# SOME DESCRIPTIVE TITLE # Copyright (C) YEAR The FreeBSD Project # This file is distributed under the same license as the FreeBSD Documentation package. # FIRST AUTHOR , YEAR. # #, fuzzy msgid "" msgstr "" "Project-Id-Version: FreeBSD Documentation VERSION\n" "POT-Creation-Date: 2025-11-08 16:17+0000\n" "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n" "Last-Translator: FULL NAME \n" "Language-Team: LANGUAGE \n" "Language: \n" "MIME-Version: 1.0\n" "Content-Type: text/plain; charset=UTF-8\n" "Content-Transfer-Encoding: 8bit\n" #. type: YAML Front Matter: description #: documentation/content/en/articles/geom-class/_index.adoc:1 #, no-wrap msgid "A guide to GEOM internals, and writing your own GEOM class" msgstr "" #. type: Title = #: documentation/content/en/articles/geom-class/_index.adoc:1 #: documentation/content/en/articles/geom-class/_index.adoc:11 #, no-wrap msgid "Writing a GEOM Class" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:44 msgid "Abstract" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:47 msgid "" "This text documents some starting points in developing GEOM classes, and " "kernel modules in general. It is assumed that the reader is familiar with C " "userland programming." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:49 msgid "'''" msgstr "" #. type: Title == #: documentation/content/en/articles/geom-class/_index.adoc:53 #, no-wrap msgid "Introduction" msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:56 #, no-wrap msgid "Documentation" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:60 msgid "" "Documentation on kernel programming is scarce - it is one of few areas where " "there is nearly nothing in the way of friendly tutorials, and the phrase " "\"use the source!\" really holds true. However, there are some bits and " "pieces (some of them seriously outdated) floating around that should be " "studied before beginning to code:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:62 msgid "" "The extref:{developers-handbook}[FreeBSD Developer's Handbook] - part of the " "documentation project, it does not contain anything specific to kernel " "programming, but rather some general useful information." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:63 msgid "" "The extref:{arch-handbook}[FreeBSD Architecture Handbook] - also from the " "documentation project, contains descriptions of several low-level facilities " "and procedures. The most important chapter is 13, extref:{arch-handbook}" "[Writing FreeBSD device drivers, driverbasics]." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:64 msgid "" "The Blueprints section of http://www.freebsddiary.org[FreeBSD Diary] web " "site - contains several interesting articles on kernel facilities." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:65 msgid "" "The man pages in section 9 - for important documentation on kernel functions." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:66 msgid "" "The man:geom[4] man page and http://phk.freebsd.dk/pubs/[PHK's GEOM slides] " "- for general introduction of the GEOM subsystem." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:67 msgid "" "Man pages man:g_bio[9], man:g_event[9], man:g_data[9], man:g_geom[9], " "man:g_provider[9], man:g_consumer[9], man:g_access[9] & others linked from " "those, for documentation on specific functionalities." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:68 msgid "" "The man:style[9] man page - for documentation on the coding-style " "conventions which must be followed for any code which is to be committed to " "the FreeBSD tree." msgstr "" #. type: Title == #: documentation/content/en/articles/geom-class/_index.adoc:70 #, no-wrap msgid "Preliminaries" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:77 msgid "" "The best way to do kernel development is to have (at least) two separate " "computers. One of these would contain the development environment and " "sources, and the other would be used to test the newly written code by " "network-booting and network-mounting filesystems from the first one. This " "way if the new code contains bugs and crashes the machine, it will not mess " "up the sources (and other \"live\" data). The second system does not even " "require a proper display. Instead, it could be connected with a serial " "cable or KVM to the first one." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:80 msgid "" "But, since not everybody has two or more computers handy, there are a few " "things that can be done to prepare an otherwise \"live\" system for " "developing kernel code. This setup is also applicable for developing in a " "http://www.vmware.com/[VMWare] or http://www.qemu.org/[QEmu] virtual machine " "(the next best thing after a dedicated development machine)." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:82 #, no-wrap msgid "Modifying a System for Development" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:85 msgid "" "For any kernel programming a kernel with `INVARIANTS` enabled is a must-" "have. So enter these in your kernel configuration file:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:90 #, no-wrap msgid "" "options INVARIANT_SUPPORT\n" "options INVARIANTS\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:93 msgid "" "For more debugging you should also include WITNESS support, which will alert " "you of mistakes in locking:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:98 #, no-wrap msgid "" "options WITNESS_SUPPORT\n" "options WITNESS\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:101 msgid "For debugging crash dumps, a kernel with debug symbols is needed:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:105 #, no-wrap msgid " makeoptions DEBUG=-g\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:110 msgid "" "With the usual way of installing the kernel (`make installkernel`) the debug " "kernel will not be automatically installed. It is called " "[.filename]#kernel.debug# and located in [.filename]#/usr/obj/usr/src/sys/" "KERNELNAME/#. For convenience it should be copied to [.filename]#/boot/" "kernel/#." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:113 msgid "" "Another convenience is enabling the kernel debugger so you can examine a " "kernel panic when it happens. For this, enter the following lines in your " "kernel configuration file:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:119 #, no-wrap msgid "" "options KDB\n" "options DDB\n" "options KDB_TRACE\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:122 msgid "" "For this to work you might need to set a sysctl (if it is not on by default):" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:126 #, no-wrap msgid " debug.debugger_on_panic=1\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:134 msgid "" "Kernel panics will happen, so care should be taken with the filesystem " "cache. In particular, having softupdates might mean the latest file version " "could be lost if a panic occurs before it is committed to storage. " "Disabling softupdates yields a great performance hit, and still does not " "guarantee data consistency. Mounting filesystem with the \"sync\" option is " "needed for that. For a compromise, the softupdates cache delays can be " "shortened. There are three sysctl's that are useful for this (best to be " "set in [.filename]#/etc/sysctl.conf#):" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:140 #, no-wrap msgid "" "kern.filedelay=5\n" "kern.dirdelay=4\n" "kern.metadelay=3\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:143 msgid "The numbers represent seconds." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:151 msgid "" "For debugging kernel panics, kernel core dumps are required. Since a kernel " "panic might make filesystems unusable, this crash dump is first written to a " "raw partition. Usually, this is the swap partition. This partition must be " "at least as large as the physical RAM in the machine. On the next boot, the " "dump is copied to a regular file. This happens after filesystems are " "checked and mounted, and before swap is enabled. This is controlled with " "two [.filename]#/etc/rc.conf# variables:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:156 #, no-wrap msgid "" "dumpdev=\"/dev/ad0s4b\"\n" "dumpdir=\"/usr/core\"\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:159 msgid "" "The `dumpdev` variable specifies the swap partition and `dumpdir` tells the " "system where in the filesystem to relocate the core dump on reboot." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:162 msgid "" "Writing kernel core dumps is slow and takes a long time so if you have lots " "of memory (>256M) and lots of panics it could be frustrating to sit and wait " "while it is done (twice - first to write it to swap, then to relocate it to " "filesystem). It is convenient then to limit the amount of RAM the system " "will use via a [.filename]#/boot/loader.conf# tunable:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:166 #, no-wrap msgid " hw.physmem=\"256M\"\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:169 msgid "" "If the panics are frequent and filesystems large (or you simply do not trust " "softupdates+background fsck) it is advisable to turn background fsck off via " "[.filename]#/etc/rc.conf# variable:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:173 #, no-wrap msgid " background_fsck=\"NO\"\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:178 msgid "" "This way, the filesystems will always get checked when needed. Note that " "with background fsck, a new panic could happen while it is checking the " "disks. Again, the safest way is not to have many local filesystems by using " "another computer as an NFS server." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:180 #, no-wrap msgid "Starting the Project" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:184 msgid "" "For the purpose of creating a new GEOM class, an empty subdirectory has to " "be created under an arbitrary user-accessible directory. You do not have to " "create the module directory under [.filename]#/usr/src#." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:186 #, no-wrap msgid "The Makefile" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:189 msgid "" "It is good practice to create [.filename]#Makefiles# for every nontrivial " "coding project, which of course includes kernel modules." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:192 msgid "" "Creating the [.filename]#Makefile# is simple thanks to an extensive set of " "helper routines provided by the system. In short, here is how a minimal " "[.filename]#Makefile# looks for a kernel module:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:197 #, no-wrap msgid "" "SRCS=g_journal.c\n" "KMOD=geom_journal\n" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:199 #, no-wrap msgid ".include \n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:203 msgid "" "This [.filename]#Makefile# (with changed filenames) will do for any kernel " "module, and a GEOM class can reside in just one kernel module. If more than " "one file is required, list it in the `SRCS` variable, separated with " "whitespace from other filenames." msgstr "" #. type: Title == #: documentation/content/en/articles/geom-class/_index.adoc:205 #, no-wrap msgid "On FreeBSD Kernel Programming" msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:208 #, no-wrap msgid "Memory Allocation" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:213 msgid "" "See man:malloc[9]. Basic memory allocation is only slightly different than " "its userland equivalent. Most notably, `malloc`() and `free`() accept " "additional parameters as is described in the man page." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:215 msgid "" "A \"malloc type\" must be declared in the declaration section of a source " "file, like this:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:219 #, no-wrap msgid " static MALLOC_DEFINE(M_GJOURNAL, \"gjournal data\", \"GEOM_JOURNAL Data\");\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:222 msgid "" "To use this macro, [.filename]#sys/param.h#, [.filename]#sys/kernel.h# and " "[.filename]#sys/malloc.h# headers must be included." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:225 msgid "" "There is another mechanism for allocating memory, the UMA (Universal Memory " "Allocator). See man:uma[9] for details, but it is a special type of " "allocator mainly used for speedy allocation of lists comprised of same-sized " "items (for example, dynamic arrays of structs)." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:227 #, no-wrap msgid "Lists and Queues" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:234 msgid "" "See man:queue[3]. There are a LOT of cases when a list of things needs to " "be maintained. Fortunately, this data structure is implemented (in several " "ways) by C macros included in the system. The most used list type is TAILQ " "because it is the most flexible. It is also the one with largest memory " "requirements (its elements are doubly-linked) and also the slowest (although " "the speed variation is on the order of several CPU instructions more, so it " "should not be taken seriously)." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:236 msgid "" "If data retrieval speed is very important, see man:tree[3] and " "man:hashinit[9]." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:238 #, no-wrap msgid "BIOs" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:242 msgid "" "Structure `bio` is used for any and all Input/Output operations concerning " "GEOM. It basically contains information about what device ('provider') " "should satisfy the request, request type, offset, length, pointer to a " "buffer, and a bunch of \"user-specific\" flags and fields that can help " "implement various hacks." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:246 msgid "" "The important thing here is that ``bio``s are handled asynchronously. That " "means that, in most parts of the code, there is no analogue to userland's " "man:read[2] and man:write[2] calls that do not return until a request is " "done. Rather, a developer-supplied function is called as a notification " "when the request gets completed (or results in error)." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:250 msgid "" "The asynchronous programming model (also called \"event-driven\") is " "somewhat harder than the much more used imperative one used in userland (at " "least it takes a while to get used to it). In some cases the helper " "routines `g_write_data`() and `g_read_data`() can be used, but __not " "always__. In particular, they cannot be used when a mutex is held; for " "example, the GEOM topology mutex or the internal mutex held during the " "`.start`() and `.stop`() functions." msgstr "" #. type: Title == #: documentation/content/en/articles/geom-class/_index.adoc:252 #, no-wrap msgid "On GEOM Programming" msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:255 #, no-wrap msgid "Ggate" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:259 msgid "" "If maximum performance is not needed, a much simpler way of making a data " "transformation is to implement it in userland via the ggate (GEOM gate) " "facility. Unfortunately, there is no easy way to convert between, or even " "share code between the two approaches." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:261 #, no-wrap msgid "GEOM Class" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:266 msgid "" "GEOM classes are transformations on the data. These transformations can be " "combined in a tree-like fashion. Instances of GEOM classes are called " "__geoms__." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:268 msgid "" "Each GEOM class has several \"class methods\" that get called when there is " "no geom instance available (or they are simply not bound to a single " "instance):" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:270 msgid "" "`.init` is called when GEOM becomes aware of a GEOM class (when the kernel " "module gets loaded.)" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:271 msgid "" "`.fini` gets called when GEOM abandons the class (when the module gets " "unloaded)" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:272 msgid "" "`.taste` is called next, once for each provider the system has available. If " "applicable, this function will usually create and start a geom instance." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:273 msgid "`.destroy_geom` is called when the geom should be disbanded" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:274 msgid "" "`.ctlconf` is called when user requests reconfiguration of existing geom" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:276 msgid "" "Also defined are the GEOM event functions, which will get copied to the geom " "instance." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:278 msgid "" "Field `.geom` in the `g_class` structure is a LIST of geoms instantiated " "from the class." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:280 msgid "These functions are called from the g_event kernel thread." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:282 #, no-wrap msgid "Softc" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:288 msgid "" "The name \"softc\" is a legacy term for \"driver private data\". The name " "most probably comes from the archaic term \"software control block\". In " "GEOM, it is a structure (more precise: pointer to a structure) that can be " "attached to a geom instance to hold whatever data is private to the geom " "instance. Most GEOM classes have the following members:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:290 msgid "`struct g_provider *provider` : The \"provider\" this geom instantiates" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:291 msgid "`uint16_t n_disks` : Number of consumer this geom consumes" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:292 msgid "" "`struct g_consumer \\**disks` : Array of `struct g_consumer*`. (It is not " "possible to use just single indirection because struct g_consumer* are " "created on our behalf by GEOM)." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:295 msgid "" "The `softc` structure contains all the state of geom instance. Every geom " "instance has its own softc." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:297 #, no-wrap msgid "Metadata" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:300 msgid "" "Format of metadata is more-or-less class-dependent, but MUST start with:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:302 msgid "16 byte buffer for null-terminated signature (usually the class name)" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:303 msgid "uint32 version ID" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:305 msgid "" "It is assumed that geom classes know how to handle metadata with version " "ID's lower than theirs." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:307 msgid "" "Metadata is located in the last sector of the provider (and thus must fit in " "it)." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:309 msgid "" "(All this is implementation-dependent but all existing code works like that, " "and it is supported by libraries.)" msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:311 #, no-wrap msgid "Labeling/creating a GEOM" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:314 msgid "The sequence of events is:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:316 msgid "user calls man:geom[8] utility (or one of its hardlinked friends)" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:317 msgid "" "the utility figures out which geom class it is supposed to handle and " "searches for [.filename]#geom_CLASSNAME.so# library (usually in [.filename]#/" "lib/geom#)." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:318 msgid "" "it man:dlopen[3]-s the library, extracts the definitions of command-line " "parameters and helper functions." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:320 msgid "In the case of creating/labeling a new geom, this is what happens:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:322 msgid "" "man:geom[8] looks in the command-line argument for the command (usually " "`label`), and calls a helper function." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:323 msgid "" "The helper function checks parameters and gathers metadata, which it " "proceeds to write to all concerned providers." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:324 msgid "" "This \"spoils\" existing geoms (if any) and initializes a new round of " "\"tasting\" of the providers. The intended geom class recognizes the " "metadata and brings the geom up." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:326 msgid "" "(The above sequence of events is implementation-dependent but all existing " "code works like that, and it is supported by libraries.)" msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:328 #, no-wrap msgid "GEOM Command Structure" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:332 msgid "" "The helper [.filename]#geom_CLASSNAME.so# library exports `class_commands` " "structure, which is an array of `struct g_command` elements. Commands are " "of uniform format and look like:" msgstr "" #. type: delimited block . 4 #: documentation/content/en/articles/geom-class/_index.adoc:336 #, no-wrap msgid " verb [-options] geomname [other]\n" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:339 msgid "Common verbs are:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:341 msgid "" "label - to write metadata to devices so they can be recognized at tasting " "and brought up in geoms" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:342 msgid "destroy - to destroy metadata, so the geoms get destroyed" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:344 msgid "Common options are:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:346 msgid "`-v` : be verbose" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:347 msgid "`-f` : force" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:351 msgid "" "Many actions, such as labeling and destroying metadata can be performed in " "userland. For this, `struct g_command` provides field `gc_func` that can be " "set to a function (in the same [.filename]#.so#) that will be called to " "process a verb. If `gc_func` is NULL, the command will be passed to kernel " "module, to `.ctlreq` function of the geom class." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:353 #, no-wrap msgid "Geoms" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:357 msgid "" "Geoms are instances of GEOM classes. They have internal data (a softc " "structure) and some functions with which they respond to external events." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:359 msgid "The event functions are:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:361 msgid "`.access` : calculates permissions (read/write/exclusive)" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:362 msgid "`.dumpconf` : returns XML-formatted information about the geom" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:363 msgid "`.orphan` : called when some underlying provider gets disconnected" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:364 msgid "`.spoiled` : called when some underlying provider gets written to" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:365 msgid "`.start` : handles I/O" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:367 msgid "" "These functions are called from the `g_down` kernel thread and there can be " "no sleeping in this context, (see definition of sleeping elsewhere) which " "limits what can be done quite a bit, but forces the handling to be fast." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:369 msgid "" "Of these, the most important function for doing actual useful work is the " "`.start`() function, which is called when a BIO request arrives for a " "provider managed by a instance of geom class." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:371 #, no-wrap msgid "GEOM Threads" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:374 msgid "There are three kernel threads created and run by the GEOM framework:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:376 msgid "" "`g_down` : Handles requests coming from high-level entities (such as a " "userland request) on the way to physical devices" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:377 msgid "" "`g_up` : Handles responses from device drivers to requests made by higher-" "level entities" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:378 msgid "" "`g_event` : Handles all other cases: creation of geom instances, access " "counting, \"spoil\" events, etc." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:380 msgid "" "When a user process issues \"read data X at offset Y of a file\" request, " "this is what happens:" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:382 msgid "" "The filesystem converts the request into a struct bio instance and passes it " "to the GEOM subsystem. It knows what geom instance should handle it because " "filesystems are hosted directly on a geom instance." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:383 msgid "" "The request ends up as a call to the `.start`() function made on the g_down " "thread and reaches the top-level geom instance." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:384 msgid "" "This top-level geom instance (for example the partition slicer) determines " "that the request should be routed to a lower-level instance (for example the " "disk driver). It makes a copy of the bio request (bio requests _ALWAYS_ need " "to be copied between instances, with `g_clone_bio`()!), modifies the data " "offset and target provider fields and executes the copy with `g_io_request`()" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:385 msgid "" "The disk driver gets the bio request also as a call to `.start`() on the " "`g_down` thread. It talks to hardware, gets the data back, and calls " "`g_io_deliver`() on the bio." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:386 msgid "" "Now, the notification of bio completion \"bubbles up\" in the `g_up` thread. " "First the partition slicer gets `.done`() called in the `g_up` thread, it " "uses information stored in the bio to free the cloned `bio` structure (with " "`g_destroy_bio`()) and calls `g_io_deliver`() on the original request." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:387 msgid "The filesystem gets the data and transfers it to userland." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:389 msgid "" "See man:g_bio[9] man page for information how the data is passed back and " "forth in the `bio` structure (note in particular the `bio_parent` and " "`bio_children` fields and how they are handled)." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:392 msgid "" "One important feature is: __THERE CAN BE NO SLEEPING IN G_UP AND G_DOWN " "THREADS__. This means that none of the following things can be done in " "those threads (the list is of course not complete, but only informative):" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:394 msgid "Calls to `msleep`() and `tsleep`(), obviously." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:395 msgid "" "Calls to `g_write_data`() and `g_read_data`(), because these sleep between " "passing the data to consumers and returning." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:396 msgid "Waiting for I/O." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:397 msgid "Calls to man:malloc[9] and `uma_zalloc`() with `M_WAITOK` flag set" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:398 msgid "sx and other sleepable locks" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:402 msgid "" "This restriction is here to stop GEOM code clogging the I/O request path, " "since sleeping is usually not time-bound and there can be no guarantees on " "how long will it take (there are some other, more technical reasons also). " "It also means that there is not much that can be done in those threads; for " "example, almost any complex thing requires memory allocation. Fortunately, " "there is a way out: creating additional kernel threads." msgstr "" #. type: Title === #: documentation/content/en/articles/geom-class/_index.adoc:404 #, no-wrap msgid "Kernel Threads for Use in GEOM Code" msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:407 msgid "" "Kernel threads are created with man:kthread_create[9] function, and they are " "sort of similar to userland threads in behavior, only they cannot return to " "caller to signify termination, but must call man:kthread_exit[9]." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:410 msgid "" "In GEOM code, the usual use of threads is to offload processing of requests " "from `g_down` thread (the `.start`() function). These threads look like " "\"event handlers\": they have a linked list of event associated with them " "(on which events can be posted by various functions in various threads so it " "must be protected by a mutex), take the events from the list one by one and " "process them in a big `switch`() statement." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:415 msgid "" "The main benefit of using a thread to handle I/O requests is that it can " "sleep when needed. Now, this sounds good, but should be carefully thought " "out. Sleeping is well and very convenient but can very effectively destroy " "performance of the geom transformation. Extremely performance-sensitive " "classes probably should do all the work in `.start`() function call, taking " "great care to handle out-of-memory and similar errors." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:419 msgid "" "The other benefit of having a event-handler thread like that is to serialize " "all the requests and responses coming from different geom threads into one " "thread. This is also very convenient but can be slow. In most cases, " "handling of `.done`() requests can be left to the `g_up` thread." msgstr "" #. type: Plain text #: documentation/content/en/articles/geom-class/_index.adoc:422 msgid "" "Mutexes in FreeBSD kernel (see man:mutex[9]) have one distinction from their " "more common userland cousins - the code cannot sleep while holding a " "mutex). If the code needs to sleep a lot, man:sx[9] locks may be more " "appropriate. On the other hand, if you do almost everything in a single " "thread, you may get away with no mutexes at all." msgstr ""