[U-Boot-Users] LAN chip dead on ARM Integrator/CP board

[Umar Qureshey]

Wolfgang Denk wrote:

>In message <43A9F197.9010603 at quartics.com> you wrote:
>  
>
>>1) I have tried setting up the MAC address.  However that does not fix 
>>the problem.  Output:
>>
>>SMC91111:  PHY auto-negotiate timed out
>>
>>Warning:  MAC addresses don't match:
>>          HW MAC address: 01:00:01:00:00:00
>>    
>>
>            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
>Where is this MAC address coming from? You are aware that this is  an
>address which will NOT work?
>
>  
>
Code inspection of smc91111.c shows that the driver is reading the MAC 
address from the ROM, if it detects one. 

>>          "ethaddr" value: 00:05:1B:00:10:4A
>>Using MAC Address 00:05:1B:00:10:4A
>>ping failed: host 192.168.168.157 is not alive
>>    
>>
>
>Did you see any packets on the wire? 
>  
>

I did not snoop the line but the link lights are dead.  I doubt sniffing 
packets would reveal much else.

>  
>
>>I have given the board a static IP address in order to bypass DHCP.
>>    
>>
>
>And which one was this? 
>
>  
>
>>Something is funky with the PHY layer because I do not see the link 
>>lights so I am not surprised no traffic is traversing the wireline.  It 
>>looks as if the chip is still in reset.
>>    
>>
>
>Given the fact that your hardware MAC address  is  illegal  (has  the
>multicast  bit  set)  you  should not be surprised. I'm not sure what
>your hardware does in such a situation.
>
>I recommend to fix your MAC addresses first.
>
>  
>
The 'ethaddr' environment variable would override any other MAC address, 
real or bunk.  If the chip is in a reset state then all bets are off as 
to the values being returned from the ROM.  This could explain why the 
MAC address from the ROM is bad. 
The chip looks dead.
Anyone know if there are certain FPGA settings that might need to be set 
on the Integrator/CP with ARM926EJS core module to wake up the LAN chip?

>>2) I did download u-boot 1.1.4 but the compilation failed with my 
>>arm-linux-gcc 3.4.2 (linked w/ uclibc) compiler.  Output:
>>    
>>
>
>Patches have been posted on the list; see the archives.
>
>  
>
I applied the patch but the compiler still puked when compiling 
cpu/arm926ejs/start.S.  It complained:

cc1: invalid option 'abi=apcs-gnu'

So I changed the cpu/arm926ejs/config.mk:

--- config.mk   2005-12-21 18:07:33.773204149 -0800
+++ config.mk.orig      2005-12-21 15:47:45.421844207 -0800
@@ -30,5 +30,5 @@
 # Supply options according to compiler version
 #
 # =========================================================================
-PLATFORM_CPPFLAGS +=$(call cc-option,-mapcs-32)
+PLATFORM_CPPFLAGS +=$(call cc-option,-mapcs-32,-mabi=apcs-gnu)
 PLATFORM_RELFLAGS +=$(call cc-option,-mshort-load-bytes,$(call cc-option,-malignment-traps,))


This got it to compile but 1.1.4 exhibits the same LAN problem.

Regards,
umar