linux-mips
[Top] [All Lists]

Re: linux 2.6.33 on movidis x16

To: David Daney <ddaney@caviumnetworks.com>
Subject: Re: linux 2.6.33 on movidis x16
From: Jan Rovins <janr@adax.com>
Date: Fri, 19 Mar 2010 18:38:17 -0400
Cc: Peter 'p2' De Schrijver <p2@debian.org>, linux-mips@linux-mips.org
In-reply-to: <4BA3F4AC.5000307@caviumnetworks.com>
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <20100305141113.GD2437@apfelkorn> <4B9EB45D.8050106@adax.com> <20100318181734.GG2437@apfelkorn> <4BA27048.2010707@caviumnetworks.com> <4BA277FB.6080808@adax.com> <20100319175244.GH2437@apfelkorn> <4BA3DE2D.1070801@adax.com> <4BA3F4AC.5000307@caviumnetworks.com>
Sender: linux-mips-bounce@linux-mips.org
User-agent: Thunderbird 2.0.0.24 (Windows/20100228)
David Daney wrote:
On 03/19/2010 01:27 PM, Jan Rovins wrote:


      case CVMX_BOARD_TYPE_KBP: /* JLR first Movidis X16 */
    case CVMX_BOARD_TYPE_CUST_WSX16: /* JLR second Movidis X16 */
/* Board has 8 SGMII ports. 4 connect out, two connect to a switch,
               and 2 loop to each other */
           if ((ipd_port >= 0) && (ipd_port < 4))
               return ipd_port+1;
           else
               return -1;

I don't think that is correct.  I just verified that that PHYs are on
addresses 0-7.  That code makes them start at 1.

The Octeon mdio bus driver also needs a small change.

I am working on a patch.
You are most likely right.
using ipd_port+1 is probably what made my eth0 show up on the 2nd RJ45 port. We were just using those boxes for early development, to verify basic Octeon functionality of some of our hardware drivers, so once an Ethernet port started working, I did not look into it any further, since it was a temporary setup just to get things going.

I looked for the original source from Movidis, and all I found in their kernel source were broken symlinks to a tool chain they that they did not give me :-(

Jan


David Daney



<Prev in Thread] Current Thread [Next in Thread>