440GX EMAC driver hang?

440GX EMAC driver hang?

[Scott Coulter]

[-- Attachment #1: Type: text/plain, Size: 1768 bytes --]

Hi,

I am doing a Linux 2.6.11.11 port for a CPCI board with a 440GX
processor on it.  Previously, I had successfully ported 2.6.11.11 to the
board's predecessor running on a 440GP processor.  Due to limitations in
this version of the PCB, EMAC ports 2 and 3 are not connected.  The
standalone kernel comes up and gets into a BusyBox shell.  When I try to
ifconfig the eth0 interface, the messages "eth0: Link is Up" and "eth0:
Speed: 100, Full duplex." are displayed and then the target hangs.  I
used the OCOTEA port as a pattern for my 440GP->440GX conversion.  The
PHYs which are connected to EMAC ports 0 and 1 are the same PHYs as on
the 440GP board (running in RMII mode).  I have tried to use the
function "ocp_remove_one_device" to whack EMAC2 and EMAC3.  I am pretty
sure that the removal worked as I no longer see init messages for those
EMACs (previously I saw all four devices get initialized).  The MAL
reports that it has "Initialized, 4 tx channels, 4 rx channels."  I used
the function "ibm440gx_set_eth_grp" to force the Ethernets to both group
0 and group 1 configurations.  So far, nothing has worked.  Can anyone
think of anything obvious that I am missing?  I know that the Ethernet
ports are functional as I used TFTP with our ROM monitor package to
download the zImage file.


Thanks,

Scott


___________________________________________________________________

  Scott N. Coulter
  Senior Software Engineer
  
  Cyclone Microsystems          
  370 James Street              Phone:  203.786.5536 ext. 118
  New Haven, CT 06513-3051      Email:  scott.coulter@cyclone.com
  U.S.A.                        Web:    http://www.cyclone.com
___________________________________________________________________


[-- Attachment #2: Type: text/html, Size: 4676 bytes --]

Re: 440GX EMAC driver hang?

[Eugene Surovegin]

On Wed, Jun 29, 2005 at 09:40:34AM -0400, Scott Coulter wrote:
> Hi,
> 
> I am doing a Linux 2.6.11.11 port for a CPCI board with a 440GX
> processor on it.  Previously, I had successfully ported 2.6.11.11 to the
> board's predecessor running on a 440GP processor.  Due to limitations in
> this version of the PCB, EMAC ports 2 and 3 are not connected.  The
> standalone kernel comes up and gets into a BusyBox shell.  When I try to
> ifconfig the eth0 interface, the messages "eth0: Link is Up" and "eth0:
> Speed: 100, Full duplex." are displayed and then the target hangs.  I
> used the OCOTEA port as a pattern for my 440GP->440GX conversion.  The
> PHYs which are connected to EMAC ports 0 and 1 are the same PHYs as on
> the 440GP board (running in RMII mode).  I have tried to use the
> function "ocp_remove_one_device" to whack EMAC2 and EMAC3.  I am pretty
> sure that the removal worked as I no longer see init messages for those
> EMACs (previously I saw all four devices get initialized).  The MAL
> reports that it has "Initialized, 4 tx channels, 4 rx channels."  I used
> the function "ibm440gx_set_eth_grp" to force the Ethernets to both group
> 0 and group 1 configurations.  So far, nothing has worked.  Can anyone
> think of anything obvious that I am missing?  I know that the Ethernet
> ports are functional as I used TFTP with our ROM monitor package to
> download the zImage file.

You can try using 4xx NAPI EMAC driver I wrote some time ago.
It can be found at http://kernel.ebshome.net/ (there is a patch and 
GIT tree which is kept up to date).

You can enable debug and mail me trace log privately, and maybe I can 
help you debug your problem.

-- 
Eugene