* smp + acpi
@ 2006-07-26 12:47 Marco Berizzi
2006-07-26 15:23 ` Andi Kleen
0 siblings, 1 reply; 7+ messages in thread
From: Marco Berizzi @ 2006-07-26 12:47 UTC (permalink / raw)
To: linux-kernel
Since 2.6.15 smp doesn't work anymore without ACPI
May be possible to have a note in "Symmetric multi processing
support" help dialog? Or is it possible to enable ACPI when
SMP is selected?
TIA
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: smp + acpi
2006-07-26 12:47 Marco Berizzi
@ 2006-07-26 15:23 ` Andi Kleen
2006-07-26 15:37 ` Marco Berizzi
0 siblings, 1 reply; 7+ messages in thread
From: Andi Kleen @ 2006-07-26 15:23 UTC (permalink / raw)
To: Marco Berizzi; +Cc: linux-kernel
"Marco Berizzi" <pupilla@hotmail.com> writes:
> Since 2.6.15 smp doesn't work anymore without ACPI
> May be possible to have a note in "Symmetric multi processing
> support" help dialog? Or is it possible to enable ACPI when
> SMP is selected?
It's probably specific to your system, nothing general.
-Andi
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: smp + acpi
2006-07-26 15:23 ` Andi Kleen
@ 2006-07-26 15:37 ` Marco Berizzi
0 siblings, 0 replies; 7+ messages in thread
From: Marco Berizzi @ 2006-07-26 15:37 UTC (permalink / raw)
To: ak; +Cc: linux-kernel
Andi Kleen wrote:
>"Marco Berizzi" <pupilla@hotmail.com> writes:
>
> > Since 2.6.15 smp doesn't work anymore without ACPI
> > May be possible to have a note in "Symmetric multi processing
> > support" help dialog? Or is it possible to enable ACPI when
> > SMP is selected?
>
>It's probably specific to your system, nothing general.
Hi Andi,
Thanks for the reply. I'm compiling linux on a pentinum
4 HT 3GHz. 2.6.14 did detect both processor, but all
kernels > 2.6.15 did not. (at least till 2.6.17.7)
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: smp + acpi
[not found] ` <fa.qBnHwLitfu7Jux0cIbFxbbRJ5m8@ifi.uio.no>
@ 2006-07-26 16:57 ` Robert Hancock
0 siblings, 0 replies; 7+ messages in thread
From: Robert Hancock @ 2006-07-26 16:57 UTC (permalink / raw)
To: Marco Berizzi; +Cc: ak, linux-kernel
Marco Berizzi wrote:
> Andi Kleen wrote:
>
>> "Marco Berizzi" <pupilla@hotmail.com> writes:
>>
>> > Since 2.6.15 smp doesn't work anymore without ACPI
>> > May be possible to have a note in "Symmetric multi processing
>> > support" help dialog? Or is it possible to enable ACPI when
>> > SMP is selected?
>>
>> It's probably specific to your system, nothing general.
>
> Hi Andi,
>
> Thanks for the reply. I'm compiling linux on a pentinum
> 4 HT 3GHz. 2.6.14 did detect both processor, but all
> kernels > 2.6.15 did not. (at least till 2.6.17.7)
Are you sure you are really using the same configuration? HyperThreading
has never worked without some amount of ACPI enabled.
--
Robert Hancock Saskatoon, SK, Canada
To email, remove "nospam" from hancockr@nospamshaw.ca
Home Page: http://www.roberthancock.com/
^ permalink raw reply [flat|nested] 7+ messages in thread
* RE: smp + acpi
@ 2006-07-26 17:07 Brown, Len
2006-07-26 17:13 ` Andi Kleen
0 siblings, 1 reply; 7+ messages in thread
From: Brown, Len @ 2006-07-26 17:07 UTC (permalink / raw)
To: Marco Berizzi, ak; +Cc: linux-kernel
>> > Since 2.6.15 smp doesn't work anymore without ACPI
>> > May be possible to have a note in "Symmetric multi processing
>> > support" help dialog? Or is it possible to enable ACPI when
>> > SMP is selected?
>>
>>It's probably specific to your system, nothing general.
>
>Hi Andi,
>
>Thanks for the reply. I'm compiling linux on a pentinum
>4 HT 3GHz. 2.6.14 did detect both processor, but all
>kernels > 2.6.15 did not. (at least till 2.6.17.7)
CONFIG_ACPI=y is necessary to parse the ACPI tables
and discover HT siblings. Except for the rare BIOS
that gives the option to enumerate HT via MPS
(thus breaking some versions of Windows),
enabling ACPI is the only way to enable HT.
Yes, in the distant past, CONFIG_ACPI=n did not remove
all ACPI code from your kernel, and that was a bug.
thanks,
-Len
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: smp + acpi
2006-07-26 17:07 Brown, Len
@ 2006-07-26 17:13 ` Andi Kleen
0 siblings, 0 replies; 7+ messages in thread
From: Andi Kleen @ 2006-07-26 17:13 UTC (permalink / raw)
To: Brown, Len; +Cc: Marco Berizzi, linux-kernel
> CONFIG_ACPI=y is necessary to parse the ACPI tables
> and discover HT siblings. Except for the rare BIOS
> that gives the option to enumerate HT via MPS
> (thus breaking some versions of Windows),
> enabling ACPI is the only way to enable HT.
>
> Yes, in the distant past, CONFIG_ACPI=n did not remove
> all ACPI code from your kernel, and that was a bug.
Ok thanks for the confirmation.
However the proposed change would be still wrong because
SMP can be without HT.
-Andi
^ permalink raw reply [flat|nested] 7+ messages in thread
* RE: smp + acpi
@ 2006-07-26 17:22 Brown, Len
0 siblings, 0 replies; 7+ messages in thread
From: Brown, Len @ 2006-07-26 17:22 UTC (permalink / raw)
To: Andi Kleen; +Cc: Marco Berizzi, linux-kernel
>> CONFIG_ACPI=y is necessary to parse the ACPI tables
>> and discover HT siblings. Except for the rare BIOS
>> that gives the option to enumerate HT via MPS
>> (thus breaking some versions of Windows),
>> enabling ACPI is the only way to enable HT.
>>
>> Yes, in the distant past, CONFIG_ACPI=n did not remove
>> all ACPI code from your kernel, and that was a bug.
>
>Ok thanks for the confirmation.
>
>However the proposed change would be still wrong because
>SMP can be without HT.
What proposed change?
I expect that the problem at hand is that CONFIG_SMP=y
is fine, but with CONFIG_ACPI=n, that isn't going to
find the HT threads on an HT system.
-Len
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2006-07-26 17:23 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <fa.55NzOEnhAyNTd6rw92/YGAb9CaI@ifi.uio.no>
[not found] ` <fa.qBnHwLitfu7Jux0cIbFxbbRJ5m8@ifi.uio.no>
2006-07-26 16:57 ` smp + acpi Robert Hancock
2006-07-26 17:22 Brown, Len
-- strict thread matches above, loose matches on Subject: below --
2006-07-26 17:07 Brown, Len
2006-07-26 17:13 ` Andi Kleen
2006-07-26 12:47 Marco Berizzi
2006-07-26 15:23 ` Andi Kleen
2006-07-26 15:37 ` Marco Berizzi
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox