Re: [PATCH net-next v2 3/9] net: ethernet: oa_tc6: implement OA TC6 configuration function
From: <Parthiban.Veerasooran@microchip.com>
Date: 2023-10-27 07:10:56
Also in:
linux-devicetree, linux-doc, lkml
Hi Andrew, On 27/10/23 1:36 am, Andrew Lunn wrote:
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safequoted
quoted
quoted
-struct oa_tc6 *oa_tc6_init(struct spi_device *spi, bool prote) +struct oa_tc6 *oa_tc6_init(struct spi_device *spi)Was there a reason to have prote initially, and then remove it here?The reason is, control communication uses "protect". But in the first patch there was no dt used. Later in this patch, dt used for all the configuration parameters and this also part of that. That's why removed and moved this to dt configuration. What's your opinion? shall I keep as it is like this? or remove the protect in the first two patches and introduce in this patch?It will actually depend on what goes into the DT binding. If using protections costs very little, i would just hard code it on. Maybe you can run some iperf tests and see if it makes a measurable difference. How fast an SPI bus are you using on your development board? If you have a 50Mbps SPI bus, it does not even matter, since the media bandwidth is just 10Mbps.
Actually protection is only used for control communication to read/write registers. It is not used in data communication where ethernet frame transfer performed. So it doesn't hurt data traffic. But of course in between data communication I may perform some control transfer for register read/write but they are not big and will not affect the speed. In my development board I use 15MHz speed SPI bus. As this is given as a configurable parameter in the OPEN Alliance specification, I have implemented it in the DT binding for user input/choice. Best Regards, Parthiban V
Andrew