Basics

Which Functions Can Be Ported and Replaced by Kunpeng NEON Intrinsic Functions?

The NEON intrinsic functions on the Kunpeng platform are used to replace Intel intrinsics.

Intel intrinsics are developed to fully explore the CPU performance of the x86 architecture and are closely related to the x86 CPU architecture. These functions are incompatible with the Arm architecture. To enable these functions on Arm-based platforms, the industry has created the open source sse2neon project. Huawei provides the AVX2KI open source solution to address this issue. AVX2KI covers more functions and provides better performance and compatibility than sse2neon. You can add a header file (avx2ki.h) to support intrinsic functions in the Arm systems. In addition, download the header files related to AVX2KI to the directory where modified files are located. For details, see Kunpeng System Library Developer Guide.

How to Query CVEs?

Common Vulnerabilities and Exposures (CVEs) is a catalog of publicly disclosed cyber security vulnerabilities. IT personnel and security researchers query CVEs to obtain vulnerability details and determine the priority of vulnerability fixing based on the vulnerability score. In CVE, each vulnerability is numbered in the format of CVE-1999-0067, CVE-2014-10001, or CVE-2014-100001. A CVE ID is the unique identifier of a vulnerability. CVE IDs are assigned by the CVE Numbering Authority (CNA), which is undertaken by IT suppliers, security vendors, and security research organizations.

You can query vulnerability information on the official CVE website or NVD website.

What Are the Differences Between the x86 Platform and the Kunpeng Platform?

For details about the differences, contact Huawei engineers.

What Is DFX?

DFX stands for Design for X, where "X" can represent a specific phase or feature during the product lifecycle. It is a new design technique that optimizes or redesigns products by considering factors such as performance, quality, manufacturability, assembly, testability, service, and price as early as possible in the design phase.

The basic attributes of DFX are as follows:

Attribute Type	Full Spelling	Description
DFA	Design for Assembly	Analyzes the mating relationships between components to optimize design and improve assembly efficiency.
DFA	Design for Availability	Ensures that when the device is in operation, the duration of service or functional downtime is minimized.
DFC	Design for Compatibility	Ensures that the product complies with relevant standards, interconnects with other devices, and maintains compatibility after version upgrades.
	Design for Compliance	Ensures that the product complies with relevant standards, regulations, and conventions to guarantee market access.
	Design for Cost	Minimizes product costs on the premise of meeting user requirements.
DFD	Design for Diagnosability	Improves the ability to accurately and effectively locate faults when the product is faulty.
	Design for Disassembly	Enables easy product disassembly for recycling.
	Design for Discard	Enables easy replacement of faulty components, without the need for repair.
DFE	Design for Environment	Reduces the negative impact on the environment during the product lifecycle.
DFE	Design for Extensibility	Enables easy addition of new features or modification of existing features.
DFEE	Design for Energy Efficiency	Reduces the product's power consumption and improves energy efficiency.
DFF	Design for Flexibility	Considers the flexibility of architecture and interfaces during design to adapt to system changes.
DFF	Design for Fabrication of the PCB	Ensures that the PCB design meets the requirements for manufacturability.
DFH	Design for Humanity/Ergonomics	Emphasizes that the product design should meet the spiritual and emotional needs of people.
DFI	Design for Installability	Increases the efficiency in project installation, debugging, and acceptance.
	Design for International	Ensures that the product meets internationalization requirements.
	Design for interoperability	Ensures that the product is interoperable with other related devices.
DFL	Design for Logistics	Reduces logistics costs such as product packaging, transportation, and customs clearance, and improves logistics efficiency.
DFM	Design for Migrationability	Ensures that the system is migratable and upgradable.
	Design for Maintainability	Ensures high maintainability and maintenance efficiency.
	Design for Manufacturability	Ensure a high first-pass yield (FPY) during the manufacturing phase.
DFP	Design For Packaging	Ensures easy packaging.
	Design for Portability	Ensures that the system is portable from one platform to another.
	Design for Performance	Enhances system performance by reducing latency and improving throughput and resource utilization.
	Design for Procurement	Ensures convenient and low-cost material procurement while fulfilling product functionality and performance.
	Design for Postponement	Supports delaying the fulfillment of differentiated customer requirements to the downstream stages of the supply chain.
DFQ	Design For Quality	Ensures high product quality.
DFR	Design for Recycling	Ensures easy product recycling.
	Design for Reliability	Ensures that the product meets user requirements during product running, including reducing the occurrence of faults, minimizing the impact of faults, and recovering from faults as soon as possible.
	Design for Repair	Ensures ease of product maintenance.
	Design for Reusability	Ensures that the product design or modules can be used by later versions or other products, improving development efficiency.
DFS	Design for Safety	Incorporates considerations into product design to protect personal safety during product use.
DFS	Design for Scalability	Ensures that the system capacity can be effectively scaled.
DFS	Design for Security	Minimizes the vulnerability of assets and resources. This encompasses key security dimensions, including confidentiality, integrity, availability, access control, authentication, non-repudiation, and privacy protection.
	Design for Serviceability	Improves installation, debugging, maintenance, and overall service efficiency.
	Design for Simplicity	Reduces product components and complexity, and reduces material, supply, and maintenance costs.
	Design for Sustainability	Focuses on the interaction between sustainable raw materials, production, and consumption.
DFSC	Design for Supply Chain	Enhances supply efficiency and inventory turnover and reduces delivery time.
DFT	Design for Testability	Incorporates special design into the circuitry to facilitate product testing and enhance the capabilities for fault detection, location, and isolation. Practical measures include increasing test points (via copper exposure) and providing comprehensive silkscreen markings (component identifiers, orientation, and descriptions) on the PCB.
DFU	Design for Upgradeability	Enhances the ease of upgrade.
DFU	Design for Usability	Enhances the ease of use, effectiveness, and efficiency for users.
DFV	Design for Variety	Handles diverse product requirements. It balances the diverse needs of customers with the economic benefits of large-scale supply.

The following figure shows the relationships between the features:

DFX is an advanced product development technique in which the development process and system design shall satisfy product functionality and performance requirements. Meanwhile, products shall demonstrate high quality, reliability, and cost-effectiveness. Only products developed in this manner can achieve market recognition.

Relationship and Differences Between ZAB and Paxos Algorithms

Similarities:
- Both of them have a role similar to the leader process, which coordinates the running of multiple follower processes.
- The leader process submits a proposal only after more than half of the follower processes give correct feedback.
- In the ZAB protocol, each proposal contains an epoch value that represents the current leader period. In Paxos, such value is called ballot.
Differences: ZAB is used to build a high-availability distributed primary-secondary data system (ZooKeeper), and Paxos is used to construct a distributed consensus state machine system.

Linux Disk Management

Disk management

Function: The fdisk command is used to manage disk partitions in an interactive operating environment.

Syntax: fdisk [parameter] [disk device]

Parameter description:

In Linux, IDE disks are named hda and hdb. SAS/SATA/SCSI disks are named sda and sdb. sda1, sda2, and sdb are partitions.
The fdisk command can be followed by the parameters listed in Table 1.

**Table 1** fdisk parameter description
fdisk	Description
-l	Views information about all disks and partitions.
n	Creates a new partition.
d	Deletes a partition.
p	Prints the partition table.
w	Writes the partition table to the disk and exits.
e	Makes an extended partition.

The following is an example of the fdisk -l command:

LVM mechanism

Logical Volume Manager (LVM) is an abstract layer built on physical storage devices. It implements dynamic disk capacity adjustment and allows logical volumes to be generated, which is more flexible than physical storage management. The GUI management tool is system-config-lvm. The process of creating logical volumes is as follows: Initialize disks or partitions as physical volumes (PVs), add several PVs to a volume group (VG), and then divide the VG into logical volumes (LVs). The LVs can be formatted and mounted as common partitions.

A PV is an entire disk or a common partition created using a utility like fdisk. A PV includes many physical extents (PEs), each with a default size of 4 MB.
A VG is a collection of one or more PVs, which may originate from different partitions on the same or multiple disks.
An LV is a piece of space split from a VG and used to create a file system.
A PE is the minimum unit used to store data. The PE size is configurable. The default size is 4 MB.

The /boot partition stores boot files and cannot be created on LVM.

The following table lists major commands:

Command Function	PV Management	VG Management	LV Management
Scan	pvscan	vgscan	lvscan
Create	pvcreate	vgcreate	lvcreate
Display	pvdisplay	vgdisplay	lvdisplay
Remove	pvremove	vgremove	lvremove
Extend	-	vgextend	lvextend
Reduce	-	vgreduce	lvreduce