| Message ID | 20250401145759.3253736-1-niklas.soderlund+renesas@ragnatech.se (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | dt-bindings: media: i2c: imx219: Remove redundant description of data-lanes | expand |
On Tue, Apr 01, 2025 at 04:57:58PM +0200, Niklas Söderlund wrote: > The bindings already reference video-interfaces.yaml in the endpoint > node, there is no need to duplicate the description of the data-lanes > property. > > An array of physical data lane indexes. Position of an entry determines > the logical lane number, while the value of an entry indicates physical > lane, e.g. for 2-lane MIPI CSI-2 bus we could have "data-lanes = <1 2>;", > assuming the clock lane is on hardware lane 0. If the hardware does not > support lane reordering, monotonically incremented values shall be used > from 0 or 1 onwards, depending on whether or not there is also a clock > lane. This property is valid for serial busses only (e.g. MIPI CSI-2). Please do not quote bindings in commit. It's never helpful. > > What the generic binding do not cover is the behavior if the property > would be omitted. But the imx219 driver have never agreed with the > description neither. Before commit ceddfd4493b3 ("media: i2c: imx219: It did not have to agree. See discussion for v3 of patch adding this binding. > Support four-lane operation") the driver errored out if not 2 lanes > where used, and after it if not 2 or 4 lanes where used. Then... fix the driver? This property describes hardware, not driver. Why current driver implementation, e.g. 1 year ago or now, would change the hardware (so the bindings)? > > Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> > --- > Hello, > > The data-lanes property is a common property and the driver have always > operated as the common description, it seemed silly to break the driver > to adhere to odd specification, then to correct the bindings. However a > more correct solution would be to do the work on the driver of course. > > This is just a drive-by fix in the hope of sparing others the time to > discover this oddity themself. This is only tested by using the bindings > themself and by 'make dt_binding_check'. > --- > Documentation/devicetree/bindings/media/i2c/imx219.yaml | 9 --------- > 1 file changed, 9 deletions(-) > > diff --git a/Documentation/devicetree/bindings/media/i2c/imx219.yaml b/Documentation/devicetree/bindings/media/i2c/imx219.yaml > index 07d088cf66e0..31beeb2be2ea 100644 > --- a/Documentation/devicetree/bindings/media/i2c/imx219.yaml > +++ b/Documentation/devicetree/bindings/media/i2c/imx219.yaml > @@ -55,15 +55,6 @@ properties: > unevaluatedProperties: false > > properties: > - data-lanes: > - description: |- > - The sensor supports either two-lane, or four-lane operation. > - If this property is omitted four-lane operation is assumed. > - For two-lane operation the property must be set to <1 2>. > - items: > - - const: 1 > - - const: 2 So 1 lane is also fine? 8 lanes are as well? Previously lack of the property in DTS meant 4 lanes, now lack of property means anything. Best regards, Krzysztof
Hi Krzysztof, Thanks for your feedback. On 2025-04-02 10:21:42 +0200, Krzysztof Kozlowski wrote: > On Tue, Apr 01, 2025 at 04:57:58PM +0200, Niklas Söderlund wrote: > > The bindings already reference video-interfaces.yaml in the endpoint > > node, there is no need to duplicate the description of the data-lanes > > property. > > > > An array of physical data lane indexes. Position of an entry determines > > the logical lane number, while the value of an entry indicates physical > > lane, e.g. for 2-lane MIPI CSI-2 bus we could have "data-lanes = <1 2>;", > > assuming the clock lane is on hardware lane 0. If the hardware does not > > support lane reordering, monotonically incremented values shall be used > > from 0 or 1 onwards, depending on whether or not there is also a clock > > lane. This property is valid for serial busses only (e.g. MIPI CSI-2). > > Please do not quote bindings in commit. It's never helpful. > > > > > What the generic binding do not cover is the behavior if the property > > would be omitted. But the imx219 driver have never agreed with the > > description neither. Before commit ceddfd4493b3 ("media: i2c: imx219: > > It did not have to agree. See discussion for v3 of patch adding this binding. Thar discussion was in 2020, the common definition video-interfaces.yaml was merged in 2021 AFIK. > > > Support four-lane operation") the driver errored out if not 2 lanes > > where used, and after it if not 2 or 4 lanes where used. > > Then... fix the driver? > > This property describes hardware, not driver. Why current driver > implementation, e.g. 1 year ago or now, would change the hardware (so > the bindings)? I agree, I thought that here we have a case where the bindings predate the standardisation. The driver do not match the bindings, in fact it breaks if the imx219 specific instructions are followed. So the risk of breaking stuff is likely low. And this was an opportunity to align the imx219 with video-interfaces.yaml. I wasted time trying to use the imx219 bindings when bringing up a device, only wanted to try to help others avoid that. > > > > > Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> > > --- > > Hello, > > > > The data-lanes property is a common property and the driver have always > > operated as the common description, it seemed silly to break the driver > > to adhere to odd specification, then to correct the bindings. However a > > more correct solution would be to do the work on the driver of course. > > > > This is just a drive-by fix in the hope of sparing others the time to > > discover this oddity themself. This is only tested by using the bindings > > themself and by 'make dt_binding_check'. > > --- > > Documentation/devicetree/bindings/media/i2c/imx219.yaml | 9 --------- > > 1 file changed, 9 deletions(-) > > > > diff --git a/Documentation/devicetree/bindings/media/i2c/imx219.yaml b/Documentation/devicetree/bindings/media/i2c/imx219.yaml > > index 07d088cf66e0..31beeb2be2ea 100644 > > --- a/Documentation/devicetree/bindings/media/i2c/imx219.yaml > > +++ b/Documentation/devicetree/bindings/media/i2c/imx219.yaml > > @@ -55,15 +55,6 @@ properties: > > unevaluatedProperties: false > > > > properties: > > - data-lanes: > > - description: |- > > - The sensor supports either two-lane, or four-lane operation. > > - If this property is omitted four-lane operation is assumed. > > - For two-lane operation the property must be set to <1 2>. > > - items: > > - - const: 1 > > - - const: 2 > > So 1 lane is also fine? 8 lanes are as well? Previously lack of the > property in DTS meant 4 lanes, now lack of property means anything. Good point, if this patch where to be followed the data-lanes should be made a required property. > > Best regards, > Krzysztof >
On 02/04/2025 11:57, Niklas Söderlund wrote: >> >>> Support four-lane operation") the driver errored out if not 2 lanes >>> where used, and after it if not 2 or 4 lanes where used. >> >> Then... fix the driver? >> >> This property describes hardware, not driver. Why current driver >> implementation, e.g. 1 year ago or now, would change the hardware (so >> the bindings)? > > I agree, I thought that here we have a case where the bindings predate > the standardisation. The driver do not match the bindings, in fact it > breaks if the imx219 specific instructions are followed. So the risk of > breaking stuff is likely low. And this was an opportunity to align the > imx219 with video-interfaces.yaml. I am sorry, but what breaks exactly? Is the device supporting two and four lanes setups? If yes, then the binding is correct. Best regards, Krzysztof
On 2025-04-02 12:29:17 +0200, Krzysztof Kozlowski wrote: > On 02/04/2025 11:57, Niklas Söderlund wrote: > >> > >>> Support four-lane operation") the driver errored out if not 2 lanes > >>> where used, and after it if not 2 or 4 lanes where used. > >> > >> Then... fix the driver? > >> > >> This property describes hardware, not driver. Why current driver > >> implementation, e.g. 1 year ago or now, would change the hardware (so > >> the bindings)? > > > > I agree, I thought that here we have a case where the bindings predate > > the standardisation. The driver do not match the bindings, in fact it > > breaks if the imx219 specific instructions are followed. So the risk of > > breaking stuff is likely low. And this was an opportunity to align the > > imx219 with video-interfaces.yaml. > > I am sorry, but what breaks exactly? > > Is the device supporting two and four lanes setups? If yes, then the > binding is correct. I understand that is the most correct reading, this should likely have been posted as an RFC. The commit message states this was an attempt to see if it was possible to align the imx219 binding with the standard binding. The rational being that the imx219 bindings where created before we had common bindings for this and that the driver never worked with the imx219 version of the standard bindings so likely there would be no users of it. Kind of like how similar bindings where rejected for IMX708 [1]. I will drop this patch, it was only a drive-by thing as I had to spend time fighting this when trying to use the device. 1. https://lore.kernel.org/linux-media/949e3330-8c3d-6106-fbf8-cab820801cfc@kernel.org/
diff --git a/Documentation/devicetree/bindings/media/i2c/imx219.yaml b/Documentation/devicetree/bindings/media/i2c/imx219.yaml index 07d088cf66e0..31beeb2be2ea 100644 --- a/Documentation/devicetree/bindings/media/i2c/imx219.yaml +++ b/Documentation/devicetree/bindings/media/i2c/imx219.yaml @@ -55,15 +55,6 @@ properties: unevaluatedProperties: false properties: - data-lanes: - description: |- - The sensor supports either two-lane, or four-lane operation. - If this property is omitted four-lane operation is assumed. - For two-lane operation the property must be set to <1 2>. - items: - - const: 1 - - const: 2 - clock-noncontinuous: true link-frequencies: true
The bindings already reference video-interfaces.yaml in the endpoint node, there is no need to duplicate the description of the data-lanes property. An array of physical data lane indexes. Position of an entry determines the logical lane number, while the value of an entry indicates physical lane, e.g. for 2-lane MIPI CSI-2 bus we could have "data-lanes = <1 2>;", assuming the clock lane is on hardware lane 0. If the hardware does not support lane reordering, monotonically incremented values shall be used from 0 or 1 onwards, depending on whether or not there is also a clock lane. This property is valid for serial busses only (e.g. MIPI CSI-2). What the generic binding do not cover is the behavior if the property would be omitted. But the imx219 driver have never agreed with the description neither. Before commit ceddfd4493b3 ("media: i2c: imx219: Support four-lane operation") the driver errored out if not 2 lanes where used, and after it if not 2 or 4 lanes where used. Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> --- Hello, The data-lanes property is a common property and the driver have always operated as the common description, it seemed silly to break the driver to adhere to odd specification, then to correct the bindings. However a more correct solution would be to do the work on the driver of course. This is just a drive-by fix in the hope of sparing others the time to discover this oddity themself. This is only tested by using the bindings themself and by 'make dt_binding_check'. --- Documentation/devicetree/bindings/media/i2c/imx219.yaml | 9 --------- 1 file changed, 9 deletions(-)