you think about this?
Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 20 lines
Quoted 19 lines

- somewhere else

Quoted 17 lines
Quoted 16 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);
Quoted 7 lines

<snip>

--
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

Regards
Martin

maintainer
#10 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Hi Martin,

Quoted 213 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 46 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 7 lines
Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

In fact, xhci-mtk driver always keeps link state across suspend/resume, no matter "wakeup-source" is marked or not, it's just used to enable/disable to send wakeup signal to SPM.

Quoted 60 lines
Quoted 3 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 8 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

I'm afraid it doesn't work.
If re-initialize phys when system resume, it will cause usb controller to be re-initialized too, but the driver doesn't support it currently.

Quoted 47 lines
Quoted 4 lines
Quoted 2 lines

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

@Roger: what do you think about this?
Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 20 lines
Quoted 19 lines

- somewhere else

Quoted 17 lines
Quoted 16 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);
Quoted 7 lines

<snip>

--
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

Regards
Martin

#11 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

In reply to Martin Blumenstingl
Quoted 280 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 8 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

That is about wakeup-source and not about whether remote wakeup should be enabled or not. i.e. wakeup-source can be present but wakeup can still be disabled by the user via sysfs.

Quoted 12 lines
Quoted 4 lines
Quoted 2 lines

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

Quoted 1 line

@Roger: what do you think about this?

Looks fine to me. However one thing is still not clear. How does the PHY driver decide if it needs to do a full power off or partial power off?
It doesn't know about the roothub's device_may_wakeup() status.

Quoted 23 lines

Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 20 lines
Quoted 19 lines

- somewhere else

Quoted 17 lines
Quoted 16 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);
Signature · 4 lines
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
#12 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Quoted 284 lines · Roger Quadros
Quoted 283 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 58 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

Quoted 7 lines
Quoted 5 lines

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

That is about wakeup-source and not about whether remote wakeup should be enabled or not.

correction. It indeed means enable/disable wakeup :).

Quoted 44 lines

i.e. wakeup-source can be present but wakeup can still be disabled by the user via sysfs.

Quoted 12 lines
Quoted 4 lines
Quoted 2 lines

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

Quoted 1 line

@Roger: what do you think about this?

Looks fine to me. However one thing is still not clear. How does the PHY driver decide if it needs to do a full power off or partial power off?
It doesn't know about the roothub's device_may_wakeup() status.

Quoted 23 lines

Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 20 lines
Quoted 19 lines

- somewhere else

Quoted 17 lines
Quoted 16 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);
Signature · 4 lines
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
#13 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

In reply to Martin Blumenstingl

Hi,

Quoted 289 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 51 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Quoted 7 lines

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 4 lines
Quoted 3 lines
Quoted 1 line

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:

Not in favor of adding explicit suspend/resume ops since PM framework already has those. I think we should let PHY drivers manage suspend/resume on its own (after creating the dependency between the controller device and PHY using device_link_add).

Thanks
Kishon

author reviewer
#14 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Hello Kishon,

Quoted 294 lines · Kishon Vijay Abraham I

Hi,

Quoted 291 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 51 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Quoted 7 lines

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 2 lines
Quoted 1 line

- fix this in the PHY driver

Quoted 7 lines
Quoted 2 lines
Quoted 1 line

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:

Not in favor of adding explicit suspend/resume ops since PM framework already has those. I think we should let PHY drivers manage suspend/resume on its own (after creating the dependency between the controller device and PHY using device_link_add).

even better if we can re-use some existing code!

the platform I am working on (64-bit Amlogic Meson GXL/GXM) does not support suspend/resume yet, so unfortunately I cannot test this. besides that I have zero experience with suspend/resume logic.

I'll try to read more about that topic, but I definitely need someone who could help testing if I have a patch ready
in any case: I think implementing this should not be done for v4.17 anymore - maybe we can find a small solution for v4.17 and switch to your proposed solution later (assuming nobody has arguments against that)

Regards
Martin

maintainer
#15 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Hi Martin,

Quoted 310 lines · Martin Blumenstingl

Hello Kishon,

Quoted 294 lines · Kishon Vijay Abraham I

Hi,

Quoted 291 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 51 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Quoted 7 lines

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 2 lines
Quoted 1 line

- fix this in the PHY driver

Quoted 7 lines
Quoted 2 lines
Quoted 1 line

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:

Not in favor of adding explicit suspend/resume ops since PM framework already has those. I think we should let PHY drivers manage suspend/resume on its own (after creating the dependency between the controller device and PHY using device_link_add).

even better if we can re-use some existing code!

the platform I am working on (64-bit Amlogic Meson GXL/GXM) does not support suspend/resume yet, so unfortunately I cannot test this. besides that I have zero experience with suspend/resume logic.

Quoted 7 lines

I'll try to read more about that topic, but I definitely need someone who could help testing if I have a patch ready
in any case: I think implementing this should not be done for v4.17 anymore - maybe we can find a small solution for v4.17 and switch to your proposed solution later (assuming nobody has arguments against that)

I'll test it.

Thanks

Quoted 3 lines

Regards
Martin

maintainer
#16 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

In reply to Martin Blumenstingl

Hi Martin & Roger:

Quoted 280 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 9 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Very sorry, I forgot that MTU3 & xHCI drivers always set them as wakeup devices by device_init_wakeup(dev, true),but not dependent on "wakeup-source" property, so maybe we can use device_can_wakeup() to decide whether call phy_exit()/init() or not.

Quoted 46 lines
Quoted 3 lines
Quoted 1 line

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

@Roger: what do you think about this?
Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 20 lines
Quoted 19 lines

- somewhere else

Quoted 17 lines
Quoted 16 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);
Quoted 7 lines

<snip>

--
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

Regards
Martin

author reviewer
#17 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Hi Roger, Hi Chunfeng,

Quoted 292 lines · Chunfeng Yun

Hi Martin & Roger:

Quoted 280 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 9 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 4 lines

Very sorry, I forgot that MTU3 & xHCI drivers always set them as wakeup devices by device_init_wakeup(dev, true),but not dependent on "wakeup-source" property, so maybe we can use device_can_wakeup() to decide whether call phy_exit()/init() or not.

the 64-bit Amlogic Meson GXL/GXM SoCs don't support suspend/resume yet, so I cannot test this
based on this suggestion I threw up two patches which are *compile tested only* based on Greg's usb-next branch
you can find them here: [0] (as well as attached to this mail)

@Chunfeng: can you please test this on one of your Mediatek SoCs? @Roger: can you please test this on a TI SoC?

(apologies in advance if these patches don't work)

please note that I won't have access to my computer until Saturday. if these patches need to be rewritten/replaced/etc. then feel free to send your own version to the list

Quoted 41 lines
Quoted 41 lines
Quoted 3 lines
Quoted 1 line

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

@Roger: what do you think about this?
Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 26 lines
Quoted 19 lines

- somewhere else

Quoted 17 lines
Quoted 16 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);

<snip>

--
cheers,
-roger

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

Regards
Martin

Regards
Martin

[0] https://github.com/xdarklight/linux/commits/usb-phy-roothub-suspend-rfc-v1
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-usb-core-split-usb_phy_roothub_-init-alloc.patch
Type: text/x-patch
Size: 5244 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-amlogic/attachments/20180320/57365d90/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0002-usb-core-use-phy_exit-during-suspend-if-wake-up-is-n.patch
Type: text/x-patch
Size: 4460 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-amlogic/attachments/20180320/57365d90/attachment-0001.bin>

#18 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Martin,

Quoted 305 lines · Martin Blumenstingl

Hi Roger, Hi Chunfeng,

Quoted 292 lines · Chunfeng Yun

Hi Martin & Roger:

Quoted 280 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 9 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 4 lines

Very sorry, I forgot that MTU3 & xHCI drivers always set them as wakeup devices by device_init_wakeup(dev, true),but not dependent on "wakeup-source" property, so maybe we can use device_can_wakeup() to decide whether call phy_exit()/init() or not.

the 64-bit Amlogic Meson GXL/GXM SoCs don't support suspend/resume yet, so I cannot test this
based on this suggestion I threw up two patches which are *compile tested only* based on Greg's usb-next branch
you can find them here: [0] (as well as attached to this mail)

Quoted 8 lines

@Chunfeng: can you please test this on one of your Mediatek SoCs? @Roger: can you please test this on a TI SoC?

(apologies in advance if these patches don't work)

please note that I won't have access to my computer until Saturday. if these patches need to be rewritten/replaced/etc. then feel free to send your own version to the list

Had to do the following to build

drivers/usb/core/hcd.c +3 -3
index 6d4a419..2884607 100644
@@ -2262,7 +2262,7 @@ int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)
 		hcd->state = HC_STATE_SUSPENDED;
 
 		if (!PMSG_IS_AUTO(msg))
-			usb_phy_roothub_suspend(hcd->phy_roothub);
+			usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);
 
 		/* Did we race with a root-hub wakeup event? */
 		if (rhdev->do_remote_wakeup) {
@@ -2302,7 +2302,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 	}
 
 	if (!PMSG_IS_AUTO(msg)) {
-		status = usb_phy_roothub_resume(hcd->phy_roothub);
+		status = usb_phy_roothub_resume(&rhdev->dev, hcd->phy_roothub);
 		if (status)
 			return status;
 	}
@@ -2344,7 +2344,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 		}
 	} else {
 		hcd->state = old_state;
-		usb_phy_roothub_suspend(hcd->phy_roothub);
+		usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);
 		dev_dbg(&rhdev->dev, "bus %s fail, err %d\n",
 				"resume", status);
 		if (status != -ESHUTDOWN)

And the following to fix runtime issues

drivers/usb/core/phy.c +4 -4
index 2eca371..8598906 100644
@@ -32,9 +32,9 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 			return PTR_ERR(phy);
 	}
 
-	roothub_entry = usb_phy_roothub_alloc(dev);
-	if (IS_ERR(roothub_entry))
-		return PTR_ERR(roothub_entry);
+	roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+	if (!roothub_entry)
+		return -ENOMEM;
 
 	roothub_entry->phy = phy;
 
@@ -162,7 +162,7 @@ int usb_phy_roothub_suspend(struct device *dev,
 	usb_phy_roothub_power_off(phy_roothub);
 
 	/* keep the PHYs initialized so the device can wake up the system */
-	if (device_can_wakeup(dev))
+	if (device_may_wakeup(dev))
 		return 0;
 
 	return usb_phy_roothub_exit(phy_roothub);

Here are my obervations
- if wakeup is disabled it works fine as expected, phy_exit() is called and I'm able to reach low power states.

- if wakeup is enabled (/sys/bus/usb/device/usb2/power/wakeup), then hcd_bus_suspend() is never called and so phy_power_off won't be called either.

This means that the device_may_wakeup() check is redundant. Sorry for suggesting this.

This also means that wakeup is not enabled on Chunfeng's platform. Chunfeng, can you confirm this? What does /sys/bus/usb/device/usb<?>/power/wakeup say?

Chunfeng sent a patch [1] to set shared_hcd->skip_phy_initialization for the mtk platform. Chunfeng, why did you need this patch?

[1] - https://patchwork.kernel.org/patch/10298641/

What options do we have now to fix the original issue?

Chungfeng, can you set skip_phy_initialization for the hcd that has PHY's linked to it? Then usb_phy_roothub_*() driver will be a no-op for you. And we could revert to the original approach of doing phy_power_off() as well as phy_exit() during suspend.

Alternatively, Martin, how about not relying on skip_phy_initialization flag but having default behaviour of no-op for the usb_phy_roothub_*() driver. As platforms migrate to it they can set a new flag in hcd to use it?

Quoted 40 lines
Quoted 28 lines
Quoted 28 lines
Quoted 4 lines
Quoted 2 lines

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

@Roger: what do you think about this?
Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 12 lines
Quoted 12 lines

- somewhere else

Quoted 10 lines
Quoted 10 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +2 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
Quoted 7 lines

drivers/usb/core/phy.h line 7

Quoted 6 lines
drivers/usb/core/phy.h new file +5 -0
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);

[0] https://github.com/xdarklight/linux/commits/usb-phy-roothub-suspend-rfc-v1

Signature · 4 lines
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
maintainer
#19 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Hi,

Quoted 324 lines · Roger Quadros

Martin,

Quoted 305 lines · Martin Blumenstingl

Hi Roger, Hi Chunfeng,

Quoted 292 lines · Chunfeng Yun

Hi Martin & Roger:

Quoted 280 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 9 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 4 lines

Very sorry, I forgot that MTU3 & xHCI drivers always set them as wakeup devices by device_init_wakeup(dev, true),but not dependent on "wakeup-source" property, so maybe we can use device_can_wakeup() to decide whether call phy_exit()/init() or not.

the 64-bit Amlogic Meson GXL/GXM SoCs don't support suspend/resume yet, so I cannot test this
based on this suggestion I threw up two patches which are *compile tested only* based on Greg's usb-next branch
you can find them here: [0] (as well as attached to this mail)

Quoted 8 lines

@Chunfeng: can you please test this on one of your Mediatek SoCs? @Roger: can you please test this on a TI SoC?

(apologies in advance if these patches don't work)

please note that I won't have access to my computer until Saturday. if these patches need to be rewritten/replaced/etc. then feel free to send your own version to the list

Had to do the following to build

drivers/usb/core/hcd.c +0 -0
index 6d4a419..2884607 100644

drivers/usb/core/hcd.c int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg) · line 2265

Quoted 5 lines
drivers/usb/core/hcd.c +1 -1
@@ -2262,7 +2262,7 @@ int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)
 		hcd->state = HC_STATE_SUSPENDED;
 
 		if (!PMSG_IS_AUTO(msg))
-			usb_phy_roothub_suspend(hcd->phy_roothub);
+			usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);

Try to use hcd->self.controller instead of &rhdev->dev;

drivers/usb/core/hcd.c int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg) · line 2305

Quoted 9 lines
drivers/usb/core/hcd.c +1 -1
 		/* Did we race with a root-hub wakeup event? */
 		if (rhdev->do_remote_wakeup) {
@@ -2302,7 +2302,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 	}
 
 	if (!PMSG_IS_AUTO(msg)) {
-		status = usb_phy_roothub_resume(hcd->phy_roothub);
+		status = usb_phy_roothub_resume(&rhdev->dev, hcd->phy_roothub);

ditto

drivers/usb/core/hcd.c int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg) · line 2347

Quoted 9 lines
drivers/usb/core/hcd.c +1 -1
 		if (status)
 			return status;
 	}
@@ -2344,7 +2344,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 		}
 	} else {
 		hcd->state = old_state;
-		usb_phy_roothub_suspend(hcd->phy_roothub);
+		usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);

ditto

Quoted 25 lines
drivers/usb/core/hcd.c +0 -0
 		dev_dbg(&rhdev->dev, "bus %s fail, err %d\n",
 				"resume", status);
 		if (status != -ESHUTDOWN)

And the following to fix runtime issues

drivers/usb/core/phy.c +3 -3
index 2eca371..8598906 100644
@@ -32,9 +32,9 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 			return PTR_ERR(phy);
 	}
 
-	roothub_entry = usb_phy_roothub_alloc(dev);
-	if (IS_ERR(roothub_entry))
-		return PTR_ERR(roothub_entry);
+	roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+	if (!roothub_entry)
+		return -ENOMEM;
 
 	roothub_entry->phy = phy;
 

drivers/usb/core/phy.c int usb_phy_roothub_suspend(struct device *dev, · line 165

Quoted 5 lines
drivers/usb/core/phy.c +1 -1
@@ -162,7 +162,7 @@ int usb_phy_roothub_suspend(struct device *dev,
 	usb_phy_roothub_power_off(phy_roothub);
 
 	/* keep the PHYs initialized so the device can wake up the system */
-	if (device_can_wakeup(dev))
+	if (device_may_wakeup(dev))

It's ok

Quoted 5 lines
drivers/usb/core/phy.c +0 -0
 		return 0;
 
 	return usb_phy_roothub_exit(phy_roothub);
Quoted 8 lines

Here are my obervations
- if wakeup is disabled it works fine as expected, phy_exit() is called and I'm able to reach low power states.

- if wakeup is enabled (/sys/bus/usb/device/usb2/power/wakeup), then hcd_bus_suspend() is never called and so phy_power_off won't be called either.

This means that the device_may_wakeup() check is redundant. Sorry for suggesting this.

You maybe use a wrong device.

Quoted 3 lines

This also means that wakeup is not enabled on Chunfeng's platform. Chunfeng, can you confirm this? What does /sys/bus/usb/device/usb<?>/power/wakeup say?

They are also "disabled" state;
I think we should use wakeup state of hcd controller, but not usb_device's, please test it, thanks.

Quoted 58 lines

Chunfeng sent a patch [1] to set shared_hcd->skip_phy_initialization for the mtk platform. Chunfeng, why did you need this patch?

[1] - https://patchwork.kernel.org/patch/10298641/

What options do we have now to fix the original issue?

Chungfeng, can you set skip_phy_initialization for the hcd that has PHY's linked to it? Then usb_phy_roothub_*() driver will be a no-op for you. And we could revert to the original approach of doing phy_power_off() as well as phy_exit() during suspend.

Alternatively, Martin, how about not relying on skip_phy_initialization flag but having default behaviour of no-op for the usb_phy_roothub_*() driver. As platforms migrate to it they can set a new flag in hcd to use it?

Quoted 40 lines
Quoted 28 lines
Quoted 28 lines
Quoted 4 lines
Quoted 2 lines

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

@Roger: what do you think about this?
Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 12 lines
Quoted 12 lines

- somewhere else

Quoted 10 lines
Quoted 10 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +2 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
Quoted 7 lines

drivers/usb/core/phy.h line 7

Quoted 6 lines
drivers/usb/core/phy.h new file +5 -0
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);

[0] https://github.com/xdarklight/linux/commits/usb-phy-roothub-suspend-rfc-v1

#20 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Quoted 328 lines · Chunfeng Yun

Hi,

Quoted 326 lines · Roger Quadros

Martin,

Quoted 305 lines · Martin Blumenstingl

Hi Roger, Hi Chunfeng,

Quoted 292 lines · Chunfeng Yun

Hi Martin & Roger:

Quoted 280 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 9 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 4 lines

Very sorry, I forgot that MTU3 & xHCI drivers always set them as wakeup devices by device_init_wakeup(dev, true),but not dependent on "wakeup-source" property, so maybe we can use device_can_wakeup() to decide whether call phy_exit()/init() or not.

the 64-bit Amlogic Meson GXL/GXM SoCs don't support suspend/resume yet, so I cannot test this
based on this suggestion I threw up two patches which are *compile tested only* based on Greg's usb-next branch
you can find them here: [0] (as well as attached to this mail)

Quoted 8 lines

@Chunfeng: can you please test this on one of your Mediatek SoCs? @Roger: can you please test this on a TI SoC?

(apologies in advance if these patches don't work)

please note that I won't have access to my computer until Saturday. if these patches need to be rewritten/replaced/etc. then feel free to send your own version to the list

Had to do the following to build

drivers/usb/core/hcd.c +0 -0
index 6d4a419..2884607 100644
@@ -2262,7 +2262,7 @@ int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)
 		hcd->state = HC_STATE_SUSPENDED;
 
Quoted 5 lines

drivers/usb/core/hcd.c int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg) · line 2265

Quoted 3 lines
drivers/usb/core/hcd.c +1 -1
 		if (!PMSG_IS_AUTO(msg))
-			usb_phy_roothub_suspend(hcd->phy_roothub);
+			usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);

Try to use hcd->self.controller instead of &rhdev->dev;

Actually it should be hcd->self.sysdev.

Quoted 45 lines

drivers/usb/core/hcd.c int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg) · line 2305

Quoted 10 lines
drivers/usb/core/hcd.c +1 -1
 		/* Did we race with a root-hub wakeup event? */
 		if (rhdev->do_remote_wakeup) {
@@ -2302,7 +2302,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 	}
 
 	if (!PMSG_IS_AUTO(msg)) {
-		status = usb_phy_roothub_resume(hcd->phy_roothub);
+		status = usb_phy_roothub_resume(&rhdev->dev, hcd->phy_roothub);

ditto

drivers/usb/core/hcd.c int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg) · line 2347

Quoted 9 lines
drivers/usb/core/hcd.c +1 -1
 		if (status)
 			return status;
 	}
@@ -2344,7 +2344,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 		}
 	} else {
 		hcd->state = old_state;
-		usb_phy_roothub_suspend(hcd->phy_roothub);
+		usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);

ditto

Quoted 24 lines
drivers/usb/core/hcd.c +0 -0
 		dev_dbg(&rhdev->dev, "bus %s fail, err %d\n",
 				"resume", status);
 		if (status != -ESHUTDOWN)

And the following to fix runtime issues

drivers/usb/core/phy.c +3 -3
index 2eca371..8598906 100644
@@ -32,9 +32,9 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 			return PTR_ERR(phy);
 	}
 
-	roothub_entry = usb_phy_roothub_alloc(dev);
-	if (IS_ERR(roothub_entry))
-		return PTR_ERR(roothub_entry);
+	roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+	if (!roothub_entry)
+		return -ENOMEM;
 
 	roothub_entry->phy = phy;
 
Quoted 7 lines

drivers/usb/core/phy.c int usb_phy_roothub_suspend(struct device *dev, · line 165

Quoted 6 lines
drivers/usb/core/phy.c +1 -1
@@ -162,7 +162,7 @@ int usb_phy_roothub_suspend(struct device *dev,
 	usb_phy_roothub_power_off(phy_roothub);
 
 	/* keep the PHYs initialized so the device can wake up the system */
-	if (device_can_wakeup(dev))
+	if (device_may_wakeup(dev))

It's ok

I had to additionally fix usb_phy_roothub_resume() from
if (device_can_wakeup(dev))
to
if (!device_may_wakeup(dev))

Quoted 9 lines
Quoted 9 lines
drivers/usb/core/phy.c +0 -0
 		return 0;
 
 	return usb_phy_roothub_exit(phy_roothub);

Here are my obervations
- if wakeup is disabled it works fine as expected, phy_exit() is called and I'm able to reach low power states.

Quoted 5 lines
Quoted 4 lines

- if wakeup is enabled (/sys/bus/usb/device/usb2/power/wakeup), then hcd_bus_suspend() is never called and so phy_power_off won't be called either.

This means that the device_may_wakeup() check is redundant. Sorry for suggesting this.

You maybe use a wrong device.

Yes, after using the correct device I don't see the problem.

Can you please try the below patch on top of the 2 Patches that Martin sent and the new patch you sent? Once you confirm it works for you I can send the 2 patches officially.

drivers/usb/core/hcd.c +3 -3
index 6d4a419..04cc453 100644
@@ -2262,7 +2262,7 @@ int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)
 		hcd->state = HC_STATE_SUSPENDED;
 
 		if (!PMSG_IS_AUTO(msg))
-			usb_phy_roothub_suspend(hcd->phy_roothub);
+			usb_phy_roothub_suspend(hcd->self.sysdev, hcd->phy_roothub);
 
 		/* Did we race with a root-hub wakeup event? */
 		if (rhdev->do_remote_wakeup) {
@@ -2302,7 +2302,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 	}
 
 	if (!PMSG_IS_AUTO(msg)) {
-		status = usb_phy_roothub_resume(hcd->phy_roothub);
+		status = usb_phy_roothub_resume(hcd->self.sysdev, hcd->phy_roothub);
 		if (status)
 			return status;
 	}
@@ -2344,7 +2344,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 		}
 	} else {
 		hcd->state = old_state;
-		usb_phy_roothub_suspend(hcd->phy_roothub);
+		usb_phy_roothub_suspend(hcd->self.sysdev, hcd->phy_roothub);
 		dev_dbg(&rhdev->dev, "bus %s fail, err %d\n",
 				"resume", status);
 		if (status != -ESHUTDOWN)
drivers/usb/core/phy.c +5 -5
index 2eca371..25fe729 100644
@@ -32,9 +32,9 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 			return PTR_ERR(phy);
 	}
 
-	roothub_entry = usb_phy_roothub_alloc(dev);
-	if (IS_ERR(roothub_entry))
-		return PTR_ERR(roothub_entry);
+	roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+	if (!roothub_entry)
+		return -ENOMEM;
 
 	roothub_entry->phy = phy;
 
@@ -162,7 +162,7 @@ int usb_phy_roothub_suspend(struct device *dev,
 	usb_phy_roothub_power_off(phy_roothub);
 
 	/* keep the PHYs initialized so the device can wake up the system */
-	if (device_can_wakeup(dev))
+	if (device_may_wakeup(dev))
 		return 0;
 
 	return usb_phy_roothub_exit(phy_roothub);
@@ -175,7 +175,7 @@ int usb_phy_roothub_resume(struct device *dev,
 	int err;
 
 	/* if the device can't wake up the system _exit was called */
-	if (device_can_wakeup(dev)) {
+	if (!device_may_wakeup(dev)) {
 		err = usb_phy_roothub_init(phy_roothub);
 		if (err)
 			return err;
Signature · 4 lines
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
maintainer
#21 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

Quoted 329 lines · Roger Quadros
Quoted 328 lines · Chunfeng Yun

Hi,

Quoted 326 lines · Roger Quadros

Martin,

Quoted 305 lines · Martin Blumenstingl

Hi Roger, Hi Chunfeng,

Quoted 292 lines · Chunfeng Yun

Hi Martin & Roger:

Quoted 280 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 9 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 4 lines

Very sorry, I forgot that MTU3 & xHCI drivers always set them as wakeup devices by device_init_wakeup(dev, true),but not dependent on "wakeup-source" property, so maybe we can use device_can_wakeup() to decide whether call phy_exit()/init() or not.

the 64-bit Amlogic Meson GXL/GXM SoCs don't support suspend/resume yet, so I cannot test this
based on this suggestion I threw up two patches which are *compile tested only* based on Greg's usb-next branch
you can find them here: [0] (as well as attached to this mail)

Quoted 8 lines

@Chunfeng: can you please test this on one of your Mediatek SoCs? @Roger: can you please test this on a TI SoC?

(apologies in advance if these patches don't work)

please note that I won't have access to my computer until Saturday. if these patches need to be rewritten/replaced/etc. then feel free to send your own version to the list

Had to do the following to build

drivers/usb/core/hcd.c +0 -0
index 6d4a419..2884607 100644
@@ -2262,7 +2262,7 @@ int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)
 		hcd->state = HC_STATE_SUSPENDED;
 
Quoted 7 lines
Quoted 5 lines

drivers/usb/core/hcd.c int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)

Quoted 3 lines
drivers/usb/core/hcd.c +1 -1
 		if (!PMSG_IS_AUTO(msg))
-			usb_phy_roothub_suspend(hcd->phy_roothub);
+			usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);

Try to use hcd->self.controller instead of &rhdev->dev;

Actually it should be hcd->self.sysdev.

I find that xhci-plat.c enables wakeup by hcd->self.controller which is different from hcd->self.sysdev sometimes.

And hcd->self.controller is the same as hcd->self.sysdev on MTK's platform.

Quoted 70 lines
Quoted 45 lines

drivers/usb/core/hcd.c int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg) · line 2305

Quoted 10 lines
drivers/usb/core/hcd.c +1 -1
 		/* Did we race with a root-hub wakeup event? */
 		if (rhdev->do_remote_wakeup) {
@@ -2302,7 +2302,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 	}
 
 	if (!PMSG_IS_AUTO(msg)) {
-		status = usb_phy_roothub_resume(hcd->phy_roothub);
+		status = usb_phy_roothub_resume(&rhdev->dev, hcd->phy_roothub);

ditto

drivers/usb/core/hcd.c int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg) · line 2347

Quoted 9 lines
drivers/usb/core/hcd.c +1 -1
 		if (status)
 			return status;
 	}
@@ -2344,7 +2344,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 		}
 	} else {
 		hcd->state = old_state;
-		usb_phy_roothub_suspend(hcd->phy_roothub);
+		usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);

ditto

Quoted 24 lines
drivers/usb/core/hcd.c +0 -0
 		dev_dbg(&rhdev->dev, "bus %s fail, err %d\n",
 				"resume", status);
 		if (status != -ESHUTDOWN)

And the following to fix runtime issues

drivers/usb/core/phy.c +3 -3
index 2eca371..8598906 100644
@@ -32,9 +32,9 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 			return PTR_ERR(phy);
 	}
 
-	roothub_entry = usb_phy_roothub_alloc(dev);
-	if (IS_ERR(roothub_entry))
-		return PTR_ERR(roothub_entry);
+	roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+	if (!roothub_entry)
+		return -ENOMEM;
 
 	roothub_entry->phy = phy;
 
Quoted 7 lines

drivers/usb/core/phy.c int usb_phy_roothub_suspend(struct device *dev, · line 165

Quoted 6 lines
drivers/usb/core/phy.c +1 -1
@@ -162,7 +162,7 @@ int usb_phy_roothub_suspend(struct device *dev,
 	usb_phy_roothub_power_off(phy_roothub);
 
 	/* keep the PHYs initialized so the device can wake up the system */
-	if (device_can_wakeup(dev))
+	if (device_may_wakeup(dev))

It's ok

I had to additionally fix usb_phy_roothub_resume() from
if (device_can_wakeup(dev))
to
if (!device_may_wakeup(dev))

Quoted 12 lines
Quoted 12 lines
drivers/usb/core/phy.c +0 -0
 		return 0;
 
 	return usb_phy_roothub_exit(phy_roothub);

Here are my obervations
- if wakeup is disabled it works fine as expected, phy_exit() is called and I'm able to reach low power states.

- if wakeup is enabled (/sys/bus/usb/device/usb2/power/wakeup), then hcd_bus_suspend() is never called and so phy_power_off won't be called either.

Quoted 7 lines
Quoted 2 lines
Quoted 1 line

This means that the device_may_wakeup() check is redundant. Sorry for suggesting this.

You maybe use a wrong device.

Yes, after using the correct device I don't see the problem.

Can you please try the below patch on top of the 2 Patches that Martin sent and the new patch you sent? Once you confirm it works for you I can send the 2 patches officially.

It works on MTK's platform, I'll test it again when official patches are sent out.

Thanks a lot

Quoted 67 lines
drivers/usb/core/hcd.c +3 -3
index 6d4a419..04cc453 100644
@@ -2262,7 +2262,7 @@ int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)
 		hcd->state = HC_STATE_SUSPENDED;
 
 		if (!PMSG_IS_AUTO(msg))
-			usb_phy_roothub_suspend(hcd->phy_roothub);
+			usb_phy_roothub_suspend(hcd->self.sysdev, hcd->phy_roothub);
 
 		/* Did we race with a root-hub wakeup event? */
 		if (rhdev->do_remote_wakeup) {
@@ -2302,7 +2302,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 	}
 
 	if (!PMSG_IS_AUTO(msg)) {
-		status = usb_phy_roothub_resume(hcd->phy_roothub);
+		status = usb_phy_roothub_resume(hcd->self.sysdev, hcd->phy_roothub);
 		if (status)
 			return status;
 	}
@@ -2344,7 +2344,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 		}
 	} else {
 		hcd->state = old_state;
-		usb_phy_roothub_suspend(hcd->phy_roothub);
+		usb_phy_roothub_suspend(hcd->self.sysdev, hcd->phy_roothub);
 		dev_dbg(&rhdev->dev, "bus %s fail, err %d\n",
 				"resume", status);
 		if (status != -ESHUTDOWN)
drivers/usb/core/phy.c +5 -5
index 2eca371..25fe729 100644
@@ -32,9 +32,9 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 			return PTR_ERR(phy);
 	}
 
-	roothub_entry = usb_phy_roothub_alloc(dev);
-	if (IS_ERR(roothub_entry))
-		return PTR_ERR(roothub_entry);
+	roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+	if (!roothub_entry)
+		return -ENOMEM;
 
 	roothub_entry->phy = phy;
 
@@ -162,7 +162,7 @@ int usb_phy_roothub_suspend(struct device *dev,
 	usb_phy_roothub_power_off(phy_roothub);
 
 	/* keep the PHYs initialized so the device can wake up the system */
-	if (device_can_wakeup(dev))
+	if (device_may_wakeup(dev))
 		return 0;
 
 	return usb_phy_roothub_exit(phy_roothub);
@@ -175,7 +175,7 @@ int usb_phy_roothub_resume(struct device *dev,
 	int err;
 
 	/* if the device can't wake up the system _exit was called */
-	if (device_can_wakeup(dev)) {
+	if (!device_may_wakeup(dev)) {
 		err = usb_phy_roothub_init(phy_roothub);
 		if (err)
 			return err;
author reviewer
#22 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

In reply to Roger Quadros

Hi Roger,

Quoted 401 lines · Roger Quadros

Martin,

Quoted 305 lines · Martin Blumenstingl

Hi Roger, Hi Chunfeng,

Quoted 292 lines · Chunfeng Yun

Hi Martin & Roger:

Quoted 280 lines · Martin Blumenstingl

Hi Roger,

Quoted 35 lines · Roger Quadros

Hi,

Quoted 32 lines · Martin Blumenstingl

Hi Roger,

Quoted 29 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 23 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

based on your explanation and reading the TI PHY drivers I am assuming that the affected SoCs are using the "phy-omap-usb2" driver

yes and phy-ti-pipe3 as well i.e. "ti,phy-usb3" and "ti,omap-usb3"

I missed that, thanks

Quoted 167 lines
Quoted 114 lines
Quoted 106 lines
Quoted 106 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)         += of.o
 usbcore-$(CONFIG_USB_PCI)            += hcd-pci.o
drivers/usb/core/phy.c new file +79 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+     struct phy              *phy;
+     struct list_head        list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+     struct usb_phy_roothub *roothub_entry;
+
+     roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+     if (!roothub_entry)
+             return ERR_PTR(-ENOMEM);
+
+     INIT_LIST_HEAD(&roothub_entry->list);
+
+     return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+                                struct list_head *list)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+     if (IS_ERR_OR_NULL(phy)) {
+             if (!phy || PTR_ERR(phy) == -ENODEV)
+                     return 0;
+             else
+                     return PTR_ERR(phy);
+     }
+
+     roothub_entry = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(roothub_entry))
+             return PTR_ERR(roothub_entry);
+
+     roothub_entry->phy = phy;
+
+     list_add_tail(&roothub_entry->list, list);
+
+     return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+     struct usb_phy_roothub *phy_roothub;
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int i, num_phys, err;
+
+     num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+                                           "#phy-cells");
+     if (num_phys <= 0)
+             return NULL;
+
+     phy_roothub = usb_phy_roothub_alloc(dev);
+     if (IS_ERR(phy_roothub))
+             return phy_roothub;
+
+     for (i = 0; i < num_phys; i++) {
+             err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+             if (err)
+                     goto err_out;
+     }
Quoted 8 lines

drivers/usb/core/phy.c line 84

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 3 lines

do you mean that phy_init should be moved to usb_phy_roothub_power_on (just before phy_power_on is called within usb_phy_roothub_power_on)?

Yes.

Quoted 42 lines

an earlier version of my patch did exactly this, but it caused problems during a suspend/resume cycle on Mediatek devices Chunfeng Yun reported that issue here [0], quote from that mail for easier reading:
"In order to keep link state on mt8173, we just power off all phys(not exit) when system enter suspend, then power on them again (needn't init, otherwise device will be disconnected) when system resume, this can avoid re-enumerating device."

Quoted 25 lines
Quoted 25 lines
drivers/usb/core/phy.c new file +25 -0
+             if (err)
+                     goto err_exit_phys;
+     }
+
+     return phy_roothub;
+
+err_exit_phys:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_exit(roothub_entry->phy);
+
+err_out:
+     return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err, ret = 0;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
Quoted 8 lines

drivers/usb/core/phy.c line 115

Quoted 6 lines
drivers/usb/core/phy.c new file +6 -0
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_exit(roothub_entry->phy);
+             if (err)
+                     ret = ret;
+     }

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

Quoted 4 lines

if I understood Chunfeng Yun correctly this will require re-enumeration of the USB devices after a suspend/resume cycle on Mediatek SoCs

Quoted 5 lines

OK. I suppose that there are 2 cases
1) Mediatek's case: USB controller context retained across suspend/resume. Remote wakeup probably required.
No re-enumeration preferred after resume. phy_exit()/phy_init() must not be called during suspend/resume to keep PHY link active.

Documentation/devicetree/bindings/usb/mediatek,mtu3.txt indeed shows that the parent of the USB controller can be marked as "wakeup-source"

Quoted 2 lines

2) TI's case: low power important: USB context is lost, OK to re-enumerate. phy_exit()/phy_init() must be called during suspend/resume.

ACK

Quoted 55 lines
Quoted 47 lines
Quoted 39 lines

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+     return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+     struct list_head *head;
+     int err;
+
+     if (!phy_roothub)
+             return 0;
+
+     head = &phy_roothub->list;
+
+     list_for_each_entry(roothub_entry, head, list) {
+             err = phy_power_on(roothub_entry->phy);
+             if (err)
+                     goto err_out;
+     }
+
+     return 0;
+
+err_out:
+     list_for_each_entry_continue_reverse(roothub_entry, head, list)
+             phy_power_off(roothub_entry->phy);
+
+     return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+     struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+     if (!phy_roothub)
+             return;
+
+     list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+             phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Quoted 3 lines

I'm not sure where this problem should be solved:
- set skip_phy_initialization in struct usb_hcd to 1 for the affected TI platforms

Many TI platforms are affected, omap5*, dra7*, am43*

Quoted 1 line

- fix this in the usb_phy_roothub code

Quoted 9 lines
Quoted 3 lines

I'd vote for fixing it in the usb_phy_roothub code. How?
How about using the device_can_wakeup() to decide if we should call phy_exit()/init() or not? If the USB device can't wakeup the system there is no point in keeping it powered/clocked right?

@Chunfeng: can you confirm Roger's idea that we could call phy_exit if the controller is *NOT* marked as "wakeup-source"?
I am also not sure if it would work, since the "wakeup-source" property is defined on the USB controller's parent node in case of the Mediatek MTU3 (Mediatek USB3.0 DRD) controller

Quoted 4 lines

Very sorry, I forgot that MTU3 & xHCI drivers always set them as wakeup devices by device_init_wakeup(dev, true),but not dependent on "wakeup-source" property, so maybe we can use device_can_wakeup() to decide whether call phy_exit()/init() or not.

the 64-bit Amlogic Meson GXL/GXM SoCs don't support suspend/resume yet, so I cannot test this
based on this suggestion I threw up two patches which are *compile tested only* based on Greg's usb-next branch
you can find them here: [0] (as well as attached to this mail)

Quoted 8 lines

@Chunfeng: can you please test this on one of your Mediatek SoCs? @Roger: can you please test this on a TI SoC?

(apologies in advance if these patches don't work)

please note that I won't have access to my computer until Saturday. if these patches need to be rewritten/replaced/etc. then feel free to send your own version to the list

Had to do the following to build

drivers/usb/core/hcd.c +3 -3
index 6d4a419..2884607 100644
@@ -2262,7 +2262,7 @@ int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)
                hcd->state = HC_STATE_SUSPENDED;
 
                if (!PMSG_IS_AUTO(msg))
-                       usb_phy_roothub_suspend(hcd->phy_roothub);
+                       usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);
 
                /* Did we race with a root-hub wakeup event? */
                if (rhdev->do_remote_wakeup) {
@@ -2302,7 +2302,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
        }
 
        if (!PMSG_IS_AUTO(msg)) {
-               status = usb_phy_roothub_resume(hcd->phy_roothub);
+               status = usb_phy_roothub_resume(&rhdev->dev, hcd->phy_roothub);
                if (status)
                        return status;
        }
@@ -2344,7 +2344,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
                }
        } else {
                hcd->state = old_state;
-               usb_phy_roothub_suspend(hcd->phy_roothub);
+               usb_phy_roothub_suspend(&rhdev->dev, hcd->phy_roothub);
                dev_dbg(&rhdev->dev, "bus %s fail, err %d\n",
                                "resume", status);
                if (status != -ESHUTDOWN)

And the following to fix runtime issues

drivers/usb/core/phy.c +4 -4
index 2eca371..8598906 100644
@@ -32,9 +32,9 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
                        return PTR_ERR(phy);
        }
 
-       roothub_entry = usb_phy_roothub_alloc(dev);
-       if (IS_ERR(roothub_entry))
-               return PTR_ERR(roothub_entry);
+       roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+       if (!roothub_entry)
+               return -ENOMEM;
 
        roothub_entry->phy = phy;
 
@@ -162,7 +162,7 @@ int usb_phy_roothub_suspend(struct device *dev,
        usb_phy_roothub_power_off(phy_roothub);
 
        /* keep the PHYs initialized so the device can wake up the system */
-       if (device_can_wakeup(dev))
+       if (device_may_wakeup(dev))
                return 0;
 
        return usb_phy_roothub_exit(phy_roothub);

Here are my obervations
- if wakeup is disabled it works fine as expected, phy_exit() is called and I'm able to reach low power states.

- if wakeup is enabled (/sys/bus/usb/device/usb2/power/wakeup), then hcd_bus_suspend() is never called and so phy_power_off won't be called either.

This means that the device_may_wakeup() check is redundant. Sorry for suggesting this.

This also means that wakeup is not enabled on Chunfeng's platform. Chunfeng, can you confirm this? What does /sys/bus/usb/device/usb<?>/power/wakeup say?

Chunfeng sent a patch [1] to set shared_hcd->skip_phy_initialization for the mtk platform. Chunfeng, why did you need this patch?

[1] - https://patchwork.kernel.org/patch/10298641/

What options do we have now to fix the original issue?

Quoted 7 lines

Chungfeng, can you set skip_phy_initialization for the hcd that has PHY's linked to it? Then usb_phy_roothub_*() driver will be a no-op for you. And we could revert to the original approach of doing phy_power_off() as well as phy_exit() during suspend.

Alternatively, Martin, how about not relying on skip_phy_initialization flag but having default behaviour of no-op for the usb_phy_roothub_*() driver. As platforms migrate to it they can set a new flag in hcd to use it?

I added this code to add support for the USB controller on the Amlogic Meson GXL and GXM SoCs:
these use a dwc3 controller for host mode

so while this approach would reduce the risk of breaking other controllers your TI SoCs would still be "affected" since I would like to enable the new logic for the dwc3 controller (which the TI SoCs use as well).

I will come up with an RFC patch series based on your and Chunfeng's suggestions thank you very much for working it out while I was away!

Quoted 38 lines
Quoted 38 lines
Quoted 36 lines
Quoted 36 lines
Quoted 4 lines
Quoted 2 lines

- fix this in the PHY driver

There is nothing to fix in the PHY driver. It is doing what it is supposed to do.

I actually wonder if phy_ops should have explicit suspend/resume support:
- assuming we define two new callbacks: .suspend and .resume
- the PHY framework could call .power_off by default if .suspend is not defined
- the PHY framework could call .power_on by default if .resume is not defined
- drivers could set .suspend and .resume on their own, allowing more fine-grained control by for example *only* stopping the clock (but not re-initializing the registers, etc.)

@Roger: what do you think about this?
Kishon (the PHY framework maintainer) is also CC'ed - I would like to hear his opinion too

Quoted 20 lines
Quoted 19 lines

- somewhere else

Quoted 17 lines
Quoted 16 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +7 -0
new file mode 100644
index 000000000000..6fde59bfbff8
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);
Quoted 8 lines
Quoted 2 lines

[0] https://github.com/xdarklight/linux/commits/usb-phy-roothub-suspend-rfc-v1

--
cheers,
-roger

Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki

Regards
Martin

maintainer
#23 lore raw

PATCH v10 3/8 usb: core: add a wrapper for the USB PHYs on the HCD

In reply to Roger Quadros
Quoted 220 lines · Roger Quadros

+some TI folks

Hi Martin,

Quoted 19 lines · Martin Blumenstingl

Many SoC platforms have separate devices for the USB PHY which are registered through the generic PHY framework. These PHYs have to be enabled to make the USB controller actually work. They also have to be disabled again on shutdown/suspend.

Currently (at least) the following HCI platform drivers are using custom code to obtain all PHYs via devicetree for the roothub/controller and disable/enable them when required:
- ehci-platform.c has ehci_platform_power_{on,off}
- xhci-mtk.c has xhci_mtk_phy_{init,exit,power_on,power_off}
- ohci-platform.c has ohci_platform_power_{on,off}

With this new wrapper the USB PHYs can be specified directly in the USB controller's devicetree node (just like on the drivers listed above). This allows SoCs like the Amlogic Meson GXL family to operate correctly once this is wired up correctly. These SoCs use a dwc3 controller and require all USB PHYs to be initialized (if one of the USB PHYs it not initialized then none of USB port works at all).

Quoted 4 lines

This patch is breaking low power cases on TI SoCs when USB is in host mode. I'll explain why below.

Quoted 101 lines
---
 drivers/usb/core/Makefile |   2 +-
 drivers/usb/core/phy.c    | 158 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/phy.h    |   7 ++
 3 files changed, 166 insertions(+), 1 deletion(-)
 create mode 100644 drivers/usb/core/phy.c
 create mode 100644 drivers/usb/core/phy.h
drivers/usb/core/Makefile +1 -1
index 92c9cefb4317..18e874b0441e 100644
@@ -6,7 +6,7 @@
 usbcore-y := usb.o hub.o hcd.o urb.o message.o driver.o
 usbcore-y += config.o file.o buffer.o sysfs.o endpoint.o
 usbcore-y += devio.o notify.o generic.o quirks.o devices.o
-usbcore-y += port.o
+usbcore-y += phy.o port.o
 
 usbcore-$(CONFIG_OF)		+= of.o
 usbcore-$(CONFIG_USB_PCI)		+= hcd-pci.o
drivers/usb/core/phy.c new file +74 -0
new file mode 100644
index 000000000000..09b7c43c0ea4
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * A wrapper for multiple PHYs which passes all phy_* function calls to
+ * multiple (actual) PHY devices. This is comes handy when initializing
+ * all PHYs on a HCD and to keep them all in the same state.
+ *
+ * Copyright (C) 2018 Martin Blumenstingl <[email protected]>
+ */
+
+#include <linux/device.h>
+#include <linux/list.h>
+#include <linux/phy/phy.h>
+#include <linux/of.h>
+
+#include "phy.h"
+
+struct usb_phy_roothub {
+	struct phy		*phy;
+	struct list_head	list;
+};
+
+static struct usb_phy_roothub *usb_phy_roothub_alloc(struct device *dev)
+{
+	struct usb_phy_roothub *roothub_entry;
+
+	roothub_entry = devm_kzalloc(dev, sizeof(*roothub_entry), GFP_KERNEL);
+	if (!roothub_entry)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&roothub_entry->list);
+
+	return roothub_entry;
+}
+
+static int usb_phy_roothub_add_phy(struct device *dev, int index,
+				   struct list_head *list)
+{
+	struct usb_phy_roothub *roothub_entry;
+	struct phy *phy = devm_of_phy_get_by_index(dev, dev->of_node, index);
+
+	if (IS_ERR_OR_NULL(phy)) {
+		if (!phy || PTR_ERR(phy) == -ENODEV)
+			return 0;
+		else
+			return PTR_ERR(phy);
+	}
+
+	roothub_entry = usb_phy_roothub_alloc(dev);
+	if (IS_ERR(roothub_entry))
+		return PTR_ERR(roothub_entry);
+
+	roothub_entry->phy = phy;
+
+	list_add_tail(&roothub_entry->list, list);
+
+	return 0;
+}
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
+{
+	struct usb_phy_roothub *phy_roothub;
+	struct usb_phy_roothub *roothub_entry;
+	struct list_head *head;
+	int i, num_phys, err;
+
+	num_phys = of_count_phandle_with_args(dev->of_node, "phys",
+					      "#phy-cells");
+	if (num_phys <= 0)
+		return NULL;
+
+	phy_roothub = usb_phy_roothub_alloc(dev);
+	if (IS_ERR(phy_roothub))
+		return phy_roothub;
+

drivers/usb/core/phy.c line 84

Quoted 10 lines
drivers/usb/core/phy.c new file +10 -0
+	for (i = 0; i < num_phys; i++) {
+		err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
+		if (err)
+			goto err_out;
+	}
+
+	head = &phy_roothub->list;
+
+	list_for_each_entry(roothub_entry, head, list) {
+		err = phy_init(roothub_entry->phy);

The phy_init() function actually enables the PHY clocks.
It should be moved to the usb_phy_roothub_exit() routine just before calling phy_power_on().

Quoted 21 lines
drivers/usb/core/phy.c new file +21 -0
+		if (err)
+			goto err_exit_phys;
+	}
+
+	return phy_roothub;
+
+err_exit_phys:
+	list_for_each_entry_continue_reverse(roothub_entry, head, list)
+		phy_exit(roothub_entry->phy);
+
+err_out:
+	return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
+
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
+{
+	struct usb_phy_roothub *roothub_entry;
+	struct list_head *head;
+	int err, ret = 0;
+

drivers/usb/core/phy.c line 115

Quoted 10 lines
drivers/usb/core/phy.c new file +10 -0
+	if (!phy_roothub)
+		return 0;
+
+	head = &phy_roothub->list;
+
+	list_for_each_entry(roothub_entry, head, list) {
+		err = phy_exit(roothub_entry->phy);
+		if (err)
+			ret = ret;
+	}

phy_exit() should be moved to usb_phy_roothub_poweroff() just after calling phy_power_off().

With that there is nothing else being done here. Shouldn't we be doing the equivalent of usb_phy_roothub_del_phy() and usb_phy_roothub_free() here?

Quoted 36 lines
drivers/usb/core/phy.c new file +36 -0
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_exit);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
+{
+	struct usb_phy_roothub *roothub_entry;
+	struct list_head *head;
+	int err;
+
+	if (!phy_roothub)
+		return 0;
+
+	head = &phy_roothub->list;
+
+	list_for_each_entry(roothub_entry, head, list) {
+		err = phy_power_on(roothub_entry->phy);
+		if (err)
+			goto err_out;
+	}
+
+	return 0;
+
+err_out:
+	list_for_each_entry_continue_reverse(roothub_entry, head, list)
+		phy_power_off(roothub_entry->phy);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_on);
+
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
+{
+	struct usb_phy_roothub *roothub_entry;
+
Quoted 8 lines

drivers/usb/core/phy.c line 156

Quoted 5 lines
drivers/usb/core/phy.c new file +5 -0
+	if (!phy_roothub)
+		return;
+
+	list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+		phy_power_off(roothub_entry->phy);

Not doing the phy_exit() here leaves the clocks enabled on our SoC and we're no longer able to reach low power states on system suspend.

Maybe you can modify phy driver if low power state is only blocked by enabled clocks, try to disable clocks in phy_power_off, and enable them again in phy_power_on()

Quoted 115 lines
Quoted 10 lines
drivers/usb/core/phy.c new file +2 -0
+}
+EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
drivers/usb/core/phy.h new file +0 -0
new file mode 100644
index 000000000000..6fde59bfbff8

drivers/usb/core/phy.h line 7

Quoted 7 lines
drivers/usb/core/phy.h new file +7 -0
@@ -0,0 +1,7 @@
+struct usb_phy_roothub;
+
+struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev);
+int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub);
+
+int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub);
+void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub);

The following patch fixes the issue for me. Let me know what you think and I can post it officially.

drivers/usb/core/phy.c +15 -28
index 09b7c43..23232d3 100644
@@ -59,8 +59,6 @@ static int usb_phy_roothub_add_phy(struct device *dev, int index,
 struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
 {
 	struct usb_phy_roothub *phy_roothub;
-	struct usb_phy_roothub *roothub_entry;
-	struct list_head *head;
 	int i, num_phys, err;
 
 	num_phys = of_count_phandle_with_args(dev->of_node, "phys",
@@ -75,25 +73,10 @@ struct usb_phy_roothub *usb_phy_roothub_init(struct device *dev)
 	for (i = 0; i < num_phys; i++) {
 		err = usb_phy_roothub_add_phy(dev, i, &phy_roothub->list);
 		if (err)
-			goto err_out;
-	}
-
-	head = &phy_roothub->list;
-
-	list_for_each_entry(roothub_entry, head, list) {
-		err = phy_init(roothub_entry->phy);
-		if (err)
-			goto err_exit_phys;
+			return ERR_PTR(err);
 	}
 
 	return phy_roothub;
-
-err_exit_phys:
-	list_for_each_entry_continue_reverse(roothub_entry, head, list)
-		phy_exit(roothub_entry->phy);
-
-err_out:
-	return ERR_PTR(err);
 }
 EXPORT_SYMBOL_GPL(usb_phy_roothub_init);
 
@@ -106,13 +89,8 @@ int usb_phy_roothub_exit(struct usb_phy_roothub *phy_roothub)
 	if (!phy_roothub)
 		return 0;
 
-	head = &phy_roothub->list;
-
-	list_for_each_entry(roothub_entry, head, list) {
-		err = phy_exit(roothub_entry->phy);
-		if (err)
-			ret = ret;
-	}
+	/* TODO: usb_phy_roothub_del_phy */
+	/* TODO: usb_phy_roothub_free */
 
 	return ret;
 }
@@ -130,16 +108,23 @@ int usb_phy_roothub_power_on(struct usb_phy_roothub *phy_roothub)
 	head = &phy_roothub->list;
 
 	list_for_each_entry(roothub_entry, head, list) {
-		err = phy_power_on(roothub_entry->phy);
+		err = phy_init(roothub_entry->phy);
 		if (err)
 			goto err_out;
+		err = phy_power_on(roothub_entry->phy);
+		if (err) {
+			phy_exit(roothub_entry->phy);
+			goto err_out;
+		}
 	}
 
 	return 0;
 
 err_out:
-	list_for_each_entry_continue_reverse(roothub_entry, head, list)
+	list_for_each_entry_continue_reverse(roothub_entry, head, list) {
 		phy_power_off(roothub_entry->phy);
+		phy_exit(roothub_entry->phy);
+	}
 
 	return err;
 }
@@ -152,7 +137,9 @@ void usb_phy_roothub_power_off(struct usb_phy_roothub *phy_roothub)
 	if (!phy_roothub)
 		return;
 
-	list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list)
+	list_for_each_entry_reverse(roothub_entry, &phy_roothub->list, list) {
 		phy_power_off(roothub_entry->phy);
+		phy_exit(roothub_entry->phy);
+	}
 }
 EXPORT_SYMBOL_GPL(usb_phy_roothub_power_off);
author reviewer
#24 lore raw

PATCH v10 4/8 usb: core: hcd: integrate the PHY wrapper into the HCD core

In reply to Martin Blumenstingl

This integrates the PHY wrapper into the core hcd infrastructure. Multiple PHYs which are part of the HCD's device tree node are now managed (= powered on/off when needed), by the new usb_phy_roothub code.

Suspend and resume is also supported, however not for runtime/auto-suspend (which is triggered for example when no devices are connected to the USB bus). This is needed on some SoCs (for example Amlogic Meson GXL) because if the PHYs are disabled during auto-suspend then devices which are plugged in afterwards are not seen by the host.

One example where this is required is the Amlogic GXL and GXM SoCs:
They are using a dwc3 USB controller with up to three ports enabled on the internal roothub. Each port has it's own PHY which must be enabled (if one of the PHYs is left disabled then none of the USB ports works at all).
The new logic works on the Amlogic GXL and GXM SoCs because the dwc3 driver internally creates a xhci-hcd which then registers a HCD which then triggers our new PHY wrapper.

USB controller drivers can opt out of this by setting "skip_phy_initialization" in struct usb_hcd to true. This is identical to how it works for a single USB PHY, so the "multiple PHY" handling is disabled for drivers that opted out of the management logic of a single PHY.

---
 drivers/usb/core/hcd.c  | 31 +++++++++++++++++++++++++++++++
 include/linux/usb/hcd.h |  1 +
 2 files changed, 32 insertions(+)
drivers/usb/core/hcd.c +31 -0
index f2307470a31e..32797c25ac3b 100644
@@ -37,6 +37,7 @@
 #include <linux/usb/otg.h>
 
 #include "usb.h"
+#include "phy.h"
 
 
 /*-------------------------------------------------------------------------*/
@@ -2260,6 +2261,9 @@ int hcd_bus_suspend(struct usb_device *rhdev, pm_message_t msg)
 		usb_set_device_state(rhdev, USB_STATE_SUSPENDED);
 		hcd->state = HC_STATE_SUSPENDED;
 
+		if (!PMSG_IS_AUTO(msg))
+			usb_phy_roothub_power_off(hcd->phy_roothub);
+
 		/* Did we race with a root-hub wakeup event? */
 		if (rhdev->do_remote_wakeup) {
 			char	buffer[6];
@@ -2296,6 +2300,13 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 		dev_dbg(&rhdev->dev, "skipped %s of dead bus\n", "resume");
 		return 0;
 	}
+
+	if (!PMSG_IS_AUTO(msg)) {
+		status = usb_phy_roothub_power_on(hcd->phy_roothub);
+		if (status)
+			return status;
+	}
+
 	if (!hcd->driver->bus_resume)
 		return -ENOENT;
 	if (HCD_RH_RUNNING(hcd))
@@ -2333,6 +2344,7 @@ int hcd_bus_resume(struct usb_device *rhdev, pm_message_t msg)
 		}
 	} else {
 		hcd->state = old_state;
+		usb_phy_roothub_power_off(hcd->phy_roothub);
 		dev_dbg(&rhdev->dev, "bus %s fail, err %d\n",
 				"resume", status);
 		if (status != -ESHUTDOWN)
@@ -2769,6 +2781,18 @@ int usb_add_hcd(struct usb_hcd *hcd,
 		}
 	}
 
+	if (!hcd->skip_phy_initialization) {
+		hcd->phy_roothub = usb_phy_roothub_init(hcd->self.sysdev);
+		if (IS_ERR(hcd->phy_roothub)) {
+			retval = PTR_ERR(hcd->phy_roothub);
+			goto err_phy_roothub_init;
+		}
+
+		retval = usb_phy_roothub_power_on(hcd->phy_roothub);
+		if (retval)
+			goto err_usb_phy_roothub_power_on;
+	}
+
 	dev_info(hcd->self.controller, "%s\n", hcd->product_desc);
 
 	/* Keep old behaviour if authorized_default is not in [0, 1]. */
@@ -2933,6 +2957,10 @@ int usb_add_hcd(struct usb_hcd *hcd,
 err_register_bus:
 	hcd_buffer_destroy(hcd);
 err_create_buf:
+	usb_phy_roothub_power_off(hcd->phy_roothub);
+err_usb_phy_roothub_power_on:
+	usb_phy_roothub_exit(hcd->phy_roothub);
+err_phy_roothub_init:
 	if (IS_ENABLED(CONFIG_GENERIC_PHY) && hcd->remove_phy && hcd->phy) {
 		phy_power_off(hcd->phy);
 		phy_exit(hcd->phy);
@@ -3017,6 +3045,9 @@ void usb_remove_hcd(struct usb_hcd *hcd)
 	usb_deregister_bus(&hcd->self);
 	hcd_buffer_destroy(hcd);
 
+	usb_phy_roothub_power_off(hcd->phy_roothub);
+	usb_phy_roothub_exit(hcd->phy_roothub);
+
 	if (IS_ENABLED(CONFIG_GENERIC_PHY) && hcd->remove_phy && hcd->phy) {
 		phy_power_off(hcd->phy);
 		phy_exit(hcd->phy);
include/linux/usb/hcd.h +1 -0
index 693502c84c04..a042675e03ba 100644
@@ -104,6 +104,7 @@ struct usb_hcd {
 	 */
 	struct usb_phy		*usb_phy;
 	struct phy		*phy;
+	struct usb_phy_roothub	*phy_roothub;
 
 	/* Flags that need to be manipulated atomically because they can
 	 * change while the host controller is running.  Always use
Signature · 1 line
2.16.1
author reviewer
#25 lore raw

PATCH v10 5/8 usb: host: xhci-mtk: remove custom USB PHY handling

In reply to Martin Blumenstingl

The new PHY wrapper is now wired up in the core HCD code. This means that PHYs are now controlled (initialized, enabled, disabled, exited) without requiring any host-driver specific code.
Remove the custom USB PHY handling from the xhci-mtk driver as the core HCD code now handles this.

---
 drivers/usb/host/xhci-mtk.c | 98 +--------------------------------------------
 1 file changed, 2 insertions(+), 96 deletions(-)
drivers/usb/host/xhci-mtk.c +2 -96
index b0ab4d5e2751..7334da9e9779 100644
@@ -14,7 +14,6 @@
 #include <linux/mfd/syscon.h>
 #include <linux/module.h>
 #include <linux/of.h>
-#include <linux/phy/phy.h>
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
 #include <linux/regmap.h>
@@ -352,62 +351,6 @@ static const struct xhci_driver_overrides xhci_mtk_overrides __initconst = {
 
 static struct hc_driver __read_mostly xhci_mtk_hc_driver;
 
-static int xhci_mtk_phy_init(struct xhci_hcd_mtk *mtk)
-{
-	int i;
-	int ret;
-
-	for (i = 0; i < mtk->num_phys; i++) {
-		ret = phy_init(mtk->phys[i]);
-		if (ret)
-			goto exit_phy;
-	}
-	return 0;
-
-exit_phy:
-	for (; i > 0; i--)
-		phy_exit(mtk->phys[i - 1]);
-
-	return ret;
-}
-
-static int xhci_mtk_phy_exit(struct xhci_hcd_mtk *mtk)
-{
-	int i;
-
-	for (i = 0; i < mtk->num_phys; i++)
-		phy_exit(mtk->phys[i]);
-
-	return 0;
-}
-
-static int xhci_mtk_phy_power_on(struct xhci_hcd_mtk *mtk)
-{
-	int i;
-	int ret;
-
-	for (i = 0; i < mtk->num_phys; i++) {
-		ret = phy_power_on(mtk->phys[i]);
-		if (ret)
-			goto power_off_phy;
-	}
-	return 0;
-
-power_off_phy:
-	for (; i > 0; i--)
-		phy_power_off(mtk->phys[i - 1]);
-
-	return ret;
-}
-
-static void xhci_mtk_phy_power_off(struct xhci_hcd_mtk *mtk)
-{
-	unsigned int i;
-
-	for (i = 0; i < mtk->num_phys; i++)
-		phy_power_off(mtk->phys[i]);
-}
-
 static int xhci_mtk_ldos_enable(struct xhci_hcd_mtk *mtk)
 {
 	int ret;
@@ -488,8 +431,6 @@ static int xhci_mtk_probe(struct platform_device *pdev)
 	struct xhci_hcd *xhci;
 	struct resource *res;
 	struct usb_hcd *hcd;
-	struct phy *phy;
-	int phy_num;
 	int ret = -ENODEV;
 	int irq;
 
@@ -529,16 +470,6 @@ static int xhci_mtk_probe(struct platform_device *pdev)
 		return ret;
 	}
 
-	mtk->num_phys = of_count_phandle_with_args(node,
-			"phys", "#phy-cells");
-	if (mtk->num_phys > 0) {
-		mtk->phys = devm_kcalloc(dev, mtk->num_phys,
-					sizeof(*mtk->phys), GFP_KERNEL);
-		if (!mtk->phys)
-			return -ENOMEM;
-	} else {
-		mtk->num_phys = 0;
-	}
 	pm_runtime_enable(dev);
 	pm_runtime_get_sync(dev);
 	device_enable_async_suspend(dev);
@@ -596,23 +527,6 @@ static int xhci_mtk_probe(struct platform_device *pdev)
 		mtk->has_ippc = false;
 	}
 
-	for (phy_num = 0; phy_num < mtk->num_phys; phy_num++) {
-		phy = devm_of_phy_get_by_index(dev, node, phy_num);
-		if (IS_ERR(phy)) {
-			ret = PTR_ERR(phy);
-			goto put_usb2_hcd;
-		}
-		mtk->phys[phy_num] = phy;
-	}
-
-	ret = xhci_mtk_phy_init(mtk);
-	if (ret)
-		goto put_usb2_hcd;
-
-	ret = xhci_mtk_phy_power_on(mtk);
-	if (ret)
-		goto exit_phys;
-
 	device_init_wakeup(dev, true);
 
 	xhci = hcd_to_xhci(hcd);
@@ -630,7 +544,7 @@ static int xhci_mtk_probe(struct platform_device *pdev)
 			dev_name(dev), hcd);
 	if (!xhci->shared_hcd) {
 		ret = -ENOMEM;
-		goto power_off_phys;
+		goto disable_device_wakeup;
 	}
 
 	ret = usb_add_hcd(hcd, irq, IRQF_SHARED);
@@ -653,13 +567,9 @@ static int xhci_mtk_probe(struct platform_device *pdev)
 	xhci_mtk_sch_exit(mtk);
 	usb_put_hcd(xhci->shared_hcd);
 
-power_off_phys:
-	xhci_mtk_phy_power_off(mtk);
+disable_device_wakeup:
 	device_init_wakeup(dev, false);
 
-exit_phys:
-	xhci_mtk_phy_exit(mtk);
-
 put_usb2_hcd:
 	usb_put_hcd(hcd);
 
@@ -682,8 +592,6 @@ static int xhci_mtk_remove(struct platform_device *dev)
 	struct xhci_hcd	*xhci = hcd_to_xhci(hcd);
 
 	usb_remove_hcd(xhci->shared_hcd);
-	xhci_mtk_phy_power_off(mtk);
-	xhci_mtk_phy_exit(mtk);
 	device_init_wakeup(&dev->dev, false);
 
 	usb_remove_hcd(hcd);
@@ -718,7 +626,6 @@ static int __maybe_unused xhci_mtk_suspend(struct device *dev)
 	del_timer_sync(&xhci->shared_hcd->rh_timer);
 
 	xhci_mtk_host_disable(mtk);
-	xhci_mtk_phy_power_off(mtk);
 	xhci_mtk_clks_disable(mtk);
 	usb_wakeup_set(mtk, true);
 	return 0;
@@ -732,7 +639,6 @@ static int __maybe_unused xhci_mtk_resume(struct device *dev)
 
 	usb_wakeup_set(mtk, false);
 	xhci_mtk_clks_enable(mtk);
-	xhci_mtk_phy_power_on(mtk);
 	xhci_mtk_host_enable(mtk);
 
 	xhci_dbg(xhci, "%s: restart port polling\n", __func__);
Signature · 1 line
2.16.1
author reviewer
#26 lore raw

PATCH v10 6/8 usb: host: ehci-platform: remove custom USB PHY handling

In reply to Martin Blumenstingl

The new PHY wrapper is now wired up in the core HCD code. This means that PHYs are now controlled (initialized, enabled, disabled, exited) without requiring any host-driver specific code.
Remove the custom USB PHY handling from the ehci-platform driver as the core HCD code now handles this.

---
 drivers/usb/host/ehci-platform.c | 55 +++-------------------------------------
 1 file changed, 4 insertions(+), 51 deletions(-)
drivers/usb/host/ehci-platform.c +4 -51
index b91eea8c73ae..4c306fb6b069 100644
@@ -27,7 +27,6 @@
 #include <linux/io.h>
 #include <linux/module.h>
 #include <linux/of.h>
-#include <linux/phy/phy.h>
 #include <linux/platform_device.h>
 #include <linux/reset.h>
 #include <linux/usb.h>
@@ -44,8 +43,6 @@
 struct ehci_platform_priv {
 	struct clk *clks[EHCI_MAX_CLKS];
 	struct reset_control *rsts;
-	struct phy **phys;
-	int num_phys;
 	bool reset_on_resume;
 };
 
@@ -80,7 +77,7 @@ static int ehci_platform_power_on(struct platform_device *dev)
 {
 	struct usb_hcd *hcd = platform_get_drvdata(dev);
 	struct ehci_platform_priv *priv = hcd_to_ehci_priv(hcd);
-	int clk, ret, phy_num;
+	int clk, ret;
 
 	for (clk = 0; clk < EHCI_MAX_CLKS && priv->clks[clk]; clk++) {
 		ret = clk_prepare_enable(priv->clks[clk]);
@@ -88,24 +85,8 @@ static int ehci_platform_power_on(struct platform_device *dev)
 			goto err_disable_clks;
 	}
 
-	for (phy_num = 0; phy_num < priv->num_phys; phy_num++) {
-		ret = phy_init(priv->phys[phy_num]);
-		if (ret)
-			goto err_exit_phy;
-		ret = phy_power_on(priv->phys[phy_num]);
-		if (ret) {
-			phy_exit(priv->phys[phy_num]);
-			goto err_exit_phy;
-		}
-	}
-
 	return 0;
 
-err_exit_phy:
-	while (--phy_num >= 0) {
-		phy_power_off(priv->phys[phy_num]);
-		phy_exit(priv->phys[phy_num]);
-	}
 err_disable_clks:
 	while (--clk >= 0)
 		clk_disable_unprepare(priv->clks[clk]);
@@ -117,12 +98,7 @@ static void ehci_platform_power_off(struct platform_device *dev)
 {
 	struct usb_hcd *hcd = platform_get_drvdata(dev);
 	struct ehci_platform_priv *priv = hcd_to_ehci_priv(hcd);
-	int clk, phy_num;
-
-	for (phy_num = 0; phy_num < priv->num_phys; phy_num++) {
-		phy_power_off(priv->phys[phy_num]);
-		phy_exit(priv->phys[phy_num]);
-	}
+	int clk;
 
 	for (clk = EHCI_MAX_CLKS - 1; clk >= 0; clk--)
 		if (priv->clks[clk])
@@ -149,7 +125,7 @@ static int ehci_platform_probe(struct platform_device *dev)
 	struct usb_ehci_pdata *pdata = dev_get_platdata(&dev->dev);
 	struct ehci_platform_priv *priv;
 	struct ehci_hcd *ehci;
-	int err, irq, phy_num, clk = 0;
+	int err, irq, clk = 0;
 
 	if (usb_disabled())
 		return -ENODEV;
@@ -202,29 +178,6 @@ static int ehci_platform_probe(struct platform_device *dev)
 					  "has-transaction-translator"))
 			hcd->has_tt = 1;
 
-		priv->num_phys = of_count_phandle_with_args(dev->dev.of_node,
-				"phys", "#phy-cells");
-
-		if (priv->num_phys > 0) {
-			priv->phys = devm_kcalloc(&dev->dev, priv->num_phys,
-					    sizeof(struct phy *), GFP_KERNEL);
-			if (!priv->phys)
-				return -ENOMEM;
-		} else
-			priv->num_phys = 0;
-
-		for (phy_num = 0; phy_num < priv->num_phys; phy_num++) {
-			priv->phys[phy_num] = devm_of_phy_get_by_index(
-					&dev->dev, dev->dev.of_node, phy_num);
-			if (IS_ERR(priv->phys[phy_num])) {
-				err = PTR_ERR(priv->phys[phy_num]);
-					goto err_put_hcd;
-			} else {
-				/* Avoiding phy_get() in usb_add_hcd() */
-				hcd->skip_phy_initialization = 1;
-			}
-		}
-
 		for (clk = 0; clk < EHCI_MAX_CLKS; clk++) {
 			priv->clks[clk] = of_clk_get(dev->dev.of_node, clk);
 			if (IS_ERR(priv->clks[clk])) {
@@ -306,7 +259,7 @@ static int ehci_platform_probe(struct platform_device *dev)
 err_put_clks:
 	while (--clk >= 0)
 		clk_put(priv->clks[clk]);
-err_put_hcd:
+
 	if (pdata == &ehci_platform_defaults)
 		dev->dev.platform_data = NULL;
 
Signature · 1 line
2.16.1
author reviewer
#27 lore raw

PATCH v10 7/8 usb: host: ohci-platform: remove custom USB PHY handling

In reply to Martin Blumenstingl

The new PHY wrapper is now wired up in the core HCD code. This means that PHYs are now controlled (initialized, enabled, disabled, exited) without requiring any host-driver specific code.
Remove the custom USB PHY handling from the ohci-platform driver as the core HCD code now handles this.

---
 drivers/usb/host/ohci-platform.c | 56 ++++------------------------------------
 1 file changed, 5 insertions(+), 51 deletions(-)
drivers/usb/host/ohci-platform.c +5 -51
index 62ef36a9333f..65a1c3fdc88c 100644
@@ -21,7 +21,7 @@
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/err.h>
-#include <linux/phy/phy.h>
+#include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
 #include <linux/reset.h>
@@ -38,8 +38,6 @@
 struct ohci_platform_priv {
 	struct clk *clks[OHCI_MAX_CLKS];
 	struct reset_control *resets;
-	struct phy **phys;
-	int num_phys;
 };
 
 static const char hcd_name[] = "ohci-platform";
@@ -48,7 +46,7 @@ static int ohci_platform_power_on(struct platform_device *dev)
 {
 	struct usb_hcd *hcd = platform_get_drvdata(dev);
 	struct ohci_platform_priv *priv = hcd_to_ohci_priv(hcd);
-	int clk, ret, phy_num;
+	int clk, ret;
 
 	for (clk = 0; clk < OHCI_MAX_CLKS && priv->clks[clk]; clk++) {
 		ret = clk_prepare_enable(priv->clks[clk]);
@@ -56,24 +54,8 @@ static int ohci_platform_power_on(struct platform_device *dev)
 			goto err_disable_clks;
 	}
 
-	for (phy_num = 0; phy_num < priv->num_phys; phy_num++) {
-		ret = phy_init(priv->phys[phy_num]);
-		if (ret)
-			goto err_exit_phy;
-		ret = phy_power_on(priv->phys[phy_num]);
-		if (ret) {
-			phy_exit(priv->phys[phy_num]);
-			goto err_exit_phy;
-		}
-	}
-
 	return 0;
 
-err_exit_phy:
-	while (--phy_num >= 0) {
-		phy_power_off(priv->phys[phy_num]);
-		phy_exit(priv->phys[phy_num]);
-	}
 err_disable_clks:
 	while (--clk >= 0)
 		clk_disable_unprepare(priv->clks[clk]);
@@ -85,12 +67,7 @@ static void ohci_platform_power_off(struct platform_device *dev)
 {
 	struct usb_hcd *hcd = platform_get_drvdata(dev);
 	struct ohci_platform_priv *priv = hcd_to_ohci_priv(hcd);
-	int clk, phy_num;
-
-	for (phy_num = 0; phy_num < priv->num_phys; phy_num++) {
-		phy_power_off(priv->phys[phy_num]);
-		phy_exit(priv->phys[phy_num]);
-	}
+	int clk;
 
 	for (clk = OHCI_MAX_CLKS - 1; clk >= 0; clk--)
 		if (priv->clks[clk])
@@ -117,7 +94,7 @@ static int ohci_platform_probe(struct platform_device *dev)
 	struct usb_ohci_pdata *pdata = dev_get_platdata(&dev->dev);
 	struct ohci_platform_priv *priv;
 	struct ohci_hcd *ohci;
-	int err, irq, phy_num, clk = 0;
+	int err, irq, clk = 0;
 
 	if (usb_disabled())
 		return -ENODEV;
@@ -169,29 +146,6 @@ static int ohci_platform_probe(struct platform_device *dev)
 		of_property_read_u32(dev->dev.of_node, "num-ports",
 				     &ohci->num_ports);
 
-		priv->num_phys = of_count_phandle_with_args(dev->dev.of_node,
-				"phys", "#phy-cells");
-
-		if (priv->num_phys > 0) {
-			priv->phys = devm_kcalloc(&dev->dev, priv->num_phys,
-					    sizeof(struct phy *), GFP_KERNEL);
-			if (!priv->phys)
-				return -ENOMEM;
-		} else
-			priv->num_phys = 0;
-
-		for (phy_num = 0; phy_num < priv->num_phys; phy_num++) {
-			priv->phys[phy_num] = devm_of_phy_get_by_index(
-					&dev->dev, dev->dev.of_node, phy_num);
-			if (IS_ERR(priv->phys[phy_num])) {
-				err = PTR_ERR(priv->phys[phy_num]);
-				goto err_put_hcd;
-			} else {
-				/* Avoiding phy_get() in usb_add_hcd() */
-				hcd->skip_phy_initialization = 1;
-			}
-		}
-
 		for (clk = 0; clk < OHCI_MAX_CLKS; clk++) {
 			priv->clks[clk] = of_clk_get(dev->dev.of_node, clk);
 			if (IS_ERR(priv->clks[clk])) {
@@ -277,7 +231,7 @@ static int ohci_platform_probe(struct platform_device *dev)
 err_put_clks:
 	while (--clk >= 0)
 		clk_put(priv->clks[clk]);
-err_put_hcd:
+
 	if (pdata == &ohci_platform_defaults)
 		dev->dev.platform_data = NULL;
 
Signature · 1 line
2.16.1
author reviewer
#28 lore raw

PATCH v10 8/8 usb: core: hcd: remove support for initializing a single PHY

In reply to Martin Blumenstingl

With the new PHY wrapper in place we can now handle multiple PHYs. Remove the code which handles only one generic PHY as this is now covered (with support for multiple PHYs as well as suspend/resume support) by the new PHY wrapper.

---
 drivers/usb/core/hcd.c  | 37 -------------------------------------
 include/linux/usb/hcd.h |  1 -
 2 files changed, 38 deletions(-)
drivers/usb/core/hcd.c +0 -37
index 32797c25ac3b..5a92d8f7c484 100644
@@ -2757,30 +2757,6 @@ int usb_add_hcd(struct usb_hcd *hcd,
 		}
 	}
 
-	if (IS_ENABLED(CONFIG_GENERIC_PHY) && !hcd->skip_phy_initialization) {
-		struct phy *phy = phy_get(hcd->self.sysdev, "usb");
-
-		if (IS_ERR(phy)) {
-			retval = PTR_ERR(phy);
-			if (retval == -EPROBE_DEFER)
-				goto err_phy;
-		} else {
-			retval = phy_init(phy);
-			if (retval) {
-				phy_put(phy);
-				goto err_phy;
-			}
-			retval = phy_power_on(phy);
-			if (retval) {
-				phy_exit(phy);
-				phy_put(phy);
-				goto err_phy;
-			}
-			hcd->phy = phy;
-			hcd->remove_phy = 1;
-		}
-	}
-
 	if (!hcd->skip_phy_initialization) {
 		hcd->phy_roothub = usb_phy_roothub_init(hcd->self.sysdev);
 		if (IS_ERR(hcd->phy_roothub)) {
@@ -2961,13 +2937,6 @@ int usb_add_hcd(struct usb_hcd *hcd,
 err_usb_phy_roothub_power_on:
 	usb_phy_roothub_exit(hcd->phy_roothub);
 err_phy_roothub_init:
-	if (IS_ENABLED(CONFIG_GENERIC_PHY) && hcd->remove_phy && hcd->phy) {
-		phy_power_off(hcd->phy);
-		phy_exit(hcd->phy);
-		phy_put(hcd->phy);
-		hcd->phy = NULL;
-	}
-err_phy:
 	if (hcd->remove_phy && hcd->usb_phy) {
 		usb_phy_shutdown(hcd->usb_phy);
 		usb_put_phy(hcd->usb_phy);
@@ -3048,12 +3017,6 @@ void usb_remove_hcd(struct usb_hcd *hcd)
 	usb_phy_roothub_power_off(hcd->phy_roothub);
 	usb_phy_roothub_exit(hcd->phy_roothub);
 
-	if (IS_ENABLED(CONFIG_GENERIC_PHY) && hcd->remove_phy && hcd->phy) {
-		phy_power_off(hcd->phy);
-		phy_exit(hcd->phy);
-		phy_put(hcd->phy);
-		hcd->phy = NULL;
-	}
 	if (hcd->remove_phy && hcd->usb_phy) {
 		usb_phy_shutdown(hcd->usb_phy);
 		usb_put_phy(hcd->usb_phy);
include/linux/usb/hcd.h +0 -1
index a042675e03ba..aef50cb2ed1b 100644
@@ -103,7 +103,6 @@ struct usb_hcd {
 	 * other external phys should be software-transparent
 	 */
 	struct usb_phy		*usb_phy;
-	struct phy		*phy;
 	struct usb_phy_roothub	*phy_roothub;
 
 	/* Flags that need to be manipulated atomically because they can
Signature · 1 line
2.16.1