Commit Graph

53672 Commits

Author SHA1 Message Date
Eric Dumazet
d9f92772e8 xfrm6: avoid potential infinite loop in _decode_session6()
syzbot found a way to trigger an infinitie loop by overflowing
@offset variable that has been forced to use u16 for some very
obscure reason in the past.

We probably want to look at NEXTHDR_FRAGMENT handling which looks
wrong, in a separate patch.

In net-next, we shall try to use skb_header_pointer() instead of
pskb_may_pull().

watchdog: BUG: soft lockup - CPU#1 stuck for 134s! [syz-executor738:4553]
Modules linked in:
irq event stamp: 13885653
hardirqs last  enabled at (13885652): [<ffffffff878009d5>] restore_regs_and_return_to_kernel+0x0/0x2b
hardirqs last disabled at (13885653): [<ffffffff87800905>] interrupt_entry+0xb5/0xf0 arch/x86/entry/entry_64.S:625
softirqs last  enabled at (13614028): [<ffffffff84df0809>] tun_napi_alloc_frags drivers/net/tun.c:1478 [inline]
softirqs last  enabled at (13614028): [<ffffffff84df0809>] tun_get_user+0x1dd9/0x4290 drivers/net/tun.c:1825
softirqs last disabled at (13614032): [<ffffffff84df1b6f>] tun_get_user+0x313f/0x4290 drivers/net/tun.c:1942
CPU: 1 PID: 4553 Comm: syz-executor738 Not tainted 4.17.0-rc3+ #40
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:check_kcov_mode kernel/kcov.c:67 [inline]
RIP: 0010:__sanitizer_cov_trace_pc+0x20/0x50 kernel/kcov.c:101
RSP: 0018:ffff8801d8cfe250 EFLAGS: 00000246 ORIG_RAX: ffffffffffffff13
RAX: ffff8801d88a8080 RBX: ffff8801d7389e40 RCX: 0000000000000006
RDX: 0000000000000000 RSI: ffffffff868da4ad RDI: ffff8801c8a53277
RBP: ffff8801d8cfe250 R08: ffff8801d88a8080 R09: ffff8801d8cfe3e8
R10: ffffed003b19fc87 R11: ffff8801d8cfe43f R12: ffff8801c8a5327f
R13: 0000000000000000 R14: ffff8801c8a4e5fe R15: ffff8801d8cfe3e8
FS:  0000000000d88940(0000) GS:ffff8801daf00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffffffffff600400 CR3: 00000001acab3000 CR4: 00000000001406e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 _decode_session6+0xc1d/0x14f0 net/ipv6/xfrm6_policy.c:150
 __xfrm_decode_session+0x71/0x140 net/xfrm/xfrm_policy.c:2368
 xfrm_decode_session_reverse include/net/xfrm.h:1213 [inline]
 icmpv6_route_lookup+0x395/0x6e0 net/ipv6/icmp.c:372
 icmp6_send+0x1982/0x2da0 net/ipv6/icmp.c:551
 icmpv6_send+0x17a/0x300 net/ipv6/ip6_icmp.c:43
 ip6_input_finish+0x14e1/0x1a30 net/ipv6/ip6_input.c:305
 NF_HOOK include/linux/netfilter.h:288 [inline]
 ip6_input+0xe1/0x5e0 net/ipv6/ip6_input.c:327
 dst_input include/net/dst.h:450 [inline]
 ip6_rcv_finish+0x29c/0xa10 net/ipv6/ip6_input.c:71
 NF_HOOK include/linux/netfilter.h:288 [inline]
 ipv6_rcv+0xeb8/0x2040 net/ipv6/ip6_input.c:208
 __netif_receive_skb_core+0x2468/0x3650 net/core/dev.c:4646
 __netif_receive_skb+0x2c/0x1e0 net/core/dev.c:4711
 netif_receive_skb_internal+0x126/0x7b0 net/core/dev.c:4785
 napi_frags_finish net/core/dev.c:5226 [inline]
 napi_gro_frags+0x631/0xc40 net/core/dev.c:5299
 tun_get_user+0x3168/0x4290 drivers/net/tun.c:1951
 tun_chr_write_iter+0xb9/0x154 drivers/net/tun.c:1996
 call_write_iter include/linux/fs.h:1784 [inline]
 do_iter_readv_writev+0x859/0xa50 fs/read_write.c:680
 do_iter_write+0x185/0x5f0 fs/read_write.c:959
 vfs_writev+0x1c7/0x330 fs/read_write.c:1004
 do_writev+0x112/0x2f0 fs/read_write.c:1039
 __do_sys_writev fs/read_write.c:1112 [inline]
 __se_sys_writev fs/read_write.c:1109 [inline]
 __x64_sys_writev+0x75/0xb0 fs/read_write.c:1109
 do_syscall_64+0x1b1/0x800 arch/x86/entry/common.c:287
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Reported-by: syzbot+0053c8...@syzkaller.appspotmail.com
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
2018-05-14 09:42:07 +02:00
Marek Lindner
18cfb44cf1 batman-adv: enable B.A.T.M.A.N. V compilation by default
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Acked-by: Antonio Quartulli <a@unstable.cc>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
2018-05-14 09:31:17 +02:00
Marek Lindner
8c46fcd783 batman-adv: disable ethtool link speed detection when auto negotiation off
Virtual interface drivers such as tun / tap interfaces, VLAN, etc tend
to initialize the interface throughput with some value for the sake of
having a throughput number to export via ethtool. This exported
throughput leaves batman-adv to conclude the interface throughput is
genuine (reflecting reality), thus no measurements are necessary.

Based on the observation that those interface types also tend to set
the link auto-negotiation to 'off', batman-adv shall check this
setting to differentiate between genuine link throughput information
and placeholders installed by virtual interfaces.

The "default throughput" setting exported via sysfs still allows to
configure the batman-adv throughput for the interface, thus disabling
the measurements.

Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Acked-by: Antonio Quartulli <a@unstable.cc>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
2018-05-14 09:30:57 +02:00
David S. Miller
4f6b15c3a6 Merge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf
Pablo Neira Ayuso says:

====================
Netfilter/IPVS fixes for net

The following patchset contains Netfilter/IPVS fixes for your net tree,
they are:

1) Fix handling of simultaneous open TCP connection in conntrack,
   from Jozsef Kadlecsik.

2) Insufficient sanitify check of xtables extension names, from
   Florian Westphal.

3) Skip unnecessary synchronize_rcu() call when transaction log
   is already empty, from Florian Westphal.

4) Incorrect destination mac validation in ebt_stp, from Stephen
   Hemminger.

5) xtables module reference counter leak in nft_compat, from
   Florian Westphal.

6) Incorrect connection reference counting logic in IPVS
   one-packet scheduler, from Julian Anastasov.

7) Wrong stats for 32-bits CPU in IPVS, also from Julian.

8) Calm down sparse error in netfilter core, also from Florian.

9) Use nla_strlcpy to fix compilation warning in nfnetlink_acct
   and nfnetlink_cthelper, again from Florian.

10) Missing module alias in icmp and icmp6 xtables extensions,
    from Florian Westphal.

11) Base chain statistics in nf_tables may be unset/null, from Florian.

12) Fix handling of large matchinfo size in nft_compat, this includes
    one preparation for before this fix. From Florian.

13) Fix bogus EBUSY error when deleting chains due to incorrect reference
    counting from the preparation phase of the two-phase commit protocol.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-13 20:28:47 -04:00
Anders Roxell
289e1f4e9e net: ipv4: ipconfig: fix unused variable
When CONFIG_PROC_FS isn't set, variable ipconfig_dir isn't used.
net/ipv4/ipconfig.c:167:31: warning: ‘ipconfig_dir’ defined but not used [-Wunused-variable]
 static struct proc_dir_entry *ipconfig_dir;
                               ^~~~~~~~~~~~
Move the declaration of ipconfig_dir inside the CONFIG_PROC_FS ifdef to
fix the warning.

Fixes: c04d2cb200 ("ipconfig: Write NTP server IPs to /proc/net/ipconfig/ntp_servers")
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-13 20:27:25 -04:00
Willem de Bruijn
b84bbaf7a6 packet: in packet_snd start writing at link layer allocation
Packet sockets allow construction of packets shorter than
dev->hard_header_len to accommodate protocols with variable length
link layer headers. These packets are padded to dev->hard_header_len,
because some device drivers interpret that as a minimum packet size.

packet_snd reserves dev->hard_header_len bytes on allocation.
SOCK_DGRAM sockets call skb_push in dev_hard_header() to ensure that
link layer headers are stored in the reserved range. SOCK_RAW sockets
do the same in tpacket_snd, but not in packet_snd.

Syzbot was able to send a zero byte packet to a device with massive
116B link layer header, causing padding to cross over into skb_shinfo.
Fix this by writing from the start of the llheader reserved range also
in the case of packet_snd/SOCK_RAW.

Update skb_set_network_header to the new offset. This also corrects
it for SOCK_DGRAM, where it incorrectly double counted reserve due to
the skb_push in dev_hard_header.

Fixes: 9ed988cd59 ("packet: validate variable length ll headers")
Reported-by: syzbot+71d74a5406d02057d559@syzkaller.appspotmail.com
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-13 20:19:27 -04:00
Colin Ian King
f0dfd7a2b3 netfilter: nf_tables: fix memory leak on error exit return
Currently the -EBUSY error return path is not free'ing resources
allocated earlier, leaving a memory leak. Fix this by exiting via the
error exit label err5 that performs the necessary resource clean
up.

Detected by CoverityScan, CID#1432975 ("Resource leak")

Fixes: 9744a6fcef ("netfilter: nf_tables: check if same extensions are set when adding elements")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2018-05-14 00:21:59 +02:00
Marek Lindner
16116dac23 batman-adv: prevent TT request storms by not sending inconsistent TT TLVLs
A translation table TVLV changset sent with an OGM consists
of a number of headers (one per VLAN) plus the changeset
itself (addition and/or deletion of entries).

The per-VLAN headers are used by OGM recipients for consistency
checks. Said consistency check might determine that a full
translation table request is needed to restore consistency. If
the TT sender adds per-VLAN headers of empty VLANs into the OGM,
recipients are led to believe to have reached an inconsistent
state and thus request a full table update. The full table does
not contain empty VLANs (due to missing entries) the cycle
restarts when the next OGM is issued.

Consequently, when the translation table TVLV headers are
composed, empty VLANs are to be excluded.

Fixes: 21a57f6e7a3b ("batman-adv: make the TT CRC logic VLAN specific")
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
2018-05-12 18:53:08 +02:00
Linus Lüssing
7072337e52 batman-adv: Fix TT sync flags for intermediate TT responses
The previous TT sync fix so far only fixed TT responses issued by the
target node directly. So far, TT responses issued by intermediate nodes
still lead to the wrong flags being added, leading to CRC mismatches.

This behaviour was observed at Freifunk Hannover in a 800 nodes setup
where a considerable amount of nodes were still infected with 'WI'
TT flags even with (most) nodes having the previous TT sync fix applied.

I was able to reproduce the issue with intermediate TT responses in a
four node test setup and this patch fixes this issue by ensuring to
use the per originator instead of the summarized, OR'd ones.

Fixes: e9c00136a4 ("batman-adv: fix tt_global_entries flags update")
Reported-by: Leonardo Mörlein <me@irrelefant.net>
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
2018-05-12 18:52:19 +02:00
David S. Miller
b2d6cee117 Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
The bpf syscall and selftests conflicts were trivial
overlapping changes.

The r8169 change involved moving the added mdelay from 'net' into a
different function.

A TLS close bug fix overlapped with the splitting of the TLS state
into separate TX and RX parts.  I just expanded the tests in the bug
fix from "ctx->conf == X" into "ctx->tx_conf == X && ctx->rx_conf
== X".

Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 20:53:22 -04:00
Linus Torvalds
4bc871984f Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Pull networking fixes from David Miller:

 1) Verify lengths of keys provided by the user is AF_KEY, from Kevin
    Easton.

 2) Add device ID for BCM89610 PHY. Thanks to Bhadram Varka.

 3) Add Spectre guards to some ATM code, courtesy of Gustavo A. R.
    Silva.

 4) Fix infinite loop in NSH protocol code. To Eric Dumazet we are most
    grateful for this fix.

 5) Line up /proc/net/netlink headers properly. This fix from YU Bo, we
    do appreciate.

 6) Use after free in TLS code. Once again we are blessed by the
    honorable Eric Dumazet with this fix.

 7) Fix regression in TLS code causing stalls on partial TLS records.
    This fix is bestowed upon us by Andrew Tomt.

 8) Deal with too small MTUs properly in LLC code, another great gift
    from Eric Dumazet.

 9) Handle cached route flushing properly wrt. MTU locking in ipv4, to
    Hangbin Liu we give thanks for this.

10) Fix regression in SO_BINDTODEVIC handling wrt. UDP socket demux.
    Paolo Abeni, he gave us this.

11) Range check coalescing parameters in mlx4 driver, thank you Moshe
    Shemesh.

12) Some ipv6 ICMP error handling fixes in rxrpc, from our good brother
    David Howells.

13) Fix kexec on mlx5 by freeing IRQs in shutdown path. Daniel Juergens,
    you're the best!

14) Don't send bonding RLB updates to invalid MAC addresses. Debabrata
    Benerjee saved us!

15) Uh oh, we were leaking in udp_sendmsg and ping_v4_sendmsg. The ship
    is now water tight, thanks to Andrey Ignatov.

16) IPSEC memory leak in ixgbe from Colin Ian King, man we've got holes
    everywhere!

17) Fix error path in tcf_proto_create, Jiri Pirko what would we do
    without you!

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (92 commits)
  net sched actions: fix refcnt leak in skbmod
  net: sched: fix error path in tcf_proto_create() when modules are not configured
  net sched actions: fix invalid pointer dereferencing if skbedit flags missing
  ixgbe: fix memory leak on ipsec allocation
  ixgbevf: fix ixgbevf_xmit_frame()'s return type
  ixgbe: return error on unsupported SFP module when resetting
  ice: Set rq_last_status when cleaning rq
  ipv4: fix memory leaks in udp_sendmsg, ping_v4_sendmsg
  mlxsw: core: Fix an error handling path in 'mlxsw_core_bus_device_register()'
  bonding: send learning packets for vlans on slave
  bonding: do not allow rlb updates to invalid mac
  net/mlx5e: Err if asked to offload TC match on frag being first
  net/mlx5: E-Switch, Include VF RDMA stats in vport statistics
  net/mlx5: Free IRQs in shutdown path
  rxrpc: Trace UDP transmission failure
  rxrpc: Add a tracepoint to log ICMP/ICMP6 and error messages
  rxrpc: Fix the min security level for kernel calls
  rxrpc: Fix error reception on AF_INET6 sockets
  rxrpc: Fix missing start of call timeout
  qed: fix spelling mistake: "taskelt" -> "tasklet"
  ...
2018-05-11 14:14:46 -07:00
Linus Torvalds
a1f45efbb9 Merge tag 'nfs-for-4.17-2' of git://git.linux-nfs.org/projects/anna/linux-nfs
Pull NFS client fixes from Anna Schumaker:
 "These patches fix both a possible corruption during NFSoRDMA MR
  recovery, and a sunrpc tracepoint crash.

  Additionally, Trond has a new email address to put in the MAINTAINERS
  file"

* tag 'nfs-for-4.17-2' of git://git.linux-nfs.org/projects/anna/linux-nfs:
  Change Trond's email address in MAINTAINERS
  sunrpc: Fix latency trace point crashes
  xprtrdma: Fix list corruption / DMAR errors during MR recovery
2018-05-11 13:56:43 -07:00
Roman Mashak
a52956dfc5 net sched actions: fix refcnt leak in skbmod
When application fails to pass flags in netlink TLV when replacing
existing skbmod action, the kernel will leak refcnt:

$ tc actions get action skbmod index 1
total acts 0

        action order 0: skbmod pipe set smac 00:11:22:33:44:55
         index 1 ref 1 bind 0

For example, at this point a buggy application replaces the action with
index 1 with new smac 00:aa:22:33:44:55, it fails because of zero flags,
however refcnt gets bumped:

$ tc actions get actions skbmod index 1
total acts 0

        action order 0: skbmod pipe set smac 00:11:22:33:44:55
         index 1 ref 2 bind 0
$

Tha patch fixes this by calling tcf_idr_release() on existing actions.

Fixes: 86da71b573 ("net_sched: Introduce skbmod action")
Signed-off-by: Roman Mashak <mrv@mojatatu.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 16:37:03 -04:00
Jiri Pirko
d68d75fdc3 net: sched: fix error path in tcf_proto_create() when modules are not configured
In case modules are not configured, error out when tp->ops is null
and prevent later null pointer dereference.

Fixes: 33a48927c1 ("sched: push TC filter protocol creation into a separate function")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 16:34:38 -04:00
William Tu
d5db21a3e6 erspan: auto detect truncated ipv6 packets.
Currently the truncated bit is set only when 1) the mirrored packet
is larger than mtu and 2) the ipv4 packet tot_len is larger than
the actual skb->len.  This patch adds another case for detecting
whether ipv6 packet is truncated or not, by checking the ipv6 header
payload_len and the skb->len.

Reported-by: Xiaoyan Jin <xiaoyanj@vmware.com>
Signed-off-by: William Tu <u9012063@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 16:03:49 -04:00
David S. Miller
f01008916f Merge tag 'rxrpc-fixes-20180510' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs
David Howells says:

====================
rxrpc: Fixes

Here are three fixes for AF_RXRPC and two tracepoints that were useful for
finding them:

 (1) Fix missing start of expect-Rx-by timeout on initial packet
     transmission so that calls will time out if the peer doesn't respond.

 (2) Fix error reception on AF_INET6 sockets by using the correct family of
     sockopts on the UDP transport socket.

 (3) Fix setting the minimum security level on kernel calls so that they
     can be encrypted.

 (4) Add a tracepoint to log ICMP/ICMP6 and other error reports from the
     transport socket.

 (5) Add a tracepoint to log UDP sendmsg failure so that we can find out if
     transmission failure occurred on the UDP socket.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 15:55:57 -04:00
Roman Mashak
af5d01842f net sched actions: fix invalid pointer dereferencing if skbedit flags missing
When application fails to pass flags in netlink TLV for a new skbedit action,
the kernel results in the following oops:

[    8.307732] BUG: unable to handle kernel paging request at 0000000000021130
[    8.309167] PGD 80000000193d1067 P4D 80000000193d1067 PUD 180e0067 PMD 0
[    8.310595] Oops: 0000 [#1] SMP PTI
[    8.311334] Modules linked in: kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel pcbc aesni_intel aes_x86_64 crypto_simd cryptd glue_helper serio_raw
[    8.314190] CPU: 1 PID: 397 Comm: tc Not tainted 4.17.0-rc3+ #357
[    8.315252] RIP: 0010:__tcf_idr_release+0x33/0x140
[    8.316203] RSP: 0018:ffffa0718038f840 EFLAGS: 00010246
[    8.317123] RAX: 0000000000000001 RBX: 0000000000021100 RCX: 0000000000000000
[    8.319831] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000021100
[    8.321181] RBP: 0000000000000000 R08: 000000000004adf8 R09: 0000000000000122
[    8.322645] R10: 0000000000000000 R11: ffffffff9e5b01ed R12: 0000000000000000
[    8.324157] R13: ffffffff9e0d3cc0 R14: 0000000000000000 R15: 0000000000000000
[    8.325590] FS:  00007f591292e700(0000) GS:ffff8fcf5bc40000(0000) knlGS:0000000000000000
[    8.327001] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    8.327987] CR2: 0000000000021130 CR3: 00000000180e6004 CR4: 00000000001606a0
[    8.329289] Call Trace:
[    8.329735]  tcf_skbedit_init+0xa7/0xb0
[    8.330423]  tcf_action_init_1+0x362/0x410
[    8.331139]  ? try_to_wake_up+0x44/0x430
[    8.331817]  tcf_action_init+0x103/0x190
[    8.332511]  tc_ctl_action+0x11a/0x220
[    8.333174]  rtnetlink_rcv_msg+0x23d/0x2e0
[    8.333902]  ? _cond_resched+0x16/0x40
[    8.334569]  ? __kmalloc_node_track_caller+0x5b/0x2c0
[    8.335440]  ? rtnl_calcit.isra.31+0xf0/0xf0
[    8.336178]  netlink_rcv_skb+0xdb/0x110
[    8.336855]  netlink_unicast+0x167/0x220
[    8.337550]  netlink_sendmsg+0x2a7/0x390
[    8.338258]  sock_sendmsg+0x30/0x40
[    8.338865]  ___sys_sendmsg+0x2c5/0x2e0
[    8.339531]  ? pagecache_get_page+0x27/0x210
[    8.340271]  ? filemap_fault+0xa2/0x630
[    8.340943]  ? page_add_file_rmap+0x108/0x200
[    8.341732]  ? alloc_set_pte+0x2aa/0x530
[    8.342573]  ? finish_fault+0x4e/0x70
[    8.343332]  ? __handle_mm_fault+0xbc1/0x10d0
[    8.344337]  ? __sys_sendmsg+0x53/0x80
[    8.345040]  __sys_sendmsg+0x53/0x80
[    8.345678]  do_syscall_64+0x4f/0x100
[    8.346339]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[    8.347206] RIP: 0033:0x7f591191da67
[    8.347831] RSP: 002b:00007fff745abd48 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
[    8.349179] RAX: ffffffffffffffda RBX: 00007fff745abe70 RCX: 00007f591191da67
[    8.350431] RDX: 0000000000000000 RSI: 00007fff745abdc0 RDI: 0000000000000003
[    8.351659] RBP: 000000005af35251 R08: 0000000000000001 R09: 0000000000000000
[    8.352922] R10: 00000000000005f1 R11: 0000000000000246 R12: 0000000000000000
[    8.354183] R13: 00007fff745afed0 R14: 0000000000000001 R15: 00000000006767c0
[    8.355400] Code: 41 89 d4 53 89 f5 48 89 fb e8 aa 20 fd ff 85 c0 0f 84 ed 00
00 00 48 85 db 0f 84 cf 00 00 00 40 84 ed 0f 85 cd 00 00 00 45 84 e4 <8b> 53 30
74 0d 85 d2 b8 ff ff ff ff 0f 8f b3 00 00 00 8b 43 2c
[    8.358699] RIP: __tcf_idr_release+0x33/0x140 RSP: ffffa0718038f840
[    8.359770] CR2: 0000000000021130
[    8.360438] ---[ end trace 60c66be45dfc14f0 ]---

The caller calls action's ->init() and passes pointer to "struct tc_action *a",
which later may be initialized to point at the existing action, otherwise
"struct tc_action *a" is still invalid, and therefore dereferencing it is an
error as happens in tcf_idr_release, where refcnt is decremented.

So in case of missing flags tcf_idr_release must be called only for
existing actions.

v2:
    - prepare patch for net tree

Fixes: 5e1567aeb7 ("net sched: skbedit action fix late binding")
Signed-off-by: Roman Mashak <mrv@mojatatu.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 15:52:43 -04:00
Chuck Lever
99722fe4d5 svcrdma: Persistently allocate and DMA-map Send buffers
While sending each RPC Reply, svc_rdma_sendto allocates and DMA-
maps a separate buffer where the RPC/RDMA transport header is
constructed. The buffer is unmapped and released in the Send
completion handler. This is significant per-RPC overhead,
especially for small RPCs.

Instead, allocate and DMA-map a buffer, and cache it in each
svc_rdma_send_ctxt. This buffer and its mapping can be re-used
for each RPC, saving the cost of memory allocation and DMA
mapping.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
3abb03face svcrdma: Simplify svc_rdma_send()
Clean up: No current caller of svc_rdma_send's passes in a chained
WR. The logic that counts the chain length can be replaced with a
constant (1).

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
986b78894b svcrdma: Remove post_send_wr
Clean up: Now that the send_wr is part of the svc_rdma_send_ctxt,
svc_rdma_post_send_wr is nearly empty.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
25fd86eca1 svcrdma: Don't overrun the SGE array in svc_rdma_send_ctxt
Receive buffers are always the same size, but each Send WR has a
variable number of SGEs, based on the contents of the xdr_buf being
sent.

While assembling a Send WR, keep track of the number of SGEs so that
we don't exceed the device's maximum, or walk off the end of the
Send SGE array.

For now the Send path just fails if it exceeds the maximum.

The current logic in svc_rdma_accept bases the maximum number of
Send SGEs on the largest NFS request that can be sent or received.
In the transport layer, the limit is actually based on the
capabilities of the underlying device, not on properties of the
Upper Layer Protocol.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
4201c74647 svcrdma: Introduce svc_rdma_send_ctxt
svc_rdma_op_ctxt's are pre-allocated and maintained on a per-xprt
free list. This eliminates the overhead of calling kmalloc / kfree,
both of which grab a globally shared lock that disables interrupts.
Introduce a replacement to svc_rdma_op_ctxt's that is built
especially for the svcrdma Send path.

Subsequent patches will take advantage of this new structure by
allocating real resources which are then cached in these objects.
The allocations are freed when the transport is torn down.

I've renamed the structure so that static type checking can be used
to ensure that uses of op_ctxt and send_ctxt are not confused. As an
additional clean up, structure fields are renamed to conform with
kernel coding conventions.

Additional clean ups:
- Handle svc_rdma_send_ctxt_get allocation failure at each call
  site, rather than pre-allocating and hoping we guessed correctly
- All send_ctxt_put call-sites request page freeing, so remove
  the @free_pages argument
- All send_ctxt_put call-sites unmap SGEs, so fold that into
  svc_rdma_send_ctxt_put

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
232627905f svcrdma: Clean up Send SGE accounting
Clean up: Since there's already a svc_rdma_op_ctxt being passed
around with the running count of mapped SGEs, drop unneeded
parameters to svc_rdma_post_send_wr().

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
f016f305f9 svcrdma: Refactor svc_rdma_dma_map_buf
Clean up: svc_rdma_dma_map_buf does mostly the same thing as
svc_rdma_dma_map_page, so let's fold these together.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
eb5d7a622e svcrdma: Allocate recv_ctxt's on CPU handling Receives
There is a significant latency penalty when processing an ingress
Receive if the Receive buffer resides in memory that is not on the
same NUMA node as the the CPU handling completions for a CQ.

The system administrator and the device driver determine which CPU
handles completions. This CPU does not change during life of the CQ.
Further the Upper Layer does not have any visibility of which CPU it
is.

Allocating Receive buffers in the Receive completion handler
guarantees that Receive buffers are allocated on the preferred NUMA
node for that CQ.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
3316f06311 svcrdma: Persistently allocate and DMA-map Receive buffers
The current Receive path uses an array of pages which are allocated
and DMA mapped when each Receive WR is posted, and then handed off
to the upper layer in rqstp::rq_arg. The page flip releases unused
pages in the rq_pages pagelist. This mechanism introduces a
significant amount of overhead.

So instead, kmalloc the Receive buffer, and leave it DMA-mapped
while the transport remains connected. This confers a number of
benefits:

* Each Receive WR requires only one receive SGE, no matter how large
  the inline threshold is. This helps the server-side NFS/RDMA
  transport operate on less capable RDMA devices.

* The Receive buffer is left allocated and mapped all the time. This
  relieves svc_rdma_post_recv from the overhead of allocating and
  DMA-mapping a fresh buffer.

* svc_rdma_wc_receive no longer has to DMA unmap the Receive buffer.
  It has to DMA sync only the number of bytes that were received.

* svc_rdma_build_arg_xdr no longer has to free a page in rq_pages
  for each page in the Receive buffer, making it a constant-time
  function.

* The Receive buffer is now plugged directly into the rq_arg's
  head[0].iov_vec, and can be larger than a page without spilling
  over into rq_arg's page list. This enables simplification of
  the RDMA Read path in subsequent patches.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
3a88092ee3 svcrdma: Preserve Receive buffer until svc_rdma_sendto
Rather than releasing the incoming svc_rdma_recv_ctxt at the end of
svc_rdma_recvfrom, hold onto it until svc_rdma_sendto.

This permits the contents of the Receive buffer to be preserved
through svc_process and then referenced directly in sendto as it
constructs Write and Reply chunks to return to the client.

The real changes will come in subsequent patches.

Note: I cannot use ->xpo_release_rqst for this purpose because that
is called _before_ ->xpo_sendto. svc_rdma_sendto uses information in
the received Call transport header to construct the Reply transport
header, which is preserved in the RPC's Receive buffer.

The historical comment in svc_send() isn't helpful: it is already
obvious that ->xpo_release_rqst is being called before ->xpo_sendto,
but there is no explanation for this ordering going back to the
beginning of the git era.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
1e5f416074 svcrdma: Simplify svc_rdma_recv_ctxt_put
Currently svc_rdma_recv_ctxt_put's callers have to know whether they
want to free the ctxt's pages or not. This means the human
developers have to know when and why to set that free_pages
argument.

Instead, the ctxt should carry that information with it so that
svc_rdma_recv_ctxt_put does the right thing no matter who is
calling.

We want to keep track of the number of pages in the Receive buffer
separately from the number of pages pulled over by RDMA Read. This
is so that the correct number of pages can be freed properly and
that number is well-documented.

So now, rc_hdr_count is the number of pages consumed by head[0]
(ie., the page index where the Read chunk should start); and
rc_page_count is always the number of pages that need to be released
when the ctxt is put.

The @free_pages argument is no longer needed.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
2c577bfea8 svcrdma: Remove sc_rq_depth
Clean up: No need to retain rq_depth in struct svcrdma_xprt, it is
used only in svc_rdma_accept().

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
ecf85b2384 svcrdma: Introduce svc_rdma_recv_ctxt
svc_rdma_op_ctxt's are pre-allocated and maintained on a per-xprt
free list. This eliminates the overhead of calling kmalloc / kfree,
both of which grab a globally shared lock that disables interrupts.
To reduce contention further, separate the use of these objects in
the Receive and Send paths in svcrdma.

Subsequent patches will take advantage of this separation by
allocating real resources which are then cached in these objects.
The allocations are freed when the transport is torn down.

I've renamed the structure so that static type checking can be used
to ensure that uses of op_ctxt and recv_ctxt are not confused. As an
additional clean up, structure fields are renamed to conform with
kernel coding conventions.

As a final clean up, helpers related to recv_ctxt are moved closer
to the functions that use them.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
bd2abef333 svcrdma: Trace key RDMA API events
This includes:
  * Posting on the Send and Receive queues
  * Send, Receive, Read, and Write completion
  * Connect upcalls
  * QP errors

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
98895edbe3 svcrdma: Trace key RPC/RDMA protocol events
This includes:
  * Transport accept and tear-down
  * Decisions about using Write and Reply chunks
  * Each RDMA segment that is handled
  * Whenever an RDMA_ERR is sent

As a clean-up, I've standardized the order of the includes, and
removed some now redundant dprintk call sites.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
b6e717cbf2 xprtrdma: Prepare RPC/RDMA includes for server-side trace points
Clean up: Move #include <trace/events/rpcrdma.h> into source files,
similar to how it is done with trace/events/sunrpc.h.

Server-side trace points will be part of the rpcrdma subsystem,
just like the client-side trace points.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
8dafcbee41 svcrdma: Use passed-in net namespace when creating RDMA listener
Ensure each RDMA listener and its children transports are created in
the same net namespace as the user that started the NFS service.
This is similar to how listener sockets are created in
svc_create_socket, required for enabling support for containers.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Chuck Lever
bcf3ffd405 svcrdma: Add proper SPDX tags for NetApp-contributed source
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
2018-05-11 15:48:57 -04:00
Eric Dumazet
575b65bc5b udp: avoid refcount_t saturation in __udp_gso_segment()
For some reason, Willem thought that the issue we fixed for TCP
in commit 7ec318feee ("tcp: gso: avoid refcount_t warning from
tcp_gso_segment()") was not relevant for UDP GSO.

But syzbot found its way.

refcount_t: saturated; leaking memory.
WARNING: CPU: 0 PID: 10261 at lib/refcount.c:78 refcount_add_not_zero+0x2d4/0x320 lib/refcount.c:78
Kernel panic - not syncing: panic_on_warn set ...

CPU: 0 PID: 10261 Comm: syz-executor5 Not tainted 4.17.0-rc3+ #38
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:77 [inline]
 dump_stack+0x1b9/0x294 lib/dump_stack.c:113
 panic+0x22f/0x4de kernel/panic.c:184
 __warn.cold.8+0x163/0x1b3 kernel/panic.c:536
 report_bug+0x252/0x2d0 lib/bug.c:186
 fixup_bug arch/x86/kernel/traps.c:178 [inline]
 do_error_trap+0x1de/0x490 arch/x86/kernel/traps.c:296
 do_invalid_op+0x1b/0x20 arch/x86/kernel/traps.c:315
 invalid_op+0x14/0x20 arch/x86/entry/entry_64.S:992
RIP: 0010:refcount_add_not_zero+0x2d4/0x320 lib/refcount.c:78
RSP: 0018:ffff880196db6b90 EFLAGS: 00010282
RAX: 0000000000000026 RBX: 00000000ffffff01 RCX: ffffc900040d9000
RDX: 0000000000004a29 RSI: ffffffff8160f6f1 RDI: ffff880196db66f0
RBP: ffff880196db6c78 R08: ffff8801b33d6740 R09: 0000000000000002
R10: ffff8801b33d6740 R11: 0000000000000000 R12: 0000000000000000
R13: 00000000ffffffff R14: ffff880196db6c50 R15: 0000000000020101
 refcount_add+0x1b/0x70 lib/refcount.c:102
 __udp_gso_segment+0xaa5/0xee0 net/ipv4/udp_offload.c:272
 udp4_ufo_fragment+0x592/0x7a0 net/ipv4/udp_offload.c:301
 inet_gso_segment+0x639/0x12b0 net/ipv4/af_inet.c:1342
 skb_mac_gso_segment+0x3ad/0x720 net/core/dev.c:2792
 __skb_gso_segment+0x3bb/0x870 net/core/dev.c:2865
 skb_gso_segment include/linux/netdevice.h:4050 [inline]
 validate_xmit_skb+0x54d/0xd90 net/core/dev.c:3122
 __dev_queue_xmit+0xbf8/0x34c0 net/core/dev.c:3579
 dev_queue_xmit+0x17/0x20 net/core/dev.c:3620
 neigh_direct_output+0x15/0x20 net/core/neighbour.c:1401
 neigh_output include/net/neighbour.h:483 [inline]
 ip_finish_output2+0xa5f/0x1840 net/ipv4/ip_output.c:229
 ip_finish_output+0x828/0xf80 net/ipv4/ip_output.c:317
 NF_HOOK_COND include/linux/netfilter.h:277 [inline]
 ip_output+0x21b/0x850 net/ipv4/ip_output.c:405
 dst_output include/net/dst.h:444 [inline]
 ip_local_out+0xc5/0x1b0 net/ipv4/ip_output.c:124
 ip_send_skb+0x40/0xe0 net/ipv4/ip_output.c:1434
 udp_send_skb.isra.37+0x5eb/0x1000 net/ipv4/udp.c:825
 udp_push_pending_frames+0x5c/0xf0 net/ipv4/udp.c:853
 udp_v6_push_pending_frames+0x380/0x3e0 net/ipv6/udp.c:1105
 udp_lib_setsockopt+0x59a/0x600 net/ipv4/udp.c:2403
 udpv6_setsockopt+0x95/0xa0 net/ipv6/udp.c:1447
 sock_common_setsockopt+0x9a/0xe0 net/core/sock.c:3046
 __sys_setsockopt+0x1bd/0x390 net/socket.c:1903
 __do_sys_setsockopt net/socket.c:1914 [inline]
 __se_sys_setsockopt net/socket.c:1911 [inline]
 __x64_sys_setsockopt+0xbe/0x150 net/socket.c:1911
 do_syscall_64+0x1b1/0x800 arch/x86/entry/common.c:287
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Fixes: ad405857b1 ("udp: better wmem accounting on gso")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 12:29:42 -04:00
Eric Dumazet
73a6bab5aa tcp: switch pacing timer to softirq based hrtimer
linux-4.16 got support for softirq based hrtimers.
TCP can switch its pacing hrtimer to this variant, since this
avoids going through a tasklet and some atomic operations.

pacing timer logic looks like other (jiffies based) tcp timers.

v2: use hrtimer_try_to_cancel() in tcp_clear_xmit_timers()
    to correctly release reference on socket if needed.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 12:24:37 -04:00
Florian Fainelli
aab9c4067d net: dsa: Plug in PHYLINK support
Add support for PHYLINK within the DSA subsystem in order to support more
complex devices such as pluggable (SFP) and non-pluggable (SFF) modules, 10G
PHYs, and traditional PHYs. Using PHYLINK allows us to drop some amount of
complexity we had while probing fixed and non-fixed PHYs using Device Tree.

Because PHYLINK separates the Ethernet MAC/port configuration into different
stages, we let switch drivers implement those, and for now, we maintain
functionality by calling dsa_slave_adjust_link() during
phylink_mac_link_{up,down} which provides semantically equivalent steps.

Drivers willing to take advantage of PHYLINK should implement the phylink_mac_*
operations that DSA wraps.

We cannot quite remove the adjust_link() callback just yet, because a number of
drivers rely on that for configuring their "CPU" and "DSA" ports, this is done
dsa_port_setup_phy_of() and dsa_port_fixed_link_register_of() still.

Drivers that utilize fixed links for user-facing ports (e.g: bcm_sf2) will need
to implement phylink_mac_ops from now on to preserve functionality, since PHYLINK
*does not* create a phy_device instance for fixed links.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 12:03:06 -04:00
Florian Fainelli
c4aef9fc0d net: dsa: Eliminate dsa_slave_get_link()
Since we use PHYLIB to manage the per-port link indication, this will
also be reflected correctly in the network device's carrier state, so we
can use ethtool_op_get_link() instead.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 12:03:06 -04:00
Florian Fainelli
11d8f3ddab net: dsa: Add PHYLINK switch operations
In preparation for adding support for PHYLINK within DSA, define a number of
operations that we will need and that switch drivers can start implementing.
Proper integration with PHYLINK will follow in subsequent patches.

We start selecting PHYLINK (which implies PHYLIB) in net/dsa/Kconfig
such that drivers can be guaranteed that this dependency is properly
taken care of and can start referencing PHYLINK helper functions without
requiring stubs or anything.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 12:03:05 -04:00
Andrey Ignatov
1b97013bfb ipv4: fix memory leaks in udp_sendmsg, ping_v4_sendmsg
Fix more memory leaks in ip_cmsg_send() callers. Part of them were fixed
earlier in 919483096b.

* udp_sendmsg one was there since the beginning when linux sources were
  first added to git;
* ping_v4_sendmsg one was copy/pasted in c319b4d76b.

Whenever return happens in udp_sendmsg() or ping_v4_sendmsg() IP options
have to be freed if they were allocated previously.

Add label so that future callers (if any) can use it instead of kfree()
before return that is easy to forget.

Fixes: c319b4d76b (net: ipv4: add IPPROTO_ICMP socket kind)
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-11 12:00:58 -04:00
David Howells
6b47fe1d1c rxrpc: Trace UDP transmission failure
Add a tracepoint to log transmission failure from the UDP transport socket
being used by AF_RXRPC.

Signed-off-by: David Howells <dhowells@redhat.com>
2018-05-10 23:26:01 +01:00
David Howells
494337c918 rxrpc: Add a tracepoint to log ICMP/ICMP6 and error messages
Add a tracepoint to log received ICMP/ICMP6 events and other error
messages.

Signed-off-by: David Howells <dhowells@redhat.com>
2018-05-10 23:26:01 +01:00
David Howells
93864fc3ff rxrpc: Fix the min security level for kernel calls
Fix the kernel call initiation to set the minimum security level for kernel
initiated calls (such as from kAFS) from the sockopt value.

Fixes: 19ffa01c9c ("rxrpc: Use structs to hold connection params and protocol info")
Signed-off-by: David Howells <dhowells@redhat.com>
2018-05-10 23:26:01 +01:00
David Howells
f2aeed3a59 rxrpc: Fix error reception on AF_INET6 sockets
AF_RXRPC tries to turn on IP_RECVERR and IP_MTU_DISCOVER on the UDP socket
it just opened for communications with the outside world, regardless of the
type of socket.  Unfortunately, this doesn't work with an AF_INET6 socket.

Fix this by turning on IPV6_RECVERR and IPV6_MTU_DISCOVER instead if the
socket is of the AF_INET6 family.

Without this, kAFS server and address rotation doesn't work correctly
because the algorithm doesn't detect received network errors.

Fixes: 75b54cb57c ("rxrpc: Add IPv6 support")
Signed-off-by: David Howells <dhowells@redhat.com>
2018-05-10 23:26:00 +01:00
David Howells
c54e43d752 rxrpc: Fix missing start of call timeout
The expect_rx_by call timeout is supposed to be set when a call is started
to indicate that we need to receive a packet by that point.  This is
currently put back every time we receive a packet, but it isn't started
when we first send a packet.  Without this, the call may wait forever if
the server doesn't deign to reply.

Fix this by setting the timeout upon a successful UDP sendmsg call for the
first DATA packet.  The timeout is initiated only for initial transmission
and not for subsequent retries as we don't want the retry mechanism to
extend the timeout indefinitely.

Fixes: a158bdd324 ("rxrpc: Fix call timeouts")
Reported-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
2018-05-10 23:26:00 +01:00
David Ahern
87f5fc7e48 bpf: Provide helper to do forwarding lookups in kernel FIB table
Provide a helper for doing a FIB and neighbor lookup in the kernel
tables from an XDP program. The helper provides a fastpath for forwarding
packets. If the packet is a local delivery or for any reason is not a
simple lookup and forward, the packet continues up the stack.

If it is to be forwarded, the forwarding can be done directly if the
neighbor is already known. If the neighbor does not exist, the first
few packets go up the stack for neighbor resolution. Once resolved, the
xdp program provides the fast path.

On successful lookup the nexthop dmac, current device smac and egress
device index are returned.

The API supports IPv4, IPv6 and MPLS protocols, but only IPv4 and IPv6
are implemented in this patch. The API includes layer 4 parameters if
the XDP program chooses to do deep packet inspection to allow compare
against ACLs implemented as FIB rules.

Header rewrite is left to the XDP program.

The lookup takes 2 flags:
- BPF_FIB_LOOKUP_DIRECT to do a lookup that bypasses FIB rules and goes
  straight to the table associated with the device (expert setting for
  those looking to maximize throughput)

- BPF_FIB_LOOKUP_OUTPUT to do a lookup from the egress perspective.
  Default is an ingress lookup.

Initial performance numbers collected by Jesper, forwarded packets/sec:

       Full stack    XDP FIB lookup    XDP Direct lookup
IPv4   1,947,969       7,074,156          7,415,333
IPv6   1,728,000       6,165,504          7,262,720

These number are single CPU core forwarding on a Broadwell
E5-1650 v4 @ 3.60GHz.

Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-11 00:10:57 +02:00
David Ahern
65a2022e89 net/ipv6: Add fib lookup stubs for use in bpf helper
Add stubs to retrieve a handle to an IPv6 FIB table, fib6_get_table,
a stub to do a lookup in a specific table, fib6_table_lookup, and
a stub for a full route lookup.

The stubs are needed for core bpf code to handle the case when the
IPv6 module is not builtin.

Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-11 00:10:57 +02:00
David Ahern
d4bea421f7 net/ipv6: Update fib6 tracepoint to take fib6_info
Similar to IPv4, IPv6 should use the FIB lookup result in the
tracepoint.

Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-11 00:10:57 +02:00
David Ahern
138118ec96 net/ipv6: Add fib6_lookup
Add IPv6 equivalent to fib_lookup. Does a fib lookup, including rules,
but returns a FIB entry, fib6_info, rather than a dst based rt6_info.
fib6_lookup is any where from 140% (MULTIPLE_TABLES config disabled)
to 60% faster than any of the dst based lookup methods (without custom
rules) and 25% faster with custom rules (e.g., l3mdev rule).

Since the lookup function has a completely different signature,
fib6_rule_action is split into 2 paths: the existing one is
renamed __fib6_rule_action and a new one for the fib6_info path
is added. fib6_rule_action decides which to call based on the
lookup_ptr. If it is fib6_table_lookup then the new path is taken.

Caller must hold rcu lock as no reference is taken on the returned
fib entry.

Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-11 00:10:56 +02:00