| Message ID | 20240517171103.221856-1-andriy.shevchenko@linux.intel.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 3aac9f4885922ad0fc01b86f85903768219475a3 |
| Headers | show |
| Series | [v1,1/1] spi: pxa2xx: Move PXA SSP bindings to the correct folder | expand |
On Fri, May 17, 2024 at 08:11:03PM +0300, Andy Shevchenko wrote: > SSP stands for Serial Synchronous Protocol and has nothing to do with > UART, also known as USART, where 'A' stands for Asynchronous. > > Move the SSP bindings to where it belongs. It's a serial device which is also used for other applications (the other one upstream being audio) so I can see where the current binding comes from and it's not super obvious that spi is especially better here.
On Fri, May 17, 2024 at 06:24:37PM +0100, Mark Brown wrote: > On Fri, May 17, 2024 at 08:11:03PM +0300, Andy Shevchenko wrote: > > SSP stands for Serial Synchronous Protocol and has nothing to do with > > UART, also known as USART, where 'A' stands for Asynchronous. > > > > Move the SSP bindings to where it belongs. > > It's a serial device which is also used for other applications (the > other one upstream being audio) so I can see where the current binding > comes from and it's not super obvious that spi is especially better > here. Hmm... okay. Then it's question to DT people. Consider this as a report. Because UART (aka serial) is definitely not the place for SPI/SSP bindings either.
On Fri, May 17, 2024 at 2:58 PM Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote: > > On Fri, May 17, 2024 at 06:24:37PM +0100, Mark Brown wrote: > > On Fri, May 17, 2024 at 08:11:03PM +0300, Andy Shevchenko wrote: > > > SSP stands for Serial Synchronous Protocol and has nothing to do with > > > UART, also known as USART, where 'A' stands for Asynchronous. > > > > > > Move the SSP bindings to where it belongs. > > > > It's a serial device which is also used for other applications (the > > other one upstream being audio) so I can see where the current binding > > comes from and it's not super obvious that spi is especially better > > here. > > Hmm... okay. Then it's question to DT people. Consider this as a report. > Because UART (aka serial) is definitely not the place for SPI/SSP bindings > either. Move it when it is converted. Until then, I don't care too much. SPI seems better than serial at least. The sound part is its own binding/node (something we wouldn't do today). Rob
On Fri, May 17, 2024 at 03:19:51PM -0500, Rob Herring wrote: > On Fri, May 17, 2024 at 2:58 PM Andy Shevchenko > <andriy.shevchenko@linux.intel.com> wrote: > > On Fri, May 17, 2024 at 06:24:37PM +0100, Mark Brown wrote: > > > On Fri, May 17, 2024 at 08:11:03PM +0300, Andy Shevchenko wrote: ... > > > > SSP stands for Serial Synchronous Protocol and has nothing to do with > > > > UART, also known as USART, where 'A' stands for Asynchronous. > > > > > > > > Move the SSP bindings to where it belongs. > > > > > > It's a serial device which is also used for other applications (the > > > other one upstream being audio) so I can see where the current binding > > > comes from and it's not super obvious that spi is especially better > > > here. > > > > Hmm... okay. Then it's question to DT people. Consider this as a report. > > Because UART (aka serial) is definitely not the place for SPI/SSP bindings > > either. > > Move it when it is converted. The problem is that somebody added a binding (in YAML) for SPI PXA2xx in the spi/ folder while this one kept unconverted. If it dangles more, it might be that we will have two asynchronous bindings for the co-existed drivers. > Until then, I don't care too much. SPI seems better than serial at least. > The sound part is its own binding/node (something we wouldn't do today).
On Mon, May 20, 2024 at 01:47:26PM +0300, Andy Shevchenko wrote: > On Fri, May 17, 2024 at 03:19:51PM -0500, Rob Herring wrote: > > On Fri, May 17, 2024 at 2:58 PM Andy Shevchenko > > <andriy.shevchenko@linux.intel.com> wrote: > > > On Fri, May 17, 2024 at 06:24:37PM +0100, Mark Brown wrote: > > > > On Fri, May 17, 2024 at 08:11:03PM +0300, Andy Shevchenko wrote: > > ... > > > > > > SSP stands for Serial Synchronous Protocol and has nothing to do with > > > > > UART, also known as USART, where 'A' stands for Asynchronous. > > > > > > > > > > Move the SSP bindings to where it belongs. > > > > > > > > It's a serial device which is also used for other applications (the > > > > other one upstream being audio) so I can see where the current binding > > > > comes from and it's not super obvious that spi is especially better > > > > here. > > > > > > Hmm... okay. Then it's question to DT people. Consider this as a report. > > > Because UART (aka serial) is definitely not the place for SPI/SSP bindings > > > either. > > > > Move it when it is converted. > > The problem is that somebody added a binding (in YAML) for SPI PXA2xx > in the spi/ folder while this one kept unconverted. Ah, well that detail was missed. > > If it dangles more, it might be that we will have two asynchronous bindings > for the co-existed drivers. Looks like all that is needed is adding the compatible strings and 'dmas' property to spi/marvell,mmp2-ssp.yaml. The examples in the old binding have other stuff, but looks like that's garbage. Rob
On Mon, May 20, 2024 at 03:36:00PM -0500, Rob Herring wrote: > On Mon, May 20, 2024 at 01:47:26PM +0300, Andy Shevchenko wrote: > > On Fri, May 17, 2024 at 03:19:51PM -0500, Rob Herring wrote: > > > On Fri, May 17, 2024 at 2:58 PM Andy Shevchenko > > > <andriy.shevchenko@linux.intel.com> wrote: > > > > On Fri, May 17, 2024 at 06:24:37PM +0100, Mark Brown wrote: > > > > > On Fri, May 17, 2024 at 08:11:03PM +0300, Andy Shevchenko wrote: ... > > > > > > SSP stands for Serial Synchronous Protocol and has nothing to do with > > > > > > UART, also known as USART, where 'A' stands for Asynchronous. > > > > > > > > > > > > Move the SSP bindings to where it belongs. > > > > > > > > > > It's a serial device which is also used for other applications (the > > > > > other one upstream being audio) so I can see where the current binding > > > > > comes from and it's not super obvious that spi is especially better > > > > > here. > > > > > > > > Hmm... okay. Then it's question to DT people. Consider this as a report. > > > > Because UART (aka serial) is definitely not the place for SPI/SSP bindings > > > > either. > > > > > > Move it when it is converted. > > > > The problem is that somebody added a binding (in YAML) for SPI PXA2xx > > in the spi/ folder while this one kept unconverted. > > Ah, well that detail was missed. > > > > > If it dangles more, it might be that we will have two asynchronous bindings > > for the co-existed drivers. > > Looks like all that is needed is adding the compatible strings and > 'dmas' property to spi/marvell,mmp2-ssp.yaml. The examples in the old > binding have other stuff, but looks like that's garbage. I'm not an expert in DT, anybody to join them in a nicest possible way?
diff --git a/Documentation/devicetree/bindings/serial/mrvl,pxa-ssp.txt b/Documentation/devicetree/bindings/spi/mrvl,pxa-ssp.txt
similarity index 100%
rename from Documentation/devicetree/bindings/serial/mrvl,pxa-ssp.txt
rename to Documentation/devicetree/bindings/spi/mrvl,pxa-ssp.txt
SSP stands for Serial Synchronous Protocol and has nothing to do with UART, also known as USART, where 'A' stands for Asynchronous. Move the SSP bindings to where it belongs. Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> --- .../devicetree/bindings/{serial => spi}/mrvl,pxa-ssp.txt | 0 1 file changed, 0 insertions(+), 0 deletions(-) rename Documentation/devicetree/bindings/{serial => spi}/mrvl,pxa-ssp.txt (100%)