kern/39864: iwn network interface driver broken in current & -5 (i386 - others??)

[kre%munnari.OZ.AU@localhost]

>Number:         39864
>Category:       kern
>Synopsis:       iwn network interface driver broken
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Nov 04 07:25:01 +0000 2008
>Originator:     Robert Elz
>Release:        NetBSD -current (4.99.73, 5.0_BETA 5.99.*) since 2008-10-23 or 
>earlier
>Organization:
        Prince of Songkla University
>Environment:
System: NetBSD jade.coe.psu.ac.th 4.0_STABLE NetBSD 4.0_STABLE 
(JADE-1.696-20080517) #9: Fri May 23 18:55:13 ICT 2008 
kre%jade.coe.psu.ac.th@localhost:/usr/obj/4/kernels/JADE i386
Architecture: i386
Machine: i386
>Description:
        An i386 kernel (-current) I compiled from -current sources on
        2008-10-16 supports the iwn (Intel 4965) driver just fine
        (that is, it works well enough whether or not it is perfect).

        A kernel (-current) I compiled from -current sources on
        2008-10-23 fails.   The same was true for a kernel I
        compiled 2008-10-29 (that was just a day or so before the
        netbsd-5 tag was applied).   And since there have been
        no if_iwn.c commits since then, I'm assuming that if I
        were to run the 5.0_BETA I compiled last night, it would
        fail the same way, or that if I build current (5.99.*) now
        it would also fail.

        I know that "fails" is not much of a useful problem
        report - someone else mentioned the problems on the
        mailing list earlier, I had assumed there would be an
        open PR about it, so when I couldn't find one, I
        decided I should make sure the issue doesn't get
        forgotten (this one should be considered a 5.0 blocker IMO).

>How-To-Repeat:
        Attempt to use any recent NetBSD system on a laptop with an
        Intel 4965 wireless card.   Expect kernel dagnostics and
        no useful communication (through the wireless at all).

>Fix:
        No idea yet.  I am assuming that one of these two commits ...

revision 1.22
date: 2008/10/20 09:33:48;  author: rtr;  state: Exp;  lines: +3 -4
- fix crazy condition that always evaluates to false without this
  iwn_fix_channel() never gets called.
- remove hardcode to ic->ic_channels[11]

fixes a problem in associating to access points
----------------------------
revision 1.21
date: 2008/10/18 15:55:27;  author: blymn;  state: Exp;  lines: +3 -3
Make scan abort command async as newstate is called from an interrupt
context so we cannot ltsleep.

        to if_iwn.c is the cause of the problem, the previous
        commit was 2008/10/13 and would have been included in
        my 2008-10-16 kernel that works.

        That is, for now, I'm assuming that this file is the
        one that broke things (it is kind of the logical choice).

        When I get time I'll try builds with each of those changes
        (and with both) reverted, and see which one is actually causing
        the problem, or perhaps that neither is and so I (or we) need
        to look further afield.  If I can I'll do that today.

        I'll also (eventually, assuming a fix doesnt happen first)
        append more info on what actually happens (just what the
        kernel complains about while it isn't working...)