when power saving is on wireless RA2860 doesnt associate on Natty

Does this bug happen when you always stay connected to mains while booting and connecting?

Yes, I have done a few test boots now and it seems consistent.
When I have power plugged in during boot the wireless will connect, but on battery if fails to authenticate.

Ok. Sounds like the same bug(s) I have, stemming from power saving attempts. There are two choices to fix that:

a)
The more conservative choice is to keep your supported ubuntu kernel and just disable those power saving measures.
You can do so by creating an empty file in /etc/pm/power.d/ with the name of those power saving scripts and without the executable bit set (see man pm-powersave):

sudo touch /etc/pm/power.d/wireless
sudo touch /etc/pm/power.d/pcie_aspm

Just the wireless part may be sufficient for you. The pcie_aspm issue is _very_ dependent on the details of the hardware. For me the wireless will stop working after a suspend/resume cycle if I go to battery then.

b)
You could also run a 2.6.39 mainline kernel if you do not need anything that only the ubuntu kernel has. Currently the newest ones available pre-compiled are:
for 64 bit: http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.39-rc5-oneiric/
or for 32 bit:
http://kernel.ubuntu.com/~kernel-ppa/mainline/daily/2011-04-19-natty/
They are not final and not "supported". The kernels of choice for curious people like me :) . All data not backed up is disposable - but then it always is.

I reported a question with similar issues... however, Natty doesnt recognize the Broadcom hardware. I deleted the original proprietary software and reinstalled. Still didnt work. I used a bootable CD/DVD with 10.04 on it and it recognizes the hardware.. The broadcom hardware with the proprietary software was working on the most recent 64bit version of Ubuntu before the update to Natty Narwhal...

Had to use my Cell tethered last evening to get up.. i have wired connection at the office when installed Natty .. and didnt know that the wireless wasnt working till later at home... I am now back at the office trying to fix the issue.

Keith, did you try any of the solutions I wrote about? What were the results?

Sorry, not yet .. on the EEE as mentioned I am still using the liveCD so can't adopt a new kernel, or place new files.

Would the touch option work if I allow it to sleep instead of a reboot ?

Should work, yes. It has to be done before the system goes to battery for the first time.

Ok, this looks like it confirms it.

1st boot:
On cable, connect to wireless, all ok
Put lid down to go into standby
Lid up, wireless reconnects correctly

2nd Boot:
On cable, connect to wireless, all ok
Disconnect power, connection remains stable
Put lid down to go into standby
Lid up, wireless fails to re-connect (device not present message shown in the panel applet)

3rd Boot:
On cable, connect to wireless, all ok
Disconnect power, connection remains stable
Run the touch commands
Put lid down to go into standby
Lid up, wireless reconnects correctly

Oh, "On cable" means power cable

I can confirm this on an EeePC 901 XP with stock Ubuntu 11.04.

This is what I found in dmesg:

[ 615.852106] phy0 -> rt2800_wait_csr_ready: Error - Unstable hardware.
[ 615.862286] phy0 -> rt2x00pci_regbusy_read: Error - Indirect register access failed: offset=0x00007010, value=0xffffffff
[ 616.764092] phy0 -> rt2800_wait_csr_ready: Error - Unstable hardware.
[ 616.774227] phy0 -> rt2x00pci_regbusy_read: Error - Indirect register access failed: offset=0x00007010, value=0xffffffff

Addendum, newest firmware v26 from the RALink website solves the problem.

Hm, I would have thought you would have to run those touch commands before disconnecting power (in the 3rd boot). Anyway, if this fixes it it's fine.

I think the kernel currently in natty-proposed 2.6.38-9.43 might fix this issue. Could any of you please test this kernel? Enable proposed, install that kernel (upgrading should do that), shut down, boot, check what kernel you are running with uname -a and test whether the bug is gone (remove any workarounds like those from comment #3 a before that).

Didn't see that comment about the firmware before. Makes sense to me that a pcie powersaving (pcie_aspm) bug is fixed by a new firmware. So maybe pcie_aspm is fixed with the new firmware - or the device now tells the system that it can't do it and it is not attempted anymore.

Even newer firmware[1] has made it into ubuntu oneiric and is on its way into natty (SRU in bug 762987). The version number for rt2860.bin is 34.00. So this should fix the issue like version 26 from the Ralink website does.

From the changelog[2]:
linux-firmware: rt2800pci: rt2800usb: Firmware update

It would be very interesting to know whether this really fixes this issue for everyone. Could you find the time to test this? Simply install the new linux-firmware deb (dpkg -i) or manually grab rt2860.bin from the tar file[3] and put it in /lib/firmware. Then unload (sudo modprobe -rv rt2800pci) and reload (sudo modprobe -v rt2800pci) the driver so that it loads the new firmware and test. This can also be done on a live CD system.

[1] https://launchpad.net/ubuntu/+source/linux-firmware/1.54/+build/2552738/+files/linux-firmware_1.54_all.deb
[2] https://launchpad.net/ubuntu/+source/linux-firmware/+changelog
[3] https://launchpad.net/ubuntu/+archive/primary/+files/linux-firmware_1.54.tar.gz

I was lazy and tried the .deb in [1] and it works fine so far.

Thanks a lot for testing.

Since both the new firmware that is already in oneiric and in the process of being SRUed into natty and the kernel from natty-proposed (as well as kernels >2.6.38) fix this issue I'll close this bug.

Thanks for reporting. If this somehow does turn out not to be fixed give a comment here. For any other issues, as always, please open a new bug.