Re: 7.0_BETA feedback sought

[David Brownlee <abs%absd.org@localhost>]

Just had a panic on an amd64 laptop running netbsd-7 as on 7th Jan.

System was running a wildfly-8 java server, firefox browser,
postgresql, (idle) mysql and a few xterms. Using the intel DRMKMS
driver with NoAccel.

panic below - I appear to have a netbsd.0.core.gz & netbsd.0.gz

panic: kernel diagnostic assertion "uvmexp.swpgonly + npages <= uvmexp.swpginuse
" failed: file "/opt/netbsd/7/sys/uvm/uvm_pager.c", line 472
cpu0: BAegRiNnI NtGr: aScPeLb aNcOkT .LO.W.E
RED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
vpanic() at WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
WARNING: SPL NOT LOWERED ON TRAP EXIT 6 0
netbsd:vpanic+0x13c
uvm_fault(0xfffffe81f2ba75d8, 0x0, 2) -> e
fatal page fault in supervisor mode
trap type 6 code 2 rip ffffffff80a78b01 cs 8 rflags 10202 cr2 8 ilevel 8 rsp fff
ffe8064358b08
curlwp 0xfffffe8064590860 pid 7919.5 lowest kstack 0xfffffe80643552c0
kern_assert() at netbsd:kern_assert+0x4f
uvm_aio_aiodone_pages() at netbsd:uvm_aio_aiodone_pages+0x51b
uvm_aio_aiodone() at netbsd:uvm_aio_aiodone+0x89
workqueue_worker() at netbsd:workqueue_worker+0x89
cpu0: End traceback...

On 8 January 2015 at 14:32, Havard Eidnes <he%netbsd.org@localhost> wrote:
>> Unsurprisingly, there appears to be support for it in the Linux
>> camp, via among others the "wl" driver, ref.
>>
>>   https://wiki.debian.org/wl
>>
>> and that driver supports some of the same chips our "bwi" driver
>> does.
>
> Unfortunately, that driver (the Linux "wl" driver) is closed-source.
> However, it appears that Broadcom has a BSD-licensed Linux driver for
> these newer chips, but that driver doesn't handle the old chips, so
> it's probably going to be difficult to see what the differences are,
> ref.
>
>   https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=154851
>
> FreeBSD has the bwn driver for cards using "v4" Broadcom firmware, but
> we don't have it (yet).  And it also doesn't support the 43224 chip,
> and this time I'll refrain from guessing how far it is from supporting
> that chip.
>
> So it looks like I'll have to eat my words:
>
>> ...it's my guess is that it should not be too difficult (famous last
>> words...) to get it working at least for the b/g protocols.
>
> It certainly doesn't look quite so trivial anymore...
>
> Regards,
>
> - Håvard