Hello,
I have two revisions of the 965WH motherboard, one that supports dual-core
only and another quad-core.
I have an E6300 in the old one and a Q6600 in the new one with 4GiB and
8GiB of RAM respectively.
I would like to report a success and also a bug with BIOS 1738.
The success is this case can be resolved, the Linux/memory/mapping is
correct now and I no longer have to use mem="8832M" in my LILO
configuration to avoid the bios setting memory that is not cachable.
However, on my new revision 965WH (I did not check with my old motherboard),
that supports the quad-core CPU, I went into the BIOS and went to the
[Hard Disks] option and when I selected it, the part of the screen that should
update showing me the hard disks in the system was blank, I could not escape
out of the bios or go to another menu, I had to reboot to get back into the
bios. This is not normal behavior, can someone at Intel attach 1-6 (I use 6)
WD Veliciraptor 300GiB drives to this motherboard to replicate the problem if
it is not obvious in the code/bios how to fix it? I cannot view the hard disks in that menu, I believe its the 2nd or 3rd menu over when you initially enter
the BIOS. I do not recall if this problem existed prior to 1738.
Also, people have e-mailed me in relation to my posts and I am curious as well
to follow up, when will the HECI/fan/voltage sensor support (that is a project
at Intel) be merged into the kernel tree so that lm sensors will eventually be
able to read these values?
This is from 2007:
http://softwareblogs.intel.com/2007/04/13/intel-me-interface-heci-now-available/
Their updated project page:
http://www.openamt.org/
Justin.
On Mon, 30 Jun 2008, Justin Piszcz wrote:
> Hello,
>
> I have two revisions of the 965WH motherboard, one that supports dual-core
> only and another quad-core.
>
> I have an E6300 in the old one and a Q6600 in the new one with 4GiB and 8GiB
> of RAM respectively.
>
> I would like to report a success and also a bug with BIOS 1738.
>
> The success is this case can be resolved, the Linux/memory/mapping is correct
> now and I no longer have to use mem="8832M" in my LILO configuration to avoid
> the bios setting memory that is not cachable.
>
> However, on my new revision 965WH (I did not check with my old motherboard),
> that supports the quad-core CPU, I went into the BIOS and went to the [Hard
> Disks] option and when I selected it, the part of the screen that should
> update showing me the hard disks in the system was blank, I could not escape
> out of the bios or go to another menu, I had to reboot to get back into the
> bios. This is not normal behavior, can someone at Intel attach 1-6 (I use 6)
> WD Veliciraptor 300GiB drives to this motherboard to replicate the problem if
> it is not obvious in the code/bios how to fix it? I cannot view the hard
> disks in that menu, I believe its the 2nd or 3rd menu over when you initially
> enter the BIOS. I do not recall if this problem existed prior to 1738.
>
> Also, people have e-mailed me in relation to my posts and I am curious as
> well
> to follow up, when will the HECI/fan/voltage sensor support (that is a
> project
> at Intel) be merged into the kernel tree so that lm sensors will eventually
> be
> able to read these values?
>
> This is from 2007:
> http://softwareblogs.intel.com/2007/04/13/intel-me-interface-heci-now-available/
>
> Their updated project page:
> http://www.openamt.org/
>
> Justin.
>
The issue regarding accessing the hard drive menu was fixed by simpling
power cycling/rebooting the host.
All issues are now resolved (regarding the slow speed/memory/etc)..
Only other question left is the HECI support in the kernel.
Justin.
On Tuesday, July 01, 2008 11:43 am Justin Piszcz wrote:
> > This is from 2007:
> > http://softwareblogs.intel.com/2007/04/13/intel-me-interface-heci-now-ava
> >ilable/
> >
> > Their updated project page:
> > http://www.openamt.org/
> >
> > Justin.
>
> The issue regarding accessing the hard drive menu was fixed by simpling
> power cycling/rebooting the host.
And the problem hasn't come back?
> All issues are now resolved (regarding the slow speed/memory/etc)..
Hopefully there's a more recent BIOS available that doesn't need to trim the
memory for you?
> Only other question left is the HECI support in the kernel.
I'll defer that question to Anas, he's been handling the HECI stuff.
Jesse
On Tue, 1 Jul 2008, Jesse Barnes wrote:
> On Tuesday, July 01, 2008 11:43 am Justin Piszcz wrote:
>>> This is from 2007:
>>> http://softwareblogs.intel.com/2007/04/13/intel-me-interface-heci-now-ava
>>> ilable/
>>>
>>> Their updated project page:
>>> http://www.openamt.org/
>>>
>>> Justin.
>>
>> The issue regarding accessing the hard drive menu was fixed by simpling
>> power cycling/rebooting the host.
>
> And the problem hasn't come back?
Correct, after several reboots, the problem has not come back, I can
access the HDD part of the BIOS w/out any issues.
>
>> All issues are now resolved (regarding the slow speed/memory/etc)..
>
> Hopefully there's a more recent BIOS available that doesn't need to trim the
> memory for you?
Erm, that is what I am trying to get across, version 1738 of the BIOS has
successfully solved the memory mapping/non-cachable memory issue.
>
>> Only other question left is the HECI support in the kernel.
>
> I'll defer that question to Anas, he's been handling the HECI stuff.
Thanks, a lot of people are probably curious where this is heading as
well, as it impacts purchasing decisions of people and whether one should
buy an Intel board if one cannot monitor the sensors on it (voltages/fan
speeds, etc)..
>
> Jesse
>
On Tuesday, July 01, 2008 12:08 pm Justin Piszcz wrote:
> >> All issues are now resolved (regarding the slow speed/memory/etc)..
> >
> > Hopefully there's a more recent BIOS available that doesn't need to trim
> > the memory for you?
>
> Erm, that is what I am trying to get across, version 1738 of the BIOS has
> successfully solved the memory mapping/non-cachable memory issue.
Oh, great. Good to hear we actually fix those bugs! :)
Thanks,
Jesse
On Tue, Jul 01, 2008 at 12:13:03PM -0700, Jesse Barnes wrote:
> On Tuesday, July 01, 2008 12:08 pm Justin Piszcz wrote:
> > Erm, that is what I am trying to get across, version 1738 of the BIOS has
> > successfully solved the memory mapping/non-cachable memory issue.
>
> Oh, great. Good to hear we actually fix those bugs! :)
Yeah, intel was starting to look hopeless as a board vendor. This only
took 10 months or so to fix while others like Gigabyte appear to have
done it in a few days as far as I can tell. :)
Oh well.
--
Len Sorensen