This patch makes cpumap functional, by adding SKB allocation and
invoking the network stack on the dequeuing CPU.
For constructing the SKB on the remote CPU, the xdp_buff in converted
into a struct xdp_pkt, and it mapped into the top headroom of the
packet, to avoid allocating separate mem. For now, struct xdp_pkt is
just a cpumap internal data structure, with info carried between
enqueue to dequeue.
If a driver doesn't have enough headroom it is simply dropped, with
return code -EOVERFLOW. This will be picked up the xdp tracepoint
infrastructure, to allow users to catch this.
V2: take into account xdp->data_meta
V4:
- Drop busypoll tricks, keeping it more simple.
- Skip RPS and Generic-XDP-recursive-reinjection, suggested by Alexei
V5: correct RCU read protection around __netif_receive_skb_core.
V6: Setting TASK_RUNNING vs TASK_INTERRUPTIBLE based on talk with Rik van Riel
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch connects cpumap to the xdp_do_redirect_map infrastructure.
Still no SKB allocation are done yet. The XDP frames are transferred
to the other CPU, but they are simply refcnt decremented on the remote
CPU. This served as a good benchmark for measuring the overhead of
remote refcnt decrement. If driver page recycle cache is not
efficient then this, exposes a bottleneck in the page allocator.
A shout-out to MST's ptr_ring, which is the secret behind is being so
efficient to transfer memory pointers between CPUs, without constantly
bouncing cache-lines between CPUs.
V3: Handle !CONFIG_BPF_SYSCALL pointed out by kbuild test robot.
V4: Make Generic-XDP aware of cpumap type, but don't allow redirect yet,
as implementation require a separate upstream discussion.
V5:
- Fix a maybe-uninitialized pointed out by kbuild test robot.
- Restrict bpf-prog side access to cpumap, open when use-cases appear
- Implement cpu_map_enqueue() as a more simple void pointer enqueue
V6:
- Allow cpumap type for usage in helper bpf_redirect_map,
general bpf-prog side restriction moved to earlier patch.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 32302902ff ("mqprio: Reserve last 32 classid values for HW
traffic classes and misc IDs") sch_mqprio started using netdev_txq_to_tc
to find the correct tc instead of dev->tc_to_txq[]
However, when mqprio is compiled as a module, it cannot resolve the
symbol, leading to this error:
ERROR: "netdev_txq_to_tc" [net/sched/sch_mqprio.ko] undefined!
This adds an EXPORT_SYMBOL() since the other user in the kernel
(netif_set_xps_queue) is also EXPORT_SYMBOL() (and not _GPL) or in a
sysfs-callback.
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
Cc: Jesus Sanchez-Palencia <jesus.sanchez-palencia@intel.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Henrik Austad <haustad@cisco.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rtnl af_ops currently rely on rtnl mutex: unregister (called from module
exit functions) takes the rtnl mutex and all users that do af_ops lookup
also take the rtnl mutex. IOW, parallel rmmod will block until doit()
callback is done.
As none of the af_ops implementation sleep we can use rcu instead.
doit functions that need the af_ops can now use rcu instead of the
rtnl mutex provided the mutex isn't needed for other reasons.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
next patch will rcu-ify rtnl af_ops, i.e. allow af_ops
lookup and function calls with rcu read lock held instead
of rtnl mutex.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
We need a real-time notification for tcp retransmission
for monitoring.
Of course we could use ftrace to dynamically instrument this
kernel function too, however we can't retrieve the connection
information at the same time, for example perf-tools [1] reads
/proc/net/tcp for socket details, which is slow when we have
a lots of connections.
Therefore, this patch adds a tracepoint for __tcp_retransmit_skb()
and exposes src/dst IP addresses and ports of the connection.
This also makes it easier to integrate into perf.
Note, I expose both IPv4 and IPv6 addresses at the same time:
for a IPv4 socket, v4 mapped address is used as IPv6 addresses,
for a IPv6 socket, LOOPBACK4_IPV6 is already filled by kernel.
Also, add sk and skb pointers as they are useful for BPF.
1. https://github.com/brendangregg/perf-tools/blob/master/net/tcpretrans
Cc: Eric Dumazet <edumazet@google.com>
Cc: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Cc: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Brendan Gregg <bgregg@netflix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If CONFIG_DST_CACHE is not selected cpu variable
will be unused and we will see a compilation warning.
Move it under the ifdef.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Fixes: d66f2b91f9 ("bpf: don't rely on the verifier lock for metadata_dst allocation")
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We can now piggyback error strings to userspace via extended acks
rather than using printk.
Before:
bridge fdb add 01:02:03:04:05:06 dev br0 vlan 4095
RTNETLINK answers: Invalid argument
After:
bridge fdb add 01:02:03:04:05:06 dev br0 vlan 4095
Error: invalid vlan id.
v3: drop 'RTM_' prefixes, suggested by David Ahern, they
are not useful, the add/del in bridge command line is enough.
Also reword error in response to malformed/bad vlan id attribute
size.
Cc: David Ahern <dsahern@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix BUG() calls to use BUG_ON(conditional) macros.
This was found using make coccicheck M=net/core on linux next
tag next-2017092
Signed-off-by: Tim Hansen <devtimhansen@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
bpf_skb_set_tunnel_*() functions require allocation of per-cpu
metadata_dst. The allocation happens upon verification of the
first program using those helpers. In preparation for removing
the verifier lock, use cmpxchg() to make sure we only allocate
the metadata_dsts once.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add extack arg to netdev_upper_dev_link and netdev_master_upper_dev_link
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pass extack to do_set_master and down to ndo_add_slave
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add netlink_ext_ack to netdev_notifier_info to allow notifier
handlers to return errors to userspace.
Clean up the initialization in dev.c such that extack is easily
added in subsequent patches where relevant. Specifically, remove
the init call in call_netdevice_notifiers_info and have callers
initalize on stack when info is declared.
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
x-netns interfaces are bound to two netns: the link netns and the upper
netns. Usually, this kind of interfaces is created in the link netns and
then moved to the upper netns. At the end, the interface is visible only
in the upper netns. The link nsid is advertised via netlink in the upper
netns, thus the user always knows where is the link part.
There is no such mechanism in the link netns. When the interface is moved
to another netns, the user cannot "follow" it.
This patch adds a new netlink attribute which helps to follow an interface
which moves to another netns. When the interface is unregistered, the new
nsid is advertised. If the interface is a x-netns interface (ie
rtnl_link_ops->get_link_net is defined), the nsid is allocated if needed.
CC: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Compiler does not really know that skb_shinfo(to|from) are constants
in skb_try_coalesce(), lets cache their values to shrink code.
We might even take care of skb_zcopy() calls later.
$ size net/core/skbuff.o.before net/core/skbuff.o
text data bss dec hex filename
40727 1298 0 42025 a429 net/core/skbuff.o.before
40631 1298 0 41929 a3c9 net/core/skbuff.o
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
no users in the tree.
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/core/dev.c:1306: warning: No description found for parameter 'name'
net/core/dev.c:1306: warning: Excess function parameter 'alias' description in 'dev_get_alias'
Fixes: 6c5570016b ("net: core: decouple ifalias get/set from rtnl lock")
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Device alias can be set by either rtnetlink (rtnl is held) or sysfs.
rtnetlink hold the rtnl mutex, sysfs acquires it for this purpose.
Add an extra mutex for it and use rcu to protect concurrent accesses.
This allows the sysfs path to not take rtnl and would later allow
to not hold it when dumping ifalias.
Based on suggestion from Eric Dumazet.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
When RTM_GETSTATS was added the fields of its header struct were not all
initialized when returning the result thus leaking 4 bytes of information
to user-space per rtnl_fill_statsinfo call, so initialize them now. Thanks
to Alexander Potapenko for the detailed report and bisection.
Reported-by: Alexander Potapenko <glider@google.com>
Fixes: 10c9ead9f3 ("rtnetlink: add new RTM_GETSTATS message to dump link stats")
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Starting from linux-4.4, 3WHS no longer takes the listener lock.
Since this time, we might hit a use-after-free in sk_filter_charge(),
if the filter we got in the memcpy() of the listener content
just happened to be replaced by a thread changing listener BPF filter.
To fix this, we need to make sure the filter refcount is not already
zero before incrementing it again.
Fixes: e994b2f0fb ("tcp: do not lock listener to process SYN packets")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move dissection of tunnel info from the flower classifier to the flow
dissector where all other dissection occurs. This should not have any
behavioural affect on other users of the flow dissector.
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sk->sk_prot and sk->sk_prot_creator can differ when the app uses
IPV6_ADDRFORM (transforming an IPv6-socket to an IPv4-one).
Which is why sk_prot_creator is there to make sure that sk_prot_free()
does the kmem_cache_free() on the right kmem_cache slab.
Now, if such a socket gets transformed back to a listening socket (using
connect() with AF_UNSPEC) we will allocate an IPv4 tcp_sock through
sk_clone_lock() when a new connection comes in. But sk_prot_creator will
still point to the IPv6 kmem_cache (as everything got copied in
sk_clone_lock()). When freeing, we will thus put this
memory back into the IPv6 kmem_cache although it was allocated in the
IPv4 cache. I have seen memory corruption happening because of this.
With slub-debugging and MEMCG_KMEM enabled this gives the warning
"cache_from_obj: Wrong slab cache. TCPv6 but object is from TCP"
A C-program to trigger this:
void main(void)
{
int fd = socket(AF_INET6, SOCK_STREAM, IPPROTO_TCP);
int new_fd, newest_fd, client_fd;
struct sockaddr_in6 bind_addr;
struct sockaddr_in bind_addr4, client_addr1, client_addr2;
struct sockaddr unsp;
int val;
memset(&bind_addr, 0, sizeof(bind_addr));
bind_addr.sin6_family = AF_INET6;
bind_addr.sin6_port = ntohs(42424);
memset(&client_addr1, 0, sizeof(client_addr1));
client_addr1.sin_family = AF_INET;
client_addr1.sin_port = ntohs(42424);
client_addr1.sin_addr.s_addr = inet_addr("127.0.0.1");
memset(&client_addr2, 0, sizeof(client_addr2));
client_addr2.sin_family = AF_INET;
client_addr2.sin_port = ntohs(42421);
client_addr2.sin_addr.s_addr = inet_addr("127.0.0.1");
memset(&unsp, 0, sizeof(unsp));
unsp.sa_family = AF_UNSPEC;
bind(fd, (struct sockaddr *)&bind_addr, sizeof(bind_addr));
listen(fd, 5);
client_fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
connect(client_fd, (struct sockaddr *)&client_addr1, sizeof(client_addr1));
new_fd = accept(fd, NULL, NULL);
close(fd);
val = AF_INET;
setsockopt(new_fd, SOL_IPV6, IPV6_ADDRFORM, &val, sizeof(val));
connect(new_fd, &unsp, sizeof(unsp));
memset(&bind_addr4, 0, sizeof(bind_addr4));
bind_addr4.sin_family = AF_INET;
bind_addr4.sin_port = ntohs(42421);
bind(new_fd, (struct sockaddr *)&bind_addr4, sizeof(bind_addr4));
listen(new_fd, 5);
client_fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
connect(client_fd, (struct sockaddr *)&client_addr2, sizeof(client_addr2));
newest_fd = accept(new_fd, NULL, NULL);
close(new_fd);
close(client_fd);
close(new_fd);
}
As far as I can see, this bug has been there since the beginning of the
git-days.
Signed-off-by: Christoph Paasch <cpaasch@apple.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
it can be switched to rcu.
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
similar to earlier patches, split out more parts of this function to
better see what is happening and where we assume rtnl is locked.
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
rtnl_fill_ifinfo currently requires caller to hold the rtnl mutex.
Unfortunately the function is quite large which makes it harder to see
which spots require the lock, which spots assume it and which ones could
do without.
Add helpers to factor out the ifindex dumping, one can use rcu to avoid
rtnl dependency.
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
This work enables generic transfer of metadata from XDP into skb. The
basic idea is that we can make use of the fact that the resulting skb
must be linear and already comes with a larger headroom for supporting
bpf_xdp_adjust_head(), which mangles xdp->data. Here, we base our work
on a similar principle and introduce a small helper bpf_xdp_adjust_meta()
for adjusting a new pointer called xdp->data_meta. Thus, the packet has
a flexible and programmable room for meta data, followed by the actual
packet data. struct xdp_buff is therefore laid out that we first point
to data_hard_start, then data_meta directly prepended to data followed
by data_end marking the end of packet. bpf_xdp_adjust_head() takes into
account whether we have meta data already prepended and if so, memmove()s
this along with the given offset provided there's enough room.
xdp->data_meta is optional and programs are not required to use it. The
rationale is that when we process the packet in XDP (e.g. as DoS filter),
we can push further meta data along with it for the XDP_PASS case, and
give the guarantee that a clsact ingress BPF program on the same device
can pick this up for further post-processing. Since we work with skb
there, we can also set skb->mark, skb->priority or other skb meta data
out of BPF, thus having this scratch space generic and programmable
allows for more flexibility than defining a direct 1:1 transfer of
potentially new XDP members into skb (it's also more efficient as we
don't need to initialize/handle each of such new members). The facility
also works together with GRO aggregation. The scratch space at the head
of the packet can be multiple of 4 byte up to 32 byte large. Drivers not
yet supporting xdp->data_meta can simply be set up with xdp->data_meta
as xdp->data + 1 as bpf_xdp_adjust_meta() will detect this and bail out,
such that the subsequent match against xdp->data for later access is
guaranteed to fail.
The verifier treats xdp->data_meta/xdp->data the same way as we treat
xdp->data/xdp->data_end pointer comparisons. The requirement for doing
the compare against xdp->data is that it hasn't been modified from it's
original address we got from ctx access. It may have a range marking
already from prior successful xdp->data/xdp->data_end pointer comparisons
though.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Just do the rename into bpf_compute_data_pointers() as we'll add
one more pointer here to recompute.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
IS_ERR() already implies unlikely(), so it can be omitted.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Key length can't be negative.
Leave comparisons against nla_len() signed just in case truncated attribute
can sneak in there.
Space savings:
add/remove: 0/0 grow/shrink: 0/7 up/down: 0/-7 (-7)
function old new delta
pneigh_delete 273 272 -1
mlx5e_rep_netevent_event 1415 1414 -1
mlx5e_create_encap_header_ipv6 1194 1193 -1
mlx5e_create_encap_header_ipv4 1071 1070 -1
cxgb4_l2t_get 1104 1103 -1
__pneigh_lookup 69 68 -1
__neigh_create 2452 2451 -1
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As measured in my prior patch ("sch_netem: faster rb tree removal"),
rbtree_postorder_for_each_entry_safe() is nice looking but much slower
than using rb_next() directly, except when tree is small enough
to fit in CPU caches (then the cost is the same)
Also note that there is not even an increase of text size :
$ size net/core/skbuff.o.before net/core/skbuff.o
text data bss dec hex filename
40711 1298 0 42009 a419 net/core/skbuff.o.before
40711 1298 0 42009 a419 net/core/skbuff.o
From: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Private part of allocation is never big enough to warrant size_t.
Space savings:
add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-10 (-10)
function old new delta
alloc_netdev_mqs 1120 1110 -10
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Zerocopy skbs frags are copied when the skb is looped to a local sock.
Commit 1080e512d4 ("net: orphan frags on receive") introduced calls
to skb_orphan_frags to deliver_skb and __netif_receive_skb for this.
With msg_zerocopy, these skbs can also exist in the tx path and thus
loop from dev_queue_xmit_nit. This already calls deliver_skb in its
loop. But it does not orphan before a separate pt_prev->func().
Add the missing skb_orphan_frags_rx.
Changes
v1->v2: handle skb_orphan_frags_rx failure
Fixes: 1f8b977ab3 ("sock: enable MSG_ZEROCOPY")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 3f1ac7a700 ("net: ethtool: add new ETHTOOL_xLINKSETTINGS API")
deprecated the ethtool_cmd::transceiver field, which was fine in
premise, except that the PHY library was actually using it to report the
type of transceiver: internal or external.
Use the first word of the reserved field to put this __u8 transceiver
field back in. It is made read-only, and we don't expect the
ETHTOOL_xLINKSETTINGS API to be doing anything with this anyway, so this
is mostly for the legacy path where we do:
ethtool_get_settings()
-> dev->ethtool_ops->get_link_ksettings()
-> convert_link_ksettings_to_legacy_settings()
to have no information loss compared to the legacy get_settings API.
Fixes: 3f1ac7a700 ("net: ethtool: add new ETHTOOL_xLINKSETTINGS API")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since XDP's view of the packet includes the MAC header, moving the start-
of-packet with bpf_xdp_adjust_head needs to also update the offset of the
MAC header (which is relative to skb->head, not to the skb->data that was
changed).
Without this, tcpdump sees packets starting from the old MAC header rather
than the new one, at least in my tests on the loopback device.
Fixes: b5cdae3291 ("net: Generic XDP")
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 109980b894 ("bpf: don't select potentially stale
ri->map from buggy xdp progs") passed the pointer to the prog
itself to be loaded into r4 prior on bpf_redirect_map() helper
call, so that we can store the owner into ri->map_owner out of
the helper.
Issue with that is that the actual address of the prog is still
subject to change when subsequent rewrites occur that require
slow path in bpf_prog_realloc() to alloc more memory, e.g. from
patching inlining helper functions or constant blinding. Thus,
we really need to take prog->aux as the address we're holding,
which also works with prog clones as they share the same aux
object.
Instead of then fetching aux->prog during runtime, which could
potentially incur cache misses due to false sharing, we are
going to just use aux for comparison on the map owner. This
will also keep the patchlet of the same size, and later check
in xdp_map_invalid() only accesses read-only aux pointer from
the prog, it's also in the same cacheline already from prior
access when calling bpf_func.
Fixes: 109980b894 ("bpf: don't select potentially stale ri->map from buggy xdp progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Denys reported wrong rate estimations with HTB classes.
It appears the bug was added in linux-4.10, since my tests
where using intervals of one second only.
HTB using 4 sec default rate estimators, reported rates
were 4x higher.
We need to properly scale the bytes/packets samples before
integrating them in EWMA.
Tested:
echo 1 >/sys/module/sch_htb/parameters/htb_rate_est
Setup HTB with one class with a rate/cail of 5Gbit
Generate traffic on this class
tc -s -d cl sh dev eth0 classid 7002:11
class htb 7002:11 parent 7002:1 prio 5 quantum 200000 rate 5Gbit ceil
5Gbit linklayer ethernet burst 80000b/1 mpu 0b cburst 80000b/1 mpu 0b
level 0 rate_handle 1
Sent 1488215421648 bytes 982969243 pkt (dropped 0, overlimits 0
requeues 0)
rate 5Gbit 412814pps backlog 136260b 2p requeues 0
TCP pkts/rtx 982969327/45 bytes 1488215557414/68130
lended: 22732826 borrowed: 0 giants: 0
tokens: -1684 ctokens: -1684
Fixes: 1c0d32fde5 ("net_sched: gen_estimator: complete rewrite of rate estimators")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Denys Fedoryshchenko <nuclearcat@nuclearcat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Using bpf_redirect_map is allowed for generic XDP programs, but the
appropriate map lookup was never performed in xdp_do_generic_redirect().
Instead the map-index is directly used as the ifindex. For the
xdp_redirect_map sample in SKB-mode '-S', this resulted in trying
sending on ifindex 0 which isn't valid, resulting in getting SKB
packets dropped. Thus, the reported performance numbers are wrong in
commit 24251c2647 ("samples/bpf: add option for native and skb mode
for redirect apps") for the 'xdp_redirect_map -S' case.
Before commit 109980b894 ("bpf: don't select potentially stale
ri->map from buggy xdp progs") it could crash the kernel. Like this
commit also check that the map_owner owner is correct before
dereferencing the map pointer. But make sure that this API misusage
can be caught by a tracepoint. Thus, allowing userspace via
tracepoints to detect misbehaving bpf_progs.
Fixes: 6103aa96ec ("net: implement XDP_REDIRECT for xdp generic")
Fixes: 24251c2647 ("samples/bpf: add option for native and skb mode for redirect apps")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
"The iwlwifi firmware compat fix is in here as well as some other
stuff:
1) Fix request socket leak introduced by BPF deadlock fix, from Eric
Dumazet.
2) Fix VLAN handling with TXQs in mac80211, from Johannes Berg.
3) Missing __qdisc_drop conversions in prio and qfq schedulers, from
Gao Feng.
4) Use after free in netlink nlk groups handling, from Xin Long.
5) Handle MTU update properly in ipv6 gre tunnels, from Xin Long.
6) Fix leak of ipv6 fib tables on netns teardown, from Sabrina Dubroca
with follow-on fix from Eric Dumazet.
7) Need RCU and preemption disabled during generic XDP data patch,
from John Fastabend"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (54 commits)
bpf: make error reporting in bpf_warn_invalid_xdp_action more clear
Revert "mdio_bus: Remove unneeded gpiod NULL check"
bpf: devmap, use cond_resched instead of cpu_relax
bpf: add support for sockmap detach programs
net: rcu lock and preempt disable missing around generic xdp
bpf: don't select potentially stale ri->map from buggy xdp progs
net: tulip: Constify tulip_tbl
net: ethernet: ti: netcp_core: no need in netif_napi_del
davicom: Display proper debug level up to 6
net: phy: sfp: rename dt properties to match the binding
dt-binding: net: sfp binding documentation
dt-bindings: add SFF vendor prefix
dt-bindings: net: don't confuse with generic PHY property
ip6_tunnel: fix setting hop_limit value for ipv6 tunnel
ip_tunnel: fix setting ttl and tos value in collect_md mode
ipv6: fix typo in fib6_net_exit()
tcp: fix a request socket leak
sctp: fix missing wake ups in some situations
netfilter: xt_hashlimit: fix build error caused by 64bit division
netfilter: xt_hashlimit: alloc hashtable with right size
...
Differ between illegal XDP action code and just driver
unsupported one to provide better feedback when we throw
a one-time warning here. Reason is that with 814abfabef
("xdp: add bpf_redirect helper function") not all drivers
support the new XDP return code yet and thus they will
fall into their 'default' case when checking for return
codes after program return, which then triggers a
bpf_warn_invalid_xdp_action() stating that the return
code is illegal, but from XDP perspective it's not.
I decided not to place something like a XDP_ACT_MAX define
into uapi i) given we don't have this either for all other
program types, ii) future action codes could have further
encoding there, which would render such define unsuitable
and we wouldn't be able to rip it out again, and iii) we
rarely add new action codes.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
do_xdp_generic must be called inside rcu critical section with preempt
disabled to ensure BPF programs are valid and per-cpu variables used
for redirect operations are consistent. This patch ensures this is true
and fixes the splat below.
The netif_receive_skb_internal() code path is now broken into two rcu
critical sections. I decided it was better to limit the preempt_enable/disable
block to just the xdp static key portion and the fallout is more
rcu_read_lock/unlock calls. Seems like the best option to me.
[ 607.596901] =============================
[ 607.596906] WARNING: suspicious RCU usage
[ 607.596912] 4.13.0-rc4+ #570 Not tainted
[ 607.596917] -----------------------------
[ 607.596923] net/core/dev.c:3948 suspicious rcu_dereference_check() usage!
[ 607.596927]
[ 607.596927] other info that might help us debug this:
[ 607.596927]
[ 607.596933]
[ 607.596933] rcu_scheduler_active = 2, debug_locks = 1
[ 607.596938] 2 locks held by pool/14624:
[ 607.596943] #0: (rcu_read_lock_bh){......}, at: [<ffffffff95445ffd>] ip_finish_output2+0x14d/0x890
[ 607.596973] #1: (rcu_read_lock_bh){......}, at: [<ffffffff953c8e3a>] __dev_queue_xmit+0x14a/0xfd0
[ 607.597000]
[ 607.597000] stack backtrace:
[ 607.597006] CPU: 5 PID: 14624 Comm: pool Not tainted 4.13.0-rc4+ #570
[ 607.597011] Hardware name: Dell Inc. Precision Tower 5810/0HHV7N, BIOS A17 03/01/2017
[ 607.597016] Call Trace:
[ 607.597027] dump_stack+0x67/0x92
[ 607.597040] lockdep_rcu_suspicious+0xdd/0x110
[ 607.597054] do_xdp_generic+0x313/0xa50
[ 607.597068] ? time_hardirqs_on+0x5b/0x150
[ 607.597076] ? mark_held_locks+0x6b/0xc0
[ 607.597088] ? netdev_pick_tx+0x150/0x150
[ 607.597117] netif_rx_internal+0x205/0x3f0
[ 607.597127] ? do_xdp_generic+0xa50/0xa50
[ 607.597144] ? lock_downgrade+0x2b0/0x2b0
[ 607.597158] ? __lock_is_held+0x93/0x100
[ 607.597187] netif_rx+0x119/0x190
[ 607.597202] loopback_xmit+0xfd/0x1b0
[ 607.597214] dev_hard_start_xmit+0x127/0x4e0
Fixes: d445516966 ("net: xdp: support xdp generic on virtual devices")
Fixes: b5cdae3291 ("net: Generic XDP")
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
We can potentially run into a couple of issues with the XDP
bpf_redirect_map() helper. The ri->map in the per CPU storage
can become stale in several ways, mostly due to misuse, where
we can then trigger a use after free on the map:
i) prog A is calling bpf_redirect_map(), returning XDP_REDIRECT
and running on a driver not supporting XDP_REDIRECT yet. The
ri->map on that CPU becomes stale when the XDP program is unloaded
on the driver, and a prog B loaded on a different driver which
supports XDP_REDIRECT return code. prog B would have to omit
calling to bpf_redirect_map() and just return XDP_REDIRECT, which
would then access the freed map in xdp_do_redirect() since not
cleared for that CPU.
ii) prog A is calling bpf_redirect_map(), returning a code other
than XDP_REDIRECT. prog A is then detached, which triggers release
of the map. prog B is attached which, similarly as in i), would
just return XDP_REDIRECT without having called bpf_redirect_map()
and thus be accessing the freed map in xdp_do_redirect() since
not cleared for that CPU.
iii) prog A is attached to generic XDP, calling the bpf_redirect_map()
helper and returning XDP_REDIRECT. xdp_do_generic_redirect() is
currently not handling ri->map (will be fixed by Jesper), so it's
not being reset. Later loading a e.g. native prog B which would,
say, call bpf_xdp_redirect() and then returns XDP_REDIRECT would
find in xdp_do_redirect() that a map was set and uses that causing
use after free on map access.
Fix thus needs to avoid accessing stale ri->map pointers, naive
way would be to call a BPF function from drivers that just resets
it to NULL for all XDP return codes but XDP_REDIRECT and including
XDP_REDIRECT for drivers not supporting it yet (and let ri->map
being handled in xdp_do_generic_redirect()). There is a less
intrusive way w/o letting drivers call a reset for each BPF run.
The verifier knows we're calling into bpf_xdp_redirect_map()
helper, so it can do a small insn rewrite transparent to the prog
itself in the sense that it fills R4 with a pointer to the own
bpf_prog. We have that pointer at verification time anyway and
R4 is allowed to be used as per calling convention we scratch
R0 to R5 anyway, so they become inaccessible and program cannot
read them prior to a write. Then, the helper would store the prog
pointer in the current CPUs struct redirect_info. Later in
xdp_do_*_redirect() we check whether the redirect_info's prog
pointer is the same as passed xdp_prog pointer, and if that's
the case then all good, since the prog holds a ref on the map
anyway, so it is always valid at that point in time and must
have a reference count of at least 1. If in the unlikely case
they are not equal, it means we got a stale pointer, so we clear
and bail out right there. Also do reset map and the owning prog
in bpf_xdp_redirect(), so that bpf_xdp_redirect_map() and
bpf_xdp_redirect() won't get mixed up, only the last call should
take precedence. A tc bpf_redirect() doesn't use map anywhere
yet, so no need to clear it there since never accessed in that
layer.
Note that in case the prog is released, and thus the map as
well we're still under RCU read critical section at that time
and have preemption disabled as well. Once we commit with the
__dev_map_insert_ctx() from xdp_do_redirect_map() and set the
map to ri->map_to_flush, we still wait for a xdp_do_flush_map()
to finish in devmap dismantle time once flush_needed bit is set,
so that is fine.
Fixes: 97f91a7cf0 ("bpf: add bpf_redirect_map helper routine")
Reported-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
After commit 0ddf3fb2c4 ("udp: preserve skb->dst if required
for IP options processing") we clear the skb head state as soon
as the skb carrying them is first processed.
Since the same skb can be processed several times when MSG_PEEK
is used, we can end up lacking the required head states, and
eventually oopsing.
Fix this clearing the skb head state only when processing the
last skb reference.
Reported-by: Eric Dumazet <edumazet@google.com>
Fixes: 0ddf3fb2c4 ("udp: preserve skb->dst if required for IP options processing")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull cgroup updates from Tejun Heo:
"Several notable changes this cycle:
- Thread mode was merged. This will be used for cgroup2 support for
CPU and possibly other controllers. Unfortunately, CPU controller
cgroup2 support didn't make this pull request but most contentions
have been resolved and the support is likely to be merged before
the next merge window.
- cgroup.stat now shows the number of descendant cgroups.
- cpuset now can enable the easier-to-configure v2 behavior on v1
hierarchy"
* 'for-4.14' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup: (21 commits)
cpuset: Allow v2 behavior in v1 cgroup
cgroup: Add mount flag to enable cpuset to use v2 behavior in v1 cgroup
cgroup: remove unneeded checks
cgroup: misc changes
cgroup: short-circuit cset_cgroup_from_root() on the default hierarchy
cgroup: re-use the parent pointer in cgroup_destroy_locked()
cgroup: add cgroup.stat interface with basic hierarchy stats
cgroup: implement hierarchy limits
cgroup: keep track of number of descent cgroups
cgroup: add comment to cgroup_enable_threaded()
cgroup: remove unnecessary empty check when enabling threaded mode
cgroup: update debug controller to print out thread mode information
cgroup: implement cgroup v2 thread support
cgroup: implement CSS_TASK_ITER_THREADED
cgroup: introduce cgroup->dom_cgrp and threaded css_set handling
cgroup: add @flags to css_task_iter_start() and implement CSS_TASK_ITER_PROCS
cgroup: reorganize cgroup.procs / task write path
cgroup: replace css_set walking populated test with testing cgrp->nr_populated_csets
cgroup: distinguish local and children populated states
cgroup: remove now unused list_head @pending in cgroup_apply_cftypes()
...
In flow dissector there are no limits to the number of nested
encapsulations or headers that might be dissected which makes for a
nice DOS attack. This patch sets a limit of the number of headers
that flow dissector will parse.
Headers includes network layer headers, transport layer headers, shim
headers for encapsulation, IPv6 extension headers, etc. The limit for
maximum number of headers to parse has be set to fifteen to account for
a reasonable number of encapsulations, extension headers, VLAN,
in a packet. Note that this limit does not supercede the STOP_AT_*
flags which may stop processing before the headers limit is reached.
Reported-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Tom Herbert <tom@quantonium.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
__skb_flow_dissect is riddled with gotos that make discerning the flow,
debugging, and extending the capability difficult. This patch
reorganizes things so that we only perform goto's after the two main
switch statements (no gotos within the cases now). It also eliminates
several goto labels so that there are only two labels that can be target
for goto.
Reported-by: Alexander Popov <alex.popov@linux.com>
Signed-off-by: Tom Herbert <tom@quantonium.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a listener registers to the FIB notification chain it receives a
dump of the FIB entries and rules from existing address families by
invoking their dump operations.
While we call into these modules we need to make sure they aren't
removed. Do that by increasing their reference count before invoking
their dump operations and decrease it afterwards.
Fixes: 04b1d4e50e ("net: core: Make the FIB notification chain generic")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
