IMX219: Adjust PLL settings based on the number of MIPI lanes

[peyton-howe]

Still can tweak the exact PLL settings, but this adds the overlays and driver support needed for 4-lane operation. I have tested on my Pi5 and run with both 4-lane and 2-lane cameras simultaneously.

[6by9]

Links to https://forums.raspberrypi.com/viewtopic.php?t=381663

For reference, driver changes and overlay changes need to be in separate patches.

The overlay can be an override on the existing imx219 overlay to reduce duplication. My diff (which I can't test) doing that was

diff --git a/arch/arm/boot/dts/overlays/imx219-overlay.dts b/arch/arm/boot/dts/overlays/imx219-overlay.dts
index 4c4bcd309a3d..3411d5bbd8ce 100644
--- a/arch/arm/boot/dts/overlays/imx219-overlay.dts
+++ b/arch/arm/boot/dts/overlays/imx219-overlay.dts
@@ -65,6 +65,21 @@ csi_ep: endpoint {
                };
        };
 
+       fragment@201 {
+               target = <&csi_ep>;
+               __dormant__ {
+                       data-lanes = <1 2 3 4>;
+               };
+       };
+
+       fragment@202 {
+               target = <&cam_endpoint>;
+               __dormant__ {
+                       data-lanes = <1 2 3 4>;
+                       link-frequencies =
+                               /bits/ 64 <363000000>;
+       };
+
        __overrides__ {
                rotation = <&cam_node>,"rotation:0";
                orientation = <&cam_node>,"orientation:0";
@@ -77,6 +92,7 @@ __overrides__ {
                       <&vcm>, "VANA-supply:0=", <&cam0_reg>;
                vcm = <&vcm>, "status=okay",
                      <&cam_node>,"lens-focus:0=", <&vcm>;
+               4lane = <0>, "+201+202";
        };
 };

[6by9]

You're not handling any failures from this table write.

TBH you may as well move the IMX219_REG_CSI_LANE_MODE register into the array and just return the value from cci_multi_reg_write, or pass &ret as the last argument to cci_write below.

[peyton-howe]

Applied your recommended changes in the latest commit

[pelwell]

There's a missing line (another closing brace and semicolon) in the proposed fragment 202, but otherwise it looks okay.

[peyton-howe]

Removed overlay changes from this PR and updated driver based on 6by9's suggestion

[pelwell]

Can you give us a Signed-off-by line for the changes in your PRs (this and #6580)? Mine would be: Signed-off-by: Phil Elwell <[email protected]>

[peyton-howe]

Done for both!