David S. Miller [Fri, 8 Dec 2017 18:27:44 +0000 (13:27 -0500)]
Merge branch 'tcp-bbr-sampling-fixes'
Neal Cardwell says:
====================
TCP BBR sampling fixes for loss recovery undo
This patch series has a few minor bug fixes for cases where spurious
loss recoveries can trick BBR estimators into estimating that the
available bandwidth is much lower than the true available bandwidth.
In both cases the fix here is to just reset the estimator upon loss
recovery undo.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Neal Cardwell [Thu, 7 Dec 2017 17:43:32 +0000 (12:43 -0500)]
tcp_bbr: reset long-term bandwidth sampling on loss recovery undo
Fix BBR so that upon notification of a loss recovery undo BBR resets
long-term bandwidth sampling.
Under high reordering, reordering events can be interpreted as loss.
If the reordering and spurious loss estimates are high enough, this
can cause BBR to spuriously estimate that we are seeing loss rates
high enough to trigger long-term bandwidth estimation. To avoid that
problem, this commit resets long-term bandwidth sampling on loss
recovery undo events.
Signed-off-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Yuchung Cheng <ycheng@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Neal Cardwell [Thu, 7 Dec 2017 17:43:31 +0000 (12:43 -0500)]
tcp_bbr: reset full pipe detection on loss recovery undo
Fix BBR so that upon notification of a loss recovery undo BBR resets
the full pipe detection (STARTUP exit) state machine.
Under high reordering, reordering events can be interpreted as loss.
If the reordering and spurious loss estimates are high enough, this
could previously cause BBR to spuriously estimate that the pipe is
full.
Since spurious loss recovery means that our overall sending will have
slowed down spuriously, this commit gives a flow more time to probe
robustly for bandwidth and decide the pipe is really full.
Signed-off-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Yuchung Cheng <ycheng@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Neal Cardwell [Thu, 7 Dec 2017 17:43:30 +0000 (12:43 -0500)]
tcp_bbr: record "full bw reached" decision in new full_bw_reached bit
This commit records the "full bw reached" decision in a new
full_bw_reached bit. This is a pure refactor that does not change the
current behavior, but enables subsequent fixes and improvements.
In particular, this enables simple and clean fixes because the full_bw
and full_bw_cnt can be unconditionally zeroed without worrying about
forgetting that we estimated we filled the pipe in Startup. And it
enables future improvements because multiple code paths can be used
for estimating that we filled the pipe in Startup; any new code paths
only need to set this bit when they think the pipe is full.
Note that this fix intentionally reduces the width of the full_bw_cnt
counter, since we have never used the most significant bit.
Signed-off-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Yuchung Cheng <ycheng@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Bert Kenward [Thu, 7 Dec 2017 17:18:58 +0000 (17:18 +0000)]
sfc: pass valid pointers from efx_enqueue_unwind
The bytes_compl and pkts_compl pointers passed to efx_dequeue_buffers
cannot be NULL. Add a paranoid warning to check this condition and fix
the one case where they were NULL.
efx_enqueue_unwind() is called very rarely, during error handling.
Without this fix it would fail with a NULL pointer dereference in
efx_dequeue_buffer, with efx_enqueue_skb in the call stack.
Fixes: e9117e5099ea ("sfc: Firmware-Assisted TSO version 2") Reported-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Bert Kenward <bkenward@solarflare.com> Tested-by: Jarod Wilson <jarod@redhat.com> Acked-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Claudiu Manoil [Thu, 7 Dec 2017 16:44:23 +0000 (18:44 +0200)]
gianfar: Disable EEE autoneg by default
This controller does not support EEE, but it may connect to a PHY
which supports EEE and advertises EEE by default, while its link
partner also advertises EEE. If this happens, the PHY enters low
power mode when the traffic rate is low and causes packet loss.
This patch disables EEE advertisement by default for any PHY that
gianfar connects to, to prevent the above unwanted outcome.
Signed-off-by: Shaohui Xie <Shaohui.Xie@nxp.com> Tested-by: Yangbo Lu <Yangbo.lu@nxp.com> Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Yousuk Seung [Thu, 7 Dec 2017 21:41:34 +0000 (13:41 -0800)]
tcp: invalidate rate samples during SACK reneging
Mark tcp_sock during a SACK reneging event and invalidate rate samples
while marked. Such rate samples may overestimate bw by including packets
that were SACKed before reneging.
In above example the rate sample taken after the last ack will count
7001-38001 as delivered while the actual delivery rate likely could
be much lower i.e. 7001-8001.
This patch adds a new field tcp_sock.sack_reneg and marks it when we
declare SACK reneging and entering TCP_CA_Loss, and unmarks it after
the last rate sample was taken before moving back to TCP_CA_Open. This
patch also invalidates rate samples taken while tcp_sock.is_sack_reneg
is set.
Fixes: b9f64820fb22 ("tcp: track data delivery rate for a TCP connection") Signed-off-by: Yousuk Seung <ysseung@google.com> Signed-off-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: Yuchung Cheng <ycheng@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Acked-by: Eric Dumazet <edumazet@google.com> Acked-by: Priyaranjan Jha <priyarjha@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The following pull-request contains BPF updates for your *net* tree.
The main changes are:
1) Fixing broken uapi for BPF tracing programs for s390 and arm64
architectures due to pt_regs being in-kernel only, and not part
of uapi right now. A wrapper is added that exports pt_regs in
an asm-generic way. For arm64 this maps to existing user_pt_regs
structure and for s390 a user_pt_regs structure exporting the
beginning of pt_regs is added and uapi-exported, thus fixing the
BPF issues seen in perf (and BPF selftests), all from Hendrik.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Bjørn Mork [Wed, 6 Dec 2017 19:21:24 +0000 (20:21 +0100)]
usbnet: fix alignment for frames with no ethernet header
The qmi_wwan minidriver support a 'raw-ip' mode where frames are
received without any ethernet header. This causes alignment issues
because the skbs allocated by usbnet are "IP aligned".
Fix by allowing minidrivers to disable the additional alignment
offset. This is implemented using a per-device flag, since the same
minidriver also supports 'ethernet' mode.
Fixes: 32f7adf633b9 ("net: qmi_wwan: support "raw IP" mode") Reported-and-tested-by: Jay Foster <jay@systech.com> Signed-off-by: Bjørn Mork <bjorn@mork.no> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 6 Dec 2017 19:08:19 +0000 (11:08 -0800)]
tcp: use current time in tcp_rcv_space_adjust()
When I switched rcv_rtt_est to high resolution timestamps, I forgot
that tp->tcp_mstamp needed to be refreshed in tcp_rcv_space_adjust()
Using an old timestamp leads to autotuning lags.
Fixes: 645f4c6f2ebd ("tcp: switch rcv_rtt_est and rcvq_space to high resolution timestamps") Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Wei Wang <weiwan@google.com> Cc: Neal Cardwell <ncardwell@google.com> Cc: Yuchung Cheng <ycheng@google.com> Acked-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David Ahern [Thu, 7 Dec 2017 04:09:12 +0000 (20:09 -0800)]
netlink: Relax attr validation for fixed length types
Commit 28033ae4e0f5 ("net: netlink: Update attr validation to require
exact length for some types") requires attributes using types NLA_U* and
NLA_S* to have an exact length. This change is exposing bugs in various
userspace commands that are sending attributes with an invalid length
(e.g., attribute has type NLA_U8 and userspace sends NLA_U32). While
the commands are clearly broken and need to be fixed, users are arguing
that the sudden change in enforcement is breaking older commands on
newer kernels for use cases that otherwise "worked".
Relax the validation to print a warning mesage similar to what is done
for messages containing extra bytes after parsing.
Fixes: 28033ae4e0f5 ("net: netlink: Update attr validation to require exact length for some types") Signed-off-by: David Ahern <dsahern@gmail.com> Reviewed-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: David S. Miller <davem@davemloft.net>
commit 8d79266bc48c ("ip6_tunnel: add collect_md mode to IPv6 tunnels")
introduced new exit point in ipxip6_rcv. however rcu_read_unlock is
missing there. this diff is fixing this
v1->v2:
instead of doing rcu_read_unlock in place, we are going to "drop"
section (to prevent skb leakage)
Fixes: 8d79266bc48c ("ip6_tunnel: add collect_md mode to IPv6 tunnels") Signed-off-by: Nikita V. Shirokov <tehnerd@fb.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 7 Dec 2017 18:53:05 +0000 (13:53 -0500)]
Merge branch 'mv88e6xxx-error-patch-fixes'
Andrew Lunn says:
====================
mv88e6xxx error patch fixes
While trying to bring up a new PHY on a board, i exercised the error
paths a bit, and discovered some bugs. The unwind for interrupt
handling deadlocks, and the MDIO code hits a BUG() when a registered
MDIO device is freed without first being unregistered.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Thu, 7 Dec 2017 00:05:57 +0000 (01:05 +0100)]
net: dsa: mv88e6xxx: Unregister MDIO bus on error path
The MDIO busses need to be unregistered before they are freed,
otherwise BUG() is called. Add a call to the unregister code if the
registration fails, since we can have multiple busses, of which some
may correctly register before one fails. This requires moving the code
around a little.
Fixes: a3c53be55c95 ("net: dsa: mv88e6xxx: Support multiple MDIO busses") Signed-off-by: Andrew Lunn <andrew@lunn.ch> Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Thu, 7 Dec 2017 00:05:56 +0000 (01:05 +0100)]
net: dsa: mv88e6xxx: Fix interrupt masking on removal
When removing the interrupt handling code, we should mask the
generation of interrupts. The code however unmasked all
interrupts. This can then cause a new interrupt. We then get into a
deadlock where the interrupt thread is waiting to run, and the code
continues, trying to remove the interrupt handler, which means waiting
for the thread to complete. On a UP machine this deadlocks.
Fix so we really mask interrupts in the hardware. The same error is
made in the error path when install the interrupt handling code.
Fixes: 3460a5770ce9 ("net: dsa: mv88e6xxx: Mask g1 interrupts and free interrupt") Signed-off-by: Andrew Lunn <andrew@lunn.ch> Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Tobias Jordan [Wed, 6 Dec 2017 14:23:23 +0000 (15:23 +0100)]
net: mvmdio: disable/unprepare clocks in EPROBE_DEFER case
add appropriate calls to clk_disable_unprepare() by jumping to out_mdio
in case orion_mdio_probe() returns -EPROBE_DEFER.
Found by Linux Driver Verification project (linuxtesting.org).
Fixes: 3d604da1e954 ("net: mvmdio: get and enable optional clock") Signed-off-by: Tobias Jordan <Tobias.Jordan@elektrobit.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Pravin Shedge [Wed, 6 Dec 2017 16:58:40 +0000 (22:28 +0530)]
drivers: net: dsa: remove duplicate includes
These duplicate includes have been found with scripts/checkincludes.pl but
they have been removed manually to avoid removing false positives.
Signed-off-by: Pravin Shedge <pravin.shedge4linux@gmail.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Chris Dion [Wed, 6 Dec 2017 15:50:28 +0000 (10:50 -0500)]
net_sched: use macvlan real dev trans_start in dev_trans_start()
Macvlan devices are similar to vlans and do not update their
own trans_start. In order for arp monitoring to work for a bond device
when the slaves are macvlans, obtain its real device.
Signed-off-by: Chris Dion <christopher.dion@dell.com> Signed-off-by: David S. Miller <davem@davemloft.net>
net: thunderx: Fix TCP/UDP checksum offload for IPv4 pkts
Offload IP header checksum to NIC.
This fixes a previous patch which disabled checksum offloading
for both IPv4 and IPv6 packets. So L3 checksum offload was
getting disabled for IPv4 pkts. And HW is dropping these pkts
for some reason.
Without this patch, IPv4 TSO appears to be broken:
WIthout this patch I get ~16kbyte/s, with patch close to 2mbyte/s
when copying files via scp from test box to my home workstation.
Looking at tcpdump on sender it looks like hardware drops IPv4 TSO skbs.
This patch restores performance for me, ipv6 looks good too.
Fixes: fa6d7cb5d76c ("net: thunderx: Fix TCP/UDP checksum offload for IPv6 pkts") Cc: Sunil Goutham <sgoutham@cavium.com> Cc: Aleksey Makarov <aleksey.makarov@auriga.com> Cc: Eric Dumazet <edumazet@google.com> Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: David S. Miller <davem@davemloft.net>
Al Viro [Tue, 5 Dec 2017 23:29:09 +0000 (23:29 +0000)]
make sock_alloc_file() do sock_release() on failures
This changes calling conventions (and simplifies the hell out
the callers). New rules: once struct socket had been passed
to sock_alloc_file(), it's been consumed either by struct file
or by sock_release() done by sock_alloc_file(). Either way
the caller should not do sock_release() after that point.
Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Al Viro [Tue, 5 Dec 2017 23:27:57 +0000 (23:27 +0000)]
fix kcm_clone()
1) it's fput() or sock_release(), not both
2) don't do fd_install() until the last failure exit.
3) not a bug per se, but... don't attach socket to struct file
until it's set up.
Take reserving descriptor into the caller, move fd_install() to the
caller, sanitize failure exits and calling conventions.
Cc: stable@vger.kernel.org # v4.6+ Acked-by: Tom Herbert <tom@herbertland.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Mohamed Ghannam [Tue, 5 Dec 2017 20:58:35 +0000 (20:58 +0000)]
dccp: CVE-2017-8824: use-after-free in DCCP code
Whenever the sock object is in DCCP_CLOSED state,
dccp_disconnect() must free dccps_hc_tx_ccid and
dccps_hc_rx_ccid and set to NULL.
Signed-off-by: Mohamed Ghannam <simo.ghannam@gmail.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 5 Dec 2017 20:45:56 +0000 (12:45 -0800)]
net: remove hlist_nulls_add_tail_rcu()
Alexander Potapenko reported use of uninitialized memory [1]
This happens when inserting a request socket into TCP ehash,
in __sk_nulls_add_node_rcu(), since sk_reuseport is not initialized.
Bug was added by commit d894ba18d4e4 ("soreuseport: fix ordering for
mixed v4/v6 sockets")
Note that d296ba60d8e2 ("soreuseport: Resolve merge conflict for v4/v6
ordering fix") missed the opportunity to get rid of
hlist_nulls_add_tail_rcu() :
Both UDP sockets and TCP/DCCP listeners no longer use
__sk_nulls_add_node_rcu() for their hash insertion.
Since all other sockets have unique 4-tuple, the reuseport status
has no special meaning, so we can always use hlist_nulls_add_head_rcu()
for them and save few cycles/instructions.
====================
net: qualcomm: rmnet: Fix leaks in failure scenarios
Patch 1 fixes a leak in transmit path where a skb cannot be
transmitted due to insufficient headroom to stamp the map header.
Patch 2 fixes a leak in rmnet_newlink() failure because the
rmnet endpoint was never freed
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
net: qualcomm: rmnet: Fix leak in device creation failure
If the rmnet device creation fails in the newlink either while
registering with the physical device or after subsequent
operations, the rmnet endpoint information is never freed.
Fixes: ceed73a2cf4a ("drivers: net: ethernet: qualcomm: rmnet: Initial implementation") Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org> Signed-off-by: David S. Miller <davem@davemloft.net>
This patch replaces the old SS_DISCONNECTING with the new TCP_CLOSING
constant.
CC: Dexuan Cui <decui@microsoft.com> CC: Cathy Avery <cavery@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Jorgen Hansen <jhansen@vmware.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Maloy [Mon, 4 Dec 2017 21:00:20 +0000 (22:00 +0100)]
tipc: fix memory leak in tipc_accept_from_sock()
When the function tipc_accept_from_sock() fails to create an instance of
struct tipc_subscriber it omits to free the already created instance of
struct tipc_conn instance before it returns.
We fix that with this commit.
Reported-by: David S. Miller <davem@davemloft.net> Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes: 14c04493cb77 ("tipc: add ability to order and receive topology events in driver") Reported-by: syzbot <syzkaller@googlegroups.com> Cc: Jon Maloy <jon.maloy@ericsson.com> Cc: Ying Xue <ying.xue@windriver.com> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 5 Dec 2017 19:40:35 +0000 (14:40 -0500)]
Merge branch 'sh_eth-dma-mapping-fixes'
Thomas Petazzoni says:
====================
net: sh_eth: DMA mapping API fixes
Here are two patches that fix how the sh_eth driver is using the DMA
mapping API: a bogus struct device is used in some places, or a NULL
struct device is used.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
net: sh_eth: don't use NULL as "struct device" for the DMA mapping API
Using NULL as argument for the DMA mapping API is bogus, as the DMA
mapping API may use information from the "struct device" to perform
the DMA mapping operation. Therefore, pass the appropriate "struct
device".
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Acked-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: David S. Miller <davem@davemloft.net>
net: sh_eth: use correct "struct device" when calling DMA mapping functions
There are two types of "struct device": the one representing the
physical device on its physical bus (platform, SPI, PCI, etc.), and
the one representing the logical device in its device class (net,
etc.).
The DMA mapping API expects to receive as argument a "struct device"
representing the physical device, as the "struct device" contains
information about the bus that the DMA API needs.
However, the sh_eth driver mistakenly uses the "struct device"
representing the logical device (embedded in "struct net_device")
rather than the "struct device" representing the physical device on
its bus.
This commit fixes that by adjusting all calls to the DMA mapping API.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Acked-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: David S. Miller <davem@davemloft.net>
Nogah Frankel [Mon, 4 Dec 2017 11:31:11 +0000 (13:31 +0200)]
net_sched: red: Avoid illegal values
Check the qmin & qmax values doesn't overflow for the given Wlog value.
Check that qmin <= qmax.
Fixes: a783474591f2 ("[PKT_SCHED]: Generic RED layer") Signed-off-by: Nogah Frankel <nogahf@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Nogah Frankel [Mon, 4 Dec 2017 11:31:10 +0000 (13:31 +0200)]
net_sched: red: Avoid devision by zero
Do not allow delta value to be zero since it is used as a divisor.
Fixes: 8af2a218de38 ("sch_red: Adaptative RED AQM") Signed-off-by: Nogah Frankel <nogahf@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Zumeng Chen [Mon, 4 Dec 2017 03:22:02 +0000 (11:22 +0800)]
gianfar: fix a flooded alignment reports because of padding issue.
According to LS1021A RM, the value of PAL can be set so that the start of the
IP header in the receive data buffer is aligned to a 32-bit boundary. Normally,
setting PAL = 2 provides minimal padding to ensure such alignment of the IP
header.
However every incoming packet's 8-byte time stamp will be inserted into the
packet data buffer as padding alignment bytes when hardware time stamping is
enabled.
So we set the padding 8+2 here to avoid the flooded alignment faults:
Johannes Berg [Sat, 2 Dec 2017 07:41:55 +0000 (08:41 +0100)]
Revert "net: core: maybe return -EEXIST in __dev_alloc_name"
This reverts commit d6f295e9def0; some userspace (in the case
we noticed it's wpa_supplicant), is relying on the current
error code to determine that a fixed name interface already
exists.
Reported-by: Jouni Malinen <j@w1.fi> Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Previously we swapped the tx_packets, tx_bytes and tx_dropped counters
with rx_packets, rx_bytes and rx_dropped counters, respectively. This
behaviour is correct and expected for VF representors but it should not
be swapped for physical port mac representors.
Fixes: eadfa4c3be99 ("nfp: add stats and xmit helpers for representors") Signed-off-by: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com> Reviewed-by: Simon Horman <simon.horman@netronome.com> Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Tue, 5 Dec 2017 14:02:41 +0000 (15:02 +0100)]
Merge branch 'bpf-fix-broken-uapi-for-pt-regs'
Hendrik Brueckner says:
====================
Perf tool bpf selftests revealed a broken uapi for s390 and arm64.
With the BPF_PROG_TYPE_PERF_EVENT program type the bpf_perf_event
structure exports the pt_regs structure for all architectures.
This fails for s390 and arm64 because pt_regs are not part of the
user api and kept in-kernel only. To mitigate the broken uapi,
introduce a wrapper that exports pt_regs in an asm-generic way.
For arm64, export the exising user_pt_regs structure. For s390,
introduce a user_pt_regs structure that exports the beginning of
pt_regs.
Note that user_pt_regs must export from the beginning of pt_regs
as BPF_PROG_TYPE_PERF_EVENT program type is not the only type for
running BPF programs.
Some more background:
For the bpf_perf_event, there is a uapi definition that is
passed to the BPF program. For other "probe" points like
trace points, kprobes, and uprobes, there is no uapi and the
BPF program is always passed pt_regs (which is OK as the BPF
program runs in the kernel context). The perf tool can attach
BPF programs to all of these "probe" points and, optionally,
can create a BPF prologue to access particular arguments
(passed as registers). For this, it uses DWARF/CFI
information to obtain the register and calls a perf-arch
backend function, regs_query_register_offset(). This function
returns the index into (user_)pt_regs for a particular
register. Then, perf creates a BPF prologue that accesses
this register based on the passed stucture from the "probe"
point.
Part of this series, are also updates to the testing and bpf selftest
to deal with asm-specifics. To complete the bpf support in perf, the
the regs_query_register_offset function is added for s390 to support
BPF prologue creation.
Changelog v1 -> v2:
- Correct kbuild test bot issues by including
asm-generic/bpf_perf_event.h for archictectures that do not have
their own asm version.
- Added patch to clean-up whitespace and coding style issues in s390
asm/ptrace.h (#4/6) as suggested by Alexei.
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
The regs_query_register_offset() helper function converts
register name like "%r0" to an offset of a register in user_pt_regs
It is required by the BPF prologue generator.
The user_pt_regs structure was recently added to "asm/ptrace.h".
Hence, update tools/perf/check-headers.sh to keep the header file
in sync with kernel changes.
Suggested-by: Thomas Richter <tmricht@linux.vnet.ibm.com> Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-and-tested-by: Thomas Richter <tmricht@linux.vnet.ibm.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
selftests/bpf: sync kernel headers and introduce arch support in Makefile
Synchronize the uapi kernel header files which solves the broken
uapi export of pt_regs. Because of arch-specific uapi headers,
extended the include path in the Makefile.
With this change, the test_verifier program compiles and runs successfully
on s390.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-and-tested-by: Thomas Richter <tmricht@linux.vnet.ibm.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Cc: Daniel Borkmann <daniel@iogearbox.net> Cc: Shuah Khan <shuah@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
s390/uapi: correct whitespace & coding style in asm/ptrace.h
Correct whitespace and coding style issues in the s390 asm/ptrace.h
uapi header file. This is preparatory work to copy it to the tools/
directory for inclusion by selftests and perf.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
arm64/bpf: correct broken uapi for BPF_PROG_TYPE_PERF_EVENT program type
Correct the broken uapi for the BPF_PROG_TYPE_PERF_EVENT program type
by exporting the user_pt_regs structure instead of the pt_regs structure
that is in-kernel only.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-by: Thomas Richter <tmricht@linux.vnet.ibm.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Cc: Will Deacon <will.deacon@arm.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
s390/bpf: correct broken uapi for BPF_PROG_TYPE_PERF_EVENT program type
To mitigate and correct the broken uapi for the BPF_PROG_TYPE_PERF_EVENT
program type, introduce a user_pt_regs structure (similar to arm64) that
exports parts from the beginnig of the pt_regs structure.
The export must start with the beginning of the pt_regs structure because
to correctly calculate BPF prologues for perf (regs_query_register_offset()).
For BPF_PROG_TYPE_PERF_EVENT program types, the BPF program is then passed
a user_pt_regs structure.
Note: Depending on future changes to the s390 pt_regs structure, consider
the user_pt_regs structure to be stable for a particular kernel version
only. (Of course, s390 tries to ensure keep it stable as much as possible.)
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-and-tested-by: Thomas Richter <tmricht@linux.vnet.ibm.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
bpf: correct broken uapi for BPF_PROG_TYPE_PERF_EVENT program type
Commit 0515e5999a466dfe ("bpf: introduce BPF_PROG_TYPE_PERF_EVENT
program type") introduced the bpf_perf_event_data structure which
exports the pt_regs structure. This is OK for multiple architectures
but fail for s390 and arm64 which do not export pt_regs. Programs
using them, for example, the bpf selftest fail to compile on these
architectures.
For s390, exporting the pt_regs is not an option because s390 wants
to allow changes to it. For arm64, there is a user_pt_regs structure
that covers parts of the pt_regs structure for use by user space.
To solve the broken uapi for s390 and arm64, introduce an abstract
type for pt_regs and add an asm/bpf_perf_event.h file that concretes
the type. An asm-generic header file covers the architectures that
export pt_regs today.
The arch-specific enablement for s390 and arm64 follows in separate
commits.
Reported-by: Thomas Richter <tmricht@linux.vnet.ibm.com> Fixes: 0515e5999a466dfe ("bpf: introduce BPF_PROG_TYPE_PERF_EVENT program type") Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Reviewed-and-tested-by: Thomas Richter <tmricht@linux.vnet.ibm.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Linus Torvalds [Mon, 4 Dec 2017 19:32:02 +0000 (11:32 -0800)]
Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Pull virtio fixes from Michael Tsirkin:
"virtio and qemu bugfixes
A couple of bugfixes that just became ready"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
virtio_balloon: fix increment of vb->num_pfns in fill_balloon()
virtio: release virtio index when fail to device_register
fw_cfg: fix driver remove
1) Various TCP control block fixes, including one that crashes with
SELinux, from David Ahern and Eric Dumazet.
2) Fix ACK generation in rxrpc, from David Howells.
3) ipvlan doesn't set the mark properly in the ipv4 route lookup key,
from Gao Feng.
4) SIT configuration doesn't take on the frag_off ipv4 field
configuration properly, fix from Hangbin Liu.
5) TSO can fail after device down/up on stmmac, fix from Lars Persson.
6) Various bpftool fixes (mostly in JSON handling) from Quentin Monnet.
7) Various SKB leak fixes in vhost/tun/tap (mostly observed as
performance problems). From Wei Xu.
8) mvpps's TX descriptors were not zero initialized, from Yan Markman.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (57 commits)
tcp: use IPCB instead of TCP_SKB_CB in inet_exact_dif_match()
tcp: add tcp_v4_fill_cb()/tcp_v4_restore_cb()
rxrpc: Fix the MAINTAINERS record
rxrpc: Use correct netns source in rxrpc_release_sock()
liquidio: fix incorrect indentation of assignment statement
stmmac: reset last TSO segment size after device open
ipvlan: Add the skb->mark as flow4's member to lookup route
s390/qeth: build max size GSO skbs on L2 devices
s390/qeth: fix GSO throughput regression
s390/qeth: fix thinko in IPv4 multicast address tracking
tap: free skb if flags error
tun: free skb in early errors
vhost: fix skb leak in handle_rx()
bnxt_en: Fix a variable scoping in bnxt_hwrm_do_send_msg()
bnxt_en: fix dst/src fid for vxlan encap/decap actions
bnxt_en: wildcard smac while creating tunnel decap filter
bnxt_en: Need to unconditionally shut down RoCE in bnxt_shutdown
phylink: ensure we take the link down when phylink_stop() is called
sfp: warn about modules requiring address change sequence
sfp: improve RX_LOS handling
...
Chris Metcalf [Mon, 4 Dec 2017 15:52:40 +0000 (10:52 -0500)]
arch/tile: mark as orphaned
The chip family of TILEPro and TILE-Gx was developed by Tilera, which
was eventually acquired by Mellanox. The tile architecture was added to
the kernel in 2010 and first appeared in 2.6.36.
Now at Mellanox we are developing new chips based on the ARM64
architecture; our last TILE-Gx chip (the Gx72) was released in 2013, and
our customers using tile architecture products are not, as far as we
know, looking to upgrade to newer kernel releases. In the absence of
someone in the community stepping up to take over maintainership, this
commit marks the architecture as orphaned.
Cc: Chris Metcalf <metcalf@alum.mit.edu> Signed-off-by: Chris Metcalf <cmetcalf@mellanox.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The following pull-request contains BPF updates for your *net* tree.
The main changes are:
1) Fix a compilation warning in xdp redirect tracepoint due to
missing bpf.h include that pulls in struct bpf_map, from Xie.
2) Limit the maximum number of attachable BPF progs for a given
perf event as long as uabi is not frozen yet. The hard upper
limit is now 64 and therefore the same as with BPF multi-prog
for cgroups. Also add related error checking for the sample
BPF loader when enabling and attaching to the perf event, from
Yonghong.
3) Specifically set the RLIMIT_MEMLOCK for the test_verifier_log
case, so that the test case can always pass and not fail in
some environments due to too low default limit, also from
Yonghong.
4) Fix up a missing license header comment for kernel/bpf/offload.c,
from Jakub.
5) Several fixes for bpftool, among others a crash on incorrect
arguments when json output is used, error message handling
fixes on unknown options and proper destruction of json writer
for some exit cases, all from Quentin.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
James Morris reported kernel stack corruption bug that
we tracked back to commit 971f10eca186 ("tcp: better TCP_SKB_CB
layout to reduce cache line misses")
First patch needs to be backported to kernels >= 3.18,
while second patch needs to be backported to kernels >= 4.9, since
this was the time when inet_exact_dif_match appeared.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
David Ahern [Sun, 3 Dec 2017 17:33:00 +0000 (09:33 -0800)]
tcp: use IPCB instead of TCP_SKB_CB in inet_exact_dif_match()
After this fix : ("tcp: add tcp_v4_fill_cb()/tcp_v4_restore_cb()"),
socket lookups happen while skb->cb[] has not been mangled yet by TCP.
Fixes: a04a480d4392 ("net: Require exact match for TCP socket lookups if dif is l3mdev") Signed-off-by: David Ahern <dsahern@gmail.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Sun, 3 Dec 2017 17:32:59 +0000 (09:32 -0800)]
tcp: add tcp_v4_fill_cb()/tcp_v4_restore_cb()
James Morris reported kernel stack corruption bug [1] while
running the SELinux testsuite, and bisected to a recent
commit bffa72cf7f9d ("net: sk_buff rbnode reorg")
We believe this commit is fine, but exposes an older bug.
SELinux code runs from tcp_filter() and might send an ICMP,
expecting IP options to be found in skb->cb[] using regular IPCB placement.
We need to defer TCP mangling of skb->cb[] after tcp_filter() calls.
This patch adds tcp_v4_fill_cb()/tcp_v4_restore_cb() in a very
similar way we added them for IPv6.
Fixes: 971f10eca186 ("tcp: better TCP_SKB_CB layout to reduce cache line misses") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: James Morris <james.l.morris@oracle.com> Tested-by: James Morris <james.l.morris@oracle.com> Tested-by: Casey Schaufler <casey@schaufler-ca.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Sun, 3 Dec 2017 15:51:08 +0000 (10:51 -0500)]
Merge branch 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm
Pull ARM fix from Russell King:
"Just one fix this time around, for the late commit in the merge window
that triggered a problem with qemu. Qemu is apparently also going to
receive a fix for the discovered issue"
* 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm:
ARM: avoid faulting on qemu
Linus Torvalds [Sun, 3 Dec 2017 15:46:16 +0000 (10:46 -0500)]
Merge tag 'hwmon-for-linus-v4.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging
Pull hwmon fixes from Guenter Roeck:
"Fixes:
- Drop reference to obsolete maintainer tree
- Fix overflow bug in pmbus driver
- Fix SMBUS timeout problem in jc42 driver
For the SMBUS timeout handling, we had a brief discussion if this
should be considered a bug fix or a feature. Peter says "it fixes real
problems where the application misbehave due to faulty content when
reading from an eeprom", and he needs the patch in his company's v4.14
images. This is good enough for me and warrants backport to stable
kernels"
* tag 'hwmon-for-linus-v4.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
hwmon: (jc42) optionally try to disable the SMBUS timeout
hwmon: (pmbus) Use 64bit math for DIRECT format values
hwmon: Drop reference to Jean's tree
David Howells [Fri, 1 Dec 2017 11:10:37 +0000 (11:10 +0000)]
rxrpc: Fix the MAINTAINERS record
Fix the MAINTAINERS record so that it's more obvious who the maintainer for
AF_RXRPC is.
Reported-by: Joe Perches <joe@perches.com> Reported-by: David Miller <davem@davemloft.net> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David Howells [Fri, 1 Dec 2017 11:09:53 +0000 (11:09 +0000)]
rxrpc: Use correct netns source in rxrpc_release_sock()
In rxrpc_release_sock() there may be no rx->local value to access, so we
can't unconditionally follow it to the rxrpc network namespace information
to poke the connection reapers.
Instead, use the socket's namespace pointer to find the namespace.
This unfixed code causes the following static checker warning:
net/rxrpc/af_rxrpc.c:898 rxrpc_release_sock()
error: we previously assumed 'rx->local' could be null (see line 887)
Fixes: 3d18cbb7fd0c ("rxrpc: Fix conn expiry timers") Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 3 Dec 2017 14:57:54 +0000 (09:57 -0500)]
Merge tag 'linux-can-fixes-for-4.15-20171201' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can
Marc Kleine-Budde says:
====================
pull-request: can 2017-12-01
this is a pull for net consisting of nine patches.
The first three patches are by Jimmy Assarsson for the kvaser_usb driver
and add the missing free()s in some error path, a signed/unsigned
comparison and ratelimit the error messages in case of incomplete
messages. Oliver Stäbler's patch for the ti_hecc driver fix the napi
poll function's return value. The return values of the probe function of
the peak_canfd and peak_pci PCI drivers are fixed by Stephane Grosjean's
patch. Two patches by me for the flexcan driver update the
bugs/features/quirks overview table and fix the error state transition
for the VF610 SoC. The two patches by Martin Kelly for the mcba_usb
driver fix a typo and a device disconnect bug.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Wolfram Sang [Sat, 2 Dec 2017 22:32:13 +0000 (23:32 +0100)]
Merge tag 'at24-4.15-fixes-for-wolfram' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux into i2c/for-current
Please consider pulling the following fixes for v4.15. While it doesn't
fix any regression introduced in the v4.15 merge window, we have a
feature in at24 since linux v4.8 - reading the mac address block from
at24mac series - which turned out to be not working.
This pull request contains changes that fix it together with a patch
that hardens the read and write argument sanitization with
out-of-bounds checks that were missing.
Lars Persson [Fri, 1 Dec 2017 10:12:44 +0000 (11:12 +0100)]
stmmac: reset last TSO segment size after device open
The mss variable tracks the last max segment size sent to the TSO
engine. We do not update the hardware as long as we receive skb:s with
the same value in gso_size.
During a network device down/up cycle (mapped to stmmac_release() and
stmmac_open() callbacks) we issue a reset to the hardware and it
forgets the setting for mss. However we did not zero out our mss
variable so the next transmission of a gso packet happens with an
undefined hardware setting.
This triggers a hang in the TSO engine and eventuelly the netdev
watchdog will bark.
Fixes: f748be531d70 ("stmmac: support new GMAC4") Signed-off-by: Lars Persson <larper@axis.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Julian Wiedmann [Fri, 1 Dec 2017 09:14:51 +0000 (10:14 +0100)]
s390/qeth: build max size GSO skbs on L2 devices
The current GSO skb size limit was copy&pasted over from the L3 path,
where it is needed due to a TSO limitation.
As L2 devices don't offer TSO support (and thus all GSO skbs are
segmented before they reach the driver), there's no reason to restrict
the stack in how large it may build the GSO skbs.
Fixes: d52aec97e5bc ("qeth: enable scatter/gather in layer 2 mode") Signed-off-by: Julian Wiedmann <jwi@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Julian Wiedmann [Fri, 1 Dec 2017 09:14:50 +0000 (10:14 +0100)]
s390/qeth: fix GSO throughput regression
Using GSO with small MTUs currently results in a substantial throughput
regression - which is caused by how qeth needs to map non-linear skbs
into its IO buffer elements:
compared to a linear skb, each GSO-segmented skb effectively consumes
twice as many buffer elements (ie two instead of one) due to the
additional header-only part. This causes the Output Queue to be
congested with low-utilized IO buffers.
Fix this as follows:
If the MSS is low enough so that a non-SG GSO segmentation produces
order-0 skbs (currently ~3500 byte), opt out from NETIF_F_SG. This is
where we anticipate the biggest savings, since an SG-enabled
GSO segmentation produces skbs that always consume at least two
buffer elements.
Larger MSS values continue to get a SG-enabled GSO segmentation, since
1) the relative overhead of the additional header-only buffer element
becomes less noticeable, and
2) the linearization overhead increases.
With the throughput regression fixed, re-enable NETIF_F_SG by default to
reap the significant CPU savings of GSO.
Fixes: 5722963a8e83 ("qeth: do not turn on SG per default") Reported-by: Nils Hoppmann <niho@de.ibm.com> Signed-off-by: Julian Wiedmann <jwi@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Julian Wiedmann [Fri, 1 Dec 2017 09:14:49 +0000 (10:14 +0100)]
s390/qeth: fix thinko in IPv4 multicast address tracking
Commit 5f78e29ceebf ("qeth: optimize IP handling in rx_mode callback")
reworked how secondary addresses are managed for qeth devices.
Instead of dropping & subsequently re-adding all addresses on every
ndo_set_rx_mode() call, qeth now keeps track of the addresses that are
currently registered with the HW.
On a ndo_set_rx_mode(), we thus only need to do (de-)registration
requests for the addresses that have actually changed.
On L3 devices, the lookup for IPv4 Multicast addresses checks the wrong
hashtable - and thus never finds a match. As a result, we first delete
*all* such addresses, and then re-add them again. So each set_rx_mode()
causes a short period where the IPv4 Multicast addresses are not
registered, and the card stops forwarding inbound traffic for them.
Fix this by setting the ->is_multicast flag on the lookup object, thus
enabling qeth_l3_ip_from_hash() to search the correct hashtable and
find a match there.
Fixes: 5f78e29ceebf ("qeth: optimize IP handling in rx_mode callback") Signed-off-by: Julian Wiedmann <jwi@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 3 Dec 2017 02:31:04 +0000 (21:31 -0500)]
Merge branch 'vhost-skb-leaks'
Wei Xu says:
====================
vhost: fix a few skb leaks
Matthew found a roughly 40% tcp throughput regression with commit c67df11f(vhost_net: try batch dequing from skb array) as discussed
in the following thread:
https://www.mail-archive.com/netdev@vger.kernel.org/msg187936.html
v4:
- fix zero iov iterator count in tap/tap_do_read()(Jason)
- don't put tun in case of EBADFD(Jason)
- Replace msg->msg_control with new 'skb' when calling tun/tap_do_read()
v3:
- move freeing skb from vhost to tun/tap recvmsg() to not
confuse the callers.
v2:
- add Matthew as the reporter, thanks matthew.
- moving zero headcount check ahead instead of defer consuming skb
due to jason and mst's comment.
- add freeing skb in favor of recvmsg() fails.
====================
Acked-by: Michael S. Tsirkin <mst@redhat.com> Tested-by: Matthew Rosato <mjrosato@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Wei Xu [Fri, 1 Dec 2017 10:10:38 +0000 (05:10 -0500)]
tap: free skb if flags error
tap_recvmsg() supports accepting skb by msg_control after
commit 3b4ba04acca8 ("tap: support receiving skb from msg_control"),
the skb if presented should be freed within the function, otherwise
it would be leaked.
Signed-off-by: Wei Xu <wexu@redhat.com> Reported-by: Matthew Rosato <mjrosato@linux.vnet.ibm.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Wei Xu [Fri, 1 Dec 2017 10:10:37 +0000 (05:10 -0500)]
tun: free skb in early errors
tun_recvmsg() supports accepting skb by msg_control after
commit ac77cfd4258f ("tun: support receiving skb through msg_control"),
the skb if presented should be freed no matter how far it can go
along, otherwise it would be leaked.
This patch fixes several missed cases.
Signed-off-by: Wei Xu <wexu@redhat.com> Reported-by: Matthew Rosato <mjrosato@linux.vnet.ibm.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Wei Xu [Fri, 1 Dec 2017 10:10:36 +0000 (05:10 -0500)]
vhost: fix skb leak in handle_rx()
Matthew found a roughly 40% tcp throughput regression with commit c67df11f(vhost_net: try batch dequing from skb array) as discussed
in the following thread:
https://www.mail-archive.com/netdev@vger.kernel.org/msg187936.html
Eventually we figured out that it was a skb leak in handle_rx()
when sending packets to the VM. This usually happens when a guest
can not drain out vq as fast as vhost fills in, afterwards it sets
off the traffic jam and leaks skb(s) which occurs as no headcount
to send on the vq from vhost side.
This can be avoided by making sure we have got enough headcount
before actually consuming a skb from the batched rx array while
transmitting, which is simply done by moving checking the zero
headcount a bit ahead.
Signed-off-by: Wei Xu <wexu@redhat.com> Reported-by: Matthew Rosato <mjrosato@linux.vnet.ibm.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 3 Dec 2017 02:25:39 +0000 (21:25 -0500)]
Merge branch 'bnxt_en-fixes'
Michael Chan says:
====================
bnxt_en: Fixes.
A shutdown fix for SMARTNIC, 2 fixes related to TC Flower vxlan
filters, and the last one fixes an out-of-scope variable when sending
short firmware messages.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
bnxt_en: Fix a variable scoping in bnxt_hwrm_do_send_msg()
short_input variable is assigned to another data pointer which is
referred out of its scope. Fix it by moving short_input definition
to the beginning of bnxt_hwrm_do_send_msg() function.
No failure has been reported so far due to this issue.
Fixes: e605db801bde ("bnxt_en: Support for Short Firmware Message") Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Sathya Perla [Fri, 1 Dec 2017 08:13:04 +0000 (03:13 -0500)]
bnxt_en: fix dst/src fid for vxlan encap/decap actions
For flows that involve a vxlan encap action, the vxlan sock
interface may be specified as the outgoing interface. The driver
must resolve the outgoing PF interface used by this socket and
use the dst_fid of the PF in the hwrm_cfa_encap_record_alloc cmd.
Similarily for flows that have a vxlan decap action, the
fid of the incoming PF interface must be used as the src_fid in
the hwrm_cfa_decap_filter_alloc cmd.
Fixes: 8c95f773b4a3 ("bnxt_en: add support for Flower based vxlan encap/decap offload") Signed-off-by: Sathya Perla <sathya.perla@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Sunil Challa [Fri, 1 Dec 2017 08:13:03 +0000 (03:13 -0500)]
bnxt_en: wildcard smac while creating tunnel decap filter
While creating a decap filter the tunnel smac need not (and must not) be
specified as we cannot ascertain the neighbor in the recv path. 'ttl'
match is also not needed for the decap filter and must be wild-carded.
Fixes: f484f6782e01 ("bnxt_en: add hwrm FW cmds for cfa_encap_record and decap_filter") Signed-off-by: Sunil Challa <sunilkumar.challa@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ray Jui [Fri, 1 Dec 2017 08:13:02 +0000 (03:13 -0500)]
bnxt_en: Need to unconditionally shut down RoCE in bnxt_shutdown
The current 'bnxt_shutdown' implementation only invokes
'bnxt_ulp_shutdown' to shut down RoCE in the case when the system is in
the path of power off (SYSTEM_POWER_OFF). While this may work in most
cases, it does not work in the smart NIC case, when Linux 'reboot'
command is initiated from the Linux that runs on the ARM cores of the
NIC card. In this particular case, Linux 'reboot' results in a system
'L3' level reset where the entire ARM and associated subsystems are
being reset, but at the same time, Nitro core is being kept in sane state
(to allow external PCIe connected servers to continue to work). Without
properly shutting down RoCE and freeing all associated resources, it
results in the ARM core to hang immediately after the 'reboot'
By always invoking 'bnxt_ulp_shutdown' in 'bnxt_shutdown', it fixes the
above issue
Fixes: 0efd2fc65c92 ("bnxt_en: Add a callback to inform RDMA driver during PCI shutdown.") Signed-off-by: Ray Jui <ray.jui@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Sat, 2 Dec 2017 01:04:20 +0000 (20:04 -0500)]
Merge tag 'nfs-for-4.15-2' of git://git.linux-nfs.org/projects/anna/linux-nfs
Pull NFS client fixes from Anna Schumaker:
"These patches fix a problem with compiling using an old version of
gcc, and also fix up error handling in the SUNRPC layer.
- NFSv4: Ensure gcc 4.4.4 can compile initialiser for
"invalid_stateid"
- SUNRPC: Allow connect to return EHOSTUNREACH
- SUNRPC: Handle ENETDOWN errors"
* tag 'nfs-for-4.15-2' of git://git.linux-nfs.org/projects/anna/linux-nfs:
SUNRPC: Handle ENETDOWN errors
SUNRPC: Allow connect to return EHOSTUNREACH
NFSv4: Ensure gcc 4.4.4 can compile initialiser for "invalid_stateid"
Linus Torvalds [Sat, 2 Dec 2017 01:00:19 +0000 (20:00 -0500)]
Merge tag 'xfs-4.15-fixes-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Pull xfs fixes from Darrick Wong:
"Here are some bug fixes for 4.15-rc2.
- fix memory leaks that appeared after removing ifork inline data
buffer
- recover deferred rmap update log items in correct order
- fix memory leaks when buffer construction fails
- fix memory leaks when bmbt is corrupt
- fix some uninitialized variables and math problems in the quota
scrubber
- add some omitted attribution tags on the log replay commit
- fix some UBSAN complaints about integer overflows with large sparse
files
- implement an effective inode mode check in online fsck
- fix log's inability to retry quota item writeout due to transient
errors"
* tag 'xfs-4.15-fixes-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
xfs: Properly retry failed dquot items in case of error during buffer writeback
xfs: scrub inode mode properly
xfs: remove unused parameter from xfs_writepage_map
xfs: ubsan fixes
xfs: calculate correct offset in xfs_scrub_quota_item
xfs: fix uninitialized variable in xfs_scrub_quota
xfs: fix leaks on corruption errors in xfs_bmap.c
xfs: fortify xfs_alloc_buftarg error handling
xfs: log recovery should replay deferred ops in order
xfs: always free inline data before resetting inode fork during ifree
Linus Torvalds [Sat, 2 Dec 2017 00:39:12 +0000 (19:39 -0500)]
Merge tag 'riscv-for-linus-4.15-rc2_cleanups' of git://git.kernel.org/pub/scm/linux/kernel/git/palmer/linux
Pull RISC-V cleanups and ABI fixes from Palmer Dabbelt:
"This contains a handful of small cleanups that are a result of
feedback that didn't make it into our original patch set, either
because the feedback hadn't been given yet, I missed the original
emails, or we weren't ready to submit the changes yet.
I've been maintaining the various cleanup patch sets I have as their
own branches, which I then merged together and signed. Each merge
commit has a short summary of the changes, and each branch is based on
your latest tag (4.15-rc1, in this case). If this isn't the right way
to do this then feel free to suggest something else, but it seems sane
to me.
Here's a short summary of the changes, roughly in order of how
interesting they are.
- libgcc.h has been moved from include/lib, where it's the only
member, to include/linux. This is meant to avoid tab completion
conflicts.
- VDSO entries for clock_get/gettimeofday/getcpu have been added.
These are simple syscalls now, but we want to let glibc use them
from the start so we can make them faster later.
- A VDSO entry for instruction cache flushing has been added so
userspace can flush the instruction cache.
- The VDSO symbol versions for __vdso_cmpxchg{32,64} have been
removed, as those VDSO entries don't actually exist.
- __io_writes has been corrected to respect the given type.
- A new READ_ONCE in arch_spin_is_locked().
- __test_and_op_bit_ord() is now actually ordered.
- Various small fixes throughout the tree to enable allmodconfig to
build cleanly.
- Removal of some dead code in our atomic support headers.
- Improvements to various comments in our atomic support headers"
* tag 'riscv-for-linus-4.15-rc2_cleanups' of git://git.kernel.org/pub/scm/linux/kernel/git/palmer/linux: (23 commits)
RISC-V: __io_writes should respect the length argument
move libgcc.h to include/linux
RISC-V: Clean up an unused include
RISC-V: Allow userspace to flush the instruction cache
RISC-V: Flush I$ when making a dirty page executable
RISC-V: Add missing include
RISC-V: Use define for get_cycles like other architectures
RISC-V: Provide stub of setup_profiling_timer()
RISC-V: Export some expected symbols for modules
RISC-V: move empty_zero_page definition to C and export it
RISC-V: io.h: type fixes for warnings
RISC-V: use RISCV_{INT,SHORT} instead of {INT,SHORT} for asm macros
RISC-V: use generic serial.h
RISC-V: remove spin_unlock_wait()
RISC-V: `sfence.vma` orderes the instruction cache
RISC-V: Add READ_ONCE in arch_spin_is_locked()
RISC-V: __test_and_op_bit_ord should be strongly ordered
RISC-V: Remove smb_mb__{before,after}_spinlock()
RISC-V: Remove __smp_bp__{before,after}_atomic
RISC-V: Comment on why {,cmp}xchg is ordered how it is
...
Linus Torvalds [Sat, 2 Dec 2017 00:37:03 +0000 (19:37 -0500)]
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 fixes from Will Deacon:
"The critical one here is a fix for fpsimd register corruption across
signals which was introduced by the SVE support code (the register
files overlap), but the others are worth having as well.
Summary:
- Fix FP register corruption when SVE is not available or in use
- Fix out-of-tree module build failure when CONFIG_ARM64_MODULE_PLTS=y
- Missing 'const' generating errors with LTO builds
- Remove unsupported events from Cortex-A73 PMU description
- Removal of stale and incorrect comments"
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
arm64: context: Fix comments and remove pointless smp_wmb()
arm64: cpu_ops: Add missing 'const' qualifiers
arm64: perf: remove unsupported events for Cortex-A73
arm64: fpsimd: Fix failure to restore FPSIMD state after signals
arm64: pgd: Mark pgd_cache as __ro_after_init
arm64: ftrace: emit ftrace-mod.o contents through code
arm64: module-plts: factor out PLT generation code for ftrace
arm64: mm: cleanup stale AIVIVT references
Palmer Dabbelt [Fri, 1 Dec 2017 21:31:31 +0000 (13:31 -0800)]
RISC-V: Fixes for clean allmodconfig build
Olaf said: Here's a short series of patches that produces a working
allmodconfig. Would be nice to see them go in so we can add build
coverage.
I've dropped patches 8 and 10 from the original set:
* [PATCH 08/10] (RISC-V: Set __ARCH_WANT_RENAMEAT to pick up generic
version) has a better fix that I've sent out for review, we don't want
renameat.
* [PATCH 10/10] (input: joystick: riscv has get_cycles) has already been
taken into Dmitry Torokhov's tree.
Palmer Dabbelt [Fri, 1 Dec 2017 21:12:10 +0000 (13:12 -0800)]
RISC-V: User-Visible Changes
This merge contains the user-visible, ABI-breaking changes that we want
to make sure we have in Linux before our first release. Highlights
include:
* VDSO entries for clock_get/gettimeofday/getcpu have been added. These
are simple syscalls now, but we want to let glibc use them from the
start so we can make them faster later.
* A VDSO entry for instruction cache flushing has been added so
userspace can flush the instruction cache.
* The VDSO symbol versions for __vdso_cmpxchg{32,64} have been removed,
as those VDSO entries don't actually exist.
Palmer Dabbelt [Fri, 1 Dec 2017 21:10:42 +0000 (13:10 -0800)]
RISC-V Atomic Cleanups
This patch set is the result of some feedback that filtered through
after our original patch set was reviewed, some of which was the result
of me missing some email. It contains:
* A new READ_ONCE in arch_spin_is_locked()
* __test_and_op_bit_ord() is now actually ordered
* Improvements to various comments
* Removal of some dead code
David S. Miller [Fri, 1 Dec 2017 20:18:42 +0000 (15:18 -0500)]
Merge branch 'sfp-phylink-fixes'
Russell King says:
====================
SFP/phylink fixes
Here are four phylink fixes:
- the "options" is a big-endian value, we must test the bits taking the
endian-ness into account.
- improve the handling of RX_LOS polarity, taking no RX_LOS polarity
bits set to mean there is no RX_LOS functionality provided.
- do not report modules that require the address mode switching as
supporting SFF8472.
- ensure that the mac_link_down() function is called when phylink_stop()
is called.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Thu, 30 Nov 2017 13:59:26 +0000 (13:59 +0000)]
phylink: ensure we take the link down when phylink_stop() is called
Ensure that we tell the MAC to take the link down when phylink_stop()
is called, and that this completes prior to phylink_stop() returns.
Reported-by: Florian Fainelli <f.fainelli@gmail.com> Tested-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Thu, 30 Nov 2017 13:59:21 +0000 (13:59 +0000)]
sfp: warn about modules requiring address change sequence
We do not support SFP modules which require the address change sequence
as detailed by SFF 8472 revision 1.22 section 8.9. Warn when these
modules are inserted, and treat them as SFF8079 modules for ethtool.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Thu, 30 Nov 2017 13:59:16 +0000 (13:59 +0000)]
sfp: improve RX_LOS handling
There are two bits in the option word for the RX_LOS signal. One
reports that the RX_LOS signal is active high, the other reports that
it is active low. When both or neither are set, the result is not
well defined in the specification.
Rather than assuming that neither set means normal RX_LOS, take this
as meaning no RX_LOS signal available, thereby ignoring the signal.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Thu, 30 Nov 2017 13:59:11 +0000 (13:59 +0000)]
sfp: fix RX_LOS signal handling
The options word is a be16 quantity, so we need to test the flags
having converted the endian-ness. Convert the flag bits to be16,
which can be optimised by the compiler, rather than converting a
variable at runtime.
Reported-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Tommi Rantala [Wed, 29 Nov 2017 10:48:42 +0000 (12:48 +0200)]
tipc: call tipc_rcv() only if bearer is up in tipc_udp_recv()
Remove the second tipc_rcv() call in tipc_udp_recv(). We have just
checked that the bearer is not up, and calling tipc_rcv() with a bearer
that is not up leads to a TIPC div-by-zero crash in
tipc_node_calculate_timer(). The crash is rare in practice, but can
happen like this:
We're enabling a bearer, but it's not yet up and fully initialized.
At the same time we receive a discovery packet, and in tipc_udp_recv()
we end up calling tipc_rcv() with the not-yet-initialized bearer,
causing later the div-by-zero crash in tipc_node_calculate_timer().
Jon Maloy explains the impact of removing the second tipc_rcv() call:
"link setup in the worst case will be delayed until the next arriving
discovery messages, 1 sec later, and this is an acceptable delay."
As the tipc_rcv() call is removed, just leave the function via the
rcu_out label, so that we will kfree_skb().
Fixes: c9b64d492b1f ("tipc: add replicast peer discovery") Signed-off-by: Tommi Rantala <tommi.t.rantala@nokia.com> Cc: Jon Maloy <jon.maloy@ericsson.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Fri, 1 Dec 2017 18:06:56 +0000 (10:06 -0800)]
tcp/dccp: block bh before arming time_wait timer
Maciej Żenczykowski reported some panics in tcp_twsk_destructor()
that might be caused by the following bug.
timewait timer is pinned to the cpu, because we want to transition
timwewait refcount from 0 to 4 in one go, once everything has been
initialized.
At the time commit ed2e92394589 ("tcp/dccp: fix timewait races in timer
handling") was merged, TCP was always running from BH habdler.
After commit 5413d1babe8f ("net: do not block BH while processing
socket backlog") we definitely can run tcp_time_wait() from process
context.
We need to block BH in the critical section so that the pinned timer
has still its purpose.
This bug is more likely to happen under stress and when very small RTO
are used in datacenter flows.
Fixes: 5413d1babe8f ("net: do not block BH while processing socket backlog") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: Maciej Żenczykowski <maze@google.com> Acked-by: Maciej Żenczykowski <maze@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / mirror_ubuntu-kernels.git / log