ksz8041 bring up

[MWoło.2]

It seems that I have a problem with DMA, but I'm not sure. Would you please take a look at the below screen

Dts looks

I intend to achieve MII with self-generated quartz.

Unfortunately, I'm stuck on this "Failed to reset the DMA".

[MWoło.2]

Hi Olivier, Hi Patrick,

I think I got what is wrong. I experimented with RMII, and when no 50MHz appears on the clock "...fb," I have a DMA problem. It means that this is not a problem with DMA but with the missing clock. Additionally, I tested DIV, MUL in PLL, and when set 100MHz or 25MHz for RMII, it also indicated a DMA problem. Only 50MHz pass for RMII. So after reading the documentation RM0436 Rev 5 section Clock distribution for Ethernet (ETH), everything seems precise. Unfortunately, I can't understand why I have no clock on lines.

Do You have a receipt for that? Why on RXC and TXC I notice nothing, it should be 25MHz on both, am I right? If Yes, how to achieve this?

[MWoło.2]

I measured both RXC and TXC using an oscilloscope, and is no sine waveform on both what I do wrong.

[PatrickF]

Hi,

both RXC and TXC come from PHY. So, some investigations are needed around PHY:

• double check you are really using the KSZ8041NL (and not RNL which does not support MII)
• thoroughly check each PHY pin (In particular, correct pin number, supplies and strapping options) on your schematics with :
  • PHY datasheet
  • PHY errata (e.g. https://ww1.microchip.com/downloads/en/DeviceDoc/80000700A.pdf)
  • PHY recommendations (e.g. https://ww1.microchip.com/downloads/en/DeviceDoc/KSZ8041NL-HW-Design-Checklist-00002498B.pdf)
  • PHY reference design (e.g. https://ww1.microchip.com/downloads/en/DeviceDoc/KSZ8041NL%20Eval%20Board%20rev1.1%20Schematic.zip).
• check correct PHY supplies (including the 1.8V output)
• Check 25MHz clock is present on PHY XI pin
• Check RST# pin is high
• etc...

From SW log, I understand KSZ8041 is discovered, so the MDIO bus is working. Maybe also linked to a microchip driver issue in MII mode.

Regards.