Hello I would like to share my experience with ram issue on gentoo (1.12.13) I am running 64bit xen fedora kernel 2.6.31-6 (http://wiki.xensource.com/xenwiki/XenParavirtOps) Why? - because previous 2.6.31-r7 from http://code.google.com/p/gentoo-xen-kernel/downloads/list was unstable and also only 3GB ram were visible.
There are 4GB ram installed on server. With standard kernel (2.6.30-r5), I am able to see it all. Because of this, I disqualify grub bug. With xen kernel, I saw only 3GB. xm dmesg: (XEN) Xen-e820 RAM map: (XEN) 0000000000000000 - 000000000009f000 (usable) (XEN) 000000000009f000 - 00000000000a0000 (reserved) (XEN) 00000000000f0000 - 0000000000100000 (reserved) (XEN) 0000000000100000 - 00000000bbeb0000 (usable) <----------- 3GB (XEN) 00000000bbeb0000 - 00000000bbee3000 (ACPI NVS) (XEN) 00000000bbee3000 - 00000000bbef0000 (ACPI data) (XEN) 00000000bbef0000 - 00000000bbf00000 (reserved) (XEN) 00000000bc000000 - 00000000c0000000 (reserved) (XEN) 00000000f0000000 - 00000000f4000000 (reserved) (XEN) 00000000fec00000 - 0000000100000000 (reserved) (XEN) 0000000100000000 - 0000000140000000 (usable) <--------- 1 GB (XEN) System RAM: 4030MB (4127036kB) I do not know how this is possible: 140000000 = 5120MB and only 4030MB are installed. In standard dmesg output (just "dmesg" in dom0, not "xm dmesg") I saw only 3GB address space. I double checked kernel configuration and compared with standard kernel config, however no dirrefence found (except xen options). lnx2 linux-2.6.31-6-xen # grep -i mem .config | grep -v "#" CONFIG_SHMEM=y CONFIG_XEN_MAX_DOMAIN_MEMORY=32 CONFIG_ARCH_SPARSEMEM_DEFAULT=y CONFIG_ARCH_SPARSEMEM_ENABLE=y CONFIG_ARCH_SELECT_MEMORY_MODEL=y CONFIG_SELECT_MEMORY_MODEL=y CONFIG_SPARSEMEM_MANUAL=y CONFIG_SPARSEMEM=y CONFIG_HAVE_MEMORY_PRESENT=y CONFIG_SPARSEMEM_EXTREME=y CONFIG_SPARSEMEM_VMEMMAP_ENABLE=y CONFIG_SPARSEMEM_VMEMMAP=y CONFIG_X86_BOOTPARAM_MEMORY_CORRUPTION_CHECK=y CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y CONFIG_INPUT_FF_MEMLESS=y CONFIG_DEVKMEM=y CONFIG_FIX_EARLYCON_MEM=y CONFIG_FIRMWARE_MEMMAP=y CONFIG_DEBUG_MEMORY_INIT=y CONFIG_HAVE_ARCH_KMEMCHECK=y CONFIG_ASYNC_MEMCPY=y CONFIG_HAS_IOMEM=y lnx2 linux-2.6.31-6-xen # grep -i ram .config | grep -v "#" CONFIG_X86_TRAMPOLINE=y CONFIG_SCHED_OMIT_FRAME_POINTER=y CONFIG_X86_BOOTPARAM_MEMORY_CORRUPTION_CHECK=y CONFIG_BLK_DEV_RAM=y CONFIG_BLK_DEV_RAM_COUNT=16 CONFIG_BLK_DEV_RAM_SIZE=16384 CONFIG_NVRAM=y CONFIG_FRAMEBUFFER_CONSOLE=y CONFIG_FRAME_WARN=2048 CONFIG_ARCH_WANT_FRAME_POINTERS=y CONFIG_FRAME_POINTER=y lnx2 linux-2.6.31-6-xen # Another symptom: xm top: ======= dom0 = 4GB xm list: ======== dom0= 3GB Once VPS was started, dom0 ram size was decreased. So only 3GB were available for dom0 + VPS As a workaround I added dom0_mem=512M to kernel line in grub.conf title XEN-6-fedora root (hd0,0) kernel /boot/xen.gz dom0_mem=512M module /boot/vmlinuz-2.6.31-6-xen-dom0 root=/dev/md0 acpi_enforce_resources=lax xm top: ======= dom0 = 512MB xm list: ======== dom0= 512MB Now, I can see only 512MB in dom0, however rest of memory (3,5GB) is usable for VPS :) :) :) I was searching on google over 2 months, but noone report this issue (some users had turned off PAE on 32bit, some had wrongly configured BIOS). So I hope that this mail will help to someone who have the same problem, or that someone will find reason of this strangeness. Thanks and regards samox86 l...@lnx.sk
