Fedora Linux Support Community & Resources Center
  #1  
Old 28th October 2008, 03:26 AM
walshy007 Offline
Registered User
 
Join Date: Jun 2005
Posts: 30
Ethernet dropping

Hello, recently upgraded to fedora 9, and it seems that periodically my ethernet drops, and then reconnects within about 20 seconds, happens every 20 minutes or so thereabouts.

it's using the skge driver with a 3Com Corporation 3c940 10/100/1000Base-T [Marvell]

related dmesg entries as follows

skge eth0: Link is up at 100 Mbps, full duplex, flow control both
skge eth0: Link is down.
skge eth0: Link is up at 100 Mbps, full duplex, flow control both
skge eth0: Link is down.
skge eth0: Link is up at 100 Mbps, full duplex, flow control both
skge eth0: Link is down.
skge eth0: Link is up at 100 Mbps, full duplex, flow control both
skge eth0: Link is down.
skge eth0: Link is up at 100 Mbps, full duplex, flow control both
skge eth0: Link is down.

fairly useless, so onto /var/log/messages

Oct 28 20:41:42 localhost kernel: skge eth0: Link is down.
Oct 28 20:41:42 localhost NetworkManager: <info> (eth0): carrier now OFF (device state 8)
Oct 28 20:41:42 localhost NetworkManager: <info> (eth0): device state change: 8 -> 2
Oct 28 20:41:42 localhost NetworkManager: <info> (eth0): deactivating device.
Oct 28 20:41:42 localhost NetworkManager: <info> eth0: canceled DHCP transaction, dhcp client pid 8359
Oct 28 20:41:42 localhost NetworkManager: <WARN> check_one_route(): (eth0) error -34 returned from rtnl_route_del(): Missing link name TLV (errno = Invalid argument)
Oct 28 20:41:44 localhost kernel: skge eth0: Link is up at 100 Mbps, full duplex, flow control both


after that it just goes into dhcp etc. which is still fairly useless,

anyone know how to get a more verbose log from the skge driver?

others seem to have had my problem before, found http://www.linuxquestions.org/questi...-up....-566342 others have had the same problem before, but no resolution found
Reply With Quote
  #2  
Old 28th October 2008, 09:10 PM
scotty38's Avatar
scotty38 Offline
Registered User
 
Join Date: Mar 2008
Location: Nottingham, UK
Age: 18
Posts: 270
Given you have a wired ethernet connection I'd give these a go and see how you get on:

chkconfig NetworkManager off
chkconfig network on
Reply With Quote
  #3  
Old 29th October 2008, 02:24 AM
walshy007 Offline
Registered User
 
Join Date: Jun 2005
Posts: 30
I wound up going

rmmod skge
modprobe skge debug=16

and while it usually takes about 20-30 seconds to reconnect, it now takes about two seconds.
here was the debug output


eth0: tx queued, slot 107, len 66
skge eth0: tx done slot 107
skge eth0: rx slot 329 status 0x580100 len 88
eth0: tx queued, slot 108, len 66
skge eth0: tx done slot 108
eth0: tx queued, slot 109, len 42
skge eth0: tx done slot 109
skge eth0: rx slot 330 status 0x3c0100 len 60
skge eth0: phy interrupt status 0x700 0x814c
skge eth0: Link is down.
skge eth0: phy interrupt status 0x1c40 0x7d4c
skge eth0: Link is up at 100 Mbps, full duplex, flow control both

my new mission is to understand what triggers interrupt 0x700 0x814c

I don't really expect anyone here to know about this, but I'll post here with what I find, if I find anything, and any assistance is welcome
Reply With Quote
  #4  
Old 29th October 2008, 04:37 AM
walshy007 Offline
Registered User
 
Join Date: Jun 2005
Posts: 30
I have discovered, that of
skge eth0: phy interrupt status 0x700 0x814c

0x700 is the istatus, read from 16 bit register PHY_MARV_INT_STAT, as defined in skge.h
0x814c is the phystat read from 16 bit register PHY_MARV_PHY_STAT in skge.h

they are read by gm_phy_read()

while the definitions of the registers yield the memory offset from gmac to allow indirect addressing, I have yet to find definitions of what each bit signifies in the status registers mentioned.

I am sure it must be defined in skge.h, any assistance is appreciated.
Reply With Quote
  #5  
Old 29th October 2008, 05:05 AM
walshy007 Offline
Registered User
 
Join Date: Jun 2005
Posts: 30
Code:
/*****  PHY_MARV_PHY_STAT	16 bit r/o	PHY Specific Status Reg *****/
enum {
	PHY_M_PS_SPEED_MSK	= 3<<14, /* Bit 15..14: Speed Mask */
	PHY_M_PS_SPEED_1000	= 1<<15, /*		10 = 1000 Mbps */
	PHY_M_PS_SPEED_100	= 1<<14, /*		01 =  100 Mbps */
	PHY_M_PS_SPEED_10	= 0,	 /*		00 =   10 Mbps */
	PHY_M_PS_FULL_DUP	= 1<<13, /* Full Duplex */
	PHY_M_PS_PAGE_REC	= 1<<12, /* Page Received */
	PHY_M_PS_SPDUP_RES	= 1<<11, /* Speed & Duplex Resolved */
	PHY_M_PS_LINK_UP	= 1<<10, /* Link Up */
	PHY_M_PS_CABLE_MSK	= 7<<7,  /* Bit  9.. 7: Cable Length Mask */
	PHY_M_PS_MDI_X_STAT	= 1<<6,  /* MDI Crossover Stat (1=MDIX) */
	PHY_M_PS_DOWNS_STAT	= 1<<5,  /* Downshift Status (1=downsh.) */
	PHY_M_PS_ENDET_STAT	= 1<<4,  /* Energy Detect Status (1=act) */
	PHY_M_PS_TX_P_EN	= 1<<3,  /* Tx Pause Enabled */
	PHY_M_PS_RX_P_EN	= 1<<2,  /* Rx Pause Enabled */
	PHY_M_PS_POL_REV	= 1<<1,  /* Polarity Reversed */
	PHY_M_PS_JABBER		= 1<<0,  /* Jabber */
};
definition of one of them was right in front of me

physical status definition found, plugging 814c into that yields
speed and duplex not resolved and no link up, rest of info doesnt matter when there's no link

so I've found the physical state when it first warns of going down (goes down.. obviously) so all that's left is the finding of what the interrupt status yields, if that doesn't give any useful information I'll have to dig deeper

Last edited by walshy007; 29th October 2008 at 05:06 AM. Reason: code tags didn't work
Reply With Quote
Reply

Tags
dropping, ethernet

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
How to know if my ISP is dropping packets savage Wibble 0 15th February 2009 06:38 PM
Azureus Dropping FC5 onthefritz Servers & Networking 22 7th April 2006 05:35 AM
FC4 - Dropping all TX packets emzer Servers & Networking 0 29th June 2005 01:34 PM
Jaw-Dropping tejas Linux Chat 1 10th May 2005 08:00 AM
Using APF for MAC address dropping? copland007 Security and Privacy 0 12th September 2004 10:01 PM


Current GMT-time: 15:13 (Saturday, 25-05-2013)

TopSubscribe to XML RSS for all Threads in all ForumsFedoraForumDotOrg Archive
logo

All trademarks, and forum posts in this site are property of their respective owner(s).
FedoraForum.org is privately owned and is not directly sponsored by the Fedora Project or Red Hat, Inc.

Privacy Policy | Term of Use | Posting Guidelines | Archive | Contact Us | Founding Members

Powered by vBulletin® Copyright ©2000 - 2012, vBulletin Solutions, Inc.

FedoraForum is Powered by RedHat