RAM: 24GB is not recognized (only 12)
 

Hi,

i have ASRock X58 Supercomputer mobo, which supports 24GB , DDR3. I have 6 DIMMs (each of 4GB) installed, i can see them all recognized in BIOS menu. However when i boot Linux kernel 2.6.28 (Gentoo R5) only 12GB are recognized. I tried mem=24GB on boot, but no luck. What could it be?
Could it be the page size? Or maybe because i only have 4GB swap ?

Thanks in advance.

Attached is the dmesg output for ram maping:

[ 0.000000] BIOS EBDA/lowmem at: 0009b800/0009b800
[ 0.000000] Initializing cgroup subsys cpuset
[ 0.000000] Initializing cgroup subsys cpu
[ 0.000000] Linux version 2.6.28-gentoo-r5 (root@master) (gcc version 4.3.2 (Gentoo 4.3.2-r3 p1.6, pie-10.1.5) ) #4 SMP
Tue May 26 11:27:16 Local time zone must be set--see zic
[ 0.000000] Command line: root=/dev/sda2 mem=24GB
[ 0.000000] KERNEL supported cpus:
[ 0.000000] Intel GenuineIntel
[ 0.000000] AMD AuthenticAMD
[ 0.000000] Centaur CentaurHauls
[ 0.000000] BIOS-provided physical RAM map:
[ 0.000000] BIOS-e820: 0000000000000000 - 000000000009b800 (usable)
[ 0.000000] BIOS-e820: 000000000009b800 - 00000000000a0000 (reserved)
[ 0.000000] BIOS-e820: 00000000000e4000 - 0000000000100000 (reserved)
[ 0.000000] BIOS-e820: 0000000000100000 - 00000000bbe90000 (usable)
[ 0.000000] BIOS-e820: 00000000bbe90000 - 00000000bbea0000 (ACPI data)
[ 0.000000] BIOS-e820: 00000000bbea0000 - 00000000bbed0000 (ACPI NVS)
[ 0.000000] BIOS-e820: 00000000bbed0000 - 00000000bbee0000 (reserved)
[ 0.000000] BIOS-e820: 00000000bbeeb800 - 00000000bc000000 (reserved)
[ 0.000000] BIOS-e820: 00000000fee00000 - 00000000fee01000 (reserved)
[ 0.000000] BIOS-e820: 00000000ffb00000 - 0000000100000000 (reserved)
[ 0.000000] BIOS-e820: 0000000100000000 - 0000000344000000 (usable)
[ 0.000000] user-defined physical RAM map:
[ 0.000000] user: 0000000000000000 - 000000000009b800 (usable)
[ 0.000000] user: 000000000009b800 - 00000000000a0000 (reserved)
[ 0.000000] user: 00000000000e4000 - 0000000000100000 (reserved)
[ 0.000000] user: 0000000000100000 - 00000000bbe90000 (usable)
[ 0.000000] user: 00000000bbe90000 - 00000000bbea0000 (ACPI data)
[ 0.000000] user: 00000000bbea0000 - 00000000bbed0000 (ACPI NVS)
[ 0.000000] user: 00000000bbed0000 - 00000000bbee0000 (reserved)
[ 0.000000] user: 00000000bbeeb800 - 00000000bc000000 (reserved)
[ 0.000000] user: 00000000fee00000 - 00000000fee01000 (reserved)
[ 0.000000] user: 00000000ffb00000 - 0000000100000000 (reserved)
[ 0.000000] user: 0000000100000000 - 0000000344000000 (usable)
[ 0.000000] DMI present.
[ 0.000000] AMI BIOS detected: BIOS may corrupt low RAM, working it around.
[ 0.000000] last_pfn = 0x344000 max_arch_pfn = 0x3ffffffff
[ 0.000000] x86 PAT enabled: cpu 0, old 0x7040600070406, new 0x7010600070106
[ 0.000000] last_pfn = 0xbbe90 max_arch_pfn = 0x3ffffffff
[ 0.000000] init_memory_mapping: 0000000000000000-00000000bbe90000
[ 0.000000] 0000000000 - 00bbe00000 page 2M
[ 0.000000] 00bbe00000 - 00bbe90000 page 4k
[ 0.000000] kernel direct mapping tables up to bbe90000 @ 10000-15000
[ 0.000000] last_map_addr: bbe90000 end: bbe90000
[ 0.000000] init_memory_mapping: 0000000100000000-0000000344000000
[ 0.000000] 0100000000 - 0344000000 page 2M
[ 0.000000] kernel direct mapping tables up to 344000000 @ 13000-22000
[ 0.000000] last_map_addr: 344000000 end: 344000000

Are you using a 64bit kernel?

of course, otherwise i couldn't have more than 4GB
master # uname -m
x86_64
master #

edited: ignore my response :)

damn, i put wrong dimms , everything with linux is fine , forget it.

The BIOS is telling Linux that you have approximately 12GB of ram.

I don't know what the problem is, but focus on BIOS and/or motherboard issues. The problem is not in Linux.

Edit: I didn't see your post above until after I posted.

Actually you can use over 4GB with a 32 bit (PAE) kernel. For 24GB of ram, I'm not sure whether it would be entirely unworkable or just a very poor choice. So no one is suggesting you should use a 32 bit bit kernel. But from your first post, detecting that you are not using a 32 bit kernel is beyond my expertise level. (I expect that if I understood the "kernel direct mapping tables" lines in your first post, that would have indicated a 64 bit kernel).

I assume that this means you have solved the problem, & that the solution was replacing incompatible RAM sticks (DIMMs).

Please thank those who helped you by:

1. Confirm that this is solved.
2. Explain in a little more detail:
   1. How you realized that DIMM replacement was the solution?
   2. What were the wrong DIMMs?
   3. What did you replace them w/?
   This is the most important of my requests, because it adds to the LQ archive of answers.
3. Mark this thread solved using the "Thread Tools".