I'm not sure this is completely correct but at least this
is consistent with the accounting of incoming broadcasts.
PR: kern/65273
Submitted by: David J Duchscher <daved@tamu.edu>
For now, preserve the gif_called functionality to limit the nesting
level because uncontrolled nesting can easily cause the kernel stack
exhaustion. Rumors are it should be shot to allow people to easily
shoot themselves in the foot, but I have ran out of cartridges. ;)
+ struct ifnet: remove unused fields, move ipv6-related field close
to each other, add a pointer to l3<->l2 translation tables (arp,nd6,
etc.) for future use.
+ struct route: remove an unused field, move close to each
other some fields that might likely go away in the future
- Add tun_mtx to tun_softc. Annotate what is (and isn't) locked by it.
- Lock down tun_flags, tun_pid.
- In the output path, cache the value of tun_flags so it's consistent
when processing a particular packet rather than re-reading the field.
- In general, use unlocked reads for debugging.
- Annotate a couple of places where additional unlocked reads may be
possible.
- Annotate that tun_pid is used as a bug in tunopen().
if_tun is now largely MPSAFE, although questions remain about some of
the cdevsw fields and how they are synchronized.
other pseudo-interfaces, break out tear-down of a softc into a
separate tun_destroy() function, and invoke that from the module
unloader. Hold tunmtx across manipulations of the global softc list.
- Add gre_mtx to protect global softc list.
- Hold gre_mtx over various list operations (insert, delete).
- Centralize if_gre interface teardown in gre_destroy(), and call this
from modevent unload and gre_clone_destroy().
- Export gre_mtx to ip_gre.c, which walks the gre list to look up gre
interfaces during encapsulation. Add a wonking comment on how we need
some sort of drain/reference count mechanism to keep gre references
alive while in use and simultaneous destroy.
This commit does not lockdown softc data, which follows in a future
commit.
- Add gif_mtx, which protects globals.
- Hold gif_mtx around manipulation of gif_softc_list.
- Abstract gif destruction code into gif_destroy(), which tears down
a softc after it's been removed from the global list by either module
unload or clone destroy.
- Lock gif_called, even though we know gif_called is broken with reentrant
network processing.
- Document an event ordering problem in gif_set_tunnel() that will need
to be fixed.
gif_softc fields not locked down in this commit.
processing with gif interfaces, to a global variable named "gif_called".
Add an annotation that this approach will not work with a reentrant
network stack, and that we should instead use packet tags to detect
excessive recursive processing.
the tap driver, even with Giant over the cdev operation vector, due to
a non-atomic test-and-set of the si_drv1 field in the dev_t. This bug
exists with Giant under high memory pressure, as malloc() may sleep
in tapcreate(), but is less likely to occur. The resolution will
probably be to cover si_drv1 using the global tapmtx since no softc is
available, but I need to think about this problem more generally
across a range of drivers using si_drv1 in combination with SI_CHEAPCLONE
to defer expensive allocation to open().
Correct what appears to be a bug in the original if_tap implementation,
in which tapopen() will panic if a tap device instance is opened more
than once due to an incorrect assertion -- only triggered if INVARIANTS
is compiled in (i.e., when built into a kernel). Return EBUSY instead.
Expand mtx_lock() coverage using tp->tap_mtx to include tp->ether_addr.
as the process that opens tun_softc can exit before the file
descriptor is closed.
Taiwan experience provided by: keichii
Crashing breakers provided by: Chia-liang Kao <clkao@clkao.org>
(tap_pid, tap_flags). if_tap should now be entirely MPSAFE.
Committed from: Bamboo house by ocean in Taiwan
Tropical paradise provided by: Chia-liang Kao <clkao@clkao.org>
add tapmtx, which protects globale variables.
Notes:
- The EBUSY check in MOD_UNLOAD may be subject to a race. Moving the
event handler unregister inside the mutex grab may prevent that race.
- Locking of global variables safely is now possible because tapclones
is only modified when the module is loading or unloading, thanks to
phk's recent chang to clone_setup().
- softc locking to follow.
stf_destroy() to handle the common softc destruction path for the
two destruction sources: interface cloning destroy, and module
unload.
NOTE: sc_ro, the cached route for stf conversion, is not synchronized
against concurrent access in this change, that will follow in a future
change.
Reviewed by: pjd
Push if_faith softc destruction logic into faith_destroy() so that
it can be called after softc list removal in both the clone destroy
and module unload paths.
really sure why we have a softc list for if_loop, given that it
can't be unloaded, but that's an issue to revisit in the future as
corrupting the softc list would still cause panics.
Reviewed by: benno
Since there are two destroy paths for if_disc interfaces --
module unload and cloan interface destroy, create a new utility
function disc_destroy(), which is callded on a softc after it
has been removed from the global softc list; the cloaner and
module unload entry paths will both remove it before calling
disc_destroy().
Reviewed by: pjd
ifp is now passed explicitly to ether_demux; no need to look it up again.
Make mtag a global var in ip_input.
Noticed by: rwatson
Approved by: bms(mentor)
functions in kern_socket.c.
Rename the "canwait" field to "mflags" and pass M_WAITOK and M_NOWAIT
in from the caller context rather than "1" or "0".
Correct mflags pass into mac_init_socket() from previous commit to not
include M_ZERO.
Submitted by: sam
Assert the BPF descriptor lock in the MAC calls referencing live
BPF descriptors.
Obtained from: TrustedBSD Project
Sponsored by: DARPA, McAfee Research
BPF. Grab the BPF descriptor lock before entering MAC since the MAC
Framework references BPF descriptor fields, including the BPF
descriptor label.
Submitted by: sam
This enables pf to track dynamic address changes on interfaces (dailup) with
the "on (<ifname>)"-syntax. This also brings hooks in anticipation of
tracking cloned interfaces, which will be in future versions of pf.
Approved by: bms(mentor)
pf/pflog/pfsync as modules. Do not list them in NOTES or modules/Makefile
(i.e. do not connect it to any (automatic) builds - yet).
Approved by: bms(mentor)
Introduce d_version field in struct cdevsw, this must always be
initialized to D_VERSION.
Flip sense of D_NOGIANT flag to D_NEEDGIANT, this involves removing
four D_NOGIANT flags and adding 145 D_NEEDGIANT flags.
This commit adds a couple of functions for pseudodrivers to use for
implementing cloning in a manner we will be able to lock down (shortly).
Basically what happens is that pseudo drivers get a way to ask for
"give me the dev_t with this unit number" or alternatively "give
me a dev_t with the lowest guaranteed free unit number" (there is
unfortunately a lot of non-POLA in the exact numeric value of this
number, just live with it for now)
Managing the unit number space this way removes the need to use
rman(9) to do so in the drivers this greatly simplifies the code in
the drivers because even using rman(9) they still needed to manage
their dev_t's anyway.
I have taken the if_tun, if_tap, snp and nmdm drivers through the
mill, partly because they (ab)used makedev(), but mostly because
together they represent three different problems for device-cloning:
if_tun and snp is the plain case: just give me a device.
if_tap has two kinds of devices, with a flag for device type.
nmdm has paired devices (ala pty) can you can clone either of them.
Free approx 86 major numbers with a mostly automatically generated patch.
A number of strategic drivers have been left behind by caution, and a few
because they still (ab)use their major number.
- allow for ifp->if_ioctl being NULL, as the rest of ifioctl() does;
- give the interface driver a chance to report a error to the caller;
- don't forget to update ifp->if_lastchange upon successful modification
of interface operation parameters.
them mostly with packet tags (one case is handled by using an mbuf flag
since the linkage between "caller" and "callee" is direct and there's no
need to incur the overhead of a packet tag).
This is (mostly) work from: sam
Silence from: -arch
Approved by: bms(mentor), sam, rwatson
This is the first of two commits; bringing in the kernel support first.
This can be enabled by compiling a kernel with options TCP_SIGNATURE
and FAST_IPSEC.
For the uninitiated, this is a TCP option which provides for a means of
authenticating TCP sessions which came into being before IPSEC. It is
still relevant today, however, as it is used by many commercial router
vendors, particularly with BGP, and as such has become a requirement for
interconnect at many major Internet points of presence.
Several parts of the TCP and IP headers, including the segment payload,
are digested with MD5, including a shared secret. The PF_KEY interface
is used to manage the secrets using security associations in the SADB.
There is a limitation here in that as there is no way to map a TCP flow
per-port back to an SPI without polluting tcpcb or using the SPD; the
code to do the latter is unstable at this time. Therefore this code only
supports per-host keying granularity.
Whilst FAST_IPSEC is mutually exclusive with KAME IPSEC (and thus IPv6),
TCP_SIGNATURE applies only to IPv4. For the vast majority of prospective
users of this feature, this will not pose any problem.
This implementation is output-only; that is, the option is honoured when
responding to a host initiating a TCP session, but no effort is made
[yet] to authenticate inbound traffic. This is, however, sufficient to
interwork with Cisco equipment.
Tested with a Cisco 2501 running IOS 12.0(27), and Quagga 0.96.4 with
local patches. Patches for tcpdump to validate TCP-MD5 sessions are also
available from me upon request.
Sponsored by: sentex.net
The basic process is to send a routing socket announcement that the
interface has departed, change if_xname, update the sockaddr_dl
associated with the interface, and announce the arrival of the interface
on the routing socket.
As part of this change, ifunit() is greatly simplified by testing
if_xname directly. if_clone_destroy() now uses if_dname to look up the
cloner for the interface and if_dunit to identify the unit number.
Reviewed by: ru, sam (concept)
Vincent Jardin <vjardin AT free.fr>
Max Laier <max AT love2party.net>
- malloc() returns a void* and does not need a cast
- when called with M_WAITOK, malloc() can not return NULL so don't
check for that case. The result of the check was bogus anyway since
it would leave the interface broken.
ifconfig(8) flag since header for version 2 is the same but IP payload
is prepended with additional 4-bytes field.
Inspired by: Roman Synyuk <roman@univ.kiev.ua>
MFC after: 2 weeks
without IFCAP_VLAN_HWTAGGING. The previous version of the
leading comment in this file could lead to the opposite conclusion.
Fix some typos in the comment as well.
if_gre.c rev.1.41-1.49
o Spell output with two ts.
o Remove assigned-to but not used variable.
o fix grammatical error in a diagnostic message.
o u_short -> u_int16_t.
o gi_len is ip_len, so it has to be network byteorder.
if_gre.h rev.1.11-1.13
o prototype must not have variable name.
o u_short -> u_int16_t.
o Spell address with two d's.
ip_gre.c rev.1.22-1.29
o KNF - return is not a function.
o The "osrc" variable in gre_mobile_input() is only ever set but not
referenced; remove it.
o correct (false) assumptions on mbuf chain. not sure if it really helps, but
anyways, it is necessary to perform m_pullup.
o correct arg to m_pullup (need to count IP header size as well).
o remove redundant adjustment of m->m_pkthdr.len.
o clear m_flags just for safety.
o tabify.
o u_short -> u_int16_t.
MFC after: 2 weeks
a new bpf_mtap2 routine that does the right thing for an mbuf
and a variable-length chunk of data that should be prepended.
o while we're sweeping the drivers, use u_int32_t uniformly when
when prepending the address family (several places were assuming
sizeof(int) was 4)
o return M_ASSERTVALID to BPF_MTAP* now that all stack-allocated
mbufs have been eliminated; this may better be moved to the bpf
routines
Reviewed by: arch@ and several others
violated the constness were corrected before the freeze. This was
suggested by mdodd@, I think, and sam@ and others have signed off on
this if I recall my conversations with them correctly.
mpf are allocated on the stack, which causes this check to falsely trigger.
A new check which takes on-stack mbufs into account will be reintroduced
after 5.2 is out the door.
Approved by: re (watson)
Requested by: many
the routing table. Move all usage and references in the tcp stack
from the routing table metrics to the tcp hostcache.
It caches measured parameters of past tcp sessions to provide better
initial start values for following connections from or to the same
source or destination. Depending on the network parameters to/from
the remote host this can lead to significant speedups for new tcp
connections after the first one because they inherit and shortcut
the learning curve.
tcp_hostcache is designed for multiple concurrent access in SMP
environments with high contention and is hash indexed by remote
ip address.
It removes significant locking requirements from the tcp stack with
regard to the routing table.
Reviewed by: sam (mentor), bms
Reviewed by: -net, -current, core@kame.net (IPv6 parts)
Approved by: re (scottl)
accordingly. The define is left intact for ABI compatibility
with userland.
This is a pre-step for the introduction of tcp_hostcache. The
network stack remains fully useable with this change.
Reviewed by: sam (mentor), bms
Reviewed by: -net, -current, core@kame.net (IPv6 parts)
Approved by: re (scottl)
the MAC label referenced from 'struct socket' in the IPv4 and
IPv6-based protocols. This permits MAC labels to be checked during
network delivery operations without dereferencing inp->inp_socket
to get to so->so_label, which will eventually avoid our having to
grab the socket lock during delivery at the network layer.
This change introduces 'struct inpcb' as a labeled object to the
MAC Framework, along with the normal circus of entry points:
initialization, creation from socket, destruction, as well as a
delivery access control check.
For most policies, the inpcb label will simply be a cache of the
socket label, so a new protocol switch method is introduced,
pr_sosetlabel() to notify protocols that the socket layer label
has been updated so that the cache can be updated while holding
appropriate locks. Most protocols implement this using
pru_sosetlabel_null(), but IPv4/IPv6 protocols using inpcbs use
the the worker function in_pcbsosetlabel(), which calls into the
MAC Framework to perform a cache update.
Biba, LOMAC, and MLS implement these entry points, as do the stub
policy, and test policy.
Reviewed by: sam, bms
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories
Short description of ip_fastforward:
o adds full direct process-to-completion IPv4 forwarding code
o handles ip fragmentation incl. hw support (ip_flow did not)
o sends icmp needfrag to source if DF is set (ip_flow did not)
o supports ipfw and ipfilter (ip_flow did not)
o supports divert, ipfw fwd and ipfilter nat (ip_flow did not)
o returns anything it can't handle back to normal ip_input
Enable with sysctl -w net.inet.ip.fastforwarding=1
Reviewed by: sam (mentor)
be printed, if the module were loaded into a kernel which had INET6 enabled.
The gre(4) driver does not use INET6, nor is it specified for IPv6. The
tunnel_status() function in ifconfig(8) is somewhat overzealous and assumes
that all tunnel interfaces speak KAME ifioctls.
This fix follows the path of least resistance, by teaching gre(4) about
the two KAME ifioctls concerned.
PR: bin/56341
in various kernel objects to represent security data, we embed a
(struct label *) pointer, which now references labels allocated using
a UMA zone (mac_label.c). This allows the size and shape of struct
label to be varied without changing the size and shape of these kernel
objects, which become part of the frozen ABI with 5-STABLE. This opens
the door for boot-time selection of the number of label slots, and hence
changes to the bound on the number of simultaneous labeled policies
at boot-time instead of compile-time. This also makes it easier to
embed label references in new objects as required for locking/caching
with fine-grained network stack locking, such as inpcb structures.
This change also moves us further in the direction of hiding the
structure of kernel objects from MAC policy modules, not to mention
dramatically reducing the number of '&' symbols appearing in both the
MAC Framework and MAC policy modules, and improving readability.
While this results in minimal performance change with MAC enabled, it
will observably shrink the size of a number of critical kernel data
structures for the !MAC case, and should have a small (but measurable)
performance benefit (i.e., struct vnode, struct socket) do to memory
conservation and reduced cost of zeroing memory.
NOTE: Users of MAC must recompile their kernel and all MAC modules as a
result of this change. Because this is an API change, third party
MAC modules will also need to be updated to make less use of the '&'
symbol.
Suggestions from: bmilekic
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories
thread being waken up. The thread waken up can run at a priority as
high as after tsleep().
- Replace selwakeup()s with selwakeuppri()s and pass appropriate
priorities.
- Add cv_broadcastpri() which raises the priority of the broadcast
threads. Used by selwakeuppri() if collision occurs.
Not objected in: -arch, -current
whether or not the isr needs to hold Giant when running; Giant-less
operation is also controlled by the setting of debug_mpsafenet
o mark all netisr's except NETISR_IP as needing Giant
o add a GIANT_REQUIRED assertion to the top of netisr's that need Giant
o pickup Giant (when debug_mpsafenet is 1) inside ip_input before
calling up with a packet
o change netisr handling so swi_net runs w/o Giant; instead we grab
Giant before invoking handlers based on whether the handler needs Giant
o change netisr handling so that netisr's that are marked MPSAFE may
have multiple instances active at a time
o add netisr statistics for packets dropped because the isr is inactive
Supported by: FreeBSD Foundation
o move it from subr_bus.c to netisr.c where it more properly belongs
o add NET_PICKUP_GIANT and NET_DROP_GIANT macros that will be used to
grab Giant as needed when MPSAFE operation is enabled
Supported by: FreeBSD Foundation
if_xname, if_dname, and if_dunit. if_xname is the name of the interface
and if_dname/unit are the driver name and instance.
This change paves the way for interface renaming and enhanced pseudo
device creation and configuration symantics.
Approved By: re (in principle)
Reviewed By: njl, imp
Tested On: i386, amd64, sparc64
Obtained From: NetBSD (if_xname)
routine that takes a locked routing table reference and removes all
references to the entry in the various data structures. This
eliminates instances of recursive locking and also closes races
where the lock on the entry had to be dropped prior to calling
rtrequest(RTM_DELETE). This also cleans up confusion where the
caller held a reference to an entry that might have been reclaimed
(and in some cases used that reference).
Supported by: FreeBSD Foundation
direct dispatch) to avoid extensive kernel stack usage and to
avoid directly re-entering the network stack. The latter causes
locking problems when, for example, a complete TCP handshake`
happens w/o a context switch.
with an mbuf until it is reclaimed. This is in contrast to tags that
vanish when an mbuf chain passes through an interface. Persistent tags
are used, for example, by MAC labels.
Add an m_tag_delete_nonpersistent function to strip non-persistent tags
from mbufs and use it to strip such tags from packets as they pass through
the loopback interface and when turned around by icmp. This fixes problems
with "tag leakage".
Pointed out by: Jonathan Stone
Reviewed by: Robert Watson
a minor conflict):
o Use ETHER_ADDR_LEN in preference to '6'.
o Remove two unnecessary (caddr_t) casts. One of them causes problems in
my tree where etherbroadcastaddr is const, and (caddr_t) casts the const
away.
is non-free. (More checks can/should be added in the future.)
Use M_ASSERTVALID in BPF_MTAP so that we catch when freed mbufs are
passed in, even if no bpf listeners are active.
Inspired by a bug in if_dc caught by Kenjiro Cho.
that covers updates to the contents. Note this is separate from holding
a reference and/or locking the routing table itself.
Other/related changes:
o rtredirect loses the final parameter by which an rtentry reference
may be returned; this was never used and added unwarranted complexity
for locking.
o minor style cleanups to routing code (e.g. ansi-fy function decls)
o remove the logic to bump the refcnt on the parent of cloned routes,
we assume the parent will remain as long as the clone; doing this avoids
a circularity in locking during delete
o convert some timeouts to MPSAFE callouts
Notes:
1. rt_mtx in struct rtentry is guarded by #ifdef _KERNEL as user-level
applications cannot/do-no know about mutex's. Doing this requires
that the mutex be the last element in the structure. A better solution
is to introduce an externalized version of struct rtentry but this is
a major task because of the intertwining of rtentry and other data
structures that are visible to user applications.
2. There are known LOR's that are expected to go away with forthcoming
work to eliminate many held references. If not these will be resolved
prior to release.
3. ATM changes are untested.
Sponsored by: FreeBSD Foundation
Obtained from: BSD/OS (partly)
any queued packets for the isr, process those packets before the newly
submitted packet, maintaining ordering of all packets being delivered
to the netisr. Remove the bypass counter since we don't bypass anymore.
Leave the comment about possible problems and options since later
performance optimization may change the strategy for addressing ordering
problems here.
Specifically, this maintains the strong isr ordering guarantee; additional
parallelism and lower latency may be possible by moving to weaker
guarantees (per-interface, for example). We will probably at some point
also want to remove the one instance netisr dispatch limit currently
enforced by a mutex, but it's not clear that's 100% safe yet, even in
the netperf branch.
Reviewed by: sam, others
o move route_cb to be private to rtsock.c
o replace global static route_proto by locals
o eliminate global #define shorthands for info references
o remove some register decls
o ansi-fy function decls
o move items to be close in scope to their usage
o add rt_dispatch function for dispatching the actual message
o cleanup tangled logic for doing all-but-me msg send
Support by: FreeBSD Foundation
(direct dispatch) in interrupt threads when the netisr in question
isn't already active. If a netisr is already active, or direct
dispatch is already in progress, we queue the packet for later
delivery. Previously, this option was disabled by default. I have
measured 20%+ performance improvements in IP packet forwarding with
this enabled.
Please report any problems ASAP, especially relating to stack depth or
out-of-order packet processing.
Discussed with: jlemon, peter
Sponsored by: DARPA, Network Associates Laboratories
then the mbuf has been consumed by a hook; otherwise beware of a null
mbuf return (gack). In particular the bridge was doing the wrong thing.
While in the ipv6 code make it's handling of pfil_run_hooks identical
to netbsd.
Pointed out by: Pyun YongHyeon <yongari@kt-is.co.kr>
even could call VOP_REVOKE() on vnodes associated with its dev_t's
has originated, but it stops right here.
If there are things people belive destroy_dev() needs to learn how to
do, please tell me about it, preferably with a reproducible test case.
Include <sys/uio.h> in bluetooth code rather than rely on <sys/vnode.h>
to do so.
The fact that some of the USB code needs to include <sys/vnode.h>
still disturbs me greatly, but I do not have time to chase that.
o revamp IPv4+IPv6+bridge usage to match API changes
o remove pfil_head instances from protosw entries (no longer used)
o add locking
o bump FreeBSD version for 3rd party modules
Heavy lifting by: "Max Laier" <max@love2party.net>
Supported by: FreeBSD Foundation
Obtained from: NetBSD (bits of pfil.h and pfil.c)
o replace magic constants with #defines (e.g. ETHER_ADDR_LEN)
o move mib variables to net.link.ether.bridge with backwards compatible
entries for well-known items maintained under BURN_BRIDGES
o revamp debugging support so it is conditioanlly compiled with BRIDGE_DEBUG
(on currently) and runtime controlled by net.link.ether.bridge.debug
o change timeout to MPSAFE callout
o optimize lookup for common case of two interfaces
o optimize forwarding path to take IFNET lock only when needed
o make boot-time printf dependent on bootverbose
o sundry style changes (ANSI decls, extraneous spaces, etc.)
Sponsored by: FreeBSD Foundation
