mbox series

[v5,0/6] usb: typec: Add new driver for Parade PS8830 Type-C Retimer

Message ID 20241112-x1e80100-ps8830-v5-0-4ad83af4d162@linaro.org
Headers show
Series usb: typec: Add new driver for Parade PS8830 Type-C Retimer | expand

Message

Abel Vesa Nov. 12, 2024, 5:01 p.m. UTC
The Parade PS8830 is a Type-C multi-protocol retimer that is controlled
via I2C. It provides altmode and orientation handling and usually sits
between the Type-C port and the PHY.

It is currently used alongside Qualcomm Snapdragon X Elite SoCs on quite
a few laptops already.

This new driver adds support for the following 3 modes:
 - DP 4lanes (pin assignments C and E)
 - DP 2lanes + USB3 (pin assignment D)
 - USB3

This retimer is a LTTPR (Link-Training Tunable PHY Repeater) which means
it can support link training from source to itself. This means that the
DP driver needs to be aware of the repeater presence and to handle
the link training accordingly. This is currently missing from msm dp
driver, but there is already a patchset [1] on the list that adds it.
Once done, full external DP will be working on all X1E laptops that make
use of this retimer.

NOTE: Currently, due to both LTTPR missing support in msm DP and a
reported crash that can happen on DP unplug, the DP DT patches are not
supposed to be merged yet. That patch is only shared for testing purposes.
Once those 2 issues have been resolved, the MDSS DP 0-2 enablement patch
will be respun.

The LTTPR patchset is already on the list:
[1] https://lore.kernel.org/all/20241031-drm-dp-msm-add-lttpr-transparent-mode-set-v1-0-cafbb9855f40@linaro.org/

Signed-off-by: Abel Vesa <abel.vesa@linaro.org>
---
Changes in v5:
- Renamed binding schema to be the same as the compatible.
- Dropped the ps8830,boot-on from the schema.
- Added register offsets and bits names to the driver, like Konrad
  suggested.
- Reordered the vregs and clocks enabling, as per Johan's request.
- Used the dev_err_probe for regmap init failure and dropped the
  multiple regulator disable calls, replacing it with single call to
  helper, as Christophe suggested. Also replaced dev_err with
  dev_err_probe on typec_switch_register and typec_mux_register failure.
- Added some new pinctrl specific properties to all pmic provided
  gpios that control retimer vregs.
- Re-ordered alphabetically the retimers default state pinconfs.
- Added the T14s patches with same exact support, as per Johan's
  request.
- Link to v4: https://lore.kernel.org/r/20241101-x1e80100-ps8830-v4-0-f0f7518b263e@linaro.org

Changes in v4:
- Renamed the driver and bindings schema to ps883x to allow future
  support for the PS8833.
- Dropped the dedicated DT property for keeping the retimers from
  resetting on probe, and replaced that with a read to figure out
  if it has been already configured or not. This involves leaving the
  reset gpio as-is on probe if the retimer has been already configured.
- Replaced the fwnode_typec_switch_get() call with typec_switch_get()
- Replaced the fwnode_typec_mux_get() call with typec_mux_get()
- Dropped the clock name, as there is only one clock. As per Bjorn's
  suggestion.
- Dropped regcache as it seems it is not needed.
- Re-worded all commit messages to explain better the problem and the
  proposed changes.
- Link to v3: https://lore.kernel.org/r/20241022-x1e80100-ps8830-v3-0-68a95f351e99@linaro.org

Changes in v3:
- Reworked the schema binding by using the usb/usb-switch.yaml defined
  port graph and properties. Addressed all comments from Johan and
  Dmitry.
- Dropped the manual caching of the config values on regmap write in the
  driver.
- Reordered the DP pin assignment states within the switch clause, as
  Dmitry suggested.
- Added SVID check to not allow any altmode other than DP.
- Added DT patches (retimer for USB orientation handling and DP
  enablement). Did this in order to offer a full picture of how it all
  fits together.
- Split the DP enablement in DT in a separate patchset so the USB
  handling can be merged separately. 
- Added ps8830,boot-on to let the driver know it is supposed to skip
  resetting the retimer on driver probe, as the bootloader might already
  let it in a pre-configured state.
- Marked all retimer voltage regulators as boot-on since we want to
  maintain the state for coldplug orientation.
- Added pinconf for all retimer0 gpios.
- Didn't pick up Konrad's T-b tags and Krzysztof's R-b tag as the rework
  is quite extensive. Especially because of the ps8830,boot-on and what
  it does.
- Link to v2: https://lore.kernel.org/r/20241004-x1e80100-ps8830-v2-0-5cd8008c8c40@linaro.org

Changes in v2:
- Addressed all comments from Johan and Konrad.
- Reworked the handling of the vregs so it would be more cleaner.
  Dropped the usage of bulk regulators API and handled them separately.
  Also discribed all regulators according to data sheet.
- Added all delays according to data sheet.
- Fixed coldplug (on boot) orientation detection.
- Didn't pick Krzysztof's R-b tag because the bindings changed w.r.t
  supplies.
- Link to v1: https://lore.kernel.org/r/20240829-x1e80100-ps8830-v1-0-bcc4790b1d45@linaro.org

---
Abel Vesa (6):
      dt-bindings: usb: Add Parade PS8830 Type-C retimer bindings
      usb: typec: Add support for Parade PS8830 Type-C Retimer
      arm64: dts: qcom: x1e80100-crd: Describe the Parade PS8830 retimers
      arm64: dts: qcom: x1e80100-crd: Enable external DisplayPort support
      arm64: dts: qcom: x1e80100-t14s: Describe the Parade PS8830 retimers
      arm64: dts: qcom: x1e80100-t14s: Enable external DisplayPort support

 .../devicetree/bindings/usb/parade,ps8830.yaml     | 119 ++++++
 .../dts/qcom/x1e78100-lenovo-thinkpad-t14s.dts     | 321 +++++++++++++-
 arch/arm64/boot/dts/qcom/x1e80100-crd.dts          | 474 ++++++++++++++++++++-
 drivers/usb/typec/mux/Kconfig                      |  10 +
 drivers/usb/typec/mux/Makefile                     |   1 +
 drivers/usb/typec/mux/ps883x.c                     | 437 +++++++++++++++++++
 6 files changed, 1352 insertions(+), 10 deletions(-)
---
base-commit: 28955f4fa2823e39f1ecfb3a37a364563527afbc
change-id: 20240521-x1e80100-ps8830-d5ccca95b557

Best regards,

Comments

Abel Vesa Nov. 12, 2024, 5:05 p.m. UTC | #1
On 24-11-12 19:01:13, Abel Vesa wrote:
> The X Elite CRD provides external DisplayPort on all 3 USB Type-C ports.
> Each one of this ports is connected to a dedicated DisplayPort
> controller.
> 
> Due to support missing in the USB/DisplayPort combo PHY driver,
> the external DisplayPort is limited to 2 lanes.
> 
> So enable all 3 remaining DisplayPort controllers and limit their data
> lanes number to 2.
> 
> Signed-off-by: Abel Vesa <abel.vesa@linaro.org>

Please do not merge this specific patch.

It is provided just for context and testing purposes.

See the cover note for more details.

> ---
>  arch/arm64/boot/dts/qcom/x1e80100-crd.dts | 24 ++++++++++++++++++++++++
>  1 file changed, 24 insertions(+)
> 
> diff --git a/arch/arm64/boot/dts/qcom/x1e80100-crd.dts b/arch/arm64/boot/dts/qcom/x1e80100-crd.dts
> index 659520404adec33c3551f8d0a5ae3db9e0a18d44..6dc2ebbf6d27fbbf0f224e58cd39ffd33792c6a1 100644
> --- a/arch/arm64/boot/dts/qcom/x1e80100-crd.dts
> +++ b/arch/arm64/boot/dts/qcom/x1e80100-crd.dts
> @@ -1125,6 +1125,30 @@ &mdss {
>  	status = "okay";
>  };
>  
> +&mdss_dp0 {
> +	status = "okay";
> +};
> +
> +&mdss_dp0_out {
> +	data-lanes = <0 1>;
> +};
> +
> +&mdss_dp1 {
> +	status = "okay";
> +};
> +
> +&mdss_dp1_out {
> +	data-lanes = <0 1>;
> +};
> +
> +&mdss_dp2 {
> +	status = "okay";
> +};
> +
> +&mdss_dp2_out {
> +	data-lanes = <0 1>;
> +};
> +
>  &mdss_dp3 {
>  	compatible = "qcom,x1e80100-dp";
>  	/delete-property/ #sound-dai-cells;
> 
> -- 
> 2.34.1
>
gregkh@linuxfoundation.org Nov. 13, 2024, 6:03 a.m. UTC | #2
On Tue, Nov 12, 2024 at 07:05:29PM +0200, Abel Vesa wrote:
> On 24-11-12 19:01:13, Abel Vesa wrote:
> > The X Elite CRD provides external DisplayPort on all 3 USB Type-C ports.
> > Each one of this ports is connected to a dedicated DisplayPort
> > controller.
> > 
> > Due to support missing in the USB/DisplayPort combo PHY driver,
> > the external DisplayPort is limited to 2 lanes.
> > 
> > So enable all 3 remaining DisplayPort controllers and limit their data
> > lanes number to 2.
> > 
> > Signed-off-by: Abel Vesa <abel.vesa@linaro.org>
> 
> Please do not merge this specific patch.

Please do not mix patches that should, and should not, be applied in the
same series as that plays havoc with our tools that want to take a whole
series at once.  Stuff like this just makes me delete the whole series
from my review queue, and if you got sent something like this, I imagine
you would do the same :(

thanks,

greg k-h