Mac Pro and VT-x
Ian Robinson
VT-x extensions are supported on the CPU's but disabled (locked by
Firmware).
Bugger. Hopefully this'll be enabled by a software update or in 10.5 so
that VMware & Parallels can use it.
Ian
--
Ian Robinson, Belfast, UK
<http://www.canicula.com/wp/>
SimonL
Or they've made a huge boob!
SimonL
Keith Pritchard
<ju...@canicula.invalid> wrote:
>When I start a VM under Parallels on a Mac Pro it tells me that Intel
>VT-x extensions are supported on the CPU's but disabled (locked by
>Firmware).
>
>Bugger. Hopefully this'll be enabled by a software update or in 10.5 so
>that VMware & Parallels can use it.
If I exit the VM, sleep the mac pro, wake it up again and then call up
the VM, I don't get the message about vt-x extensions being disabled.
I think it's just another symptom of parallels on the macpro not being
quite ready yet, I'm not too worried :) YET!
Keith
--
Zimbra V4 now available at www.simplymailsolutions.com
Including free mobile sync & macosx email/isync/addressbook/ical
compatibility
Chris Ridd
> Might be a sign that Apple is trying to force people to use Bootcamp?
> Or they've made a huge boob!
I read earlier that *some* Mac Pros had the VT extensions disabled by
firmware. It isn't clear why this is.
Since VMware doesn't use or need the VT extensions, this only currently
buggers up Parallels.
Cheers,
Chris
Graham J Lee
Does it bugger it up? On the (earlier) iMac, you're free to turn
Vanderpool off in Parallels.
--
Graham Lee
http://www.thaesofereode.info
Chris Ridd
> On Sat, 9 Sep 2006 12:20:20 +0100, Ian Robinson
> <ju...@canicula.invalid> wrote:
>
>> When I start a VM under Parallels on a Mac Pro it tells me that Intel
>> VT-x extensions are supported on the CPU's but disabled (locked by
>> Firmware).
>>
>> Bugger. Hopefully this'll be enabled by a software update or in 10.5 so
>> that VMware & Parallels can use it.
>
> If I exit the VM, sleep the mac pro, wake it up again and then call up
> the VM, I don't get the message about vt-x extensions being disabled.
> I think it's just another symptom of parallels on the macpro not being
> quite ready yet, I'm not too worried :) YET!
In another report I just read, it seems as though Apple only enables VT
on one of the processor cores - so running Parallels will produce this
error 50% of the time. I don't know if xnu migrates tasks between cores
or not, but if it does then you better forcibly disable VT in Parallels.
<http://arstechnica.com/journals/apple.ars/2006/9/8/5228>
Cheers,
Chris
Ian Robinson
(in article <4mft0bF...@individual.net>):
> this only currently
> buggers up Parallels.
Parallels still runs. It just tells you it's not using VT-x.
Graham J Lee
> On 2006-09-09 14:03:09 +0100, Keith Pritchard <kei...@gmail.com> said:
>
>>
>> If I exit the VM, sleep the mac pro, wake it up again and then call up
>> the VM, I don't get the message about vt-x extensions being disabled.
>> I think it's just another symptom of parallels on the macpro not being
>> quite ready yet, I'm not too worried :) YET!
>
> In another report I just read, it seems as though Apple only enables VT
> on one of the processor cores - so running Parallels will produce this
> error 50% of the time.
Or 25%, on a Mac Pro. Unless it's core 0, in which case probably a bit
less than 25% ;-). That sounds like an honest-to-goodness firmware bug,
as it takes the S out of SMP.
> I don't know if xnu migrates tasks between cores
> or not
>
Hmmm...you pose an interesting point there. I think, based on a quick
read, that it is possible but doesn't often happen (except when, say,
shutting off a processor). I want to look into that a bit deeper :-)
Chris Ridd
<info.thae...@leeg.invalid> said:
> On 9/9/06 16:43, Chris Ridd wrote:
>> On 2006-09-09 14:03:09 +0100, Keith Pritchard <kei...@gmail.com> said:
>>
>>>
>>> If I exit the VM, sleep the mac pro, wake it up again and then call up
>>> the VM, I don't get the message about vt-x extensions being disabled.
>>> I think it's just another symptom of parallels on the macpro not being
>>> quite ready yet, I'm not too worried :) YET!
>>
>> In another report I just read, it seems as though Apple only enables VT
>> on one of the processor cores - so running Parallels will produce this
>> error 50% of the time.
>
> Or 25%, on a Mac Pro.
Still 50%, I think - two CPUs each having one core with VT and one without?
I've no Mac Pro to test this out on. Donations gratefully received :-)
>> I don't know if xnu migrates tasks between cores or not
>>
>
> Hmmm...you pose an interesting point there. I think, based on a quick
> read, that it is possible but doesn't often happen (except when, say,
> shutting off a processor). I want to look into that a bit deeper :-)
Yay kernel source!
Cheers,
Chris
Graham J Lee
> On 2006-09-09 19:25:17 +0100, Graham J Lee
> <info.thae...@leeg.invalid> said:
>
>> On 9/9/06 16:43, Chris Ridd wrote:
>>> On 2006-09-09 14:03:09 +0100, Keith Pritchard <kei...@gmail.com> said:
>>>
>>>>
>>>> If I exit the VM, sleep the mac pro, wake it up again and then call up
>>>> the VM, I don't get the message about vt-x extensions being disabled.
>>>> I think it's just another symptom of parallels on the macpro not being
>>>> quite ready yet, I'm not too worried :) YET!
>>>
>>> In another report I just read, it seems as though Apple only enables
>>> VT on one of the processor cores - so running Parallels will produce
>>> this error 50% of the time.
>>
>> Or 25%, on a Mac Pro.
>
> Still 50%, I think - two CPUs each having one core with VT and one without?
>
I read it as one of the four cores. 'Rudolph lays the responsibility
for that at the feet of Apple, telling us that "VT support is disabled
because of an Apple bug." He told us that it is enabled on one core on
not on other three.' Definitely does look like a bug, there's no way
you'd intentionally make one of the CPUs run a different instruction set
from the others.
Especially when you've hobbled your OS to longer have a concept of
multiple processor sets.
> I've no Mac Pro to test this out on. Donations gratefully received :-)
>
I'll take donations for the 24" C2D iMac then, so we can test both new
chips.
>>> I don't know if xnu migrates tasks between cores or not
>>>
>>
>> Hmmm...you pose an interesting point there. I think, based on a quick
>> read, that it is possible but doesn't often happen (except when, say,
>> shutting off a processor). I want to look into that a bit deeper :-)
>
> Yay kernel source!
>
Indeed. Investigation shall occur mañana (though I shall have to break
one of my production boxen in testing this, because I only have a
single-core iBoko at home).