Kernel module rt61 freeze the computer

[Gianni Moschini]

Binary package hint: linux-amd64-generic

In feisty beta 20070322 and earlier versions (ubuntu and kubuntu), with a PCI wireless card based on a ralink rt61 chipset the system freeze randomly.

If I boot the live session and do "sudo rmmod rt61" and make sure the module is unloaded, then I can use the live session during hours without freezing. So I'm sure that it's coming from this particular module but can't find any related package here (except linux-amd64-generic because I tried with both ubuntu and kubuntu amd64 versions)

[Gianni Moschini (gianni.net)]

Here I'm using the rt61-1.1.0-b1 kernel module compiled by hand on ubuntu dapper and it's working flawlessly.

I have no file to attach, but if you have any idea how I can get more informations I will be glad to give you any details.

Output of "lspci -vv"
_________________

0000:00:09.0 Network controller: RaLink: Unknown device 0301
        Subsystem: Hercules: Unknown device 0050
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=slow >TAbort- <TAbort- <MAbort- >SERR- <PERR-
        Latency: 32, Cache Line Size: 0x08 (32 bytes)
        Interrupt: pin A routed to IRQ 193
        Region 0: Memory at fb000000 (32-bit, non-prefetchable) [size=32K]
        Capabilities: [40] Power Management version 2
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-

[Gianni Moschini (gianni.net)]

I see the exact same problem reported on the rt2x00 project forum:
http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?t=3509

user valcom seems to suffer from the freeze with "soft loop" error output than this bug:
https://bugs.launchpad.net/ubuntu/+bug/34902

I will try to boot the feisty live session, exiting the Xorg, unloading the kernel module and reloading it with debug=1 to get more information

[Gianni Moschini (gianni.net)]

found another post with system freeze on rt61. May be related to SMP/PREEMPT in the kernel.

http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?t=1868

Does feisty for amd64 use SMP ? I think so but needs to verify.

[Gianni Moschini (gianni.net)]

Ok so this is caused by the combination of an SMP enabled kernel and the rt61 driver which seems to be both incompatible (see the link in the previous post).

The rt61 driver in feisty (20070322 beta) is :
http://boby.joe.free.fr/files/DSC01349.png

the kernel on feisty amd64 (kubuntu and ubuntu) is :
2.6.20-12-generic #2 SMP

PS: The PREEMPT does not seems to be a problem as I use 2.6.15-23-386 PREEMPT on Dapper with the rt61-1.1.0-b1 and I do not experience the same problem.

Also I got the same output as the bug related to the rt2500 (this is not the same kernel module but it probably shares a part of the code) which is:

http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?t=3509

A call trace can be found here:
http://boby.joe.free.fr/files/DSC01349.png

[Gianni Moschini (gianni.net)]

The first picture link should have been this one (rt61 modinfo)
http://boby.joe.free.fr/files/DSC01350.png

[Gianni Moschini (gianni.net)]

As it has been already discussed in the forum of the rt2x00 project, I'm reporting the bug in the project bugzilla.

http://sourceforge.net/tracker/index.php?func=detail&aid=1687350&group_id=107832&atid=648844

[Gianni Moschini (gianni.net)]

switched from linux-meta to linux-source-2.6.20

[Lennart Hansen (lahansen)]

I can confirm this also happens on my pc with a Sitecom - Model: WL-112 - Card

Here is a pic of the crash. - Sorry the bad quality

Will upload more info about the card + env later when I have it ready.

[Lennart Hansen (lahansen)]

Here is lspci

[Lennart Hansen (lahansen)]

lspcmcia

[Lennart Hansen (lahansen)]

modinfo

[Lennart Hansen (lahansen)]

If any additional info is needed I'll gladly provide it

Thanks ;-)

[Lennart Hansen (lahansen)]

How come no developer have been assigned this bug?

[ttoine (ttoine)]

I can confirm that the RT61 module in standard with the feisty kernell (generic, lowlatency, realtime, for i386 or amd64) is generating problems. The card is detected and seems to work with the standards kernell, will appear in the network manager, and it will be possible to see the wifi network around the computer. But when you try to get some network connection (wep ascii or hexa), it will not be possible to obtain an ip address. The most important bug for me is that if i try to use the Gnome Network Manager to set up a network, first, it does not work, and second, the system will hang on at next start-up/reboot.

I had to pick off the card, and reconfigure the kernell using synaptic.....

Some people have find different ways on the official ubuntuforums. For what I looked, it seems that the current driver with the kernell is bugued for Ralink rt2561 chipset, but works well with rt2500... as new wifi cards of a lot of brand like D-Link, Netgear or Linksys are now distributed with the new chipset. And of course, as the revision and chipset used do not appear on the packaging box, it is impossible for user to now in advance wich one they will have. And it is very strange to see the card in the network manager and not be able to use it at all.

So, I saw that there are other bug reports for Dapper, and Edgy with the same problem, where Ben Collins took time to see... It means that there might be some work around this problem. But I would like to know if some one can do something quick before the release of Ubuntu Feisty as it is a quite important hardware support bug.

Toine

[Gianni Moschini (gianni.net)]

They are trying to fix that upstream, but I have no time to try.

[QUOTE]
Date: 2007-04-11 17:09
Sender: markwallisProject Admin
Logged In: YES
user_id=342445
Originator: NO

Another patch has made it into CVS - can you please retest and report
back.
[/QUOTE]

You can get more details there:
http://sourceforge.net/tracker/index.php?func=detail&aid=1687350&group_id=107832&atid=648844

[Gianni Moschini (gianni.net)]

Could someone please confirm this, as it has been reported by many different users:
https://bugs.launchpad.net/bugs/+bugs?field.searchtext=rt61

[Gianni Moschini (gianni.net)]

OOps, done :-)

[Gianni Moschini (gianni.net)]

I reject my own bug report because I can find an older one, related to this issue with the same version of the kernel but with older patch version (2.6.20-9)