Message-Id: <20250226-dwc3-refactor-v4-0-4415e7111e49@oss.qualcomm.com>
Date: Wed, 26 Feb 2025 16:17:47 -0800
From: Bjorn Andersson <bjorn.andersson@....qualcomm.com>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Rob Herring <robh@...nel.org>,
        Krzysztof Kozlowski <krzk+dt@...nel.org>,
        Conor Dooley <conor+dt@...nel.org>, Felipe Balbi <balbi@...nel.org>,
        Wesley Cheng <quic_wcheng@...cinc.com>,
        Saravana Kannan <saravanak@...gle.com>,
        Thinh Nguyen <Thinh.Nguyen@...opsys.com>,
        Philipp Zabel <p.zabel@...gutronix.de>,
        Bjorn Andersson <andersson@...nel.org>,
        Konrad Dybcio <konradybcio@...nel.org>, Frank Li <Frank.li@....com>
Cc: linux-arm-msm@...r.kernel.org, linux-usb@...r.kernel.org,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        Bjorn Andersson <bjorn.andersson@....qualcomm.com>
Subject: [PATCH v4 0/7] usb: dwc3: qcom: Flatten dwc3 structure

The USB IP-block found in most Qualcomm platforms is modelled in the
Linux kernel as 3 different independent device drivers, but as shown by
the already existing layering violations in the Qualcomm glue driver
they can not be operated independently.

With the current implementation, the glue driver registers the core and
has no way to know when this is done. As a result, e.g. the suspend
callbacks needs to guard against NULL pointer dereferences when trying
to peek into the struct dwc3 found in the drvdata of the child.

Missing from the upstream Qualcomm USB support is proper handling of
role switching, in which the glue needs to be notified upon DRD mode
changes. Several attempts has been made through the years to register
callbacks etc, but they always fall short when it comes to handling of
the core's probe deferral on resources etc.

Furhtermore, the DeviceTree binding is a direct representation of the
Linux driver model, and doesn't necessarily describe "the USB IP-block".

This series therefor attempts to flatten the driver split, and operate
the glue and core out of the same platform_device instance. And in order
to do this, the DeviceTree representation of the IP block is flattened.

Departing from previous versions' attempts at runtime-convert the
Devicetree representation is swapped out and instead a snapshot of the
current dwc3-qcom driver is proposed to be carried for a limited time.

A patch to convert a single platform - sc8280xp - is included in the
series. The broader conversion will be submitted in a follow up series.

---
Changes in v4:
- dwc3_{init,uninit}() renamed to dwc3_core_probe() and dwc3_core_remove()
- dwc3_{suspend, resume, complete}() changed to dwc3_pm_*()
- Arguments to dwc3_core_probe() are wrapped in a struct to better
  handle the expected growing list of parameters.
- Add the lost call to dwc3_core_remove() from the Qualcomm glue driver
- Removed now unused cleanup.h, of_address.h, and of_irq.h includes from
  dwc3-qcom.c
- Link to v3: https://lore.kernel.org/r/20250113-dwc3-refactor-v3-0-d1722075df7b@oss.qualcomm.com

Changes in v3:
- Replaced the handcoded migration logic of compatible, reg, interrupts,
  phys with overlays.
- Move the migration logic (and overlays) to a new drivers/of/overlays
  directory and apply this at postcore, so that it takes effect prior to
  the relevant platform_devices are created
- struct dwc3 is embedded in the glue context, rather than having a
  separate object allocated
- The hack of using of_address_to_resource() to avoid platform_resource
  being stale is removed (thanks to applying migration at postcore)
- Link to v2: https://lore.kernel.org/r/20240811-dwc3-refactor-v2-0-91f370d61ad2@quicinc.com

Changes in v2:
- Rewrite after ACPI removal, multiport support and interrupt fixes
- Completely changed strategy for DeviceTree binding, as previous idea
  of using snps,dwc3 as a generic fallback required unreasonable changes
  to that binding.
- Abandoned idea of supporting both flattened and unflattened device
  model in the one driver. As Johan pointed out, it will leave the race
  condition holes and will make the code harder to understand.
  Furthermore, the role switching logic that we intend to introduce
  following this would have depended on the user updating their
  DeviceTree blobs.
- Per above, introduced the dynamic DeviceTree rewrite
- Link to v1: https://lore.kernel.org/all/20231016-dwc3-refactor-v1-0-ab4a84165470@quicinc.com/

---
Bjorn Andersson (7):
      dt-bindings: usb: Introduce qcom,snps-dwc3
      usb: dwc3: core: Expose core driver as library
      usb: dwc3: core: Don't touch resets and clocks
      usb: dwc3: qcom: Don't rely on drvdata during probe
      usb: dwc3: qcom: Snapshot driver for backwards compatibilty
      usb: dwc3: qcom: Transition to flattened model
      arm64: dts: qcom: sc8280x: Flatten the USB nodes

 .../devicetree/bindings/usb/qcom,dwc3.yaml         |  13 +-
 .../devicetree/bindings/usb/qcom,snps-dwc3.yaml    | 619 ++++++++++++++
 arch/arm64/boot/dts/qcom/sa8295p-adp.dts           |  12 +-
 arch/arm64/boot/dts/qcom/sa8540p-ride.dts          |   5 +-
 arch/arm64/boot/dts/qcom/sc8280xp-crd.dts          |  12 +-
 .../boot/dts/qcom/sc8280xp-huawei-gaokun3.dts      |  10 +-
 .../dts/qcom/sc8280xp-lenovo-thinkpad-x13s.dts     |  11 +-
 .../boot/dts/qcom/sc8280xp-microsoft-arcata.dts    |  10 +-
 .../boot/dts/qcom/sc8280xp-microsoft-blackrock.dts |  18 +-
 arch/arm64/boot/dts/qcom/sc8280xp.dtsi             | 157 ++--
 drivers/usb/dwc3/Makefile                          |   1 +
 drivers/usb/dwc3/core.c                            | 174 ++--
 drivers/usb/dwc3/dwc3-qcom-legacy.c                | 934 +++++++++++++++++++++
 drivers/usb/dwc3/dwc3-qcom.c                       | 152 ++--
 drivers/usb/dwc3/glue.h                            |  33 +
 15 files changed, 1881 insertions(+), 280 deletions(-)
---
base-commit: c1136f35c7d1ce2517e875884644a44da6121c35
change-id: 20231016-dwc3-refactor-931e3b08a8b9

Best regards,
-- 
Bjorn Andersson <bjorn.andersson@....qualcomm.com>

Powered by blists - more mailing lists