2022-09 Monthly Archive

Misc ideas

Pressing the Enter key by itself will repeat the last command.

Linux has ELF, Windows has PE (Portable Executable), Apple (MacOS and iOS) has Mach-O and Unix System V has COFF.

root account always have UID/GID 0 linux - Does the root account always have UID/GID 0? - Super User.

At the file descriptor level, stdin* is defined to be file descriptor 0**, **stdout* is defined to be file descriptor 1; and *stderr* is defined to be file descriptor 2.

EPTP is like CR3 but for EPT. So constructing an extended page table is the responsibility of VMM just like constructing a normal page table is the responsibility of OS.

iTLB / dTLB / TLB Hierarchy

(2022) Modern processors typically use a two-level TLB hierarchy:

The first level (L1) is split in two components: the dTLB is used to cache translations triggered by data loads while the iTLB stores those triggered by instruction fetches.（也就是说取指产生的 TLB miss 放在 iTLB 里面而读数据产生的 TLB miss 放在 dTLB 里面）。
A second level (L2) TLB is consulted after an L1 miss and can be either shared (sTLB), storing translations from either memory access type, or itself split, handling data loads and instruction fetches independently。

sec22fall_tatar.pdf

Similar to caches, TLBs may have multiple levels. CPUs can be (and nowadays usually are) built with multiple TLBs, for example a small L1 TLB (potentially fully associative) that is extremely fast, and a larger L2 TLB that is somewhat slower. When instruction-TLB (ITLB) and data-TLB (DTLB) are used, a CPU can have three (ITLB1, DTLB1, TLB2) or four TLBs.

TLB 本身是缓存页表的。

https://en.wikipedia.org/wiki/Translation_lookaside_buffer#:~:text=Similar%20to%20caches,or%20four%20TLBs.

`arch/x86/include`

It has 2 sub-folders: asm and uapi/asm.

The asm denotes architecture-specific-macros, so it is put under this directory is not strange. It is not designed to be interacted with the userspace, so it is not in uapi.

uapi/asm : provide a standardized way for user-space applications to access the system call interface and other kernel features on x86-based systems. These header files define system call numbers, function prototypes, and data structures that are needed by user-space programs to interact with the kernel.

`include` Directory in the kernel source code

All the common headers are in include and its sub-directories, and the architecture-specific headers are in sub-directories of the arch/x86/include directory. In most cases you’d include headers from include.

When you compile the kernel source code, the compiler will automatically search for header files in the appropriate directories, including both the generic /include folder and the architecture-specific arch/x86/include folder.

Why kvm_host.h is located under the /include/linux folder?

Because it needs to be included, and it is not arch-specific.

When VCPU is in low overhead, can Host use the CPU resource to do other things?

Guest Non-Register State -> Activity state may has the value HLT, it denotes the logical processor is inactive because it executed the HLT instruction. (SDM 24.4.2 Guest Non-Register State)

I think the Host couldn't do the dynamic sharing this way, it can only do some statically sharing by the scheduling algorithm. For an example, if we use the CFS algorithm, then all the VMs will share the vcpu without regarding the real CPU utilization of each VM.

Need to be confirmed: virtualization - KVM CPU share / priority / overselling - Stack Overflow

How do we kick a vcpu down if it just don't VMexit?

Remember we still have the task switch function which is enforced by the clocking hardware. So it will vmexit when task switch happened and the VMM can determine which task to run next. (The VM is managed by QEMU for VMLAUNCH and VMRESUME, and QEMU is managed by the OS).

virtualization - KVM CPU share / priority / overselling - Stack Overflow

What does "Idle" mean for CPU? / HLT instruction / Halt instruction

TLDR; by halt (HLT).

This instruction is a privileged instruction, so userspace cannot execute this instruction directly.

There is an always ready process named "System Idle Process" (Windows), you can see it in task manager -> Details, and sort the process by the CPU utilization.

This process is essentially a wrapper for halt.

不要把进程挂起和 halt 指令混淆，当我们调用 sleep 之类函数时，暂停运行的只是进程，此时如果还有其它进程可以运行那么 CPU 是不会空闲下来的，当 CPU 开始执行 halt 指令时就意味着系统中所有进程都已经暂停运行。

CPU 空闲时在干嘛？ - 知乎

CPU 如何从 HLT 中恢复？

HALT 时，CPU 的 LAPIC 并不会停止工作，即 CPU 将会继续接收外部 中断、异常 等事件（事实上，CPU HALT 状态的退出将由外部事件触发）。当中断服务函数执行完毕后，RIP 会指向 HLT 的 下一条指令，即 CPU 继续执行 HLT 之后的程序。

CPU在空闲的时候做什么_河马虚拟化的博客-CSDN博客_native_safe_halt

TPM

The TPM is an actual co-processor living on the motherboard of your device.

TPM can store and encrypt data based on these measurements.

If you take the output generated by this TPM chip and that output was generated against specific measurements then that output can never be decrypted on another TPM unless it was somehow identical to your current device -- I'll demo this in a minute. This should never be the case even if it's the same model of laptop because the manufacturer would have put in different serial numbers.

TPM 芯片的工作原理是什么？ - furious 的回答 - 知乎 https://www.zhihu.com/question/35496371/answer/856294314

Style customization for Thunderbird

If you just want to make the "Unread" explict, you could just change the theme to "Dark".

Note: should first change toolkit.legacyUserProfileCustomizations.stylesheets to true.

Aris-t2/CustomCSSforTb: Custom user styles for Thunderbird 68+

Attestation Service

An attest service, or attestation service, is an independent review of a company's financial statement conducted by a certified public accountant (CPA).

Intel vs. AMD in Features

Intel	AMD	Who First?
Total Memory Encryption (TME)	Secure Memory Encryption (SME)	AMD
MKTME	Secure Encrypted Virtualization (SEV)	AMD
Extended Page Table (EPT)	Nested Page Table (NPT)	Don't know

Where to find the spec?

Index of /x86docs/

ELF symbols

Assembly is not ELF!

For example, in QEMU: they are defined in target/i386/cpu.h.

in KVM, they are defined under arch/x86/include/asm/cpufeatures.h and tools/arch/x86/include/asm/cpufeatures.h, which means some features are shared between Intel and AMD.

Authenticated Code Module (ACM)

A code module that is designed to be loaded, verified and executed by the CPU in on-chip memory (CRAM).

There are two types of ACM signing keys:

Debug key
Production key

There are three possible types of signed AC Modules:

Debug signed: signed by debug ACM key. Used for debugging and validation.
Production signed: signed by production ACM key. Can run only on production HW.
Production signed Non-production worthy (NPW): signed by production ACM key. Can run only on production HW. These modules are intended for debugging on production HW.

What is on-chip memory (CRAM)?

The on-chip SRAM, refers to data memory residing on-chip, that is mapped into an address space disjoint from the off-chip memory, but connected to the same address and data buses.

The primary characteristic of an OCRAM is its size. The amount of available on chip RAM can vary considerably between SoCs from a few kilobytes to many megabytes.

On chip RAM is, most often than not, implemented as a fast low-latency static RAM.

It usually requires little to no configuration before use meaning it can be available early within the boot sequence. OCRAM is often used as an important component of the boot process, providing a small quantity of RAM resources to the bootloader before the main external SDRAM is initialized.

Introduction to On-Chip RAM - JBLopen

Wrapper: failed with: gRPC call failed: Connection reset by peer

Turn off the network proxy.

5800X3D/3D V-Cache

V stands for "vertical", not "virtual".

3D V-Cache 是 AMD 开发的一项新技术，可以让 L3 缓存垂直堆叠，这样可以在占用很少空间的情况下显着增加缓存的大小。

How does shell work when you executing commands?

Run ls in shell process

I think P2 is a non-login, non-interactive shell.

Run exec ls in shell process

Create a new user / Add a new user (tested on Ubuntu, CentOS)

First, add new user lei:

useradd -ou 0 -g 0 -m lei
# if the user exists...
usermod -ou 0 -g 0 lei

When finished, execute:

visudo

Then add:

lei ALL=(ALL:ALL) ALL

alongside the line root ALL=(ALL:ALL) ALL.

Then

passwd lei

Then (On CentOS, you may NOT need this)

sudo usermod -aG sudo lei

Create a sudo user in Ubuntu -

Python LSP server parameters

python-lsp-server/CONFIGURATION.md at develop · python-lsp/python-lsp-server

Difference between `(1ULL << 2)` and `0x00000004`

The first one is 64 bit and is for 64 bits long register, the latter is for 32 bit long register.

Another advantange of 0x00000004 it can represent multi-bits in one time, and (1ULL << 2) is always used to represents a single bit value.

code-snippets/bit_definition.c at main · tristone13th/code-snippets

Tabstop, shiftwidth, softtabstop, expandtab, smarttab

tabstop: Tab width relative to a space char.

shiftwidth: When you indent a line, the width you indented is determined by shiftwidth.

if you set expandtab, then it will indent with spaces;
if you set noexpandtab:
- if the value of shiftwidth is an integer multiple of tabstop, then it will appear as several tabs;
- if not, then it will fill with tabs first, the remaining will be spaces.

softtabstop: The value is converted to tabs and spaces when you insert a tab. The only difference with shiftwidth is that it is triggered by inserting a Tab, not by indentation.

smarttab: determines whether shiftwidth is triggered or softtabstop is triggered.

expandtab: determines if a tab is expanded to spaces.

How to fix "You’ll Need a New App to Open This ms-gaming overlay"

How to Fix the 'You'll need a new app to open this ms-gaming overlay' Error on Windows 10? - Appuals.com

The Shadow Suite

encode password: one way hash function, This is an algorithm that is easy to compute in one direction, but very difficult to calculate in the reverse direction.

When a user picks or is assigned a password, it is encoded with a randomly generated value called the salt. The salt value is then stored with the encoded password.

prone to dictionary attack: an 8 character password encodes to 4096 * 13 character strings. So a dictionary of say 400,000 common words, names, passwords, and simple variations would easily fit on a 4GB hard drive. The attacker need only sort them, and then check for matches. Since a 4GB hard drive can be had for under $1000.00, this is well within the means of most system crackers.

The Shadow Suite solves the problem by relocating the passwords to another file (usually /etc/shadow). The /etc/shadow file is set so that it cannot be read by just anyone. Only root will be able to read and write to the /etc/shadow file.

Linux Shadow Password HOWTO: Why shadow your passwd file?

How to see which version of softwares Ubuntu is using

Open the manifest file in:

Ubuntu 22.04.1 LTS

if you also want to get the source code, you can:

How to get source code of package using apt on Debian/Ubuntu - nixCraft

How do I get and modify the source code of packages installed through apt-get? - Ask Ubuntu

Linux PAM

Pluggable Authentication Modules

Will be used by (not only) /usr/bin/login.

Understanding PAM - Linux.com

Linux utils like /usr/bin/login

util-linux/util-linux

Usermode-Helper API

IBM Developer

Shell type

A login shell is the first process that executes under your user ID when you log in for an interactive session. (You needn't to login again).

A Non-login shell needn't you to login again because you have logined.

Examples:

	Interactive	Non-interactive
Login	ssh/su - (bash_profile)	Rare (Non-interactive, then how to login?)
Non-login	Start a shell in a shell (bashrc)	.sh file

bashrc - What is the difference between interactive shells, login shells, non-login shell and their use cases? - Unix & Linux Stack Exchange

Linux init 进程

init 比较特殊，两个状态，刚开始是内核态，属于内核线程（kernel_init），然后转成用户态（init）（通过一个系统调用，开始执行用户空间的 /sbin/init 程序）。

Linux下1号进程的前世今生----Linux进程的管理与调度_CHENG Jian的博客-CSDN博客

Linux的init进程_开发联盟的博客-CSDN博客_状态init

Idle process

idle 进程，也就是 swapper 进程，其 pid 是 0，是所有进程的祖先。每个 cpu 都有一个 idle 进程。

// in kernel/sched/idle.c
// start_kernel –> rest_init
void cpu_startup_entry(enum cpuhp_state state)
{
	arch_cpu_idle_prepare();
	cpuhp_online_idle(state);
	while (1)
         //idle 进程是一个死循环，当没有 thread 处于 TASK_RUNNING 的时候，CPU 会跑这个 idle 进程，idle 的优先级最低
		do_idle();
}

每个cpu 都有一个idle进程._tiantao2012的博客-CSDN博客_do_idle

Change mermaid in typora to dark theme

Add %%{init: { 'theme': 'dark' } }%% to the first line.

mermaid - Markdownish syntax for generating flowcharts, sequence diagrams, class diagrams, gantt charts and git graphs.

What is gdb MI?

GNU Debugger Machine-Interface.

A frontend for gdb.

GDB/MI

Not only commits are recognized by Github as contribution to show up in the graph.

Issues, pull requests and discussions will appear on your contribution graph if they were opened in a standalone repository, not a fork.

Why are my contributions not showing up on my profile? - GitHub Docs

CONFIG_=n And # CONFIG_ is not set

CONFIG_=n It means that you are explicitly disabling this config item from your defconfig file. So in your .config file it will show like # CONFIG_ is not set.

# CONFIG_ is not set In this case also you are explicitly disabling this config. So generated config file (.config) will show like # CONFIG_ is not set.

However with Any of these options you can disable the config item; but as per @ Alexandre comment it seems that "not set" is the proper way.

c - Which approach to use for Kernel CONFIG_? - Stack Overflow

Uncore

"Uncore" is a term used by Intel to describe the functions of a microprocessor that are not in the core, but which must be closely connected to the core to achieve high performance.

Core 的内容：ALU, FPU, L1, L2 Cache；
Uncore 的内容：QPI controller, L3 Cache, On-die 内存控制器, On-die PCIE Root Complex 等等。

英特尔 Uncore 设计根源，来自于北桥芯片。Uncore 的设计是，将对于处理器核心有关键作用（不包括处理器本身比如 ALU） 的功能重新组合编排，从物理上使它们更靠近核心（集成至处理器芯片上，而它们有部分原来是位于北桥上），以降低它们的访问延时。

Intel微处理器Uncore架构简介-CSDN博客

Uncore - Wikipedia

How to examine an object file?

objdump and nm.

nm can list all symbols of an object file.

AT&T Assembly syntax

Use "#" as the comments mark.

https://gist.github.com/mishurov/6bcf04df329973c15044

Assembler Directives

All assembler directives have names that begin with a period (`.').

Using as - Assembler Directives

Kernel assembly style

Kernel use AT&T as the standard syntax.

GAS

Kernel Address Space Layout Randomization (or KASLR)

Two approachs for adding additional security for kernel:

Hiding the kernel addresses;
1. You cannot really hide it, because most systems out there are running stock distribution kernels. Getting the needed address information from the distributor's kernel package is not a particularly challenging task.
2. So, on these systems, there is no real mystery about the layout of the kernel, regardless of whether pointer values are allowed to leak to user space or not.
Randomizing the kernel addresses.
1. If the kernel calculated a different offset for itself at every boot, it could subtract that offset from pointer values before passing them to user space.
2. Those pointers could then be used by tools like perf, but they would no longer be helpful for anybody seeking to overwrite kernel data structures.
3. ASLR is a "statistical defense", because brute force methods can generally be used to overcome it.

Kernel address randomization [LWN.net]

Kernel address space layout randomization [LWN.net]

Exploit (security)

An exploit (from the English verb to exploit, meaning "to use something to one’s own advantage") is a piece of software, a chunk of data, or a sequence of commands that takes advantage of a bug or vulnerability to cause unintended or unanticipated behavior to occur on computer software, hardware, or something electronic.

Exploit - Wikipedia

Kick?

The goal of a VCPU kick is to bring a VCPU thread out of guest mode in order to perform some KVM maintenance. To do so, an IPI is sent, forcing a guest mode exit.

KVM VCPU Requests — The Linux Kernel documentation

Microcode (微码映射, 微程序), microops (micro-operations, uops, $\mu$ops, μops, 微操作), microinstructions (微指令)

Microcode 是一个从 instruction 到 micro-operation 的映射。

原始的 microcode 映射是从一个 ROM 中来的，为了能在 ROM 上动态打补丁，新设计了一块相对比较小的 SRAM。既然是在 SRAM 上进行，掉电以后就需要重新打补丁了，每次重启都要重新打补丁。打补丁的方式也很简单，就是把 Microcode 放在某个内存中，写个 MSR 就行了。 An update to microcode can allow a CPU's low-level behaviour to be modified to work around certain yet to be discovered bugs, without needing to replace the CPU hardware.

Microcode 更新可以通过 BIOS 来进行，也可以通过操作系统来进行。BIOS 更新的好处是一更新，上面安装的操作系统（Linux，Windows）都可以收益；操作系统更新比如 Windows 一般通过 Windows Update 来更新 Microcode。

x86 instructions are internally converted into simpler RISC-style micro-operations that are specific to a particular processor and stepping level. RISC computers do not use microcode, which is the reason why RISC compilers generate more instructions than CISC compilers.

直接更新 ROM 里的微码映射，和每次重启把更新写在 SRAM 里的区别是什么呢？

ROM 里的是不可更改的，所以不存在更新映射这一说，只可以通过 SRAM 来打补丁。

Microcode是什么？它为什么能修正CPU硬件错误？ - 知乎

微操作 micro-operations (also known as micro-ops or μops) 是处理器前端和后端的分界线，处理器前端会把指令解码成微操作并放入微操作队列，后端从队列中取出微操作并执行。

terminology - What exactly is microcode and how does it differ from firmware? - Super User

Micro-operations are detailed low-level instructions used in some designs to implement complex machine instructions. Each micro-op would execute in a single clock cycle. As with traditional microcode, one or more table lookups (or equivalent) is done to locate the appropriate μop-sequence based on the encoding and semantics of the machine instruction. Execution engine (Backend) inside the cpu can only execute so-called “micro-ops” (uops), that were provided by the Frontend.

A microinstruction contains all control information required to operate all involved functional units for one clock cycle. We refer to a plurality of microinstructions as microcode. 微指令是指在机器的一个 CPU 周期中，一组实现一定操作功能的微命令的组合。以上都能印证，微指令是对应一个时钟周期的。

微操作：完成一条机器指令是通过完成一个微操作序列实现的，这些微操作是计算机中最基本、不可再分解的操作。比如：取指周期中，我们需要将程序计数器 PC 的内容送入到 MAR 中，CU 向存储器发出读信号: 1->R,再比如执行周期中清空 ACC 的操作：0->ACC,这些都是微操作。

微命令：微命令是微操作的控制信号，每一个微操作都需要再对应的微命令控制下完成。所以微命令和微操作是一一对应的。

微指令：微指令是若干微命令的集合。一条微指令通常至少包含两大部分信息：

操作控制字段: 该字段显示了这条微指令是由那些微命令构成的。
顺序控制字段：用于控制产生下一条将要执行的微指令的地址。

微程序（microcode）：微程序是微指令的有序集合，一条机器指令的功能由一段微程序实现。

所以，整个故事开始浮现…，对于一条指令，CPU 将其翻译为一个 microcode，microcode 由一些微指令组成，每一个微指令又由一些微操作组成。

Difference between #AC and #DB and the handling of them

25. Bus lock detection and handling — The Linux Kernel documentation

Relationship between TLB and MMU

Note that usually the TLB is considered to be part of the MMU. The Wikipedia article on MMU shows a figure from an old version of a book that indicates that they are separate. However, the most recent version of the book has removed the figure and says that the TLB is part of the MMU.

x86 - Does a hyper-threaded core share MMU and TLB? - Stack Overflow

`CONFIG_HZ`

The frequency of interrupt.

x86 supports 100 Hz, 250 Hz and 1000 Hz.

Linux中的tick模式 - 知乎

Linux中的tick模式 - L

How does hardware be configured by this values?

For periodic mode, software sets a "initial count" and the local APIC uses it for a "current count". The local APIC decrements the current count until it reaches zero, then generates a timer IRQ and resets the current count to the initial count and begins decrementing the current count again. In this way the local APIC generates IRQs at a fixed rate depending on the initial count. The current count is decremented at a rate that depends on the CPU's external frequency ("bus frequency") divided by the value in the local APIC's "Divide Configuration Register".

For example, for a 2.4 GHz CPU with an external/bus frequency of 800 MHz, if the Divide Configuration Register is set to "divide by 4" and the initial count is set to 123456; then the local APIC timer would decrement the count at a rate of 200 MHz and generate a timer IRQ every 617.28 us, giving a rate of IRQs of 1620.01 Hz.

APIC timer - OSDev Wiki

笔记本合盖子不休眠

https://man7.org/linux/man-pages/man5/logind.conf.5.html

笔记本安装ubuntu关闭盖子不休眠设置_icodekang的博客-CSDN博客_ubuntu关闭盖子不挂起

Bus from southbridge to CPU

其中南桥跟 CPU 之间的通信 DMI 总线（Intel platform, on AMD, it is just PCIe），也就是 PCIe 3.0 x4 改了个名字。

On Intel platforms, all southbridge features and remaining I/O functions are managed by the PCH which is directly connected to the CPU via the Direct Media Interface (DMI).

DMI shares many characteristics with PCI Express, using multiple lanes and differential signaling to form a point-to-point link. Most implementations use a ×4 link.

什么是“可信“ (trusted, Not equals to trustworthy)

1990 年，国际标准化组织与国际电子技术委员会 ISO/IEC 在其发布的目录服务系列标准中基于行为预期性定义了可信性：如果第 2 个实体完全按照第 1 个实体的预期行动时，则第 1 个实体认为第 2 个实体是可信的。

1999 年，国际标准化组织与国际电子技术委员会在 ISO/IEC15408 标准中定义可信为：参与计算的组件、操作或过程在任意的条件下是可预测的，并能够抵御病毒和一定程度的物理干扰。

2002 年，TCG 用实体行为的预期性来定义可信：一个实体是可信的，如果它的行为总是以预期的方式，朝着预期的目标。这一定义的优点是抓住了实体的行为特征，符合哲学上实践是检验真理的唯一标准的基本原则。

IEEE 可信计算技术委员会认为，可信是指计算机系统所提供的服务是可信赖的，而且这种可信赖是可论证的。

我国沈昌祥院士认为，可信计算系统是能够提供系统的可靠性、可用性、信息和行为安全性的计算机系统。系统的可靠性和安全性是现阶段可信计算最主要的两个属性。因此，可信可简单表述为可信≈可靠 + 安全。

可信计算概述 - 知乎

QEMU trival patches

for small patches only touch a few lines of code.

Trivial Patches — QEMU documentation

The meaning of `vim.schedule`

https://www.youtube.com/watch?v=GMS0JvS7W1Y

Wrapper: failed with: gRPC call failed: Connection reset by peer

setup proxy to none.

Sed

-s: it only matter if you give sed multiple files.

Clock

Look inside qemu's definition of clock types:

/**
 * QEMUClockType:
 *
 * The following clock types are available:
 *
 * @QEMU_CLOCK_REALTIME: Real time clock
 *
 * The real time clock should be used only for stuff which does not
 * change the virtual machine state, as it runs even if the virtual
 * machine is stopped.
 *
 * @QEMU_CLOCK_VIRTUAL: virtual clock
 *
 * The virtual clock only runs during the emulation. It stops
 * when the virtual machine is stopped.
 *
 * @QEMU_CLOCK_HOST: host clock
 *
 * The host clock should be used for device models that emulate accurate
 * real time sources. It will continue to run when the virtual machine
 * is suspended, and it will reflect system time changes the host may
 * undergo (e.g. due to NTP).
 *
 * @QEMU_CLOCK_VIRTUAL_RT: realtime clock used for icount warp
 *
 * Outside icount mode, this clock is the same as @QEMU_CLOCK_VIRTUAL.
 * In icount mode, this clock counts nanoseconds while the virtual
 * machine is running.  It is used to increase @QEMU_CLOCK_VIRTUAL
 * while the CPUs are sleeping and thus not executing instructions.
 */

typedef enum {
    QEMU_CLOCK_REALTIME = 0,
    QEMU_CLOCK_VIRTUAL = 1,
    QEMU_CLOCK_HOST = 2,
    QEMU_CLOCK_VIRTUAL_RT = 3,
    QEMU_CLOCK_MAX
} QEMUClockType;

Clock generator/Electronic oscillator

MCUs have at least one clock to keep timing going at a regular pace, much like you would use a metronome to keep pace in piano playing.

Here is a good explanation of clock:

CPU Clocks and Clock Interrupts, and Their Effects on Schedulers

Clock interrupt

Aka. timer interrupt.

The clock interrupt is a regular interrupt, often every millisecond, which is used by the Operating System for things like sharing the CPU. All sorts of internal operations are timed by this clock - timeouts waiting for events, movements on the screen, and so on.

The CPU clock and the clock interrupt are probably derived, distantly, from the same crystal.

Processor, Die, Core, Thread (logical core)

The instruction cache TLB may be duplicated or shared in each logical processor, depending on implementation specifics of different processor families.

MMU? per thread or per core?

L2 maybe shared by all cores such as the e-cores in Alder Lake.

TLB is in logical core?

SDM 28.4.1:

… identifies two kinds of translation-related information that may be cached by a logical processor: translations, which are mappings from linear page numbers to physical page frames, and pagingstructure caches, …

L1 Cache sharing strategy in the context of Hyperthreading

Shared memory in Hyperthreading - Intel Communities

Code debugging

Kernel debugging

First, make kernel config allow debugging (The navi command in my XDG config):

CONFIG_DEBUG_INFO=y
CONFIG_DEBUG_INFO_DWARF5=y
CONFIG_GDB_SCRIPTS=y

or gdb will complaint that there is no symbol when you start it using gdb vmlinux.

Second (The navi command in my XDG config)

# make an initrd (one time)
mkinitramfs -o ramdisk.img

# without kernel modules
sudo qemu-system-x86_64 -s -S -m 16G -smp 16 -chardev stdio,id=gdb0 -device isa-debugcon,iobase=0x402,chardev=gdb0,id=d1 -enable-kvm -vnc :4 -net nic,model=virtio -net user,hostfwd=tcp::8022-:22 -kernel arch/x86_64/boot/bzImage -append "nokaslr" -initrd ~/ramdisk.img

# start a gdb
gdb vmlinux -ex "target remote :1234" -ex "hbreak start_kernel" -ex "c"

Third, start debug adaptor in Neovim.

Kernel debugging with kernel modules (e.g., KVM)

In this case, you need a virtual machine because you need to install the kernel module.

# with kernel modules
sudo qemu-system-x86_64 -s -S -m 16G -smp 16 -chardev stdio,id=gdb0 -device isa-debugcon,iobase=0x402,chardev=gdb0,id=d1 -enable-kvm ~/lei.img -vnc :4 -net nic,model=virtio -net user,hostfwd=tcp::8022-:22

# start a gdb
gdb vmlinux -ex "target remote :1234" -ex "hbreak start_kernel" -ex "c"

Qemu debugging

gdb --args qemu-system-x86_64 --m 16G -smp 16 -enable-kvm /home/lei/p/lei.img -vnc :4 -net nic,model=virtio -net user,hostfwd=tcp::8022-:22

Linux Process/Thread model

在 Linux 内核中，其实是没有线程的概念的，它把所有的线程当做标准的进程来实现。

对于 Linux 而言，线程只是进程间共享资源的一种方式，非常轻量。举个简单例子，假设有一个进程包含了 N 个线程。对于那些显示支持线程的操作系统而言，应该是存在一个进程描述符，依次轮流指向 N 个线程。进程描述符指明共享资源，包括内存空间和打开的文件，然后线程描述它们自己独享的资源。相反的是在 Linux 中，只有 N 个进程，因此有 N 个 task_struct 数据结构，只是这些数据结构的某些资源项是共享的。

在 Linux 中，还存在一个 Kernel Thread 的概念，也就是内核线程。内核创建一些内核线程来执行一些后台任务。相对于普通的进程，内核线程完整的存在于内核空间，是没有自己的地址空间的，也就是 mm 指针为空，它的操作仅存在于内核态，并且也不会上下文切换到用户态。不过内核线程和普通进程类似的是，是可调度和可抢占的。

任何一个进程，如果只有主线程，那 pid 是自己，tgid (thread group) 是自己，pgid (process group) 指向的还是自己。

The PGID is the same as the PID of the first process.

Thread groups were a feature added in Linux 2.4 to support the POSIX threads notion of a set of threads that share a single PID. Internally, this shared PID is the so-called thread group identifier (TGID) for the thread group. Since Linux 2.4, calls to getpid(2) return the TGID of the caller.¹

the separation of process ID and thread ID make the Linux behaviour look more like other systems and more like the POSIX requirements in this sense. Though technically the OS still doesn't have separate implementations for threads and processes.[^2]

但是，如果一个进程创建了其他线程，那就会有所变化了。线程有自己的 pid，tgid 就是进程的主线程的 pid，group_leader 指向的就是进程的主线程。

有了 tgid 之后，我们就可以判断一个 task 是线程还是进程了。

在 Linux 上，进程和线程的相同点要远远大于不同点。在 Linux 下的线程甚至都被称为了轻量级进程。

对于线程来说，我们假如一个进程下创建了多个线程出来。那么每个线程的 pid 都是不同的。但是我们一般又需要记录线程是属于哪个进程的。这时候，tgid 就派上用场了，通过 tgid 字段来表示自己所归属的进程 ID。

事实上，进程线程创建的时候，使用的函数看起来不一样。但实际在底层实现上，最终都是使用同一个函数来实现的。

内存虚拟地址空间是最重要的。因此区分一个 Task 任务该叫线程还是该叫进程，一般习惯上就看它是否有独立的地址空间。

这里展开多说一句，对于内核任务来说，无论有多少个任务，其使用地址空间都是同一个。所以一般都叫内核线程，而不是内核进程。

浅谈 Linux 线程模型 - 知乎

聊聊Linux中线程和进程的联系与区别！

Typical C headers

unistd.h

Unix standard: On Unix-like systems, the interface defined by unistd.h is typically made up largely of system call wrapper functions such as fork, pipe and I/O primitives (read, write, close, etc.).

Are threads implemented as processes on Linux? - Unix & Linux Stack Exchange ↩

Misc ideas

iTLB / dTLB / TLB Hierarchy

arch/x86/include

include Directory in the kernel source code

When VCPU is in low overhead, can Host use the CPU resource to do other things?

How do we kick a vcpu down if it just don't VMexit?

What does "Idle" mean for CPU? / HLT instruction / Halt instruction

TPM

Style customization for Thunderbird

Attestation Service

Intel vs. AMD in Features

Where to find the spec?

ELF symbols

CPUID is x86 related rather than vendor (Intel, AMD) related

Authenticated Code Module (ACM)

What is on-chip memory (CRAM)?

Wrapper: failed with: gRPC call failed: Connection reset by peer

5800X3D/3D V-Cache

How does shell work when you executing commands?

Create a new user / Add a new user (tested on Ubuntu, CentOS)

Python LSP server parameters

Difference between (1ULL << 2) and 0x00000004

Tabstop, shiftwidth, softtabstop, expandtab, smarttab

How to fix "You’ll Need a New App to Open This ms-gaming overlay"

The Shadow Suite

How to see which version of softwares Ubuntu is using

Linux PAM

Linux utils like /usr/bin/login

Usermode-Helper API

Shell type

Linux init 进程

Idle process

Change mermaid in typora to dark theme

What is gdb MI?

Not only commits are recognized by Github as contribution to show up in the graph.

CONFIG_=n And # CONFIG_ is not set

Uncore

How to examine an object file?

AT&T Assembly syntax

Assembler Directives

Kernel assembly style

Kernel Address Space Layout Randomization (or KASLR)

Exploit (security)

Kick?

Microcode (微码映射, 微程序), microops (micro-operations, uops, $\mu$ops, μops, 微操作), microinstructions (微指令)

Difference between #AC and #DB and the handling of them

Relationship between TLB and MMU

CONFIG_HZ

笔记本合盖子不休眠

Bus from southbridge to CPU

什么是“可信“ (trusted, Not equals to trustworthy)

QEMU trival patches

The meaning of vim.schedule

Wrapper: failed with: gRPC call failed: Connection reset by peer

Sed

Clock

Clock generator/Electronic oscillator

Clock interrupt

Processor, Die, Core, Thread (logical core)

Code debugging

Kernel debugging

Kernel debugging with kernel modules (e.g., KVM)

Qemu debugging

Linux Process/Thread model

Typical C headers

unistd.h

`arch/x86/include`

`include` Directory in the kernel source code

Difference between `(1ULL << 2)` and `0x00000004`

`CONFIG_HZ`

The meaning of `vim.schedule`