Skip to main content

CPU Virtualization Deep Drive

Understand CPU Virtualization:

At this time most modern CPU architectures support different levels of privilege, called rings. Ring 0 has the highest privilege; processes running in this ring are referred to as running in supervisor or kernel mode. Processes running in higher rings are referred to as running in user mode. The higher the number of the ring, the less privilege a process running in this ring has. An x86-based CPU supports four levels of privilege, but in practice only rings 0 and 3 are used. In Linux and Windows, the operating system kernel and hardware drivers run in Ring 0, and user processes run in Ring 3. Only processes in Ring 0 can access hardware. If a process in a higher ring needs to access hardware, such as the hard disk, it has to use the APIs of the kernel.

Here is the demonstration of CPU Rings with Native Operating System:

When we will use Xen, the Xen Hypervisor runs in Ring 0, and the operating system kernel is moved to Ring 1. This is the reason why the operating system has to be modified to run para-virtualized—the kernel has to “know” that it is running in Ring 1 instead of Ring 0. 

Here is the demonstration of CPU Rings with Xen Hypervisor:
When we will have "VT-enabled hardware" allows the Hypervisor to emulate a Ring 0 for the operating system kernel, allowing it to run unmodified:

Here is the damonstration of CPU Rings and VT-Enabled Hardware:
That is done about CPU Virtualization: 


Comments

Post a Comment

Popular posts from this blog

Docker Container Management from Cockpit

Cockpit can manage containers via docker. This functionality is present in the Cockpit docker package. Cockpit communicates with docker via its API via the /var/run/docker.sock unix socket. The docker API is root equivalent, and on a properly configured system, only root can access the docker API. If the currently logged in user is not root then Cockpit will try to escalate the user’s privileges via Polkit or sudo before connecting to the socket. Alternatively, we can create a docker Unix group. Anyone in that docker group can then access the docker API, and gain root privileges on the system. [root@rhel8 ~] #  yum install cockpit-docker    -y  Once the package installed then "containers" section would be added in the dashboard and we can manage the containers and images from the console. We can search or pull an image from docker hub just by searching with the keyword like nginx centos.   Once the Image downloaded we can start a contai
Post a Comment
Read more

Remote Systems Management With Cockpit

The cockpit is a Red Hat Enterprise Linux web-based interface designed for managing and monitoring your local system, as well as Linux servers located in your network environment. In RHEL 8 Cockpit is the default installation candidate we can just start the service and then can start the management of machines. For RHEL7 or Fedora based machines we can follow steps to install and configure the cockpit.  Following are the few features of cockpit.  Managing services Managing user accounts Managing and monitoring system services Configuring network interfaces and firewall Reviewing system logs Managing virtual machines Creating diagnostic reports Setting kernel dump configuration Configuring SELinux Updating software Managing system subscriptions Installation of cockpit package.  [root@rhel8 ~] #  dnf   install cockpit cockpit-dashboard  -y  We need to enable the socket.  [root@rhel8 ~] #  systemctl enable --now cockpit.socket If firewall is runnin
Post a Comment
Read more

Containers Without Docker on RHEL/Fedora

Docker is perfectly doing well with the containerization. Since docker uses the Server/Client architecture to run the containers. So, even if I am a client or developer who just wants to create a docker image from Dockerfile I need to start the docker daemon which of course generates some extra overhead on the machine.  Also, a daemon that needs to run on your system, and it needs to run with root privileges which might have certain security implications. Here now the solution is available where we do not need to start the daemon to create the containers. We can create the images and push them any of the repositories and images are fully compatible to run on any of the environment.  Podman is an open-source Linux tool for working with containers. That includes containers in registries such as docker.io and quay.io. let's start with the podman to manage the containers.  Install the package  [root@rhel8 ~] # dnf install podman -y  OR [root@rhel8 ~] # yum
1 comment
Read more