Hi, BIOS Settings show ===================================== Total Memory : 4096 MB with 8 MB Shared Memory : Dual - Channel Memory Mode DDRII1 : 2048 MB/266 MHz (DDRII533) DDRII2 : 2048 MB/266 MHz (DDRII533) ====================================== There are no setting to limit RAM in BIOS and so on also. but OpenBSD 4.7/amd64 SMP detects only 3 GB. Is there anything more I should do to get the other 1 GB of RAM recognized by the System? Thanks --Siju dmesg below-------------- OpenBSD 4.7 (GENERIC.MP) #130: Wed Mar 17 20:48:50 MDT 2010 email@example.com:/usr/src/sys/arch/amd64/compile/GENERIC.MP real mem = 3077242880 (2934MB) avail mem = 2987610112 (2849MB) mainbus0 at root bios0 at mainbus0: SMBIOS rev. 2.3 @ 0xfbc30 (12 entries) bios0: vendor American Megatrends Inc. version "P1.00" date 06/12/2006 acpi0 at bios0: rev 0 acpi0: tables DSDT FACP APIC MCFG OEMB acpi0: wakeup devices P0P2(S4) P0P1(S4) PS2K(S4) PS2M(S4) UAR1(S4) EUSB(S4) MC97(S4) HDAC(S4) P0P4(S4) P0P5(S4) P0P6(S4) P0P7(S4) P0P8(S4) P0P9(S4) USB0(S4) USB1(S4) USB2(S4) USB3(S4) SLPB(S4) acpitimer0 at acpi0: 3579545 Hz, 24 bits acpimadt0 at acpi0 addr 0xfee00000: PC-AT compat cpu0 at mainbus0: apid 0 (boot processor) cpu0: Intel(R) Pentium(R) D CPU 2.66GHz, 2659.72 MHz cpu0: FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CNXT-ID,CX16,xTPR,LONG cpu0: 1MB 64b/line 8-way L2 cache cpu0: apic clock running at 132MHz cpu1 at mainbus0: apid 1 (application processor) cpu1: Intel(R) Pentium(R) D CPU 2.66GHz, 2659.31 MHz cpu1: FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,TM2,CNXT-ID,CX16,xTPR,LONG cpu1: 1MB 64b/line 8-way L2 cache ioapic0 at mainbus0: apid 2 pa 0xfec00000, version 20, 24 pins ioapic0: misconfigured as apic 1, remapped ...
This is normal. Large memory support is not yet included in OpenBSD by default for amd64.
On Thu, May 27, 2010 at 3:22 PM, Peter Kay (Syllopsium) Is there anything I can do to get this by recompiling the kernel or something? Thanks --Siju
Yes and no; google for 'bigmem' in the misc list for marc.info. You have to edit a source file and recompile the kernel. However, it won't work (to be precise, it will probably crash on boot, or possibly afterwards) unless you have an IOMMU, and most Intel systems don't. OpenBSD doesn't support the AGP/PCI-e GART as an IOMMU, and I'm not sure if it supports VT-d platforms (which you probably aren't running anyway). The only option here is AMD. Until the devs tell us it's working, it's not worth persevering with. PK
bad advice. Don't do it, there is a reason it isn't disabled. send oga some beer for c2k10 instead he might feel more pressure that way :-)
Forgive me for the noob question (i'm a newbie at openbsd), but if i want to build, for example, a large squid cache using openbsd, in a server with BIIIG ram (12gb+), i will no be able to use the full memory space? is this what you guys are saying?
That is correct. 12G of ram for web cache? Why not a better disk instead?
Hi Marco, tks for your anwser. Maybe i have not choosed the best example, but the question is answered. IMHO, this serious limits the use of OpenBSD in tasks that takes huge memory space to execute. There is an serious effort to avoid this limitation? I'm not pushing anything, and i'm fully aware that developers does this system because they want, and not to satisfy my personal needs. I just want to be well informed to not talk any bulls**t. Tks in advance and sorry for my poor english.
Yep, there are plans to work on it during c2k10 hackathon. Support for big mem was enabled in the past (in 4.4 for some short time???), but there were problems with it. Of course you can try to enable it by your self. It's easy change and how to do that is in couple of threads in mailing list archives. So if you need this amount of memory then you can try to enable it or maybe use different OS for now. No one OS is best for everything. Not even OpenBSD which is great in so many areas. On Thu, May 27, 2010 at 7:48 PM, Leonardo Carneiro - Veltrac
Maybe people need to sit down and invent cases in order to figure out why this is a REAL, SERIOUS ISSUE(tm) and later figure that the invented case wasn't valid, but there must be an ISSUE anyhow. Sure, everybody wants to use all ram, but unless someone says "it might blow up if you experiment", I'd rather stick to a 3+ G squid cache than experiment with bouncebuffered devices and I-dont-know-how-well-it-works-IOMMUs if I was serious about web caching. -- To our sweethearts and wives. May they never meet. -- 19th century toast
It's not because you or me didn't said about an example that use tons of RAM, doesn't mean that such scenarios does not exist. I develop a application in java that serves more than ten thousand clients, and in top hours it's sure to use at least 2,5gb (this single process). I didn't want to expose my personal example, but if you want one, here it is.
Incorrect, that is all we support on amd64 right now. there are a few machines where it mystically does not work, though. That Being worked on, needs some changes that hopefully will be done at c2k10, we definitely have access to a machine with the intel VT stuff, the amd64 new shiny iommu i'm not sure if we have hardware yet. That is the best approach for now. -0- -- Stupid, adj.: Losing $25 on the game and $25 on the instant replay.
Thanks to all of you for the latest info :-) --Siju
|Rafael J. Wysocki||[Bug #11409] build issue #564 for v2.6.27-rc4 : undefined reference to `NS8390p_in...|
|Greg Kroah-Hartman||[PATCH 20/36] Driver core: Call device_pm_add() after bus_add_device() in device_a...|
|Aaron Straus||Re: [NFS] blocks of zeros (NULLs) in NFS files in kernels >= 2.6.20|
|Avi Kivity||Re: [kvm-ia64-devel] [PATCH] ia64 kvm fixes for O=... builds|
|Gerrit Renker||v2 [PATCH 1/4] dccp: Limit feature negotiation to connection setup phase|
|Daniel Lezcano||getsockopt(TCP_DEFER_ACCEPT) value change|
|David Miller||Re: 220.127.116.11: bnx2/tg3: BUG: "scheduling while atomic" trying to ifenslave a seco...|
|Ingo Molnar||Re: [regression] nf_iterate(), BUG: unable to handle kernel NULL pointer dereference|
|Gerrit Renker||[PATCH 37/37] dccp: Debugging functions for feature negotiation|
|Junio C Hamano||Re: git-svnimport|
|Anuj Gakhar||Git Architecture Question|
|Johannes Schindelin||Re: [PATCH] Fix approxidate("never") to always return 0|
|A Large Angry SCM||Re: [RFC] origin link for cherry-pick and revert|
|Gabriel||[PATCH] When a remote is added but not fetched, tell the user.|
|Jason Dixon||Re: any web management gui for pf ?|
|Christophe Rioux||Implementation example of snmp|
|Nick Holland||Re: booting openbsd on eee without cd-rom|
|Bryan Irvine||Re: OpenBSD 4.7 Released, May 19 2010|
|Marco Peereboom||Re: Singularity OS|
|Alexander Graf||Re: [Qemu-devel] [PATCH] Inter-VM shared memory PCI device|
|Takuya Yoshikawa||Re: [Qemu-devel] [RFC PATCH 00/20] Kemari for KVM v0.1|
|Marcelo Tosatti||Re: [PATCH v4 0/2] qemu-kvm: Save&restore debug registers|
|Avi Kivity||Re: [patch 6/6] kvm: remove explicit kvm_arch_reset_vcpu from kvm_init_vcpu|
|David S. Ahern||RHEL5.5, 32-bit VM repeatedly locks up due to kvmclock|