This patch set add USB device controller for the NPCM Baseboard
Management Controllers (BMC).
NPCM UDC driver is a part of the USB ChipIdea driver.
The NPCM UDC is using the CI_HDRC_FORCE_VBUS_ACTIVE_ALWAYS flag
therefor, this patch set should apply after
usb: chipidea: add CI_HDRC_FORCE_VBUS_ACTIVE_ALWAYS flag commit will
apply.
BMC NPCM7XX and BMC NPCM8XX has ten identical NPCM UDC modules,
The NPCM UDC were tested on NPCM845 evaluation board.
Tomer Maimon (2):
dt-binding: usb: ci-hdrc-usb2: document Nuvoton NPCM supprt
usb: chipidea: Add support for NPCM
.../devicetree/bindings/usb/ci-hdrc-usb2.yaml | 16 +++
drivers/usb/chipidea/Kconfig | 4 +
drivers/usb/chipidea/Makefile | 1 +
drivers/usb/chipidea/ci_hdrc_npcm.c | 126 ++++++++++++++++++
4 files changed, 147 insertions(+)
create mode 100644 drivers/usb/chipidea/ci_hdrc_npcm.c
--
2.33.0
Nuvoton NPCM BMC SoCs use ChipIdea silicon IP for the USB device controller.
Signed-off-by: Tomer Maimon <[email protected]>
---
.../devicetree/bindings/usb/ci-hdrc-usb2.yaml | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/Documentation/devicetree/bindings/usb/ci-hdrc-usb2.yaml b/Documentation/devicetree/bindings/usb/ci-hdrc-usb2.yaml
index 1394557517b1..9de4dfe004d1 100644
--- a/Documentation/devicetree/bindings/usb/ci-hdrc-usb2.yaml
+++ b/Documentation/devicetree/bindings/usb/ci-hdrc-usb2.yaml
@@ -16,6 +16,7 @@ properties:
- enum:
- chipidea,usb2
- lsi,zevio-usb
+ - nuvoton,npcm-udc
- nvidia,tegra20-ehci
- nvidia,tegra20-udc
- nvidia,tegra30-ehci
@@ -325,6 +326,20 @@ properties:
type: boolean
deprecated: true
+ nuvoton,sysgcr:
+ $ref: /schemas/types.yaml#/definitions/phandle-array
+ items:
+ - items:
+ - description: phandle to syscon that configures usb phy mux.
+ - description: offset of usb phy mux selection.
+ - description: mask usb phy mux selection.
+ - description: value usb phy mux selection.
+ description:
+ A phandle to syscon with three arguments that configure usb phy mux.
+ The argument one is the offset of usb phy mux selection, the argument two
+ is the mask usb phy mux selection, the argument three is the mask usb phy
+ mux selection.
+
port:
description:
Any connector to the data bus of this controller should be modelled
@@ -388,6 +403,7 @@ allOf:
enum:
- chipidea,usb2
- lsi,zevio-usb
+ - nuvoton,npcm-udc
- nvidia,tegra20-udc
- nvidia,tegra30-udc
- nvidia,tegra114-udc
--
2.33.0
On Tue, Sep 19, 2023 at 04:31:56PM +0300, Tomer Maimon wrote:
> On Tue, 19 Sept 2023 at 15:39, Krzysztof Kozlowski
> <[email protected]> wrote:
> >
> > On 19/09/2023 07:14, Tomer Maimon wrote:
> > >>> - nvidia,tegra20-ehci
> > >>> - nvidia,tegra20-udc
> > >>> - nvidia,tegra30-ehci
> > >>> @@ -325,6 +326,20 @@ properties:
> > >>> type: boolean
> > >>> deprecated: true
> > >>>
> > >>> + nuvoton,sysgcr:
> > >>> + $ref: /schemas/types.yaml#/definitions/phandle-array
> > >>> + items:
> > >>> + - items:
> > >>> + - description: phandle to syscon that configures usb phy mux.
> > >>> + - description: offset of usb phy mux selection.
> > >>> + - description: mask usb phy mux selection.
> > >>> + - description: value usb phy mux selection.
> > >>> + description:
> > >>> + A phandle to syscon with three arguments that configure usb phy mux.
> > >>> + The argument one is the offset of usb phy mux selection, the argument two
> > >>> + is the mask usb phy mux selection, the argument three is the mask usb phy
> > >>> + mux selection.
> > >>
> > >> Sorry, you miss phy driver. Don't use syscon instead of proper hardware
> > >> devices.
> > > Sorry the role of nuvoton,sysgcr property is to handle a mux between
> > > the different devices and not the handle the phy itself, handle the
> > > mux done in the GCR.
> > > Should we move the nuvoton,sysgcr description to another place in the
> > > ci-hdrc-usb2.yaml
> > > or
> > > Should we use a different driver to handle the mux and call it from
> > > the ci-hdrc-npcm driver, If yes which driver should we use?
> >
> > What is an "usb phy mux"?
> We have USB phy that could be connected to USB host (different driver)
> or it can be connected to the UDC driver(ChipIdea)
Isn't that just role switching? There is a driver framework for that in
drivers/usb/roles/. Though it doesn't seem widely used yet.
Rob
Thanks Rob,
I will check drivers/usb/roles
On Tue, 19 Sept 2023 at 19:28, Rob Herring <[email protected]> wrote:
>
> On Tue, Sep 19, 2023 at 04:31:56PM +0300, Tomer Maimon wrote:
> > On Tue, 19 Sept 2023 at 15:39, Krzysztof Kozlowski
> > <[email protected]> wrote:
> > >
> > > On 19/09/2023 07:14, Tomer Maimon wrote:
> > > >>> - nvidia,tegra20-ehci
> > > >>> - nvidia,tegra20-udc
> > > >>> - nvidia,tegra30-ehci
> > > >>> @@ -325,6 +326,20 @@ properties:
> > > >>> type: boolean
> > > >>> deprecated: true
> > > >>>
> > > >>> + nuvoton,sysgcr:
> > > >>> + $ref: /schemas/types.yaml#/definitions/phandle-array
> > > >>> + items:
> > > >>> + - items:
> > > >>> + - description: phandle to syscon that configures usb phy mux.
> > > >>> + - description: offset of usb phy mux selection.
> > > >>> + - description: mask usb phy mux selection.
> > > >>> + - description: value usb phy mux selection.
> > > >>> + description:
> > > >>> + A phandle to syscon with three arguments that configure usb phy mux.
> > > >>> + The argument one is the offset of usb phy mux selection, the argument two
> > > >>> + is the mask usb phy mux selection, the argument three is the mask usb phy
> > > >>> + mux selection.
> > > >>
> > > >> Sorry, you miss phy driver. Don't use syscon instead of proper hardware
> > > >> devices.
> > > > Sorry the role of nuvoton,sysgcr property is to handle a mux between
> > > > the different devices and not the handle the phy itself, handle the
> > > > mux done in the GCR.
> > > > Should we move the nuvoton,sysgcr description to another place in the
> > > > ci-hdrc-usb2.yaml
> > > > or
> > > > Should we use a different driver to handle the mux and call it from
> > > > the ci-hdrc-npcm driver, If yes which driver should we use?
> > >
> > > What is an "usb phy mux"?
> > We have USB phy that could be connected to USB host (different driver)
> > or it can be connected to the UDC driver(ChipIdea)
>
> Isn't that just role switching? There is a driver framework for that in
> drivers/usb/roles/. Though it doesn't seem widely used yet.
>
> Rob
Hi Rob,
Again thanks for your suggestion I took a look at the
drivers/usb/roles and I am not sure it answers NPCM ChipIdae case.
For example, in the NPCM845 we have ten UDC ChipIdea modules (UDC0-9).
Only UDC8 and UDC9 are muxed with USB host0 and USB host1.
NPCM UDC and NPCM USB host are different HW modules therefore different drivers.
The ChipIdea driver uses its own internal USB role function to switch
between ChipIdea UDC and ChipIdea USB host and we can't replace it
with a unique NPCM USB role function also we need to set the mux only
at the probe stage.
This is why I have added nuvoton,sysgcr property to the NPCM ChipIdea driver.
With the above do you think I should do it differently?
On Wed, 20 Sept 2023 at 12:11, Tomer Maimon <[email protected]> wrote:
>
> Thanks Rob,
>
> I will check drivers/usb/roles
>
> On Tue, 19 Sept 2023 at 19:28, Rob Herring <[email protected]> wrote:
> >
> > On Tue, Sep 19, 2023 at 04:31:56PM +0300, Tomer Maimon wrote:
> > > On Tue, 19 Sept 2023 at 15:39, Krzysztof Kozlowski
> > > <[email protected]> wrote:
> > > >
> > > > On 19/09/2023 07:14, Tomer Maimon wrote:
> > > > >>> - nvidia,tegra20-ehci
> > > > >>> - nvidia,tegra20-udc
> > > > >>> - nvidia,tegra30-ehci
> > > > >>> @@ -325,6 +326,20 @@ properties:
> > > > >>> type: boolean
> > > > >>> deprecated: true
> > > > >>>
> > > > >>> + nuvoton,sysgcr:
> > > > >>> + $ref: /schemas/types.yaml#/definitions/phandle-array
> > > > >>> + items:
> > > > >>> + - items:
> > > > >>> + - description: phandle to syscon that configures usb phy mux.
> > > > >>> + - description: offset of usb phy mux selection.
> > > > >>> + - description: mask usb phy mux selection.
> > > > >>> + - description: value usb phy mux selection.
> > > > >>> + description:
> > > > >>> + A phandle to syscon with three arguments that configure usb phy mux.
> > > > >>> + The argument one is the offset of usb phy mux selection, the argument two
> > > > >>> + is the mask usb phy mux selection, the argument three is the mask usb phy
> > > > >>> + mux selection.
> > > > >>
> > > > >> Sorry, you miss phy driver. Don't use syscon instead of proper hardware
> > > > >> devices.
> > > > > Sorry the role of nuvoton,sysgcr property is to handle a mux between
> > > > > the different devices and not the handle the phy itself, handle the
> > > > > mux done in the GCR.
> > > > > Should we move the nuvoton,sysgcr description to another place in the
> > > > > ci-hdrc-usb2.yaml
> > > > > or
> > > > > Should we use a different driver to handle the mux and call it from
> > > > > the ci-hdrc-npcm driver, If yes which driver should we use?
> > > >
> > > > What is an "usb phy mux"?
> > > We have USB phy that could be connected to USB host (different driver)
> > > or it can be connected to the UDC driver(ChipIdea)
> >
> > Isn't that just role switching? There is a driver framework for that in
> > drivers/usb/roles/. Though it doesn't seem widely used yet.
> >
> > Rob
Best regards,
Tomer
On 21/09/2023 15:17, Tomer Maimon wrote:
> Hi Rob,
>
> Again thanks for your suggestion I took a look at the
> drivers/usb/roles and I am not sure it answers NPCM ChipIdae case.
> For example, in the NPCM845 we have ten UDC ChipIdea modules (UDC0-9).
> Only UDC8 and UDC9 are muxed with USB host0 and USB host1.
> NPCM UDC and NPCM USB host are different HW modules therefore different drivers.
>
> The ChipIdea driver uses its own internal USB role function to switch
> between ChipIdea UDC and ChipIdea USB host and we can't replace it
> with a unique NPCM USB role function also we need to set the mux only
> at the probe stage.
>
> This is why I have added nuvoton,sysgcr property to the NPCM ChipIdea driver.
>
> With the above do you think I should do it differently?
Your entire rationale above explains drivers, so it is not proper
justification for bindings. How your probe stage has anything to do with
bindings? If drivers do not work or miss something, change them.
Best regards,
Krzysztof