Deep dive into Kernel

static void continue_new_vcpu(struct vcpu *prev){ schedule_tail(prev); if ( is_idle_vcpu(current) ) reset_stack_and_jump(idle_loop); else /* check_wakeup_from_wait(); */ reset_stack_and_jump(return_to_new_vcpu); return_to_new_vcpu -> return_to_hypervisor ENTRY(return_to_hypervisor) cpsid i ldr lr, [sp, #UREGS_lr] ldr r11, [sp, #UREGS_pc] msr ELR_hyp, r11 ldr r11, [sp, #UREGS_cpsr] msr SPSR_hyp, ..

o CentOS 6.3에서 Xen Hypervisor 빌드 CentOS 6.3 x86_64 버전을 minimal 로 설치함 ¡ Xen 소스 기반 빌드 빌드 툴 설치 #(CentOS5)용 yum groupinstall "Development Libraries" -> CentOS6에서는 "Additional Development"로 변경 yum groupinstall "Additional Development" yum groupinstall "Development Tools" yum install transfig wget tar less texi2html libaio-devel dev86 glibc-devel e2fsprogs-devel gitk mkinitrd iasl xz-devel bzip2-devel..

1. Xen Interface Manual 1.1 Virtual Architecture o CPU state 모든 privileged state 는 Xen에 의해서 처리된다. 따라서 Guest OS는 CR3[1] 나 EFLAGS 에 대한 direct access가 제한된다. 이를 해결하기 위해선 Guest OS는 Hypercall을 통해서 필요한 작업들을 수행할 수 있다. ToDO: CPU ring 값과 의미 o Exception Guest OS는 Exception(Trap)을 처리하기 위해서 virtual IDT[2]를 가지고 있다. Domain는 자신의 Trap handler table을 set_trap_table Hypercall을 이용해서 Hypervisor에게 전달한다. o Interrupts ..

On Mon, Mar 04, 2013 at 02:09:06PM +0900, Choonho Son wrote: > Hi all, > > I have a machine with Intel Xeon E5-2630. Intel E5-series supports > AVX(Advanced Vector eXtension). And Xen Hypervisor 4.1 also supports > AVX feature. Lets CC the Intel folks since these two git commits point to them: commit 25050f53be29518a3c9bdd0628333cd3121d2750 Author: Liu, Jinsong Date: Tue Dec 6 11:27:18 2011 +000..

Xen Network I/O Architecture TX 방식 [DomU]1. DomU creates a grant reference providing access to a page2. DomU forwards the reference as part of the I/O buffer descriptor 3. Invoke a hypervall [Dom0=driver domain]1. map the guest page into its address space before sending the request through bridge2. physical device send the packet3. device driver frees the pages a callback function is automatical..

Xen Network의 overhead1. Each received packet must traverse the ethernet bridge to arrive ata the appropriate back-end driver2. additional data copy required to copy a received packet from driver domain to guest memory3. substantial overheads associated with the grant mechanism Paper: @inproceedings{ram2009achieving, title={Achieving 10 Gb/s using safe and transparent network interface virtualiza..

iface eth0 inet static address x.x.x.x post-up ethtool -K eth0 tx off

Oracle VM 3.0 버전이 릴리스 되었네요. News Facts Today at an event for customers, partners and industry experts, Oracle announced Oracle VM 3.0, the latest release of Oracle’s server virtualization and management solution. Oracle VM 3.0 is suitable for all datacenter workloads and features new policy-based management capabilities, advanced storage management via the Oracle VM Storage Connect plug-in API; ..

mount.ocfs2: Unable to access cluster service while trying to join the group In this error, start cluster service first, /etc/init.d/o2cb start ocfs2 the last word "ocfs2" is same with cluster name in /etc/ocfs2/cluster.conf

I failed to add OVM server in a cluster. After debugging log file, I found the problem. The problem is: 1) when adding new node, ovs-agent parses /etc/ocfs2/cluster.conf - using /opt/ovs-agent-latest/utils/config_o2cb.sh 2) This script has problem - it parse some keyword like 'node' , 'cluster' - if hostname has the word like 'node', it fail to parse - and ocfs service fails The solution is: 1) ..