• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 97 subscribers
  • Views 1820 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Boot, nfs server not responding

Helmut Forren
Mastermind 6805 points
Other Parts Discussed in Thread: THS8200

I'm using a DM6467 processor on a custom board derived from the EVM.

My prior prototype successfully boots using an nfs server served by Ubuntu running in a VM under Windows XP(MCE).

My new, modified prototype isn't being so successful. 

I switched back to the old one to test and make sure my Ubuntu server was working correctly, and it is.  Below are "works" and "does not work" console logs.

OLD WORKS:

[Wed Nov 23 15:18:46.187 2011] eth0: attached PHY driver [ET1011C] (mii_bus:phy_addr=1:01, id=282f014)

[Wed Nov 23 15:18:46.187 2011] Sending DHCP requests ....

[Wed Nov 23 15:19:01.125 2011] PHY: 1:01 - Link is Up - 100/Full

[Wed Nov 23 15:19:11.421 2011] .. timed out!

[Wed Nov 23 15:20:13.343 2011] IP-Config: Retrying forever (NFS root)...

[Wed Nov 23 15:20:13.968 2011] eth0: attached PHY driver [ET1011C] (mii_bus:phy_addr=1:01, id=282f014)

[Wed Nov 23 15:20:15.187 2011] Sending DHCP requests ....

[Wed Nov 23 15:20:35.953 2011] PHY: 1:01 - Link is Up - 100/Full

[Wed Nov 23 15:20:46.093 2011] ., OK

[Wed Nov 23 15:20:46.718 2011] IP-Config: Got DHCP answer from 192.168.123.33, my address is 192.168.123.19

[Wed Nov 23 15:20:46.796 2011] IP-Config: Complete:

[Wed Nov 23 15:20:46.796 2011]      device=eth0, addr=192.168.123.19, mask=255.255.255.0, gw=192.168.123.33,

[Wed Nov 23 15:20:46.796 2011]      host=192.168.123.19, domain=domain.invalid, nis-domain=(none),

[Wed Nov 23 15:20:46.796 2011]      bootserver=192.168.123.33, rootserver=192.168.123.18, rootpath=

[Wed Nov 23 15:20:46.796 2011] Looking up port of RPC 100003/2 on 192.168.123.18

[Wed Nov 23 15:20:46.796 2011] Looking up port of RPC 100005/1 on 192.168.123.18

[Wed Nov 23 15:20:46.796 2011] VFS: Mounted root (nfs filesystem) on device 0:13.

[Wed Nov 23 15:20:46.796 2011] Freeing init memory: 144K

[Wed Nov 23 15:20:47.140 2011] INIT: version 2.86 booting

[Wed Nov 23 15:20:47.625 2011] Please wait: booting...

NEW FAILS:

[Wed Nov 23 15:10:24.406 2011] eth0: attached PHY driver [ET1011C] (mii_bus:phy_addr=1:01, id=282f014)

[Wed Nov 23 15:10:25.625 2011] Sending DHCP requests ..

[Wed Nov 23 15:10:32.953 2011] PHY: 1:01 - Link is Up - 100/Full

[Wed Nov 23 15:10:33.968 2011] .... timed out!

[Wed Nov 23 15:11:59.156 2011] IP-Config: Retrying forever (NFS root)...

[Wed Nov 23 15:11:59.781 2011] eth0: attached PHY driver [ET1011C] (mii_bus:phy_addr=1:01, id=282f014)

[Wed Nov 23 15:12:01.015 2011] Sending DHCP requests ...

[Wed Nov 23 15:12:21.781 2011] PHY: 1:01 - Link is Up - 100/Full

[Wed Nov 23 15:12:22.734 2011] .,. OK

[Wed Nov 23 15:12:42.343 2011] IP-Config: Got DHCP answer from 192.168.123.33, my address is 192.168.123.19

[Wed Nov 23 15:12:42.343 2011] IP-Config: Complete:

[Wed Nov 23 15:12:42.359 2011]      device=eth0, addr=192.168.123.19, mask=255.255.255.0, gw=192.168.123.33,

[Wed Nov 23 15:12:42.359 2011]      host=192.168.123.19, domain=domain.invalid, nis-domain=(none),

[Wed Nov 23 15:12:42.359 2011]      bootserver=192.168.123.33, rootserver=192.168.123.18, rootpath=

[Wed Nov 23 15:12:42.359 2011] Looking up port of RPC 100003/2 on 192.168.123.18

[Wed Nov 23 15:12:42.359 2011] Looking up port of RPC 100005/1 on 192.168.123.18

[Wed Nov 23 15:12:58.156 2011] VFS: Mounted root (nfs filesystem) on device 0:13.

[Wed Nov 23 15:12:58.156 2011] Freeing init memory: 144K

[Wed Nov 23 15:13:15.593 2011] nfs: server 192.168.123.18 not responding, still trying

[Wed Nov 23 15:13:20.968 2011] nfs: server 192.168.123.18 not responding, still trying

Any suggestions?  While the old board started before and then keeps on going after its snippet I provided above, the new board starts before but then stops at the end of its snippet above.  Just in case, note that the hardware is not identical.  The old hardware had an ADV7171 for NTSC output, while the new hardware has a THS8200 for VGA output.  The kernel being booted is still the one for the old hardware.  But I wouldn't think that would matter. 
Thanks very much.

 

  • 0
    Guru 26430 points

    Seems the experts aren't responding. I'll state the obviously and say that the network layer is going down for some reason. In both cases, the RPC calls suceed which means NFS is good. The few posts on this forum about a similiar problem would imply some sort of mismatch between U-Boot, Linux network boot arg settings and maybe the /etc/network/interfaces file. Something to do with the network being reinitialized. I can't quite figure that out the posts. Perhaps you should compare the U-Boot environment variables on the two platforms and the common /etc/network/interfaces file on the NFS share.

     

  • 0 in reply to Norman Wong
    Mastermind 6805 points

    Norman, thanks very much for your assistance.  I figured out the solution early this morning.  Note that all settings were identical between the old and new boards, and I tested against the same server.  The old board worked successfully both before and after the new board failed.

    RESOLVED

    One of the considerations on my to-do list was the digital signal quality between the DaVinci proc and the PHY chip.  I had only ever inherited the EVM design.  Looking into it, 125MHz for 1000Base-T?  Darn, that's fast.  I new I had some signal routing meandering on the new board that didn't exist on the old board.  But 125MHz with 1ns rise time and 2ns setup time (whoosh!) is just too darned fast for my very nice 60MHz digital scope.  Comparative probing didn't reveal anything of use.  

    But 10Base-T runs at only 2.5MHz.  I couldn't find a cat-3 cable, and after all the higher speeds might barely run on that anyway.  I did find a 20 year old hub that supports only 10Base-T, however.  IT WORKED!

    So, my new board now works at 10Base-T, and I'm satisfied with that for now.  The likely culprit is those signal trace meanderings that grew between prototype revisions.  I can fix that on the next rev.