Comment 0 for bug 373034

Hi Guys,

Since we are no longer taking the 2.6.27.y stable updates wholesale for Intrepid, I'm reporting this on behalf of Luis Rodriguez:

"These are worthy for Intrepid. These patches are now part of 2.6.27.22. Without these on SMP machines some PCI 11n atheros devices will hang a box. This is more noticeable on MacBook Pros, but I was finally able to reproduce with a regular card. All PCI express 11n devices from us are OK as PCI express is already serialized. Oh and these are also on 2.6.28.10 and made it into the first release of 2.6.29."

commit e06377f6790bab20549524066b493f36f88e806e
Author: Luis R. Rodriguez <email address hidden>
Date: Mon Mar 23 19:03:30 2009 -0400

    ath9k: AR9280 PCI devices must serialize IO as well

    This is a port of:
    commit SHA1 5ec905a8df3fa877566ba98298433fbfb3d688cc
    for 2.6.27

    Signed-off-by: Luis R. Rodriguez <email address hidden>
    Signed-off-by: Greg Kroah-Hartman <email address hidden>

commit 7a1ed3d6dccfd8a5c485c8396397f307ca26bb8f
Author: Luis R. Rodriguez <email address hidden>
Date: Mon Mar 23 19:03:29 2009 -0400

    ath9k: implement IO serialization

    This is a port of:
    commit SHA1 6158425be398936af1fd04451f78ffad01529cb0
    for 2.6.27

    All 802.11n PCI devices (Cardbus, PCI, mini-PCI) require
    serialization of IO when on non-uniprocessor systems. PCI
    express devices not not require this.

    This should fix our only last standing open ath9k kernel.org
    bugzilla bug report:

    http://bugzilla.kernel.org/show_bug.cgi?id=12110

    Signed-off-by: Luis R. Rodriguez <email address hidden>
    Signed-off-by: Greg Kroah-Hartman <email address hidden>