diff --git a/sys/conf/NOTES b/sys/conf/NOTES
index 5819eeb57b2d..fec65452ba0f 100644
--- a/sys/conf/NOTES
+++ b/sys/conf/NOTES
@@ -1,2872 +1,2874 @@
#
# NOTES -- Lines that can be cut/pasted into kernel and hints configs.
#
# Lines that begin with 'device', 'options', 'machine', 'ident', 'maxusers',
# 'makeoptions', 'hints', etc. go into the kernel configuration that you
# run config(8) with.
#
# Lines that begin with 'envvar hint.' should go into your hints file.
# See /boot/device.hints and/or the 'hints' config(8) directive.
#
# Please use ``make LINT'' to create an old-style LINT file if you want to
# do kernel test-builds.
#
# This file contains machine independent kernel configuration notes. For
# machine dependent notes, look in /sys/<arch>/conf/NOTES.
#
#
# NOTES conventions and style guide:
#
# Large block comments should begin and end with a line containing only a
# comment character.
#
# To describe a particular object, a block comment (if it exists) should
# come first. Next should come device, options, and hints lines in that
# order. All device and option lines must be described by a comment that
# doesn't just expand the device or option name. Use only a concise
# comment on the same line if possible. Very detailed descriptions of
# devices and subsystems belong in man pages.
#
# A space followed by a tab separates 'options' from an option name. Two
# spaces followed by a tab separate 'device' from a device name. Comments
# after an option or device should use one space after the comment character.
# To comment out a negative option that disables code and thus should not be
# enabled for LINT builds, precede 'options' with "#!".
#
#
# This is the ``identification'' of the kernel. Usually this should
# be the same as the name of your kernel.
#
ident LINT
#
# The `maxusers' parameter controls the static sizing of a number of
# internal system tables by a formula defined in subr_param.c.
# Omitting this parameter or setting it to 0 will cause the system to
# auto-size based on physical memory.
#
maxusers 10
# To statically compile in device wiring instead of /boot/device.hints
#hints "LINT.hints" # Default places to look for devices.
# Use the following to compile in values accessible to the kernel
# through getenv() (or kenv(1) in userland). The format of the file
# is 'variable=value', see kenv(1)
#
#env "LINT.env"
#
# The `makeoptions' parameter allows variables to be passed to the
# generated Makefile in the build area.
#
# CONF_CFLAGS gives some extra compiler flags that are added to ${CFLAGS}
# after most other flags. Here we use it to inhibit use of non-optimal
# gcc built-in functions (e.g., memcmp).
#
# DEBUG happens to be magic.
# The following is equivalent to 'config -g KERNELNAME' and creates
# 'kernel.debug' compiled with -g debugging as well as a normal
# 'kernel'. Use 'make install.debug' to install the debug kernel
# but that isn't normally necessary as the debug symbols are not loaded
# by the kernel and are not useful there anyway.
#
# KERNEL can be overridden so that you can change the default name of your
# kernel.
#
# MODULES_OVERRIDE can be used to limit modules built to a specific list.
#
makeoptions CONF_CFLAGS=-fno-builtin #Don't allow use of memcmp, etc.
#makeoptions DEBUG=-g #Build kernel with gdb(1) debug symbols
#makeoptions KERNEL=foo #Build kernel "foo" and install "/foo"
# Only build ext2fs module plus those parts of the sound system I need.
#makeoptions MODULES_OVERRIDE="ext2fs sound/sound sound/driver/maestro3"
makeoptions DESTDIR=/tmp
#
# FreeBSD processes are subject to certain limits to their consumption
# of system resources. See getrlimit(2) for more details. Each
# resource limit has two values, a "soft" limit and a "hard" limit.
# The soft limits can be modified during normal system operation, but
# the hard limits are set at boot time. Their default values are
# in sys/<arch>/include/vmparam.h. There are two ways to change them:
#
# 1. Set the values at kernel build time. The options below are one
# way to allow that limit to grow to 1GB. They can be increased
# further by changing the parameters:
#
# 2. In /boot/loader.conf, set the tunables kern.maxswzone,
# kern.maxbcache, kern.maxtsiz, kern.dfldsiz, kern.maxdsiz,
# kern.dflssiz, kern.maxssiz and kern.sgrowsiz.
#
# The options in /boot/loader.conf override anything in the kernel
# configuration file. See the function init_param1 in
# sys/kern/subr_param.c for more details.
#
options MAXDSIZ=(1024UL*1024*1024)
options MAXSSIZ=(128UL*1024*1024)
options DFLDSIZ=(1024UL*1024*1024)
#
# BLKDEV_IOSIZE sets the default block size used in user block
# device I/O. Note that this value will be overridden by the label
# when specifying a block device from a label with a non-0
# partition blocksize. The default is PAGE_SIZE.
#
options BLKDEV_IOSIZE=8192
#
# MAXPHYS and DFLTPHYS
#
# These are the maximal and safe 'raw' I/O block device access sizes.
# Reads and writes will be split into MAXPHYS chunks for known good
# devices and DFLTPHYS for the rest. Some applications have better
# performance with larger raw I/O access sizes. Note that certain VM
# parameters are derived from these values and making them too large
# can make an unbootable kernel.
#
# The defaults are 64K and 128K respectively.
options DFLTPHYS=(64*1024)
options MAXPHYS=(128*1024)
# This allows you to actually store this configuration file into
# the kernel binary itself. See config(8) for more details.
#
options INCLUDE_CONFIG_FILE # Include this file in kernel
#
# Compile-time defaults for various boot parameters
#
options BOOTVERBOSE=1
options BOOTHOWTO=RB_MULTIPLE
#
# Compile-time defaults for dmesg boot tagging
#
# Default boot tag; may use 'kern.boot_tag' loader tunable to override. The
# current boot's tag is also exposed via the 'kern.boot_tag' sysctl.
options BOOT_TAG=\"\"
# Maximum boot tag size the kernel's static buffer should accommodate. Maximum
# size for both BOOT_TAG and the assocated tunable.
options BOOT_TAG_SZ=32
options GEOM_BDE # Disk encryption.
options GEOM_CACHE # Disk cache.
options GEOM_CONCAT # Disk concatenation.
options GEOM_ELI # Disk encryption.
options GEOM_GATE # Userland services.
options GEOM_JOURNAL # Journaling.
options GEOM_LABEL # Providers labelization.
options GEOM_LINUX_LVM # Linux LVM2 volumes
options GEOM_MAP # Map based partitioning
options GEOM_MIRROR # Disk mirroring.
options GEOM_MULTIPATH # Disk multipath
options GEOM_NOP # Test class.
options GEOM_PART_APM # Apple partitioning
options GEOM_PART_BSD # BSD disklabel
options GEOM_PART_BSD64 # BSD disklabel64
options GEOM_PART_EBR # Extended Boot Records
options GEOM_PART_GPT # GPT partitioning
options GEOM_PART_LDM # Logical Disk Manager
options GEOM_PART_MBR # MBR partitioning
options GEOM_RAID # Soft RAID functionality.
options GEOM_RAID3 # RAID3 functionality.
options GEOM_SHSEC # Shared secret.
options GEOM_STRIPE # Disk striping.
options GEOM_UZIP # Read-only compressed disks
options GEOM_VINUM # Vinum logical volume manager
options GEOM_VIRSTOR # Virtual storage.
options GEOM_ZERO # Performance testing helper.
#
# The root device and filesystem type can be compiled in;
# this provides a fallback option if the root device cannot
# be correctly guessed by the bootstrap code, or an override if
# the RB_DFLTROOT flag (-r) is specified when booting the kernel.
#
options ROOTDEVNAME=\"ufs:da0s2e\"
#####################################################################
# Scheduler options:
#
# Specifying one of SCHED_4BSD or SCHED_ULE is mandatory. These options
# select which scheduler is compiled in.
#
# SCHED_4BSD is the historical, proven, BSD scheduler. It has a global run
# queue and no CPU affinity which makes it suboptimal for SMP. It has very
# good interactivity and priority selection.
#
# SCHED_ULE provides significant performance advantages over 4BSD on many
# workloads on SMP machines. It supports cpu-affinity, per-cpu runqueues
# and scheduler locks. It also has a stronger notion of interactivity
# which leads to better responsiveness even on uniprocessor machines. This
# is the default scheduler.
#
# SCHED_STATS is a debugging option which keeps some stats in the sysctl
# tree at 'kern.sched.stats' and is useful for debugging scheduling decisions.
#
options SCHED_4BSD
options SCHED_STATS
#options SCHED_ULE
#####################################################################
# SMP OPTIONS:
#
# SMP enables building of a Symmetric MultiProcessor Kernel.
# Mandatory:
options SMP # Symmetric MultiProcessor Kernel
# EARLY_AP_STARTUP releases the Application Processors earlier in the
# kernel startup process (before devices are probed) rather than at the
# end. This is a temporary option for use during the transition from
# late to early AP startup.
options EARLY_AP_STARTUP
# MAXCPU defines the maximum number of CPUs that can boot in the system.
# A default value should be already present, for every architecture.
options MAXCPU=32
# NUMA enables use of Non-Uniform Memory Access policies in various kernel
# subsystems.
options NUMA
# MAXMEMDOM defines the maximum number of memory domains that can boot in the
# system. A default value should already be defined by every architecture.
options MAXMEMDOM=2
# ADAPTIVE_MUTEXES changes the behavior of blocking mutexes to spin
# if the thread that currently owns the mutex is executing on another
# CPU. This behavior is enabled by default, so this option can be used
# to disable it.
options NO_ADAPTIVE_MUTEXES
# ADAPTIVE_RWLOCKS changes the behavior of reader/writer locks to spin
# if the thread that currently owns the rwlock is executing on another
# CPU. This behavior is enabled by default, so this option can be used
# to disable it.
options NO_ADAPTIVE_RWLOCKS
# ADAPTIVE_SX changes the behavior of sx locks to spin if the thread that
# currently owns the sx lock is executing on another CPU.
# This behavior is enabled by default, so this option can be used to
# disable it.
options NO_ADAPTIVE_SX
# MUTEX_NOINLINE forces mutex operations to call functions to perform each
# operation rather than inlining the simple cases. This can be used to
# shrink the size of the kernel text segment. Note that this behavior is
# already implied by the INVARIANT_SUPPORT, INVARIANTS, KTR, LOCK_PROFILING,
# and WITNESS options.
options MUTEX_NOINLINE
# RWLOCK_NOINLINE forces rwlock operations to call functions to perform each
# operation rather than inlining the simple cases. This can be used to
# shrink the size of the kernel text segment. Note that this behavior is
# already implied by the INVARIANT_SUPPORT, INVARIANTS, KTR, LOCK_PROFILING,
# and WITNESS options.
options RWLOCK_NOINLINE
# SX_NOINLINE forces sx lock operations to call functions to perform each
# operation rather than inlining the simple cases. This can be used to
# shrink the size of the kernel text segment. Note that this behavior is
# already implied by the INVARIANT_SUPPORT, INVARIANTS, KTR, LOCK_PROFILING,
# and WITNESS options.
options SX_NOINLINE
# SMP Debugging Options:
#
# CALLOUT_PROFILING enables rudimentary profiling of the callwheel data
# structure used as backend in callout(9).
# PREEMPTION allows the threads that are in the kernel to be preempted by
# higher priority [interrupt] threads. It helps with interactivity
# and allows interrupt threads to run sooner rather than waiting.
# FULL_PREEMPTION instructs the kernel to preempt non-realtime kernel
# threads. Its sole use is to expose race conditions and other
# bugs during development. Enabling this option will reduce
# performance and increase the frequency of kernel panics by
# design. If you aren't sure that you need it then you don't.
# Relies on the PREEMPTION option. DON'T TURN THIS ON.
# SLEEPQUEUE_PROFILING enables rudimentary profiling of the hash table
# used to hold active sleep queues as well as sleep wait message
# frequency.
# TURNSTILE_PROFILING enables rudimentary profiling of the hash table
# used to hold active lock queues.
# UMTX_PROFILING enables rudimentary profiling of the hash table used
# to hold active lock queues.
# WITNESS enables the witness code which detects deadlocks and cycles
# during locking operations.
# WITNESS_KDB causes the witness code to drop into the kernel debugger if
# a lock hierarchy violation occurs or if locks are held when going to
# sleep.
# WITNESS_SKIPSPIN disables the witness checks on spin mutexes.
options PREEMPTION
options FULL_PREEMPTION
options WITNESS
options WITNESS_KDB
options WITNESS_SKIPSPIN
# LOCK_PROFILING - Profiling locks. See LOCK_PROFILING(9) for details.
options LOCK_PROFILING
# Set the number of buffers and the hash size. The hash size MUST be larger
# than the number of buffers. Hash size should be prime.
options MPROF_BUFFERS="1536"
options MPROF_HASH_SIZE="1543"
# Profiling for the callout(9) backend.
options CALLOUT_PROFILING
# Profiling for internal hash tables.
options SLEEPQUEUE_PROFILING
options TURNSTILE_PROFILING
options UMTX_PROFILING
# Debugging traces for epoch(9) misuse
options EPOCH_TRACE
#####################################################################
# COMPATIBILITY OPTIONS
# Old tty interface.
options COMPAT_43TTY
# Note that as a general rule, COMPAT_FREEBSD<n> depends on
# COMPAT_FREEBSD<n+1>, COMPAT_FREEBSD<n+2>, etc.
# Enable FreeBSD4 compatibility syscalls
options COMPAT_FREEBSD4
# Enable FreeBSD5 compatibility syscalls
options COMPAT_FREEBSD5
# Enable FreeBSD6 compatibility syscalls
options COMPAT_FREEBSD6
# Enable FreeBSD7 compatibility syscalls
options COMPAT_FREEBSD7
# Enable FreeBSD9 compatibility syscalls
options COMPAT_FREEBSD9
# Enable FreeBSD10 compatibility syscalls
options COMPAT_FREEBSD10
# Enable FreeBSD11 compatibility syscalls
options COMPAT_FREEBSD11
# Enable FreeBSD12 compatibility syscalls
options COMPAT_FREEBSD12
# Enable FreeBSD13 compatibility syscalls
options COMPAT_FREEBSD13
# Enable FreeBSD14 compatibility syscalls
options COMPAT_FREEBSD14
# Enable Linux Kernel Programming Interface
options COMPAT_LINUXKPI
#
# These three options provide support for System V Interface
# Definition-style interprocess communication, in the form of shared
# memory, semaphores, and message queues, respectively.
#
options SYSVSHM
options SYSVSEM
options SYSVMSG
#####################################################################
# DEBUGGING OPTIONS
#
# Compile with kernel debugger related code.
#
options KDB
#
# Print a stack trace of the current thread on the console for a panic.
#
options KDB_TRACE
#
# Don't enter the debugger for a panic. Intended for unattended operation
# where you may want to enter the debugger from the console, but still want
# the machine to recover from a panic.
#
options KDB_UNATTENDED
#
# Enable the ddb debugger backend.
#
options DDB
#
# Print the numerical value of symbols in addition to the symbolic
# representation.
#
options DDB_NUMSYM
#
# Enable the remote gdb debugger backend.
#
options GDB
#
# Trashes list pointers when they become invalid (i.e., the element is
# removed from a list). Relatively inexpensive to enable.
#
options QUEUE_MACRO_DEBUG_TRASH
#
# Stores information about the last caller to modify the list object
# in the list object. Requires additional memory overhead.
#
#options QUEUE_MACRO_DEBUG_TRACE
#
# SYSCTL_DEBUG enables a 'sysctl' debug tree that can be used to dump the
# contents of the registered sysctl nodes on the console. It is disabled by
# default because it generates excessively verbose console output that can
# interfere with serial console operation.
#
options SYSCTL_DEBUG
#
# Enable textdump by default, this disables kernel core dumps.
#
options TEXTDUMP_PREFERRED
#
# Enable extra debug messages while performing textdumps.
#
options TEXTDUMP_VERBOSE
#
# NO_SYSCTL_DESCR omits the sysctl node descriptions to save space in the
# resulting kernel.
options NO_SYSCTL_DESCR
#
# MALLOC_DEBUG_MAXZONES enables multiple uma zones for malloc(9)
# allocations that are smaller than a page. The purpose is to isolate
# different malloc types into hash classes, so that any buffer
# overruns or use-after-free will usually only affect memory from
# malloc types in that hash class. This is purely a debugging tool;
# by varying the hash function and tracking which hash class was
# corrupted, the intersection of the hash classes from each instance
# will point to a single malloc type that is being misused. At this
# point inspection or memguard(9) can be used to catch the offending
# code.
#
options MALLOC_DEBUG_MAXZONES=8
#
# DEBUG_MEMGUARD builds and enables memguard(9), a replacement allocator
# for the kernel used to detect modify-after-free scenarios. See the
# memguard(9) man page for more information on usage.
#
options DEBUG_MEMGUARD
#
# DEBUG_REDZONE enables buffer underflows and buffer overflows detection for
# malloc(9).
#
options DEBUG_REDZONE
#
# EARLY_PRINTF enables support for calling a special printf (eprintf)
# very early in the kernel (before cn_init() has been called). This
# should only be used for debugging purposes early in boot. Normally,
# it is not defined. It is commented out here because this feature
# isn't generally available. And the required eputc() isn't defined.
#
#options EARLY_PRINTF
#
# KTRACE enables the system-call tracing facility ktrace(2). To be more
# SMP-friendly, KTRACE uses a worker thread to process most trace events
# asynchronously to the thread generating the event. This requires a
# pre-allocated store of objects representing trace events. The
# KTRACE_REQUEST_POOL option specifies the initial size of this store.
# The size of the pool can be adjusted both at boottime and runtime via
# the kern.ktrace_request_pool tunable and sysctl.
#
options KTRACE #kernel tracing
options KTRACE_REQUEST_POOL=101
#
# KTR is a kernel tracing facility imported from BSD/OS. It is
# enabled with the KTR option. KTR_ENTRIES defines the number of
# entries in the circular trace buffer; it may be an arbitrary number.
# KTR_BOOT_ENTRIES defines the number of entries during the early boot,
# before malloc(9) is functional.
# KTR_COMPILE defines the mask of events to compile into the kernel as
# defined by the KTR_* constants in <sys/ktr.h>. KTR_MASK defines the
# initial value of the ktr_mask variable which determines at runtime
# what events to trace. KTR_CPUMASK determines which CPU's log
# events, with bit X corresponding to CPU X. The layout of the string
# passed as KTR_CPUMASK must match a series of bitmasks each of them
# separated by the "," character (ie:
# KTR_CPUMASK=0xAF,0xFFFFFFFFFFFFFFFF). KTR_VERBOSE enables
# dumping of KTR events to the console by default. This functionality
# can be toggled via the debug.ktr_verbose sysctl and defaults to off
# if KTR_VERBOSE is not defined. See ktr(4) and ktrdump(8) for details.
#
options KTR
options KTR_BOOT_ENTRIES=1024
options KTR_ENTRIES=(128*1024)
options KTR_COMPILE=(KTR_ALL)
options KTR_MASK=KTR_INTR
options KTR_CPUMASK=0x3
options KTR_VERBOSE
#
# ALQ(9) is a facility for the asynchronous queuing of records from the kernel
# to a vnode, and is employed by services such as ktr(4) to produce trace
# files based on a kernel event stream. Records are written asynchronously
# in a worker thread.
#
options ALQ
options KTR_ALQ
#
# The INVARIANTS option is used in a number of source files to enable
# extra sanity checking of internal structures. This support is not
# enabled by default because of the extra time it would take to check
# for these conditions, which can only occur as a result of
# programming errors.
#
options INVARIANTS
#
# The INVARIANT_SUPPORT option makes us compile in support for
# verifying some of the internal structures. It is a prerequisite for
# 'INVARIANTS', as enabling 'INVARIANTS' will make these functions be
# called. The intent is that you can set 'INVARIANTS' for single
# source files (by changing the source file or specifying it on the
# command line) if you have 'INVARIANT_SUPPORT' enabled. Also, if you
# wish to build a kernel module with 'INVARIANTS', then adding
# 'INVARIANT_SUPPORT' to your kernel will provide all the necessary
# infrastructure without the added overhead.
#
options INVARIANT_SUPPORT
#
# The KASSERT_PANIC_OPTIONAL option allows kasserts to fire without
# necessarily inducing a panic. Panic is the default behavior, but
# runtime options can configure it either entirely off, or off with a
# limit.
#
options KASSERT_PANIC_OPTIONAL
#
# The DIAGNOSTIC option is used to enable extra debugging information
# and invariants checking. The added checks are too expensive or noisy
# for an INVARIANTS kernel and thus are disabled by default. It is
# expected that a kernel configured with DIAGNOSTIC will also have the
# INVARIANTS option enabled.
#
options DIAGNOSTIC
#
# REGRESSION causes optional kernel interfaces necessary only for regression
# testing to be enabled. These interfaces may constitute security risks
# when enabled, as they permit processes to easily modify aspects of the
# run-time environment to reproduce unlikely or unusual (possibly normally
# impossible) scenarios.
#
options REGRESSION
#
# This option lets some drivers co-exist that can't co-exist in a running
# system. This is used to be able to compile all kernel code in one go for
# quality assurance purposes (like this file, which the option takes it name
# from.)
#
options COMPILING_LINT
#
# STACK enables the stack(9) facility, allowing the capture of kernel stack
# for the purpose of procinfo(1), etc. stack(9) will also be compiled in
# automatically if DDB(4) is compiled into the kernel.
#
options STACK
#
# The NUM_CORE_FILES option specifies the limit for the number of core
# files generated by a particular process, when the core file format
# specifier includes the %I pattern. Since we only have 1 character for
# the core count in the format string, meaning the range will be 0-9, the
# maximum value allowed for this option is 10.
# This core file limit can be adjusted at runtime via the debug.ncores
# sysctl.
#
options NUM_CORE_FILES=5
#
# The TSLOG option enables timestamped logging of events, especially
# function entries/exits, in order to track the time spent by the kernel.
# In particular, this is useful when investigating the early boot process,
# before it is possible to use more sophisticated tools like DTrace.
# The TSLOGSIZE option controls the size of the (preallocated, fixed
# length) buffer used for storing these events (default: 262144 records).
# The TSLOG_PAGEZERO option enables TSLOG of pmap_zero_page; this must be
# enabled separately since it typically generates too many records to be
# useful.
#
# For security reasons the TSLOG option should not be enabled on systems
# used in production.
#
options TSLOG
options TSLOGSIZE=262144
#####################################################################
# PERFORMANCE MONITORING OPTIONS
#
# The hwpmc driver that allows the use of in-CPU performance monitoring
# counters for performance monitoring. The base kernel needs to be configured
# with the 'options' line, while the hwpmc device can be either compiled
# in or loaded as a loadable kernel module.
#
# Additional configuration options may be required on specific architectures,
# please see hwpmc(4).
device hwpmc # Driver (also a loadable module)
options HWPMC_DEBUG
options HWPMC_HOOKS # Other necessary kernel hooks
#####################################################################
# NETWORKING OPTIONS
#
# Protocol families
#
options INET #Internet communications protocols
options INET6 #IPv6 communications protocols
#
# Note if you include INET/INET6 or both options
# You *must* define at least one of the congestion control
# options or the compile will fail. GENERIC defines
# options CC_CUBIC. You may want to specify a default
# if multiple congestion controls are compiled in.
# The string in default is the name of the
# cc module as it would appear in the sysctl for
# setting the default. The code defines CUBIC
# as default, or the sole cc_module compiled in.
#
options CC_CDG
options CC_CHD
options CC_CUBIC
options CC_DCTCP
options CC_HD
options CC_HTCP
options CC_NEWRENO
options CC_VEGAS
options CC_DEFAULT=\"cubic\"
options RATELIMIT # TX rate limiting support
options ROUTETABLES=2 # allocated fibs up to 65536. default is 1.
# but that would be a bad idea as they are large.
options TCP_OFFLOAD # TCP offload support.
options TCP_RFC7413 # TCP Fast Open
options TCPHPTS
# In order to enable IPSEC you MUST also add device crypto to
# your kernel configuration
options IPSEC #IP security (requires device crypto)
# Option IPSEC_SUPPORT does not enable IPsec, but makes it possible to
# load it as a kernel module. You still MUST add device crypto to your kernel
# configuration.
options IPSEC_SUPPORT
#options IPSEC_DEBUG #debug for IP security
# Alternative TCP stacks
options TCP_BBR
options TCP_RACK
# TLS framing and encryption/decryption of data over TCP sockets.
options KERN_TLS # TLS transmit and receive offload
# Netlink kernel/user<>kernel/user messaging interface
options NETLINK
#
# SMB/CIFS requester
# NETSMB enables support for SMB protocol, it requires LIBMCHAIN and LIBICONV
# options.
options NETSMB #SMB/CIFS requester
# mchain library. It can be either loaded as KLD or compiled into kernel
options LIBMCHAIN
# libalias library, performing NAT
options LIBALIAS
#
# SCTP is a NEW transport protocol defined by
# RFC2960 updated by RFC3309 and RFC3758.. and
# soon to have a new base RFC and many many more
# extensions. This release supports all the extensions
# including many drafts (most about to become RFC's).
# It is the reference implementation of SCTP
# and is quite well tested.
#
# Note YOU MUST have both INET and INET6 defined.
# You don't have to enable V6, but SCTP is
# dual stacked and so far we have not torn apart
# the V6 and V4.. since an association can span
# both a V6 and V4 address at the SAME time :-)
#
# The SCTP_SUPPORT option does not enable SCTP, but provides the necessary
# support for loading SCTP as a loadable kernel module.
#
options SCTP
options SCTP_SUPPORT
# There are bunches of options:
# this one turns on all sorts of
# nastily printing that you can
# do. It's all controlled by a
# bit mask (settable by socket opt and
# by sysctl). Including will not cause
# logging until you set the bits.. but it
# can be quite verbose.. so without this
# option we don't do any of the tests for
# bits and prints.. which makes the code run
# faster.. if you are not debugging don't use.
options SCTP_DEBUG
#
# All that options after that turn on specific types of
# logging. You can monitor CWND growth, flight size
# and all sorts of things. Go look at the code and
# see. I have used this to produce interesting
# charts and graphs as well :->
#
# I have not yet committed the tools to get and print
# the logs, I will do that eventually .. before then
# if you want them send me an email rrs@freebsd.org
# You basically must have ktr(4) enabled for these
# and you then set the sysctl to turn on/off various
# logging bits. Use ktrdump(8) to pull the log and run
# it through a display program.. and graphs and other
# things too.
#
options SCTP_LOCK_LOGGING
options SCTP_MBUF_LOGGING
options SCTP_MBCNT_LOGGING
options SCTP_PACKET_LOGGING
options SCTP_LTRACE_CHUNKS
options SCTP_LTRACE_ERRORS
# OpenFabrics Enterprise Distribution (Infiniband).
options OFED
options OFED_DEBUG_INIT
# Sockets Direct Protocol
options SDP
options SDP_DEBUG
# IP over Infiniband
options IPOIB
options IPOIB_DEBUG
options IPOIB_CM
# altq(9). Enable the base part of the hooks with the ALTQ option.
# Individual disciplines must be built into the base system and can not be
# loaded as modules at this point. ALTQ requires a stable TSC so if yours is
# broken or changes with CPU throttling then you must also have the ALTQ_NOPCC
# option.
options ALTQ
options ALTQ_CBQ # Class Based Queueing
options ALTQ_RED # Random Early Detection
options ALTQ_RIO # RED In/Out
options ALTQ_CODEL # CoDel Active Queueing
options ALTQ_HFSC # Hierarchical Packet Scheduler
options ALTQ_FAIRQ # Fair Packet Scheduler
options ALTQ_CDNR # Traffic conditioner
options ALTQ_PRIQ # Priority Queueing
options ALTQ_NOPCC # Required if the TSC is unusable
options ALTQ_DEBUG
# netgraph(4). Enable the base netgraph code with the NETGRAPH option.
# Individual node types can be enabled with the corresponding option
# listed below; however, this is not strictly necessary as netgraph
# will automatically load the corresponding KLD module if the node type
# is not already compiled into the kernel. Each type below has a
# corresponding man page, e.g., ng_async(8).
options NETGRAPH # netgraph(4) system
options NETGRAPH_DEBUG # enable extra debugging, this
# affects netgraph(4) and nodes
# Node types
options NETGRAPH_ASYNC
options NETGRAPH_BLUETOOTH # ng_bluetooth(4)
options NETGRAPH_BLUETOOTH_HCI # ng_hci(4)
options NETGRAPH_BLUETOOTH_L2CAP # ng_l2cap(4)
options NETGRAPH_BLUETOOTH_SOCKET # ng_btsocket(4)
options NETGRAPH_BLUETOOTH_UBT # ng_ubt(4)
options NETGRAPH_BLUETOOTH_UBTBCMFW # ubtbcmfw(4)
options NETGRAPH_BPF
options NETGRAPH_BRIDGE
options NETGRAPH_CAR
options NETGRAPH_CHECKSUM
options NETGRAPH_CISCO
options NETGRAPH_DEFLATE
options NETGRAPH_DEVICE
options NETGRAPH_ECHO
options NETGRAPH_EIFACE
options NETGRAPH_ETHER
options NETGRAPH_FRAME_RELAY
options NETGRAPH_GIF
options NETGRAPH_GIF_DEMUX
options NETGRAPH_HOLE
options NETGRAPH_IFACE
options NETGRAPH_IP_INPUT
options NETGRAPH_IPFW
options NETGRAPH_KSOCKET
options NETGRAPH_L2TP
options NETGRAPH_LMI
options NETGRAPH_MPPC_COMPRESSION
options NETGRAPH_MPPC_ENCRYPTION
options NETGRAPH_NETFLOW
options NETGRAPH_NAT
options NETGRAPH_ONE2MANY
options NETGRAPH_PATCH
options NETGRAPH_PIPE
options NETGRAPH_PPP
options NETGRAPH_PPPOE
options NETGRAPH_PPTPGRE
options NETGRAPH_PRED1
options NETGRAPH_RFC1490
options NETGRAPH_SOCKET
options NETGRAPH_SPLIT
options NETGRAPH_TAG
options NETGRAPH_TCPMSS
options NETGRAPH_TEE
options NETGRAPH_UI
options NETGRAPH_VJC
options NETGRAPH_VLAN
# Network stack virtualization.
options VIMAGE
options VNET_DEBUG # debug for VIMAGE
#
# Network interfaces:
# The `loop' device is MANDATORY when networking is enabled.
device loop
# The `ether' device provides generic code to handle
# Ethernets; it is MANDATORY when an Ethernet device driver is
# configured.
device ether
# The `vlan' device implements the VLAN tagging of Ethernet frames
# according to IEEE 802.1Q.
device vlan
# The `vxlan' device implements the VXLAN encapsulation of Ethernet
# frames in UDP packets according to RFC7348.
device vxlan
# The `wlan' device provides generic code to support 802.11
# drivers, including host AP mode; it is MANDATORY for the wi,
# and ath drivers and will eventually be required by all 802.11 drivers.
device wlan
options IEEE80211_DEBUG #enable debugging msgs
options IEEE80211_DEBUG_REFCNT
options IEEE80211_SUPPORT_MESH #enable 802.11s D3.0 support
options IEEE80211_SUPPORT_TDMA #enable TDMA support
# The `wlan_wep', `wlan_tkip', and `wlan_ccmp' devices provide
# support for WEP, TKIP, and AES-CCMP crypto protocols optionally
# used with 802.11 devices that depend on the `wlan' module.
device wlan_wep
device wlan_ccmp
device wlan_tkip
# The `wlan_xauth' device provides support for external (i.e. user-mode)
# authenticators for use with 802.11 drivers that use the `wlan'
# module and support 802.1x and/or WPA security protocols.
device wlan_xauth
# The `wlan_acl' device provides a MAC-based access control mechanism
# for use with 802.11 drivers operating in ap mode and using the
# `wlan' module.
# The 'wlan_amrr' device provides AMRR transmit rate control algorithm
device wlan_acl
device wlan_amrr
# The `bpf' device enables the Berkeley Packet Filter. Be
# aware of the legal and administrative consequences of enabling this
# option. DHCP requires bpf.
device bpf
# The `netmap' device implements memory-mapped access to network
# devices from userspace, enabling wire-speed packet capture and
# generation even at 10Gbit/s. Requires support in the device
# driver. Supported drivers are ixgbe, e1000, re.
device netmap
# The `disc' device implements a minimal network interface,
# which throws away all packets sent and never receives any. It is
# included for testing and benchmarking purposes.
device disc
# The `epair' device implements a virtual back-to-back connected Ethernet
# like interface pair.
device epair
# The `edsc' device implements a minimal Ethernet interface,
# which discards all packets sent and receives none.
device edsc
# The `tuntap' device implements (user-)ppp, nos-tun(8) and a pty-like virtual
# Ethernet interface
device tuntap
# The `gif' device implements IPv6 over IP4 tunneling,
# IPv4 over IPv6 tunneling, IPv4 over IPv4 tunneling and
# IPv6 over IPv6 tunneling.
# The `gre' device implements GRE (Generic Routing Encapsulation) tunneling,
# as specified in the RFC 2784 and RFC 2890.
# The `me' device implements Minimal Encapsulation within IPv4 as
# specified in the RFC 2004.
# The XBONEHACK option allows the same pair of addresses to be configured on
# multiple gif interfaces.
device gif
device gre
device me
options XBONEHACK
# The `stf' device implements 6to4 encapsulation.
device stf
# The pf packet filter consists of three devices:
# The `pf' device provides /dev/pf and the firewall code itself.
# The `pflog' device provides the pflog0 interface which logs packets.
# The `pfsync' device provides the pfsync0 interface used for
# synchronization of firewall state tables (over the net).
device pf
device pflog
device pfsync
# Bridge interface.
device if_bridge
# Common Address Redundancy Protocol. See carp(4) for more details.
device carp
# IPsec interface.
device enc
# Link aggregation interface.
device lagg
# WireGuard interface.
device wg
#
# Internet family options:
#
# MROUTING enables the kernel multicast packet forwarder, which works
# with mrouted and XORP.
#
# IPFIREWALL enables support for IP firewall construction, in
# conjunction with the `ipfw' program. IPFIREWALL_VERBOSE sends
# logged packets to the system logger. IPFIREWALL_VERBOSE_LIMIT
# limits the number of times a matching entry can be logged.
#
# WARNING: IPFIREWALL defaults to a policy of "deny ip from any to any"
# and if you do not add other rules during startup to allow access,
# YOU WILL LOCK YOURSELF OUT. It is suggested that you set firewall_type=open
# in /etc/rc.conf when first enabling this feature, then refining the
# firewall rules in /etc/rc.firewall after you've tested that the new kernel
# feature works properly.
#
# IPFIREWALL_DEFAULT_TO_ACCEPT causes the default rule (at boot) to
# allow everything. Use with care, if a cracker can crash your
# firewall machine, they can get to your protected machines. However,
# if you are using it as an as-needed filter for specific problems as
# they arise, then this may be for you. Changing the default to 'allow'
# means that you won't get stuck if the kernel and /sbin/ipfw binary get
# out of sync.
#
# IPDIVERT enables the divert IP sockets, used by ``ipfw divert''. It
# depends on IPFIREWALL if compiled into the kernel.
#
# IPFIREWALL_NAT adds support for in kernel nat in ipfw, and it requires
# LIBALIAS.
#
# IPFIREWALL_NAT64 adds support for in kernel NAT64 in ipfw.
#
# IPFIREWALL_NPTV6 adds support for in kernel NPTv6 in ipfw.
#
# IPFIREWALL_PMOD adds support for protocols modification module. Currently
# it supports only TCP MSS modification.
#
# IPSTEALTH enables code to support stealth forwarding (i.e., forwarding
# packets without touching the TTL). This can be useful to hide firewalls
# from traceroute and similar tools.
#
# PF_DEFAULT_TO_DROP causes the default pf(4) rule to deny everything.
#
# TCPPCAP enables code which keeps the last n packets sent and received
# on a TCP socket.
#
# TCP_BLACKBOX enables enhanced TCP event logging.
#
# TCP_HHOOK enables the hhook(9) framework hooks for the TCP stack.
+# SOCKET_HHOOK enables the hhook(9) framework hooks for socket operations.
#
# ROUTE_MPATH provides support for multipath routing.
#
options MROUTING # Multicast routing
options IPFIREWALL #firewall
options IPFIREWALL_VERBOSE #enable logging to syslogd(8)
options IPFIREWALL_VERBOSE_LIMIT=100 #limit verbosity
options IPFIREWALL_DEFAULT_TO_ACCEPT #allow everything by default
options IPFIREWALL_NAT #ipfw kernel nat support
options IPFIREWALL_NAT64 #ipfw kernel NAT64 support
options IPFIREWALL_NPTV6 #ipfw kernel IPv6 NPT support
options IPDIVERT #divert sockets
options IPFILTER #ipfilter support
options IPFILTER_LOG #ipfilter logging
options IPFILTER_LOOKUP #ipfilter pools
options IPFILTER_DEFAULT_BLOCK #block all packets by default
options IPSTEALTH #support for stealth forwarding
options PF_DEFAULT_TO_DROP #drop everything by default
options TCPPCAP
options TCP_BLACKBOX
options TCP_HHOOK
+options SOCKET_HHOOK
options ROUTE_MPATH
# The MBUF_STRESS_TEST option enables options which create
# various random failures / extreme cases related to mbuf
# functions. See mbuf(9) for a list of available test cases.
# MBUF_PROFILING enables code to profile the mbuf chains
# exiting the system (via participating interfaces) and
# return a logarithmic histogram of monitored parameters
# (e.g. packet size, wasted space, number of mbufs in chain).
options MBUF_STRESS_TEST
options MBUF_PROFILING
# Statically link in accept filters
options ACCEPT_FILTER_DATA
options ACCEPT_FILTER_DNS
options ACCEPT_FILTER_HTTP
options ACCEPT_FILTER_TLS
# TCP_SIGNATURE adds support for RFC 2385 (TCP-MD5) digests. These are
# carried in TCP option 19. This option is commonly used to protect
# TCP sessions (e.g. BGP) where IPSEC is not available nor desirable.
# This is enabled on a per-socket basis using the TCP_MD5SIG socket option.
# This requires the use of 'device crypto' and either 'options IPSEC' or
# 'options IPSEC_SUPPORT'.
options TCP_SIGNATURE #include support for RFC 2385
# DUMMYNET enables the "dummynet" bandwidth limiter. You need IPFIREWALL
# as well. See dummynet(4) and ipfw(8) for more info. When you run
# DUMMYNET, HZ/kern.hz should be at least 1000 for adequate response.
options DUMMYNET
# The DEBUGNET option enables a basic debug/panic-time networking API. It
# is used by NETDUMP and NETGDB.
options DEBUGNET
# The NETDUMP option enables netdump(4) client support in the kernel.
# This allows a panicking kernel to transmit a kernel dump to a remote host.
options NETDUMP
# The NETGDB option enables netgdb(4) support in the kernel. This allows a
# panicking kernel to be debugged as a GDB remote over the network.
options NETGDB
#####################################################################
# FILESYSTEM OPTIONS
#
# Only the root filesystem needs to be statically compiled or preloaded
# as module; everything else will be automatically loaded at mount
# time. Some people still prefer to statically compile other
# filesystems as well.
#
# NB: The UNION filesystem was known to be buggy in the past. It is now
# being actively maintained, although there are still some issues being
# resolved.
#
# One of these is mandatory:
options FFS #Fast filesystem
options NFSCL #Network File System client
# The rest are optional:
options AUTOFS #Automounter filesystem
options CD9660 #ISO 9660 filesystem
options FDESCFS #File descriptor filesystem
options FUSEFS #FUSEFS support module
options MSDOSFS #MS DOS File System (FAT, FAT32)
options NFSLOCKD #Network Lock Manager
options NFSD #Network Filesystem Server
options KGSSAPI #Kernel GSSAPI implementation
options NULLFS #NULL filesystem
options PROCFS #Process filesystem (requires PSEUDOFS)
options PSEUDOFS #Pseudo-filesystem framework
options PSEUDOFS_TRACE #Debugging support for PSEUDOFS
options SMBFS #SMB/CIFS filesystem
options TMPFS #Efficient memory filesystem
options UDF #Universal Disk Format
options UNIONFS #Union filesystem
# The xFS_ROOT options REQUIRE the associated ``options xFS''
options NFS_ROOT #NFS usable as root device
# Soft updates is a technique for improving filesystem speed and
# making abrupt shutdown less risky.
#
options SOFTUPDATES
# Extended attributes allow additional data to be associated with files,
# and is used for ACLs, Capabilities, and MAC labels.
# See src/sys/ufs/ufs/README.extattr for more information.
options UFS_EXTATTR
options UFS_EXTATTR_AUTOSTART
# Access Control List support for UFS filesystems. The current ACL
# implementation requires extended attribute support, UFS_EXTATTR,
# for the underlying filesystem.
# See src/sys/ufs/ufs/README.acls for more information.
options UFS_ACL
# Directory hashing improves the speed of operations on very large
# directories at the expense of some memory.
options UFS_DIRHASH
# Gjournal-based UFS journaling support.
options UFS_GJOURNAL
# Make space in the kernel for a root filesystem on a md device.
# Define to the number of kilobytes to reserve for the filesystem.
# This is now optional.
# If not defined, the root filesystem passed in as the MFS_IMAGE makeoption
# will be automatically embedded in the kernel during linking. Its exact size
# will be consumed within the kernel.
# If defined, the old way of embedding the filesystem in the kernel will be
# used. That is to say MD_ROOT_SIZE KB will be allocated in the kernel and
# later, the filesystem image passed in as the MFS_IMAGE makeoption will be
# dd'd into the reserved space if it fits.
options MD_ROOT_SIZE=10
# Make the md device a potential root device, either with preloaded
# images of type mfs_root or md_root.
options MD_ROOT
# Write-protect the md root device so that it may not be mounted writeable.
options MD_ROOT_READONLY
# Allow to read MD image from external memory regions
options MD_ROOT_MEM
# Disk quotas are supported when this option is enabled.
options QUOTA #enable disk quotas
# If you are running a machine just as a fileserver for PC and MAC
# users, using SAMBA, you may consider setting this option
# and keeping all those users' directories on a filesystem that is
# mounted with the suiddir option. This gives new files the same
# ownership as the directory (similar to group). It's a security hole
# if you let these users run programs, so confine it to file-servers
# (but it'll save you lots of headaches in those cases). Root owned
# directories are exempt and X bits are cleared. The suid bit must be
# set on the directory as well; see chmod(1). PC owners can't see/set
# ownerships so they keep getting their toes trodden on. This saves
# you all the support calls as the filesystem it's used on will act as
# they expect: "It's my dir so it must be my file".
#
options SUIDDIR
# NFS options:
options NFS_MINATTRTIMO=3 # VREG attrib cache timeout in sec
options NFS_MAXATTRTIMO=60
options NFS_MINDIRATTRTIMO=30 # VDIR attrib cache timeout in sec
options NFS_MAXDIRATTRTIMO=60
options NFS_DEBUG # Enable NFS Debugging
#
# Add support for the EXT2FS filesystem of Linux fame. Be a bit
# careful with this - the ext2fs code has a tendency to lag behind
# changes and not be exercised very much, so mounting read/write could
# be dangerous (and even mounting read only could result in panics.)
#
options EXT2FS
# The system memory devices; /dev/mem, /dev/kmem
device mem
# The kernel symbol table device; /dev/ksyms
device ksyms
# Optional character code conversion support with LIBICONV.
# Each option requires their base file system and LIBICONV.
options CD9660_ICONV
options MSDOSFS_ICONV
options UDF_ICONV
#####################################################################
# POSIX P1003.1B
# Real time extensions added in the 1993 POSIX
# _KPOSIX_PRIORITY_SCHEDULING: Build in _POSIX_PRIORITY_SCHEDULING
options _KPOSIX_PRIORITY_SCHEDULING
# p1003_1b_semaphores are very experimental,
# user should be ready to assist in debugging if problems arise.
options P1003_1B_SEMAPHORES
# POSIX message queue
options P1003_1B_MQUEUE
#####################################################################
# SECURITY POLICY PARAMETERS
# Support for BSM audit
options AUDIT
# Support for Mandatory Access Control (MAC):
options MAC
options MAC_BIBA
options MAC_BSDEXTENDED
options MAC_DDB
options MAC_IFOFF
options MAC_IPACL
options MAC_LOMAC
options MAC_MLS
options MAC_NONE
options MAC_NTPD
options MAC_PARTITION
options MAC_PORTACL
options MAC_PRIORITY
options MAC_SEEOTHERUIDS
options MAC_STUB
options MAC_TEST
options MAC_VERIEXEC
options MAC_VERIEXEC_SHA1
options MAC_VERIEXEC_SHA256
options MAC_VERIEXEC_SHA384
options MAC_VERIEXEC_SHA512
device mac_veriexec_parser
# Support for Capsicum
options CAPABILITIES # fine-grained rights on file descriptors
options CAPABILITY_MODE # sandboxes with no global namespace access
#####################################################################
# CLOCK OPTIONS
# The granularity of operation is controlled by the kernel option HZ (default
# frequency of 1000 Hz or a period 1ms between calls). Virtual machine guests
# use a value of 100. Lower values may lower overhead at the expense of accuracy
# of scheduling, though the adaptive tick code reduces that overhead.
options HZ=100
# Enable support for the kernel PLL to use an external PPS signal,
# under supervision of [x]ntpd(8)
# More info in ntpd documentation: http://www.eecis.udel.edu/~ntp
options PPS_SYNC
# Enable support for generic feed-forward clocks in the kernel.
# The feed-forward clock support is an alternative to the feedback oriented
# ntpd/system clock approach, and is to be used with a feed-forward
# synchronization algorithm such as the RADclock:
# More info here: http://www.synclab.org/radclock
options FFCLOCK
#####################################################################
# SCSI DEVICES
# SCSI DEVICE CONFIGURATION
# The SCSI subsystem consists of the `base' SCSI code, a number of
# high-level SCSI device `type' drivers, and the low-level host-adapter
# device drivers. The host adapters are listed in the ISA and PCI
# device configuration sections below.
#
# It is possible to wire down your SCSI devices so that a given bus,
# target, and LUN always come on line as the same device unit. In
# earlier versions the unit numbers were assigned in the order that
# the devices were probed on the SCSI bus. This means that if you
# removed a disk drive, you may have had to rewrite your /etc/fstab
# file, and also that you had to be careful when adding a new disk
# as it may have been probed earlier and moved your device configuration
# around.
# This old behavior is maintained as the default behavior. The unit
# assignment begins with the first non-wired down unit for a device
# type. For example, if you wire a disk as "da3" then the first
# non-wired disk will be assigned da4.
# The syntax for wiring down devices is:
envvar hint.scbus.0.at="ahc0"
envvar hint.scbus.1.at="ahc1"
envvar hint.scbus.1.bus="0"
envvar hint.scbus.3.at="ahc2"
envvar hint.scbus.3.bus="0"
envvar hint.scbus.2.at="ahc2"
envvar hint.scbus.2.bus="1"
envvar hint.da.0.at="scbus0"
envvar hint.da.0.target="0"
envvar hint.da.0.unit="0"
envvar hint.da.1.at="scbus3"
envvar hint.da.1.target="1"
envvar hint.da.2.at="scbus2"
envvar hint.da.2.target="3"
envvar hint.sa.1.at="scbus1"
envvar hint.sa.1.target="6"
# "units" (SCSI logical unit number) that are not specified are
# treated as if specified as LUN 0.
# All SCSI devices allocate as many units as are required.
# The ch driver drives SCSI Media Changer ("jukebox") devices.
#
# The da driver drives SCSI Direct Access ("disk") and Optical Media
# ("WORM") devices.
#
# The sa driver drives SCSI Sequential Access ("tape") devices.
#
# The cd driver drives SCSI Read Only Direct Access ("cd") devices.
#
# The ses driver drives SCSI Environment Services ("ses") and
# SAF-TE ("SCSI Accessible Fault-Tolerant Enclosure") devices.
#
# The pt driver drives SCSI Processor devices.
#
# The sg driver provides a passthrough API that is compatible with the
# Linux SG driver. It will work in conjunction with the Linuxulator
# to run linux SG apps. It can also stand on its own and provide
# source level API compatibility for porting apps to FreeBSD.
#
# Target Mode support is provided here but also requires that a SIM
# (SCSI Host Adapter Driver) provide support as well.
#
# The targ driver provides target mode support as a Processor type device.
# It exists to give the minimal context necessary to respond to Inquiry
# commands. There is a sample user application that shows how the rest
# of the command support might be done in /usr/share/examples/scsi_target.
#
# The targbh driver provides target mode support and exists to respond
# to incoming commands that do not otherwise have a logical unit assigned
# to them.
#
# The pass driver provides a passthrough API to access the CAM subsystem.
device scbus #base SCSI code
device ch #SCSI media changers
device da #SCSI direct access devices (aka disks)
device sa #SCSI tapes
device cd #SCSI CD-ROMs
device ses #Enclosure Services (SES and SAF-TE)
device pt #SCSI processor
device targ #SCSI Target Mode Code
device targbh #SCSI Target Mode Blackhole Device
device pass #CAM passthrough driver
device sg #Linux SCSI passthrough
device ctl #CAM Target Layer
# CAM OPTIONS:
# debugging options:
# CAMDEBUG Compile in all possible debugging.
# CAM_DEBUG_COMPILE Debug levels to compile in.
# CAM_DEBUG_FLAGS Debug levels to enable on boot.
# CAM_DEBUG_BUS Limit debugging to the given bus.
# CAM_DEBUG_TARGET Limit debugging to the given target.
# CAM_DEBUG_LUN Limit debugging to the given lun.
# CAM_DEBUG_DELAY Delay in us after printing each debug line.
# CAM_IO_STATS Publish additional CAM device statics by sysctl
#
# CAM_MAX_HIGHPOWER: Maximum number of concurrent high power (start unit) cmds
# SCSI_NO_SENSE_STRINGS: When defined disables sense descriptions
# SCSI_NO_OP_STRINGS: When defined disables opcode descriptions
# SCSI_DELAY: The number of MILLISECONDS to freeze the SIM (scsi adapter)
# queue after a bus reset, and the number of milliseconds to
# freeze the device queue after a bus device reset. This
# can be changed at boot and runtime with the
# kern.cam.scsi_delay tunable/sysctl.
options CAMDEBUG
options CAM_DEBUG_COMPILE=-1
options CAM_DEBUG_FLAGS=(CAM_DEBUG_INFO|CAM_DEBUG_PROBE|CAM_DEBUG_PERIPH)
options CAM_DEBUG_BUS=-1
options CAM_DEBUG_TARGET=-1
options CAM_DEBUG_LUN=-1
options CAM_DEBUG_DELAY=1
options CAM_MAX_HIGHPOWER=4
options SCSI_NO_SENSE_STRINGS
options SCSI_NO_OP_STRINGS
options SCSI_DELAY=5000 # Be pessimistic about Joe SCSI device
options CAM_IOSCHED_DYNAMIC
options CAM_IO_STATS
options CAM_TEST_FAILURE
# Options for the CAM CDROM driver:
# CHANGER_MIN_BUSY_SECONDS: Guaranteed minimum time quantum for a changer LUN
# CHANGER_MAX_BUSY_SECONDS: Maximum time quantum per changer LUN, only
# enforced if there is I/O waiting for another LUN
# The compiled in defaults for these variables are 2 and 10 seconds,
# respectively.
#
# These can also be changed on the fly with the following sysctl variables:
# kern.cam.cd.changer.min_busy_seconds
# kern.cam.cd.changer.max_busy_seconds
#
options CHANGER_MIN_BUSY_SECONDS=2
options CHANGER_MAX_BUSY_SECONDS=10
# Options for the CAM sequential access driver:
# SA_IO_TIMEOUT: Timeout for read/write/wfm operations, in minutes
# SA_SPACE_TIMEOUT: Timeout for space operations, in minutes
# SA_REWIND_TIMEOUT: Timeout for rewind operations, in minutes
# SA_ERASE_TIMEOUT: Timeout for erase operations, in minutes
# SA_1FM_AT_EOD: Default to model which only has a default one filemark at EOT.
options SA_IO_TIMEOUT=4
options SA_SPACE_TIMEOUT=60
options SA_REWIND_TIMEOUT=(2*60)
options SA_ERASE_TIMEOUT=(4*60)
options SA_1FM_AT_EOD
# Optional timeout for the CAM processor target (pt) device
# This is specified in seconds. The default is 60 seconds.
options SCSI_PT_DEFAULT_TIMEOUT=60
# Optional enable of doing SES passthrough on other devices (e.g., disks)
#
# Normally disabled because a lot of newer SCSI disks report themselves
# as having SES capabilities, but this can then clot up attempts to build
# a topology with the SES device that's on the box these drives are in....
options SES_ENABLE_PASSTHROUGH
# iSCSI
#
# iSCSI permits access to SCSI peripherals over a network connection
# (e.g. via a TCP/IP socket)
device cfiscsi # CAM Target Layer iSCSI target frontend
device iscsi # iSCSI initiator
device iser # iSCSI Extensions for RDMA (iSER) initiator
#####################################################################
# MISCELLANEOUS DEVICES AND OPTIONS
device pty #BSD-style compatibility pseudo ttys
device nmdm #back-to-back tty devices
device md #Memory/malloc disk
device snp #Snoop device - to look at pty/vty/etc..
device ccd #Concatenated disk driver
device firmware #firmware(9) support
# Kernel side iconv library
options LIBICONV
# Size of the kernel message buffer. Should be N * pagesize.
options MSGBUF_SIZE=40960
#####################################################################
# HARDWARE BUS CONFIGURATION
#
# PCI bus & PCI options:
#
device pci
options PCI_HP # PCI-Express native HotPlug
options PCI_IOV # PCI SR-IOV support
#####################################################################
# HARDWARE DEVICE CONFIGURATION
# For ISA the required hints are listed.
# PCI, CardBus, and SD/MMC are self identifying buses, so
# no hints are needed.
#
# Mandatory devices:
#
# These options are valid for other keyboard drivers as well.
options KBD_DISABLE_KEYMAP_LOAD # refuse to load a keymap
options KBD_INSTALL_CDEV # install a CDEV entry in /dev
# Define keyboard latency (try 200/15 for a snappy interactive console)
options KBD_DELAY1=200 # define initial key delay
options KBD_DELAY2=15 # define key delay
device kbdmux # keyboard multiplexer
options KBDMUX_DFLT_KEYMAP # specify the built-in keymap
makeoptions KBDMUX_DFLT_KEYMAP=it.iso
options FB_DEBUG # Frame buffer debugging
# Enable experimental features of the syscons terminal emulator (teken).
options TEKEN_CONS25 # cons25-style terminal emulation
options TEKEN_UTF8 # UTF-8 output handling
# The vt video console driver.
device vt
options VT_ALT_TO_ESC_HACK=1 # Prepend ESC sequence to ALT keys
options VT_MAXWINDOWS=16 # Number of virtual consoles
options VT_TWOBUTTON_MOUSE # Use right mouse button to paste
# The following options set the maximum framebuffer size.
options VT_FB_MAX_HEIGHT=480
options VT_FB_MAX_WIDTH=640
# The following options will let you change the default vt terminal colors.
options TERMINAL_NORM_ATTR=(FG_GREEN|BG_BLACK)
options TERMINAL_KERN_ATTR=(FG_LIGHTRED|BG_BLACK)
#
# Optional devices:
#
#
# SCSI host adapters:
#
# aacraid: Adaptec by PMC RAID controllers, Series 6/7/8 and upcoming
# families. Container interface, CAM required.
# ahc: Adaptec 274x/284x/2910/293x/294x/394x/3950x/3960x/398X/4944/
# 19160x/29160x, aic7770/aic78xx
# ahd: Adaptec 29320/39320 Controllers.
# isp: Qlogic ISP 1020, 1040 and 1040B PCI SCSI host adapters,
# ISP 1240 Dual Ultra SCSI, ISP 1080 and 1280 (Dual) Ultra2,
# ISP 12160 Ultra3 SCSI,
# Qlogic ISP 2100 and ISP 2200 1Gb Fibre Channel host adapters.
# Qlogic ISP 2300 and ISP 2312 2Gb Fibre Channel host adapters.
# Qlogic ISP 2322 and ISP 6322 2Gb Fibre Channel host adapters.
# ispfw: Firmware module for Qlogic host adapters
# mpr: LSI-Logic MPT/Fusion Gen 3
# mps: LSI-Logic MPT/Fusion Gen 2
# mpt: LSI-Logic MPT/Fusion 53c1020 or 53c1030 Ultra4
# or FC9x9 Fibre Channel host adapters.
# sym: Symbios/Logic 53C8XX family of PCI-SCSI I/O processors:
# 53C810, 53C810A, 53C815, 53C825, 53C825A, 53C860, 53C875,
# 53C876, 53C885, 53C895, 53C895A, 53C896, 53C897, 53C1510D,
# 53C1010-33, 53C1010-66.
device aacraid
device ahc
device ahd
device isp
envvar hint.isp.0.disable="1"
envvar hint.isp.0.role="3"
envvar hint.isp.0.prefer_iomap="1"
envvar hint.isp.0.prefer_memmap="1"
envvar hint.isp.0.fwload_disable="1"
envvar hint.isp.0.ignore_nvram="1"
envvar hint.isp.0.fullduplex="1"
envvar hint.isp.0.topology="lport"
envvar hint.isp.0.topology="nport"
envvar hint.isp.0.topology="lport-only"
envvar hint.isp.0.topology="nport-only"
# we can't get u_int64_t types, nor can we get strings if it's got
# a leading 0x, hence this silly dodge.
envvar hint.isp.0.portwnn="w50000000aaaa0000"
envvar hint.isp.0.nodewnn="w50000000aaaa0001"
device ispfw
device mpr # LSI-Logic MPT-Fusion 3
device mps # LSI-Logic MPT-Fusion 2
device mpt # LSI-Logic MPT-Fusion
device sym
# The aic7xxx driver will attempt to use memory mapped I/O for all PCI
# controllers that have it configured only if this option is set. Unfortunately,
# this doesn't work on some motherboards, which prevents it from being the
# default.
options AHC_ALLOW_MEMIO
# Dump the contents of the ahc controller configuration PROM.
options AHC_DUMP_EEPROM
# Bitmap of units to enable targetmode operations.
options AHC_TMODE_ENABLE
# Compile in Aic7xxx Debugging code.
options AHC_DEBUG
# Aic7xxx driver debugging options. See sys/dev/aic7xxx/aic7xxx.h
options AHC_DEBUG_OPTS
# Print register bitfields in debug output. Adds ~128k to driver
# See ahc(4).
options AHC_REG_PRETTY_PRINT
# Compile in aic79xx debugging code.
options AHD_DEBUG
# Aic79xx driver debugging options. Adds ~215k to driver. See ahd(4).
options AHD_DEBUG_OPTS=0xFFFFFFFF
# Print human-readable register definitions when debugging
options AHD_REG_PRETTY_PRINT
# Bitmap of units to enable targetmode operations.
options AHD_TMODE_ENABLE
# Options used in dev/isp/ (Qlogic SCSI/FC driver).
#
# ISP_TARGET_MODE - enable target mode operation
#
options ISP_TARGET_MODE=1
#
# ISP_DEFAULT_ROLES - default role
# none=0
# target=1
# initiator=2
# both=3 (not supported currently)
#
# ISP_INTERNAL_TARGET (trivial internal disk target, for testing)
#
options ISP_DEFAULT_ROLES=0
#options SYM_SETUP_SCSI_DIFF #-HVD support for 825a, 875, 885
# disabled:0 (default), enabled:1
#options SYM_SETUP_PCI_PARITY #-PCI parity checking
# disabled:0, enabled:1 (default)
#options SYM_SETUP_MAX_LUN #-Number of LUNs supported
# default:8, range:[1..64]
#
# Compaq "CISS" RAID controllers (SmartRAID 5* series)
# These controllers have a SCSI-like interface, and require the
# CAM infrastructure.
#
device ciss
#
# Compaq Smart RAID, Mylex DAC960 and AMI MegaRAID controllers. Only
# one entry is needed; the code will find and configure all supported
# controllers.
#
device ida # Compaq Smart RAID
device mlx # Mylex DAC960
device mfi # LSI MegaRAID SAS
device mfip # LSI MegaRAID SAS passthrough, requires CAM
options MFI_DEBUG
device mrsas # LSI/Avago MegaRAID SAS/SATA, 6Gb/s and 12Gb/s
# NVM Express
#
# nvme: PCI-express NVM Express host controllers
# nvmf: NVM Express over Fabrics host
# nvmft: NVM Express over Fabrics CAM Target Layer frontend
# nvmf_tcp: TCP transport for NVM Express over Fabrics
# nda: CAM NVMe disk driver
# nvd: non-CAM NVMe disk driver
device nvme # PCI-express NVMe host driver
options NVME_USE_NVD=1 # Use nvd(4) instead of the CAM nda(4) driver
device nvmf # NVMeoF host driver
device nvmft # NVMeoF ctl(4) frontend
device nvmf_tcp # NVMeoF TCP transport
device nda # NVMe direct access devices (aka disks)
device nvd # expose NVMe namespaces as disks, depends on nvme
#
# Serial ATA host controllers:
#
# ahci: Advanced Host Controller Interface (AHCI) compatible
# mvs: Marvell 88SX50XX/88SX60XX/88SX70XX/SoC controllers
# siis: SiliconImage SiI3124/SiI3132/SiI3531 controllers
#
# These drivers are part of cam(4) subsystem. They supersede less featured
# ata(4) subsystem drivers, supporting same hardware.
device ahci # AHCI-compatible SATA controllers
device mvs # Marvell 88SX50XX/88SX60XX/88SX70XX/SoC SATA
device siis # SiliconImage SiI3124/SiI3132/SiI3531 SATA
device ada # ATA/SATA direct access devices (aka disks)
#
# The 'ATA' driver supports all legacy ATA/ATAPI controllers, including
# PC Card devices. You only need one "device ata" for it to find all
# PCI and PC Card ATA/ATAPI devices on modern machines.
# Alternatively, individual bus and chipset drivers may be chosen by using
# the 'atacore' driver then selecting the drivers on a per vendor basis.
# For example to build a system which only supports a VIA chipset,
# omit 'ata' and include the 'atacore', 'atapci' and 'atavia' drivers.
device ata # Legacy ATA/SATA controllers
# Modular ATA
#device atacore # Core ATA functionality
#device ataisa # ISA bus support
#device atapci # PCI bus support; only generic chipset support
# PCI ATA chipsets
#device ataacard # ACARD
#device ataacerlabs # Acer Labs Inc. (ALI)
#device ataamd # American Micro Devices (AMD)
#device ataati # ATI
#device atacenatek # Cenatek
#device atacypress # Cypress
#device atacyrix # Cyrix
#device atahighpoint # HighPoint
#device ataintel # Intel
#device ataite # Integrated Technology Inc. (ITE)
#device atajmicron # JMicron
#device atamarvell # Marvell
#device atamicron # Micron
#device atanational # National
#device atanetcell # NetCell
#device atanvidia # nVidia
#device atapromise # Promise
#device ataserverworks # ServerWorks
#device atasiliconimage # Silicon Image Inc. (SiI) (formerly CMD)
#device atasis # Silicon Integrated Systems Corp.(SiS)
#device atavia # VIA Technologies Inc.
#
# For older non-PCI, non-PnPBIOS systems, these are the hints lines to add:
envvar hint.ata.0.at="isa"
envvar hint.ata.0.port="0x1f0"
envvar hint.ata.0.irq="14"
envvar hint.ata.1.at="isa"
envvar hint.ata.1.port="0x170"
envvar hint.ata.1.irq="15"
#
# uart: generic driver for serial interfaces.
#
device uart
# Options for uart(4)
options UART_PPS_ON_CTS # Do time pulse capturing using CTS
# instead of DCD.
options UART_POLL_FREQ # Set polling rate, used when hw has
# no interrupt support (50 Hz default).
# The following hint should only be used for pure ISA devices. It is not
# needed otherwise. Use of hints is strongly discouraged.
envvar hint.uart.0.at="isa"
# The following 3 hints are used when the UART is a system device (i.e., a
# console or debug port), but only on platforms that don't have any other
# means to pass the information to the kernel. The unit number of the hint
# is only used to bundle the hints together. There is no relation to the
# unit number of the probed UART.
envvar hint.uart.0.port="0x3f8"
envvar hint.uart.0.flags="0x10"
envvar hint.uart.0.baud="115200"
# `flags' for serial drivers that support consoles, like uart(4):
# 0x10 enable console support for this unit. Other console flags
# (if applicable) are ignored unless this is set. Enabling
# console support does not make the unit the preferred console.
# Boot with -h or set boot_serial=YES in the loader.
# Currently, at most one unit can have console support; the
# first one (in config file order) with this flag set is
# preferred.
# 0x80 use this port for serial line gdb support in ddb. Also known
# as debug port.
#
# Options for serial drivers that support consoles:
options BREAK_TO_DEBUGGER # A BREAK/DBG on the console goes to
# ddb, if available.
# Solaris implements a new BREAK which is initiated by a character
# sequence CR ~ ^b which is similar to a familiar pattern used on
# Sun servers by the Remote Console. There are FreeBSD extensions:
# CR ~ ^p requests force panic and CR ~ ^r requests a clean reboot.
options ALT_BREAK_TO_DEBUGGER
# Serial Communications Controller
# Supports the Freescale/NXP QUad Integrated and Zilog Z8530 multi-channel
# communications controllers.
device scc
# PCI Universal Communications driver
# Supports various multi port PCI I/O cards.
device puc
#
# Network interfaces:
#
# MII bus support is required for many PCI Ethernet NICs,
# namely those which use MII-compliant transceivers or implement
# transceiver control interfaces that operate like an MII. Adding
# "device miibus" to the kernel config pulls in support for the generic
# miibus API, the common support for bit-bang'ing the MII and all
# of the PHY drivers, including a generic one for PHYs that aren't
# specifically handled by an individual driver. Support for specific
# PHYs may be built by adding "device mii", "device mii_bitbang" if
# needed by the NIC driver and then adding the appropriate PHY driver.
device mii # Minimal MII support
device mii_bitbang # Common module for bit-bang'ing the MII
device miibus # MII support w/ bit-bang'ing and all PHYs
device acphy # Altima Communications AC101
device amphy # AMD AM79c873 / Davicom DM910{1,2}
device atphy # Attansic/Atheros F1
device axphy # Asix Semiconductor AX88x9x
device bmtphy # Broadcom BCM5201/BCM5202 and 3Com 3c905C
device bnxt # Broadcom NetXtreme-C/NetXtreme-E
device brgphy # Broadcom BCM54xx/57xx 1000baseTX
device cgem # Cadence GEM Gigabit Ethernet
device ciphy # Cicada/Vitesse CS/VSC8xxx
device e1000phy # Marvell 88E1000 1000/100/10-BT
device gentbi # Generic 10-bit 1000BASE-{LX,SX} fiber ifaces
device icsphy # ICS ICS1889-1893
device ip1000phy # IC Plus IP1000A/IP1001
device jmphy # JMicron JMP211/JMP202
device lxtphy # Level One LXT-970
device nsgphy # NatSemi DP8361/DP83865/DP83891
device nsphy # NatSemi DP83840A
device nsphyter # NatSemi DP83843/DP83815
device pnaphy # HomePNA
device qsphy # Quality Semiconductor QS6612
device rdcphy # RDC Semiconductor R6040
device rgephy # RealTek 8169S/8110S/8211B/8211C
device rlphy # RealTek 8139
device rlswitch # RealTek 8305
device smcphy # SMSC LAN91C111
device tdkphy # TDK 89Q2120
device truephy # LSI TruePHY
device xmphy # XaQti XMAC II
# ae: Support for gigabit ethernet adapters based on the Attansic/Atheros
# L2 PCI-Express FastEthernet controllers.
# age: Support for gigabit ethernet adapters based on the Attansic/Atheros
# L1 PCI express gigabit ethernet controllers.
# alc: Support for Atheros AR8131/AR8132 PCIe ethernet controllers.
# ale: Support for Atheros AR8121/AR8113/AR8114 PCIe ethernet controllers.
# ath: Atheros a/b/g WiFi adapters (requires ath_hal and wlan)
# bce: Broadcom NetXtreme II (BCM5706/BCM5708) PCI/PCIe Gigabit Ethernet
# adapters.
# bfe: Broadcom BCM4401 Ethernet adapter.
# bge: Support for gigabit ethernet adapters based on the Broadcom
# BCM570x family of controllers, including the 3Com 3c996-T,
# the Netgear GA302T, the SysKonnect SK-9D21 and SK-9D41, and
# the embedded gigE NICs on Dell PowerEdge 2550 servers.
# bnxt: Broadcom NetXtreme-C and NetXtreme-E PCIe 10/25/50G Ethernet adapters.
# bxe: Broadcom NetXtreme II (BCM5771X/BCM578XX) PCIe 10Gb Ethernet
# adapters.
# bwi: Broadcom BCM430* and BCM431* family of wireless adapters.
# bwn: Broadcom BCM43xx family of wireless adapters.
# cas: Sun Cassini/Cassini+ and National Semiconductor DP83065 Saturn
# cxgb: Chelsio T3 based 1GbE/10GbE PCIe Ethernet adapters.
# cxgbe:Chelsio T4, T5, and T6-based 1/10/25/40/100GbE PCIe Ethernet
# adapters.
# cxgbev: Chelsio T4, T5, and T6-based PCIe Virtual Functions.
# dc: Support for PCI fast ethernet adapters based on the DEC/Intel 21143
# and various workalikes including:
# the ADMtek AL981 Comet and AN985 Centaur, the ASIX Electronics
# AX88140A and AX88141, the Davicom DM9100 and DM9102, the Lite-On
# 82c168 and 82c169 PNIC, the Lite-On/Macronix LC82C115 PNIC II
# and the Macronix 98713/98713A/98715/98715A/98725 PMAC. This driver
# replaces the old al, ax, dm, pn and mx drivers. List of brands:
# Digital DE500-BA, Kingston KNE100TX, D-Link DFE-570TX, SOHOware SFA110,
# SVEC PN102-TX, CNet Pro110B, 120A, and 120B, Compex RL100-TX,
# LinkSys LNE100TX, LNE100TX V2.0, Jaton XpressNet, Alfa Inc GFC2204,
# KNE110TX.
# em: Intel Pro/1000 Gigabit Ethernet 82542, 82543, 82544 based adapters.
# fxp: Intel EtherExpress Pro/100B
# (hint of prefer_iomap can be done to prefer I/O instead of Mem mapping)
# gem: Apple GMAC/Sun ERI/Sun GEM
# jme: JMicron JMC260 Fast Ethernet/JMC250 Gigabit Ethernet based adapters.
# le: AMD Am7900 LANCE and Am79C9xx PCnet
# lge: Support for PCI gigabit ethernet adapters based on the Level 1
# LXT1001 NetCellerator chipset. This includes the D-Link DGE-500SX,
# SMC TigerCard 1000 (SMC9462SX), and some Addtron cards.
# lio: Support for Cavium 23XX Ethernet adapters
# malo: Marvell Libertas wireless NICs.
# mwl: Marvell 88W8363 802.11n wireless NICs.
# Requires the mwl firmware module
# mwlfw: Marvell 88W8363 firmware
# msk: Support for gigabit ethernet adapters based on the Marvell/SysKonnect
# Yukon II Gigabit controllers, including 88E8021, 88E8022, 88E8061,
# 88E8062, 88E8035, 88E8036, 88E8038, 88E8050, 88E8052, 88E8053,
# 88E8055, 88E8056 and D-Link 560T/550SX.
# mlxfw: Mellanox firmware update module.
# mlx5: Mellanox ConnectX-4 and ConnectX-4 LX IB and Eth shared code module.
# mlx5en:Mellanox ConnectX-4 and ConnectX-4 LX PCIe Ethernet adapters.
# my: Myson Fast Ethernet (MTD80X, MTD89X)
# nge: Support for PCI gigabit ethernet adapters based on the National
# Semiconductor DP83820 and DP83821 chipset. This includes the
# SMC EZ Card 1000 (SMC9462TX), D-Link DGE-500T, Asante FriendlyNet
# GigaNIX 1000TA and 1000TPC, the Addtron AEG320T, the Surecom
# EP-320G-TX and the Netgear GA622T.
# oce: Emulex 10 Gbit adapters (OneConnect Ethernet)
# ral: Ralink Technology IEEE 802.11 wireless adapter
# re: RealTek 8139C+/8169/816xS/811xS/8101E PCI/PCIe Ethernet adapter
# rl: Support for PCI fast ethernet adapters based on the RealTek 8129/8139
# chipset. Note that the RealTek driver defaults to using programmed
# I/O to do register accesses because memory mapped mode seems to cause
# severe lockups on SMP hardware. This driver also supports the
# Accton EN1207D `Cheetah' adapter, which uses a chip called
# the MPX 5030/5038, which is either a RealTek in disguise or a
# RealTek workalike. Note that the D-Link DFE-530TX+ uses the RealTek
# chipset and is supported by this driver, not the 'vr' driver.
# rtwn: RealTek wireless adapters.
# rtwnfw: RealTek wireless firmware.
# sge: Silicon Integrated Systems SiS190/191 Fast/Gigabit Ethernet adapter
# sis: Support for NICs based on the Silicon Integrated Systems SiS 900,
# SiS 7016 and NS DP83815 PCI fast ethernet controller chips.
# sk: Support for the SysKonnect SK-984x series PCI gigabit ethernet NICs.
# This includes the SK-9841 and SK-9842 single port cards (single mode
# and multimode fiber) and the SK-9843 and SK-9844 dual port cards
# (also single mode and multimode).
# The driver will autodetect the number of ports on the card and
# attach each one as a separate network interface.
# ste: Sundance Technologies ST201 PCI fast ethernet controller, includes
# the D-Link DFE-550TX.
# stge: Support for gigabit ethernet adapters based on the Sundance/Tamarack
# TC9021 family of controllers, including the Sundance ST2021/ST2023,
# the Sundance/Tamarack TC9021, the D-Link DL-4000 and ASUS NX1101.
# ti: Support for PCI gigabit ethernet NICs based on the Alteon Networks
# Tigon 1 and Tigon 2 chipsets. This includes the Alteon AceNIC, the
# 3Com 3c985, the Netgear GA620 and various others. Note that you will
# probably want to bump up kern.ipc.nmbclusters a lot to use this driver.
# vr: Support for various fast ethernet adapters based on the VIA
# Technologies VT3043 `Rhine I' and VT86C100A `Rhine II' chips,
# including the D-Link DFE520TX and D-Link DFE530TX (see 'rl' for
# DFE530TX+), the Hawking Technologies PN102TX, and the AOpen/Acer ALN-320.
# vte: DM&P Vortex86 RDC R6040 Fast Ethernet
# xl: Support for the 3Com 3c900, 3c905, 3c905B and 3c905C (Fast)
# Etherlink XL cards and integrated controllers. This includes the
# integrated 3c905B-TX chips in certain Dell Optiplex and Dell
# Precision desktop machines and the integrated 3c905-TX chips
# in Dell Latitude laptop docking stations.
# Also supported: 3Com 3c980(C)-TX, 3Com 3cSOHO100-TX, 3Com 3c450-TX
# PCI Ethernet NICs that use the common MII bus controller code.
device ae # Attansic/Atheros L2 FastEthernet
device age # Attansic/Atheros L1 Gigabit Ethernet
device alc # Atheros AR8131/AR8132 Ethernet
device ale # Atheros AR8121/AR8113/AR8114 Ethernet
device bce # Broadcom BCM5706/BCM5708 Gigabit Ethernet
device bfe # Broadcom BCM440x 10/100 Ethernet
device bge # Broadcom BCM570xx Gigabit Ethernet
device cas # Sun Cassini/Cassini+ and NS DP83065 Saturn
device dc # DEC/Intel 21143 and various workalikes
device et # Agere ET1310 10/100/Gigabit Ethernet
device fxp # Intel EtherExpress PRO/100B (82557, 82558)
envvar hint.fxp.0.prefer_iomap="0"
device gem # Apple GMAC/Sun ERI/Sun GEM
device jme # JMicron JMC250 Gigabit/JMC260 Fast Ethernet
device lge # Level 1 LXT1001 gigabit Ethernet
device lio # Support for Cavium 23XX Ethernet adapters
device mlxfw # Mellanox firmware update module
device mlx5 # Shared code module between IB and Ethernet
device mlx5en # Mellanox ConnectX-4 and ConnectX-4 LX
device msk # Marvell/SysKonnect Yukon II Gigabit Ethernet
device my # Myson Fast Ethernet (MTD80X, MTD89X)
device nge # NatSemi DP83820 gigabit Ethernet
device re # RealTek 8139C+/8169/8169S/8110S
device rl # RealTek 8129/8139
device sge # Silicon Integrated Systems SiS190/191
device sis # Silicon Integrated Systems SiS 900/SiS 7016
device sk # SysKonnect SK-984x & SK-982x gigabit Ethernet
device ste # Sundance ST201 (D-Link DFE-550TX)
device stge # Sundance/Tamarack TC9021 gigabit Ethernet
device vr # VIA Rhine, Rhine II
device vte # DM&P Vortex86 RDC R6040 Fast Ethernet
device xl # 3Com 3c90x (``Boomerang'', ``Cyclone'')
# PCI/PCI-X/PCIe Ethernet NICs that use iflib infrastructure
device iflib
device em # Intel Pro/1000 Gigabit Ethernet
device ix # Intel Pro/10Gbe PCIE Ethernet
device ixv # Intel Pro/10Gbe PCIE Ethernet VF
# PCI Ethernet NICs.
device cxgb # Chelsio T3 10 Gigabit Ethernet
device cxgb_t3fw # Chelsio T3 10 Gigabit Ethernet firmware
device cxgbe # Chelsio T4-T6 1/10/25/40/100 Gigabit Ethernet
device cxgbev # Chelsio T4-T6 Virtual Functions
device le # AMD Am7900 LANCE and Am79C9xx PCnet
device mxge # Myricom Myri-10G 10GbE NIC
device oce # Emulex 10 GbE (OneConnect Ethernet)
device ti # Alteon Networks Tigon I/II gigabit Ethernet
# PCI IEEE 802.11 Wireless NICs
device ath # Atheros pci/cardbus NIC's
device ath_hal # pci/cardbus chip support
#device ath_ar5210 # AR5210 chips
#device ath_ar5211 # AR5211 chips
#device ath_ar5212 # AR5212 chips
#device ath_rf2413
#device ath_rf2417
#device ath_rf2425
#device ath_rf5111
#device ath_rf5112
#device ath_rf5413
#device ath_ar5416 # AR5416 chips
# All of the AR5212 parts have a problem when paired with the AR71xx
# CPUS. These parts have a bug that triggers a fatal bus error on the AR71xx
# only. Details of the exact nature of the bug are sketchy, but some can be
# found at https://forum.openwrt.org/viewtopic.php?pid=70060 on pages 4, 5 and
# 6. This option enables this workaround. There is a performance penalty
# for this work around, but without it things don't work at all. The DMA
# from the card usually bursts 128 bytes, but on the affected CPUs, only
# 4 are safe.
options AH_RXCFG_SDMAMW_4BYTES
#device ath_ar9160 # AR9160 chips
#device ath_ar9280 # AR9280 chips
#device ath_ar9285 # AR9285 chips
device ath_rate_sample # SampleRate tx rate control for ath
device bwi # Broadcom BCM430* BCM431*
device bwn # Broadcom BCM43xx
device malo # Marvell Libertas wireless NICs.
device mwl # Marvell 88W8363 802.11n wireless NICs.
device mwlfw
device ral # Ralink Technology RT2500 wireless NICs.
device rtwn # Realtek wireless NICs
device rtwnfw
# Use sf_buf(9) interface for jumbo buffers on ti(4) controllers.
#options TI_SF_BUF_JUMBO
# Turn on the header splitting option for the ti(4) driver firmware. This
# only works for Tigon II chips, and has no effect for Tigon I chips.
# This option requires the TI_SF_BUF_JUMBO option above.
#options TI_JUMBO_HDRSPLIT
# These two options allow manipulating the mbuf cluster size and mbuf size,
# respectively. Be very careful with NIC driver modules when changing
# these from their default values, because that can potentially cause a
# mismatch between the mbuf size assumed by the kernel and the mbuf size
# assumed by a module. The only driver that currently has the ability to
# detect a mismatch is ti(4).
options MCLSHIFT=11 # mbuf cluster shift in bits, 11 == 2KB
options MSIZE=256 # mbuf size in bytes
#
# Sound drivers
#
# sound: The generic sound driver.
#
device sound
#
# snd_*: Device-specific drivers.
#
# The flags of the device tell the device a bit more info about the
# device that normally is obtained through the PnP interface.
# bit 2..0 secondary DMA channel;
# bit 4 set if the board uses two dma channels;
# bit 15..8 board type, overrides autodetection; leave it
# zero if don't know what to put in (and you don't,
# since this is unsupported at the moment...).
#
# snd_als4000: Avance Logic ALS4000 PCI.
# snd_atiixp: ATI IXP 200/300/400 PCI.
# snd_cmi: CMedia CMI8338/CMI8738 PCI.
# snd_cs4281: Crystal Semiconductor CS4281 PCI.
# snd_csa: Crystal Semiconductor CS461x/428x PCI. (except
# 4281)
# snd_emu10k1: Creative EMU10K1 PCI and EMU10K2 (Audigy) PCI.
# snd_emu10kx: Creative SoundBlaster Live! and Audigy
# snd_envy24: VIA Envy24 and compatible, needs snd_spicds.
# snd_envy24ht: VIA Envy24HT and compatible, needs snd_spicds.
# snd_es137x: Ensoniq AudioPCI ES137x PCI.
# snd_fm801: Forte Media FM801 PCI.
# snd_hda: Intel High Definition Audio (Controller) and
# compatible.
# snd_hdspe: RME HDSPe AIO and RayDAT.
# snd_ich: Intel ICH AC'97 and some more audio controllers
# embedded in a chipset, for example nVidia
# nForce controllers.
# snd_maestro3: ESS Technology Maestro-3/Allegro PCI.
# snd_neomagic: Neomagic 256 AV/ZX PCI.
# snd_solo: ESS Solo-1x PCI.
# snd_spicds: SPI codec driver, needed by Envy24/Envy24HT drivers.
# snd_t4dwave: Trident 4DWave DX/NX PCI, Sis 7018 PCI and Acer Labs
# M5451 PCI.
# snd_uaudio: USB audio.
# snd_via8233: VIA VT8233x PCI.
# snd_via82c686: VIA VT82C686A PCI.
# snd_vibes: S3 Sonicvibes PCI.
device snd_als4000
device snd_atiixp
device snd_cmi
device snd_cs4281
device snd_csa
device snd_emu10k1
device snd_emu10kx
device snd_envy24
device snd_envy24ht
device snd_es137x
device snd_fm801
device snd_hda
device snd_hdspe
device snd_ich
device snd_maestro3
device snd_neomagic
device snd_solo
device snd_spicds
device snd_t4dwave
device snd_uaudio
device snd_via8233
device snd_via82c686
device snd_vibes
# For non-PnP sound cards:
envvar hint.pcm.0.at="isa"
envvar hint.pcm.0.irq="10"
envvar hint.pcm.0.drq="1"
envvar hint.pcm.0.flags="0x0"
envvar hint.sbc.0.at="isa"
envvar hint.sbc.0.port="0x220"
envvar hint.sbc.0.irq="5"
envvar hint.sbc.0.drq="1"
envvar hint.sbc.0.flags="0x15"
envvar hint.gusc.0.at="isa"
envvar hint.gusc.0.port="0x220"
envvar hint.gusc.0.irq="5"
envvar hint.gusc.0.drq="1"
envvar hint.gusc.0.flags="0x13"
#
# Following options are intended for debugging/testing purposes:
#
# SND_DEBUG Enable extra debugging code that includes
# sanity checking and possible increase of
# verbosity.
#
# SND_DIAGNOSTIC Similar in a spirit of INVARIANTS/DIAGNOSTIC,
# zero tolerance against inconsistencies.
#
# SND_FEEDER_MULTIFORMAT By default, only 16/32 bit feeders are compiled
# in. This options enable most feeder converters
# except for 8bit. WARNING: May bloat the kernel.
#
# SND_FEEDER_FULL_MULTIFORMAT Ditto, but includes 8bit feeders as well.
#
# SND_FEEDER_RATE_HP (feeder_rate) High precision 64bit arithmetic
# as much as possible (the default trying to
# avoid it). Possible slowdown.
#
# SND_PCM_64 (Only applicable for i386/32bit arch)
# Process 32bit samples through 64bit
# integer/arithmetic. Slight increase of dynamic
# range at a cost of possible slowdown.
#
# SND_OLDSTEREO Only 2 channels are allowed, effectively
# disabling multichannel processing.
#
options SND_DEBUG
options SND_DIAGNOSTIC
options SND_FEEDER_MULTIFORMAT
options SND_FEEDER_FULL_MULTIFORMAT
options SND_FEEDER_RATE_HP
options SND_PCM_64
options SND_OLDSTEREO
#
# Cardbus
#
# cbb: pci/CardBus bridge implementing YENTA interface
# cardbus: CardBus slots
device cbb
device cardbus
#
# MMC/SD
#
# mmc MMC/SD bus
# mmcsd MMC/SD memory card
# sdhci Generic PCI SD Host Controller
# rtsx Realtek SD card reader (RTS5209, RTS5227, ...)
device mmc
device mmcsd
device sdhci
device rtsx
#
# SMB bus
#
# System Management Bus support is provided by the 'smbus' device.
# Access to the SMBus device is via the 'smb' device (/dev/smb*),
# which is a child of the 'smbus' device.
#
# Supported devices:
# smb standard I/O through /dev/smb*
#
# Supported SMB interfaces:
# iicsmb I2C to SMB bridge with any iicbus interface
# intpm Intel PIIX4 (82371AB, 82443MX) Power Management Unit
# alpm Acer Aladdin-IV/V/Pro2 Power Management Unit
# ichsmb Intel ICH SMBus controller chips (82801AA, 82801AB, 82801BA)
# viapm VIA VT82C586B/596B/686A and VT8233 Power Management Unit
# amdpm AMD 756 Power Management Unit
# amdsmb AMD 8111 SMBus 2.0 Controller
# nfpm NVIDIA nForce Power Management Unit
# nfsmb NVIDIA nForce2/3/4 MCP SMBus 2.0 Controller
# ismt Intel SMBus 2.0 controller chips (on Atom S1200, C2000)
#
device smbus # Bus support, required for smb below.
device intpm
options ENABLE_ALART # Control alarm on Intel intpm driver
device alpm
device ichsmb
device viapm
device amdpm
device amdsmb
device nfpm
device nfsmb
device ismt
device smb
# SMBus peripheral devices
#
# jedec_dimm Asset and temperature reporting for DDR3 and DDR4 DIMMs
#
device jedec_dimm
# I2C Bus
#
# Philips i2c bus support is provided by the `iicbus' device.
#
# Supported devices:
# ic i2c network interface
# iic i2c standard io
# iicsmb i2c to smb bridge. Allow i2c i/o with smb commands.
# iicoc simple polling driver for OpenCores I2C controller
#
# Other:
# iicbb generic I2C bit-banging code (needed by lpbb)
#
device iicbus # Bus support, required for ic/iic/iicsmb below.
device iicbb # bitbang driver; implements i2c on a pair of gpio pins
device ic
device iic # userland access to i2c slave devices via ioctl(8)
device iicsmb # smb over i2c bridge
device iicoc # OpenCores I2C controller support
# I2C bus multiplexer (mux) devices
device iicmux # i2c mux core driver
device iic_gpiomux # i2c mux hardware controlled via gpio pins
device ltc430x # LTC4305 and LTC4306 i2c mux chips
# I2C peripheral devices
#
device ad7418 # Analog Devices temp and voltage sensor
device ads111x # Texas Instruments ADS101x and ADS111x ADCs
device ds1307 # Dallas DS1307 RTC and compatible
device ds13rtc # All Dallas/Maxim ds13xx chips
device ds1672 # Dallas DS1672 RTC
device ds3231 # Dallas DS3231 RTC + temperature
device fan53555 # Fairchild Semi FAN53555/SYR82x Regulator
device icee # AT24Cxxx and compatible EEPROMs
device isl12xx # Intersil ISL12xx RTC
device lm75 # LM75 compatible temperature sensor
device nxprtc # NXP RTCs: PCA/PFC212x PCA/PCF85xx
device rtc8583 # Epson RTC-8583
device s35390a # Seiko Instruments S-35390A RTC
device sy8106a # Silergy Corp. SY8106A buck regulator
# Parallel-Port Bus
#
# Parallel port bus support is provided by the `ppbus' device.
# Multiple devices may be attached to the parallel port, devices
# are automatically probed and attached when found.
#
# Supported devices:
# lpt Parallel Printer
# plip Parallel network interface
# ppi General-purpose I/O ("Geek Port") + IEEE1284 I/O
# pps Pulse per second Timing Interface
# lpbb Philips official parallel port I2C bit-banging interface
# pcfclock Parallel port clock driver.
#
# Supported interfaces:
# ppc ISA-bus parallel port interfaces.
#
options PPC_PROBE_CHIPSET # Enable chipset specific detection
# (see flags in ppc(4))
options DEBUG_1284 # IEEE1284 signaling protocol debug
options PERIPH_1284 # Makes your computer act as an IEEE1284
# compliant peripheral
options DONTPROBE_1284 # Avoid boot detection of PnP parallel devices
options LPT_DEBUG # Printer driver debug
options PPC_DEBUG # Parallel chipset level debug
options PLIP_DEBUG # Parallel network IP interface debug
options PCFCLOCK_VERBOSE # Verbose pcfclock driver
options PCFCLOCK_MAX_RETRIES=5 # Maximum read tries (default 10)
device ppc
envvar hint.ppc.0.at="isa"
envvar hint.ppc.0.irq="7"
device ppbus
device lpt
device plip
device ppi
device pps
device lpbb
device pcfclock
# General Purpose I/O pins
device dwgpio # Synopsys DesignWare APB GPIO Controller
device gpio # gpio interfaces and bus support
device gpiobacklight # sysctl control of gpio-based backlight
device gpioiic # i2c via gpio bitbang
device gpiokeys # kbd(4) glue for gpio-based key input
device gpioled # led(4) gpio glue
device gpiopower # event handler for gpio-based powerdown
device gpiopps # Pulse per second input from gpio pin
device gpioregulator # extres/regulator glue for gpio pin
device gpiospi # SPI via gpio bitbang
device gpioths # 1-wire temp/humidity sensor on gpio pin
# Pulse width modulation
device pwmbus # pwm interface and bus support
device pwmc # userland control access to pwm outputs
#
# Etherswitch framework and drivers
#
# etherswitch The etherswitch(4) framework
# miiproxy Proxy device for miibus(4) functionality
#
# Switch hardware support:
# arswitch Atheros switches
# ip17x IC+ 17x family switches
# rtl8366r Realtek RTL8366 switches
# ukswitch Multi-PHY switches
#
device etherswitch
device miiproxy
device arswitch
device ip17x
device rtl8366rb
device ukswitch
# Kernel BOOTP support
options BOOTP # Use BOOTP to obtain IP address/hostname
# Requires NFSCL and NFS_ROOT
options BOOTP_NFSROOT # NFS mount root filesystem using BOOTP info
options BOOTP_NFSV3 # Use NFS v3 to NFS mount root
options BOOTP_COMPAT # Workaround for broken bootp daemons.
options BOOTP_WIRED_TO=fxp0 # Use interface fxp0 for BOOTP
options BOOTP_BLOCKSIZE=8192 # Override NFS block size
#
# Enable software watchdog routines, even if hardware watchdog is present.
# By default, software watchdog timer is enabled only if no hardware watchdog
# is present.
#
options SW_WATCHDOG
#
# Add the software deadlock resolver thread.
#
options DEADLKRES
#
# Disable swapping of stack pages. This option removes all
# code which actually performs swapping, so it's not possible to turn
# it back on at run-time.
#
# This is sometimes usable for systems which don't have any swap space
# (see also sysctl "vm.disable_swapspace_pageouts")
#
#options NO_SWAPPING
# Set the number of sf_bufs to allocate. sf_bufs are virtual buffers
# for sendfile(2) that are used to map file VM pages, and normally
# default to a quantity that is roughly 16*MAXUSERS+512. You would
# typically want about 4 of these for each simultaneous file send.
#
options NSFBUFS=1024
#
# Enable extra debugging code for locks. This stores the filename and
# line of whatever acquired the lock in the lock itself, and changes a
# number of function calls to pass around the relevant data. This is
# not at all useful unless you are debugging lock code. Note that
# modules should be recompiled as this option modifies KBI.
#
options DEBUG_LOCKS
#
# VirtIO support
#
# The virtio entry provides a generic bus for use by the device drivers.
# It must be combined with an interface that communicates with the host.
# Multiple such interfaces are defined by the VirtIO specification
# including PCI and MMIO.
#
device virtio # Generic VirtIO bus (required)
device virtio_mmio # VirtIO MMIO Interface
device virtio_pci # VirtIO PCI Interface
device vtnet # VirtIO Ethernet device
device virtio_balloon # VirtIO Memory Balloon device
device virtio_blk # VirtIO Block device
device virtio_console # VirtIO Console device
device virtio_gpu # VirtIO GPU device
device virtio_random # VirtIO Entropy device
device virtio_scmi # VirtIO SCSI device
device virtio_scsi # VirtIO SCMI device
#####################################################################
# HID support
device hid # Generic HID support
options HID_DEBUG # enable debug msgs
device hidbus # HID bus
device hidmap # HID to evdev mapping
device hidraw # Raw access driver
options HIDRAW_MAKE_UHID_ALIAS # install /dev/uhid alias
device hconf # Multitouch configuration TLC
device hcons # Consumer controls
device hgame # Generic game controllers
device hkbd # HID keyboard
device hms # HID mouse
device hmt # HID multitouch (MS-compatible)
device hpen # Generic pen driver
device hsctrl # System controls
device ps4dshock # Sony PS4 DualShock 4 gamepad driver
device xb360gp # XBox 360 gamepad driver
#####################################################################
# USB support
# UHCI controller
device uhci
# OHCI controller
device ohci
# EHCI controller
device ehci
# XHCI controller
device xhci
# SL811 Controller
#device slhci
# General USB code (mandatory for USB)
device usb
#
# USB Double Bulk Pipe devices
device udbp
# USB temperature meter
device ugold
# USB LED
device uled
# Human Interface Device (anything with buttons and dials)
device uhid
# USB keyboard
device ukbd
# USB printer
device ulpt
# USB mass storage driver (Requires scbus and da)
device umass
# USB mass storage driver for device-side mode
device usfs
# USB support for Belkin F5U109 and Magic Control Technology serial adapters
device umct
# USB modem support
device umodem
# USB mouse
device ums
# USB touchpad(s)
device atp
device wsp
# eGalax USB touch screen
device uep
# Diamond Rio 500 MP3 player
device urio
# HID-over-USB driver
device usbhid
#
# USB serial support
device ucom
# USB support for 3G modem cards by Option, Novatel, Huawei and Sierra
device u3g
# USB support for Technologies ARK3116 based serial adapters
device uark
# USB support for Belkin F5U103 and compatible serial adapters
device ubsa
# USB support for serial adapters based on the FT8U100AX and FT8U232AM
device uftdi
# USB support for some Windows CE based serial communication.
device uipaq
# USB support for Prolific PL-2303 serial adapters
device uplcom
# USB support for Silicon Laboratories CP2101/CP2102 based USB serial adapters
device uslcom
# USB Visor and Palm devices
device uvisor
# USB serial support for DDI pocket's PHS
device uvscom
#
# USB ethernet support
device uether
# ADMtek USB ethernet. Supports the LinkSys USB100TX,
# the Billionton USB100, the Melco LU-ATX, the D-Link DSB-650TX
# and the SMC 2202USB. Also works with the ADMtek AN986 Pegasus
# eval board.
device aue
# ASIX Electronics AX88172 USB 2.0 ethernet driver. Used in the
# LinkSys USB200M and various other adapters.
device axe
# ASIX Electronics AX88178A/AX88179 USB 2.0/3.0 gigabit ethernet driver.
device axge
#
# Devices which communicate using Ethernet over USB, particularly
# Communication Device Class (CDC) Ethernet specification. Supports
# Sharp Zaurus PDAs, some DOCSIS cable modems and so on.
device cdce
#
# CATC USB-EL1201A USB ethernet. Supports the CATC Netmate
# and Netmate II, and the Belkin F5U111.
device cue
#
# Kawasaki LSI ethernet. Supports the LinkSys USB10T,
# Entrega USB-NET-E45, Peracom Ethernet Adapter, the
# 3Com 3c19250, the ADS Technologies USB-10BT, the ATen UC10T,
# the Netgear EA101, the D-Link DSB-650, the SMC 2102USB
# and 2104USB, and the Corega USB-T.
device kue
#
# RealTek RTL8150 USB to fast ethernet. Supports the Melco LUA-KTX
# and the GREEN HOUSE GH-USB100B.
device rue
#
# Davicom DM9601E USB to fast ethernet. Supports the Corega FEther USB-TXC.
device udav
#
# RealTek RTL8152/RTL8153 USB Ethernet driver
device ure
#
# Moschip MCS7730/MCS7840 USB to fast ethernet. Supports the Sitecom LN030.
device mos
#
# HSxPA devices from Option N.V
device uhso
# Realtek RTL8188SU/RTL8191SU/RTL8192SU wireless driver
device rsu
#
# Ralink Technology RT2501USB/RT2601USB wireless driver
device rum
# Ralink Technology RT2700U/RT2800U/RT3000U wireless driver
device run
#
# Atheros AR5523 wireless driver
device uath
#
# Conexant/Intersil PrismGT wireless driver
device upgt
#
# Ralink Technology RT2500USB wireless driver
device ural
#
# RNDIS USB ethernet driver
device urndis
# Realtek RTL8187B/L wireless driver
device urtw
#
# ZyDas ZD1211/ZD1211B wireless driver
device zyd
#
# Sierra USB wireless driver
device usie
#
# debugging options for the USB subsystem
#
options USB_DEBUG
options U3G_DEBUG
# options for ukbd:
options UKBD_DFLT_KEYMAP # specify the built-in keymap
makeoptions UKBD_DFLT_KEYMAP=jp.106
# options for uplcom:
options UPLCOM_INTR_INTERVAL=100 # interrupt pipe interval
# in milliseconds
# options for uvscom:
options UVSCOM_DEFAULT_OPKTSIZE=8 # default output packet size
options UVSCOM_INTR_INTERVAL=100 # interrupt pipe interval
# in milliseconds
#####################################################################
# FireWire support
device firewire # FireWire bus code
device sbp # SCSI over Firewire (Requires scbus and da)
device sbp_targ # SBP-2 Target mode (Requires scbus and targ)
device fwe # Ethernet over FireWire (non-standard!)
device fwip # IP over FireWire (RFC2734 and RFC3146)
#####################################################################
# dcons support (Dumb Console Device)
device dcons # dumb console driver
device dcons_crom # FireWire attachment
options DCONS_BUF_SIZE=16384 # buffer size
options DCONS_POLL_HZ=100 # polling rate
options DCONS_FORCE_CONSOLE=0 # force to be the primary console
options DCONS_FORCE_GDB=1 # force to be the gdb device
#####################################################################
# crypto subsystem
#
# This is a port of the OpenBSD crypto framework. Include this when
# configuring IPSEC and when you have a h/w crypto device to accelerate
# user applications that link to OpenSSL.
#
# Drivers are ports from OpenBSD with some simple enhancements that have
# been fed back to OpenBSD.
device crypto # core crypto support
# Only install the cryptodev device if you are running tests, or know
# specifically why you need it. In most cases, it is not needed and
# will make things slower.
device cryptodev # /dev/crypto for access to h/w
device rndtest # FIPS 140-2 entropy tester
device ccr # Chelsio T6
device hifn # Hifn 7951, 7781, etc.
options HIFN_DEBUG # enable debugging support: hw.hifn.debug
options HIFN_RNDTEST # enable rndtest support
device safe # SafeNet 1141
options SAFE_DEBUG # enable debugging support: hw.safe.debug
options SAFE_RNDTEST # enable rndtest support
#####################################################################
#
# Embedded system options:
#
# An embedded system might want to run something other than init.
options INIT_PATH=/sbin/init:/rescue/init
# Debug options
options BUS_DEBUG # enable newbus debugging
options DEBUG_VFS_LOCKS # enable VFS lock debugging
options SOCKBUF_DEBUG # enable sockbuf last record/mb tail checking
options IFMEDIA_DEBUG # enable debugging in net/if_media.c
#
# Verbose SYSINIT
#
# Make the SYSINIT process performed by mi_startup() verbose. This is very
# useful when porting to a new architecture. If DDB is also enabled, this
# will print function names instead of addresses. If defined with a value
# of zero, the verbose code is compiled-in but disabled by default, and can
# be enabled with the debug.verbose_sysinit=1 tunable.
options VERBOSE_SYSINIT
#####################################################################
# SYSV IPC KERNEL PARAMETERS
#
# Maximum number of System V semaphores that can be used on the system at
# one time.
options SEMMNI=11
# Total number of semaphores system wide
options SEMMNS=61
# Total number of undo structures in system
options SEMMNU=31
# Maximum number of System V semaphores that can be used by a single process
# at one time.
options SEMMSL=61
# Maximum number of operations that can be outstanding on a single System V
# semaphore at one time.
options SEMOPM=101
# Maximum number of undo operations that can be outstanding on a single
# System V semaphore at one time.
options SEMUME=11
# Maximum number of shared memory pages system wide.
options SHMALL=1025
# Maximum size, in bytes, of a single System V shared memory region.
options SHMMAX=(SHMMAXPGS*PAGE_SIZE+1)
options SHMMAXPGS=1025
# Minimum size, in bytes, of a single System V shared memory region.
options SHMMIN=2
# Maximum number of shared memory regions that can be used on the system
# at one time.
options SHMMNI=33
# Maximum number of System V shared memory regions that can be attached to
# a single process at one time.
options SHMSEG=9
# Set the amount of time (in seconds) the system will wait before
# rebooting automatically when a kernel panic occurs. If set to (-1),
# the system will wait indefinitely until a key is pressed on the
# console.
options PANIC_REBOOT_WAIT_TIME=16
# Attempt to bypass the buffer cache and put data directly into the
# userland buffer for read operation when O_DIRECT flag is set on the
# file. Both offset and length of the read operation must be
# multiples of the physical media sector size.
#
options DIRECTIO
# Specify a lower limit for the number of swap I/O buffers. They are
# (among other things) used when bypassing the buffer cache due to
# DIRECTIO kernel option enabled and O_DIRECT flag set on file.
#
options NSWBUF_MIN=120
#####################################################################
# More undocumented options for linting.
# Note that documenting these is not considered an affront.
options CAM_DEBUG_DELAY
options DEBUG
# Kernel filelock debugging.
options LOCKF_DEBUG
# System V compatible message queues
# Please note that the values provided here are used to test kernel
# building. The defaults in the sources provide almost the same numbers.
# MSGSSZ must be a power of 2 between 8 and 1024.
options MSGMNB=2049 # Max number of chars in queue
options MSGMNI=41 # Max number of message queue identifiers
options MSGSEG=2049 # Max number of message segments
options MSGSSZ=16 # Size of a message segment
options MSGTQL=41 # Max number of messages in system
options NBUF=512 # Number of buffer headers
options SC_DEBUG_LEVEL=5 # Syscons debug level
options SC_RENDER_DEBUG # syscons rendering debugging
options VFS_BIO_DEBUG # VFS buffer I/O debugging
options KSTACK_MAX_PAGES=32 # Maximum pages to give the kernel stack
options KSTACK_USAGE_PROF
# Adaptec Array Controller driver options
options AAC_DEBUG # Debugging levels:
# 0 - quiet, only emit warnings
# 1 - noisy, emit major function
# points and things done
# 2 - extremely noisy, emit trace
# items in loops, etc.
# Resource Accounting
options RACCT
# Resource Limits
options RCTL
# Yet more undocumented options for linting.
options MAXFILES=999
# Random number generator
# Alternative algorithm.
#options RANDOM_FENESTRASX
# Allow the CSPRNG algorithm to be loaded as a module.
#options RANDOM_LOADABLE
# Select this to allow high-rate but potentially expensive
# harvesting of Slab-Allocator entropy. In very high-rate
# situations the value of doing this is dubious at best.
options RANDOM_ENABLE_UMA # slab allocator
# Select this to allow high-rate but potentially expensive
# harvesting of the m_next pointer in the mbuf. Note that
# the m_next pointer is NULL except when receiving > 4K
# jumbo frames or sustained bursts by way of LRO. Thus in
# the common case it is stirring zero in to the entropy
# pool. In cases where it is not NULL it is pointing to one
# of a small (in the thousands to 10s of thousands) number
# of 256 byte aligned mbufs. Hence it is, even in the best
# case, a poor source of entropy. And in the absence of actual
# runtime analysis of entropy collection may mislead the user in
# to believe that substantially more entropy is being collected
# than in fact is - leading to a different class of security
# risk. In high packet rate situations ethernet entropy
# collection is also very expensive, possibly leading to as
# much as a 50% drop in packets received.
# This option is present to maintain backwards compatibility
# if desired, however it cannot be recommended for use in any
# environment.
options RANDOM_ENABLE_ETHER # ether_input
# Module to enable execution of application via emulators like QEMU
options IMGACT_BINMISC
# zlib I/O stream support
# This enables support for compressed core dumps.
options GZIO
# zstd support
# This enables support for Zstd compressed core dumps, GEOM_UZIP images,
# and is required by zfs if statically linked.
options ZSTDIO
# BHND(4) drivers
options BHND_LOGLEVEL # Logging threshold level
# evdev interface
device evdev # input event device support
options EVDEV_SUPPORT # evdev support in legacy drivers
options EVDEV_DEBUG # enable event debug msgs
device uinput # install /dev/uinput cdev
options UINPUT_DEBUG # enable uinput debug msgs
# Encrypted kernel crash dumps.
options EKCD
# Serial Peripheral Interface (SPI) support.
device spibus # Bus support.
device at45d # DataFlash driver
device cqspi #
device mx25l # SPIFlash driver
device n25q #
device spigen # Generic access to SPI devices from userland.
# Enable legacy /dev/spigenN name aliases for /dev/spigenX.Y devices.
options SPIGEN_LEGACY_CDEVNAME # legacy device names for spigen
# Compression supports.
device zlib # gzip/zlib compression/decompression library
device xz # xz_embedded LZMA de-compression library
# Kernel support for stats(3).
options STATS
diff --git a/sys/conf/options b/sys/conf/options
index fcab21ad7e78..40c24799aa0c 100644
--- a/sys/conf/options
+++ b/sys/conf/options
@@ -1,1025 +1,1026 @@
#
# On the handling of kernel options
#
# All kernel options should be listed in NOTES, with suitable
# descriptions. Negative options (options that make some code not
# compile) should be commented out; LINT (generated from NOTES) should
# compile as much code as possible. Try to structure option-using
# code so that a single option only switch code on, or only switch
# code off, to make it possible to have a full compile-test. If
# necessary, you can check for COMPILING_LINT to get maximum code
# coverage.
#
# All new options shall also be listed in either "conf/options" or
# "conf/options.<machine>". Options that affect a single source-file
# <xxx>.[c|s] should be directed into "opt_<xxx>.h", while options
# that affect multiple files should either go in "opt_global.h" if
# this is a kernel-wide option (used just about everywhere), or in
# "opt_<option-name-in-lower-case>.h" if it affects only some files.
# Note that the effect of listing only an option without a
# header-file-name in conf/options (and cousins) is that the last
# convention is followed.
#
# This handling scheme is not yet fully implemented.
#
#
# Format of this file:
# Option name filename
#
# If filename is missing, the default is
# opt_<name-of-option-in-lower-case>.h
AAC_DEBUG opt_aac.h
AACRAID_DEBUG opt_aacraid.h
AHC_ALLOW_MEMIO opt_aic7xxx.h
AHC_TMODE_ENABLE opt_aic7xxx.h
AHC_DUMP_EEPROM opt_aic7xxx.h
AHC_DEBUG opt_aic7xxx.h
AHC_DEBUG_OPTS opt_aic7xxx.h
AHC_REG_PRETTY_PRINT opt_aic7xxx.h
AHD_DEBUG opt_aic79xx.h
AHD_DEBUG_OPTS opt_aic79xx.h
AHD_TMODE_ENABLE opt_aic79xx.h
AHD_REG_PRETTY_PRINT opt_aic79xx.h
# Debugging options.
ALT_BREAK_TO_DEBUGGER opt_kdb.h
BREAK_TO_DEBUGGER opt_kdb.h
BUF_TRACKING opt_global.h
DDB
DDB_BUFR_SIZE opt_ddb.h
DDB_CAPTURE_DEFAULTBUFSIZE opt_ddb.h
DDB_CAPTURE_MAXBUFSIZE opt_ddb.h
DDB_CTF opt_ddb.h
DDB_NUMSYM opt_ddb.h
EARLY_PRINTF opt_global.h
FULL_BUF_TRACKING opt_global.h
GDB
KDB opt_global.h
KDB_TRACE opt_kdb.h
KDB_UNATTENDED opt_kdb.h
KLD_DEBUG opt_kld.h
NUM_CORE_FILES opt_global.h
QUEUE_MACRO_DEBUG_TRACE opt_global.h
QUEUE_MACRO_DEBUG_TRASH opt_global.h
SYSCTL_DEBUG opt_sysctl.h
TEXTDUMP_PREFERRED opt_ddb.h
TEXTDUMP_VERBOSE opt_ddb.h
TSLOG opt_global.h
TSLOG_PAGEZERO opt_global.h
TSLOGSIZE opt_global.h
# Miscellaneous options.
ALQ
ALTERA_SDCARD_FAST_SIM opt_altera_sdcard.h
ATSE_CFI_HACK opt_cfi.h
AUDIT opt_global.h
BOOTHOWTO opt_global.h
BOOTVERBOSE opt_global.h
CALLOUT_PROFILING
CAPABILITIES opt_capsicum.h
CAPABILITY_MODE opt_capsicum.h
CC_CDG opt_global.h
CC_CHD opt_global.h
CC_CUBIC opt_global.h
CC_DEFAULT opt_cc.h
CC_DCTCP opt_global.h
CC_HD opt_global.h
CC_HTCP opt_global.h
CC_NEWRENO opt_global.h
CC_VEGAS opt_global.h
COMPAT_43 opt_global.h
COMPAT_43TTY opt_global.h
COMPAT_FREEBSD4 opt_global.h
COMPAT_FREEBSD5 opt_global.h
COMPAT_FREEBSD6 opt_global.h
COMPAT_FREEBSD7 opt_global.h
COMPAT_FREEBSD9 opt_global.h
COMPAT_FREEBSD10 opt_global.h
COMPAT_FREEBSD11 opt_global.h
COMPAT_FREEBSD12 opt_global.h
COMPAT_FREEBSD13 opt_global.h
COMPAT_FREEBSD14 opt_global.h
COMPAT_LINUXKPI opt_dontuse.h
COMPILING_LINT opt_global.h
CY_PCI_FASTINTR
DEADLKRES opt_watchdog.h
EXPERIMENTAL opt_global.h
DIRECTIO
FILEMON opt_dontuse.h
FFCLOCK
FULL_PREEMPTION opt_sched.h
GZIO opt_gzio.h
IMGACT_BINMISC opt_dontuse.h
IPI_PREEMPTION opt_sched.h
GEOM_BDE opt_geom.h
GEOM_CACHE opt_geom.h
GEOM_CONCAT opt_geom.h
GEOM_ELI opt_geom.h
GEOM_GATE opt_geom.h
GEOM_JOURNAL opt_geom.h
GEOM_LABEL opt_geom.h
GEOM_LABEL_GPT opt_geom.h
GEOM_LINUX_LVM opt_geom.h
GEOM_MAP opt_geom.h
GEOM_MIRROR opt_geom.h
GEOM_MOUNTVER opt_geom.h
GEOM_MULTIPATH opt_geom.h
GEOM_NOP opt_geom.h
GEOM_PART_APM opt_geom.h
GEOM_PART_BSD opt_geom.h
GEOM_PART_BSD64 opt_geom.h
GEOM_PART_EBR opt_geom.h
GEOM_PART_GPT opt_geom.h
GEOM_PART_LDM opt_geom.h
GEOM_PART_MBR opt_geom.h
GEOM_RAID opt_geom.h
GEOM_RAID3 opt_geom.h
GEOM_SHSEC opt_geom.h
GEOM_STRIPE opt_geom.h
GEOM_UZIP opt_geom.h
GEOM_UZIP_DEBUG opt_geom.h
GEOM_VINUM opt_geom.h
GEOM_VIRSTOR opt_geom.h
GEOM_ZERO opt_geom.h
IFLIB opt_iflib.h
KDTRACE_HOOKS opt_global.h
KDTRACE_FRAME opt_kdtrace.h
KDTRACE_NO_MIB_SDT opt_global.h
KN_HASHSIZE opt_kqueue.h
KSTACK_MAX_PAGES
KSTACK_PAGES
KSTACK_USAGE_PROF
KTRACE
KTRACE_REQUEST_POOL opt_ktrace.h
LIBICONV
MAC opt_global.h
MAC_BIBA opt_dontuse.h
MAC_BSDEXTENDED opt_dontuse.h
MAC_DDB opt_dontuse.h
MAC_DEBUG opt_mac.h
MAC_IFOFF opt_dontuse.h
MAC_IPACL opt_dontuse.h
MAC_LOMAC opt_dontuse.h
MAC_MLS opt_dontuse.h
MAC_NONE opt_dontuse.h
MAC_NTPD opt_dontuse.h
MAC_PARTITION opt_dontuse.h
MAC_PORTACL opt_dontuse.h
MAC_PRIORITY opt_dontuse.h
MAC_SEEOTHERUIDS opt_dontuse.h
MAC_STATIC opt_mac.h
MAC_STUB opt_dontuse.h
MAC_TEST opt_dontuse.h
MAC_GRANTBYLABEL opt_dontuse.h
MAC_VERIEXEC opt_dontuse.h
MAC_VERIEXEC_DEBUG opt_mac.h
MAC_VERIEXEC_SHA1 opt_dontuse.h
MAC_VERIEXEC_SHA256 opt_dontuse.h
MAC_VERIEXEC_SHA384 opt_dontuse.h
MAC_VERIEXEC_SHA512 opt_dontuse.h
MD_ROOT opt_md.h
MD_ROOT_FSTYPE opt_md.h
MD_ROOT_READONLY opt_md.h
MD_ROOT_SIZE opt_md.h
MD_ROOT_MEM opt_md.h
MFI_DEBUG opt_mfi.h
MFI_DECODE_LOG opt_mfi.h
MPROF_BUFFERS opt_mprof.h
MPROF_HASH_SIZE opt_mprof.h
NEW_PCIB opt_global.h
NO_ADAPTIVE_MUTEXES opt_adaptive_mutexes.h
NO_ADAPTIVE_RWLOCKS
NO_ADAPTIVE_SX
NO_OBSOLETE_CODE opt_global.h
NO_SYSCTL_DESCR opt_global.h
NSWBUF_MIN opt_param.h
MBUF_PACKET_ZONE_DISABLE opt_global.h
PANIC_REBOOT_WAIT_TIME opt_panic.h
PCI_HP opt_pci.h
PCI_IOV opt_global.h
PPC_DEBUG opt_ppc.h
PPC_PROBE_CHIPSET opt_ppc.h
PPS_SYNC opt_ntp.h
PREEMPTION opt_sched.h
QUOTA
SCHED_4BSD opt_sched.h
SCHED_STATS opt_sched.h
SCHED_ULE opt_sched.h
SLEEPQUEUE_PROFILING
SLHCI_DEBUG opt_slhci.h
+SOCKET_HHOOK opt_global.h
STACK opt_stack.h
SUIDDIR
MSGMNB opt_sysvipc.h
MSGMNI opt_sysvipc.h
MSGSEG opt_sysvipc.h
MSGSSZ opt_sysvipc.h
MSGTQL opt_sysvipc.h
SEMMNI opt_sysvipc.h
SEMMNS opt_sysvipc.h
SEMMNU opt_sysvipc.h
SEMMSL opt_sysvipc.h
SEMOPM opt_sysvipc.h
SEMUME opt_sysvipc.h
SHMALL opt_sysvipc.h
SHMMAX opt_sysvipc.h
SHMMAXPGS opt_sysvipc.h
SHMMIN opt_sysvipc.h
SHMMNI opt_sysvipc.h
SHMSEG opt_sysvipc.h
SYSVMSG opt_sysvipc.h
SYSVSEM opt_sysvipc.h
SYSVSHM opt_sysvipc.h
SW_WATCHDOG opt_watchdog.h
TCPHPTS
TCP_REQUEST_TRK opt_global.h
TCP_ACCOUNTING opt_global.h
TCP_BBR opt_inet.h
TCP_RACK opt_inet.h
#
# TCP SaD Detection is an experimental Sack attack Detection (SaD)
# algorithm that uses "normal" behaviour with SACK's to detect
# a possible attack. It is strictly experimental at this point.
#
TCP_SAD_DETECTION opt_inet.h
TURNSTILE_PROFILING
UMTX_PROFILING
UMTX_CHAINS opt_global.h
VERBOSE_SYSINIT
ZSTDIO opt_zstdio.h
# Sanitizers
COVERAGE opt_global.h
KASAN opt_global.h
KCOV
KCSAN opt_global.h
KMSAN opt_global.h
KUBSAN opt_global.h
# POSIX kernel options
P1003_1B_MQUEUE opt_posix.h
P1003_1B_SEMAPHORES opt_posix.h
_KPOSIX_PRIORITY_SCHEDULING opt_posix.h
# Do we want the config file compiled into the kernel?
INCLUDE_CONFIG_FILE opt_config.h
# Options for static filesystems. These should only be used at config
# time, since the corresponding lkms cannot work if there are any static
# dependencies. Unusability is enforced by hiding the defines for the
# options in a never-included header.
AUTOFS opt_dontuse.h
CD9660 opt_dontuse.h
EXT2FS opt_dontuse.h
FDESCFS opt_dontuse.h
FFS opt_dontuse.h
FUSEFS opt_dontuse.h
MSDOSFS opt_dontuse.h
NULLFS opt_dontuse.h
PROCFS opt_dontuse.h
PSEUDOFS opt_dontuse.h
SMBFS opt_dontuse.h
TARFS opt_dontuse.h
TMPFS opt_dontuse.h
UDF opt_dontuse.h
UNIONFS opt_dontuse.h
ZFS opt_dontuse.h
# Pseudofs debugging
PSEUDOFS_TRACE opt_pseudofs.h
# Tarfs debugging
TARFS_DEBUG opt_tarfs.h
# In-kernel GSS-API
KGSSAPI opt_kgssapi.h
KGSSAPI_DEBUG opt_kgssapi.h
# These static filesystems have one slightly bogus static dependency in
# sys/i386/i386/autoconf.c. If any of these filesystems are
# statically compiled into the kernel, code for mounting them as root
# filesystems will be enabled - but look below.
# NFSCL - client
# NFSD - server
NFSCL opt_nfs.h
NFSD opt_nfs.h
# filesystems and libiconv bridge
CD9660_ICONV opt_dontuse.h
MSDOSFS_ICONV opt_dontuse.h
UDF_ICONV opt_dontuse.h
# If you are following the conditions in the copyright,
# you can enable soft-updates which will speed up a lot of thigs
# and make the system safer from crashes at the same time.
# otherwise a STUB module will be compiled in.
SOFTUPDATES opt_ffs.h
# On small, embedded systems, it can be useful to turn off support for
# snapshots. It saves about 30-40k for a feature that would be lightly
# used, if it is used at all.
NO_FFS_SNAPSHOT opt_ffs.h
# Enabling this option turns on support for Access Control Lists in UFS,
# which can be used to support high security configurations. Depends on
# UFS_EXTATTR.
UFS_ACL opt_ufs.h
# Enabling this option turns on support for extended attributes in UFS-based
# filesystems, which can be used to support high security configurations
# as well as new filesystem features.
UFS_EXTATTR opt_ufs.h
UFS_EXTATTR_AUTOSTART opt_ufs.h
# Enable fast hash lookups for large directories on UFS-based filesystems.
UFS_DIRHASH opt_ufs.h
# Enable gjournal-based UFS journal.
UFS_GJOURNAL opt_ufs.h
# The below sentence is not in English, and neither is this one.
# We plan to remove the static dependences above, with a
# <filesystem>_ROOT option to control if it usable as root. This list
# allows these options to be present in config files already (though
# they won't make any difference yet).
NFS_ROOT opt_nfsroot.h
# SMB/CIFS requester
NETSMB opt_netsmb.h
# Enable debugnet(4) networking support.
DEBUGNET opt_global.h
# Enable netdump(4) client support.
NETDUMP opt_global.h
# Enable netgdb(4) support.
NETGDB opt_global.h
# Options used only in subr_param.c.
HZ opt_param.h
MAXFILES opt_param.h
NBUF opt_param.h
NSFBUFS opt_param.h
VM_BCACHE_SIZE_MAX opt_param.h
VM_SWZONE_SIZE_MAX opt_param.h
MAXUSERS
DFLDSIZ opt_param.h
MAXDSIZ opt_param.h
MAXSSIZ opt_param.h
# Generic SCSI options.
CAM_MAX_HIGHPOWER opt_cam.h
CAMDEBUG opt_cam.h
CAM_DEBUG_COMPILE opt_cam.h
CAM_DEBUG_DELAY opt_cam.h
CAM_DEBUG_BUS opt_cam.h
CAM_DEBUG_TARGET opt_cam.h
CAM_DEBUG_LUN opt_cam.h
CAM_DEBUG_FLAGS opt_cam.h
CAM_BOOT_DELAY opt_cam.h
CAM_IOSCHED_DYNAMIC opt_cam.h
CAM_IO_STATS opt_cam.h
CAM_TEST_FAILURE opt_cam.h
SCSI_DELAY opt_scsi.h
SCSI_NO_SENSE_STRINGS opt_scsi.h
SCSI_NO_OP_STRINGS opt_scsi.h
# Options used only in cam/ata/ata_da.c
ATA_STATIC_ID opt_ada.h
# Options used only in cam/scsi/scsi_cd.c
CHANGER_MIN_BUSY_SECONDS opt_cd.h
CHANGER_MAX_BUSY_SECONDS opt_cd.h
# Options used only in cam/scsi/scsi_da.c
DA_TRACK_REFS opt_da.h
# Options used only in cam/scsi/scsi_sa.c.
SA_IO_TIMEOUT opt_sa.h
SA_SPACE_TIMEOUT opt_sa.h
SA_REWIND_TIMEOUT opt_sa.h
SA_ERASE_TIMEOUT opt_sa.h
SA_1FM_AT_EOD opt_sa.h
# Options used only in cam/scsi/scsi_pt.c
SCSI_PT_DEFAULT_TIMEOUT opt_pt.h
# Options used only in cam/scsi/scsi_ses.c
SES_ENABLE_PASSTHROUGH opt_ses.h
# Options used in dev/sym/ (Symbios SCSI driver).
SYM_SETUP_SCSI_DIFF opt_sym.h #-HVD support for 825a, 875, 885
# disabled:0 (default), enabled:1
SYM_SETUP_PCI_PARITY opt_sym.h #-PCI parity checking
# disabled:0, enabled:1 (default)
SYM_SETUP_MAX_LUN opt_sym.h #-Number of LUNs supported
# default:8, range:[1..64]
# Options used only in dev/isp/*
ISP_TARGET_MODE opt_isp.h
ISP_FW_CRASH_DUMP opt_isp.h
ISP_DEFAULT_ROLES opt_isp.h
ISP_INTERNAL_TARGET opt_isp.h
ISP_FCTAPE_OFF opt_isp.h
# Net stuff.
ACCEPT_FILTER_DATA
ACCEPT_FILTER_DNS
ACCEPT_FILTER_HTTP
ACCEPT_FILTER_TLS
ALTQ opt_global.h
ALTQ_CBQ opt_altq.h
ALTQ_CDNR opt_altq.h
ALTQ_CODEL opt_altq.h
ALTQ_DEBUG opt_altq.h
ALTQ_HFSC opt_altq.h
ALTQ_FAIRQ opt_altq.h
ALTQ_NOPCC opt_altq.h
ALTQ_PRIQ opt_altq.h
ALTQ_RED opt_altq.h
ALTQ_RIO opt_altq.h
BOOTP opt_bootp.h
BOOTP_BLOCKSIZE opt_bootp.h
BOOTP_COMPAT opt_bootp.h
BOOTP_NFSROOT opt_bootp.h
BOOTP_NFSV3 opt_bootp.h
BOOTP_WIRED_TO opt_bootp.h
DEVICE_POLLING
DUMMYNET opt_ipdn.h
RATELIMIT opt_ratelimit.h
RATELIMIT_DEBUG opt_ratelimit.h
INET opt_inet.h
INET6 opt_inet6.h
STATS opt_global.h
IPDIVERT
IPFILTER opt_ipfilter.h
IPFILTER_DEFAULT_BLOCK opt_ipfilter.h
IPFILTER_LOG opt_ipfilter.h
IPFILTER_LOOKUP opt_ipfilter.h
IPFIREWALL opt_ipfw.h
IPFIREWALL_DEFAULT_TO_ACCEPT opt_ipfw.h
IPFIREWALL_NAT opt_ipfw.h
IPFIREWALL_NAT64 opt_ipfw.h
IPFIREWALL_NPTV6 opt_ipfw.h
IPFIREWALL_VERBOSE opt_ipfw.h
IPFIREWALL_VERBOSE_LIMIT opt_ipfw.h
IPFIREWALL_PMOD opt_ipfw.h
IPSEC opt_ipsec.h
IPSEC_DEBUG opt_ipsec.h
IPSEC_SUPPORT opt_ipsec.h
IPSTEALTH
KERN_TLS
KRPC
LIBALIAS
LIBMCHAIN
MBUF_PROFILING
MBUF_STRESS_TEST
MROUTING opt_mrouting.h
NFSLOCKD
NETLINK opt_global.h
PF_DEFAULT_TO_DROP opt_pf.h
ROUTE_MPATH opt_route.h
ROUTETABLES opt_route.h
FIB_ALGO opt_route.h
RSS opt_rss.h
SLIP_IFF_OPTS opt_slip.h
TCPPCAP opt_global.h
SIFTR
TCP_BLACKBOX opt_global.h
TCP_HHOOK opt_global.h
TCP_OFFLOAD opt_inet.h # Enable code to dispatch TCP offloading
TCP_RFC7413 opt_inet.h
TCP_RFC7413_MAX_KEYS opt_inet.h
TCP_RFC7413_MAX_PSKS opt_inet.h
TCP_SIGNATURE opt_ipsec.h
VLAN_ARRAY opt_vlan.h
XDR
XBONEHACK
#
# SCTP
#
SCTP opt_sctp.h
SCTP_SUPPORT opt_sctp.h
SCTP_DEBUG opt_sctp.h # Enable debug printfs
SCTP_LOCK_LOGGING opt_sctp.h # Log to KTR lock activity
SCTP_MBUF_LOGGING opt_sctp.h # Log to KTR general mbuf aloc/free
SCTP_MBCNT_LOGGING opt_sctp.h # Log to KTR mbcnt activity
SCTP_PACKET_LOGGING opt_sctp.h # Log to a packet buffer last N packets
SCTP_LTRACE_CHUNKS opt_sctp.h # Log to KTR chunks processed
SCTP_LTRACE_ERRORS opt_sctp.h # Log to KTR error returns.
SCTP_USE_PERCPU_STAT opt_sctp.h # Use per cpu stats.
SCTP_MCORE_INPUT opt_sctp.h # Have multiple input threads for input mbufs
SCTP_LOCAL_TRACE_BUF opt_sctp.h # Use tracebuffer exported via sysctl
SCTP_DETAILED_STR_STATS opt_sctp.h # Use per PR-SCTP policy stream stats
#
#
#
# Netgraph(4). Use option NETGRAPH to enable the base netgraph code.
# Each netgraph node type can be either be compiled into the kernel
# or loaded dynamically. To get the former, include the corresponding
# option below. Each type has its own man page, e.g. ng_async(4).
NETGRAPH
NETGRAPH_DEBUG opt_netgraph.h
NETGRAPH_ASYNC opt_netgraph.h
NETGRAPH_BLUETOOTH opt_netgraph.h
NETGRAPH_BLUETOOTH_BT3C opt_netgraph.h
NETGRAPH_BLUETOOTH_H4 opt_netgraph.h
NETGRAPH_BLUETOOTH_HCI opt_netgraph.h
NETGRAPH_BLUETOOTH_L2CAP opt_netgraph.h
NETGRAPH_BLUETOOTH_SOCKET opt_netgraph.h
NETGRAPH_BLUETOOTH_UBT opt_netgraph.h
NETGRAPH_BLUETOOTH_UBTBCMFW opt_netgraph.h
NETGRAPH_BPF opt_netgraph.h
NETGRAPH_BRIDGE opt_netgraph.h
NETGRAPH_CAR opt_netgraph.h
NETGRAPH_CHECKSUM opt_netgraph.h
NETGRAPH_CISCO opt_netgraph.h
NETGRAPH_DEFLATE opt_netgraph.h
NETGRAPH_DEVICE opt_netgraph.h
NETGRAPH_ECHO opt_netgraph.h
NETGRAPH_EIFACE opt_netgraph.h
NETGRAPH_ETHER opt_netgraph.h
NETGRAPH_ETHER_ECHO opt_netgraph.h
NETGRAPH_FEC opt_netgraph.h
NETGRAPH_FRAME_RELAY opt_netgraph.h
NETGRAPH_GIF opt_netgraph.h
NETGRAPH_GIF_DEMUX opt_netgraph.h
NETGRAPH_HOLE opt_netgraph.h
NETGRAPH_IFACE opt_netgraph.h
NETGRAPH_IP_INPUT opt_netgraph.h
NETGRAPH_IPFW opt_netgraph.h
NETGRAPH_KSOCKET opt_netgraph.h
NETGRAPH_L2TP opt_netgraph.h
NETGRAPH_LMI opt_netgraph.h
NETGRAPH_MPPC_COMPRESSION opt_netgraph.h
NETGRAPH_MPPC_ENCRYPTION opt_netgraph.h
NETGRAPH_NAT opt_netgraph.h
NETGRAPH_NETFLOW opt_netgraph.h
NETGRAPH_ONE2MANY opt_netgraph.h
NETGRAPH_PATCH opt_netgraph.h
NETGRAPH_PIPE opt_netgraph.h
NETGRAPH_PPP opt_netgraph.h
NETGRAPH_PPPOE opt_netgraph.h
NETGRAPH_PPTPGRE opt_netgraph.h
NETGRAPH_PRED1 opt_netgraph.h
NETGRAPH_RFC1490 opt_netgraph.h
NETGRAPH_SOCKET opt_netgraph.h
NETGRAPH_SPLIT opt_netgraph.h
NETGRAPH_SPPP opt_netgraph.h
NETGRAPH_TAG opt_netgraph.h
NETGRAPH_TCPMSS opt_netgraph.h
NETGRAPH_TEE opt_netgraph.h
NETGRAPH_TTY opt_netgraph.h
NETGRAPH_UI opt_netgraph.h
NETGRAPH_VJC opt_netgraph.h
NETGRAPH_VLAN opt_netgraph.h
# DRM options
DRM_DEBUG opt_drm.h
TI_SF_BUF_JUMBO opt_ti.h
TI_JUMBO_HDRSPLIT opt_ti.h
# Misc debug flags. Most of these should probably be replaced with
# 'DEBUG', and then let people recompile just the interesting modules
# with 'make CC="cc -DDEBUG"'.
DEBUG_1284 opt_ppb_1284.h
LPT_DEBUG opt_lpt.h
PLIP_DEBUG opt_plip.h
LOCKF_DEBUG opt_debug_lockf.h
SI_DEBUG opt_debug_si.h
IFMEDIA_DEBUG opt_ifmedia.h
# Fb options
FB_DEBUG opt_fb.h
# ppbus related options
PERIPH_1284 opt_ppb_1284.h
DONTPROBE_1284 opt_ppb_1284.h
# smbus related options
ENABLE_ALART opt_intpm.h
# These cause changes all over the kernel
BLKDEV_IOSIZE opt_global.h
BURN_BRIDGES opt_global.h
DEBUG opt_global.h
DEBUG_LOCKS opt_global.h
DEBUG_VFS_LOCKS opt_global.h
DFLTPHYS opt_global.h
DIAGNOSTIC opt_global.h
INVARIANT_SUPPORT opt_global.h
INVARIANTS opt_global.h
KASSERT_PANIC_OPTIONAL opt_global.h
MAXCPU opt_global.h
MAXMEMDOM opt_global.h
MAXPHYS opt_maxphys.h
MCLSHIFT opt_global.h
MUTEX_NOINLINE opt_global.h
LOCK_PROFILING opt_global.h
MSIZE opt_global.h
REGRESSION opt_global.h
RWLOCK_NOINLINE opt_global.h
SX_NOINLINE opt_global.h
VFS_BIO_DEBUG opt_global.h
# These are VM related options
VM_KMEM_SIZE opt_vm.h
VM_KMEM_SIZE_SCALE opt_vm.h
VM_KMEM_SIZE_MAX opt_vm.h
VM_NRESERVLEVEL opt_vm.h
VM_LEVEL_0_ORDER opt_vm.h
NO_SWAPPING opt_vm.h
MALLOC_MAKE_FAILURES opt_vm.h
MALLOC_PROFILE opt_vm.h
MALLOC_DEBUG_MAXZONES opt_vm.h
# The MemGuard replacement allocator used for tamper-after-free detection
DEBUG_MEMGUARD opt_vm.h
# The RedZone malloc(9) protection
DEBUG_REDZONE opt_vm.h
# Standard SMP options
EARLY_AP_STARTUP opt_global.h
SMP opt_global.h
NUMA opt_global.h
# Size of the kernel message buffer
MSGBUF_SIZE opt_msgbuf.h
# NFS options
NFS_MINATTRTIMO opt_nfs.h
NFS_MAXATTRTIMO opt_nfs.h
NFS_MINDIRATTRTIMO opt_nfs.h
NFS_MAXDIRATTRTIMO opt_nfs.h
NFS_DEBUG opt_nfs.h
# TMPFS options
TMPFS_PAGES_MINRESERVED opt_tmpfs.h
# Options for uart(4)
UART_PPS_ON_CTS opt_uart.h
UART_POLL_FREQ opt_uart.h
UART_DEV_TOLERANCE_PCT opt_uart.h
# options for bus/device framework
BUS_DEBUG opt_bus.h
# options for USB support
USB_DEBUG opt_usb.h
USB_HOST_ALIGN opt_usb.h
USB_REQ_DEBUG opt_usb.h
USB_TEMPLATE opt_usb.h
USB_VERBOSE opt_usb.h
USB_DMA_SINGLE_ALLOC opt_usb.h
USB_EHCI_BIG_ENDIAN_DESC opt_usb.h
U3G_DEBUG opt_u3g.h
UKBD_DFLT_KEYMAP opt_ukbd.h
UPLCOM_INTR_INTERVAL opt_uplcom.h
UVSCOM_DEFAULT_OPKTSIZE opt_uvscom.h
UVSCOM_INTR_INTERVAL opt_uvscom.h
# options for the Realtek rtwn driver
RTWN_DEBUG opt_rtwn.h
RTWN_WITHOUT_UCODE opt_rtwn.h
# Embedded system options
INIT_PATH
ROOTDEVNAME
FDC_DEBUG opt_fdc.h
PCFCLOCK_VERBOSE opt_pcfclock.h
PCFCLOCK_MAX_RETRIES opt_pcfclock.h
KTR opt_global.h
KTR_ALQ opt_ktr.h
KTR_MASK opt_ktr.h
KTR_CPUMASK opt_ktr.h
KTR_COMPILE opt_global.h
KTR_BOOT_ENTRIES opt_global.h
KTR_ENTRIES opt_global.h
KTR_VERBOSE opt_ktr.h
WITNESS opt_global.h
WITNESS_KDB opt_witness.h
WITNESS_NO_VNODE opt_witness.h
WITNESS_SKIPSPIN opt_witness.h
WITNESS_COUNT opt_witness.h
OPENSOLARIS_WITNESS opt_global.h
EPOCH_TRACE opt_global.h
# options for ACPI support
ACPI_DEBUG opt_acpi.h
ACPI_MAX_TASKS opt_acpi.h
ACPI_MAX_THREADS opt_acpi.h
DEV_ACPI opt_acpi.h
ACPI_EARLY_EPYC_WAR opt_acpi.h
# options for IOMMU support
IOMMU opt_iommu.h
# ISA support
DEV_ISA opt_isa.h
ISAPNP opt_dontuse.h
# various 'device presence' options.
DEV_BPF opt_bpf.h
DEV_CARP opt_carp.h
DEV_NETMAP opt_global.h
DEV_PCI opt_pci.h
DEV_PF opt_pf.h
DEV_PFLOG opt_pf.h
DEV_PFSYNC opt_pf.h
DEV_SPLASH opt_splash.h
DEV_VLAN opt_vlan.h
# bce driver
BCE_DEBUG opt_bce.h
BCE_NVRAM_WRITE_SUPPORT opt_bce.h
SOCKBUF_DEBUG opt_global.h
# options for hifn driver
HIFN_DEBUG opt_hifn.h
HIFN_RNDTEST opt_hifn.h
# options for safenet driver
SAFE_DEBUG opt_safe.h
SAFE_NO_RNG opt_safe.h
SAFE_RNDTEST opt_safe.h
# syscons/vt options
MAXCONS opt_syscons.h
SC_ALT_MOUSE_IMAGE opt_syscons.h
SC_CUT_SPACES2TABS opt_syscons.h
SC_CUT_SEPCHARS opt_syscons.h
SC_DEBUG_LEVEL opt_syscons.h
SC_DFLT_FONT opt_syscons.h
SC_DFLT_TERM opt_syscons.h
SC_DISABLE_KDBKEY opt_syscons.h
SC_DISABLE_REBOOT opt_syscons.h
SC_HISTORY_SIZE opt_syscons.h
SC_KERNEL_CONS_ATTR opt_syscons.h
SC_KERNEL_CONS_ATTRS opt_syscons.h
SC_KERNEL_CONS_REV_ATTR opt_syscons.h
SC_MOUSE_CHAR opt_syscons.h
SC_NO_CUTPASTE opt_syscons.h
SC_NO_FONT_LOADING opt_syscons.h
SC_NO_HISTORY opt_syscons.h
SC_NO_MODE_CHANGE opt_syscons.h
SC_NO_SUSPEND_VTYSWITCH opt_syscons.h
SC_NO_SYSMOUSE opt_syscons.h
SC_NO_TERM_DUMB opt_syscons.h
SC_NO_TERM_SC opt_syscons.h
SC_NO_TERM_TEKEN opt_syscons.h
SC_NORM_ATTR opt_syscons.h
SC_NORM_REV_ATTR opt_syscons.h
SC_PIXEL_MODE opt_syscons.h
SC_RENDER_DEBUG opt_syscons.h
SC_TWOBUTTON_MOUSE opt_syscons.h
VT_ALT_TO_ESC_HACK opt_syscons.h
VT_FB_MAX_WIDTH opt_syscons.h
VT_FB_MAX_HEIGHT opt_syscons.h
VT_MAXWINDOWS opt_syscons.h
VT_TWOBUTTON_MOUSE opt_syscons.h
DEV_SC opt_syscons.h
DEV_VT opt_syscons.h
# teken terminal emulator options
TEKEN_CONS25 opt_teken.h
TEKEN_UTF8 opt_teken.h
TERMINAL_KERN_ATTR opt_teken.h
TERMINAL_NORM_ATTR opt_teken.h
# options for printf
PRINTF_BUFR_SIZE opt_printf.h
BOOT_TAG opt_printf.h
BOOT_TAG_SZ opt_printf.h
# kbd options
KBD_DISABLE_KEYMAP_LOAD opt_kbd.h
KBD_INSTALL_CDEV opt_kbd.h
KBD_MAXRETRY opt_kbd.h
KBD_MAXWAIT opt_kbd.h
KBD_RESETDELAY opt_kbd.h
KBD_DELAY1 opt_kbd.h
KBD_DELAY2 opt_kbd.h
KBDIO_DEBUG opt_kbd.h
KBDMUX_DFLT_KEYMAP opt_kbdmux.h
# options for the Atheros driver
ATH_DEBUG opt_ath.h
ATH_TXBUF opt_ath.h
ATH_RXBUF opt_ath.h
ATH_DIAGAPI opt_ath.h
ATH_TX99_DIAG opt_ath.h
ATH_ENABLE_DFS opt_ath.h
ATH_EEPROM_FIRMWARE opt_ath.h
ATH_ENABLE_RADIOTAP_VENDOR_EXT opt_ath.h
ATH_DEBUG_ALQ opt_ath.h
ATH_KTR_INTR_DEBUG opt_ath.h
AH_DEBUG opt_ah.h
AH_ASSERT opt_ah.h
AH_DEBUG_ALQ opt_ah.h
AH_REGOPS_FUNC opt_ah.h
AH_WRITE_REGDOMAIN opt_ah.h
AH_DEBUG_COUNTRY opt_ah.h
AH_WRITE_EEPROM opt_ah.h
AH_PRIVATE_DIAG opt_ah.h
AH_NEED_DESC_SWAP opt_ah.h
AH_USE_INIPDGAIN opt_ah.h
AH_MAXCHAN opt_ah.h
AH_RXCFG_SDMAMW_4BYTES opt_ah.h
AH_INTERRUPT_DEBUGGING opt_ah.h
# AR5416 and later interrupt mitigation
# XXX do not use this for AR9130
AH_AR5416_INTERRUPT_MITIGATION opt_ah.h
# options for the Altera mSGDMA driver (altera_msgdma)
ALTERA_MSGDMA_DESC_STD opt_altera_msgdma.h
ALTERA_MSGDMA_DESC_EXT opt_altera_msgdma.h
ALTERA_MSGDMA_DESC_PF_STD opt_altera_msgdma.h
ALTERA_MSGDMA_DESC_PF_EXT opt_altera_msgdma.h
# options for the Broadcom BCM43xx driver (bwi)
BWI_DEBUG opt_bwi.h
BWI_DEBUG_VERBOSE opt_bwi.h
# options for the Brodacom BCM43xx driver (bwn)
BWN_DEBUG opt_bwn.h
BWN_GPL_PHY opt_bwn.h
BWN_USE_SIBA opt_bwn.h
# Options for the SIBA driver
SIBA_DEBUG opt_siba.h
# options for the Marvell 8335 wireless driver
MALO_DEBUG opt_malo.h
MALO_TXBUF opt_malo.h
MALO_RXBUF opt_malo.h
# options for the Marvell wireless driver
MWL_DEBUG opt_mwl.h
MWL_TXBUF opt_mwl.h
MWL_RXBUF opt_mwl.h
MWL_DIAGAPI opt_mwl.h
MWL_AGGR_SIZE opt_mwl.h
MWL_TX_NODROP opt_mwl.h
# Options for the Marvell NETA driver
MVNETA_MULTIQUEUE opt_mvneta.h
MVNETA_KTR opt_mvneta.h
# Options for the Intel 802.11ac wireless driver
IWM_DEBUG opt_iwm.h
# Options for the Intel 802.11n wireless driver
IWN_DEBUG opt_iwn.h
# Options for the Intel 3945ABG wireless driver
WPI_DEBUG opt_wpi.h
# dcons options
DCONS_BUF_SIZE opt_dcons.h
DCONS_POLL_HZ opt_dcons.h
DCONS_FORCE_CONSOLE opt_dcons.h
DCONS_FORCE_GDB opt_dcons.h
# HWPMC options
HWPMC_DEBUG opt_global.h
HWPMC_HOOKS
# 802.11 support layer
IEEE80211_DEBUG opt_wlan.h
IEEE80211_DEBUG_REFCNT opt_wlan.h
IEEE80211_SUPPORT_MESH opt_wlan.h
IEEE80211_SUPPORT_SUPERG opt_wlan.h
IEEE80211_SUPPORT_TDMA opt_wlan.h
IEEE80211_ALQ opt_wlan.h
IEEE80211_DFS_DEBUG opt_wlan.h
# 802.11 TDMA support
TDMA_SLOTLEN_DEFAULT opt_tdma.h
TDMA_SLOTCNT_DEFAULT opt_tdma.h
TDMA_BINTVAL_DEFAULT opt_tdma.h
TDMA_TXRATE_11B_DEFAULT opt_tdma.h
TDMA_TXRATE_11G_DEFAULT opt_tdma.h
TDMA_TXRATE_11A_DEFAULT opt_tdma.h
TDMA_TXRATE_TURBO_DEFAULT opt_tdma.h
TDMA_TXRATE_HALF_DEFAULT opt_tdma.h
TDMA_TXRATE_QUARTER_DEFAULT opt_tdma.h
TDMA_TXRATE_11NA_DEFAULT opt_tdma.h
TDMA_TXRATE_11NG_DEFAULT opt_tdma.h
# VideoMode
PICKMODE_DEBUG opt_videomode.h
# Network stack virtualization options
VIMAGE opt_global.h
VNET_DEBUG opt_global.h
# Common Flash Interface (CFI) options
CFI_SUPPORT_STRATAFLASH opt_cfi.h
CFI_ARMEDANDDANGEROUS opt_cfi.h
CFI_HARDWAREBYTESWAP opt_cfi.h
# Sound options
SND_DEBUG opt_snd.h
SND_DIAGNOSTIC opt_snd.h
SND_FEEDER_MULTIFORMAT opt_snd.h
SND_FEEDER_FULL_MULTIFORMAT opt_snd.h
SND_FEEDER_RATE_HP opt_snd.h
SND_PCM_64 opt_snd.h
SND_OLDSTEREO opt_snd.h
X86BIOS
# Flattened device tree options
FDT opt_platform.h
FDT_DTB_STATIC opt_platform.h
# OFED Infiniband stack
OFED opt_ofed.h
OFED_DEBUG_INIT opt_ofed.h
SDP opt_ofed.h
SDP_DEBUG opt_ofed.h
IPOIB opt_ofed.h
IPOIB_DEBUG opt_ofed.h
IPOIB_CM opt_ofed.h
# Resource Accounting
RACCT opt_global.h
RACCT_DEFAULT_TO_DISABLED opt_global.h
# Resource Limits
RCTL opt_global.h
# Random number generator(s)
# Alternative RNG algorithm.
RANDOM_FENESTRASX opt_global.h
# With this, no entropy processor is loaded, but the entropy
# harvesting infrastructure is present. This means an entropy
# processor may be loaded as a module.
RANDOM_LOADABLE opt_global.h
# This turns on high-rate and potentially expensive harvesting in
# the uma slab allocator.
RANDOM_ENABLE_UMA opt_global.h
RANDOM_ENABLE_ETHER opt_global.h
# This options turns TPM into entropy source.
TPM_HARVEST opt_tpm.h
# BHND(4) driver
BHND_LOGLEVEL opt_global.h
# GPIO and child devices
GPIO_SPI_DEBUG opt_gpio.h
# SPI devices
SPIGEN_LEGACY_CDEVNAME opt_spi.h
# etherswitch(4) driver
RTL8366_SOFT_RESET opt_etherswitch.h
# evdev protocol support
EVDEV_SUPPORT opt_evdev.h
EVDEV_DEBUG opt_evdev.h
UINPUT_DEBUG opt_evdev.h
# Hyper-V network driver
HN_DEBUG opt_hn.h
# CAM-based MMC stack
MMCCAM
# Encrypted kernel crash dumps
EKCD opt_ekcd.h
# NVME options
NVME_USE_NVD opt_nvme.h
# amdsbwd options
AMDSBWD_DEBUG opt_amdsbwd.h
# gcov support
GCOV opt_global.h
LINDEBUGFS
# options for HID support
HID_DEBUG opt_hid.h
IICHID_DEBUG opt_hid.h
IICHID_SAMPLING opt_hid.h
HKBD_DFLT_KEYMAP opt_hkbd.h
HIDRAW_MAKE_UHID_ALIAS opt_hid.h
# kenv options
# The early kernel environment (loader environment, config(8)-provided static)
# is typically cleared after the dynamic environment comes up to ensure that
# we're not inadvertently holding on to 'secret' values in these stale envs.
# This option is insecure except in controlled environments where the static
# environment's contents are known to be safe.
PRESERVE_EARLY_KENV opt_global.h
diff --git a/sys/kern/uipc_socket.c b/sys/kern/uipc_socket.c
index f0b36fc5595e..f0d22b3887a4 100644
--- a/sys/kern/uipc_socket.c
+++ b/sys/kern/uipc_socket.c
@@ -1,4300 +1,4322 @@
/*-
* SPDX-License-Identifier: BSD-3-Clause
*
* Copyright (c) 1982, 1986, 1988, 1990, 1993
* The Regents of the University of California.
* Copyright (c) 2004 The FreeBSD Foundation
* Copyright (c) 2004-2008 Robert N. M. Watson
* 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.
*
* 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.
*/
/*
* Comments on the socket life cycle:
*
* soalloc() sets of socket layer state for a socket, called only by
* socreate() and sonewconn(). Socket layer private.
*
* sodealloc() tears down socket layer state for a socket, called only by
* sofree() and sonewconn(). Socket layer private.
*
* pru_attach() associates protocol layer state with an allocated socket;
* called only once, may fail, aborting socket allocation. This is called
* from socreate() and sonewconn(). Socket layer private.
*
* pru_detach() disassociates protocol layer state from an attached socket,
* and will be called exactly once for sockets in which pru_attach() has
* been successfully called. If pru_attach() returned an error,
* pru_detach() will not be called. Socket layer private.
*
* pru_abort() and pru_close() notify the protocol layer that the last
* consumer of a socket is starting to tear down the socket, and that the
* protocol should terminate the connection. Historically, pru_abort() also
* detached protocol state from the socket state, but this is no longer the
* case.
*
* socreate() creates a socket and attaches protocol state. This is a public
* interface that may be used by socket layer consumers to create new
* sockets.
*
* sonewconn() creates a socket and attaches protocol state. This is a
* public interface that may be used by protocols to create new sockets when
* a new connection is received and will be available for accept() on a
* listen socket.
*
* soclose() destroys a socket after possibly waiting for it to disconnect.
* This is a public interface that socket consumers should use to close and
* release a socket when done with it.
*
* soabort() destroys a socket without waiting for it to disconnect (used
* only for incoming connections that are already partially or fully
* connected). This is used internally by the socket layer when clearing
* listen socket queues (due to overflow or close on the listen socket), but
* is also a public interface protocols may use to abort connections in
* their incomplete listen queues should they no longer be required. Sockets
* placed in completed connection listen queues should not be aborted for
* reasons described in the comment above the soclose() implementation. This
* is not a general purpose close routine, and except in the specific
* circumstances described here, should not be used.
*
* sofree() will free a socket and its protocol state if all references on
* the socket have been released, and is the public interface to attempt to
* free a socket when a reference is removed. This is a socket layer private
* interface.
*
* NOTE: In addition to socreate() and soclose(), which provide a single
* socket reference to the consumer to be managed as required, there are two
* calls to explicitly manage socket references, soref(), and sorele().
* Currently, these are generally required only when transitioning a socket
* from a listen queue to a file descriptor, in order to prevent garbage
* collection of the socket at an untimely moment. For a number of reasons,
* these interfaces are not preferred, and should be avoided.
*
* NOTE: With regard to VNETs the general rule is that callers do not set
* curvnet. Exceptions to this rule include soabort(), sodisconnect(),
* sofree(), sorele(), sonewconn() and sorflush(), which are usually called
* from a pre-set VNET context. sopoll() currently does not need a VNET
* context to be set.
*/
#include <sys/cdefs.h>
#include "opt_inet.h"
#include "opt_inet6.h"
#include "opt_kern_tls.h"
#include "opt_ktrace.h"
#include "opt_sctp.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/capsicum.h>
#include <sys/fcntl.h>
#include <sys/limits.h>
#include <sys/lock.h>
#include <sys/mac.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/mutex.h>
#include <sys/domain.h>
#include <sys/file.h> /* for struct knote */
#include <sys/hhook.h>
#include <sys/kernel.h>
#include <sys/khelp.h>
#include <sys/ktls.h>
#include <sys/event.h>
#include <sys/eventhandler.h>
#include <sys/poll.h>
#include <sys/proc.h>
#include <sys/protosw.h>
#include <sys/sbuf.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/resourcevar.h>
#include <net/route.h>
#include <sys/signalvar.h>
#include <sys/stat.h>
#include <sys/sx.h>
#include <sys/sysctl.h>
#include <sys/taskqueue.h>
#include <sys/uio.h>
#include <sys/un.h>
#include <sys/unpcb.h>
#include <sys/jail.h>
#include <sys/syslog.h>
#include <netinet/in.h>
#include <netinet/in_pcb.h>
#include <netinet/tcp.h>
#include <net/vnet.h>
#include <security/mac/mac_framework.h>
#include <vm/uma.h>
#ifdef COMPAT_FREEBSD32
#include <sys/mount.h>
#include <sys/sysent.h>
#include <compat/freebsd32/freebsd32.h>
#endif
static int soreceive_rcvoob(struct socket *so, struct uio *uio,
int flags);
static void so_rdknl_lock(void *);
static void so_rdknl_unlock(void *);
static void so_rdknl_assert_lock(void *, int);
static void so_wrknl_lock(void *);
static void so_wrknl_unlock(void *);
static void so_wrknl_assert_lock(void *, int);
static void filt_sordetach(struct knote *kn);
static int filt_soread(struct knote *kn, long hint);
static void filt_sowdetach(struct knote *kn);
static int filt_sowrite(struct knote *kn, long hint);
static int filt_soempty(struct knote *kn, long hint);
-static int inline hhook_run_socket(struct socket *so, void *hctx, int32_t h_id);
fo_kqfilter_t soo_kqfilter;
static struct filterops soread_filtops = {
.f_isfd = 1,
.f_detach = filt_sordetach,
.f_event = filt_soread,
};
static struct filterops sowrite_filtops = {
.f_isfd = 1,
.f_detach = filt_sowdetach,
.f_event = filt_sowrite,
};
static struct filterops soempty_filtops = {
.f_isfd = 1,
.f_detach = filt_sowdetach,
.f_event = filt_soempty,
};
so_gen_t so_gencnt; /* generation count for sockets */
MALLOC_DEFINE(M_SONAME, "soname", "socket name");
MALLOC_DEFINE(M_PCB, "pcb", "protocol control block");
#define VNET_SO_ASSERT(so) \
VNET_ASSERT(curvnet != NULL, \
("%s:%d curvnet is NULL, so=%p", __func__, __LINE__, (so)));
+#ifdef SOCKET_HHOOK
VNET_DEFINE(struct hhook_head *, socket_hhh[HHOOK_SOCKET_LAST + 1]);
#define V_socket_hhh VNET(socket_hhh)
+static inline int hhook_run_socket(struct socket *, void *, int32_t);
+#endif
/*
* Limit on the number of connections in the listen queue waiting
* for accept(2).
* NB: The original sysctl somaxconn is still available but hidden
* to prevent confusion about the actual purpose of this number.
*/
static u_int somaxconn = SOMAXCONN;
static int
sysctl_somaxconn(SYSCTL_HANDLER_ARGS)
{
int error;
int val;
val = somaxconn;
error = sysctl_handle_int(oidp, &val, 0, req);
if (error || !req->newptr )
return (error);
/*
* The purpose of the UINT_MAX / 3 limit, is so that the formula
* 3 * so_qlimit / 2
* below, will not overflow.
*/
if (val < 1 || val > UINT_MAX / 3)
return (EINVAL);
somaxconn = val;
return (0);
}
SYSCTL_PROC(_kern_ipc, OID_AUTO, soacceptqueue,
CTLTYPE_UINT | CTLFLAG_RW | CTLFLAG_MPSAFE, 0, sizeof(int),
sysctl_somaxconn, "I",
"Maximum listen socket pending connection accept queue size");
SYSCTL_PROC(_kern_ipc, KIPC_SOMAXCONN, somaxconn,
CTLTYPE_UINT | CTLFLAG_RW | CTLFLAG_SKIP | CTLFLAG_MPSAFE, 0,
sizeof(int), sysctl_somaxconn, "I",
"Maximum listen socket pending connection accept queue size (compat)");
static int numopensockets;
SYSCTL_INT(_kern_ipc, OID_AUTO, numopensockets, CTLFLAG_RD,
&numopensockets, 0, "Number of open sockets");
/*
* so_global_mtx protects so_gencnt, numopensockets, and the per-socket
* so_gencnt field.
*/
static struct mtx so_global_mtx;
MTX_SYSINIT(so_global_mtx, &so_global_mtx, "so_glabel", MTX_DEF);
/*
* General IPC sysctl name space, used by sockets and a variety of other IPC
* types.
*/
SYSCTL_NODE(_kern, KERN_IPC, ipc, CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
"IPC");
/*
* Initialize the socket subsystem and set up the socket
* memory allocator.
*/
static uma_zone_t socket_zone;
int maxsockets;
static void
socket_zone_change(void *tag)
{
maxsockets = uma_zone_set_max(socket_zone, maxsockets);
}
+static void
+socket_init(void *tag)
+{
+
+ socket_zone = uma_zcreate("socket", sizeof(struct socket), NULL, NULL,
+ NULL, NULL, UMA_ALIGN_PTR, 0);
+ maxsockets = uma_zone_set_max(socket_zone, maxsockets);
+ uma_zone_set_warning(socket_zone, "kern.ipc.maxsockets limit reached");
+ EVENTHANDLER_REGISTER(maxsockets_change, socket_zone_change, NULL,
+ EVENTHANDLER_PRI_FIRST);
+}
+SYSINIT(socket, SI_SUB_PROTO_DOMAININIT, SI_ORDER_ANY, socket_init, NULL);
+
+#ifdef SOCKET_HHOOK
static void
socket_hhook_register(int subtype)
{
if (hhook_head_register(HHOOK_TYPE_SOCKET, subtype,
&V_socket_hhh[subtype],
HHOOK_NOWAIT|HHOOK_HEADISINVNET) != 0)
printf("%s: WARNING: unable to register hook\n", __func__);
}
static void
socket_hhook_deregister(int subtype)
{
if (hhook_head_deregister(V_socket_hhh[subtype]) != 0)
printf("%s: WARNING: unable to deregister hook\n", __func__);
}
-static void
-socket_init(void *tag)
-{
-
- socket_zone = uma_zcreate("socket", sizeof(struct socket), NULL, NULL,
- NULL, NULL, UMA_ALIGN_PTR, 0);
- maxsockets = uma_zone_set_max(socket_zone, maxsockets);
- uma_zone_set_warning(socket_zone, "kern.ipc.maxsockets limit reached");
- EVENTHANDLER_REGISTER(maxsockets_change, socket_zone_change, NULL,
- EVENTHANDLER_PRI_FIRST);
-}
-SYSINIT(socket, SI_SUB_PROTO_DOMAININIT, SI_ORDER_ANY, socket_init, NULL);
-
static void
socket_vnet_init(const void *unused __unused)
{
int i;
/* We expect a contiguous range */
for (i = 0; i <= HHOOK_SOCKET_LAST; i++)
socket_hhook_register(i);
}
VNET_SYSINIT(socket_vnet_init, SI_SUB_PROTO_DOMAININIT, SI_ORDER_ANY,
socket_vnet_init, NULL);
static void
socket_vnet_uninit(const void *unused __unused)
{
int i;
for (i = 0; i <= HHOOK_SOCKET_LAST; i++)
socket_hhook_deregister(i);
}
VNET_SYSUNINIT(socket_vnet_uninit, SI_SUB_PROTO_DOMAININIT, SI_ORDER_ANY,
socket_vnet_uninit, NULL);
+#endif /* SOCKET_HHOOK */
/*
* Initialise maxsockets. This SYSINIT must be run after
* tunable_mbinit().
*/
static void
init_maxsockets(void *ignored)
{
TUNABLE_INT_FETCH("kern.ipc.maxsockets", &maxsockets);
maxsockets = imax(maxsockets, maxfiles);
}
SYSINIT(param, SI_SUB_TUNABLES, SI_ORDER_ANY, init_maxsockets, NULL);
/*
* Sysctl to get and set the maximum global sockets limit. Notify protocols
* of the change so that they can update their dependent limits as required.
*/
static int
sysctl_maxsockets(SYSCTL_HANDLER_ARGS)
{
int error, newmaxsockets;
newmaxsockets = maxsockets;
error = sysctl_handle_int(oidp, &newmaxsockets, 0, req);
if (error == 0 && req->newptr && newmaxsockets != maxsockets) {
if (newmaxsockets > maxsockets &&
newmaxsockets <= maxfiles) {
maxsockets = newmaxsockets;
EVENTHANDLER_INVOKE(maxsockets_change);
} else
error = EINVAL;
}
return (error);
}
SYSCTL_PROC(_kern_ipc, OID_AUTO, maxsockets,
CTLTYPE_INT | CTLFLAG_RWTUN | CTLFLAG_NOFETCH | CTLFLAG_MPSAFE,
&maxsockets, 0, sysctl_maxsockets, "IU",
"Maximum number of sockets available");
/*
* Socket operation routines. These routines are called by the routines in
* sys_socket.c or from a system process, and implement the semantics of
* socket operations by switching out to the protocol specific routines.
*/
/*
* Get a socket structure from our zone, and initialize it. Note that it
* would probably be better to allocate socket and PCB at the same time, but
* I'm not convinced that all the protocols can be easily modified to do
* this.
*
* soalloc() returns a socket with a ref count of 0.
*/
static struct socket *
soalloc(struct vnet *vnet)
{
struct socket *so;
so = uma_zalloc(socket_zone, M_NOWAIT | M_ZERO);
if (so == NULL)
return (NULL);
#ifdef MAC
if (mac_socket_init(so, M_NOWAIT) != 0) {
uma_zfree(socket_zone, so);
return (NULL);
}
#endif
if (khelp_init_osd(HELPER_CLASS_SOCKET, &so->osd)) {
uma_zfree(socket_zone, so);
return (NULL);
}
/*
* The socket locking protocol allows to lock 2 sockets at a time,
* however, the first one must be a listening socket. WITNESS lacks
* a feature to change class of an existing lock, so we use DUPOK.
*/
mtx_init(&so->so_lock, "socket", NULL, MTX_DEF | MTX_DUPOK);
mtx_init(&so->so_snd_mtx, "so_snd", NULL, MTX_DEF);
mtx_init(&so->so_rcv_mtx, "so_rcv", NULL, MTX_DEF);
so->so_rcv.sb_sel = &so->so_rdsel;
so->so_snd.sb_sel = &so->so_wrsel;
sx_init(&so->so_snd_sx, "so_snd_sx");
sx_init(&so->so_rcv_sx, "so_rcv_sx");
TAILQ_INIT(&so->so_snd.sb_aiojobq);
TAILQ_INIT(&so->so_rcv.sb_aiojobq);
TASK_INIT(&so->so_snd.sb_aiotask, 0, soaio_snd, so);
TASK_INIT(&so->so_rcv.sb_aiotask, 0, soaio_rcv, so);
#ifdef VIMAGE
VNET_ASSERT(vnet != NULL, ("%s:%d vnet is NULL, so=%p",
__func__, __LINE__, so));
so->so_vnet = vnet;
#endif
+#ifdef SOCKET_HHOOK
/* We shouldn't need the so_global_mtx */
if (hhook_run_socket(so, NULL, HHOOK_SOCKET_CREATE)) {
/* Do we need more comprehensive error returns? */
uma_zfree(socket_zone, so);
return (NULL);
}
+#endif
mtx_lock(&so_global_mtx);
so->so_gencnt = ++so_gencnt;
++numopensockets;
#ifdef VIMAGE
vnet->vnet_sockcnt++;
#endif
mtx_unlock(&so_global_mtx);
return (so);
}
/*
* Free the storage associated with a socket at the socket layer, tear down
* locks, labels, etc. All protocol state is assumed already to have been
* torn down (and possibly never set up) by the caller.
*/
void
sodealloc(struct socket *so)
{
KASSERT(so->so_count == 0, ("sodealloc(): so_count %d", so->so_count));
KASSERT(so->so_pcb == NULL, ("sodealloc(): so_pcb != NULL"));
mtx_lock(&so_global_mtx);
so->so_gencnt = ++so_gencnt;
--numopensockets; /* Could be below, but faster here. */
#ifdef VIMAGE
VNET_ASSERT(so->so_vnet != NULL, ("%s:%d so_vnet is NULL, so=%p",
__func__, __LINE__, so));
so->so_vnet->vnet_sockcnt--;
#endif
mtx_unlock(&so_global_mtx);
#ifdef MAC
mac_socket_destroy(so);
#endif
+#ifdef SOCKET_HHOOK
hhook_run_socket(so, NULL, HHOOK_SOCKET_CLOSE);
+#endif
khelp_destroy_osd(&so->osd);
if (SOLISTENING(so)) {
if (so->sol_accept_filter != NULL)
accept_filt_setopt(so, NULL);
} else {
if (so->so_rcv.sb_hiwat)
(void)chgsbsize(so->so_cred->cr_uidinfo,
&so->so_rcv.sb_hiwat, 0, RLIM_INFINITY);
if (so->so_snd.sb_hiwat)
(void)chgsbsize(so->so_cred->cr_uidinfo,
&so->so_snd.sb_hiwat, 0, RLIM_INFINITY);
sx_destroy(&so->so_snd_sx);
sx_destroy(&so->so_rcv_sx);
mtx_destroy(&so->so_snd_mtx);
mtx_destroy(&so->so_rcv_mtx);
}
crfree(so->so_cred);
mtx_destroy(&so->so_lock);
uma_zfree(socket_zone, so);
}
/*
* socreate returns a socket with a ref count of 1 and a file descriptor
* reference. The socket should be closed with soclose().
*/
int
socreate(int dom, struct socket **aso, int type, int proto,
struct ucred *cred, struct thread *td)
{
struct protosw *prp;
struct socket *so;
int error;
/*
* XXX: divert(4) historically abused PF_INET. Keep this compatibility
* shim until all applications have been updated.
*/
if (__predict_false(dom == PF_INET && type == SOCK_RAW &&
proto == IPPROTO_DIVERT)) {
dom = PF_DIVERT;
printf("%s uses obsolete way to create divert(4) socket\n",
td->td_proc->p_comm);
}
prp = pffindproto(dom, type, proto);
if (prp == NULL) {
/* No support for domain. */
if (pffinddomain(dom) == NULL)
return (EAFNOSUPPORT);
/* No support for socket type. */
if (proto == 0 && type != 0)
return (EPROTOTYPE);
return (EPROTONOSUPPORT);
}
MPASS(prp->pr_attach);
if ((prp->pr_flags & PR_CAPATTACH) == 0) {
if (CAP_TRACING(td))
ktrcapfail(CAPFAIL_PROTO, &proto);
if (IN_CAPABILITY_MODE(td))
return (ECAPMODE);
}
if (prison_check_af(cred, prp->pr_domain->dom_family) != 0)
return (EPROTONOSUPPORT);
so = soalloc(CRED_TO_VNET(cred));
if (so == NULL)
return (ENOBUFS);
so->so_type = type;
so->so_cred = crhold(cred);
if ((prp->pr_domain->dom_family == PF_INET) ||
(prp->pr_domain->dom_family == PF_INET6) ||
(prp->pr_domain->dom_family == PF_ROUTE))
so->so_fibnum = td->td_proc->p_fibnum;
else
so->so_fibnum = 0;
so->so_proto = prp;
#ifdef MAC
mac_socket_create(cred, so);
#endif
knlist_init(&so->so_rdsel.si_note, so, so_rdknl_lock, so_rdknl_unlock,
so_rdknl_assert_lock);
knlist_init(&so->so_wrsel.si_note, so, so_wrknl_lock, so_wrknl_unlock,
so_wrknl_assert_lock);
if ((prp->pr_flags & PR_SOCKBUF) == 0) {
so->so_snd.sb_mtx = &so->so_snd_mtx;
so->so_rcv.sb_mtx = &so->so_rcv_mtx;
}
/*
* Auto-sizing of socket buffers is managed by the protocols and
* the appropriate flags must be set in the pru_attach function.
*/
CURVNET_SET(so->so_vnet);
error = prp->pr_attach(so, proto, td);
CURVNET_RESTORE();
if (error) {
sodealloc(so);
return (error);
}
soref(so);
*aso = so;
return (0);
}
#ifdef REGRESSION
static int regression_sonewconn_earlytest = 1;
SYSCTL_INT(_regression, OID_AUTO, sonewconn_earlytest, CTLFLAG_RW,
®ression_sonewconn_earlytest, 0, "Perform early sonewconn limit test");
#endif
static int sooverprio = LOG_DEBUG;
SYSCTL_INT(_kern_ipc, OID_AUTO, sooverprio, CTLFLAG_RW,
&sooverprio, 0, "Log priority for listen socket overflows: 0..7 or -1 to disable");
static struct timeval overinterval = { 60, 0 };
SYSCTL_TIMEVAL_SEC(_kern_ipc, OID_AUTO, sooverinterval, CTLFLAG_RW,
&overinterval,
"Delay in seconds between warnings for listen socket overflows");
/*
* When an attempt at a new connection is noted on a socket which supports
* accept(2), the protocol has two options:
* 1) Call legacy sonewconn() function, which would call protocol attach
* method, same as used for socket(2).
* 2) Call solisten_clone(), do attach that is specific to a cloned connection,
* and then call solisten_enqueue().
*
* Note: the ref count on the socket is 0 on return.
*/
struct socket *
solisten_clone(struct socket *head)
{
struct sbuf descrsb;
struct socket *so;
int len, overcount;
u_int qlen;
const char localprefix[] = "local:";
char descrbuf[SUNPATHLEN + sizeof(localprefix)];
#if defined(INET6)
char addrbuf[INET6_ADDRSTRLEN];
#elif defined(INET)
char addrbuf[INET_ADDRSTRLEN];
#endif
bool dolog, over;
SOLISTEN_LOCK(head);
over = (head->sol_qlen > 3 * head->sol_qlimit / 2);
#ifdef REGRESSION
if (regression_sonewconn_earlytest && over) {
#else
if (over) {
#endif
head->sol_overcount++;
dolog = (sooverprio >= 0) &&
!!ratecheck(&head->sol_lastover, &overinterval);
/*
* If we're going to log, copy the overflow count and queue
* length from the listen socket before dropping the lock.
* Also, reset the overflow count.
*/
if (dolog) {
overcount = head->sol_overcount;
head->sol_overcount = 0;
qlen = head->sol_qlen;
}
SOLISTEN_UNLOCK(head);
if (dolog) {
/*
* Try to print something descriptive about the
* socket for the error message.
*/
sbuf_new(&descrsb, descrbuf, sizeof(descrbuf),
SBUF_FIXEDLEN);
switch (head->so_proto->pr_domain->dom_family) {
#if defined(INET) || defined(INET6)
#ifdef INET
case AF_INET:
#endif
#ifdef INET6
case AF_INET6:
if (head->so_proto->pr_domain->dom_family ==
AF_INET6 ||
(sotoinpcb(head)->inp_inc.inc_flags &
INC_ISIPV6)) {
ip6_sprintf(addrbuf,
&sotoinpcb(head)->inp_inc.inc6_laddr);
sbuf_printf(&descrsb, "[%s]", addrbuf);
} else
#endif
{
#ifdef INET
inet_ntoa_r(
sotoinpcb(head)->inp_inc.inc_laddr,
addrbuf);
sbuf_cat(&descrsb, addrbuf);
#endif
}
sbuf_printf(&descrsb, ":%hu (proto %u)",
ntohs(sotoinpcb(head)->inp_inc.inc_lport),
head->so_proto->pr_protocol);
break;
#endif /* INET || INET6 */
case AF_UNIX:
sbuf_cat(&descrsb, localprefix);
if (sotounpcb(head)->unp_addr != NULL)
len =
sotounpcb(head)->unp_addr->sun_len -
offsetof(struct sockaddr_un,
sun_path);
else
len = 0;
if (len > 0)
sbuf_bcat(&descrsb,
sotounpcb(head)->unp_addr->sun_path,
len);
else
sbuf_cat(&descrsb, "(unknown)");
break;
}
/*
* If we can't print something more specific, at least
* print the domain name.
*/
if (sbuf_finish(&descrsb) != 0 ||
sbuf_len(&descrsb) <= 0) {
sbuf_clear(&descrsb);
sbuf_cat(&descrsb,
head->so_proto->pr_domain->dom_name ?:
"unknown");
sbuf_finish(&descrsb);
}
KASSERT(sbuf_len(&descrsb) > 0,
("%s: sbuf creation failed", __func__));
/*
* Preserve the historic listen queue overflow log
* message, that starts with "sonewconn:". It has
* been known to sysadmins for years and also test
* sys/kern/sonewconn_overflow checks for it.
*/
if (head->so_cred == 0) {
log(LOG_PRI(sooverprio),
"sonewconn: pcb %p (%s): "
"Listen queue overflow: %i already in "
"queue awaiting acceptance (%d "
"occurrences)\n", head->so_pcb,
sbuf_data(&descrsb),
qlen, overcount);
} else {
log(LOG_PRI(sooverprio),
"sonewconn: pcb %p (%s): "
"Listen queue overflow: "
"%i already in queue awaiting acceptance "
"(%d occurrences), euid %d, rgid %d, jail %s\n",
head->so_pcb, sbuf_data(&descrsb), qlen,
overcount, head->so_cred->cr_uid,
head->so_cred->cr_rgid,
head->so_cred->cr_prison ?
head->so_cred->cr_prison->pr_name :
"not_jailed");
}
sbuf_delete(&descrsb);
overcount = 0;
}
return (NULL);
}
SOLISTEN_UNLOCK(head);
VNET_ASSERT(head->so_vnet != NULL, ("%s: so %p vnet is NULL",
__func__, head));
so = soalloc(head->so_vnet);
if (so == NULL) {
log(LOG_DEBUG, "%s: pcb %p: New socket allocation failure: "
"limit reached or out of memory\n",
__func__, head->so_pcb);
return (NULL);
}
so->so_listen = head;
so->so_type = head->so_type;
/*
* POSIX is ambiguous on what options an accept(2)ed socket should
* inherit from the listener. Words "create a new socket" may be
* interpreted as not inheriting anything. Best programming practice
* for application developers is to not rely on such inheritance.
* FreeBSD had historically inherited all so_options excluding
* SO_ACCEPTCONN, which virtually means all SOL_SOCKET level options,
* including those completely irrelevant to a new born socket. For
* compatibility with older versions we will inherit a list of
* meaningful options.
* The crucial bit to inherit is SO_ACCEPTFILTER. We need it present
* in the child socket for soisconnected() promoting socket from the
* incomplete queue to complete. It will be cleared before the child
* gets available to accept(2).
*/
so->so_options = head->so_options & (SO_ACCEPTFILTER | SO_KEEPALIVE |
SO_DONTROUTE | SO_LINGER | SO_OOBINLINE | SO_NOSIGPIPE);
so->so_linger = head->so_linger;
so->so_state = head->so_state;
so->so_fibnum = head->so_fibnum;
so->so_proto = head->so_proto;
so->so_cred = crhold(head->so_cred);
+#ifdef SOCKET_HHOOK
if (V_socket_hhh[HHOOK_SOCKET_NEWCONN]->hhh_nhooks > 0) {
if (hhook_run_socket(so, head, HHOOK_SOCKET_NEWCONN)) {
sodealloc(so);
log(LOG_DEBUG, "%s: hhook run failed\n", __func__);
return (NULL);
}
}
+#endif
#ifdef MAC
mac_socket_newconn(head, so);
#endif
knlist_init(&so->so_rdsel.si_note, so, so_rdknl_lock, so_rdknl_unlock,
so_rdknl_assert_lock);
knlist_init(&so->so_wrsel.si_note, so, so_wrknl_lock, so_wrknl_unlock,
so_wrknl_assert_lock);
VNET_SO_ASSERT(head);
if (soreserve(so, head->sol_sbsnd_hiwat, head->sol_sbrcv_hiwat)) {
sodealloc(so);
log(LOG_DEBUG, "%s: pcb %p: soreserve() failed\n",
__func__, head->so_pcb);
return (NULL);
}
so->so_rcv.sb_lowat = head->sol_sbrcv_lowat;
so->so_snd.sb_lowat = head->sol_sbsnd_lowat;
so->so_rcv.sb_timeo = head->sol_sbrcv_timeo;
so->so_snd.sb_timeo = head->sol_sbsnd_timeo;
so->so_rcv.sb_flags = head->sol_sbrcv_flags & SB_AUTOSIZE;
so->so_snd.sb_flags = head->sol_sbsnd_flags & SB_AUTOSIZE;
if ((so->so_proto->pr_flags & PR_SOCKBUF) == 0) {
so->so_snd.sb_mtx = &so->so_snd_mtx;
so->so_rcv.sb_mtx = &so->so_rcv_mtx;
}
return (so);
}
/* Connstatus may be 0 or SS_ISCONNECTED. */
struct socket *
sonewconn(struct socket *head, int connstatus)
{
struct socket *so;
if ((so = solisten_clone(head)) == NULL)
return (NULL);
if (so->so_proto->pr_attach(so, 0, NULL) != 0) {
sodealloc(so);
log(LOG_DEBUG, "%s: pcb %p: pr_attach() failed\n",
__func__, head->so_pcb);
return (NULL);
}
(void)solisten_enqueue(so, connstatus);
return (so);
}
/*
* Enqueue socket cloned by solisten_clone() to the listen queue of the
* listener it has been cloned from.
*
* Return 'true' if socket landed on complete queue, otherwise 'false'.
*/
bool
solisten_enqueue(struct socket *so, int connstatus)
{
struct socket *head = so->so_listen;
MPASS(refcount_load(&so->so_count) == 0);
refcount_init(&so->so_count, 1);
SOLISTEN_LOCK(head);
if (head->sol_accept_filter != NULL)
connstatus = 0;
so->so_state |= connstatus;
soref(head); /* A socket on (in)complete queue refs head. */
if (connstatus) {
TAILQ_INSERT_TAIL(&head->sol_comp, so, so_list);
so->so_qstate = SQ_COMP;
head->sol_qlen++;
solisten_wakeup(head); /* unlocks */
return (true);
} else {
/*
* Keep removing sockets from the head until there's room for
* us to insert on the tail. In pre-locking revisions, this
* was a simple if(), but as we could be racing with other
* threads and soabort() requires dropping locks, we must
* loop waiting for the condition to be true.
*/
while (head->sol_incqlen > head->sol_qlimit) {
struct socket *sp;
sp = TAILQ_FIRST(&head->sol_incomp);
TAILQ_REMOVE(&head->sol_incomp, sp, so_list);
head->sol_incqlen--;
SOCK_LOCK(sp);
sp->so_qstate = SQ_NONE;
sp->so_listen = NULL;
SOCK_UNLOCK(sp);
sorele_locked(head); /* does SOLISTEN_UNLOCK, head stays */
soabort(sp);
SOLISTEN_LOCK(head);
}
TAILQ_INSERT_TAIL(&head->sol_incomp, so, so_list);
so->so_qstate = SQ_INCOMP;
head->sol_incqlen++;
SOLISTEN_UNLOCK(head);
return (false);
}
}
#if defined(SCTP) || defined(SCTP_SUPPORT)
/*
* Socket part of sctp_peeloff(). Detach a new socket from an
* association. The new socket is returned with a reference.
*
* XXXGL: reduce copy-paste with solisten_clone().
*/
struct socket *
sopeeloff(struct socket *head)
{
struct socket *so;
VNET_ASSERT(head->so_vnet != NULL, ("%s:%d so_vnet is NULL, head=%p",
__func__, __LINE__, head));
so = soalloc(head->so_vnet);
if (so == NULL) {
log(LOG_DEBUG, "%s: pcb %p: New socket allocation failure: "
"limit reached or out of memory\n",
__func__, head->so_pcb);
return (NULL);
}
so->so_type = head->so_type;
so->so_options = head->so_options;
so->so_linger = head->so_linger;
so->so_state = (head->so_state & SS_NBIO) | SS_ISCONNECTED;
so->so_fibnum = head->so_fibnum;
so->so_proto = head->so_proto;
so->so_cred = crhold(head->so_cred);
#ifdef MAC
mac_socket_newconn(head, so);
#endif
knlist_init(&so->so_rdsel.si_note, so, so_rdknl_lock, so_rdknl_unlock,
so_rdknl_assert_lock);
knlist_init(&so->so_wrsel.si_note, so, so_wrknl_lock, so_wrknl_unlock,
so_wrknl_assert_lock);
VNET_SO_ASSERT(head);
if (soreserve(so, head->so_snd.sb_hiwat, head->so_rcv.sb_hiwat)) {
sodealloc(so);
log(LOG_DEBUG, "%s: pcb %p: soreserve() failed\n",
__func__, head->so_pcb);
return (NULL);
}
if ((*so->so_proto->pr_attach)(so, 0, NULL)) {
sodealloc(so);
log(LOG_DEBUG, "%s: pcb %p: pru_attach() failed\n",
__func__, head->so_pcb);
return (NULL);
}
so->so_rcv.sb_lowat = head->so_rcv.sb_lowat;
so->so_snd.sb_lowat = head->so_snd.sb_lowat;
so->so_rcv.sb_timeo = head->so_rcv.sb_timeo;
so->so_snd.sb_timeo = head->so_snd.sb_timeo;
so->so_rcv.sb_flags |= head->so_rcv.sb_flags & SB_AUTOSIZE;
so->so_snd.sb_flags |= head->so_snd.sb_flags & SB_AUTOSIZE;
if ((so->so_proto->pr_flags & PR_SOCKBUF) == 0) {
so->so_snd.sb_mtx = &so->so_snd_mtx;
so->so_rcv.sb_mtx = &so->so_rcv_mtx;
}
soref(so);
return (so);
}
#endif /* SCTP */
int
sobind(struct socket *so, struct sockaddr *nam, struct thread *td)
{
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_bind(so, nam, td);
CURVNET_RESTORE();
return (error);
}
int
sobindat(int fd, struct socket *so, struct sockaddr *nam, struct thread *td)
{
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_bindat(fd, so, nam, td);
CURVNET_RESTORE();
return (error);
}
/*
* solisten() transitions a socket from a non-listening state to a listening
* state, but can also be used to update the listen queue depth on an
* existing listen socket. The protocol will call back into the sockets
* layer using solisten_proto_check() and solisten_proto() to check and set
* socket-layer listen state. Call backs are used so that the protocol can
* acquire both protocol and socket layer locks in whatever order is required
* by the protocol.
*
* Protocol implementors are advised to hold the socket lock across the
* socket-layer test and set to avoid races at the socket layer.
*/
int
solisten(struct socket *so, int backlog, struct thread *td)
{
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_listen(so, backlog, td);
CURVNET_RESTORE();
return (error);
}
/*
* Prepare for a call to solisten_proto(). Acquire all socket buffer locks in
* order to interlock with socket I/O.
*/
int
solisten_proto_check(struct socket *so)
{
SOCK_LOCK_ASSERT(so);
if ((so->so_state & (SS_ISCONNECTED | SS_ISCONNECTING |
SS_ISDISCONNECTING)) != 0)
return (EINVAL);
/*
* Sleeping is not permitted here, so simply fail if userspace is
* attempting to transmit or receive on the socket. This kind of
* transient failure is not ideal, but it should occur only if userspace
* is misusing the socket interfaces.
*/
if (!sx_try_xlock(&so->so_snd_sx))
return (EAGAIN);
if (!sx_try_xlock(&so->so_rcv_sx)) {
sx_xunlock(&so->so_snd_sx);
return (EAGAIN);
}
mtx_lock(&so->so_snd_mtx);
mtx_lock(&so->so_rcv_mtx);
/* Interlock with soo_aio_queue() and KTLS. */
if (!SOLISTENING(so)) {
bool ktls;
#ifdef KERN_TLS
ktls = so->so_snd.sb_tls_info != NULL ||
so->so_rcv.sb_tls_info != NULL;
#else
ktls = false;
#endif
if (ktls ||
(so->so_snd.sb_flags & (SB_AIO | SB_AIO_RUNNING)) != 0 ||
(so->so_rcv.sb_flags & (SB_AIO | SB_AIO_RUNNING)) != 0) {
solisten_proto_abort(so);
return (EINVAL);
}
}
return (0);
}
/*
* Undo the setup done by solisten_proto_check().
*/
void
solisten_proto_abort(struct socket *so)
{
mtx_unlock(&so->so_snd_mtx);
mtx_unlock(&so->so_rcv_mtx);
sx_xunlock(&so->so_snd_sx);
sx_xunlock(&so->so_rcv_sx);
}
void
solisten_proto(struct socket *so, int backlog)
{
int sbrcv_lowat, sbsnd_lowat;
u_int sbrcv_hiwat, sbsnd_hiwat;
short sbrcv_flags, sbsnd_flags;
sbintime_t sbrcv_timeo, sbsnd_timeo;
SOCK_LOCK_ASSERT(so);
KASSERT((so->so_state & (SS_ISCONNECTED | SS_ISCONNECTING |
SS_ISDISCONNECTING)) == 0,
("%s: bad socket state %p", __func__, so));
if (SOLISTENING(so))
goto listening;
/*
* Change this socket to listening state.
*/
sbrcv_lowat = so->so_rcv.sb_lowat;
sbsnd_lowat = so->so_snd.sb_lowat;
sbrcv_hiwat = so->so_rcv.sb_hiwat;
sbsnd_hiwat = so->so_snd.sb_hiwat;
sbrcv_flags = so->so_rcv.sb_flags;
sbsnd_flags = so->so_snd.sb_flags;
sbrcv_timeo = so->so_rcv.sb_timeo;
sbsnd_timeo = so->so_snd.sb_timeo;
if (!(so->so_proto->pr_flags & PR_SOCKBUF)) {
sbdestroy(so, SO_SND);
sbdestroy(so, SO_RCV);
}
#ifdef INVARIANTS
bzero(&so->so_rcv,
sizeof(struct socket) - offsetof(struct socket, so_rcv));
#endif
so->sol_sbrcv_lowat = sbrcv_lowat;
so->sol_sbsnd_lowat = sbsnd_lowat;
so->sol_sbrcv_hiwat = sbrcv_hiwat;
so->sol_sbsnd_hiwat = sbsnd_hiwat;
so->sol_sbrcv_flags = sbrcv_flags;
so->sol_sbsnd_flags = sbsnd_flags;
so->sol_sbrcv_timeo = sbrcv_timeo;
so->sol_sbsnd_timeo = sbsnd_timeo;
so->sol_qlen = so->sol_incqlen = 0;
TAILQ_INIT(&so->sol_incomp);
TAILQ_INIT(&so->sol_comp);
so->sol_accept_filter = NULL;
so->sol_accept_filter_arg = NULL;
so->sol_accept_filter_str = NULL;
so->sol_upcall = NULL;
so->sol_upcallarg = NULL;
so->so_options |= SO_ACCEPTCONN;
listening:
if (backlog < 0 || backlog > somaxconn)
backlog = somaxconn;
so->sol_qlimit = backlog;
mtx_unlock(&so->so_snd_mtx);
mtx_unlock(&so->so_rcv_mtx);
sx_xunlock(&so->so_snd_sx);
sx_xunlock(&so->so_rcv_sx);
}
/*
* Wakeup listeners/subsystems once we have a complete connection.
* Enters with lock, returns unlocked.
*/
void
solisten_wakeup(struct socket *sol)
{
if (sol->sol_upcall != NULL)
(void )sol->sol_upcall(sol, sol->sol_upcallarg, M_NOWAIT);
else {
selwakeuppri(&sol->so_rdsel, PSOCK);
KNOTE_LOCKED(&sol->so_rdsel.si_note, 0);
}
SOLISTEN_UNLOCK(sol);
wakeup_one(&sol->sol_comp);
if ((sol->so_state & SS_ASYNC) && sol->so_sigio != NULL)
pgsigio(&sol->so_sigio, SIGIO, 0);
}
/*
* Return single connection off a listening socket queue. Main consumer of
* the function is kern_accept4(). Some modules, that do their own accept
* management also use the function. The socket reference held by the
* listen queue is handed to the caller.
*
* Listening socket must be locked on entry and is returned unlocked on
* return.
* The flags argument is set of accept4(2) flags and ACCEPT4_INHERIT.
*/
int
solisten_dequeue(struct socket *head, struct socket **ret, int flags)
{
struct socket *so;
int error;
SOLISTEN_LOCK_ASSERT(head);
while (!(head->so_state & SS_NBIO) && TAILQ_EMPTY(&head->sol_comp) &&
head->so_error == 0) {
error = msleep(&head->sol_comp, SOCK_MTX(head), PSOCK | PCATCH,
"accept", 0);
if (error != 0) {
SOLISTEN_UNLOCK(head);
return (error);
}
}
if (head->so_error) {
error = head->so_error;
head->so_error = 0;
} else if ((head->so_state & SS_NBIO) && TAILQ_EMPTY(&head->sol_comp))
error = EWOULDBLOCK;
else
error = 0;
if (error) {
SOLISTEN_UNLOCK(head);
return (error);
}
so = TAILQ_FIRST(&head->sol_comp);
SOCK_LOCK(so);
KASSERT(so->so_qstate == SQ_COMP,
("%s: so %p not SQ_COMP", __func__, so));
head->sol_qlen--;
so->so_qstate = SQ_NONE;
so->so_listen = NULL;
TAILQ_REMOVE(&head->sol_comp, so, so_list);
if (flags & ACCEPT4_INHERIT)
so->so_state |= (head->so_state & SS_NBIO);
else
so->so_state |= (flags & SOCK_NONBLOCK) ? SS_NBIO : 0;
SOCK_UNLOCK(so);
sorele_locked(head);
*ret = so;
return (0);
}
/*
* Free socket upon release of the very last reference.
*/
static void
sofree(struct socket *so)
{
struct protosw *pr = so->so_proto;
SOCK_LOCK_ASSERT(so);
KASSERT(refcount_load(&so->so_count) == 0,
("%s: so %p has references", __func__, so));
KASSERT(SOLISTENING(so) || so->so_qstate == SQ_NONE,
("%s: so %p is on listen queue", __func__, so));
SOCK_UNLOCK(so);
if (so->so_dtor != NULL)
so->so_dtor(so);
VNET_SO_ASSERT(so);
if (pr->pr_detach != NULL)
pr->pr_detach(so);
/*
* From this point on, we assume that no other references to this
* socket exist anywhere else in the stack. Therefore, no locks need
* to be acquired or held.
*/
if (!(pr->pr_flags & PR_SOCKBUF) && !SOLISTENING(so)) {
sbdestroy(so, SO_SND);
sbdestroy(so, SO_RCV);
}
seldrain(&so->so_rdsel);
seldrain(&so->so_wrsel);
knlist_destroy(&so->so_rdsel.si_note);
knlist_destroy(&so->so_wrsel.si_note);
sodealloc(so);
}
/*
* Release a reference on a socket while holding the socket lock.
* Unlocks the socket lock before returning.
*/
void
sorele_locked(struct socket *so)
{
SOCK_LOCK_ASSERT(so);
if (refcount_release(&so->so_count))
sofree(so);
else
SOCK_UNLOCK(so);
}
/*
* Close a socket on last file table reference removal. Initiate disconnect
* if connected. Free socket when disconnect complete.
*
* This function will sorele() the socket. Note that soclose() may be called
* prior to the ref count reaching zero. The actual socket structure will
* not be freed until the ref count reaches zero.
*/
int
soclose(struct socket *so)
{
struct accept_queue lqueue;
int error = 0;
bool listening, last __diagused;
CURVNET_SET(so->so_vnet);
funsetown(&so->so_sigio);
if (so->so_state & SS_ISCONNECTED) {
if ((so->so_state & SS_ISDISCONNECTING) == 0) {
error = sodisconnect(so);
if (error) {
if (error == ENOTCONN)
error = 0;
goto drop;
}
}
if ((so->so_options & SO_LINGER) != 0 && so->so_linger != 0) {
if ((so->so_state & SS_ISDISCONNECTING) &&
(so->so_state & SS_NBIO))
goto drop;
while (so->so_state & SS_ISCONNECTED) {
error = tsleep(&so->so_timeo,
PSOCK | PCATCH, "soclos",
so->so_linger * hz);
if (error)
break;
}
}
}
drop:
if (so->so_proto->pr_close != NULL)
so->so_proto->pr_close(so);
SOCK_LOCK(so);
if ((listening = SOLISTENING(so))) {
struct socket *sp;
TAILQ_INIT(&lqueue);
TAILQ_SWAP(&lqueue, &so->sol_incomp, socket, so_list);
TAILQ_CONCAT(&lqueue, &so->sol_comp, so_list);
so->sol_qlen = so->sol_incqlen = 0;
TAILQ_FOREACH(sp, &lqueue, so_list) {
SOCK_LOCK(sp);
sp->so_qstate = SQ_NONE;
sp->so_listen = NULL;
SOCK_UNLOCK(sp);
last = refcount_release(&so->so_count);
KASSERT(!last, ("%s: released last reference for %p",
__func__, so));
}
}
sorele_locked(so);
if (listening) {
struct socket *sp, *tsp;
TAILQ_FOREACH_SAFE(sp, &lqueue, so_list, tsp)
soabort(sp);
}
CURVNET_RESTORE();
return (error);
}
/*
* soabort() is used to abruptly tear down a connection, such as when a
* resource limit is reached (listen queue depth exceeded), or if a listen
* socket is closed while there are sockets waiting to be accepted.
*
* This interface is tricky, because it is called on an unreferenced socket,
* and must be called only by a thread that has actually removed the socket
* from the listen queue it was on. Likely this thread holds the last
* reference on the socket and soabort() will proceed with sofree(). But
* it might be not the last, as the sockets on the listen queues are seen
* from the protocol side.
*
* This interface will call into the protocol code, so must not be called
* with any socket locks held. Protocols do call it while holding their own
* recursible protocol mutexes, but this is something that should be subject
* to review in the future.
*
* Usually socket should have a single reference left, but this is not a
* requirement. In the past, when we have had named references for file
* descriptor and protocol, we asserted that none of them are being held.
*/
void
soabort(struct socket *so)
{
VNET_SO_ASSERT(so);
if (so->so_proto->pr_abort != NULL)
so->so_proto->pr_abort(so);
SOCK_LOCK(so);
sorele_locked(so);
}
int
soaccept(struct socket *so, struct sockaddr *sa)
{
#ifdef INVARIANTS
u_char len = sa->sa_len;
#endif
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_accept(so, sa);
KASSERT(sa->sa_len <= len,
("%s: protocol %p sockaddr overflow", __func__, so->so_proto));
CURVNET_RESTORE();
return (error);
}
int
sopeeraddr(struct socket *so, struct sockaddr *sa)
{
#ifdef INVARIANTS
u_char len = sa->sa_len;
#endif
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_peeraddr(so, sa);
KASSERT(sa->sa_len <= len,
("%s: protocol %p sockaddr overflow", __func__, so->so_proto));
CURVNET_RESTORE();
return (error);
}
int
sosockaddr(struct socket *so, struct sockaddr *sa)
{
#ifdef INVARIANTS
u_char len = sa->sa_len;
#endif
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_sockaddr(so, sa);
KASSERT(sa->sa_len <= len,
("%s: protocol %p sockaddr overflow", __func__, so->so_proto));
CURVNET_RESTORE();
return (error);
}
int
soconnect(struct socket *so, struct sockaddr *nam, struct thread *td)
{
return (soconnectat(AT_FDCWD, so, nam, td));
}
int
soconnectat(int fd, struct socket *so, struct sockaddr *nam, struct thread *td)
{
int error;
CURVNET_SET(so->so_vnet);
/*
* If protocol is connection-based, can only connect once.
* Otherwise, if connected, try to disconnect first. This allows
* user to disconnect by connecting to, e.g., a null address.
*
* Note, this check is racy and may need to be re-evaluated at the
* protocol layer.
*/
if (so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING) &&
((so->so_proto->pr_flags & PR_CONNREQUIRED) ||
(error = sodisconnect(so)))) {
error = EISCONN;
} else {
/*
* Prevent accumulated error from previous connection from
* biting us.
*/
so->so_error = 0;
if (fd == AT_FDCWD) {
error = so->so_proto->pr_connect(so, nam, td);
} else {
error = so->so_proto->pr_connectat(fd, so, nam, td);
}
}
CURVNET_RESTORE();
return (error);
}
int
soconnect2(struct socket *so1, struct socket *so2)
{
int error;
CURVNET_SET(so1->so_vnet);
error = so1->so_proto->pr_connect2(so1, so2);
CURVNET_RESTORE();
return (error);
}
int
sodisconnect(struct socket *so)
{
int error;
if ((so->so_state & SS_ISCONNECTED) == 0)
return (ENOTCONN);
if (so->so_state & SS_ISDISCONNECTING)
return (EALREADY);
VNET_SO_ASSERT(so);
error = so->so_proto->pr_disconnect(so);
return (error);
}
int
sosend_dgram(struct socket *so, struct sockaddr *addr, struct uio *uio,
struct mbuf *top, struct mbuf *control, int flags, struct thread *td)
{
long space;
ssize_t resid;
int clen = 0, error, dontroute;
KASSERT(so->so_type == SOCK_DGRAM, ("sosend_dgram: !SOCK_DGRAM"));
KASSERT(so->so_proto->pr_flags & PR_ATOMIC,
("sosend_dgram: !PR_ATOMIC"));
if (uio != NULL)
resid = uio->uio_resid;
else
resid = top->m_pkthdr.len;
/*
* In theory resid should be unsigned. However, space must be
* signed, as it might be less than 0 if we over-committed, and we
* must use a signed comparison of space and resid. On the other
* hand, a negative resid causes us to loop sending 0-length
* segments to the protocol.
*/
if (resid < 0) {
error = EINVAL;
goto out;
}
dontroute =
(flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0;
if (td != NULL)
td->td_ru.ru_msgsnd++;
if (control != NULL)
clen = control->m_len;
SOCKBUF_LOCK(&so->so_snd);
if (so->so_snd.sb_state & SBS_CANTSENDMORE) {
SOCKBUF_UNLOCK(&so->so_snd);
error = EPIPE;
goto out;
}
if (so->so_error) {
error = so->so_error;
so->so_error = 0;
SOCKBUF_UNLOCK(&so->so_snd);
goto out;
}
if ((so->so_state & SS_ISCONNECTED) == 0) {
/*
* `sendto' and `sendmsg' is allowed on a connection-based
* socket if it supports implied connect. Return ENOTCONN if
* not connected and no address is supplied.
*/
if ((so->so_proto->pr_flags & PR_CONNREQUIRED) &&
(so->so_proto->pr_flags & PR_IMPLOPCL) == 0) {
if (!(resid == 0 && clen != 0)) {
SOCKBUF_UNLOCK(&so->so_snd);
error = ENOTCONN;
goto out;
}
} else if (addr == NULL) {
if (so->so_proto->pr_flags & PR_CONNREQUIRED)
error = ENOTCONN;
else
error = EDESTADDRREQ;
SOCKBUF_UNLOCK(&so->so_snd);
goto out;
}
}
/*
* Do we need MSG_OOB support in SOCK_DGRAM? Signs here may be a
* problem and need fixing.
*/
space = sbspace(&so->so_snd);
if (flags & MSG_OOB)
space += 1024;
space -= clen;
SOCKBUF_UNLOCK(&so->so_snd);
if (resid > space) {
error = EMSGSIZE;
goto out;
}
if (uio == NULL) {
resid = 0;
if (flags & MSG_EOR)
top->m_flags |= M_EOR;
} else {
/*
* Copy the data from userland into a mbuf chain.
* If no data is to be copied in, a single empty mbuf
* is returned.
*/
top = m_uiotombuf(uio, M_WAITOK, space, max_hdr,
(M_PKTHDR | ((flags & MSG_EOR) ? M_EOR : 0)));
if (top == NULL) {
error = EFAULT; /* only possible error */
goto out;
}
space -= resid - uio->uio_resid;
resid = uio->uio_resid;
}
KASSERT(resid == 0, ("sosend_dgram: resid != 0"));
/*
* XXXRW: Frobbing SO_DONTROUTE here is even worse without sblock
* than with.
*/
if (dontroute) {
SOCK_LOCK(so);
so->so_options |= SO_DONTROUTE;
SOCK_UNLOCK(so);
}
/*
* XXX all the SBS_CANTSENDMORE checks previously done could be out
* of date. We could have received a reset packet in an interrupt or
* maybe we slept while doing page faults in uiomove() etc. We could
* probably recheck again inside the locking protection here, but
* there are probably other places that this also happens. We must
* rethink this.
*/
VNET_SO_ASSERT(so);
error = so->so_proto->pr_send(so, (flags & MSG_OOB) ? PRUS_OOB :
/*
* If the user set MSG_EOF, the protocol understands this flag and
* nothing left to send then use PRU_SEND_EOF instead of PRU_SEND.
*/
((flags & MSG_EOF) &&
(so->so_proto->pr_flags & PR_IMPLOPCL) &&
(resid <= 0)) ?
PRUS_EOF :
/* If there is more to send set PRUS_MORETOCOME */
(flags & MSG_MORETOCOME) ||
(resid > 0 && space > 0) ? PRUS_MORETOCOME : 0,
top, addr, control, td);
if (dontroute) {
SOCK_LOCK(so);
so->so_options &= ~SO_DONTROUTE;
SOCK_UNLOCK(so);
}
clen = 0;
control = NULL;
top = NULL;
out:
if (top != NULL)
m_freem(top);
if (control != NULL)
m_freem(control);
return (error);
}
/*
* Send on a socket. If send must go all at once and message is larger than
* send buffering, then hard error. Lock against other senders. If must go
* all at once and not enough room now, then inform user that this would
* block and do nothing. Otherwise, if nonblocking, send as much as
* possible. The data to be sent is described by "uio" if nonzero, otherwise
* by the mbuf chain "top" (which must be null if uio is not). Data provided
* in mbuf chain must be small enough to send all at once.
*
* Returns nonzero on error, timeout or signal; callers must check for short
* counts if EINTR/ERESTART are returned. Data and control buffers are freed
* on return.
*/
int
sosend_generic(struct socket *so, struct sockaddr *addr, struct uio *uio,
struct mbuf *top, struct mbuf *control, int flags, struct thread *td)
{
long space;
ssize_t resid;
int clen = 0, error, dontroute;
int atomic = sosendallatonce(so) || top;
int pr_send_flag;
#ifdef KERN_TLS
struct ktls_session *tls;
int tls_enq_cnt, tls_send_flag;
uint8_t tls_rtype;
tls = NULL;
tls_rtype = TLS_RLTYPE_APP;
#endif
if (uio != NULL)
resid = uio->uio_resid;
else if ((top->m_flags & M_PKTHDR) != 0)
resid = top->m_pkthdr.len;
else
resid = m_length(top, NULL);
/*
* In theory resid should be unsigned. However, space must be
* signed, as it might be less than 0 if we over-committed, and we
* must use a signed comparison of space and resid. On the other
* hand, a negative resid causes us to loop sending 0-length
* segments to the protocol.
*
* Also check to make sure that MSG_EOR isn't used on SOCK_STREAM
* type sockets since that's an error.
*/
if (resid < 0 || (so->so_type == SOCK_STREAM && (flags & MSG_EOR))) {
error = EINVAL;
goto out;
}
dontroute =
(flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0 &&
(so->so_proto->pr_flags & PR_ATOMIC);
if (td != NULL)
td->td_ru.ru_msgsnd++;
if (control != NULL)
clen = control->m_len;
error = SOCK_IO_SEND_LOCK(so, SBLOCKWAIT(flags));
if (error)
goto out;
#ifdef KERN_TLS
tls_send_flag = 0;
tls = ktls_hold(so->so_snd.sb_tls_info);
if (tls != NULL) {
if (tls->mode == TCP_TLS_MODE_SW)
tls_send_flag = PRUS_NOTREADY;
if (control != NULL) {
struct cmsghdr *cm = mtod(control, struct cmsghdr *);
if (clen >= sizeof(*cm) &&
cm->cmsg_type == TLS_SET_RECORD_TYPE) {
tls_rtype = *((uint8_t *)CMSG_DATA(cm));
clen = 0;
m_freem(control);
control = NULL;
atomic = 1;
}
}
if (resid == 0 && !ktls_permit_empty_frames(tls)) {
error = EINVAL;
goto release;
}
}
#endif
restart:
do {
SOCKBUF_LOCK(&so->so_snd);
if (so->so_snd.sb_state & SBS_CANTSENDMORE) {
SOCKBUF_UNLOCK(&so->so_snd);
error = EPIPE;
goto release;
}
if (so->so_error) {
error = so->so_error;
so->so_error = 0;
SOCKBUF_UNLOCK(&so->so_snd);
goto release;
}
if ((so->so_state & SS_ISCONNECTED) == 0) {
/*
* `sendto' and `sendmsg' is allowed on a connection-
* based socket if it supports implied connect.
* Return ENOTCONN if not connected and no address is
* supplied.
*/
if ((so->so_proto->pr_flags & PR_CONNREQUIRED) &&
(so->so_proto->pr_flags & PR_IMPLOPCL) == 0) {
if (!(resid == 0 && clen != 0)) {
SOCKBUF_UNLOCK(&so->so_snd);
error = ENOTCONN;
goto release;
}
} else if (addr == NULL) {
SOCKBUF_UNLOCK(&so->so_snd);
if (so->so_proto->pr_flags & PR_CONNREQUIRED)
error = ENOTCONN;
else
error = EDESTADDRREQ;
goto release;
}
}
space = sbspace(&so->so_snd);
if (flags & MSG_OOB)
space += 1024;
if ((atomic && resid > so->so_snd.sb_hiwat) ||
clen > so->so_snd.sb_hiwat) {
SOCKBUF_UNLOCK(&so->so_snd);
error = EMSGSIZE;
goto release;
}
if (space < resid + clen &&
(atomic || space < so->so_snd.sb_lowat || space < clen)) {
if ((so->so_state & SS_NBIO) ||
(flags & (MSG_NBIO | MSG_DONTWAIT)) != 0) {
SOCKBUF_UNLOCK(&so->so_snd);
error = EWOULDBLOCK;
goto release;
}
error = sbwait(so, SO_SND);
SOCKBUF_UNLOCK(&so->so_snd);
if (error)
goto release;
goto restart;
}
SOCKBUF_UNLOCK(&so->so_snd);
space -= clen;
do {
if (uio == NULL) {
resid = 0;
if (flags & MSG_EOR)
top->m_flags |= M_EOR;
#ifdef KERN_TLS
if (tls != NULL) {
ktls_frame(top, tls, &tls_enq_cnt,
tls_rtype);
tls_rtype = TLS_RLTYPE_APP;
}
#endif
} else {
/*
* Copy the data from userland into a mbuf
* chain. If resid is 0, which can happen
* only if we have control to send, then
* a single empty mbuf is returned. This
* is a workaround to prevent protocol send
* methods to panic.
*/
#ifdef KERN_TLS
if (tls != NULL) {
top = m_uiotombuf(uio, M_WAITOK, space,
tls->params.max_frame_len,
M_EXTPG |
((flags & MSG_EOR) ? M_EOR : 0));
if (top != NULL) {
ktls_frame(top, tls,
&tls_enq_cnt, tls_rtype);
}
tls_rtype = TLS_RLTYPE_APP;
} else
#endif
top = m_uiotombuf(uio, M_WAITOK, space,
(atomic ? max_hdr : 0),
(atomic ? M_PKTHDR : 0) |
((flags & MSG_EOR) ? M_EOR : 0));
if (top == NULL) {
error = EFAULT; /* only possible error */
goto release;
}
space -= resid - uio->uio_resid;
resid = uio->uio_resid;
}
if (dontroute) {
SOCK_LOCK(so);
so->so_options |= SO_DONTROUTE;
SOCK_UNLOCK(so);
}
/*
* XXX all the SBS_CANTSENDMORE checks previously
* done could be out of date. We could have received
* a reset packet in an interrupt or maybe we slept
* while doing page faults in uiomove() etc. We
* could probably recheck again inside the locking
* protection here, but there are probably other
* places that this also happens. We must rethink
* this.
*/
VNET_SO_ASSERT(so);
pr_send_flag = (flags & MSG_OOB) ? PRUS_OOB :
/*
* If the user set MSG_EOF, the protocol understands
* this flag and nothing left to send then use
* PRU_SEND_EOF instead of PRU_SEND.
*/
((flags & MSG_EOF) &&
(so->so_proto->pr_flags & PR_IMPLOPCL) &&
(resid <= 0)) ?
PRUS_EOF :
/* If there is more to send set PRUS_MORETOCOME. */
(flags & MSG_MORETOCOME) ||
(resid > 0 && space > 0) ? PRUS_MORETOCOME : 0;
#ifdef KERN_TLS
pr_send_flag |= tls_send_flag;
#endif
error = so->so_proto->pr_send(so, pr_send_flag, top,
addr, control, td);
if (dontroute) {
SOCK_LOCK(so);
so->so_options &= ~SO_DONTROUTE;
SOCK_UNLOCK(so);
}
#ifdef KERN_TLS
if (tls != NULL && tls->mode == TCP_TLS_MODE_SW) {
if (error != 0) {
m_freem(top);
top = NULL;
} else {
soref(so);
ktls_enqueue(top, so, tls_enq_cnt);
}
}
#endif
clen = 0;
control = NULL;
top = NULL;
if (error)
goto release;
} while (resid && space > 0);
} while (resid);
release:
SOCK_IO_SEND_UNLOCK(so);
out:
#ifdef KERN_TLS
if (tls != NULL)
ktls_free(tls);
#endif
if (top != NULL)
m_freem(top);
if (control != NULL)
m_freem(control);
return (error);
}
/*
* Send to a socket from a kernel thread.
*
* XXXGL: in almost all cases uio is NULL and the mbuf is supplied.
* Exception is nfs/bootp_subr.c. It is arguable that the VNET context needs
* to be set at all. This function should just boil down to a static inline
* calling the protocol method.
*/
int
sosend(struct socket *so, struct sockaddr *addr, struct uio *uio,
struct mbuf *top, struct mbuf *control, int flags, struct thread *td)
{
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_sosend(so, addr, uio,
top, control, flags, td);
CURVNET_RESTORE();
return (error);
}
/*
* send(2), write(2) or aio_write(2) on a socket.
*/
int
sousrsend(struct socket *so, struct sockaddr *addr, struct uio *uio,
struct mbuf *control, int flags, struct proc *userproc)
{
struct thread *td;
ssize_t len;
int error;
td = uio->uio_td;
len = uio->uio_resid;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_sosend(so, addr, uio, NULL, control, flags,
td);
CURVNET_RESTORE();
if (error != 0) {
/*
* Clear transient errors for stream protocols if they made
* some progress. Make exclusion for aio(4) that would
* schedule a new write in case of EWOULDBLOCK and clear
* error itself. See soaio_process_job().
*/
if (uio->uio_resid != len &&
(so->so_proto->pr_flags & PR_ATOMIC) == 0 &&
userproc == NULL &&
(error == ERESTART || error == EINTR ||
error == EWOULDBLOCK))
error = 0;
/* Generation of SIGPIPE can be controlled per socket. */
if (error == EPIPE && (so->so_options & SO_NOSIGPIPE) == 0 &&
(flags & MSG_NOSIGNAL) == 0) {
if (userproc != NULL) {
/* aio(4) job */
PROC_LOCK(userproc);
kern_psignal(userproc, SIGPIPE);
PROC_UNLOCK(userproc);
} else {
PROC_LOCK(td->td_proc);
tdsignal(td, SIGPIPE);
PROC_UNLOCK(td->td_proc);
}
}
}
return (error);
}
/*
* The part of soreceive() that implements reading non-inline out-of-band
* data from a socket. For more complete comments, see soreceive(), from
* which this code originated.
*
* Note that soreceive_rcvoob(), unlike the remainder of soreceive(), is
* unable to return an mbuf chain to the caller.
*/
static int
soreceive_rcvoob(struct socket *so, struct uio *uio, int flags)
{
struct protosw *pr = so->so_proto;
struct mbuf *m;
int error;
KASSERT(flags & MSG_OOB, ("soreceive_rcvoob: (flags & MSG_OOB) == 0"));
VNET_SO_ASSERT(so);
m = m_get(M_WAITOK, MT_DATA);
error = pr->pr_rcvoob(so, m, flags & MSG_PEEK);
if (error)
goto bad;
do {
error = uiomove(mtod(m, void *),
(int) min(uio->uio_resid, m->m_len), uio);
m = m_free(m);
} while (uio->uio_resid && error == 0 && m);
bad:
if (m != NULL)
m_freem(m);
return (error);
}
/*
* Following replacement or removal of the first mbuf on the first mbuf chain
* of a socket buffer, push necessary state changes back into the socket
* buffer so that other consumers see the values consistently. 'nextrecord'
* is the callers locally stored value of the original value of
* sb->sb_mb->m_nextpkt which must be restored when the lead mbuf changes.
* NOTE: 'nextrecord' may be NULL.
*/
static __inline void
sockbuf_pushsync(struct sockbuf *sb, struct mbuf *nextrecord)
{
SOCKBUF_LOCK_ASSERT(sb);
/*
* First, update for the new value of nextrecord. If necessary, make
* it the first record.
*/
if (sb->sb_mb != NULL)
sb->sb_mb->m_nextpkt = nextrecord;
else
sb->sb_mb = nextrecord;
/*
* Now update any dependent socket buffer fields to reflect the new
* state. This is an expanded inline of SB_EMPTY_FIXUP(), with the
* addition of a second clause that takes care of the case where
* sb_mb has been updated, but remains the last record.
*/
if (sb->sb_mb == NULL) {
sb->sb_mbtail = NULL;
sb->sb_lastrecord = NULL;
} else if (sb->sb_mb->m_nextpkt == NULL)
sb->sb_lastrecord = sb->sb_mb;
}
/*
* Implement receive operations on a socket. We depend on the way that
* records are added to the sockbuf by sbappend. In particular, each record
* (mbufs linked through m_next) must begin with an address if the protocol
* so specifies, followed by an optional mbuf or mbufs containing ancillary
* data, and then zero or more mbufs of data. In order to allow parallelism
* between network receive and copying to user space, as well as avoid
* sleeping with a mutex held, we release the socket buffer mutex during the
* user space copy. Although the sockbuf is locked, new data may still be
* appended, and thus we must maintain consistency of the sockbuf during that
* time.
*
* The caller may receive the data as a single mbuf chain by supplying an
* mbuf **mp0 for use in returning the chain. The uio is then used only for
* the count in uio_resid.
*/
int
soreceive_generic(struct socket *so, struct sockaddr **psa, struct uio *uio,
struct mbuf **mp0, struct mbuf **controlp, int *flagsp)
{
struct mbuf *m, **mp;
int flags, error, offset;
ssize_t len;
struct protosw *pr = so->so_proto;
struct mbuf *nextrecord;
int moff, type = 0;
ssize_t orig_resid = uio->uio_resid;
bool report_real_len = false;
mp = mp0;
if (psa != NULL)
*psa = NULL;
if (controlp != NULL)
*controlp = NULL;
if (flagsp != NULL) {
report_real_len = *flagsp & MSG_TRUNC;
*flagsp &= ~MSG_TRUNC;
flags = *flagsp &~ MSG_EOR;
} else
flags = 0;
if (flags & MSG_OOB)
return (soreceive_rcvoob(so, uio, flags));
if (mp != NULL)
*mp = NULL;
error = SOCK_IO_RECV_LOCK(so, SBLOCKWAIT(flags));
if (error)
return (error);
restart:
SOCKBUF_LOCK(&so->so_rcv);
m = so->so_rcv.sb_mb;
/*
* If we have less data than requested, block awaiting more (subject
* to any timeout) if:
* 1. the current count is less than the low water mark, or
* 2. MSG_DONTWAIT is not set
*/
if (m == NULL || (((flags & MSG_DONTWAIT) == 0 &&
sbavail(&so->so_rcv) < uio->uio_resid) &&
sbavail(&so->so_rcv) < so->so_rcv.sb_lowat &&
m->m_nextpkt == NULL && (pr->pr_flags & PR_ATOMIC) == 0)) {
KASSERT(m != NULL || !sbavail(&so->so_rcv),
("receive: m == %p sbavail == %u",
m, sbavail(&so->so_rcv)));
if (so->so_error || so->so_rerror) {
if (m != NULL)
goto dontblock;
if (so->so_error)
error = so->so_error;
else
error = so->so_rerror;
if ((flags & MSG_PEEK) == 0) {
if (so->so_error)
so->so_error = 0;
else
so->so_rerror = 0;
}
SOCKBUF_UNLOCK(&so->so_rcv);
goto release;
}
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (so->so_rcv.sb_state & SBS_CANTRCVMORE) {
if (m != NULL)
goto dontblock;
#ifdef KERN_TLS
else if (so->so_rcv.sb_tlsdcc == 0 &&
so->so_rcv.sb_tlscc == 0) {
#else
else {
#endif
SOCKBUF_UNLOCK(&so->so_rcv);
goto release;
}
}
for (; m != NULL; m = m->m_next)
if (m->m_type == MT_OOBDATA || (m->m_flags & M_EOR)) {
m = so->so_rcv.sb_mb;
goto dontblock;
}
if ((so->so_state & (SS_ISCONNECTING | SS_ISCONNECTED |
SS_ISDISCONNECTING | SS_ISDISCONNECTED)) == 0 &&
(so->so_proto->pr_flags & PR_CONNREQUIRED) != 0) {
SOCKBUF_UNLOCK(&so->so_rcv);
error = ENOTCONN;
goto release;
}
if (uio->uio_resid == 0 && !report_real_len) {
SOCKBUF_UNLOCK(&so->so_rcv);
goto release;
}
if ((so->so_state & SS_NBIO) ||
(flags & (MSG_DONTWAIT|MSG_NBIO))) {
SOCKBUF_UNLOCK(&so->so_rcv);
error = EWOULDBLOCK;
goto release;
}
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
error = sbwait(so, SO_RCV);
SOCKBUF_UNLOCK(&so->so_rcv);
if (error)
goto release;
goto restart;
}
dontblock:
/*
* From this point onward, we maintain 'nextrecord' as a cache of the
* pointer to the next record in the socket buffer. We must keep the
* various socket buffer pointers and local stack versions of the
* pointers in sync, pushing out modifications before dropping the
* socket buffer mutex, and re-reading them when picking it up.
*
* Otherwise, we will race with the network stack appending new data
* or records onto the socket buffer by using inconsistent/stale
* versions of the field, possibly resulting in socket buffer
* corruption.
*
* By holding the high-level sblock(), we prevent simultaneous
* readers from pulling off the front of the socket buffer.
*/
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (uio->uio_td)
uio->uio_td->td_ru.ru_msgrcv++;
KASSERT(m == so->so_rcv.sb_mb, ("soreceive: m != so->so_rcv.sb_mb"));
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
nextrecord = m->m_nextpkt;
if (pr->pr_flags & PR_ADDR) {
KASSERT(m->m_type == MT_SONAME,
("m->m_type == %d", m->m_type));
orig_resid = 0;
if (psa != NULL)
*psa = sodupsockaddr(mtod(m, struct sockaddr *),
M_NOWAIT);
if (flags & MSG_PEEK) {
m = m->m_next;
} else {
sbfree(&so->so_rcv, m);
so->so_rcv.sb_mb = m_free(m);
m = so->so_rcv.sb_mb;
sockbuf_pushsync(&so->so_rcv, nextrecord);
}
}
/*
* Process one or more MT_CONTROL mbufs present before any data mbufs
* in the first mbuf chain on the socket buffer. If MSG_PEEK, we
* just copy the data; if !MSG_PEEK, we call into the protocol to
* perform externalization (or freeing if controlp == NULL).
*/
if (m != NULL && m->m_type == MT_CONTROL) {
struct mbuf *cm = NULL, *cmn;
struct mbuf **cme = &cm;
#ifdef KERN_TLS
struct cmsghdr *cmsg;
struct tls_get_record tgr;
/*
* For MSG_TLSAPPDATA, check for an alert record.
* If found, return ENXIO without removing
* it from the receive queue. This allows a subsequent
* call without MSG_TLSAPPDATA to receive it.
* Note that, for TLS, there should only be a single
* control mbuf with the TLS_GET_RECORD message in it.
*/
if (flags & MSG_TLSAPPDATA) {
cmsg = mtod(m, struct cmsghdr *);
if (cmsg->cmsg_type == TLS_GET_RECORD &&
cmsg->cmsg_len == CMSG_LEN(sizeof(tgr))) {
memcpy(&tgr, CMSG_DATA(cmsg), sizeof(tgr));
if (__predict_false(tgr.tls_type ==
TLS_RLTYPE_ALERT)) {
SOCKBUF_UNLOCK(&so->so_rcv);
error = ENXIO;
goto release;
}
}
}
#endif
do {
if (flags & MSG_PEEK) {
if (controlp != NULL) {
*controlp = m_copym(m, 0, m->m_len,
M_NOWAIT);
controlp = &(*controlp)->m_next;
}
m = m->m_next;
} else {
sbfree(&so->so_rcv, m);
so->so_rcv.sb_mb = m->m_next;
m->m_next = NULL;
*cme = m;
cme = &(*cme)->m_next;
m = so->so_rcv.sb_mb;
}
} while (m != NULL && m->m_type == MT_CONTROL);
if ((flags & MSG_PEEK) == 0)
sockbuf_pushsync(&so->so_rcv, nextrecord);
while (cm != NULL) {
cmn = cm->m_next;
cm->m_next = NULL;
if (pr->pr_domain->dom_externalize != NULL) {
SOCKBUF_UNLOCK(&so->so_rcv);
VNET_SO_ASSERT(so);
error = (*pr->pr_domain->dom_externalize)
(cm, controlp, flags);
SOCKBUF_LOCK(&so->so_rcv);
} else if (controlp != NULL)
*controlp = cm;
else
m_freem(cm);
if (controlp != NULL) {
while (*controlp != NULL)
controlp = &(*controlp)->m_next;
}
cm = cmn;
}
if (m != NULL)
nextrecord = so->so_rcv.sb_mb->m_nextpkt;
else
nextrecord = so->so_rcv.sb_mb;
orig_resid = 0;
}
if (m != NULL) {
if ((flags & MSG_PEEK) == 0) {
KASSERT(m->m_nextpkt == nextrecord,
("soreceive: post-control, nextrecord !sync"));
if (nextrecord == NULL) {
KASSERT(so->so_rcv.sb_mb == m,
("soreceive: post-control, sb_mb!=m"));
KASSERT(so->so_rcv.sb_lastrecord == m,
("soreceive: post-control, lastrecord!=m"));
}
}
type = m->m_type;
if (type == MT_OOBDATA)
flags |= MSG_OOB;
} else {
if ((flags & MSG_PEEK) == 0) {
KASSERT(so->so_rcv.sb_mb == nextrecord,
("soreceive: sb_mb != nextrecord"));
if (so->so_rcv.sb_mb == NULL) {
KASSERT(so->so_rcv.sb_lastrecord == NULL,
("soreceive: sb_lastercord != NULL"));
}
}
}
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
/*
* Now continue to read any data mbufs off of the head of the socket
* buffer until the read request is satisfied. Note that 'type' is
* used to store the type of any mbuf reads that have happened so far
* such that soreceive() can stop reading if the type changes, which
* causes soreceive() to return only one of regular data and inline
* out-of-band data in a single socket receive operation.
*/
moff = 0;
offset = 0;
while (m != NULL && !(m->m_flags & M_NOTAVAIL) && uio->uio_resid > 0
&& error == 0) {
/*
* If the type of mbuf has changed since the last mbuf
* examined ('type'), end the receive operation.
*/
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (m->m_type == MT_OOBDATA || m->m_type == MT_CONTROL) {
if (type != m->m_type)
break;
} else if (type == MT_OOBDATA)
break;
else
KASSERT(m->m_type == MT_DATA,
("m->m_type == %d", m->m_type));
so->so_rcv.sb_state &= ~SBS_RCVATMARK;
len = uio->uio_resid;
if (so->so_oobmark && len > so->so_oobmark - offset)
len = so->so_oobmark - offset;
if (len > m->m_len - moff)
len = m->m_len - moff;
/*
* If mp is set, just pass back the mbufs. Otherwise copy
* them out via the uio, then free. Sockbuf must be
* consistent here (points to current mbuf, it points to next
* record) when we drop priority; we must note any additions
* to the sockbuf when we block interrupts again.
*/
if (mp == NULL) {
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
SOCKBUF_UNLOCK(&so->so_rcv);
if ((m->m_flags & M_EXTPG) != 0)
error = m_unmapped_uiomove(m, moff, uio,
(int)len);
else
error = uiomove(mtod(m, char *) + moff,
(int)len, uio);
SOCKBUF_LOCK(&so->so_rcv);
if (error) {
/*
* The MT_SONAME mbuf has already been removed
* from the record, so it is necessary to
* remove the data mbufs, if any, to preserve
* the invariant in the case of PR_ADDR that
* requires MT_SONAME mbufs at the head of
* each record.
*/
if (pr->pr_flags & PR_ATOMIC &&
((flags & MSG_PEEK) == 0))
(void)sbdroprecord_locked(&so->so_rcv);
SOCKBUF_UNLOCK(&so->so_rcv);
goto release;
}
} else
uio->uio_resid -= len;
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (len == m->m_len - moff) {
if (m->m_flags & M_EOR)
flags |= MSG_EOR;
if (flags & MSG_PEEK) {
m = m->m_next;
moff = 0;
} else {
nextrecord = m->m_nextpkt;
sbfree(&so->so_rcv, m);
if (mp != NULL) {
m->m_nextpkt = NULL;
*mp = m;
mp = &m->m_next;
so->so_rcv.sb_mb = m = m->m_next;
*mp = NULL;
} else {
so->so_rcv.sb_mb = m_free(m);
m = so->so_rcv.sb_mb;
}
sockbuf_pushsync(&so->so_rcv, nextrecord);
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
}
} else {
if (flags & MSG_PEEK)
moff += len;
else {
if (mp != NULL) {
if (flags & MSG_DONTWAIT) {
*mp = m_copym(m, 0, len,
M_NOWAIT);
if (*mp == NULL) {
/*
* m_copym() couldn't
* allocate an mbuf.
* Adjust uio_resid back
* (it was adjusted
* down by len bytes,
* which we didn't end
* up "copying" over).
*/
uio->uio_resid += len;
break;
}
} else {
SOCKBUF_UNLOCK(&so->so_rcv);
*mp = m_copym(m, 0, len,
M_WAITOK);
SOCKBUF_LOCK(&so->so_rcv);
}
}
sbcut_locked(&so->so_rcv, len);
}
}
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (so->so_oobmark) {
if ((flags & MSG_PEEK) == 0) {
so->so_oobmark -= len;
if (so->so_oobmark == 0) {
so->so_rcv.sb_state |= SBS_RCVATMARK;
break;
}
} else {
offset += len;
if (offset == so->so_oobmark)
break;
}
}
if (flags & MSG_EOR)
break;
/*
* If the MSG_WAITALL flag is set (for non-atomic socket), we
* must not quit until "uio->uio_resid == 0" or an error
* termination. If a signal/timeout occurs, return with a
* short count but without error. Keep sockbuf locked
* against other readers.
*/
while (flags & MSG_WAITALL && m == NULL && uio->uio_resid > 0 &&
!sosendallatonce(so) && nextrecord == NULL) {
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (so->so_error || so->so_rerror ||
so->so_rcv.sb_state & SBS_CANTRCVMORE)
break;
/*
* Notify the protocol that some data has been
* drained before blocking.
*/
if (pr->pr_flags & PR_WANTRCVD) {
SOCKBUF_UNLOCK(&so->so_rcv);
VNET_SO_ASSERT(so);
pr->pr_rcvd(so, flags);
SOCKBUF_LOCK(&so->so_rcv);
if (__predict_false(so->so_rcv.sb_mb == NULL &&
(so->so_error || so->so_rerror ||
so->so_rcv.sb_state & SBS_CANTRCVMORE)))
break;
}
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
/*
* We could receive some data while was notifying
* the protocol. Skip blocking in this case.
*/
if (so->so_rcv.sb_mb == NULL) {
error = sbwait(so, SO_RCV);
if (error) {
SOCKBUF_UNLOCK(&so->so_rcv);
goto release;
}
}
m = so->so_rcv.sb_mb;
if (m != NULL)
nextrecord = m->m_nextpkt;
}
}
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (m != NULL && pr->pr_flags & PR_ATOMIC) {
if (report_real_len)
uio->uio_resid -= m_length(m, NULL) - moff;
flags |= MSG_TRUNC;
if ((flags & MSG_PEEK) == 0)
(void) sbdroprecord_locked(&so->so_rcv);
}
if ((flags & MSG_PEEK) == 0) {
if (m == NULL) {
/*
* First part is an inline SB_EMPTY_FIXUP(). Second
* part makes sure sb_lastrecord is up-to-date if
* there is still data in the socket buffer.
*/
so->so_rcv.sb_mb = nextrecord;
if (so->so_rcv.sb_mb == NULL) {
so->so_rcv.sb_mbtail = NULL;
so->so_rcv.sb_lastrecord = NULL;
} else if (nextrecord->m_nextpkt == NULL)
so->so_rcv.sb_lastrecord = nextrecord;
}
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
/*
* If soreceive() is being done from the socket callback,
* then don't need to generate ACK to peer to update window,
* since ACK will be generated on return to TCP.
*/
if (!(flags & MSG_SOCALLBCK) &&
(pr->pr_flags & PR_WANTRCVD)) {
SOCKBUF_UNLOCK(&so->so_rcv);
VNET_SO_ASSERT(so);
pr->pr_rcvd(so, flags);
SOCKBUF_LOCK(&so->so_rcv);
}
}
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (orig_resid == uio->uio_resid && orig_resid &&
(flags & MSG_EOR) == 0 && (so->so_rcv.sb_state & SBS_CANTRCVMORE) == 0) {
SOCKBUF_UNLOCK(&so->so_rcv);
goto restart;
}
SOCKBUF_UNLOCK(&so->so_rcv);
if (flagsp != NULL)
*flagsp |= flags;
release:
SOCK_IO_RECV_UNLOCK(so);
return (error);
}
/*
* Optimized version of soreceive() for stream (TCP) sockets.
*/
int
soreceive_stream(struct socket *so, struct sockaddr **psa, struct uio *uio,
struct mbuf **mp0, struct mbuf **controlp, int *flagsp)
{
int len = 0, error = 0, flags, oresid;
struct sockbuf *sb;
struct mbuf *m, *n = NULL;
/* We only do stream sockets. */
if (so->so_type != SOCK_STREAM)
return (EINVAL);
if (psa != NULL)
*psa = NULL;
if (flagsp != NULL)
flags = *flagsp &~ MSG_EOR;
else
flags = 0;
if (controlp != NULL)
*controlp = NULL;
if (flags & MSG_OOB)
return (soreceive_rcvoob(so, uio, flags));
if (mp0 != NULL)
*mp0 = NULL;
sb = &so->so_rcv;
#ifdef KERN_TLS
/*
* KTLS store TLS records as records with a control message to
* describe the framing.
*
* We check once here before acquiring locks to optimize the
* common case.
*/
if (sb->sb_tls_info != NULL)
return (soreceive_generic(so, psa, uio, mp0, controlp,
flagsp));
#endif
/* Prevent other readers from entering the socket. */
error = SOCK_IO_RECV_LOCK(so, SBLOCKWAIT(flags));
if (error)
return (error);
SOCKBUF_LOCK(sb);
#ifdef KERN_TLS
if (sb->sb_tls_info != NULL) {
SOCKBUF_UNLOCK(sb);
SOCK_IO_RECV_UNLOCK(so);
return (soreceive_generic(so, psa, uio, mp0, controlp,
flagsp));
}
#endif
/* Easy one, no space to copyout anything. */
if (uio->uio_resid == 0) {
error = EINVAL;
goto out;
}
oresid = uio->uio_resid;
/* We will never ever get anything unless we are or were connected. */
if (!(so->so_state & (SS_ISCONNECTED|SS_ISDISCONNECTED))) {
error = ENOTCONN;
goto out;
}
restart:
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
/* Abort if socket has reported problems. */
if (so->so_error) {
if (sbavail(sb) > 0)
goto deliver;
if (oresid > uio->uio_resid)
goto out;
error = so->so_error;
if (!(flags & MSG_PEEK))
so->so_error = 0;
goto out;
}
/* Door is closed. Deliver what is left, if any. */
if (sb->sb_state & SBS_CANTRCVMORE) {
if (sbavail(sb) > 0)
goto deliver;
else
goto out;
}
/* Socket buffer is empty and we shall not block. */
if (sbavail(sb) == 0 &&
((so->so_state & SS_NBIO) || (flags & (MSG_DONTWAIT|MSG_NBIO)))) {
error = EAGAIN;
goto out;
}
/* Socket buffer got some data that we shall deliver now. */
if (sbavail(sb) > 0 && !(flags & MSG_WAITALL) &&
((so->so_state & SS_NBIO) ||
(flags & (MSG_DONTWAIT|MSG_NBIO)) ||
sbavail(sb) >= sb->sb_lowat ||
sbavail(sb) >= uio->uio_resid ||
sbavail(sb) >= sb->sb_hiwat) ) {
goto deliver;
}
/* On MSG_WAITALL we must wait until all data or error arrives. */
if ((flags & MSG_WAITALL) &&
(sbavail(sb) >= uio->uio_resid || sbavail(sb) >= sb->sb_hiwat))
goto deliver;
/*
* Wait and block until (more) data comes in.
* NB: Drops the sockbuf lock during wait.
*/
error = sbwait(so, SO_RCV);
if (error)
goto out;
goto restart;
deliver:
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
KASSERT(sbavail(sb) > 0, ("%s: sockbuf empty", __func__));
KASSERT(sb->sb_mb != NULL, ("%s: sb_mb == NULL", __func__));
/* Statistics. */
if (uio->uio_td)
uio->uio_td->td_ru.ru_msgrcv++;
/* Fill uio until full or current end of socket buffer is reached. */
len = min(uio->uio_resid, sbavail(sb));
if (mp0 != NULL) {
/* Dequeue as many mbufs as possible. */
if (!(flags & MSG_PEEK) && len >= sb->sb_mb->m_len) {
if (*mp0 == NULL)
*mp0 = sb->sb_mb;
else
m_cat(*mp0, sb->sb_mb);
for (m = sb->sb_mb;
m != NULL && m->m_len <= len;
m = m->m_next) {
KASSERT(!(m->m_flags & M_NOTAVAIL),
("%s: m %p not available", __func__, m));
len -= m->m_len;
uio->uio_resid -= m->m_len;
sbfree(sb, m);
n = m;
}
n->m_next = NULL;
sb->sb_mb = m;
sb->sb_lastrecord = sb->sb_mb;
if (sb->sb_mb == NULL)
SB_EMPTY_FIXUP(sb);
}
/* Copy the remainder. */
if (len > 0) {
KASSERT(sb->sb_mb != NULL,
("%s: len > 0 && sb->sb_mb empty", __func__));
m = m_copym(sb->sb_mb, 0, len, M_NOWAIT);
if (m == NULL)
len = 0; /* Don't flush data from sockbuf. */
else
uio->uio_resid -= len;
if (*mp0 != NULL)
m_cat(*mp0, m);
else
*mp0 = m;
if (*mp0 == NULL) {
error = ENOBUFS;
goto out;
}
}
} else {
/* NB: Must unlock socket buffer as uiomove may sleep. */
SOCKBUF_UNLOCK(sb);
error = m_mbuftouio(uio, sb->sb_mb, len);
SOCKBUF_LOCK(sb);
if (error)
goto out;
}
SBLASTRECORDCHK(sb);
SBLASTMBUFCHK(sb);
/*
* Remove the delivered data from the socket buffer unless we
* were only peeking.
*/
if (!(flags & MSG_PEEK)) {
if (len > 0)
sbdrop_locked(sb, len);
/* Notify protocol that we drained some data. */
if ((so->so_proto->pr_flags & PR_WANTRCVD) &&
(((flags & MSG_WAITALL) && uio->uio_resid > 0) ||
!(flags & MSG_SOCALLBCK))) {
SOCKBUF_UNLOCK(sb);
VNET_SO_ASSERT(so);
so->so_proto->pr_rcvd(so, flags);
SOCKBUF_LOCK(sb);
}
}
/*
* For MSG_WAITALL we may have to loop again and wait for
* more data to come in.
*/
if ((flags & MSG_WAITALL) && uio->uio_resid > 0)
goto restart;
out:
SBLASTRECORDCHK(sb);
SBLASTMBUFCHK(sb);
SOCKBUF_UNLOCK(sb);
SOCK_IO_RECV_UNLOCK(so);
return (error);
}
/*
* Optimized version of soreceive() for simple datagram cases from userspace.
* Unlike in the stream case, we're able to drop a datagram if copyout()
* fails, and because we handle datagrams atomically, we don't need to use a
* sleep lock to prevent I/O interlacing.
*/
int
soreceive_dgram(struct socket *so, struct sockaddr **psa, struct uio *uio,
struct mbuf **mp0, struct mbuf **controlp, int *flagsp)
{
struct mbuf *m, *m2;
int flags, error;
ssize_t len;
struct protosw *pr = so->so_proto;
struct mbuf *nextrecord;
if (psa != NULL)
*psa = NULL;
if (controlp != NULL)
*controlp = NULL;
if (flagsp != NULL)
flags = *flagsp &~ MSG_EOR;
else
flags = 0;
/*
* For any complicated cases, fall back to the full
* soreceive_generic().
*/
if (mp0 != NULL || (flags & (MSG_PEEK | MSG_OOB | MSG_TRUNC)))
return (soreceive_generic(so, psa, uio, mp0, controlp,
flagsp));
/*
* Enforce restrictions on use.
*/
KASSERT((pr->pr_flags & PR_WANTRCVD) == 0,
("soreceive_dgram: wantrcvd"));
KASSERT(pr->pr_flags & PR_ATOMIC, ("soreceive_dgram: !atomic"));
KASSERT((so->so_rcv.sb_state & SBS_RCVATMARK) == 0,
("soreceive_dgram: SBS_RCVATMARK"));
KASSERT((so->so_proto->pr_flags & PR_CONNREQUIRED) == 0,
("soreceive_dgram: P_CONNREQUIRED"));
/*
* Loop blocking while waiting for a datagram.
*/
SOCKBUF_LOCK(&so->so_rcv);
while ((m = so->so_rcv.sb_mb) == NULL) {
KASSERT(sbavail(&so->so_rcv) == 0,
("soreceive_dgram: sb_mb NULL but sbavail %u",
sbavail(&so->so_rcv)));
if (so->so_error) {
error = so->so_error;
so->so_error = 0;
SOCKBUF_UNLOCK(&so->so_rcv);
return (error);
}
if (so->so_rcv.sb_state & SBS_CANTRCVMORE ||
uio->uio_resid == 0) {
SOCKBUF_UNLOCK(&so->so_rcv);
return (0);
}
if ((so->so_state & SS_NBIO) ||
(flags & (MSG_DONTWAIT|MSG_NBIO))) {
SOCKBUF_UNLOCK(&so->so_rcv);
return (EWOULDBLOCK);
}
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
error = sbwait(so, SO_RCV);
if (error) {
SOCKBUF_UNLOCK(&so->so_rcv);
return (error);
}
}
SOCKBUF_LOCK_ASSERT(&so->so_rcv);
if (uio->uio_td)
uio->uio_td->td_ru.ru_msgrcv++;
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
nextrecord = m->m_nextpkt;
if (nextrecord == NULL) {
KASSERT(so->so_rcv.sb_lastrecord == m,
("soreceive_dgram: lastrecord != m"));
}
KASSERT(so->so_rcv.sb_mb->m_nextpkt == nextrecord,
("soreceive_dgram: m_nextpkt != nextrecord"));
/*
* Pull 'm' and its chain off the front of the packet queue.
*/
so->so_rcv.sb_mb = NULL;
sockbuf_pushsync(&so->so_rcv, nextrecord);
/*
* Walk 'm's chain and free that many bytes from the socket buffer.
*/
for (m2 = m; m2 != NULL; m2 = m2->m_next)
sbfree(&so->so_rcv, m2);
/*
* Do a few last checks before we let go of the lock.
*/
SBLASTRECORDCHK(&so->so_rcv);
SBLASTMBUFCHK(&so->so_rcv);
SOCKBUF_UNLOCK(&so->so_rcv);
if (pr->pr_flags & PR_ADDR) {
KASSERT(m->m_type == MT_SONAME,
("m->m_type == %d", m->m_type));
if (psa != NULL)
*psa = sodupsockaddr(mtod(m, struct sockaddr *),
M_WAITOK);
m = m_free(m);
}
KASSERT(m, ("%s: no data or control after soname", __func__));
/*
* Packet to copyout() is now in 'm' and it is disconnected from the
* queue.
*
* Process one or more MT_CONTROL mbufs present before any data mbufs
* in the first mbuf chain on the socket buffer. We call into the
* protocol to perform externalization (or freeing if controlp ==
* NULL). In some cases there can be only MT_CONTROL mbufs without
* MT_DATA mbufs.
*/
if (m->m_type == MT_CONTROL) {
struct mbuf *cm = NULL, *cmn;
struct mbuf **cme = &cm;
do {
m2 = m->m_next;
m->m_next = NULL;
*cme = m;
cme = &(*cme)->m_next;
m = m2;
} while (m != NULL && m->m_type == MT_CONTROL);
while (cm != NULL) {
cmn = cm->m_next;
cm->m_next = NULL;
if (pr->pr_domain->dom_externalize != NULL) {
error = (*pr->pr_domain->dom_externalize)
(cm, controlp, flags);
} else if (controlp != NULL)
*controlp = cm;
else
m_freem(cm);
if (controlp != NULL) {
while (*controlp != NULL)
controlp = &(*controlp)->m_next;
}
cm = cmn;
}
}
KASSERT(m == NULL || m->m_type == MT_DATA,
("soreceive_dgram: !data"));
while (m != NULL && uio->uio_resid > 0) {
len = uio->uio_resid;
if (len > m->m_len)
len = m->m_len;
error = uiomove(mtod(m, char *), (int)len, uio);
if (error) {
m_freem(m);
return (error);
}
if (len == m->m_len)
m = m_free(m);
else {
m->m_data += len;
m->m_len -= len;
}
}
if (m != NULL) {
flags |= MSG_TRUNC;
m_freem(m);
}
if (flagsp != NULL)
*flagsp |= flags;
return (0);
}
int
soreceive(struct socket *so, struct sockaddr **psa, struct uio *uio,
struct mbuf **mp0, struct mbuf **controlp, int *flagsp)
{
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_soreceive(so, psa, uio, mp0, controlp, flagsp);
CURVNET_RESTORE();
return (error);
}
int
soshutdown(struct socket *so, enum shutdown_how how)
{
int error;
CURVNET_SET(so->so_vnet);
error = so->so_proto->pr_shutdown(so, how);
CURVNET_RESTORE();
return (error);
}
/*
* Used by several pr_shutdown implementations that use generic socket buffers.
*/
void
sorflush(struct socket *so)
{
int error;
VNET_SO_ASSERT(so);
/*
* Dislodge threads currently blocked in receive and wait to acquire
* a lock against other simultaneous readers before clearing the
* socket buffer. Don't let our acquire be interrupted by a signal
* despite any existing socket disposition on interruptable waiting.
*
* The SOCK_IO_RECV_LOCK() is important here as there some pr_soreceive
* methods that read the top of the socket buffer without acquisition
* of the socket buffer mutex, assuming that top of the buffer
* exclusively belongs to the read(2) syscall. This is handy when
* performing MSG_PEEK.
*/
socantrcvmore(so);
error = SOCK_IO_RECV_LOCK(so, SBL_WAIT | SBL_NOINTR);
if (error != 0) {
KASSERT(SOLISTENING(so),
("%s: soiolock(%p) failed", __func__, so));
return;
}
sbrelease(so, SO_RCV);
SOCK_IO_RECV_UNLOCK(so);
}
+#ifdef SOCKET_HHOOK
/*
* Wrapper for Socket established helper hook.
* Parameters: socket, context of the hook point, hook id.
*/
-static int inline
+static inline int
hhook_run_socket(struct socket *so, void *hctx, int32_t h_id)
{
struct socket_hhook_data hhook_data = {
.so = so,
.hctx = hctx,
.m = NULL,
.status = 0
};
CURVNET_SET(so->so_vnet);
HHOOKS_RUN_IF(V_socket_hhh[h_id], &hhook_data, &so->osd);
CURVNET_RESTORE();
/* Ugly but needed, since hhooks return void for now */
return (hhook_data.status);
}
+#endif
/*
* Perhaps this routine, and sooptcopyout(), below, ought to come in an
* additional variant to handle the case where the option value needs to be
* some kind of integer, but not a specific size. In addition to their use
* here, these functions are also called by the protocol-level pr_ctloutput()
* routines.
*/
int
sooptcopyin(struct sockopt *sopt, void *buf, size_t len, size_t minlen)
{
size_t valsize;
/*
* If the user gives us more than we wanted, we ignore it, but if we
* don't get the minimum length the caller wants, we return EINVAL.
* On success, sopt->sopt_valsize is set to however much we actually
* retrieved.
*/
if ((valsize = sopt->sopt_valsize) < minlen)
return EINVAL;
if (valsize > len)
sopt->sopt_valsize = valsize = len;
if (sopt->sopt_td != NULL)
return (copyin(sopt->sopt_val, buf, valsize));
bcopy(sopt->sopt_val, buf, valsize);
return (0);
}
/*
* Kernel version of setsockopt(2).
*
* XXX: optlen is size_t, not socklen_t
*/
int
so_setsockopt(struct socket *so, int level, int optname, void *optval,
size_t optlen)
{
struct sockopt sopt;
sopt.sopt_level = level;
sopt.sopt_name = optname;
sopt.sopt_dir = SOPT_SET;
sopt.sopt_val = optval;
sopt.sopt_valsize = optlen;
sopt.sopt_td = NULL;
return (sosetopt(so, &sopt));
}
int
sosetopt(struct socket *so, struct sockopt *sopt)
{
int error, optval;
struct linger l;
struct timeval tv;
sbintime_t val, *valp;
uint32_t val32;
#ifdef MAC
struct mac extmac;
#endif
CURVNET_SET(so->so_vnet);
error = 0;
if (sopt->sopt_level != SOL_SOCKET) {
if (so->so_proto->pr_ctloutput != NULL)
error = (*so->so_proto->pr_ctloutput)(so, sopt);
else
error = ENOPROTOOPT;
} else {
switch (sopt->sopt_name) {
case SO_ACCEPTFILTER:
error = accept_filt_setopt(so, sopt);
if (error)
goto bad;
break;
case SO_LINGER:
error = sooptcopyin(sopt, &l, sizeof l, sizeof l);
if (error)
goto bad;
if (l.l_linger < 0 ||
l.l_linger > USHRT_MAX ||
l.l_linger > (INT_MAX / hz)) {
error = EDOM;
goto bad;
}
SOCK_LOCK(so);
so->so_linger = l.l_linger;
if (l.l_onoff)
so->so_options |= SO_LINGER;
else
so->so_options &= ~SO_LINGER;
SOCK_UNLOCK(so);
break;
case SO_DEBUG:
case SO_KEEPALIVE:
case SO_DONTROUTE:
case SO_USELOOPBACK:
case SO_BROADCAST:
case SO_REUSEADDR:
case SO_REUSEPORT:
case SO_REUSEPORT_LB:
case SO_OOBINLINE:
case SO_TIMESTAMP:
case SO_BINTIME:
case SO_NOSIGPIPE:
case SO_NO_DDP:
case SO_NO_OFFLOAD:
case SO_RERROR:
error = sooptcopyin(sopt, &optval, sizeof optval,
sizeof optval);
if (error)
goto bad;
SOCK_LOCK(so);
if (optval)
so->so_options |= sopt->sopt_name;
else
so->so_options &= ~sopt->sopt_name;
SOCK_UNLOCK(so);
break;
case SO_SETFIB:
error = sooptcopyin(sopt, &optval, sizeof optval,
sizeof optval);
if (error)
goto bad;
if (optval < 0 || optval >= rt_numfibs) {
error = EINVAL;
goto bad;
}
if (((so->so_proto->pr_domain->dom_family == PF_INET) ||
(so->so_proto->pr_domain->dom_family == PF_INET6) ||
(so->so_proto->pr_domain->dom_family == PF_ROUTE)))
so->so_fibnum = optval;
else
so->so_fibnum = 0;
break;
case SO_USER_COOKIE:
error = sooptcopyin(sopt, &val32, sizeof val32,
sizeof val32);
if (error)
goto bad;
so->so_user_cookie = val32;
break;
case SO_SNDBUF:
case SO_RCVBUF:
case SO_SNDLOWAT:
case SO_RCVLOWAT:
error = so->so_proto->pr_setsbopt(so, sopt);
if (error)
goto bad;
break;
case SO_SNDTIMEO:
case SO_RCVTIMEO:
#ifdef COMPAT_FREEBSD32
if (SV_CURPROC_FLAG(SV_ILP32)) {
struct timeval32 tv32;
error = sooptcopyin(sopt, &tv32, sizeof tv32,
sizeof tv32);
CP(tv32, tv, tv_sec);
CP(tv32, tv, tv_usec);
} else
#endif
error = sooptcopyin(sopt, &tv, sizeof tv,
sizeof tv);
if (error)
goto bad;
if (tv.tv_sec < 0 || tv.tv_usec < 0 ||
tv.tv_usec >= 1000000) {
error = EDOM;
goto bad;
}
if (tv.tv_sec > INT32_MAX)
val = SBT_MAX;
else
val = tvtosbt(tv);
SOCK_LOCK(so);
valp = sopt->sopt_name == SO_SNDTIMEO ?
(SOLISTENING(so) ? &so->sol_sbsnd_timeo :
&so->so_snd.sb_timeo) :
(SOLISTENING(so) ? &so->sol_sbrcv_timeo :
&so->so_rcv.sb_timeo);
*valp = val;
SOCK_UNLOCK(so);
break;
case SO_LABEL:
#ifdef MAC
error = sooptcopyin(sopt, &extmac, sizeof extmac,
sizeof extmac);
if (error)
goto bad;
error = mac_setsockopt_label(sopt->sopt_td->td_ucred,
so, &extmac);
#else
error = EOPNOTSUPP;
#endif
break;
case SO_TS_CLOCK:
error = sooptcopyin(sopt, &optval, sizeof optval,
sizeof optval);
if (error)
goto bad;
if (optval < 0 || optval > SO_TS_CLOCK_MAX) {
error = EINVAL;
goto bad;
}
so->so_ts_clock = optval;
break;
case SO_MAX_PACING_RATE:
error = sooptcopyin(sopt, &val32, sizeof(val32),
sizeof(val32));
if (error)
goto bad;
so->so_max_pacing_rate = val32;
break;
default:
+#ifdef SOCKET_HHOOK
if (V_socket_hhh[HHOOK_SOCKET_OPT]->hhh_nhooks > 0)
error = hhook_run_socket(so, sopt,
HHOOK_SOCKET_OPT);
else
+#endif
error = ENOPROTOOPT;
break;
}
if (error == 0 && so->so_proto->pr_ctloutput != NULL)
(void)(*so->so_proto->pr_ctloutput)(so, sopt);
}
bad:
CURVNET_RESTORE();
return (error);
}
/*
* Helper routine for getsockopt.
*/
int
sooptcopyout(struct sockopt *sopt, const void *buf, size_t len)
{
int error;
size_t valsize;
error = 0;
/*
* Documented get behavior is that we always return a value, possibly
* truncated to fit in the user's buffer. Traditional behavior is
* that we always tell the user precisely how much we copied, rather
* than something useful like the total amount we had available for
* her. Note that this interface is not idempotent; the entire
* answer must be generated ahead of time.
*/
valsize = min(len, sopt->sopt_valsize);
sopt->sopt_valsize = valsize;
if (sopt->sopt_val != NULL) {
if (sopt->sopt_td != NULL)
error = copyout(buf, sopt->sopt_val, valsize);
else
bcopy(buf, sopt->sopt_val, valsize);
}
return (error);
}
int
sogetopt(struct socket *so, struct sockopt *sopt)
{
int error, optval;
struct linger l;
struct timeval tv;
#ifdef MAC
struct mac extmac;
#endif
CURVNET_SET(so->so_vnet);
error = 0;
if (sopt->sopt_level != SOL_SOCKET) {
if (so->so_proto->pr_ctloutput != NULL)
error = (*so->so_proto->pr_ctloutput)(so, sopt);
else
error = ENOPROTOOPT;
CURVNET_RESTORE();
return (error);
} else {
switch (sopt->sopt_name) {
case SO_ACCEPTFILTER:
error = accept_filt_getopt(so, sopt);
break;
case SO_LINGER:
SOCK_LOCK(so);
l.l_onoff = so->so_options & SO_LINGER;
l.l_linger = so->so_linger;
SOCK_UNLOCK(so);
error = sooptcopyout(sopt, &l, sizeof l);
break;
case SO_USELOOPBACK:
case SO_DONTROUTE:
case SO_DEBUG:
case SO_KEEPALIVE:
case SO_REUSEADDR:
case SO_REUSEPORT:
case SO_REUSEPORT_LB:
case SO_BROADCAST:
case SO_OOBINLINE:
case SO_ACCEPTCONN:
case SO_TIMESTAMP:
case SO_BINTIME:
case SO_NOSIGPIPE:
case SO_NO_DDP:
case SO_NO_OFFLOAD:
case SO_RERROR:
optval = so->so_options & sopt->sopt_name;
integer:
error = sooptcopyout(sopt, &optval, sizeof optval);
break;
case SO_DOMAIN:
optval = so->so_proto->pr_domain->dom_family;
goto integer;
case SO_TYPE:
optval = so->so_type;
goto integer;
case SO_PROTOCOL:
optval = so->so_proto->pr_protocol;
goto integer;
case SO_ERROR:
SOCK_LOCK(so);
if (so->so_error) {
optval = so->so_error;
so->so_error = 0;
} else {
optval = so->so_rerror;
so->so_rerror = 0;
}
SOCK_UNLOCK(so);
goto integer;
case SO_SNDBUF:
optval = SOLISTENING(so) ? so->sol_sbsnd_hiwat :
so->so_snd.sb_hiwat;
goto integer;
case SO_RCVBUF:
optval = SOLISTENING(so) ? so->sol_sbrcv_hiwat :
so->so_rcv.sb_hiwat;
goto integer;
case SO_SNDLOWAT:
optval = SOLISTENING(so) ? so->sol_sbsnd_lowat :
so->so_snd.sb_lowat;
goto integer;
case SO_RCVLOWAT:
optval = SOLISTENING(so) ? so->sol_sbrcv_lowat :
so->so_rcv.sb_lowat;
goto integer;
case SO_SNDTIMEO:
case SO_RCVTIMEO:
SOCK_LOCK(so);
tv = sbttotv(sopt->sopt_name == SO_SNDTIMEO ?
(SOLISTENING(so) ? so->sol_sbsnd_timeo :
so->so_snd.sb_timeo) :
(SOLISTENING(so) ? so->sol_sbrcv_timeo :
so->so_rcv.sb_timeo));
SOCK_UNLOCK(so);
#ifdef COMPAT_FREEBSD32
if (SV_CURPROC_FLAG(SV_ILP32)) {
struct timeval32 tv32;
CP(tv, tv32, tv_sec);
CP(tv, tv32, tv_usec);
error = sooptcopyout(sopt, &tv32, sizeof tv32);
} else
#endif
error = sooptcopyout(sopt, &tv, sizeof tv);
break;
case SO_LABEL:
#ifdef MAC
error = sooptcopyin(sopt, &extmac, sizeof(extmac),
sizeof(extmac));
if (error)
goto bad;
error = mac_getsockopt_label(sopt->sopt_td->td_ucred,
so, &extmac);
if (error)
goto bad;
/* Don't copy out extmac, it is unchanged. */
#else
error = EOPNOTSUPP;
#endif
break;
case SO_PEERLABEL:
#ifdef MAC
error = sooptcopyin(sopt, &extmac, sizeof(extmac),
sizeof(extmac));
if (error)
goto bad;
error = mac_getsockopt_peerlabel(
sopt->sopt_td->td_ucred, so, &extmac);
if (error)
goto bad;
/* Don't copy out extmac, it is unchanged. */
#else
error = EOPNOTSUPP;
#endif
break;
case SO_LISTENQLIMIT:
optval = SOLISTENING(so) ? so->sol_qlimit : 0;
goto integer;
case SO_LISTENQLEN:
optval = SOLISTENING(so) ? so->sol_qlen : 0;
goto integer;
case SO_LISTENINCQLEN:
optval = SOLISTENING(so) ? so->sol_incqlen : 0;
goto integer;
case SO_TS_CLOCK:
optval = so->so_ts_clock;
goto integer;
case SO_MAX_PACING_RATE:
optval = so->so_max_pacing_rate;
goto integer;
default:
+#ifdef SOCKET_HHOOK
if (V_socket_hhh[HHOOK_SOCKET_OPT]->hhh_nhooks > 0)
error = hhook_run_socket(so, sopt,
HHOOK_SOCKET_OPT);
else
+#endif
error = ENOPROTOOPT;
break;
}
}
#ifdef MAC
bad:
#endif
CURVNET_RESTORE();
return (error);
}
int
soopt_getm(struct sockopt *sopt, struct mbuf **mp)
{
struct mbuf *m, *m_prev;
int sopt_size = sopt->sopt_valsize;
MGET(m, sopt->sopt_td ? M_WAITOK : M_NOWAIT, MT_DATA);
if (m == NULL)
return ENOBUFS;
if (sopt_size > MLEN) {
MCLGET(m, sopt->sopt_td ? M_WAITOK : M_NOWAIT);
if ((m->m_flags & M_EXT) == 0) {
m_free(m);
return ENOBUFS;
}
m->m_len = min(MCLBYTES, sopt_size);
} else {
m->m_len = min(MLEN, sopt_size);
}
sopt_size -= m->m_len;
*mp = m;
m_prev = m;
while (sopt_size) {
MGET(m, sopt->sopt_td ? M_WAITOK : M_NOWAIT, MT_DATA);
if (m == NULL) {
m_freem(*mp);
return ENOBUFS;
}
if (sopt_size > MLEN) {
MCLGET(m, sopt->sopt_td != NULL ? M_WAITOK :
M_NOWAIT);
if ((m->m_flags & M_EXT) == 0) {
m_freem(m);
m_freem(*mp);
return ENOBUFS;
}
m->m_len = min(MCLBYTES, sopt_size);
} else {
m->m_len = min(MLEN, sopt_size);
}
sopt_size -= m->m_len;
m_prev->m_next = m;
m_prev = m;
}
return (0);
}
int
soopt_mcopyin(struct sockopt *sopt, struct mbuf *m)
{
struct mbuf *m0 = m;
if (sopt->sopt_val == NULL)
return (0);
while (m != NULL && sopt->sopt_valsize >= m->m_len) {
if (sopt->sopt_td != NULL) {
int error;
error = copyin(sopt->sopt_val, mtod(m, char *),
m->m_len);
if (error != 0) {
m_freem(m0);
return(error);
}
} else
bcopy(sopt->sopt_val, mtod(m, char *), m->m_len);
sopt->sopt_valsize -= m->m_len;
sopt->sopt_val = (char *)sopt->sopt_val + m->m_len;
m = m->m_next;
}
if (m != NULL) /* should be allocated enoughly at ip6_sooptmcopyin() */
panic("ip6_sooptmcopyin");
return (0);
}
int
soopt_mcopyout(struct sockopt *sopt, struct mbuf *m)
{
struct mbuf *m0 = m;
size_t valsize = 0;
if (sopt->sopt_val == NULL)
return (0);
while (m != NULL && sopt->sopt_valsize >= m->m_len) {
if (sopt->sopt_td != NULL) {
int error;
error = copyout(mtod(m, char *), sopt->sopt_val,
m->m_len);
if (error != 0) {
m_freem(m0);
return(error);
}
} else
bcopy(mtod(m, char *), sopt->sopt_val, m->m_len);
sopt->sopt_valsize -= m->m_len;
sopt->sopt_val = (char *)sopt->sopt_val + m->m_len;
valsize += m->m_len;
m = m->m_next;
}
if (m != NULL) {
/* enough soopt buffer should be given from user-land */
m_freem(m0);
return(EINVAL);
}
sopt->sopt_valsize = valsize;
return (0);
}
/*
* sohasoutofband(): protocol notifies socket layer of the arrival of new
* out-of-band data, which will then notify socket consumers.
*/
void
sohasoutofband(struct socket *so)
{
if (so->so_sigio != NULL)
pgsigio(&so->so_sigio, SIGURG, 0);
selwakeuppri(&so->so_rdsel, PSOCK);
}
int
sopoll(struct socket *so, int events, struct ucred *active_cred,
struct thread *td)
{
/*
* We do not need to set or assert curvnet as long as everyone uses
* sopoll_generic().
*/
return (so->so_proto->pr_sopoll(so, events, active_cred, td));
}
int
sopoll_generic(struct socket *so, int events, struct ucred *active_cred,
struct thread *td)
{
int revents;
SOCK_LOCK(so);
if (SOLISTENING(so)) {
if (!(events & (POLLIN | POLLRDNORM)))
revents = 0;
else if (!TAILQ_EMPTY(&so->sol_comp))
revents = events & (POLLIN | POLLRDNORM);
else if ((events & POLLINIGNEOF) == 0 && so->so_error)
revents = (events & (POLLIN | POLLRDNORM)) | POLLHUP;
else {
selrecord(td, &so->so_rdsel);
revents = 0;
}
} else {
revents = 0;
SOCK_SENDBUF_LOCK(so);
SOCK_RECVBUF_LOCK(so);
if (events & (POLLIN | POLLRDNORM))
if (soreadabledata(so))
revents |= events & (POLLIN | POLLRDNORM);
if (events & (POLLOUT | POLLWRNORM))
if (sowriteable(so))
revents |= events & (POLLOUT | POLLWRNORM);
if (events & (POLLPRI | POLLRDBAND))
if (so->so_oobmark ||
(so->so_rcv.sb_state & SBS_RCVATMARK))
revents |= events & (POLLPRI | POLLRDBAND);
if ((events & POLLINIGNEOF) == 0) {
if (so->so_rcv.sb_state & SBS_CANTRCVMORE) {
revents |= events & (POLLIN | POLLRDNORM);
if (so->so_snd.sb_state & SBS_CANTSENDMORE)
revents |= POLLHUP;
}
}
if (so->so_rcv.sb_state & SBS_CANTRCVMORE)
revents |= events & POLLRDHUP;
if (revents == 0) {
if (events &
(POLLIN | POLLPRI | POLLRDNORM | POLLRDBAND | POLLRDHUP)) {
selrecord(td, &so->so_rdsel);
so->so_rcv.sb_flags |= SB_SEL;
}
if (events & (POLLOUT | POLLWRNORM)) {
selrecord(td, &so->so_wrsel);
so->so_snd.sb_flags |= SB_SEL;
}
}
SOCK_RECVBUF_UNLOCK(so);
SOCK_SENDBUF_UNLOCK(so);
}
SOCK_UNLOCK(so);
return (revents);
}
int
soo_kqfilter(struct file *fp, struct knote *kn)
{
struct socket *so = kn->kn_fp->f_data;
struct sockbuf *sb;
sb_which which;
struct knlist *knl;
switch (kn->kn_filter) {
case EVFILT_READ:
kn->kn_fop = &soread_filtops;
knl = &so->so_rdsel.si_note;
sb = &so->so_rcv;
which = SO_RCV;
break;
case EVFILT_WRITE:
kn->kn_fop = &sowrite_filtops;
knl = &so->so_wrsel.si_note;
sb = &so->so_snd;
which = SO_SND;
break;
case EVFILT_EMPTY:
kn->kn_fop = &soempty_filtops;
knl = &so->so_wrsel.si_note;
sb = &so->so_snd;
which = SO_SND;
break;
default:
return (EINVAL);
}
SOCK_LOCK(so);
if (SOLISTENING(so)) {
knlist_add(knl, kn, 1);
} else {
SOCK_BUF_LOCK(so, which);
knlist_add(knl, kn, 1);
sb->sb_flags |= SB_KNOTE;
SOCK_BUF_UNLOCK(so, which);
}
SOCK_UNLOCK(so);
return (0);
}
static void
filt_sordetach(struct knote *kn)
{
struct socket *so = kn->kn_fp->f_data;
so_rdknl_lock(so);
knlist_remove(&so->so_rdsel.si_note, kn, 1);
if (!SOLISTENING(so) && knlist_empty(&so->so_rdsel.si_note))
so->so_rcv.sb_flags &= ~SB_KNOTE;
so_rdknl_unlock(so);
}
/*ARGSUSED*/
static int
filt_soread(struct knote *kn, long hint)
{
struct socket *so;
so = kn->kn_fp->f_data;
if (SOLISTENING(so)) {
SOCK_LOCK_ASSERT(so);
kn->kn_data = so->sol_qlen;
if (so->so_error) {
kn->kn_flags |= EV_EOF;
kn->kn_fflags = so->so_error;
return (1);
}
return (!TAILQ_EMPTY(&so->sol_comp));
}
SOCK_RECVBUF_LOCK_ASSERT(so);
kn->kn_data = sbavail(&so->so_rcv) - so->so_rcv.sb_ctl;
if (so->so_rcv.sb_state & SBS_CANTRCVMORE) {
kn->kn_flags |= EV_EOF;
kn->kn_fflags = so->so_error;
return (1);
} else if (so->so_error || so->so_rerror)
return (1);
if (kn->kn_sfflags & NOTE_LOWAT) {
if (kn->kn_data >= kn->kn_sdata)
return (1);
} else if (sbavail(&so->so_rcv) >= so->so_rcv.sb_lowat)
return (1);
+#ifdef SOCKET_HHOOK
/* This hook returning non-zero indicates an event, not error */
return (hhook_run_socket(so, NULL, HHOOK_FILT_SOREAD));
+#else
+ return (0);
+#endif
}
static void
filt_sowdetach(struct knote *kn)
{
struct socket *so = kn->kn_fp->f_data;
so_wrknl_lock(so);
knlist_remove(&so->so_wrsel.si_note, kn, 1);
if (!SOLISTENING(so) && knlist_empty(&so->so_wrsel.si_note))
so->so_snd.sb_flags &= ~SB_KNOTE;
so_wrknl_unlock(so);
}
/*ARGSUSED*/
static int
filt_sowrite(struct knote *kn, long hint)
{
struct socket *so;
so = kn->kn_fp->f_data;
if (SOLISTENING(so))
return (0);
SOCK_SENDBUF_LOCK_ASSERT(so);
kn->kn_data = sbspace(&so->so_snd);
+#ifdef SOCKET_HHOOK
hhook_run_socket(so, kn, HHOOK_FILT_SOWRITE);
+#endif
if (so->so_snd.sb_state & SBS_CANTSENDMORE) {
kn->kn_flags |= EV_EOF;
kn->kn_fflags = so->so_error;
return (1);
} else if (so->so_error) /* temporary udp error */
return (1);
else if (((so->so_state & SS_ISCONNECTED) == 0) &&
(so->so_proto->pr_flags & PR_CONNREQUIRED))
return (0);
else if (kn->kn_sfflags & NOTE_LOWAT)
return (kn->kn_data >= kn->kn_sdata);
else
return (kn->kn_data >= so->so_snd.sb_lowat);
}
static int
filt_soempty(struct knote *kn, long hint)
{
struct socket *so;
so = kn->kn_fp->f_data;
if (SOLISTENING(so))
return (1);
SOCK_SENDBUF_LOCK_ASSERT(so);
kn->kn_data = sbused(&so->so_snd);
if (kn->kn_data == 0)
return (1);
else
return (0);
}
int
socheckuid(struct socket *so, uid_t uid)
{
if (so == NULL)
return (EPERM);
if (so->so_cred->cr_uid != uid)
return (EPERM);
return (0);
}
/*
* These functions are used by protocols to notify the socket layer (and its
* consumers) of state changes in the sockets driven by protocol-side events.
*/
/*
* Procedures to manipulate state flags of socket and do appropriate wakeups.
*
* Normal sequence from the active (originating) side is that
* soisconnecting() is called during processing of connect() call, resulting
* in an eventual call to soisconnected() if/when the connection is
* established. When the connection is torn down soisdisconnecting() is
* called during processing of disconnect() call, and soisdisconnected() is
* called when the connection to the peer is totally severed. The semantics
* of these routines are such that connectionless protocols can call
* soisconnected() and soisdisconnected() only, bypassing the in-progress
* calls when setting up a ``connection'' takes no time.
*
* From the passive side, a socket is created with two queues of sockets:
* so_incomp for connections in progress and so_comp for connections already
* made and awaiting user acceptance. As a protocol is preparing incoming
* connections, it creates a socket structure queued on so_incomp by calling
* sonewconn(). When the connection is established, soisconnected() is
* called, and transfers the socket structure to so_comp, making it available
* to accept().
*
* If a socket is closed with sockets on either so_incomp or so_comp, these
* sockets are dropped.
*
* If higher-level protocols are implemented in the kernel, the wakeups done
* here will sometimes cause software-interrupt process scheduling.
*/
void
soisconnecting(struct socket *so)
{
SOCK_LOCK(so);
so->so_state &= ~(SS_ISCONNECTED|SS_ISDISCONNECTING);
so->so_state |= SS_ISCONNECTING;
SOCK_UNLOCK(so);
}
void
soisconnected(struct socket *so)
{
bool last __diagused;
SOCK_LOCK(so);
so->so_state &= ~(SS_ISCONNECTING|SS_ISDISCONNECTING);
so->so_state |= SS_ISCONNECTED;
if (so->so_qstate == SQ_INCOMP) {
struct socket *head = so->so_listen;
int ret;
KASSERT(head, ("%s: so %p on incomp of NULL", __func__, so));
/*
* Promoting a socket from incomplete queue to complete, we
* need to go through reverse order of locking. We first do
* trylock, and if that doesn't succeed, we go the hard way
* leaving a reference and rechecking consistency after proper
* locking.
*/
if (__predict_false(SOLISTEN_TRYLOCK(head) == 0)) {
soref(head);
SOCK_UNLOCK(so);
SOLISTEN_LOCK(head);
SOCK_LOCK(so);
if (__predict_false(head != so->so_listen)) {
/*
* The socket went off the listen queue,
* should be lost race to close(2) of sol.
* The socket is about to soabort().
*/
SOCK_UNLOCK(so);
sorele_locked(head);
return;
}
last = refcount_release(&head->so_count);
KASSERT(!last, ("%s: released last reference for %p",
__func__, head));
}
again:
if ((so->so_options & SO_ACCEPTFILTER) == 0) {
TAILQ_REMOVE(&head->sol_incomp, so, so_list);
head->sol_incqlen--;
TAILQ_INSERT_TAIL(&head->sol_comp, so, so_list);
head->sol_qlen++;
so->so_qstate = SQ_COMP;
SOCK_UNLOCK(so);
solisten_wakeup(head); /* unlocks */
} else {
SOCK_RECVBUF_LOCK(so);
soupcall_set(so, SO_RCV,
head->sol_accept_filter->accf_callback,
head->sol_accept_filter_arg);
so->so_options &= ~SO_ACCEPTFILTER;
ret = head->sol_accept_filter->accf_callback(so,
head->sol_accept_filter_arg, M_NOWAIT);
if (ret == SU_ISCONNECTED) {
soupcall_clear(so, SO_RCV);
SOCK_RECVBUF_UNLOCK(so);
goto again;
}
SOCK_RECVBUF_UNLOCK(so);
SOCK_UNLOCK(so);
SOLISTEN_UNLOCK(head);
}
return;
}
SOCK_UNLOCK(so);
wakeup(&so->so_timeo);
sorwakeup(so);
sowwakeup(so);
}
void
soisdisconnecting(struct socket *so)
{
SOCK_LOCK(so);
so->so_state &= ~SS_ISCONNECTING;
so->so_state |= SS_ISDISCONNECTING;
if (!SOLISTENING(so)) {
SOCK_RECVBUF_LOCK(so);
socantrcvmore_locked(so);
SOCK_SENDBUF_LOCK(so);
socantsendmore_locked(so);
}
SOCK_UNLOCK(so);
wakeup(&so->so_timeo);
}
void
soisdisconnected(struct socket *so)
{
SOCK_LOCK(so);
/*
* There is at least one reader of so_state that does not
* acquire socket lock, namely soreceive_generic(). Ensure
* that it never sees all flags that track connection status
* cleared, by ordering the update with a barrier semantic of
* our release thread fence.
*/
so->so_state |= SS_ISDISCONNECTED;
atomic_thread_fence_rel();
so->so_state &= ~(SS_ISCONNECTING|SS_ISCONNECTED|SS_ISDISCONNECTING);
if (!SOLISTENING(so)) {
SOCK_UNLOCK(so);
SOCK_RECVBUF_LOCK(so);
socantrcvmore_locked(so);
SOCK_SENDBUF_LOCK(so);
sbdrop_locked(&so->so_snd, sbused(&so->so_snd));
socantsendmore_locked(so);
} else
SOCK_UNLOCK(so);
wakeup(&so->so_timeo);
}
int
soiolock(struct socket *so, struct sx *sx, int flags)
{
int error;
KASSERT((flags & SBL_VALID) == flags,
("soiolock: invalid flags %#x", flags));
if ((flags & SBL_WAIT) != 0) {
if ((flags & SBL_NOINTR) != 0) {
sx_xlock(sx);
} else {
error = sx_xlock_sig(sx);
if (error != 0)
return (error);
}
} else if (!sx_try_xlock(sx)) {
return (EWOULDBLOCK);
}
if (__predict_false(SOLISTENING(so))) {
sx_xunlock(sx);
return (ENOTCONN);
}
return (0);
}
void
soiounlock(struct sx *sx)
{
sx_xunlock(sx);
}
/*
* Make a copy of a sockaddr in a malloced buffer of type M_SONAME.
*/
struct sockaddr *
sodupsockaddr(const struct sockaddr *sa, int mflags)
{
struct sockaddr *sa2;
sa2 = malloc(sa->sa_len, M_SONAME, mflags);
if (sa2)
bcopy(sa, sa2, sa->sa_len);
return sa2;
}
/*
* Register per-socket destructor.
*/
void
sodtor_set(struct socket *so, so_dtor_t *func)
{
SOCK_LOCK_ASSERT(so);
so->so_dtor = func;
}
/*
* Register per-socket buffer upcalls.
*/
void
soupcall_set(struct socket *so, sb_which which, so_upcall_t func, void *arg)
{
struct sockbuf *sb;
KASSERT(!SOLISTENING(so), ("%s: so %p listening", __func__, so));
switch (which) {
case SO_RCV:
sb = &so->so_rcv;
break;
case SO_SND:
sb = &so->so_snd;
break;
}
SOCK_BUF_LOCK_ASSERT(so, which);
sb->sb_upcall = func;
sb->sb_upcallarg = arg;
sb->sb_flags |= SB_UPCALL;
}
void
soupcall_clear(struct socket *so, sb_which which)
{
struct sockbuf *sb;
KASSERT(!SOLISTENING(so), ("%s: so %p listening", __func__, so));
switch (which) {
case SO_RCV:
sb = &so->so_rcv;
break;
case SO_SND:
sb = &so->so_snd;
break;
}
SOCK_BUF_LOCK_ASSERT(so, which);
KASSERT(sb->sb_upcall != NULL,
("%s: so %p no upcall to clear", __func__, so));
sb->sb_upcall = NULL;
sb->sb_upcallarg = NULL;
sb->sb_flags &= ~SB_UPCALL;
}
void
solisten_upcall_set(struct socket *so, so_upcall_t func, void *arg)
{
SOLISTEN_LOCK_ASSERT(so);
so->sol_upcall = func;
so->sol_upcallarg = arg;
}
static void
so_rdknl_lock(void *arg)
{
struct socket *so = arg;
retry:
if (SOLISTENING(so)) {
SOLISTEN_LOCK(so);
} else {
SOCK_RECVBUF_LOCK(so);
if (__predict_false(SOLISTENING(so))) {
SOCK_RECVBUF_UNLOCK(so);
goto retry;
}
}
}
static void
so_rdknl_unlock(void *arg)
{
struct socket *so = arg;
if (SOLISTENING(so))
SOLISTEN_UNLOCK(so);
else
SOCK_RECVBUF_UNLOCK(so);
}
static void
so_rdknl_assert_lock(void *arg, int what)
{
struct socket *so = arg;
if (what == LA_LOCKED) {
if (SOLISTENING(so))
SOLISTEN_LOCK_ASSERT(so);
else
SOCK_RECVBUF_LOCK_ASSERT(so);
} else {
if (SOLISTENING(so))
SOLISTEN_UNLOCK_ASSERT(so);
else
SOCK_RECVBUF_UNLOCK_ASSERT(so);
}
}
static void
so_wrknl_lock(void *arg)
{
struct socket *so = arg;
retry:
if (SOLISTENING(so)) {
SOLISTEN_LOCK(so);
} else {
SOCK_SENDBUF_LOCK(so);
if (__predict_false(SOLISTENING(so))) {
SOCK_SENDBUF_UNLOCK(so);
goto retry;
}
}
}
static void
so_wrknl_unlock(void *arg)
{
struct socket *so = arg;
if (SOLISTENING(so))
SOLISTEN_UNLOCK(so);
else
SOCK_SENDBUF_UNLOCK(so);
}
static void
so_wrknl_assert_lock(void *arg, int what)
{
struct socket *so = arg;
if (what == LA_LOCKED) {
if (SOLISTENING(so))
SOLISTEN_LOCK_ASSERT(so);
else
SOCK_SENDBUF_LOCK_ASSERT(so);
} else {
if (SOLISTENING(so))
SOLISTEN_UNLOCK_ASSERT(so);
else
SOCK_SENDBUF_UNLOCK_ASSERT(so);
}
}
/*
* Create an external-format (``xsocket'') structure using the information in
* the kernel-format socket structure pointed to by so. This is done to
* reduce the spew of irrelevant information over this interface, to isolate
* user code from changes in the kernel structure, and potentially to provide
* information-hiding if we decide that some of this information should be
* hidden from users.
*/
void
sotoxsocket(struct socket *so, struct xsocket *xso)
{
bzero(xso, sizeof(*xso));
xso->xso_len = sizeof *xso;
xso->xso_so = (uintptr_t)so;
xso->so_type = so->so_type;
xso->so_options = so->so_options;
xso->so_linger = so->so_linger;
xso->so_state = so->so_state;
xso->so_pcb = (uintptr_t)so->so_pcb;
xso->xso_protocol = so->so_proto->pr_protocol;
xso->xso_family = so->so_proto->pr_domain->dom_family;
xso->so_timeo = so->so_timeo;
xso->so_error = so->so_error;
xso->so_uid = so->so_cred->cr_uid;
xso->so_pgid = so->so_sigio ? so->so_sigio->sio_pgid : 0;
if (SOLISTENING(so)) {
xso->so_qlen = so->sol_qlen;
xso->so_incqlen = so->sol_incqlen;
xso->so_qlimit = so->sol_qlimit;
xso->so_oobmark = 0;
} else {
xso->so_state |= so->so_qstate;
xso->so_qlen = xso->so_incqlen = xso->so_qlimit = 0;
xso->so_oobmark = so->so_oobmark;
sbtoxsockbuf(&so->so_snd, &xso->so_snd);
sbtoxsockbuf(&so->so_rcv, &xso->so_rcv);
}
}
int
so_options_get(const struct socket *so)
{
return (so->so_options);
}
void
so_options_set(struct socket *so, int val)
{
so->so_options = val;
}
int
so_error_get(const struct socket *so)
{
return (so->so_error);
}
void
so_error_set(struct socket *so, int val)
{
so->so_error = val;
}