[upstream commit 0baf26b0fcd74bbfcef53c5d5e8bad2b99c8d0d2]
This patch makes "struct tcp_congestion_ops" to be the first user
of BPF STRUCT_OPS. It allows implementing a tcp_congestion_ops
in bpf.
The BPF implemented tcp_congestion_ops can be used like
regular kernel tcp-cc through sysctl and setsockopt. e.g.
[root@arch-fb-vm1 bpf]# sysctl -a | egrep congestion
net.ipv4.tcp_allowed_congestion_control = reno cubic bpf_cubic
net.ipv4.tcp_available_congestion_control = reno bic cubic bpf_cubic
net.ipv4.tcp_congestion_control = bpf_cubic
There has been attempt to move the TCP CC to the user space
(e.g. CCP in TCP). The common arguments are faster turn around,
get away from long-tail kernel versions in production...etc,
which are legit points.
BPF has been the continuous effort to join both kernel and
userspace upsides together (e.g. XDP to gain the performance
advantage without bypassing the kernel). The recent BPF
advancements (in particular BTF-aware verifier, BPF trampoline,
BPF CO-RE...) made implementing kernel struct ops (e.g. tcp cc)
possible in BPF. It allows a faster turnaround for testing algorithm
in the production while leveraging the existing (and continue growing)
BPF feature/framework instead of building one specifically for
userspace TCP CC.
This patch allows write access to a few fields in tcp-sock
(in bpf_tcp_ca_btf_struct_access()).
The optional "get_info" is unsupported now. It can be added
later. One possible way is to output the info with a btf-id
to describe the content.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003508.3856115-1-kafai@fb.com
[upstream commit 91cc1a99740e2ed1d903b5906afb470cc5a07379]
Annotate BPF program context types with program-side type and kernel-side type.
This type information is used by the verifier. btf_get_prog_ctx_type() is
used in the later patches to verify that BTF type of ctx in BPF program matches to
kernel expected ctx type. For example, the XDP program type is:
BPF_PROG_TYPE(BPF_PROG_TYPE_XDP, xdp, struct xdp_md, struct xdp_buff)
That means that XDP program should be written as:
int xdp_prog(struct xdp_md *ctx) { ... }
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20191114185720.1641606-16-ast@kernel.org
[upstream commit 9cc31b3a092d9bf2a18f09ad77e727ddb42a5b1e]
btf_resolve_helper_id() caching logic is a bit racy, since under root the
verifier can verify several programs in parallel. Fix it with READ/WRITE_ONCE.
Fix the type as well, since error is also recorded.
Fixes: a7658e1a4164 ("bpf: Check types of arguments passed into helpers")
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20191114185720.1641606-15-ast@kernel.org
[upstream commit a7658e1a4164ce2b9eb4a11aadbba38586e93bd6]
Introduce new helper that reuses existing skb perf_event output
implementation, but can be called from raw_tracepoint programs
that receive 'struct sk_buff *' as tracepoint argument or
can walk other kernel data structures to skb pointer.
In order to do that teach verifier to resolve true C types
of bpf helpers into in-kernel BTF ids.
The type of kernel pointer passed by raw tracepoint into bpf
program will be tracked by the verifier all the way until
it's passed into helper function.
For example:
kfree_skb() kernel function calls trace_kfree_skb(skb, loc);
bpf programs receives that skb pointer and may eventually
pass it into bpf_skb_output() bpf helper which in-kernel is
implemented via bpf_skb_event_output() kernel function.
Its first argument in the kernel is 'struct sk_buff *'.
The verifier makes sure that types match all the way.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20191016032505.2089704-11-ast@kernel.org
Currently, connect(), sendmsg(), recvmsg() and bind-related hooks
are all lacking perf event rb output in order to push notifications
or monitoring events up to user space. Back in commit a5a3a828cd00
("bpf: add perf event notificaton support for sock_ops"), I've worked
with Sowmini to enable them for sock_ops where the context part is
not used (as opposed to skbs for example where the packet data can
be appended). Make the bpf_sockopt_event_output() helper generic and
enable it for mentioned hooks.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/69c39daf87e076b31e52473c902e9bfd37559124.1585323121.git.daniel@iogearbox.net
Sometimes, gcc will optimize the function by spliting it to two or
more functions. In this case, kfree_skb() is splited to
kfree_skb and kfree_skb.part.0. However, the
function/tracepoint trace_kfree_skb() in it needs the return address
of kfree_skb().
This split makes the call chains becomes:
kfree_skb() -> kfree_skb.part.0 -> trace_kfree_skb()
which makes the return address that passed to trace_kfree_skb() be
kfree_skb().
Therefore, introduce '__fix_address', which is the combination of
'__noclone' and 'noinline', and apply it to kfree_skb() to
prevent to from being splited or made inline.
(Is it better to simply apply '__noclone oninline' to kfree_skb?
I'm thinking maybe other functions have the same problems)
Meanwhile, wrap 'skb_unref()' with 'unlikely()', as the compiler thinks
it is likely return true and splits kfree_skb().
Signed-off-by: Menglong Dong <imagedong@tencent.com>
upstream commit: ebe48d368e97d007bfeb76fcb065d6cfc4c96645
The maximum message size that can be send is bigger than
the maximum site that skb_page_frag_refill can allocate.
So it is possible to write beyond the allocated buffer.
Fix this by doing a fallback to COW in that case.
v2:
Avoid get get_order() costs as suggested by Linus Torvalds.
Fixes: cac2661c53f3 ("esp4: Avoid skb_cow_data whenever possible")
Fixes: 03e2a30f6a27 ("esp6: Avoid skb_cow_data whenever possible")
Reported-by: valis <sec@valis.email>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Vaibhav Rustagi <vaibhavrustagi@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We'd like to be able to identify netns from sk_msg hooks
to accelerate local process communication form different netns.
Signed-off-by: Xu Liu <liuxu623@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210820071712.52852-2-liuxu623@gmail.com
(cherry picked from commit fab60e29fcc6d60396da20d63d45fd0d305ba4e4)
We'd like to be able to identify netns from sockops hooks to
accelerate local process communication form different netns.
Signed-off-by: Xu Liu <liuxu623@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20210818105820.91894-2-liuxu623@gmail.com
(cherry picked from commit 6cf1770d63dd2d0d0d4048e7b3ee360336c072d9)
It's getting more common to run nested container environments for
testing cloud software. One of such examples is Kind [1] which runs a
Kubernetes cluster in Docker containers on a single host. Each container
acts as a Kubernetes node, and thus can run any Pod (aka container)
inside the former. This approach simplifies testing a lot, as it
eliminates complicated VM setups.
Unfortunately, such a setup breaks some functionality when cgroupv2 BPF
programs are used for load-balancing. The load-balancer BPF program
needs to detect whether a request originates from the host netns or a
container netns in order to allow some access, e.g. to a service via a
loopback IP address. Typically, the programs detect this by comparing
netns cookies with the one of the init ns via a call to
bpf_get_netns_cookie(NULL). However, in nested environments the latter
cannot be used given the Kubernetes node's netns is outside the init ns.
To fix this, we need to pass the Kubernetes node netns cookie to the
program in a different way: by extending getsockopt() with a
SO_NETNS_COOKIE option, the orchestrator which runs in the Kubernetes
node netns can retrieve the cookie and pass it to the program instead.
Thus, this is following up on Eric's commit 3d368ab87cf6 ("net:
initialize net->net_cookie at netns setup") to allow retrieval via
SO_NETNS_COOKIE. This is also in line in how we retrieve socket cookie
via SO_COOKIE.
[1] https://kind.sigs.k8s.io/
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Martynas Pumputis <m@lambda.lt>
Cc: Eric Dumazet <edumazet@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It is simpler to make net->net_cookie a plain u64
written once in setup_net() instead of looping
and using atomic64 helpers.
Lorenz Bauer wants to add SO_NETNS_COOKIE socket option
and this patch would makes his patch series simpler.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Lorenz Bauer <lmb@cloudflare.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With its use in BPF, the cookie generator can be called very frequently
in particular when used out of cgroup v2 hooks (e.g. connect / sendmsg)
and attached to the root cgroup, for example, when used in v1/v2 mixed
environments. In particular, when there's a high churn on sockets in the
system there can be many parallel requests to the bpf_get_socket_cookie()
and bpf_get_netns_cookie() helpers which then cause contention on the
atomic counter.
As similarly done in f991bd2e1421 ("fs: introduce a per-cpu last_ino
allocator"), add a small helper library that both can use for the 64 bit
counters. Given this can be called from different contexts, we also need
to deal with potential nested calls even though in practice they are
considered extremely rare. One idea as suggested by Eric Dumazet was
to use a reverse counter for this situation since we don't expect 64 bit
overflows anyways; that way, we can avoid bigger gaps in the 64 bit
counter space compared to just batch-wise increase. Even on machines
with small number of cores (e.g. 4) the cookie generation shrinks from
min/max/med/avg (ns) of 22/50/40/38.9 down to 10/35/14/17.3 when run
in parallel from multiple CPUs.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Link: https://lore.kernel.org/bpf/8a80b8d27d3c49f9a14e1d5213c19d8be87d1dc8.1601477936.git.daniel@iogearbox.net
Tooling that populates the SOCK{MAP,HASH} with sockets from user-space
needs a way to inspect its contents. Returning the struct sock * that the
map holds to user-space is neither safe nor useful. An approach established
by REUSEPORT_SOCKARRAY is to return a socket cookie (a unique identifier)
instead.
Since socket cookies are u64 values, SOCK{MAP,HASH} need to support such a
value size for lookup to be possible. This requires special handling on
update, though. Attempts to do a lookup on a map holding u32 values will be
met with ENOSPC error.
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200218171023.844439-7-jakub@cloudflare.com
We currently make heavy use of the socket cookie in BPF's connect(),
sendmsg() and recvmsg() hooks for load-balancing decisions. However,
it is currently not enabled/implemented in BPF {post-}bind hooks
where it can later be used in combination for correlation in the tc
egress path, for example.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/e9d71f310715332f12d238cc650c1edc5be55119.1585323121.git.daniel@iogearbox.net
Turning on GSO by default will make some services (such as TGW) not work properly,
so turn off GSO by default.Buf for some benchmarks (such as redis), turning on GSO
will improve performance.Therefore, a sysctl is provided to user.
Signed-off-by: mungerjiang<mungerjiang@tencent.com>
upstream:0a6b2a1dc2a2105f178255fe495eb914b09cb37a
Oleksandr Natalenko reported performance issues with BBR without FQ
packet scheduler that were root caused to lack of SG and GSO/TSO on
his configuration.
In this mode, TCP internal pacing has to setup a high resolution timer
for each MSS sent.
We could implement in TCP a strategy similar to the one adopted
in commit fefa569a9d4b ("net_sched: sch_fq: account for schedule/timers drifts")
or decide to finally switch TCP stack to a GSO only mode.
This has many benefits :
1) Most TCP developments are done with TSO in mind.
2) Less high-resolution timers needs to be armed for TCP-pacing
3) GSO can benefit of xmit_more hint
4) Receiver GRO is more effective (as if TSO was used for real on sender)
-> Lower ACK traffic
5) Write queues have less overhead (one skb holds about 64KB of payload)
6) SACK coalescing just works.
7) rtx rb-tree contains less packets, SACK is cheaper.
This patch implements the minimum patch, but we can remove some legacy
code as follow ups.
Tested:
On 40Gbit link, one netperf -t TCP_STREAM
BBR+fq:
sg on: 26 Gbits/sec
sg off: 15.7 Gbits/sec (was 2.3 Gbit before patch)
BBR+pfifo_fast:
sg on: 24.2 Gbits/sec
sg off: 14.9 Gbits/sec (was 0.66 Gbit before patch !!! )
BBR+fq_codel:
sg on: 24.4 Gbits/sec
sg off: 15 Gbits/sec (was 0.66 Gbit before patch !!! )
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Signed-off-by: David S. Miller <davem@davemloft.net>
The TCP_PINGPONG_THRESH parameter affects nginx performance.
When the default value is set to 1, the requested QPS is 13% higher than when it is set to 3
Use sysctl -w net.core.tcp_pingpong_thresh=%d to change the default value
Signed-off-by: Mark Whwu <markwhwu@tencent.com>
add SO_MARK2 to get the MARK of flow
backport from tkernel2:
commit c6f2e27f7ad0b97ae0c162f70b6e0375f2270f27
Signed-off-by: Hongbo Li <herberthbli@tencent.com>
Signed-off-by: brookxu <brookxu@tencent.com>
[ Upstream commit 7ad18ff6449cbd6beb26b53128ddf56d2685aa93 ]
Commit 38ec4944b593 ("gro: ensure frag0 meets IP header alignment")
did the right thing, but missed the fact that napi_gro_frags() logics
calls for skb_gro_reset_offset() *before* pulling Ethernet header
to the skb linear space.
That said, the introduced check for frag0 address being aligned to 4
always fails for it as Ethernet header is obviously 14 bytes long,
and in case with NET_IP_ALIGN its start is not aligned to 4.
Fix this by adding @nhoff argument to skb_gro_reset_offset() which
tells if an IP header is placed right at the start of frag0 or not.
This restores Fast GRO for napi_gro_frags() that became very slow
after the mentioned commit, and preserves the introduced check to
avoid silent unaligned accesses.
From v1 [0]:
- inline tiny skb_gro_reset_offset() to let the code be optimized
more efficively (esp. for the !NET_IP_ALIGN case) (Eric);
- pull in Reviewed-by from Eric.
[0] https://lore.kernel.org/netdev/20210418114200.5839-1-alobakin@pm.me
Fixes: 38ec4944b593 ("gro: ensure frag0 meets IP header alignment")
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Alexander Lobakin <alobakin@pm.me>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 38ec4944b593fd90c5ef42aaaa53e66ae5769d04 upstream.
After commit 0f6925b3e8da ("virtio_net: Do not pull payload in skb->head")
Guenter Roeck reported one failure in his tests using sh architecture.
After much debugging, we have been able to spot silent unaligned accesses
in inet_gro_receive()
The issue at hand is that upper networking stacks assume their header
is word-aligned. Low level drivers are supposed to reserve NET_IP_ALIGN
bytes before the Ethernet header to make that happen.
This patch hardens skb_gro_reset_offset() to not allow frag0 fast-path
if the fragment is not properly aligned.
Some arches like x86, arm64 and powerpc do not care and define NET_IP_ALIGN
as 0, this extra check will be a NOP for them.
Note that if frag0 is not used, GRO will call pskb_may_pull()
as many times as needed to pull network and transport headers.
Fixes: 0f6925b3e8da ("virtio_net: Do not pull payload in skb->head")
Fixes: 78a478d0efd9 ("gro: Inline skb_gro_header and cache frag0 virtual address")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Guenter Roeck <linux@roeck-us.net>
Cc: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit d47ec7a0a7271dda08932d6208e4ab65ab0c987c ]
After a short network outage, the dst_entry is timed out and put
in DST_OBSOLETE_DEAD. We are in this code because arp reply comes
from this neighbour after network recovers. There is a potential
race condition that dst_entry is still in DST_OBSOLETE_DEAD.
With that, another neighbour lookup causes more harm than good.
In best case all packets in arp_queue are lost. This is
counterproductive to the original goal of finding a better path
for those packets.
I observed a worst case with 4.x kernel where a dst_entry in
DST_OBSOLETE_DEAD state is associated with loopback net_device.
It leads to an ethernet header with all zero addresses.
A packet with all zero source MAC address is quite deadly with
mac80211, ath9k and 802.11 block ack. It fails
ieee80211_find_sta_by_ifaddr in ath9k (xmit.c). Ath9k flushes tx
queue (ath_tx_complete_aggr). BAW (block ack window) is not
updated. BAW logic is damaged and ath9k transmission is disabled.
Signed-off-by: Tong Zhu <zhutong@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 9adc89af724f12a03b47099cd943ed54e877cd59 upstream.
Currently the mentioned helper can end-up freeing the socket wmem
without waking-up any processes waiting for more write memory.
If the partially orphaned skb is attached to an UDP (or raw) socket,
the lack of wake-up can hang the user-space.
Even for TCP sockets not calling the sk destructor could have bad
effects on TSQ.
Address the issue using skb_orphan to release the sk wmem before
setting the new sock_efree destructor. Additionally bundle the
whole ownership update in a new helper, so that later other
potential users could avoid duplicate code.
v1 -> v2:
- use skb_orphan() instead of sort of open coding it (Eric)
- provide an helper for the ownership change (Eric)
Fixes: f6ba8d33cfbb ("netem: fix skb_orphan_partial()")
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6306c1189e77a513bf02720450bb43bd4ba5d8ae upstream.
Multiple BPF-helpers that can manipulate/increase the size of the SKB uses
__bpf_skb_max_len() as the max-length. This function limit size against
the current net_device MTU (skb->dev->mtu).
When a BPF-prog grow the packet size, then it should not be limited to the
MTU. The MTU is a transmit limitation, and software receiving this packet
should be allowed to increase the size. Further more, current MTU check in
__bpf_skb_max_len uses the MTU from ingress/current net_device, which in
case of redirects uses the wrong net_device.
This patch keeps a sanity max limit of SKB_MAX_ALLOC (16KiB). The real limit
is elsewhere in the system. Jesper's testing[1] showed it was not possible
to exceed 8KiB when expanding the SKB size via BPF-helper. The limiting
factor is the define KMALLOC_MAX_CACHE_SIZE which is 8192 for
SLUB-allocator (CONFIG_SLUB) in-case PAGE_SIZE is 4096. This define is
in-effect due to this being called from softirq context see code
__gfp_pfmemalloc_flags() and __do_kmalloc_node(). Jakub's testing showed
that frames above 16KiB can cause NICs to reset (but not crash). Keep this
sanity limit at this level as memory layer can differ based on kernel
config.
[1] https://github.com/xdp-project/bpf-examples/tree/master/MTU-tests
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/161287788936.790810.2937823995775097177.stgit@firesoul
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit d2126838050ccd1dadf310ffb78b2204f3b032b9 ]
the following command:
# tc filter add dev $h2 ingress protocol ip pref 1 handle 101 flower \
$tcflags dst_ip 192.0.2.2 ip_ttl 63 action drop
doesn't drop all IPv4 packets that match the configured TTL / destination
address. In particular, if "fragment offset" or "more fragments" have non
zero value in the IPv4 header, setting of FLOW_DISSECTOR_KEY_IP is simply
ignored. Fix this dissecting IPv4 TTL and TOS before fragment info; while
at it, add a selftest for tc flower's match on 'ip_ttl' that verifies the
correct behavior.
Fixes: 518d8a2e9bad ("net/flow_dissector: add support for dissection of misc ip header fields")
Reported-by: Shuang Li <shuali@redhat.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 3a5ca857079ea022e0b1b17fc154f7ad7dbc150f upstream.
When a non-initial netns is destroyed, the usual policy is to delete
all virtual network interfaces contained, but move physical interfaces
back to the initial netns. This keeps the physical interface visible
on the system.
CAN devices are somewhat special, as they define rtnl_link_ops even
if they are physical devices. If a CAN interface is moved into a
non-initial netns, destroying that netns lets the interface vanish
instead of moving it back to the initial netns. default_device_exit()
skips CAN interfaces due to having rtnl_link_ops set. Reproducer:
ip netns add foo
ip link set can0 netns foo
ip netns delete foo
WARNING: CPU: 1 PID: 84 at net/core/dev.c:11030 ops_exit_list+0x38/0x60
CPU: 1 PID: 84 Comm: kworker/u4:2 Not tainted 5.10.19 #1
Workqueue: netns cleanup_net
[<c010e700>] (unwind_backtrace) from [<c010a1d8>] (show_stack+0x10/0x14)
[<c010a1d8>] (show_stack) from [<c086dc10>] (dump_stack+0x94/0xa8)
[<c086dc10>] (dump_stack) from [<c086b938>] (__warn+0xb8/0x114)
[<c086b938>] (__warn) from [<c086ba10>] (warn_slowpath_fmt+0x7c/0xac)
[<c086ba10>] (warn_slowpath_fmt) from [<c0629f20>] (ops_exit_list+0x38/0x60)
[<c0629f20>] (ops_exit_list) from [<c062a5c4>] (cleanup_net+0x230/0x380)
[<c062a5c4>] (cleanup_net) from [<c0142c20>] (process_one_work+0x1d8/0x438)
[<c0142c20>] (process_one_work) from [<c0142ee4>] (worker_thread+0x64/0x5a8)
[<c0142ee4>] (worker_thread) from [<c0148a98>] (kthread+0x148/0x14c)
[<c0148a98>] (kthread) from [<c0100148>] (ret_from_fork+0x14/0x2c)
To properly restore physical CAN devices to the initial netns on owning
netns exit, introduce a flag on rtnl_link_ops that can be set by drivers.
For CAN devices setting this flag, default_device_exit() considers them
non-virtual, applying the usual namespace move.
The issue was introduced in the commit mentioned below, as at that time
CAN devices did not have a dellink() operation.
Fixes: e008b5fc8dc7 ("net: Simplfy default_device_exit and improve batching.")
Link: https://lore.kernel.org/r/20210302122423.872326-1-martin@strongswan.org
Signed-off-by: Martin Willi <martin@strongswan.org>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 275b1e88cabb34dbcbe99756b67e9939d34a99b6 ]
pktgen create threads for all online cpus and bond these threads to
relevant cpu repecivtily. when this thread firstly be woken up, it
will compare cpu currently running with the cpu specified at the time
of creation and if the two cpus are not equal, BUG_ON() will take effect
causing panic on the system.
Notice that these threads could be migrated to other cpus before start
running because of the cpu hotplug after these threads have created. so the
BUG_ON() used here seems unreasonable and we can replace it with WARN_ON()
to just printf a warning other than panic the system.
Signed-off-by: Di Zhu <zhudi21@huawei.com>
Link: https://lore.kernel.org/r/20210125124229.19334-1-zhudi21@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 3b23a32a63219f51a5298bc55a65ecee866e79d0 upstream.
dev_ifsioc_locked() is called with only RCU read lock, so when
there is a parallel writer changing the mac address, it could
get a partially updated mac address, as shown below:
Thread 1 Thread 2
// eth_commit_mac_addr_change()
memcpy(dev->dev_addr, addr->sa_data, ETH_ALEN);
// dev_ifsioc_locked()
memcpy(ifr->ifr_hwaddr.sa_data,
dev->dev_addr,...);
Close this race condition by guarding them with a RW semaphore,
like netdev_get_name(). We can not use seqlock here as it does not
allow blocking. The writers already take RTNL anyway, so this does
not affect the slow path. To avoid bothering existing
dev_set_mac_address() callers in drivers, introduce a new wrapper
just for user-facing callers on ioctl and rtnetlink paths.
Note, bonding also changes slave mac addresses but that requires
a separate patch due to the complexity of bonding code.
Fixes: 3710becf8a58 ("net: RCU locking for simple ioctl()")
Reported-by: "Gong, Sishuai" <sishuai@purdue.edu>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 097b9146c0e26aabaa6ff3e5ea536a53f5254a79 upstream.
Avoid the assumption that ksize(kmalloc(S)) == ksize(kmalloc(S)): when
cloning an skb, save and restore truesize after pskb_expand_head(). This
can occur if the allocator decides to service an allocation of the same
size differently (e.g. use a different size class, or pass the
allocation on to KFENCE).
Because truesize is used for bookkeeping (such as sk_wmem_queued), a
modified truesize of a cloned skb may result in corrupt bookkeeping and
relevant warnings (such as in sk_stream_kill_queues()).
Link: https://lkml.kernel.org/r/X9JR/J6dMMOy1obu@elver.google.com
Reported-by: syzbot+7b99aafdcc2eedea6178@syzkaller.appspotmail.com
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20210201160420.2826895-1-elver@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 2c0a10af688c02adcf127aad29e923e0056c6b69 ]
BPF end-user on Cilium slack-channel (Carlo Carraro) wants to use
bpf_fib_lookup for doing MTU-check, but *prior* to extending packet size,
by adjusting fib_params 'tot_len' with the packet length plus the expected
encap size. (Just like the bpf_check_mtu helper supports). He discovered
that for SKB ctx the param->tot_len was not used, instead skb->len was used
(via MTU check in is_skb_forwardable() that checks against netdev MTU).
Fix this by using fib_params 'tot_len' for MTU check. If not provided (e.g.
zero) then keep existing TC behaviour intact. Notice that 'tot_len' for MTU
check is done like XDP code-path, which checks against FIB-dst MTU.
V16:
- Revert V13 optimization, 2nd lookup is against egress/resulting netdev
V13:
- Only do ifindex lookup one time, calling dev_get_by_index_rcu().
V10:
- Use same method as XDP for 'tot_len' MTU check
Fixes: 4c79579b44b1 ("bpf: Change bpf_fib_lookup to return lookup status")
Reported-by: Carlo Carraro <colrack@gmail.com>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/161287789444.790810.15247494756551413508.stgit@firesoul
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 8dc1c444df193701910f5e80b5d4caaf705a8fb0 upstream.
Commit c80794323e82 ("net: Fix packet reordering caused by GRO and
listified RX cooperation") had the unfortunate effect of adding
latencies in common workloads.
Before the patch, GRO packets were immediately passed to
upper stacks.
After the patch, we can accumulate quite a lot of GRO
packets (depdending on NAPI budget).
My fix is counting in napi->rx_count number of segments
instead of number of logical packets.
Fixes: c80794323e82 ("net: Fix packet reordering caused by GRO and listified RX cooperation")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Bisected-by: John Sperbeck <jsperbeck@google.com>
Tested-by: Jian Yang <jianyang@google.com>
Cc: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Edward Cree <ecree.xilinx@gmail.com>
Reviewed-by: Alexander Lobakin <alobakin@pm.me>
Link: https://lore.kernel.org/r/20210204213146.4192368-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 52cbd23a119c6ebf40a527e53f3402d2ea38eccb upstream.
When iteratively computing a checksum with csum_block_add, track the
offset "pos" to correctly rotate in csum_block_add when offset is odd.
The open coded implementation of skb_copy_and_csum_datagram did this.
With the switch to __skb_datagram_iter calling csum_and_copy_to_iter,
pos was reinitialized to 0 on each call.
Bring back the pos by passing it along with the csum to the callback.
Changes v1->v2
- pass csum value, instead of csump pointer (Alexander Duyck)
Link: https://lore.kernel.org/netdev/20210128152353.GB27281@optiplex/
Fixes: 950fcaecd5cc ("datagram: consolidate datagram copy to iter helpers")
Reported-by: Oliver Graute <oliver.graute@gmail.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20210203192952.1849843-1-willemdebruijn.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit eb4e8fac00d1e01ada5e57c05d24739156086677 upstream.
Following race condition was detected:
<CPU A, t0> - neigh_flush_dev() is under execution and calls
neigh_mark_dead(n) marking the neighbour entry 'n' as dead.
<CPU B, t1> - Executing: __netif_receive_skb() ->
__netif_receive_skb_core() -> arp_rcv() -> arp_process().arp_process()
calls __neigh_lookup() which takes a reference on neighbour entry 'n'.
<CPU A, t2> - Moves further along neigh_flush_dev() and calls
neigh_cleanup_and_release(n), but since reference count increased in t2,
'n' couldn't be destroyed.
<CPU B, t3> - Moves further along, arp_process() and calls
neigh_update()-> __neigh_update() -> neigh_update_gc_list(), which adds
the neighbour entry back in gc_list(neigh_mark_dead(), removed it
earlier in t0 from gc_list)
<CPU B, t4> - arp_process() finally calls neigh_release(n), destroying
the neighbour entry.
This leads to 'n' still being part of gc_list, but the actual
neighbour structure has been freed.
The situation can be prevented from happening if we disallow a dead
entry to have any possibility of updating gc_list. This is what the
patch intends to achieve.
Fixes: 9c29a2f55ec0 ("neighbor: Fix locking order for gc_list changes")
Signed-off-by: Chinmay Agarwal <chinagar@codeaurora.org>
Reviewed-by: Cong Wang <xiyou.wangcong@gmail.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20210127165453.GA20514@chinagar-linux.qualcomm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 66c556025d687dbdd0f748c5e1df89c977b6c02a upstream.
Commit 3226b158e67c ("net: avoid 32 x truesize under-estimation for
tiny skbs") ensured that skbs with data size lower than 1025 bytes
will be kmalloc'ed to avoid excessive page cache fragmentation and
memory consumption.
However, the fix adressed only __napi_alloc_skb() (primarily for
virtio_net and napi_get_frags()), but the issue can still be achieved
through __netdev_alloc_skb(), which is still used by several drivers.
Drivers often allocate a tiny skb for headers and place the rest of
the frame to frags (so-called copybreak).
Mirror the condition to __netdev_alloc_skb() to handle this case too.
Since v1 [0]:
- fix "Fixes:" tag;
- refine commit message (mention copybreak usecase).
[0] https://lore.kernel.org/netdev/20210114235423.232737-1-alobakin@pm.me
Fixes: a1c7fff7e18f ("net: netdev_alloc_skb() use build_skb()")
Signed-off-by: Alexander Lobakin <alobakin@pm.me>
Link: https://lore.kernel.org/r/20210115150354.85967-1-alobakin@pm.me
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ no upstream commit ]
Fix a potential kernel address leakage for the prerequisite where there is
a BPF program attached to the cgroup/setsockopt hook. The latter can only
be attached under root, however, if the attached program returns 1 to then
run the related kernel handler, an unprivileged program could probe for
kernel addresses that way. The reason this is possible is that we're under
set_fs(KERNEL_DS) when running the kernel setsockopt handler. Aside from
old cBPF there is also SCTP's struct sctp_getaddrs_old which contains
pointers in the uapi struct that further need copy_from_user() inside the
handler. In the normal case this would just return -EFAULT, but under a
temporary KERNEL_DS setting the memory would be copied and we'd end up at
a different error code, that is, -EINVAL, for both cases given subsequent
validations fail, which then allows the app to distinguish and make use of
this fact for probing the address space. In case of later kernel versions
this issue won't work anymore thanks to Christoph Hellwig's work that got
rid of the various temporary set_fs() address space overrides altogether.
One potential option for 5.4 as the only affected stable kernel with the
least complexity would be to remap those affected -EFAULT copy_from_user()
error codes with -EINVAL such that they cannot be probed anymore. Risk of
breakage should be rather low for this particular error case.
Fixes: 0d01da6afc54 ("bpf: implement getsockopt and setsockopt hooks")
Reported-by: Ryota Shiga (Flatt Security)
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Stanislav Fomichev <sdf@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 3226b158e67cfaa677fd180152bfb28989cb2fac ]
Both virtio net and napi_get_frags() allocate skbs
with a very small skb->head
While using page fragments instead of a kmalloc backed skb->head might give
a small performance improvement in some cases, there is a huge risk of
under estimating memory usage.
For both GOOD_COPY_LEN and GRO_MAX_HEAD, we can fit at least 32 allocations
per page (order-3 page in x86), or even 64 on PowerPC
We have been tracking OOM issues on GKE hosts hitting tcp_mem limits
but consuming far more memory for TCP buffers than instructed in tcp_mem[2]
Even if we force napi_alloc_skb() to only use order-0 pages, the issue
would still be there on arches with PAGE_SIZE >= 32768
This patch makes sure that small skb head are kmalloc backed, so that
other objects in the slab page can be reused instead of being held as long
as skbs are sitting in socket queues.
Note that we might in the future use the sk_buff napi cache,
instead of going through a more expensive __alloc_skb()
Another idea would be to use separate page sizes depending
on the allocated length (to never have more than 4 frags per page)
I would like to thank Greg Thelen for his precious help on this matter,
analysing crash dumps is always a time consuming task.
Fixes: fd11a83dd363 ("net: Pull out core bits of __netdev_alloc_skb and add __napi_alloc_skb")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Greg Thelen <gthelen@google.com>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Link: https://lore.kernel.org/r/20210113161819.1155526-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit fd2ddef043592e7de80af53f47fa46fd3573086e ]
reuse->socks[] is modified concurrently by reuseport_add_sock. To
prevent reading values that have not been fully initialized, only read
the array up until the last known safe index instead of incorrectly
re-reading the last index of the array.
Fixes: acdcecc61285f ("udp: correct reuseport selection with connected sockets")
Signed-off-by: Baptiste Lepers <baptiste.lepers@gmail.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20210107051110.12247-1-baptiste.lepers@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 54970a2fbb673f090b7f02d7f57b10b2e0707155 upstream.
syzbot reproduces BUG_ON in skb_checksum_help():
tun creates (bogus) skb with huge partial-checksummed area and
small ip packet inside. Then ip_rcv trims the skb based on size
of internal ip packet, after that csum offset points beyond of
trimmed skb. Then checksum_tg() called via netfilter hook
triggers BUG_ON:
offset = skb_checksum_start_offset(skb);
BUG_ON(offset >= skb_headlen(skb));
To work around the problem this patch forces pskb_trim_rcsum_slow()
to return -EINVAL in described scenario. It allows its callers to
drop such kind of packets.
Link: https://syzkaller.appspot.com/bug?id=b419a5ca95062664fe1a60b764621eb4526e2cd0
Reported-by: syzbot+7010af67ced6105e5ab6@syzkaller.appspotmail.com
Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/1b2494af-2c56-8ee2-7bc0-923fcad1cdf8@virtuozzo.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 4ae2bb81649dc03dfc95875f02126b14b773f7ab ]
Accesses to dev->xps_rxqs_map (when using dev->num_tc) should be
protected by the rtnl lock, like we do for netif_set_xps_queue. I didn't
see an actual bug being triggered, but let's be safe here and take the
rtnl lock while accessing the map in sysfs.
Fixes: 8af2c06ff4b1 ("net-sysfs: Add interface for Rx queue(s) map per Tx queue")
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 2d57b4f142e0b03e854612b8e28978935414bced ]
Two race conditions can be triggered when storing xps rxqs, resulting in
various oops and invalid memory accesses:
1. Calling netdev_set_num_tc while netif_set_xps_queue:
- netif_set_xps_queue uses dev->tc_num as one of the parameters to
compute the size of new_dev_maps when allocating it. dev->tc_num is
also used to access the map, and the compiler may generate code to
retrieve this field multiple times in the function.
- netdev_set_num_tc sets dev->tc_num.
If new_dev_maps is allocated using dev->tc_num and then dev->tc_num
is set to a higher value through netdev_set_num_tc, later accesses to
new_dev_maps in netif_set_xps_queue could lead to accessing memory
outside of new_dev_maps; triggering an oops.
2. Calling netif_set_xps_queue while netdev_set_num_tc is running:
2.1. netdev_set_num_tc starts by resetting the xps queues,
dev->tc_num isn't updated yet.
2.2. netif_set_xps_queue is called, setting up the map with the
*old* dev->num_tc.
2.3. netdev_set_num_tc updates dev->tc_num.
2.4. Later accesses to the map lead to out of bound accesses and
oops.
A similar issue can be found with netdev_reset_tc.
One way of triggering this is to set an iface up (for which the driver
uses netdev_set_num_tc in the open path, such as bnx2x) and writing to
xps_rxqs in a concurrent thread. With the right timing an oops is
triggered.
Both issues have the same fix: netif_set_xps_queue, netdev_set_num_tc
and netdev_reset_tc should be mutually exclusive. We do that by taking
the rtnl lock in xps_rxqs_store.
Fixes: 8af2c06ff4b1 ("net-sysfs: Add interface for Rx queue(s) map per Tx queue")
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit fb25038586d0064123e393cadf1fadd70a9df97a ]
Accesses to dev->xps_cpus_map (when using dev->num_tc) should be
protected by the rtnl lock, like we do for netif_set_xps_queue. I didn't
see an actual bug being triggered, but let's be safe here and take the
rtnl lock while accessing the map in sysfs.
Fixes: 184c449f91fe ("net: Add support for XPS with QoS via traffic classes")
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 1ad58225dba3f2f598d2c6daed4323f24547168f ]
Two race conditions can be triggered when storing xps cpus, resulting in
various oops and invalid memory accesses:
1. Calling netdev_set_num_tc while netif_set_xps_queue:
- netif_set_xps_queue uses dev->tc_num as one of the parameters to
compute the size of new_dev_maps when allocating it. dev->tc_num is
also used to access the map, and the compiler may generate code to
retrieve this field multiple times in the function.
- netdev_set_num_tc sets dev->tc_num.
If new_dev_maps is allocated using dev->tc_num and then dev->tc_num
is set to a higher value through netdev_set_num_tc, later accesses to
new_dev_maps in netif_set_xps_queue could lead to accessing memory
outside of new_dev_maps; triggering an oops.
2. Calling netif_set_xps_queue while netdev_set_num_tc is running:
2.1. netdev_set_num_tc starts by resetting the xps queues,
dev->tc_num isn't updated yet.
2.2. netif_set_xps_queue is called, setting up the map with the
*old* dev->num_tc.
2.3. netdev_set_num_tc updates dev->tc_num.
2.4. Later accesses to the map lead to out of bound accesses and
oops.
A similar issue can be found with netdev_reset_tc.
One way of triggering this is to set an iface up (for which the driver
uses netdev_set_num_tc in the open path, such as bnx2x) and writing to
xps_cpus in a concurrent thread. With the right timing an oops is
triggered.
Both issues have the same fix: netif_set_xps_queue, netdev_set_num_tc
and netdev_reset_tc should be mutually exclusive. We do that by taking
the rtnl lock in xps_cpus_store.
Fixes: 184c449f91fe ("net: Add support for XPS with QoS via traffic classes")
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the virtualization scenario, you can use pvipi to optimize the sending of
IPI interrupts and reduce the number of vmexit.
rps_using_pvipi test
Test guest: 32 core, 64G memory, virtio-net 8 queues, rss interrupt is bound
to 24-31 core. Observe the number of vmexit of 31 cores.
1. 1000000pps
100 iperf instances, 10000pps per instance.
disable rps_using_pvipi 68077 vmexit/s
enable rps_using_pvipi 21138 vmexit/s
2. 2000000pps
1000 iperf instances, 2000pps per instance.
disable rps_using_pvipi 40879 vmexit/s
enable rps_using_pvipi 9195 vmexit/s
3. 2000000pps
2000 iperf instances, 1000pps per instance.
disable rps_using_pvipi 24746 vmexit/s
enable rps_using_pvipi 7292 vmexit/s
Enabling rps_using_pvipi can greatly reduce the number of vmexit.
Signed-off-by: yongduan <yongduan@tencent.com>
Reviewed-by: Jiang Biao <benbjiang@tencent.com>
In the virtualization scenario, you can use pvipi to optimize the sending of
IPI interrupts and reduce the number of vmexit.
rps_using_pvipi test
Test guest: 32 core, 64G memory, virtio-net 8 queues, rss interrupt is bound
to 24-31 core. Observe the number of vmexit of 31 cores.
1. 1000000pps
100 iperf instances, 10000pps per instance.
disable rps_using_pvipi 68077 vmexit/s
enable rps_using_pvipi 21138 vmexit/s
2. 2000000pps
1000 iperf instances, 2000pps per instance.
disable rps_using_pvipi 40879 vmexit/s
enable rps_using_pvipi 9195 vmexit/s
3. 2000000pps
2000 iperf instances, 1000pps per instance.
disable rps_using_pvipi 24746 vmexit/s
enable rps_using_pvipi 7292 vmexit/s
Enabling rps_using_pvipi can greatly reduce the number of vmexit.
Signed-off-by: yongduan <yongduan@tencent.com>
Reviewed-by: Jiang Biao <benbjiang@tencent.com>
Martin KaFai Lau