aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorMitchell Horne <mhorne@FreeBSD.org>2023-08-03 13:48:15 +0000
committerMitchell Horne <mhorne@FreeBSD.org>2023-09-11 22:42:15 +0000
commit5a0c410787b8b2f92a0c3f6126c95ffe34de344c (patch)
tree37a9b63ee89052ab6727b6862631bfde283464c9
parentaa26195de58d6dc8e3d989ec50f1c5934c3f0e22 (diff)
downloadsrc-5a0c410787b8.tar.gz
src-5a0c410787b8.zip
intro(9): rewrite from scratch
This page has existed as a placeholder since its creation in 1995. It does not provide a useful introduction to the content in this section. Reimagine it as a top-level overview page containing brief descriptions and links to existing pages in section 9. It is roughly organized into sub-sections, grouped by topic or subsystem. In other words, the page is meant to function as a map to other content. There is a balance to be found here between providing as many links as possible and keeping the page concise and searchable. In general the aim is to reference pages which provide the best entry point to a particular topic. For example, a link is given to locking(9), but not to the specific lock pages such as mutex(9) or rwlock(9). NetBSD has done something similar with their intro(9), so some inspiration has been taken from there, although their content doesn't align that closely with what we have. I have done a thorough review of our existing pages and formed these subsections around them, but they are meant to evolve. PR: 270481 Reviewed by: imp, emaste MFC after: 3 weeks Relnotes: yes Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D41104 (cherry picked from commit 84f9f2c5cf7841fffc03ccb1833814892ae15132)
-rw-r--r--share/man/man9/intro.9594
1 files changed, 505 insertions, 89 deletions
diff --git a/share/man/man9/intro.9 b/share/man/man9/intro.9
index 4777ee2b8504..458ee0ab4697 100644
--- a/share/man/man9/intro.9
+++ b/share/man/man9/intro.9
@@ -1,103 +1,519 @@
-.\" Copyright (c) 1983, 1991, 1993
-.\" The Regents of the University of California. All rights reserved.
.\"
-.\" Redistribution and use in source and binary forms, with or without
-.\" modification, are permitted provided that the following conditions
-.\" are met:
-.\" 1. Redistributions of source code must retain the above copyright
-.\" notice, this list of conditions and the following disclaimer.
-.\" 2. Redistributions in binary form must reproduce the above copyright
-.\" notice, this list of conditions and the following disclaimer in the
-.\" documentation and/or other materials provided with the distribution.
-.\" 3. Neither the name of the University nor the names of its contributors
-.\" may be used to endorse or promote products derived from this software
-.\" without specific prior written permission.
+.\" SPDX-License-Identifier: BSD-2-Clause
.\"
-.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
-.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
-.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
-.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
-.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
-.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
-.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
-.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
-.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
-.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
-.\" SUCH DAMAGE.
+.\" Copyright (c) 2023 The FreeBSD Foundation
.\"
-.Dd December 13, 1995
+.\" This manual page was written by Mitchell Horne <mhorne@FreeBSD.org> under
+.\" sponsorship from the FreeBSD Foundation.
+.\"
+.Dd August 2, 2023
.Dt INTRO 9
.Os
.Sh NAME
.Nm intro
-.Nd "introduction to system kernel interfaces"
+.Nd "introduction to kernel programming interfaces"
.Sh DESCRIPTION
-This section contains information about the interfaces and
-subroutines in the kernel.
-.Sh PROTOTYPES ANSI-C AND ALL THAT
-Yes please.
-.Pp
-We would like all code to be fully prototyped.
-.Pp
-If your code compiles cleanly with
-.Nm cc
-.Ar -Wall
-we would feel happy about it.
-It is important to understand that this is not a question of just shutting up
-.Nm cc ,
-it is a question about avoiding the things it complains about.
-To put it bluntly, do not hide the problem by casting and other
-obfuscating practices, solve the problem.
-.Sh INDENTATION AND STYLE
-Believe it or not, there actually exists a guide for indentation and style.
-It is not generally applied though.
-.Pp
-We would appreciate if people would pay attention to it, and at least not
-violate it blatantly.
-.Pp
-We do not mind it too badly if you have your own style, but please make
-sure we can read it too.
-.Pp
-Please take time to read
+Welcome to the
+.Fx
+kernel documentation.
+Outside of the source code itself, this set of
+.Xr man 1
+pages is the primary resource for information on usage of the numerous
+programming interfaces available within the kernel.
+In some cases, it is also a source of truth for the implementation details
+and/or design decisions behind a particular subsystem or piece of code.
+.Pp
+The intended audience of this documentation is developers, and the primary
+authors are also developers.
+It is written assuming a certain familiarity with common programming or
+OS-level concepts and practices.
+However, this documentation should also attempt to provide enough background
+information that readers approaching a particular subsystem or interface for
+the first time will be able to understand.
+.Pp
+To further set expectations, we acknowledge that kernel documentation, like the
+source code itself, is forever a work-in-progress.
+There will be large sections of the codebase whose documentation is subtly or
+severely outdated, or missing altogether.
+This documentation is a supplement to the source code, and can not always be
+taken at face value.
+.Pp
+At its best, section 9 documentation will provide a description of a particular
+piece of code that, paired with its implementation, fully informs the reader of
+the intended and realized effects.
+.Pp
+.Xr man 1
+pages in this section most frequently describe functions, but may also
+describe types, global variables, macros, or high-level concepts.
+.Sh CODING GUIDELINES
+Code written for the
+.Fx
+kernel is expected to conform to the established style and coding conventions.
+Please see
.Xr style 9
-for more information.
-.Sh NAMING THINGS
-Some general rules exist:
-.Bl -enum
-.It
-If a function is meant as a debugging aid in DDB, it should be enclosed
-in
-.Bd -literal -offset indent
-#ifdef DDB
-
-#endif /* DDB */
+for a detailed set of rules and guidelines.
+.Sh OVERVIEW
+Below is presented various subsystems.
+.Ss Data Structures
+There are implementations for many well-known data structures available in the
+kernel.
+.Bl -tag -width "Xr bitstring 3"
+.It Xr bitstring 3
+Simple bitmap implementation.
+.It Xr counter 9
+An SMP-safe general-purpose counter implementation.
+.It Xr hash 9
+Hash map implementation.
+.It Xr nv 9
+Name/value pairs.
+.It Xr queue 3
+Singly-linked and doubly-linked lists, and queues.
+.It Xr refcount 9
+An SMP-safe implementation of reference counts.
+.It Xr sbuf 9
+Dynamic string composition.
+.It Xr sglist 9
+A scatter/gather list implementation.
+.El
+.Ss Utility Functions
+Functions or facilities of general usefulness or convenience.
+See also the
+.Sx Testing and Debugging Tools
+or
+.Sx Miscellaneous
+sub-sections below.
+.Pp
+Formatted output and logging functions are described by
+.Xr printf 9 .
+.Pp
+Endian-swapping functions:
+.Xr byteorder 9 .
+.Pp
+Data output in hexadecimal format:
+.Xr hexdump 9 .
+.Pp
+A rich set of macros for declaring
+.Xr sysctl 8
+variables and functions is described by
+.Xr sysctl 9 .
+.Pp
+Non-recoverable errors in the kernel should trigger a
+.Xr panic 9 .
+Run-time assertions can be verified using the
+.Xr KASSERT 9
+macros.
+Compile-time assertions should use
+.Fn _Static_assert .
+.Pp
+The SYSINIT framework provides macros for declaring functions that will be
+executed during start-up and shutdown; see
+.Xr SYSINIT 9 .
+.Pp
+Deprecation messages may be emitted with
+.Xr gone_in 9 .
+.Pp
+A unit number facility is provided by
+.Xr unr 9 .
+.Ss Synchronization Primitives
+The
+.Xr locking 9
+man page gives an overview of the various types of locks available in the
+kernel and advice on their usage.
+.Pp
+Atomic primitives are described by
+.Xr atomic 9 .
+.Pp
+The
+.Xr epoch 9
+and
+.Xr smr 9
+facilities are used to create lock-free data structures.
+There is also
+.Xr seqc 9 .
+.Ss Memory Management
+Dynamic memory allocations inside the kernel are generally done using
+.Xr malloc 9 .
+Frequently allocated objects may prefer to use
+.Xr uma 9 .
+.Pp
+.\" MHTODO: It would be useful to have a vm_page(9) or similar
+.\" high-level page which points to the following contents instead.
+Much of the virtual memory system operates on
+.Vt vm_page_t
+structures.
+The following functions are documented:
+.Bd -ragged -offset indent
+.Xr vm_page_advise 9 ,
+.Xr vm_page_alloc 9 ,
+.Xr vm_page_bits 9 ,
+.Xr vm_page_aflag 9 ,
+.Xr vm_page_alloc 9 ,
+.Xr vm_page_bits 9 ,
+.Xr vm_page_busy 9 ,
+.Xr vm_page_deactivate 9 ,
+.Xr vm_page_free 9 ,
+.Xr vm_page_grab 9 ,
+.Xr vm_page_insert 9 ,
+.Xr vm_page_lookup 9 ,
+.Xr vm_page_rename 9 ,
+.Xr vm_page_sbusy 9 ,
+.Xr vm_page_wire 9
+.Ed
+.Pp
+Virtual address space maps are managed with the
+.Xr vm_map 9
+API.
+.Pp
+The machine-dependent portion of the virtual memory stack is the
+.Xr pmap 9
+module.
+.Pp
+Allocation policies for NUMA memory domains are managed with the
+.Xr domainset 9
+API.
+.Ss File Systems
+The kernel interface for file systems is
+.Xr VFS 9 .
+File system implementations register themselves with
+.Xr vfsconf 9 .
+.Pp
+The abstract and filesystem-independent representation of a file, directory, or
+other file-like entity within the kernel is the
+.Xr vnode 9 .
+.Pp
+The implementation of access control lists for filesystems is described by
+.Xr acl 9 .
+Also
+.Xr vaccess 9 .
+.Ss I/O and Storage
+.\" TODO: This page needs to be rewritten before it can be included here.
+.\" The buffer cache is described by:
+.\" .Xr buf 9
+.\" .Pp
+The GEOM framework represents I/O requests using the
+.Xr bio 9
+structure.
+.Pp
+Disk drivers connect themselves to GEOM using the
+.Xr disk 9
+API.
+.Pp
+The
+.Xr devstat 9
+facility provides an interface for recording device statistics in disk drivers.
+.Ss Networking
+Much of the networking stack uses the
+.Xr mbuf 9 ,
+a flexible memory management unit commonly used to store network packets.
+.Pp
+Network interfaces are implemented using the
+.Xr ifnet 9
+API, which has functions for drivers and consumers.
+.Pp
+A framework for managing packet output queues is described by
+.Xr altq 9 .
+.Pp
+To receive incoming packets, network protocols register themselves with
+.Xr netisr 9 .
+.Pp
+Virtualization of the network stack is provided by
+.Xr VNET 9 .
+.Pp
+The front-end for interfacing with network sockets from within the kernel is
+described by
+.Xr socket 9 .
+The back-end interface for socket implementations is
+.Xr domain 9 .
+.Pp
+The low-level packet filter interface is described by
+.Xr pfil 9 .
+.Pp
+The
+.Xr bpf 9
+interface provides a mechanism to redirect packets to userspace.
+.Pp
+The subsystem for IEEE 802.11 wireless networking is described by
+.Xr ieee80211 9 .
+.Pp
+A framework for modular TCP implementations is described by
+.Xr tcp_functions 9 .
+.Pp
+A framework for modular congestion control algorithms is described by
+.Xr mod_cc 9 .
+.Ss Device Drivers
+.\" TODO: a bus(9) or newbus(9) page, as well as updates to existing pages
+.\" would be helpful in laying out the high-level concepts of FreeBSD's device
+.\" structure, and explaining the organization of existing documentation.
+Consult the
+.Xr device 9
+and
+.Xr driver 9
+pages first.
+.Pp
+Most drivers act as devices, and provide a set of methods implementing the
+device interface.
+This includes methods such as
+.Xr DEVICE_PROBE 9 ,
+.Xr DEVICE_ATTACH 9 ,
+and
+.Xr DEVICE_DETACH 9 .
+.Pp
+In addition to devices, there are buses.
+Buses may have children, in the form of devices or other buses.
+Bus drivers will implement additional methods, such as
+.Xr BUS_ADD_CHILD 9 ,
+.Xr BUS_READ_IVAR 9 ,
+or
+.Xr BUS_RESCAN 9 .
+.Pp
+Buses often perform resource accounting on behalf of their children.
+For this there is the
+.Xr rman 9
+API.
+.Pp
+Drivers can request and manage their resources (e.g. memory-space or IRQ
+number) from their parent using the following sets of functions:
+.Bd -ragged -offset indent
+.Xr bus_alloc_resource 9 ,
+.Xr bus_adjust_resource 9 ,
+.Xr bus_get_resource 9 ,
+.Xr bus_map_resource 9 ,
+.Xr bus_release_resource 9 ,
+.Xr bus_set_resource 9
.Ed
.Pp
-And the name of the procedure should start with the prefix
-.Li DDB_
-to clearly identify the procedure as a debugger routine.
+Direct Memory Access (DMA) is handled using the
+.Xr busdma 9
+framework.
+.Pp
+Functions for accessing bus space (i.e. read/write) are provided by
+.Xr bus_space 9 .
+.Ss Clocks and Timekeeping
+The kernel clock frequency and overall system time model is described by
+.Xr hz 9 .
+.Pp
+A few global time variables, such as system up-time, are described by
+.Xr time 9 .
+.Pp
+Raw CPU cycles are provided by
+.Xr get_cyclecount 9 .
+.Ss Userspace Memory Access
+Direct read/write access of userspace memory from the kernel is not permitted,
+and memory transactions that cross the kernel/user boundary must go through one
+of several interfaces built for this task.
+.Pp
+Most device drivers use the
+.Xr uiomove 9
+set of routines.
+.Pp
+Simpler primitives for reading or writing smaller chunks of memory are
+described by
+.Xr casuword 9 ,
+.Xr copy 9 ,
+.Xr fetch 9 ,
+and
+.Xr store 9 .
+.Ss Kernel Threads, Tasks, and Callbacks
+Kernel threads and processes are created using the
+.Xr kthread 9
+and
+.Xr kproc 9
+interfaces, respectively.
+.Pp
+Where dedicated kernel threads are too heavyweight, there is also the
+.Xr taskqueue 9
+interface.
+.Pp
+For low-latency callback handling, the
+.Xr callout 9
+framework should be used.
+.Pp
+Dynamic handlers for pre-defined event hooks are registered and invoked using
+the
+.Xr EVENTHANDLER 9
+API.
+.Ss Thread Switching and Scheduling
+The machine-independent interface to a context switch is
+.Xr mi_switch 9 .
+.Pp
+To prevent preemption, use a
+.Xr critical 9
+section.
+.Pp
+To voluntarily yield the processor, use
+.Xr kern_yield 9 .
+.Pp
+The various functions which will deliberately put a thread to sleep are
+described by
+.Xr sleep 9 .
+Sleeping threads are removed from the scheduler and placed on a
+.Xr sleepqueue 9 .
+.\" TODO: This page is outdated and can't be included here yet.
+.\".Pp
+.\"The thread scheduler interface is described by
+.\".Xr scheduler 9 .
+.Ss Processes and Signals
+To locate a process or process group by its identifier, use
+.Xr pfind 9
+and
+.Xr pgfind 9 .
+Alternatively, the
+.Xr pget 9
+function provides additional search specificity.
+.Pp
+The "hold count" of a process can be manipulated with
+.Xr PHOLD 9 .
+.Pp
+The kernel interface for signals is described by
+.Xr signal 9 .
+.Pp
+Signals can be sent to processes or process groups using the functions
+described by
+.Xr psignal 9 .
+.Ss Security
+See the generic security overview in
+.Xr security 7 .
+.Pp
+The basic structure for user credentials is
+.Vt struct ucred ,
+managed by the
+.Xr ucred 9
+API.
+Thread credentials are verified using
+.Xr priv 9
+to allow or deny certain privileged actions.
+.Pp
+Policies influenced by
+.Va kern.securelevel
+must use the
+.Xr securelevel_gt 9
+or
+.Xr securelevel_ge 9
+functions.
+.Pp
+The Mandatory Access Control (MAC) framework provides a wide set of hooks,
+supporting dynamically-registered security modules;
+see
+.Xr mac 9 .
+.Pp
+Cryptographic services are provided by the OpenCrypto framework.
+This API provides and interface for both consumers and crypto drivers;
+see
+.Xr crypto 9 .
+.Pp
+For information on random number generation, see
+.Xr random 9
+and
+.Xr prng 9 .
+.Ss Kernel Modules
+The interfaces for declaring loadable kernel modules are described by
+.Xr module 9 .
+.Ss Interrupts
+The machine-independent portion of the interrupt framework supporting the
+registration and execution of interrupt handlers is described by
+.Xr intr_event 9 .
+.Pp
+Software interrupts are provided by
+.Xr swi 9 .
+.Pp
+Device drivers register their interrupt handlers using the
+.Xr bus_setup_intr 9
+function.
+.Ss Testing and Debugging Tools
+A kernel test framework:
+.Xr kern_testfrwk 9
+.Pp
+A facility for defining configurable fail points is described by
+.Xr fail 9 .
+.Pp
+Commands for the
+.Xr ddb 4
+kernel debugger are defined with the
+.Xr DB_COMMAND 9
+family of macros.
+.Pp
+The
+.Xr ktr 4
+tracing facility adds static tracepoints to many areas of the kernel.
+These tracepoints are defined using the macros described by
+.Xr ktr 9 .
+.Pp
+Static probes for DTrace are defined using the
+.Xr SDT 9
+macros.
+.Pp
+Stack traces can be captured and printed with the
+.Xr stack 9
+API.
+.Pp
+Kernel sanitizers can perform additional compiler-assisted checks against
+memory use/access.
+These runtimes are capable of detecting difficult-to-identify classes of bugs,
+at the cost of a large overhead.
+Supported is the Kernel Address Sanitizer
+.Xr KASAN 9 ,
+and the Kernel Memory Sanitizer
+.Xr KMSAN 9 .
+.Pp
+The
+.Cd LOCK_PROFILING
+kernel config option enables extra code to assist with profiling and/or
+debugging lock performance;
+see
+.Xr LOCK_PROFILING 9 .
+.Ss Driver Tools
+Defined functions/APIs for specific types of devices.
+.Bl -tag -width "Xr usbdi 9"
+.It Xr iflib 9
+Programming interface for
+.Xr iflib 4
+based network drivers.
+.It Xr pci 9
+Peripheral Component Interconnect (PCI) and PCI Express (PCIe) programming API.
+.It Xr pwmbus 9
+Pulse-Width Modulation (PWM) bus interface methods.
+.It Xr usbdi 9
+Universal Serial Bus programming interface.
+.It Xr superio 9
+Functions for Super I/O controller devices.
.El
-.Sh SCOPE OF SYMBOLS
-It is important to carefully consider the scope of symbols in the kernel.
-The default is to make everything static, unless some reason requires
-the opposite.
-.Pp
-There are several reasons for this policy,
-the main one is that the kernel is one monolithic name-space,
-and pollution is not a good idea here either.
-.Pp
-For device drivers and other modules that do not add new internal interfaces
-to the kernel, the entire source should be in one file if possible.
-That way all symbols can be made static.
-.Pp
-If for some reason a module is split over multiple source files, then try
-to split the module along some major fault-line and consider using the
-number of global symbols as your guide.
-The fewer the better.
+.Ss Miscellaneous
+Dynamic per-CPU variables:
+.Xr dpcpu 9 .
+.Pp
+CPU bitmap management:
+.Xr cpuset 9 .
+.Pp
+Kernel environment management:
+.Xr getenv 9 .
+.Pp
+Contexts for CPU floating-point registers are managed by the
+.Xr fpu_kern 9
+facility.
+.Pp
+For details on the shutdown/reboot procedure and available shutdown hooks, see
+.Xr reboot 9 .
+.Pp
+A facility for asynchronous logging to files from within the kernel is provided
+by
+.Xr alq 9 .
+.Pp
+The
+.Xr osd 9
+framework provides a mechanism to dynamically extend core structures in a way
+that preserves KBI.
+See the
+.Xr hhook 9
+and
+.Xr khelp 9
+APIs for information on how this is used.
+.Pp
+The kernel object implementation is described by
+.Xr kobj 9 .
.Sh SEE ALSO
+.Xr man 1 ,
.Xr style 9
-.Sh HISTORY
-The
-.Nm
-section manual page appeared in
-.Fx 2.2 .
+.Rs
+.%T "The FreeBSD Architecture Handbook"
+.%U "https://docs.freebsd.org/en/books/arch-handbook/"
+.Re