Widen the ifnet_detach_sxlock to cover the entire vnet sysuninit code.
This ensures that we can't end up having the vnet_sysuninit free the UDP
pcb while the detach code is running and trying to purge the UDP pcb.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D28530
Historically receive buffer overflows have been ignored and programs
could not tell if they missed messages or messages had been truncated
because of overflows. Since programs historically do not expect to get
receive overflow errors, this behavior is not the default.
This is really really important for programs that use route(4) to keep in sync
with the system. If we loose a message then we need to reload the full system
state, otherwise the behaviour from that point is undefined and can lead
to chasing bogus bug reports.
ROUTE_MPATH was added to the GENERIC kernel in r368648.
According to the plan in D27428, it was enabled with `net.route.multipath` sysctl set to 0.
Given enough time has passed, this change enables route multipath by default.
The goal is to ship FreeBSD 13 with multipath turned on.
Reviewed By: donner, olivier
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D28423
Memory and PCI resources are freed with no particular order. This could
cause use-after-frees when detaching following a failed attach. For
instance, iflib_tx_structures_free() frees ctx->ifc_txqs[] but
iflib_tqg_detach() attempts to access this array. Similarly, adapter
queues gets freed by IFDI_QUEUES_FREE() but IFDI_DETACH() attempts to
access adapter queues to free PCI resources.
MFC after: 2 weeks
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D27634
Currently, if the immutable algorithm like bsearch or radix_lockless
receives rtable update notification, it schedules algorithm rebuild.
This rebuild is executed by the callout after ~50 milliseconds.
It is possible that a script adding an interface address and than route
with the gateway bound to that address will fail. It can happen due
to the fact that fib is not updated by the time the route addition
request arrives.
Fix this by allowing synchronous algorithm rebuilds based on certain
conditions. By default, these conditions assume:
1) less than net.route.algo.fib_sync_limit=100 routes
2) routes without gateway.
* Move algo instance build entirely under rib WLOCK.
Rib lock is only used for control plane (except radix algo, but there
are no rebuilds).
* Add rib_walk_ext_locked() function to allow RIB iteration with
rib lock already held.
* Fix rare potential callout use-after-free for fds by binding fd
callout to the relevant rib rmlock. In that case, callout_stop()
under rib WLOCK guarantees no callout will be executed afterwards.
MFC after: 3 days
subscriptions during RIB modifications.
Add new subscriptions to the beginning of the lists instead of
the end. This fixes the situation when new subscription is created
int the callback for the existing subscription, leading to the
subscription notification handler pick it.
MFC after: 3 days
* Move per-prefix debug lines under LOG_DEBUG2
* Create fib instance counter to distingush log messages between
instances
* Add more messages on rebuild reason.
MFC after: 3 days
The initial plan was to remove rib_lookup_info() before
FreeBSD 13. As several customers are still remaining,
fix rib_lookup_info() for the multipath use case.
D26436 introduced support for stacked vlans that changed the way vlans
are configured. In particular, this change broke setups that have
same-number vlans as subinterfaces.
Vlan support was initially created assuming "vlanX" semantics. In this paradigm,
automatic number assignment supported by cloning (ifconfig vlan create) was a
natural fit.
When "ifaceX.Y" support was added, allowing to have the same vlan number on
multiple devices, cloning code became more complex, as the is no
unified "vlan" namespace anymore. Such interfaces got the first spare
index from "vlan" cloner. This, in turn, led to the following problem:
ifconfig ix0.333 create -> index 1
ifconfig ix0.444 create -> index 2
ifconfig vlan2 create -> allocation failure
This change fixes such allocations by using cloning indexes only for
"vlanX" interfaces.
Reviewed by: hselasky
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D27505
Originally IFCAP_NOMAP meant that the mbuf has external storage pointer
that points to unmapped address. Then, this was extended to array of
such pointers. Then, such mbufs were augmented with header/trailer.
Basically, extended mbufs are extended, and set of features is subject
to change. The new name should be generic enough to avoid further
renaming.
tree that fix the ratelimit code. There were several bugs
in tcp_ratelimit itself and we needed further work to support
the multiple tag format coming for the joint TLS and Ratelimit dances.
Sponsored by: Netflix Inc.
Differential Revision: https://reviews.freebsd.org/D28357
This will allow userspace to more easily figure out if ALTQ is built
into the kernel and what disciplines are supported.
Reviewed by: donner@
Differential Revision: https://reviews.freebsd.org/D28302
The NS_MOREFRAG flag can be set in a netmap slot to represent a
multi-fragment packet. Only the last fragment of a packet does
not have the flag set. On TX rings, the flag may be set by the
userspace application. The kernel will look at the flag and use it
to properly set up the NIC TX descriptors.
On RX rings, the kernel may set the flag if the packet received
was split across multiple netmap buffers. The userspace application
should look at the flag to know when the packet is complete.
Submitted by: rajesh1.kumar_amd.com
Reviewed by: vmaffione
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D27799
rxd_frag_to_sd() have pf_rv parameter as NULL with the current
code. This patch fixes the NULL pointer dereference in that
case thus avoiding a possible panic.
Submitted by: rajesh1.kumar at amd.com
Reviewed by: gallatin
Differential Revision: https://reviews.freebsd.org/D28115
it gets unused.
Currently if_purgeifaddrs() uses in6_purgeaddr() to remove IPv6
ifaddrs. in6_purgeaddr() does not trrigger prefix removal if
number of linked ifas goes to 0, as this is a low-level function.
As a result, if_purgeifaddrs() purges all IPv4/IPv6 addresses but
keeps corresponding IPv6 prefixes.
Fix this by creating higher-level wrapper which handles unused
prefix usecase and use it in if_purgeifaddrs().
Differential revision: https://reviews.freebsd.org/D28128
rtinit[1]() is a function used to add or remove interface address prefix routes,
similar to ifa_maintain_loopback_route().
It was intended to be family-agnostic. There is a problem with this approach
in reality.
1) IPv6 code does not use it for the ifa routes. There is a separate layer,
nd6_prelist_(), providing interface for maintaining interface routes. Its part,
responsible for the actual route table interaction, mimics rtenty() code.
2) rtinit tries to combine multiple actions in the same function: constructing
proper route attributes and handling iterations over multiple fibs, for the
non-zero net.add_addr_allfibs use case. It notably increases the code complexity.
3) dstaddr handling. flags parameter re-uses RTF_ flags. As there is no special flag
for p2p connections, host routes and p2p routes are handled in the same way.
Additionally, mapping IFA flags to RTF flags makes the interface pretty messy.
It make rtinit() to clash with ifa_mainain_loopback_route() for IPV4 interface
aliases.
4) rtinit() is the last customer passing non-masked prefixes to rib_action(),
complicating rib_action() implementation.
5) rtinit() coupled ifa announce/withdrawal notifications, producing "false positive"
ifa messages in certain corner cases.
To address all these points, the following has been done:
* rtinit() has been split into multiple functions:
- Route attribute construction were moved to the per-address-family functions,
dealing with (2), (3) and (4).
- funnction providing net.add_addr_allfibs handling and route rtsock notificaions
is the new routing table inteface.
- rtsock ifa notificaion has been moved out as well. resulting set of funcion are only
responsible for the actual route notifications.
Side effects:
* /32 alias does not result in interface routes (/32 route and "host" route)
* RTF_PINNED is now set for IPv6 prefixes corresponding to the interface addresses
Differential revision: https://reviews.freebsd.org/D28186
Removed code iterates over if_addrhead and tries to remove
routes for each ifa.
This is exactly the thing that if_purgeaddrs() do, and
if_purgeaddr() is already called in the end.
Reviewed by: glebius
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D28106
In e86bddea9f sys/netpfil/pf/pf.h grew a
declaration of pf_get_ruleset_number. Now delete the old declaration
from sys/net/pfvar.h.
Reviewed by: kp
Sponsored by: Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D28081
Restore the hwofs functionality temporarily disabled by
7ba6ecf216 to prevent issues with iflib.
This patch brings the necessary changes to iflib to
enable howfs to allow interface restarts without
disrupting netmap applications actively using its
rings.
After this change, it becomes possible for multiple
non-cooperating netmap applications to use non-overlapping
subsets of the available netmap rings without clashing
with each other.
PR: 252453
MFC after: 1 week
The iflib_queues_alloc() allocates isc_nrxqs iflib_dma_info structs
for each rxqset, and links each struct to a different free list.
As a result, it must be isc_nrxqs >= isc_nfl (plus the completion
queue, if present).
Add an assertion to make this constraint explicit.
MFC after: 2 weeks
Since 1d238b07d5, krings are disabled before
a reinit cycle triggered by iflib_netmap_register.
However, this operation is actually necessary also for
any interface reinit triggered by other causes (i.e.,
ifconfig commands).
We achieve this goal by moving the krings enable/disable
operation inside iflib_stop() and iflib_init_locked().
Once here, this change also removes some redundant operations
from iflib_netmap_register(), that are already performed by
iflib_stop().
PR: 252453
MFC after: 1 week
When netmap_fl_refill() is called at initialization time (e.g.,
during netmap_iflib_register()), nic_i must be 0, since the
free list is reinitialized. At the end of the refill cycle, nic_i
must still be zero, because exactly N descriptors (N is the ring size)
are refilled.
This patch therefore fixes the assertions to check on nic_i rather
than on nm_i. The current netmap_reset() may in fact cause nm_i
to be != 0 while the device is resetting: this may happen when
multiple non-cooperating processes open different subsets of the
available netmap rings.
PR: 252518
MFC after: 1 week
When different processes open separate subsets of the
available rings of a same netmap interface, a device
reset may be performed while one of the processes
is actively using some rings (e.g., caused by another
process executing a nmport_open()).
With this patch, such situation will cause the
active process to get a POLLERR, so that it can
have a chance to detect the situation.
We also guarantee that no process is running a txsync
or rxsync (ioctl or poll) while an iflib device reset
is in progress.
PR: 252453
MFC after: 1 week
Doing a 'dd' over iscsi will reliably cause stalls. Tx
cleaning _should_ reliably happen as data is sent.
However, currently if the transmit queue fills it will
wait until the iflib timer (hz/2) runs.
This change causes the the tx taskq thread to be run
if there are completed descriptors.
While here:
- make timer interrupt delay a sysctl
- simplify txd_db_check handling
- comment on INTR types
Background on the change:
Initially doorbell updates were minimized by only writing to the register
on every fourth packet. If txq_drain would return without writing to the
doorbell it scheduled a callout on the next tick to do the doorbell write
to ensure that the write otherwise happened "soon". At that time a sysctl
was added for users to avoid the potential added latency by simply writing
to the doorbell register on every packet. This worked perfectly well for
e1000 and ixgbe ... and appeared to work well on ixl. However, as it
turned out there was a race to this approach that would lockup the ixl MAC.
It was possible for a lower producer index to be written after a higher one.
On e1000 and ixgbe this was harmless - on ixl it was fatal. My initial
response was to add a lock around doorbell writes - fixing the problem but
adding an unacceptable amount of lock contention.
The next iteration was to use transmit interrupts to drive delayed doorbell
writes. If there were no packets in the queue all doorbell writes would be
immediate as the queue started to fill up we could delay doorbell writes
further and further. At the start of drain if we've cleaned any packets we
know we've moved the state machine along and we write the doorbell (an
obvious missing optimization was to skip that doorbell write if db_pending
is zero). This change required that tx interrupts be scheduled periodically
as opposed to just when the hardware txq was full. However, that just leads
to our next problem.
Initially dedicated msix vectors were used for both tx and rx. However, it
was often possible to use up all available vectors before we set up all the
queues we wanted. By having rx and tx share a vector for a given queue we
could halve the number of vectors used by a given configuration. The problem
here is that with this change only e1000 passed the necessary value to have
the fast interrupt drive tx when appropriate.
Reported by: mav@
Tested by: mav@
Reviewed by: gallatin@
MFC after: 1 month
Sponsored by: iXsystems
Differential Revision: https://reviews.freebsd.org/D27683
Improve caching behaviour by using counter_u64 rather than variables
shared between cores.
The result of converting all counters to counter(9) (i.e. this full
patch series) is a significant improvement in throughput. As tested by
olivier@, on Intel Xeon E5-2697Av4 (16Cores, 32 threads) hardware with
Mellanox ConnectX-4 MCX416A-CCAT (100GBase-SR4) nics we see:
x FreeBSD 20201223: inet packets-per-second
+ FreeBSD 20201223 with pf patches: inet packets-per-second
+--------------------------------------------------------------------------+
| + |
| xx + |
|xxx +++|
||A| |
| |A||
+--------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 5 9216962 9526356 9343902 9371057.6 116720.36
+ 5 19427190 19698400 19502922 19546509 109084.92
Difference at 95.0% confidence
1.01755e+07 +/- 164756
108.584% +/- 2.9359%
(Student's t, pooled s = 112967)
Reviewed by: philip
MFC after: 2 weeks
Sponsored by: Orange Business Services
Differential Revision: https://reviews.freebsd.org/D27763
Factor out allocating and freeing pfi_kkif structures. This will be
useful when we change the counters to be counter_u64, so we don't have
to deal with that complexity in the multiple locations where we allocate
pfi_kkif structures.
No functional change.
MFC after: 2 weeks
Sponsored by: Orange Business Services
Differential Revision: https://reviews.freebsd.org/D27762
This improves the cache behaviour of pf and results in improved
throughput.
MFC after: 2 weeks
Sponsored by: Orange Business Services
Differential Revision: https://reviews.freebsd.org/D27760
The u_* counters are used only to communicate with userspace, as
userspace cannot use counter_u64. As pf_krule is not passed to userspace
these fields are now obsolete.
MFC after: 2 weeks
Sponsored by: Orange Business Services
Differential Revision: https://reviews.freebsd.org/D27759
As part of the split between user and kernel mode structures we're
moving all user space usable definitions into pf.h.
No functional change intended.
MFC after: 2 weeks
Sponsored by: Orange Business Services
Differential Revision: https://reviews.freebsd.org/D27757
Introduce a kernel version of struct pf_src_node (pf_ksrc_node).
This will allow us to improve the in-kernel data structure without
breaking userspace compatibility.
Reviewed by: philip
MFC after: 2 weeks
Sponsored by: Orange Business Services
Differential Revision: https://reviews.freebsd.org/D27707
Specifically implement the if_requestencap callback function for infiniband.
Most of the changes are simply a cut and paste of the equivalent ethernet part.
Reviewed by: melifaro @
Differential Revision: https://reviews.freebsd.org/D27631
MFC after: 1 week
Sponsored by: Mellanox Technologies // NVIDIA Networking
Need to update both link layer address and broadcast address when active link changes for IP over infiniband.
This is because the broadcast address contains the so-called P-key, which is interface dependent.
Reviewed by: kib @
Differential Revision: https://reviews.freebsd.org/D27658
MFC after: 1 week
Sponsored by: Mellanox Technologies // NVIDIA Networking
Use recently-added combination of `fib[46]_lookup_rt()` which
returns rtentry & raw nexthop with `rt_get_inet[6]_plen()` which
returns address/prefix length of prefix inside `rt`.
Add `nhop_select_func()` wrapper around inlined `nhop_select()` to
allow callers external to the routing subsystem select the proper
nexthop from the multipath group without including internal headers.
New calls does not require reference counting objects and reduce
the amount of copied/processed rtentry data.
Differential Revision: https://reviews.freebsd.org/D27675
This change introduces framework that allows to dynamically
attach or detach longest prefix match (lpm) lookup algorithms
to speed up datapath route tables lookups.
Framework takes care of handling initial synchronisation,
route subscription, nhop/nhop groups reference and indexing,
dataplane attachments and fib instance algorithm setup/teardown.
Framework features automatic algorithm selection, allowing for
picking the best matching algorithm on-the-fly based on the
amount of routes in the routing table.
Currently framework code is guarded under FIB_ALGO config option.
An idea is to enable it by default in the next couple of weeks.
The following algorithms are provided by default:
IPv4:
* bsearch4 (lockless binary search in a special IP array), tailored for
small-fib (<16 routes)
* radix4_lockless (lockless immutable radix, re-created on every rtable change),
tailored for small-fib (<1000 routes)
* radix4 (base system radix backend)
* dpdk_lpm4 (DPDK DIR24-8-based lookups), lockless datastrucure, optimized
for large-fib (D27412)
IPv6:
* radix6_lockless (lockless immutable radix, re-created on every rtable change),
tailed for small-fib (<1000 routes)
* radix6 (base system radix backend)
* dpdk_lpm6 (DPDK DIR24-8-based lookups), lockless datastrucure, optimized
for large-fib (D27412)
Performance changes:
Micro benchmarks (I7-7660U, single-core lookups, 2048k dst, code in D27604):
IPv4:
8 routes:
radix4: ~20mpps
radix4_lockless: ~24.8mpps
bsearch4: ~69mpps
dpdk_lpm4: ~67 mpps
700k routes:
radix4_lockless: 3.3mpps
dpdk_lpm4: 46mpps
IPv6:
8 routes:
radix6_lockless: ~20mpps
dpdk_lpm6: ~70mpps
100k routes:
radix6_lockless: 13.9mpps
dpdk_lpm6: 57mpps
Forwarding benchmarks:
+ 10-15% IPv4 forwarding performance (small-fib, bsearch4)
+ 25% IPv4 forwarding performance (full-view, dpdk_lpm4)
+ 20% IPv6 forwarding performance (full-view, dpdk_lpm6)
Control:
Framwork adds the following runtime sysctls:
List algos
* net.route.algo.inet.algo_list: bsearch4, radix4_lockless, radix4
* net.route.algo.inet6.algo_list: radix6_lockless, radix6, dpdk_lpm6
Debug level (7=LOG_DEBUG, per-route)
net.route.algo.debug_level: 5
Algo selection (currently only for fib 0):
net.route.algo.inet.algo: bsearch4
net.route.algo.inet6.algo: radix6_lockless
Support for manually changing algos in non-default fib will be added
soon. Some sysctl names will be changed in the near future.
Differential Revision: https://reviews.freebsd.org/D27401
When copying sockaddrs out to userspace, we pad them to a multiple of
the platform alignment (sizeof(long)). However, some sockaddr sizes,
such as struct sockaddr_dl, are not an integer multiple of the
alignment, so we may end up copying out uninitialized bytes.
Fix this by always bouncing through a pre-zeroed sockaddr_storage.
Reported by: KASAN
Reviewed by: melifaro
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D27729
This improves cache behaviour by not writing to the same variable from
multiple cores simultaneously.
pf_state is only used in the kernel, so can be safely modified.
Reviewed by: Lutz Donnerhacke, philip
MFC after: 1 week
Sponsed by: Orange Business Services
Differential Revision: https://reviews.freebsd.org/D27661
The algorithm we use to update checksums only works correctly if the
updated data is aligned on 16-bit boundaries (relative to the start of
the packet).
Import the OpenBSD fix for this issue.
PR: 240416
Obtained from: OpenBSD
MFC after: 1 week
Reviewed by: tuexen (previous version)
Differential Revision: https://reviews.freebsd.org/D27696
And switch from int to bool while at it.
Reviewed by: melifaro@
Differential Revision: https://reviews.freebsd.org/D27725
MFC after: 1 week
Sponsored by: Mellanox Technologies // NVIDIA Networking
The packet, if processed at this point, was already parsed to be UDP
directed to a vxlan port.
Connect-X 4+ does not provide easy method to infer which parser
processed the packet, so driver cannot set the flag without a lot of
efforts which are only to satisfy the formal requirements.
Reviewed by: bryanv, np
Sponsored by: Mellanox Technologies/NVidia Networking
Differential revision: https://reviews.freebsd.org/D27449
MFC after: 1 week
rtsock code was build around the assumption that each rtentry record
in the system radix tree is a ready-to-use sockaddr. This assumptions
turned out to be not quite true:
* masks have their length tweaked, so we have rtsock_fix_netmask() hack
* IPv6 addresses have their scope embedded, so we have another explicit
deembedding hack.
Change the code to decouple rtentry internals from rtsock code using
newly-created rtentry accessors. This will allow to eventually eliminate
both of the hacks and change rtentry dst/mask format.
Differential Revision: https://reviews.freebsd.org/D27451
struct ifconf and struct ifreq use the odd style "struct<tab>foo".
struct ifdrv seems to have tried to follow this but was committed with
spaces in place of most tabs resulting in "struct<space><space>ifdrv".
MFC after: 3 days
In some error paths we would fail to detach from the iflib taskqueue
groups. Also move the detach code into its own subroutine instead of
duplicating it.
Submitted by: Sai Rajesh Tallamraju <stallamr@netapp.com>
MFC after: 2 weeks
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D27342
Multiple consumers like ipfw, netflow or new route lookup algorithms
need to get the prefix data out of struct rtentry.
Instead of providing direct access to the rtentry, create IPv4/IPv6
accessors to abstract struct rtentry internals and avoid including
internal routing headers for external consumers.
While here, move struct route_nhop_data to the public header, so external
customers can actually use lookup functions returning rt&nhop data.
Differential Revision: https://reviews.freebsd.org/D27416
Revert the mitigation code for the vnet/epair cleanup race (done in r365457).
r368237 introduced a more reliable fix.
MFC after: 2 weeks
Sponsored by: Modirum MDPay
When destroying a vnet and an epair (with one end in the vnet) we often
panicked. This was the result of the destruction of the epair, which destroys
both ends simultaneously, happening while vnet_if_return() was moving the
struct ifnet to its home vnet. This can result in a freed ifnet being re-added
to the home vnet V_ifnet list. That in turn panics the next time the ifnet is
used.
Prevent this race by ensuring that vnet_if_return() cannot run at the same time
as if_detach() or epair_clone_destroy().
PR: 238870, 234985, 244703, 250870
MFC after: 2 weeks
Sponsored by: Modirum MDPay
Differential Revision: https://reviews.freebsd.org/D27378
ROUTE_MPATH is the new config option controlling new multipath routing
implementation. Remove the last pieces of RADIX_MPATH-related code and
the config option.
Reviewed by: glebius
Differential Revision: https://reviews.freebsd.org/D27244
It allows code within routing subsystem to transparently reference nexthops
and nexthop groups, similar to nhop_free_any(), abstracting ROUTE_MPATH
details.
Differential Revision: https://reviews.freebsd.org/D27410
No functional changes.
* Make lookup path of fib<4|6>_lookup_debugnet() separate functions
(fib<46>_lookup_rt()). These will be used in the control plane code
requiring unlocked radix operations and actual prefix pointer.
* Make lookup part of fib<4|6>_check_urpf() separate functions.
This change simplifies the switch to alternative lookup implementations,
which helps algorithmic lookups introduction.
* While here, use static initializers for IPv4/IPv6 keys
Differential Revision: https://reviews.freebsd.org/D27405
The resulting KPI can be used by routing table consumers to estimate the required
scale for route table export.
* Add tracking for rib routes
* Add accessors for number of nexthops/nexthop objects
* Simplify rib_unsubscribe: store rnh we're attached to instead of requiring it up
again on destruction. This helps in the cases when rnh is not linked yet/already unlinked.
Differential Revision: https://reviews.freebsd.org/D27404
Replace MAXPHYS by runtime variable maxphys. It is initialized from
MAXPHYS by default, but can be also adjusted with the tunable kern.maxphys.
Make b_pages[] array in struct buf flexible. Size b_pages[] for buffer
cache buffers exactly to atop(maxbcachebuf) (currently it is sized to
atop(MAXPHYS)), and b_pages[] for pbufs is sized to atop(maxphys) + 1.
The +1 for pbufs allow several pbuf consumers, among them vmapbuf(),
to use unaligned buffers still sized to maxphys, esp. when such
buffers come from userspace (*). Overall, we save significant amount
of otherwise wasted memory in b_pages[] for buffer cache buffers,
while bumping MAXPHYS to desired high value.
Eliminate all direct uses of the MAXPHYS constant in kernel and driver
sources, except a place which initialize maxphys. Some random (and
arguably weird) uses of MAXPHYS, e.g. in linuxolator, are converted
straight. Some drivers, which use MAXPHYS to size embeded structures,
get private MAXPHYS-like constant; their convertion is out of scope
for this work.
Changes to cam/, dev/ahci, dev/ata, dev/mpr, dev/mpt, dev/mvs,
dev/siis, where either submitted by, or based on changes by mav.
Suggested by: mav (*)
Reviewed by: imp, mav, imp, mckusick, scottl (intermediate versions)
Tested by: pho
Sponsored by: The FreeBSD Foundation
Differential revision: https://reviews.freebsd.org/D27225
When we terminate a vnet (i.e. jail) we move interfaces back to their home
vnet. We need to protect our access to the V_ifnet CK_LIST.
We could enter NET_EPOCH, but if_detach_internal() (called from if_vmove())
waits for net epoch callback completion. That's not possible from NET_EPOCH.
Instead, we take the IFNET_WLOCK, build a list of the interfaces that need to
move and, once we've released the lock, move them back to their home vnet.
We cannot hold the IFNET_WLOCK() during if_vmove(), because that results in a
LOR between ifnet_sx, in_multi_sx and iflib ctx lock.
Separate out moving the ifp into or out of V_ifnet, so we can hold the lock as
we do the list manipulation, but do not hold it as we if_vmove().
Reviewed by: melifaro
MFC after: 2 weeks
Sponsored by: Modirum MDPay
Differential Revision: https://reviews.freebsd.org/D27279
It no longer serves any purpose, as evidenced by the fact that we never take it
without ifnet_sxlock.
Sponsored by: Modirum MDPay
Differential Revision: https://reviews.freebsd.org/D27278
* Make rib_walk() order of arguments consistent with the rest of RIB api
* Add rib_walk_ext() allowing to exec callback before/after iteration.
* Rename rt_foreach_fib_walk_del -> rib_foreach_table_walk_del
* Rename rt_forach_fib_walk -> rib_foreach_table_walk
* Move rib_foreach_table_walk{_del} to route/route_helpers.c
* Slightly refactor rib_foreach_table_walk{_del} to make the implementation
consistent and prepare for upcoming iterator optimizations.
Differential Revision: https://reviews.freebsd.org/D27219
Don't use "new" as an identifier, and add explicit casts from void *.
As a general policy, FreeBSD doesn't make any C++ compatibility
guarantees for kernel headers like it does for userland, but it is a
small effort to do so in this case, to the benefit of a downstream
consumer (NetApp).
Reviewed by: rscheff
Sponsored by: NetApp, Inc.
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D27286
When links come and go, lacp goes into a "suppress distributing" mode
where it drops traffic for 3 seconds. When in this mode, lagg/lacp
historiclally drops traffic with ENETDOWN. That return value causes TCP
to close any connection where it gets that value back from the lower
parts of the stack. This means that any TCP connection with active
traffic during a 3-second windown when an LACP link comes or goes
would get closed.
TCP treats return values of ENOBUFS as transient errors, and re-schedules
transmission later. So rather than returning ENETDOWN, lets
return ENOBUFS instead. This allows TCP connections to be preserved.
I've tested this by repeatedly bouncing links on a Netlfix CDN server
under a moderate (20Gb/s) load and overved ENOBUFS reported back to
the TCP stack (as reported by a RACK TCP sysctl).
Reviewed by: jhb, jtl, rrs
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D27188
lagg(4) replaces if_output method of its child interfaces and expects
that this method can be called only by child interfaces. But it is
possible that lagg_port_output() could be called by children of child
interfaces. In this case ifnet's if_lagg field is NULL. Add check that
lp is not NULL.
Obtained from: Yandex LLC
MFC after: 1 week
Sponsored by: Yandex LLC
Add the missing static keyword present in the declaration.
Reviewed by: melifaro
Sponsored by: NetApp, Inc.
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D27024
The goal of the fib support is to provide multiple independent
routing tables, isolated from each other.
net.add_addr_allfibs default tries to shift gears in the opposite
direction, unconditionally inserting all addresses to all of the fibs.
There are use cases when this is necessary, however this is not a
default expected behaviour, especially compared to other implementations.
Provide WARNING message for the setups with multiple fibs to notify
potential users of the feature.
Differential Revision: https://reviews.freebsd.org/D26076
Stop advancing counter past the current iteration number at the start
of iteration. This removes the need of subtracting one when
calculating index for copyout, and arguably fixes off-by-one reporting
of copied out elements when copyout failed.
Reviewed by: hselasky
Sponsored by: Mellanox Technologies / NVidia Networking
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D27073
Use consistent output format for hex.
Print both media and mask where relevant.
Reviewed by: hselasky
Sponsored by: Mellanox Technologies/NVidia Networking
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D27034
This can be used to detect if an ethernet address is specifically an
IPv6 multicast address, defined in accordance to RFC 2464.
ETHER_IS_MULTICAST is still preferred in the general case.
Reviewed by: ae
Sponsored by: NetApp, Inc.
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D26611
This gives a more uniform API for send tag life cycle management.
Reviewed by: gallatin, hselasky
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D27000
- Add a new send tag type for a send tag that supports both rate
limiting (packet pacing) and TLS offload (mostly similar to D22669
but adds a separate structure when allocating the new tag type).
- When allocating a send tag for TLS offload, check to see if the
connection already has a pacing rate. If so, allocate a tag that
supports both rate limiting and TLS offload rather than a plain TLS
offload tag.
- When setting an initial rate on an existing ifnet KTLS connection,
set the rate in the TCP control block inp and then reset the TLS
send tag (via ktls_output_eagain) to reallocate a TLS + ratelimit
send tag. This allocates the TLS send tag asynchronously from a
task queue, so the TLS rate limit tag alloc is always sleepable.
- When modifying a rate on a connection using KTLS, look for a TLS
send tag. If the send tag is only a plain TLS send tag, assume we
failed to allocate a TLS ratelimit tag (either during the
TCP_TXTLS_ENABLE socket option, or during the send tag reset
triggered by ktls_output_eagain) and ignore the new rate. If the
send tag is a ratelimit TLS send tag, change the rate on the TLS tag
and leave the inp tag alone.
- Lock the inp lock when setting sb_tls_info for a socket send buffer
so that the routines in tcp_ratelimit can safely dereference the
pointer without needing to grab the socket buffer lock.
- Add an IFCAP_TXTLS_RTLMT capability flag and associated
administrative controls in ifconfig(8). TLS rate limit tags are
only allocated if this capability is enabled. Note that TLS offload
(whether unlimited or rate limited) always requires IFCAP_TXTLS[46].
Reviewed by: gallatin, hselasky
Relnotes: yes
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D26691
Code was supposed to call callout_reset_sbt_on() rather than
callout_reset_sbt(). This resulted into passing a "cpu" value
to a "flag" argument. A recipe for subtle errors.
PR: 248652
Reported by: sg@efficientip.com
MFC with: r367093
The way netmap TX is handled in iflib when TX interrupts are not
used (IFC_NETMAP_TX_IRQ not set) has some issues:
- The netmap_tx_irq() function gets called by iflib_timer(), which
gets scheduled with tick granularity (hz). This is not frequent
enough for 10Gbps NICs and beyond (e.g., ixgbe or ixl). The end
result is that the transmitting netmap application is not woken
up fast enough to saturate the link with small packets.
- The iflib_timer() functions also calls isc_txd_credits_update()
to ask for more TX completion updates. However, this violates
the netmap requirement that only txsync can access the TX queue
for datapath operations. Only netmap_tx_irq() may be called out
of the txsync context.
This change introduces per-tx-queue netmap timers, using microsecond
granularity to ensure that netmap_tx_irq() can be called often enough
to allow for maximum packet rate. The timer routine simply calls
netmap_tx_irq() to wake up the netmap application. The latter will
wake up and call txsync to collect TX completion updates.
This change brings back line rate speed with small packets for ixgbe.
For the time being, timer expiration is hardcoded to 90 microseconds,
in order to avoid introducing a new sysctl.
We may eventually implement an adaptive expiration period or use another
deferred work mechanism in place of timers.
Also, fix the timers usage to make sure that each queue is serviced
by a different CPU.
PR: 248652
Reported by: sg@efficientip.com
MFC after: 2 weeks
the failover protocol is supported due to limitations in the IPoIB
architecture. Refer to the lagg(4) manual page for how to configure
and use this new feature. A new network interface type,
IFT_INFINIBANDLAG, has been added, similar to the existing
IFT_IEEE8023ADLAG .
ifconfig(8) has been updated to accept a new laggtype argument when
creating lagg(4) network interfaces. This new argument is used to
distinguish between ethernet and infiniband type of lagg(4) network
interface. The laggtype argument is optional and defaults to
ethernet. The lagg(4) command line syntax is backwards compatible.
Differential Revision: https://reviews.freebsd.org/D26254
Reviewed by: melifaro@
MFC after: 1 week
Sponsored by: Mellanox Technologies // NVIDIA Networking
802.1ad interfaces are created with ifconfig using the "vlanproto" parameter.
Eg., the following creates a 802.1Q VLAN (id #42) over a 802.1ad S-VLAN
(id #5) over a physical Ethernet interface (em0).
ifconfig vlan5 create vlandev em0 vlan 5 vlanproto 802.1ad up
ifconfig vlan42 create vlandev vlan5 vlan 42 inet 10.5.42.1/24
VLAN_MTU, VLAN_HWCSUM and VLAN_TSO capabilities should be properly
supported. VLAN_HWTAGGING is only partially supported, as there is
currently no IFCAP_VLAN_* denoting the possibility to set the VLAN
EtherType to anything else than 0x8100 (802.1ad uses 0x88A8).
Submitted by: Olivier Piras
Sponsored by: RG Nets
Differential Revision: https://reviews.freebsd.org/D26436
connections over multiple paths.
Multipath routing relies on mbuf flowid data for both transit
and outbound traffic. Current code fills mbuf flowid from inp_flowid
for connection-oriented sockets. However, inp_flowid is currently
not calculated for outbound connections.
This change creates simple hashing functions and starts calculating hashes
for TCP,UDP/UDP-Lite and raw IP if multipath routes are present in the
system.
Reviewed by: glebius (previous version),ae
Differential Revision: https://reviews.freebsd.org/D26523
This flag is going to be used by IKE daemon to signal if
Extended Sequence Number feature is going to be used.
Value for this flag was taken from OpenBSD source code
6b4cbaf181
Submitted by: Patryk Duda <pdk@semihalf.com>
Reviewed by: ae
Differential revision: https://reviews.freebsd.org/D22366
Obtained from: Semihalf
Sponsored by: Stormshield
This adds a new IP_PROTO / IPV6_PROTO setsockopt (getsockopt)
option IP(V6)_VLAN_PCP, which can be set to -1 (interface
default), or explicitly to any priority between 0 and 7.
Note that for untagged traffic, explicitly adding a
priority will insert a special 801.1Q vlan header with
vlan ID = 0 to carry the priority setting
Reviewed by: gallatin, rrs
MFC after: 2 weeks
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D26409
We're not allowed to hold NET_EPOCH while sleeping, so when we call ioctl()
handlers for member interfaces we cannot be in NET_EPOCH. We still need some
protection of our CK_LISTs, so hold BRIDGE_LOCK instead.
That requires changing BRIDGE_LOCK into a sleepable lock, and separating the
BRIDGE_RT_LOCK, to protect bridge_rtnode lists. That lock is taken in the data
path (while in NET_EPOCH), so it cannot be a sleepable lock.
While here document the locking strategy.
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D26418
Both cxgbe(4) and mlx5(4) wrapped the existing send tag header with
their own identical headers that stored the type that the
type-specific tag structures inherited from, so in practice it seems
drivers need this in the tag anyway. This permits removing these
extra header indirections (struct cxgbe_snd_tag and struct
mlx5e_snd_tag).
In addition, this permits driver-independent code to query the type of
a tag, e.g. to know what type of tag is being queried via
if_snd_query.
Reviewed by: gallatin, hselasky, np, kib
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D26689
Nexthop lookup was not consireding rt_flags when doing
structure comparison, which lead to an original nexthop
selection when changing flags. Fix the case by adding
rt_flags field into comparison and rearranging nhop_priv
fields to allow for efficient matching.
Fix `route change X/Y flags` case - recent changes
disallowed specifying RTF_GATEWAY flag without actual gateway.
It turns out, route(8) fills in RTF_GATEWAY by default, unless
-interface flag is specified. Fix regression by clearing
RTF_GATEWAY flag instead of failing.
Fix route flag reporting in RTM_CHANGE messages by explicitly
updating rtm_flags after operation competion.
Add IPv4/IPv6 tests for flag-only route changes.
This change is based on the nexthop objects landed in D24232.
The change introduces the concept of nexthop groups.
Each group contains the collection of nexthops with their
relative weights and a dataplane-optimized structure to enable
efficient nexthop selection.
Simular to the nexthops, nexthop groups are immutable. Dataplane part
gets compiled during group creation and is basically an array of
nexthop pointers, compiled w.r.t their weights.
With this change, `rt_nhop` field of `struct rtentry` contains either
nexthop or nexthop group. They are distinguished by the presense of
NHF_MULTIPATH flag.
All dataplane lookup functions returns pointer to the nexthop object,
leaving nexhop groups details inside routing subsystem.
User-visible changes:
The change is intended to be backward-compatible: all non-mpath operations
should work as before with ROUTE_MPATH and net.route.multipath=1.
All routes now comes with weight, default weight is 1, maximum is 2^24-1.
Current maximum multipath group width is statically set to 64.
This will become sysctl-tunable in the followup changes.
Using functionality:
* Recompile kernel with ROUTE_MPATH
* set net.route.multipath to 1
route add -6 2001:db8::/32 2001:db8::2 -weight 10
route add -6 2001:db8::/32 2001:db8::3 -weight 20
netstat -6On
Nexthop groups data
Internet6:
GrpIdx NhIdx Weight Slots Gateway Netif Refcnt
1 ------- ------- ------- --------------------------------------- --------- 1
13 10 1 2001:db8::2 vlan2
14 20 2 2001:db8::3 vlan2
Next steps:
* Land outbound hashing for locally-originated routes ( D26523 ).
* Fix net/bird multipath (net/frr seems to work fine)
* Add ROUTE_MPATH to GENERIC
* Set net.route.multipath=1 by default
Tested by: olivier
Reviewed by: glebius
Relnotes: yes
Differential Revision: https://reviews.freebsd.org/D26449
For interfaces that do not support SIOCGIFMEDIA (for which there are
quite a few) the only fallback is to query the interface for
if_data->ifi_link_state. While it's possible to get at if_data for an
interface via getifaddrs(3) or sysctl, both are heavy weight mechanisms.
SIOCGIFDATA is a simple ioctl to retrieve this fast with very little
resource use in comparison. This implementation mirrors that of other
similar ioctls in FreeBSD.
Submitted by: Roy Marples <roy@marples.name>
Reviewed by: markj
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D26538
* Split rt_setmetrics into get_info_weight() and rt_set_expire_info(),
as these two can be applied at different entities and at different times.
* Start filling route weight in route change notifications
* Pass flowid to UDP/raw IP route lookups
* Rework nd6_subscription_cb() and sysctl_dumpentry() to prepare for the fact
that rtentry can contain multiple nexthops.
Differential Revision: https://reviews.freebsd.org/D26497
* Zero gw_sdl if switching to interface route - the assumption
that underlying storage is zeroed is incorrect with route changes.
* Apply proper flag mask to rte.
Reported by: vangyzen
This lets a VXLAN pseudo-interface take advantage of hardware checksumming (tx
and rx), TSO, and RSS if the NIC is capable of performing these operations on
inner VXLAN traffic.
A VXLAN interface inherits the capabilities of its vxlandev interface if one is
specified or of the interface that hosts the vxlanlocal address. If other
interfaces will carry traffic for that VXLAN then they must have the same
hardware capabilities.
On transmit, if_vxlan verifies that the outbound interface has the required
capabilities and then translates the CSUM_ flags to their inner equivalents.
This tells the hardware ifnet that it needs to operate on the inner frame and
not the outer VXLAN headers.
An event is generated when a VXLAN ifnet starts. This allows hardware drivers to
configure their devices to expect VXLAN traffic on the specified incoming port.
On receive, the hardware does RSS and checksum verification on the inner frame.
if_vxlan now does a direct netisr dispatch to take full advantage of RSS. It is
not very clear why it didn't do this already.
Future work:
Rx: it should be possible to avoid the first trip up the protocol stack to get
the frame to if_vxlan just so it can decapsulate and requeue for a second trip
up the stack. The hardware NIC driver could directly call an if_vxlan receive
routine for VXLAN traffic instead.
Rx: LRO. depends on what happens with the previous item. There will have to to
be a mechanism to indicate that it's time for if_vxlan to flush its LRO state.
Reviewed by: kib@
Relnotes: Yes
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D25873
These are similar to the existing VLAN capabilities.
Reviewed by: kib@
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D25873
Currently kernel requests deletion for the certain routes with specified gateway,
but this gateway is not actually checked. With multipath routes, internal gateway
checking becomes mandatory. Add the logic performing this check.
Generalise RTF_PINNED routes to the generic route priorities, simplifying the logic.
Add lookup_prefix() function to perform exact match search based on data in @info.
Differential Revision: https://reviews.freebsd.org/D26356
There's a race where dying vnets move their interfaces back to their original
vnet, and if_epair cleanup (where deleting one interface also deletes the other
end of the epair). This is commonly triggered by the pf tests, but also by
cleanup of vnet jails.
As we've not yet been able to fix the root cause of the issue work around the
panic by not dereferencing a NULL softc in epair_qflush() and by not
re-attaching DYING interfaces.
This isn't a full fix, but makes a very common panic far less likely.
PR: 244703, 238870
Reviewed by: lutz_donnerhacke.de
MFC after: 4 days
Differential Revision: https://reviews.freebsd.org/D26324
Use the same link-level gateway when adding or deleting interface routes.
This helps nexthop checking in the upcoming multipath changes.
Differential Revision: https://reviews.freebsd.org/D26317
After nexthop introduction, loopback routes for the interface addresses
were created without embedding actual interface index in the gateway.
The latter is needed to pass the IPv6 scope during transmission via loopback..
Fix the regression by actually using passed gateway data with interface index.
Differential Revision: https://reviews.freebsd.org/D26306
The pidx argument of isc_rxd_flush() indicates which is the last valid
receive descriptor to be used by the NIC. However, current code has
multiple issues:
- Intel drivers write pidx to their RDT register, which means that
NICs will only use the descriptors up to pidx-1 (modulo ring size N),
and won't actually use the one pointed by pidx. This does not break
reception, but it is anyway confusing and suboptimal (the NIC will
actually see only N-2 descriptors as available, rather than N-1).
Other drivers (if_vmx, if_bnxt, if_mgb) adhere to this semantic).
- The semantic used by Intel (RDT is one descriptor past the last
valid one) is used by most (if not all) NICs, and it is also used
on the TX side (also in iflib). Since iflib is not currently
using this semantic for RX, it must decrement fl->ifl_pidx
(modulo N) before calling isc_rxd_flush(), and then the
per-driver callback implementation must increment the index
again (to match the real semantic). This is confusing and suboptimal.
- The iflib refill function is also called at initialization.
However, in case the ring size is smaller than 128 (e.g. if_mgb),
the refill function will actually prepare all the receive
descriptors (N), without leaving one unused, as most of NICs assume
(e.g. to avoid RDT to overrun RDH). I can speculate that the code
looks like this right now because this issue showed up during
testing (e.g. with if_mgb), and it was easy to workaround by
decrementing pidx before isc_rxd_flush().
The goal of this change is to simplify the code (removing a bunch
of instructions from the RX fast path), and to make the semantic of
isc_rxd_flush() consistent across drivers. To achieve this, we:
- change the semantics of the pidx argument to the usual one (that
is the index one past the last valid one), so that both iflib and
drivers avoid the decrement/increment dance.
- fix the initialization code to prepare at most N-1 descriptors.
Reviewed by: markj
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D26191
No functional changes.
Initially this function was created to perform runtime flag conversions
for the previous incarnation of fib lookup functions. As these functions
got deprecated, move the function to the file with the only remaining
caller. Lastly, rename it to convert_rt_to_nh_flags() to follow the
naming notation.
No functional changes.
net/route/shared.h was created in the inital phases of nexthop conversion.
It was intended to serve the same purpose as route_var.h - share definitions
of functions and structures between the routing subsystem components. At
that time route_var.h was included by many files external to the routing
subsystem, which largerly defeats its purpose.
As currently this is not the case anymore and amount of route_var.h includes
is roughly the same as shared.h, retire the latter in favour of the former.
As nexthops are immutable, some operations such as route attribute changes
require nexthop fetching, forking, modification and route switching.
These operations are not atomic, so they may need to be retried multiple
times in presence of multiple speakers changing the same route.
This change introduces "synchronisation" primitive: route_update_conditional(),
simplifying logic for route changes and upcoming multipath operations.
Differential Revision: https://reviews.freebsd.org/D26216
At initialization time, the netmap RX refill function used to
prepare the NIC RX ring with N-1 buffers rather than N (with
N equal to the number of descriptors in the NIC RX ring).
This is not how netmap is supposed to work, as it would keep
kring->nr_hwcur not in sync with the NIC "next index to refill"
(i.e., fl->ifl_pidx). Instead we prepare N buffers, although we
still publish (with isc_rxd_flush()) only the first N-1 buffers,
to avoid the NIC producer pointer to overrun the NIC consumer
pointer (for NICs where this is a real issue, e.g. Intel ones).
MFC after: 2 weeks
rtentry lock traditionally served 2 purposed: first was protecting refcounts,
the second was assuring consistent field access/changes.
Since route nexthop introduction, the need for the former disappeared and
the need for the latter reduced.
To be more precise, the following rte field are mutable:
rt_nhop (nexthop pointer, updated with RIB_WLOCK, passed in rib_cmd_info)
rte_flags (only RTF_HOST and RTF_UP, where RTF_UP gets changed at rte removal)
rt_weight (relative weight, updated with RIB_WLOCK, passed in rib_cmd_info)
rt_expire (time when rte deletion is scheduled, updated with RIB_WLOCK)
rt_chain (deletion chain pointer, updated with RIB_WLOCK)
All of them are updated under RIB_WLOCK, so the only remaining concern is the reading.
rt_nhop and rt_weight (addressed in this review) are read under rib lock and
stored in the rib_cmd_info, so the caller has no problem with consitency.
rte_flags is currently read unlocked in rtsock reporting (however the scope
is only RTF_UP flag, which is pretty static).
rt_expire is currently read unlocked in rtsock reporting.
rt_chain accesses are safe, as this is only used at route deletion.
rt_expire and rte_flags reads will be dealt in a separate reviews soon.
Differential Revision: https://reviews.freebsd.org/D26162
The semantic of the pidx argument of isc_rxd_flush() is the
last valid index of in the free list, rather than the next
index to be published. However, netmap was still using the
old convention. While there, also refactor the netmap_fl_refill()
to simplify a little bit and add an assertion.
MFC after: 2 weeks
No functional changes.
Most of the routing flags are stored in the netxtop instead of rtentry.
Rename rt->rt_flags to rt->rte_flags to simplify reading/modifying code
checking routing flags.
In the new multipath code, rt->rt_nhop may actually point to nexthop group
instead of nhop. To ease transition, reduce the amount of rt->rt_nhop->...
accesses.
Differential Revision: https://reviews.freebsd.org/D26156
Allow to dynamically grow the amount of fibs in each vnet.
This change alters current behavior. Currently, if one defines
ROUTETABLES > 1 in the kernel config, each vnet will be created
with the number of fibs defined in the kernel config.
After this commit vnets will be created with fibs=1.
Dynamic net.fibs is not compatible with net.add_addr_allfibs.
The plan is to deprecate the latter and make
net.add_addr_allfibs=0 default behaviour.
Reviewed by: glebius
Relnotes: yes
Differential Revision: https://reviews.freebsd.org/D26062
