Skge broken for Marvell network chip

[Ulrik Mikaelsson]

Binary package hint: linux-image-2.6.22-9-generic

Since upgrade to Gutsy, a few weeks ago, I've been having problem with random lockups in the skge driver. They occur with varying frequency, usually once every two days or so, but sometimes a little more often. Yesterday I managed to more or less exactly reproduce them after ~5 minutes of video-streaming from a Mac Mini.

When the problem arises, everything in the system looks normal, IP is assigned and so on, but nothing is reachable that should be. DHCP lease-renew don't work either.

To work-around, reloading the skge module works, making me strongly suspect the driver rather than hardware.

Found in logs:
kernel: [323028.012934] NETDEV WATCHDOG: internet: transmit timed out

[Ulrik Mikaelsson (rawler)]

Relevant lshw;

     *-pci:0
          description: PCI bridge
          product: CK804 PCI Bridge
          vendor: nVidia Corporation
          physical id: 9
          bus info: pci@0000:00:09.0
          version: a2
          width: 32 bits
          clock: 66MHz
          capabilities: pci subtractive_decode bus_master
        *-network
             description: Ethernet interface
             product: 88E8001 Gigabit Ethernet Controller
             vendor: Marvell Technology Group Ltd.
             physical id: c
             bus info: pci@0000:05:0c.0
             logical name: internet
             version: 13
             serial: 00:11:d8:94:8e:7f
             size: 100MB/s
             capacity: 1GB/s
             width: 32 bits
             clock: 66MHz
             capabilities: pm vpd bus_master cap_list ethernet physical tp 10bt 10bt-fd 100bt 100bt-fd 1000bt 1000bt-fd autonegotiation
             configuration: autonegotiation=on broadcast=yes driver=skge driverversion=1.11 duplex=full firmware=N/A ip=85.225.79.229 latency=32 link=yes maxlatency=31 mingnt=23 module=skge multicast=yes port=twisted pair speed=100MB/s

[Ulrik Mikaelsson (rawler)]

Also note; this is a regression from Feisty.

[Ulrik Mikaelsson (rawler)]

Bug #37784 seems related, although the symptoms differ quite a lot.

I have no problem with large uploads, as is described. In fact, when video-streaming I managed to hang the driver pretty repeatedly, while when copying the video material over to the Mac, everything went smooth.

Also, for me I did not have the problems in Feisty (have never run anything previous to Feisty on this HW), but started occurring on upgrade to Gutsy.

[Brian Murray (brian-murray)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. It would be helpful to have more than just the output of 'dmesg" as it seems to be missing a fair bit at the beginning. So if you could reproduce the issue and then attach '/var/log/messages' to your bug report that would be great. Thanks in advance.

[Ulrik Mikaelsson (rawler)]

Allright. The bug, after not being produced for a long while, occurred again tonight.

Here's the wanted messages-log.

[Nono (douillemaster)]

Hi ! I'm french so please scuse me for my poor english. Last night I tried to install a Feisty Fawn release on a computer with an Asus P4P800 Deluxe motherboard and I experienced a strange behavior of my NIC.. The installation procedure worked fine until the progression reached 82% : Mirrors Checking... (not sure it's the english message it was a french installation), then the cpu seemed to stop working and nothing happens after (but the pc was fine : no freeze, no bug, no error message). The only solution I've found to avoid this problem was desactivating the eth0 interface..
After the reboot I checked eth0, it seemed to work fine so I opened a terminal and launch a "apt-get update" with no results ( hanging on the first "getting headers..." with no errors messages). I opened Firefox, got google.com and launch a request : no problem at all... But after a few clicks I realised that only one website in 10 attempts really opened. Even Ubuntu.com was unreachable. I opened a Terminal and ping all of these websites and I discovered that they were all ok. The NIC of the P4P800 Deluxe is a 3Com 3c940 (10/100/1000Base-T) [Marvell] and it works with the skge driver too. There's something really strange with this driver....

[rooijan (rrossouw)]

I think Feisty 64-bit might have a problematic skge module. I can use the network through 32-bit Live CD's and installed versions from Ubuntu and Fedora with no problem. Vista 32-bit is also installed on this machine and working. With the 64-bit version though I can see the adapter and assign a static IP. I have link on the card but can not ping anything on the same network. ifconfig shows that there are no packets received.

Noteworthy that I have the same symptoms with a Fedora 64-bit live CD.

[rooijan (rrossouw)]

Sorry not sure what I was thinking. I am not using Feisty. I am using Gutsy(7.10) 64-bit version.

[rooijan (rrossouw)]

Update:
I have played around with a few things. Right now I have network traffic with reasonable speed. Not sure which of the kernel parms worked Maybe all of them but I suspect it was booting the kernel with mem=3G that worked. So must be a bug with skge or sk98lin and addressing over 3Gb memory.

1. Compiled sk98lin driver 10.21.1.3 and blacklisted skge and ipv6
root@rrosso-gutsy64:/media/disk/DriverInstall# lsmod | grep sk
sk98lin 209756 1

2. Use static ip and restart networking

3. boot kernel with the following
kernel /boot/vmlinuz-2.6.22-14-generic root=UUID=80479eaf-1974-4e71-ad71-a7982df90353 ro quiet splash noapic nolapic acpi=off mem=3G

If I have more time I will try skge again and see if just using mem= will work. That way it is narrowed down a little bit to a 3Gb+ problem.

[rooijan (rrossouw)]

OK one more update. skge driver works with mem=3G passed to the kernel.

root@rrosso-gutsy64:~# lsmod | grep sk
skge 47248 0

root@rrosso-gutsy64:~# lshw -C network
  *-network
       description: Ethernet interface
       product: 88E8001 Gigabit Ethernet Controller
       vendor: Marvell Technology Group Ltd.
       physical id: c
       bus info: pci@0000:06:0c.0
       logical name: eth0
       version: 13
       serial: 00:17:31:71:b0:a8
       size: 100MB/s
       capacity: 1GB/s
       width: 32 bits
       clock: 66MHz
       capabilities: pm vpd bus_master cap_list ethernet physical tp 10bt 10bt-fd 100bt 100bt-fd 1000bt 1000bt-fd autonegotiation
       configuration: autonegotiation=on broadcast=yes driver=skge driverversion=1.11 duplex=full firmware=N/A ip=192.168.1.3 latency=32 link=yes maxlatency=31 mingnt=23 module=skge multicast=yes port=twisted pair speed=100MB/s

kernel /boot/vmlinuz-2.6.22-14-generic root=UUID=80479eaf-1974-4e71-ad71-a7982df90353 ro quiet splash mem=3G

[rooijan (rrossouw)]

Update:
This problem still exists in Hardy 64-bit. Using skge module. Without mem=3G passed to the kernel this driver does not work.

Details of my working setup:
From /boot/grub/men.lst
kernel /boot/vmlinuz-2.6.24-2-generic root=UUID=80479eaf-1974-4e71-ad71
-a7982df90353 mem=3G ro quiet splash

lspci: 06:0c.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001 Gigabit Ethernet Controller (rev 13)
kernel: Linux rrosso-gutsy64 2.6.24-2-generic #1 SMP

root@rrosso-gutsy64:~# dmesg | grep skge
[ 34.280751] skge 1.13 addr 0xd3000000 irq 18 chip Yukon-Lite rev 9
[ 34.281006] skge eth0: addr 00:17:31:71:b0:a8
[ 44.213731] skge eth0: enabling interface
[ 45.897889] skge eth0: Link is up at 100 Mbps, full duplex, flow control both

[Leann Ogasawara (leannogasawara)]

Just a note that the kernel source package for the Hardy release is now just 'linux' not 'linux-source-2.6.24'. The naming convention changed as of this new release. Sorry for any confusion. I've gone ahead and reassigned to the appropriate package.

[Leann Ogasawara (leannogasawara)]

Per the kernel team's bug policy, can you please attach the following information for the newer Hardy kernel. Please be sure to attach each file as a separate attachment.

* uname -a > uname-a.log
* cat /proc/version_signature > version.log
* dmesg > dmesg.log
* sudo lspci -vvnn > lspci-vvnn.log

For more information regarding the kernel team bug policy, please refer to https://wiki.ubuntu.com/KernelTeamBugPolicies .

Also, just a note that Hardy Alpha3 should be released in the next day or two. You might want to wait to test this newer release and then grab the log files from this. http://cdimage.ubuntu.com/releases/hardy/

Thanks again and we appreciate your help and feedback.

[rooijan (rrossouw)]

Update on testing 2.6.24-3-generic. Without the mem=3G still not able to pass network traffic. See attached log files.

[Leann Ogasawara (leannogasawara)]

Thanks for the files. Just for future reference, if you could attach each file as a separate attachment it would be much appreciated. Could you actually try booting without the mem=3G parameter. Hopefully useful error messages may be logged which will help with debugging. Thanks.

[Connor Imes (ckimes)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. You reported this bug a while ago and there hasn't been any activity in it recently. We were wondering is this still an issue for you? Can you try with latest Ubuntu release? Thanks in advance.

[rooijan (rrossouw)]

Thanks for checking back. Iif I remember correct this was a 64-bit issue only and I don't have 64-bit installed anymore since I had this problem. I will set up a newer 64-bit version again and let you now.

-Riaan

----- Original Message ----
From: Connor Imes <email address hidden>
To: <email address hidden>
Sent: Sunday, August 24, 2008 6:33:10 AM
Subject: [Bug 131965] Re: Skge broken for Marvell network chip

Thank you for taking the time to report this bug and helping to make
Ubuntu better. You reported this bug a while ago and there hasn't been
any activity in it recently. We were wondering is this still an issue
for you? Can you try with latest Ubuntu release? Thanks in advance.

--
Skge broken for Marvell network chip
https://bugs.launchpad.net/bugs/131965
You received this bug notification because you are a direct subscriber
of the bug.

[Leann Ogasawara (leannogasawara)]

The Ubuntu Kernel Team is planning to move to the 2.6.27 kernel for the upcoming Intrepid Ibex 8.10 release. As a result, the kernel team would appreciate it if you could please test this newer 2.6.27 Ubuntu kernel. There are one of two ways you should be able to test:

1) If you are comfortable installing packages on your own, the linux-image-2.6.27-* package is currently available for you to install and test.

--or--

2) The upcoming Alpha5 for Intrepid Ibex 8.10 will contain this newer 2.6.27 Ubuntu kernel. Alpha5 is set to be released Thursday Sept 4. Please watch http://www.ubuntu.com/testing for Alpha5 to be announced. You should then be able to test via a LiveCD.

Please let us know immediately if this newer 2.6.27 kernel resolves the bug reported here or if the issue remains. More importantly, please open a new bug report for each new bug/regression introduced by the 2.6.27 kernel and tag the bug report with 'linux-2.6.27'. Also, please specifically note if the issue does or does not appear in the 2.6.26 kernel. Thanks again, we really appreicate your help and feedback.

[jszabo (jason-szabo)]

Hello,

I am finding the exact same issue with the latest 2.6.27-4 kernel (Intrepid Alpha5 -- 64-bit). Supplying the kernel with the mem=3G boot parameter allows the skge driver to pass traffic. Without it...all seems functional but no traffic is passed.

I have attached the requested outputs/logs for investigation. I am running the following hardware:
-ASUS M2N-SLI mobo
-AMD Athlon 64 x2 Dual Core 6400+
-Network Card with driver issues PCI D-Link DGE-530T
-Note: The PCI-Express Ethernet card (same model) works fine but uses the sky2 driver

Please let me know if you require any additional information.

Respectfully...J

[SupportGuy (3-admin-labs)]

Hello,

I am having the same problem with DGE-530T on 32bit x86 machine with Hardy [ 2.6.24-19-generic]. Also, as mentioned in this post I am able to reproduce this failure by using vlc - streaming. Since, there is no reported fix, I will compile a custom kernel with sk98lin (deprecated). I'm also willing to test any patches to skge.

Thanks
-JG

[Connor Imes (ckimes)]

Marked the development kernel as Triaged with Medium priority, based on the feedback and information provided in the last few posts. I will close the task against the old kernel.

Assigned to Ubuntu Kernel Team.

Thank you.

[Launchpad Janitor (janitor)]

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.

[Jim Lieb (lieb)]

Please try the Jaunty release. This update contains fixes for restarting the interface
under conditions similar to yours. Let us know the result so we can either pursue this
further or close the bug.

[StanS (stan-saraczewski-yahoo)]

I am not able to network after installing Jaunty. I've followed numerous suggestions on the forums to no avail.

DHCP appears to be working as an IP address is assigned.

What information shall I provide ?

[StanS (stan-saraczewski-yahoo)]

For the record - it was the Beta install of Jaunty. Not sure if that will make a difference ?

[Jim Lieb (lieb)]

StanS, Now that Jaunty is released, please upgrade and test again. As for what we want, see Leann's comment 13 above for pointers. This bug is for the skge driver for the Marvell NIC. Please confirm. If you managed to receive an address via DHCP, the
NIC had to be working. You can verify that by 'ifconfig -a' from the console or via the network manager settings in the desktop.

[StanS (stan-saraczewski-yahoo)]

I've managed to get around this problem by installing an old Linksys NIC in one of the PCI slots and telling bios to ignore the onboard NIC.

Apparently a different driver is installed...

I happened to grab a SUSE .iso 64 bit cd and went with it cause for once (out of numerous Ubuntu installs) the monitor was configured correctly.

Thanks for getting back to me.

[Jim Lieb (lieb)]

I have found the following but I am reluctant to shove this into the pipeline without more info. This applies to one board but there could be a common thread. I have emailed a query to the author of the patch which, along with the commentary, is here:

 http://patchwork.ozlabs.org/patch/22884/

[Julien Lavergne (gilir)]

I can confirmed the problem :
Jaunty amd64, kernel 2.6.28.

With 3 Gb of RAM, all is OK
With 4 Gb, the network is down. I can't access to the Internet, and can't even access to my router with the local adress (192.168.X.X).

I tried the fix which Jim posted a link, but recompiled the kernel with this patch doesn't seems to work.

[Jeremy Foshee (jeremyfoshee)]

This bug report was marked as Triaged a while ago but has not had any updated comments for quite some time. Please let us know if this issue remains in the current Ubuntu release, http://www.ubuntu.com/getubuntu/download . If the issue remains, click on the current status under the Status column and change the status back to "New". Thanks.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[Julien Lavergne (gilir)]

It's still a problem on Karmic amd64.

[Jeremy Foshee (jeremyfoshee)]

Julien,
    Are you able to test against Lucid? I'd like to see if it is the same in the current development version.

Thanks!

~JFo

[Leann Ogasawara (leannogasawara)]

Hi Julien,

As Jeremy mentioned, if you are able to test with the latest Lucid Beta release, that would be great. http://cdimage.ubuntu.com/releases/lucid/beta-1/ . If the issue does remain, can you reproduce the issue and then run the following command which will automatically gather and attach relevant debug information to this report.

apport-collect 131965

Thanks in advance.

[Julien Lavergne (gilir)]

AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.21.
Architecture: amd64
ArecordDevices:
 **** List of CAPTURE Hardware Devices ****
 card 0: NVidia [HDA NVidia], device 0: AD198x Analog [AD198x Analog]
   Subdevices: 1/1
   Subdevice #0: subdevice #0
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: seagle 1510 F.... pulseaudio
CRDA: Error: [Errno 2] Aucun fichier ou dossier de ce type
Card0.Amixer.info:
 Card hw:0 'NVidia'/'HDA NVidia at 0xd5000000 irq 23'
   Mixer name : 'Analog Devices AD1986A'
   Components : 'HDA:11d41986,1043818f,00100500'
   Controls : 23
   Simple ctrls : 14
DistroRelease: Ubuntu 10.04
HibernationDevice: RESUME=/dev/sda6
MachineType: System manufacturer System Product Name
Package: linux (not installed)
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-21-generic root=UUID=e1d247f0-86de-497c-b432-123da8969b38 ro quiet splash
ProcEnviron:
 LANG=fr_FR.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.32-21.31-generic 2.6.32.11+drm33.2
Regression: No
RelatedPackageVersions: linux-firmware 1.34
Reproducible: Yes
RfKill:
 0: hci0: Bluetooth
  Soft blocked: no
  Hard blocked: no
Tags: lucid kconfig needs-upstream-testing
Uname: Linux 2.6.32-21-generic x86_64
UserGroups: adm admin cdrom dialout fuse lpadmin plugdev sambashare
dmi.bios.date: 07/13/2006
dmi.bios.vendor: Phoenix Technologies, LTD
dmi.bios.version: ASUS P5NSLI ACPI BIOS Revision 0601
dmi.board.name: P5NSLI
dmi.board.vendor: ASUSTeK Computer INC.
dmi.board.version: 1.XX
dmi.chassis.asset.tag: 123456789000
dmi.chassis.type: 3
dmi.chassis.vendor: Chassis Manufacture
dmi.chassis.version: Chassis Version
dmi.modalias: dmi:bvnPhoenixTechnologies,LTD:bvrASUSP5NSLIACPIBIOSRevision0601:bd07/13/2006:svnSystemmanufacturer:pnSystemProductName:pvrSystemVersion:rvnASUSTeKComputerINC.:rnP5NSLI:rvr1.XX:cvnChassisManufacture:ct3:cvrChassisVersion:
dmi.product.name: System Product Name
dmi.product.version: System Version
dmi.sys.vendor: System manufacturer

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

apport information

[Julien Lavergne (gilir)]

I can confirm it's still an issue on lucid amd64, the i386 didn't have the problem.

[Julien Lavergne (gilir)]

Also try with a mainline 2.6.32, same problem.

[Alex Ivasyuv (industral)]

still exist in 10.04 server

[Tim Gardner (timg-tpi)]

Julien or Alex - please attach the output of 'lspci -vvnn' and 'dmesg' from the kernel version that _is_ failing to bring up your network.

[Alex Ivasyuv (industral)]

I get 100Mb/s instead of 1000Mb/s on Gigabit card.

[Tim Gardner (timg-tpi)]

Given that the original reporter hasn't commented on this bug in over 3 years I'm going to close this one and ask that you guys start your own bugs (since you all have different issues).

Julien - Use 'ubuntu-bug linux' to upload your system information whilst running the kernel causing the problem (if possible).

Alex - you have i915 issues (an oops in GEM) that may or may not affect your link speed. Search for 'Oops' in http://launchpadlibrarian.net/50402988/dmesg.log. You gotta fix the oops issue first 'cause all bets are off after that.

[Tim Gardner (timg-tpi)]

Expired, no response from the original reporter.

[Alex Ivasyuv (industral)]

Tim - OK, I've fixed Oops, but still exist issue with 100Mb/s instead of 1Gb/s.

[Julien Lavergne (gilir)]

bug reported : bug 595262

[Adam Herst (adamherst)]

I'm experiencing similar connectivity problems on Ubuntu 10.04 64-bit. Booting with mem=3G seems to fix the problem.