The xdp tracepoints for mem id disconnect don't carry information about, why
it was not safe_to_remove. The tracepoint page_pool:page_pool_inflight in
this patch can be used for extract this info for further debugging.
This patchset also adds tracepoint for the pages_state_* release/hold
transitions, including a pointer to the page. This can be used for stats
about in-flight pages, or used to debug page leakage via keeping track of
page pointer and combining this with kprobe for __put_page().
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These tracepoints make it easier to troubleshoot XDP mem id disconnect.
The xdp:mem_disconnect tracepoint cannot be replaced via kprobe. It is
placed at the last stable place for the pointer to struct xdp_mem_allocator,
just before it's scheduled for RCU removal. It also extract info on
'safe_to_remove' and 'force'.
Detailed info about in-flight pages is not available at this layer. The next
patch will added tracepoints needed at the page_pool layer for this.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If bugs exists or are introduced later e.g. by drivers misusing the API,
then we want to warn about the issue, such that developer notice. This patch
will generate a bit of noise in form of periodic pr_warn every 30 seconds.
It is not nice to have this stall warning running forever. Thus, this patch
will (after 120 attempts) force disconnect the mem id (from the rhashtable)
and free the page_pool object. This will cause fallback to the put_page() as
before, which only potentially leak DMA-mappings, if objects are really
stuck for this long. In that unlikely case, a WARN_ONCE should show us the
call stack.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is needed before we can allow drivers to use page_pool for
DMA-mappings. Today with page_pool and XDP return API, it is possible to
remove the page_pool object (from rhashtable), while there are still
in-flight packet-pages. This is safely handled via RCU and failed lookups in
__xdp_return() fallback to call put_page(), when page_pool object is gone.
In-case page is still DMA mapped, this will result in page note getting
correctly DMA unmapped.
To solve this, the page_pool is extended with tracking in-flight pages. And
XDP disconnect system queries page_pool and waits, via workqueue, for all
in-flight pages to be returned.
To avoid killing performance when tracking in-flight pages, the implement
use two (unsigned) counters, that in placed on different cache-lines, and
can be used to deduct in-flight packets. This is done by mapping the
unsigned "sequence" counters onto signed Two's complement arithmetic
operations. This is e.g. used by kernel's time_after macros, described in
kernel commit 1ba3aab303 and 5a581b367b, and also explained in RFC1982.
The trick is these two incrementing counters only need to be read and
compared, when checking if it's safe to free the page_pool structure. Which
will only happen when driver have disconnected RX/alloc side. Thus, on a
non-fast-path.
It is chosen that page_pool tracking is also enabled for the non-DMA
use-case, as this can be used for statistics later.
After this patch, using page_pool requires more strict resource "release",
e.g. via page_pool_release_page() that was introduced in this patchset, and
previous patches implement/fix this more strict requirement.
Drivers no-longer call page_pool_destroy(). Drivers already call
xdp_rxq_info_unreg() which call xdp_rxq_info_unreg_mem_model(), which will
attempt to disconnect the mem id, and if attempt fails schedule the
disconnect for later via delayed workqueue.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case driver fails to register the page_pool with XDP return API (via
xdp_rxq_info_reg_mem_model()), then the driver can free the page_pool
resources more directly than calling page_pool_destroy(), which does a
unnecessarily RCU free procedure.
This patch is preparing for removing page_pool_destroy(), from driver
invocation.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When converting an xdp_frame into an SKB, and sending this into the network
stack, then the underlying XDP memory model need to release associated
resources, because the network stack don't have callbacks for XDP memory
models. The only memory model that needs this is page_pool, when a driver
use the DMA-mapping feature.
Introduce page_pool_release_page(), which basically does the same as
page_pool_unmap_page(). Add xdp_release_frame() as the XDP memory model
interface for calling it, if the memory model match MEM_TYPE_PAGE_POOL, to
save the function call overhead for others. Have cpumap call
xdp_release_frame() before xdp_scrub_frame().
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix error handling case, where inserting ID with rhashtable_insert_slow
fails in xdp_rxq_info_reg_mem_model, which leads to never releasing the IDA
ID, as the lookup in xdp_rxq_info_unreg_mem_model fails and thus
ida_simple_remove() is never called.
Fix by releasing ID via ida_simple_remove(), and mark xdp_rxq->mem.id with
zero, which is already checked in xdp_rxq_info_unreg_mem_model().
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
On a previous patch dma addr was stored in 'struct page'.
Use that to unmap DMA addresses used by network drivers
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Implement support for previously added flow dissector meta key.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use previously introduced infra to obtain and store ingress ifindex
instead doing it locally.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add new key meta that contains ingress ifindex value and add a function
to dissect this from skb. The key and function is prepared to cover
other potential skb metadata values dissection.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
in IPoIB case we can't see a VF broadcast address for but
can see for PF
Before:
11: ib1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 2044 qdisc pfifo_fast
state UP mode DEFAULT group default qlen 256
link/infiniband
80:00:00:66:fe:80:00:00:00:00:00:00:24:8a:07:03:00:a4:3e:7c brd
00:ff:ff:ff:ff:12:40:1b:ff:ff:00:00:00:00:00:00:ff:ff:ff:ff
vf 0 MAC 14:80:00:00:66:fe, spoof checking off, link-state disable,
trust off, query_rss off
...
After:
11: ib1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 2044 qdisc pfifo_fast
state UP mode DEFAULT group default qlen 256
link/infiniband
80:00:00:66:fe:80:00:00:00:00:00:00:24:8a:07:03:00:a4:3e:7c brd
00:ff:ff:ff:ff:12:40:1b:ff:ff:00:00:00:00:00:00:ff:ff:ff:ff
vf 0 link/infiniband
80:00:00:66:fe:80:00:00:00:00:00:00:24:8a:07:03:00:a4:3e:7c brd
00:ff:ff:ff:ff:12:40:1b:ff:ff:00:00:00:00:00:00:ff:ff:ff:ff, spoof
checking off, link-state disable, trust off, query_rss off
v1->v2: add the IFLA_VF_BROADCAST constant
v2->v3: put IFLA_VF_BROADCAST at the end
to avoid KABI breakage and set NLA_REJECT
dev_setlink
Signed-off-by: Denis Kirjanov <kda@linux-powerpc.org>
Acked-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Both listeners - mlxsw and netdevsim - of IPv6 FIB notifications are now
ready to handle IPv6 multipath notifications.
Therefore, stop ignoring such notifications in both drivers and stop
sending notification for each added / deleted nexthop.
v2:
* Remove 'multipath_rt' from 'struct fib6_entry_notifier_info'
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If all the nexthops of a multipath route are being deleted, send one
notification for the entire route, instead of one per-nexthop.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Emit a notification when a multipath routes is added or replace.
Note that unlike the replace notifications sent from fib6_add_rt2node(),
it is possible we are sending a 'FIB_EVENT_ENTRY_REPLACE' when a route
was merely added and not replaced.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Extend the IPv6 FIB notifier info with number of sibling routes being
notified.
This will later allow listeners to process one notification for a
multipath routes instead of N, where N is the number of nexthops.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Causes crash when lifetime expires on an adress as garbage is
dereferenced soon after.
This used to look like this:
for (ifap = &ifa->ifa_dev->ifa_list;
*ifap != NULL; ifap = &(*ifap)->ifa_next) {
if (*ifap == ifa) ...
but this was changed to:
struct in_ifaddr *tmp;
ifap = &ifa->ifa_dev->ifa_list;
tmp = rtnl_dereference(*ifap);
while (tmp) {
tmp = rtnl_dereference(tmp->ifa_next); // Bogus
if (rtnl_dereference(*ifap) == ifa) {
...
ifap = &tmp->ifa_next; // Can be NULL
tmp = rtnl_dereference(*ifap); // Dereference
}
}
Remove the bogus assigment/list entry skip.
Fixes: 2638eb8b50 ("net: ipv4: provide __rcu annotation for ifa_list")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
"Lots of bug fixes here:
1) Out of bounds access in __bpf_skc_lookup, from Lorenz Bauer.
2) Fix rate reporting in cfg80211_calculate_bitrate_he(), from John
Crispin.
3) Use after free in psock backlog workqueue, from John Fastabend.
4) Fix source port matching in fdb peer flow rule of mlx5, from Raed
Salem.
5) Use atomic_inc_not_zero() in fl6_sock_lookup(), from Eric Dumazet.
6) Network header needs to be set for packet redirect in nfp, from
John Hurley.
7) Fix udp zerocopy refcnt, from Willem de Bruijn.
8) Don't assume linear buffers in vxlan and geneve error handlers,
from Stefano Brivio.
9) Fix TOS matching in mlxsw, from Jiri Pirko.
10) More SCTP cookie memory leak fixes, from Neil Horman.
11) Fix VLAN filtering in rtl8366, from Linus Walluij.
12) Various TCP SACK payload size and fragmentation memory limit fixes
from Eric Dumazet.
13) Use after free in pneigh_get_next(), also from Eric Dumazet.
14) LAPB control block leak fix from Jeremy Sowden"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (145 commits)
lapb: fixed leak of control-blocks.
tipc: purge deferredq list for each grp member in tipc_group_delete
ax25: fix inconsistent lock state in ax25_destroy_timer
neigh: fix use-after-free read in pneigh_get_next
tcp: fix compile error if !CONFIG_SYSCTL
hv_sock: Suppress bogus "may be used uninitialized" warnings
be2net: Fix number of Rx queues used for flow hashing
net: handle 802.1P vlan 0 packets properly
tcp: enforce tcp_min_snd_mss in tcp_mtu_probing()
tcp: add tcp_min_snd_mss sysctl
tcp: tcp_fragment() should apply sane memory limits
tcp: limit payload size of sacked skbs
Revert "net: phylink: set the autoneg state in phylink_phy_change"
bpf: fix nested bpf tracepoints with per-cpu data
bpf: Fix out of bounds memory access in bpf_sk_storage
vsock/virtio: set SOCK_DONE on peer shutdown
net: dsa: rtl8366: Fix up VLAN filtering
net: phylink: set the autoneg state in phylink_phy_change
net: add high_order_alloc_disable sysctl/static key
tcp: add tcp_tx_skb_cache sysctl
...
Currently user is unable to delete the filter. See following example:
$ tc filter add dev ens16np1 ingress pref 1 handle 1 matchall action drop
$ tc filter show dev ens16np1 ingress
filter protocol all pref 1 matchall chain 0
filter protocol all pref 1 matchall chain 0 handle 0x1
in_hw
action order 1: gact action drop
random type none pass val 0
index 1 ref 1 bind 1
$ tc filter del dev ens16np1 ingress pref 1 handle 1 matchall action drop
RTNETLINK answers: Operation not supported
Implement tcf_proto_ops->delete() op and allow user to delete the filter.
Reported-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix nla_policy definition by specifying an exact length type attribute
to CTINFO action paraneter block structure. Without this change,
netlink parsing will fail validation and the action will not be
instantiated.
8cb081746c ("netlink: make validation more configurable for future")
introduced much stricter checking to attributes being passed via
netlink. Existing actions were updated to use less restrictive
deprecated versions of nla_parse_nested.
As a new module, act_ctinfo should be designed to use the strict
checking model otherwise, well, what was the point of implementing it.
Confession time: Until very recently, development of this module has
been done on 'net-next' tree to 'clean compile' level with run-time
testing on backports to 4.14 & 4.19 kernels under openwrt. This is how
I managed to miss the run-time impacts of the new strict
nla_parse_nested function. I hopefully have learned something from this
(glances toward laptop running a net-next kernel)
There is however a still outstanding implication on iproute2 user space
in that it needs to be told to pass nested netlink messages with the
nested attribute actually set. So even with this kernel fix to do
things correctly you still cannot instantiate a new 'strict'
nla_parse_nested based action such as act_ctinfo with iproute2's tc.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use correct return value on action creation: ACT_P_CREATED.
The use of incorrect return value could result in a situation where the
system thought a ctinfo module was listening but actually wasn't
instantiated correctly leading to an OOPS in tcf_generic_walker().
Confession time: Until very recently, development of this module has
been done on 'net-next' tree to 'clean compile' level with run-time
testing on backports to 4.14 & 4.19 kernels under openwrt. During the
back & forward porting during development & testing, the critical
ACT_P_CREATED return code got missed despite being in the 4.14 & 4.19
backports. I have now gone through the init functions, using act_csum
as reference with a fine toothed comb. Bonus, no more OOPSes. I
managed to also miss this issue till now due to the new strict
nla_parse_nested function failing validation before action creation.
As an inexperienced developer I've learned that
copy/pasting/backporting/forward porting code correctly is hard. If I
ever get to a developer conference I shall don the cone of shame.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Using a bare block cipher in non-crypto code is almost always a bad idea,
not only for security reasons (and we've seen some examples of this in
the kernel in the past), but also for performance reasons.
In the TCP fastopen case, we call into the bare AES block cipher one or
two times (depending on whether the connection is IPv4 or IPv6). On most
systems, this results in a call chain such as
crypto_cipher_encrypt_one(ctx, dst, src)
crypto_cipher_crt(tfm)->cit_encrypt_one(crypto_cipher_tfm(tfm), ...);
aesni_encrypt
kernel_fpu_begin();
aesni_enc(ctx, dst, src); // asm routine
kernel_fpu_end();
It is highly unlikely that the use of special AES instructions has a
benefit in this case, especially since we are doing the above twice
for IPv6 connections, instead of using a transform which can process
the entire input in one go.
We could switch to the cbcmac(aes) shash, which would at least get
rid of the duplicated overhead in *some* cases (i.e., today, only
arm64 has an accelerated implementation of cbcmac(aes), while x86 will
end up using the generic cbcmac template wrapping the AES-NI cipher,
which basically ends up doing exactly the above). However, in the given
context, it makes more sense to use a light-weight MAC algorithm that
is more suitable for the purpose at hand, such as SipHash.
Since the output size of SipHash already matches our chosen value for
TCP_FASTOPEN_COOKIE_SIZE, and given that it accepts arbitrary input
sizes, this greatly simplifies the code as well.
NOTE: Server farms backing a single server IP for load balancing purposes
and sharing a single fastopen key will be adversely affected by
this change unless all systems in the pool receive their kernel
upgrades at the same time.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In patch series, commit 9195948fbf ("tipc: improve TIPC throughput by
Gap ACK blocks"), as for simplicity, the repeated retransmit failures'
detection in the function - "tipc_link_retrans()" was kept there for
broadcast retransmissions only.
This commit now reapplies this feature for link unicast retransmissions
that has been done via the function - "tipc_link_advance_transmq()".
Also, the "tipc_link_retrans()" is renamed to "tipc_link_bc_retrans()"
as it is used only for broadcast.
Acked-by: Jon Maloy <jon.maloy@ericsson.se>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet says:
====================
tcp: make sack processing more robust
Jonathan Looney brought to our attention multiple problems
in TCP stack at the sender side.
SACK processing can be abused by malicious peers to either
cause overflows, or increase of memory usage.
First two patches fix the immediate problems.
Since the malicious peers abuse senders by advertizing a very
small MSS in their SYN or SYNACK packet, the last two
patches add a new sysctl so that admins can chose a higher
limit for MSS clamping.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
lapb_register calls lapb_create_cb, which initializes the control-
block's ref-count to one, and __lapb_insert_cb, which increments it when
adding the new block to the list of blocks.
lapb_unregister calls __lapb_remove_cb, which decrements the ref-count
when removing control-block from the list of blocks, and calls lapb_put
itself to decrement the ref-count before returning.
However, lapb_unregister also calls __lapb_devtostruct to look up the
right control-block for the given net_device, and __lapb_devtostruct
also bumps the ref-count, which means that when lapb_unregister returns
the ref-count is still 1 and the control-block is leaked.
Call lapb_put after __lapb_devtostruct to fix leak.
Reported-by: syzbot+afb980676c836b4a0afa@syzkaller.appspotmail.com
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Syzbot reported a memleak caused by grp members' deferredq list not
purged when the grp is be deleted.
The issue occurs when more(msg_grp_bc_seqno(hdr), m->bc_rcv_nxt) in
tipc_group_filter_msg() and the skb will stay in deferredq.
So fix it by calling __skb_queue_purge for each member's deferredq
in tipc_group_delete() when a tipc sk leaves the grp.
Fixes: b87a5ea31c ("tipc: guarantee group unicast doesn't bypass group broadcast")
Reported-by: syzbot+78fbe679c8ca8d264a8d@syzkaller.appspotmail.com
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The EXPORT_SYMBOL for lapb_register was next to a different function.
Moved it to the right place.
Signed-off-by: Jeremy Sowden <jeremy@azazel.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp_tx_skb_cache_key and tcp_rx_skb_cache_key must be available
even if CONFIG_SYSCTL is not set.
Fixes: 0b7d7f6b22 ("tcp: add tcp_tx_skb_cache sysctl")
Fixes: ede61ca474 ("tcp: add tcp_rx_skb_cache sysctl")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
gcc 8.2.0 may report these bogus warnings under some condition:
warning: ‘vnew’ may be used uninitialized in this function
warning: ‘hvs_new’ may be used uninitialized in this function
Actually, the 2 pointers are only initialized and used if the variable
"conn_from_host" is true. The code is not buggy here.
Signed-off-by: Dexuan Cui <decui@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When stack receives pkt: [802.1P vlan 0][802.1AD vlan 100][IPv4],
vlan_do_receive() returns false if it does not find vlan_dev. Later
__netif_receive_skb_core() fails to find packet type handler for
skb->protocol 801.1AD and drops the packet.
801.1P header with vlan id 0 should be handled as untagged packets.
This patch fixes it by checking if vlan_id is 0 and processes next vlan
header.
Signed-off-by: Govindarajulu Varadarajan <gvaradar@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If mtu probing is enabled tcp_mtu_probing() could very well end up
with a too small MSS.
Use the new sysctl tcp_min_snd_mss to make sure MSS search
is performed in an acceptable range.
CVE-2019-11479 -- tcp mss hardcoded to 48
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Cc: Jonathan Looney <jtl@netflix.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Tyler Hicks <tyhicks@canonical.com>
Cc: Bruce Curtis <brucec@netflix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some TCP peers announce a very small MSS option in their SYN and/or
SYN/ACK messages.
This forces the stack to send packets with a very high network/cpu
overhead.
Linux has enforced a minimal value of 48. Since this value includes
the size of TCP options, and that the options can consume up to 40
bytes, this means that each segment can include only 8 bytes of payload.
In some cases, it can be useful to increase the minimal value
to a saner value.
We still let the default to 48 (TCP_MIN_SND_MSS), for compatibility
reasons.
Note that TCP_MAXSEG socket option enforces a minimal value
of (TCP_MIN_MSS). David Miller increased this minimal value
in commit c39508d6f1 ("tcp: Make TCP_MAXSEG minimum more correct.")
from 64 to 88.
We might in the future merge TCP_MIN_SND_MSS and TCP_MIN_MSS.
CVE-2019-11479 -- tcp mss hardcoded to 48
Signed-off-by: Eric Dumazet <edumazet@google.com>
Suggested-by: Jonathan Looney <jtl@netflix.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Tyler Hicks <tyhicks@canonical.com>
Cc: Bruce Curtis <brucec@netflix.com>
Cc: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jonathan Looney reported that a malicious peer can force a sender
to fragment its retransmit queue into tiny skbs, inflating memory
usage and/or overflow 32bit counters.
TCP allows an application to queue up to sk_sndbuf bytes,
so we need to give some allowance for non malicious splitting
of retransmit queue.
A new SNMP counter is added to monitor how many times TCP
did not allow to split an skb if the allowance was exceeded.
Note that this counter might increase in the case applications
use SO_SNDBUF socket option to lower sk_sndbuf.
CVE-2019-11478 : tcp_fragment, prevent fragmenting a packet when the
socket is already using more than half the allowed space
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Jonathan Looney <jtl@netflix.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Reviewed-by: Tyler Hicks <tyhicks@canonical.com>
Cc: Bruce Curtis <brucec@netflix.com>
Cc: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jonathan Looney reported that TCP can trigger the following crash
in tcp_shifted_skb() :
BUG_ON(tcp_skb_pcount(skb) < pcount);
This can happen if the remote peer has advertized the smallest
MSS that linux TCP accepts : 48
An skb can hold 17 fragments, and each fragment can hold 32KB
on x86, or 64KB on PowerPC.
This means that the 16bit witdh of TCP_SKB_CB(skb)->tcp_gso_segs
can overflow.
Note that tcp_sendmsg() builds skbs with less than 64KB
of payload, so this problem needs SACK to be enabled.
SACK blocks allow TCP to coalesce multiple skbs in the retransmit
queue, thus filling the 17 fragments to maximal capacity.
CVE-2019-11477 -- u16 overflow of TCP_SKB_CB(skb)->tcp_gso_segs
Fixes: 832d11c5cd ("tcp: Try to restore large SKBs while SACK processing")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Jonathan Looney <jtl@netflix.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Tyler Hicks <tyhicks@canonical.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Bruce Curtis <brucec@netflix.com>
Cc: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Alexei Starovoitov says:
====================
pull-request: bpf 2019-06-15
The following pull-request contains BPF updates for your *net* tree.
The main changes are:
1) fix stack layout of JITed x64 bpf code, from Alexei.
2) fix out of bounds memory access in bpf_sk_storage, from Arthur.
3) fix lpm trie walk, from Jonathan.
4) fix nested bpf_perf_event_output, from Matt.
5) and several other fixes.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
bpf_sk_storage maps use multiple spin locks to reduce contention.
The number of locks to use is determined by the number of possible CPUs.
With only 1 possible CPU, bucket_log == 0, and 2^0 = 1 locks are used.
When updating elements, the correct lock is determined with hash_ptr().
Calling hash_ptr() with 0 bits is undefined behavior, as it does:
x >> (64 - bits)
Using the value results in an out of bounds memory access.
In my case, this manifested itself as a page fault when raw_spin_lock_bh()
is called later, when running the self tests:
./tools/testing/selftests/bpf/test_verifier 773 775
[ 16.366342] BUG: unable to handle page fault for address: ffff8fe7a66f93f8
Force the minimum number of locks to two.
Signed-off-by: Arthur Fabre <afabre@cloudflare.com>
Fixes: 6ac99e8f23 ("bpf: Introduce bpf sk local storage")
Acked-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This config option makes only couple of lines optional.
Two small helpers and an int in couple of cls structs.
Remove the config option and always compile this in.
This saves the user from unexpected surprises when he adds
a filter with ingress device match which is silently ignored
in case the config option is not set.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Set the SOCK_DONE flag to match the TCP_CLOSING state when a peer has
shut down and there is nothing left to read.
This fixes the following bug:
1) Peer sends SHUTDOWN(RDWR).
2) Socket enters TCP_CLOSING but SOCK_DONE is not set.
3) read() returns -ENOTCONN until close() is called, then returns 0.
Signed-off-by: Stephen Barber <smbarber@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Get rid of the dsa_slave_switchdev_port_{attr_set,obj}_event functions
in favor of the switchdev_handle_port_{attr_set,obj_add,obj_del}
helpers which recurse into the lower devices of the target interface.
This has the benefit of being aware of the operations made on the
bridge device itself, where orig_dev is the bridge, and dev is the
slave. This can be used later to configure the hardware switches.
Only VLAN and (port) MDB objects not directly targeting the slave
device are unsupported at the moment, so skip this case in their
respective case statements.
Signed-off-by: Vivien Didelot <vivien.didelot@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The switchdev handle helpers make use of a device checking helper
requiring a const net_device. Make dsa_slave_dev_check compliant
to this.
Signed-off-by: Vivien Didelot <vivien.didelot@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The current DSA code handling switchdev objects does not recurse into
the lower devices thus is never called with an orig_dev member being
a bridge device, hence remove this useless check.
At the same time, remove the comments about the callers, which is
unlikely to be updated if the code changes and thus will be confusing.
Signed-off-by: Vivien Didelot <vivien.didelot@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
>From linux-3.7, (commit 5640f76858 "net: use a per task frag
allocator") TCP sendmsg() has preferred using order-3 allocations.
While it gives good results for most cases, we had reports
that heavy uses of TCP over loopback were hitting a spinlock
contention in page allocations/freeing.
This commits adds a sysctl so that admins can opt-in
for order-0 allocations. Hopefully mm layer might optimize
order-3 allocations in the future since it could give us
a nice boost (see 8 lines of following benchmark)
The following benchmark shows a win when more than 8 TCP_STREAM
threads are running (56 x86 cores server in my tests)
for thr in {1..30}
do
sysctl -wq net.core.high_order_alloc_disable=0
T0=`./super_netperf $thr -H 127.0.0.1 -l 15`
sysctl -wq net.core.high_order_alloc_disable=1
T1=`./super_netperf $thr -H 127.0.0.1 -l 15`
echo $thr:$T0:$T1
done
1: 49979: 37267
2: 98745: 76286
3: 141088: 110051
4: 177414: 144772
5: 197587: 173563
6: 215377: 208448
7: 241061: 234087
8: 267155: 263373
9: 295069: 297402
10: 312393: 335213
11: 340462: 368778
12: 371366: 403954
13: 412344: 443713
14: 426617: 473580
15: 474418: 507861
16: 503261: 538539
17: 522331: 563096
18: 532409: 567084
19: 550824: 605240
20: 525493: 641988
21: 564574: 665843
22: 567349: 690868
23: 583846: 710917
24: 588715: 736306
25: 603212: 763494
26: 604083: 792654
27: 602241: 796450
28: 604291: 797993
29: 611610: 833249
30: 577356: 841062
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Feng Tang reported a performance regression after introduction
of per TCP socket tx/rx caches, for TCP over loopback (netperf)
There is high chance the regression is caused by a change on
how well the 32 KB per-thread page (current->task_frag) can
be recycled, and lack of pcp caches for order-3 pages.
I could not reproduce the regression myself, cpus all being
spinning on the mm spinlocks for page allocs/freeing, regardless
of enabling or disabling the per tcp socket caches.
It seems best to disable the feature by default, and let
admins enabling it.
MM layer either needs to provide scalable order-3 pages
allocations, or could attempt a trylock on zone->lock if
the caller only attempts to get a high-order page and is
able to fallback to order-0 ones in case of pressure.
Tests run on a 56 cores host (112 hyper threads)
- 35.49% netperf [kernel.vmlinux] [k] queued_spin_lock_slowpath
- 35.49% queued_spin_lock_slowpath
- 18.18% get_page_from_freelist
- __alloc_pages_nodemask
- 18.18% alloc_pages_current
skb_page_frag_refill
sk_page_frag_refill
tcp_sendmsg_locked
tcp_sendmsg
inet_sendmsg
sock_sendmsg
__sys_sendto
__x64_sys_sendto
do_syscall_64
entry_SYSCALL_64_after_hwframe
__libc_send
+ 17.31% __free_pages_ok
+ 31.43% swapper [kernel.vmlinux] [k] intel_idle
+ 9.12% netperf [kernel.vmlinux] [k] copy_user_enhanced_fast_string
+ 6.53% netserver [kernel.vmlinux] [k] copy_user_enhanced_fast_string
+ 0.69% netserver [kernel.vmlinux] [k] queued_spin_lock_slowpath
+ 0.68% netperf [kernel.vmlinux] [k] skb_release_data
+ 0.52% netperf [kernel.vmlinux] [k] tcp_sendmsg_locked
0.46% netperf [kernel.vmlinux] [k] _raw_spin_lock_irqsave
Fixes: 472c2e07ee ("tcp: add one skb cache for tx")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Feng Tang <feng.tang@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of relying on rps_needed, it is safer to use a separate
static key, since we do not want to enable TCP rx_skb_cache
by default. This feature can cause huge increase of memory
usage on hosts with millions of sockets.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This was originally passed through to the VRF logic in compute_score().
But that logic has now been replaced by udp_sk_bound_dev_eq() and so
this code is no longer used or needed.
Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Originally this was used by the VRF logic in compute_score(), but that
was later replaced by udp_sk_bound_dev_eq() and the parameter became
unused.
Note this change adds an 'unused variable' compiler warning that will be
removed in the next patch (I've split the removal in two to make review
slightly easier).
Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we want to set a EDT time for the skb we want to send
via ip_send_unicast_reply(), we have to pass a new parameter
and initialize ipc.sockc.transmit_time with it.
This fixes the EDT time for ACK/RST packets sent on behalf of
a TIME_WAIT socket.
Fixes: a842fe1425 ("tcp: add optional per socket transmit delay")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
