Dear li guohu,

please keep the mailing list on Cc:

In message BLU166-W3176968F6747D75C62C676C6360@phx.gbl you wrote:

Thanks for your detailed help. I have solved the second question based on your answer.

Good.

...

...

1. Before the u-boot enter the command line, some error like "Error:

environment overflow " happens. The space in the ram used to save the environment variables are 0x2000. I think this is enough. I set this space in the head file like this: #define CON FIG_SYS_ENV_SIZE 0x2000.

Yes, I am using the old version of the u-boot. The version is U-Boot 2009.03. How can I deal with this warning? I mean, if I don't change a new version. Why this vwarning happens?

Eventually your assumption ("I think this is enough.") is wrong? Check the source code where and why this message is printed - grep is your friend.

...

...

3. After the u-boot enter the command line, I enter the command "mii

info". But nothing appears. If I enter the command "mii read 0x1 0x1", I can only see 0xFFFF. If I ping some address, I can see 32 lines if I use the "mii info" then. The address of th e PHY is 0x1.

You probably have some misconfiguration in your port.

I have checked the port configuration carefully. I think it is correct.

The golden rule is: if everything is correct, it will work. So there must be a problem somewhere.

The board I am using is made by myself based on the board TWR-MPC5125 made by freescale. This is a difficult question. I think maybe the hardware has some errors because I can't not see any output from the PHY. But I can see the signal from the CPU in the MDC pin and the MDIO pin of the PHY. Compared with the signals of the TWR-MPC5125, I think they are correct.

There is many things you may need (or fail) to configure for a PHY. How are clocks generated? Does it need a special reset handling, sya, through a GPIO or similar? Does it need any other initialization? Is it a supported or standard PHY type? There are so many questions we cannot help you with because we know neither the hardware nor the code.

So this is a MPC5125 based system? Well, code for this processor has never been mainlined; I remember that it is sufficiently different from the other members of the MPC512x family to make it largely incompatible. I would not be surprised if there were a number of bugs in that out-of-tree port.

--_8ea4e2c9-4b0d-4e26-9643-f0b60da03b1b_ Content-Type: text/html; charset="gb2312" Content-Transfer-Encoding: 8bit

Please stop posting HTML!!

Best regards,

Wolfgang Denk