FreeBSD 9-CURRENT (and 8.0) on MacMini (rev. 3,1)
Phil Regnauld
Wasn't sure where to post this, so I'll try here as it involves -CURRENT
as well, and effort is probably best spent there.
Have acquired a MacMini "Server" model (4 GB RAM, 2 x 500 GB disk, and
no optical drive) to build a workshop training server. I am trying
to get FreeBSD to work on this beast, and so far I've had mitigated
success.
Below is the combination of versions that work/don't work. If it hangs/panics,
it's at boot time.
Boot mode ACPI NO ACPI/SAFE
F7.2-R amd64 hang panic (in swapper)
F7.2-R i386 WORKS, but 2.7 GB visible. PAE kernel panics on starting CPU2
no-ACPI not tested (not necessary)
F8.0-R amd64 hang hang
F8.0-R i386 hang hang ohci
F8.0-S amd64 hang hang at ohci early: SMM active, request owner change
F9.0-C amd64 panic acpica hang on md0: Preloaded image (Feb 2010 SNAP)
F9.0-C i386 panic acpica panic acpica
Stopped at kbd_enter+0x3a: movl $0,kbd_why
(USB kbd dead here, cannot backtrace)
Currently I'm running 7.3-STABLE/i386, without PAE, limited to 2.6 GB
RAM. ZFS works like a charm, and so does the built-in ethernet. ACPI
works too, but asmc(4) is not available in 7. cputemp(4) works fine as
well, but I don't think the fans ever change speed (though I succeeded
in building world and kernel multiple times, and the CPU core temp never
exceeded 80 C).
On those that don't work, apart from disabling ACPI, I've attempted
disabling various bits of the HW (sio, atkbdc, fdc, ...) but that hasn't helped
so far. Also, so advice out there recommends disabling HW in the BIOS, which
a Mac doesn't have.
I checked the archives for anything relevant, and do see that older versions
of the MacMini hardware tend to work (with some quirks), but it amd64's
has always been an issue, it seems.
I've also checked out the following links:
http://wiki.freebsd.org/AppleMacbook
Including:
"If your system stops early at boot, try reverting r189055: http://svn.freebsd.org/viewvc/base?view=revision&revision=189055"
(haven't tried that yet, but since 9.0 doesn't work, I didn't see the point).
So, anyway, I'd like to get -CURRENT or even 8-STABLE to work on this.
I'm ready to spend time helping whoever can guide me in debugging this critter.
DDB is a bit tricky since the USB tends to hang as well, and I can't break
into the debugger, or the keyboard doesn't work in it.
I can include verbose boot logs, test any kernel (setting up PXE boot env).
I could file PRs (but would like to refrain from doing this until I'm sure
I've tried the right steps and avoid polluting the PR DB with redundant tickets).
Any help appreciated!
Cheers,
Phil
_______________________________________________
freebsd...@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-curre...@freebsd.org"
Attilio Rao
> Hi everyone,
>
> Wasn't sure where to post this, so I'll try here as it involves -CURRENT
> as well, and effort is probably best spent there.
>
> Have acquired a MacMini "Server" model (4 GB RAM, 2 x 500 GB disk, and
> no optical drive) to build a workshop training server. I am trying
> to get FreeBSD to work on this beast, and so far I've had mitigated
> success.
>
> Below is the combination of versions that work/don't work. If it hangs/panics,
> it's at boot time.
>
> Boot mode ACPI NO ACPI/SAFE
>
> F7.2-R amd64 hang panic (in swapper)
> F7.2-R i386 WORKS, but 2.7 GB visible. PAE kernel panics on starting CPU2
> no-ACPI not tested (not necessary)
>
> F8.0-R amd64 hang hang
> F8.0-R i386 hang hang ohci
>
> F8.0-S amd64 hang hang at ohci early: SMM active, request owner change
>
> F9.0-C amd64 panic acpica hang on md0: Preloaded image (Feb 2010 SNAP)
> F9.0-C i386 panic acpica panic acpica
> Stopped at kbd_enter+0x3a: movl $0,kbd_why
> (USB kbd dead here, cannot backtrace)
I would start by compiling a debugging kernel and using serial port
for capturing, starting reporting the ACPI bug in the latest case,
then we can get useful informations.
Attilio
--
Peace can only be achieved by understanding - A. Einstein
Phil Regnauld
>
> I would start by compiling a debugging kernel and using serial port
> for capturing, starting reporting the ACPI bug in the latest case,
> then we can get useful informations.
Hi Attilio,
Any pointers on how to achieve that on a machine with no serial ports ?
I've checked out:
http://www.freebsd.org/doc/en/books/developers-handbook/kerneldebug-online-gdb.html
and
http://wiki.freebsd.org/DebugWithDcons (there is a recognized firewire port)
I don't otherwise see how to get a core to disk halfway through the boot
process.
Cheers,
Phil
Andriy Gapon
> Attilio Rao (attilio) writes:
>> I would start by compiling a debugging kernel and using serial port
>> for capturing, starting reporting the ACPI bug in the latest case,
>> then we can get useful informations.
>
> Hi Attilio,
>
> Any pointers on how to achieve that on a machine with no serial ports ?
> I've checked out:
> http://www.freebsd.org/doc/en/books/developers-handbook/kerneldebug-online-gdb.html
> and
> http://wiki.freebsd.org/DebugWithDcons (there is a recognized firewire port)
>
> I don't otherwise see how to get a core to disk halfway through the boot
> process.
You could try to use firewire console.
See dcons(4).
Also, a good and quicker start is to report actual panics that you get, as
Attilio has suggested.
When everything else fails, a digital camera still can be used to get screen
captures:-)
--
Andriy Gapon