[U-Boot] Fw: TFTP fails when using network switch

[Loren A. Linden Levy]

Hi Jerry,

It looks to me like for the M%282 board the PHY address is determined
by querying the MII interface:

include/configs/M5282EVB.h

#define CONFIG_MCFFEC
#ifdef CONFIG_MCFFEC
#       define CONFIG_NET_MULTI         1
#       define CONFIG_MII               1
#       define CFG_DISCOVER_PHY
...

This cause the code in ./drivers/net/mcffec.c specifically
mii_discover_phy to be run, I tried to enable the debugging in this
code:

//#undef        ET_DEBUG
//#undef        MII_DEBUG

but it does not seem to print anything (:(). One interesting feature is
that once my uClinux image has booted I can talk to it over the
switched network just fin. So the issue is really something in the
setup of the MII PHY interface in u-boot. Is there any build target to
allow me to try and debug this?

Thanks.

Alex


On Mon, 19 Jan 2009 20:14:07 -0500
Jerry Van Baren <gvb.uboot at gmail.com> wrote:

> Hi Loren,
> 
> Please don't top post.  It annoys us curmudgeons.
> 
> Loren A. Linden Levy wrote:
> > Hi All,
> > 
> > I still have not found a solution and I am a bit desperate. I did not
> > seem to be able to send any packets (using ping) from the FECO on the
> > M5282Lite board to the tftp host, however pings from the host were
> > answered and I saw packets going from the coldfire board to the host
> > in wireshark. I also get a strange error when I ask for mii info:
> > 
> > uBOOT=>> mii device
> > MII devices: 'FEC0' 
> > Current device: 'FEC0
> > 
> > uBOOT=>> mii info
> > MII not complete
> > MII not complete
> > ...
> > 
> > a dump gives me:
> > 
> > uBOOT=>> mii dump
> > MII not complete
> > 0.     (ffff)                 -- PHY control register --
> >  (8000:8000) 0.15    =     1    reset
> >  (4000:4000) 0.14    =     1    loopback
> >  (2040:2040) 0. 6,13 =   b11    speed selection = ??? Mbps
> >  (1000:1000) 0.12    =     1    A/N enable
> >  (0800:0800) 0.11    =     1    power-down
> >  (0400:0400) 0.10    =     1    isolate
> >  (0200:0200) 0. 9    =     1    restart A/N
> >  (0100:0100) 0. 8    =     1    duplex = full
> >  (0080:0080) 0. 7    =     1    collision test enable
> >  (003f:003f) 0. 5- 0 =    63    (reserved)
> 
> Note that ALL the bits are '1'.  That is indicating your PHY is NOT 
> DRIVING data on the MII data line (always idle high).  Odds are really 
> good that you are not driving your MII clock and data properly, your PHY 
> isn't at the address you think it is at, or your hardware has problems. 
>   The only reason you are getting any ether activity is likely because 
> your PHY is defaulting to a configuration that is mostly (not entirely) 
> broken.
> 
> * Double-check your MII clock and data pin configurations.  Are you 
> initializing them correctly?  Error probability: 87%.
> 
> * What address is your PHY strapped to?  You need to know this.  Error 
> probability: 11% (your emails imply you don't know this - I bumped the 
> probability up because of that impression).
> 
> * Use a logic analyzer (preferred), o-scope (good) or a logic probe on 
> your PHY's MII clock and data lines and verify they are receiving the 
> right sequence (or at least wiggling when you do the MII commands). 
> With a logic analyzer you should be able to read the data you are 
> sending and verify the bit pattern.  You can do this with an o-scope 
> (especially a digital one) too, but a little trickier.  This is likely a 
> software error (above 87% + 11%) - hardware error probability: 2%
> 
> * Oh yeah, you might be clocking the MII lines too fast.  Pretty small 
> probability, though.
> 
> [snip]
> 
> Good luck and happy probing,
> gvb
> 
> P.S. I used mostly odd percentages because I read somewhere that made up 
> percentages are more believable when they are odd.  ;-)


-- 
 -- 
---------------------------------------------
Loren A. Linden Levy
Department of Physics 
390 UCB 
University of Colorado
Boulder, CO   80309-0390
Tel:   303-735-6146 (CU) / +049 040 8998 4789 (DESY) 
Fax:   303-492-3352 (CU) / +049 040 8998 4034 (DESY) 
Cell:  303-332-2768 (U.S.) / +049 (0)151 5496 1831 (Germany)
Email: Loren.Lindenlevy at colorado.edu
url:   http://up.colorado.edu/~lindenle
_____________________________________________
---------------------------------------------
This email has been cryptographically signed.
Search for "lindenle" at pgp.mit.edu
to obtain my public key which can be used 
to verify the authenticity of this message.
_____________________________________________
---------------------------------------------
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
Url : http://lists.denx.de/pipermail/u-boot/attachments/20090120/610c8b41/attachment.pgp