More RAM than MemTotal

[Nadav Har'El]

On Wed, Mar 28, 2012, Orna Agmon Ben-Yehuda wrote about "Re: More RAM than MemTotal":
> Baruch, thank you, but on my machine I do not have this line in dmesg.
> Indeed I use a x86_64 machine (also in the virtualized machine), and the
> PCI hole to which Nadav points says the problem exists only in 32 bit
> architectures.

I'm not sure this is actually true (though I'm definitely not an expert
in that area). It is possible that some devices and/or BIOSes, which are
designed to work with 32 bit operating systems, still have this problem.

My desktop machine is x86_64 running 64-bit Linux, has only 2 GB of RAM
(so you'd think there's plenty of room to map devices above this 2 GB
line), but still I see in the boot logs "BIOS-provided physical RAM map"
with various reserved ranges:

 BIOS-e820: 000000000009e000 - 00000000000a0000 (reserved)
 BIOS-e820: 00000000000ce000 - 00000000000d0000 (reserved)
 BIOS-e820: 00000000000e0000 - 0000000000100000 (reserved)
 BIOS-e820: 000000007d5b0000 - 000000007d5c9000 (ACPI data)
 BIOS-e820: 000000007d5c9000 - 000000007d5cc000 (ACPI NVS)
 BIOS-e820: 000000007d5cc000 - 0000000080000000 (reserved)
 BIOS-e820: 00000000e0000000 - 00000000f0000000 (reserved)
 BIOS-e820: 00000000fec00000 - 00000000fec10000 (reserved)
 BIOS-e820: 00000000fee00000 - 00000000fee01000 (reserved)
 BIOS-e820: 00000000ff000000 - 0000000100000000 (reserved)

I'm now thinking these might not be only PCI devices, but also other things
like interrupt tables, BIOS data, VGA memory, and who knows what else.
I'm way outside my expertise area here.
See also http://en.wikipedia.org/wiki/E820

In any case, my mail had two other suspects to consider ;-)

-- 
Nadav Har'El                        |                 Wednesday, Mar 28 2012, 
nyh at math.technion.ac.il             |-----------------------------------------
Phone +972-523-790466, ICQ 13349191 |Always borrow money from pessimists. They
http://nadav.harel.org.il           |don't expect to be paid back.