port-amd64/39758: if_iwn.c 1.20 makes iwn(4) inoperable

[tgen%deepbone.net@localhost]

>Number:         39758
>Category:       port-amd64
>Synopsis:       if_iwn.c 1.20 makes iwn(4) inoperable
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    port-amd64-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Oct 17 18:50:00 +0000 2008
>Originator:     Thomas E. Spanjaard
>Release:        NetBSD 4.99.73
>Organization:
        
>Environment:
        
        
System: NetBSD hydra 4.99.73 NetBSD 4.99.73 (HYDRA) #3: Fri Oct 17 18:22:29 UTC 
2008 tgen@hydra:/home/tgen/work/netbsd/sys/arch/amd64/compile/obj/HYDRA amd64
Architecture: x86_64
Machine: amd64
>Description:
iwn(4) stops being able to find any network at all with r1.20 of if_iwn.c (and
if_iwnreg.h r1.4). Reverting to r1.19 and r1.3 fixes the issue.

iwn0 at pci2 dev 0 function 0: vendor 0x8086 product 0x4230 (rev. 0x61)
iwn0: interrupting at ioapic0 pin 17
iwn0: MoW2, address 00:21:5c:08:32:b3
iwn0: 11a rates: 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps
iwn0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
iwn0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 
36Mbps 48Mbps 54Mbps

I have two dmesg excerpts; one for a GENERIC kernel from monday with r1.19
and one for an almost-GENERIC (only options EST_DEBUG added) from today with
r1.20. There are no significant differences, except for the number of errors
reported for iwn0, but those differ between boots of the same kernel as well
as resuming from ACPI S3 suspend. But I guess that's due to the driver being
in alpha stages, or something. If the excerpts are deemed useful, just ask.
>How-To-Repeat:
Try to connect to 802.11 networks with an iwn(4) adapter on amd64 with r1.20
of if_iwn.c.
>Fix:
Revert if_iwn.c to r1.19 and if_iwnreg.h to r1.3.

>Unformatted: