Disabling physical RAM regions for testing

Disabling physical RAM regions for testing

[martin f krafft]

[-- Attachment #1: Type: text/plain, Size: 1679 bytes --]

Hello,

A NAS seems to be having RAM problems, leading me to want to swap
out DIMMs until I found the offender; I'd rather not go to the rack
every time I need to make a change.

Is there a way to declare a DIMM unavailable to Linux with a boot
parameter? There is mem= and memmap= but i cannot figure them out.

I have two 4Gb DIMMs in there and I just want to disable one, then
the other…

This is the e820 physical RAM map:

[    0.000000] e820: BIOS-provided physical RAM map:
[    0.000000] BIOS-e820: [mem 0x0000000000000000-0x000000000009e7ff] usable
[    0.000000] BIOS-e820: [mem 0x000000000009e800-0x000000000009ffff] reserved
[    0.000000] BIOS-e820: [mem 0x00000000000e2000-0x00000000000fffff] reserved
[    0.000000] BIOS-e820: [mem 0x0000000000100000-0x00000000bff8ffff] usable
[    0.000000] BIOS-e820: [mem 0x00000000bff9e000-0x00000000bff9ffff] reserved
[    0.000000] BIOS-e820: [mem 0x00000000bffa0000-0x00000000bffaa5ff] ACPI data
[    0.000000] BIOS-e820: [mem 0x00000000bffaa600-0x00000000bffdffff] ACPI NVS
[    0.000000] BIOS-e820: [mem 0x00000000bffe0000-0x00000000bfffffff] reserved
[    0.000000] BIOS-e820: [mem 0x00000000e0000000-0x00000000efffffff] reserved
[    0.000000] BIOS-e820: [mem 0x00000000ffa00000-0x00000000ffffffff] reserved
[    0.000000] BIOS-e820: [mem 0x0000000100000000-0x000000022fffffff] usable

But what do I do with those info?

-- 
@martinkrafft | http://madduck.net/ | http://two.sentenc.es/
 
"aus der kriegsschule des lebens -
 was mich nicht umbringt, macht mich härter."
                                                 - friedrich nietzsche
 
spamtraps: madduck.bogus@madduck.net

[-- Attachment #2: Digital signature (see http://martin-krafft.net/gpg/sig-policy/999bbcc4/current) --]
[-- Type: application/pgp-signature, Size: 1107 bytes --]

Re: Disabling physical RAM regions for testing

[Austin S Hemmelgarn]

[-- Attachment #1: Type: text/plain, Size: 1707 bytes --]

On 2014-08-05 07:40, martin f krafft wrote:
> Hello,
> 
> A NAS seems to be having RAM problems, leading me to want to swap
> out DIMMs until I found the offender; I'd rather not go to the rack
> every time I need to make a change.
> 
> Is there a way to declare a DIMM unavailable to Linux with a boot
> parameter? There is mem= and memmap= but i cannot figure them out.
> 
> I have two 4Gb DIMMs in there and I just want to disable one, then
> the other…
> 
> This is the e820 physical RAM map:
> 
> [    0.000000] e820: BIOS-provided physical RAM map:
> [    0.000000] BIOS-e820: [mem 0x0000000000000000-0x000000000009e7ff] usable
> [    0.000000] BIOS-e820: [mem 0x000000000009e800-0x000000000009ffff] reserved
> [    0.000000] BIOS-e820: [mem 0x00000000000e2000-0x00000000000fffff] reserved
> [    0.000000] BIOS-e820: [mem 0x0000000000100000-0x00000000bff8ffff] usable
> [    0.000000] BIOS-e820: [mem 0x00000000bff9e000-0x00000000bff9ffff] reserved
> [    0.000000] BIOS-e820: [mem 0x00000000bffa0000-0x00000000bffaa5ff] ACPI data
> [    0.000000] BIOS-e820: [mem 0x00000000bffaa600-0x00000000bffdffff] ACPI NVS
> [    0.000000] BIOS-e820: [mem 0x00000000bffe0000-0x00000000bfffffff] reserved
> [    0.000000] BIOS-e820: [mem 0x00000000e0000000-0x00000000efffffff] reserved
> [    0.000000] BIOS-e820: [mem 0x00000000ffa00000-0x00000000ffffffff] reserved
> [    0.000000] BIOS-e820: [mem 0x0000000100000000-0x000000022fffffff] usable
> 
> But what do I do with those info?
> 
The e820 is probably not what you want to be looking at.  I would
suggest looking into memory hot-plug/hot-remove support, IIRC the
hot-remove support allows you to bring a bank offline.


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 2967 bytes --]

Re: Disabling physical RAM regions for testing

[martin f. krafft]

[-- Attachment #1: Type: text/plain, Size: 1182 bytes --]

also sprach Austin S Hemmelgarn <ahferroin7@gmail.com> [2014-08-05 14:26 +0200]:
> The e820 is probably not what you want to be looking at.  I would
> suggest looking into memory hot-plug/hot-remove support, IIRC the
> hot-remove support allows you to bring a bank offline.

Oh, good pointer, yeah, that would do and allow for online testing
to some degree.

I found http://feishare.com/linux/memory-hotplug-in-linux and
memory0 through memory69 in /sys/devices/system/memory, and only
3 of those are "removable==1".

My goal is to bring one of the two 4Gb DIMMs offline, verify that
the error I am seeing still exists, bring it back online and take
the other offline.

How do I find out which memory to hot-remove?

Am I right in assuming that I won't be able to remove
one-then-the-other as some memory regions may not be movable?

-- 
@martinkrafft | http://madduck.net/ | http://two.sentenc.es/
 
"the scientific paper in its orthodox form does embody a totally
 mistaken conception, even a travesty, of the nature of scientific
 thought."
                                                -- sir peter medawar
 
spamtraps: madduck.bogus@madduck.net

[-- Attachment #2: Digital signature (see http://martin-krafft.net/gpg/sig-policy/999bbcc4/current) --]
[-- Type: application/pgp-signature, Size: 1107 bytes --]