Bug 606243

Summary: rt2800pci driver attaches to RT2860-based devices by default
Product: [openSUSE] openSUSE 11.3 Reporter: Forgotten User cAXlJ_FoSf <forgotten_cAXlJ_FoSf>
Component: KernelAssignee: E-mail List <kernel-maintainers>
Status: VERIFIED FIXED QA Contact: E-mail List <qa-bugs>
Severity: Critical    
Priority: P2 - High CC: forgotten_1-yzHWP3HO, jeffm
Version: Factory   
Target Milestone: ---   
Hardware: i586   
OS: openSUSE 11.3   
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Bug Depends on:    
Bug Blocks: 691073    

Description Forgotten User cAXlJ_FoSf 2010-05-16 16:47:22 UTC
User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.9) Gecko/20100317 Firefox/3.5.9

The rt2800pci driver attaches to RT2860-based devices by default but it is known not to work. It should use the rt2860sta driver instead.

Reproducible: Always
Comment 1 Jeff Mahoney 2010-05-20 19:31:31 UTC
Thu May 20 21:31:10 CEST 2010 - jeffm@suse.com

- Disabled CONFIG_RT2800PCI (bnc#606243)
  - These devices are handled by the rt2860 staging driver.
Comment 2 Forgotten User cAXlJ_FoSf 2010-05-20 22:27:36 UTC
(In reply to comment #1)
> Thu May 20 21:31:10 CEST 2010 - jeffm@suse.com
> 
> - Disabled CONFIG_RT2800PCI (bnc#606243)
>   - These devices are handled by the rt2860 staging driver.

Are you sure the whole driver should be disabled? It doesn't work yet for RT2860 based hardware but it supports RT2760, RT2790, RT2880, RT2890, and RT3052 chipsets as well although I don't know what their status is. Maybe it would be better to remove just the PCI IDs of RT2860-based devices?
Comment 3 Jeff Mahoney 2010-05-20 23:53:59 UTC
The rt2860 driver handles more than the rt2860.

Also, the comment from the RT2800PCI Kconfig option:
          This adds support for rt2800/rt3000/rt3500 wireless chipset family.
          Supported chips: RT2760, RT2790, RT2860, RT2880, RT2890 & RT3052

          This driver is non-functional at the moment and is intended for
          developers.
Comment 4 Forgotten User 1-yzHWP3HO 2010-05-30 16:10:13 UTC
Jeff,

I don't easily follow your comment. M7 uses RT2800PCI and this is inteneded for developers?

Reason I ask is because M7 on an eeepc 901 only works for a few seconds.

The default gateway is pingable when it's associated (or any other address, like 194.109.6.66), but only for approx 3 to 4 ping packets. After this it dies.

This problem happens with my own network @home as well as a completely open joikuspot / unencrypted network.

Do I need to raise a new bug or keep this reopened?

Unless I'm dong something wrong, it's a show stopper I guess.
Comment 5 Forgotten User 1-yzHWP3HO 2010-06-11 19:56:05 UTC
helly, anyone alive here?

wireless works for say 3 pings and only a network reload gets it to work shortly.

we need to sort this out before it gets distributed....
Comment 6 Jeff Mahoney 2010-06-12 15:56:53 UTC
That driver is intended for debugging by developers. The rt2860sta staging driver should be used instead and has been set that way since May 20 in the kernel repo.

Grab the latest factory kernel and see if it fixes it for you.

That said, please don't adjust the priority of bugs. It's not what priority it is for you - it's what priority it is for the developers and maintainers. openSUSE bugs are *never* CritSit.
Comment 7 Forgotten User cAXlJ_FoSf 2010-06-12 18:08:43 UTC
(In reply to comment #5)
> helly, anyone alive here?
> 
> wireless works for say 3 pings and only a network reload gets it to work
> shortly.
> 
> we need to sort this out before it gets distributed....

The rt2800pci driver has been disabled in recent kernels. Your problem has nothing to do with this bug, please open a separate bug and provide some useful information.
Comment 8 Forgotten User 1-yzHWP3HO 2010-06-12 20:45:06 UTC
the setup is based on milestone 7. Pretty recent, right? The reason I changed the stuff because there is a thought behind it -- eventhough probably not correct:

I am 100% sure if the 11.3 comes out and wireless fails for a lot of people, it will do quite some damage to the product. I don't think it's something we want.

Just give me some information what to provide. I don't see a lot in the logs. The thing I see is that whe wireless is brought up, it is alive for a few seconds and then dies.
Comment 9 Jeff Mahoney 2010-06-12 20:56:35 UTC
No. The timestamp for milestone 7's kernel is May 17. You can check using rpm -qi --changelog <kernel package>
Comment 10 Forgotten User cAXlJ_FoSf 2010-06-12 21:58:11 UTC
(In reply to comment #8)
> the setup is based on milestone 7. Pretty recent, right? The reason I changed

No.
 
> I am 100% sure if the 11.3 comes out and wireless fails for a lot of people, it
> will do quite some damage to the product. I don't think it's something we want.
> 
> Just give me some information what to provide. I don't see a lot in the logs.
> The thing I see is that whe wireless is brought up, it is alive for a few
> seconds and then dies.

You have not even stated whether rt2800pci or rt2860sta attaches to your wireless card. Update your kernel to a recent version and if it still fails open a new bug. This one is about removing the rt2800pci driver and that has been fixed a while ago.
Comment 11 Forgotten User 1-yzHWP3HO 2010-06-13 06:21:09 UTC
the subject states "rt2800pci driver attaches to RT2860-based devices by default"
so yes, it indeed in M7 attaches rt2800pci to it.

(the 2860 used to work with 11.2)

I have added a new bug for M7.

Hopefully it will professionally be picked up. (e.g. not the way the last comments are)