Bug 7684 - forcedeth.c 0.59: problem with sideband managment
Summary: forcedeth.c 0.59: problem with sideband managment
Status: RESOLVED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Network (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: Ayaz Abdulla
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-12-15 01:12 UTC by Michael Reske
Modified: 2007-01-03 13:44 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.20-rc1
Subsystem:
Regression: ---
Bisected commit-id:


Attachments
ethtool -S eth0 (650 bytes, application/octet-stream)
2006-12-19 01:16 UTC, Michael Reske
Details
ethtool -d eth0 (13.99 KB, application/octet-stream)
2006-12-19 01:17 UTC, Michael Reske
Details
lspci -xxx (27.70 KB, application/octet-stream)
2006-12-19 01:17 UTC, Michael Reske
Details
dmesg (60.03 KB, application/octet-stream)
2006-12-19 01:17 UTC, Michael Reske
Details
Mgmt unit fixes (147.78 KB, text/plain)
2006-12-20 17:55 UTC, Ayaz Abdulla
Details
Mgmt unit fixes (147.77 KB, text/plain)
2006-12-20 17:58 UTC, Ayaz Abdulla
Details
Kernelmessages with the new version (58.51 KB, text/plain)
2007-01-02 01:08 UTC, Michael Reske
Details
Mgmt Fix for gigabit phy (147.95 KB, text/plain)
2007-01-03 10:54 UTC, Ayaz Abdulla
Details
This is the patch for the sideband management fixes. (6.93 KB, patch)
2007-01-03 13:41 UTC, Ayaz Abdulla
Details | Diff

Description Michael Reske 2006-12-15 01:12:17 UTC
Most recent kernel where this bug did *NOT* occur: 2.6.19-git2
Distribution:
Hardware Environment: Dual Opteron 2214 on Supermicro H8DMR-82 (nVidia MCP55)
Software Environment:
Problem Description: Since the upgrade from 0.57 to 0.59 eth0 doesn't responds 
for any request. The driver can initialize the device and this is recognizing 
link up/down but no packets can sent or received.

Steps to reproduce: Removing of DEV_HAS_MGMT_UNIT fix this problem.
Comment 1 Ayaz Abdulla 2006-12-18 15:30:48 UTC
Could you enable all dprintk messages by setting the #if 0 to #if 1 near the 
top of the source file?
Could you post output of the following:
- "ethtool -S ethX" statistics
- "ethtool -d ethX" register dump.
- dmesg
- "lspci -xxx"
Comment 2 Michael Reske 2006-12-19 01:16:38 UTC
Created attachment 9874 [details]
ethtool -S eth0
Comment 3 Michael Reske 2006-12-19 01:17:06 UTC
Created attachment 9875 [details]
ethtool -d eth0
Comment 4 Michael Reske 2006-12-19 01:17:30 UTC
Created attachment 9876 [details]
lspci -xxx
Comment 5 Michael Reske 2006-12-19 01:17:57 UTC
Created attachment 9877 [details]
dmesg
Comment 6 Ayaz Abdulla 2006-12-20 17:55:01 UTC
Created attachment 9906 [details]
Mgmt unit fixes

Can you try out this file which contains mgmt unit fixes?
Comment 7 Ayaz Abdulla 2006-12-20 17:58:24 UTC
Created attachment 9907 [details]
Mgmt unit fixes

Here is the latest file to try out.
Comment 8 Michael Reske 2007-01-02 01:08:46 UTC
Created attachment 9990 [details]
Kernelmessages with the new version

I tried the new driver, but it doesn't work to.
Comment 9 Ayaz Abdulla 2007-01-02 08:15:08 UTC
Do both ethernet ports not work? or just one of them? The sideband managment 
will only choose one of the ports to use, however, it will initialize the phy 
on both macs. It seems that something got messed up when it initalized the phy.

Can you send me output of "ethtool eth0" so I can verify speed/duplex settings 
with hw registers? Also, you are not forcing the speed/duplex right?
Comment 10 Michael Reske 2007-01-02 08:25:46 UTC
Only eth0 is affected. In the ethtool-output you can see the wrong speed 
(1000Mb/s expected).

Without managment:
Settings for eth0:
        Supported ports: [ MII ]
        Supported link modes:   10baseT/Half 10baseT/Full 
                                100baseT/Half 100baseT/Full 
                                1000baseT/Full 
        Supports auto-negotiation: Yes
        Advertised link modes:  10baseT/Half 10baseT/Full 
                                100baseT/Half 100baseT/Full 
                                1000baseT/Full 
        Advertised auto-negotiation: Yes
        Speed: 1000Mb/s
        Duplex: Full
        Port: MII
        PHYAD: 2
        Transceiver: externel
        Auto-negotiation: on
        Supports Wake-on: g
        Wake-on: d
        Link detected: yes

With managment:
Settings for eth0:
        Supported ports: [ MII ]
        Supported link modes:   10baseT/Half 10baseT/Full 
                                100baseT/Half 100baseT/Full 
        Supports auto-negotiation: Yes
        Advertised link modes:  10baseT/Half 10baseT/Full 
                                100baseT/Half 100baseT/Full 
        Advertised auto-negotiation: Yes
        Speed: 100Mb/s
        Duplex: Full
        Port: MII
        PHYAD: 2
        Transceiver: externel
        Auto-negotiation: on
        Supports Wake-on: g
        Wake-on: d
        Link detected: yes

I'm not forcing any of the speed/duplex settings.
Comment 11 Ayaz Abdulla 2007-01-03 10:54:38 UTC
Created attachment 9999 [details]
Mgmt Fix for gigabit phy

I believe I found the issue. Try this attachemnt.
Comment 12 Michael Reske 2007-01-03 11:49:15 UTC
Ok, the latest version works fine for me :-).

Thanks Ayaz .
Comment 13 Ayaz Abdulla 2007-01-03 11:50:56 UTC
Great! I will create a patch and submit to kernel maintainers.

Thanks for your help!
Comment 14 Ayaz Abdulla 2007-01-03 13:41:44 UTC
Created attachment 10000 [details]
This is the patch for the sideband management fixes.

Note You need to log in before you can comment on or make changes to this bug.