Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 22

Thread: Weird wireless failure...

  1. #11
    Join Date
    Sep 2009
    Beans
    32

    Re: Weird wireless failure...

    Quote Originally Posted by TBABill View Post
    I think you can just make sure rt2860sta is the ONLY driver in /etc/modules, then enter the following into /etc/modprobe.d/blacklist.conf (all this as root):
    While the idea that my problem is kernel drivers seems promising, stabbing in the dark blacklisting modules didn't work for me when I tried it. It is entirely possible that I didn't blacklist the right modules... but I'm not sure that rt2860 is the right driver for my hardware, in any case. I'm also bemused as to why this would suddenly become a problem for me even though wireless had worked fine for over a year - with various netbook-remix distributions.

    The evidence making me sceptical that rt2860.sta is the correct module is:

    $ lspci | grep RaLink
    01:00.0 Network controller: RaLink RT2800 802.11n PCI
    $

    How can I verify that I should be using the RT2860 driver for a RT2800 device?

    Why did this suddenly start going wrong?

  2. #12
    Join Date
    Aug 2010
    Location
    Wales
    Beans
    Hidden!
    Distro
    Ubuntu 10.04 Lucid Lynx

    Re: Weird wireless failure...

    Quote Originally Posted by aSteve View Post
    $ lspci | grep RaLink
    01:00.0 Network controller: RaLink RT2800 802.11n PCI
    $

    How can I verify that I should be using the RT2860 driver for a RT2800 device?

    Why did this suddenly start going wrong?
    The Rt2860 driver should work for your device
    Perseverance will succeed (usually)
    ubuntu user -32597 - linux user - 526569
    Wireless Info Script
    Blog

  3. #13
    Join Date
    Sep 2009
    Beans
    32

    Re: Weird wireless failure...

    Quote Originally Posted by Hippytaff View Post
    The Rt2860 driver should work for your device
    On that advice, I tried again...

    I put this in /etc/modules.d/blacklist.conf

    Code:
    blacklist rt2x00usb
    blacklist rt2x00pci
    blacklist rt28O0pci
    Reboot and the event log says:
    Code:
    Mar 28 13:31:59 eeepc kernel: [   11.165765] rt2800pci 0000:01:00.0: PCI INT A -> GSI 19 (level, low) -> IRQ 19
    Mar 28 13:31:59 eeepc kernel: [   11.165780] rt2800pci 0000:01:00.0: setting latency timer to 64
    Mar 28 13:31:59 eeepc kernel: [   11.178770] phy0 -> rt2800_init_eeprom: Error - Invalid RF chipset detected.
    Mar 28 13:31:59 eeepc kernel: [   11.178778] phy0 -> rt2x00lib_probe_dev: Error - Failed to allocate device.
    Mar 28 13:31:59 eeepc kernel: [   11.196982] rt2800pci 0000:01:00.0: PCI INT A disabled
    Mar 28 13:31:59 eeepc kernel: [   11.198876] rt2860sta: module is from the staging directory, the quality is unknown, you have been warned.
    Mar 28 13:31:59 eeepc kernel: [   11.266165] rt2860 0000:01:00.0: PCI INT A -> GSI 19 (level, low) -> IRQ 19
    Mar 28 13:31:59 eeepc kernel: [   11.266240] rt2860 0000:01:00.0: setting latency timer to 64
    Mar 28 13:31:59 eeepc kernel: [   11.267102] === pAd = f8c3a000, size = 484312 ===
    Mar 28 13:31:59 eeepc kernel: [   11.267107] <-- RTMPAllocAdapterBlock, Status=0
    Mar 28 13:31:59 eeepc kernel: [   11.267113] pAd->CSRBaseAddress =0xf8360000, csr_addr=0xf8360000!
    Mar 28 13:31:59 eeepc kernel: [   11.370359] RX DESC f2b50000  size = 2048
    Mar 28 13:31:59 eeepc kernel: [   11.371547] <-- RTMPAllocTxRxRingMemory, Status=0
    Mar 28 13:31:59 eeepc kernel: [   11.374304] 1. Phy Mode = 0
    Mar 28 13:31:59 eeepc kernel: [   11.374312] 2. Phy Mode = 0
    Mar 28 13:31:59 eeepc kernel: [   11.391944] Console: switching to colour frame buffer device 128x37
    Mar 28 13:31:59 eeepc kernel: [   11.400066] fb0: inteldrmfb frame buffer device
    Mar 28 13:31:59 eeepc kernel: [   11.400071] drm: registered panic notifier
    Mar 28 13:31:59 eeepc kernel: [   11.401991] 3. Phy Mode = 0
    Mar 28 13:31:59 eeepc kernel: [   11.401999] ERROR! ****** BBP_Write_Latch Buffer exceeds max boundry ******
    Followed by a lot of errors vaguely similar to before. If I blacklist rt2860.sta, I don't get masses of errors, but I don't get a working wireless either.

    Based upon the error "phy0 -> rt2800_init_eeprom: Error - Invalid RF chipset detected." I'd argue that it looks as if the rt2860.sta is not the right driver for my hardware...

    I'm not sure if the warning about using the 'staging area' is relevant, or not.
    Last edited by aSteve; March 28th, 2011 at 01:54 PM.

  4. #14
    Join Date
    Aug 2010
    Location
    Wales
    Beans
    Hidden!
    Distro
    Ubuntu 10.04 Lucid Lynx

    Re: Weird wireless failure...

    Ralink can be a pain. you might want to consider using the windows xp driver with ndisgtk.
    Code:
    sudo apt-get install ndisgtk
    to install it.

    It has to be the XP driver though
    Perseverance will succeed (usually)
    ubuntu user -32597 - linux user - 526569
    Wireless Info Script
    Blog

  5. #15
    Join Date
    Sep 2009
    Beans
    32

    Re: Weird wireless failure...

    It worked fine for me for many months, and - recently - failed...

    Assuming this isn't a hardware fault, or an inadvertent configuration error, I'd prefer to avoid introducing the added complexity of finding and trying to install a windows driver.

    As I understood it, the eeepc was one of the key devices that the netbook remix distribution (with roots in EasyPeasy - itself based upon Ubuntu EEE) intends to target. I can understand it when untested hardware is a problem - but this is standard hardware that has worked for ages without a glitch with the default configuration.

    Curiously, when I do 'lsmod | grep rt2' - even with the blacklisting, I still get rt2800pci and rt2x00usb and rt2x00lib listed... suggesting that these are actually dependencies of rt2860sta? If so, perhaps the blacklisting was a red-herring?

  6. #16
    Join Date
    Aug 2010
    Location
    Wales
    Beans
    Hidden!
    Distro
    Ubuntu 10.04 Lucid Lynx

    Re: Weird wireless failure...

    Have you rebooted since blacklisting them. Blacklisting just revents them from being loaded during boot, so they will still be loaded unless you reboot. Alternatively you could remove them
    Code:
    modprobe -r
    to remove the named module. And
    Code:
     modprobe rt2800
    to reinstall them.

    Assuming it isn't a hardware fault. It is strange that it would just stop working.
    Perseverance will succeed (usually)
    ubuntu user -32597 - linux user - 526569
    Wireless Info Script
    Blog

  7. #17
    Join Date
    Dec 2009
    Location
    Southern Maryland
    Beans
    1,575
    Distro
    Ubuntu

    Re: Weird wireless failure...

    I don't think it's a red-herring issue, but maybe more of a needle/haystack issue since it worked and now does not. I researched a lot and have helped quite a few get 28xx series chips going, but http://wiki.debian.org/rt2860sta suggest that perhaps trying wicd as the network manager instead could render a result?

    From the Debian wiki
    rt2860sta is a module for the Ralink RT2700P/RT2700E/RT2800P/RT2800E and RT3000E PCI 802.11 draft-n wireless LAN chipsets.
    So while the rt2800 exists, there is documentation online that suggests it is not perfected (or workable). I have not tried to use it or attempted to make it workable so i cannot confirm. However, the rt2860sta handles lots of chipsets and I think even more than listed on the Debian wiki (rt3090 comes to mind).

    It could be as simple as something with the network manager itself and trying another may offer a solution. I had to install wicd on ONE machine that is different than my others because it has an Atheros wireless that has always just worked out of the box. On my Debian Squeeze install network manager connected but showed "not managed" and I had no way of knowing I was even online till I opened a browser or something else online. So I installed wicd and problem "worked around" because I can now connect and disconnect at will instead of just always being online but not knowing it.

    Good luck. I hope it's not a compound problem with drivers AND network manager or that will be a hair puller trying to figure out the combination that works.

  8. #18
    Join Date
    Sep 2009
    Beans
    32

    Re: Weird wireless failure...

    Quote Originally Posted by Hippytaff View Post
    Have you rebooted since blacklisting them.
    Yes. The error messages I posted were the first after the reboot.

  9. #19
    Join Date
    Aug 2010
    Location
    Wales
    Beans
    Hidden!
    Distro
    Ubuntu 10.04 Lucid Lynx

    Re: Weird wireless failure...

    Quote Originally Posted by TBABill View Post
    I had to install wicd on ONE machine
    I have seen this work too - when all else has failed WICD has saved the day. worth a shot
    Perseverance will succeed (usually)
    ubuntu user -32597 - linux user - 526569
    Wireless Info Script
    Blog

  10. #20
    Join Date
    Sep 2009
    Beans
    32

    Re: Weird wireless failure...

    Quote Originally Posted by TBABill View Post
    I don't think it's a red-herring issue, but maybe more of a needle/haystack issue since it worked and now does not. I researched a lot and have helped quite a few get 28xx series chips going, but http://wiki.debian.org/rt2860sta suggest that perhaps trying wicd as the network manager instead could render a result?
    That link is helpful - it is the first 'official' source suggesting that rt2860sta is the right module for a RT2800 PCI device. My comment about a 'red herring' is basically that altering the blacklisted modules doesn't appear to affect what lsmod responds... suggesting that the automatic selection of kernel modules always did work - and the additional modules should have been in the output of lsmod in the first place.

    My hunch is that the wi-fi chipset has some sort of flashed firmware which is currently wrong/corrupted. I note that, while the rt2860sta.ko files are different, they are all the same size (from kernels which 'just worked' to recent kernels where problems were found.) This suggests (non-conclusively) that it is not this module itself that has changed... causing me grief.

    I'm slowly leaning towards concluding that the cause may be hardware, not software... it's not conclusive - but I've not yet found any software oriented justification for the failure that makes sense. The right kernel module was selected automatically - and reverting to old kernels (with old modules) didn't make the card work. It is a pity that the card is integrated and there is no easy way to diagnose by swapping the component out.

Page 2 of 3 FirstFirst 123 LastLast

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •