Paul Ripke wrote:

>   1597      1 amrctl   __sysctl(0xbfbfea3c, 0x2, 0xbbbd9a00, 0xbfbfea44, 0, 0) = 0
>   1597      1 amrctl   open("/dev/amr0", 0, 0x8049aa2) = 3
>   1597      1 amrctl   ioctl(0x3, AMR_IO_VERSION, 0xbfbfea74) = 0
>        "\M^Y\0\0\0"
>   1597      1 amrctl   ioctl(0x3, AMR_IO_COMMAND, 0xbfbfe0ec) Err#5 EIO
>        "\M-!\^N\0\0\0M\M-1\M-;\0\M-D\M->o<\M-k\M-?\M-;`\M-1\M->\M-;\0\0\0\0\^A\0\0\0\M^]q\M-?\M-;\M-@\M-"\^D\b\0\^D\0"

kauth(9) can only return either 0 or 1 (EPERM), so it's quite easy
to tell when the error comes from there. is this something that was
introduced in a recent kernel?

in the amr driver (sys/dev/pci/amr.c) there is only one kauth(9)
call, and it bails right away if failed. you could try commenting
it out, rebuilding and trying again, but I would say the problem
is elsewhere.

-e.

-- 
Elad Efrat