well as sets in some of the groundwork for committing BBR. The
hpts system is updated as well as some other needed utilities
for the entrance of BBR. This is actually part 1 of 3 more
needed commits which will finally complete with BBRv1 being
added as a new tcp stack.
Sponsored by: Netflix Inc.
Differential Revision: https://reviews.freebsd.org/D20834
Unmapped mbufs allow sendfile to carry multiple pages of data in a
single mbuf, without mapping those pages. It is a requirement for
Netflix's in-kernel TLS, and provides a 5-10% CPU savings on heavy web
serving workloads when used by sendfile, due to effectively
compressing socket buffers by an order of magnitude, and hence
reducing cache misses.
For this new external mbuf buffer type (EXT_PGS), the ext_buf pointer
now points to a struct mbuf_ext_pgs structure instead of a data
buffer. This structure contains an array of physical addresses (this
reduces cache misses compared to an earlier version that stored an
array of vm_page_t pointers). It also stores additional fields needed
for in-kernel TLS such as the TLS header and trailer data that are
currently unused. To more easily detect these mbufs, the M_NOMAP flag
is set in m_flags in addition to M_EXT.
Various functions like m_copydata() have been updated to safely access
packet contents (using uiomove_fromphys()), to make things like BPF
safe.
NIC drivers advertise support for unmapped mbufs on transmit via a new
IFCAP_NOMAP capability. This capability can be toggled via the new
'nomap' and '-nomap' ifconfig(8) commands. For NIC drivers that only
transmit packet contents via DMA and use bus_dma, adding the
capability to if_capabilities and if_capenable should be all that is
required.
If a NIC does not support unmapped mbufs, they are converted to a
chain of mapped mbufs (using sf_bufs to provide the mapping) in
ip_output or ip6_output. If an unmapped mbuf requires software
checksums, it is also converted to a chain of mapped mbufs before
computing the checksum.
Submitted by: gallatin (earlier version)
Reviewed by: gallatin, hselasky, rrs
Discussed with: ae, kp (firewalls)
Relnotes: yes
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D20616
instead of a linear array.
The multicast memberships for the inpcb structure are protected by a
non-sleepable lock, INP_WLOCK(), which needs to be dropped when
calling the underlying possibly sleeping if_ioctl() method. When using
a linear array to keep track of multicast memberships, the computed
memory location of the multicast filter may suddenly change, due to
concurrent insertion or removal of elements in the linear array. This
in turn leads to various invalid memory access issues and kernel
panics.
To avoid this problem, put all multicast memberships on a STAILQ based
list. Then the memory location of the IPv4 and IPv6 multicast filters
become fixed during their lifetime and use after free and memory leak
issues are easier to track, for example by: vmstat -m | grep multi
All list manipulation has been factored into inline functions
including some macros, to easily allow for a future hash-list
implementation, if needed.
This patch has been tested by pho@ .
Differential Revision: https://reviews.freebsd.org/D20080
Reviewed by: markj @
MFC after: 1 week
Sponsored by: Mellanox Technologies
With this opcode it is possible to match TCP packets with specified
MSS option, whose value corresponds to configured in opcode value.
It is allowed to specify single value, range of values, or array of
specific values or ranges. E.g.
# ipfw add deny log tcp from any to any tcpmss 0-500
Reviewed by: melifaro,bcr
Obtained from: Yandex LLC
MFC after: 1 week
Sponsored by: Yandex LLC
If we take the slow path for forwarding we should still tell our
firewalls (hooked through pfil(9)) that we're forwarding. Pass the
ip_output() flags to ip_output_pfil() so it can set the PFIL_FWD flag
when we're forwarding.
MFC after: 1 week
Sponsored by: Axiado
in response to SACKs. The default behavior is unchanged; however, the limit
can be activated by changing the new net.inet.tcp.rack.split_limit sysctl.
Submitted by: Peter Lei <peterlei@netflix.com>
Reported by: jtl
Reviewed by: lstewart (earlier version)
Security: CVE-2019-5599
rename the source to gsb_crc32.c.
This is a prerequisite of unifying kernel zlib instances.
PR: 229763
Submitted by: Yoshihiro Ota <ota at j.email.ne.jp>
Differential Revision: https://reviews.freebsd.org/D20193
Apply a linker script when linking i386 kernel modules to apply padding
to a set_pcpu or set_vnet section. The padding value is kind-of random
and is used to catch modules not compiled with the linker-script, so
possibly still having problems leading to kernel panics.
This is needed as the code generated on certain architectures for
non-simple-types, e.g., an array can generate an absolute relocation
on the edge (just outside) the section and thus will not be properly
relocated. Adding the padding to the end of the section will ensure
that even absolute relocations of complex types will be inside the
section, if they are the last object in there and hence relocation will
work properly and avoid panics such as observed with carp.ko or ipsec.ko.
There is a rather lengthy discussion of various options to apply in
the mentioned PRs and their depends/blocks, and the review.
There seems no best solution working across multiple toolchains and
multiple version of them, so I took the liberty of taking one,
as currently our users (and our CI system) are hitting this on
just i386 and we need some solution. I wish we would have a proper
fix rather than another "hack".
Also backout r340009 which manually, temporarily fixed CARP before 12.0-R
"by chance" after a lead-up of various other link-elf.c and related fixes.
PR: 230857,238012
With suggestions from: arichardson (originally last year)
Tested by: lwhsu
Event: Waterloo Hackathon 2019
Reported by: lwhsu, olivier
MFC after: 6 weeks
Differential Revision: https://reviews.freebsd.org/D17512
The corresponding changes for the RACK stack where missed and are added
by this commit.
Reviewed by: Richard Scheffenegger, rrs@
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D20372
another one in udp_output(). This one is a race condition.
We do check on the laddr and lport without holding a lock in
order to determine whether we want a read or a write lock
(this is in the "sendto/sendmsg" cases where addr (sin) is given).
Instrumenting the kernel showed that after taking the lock, we
had bound to a local port from a parallel thread on the same socket.
If we find that case, unlock, and retry again. Taking the write
lock would not be a problem in first place (apart from killing some
parallelism). However the retry is needed as later on based on
similar condition checks we do acquire the pcbinfo lock and if the
conditions have changed, we might find ourselves with a lock
inconsistency, hence at the end of the function when trying to
unlock, hitting the KASSERT.
Reported by: syzbot+bdf4caa36f3ceeac198f@syzkaller.appspotmail.com
Reviewed by: markj
MFC after: 6 weeks
Event: Waterloo Hackathon 2019
netdump waits for acknowledgement from the server for each write. When
dumping page table pages, we perform many small writes, limiting
throughput. Use the netdump client's buffer to buffer small contiguous
writes before calling netdump_send() to flush the MAXDUMPPGS-sized
buffer. This results in a significant reduction in the time taken to
complete a netdump.
Submitted by: Sam Gwydir <sam@samgwydir.com>
Reviewed by: cem
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D20317
received and support for time stamps was negotiated in the SYN/SYNACK
exchange, perform the PAWS check and only expand the syn cache entry if
the check is passed.
Without this check, endpoints may get stuck on the incomplete queue.
Reviewed by: jtl@
MFC after: 3 days
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D20374
- Perform ifp mismatch checks (to determine if a send tag is allocated
for a different ifp than the one the packet is being output on), in
ip_output() and ip6_output(). This avoids sending packets with send
tags to ifnet drivers that don't support send tags.
Since we are now checking for ifp mismatches before invoking
if_output, we can now try to allocate a new tag before invoking
if_output sending the original packet on the new tag if allocation
succeeds.
To avoid code duplication for the fragment and unfragmented cases,
add ip_output_send() and ip6_output_send() as wrappers around
if_output and nd6_output_ifp, respectively. All of the logic for
setting send tags and dealing with send tag-related errors is done
in these wrapper functions.
For pseudo interfaces that wrap other network interfaces (vlan and
lagg), wrapper send tags are now allocated so that ip*_output see
the wrapper ifp as the ifp in the send tag. The if_transmit
routines rewrite the send tags after performing an ifp mismatch
check. If an ifp mismatch is detected, the transmit routines fail
with EAGAIN.
- To provide clearer life cycle management of send tags, especially
in the presence of vlan and lagg wrapper tags, add a reference count
to send tags managed via m_snd_tag_ref() and m_snd_tag_rele().
Provide a helper function (m_snd_tag_init()) for use by drivers
supporting send tags. m_snd_tag_init() takes care of the if_ref
on the ifp meaning that code alloating send tags via if_snd_tag_alloc
no longer has to manage that manually. Similarly, m_snd_tag_rele
drops the refcount on the ifp after invoking if_snd_tag_free when
the last reference to a send tag is dropped.
This also closes use after free races if there are pending packets in
driver tx rings after the socket is closed (e.g. from tcpdrop).
In order for m_free to work reliably, add a new CSUM_SND_TAG flag in
csum_flags to indicate 'snd_tag' is set (rather than 'rcvif').
Drivers now also check this flag instead of checking snd_tag against
NULL. This avoids false positive matches when a forwarded packet
has a non-NULL rcvif that was treated as a send tag.
- cxgbe was relying on snd_tag_free being called when the inp was
detached so that it could kick the firmware to flush any pending
work on the flow. This is because the driver doesn't require ACK
messages from the firmware for every request, but instead does a
kind of manual interrupt coalescing by only setting a flag to
request a completion on a subset of requests. If all of the
in-flight requests don't have the flag when the tag is detached from
the inp, the flow might never return the credits. The current
snd_tag_free command issues a flush command to force the credits to
return. However, the credit return is what also frees the mbufs,
and since those mbufs now hold references on the tag, this meant
that snd_tag_free would never be called.
To fix, explicitly drop the mbuf's reference on the snd tag when the
mbuf is queued in the firmware work queue. This means that once the
inp's reference on the tag goes away and all in-flight mbufs have
been queued to the firmware, tag's refcount will drop to zero and
snd_tag_free will kick in and send the flush request. Note that we
need to avoid doing this in the middle of ethofld_tx(), so the
driver grabs a temporary reference on the tag around that loop to
defer the free to the end of the function in case it sends the last
mbuf to the queue after the inp has dropped its reference on the
tag.
- mlx5 preallocates send tags and was using the ifp pointer even when
the send tag wasn't in use. Explicitly use the ifp from other data
structures instead.
- Sprinkle some assertions in various places to assert that received
packets don't have a send tag, and that other places that overwrite
rcvif (e.g. 802.11 transmit) don't clobber a send tag pointer.
Reviewed by: gallatin, hselasky, rgrimes, ae
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D20117
that we end up in a consistent locking state at the end of
udp_output() in order to be able to see what the values are based
on which we once took a decision (note: some values may have changed).
This helped to debug a syzkaller report.
MFC after: 2 months
Event: Waterloo Hackathon 2019
These are obviously missing from the .c files, but don't show up in any
tinderbox configuration (due to latent header pollution of some kind). It
seems some configurations don't have this pollution, and the includes are
obviously missing, so go ahead and add them.
Reported by: Peter Jeremy <peter AT rulingia.com>
X-MFC-With: r347984
This allows replacing "sys/eventfilter.h" includes with "sys/_eventfilter.h"
in other header files (e.g., sys/{bus,conf,cpu}.h) and reduces header
pollution substantially.
EVENTHANDLER_DECLARE and EVENTHANDLER_LIST_DECLAREs were moved out of .c
files into appropriate headers (e.g., sys/proc.h, powernv/opal.h).
As a side effect of reduced header pollution, many .c files and headers no
longer contain needed definitions. The remainder of the patch addresses
adding appropriate includes to fix those files.
LOCK_DEBUG and LOCK_FILE_LINE_ARG are moved to sys/_lock.h, as required by
sys/mutex.h since r326106 (but silently protected by header pollution prior
to this change).
No functional change (intended). Of course, any out of tree modules that
relied on header pollution for sys/eventhandler.h, sys/lock.h, or
sys/mutex.h inclusion need to be fixed. __FreeBSD_version has been bumped.
r333175 converted the global multicast lock to a sleepable sx lock,
so the lock order with respect to the (non-sleepable) inp lock changed.
To handle this, r333175 and r333505 added code to drop the inp lock,
but this opened races that could leave multicast group description
structures in an inconsistent state. This change fixes the problem by
simply acquiring the global lock sooner. Along the way, this fixes
some LORs and bogus error handling introduced in r333175, and commits
some related cleanup.
Reported by: syzbot+ba7c4943547e0604faca@syzkaller.appspotmail.com
Reported by: syzbot+1b803796ab94d11a46f9@syzkaller.appspotmail.com
Reviewed by: ae
MFC after: 3 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D20070
Boot-time netdump configuration is much more useful if one can configure the
client and gateway addresses. Fix trivial typo.
(Long-standing bug, I believe it dates to the original netdump commit.)
Spotted by: one of vangyzen@ or markj@
Sponsored by: Dell EMC Isilon
Prior to this revision, struct diocskerneldump_arg (and struct netdump_conf
with embedded diocskerneldump_arg before r347192), were copied in their
entirety to the global 'nd_conf' variable. Also prior to this revision,
de-configuring netdump would *not* remove the the key material from global
nd_conf.
As part of Encrypted Kernel Crash Dumps (EKCD), which was developed
contemporaneously with netdump but happened to land first, the
diocskerneldump_arg structure will contain sensitive key material
(kda_key[]) when encrypted dumps are configured.
Netdump doesn't have any use for the key data -- encryption is handled in
the core dumper code -- so in this revision, we no longer store it.
Unfortunately, I think this leak dates to the initial import of netdump in
r333283; so it's present in FreeBSD 12.0.
Fortunately, the impact *seems* relatively minor. Any new *netdump*
configuration would overwrite the key material; for active encrypted netdump
configurations, the key data stored was just a duplicate of the key material
already in the core dumper code; and no user interface (other than
/dev/kmem) actually exposed the leaked material to userspace.
Reviewed by: markj, rpokala (earlier commit message)
MFC after: 2 weeks
Security: yes (minor)
Sponsored by: Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D20233
Bind the TCP pacer threads to NUMA domains and build per-domain
pacer-thread lookup tables. These tables allow us to use the
inpcb's NUMA domain information to match an inpcb with a pacer
thread on the same domain.
The motivation for this is to keep the TCP connection local to a
NUMA domain as much as possible.
Thanks to jhb for pre-reviewing an earlier version of the patch.
Reviewed by: rrs
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D20134
This adds initial support for RFC 2883.
Submitted by: Richard Scheffenegger
Reviewed by: rrs@
Differential Revision: https://reviews.freebsd.org/D19334
This is descrined in RFC 6582, which updates RFC 3782.
Submitted by: Richard Scheffenegger
Reviewed by: lstewart@
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D17614
lookup KPI in ip_output() like it is already used in ip_forward().
However, when there is no PCB provided we can use fast KPI, gaining
performance advantage.
Typical case when ip_output() is called without a PCB pointer is a
sendto(2) on a not connected UDP socket. In practice DNS servers do
this.
Reviewed by: melifaro
Differential Revision: https://reviews.freebsd.org/D19804
Allow users to specify multiple dump configurations in a prioritized list.
This enables fallback to secondary device(s) if primary dump fails. E.g.,
one might configure a preference for netdump, but fallback to disk dump as a
second choice if netdump is unavailable.
This change does not list-ify netdump configuration, which is tracked
separately from ordinary disk dumps internally; only one netdump
configuration can be made at a time, for now. It also does not implement
IPv6 netdump.
savecore(8) is already capable of scanning and iterating multiple devices
from /etc/fstab or passed on the command line.
This change doesn't update the rc or loader variables 'dumpdev' in any way;
it can still be set to configure a single dump device, and rc.d/savecore
still uses it as a single device. Only dumpon(8) is updated to be able to
configure the more complicated configurations for now.
As part of revving the ABI, unify netdump and disk dump configuration ioctl
/ structure, and leave room for ipv6 netdump as a future possibility.
Backwards-compatibility ioctls are added to smooth ABI transition,
especially for developers who may not keep kernel and userspace perfectly
synced.
Reviewed by: markj, scottl (earlier version)
Relnotes: maybe
Sponsored by: Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D19996
Turning on multicast debug made multicast failure worse
because the strings and #define values no longer matched
up. Fix them, and make sure they stay matched-up.
Submitted by: torek
MFC after: 1 week
Sponsored by: iXsystems, Inc.
Drivers can now pass up numa domain information via the
mbuf numa domain field. This information is then used
by TCP syncache_socket() to associate that information
with the inpcb. The domain information is then fed back
into transmitted mbufs in ip{6}_output(). This mechanism
is nearly identical to what is done to track RSS hash values
in the inp_flowid.
Follow on changes will use this information for lacp egress
port selection, binding TCP pacers to the appropriate NUMA
domain, etc.
Reviewed by: markj, kib, slavash, bz, scottl, jtl, tuexen
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D20028
This GRE-in-UDP encapsulation allows the UDP source port field to be
used as an entropy field for load-balancing of GRE traffic in transit
networks. Also most of multiqueue network cards are able distribute
incoming UDP datagrams to different NIC queues, while very little are
able do this for GRE packets.
When an administrator enables UDP encapsulation with command
`ifconfig gre0 udpencap`, the driver creates kernel socket, that binds
to tunnel source address and after udp_set_kernel_tunneling() starts
receiving of all UDP packets destined to 4754 port. Each kernel socket
maintains list of tunnels with different destination addresses. Thus
when several tunnels use the same source address, they all handled by
single socket. The IP[V6]_BINDANY socket option is used to be able bind
socket to source address even if it is not yet available in the system.
This may happen on system boot, when gre(4) interface is created before
source address become available. The encapsulation and sending of packets
is done directly from gre(4) into ip[6]_output() without using sockets.
Reviewed by: eugen
MFC after: 1 month
Relnotes: yes
Differential Revision: https://reviews.freebsd.org/D19921
In r297225 the initial INP_RLOCK() was replaced by an early
acquisition of an r- or w-lock depending on input variables
possibly extending the write locked area for reasons not entirely
clear but possibly to avoid a later case of unlock and relock
leading to a possible race condition and possibly in order to
allow the route cache to work for connected sockets.
Unfortunately the conditions were not 1:1 replicated (probably
because of the route cache needs). While this would not be a
problem the legacy IP code compared to IPv6 has an extra case
when dealing with IP_SENDSRCADDR. In a particular case we were
holding an exclusive inp lock and acquired the shared udbinfo
lock (now epoch).
When then running into an error case, the locking assertions
on release fired as the udpinfo and inp lock levels did not match.
Break up the special case and in that particular case acquire
and udpinfo lock depending on the exclusitivity of the inp lock.
MFC After: 9 days
Reported-by: syzbot+1f5c6800e4f99bdb1a48@syzkaller.appspotmail.com
Reviewed by: tuexen
Differential Revision: https://reviews.freebsd.org/D19594
fragmented packets.
When sending IPv4 and IPv6 fragmented packets and a fragment is lost,
the mbuf making up the fragment will remain in the temporary hashed
fragment list for a while. If the network interface departs before the
so-called slow timeout clears the packet, the fragment causes a panic
when the timeout kicks in due to accessing a freed network interface
structure.
Make sure that when a network device is departing, all hashed IPv4 and
IPv6 fragments belonging to it, get freed.
Backtrace:
panic()
icmp6_reflect()
hlim = ND_IFINFO(m->m_pkthdr.rcvif)->chlim;
^^^^ rcvif->if_afdata[AF_INET6] is NULL.
icmp6_error()
frag6_freef()
frag6_slowtimo()
pfslowtimo()
softclock_call_cc()
softclock()
ithread_loop()
Differential Revision: https://reviews.freebsd.org/D19622
Reviewed by: bz (network), adrian
MFC after: 1 week
Sponsored by: Mellanox Technologies
The logic was present for the 11 version of the DIOCSKERNELDUMP ioctl, but
had not been updated for the 12 ABI.
Reviewed by: markj
Differential Revision: https://reviews.freebsd.org/D19980
The inp lock is not needed for testing the V6 flag as that flag is set
once when the inp is created and never changes. For non-TCP socket
options the lock is immediately dropped after checking that flag.
This just pushes the lock down to only be acquired for TCP socket
options.
This isn't a hot-path, more a cosmetic cleanup I noticed while reading
the code.
Reviewed by: bz
MFC after: 1 month
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D19740
ensure that the ip_hl field is valid. Furthermore, ensure that the complete
IPv4 header is contained in the first mbuf. Finally, move the length checks
before relying on them when accessing fields of the IPv4 header.
Reported by: jtl@
Reviewed by: jtl@
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D19181