The meta_type_ops structures are never modified, so declare them as const.
Done with the help of Coccinelle.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add Byte Queue Limits (BQL) support to bcmgenet driver.
Signed-off-by: Petri Gynther <pgynther@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
bcmgenet_isr1() and bcmgenet_isr0() run in hard irq context,
we do not need to block irq again.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Petri Gynther <pgynther@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
By using napi_complete_done(), we allow fine tuning
of /sys/class/net/ethX/gro_flush_timeout for higher GRO aggregation
efficiency for a Gbit NIC.
Check commit 24d2e4a507 ("tg3: use napi_complete_done()") for details.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Petri Gynther <pgynther@google.com>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Petri Gynther <pgynther@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Marcelo Ricardo Leitner says:
====================
sctp: delay calls to sk_data_ready() as much as possible
1st patch is a preparation for the 2nd. The idea is to not call
->sk_data_ready() for every data chunk processed while processing
packets but only once before releasing the socket.
v2: patchset re-checked, small changelog fixes
v3: on patch 2, make use of local vars to make it more readable
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently processing of multiple chunks in a single SCTP packet leads to
multiple calls to sk_data_ready, causing multiple wake up signals which
are costy and doesn't make it wake up any faster.
With this patch it will note that the wake up is pending and will do it
before leaving the state machine interpreter, latest place possible to
do it realiably and cleanly.
Note that sk_data_ready events are not dependent on asocs, unlike waking
up writers.
v2: series re-checked
v3: use local vars to cleanup the code, suggested by Jakub Sitnicki
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It wastes space and gets worse as we add new flags, so convert bit-wide
flags to a bitfield.
Currently it already saves 4 bytes in sctp_sock, which are left as holes
in it for now. The whole struct needs packing, which should be done in
another patch.
Note that do_auto_asconf cannot be merged, as explained in the comment
before it.
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This function compiles to 895 bytes of machine code.
Clearly, this isn't a time-critical function.
For one, it has a number of udelay(1) calls.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
CC: David S. Miller <davem@davemloft.net>
CC: linux-kernel@vger.kernel.org
CC: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
Xin Long says:
====================
bridge: support sending rntl info when we set attributes through sysfs/ioctl
This patchset is used to support sending rntl info to user in some places,
and ensure that whenever those attributes change internally or from sysfs,
that a netlink notification is sent out to listeners.
It also make some adjustment in bridge sysfs so that we can implement this
easily.
I've done some tests on this patchset, like:
[br_sysfs]
1. change all the attribute values of br or brif:
$ echo $value > /sys/class/net/br0/bridge/{*}
$ echo $value > /sys/class/net/br0/brif/eth1/{*}
2. meanwhile, on another terminal to observe the msg:
$ bridge monitor
[br_ioctl]
1. in bridge-utils package, do some changes in br_set, let brctl command
use ioctl to set attribute:
if ((ret = set_sysfs(path, value)) < 0) { -->
if (1) {
$ brctl set*
2. meanwhile, on another terminal to observe the msg:
$ bridge monitor
This test covers all the attributes that brctl and sysfs support to set.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Now when we change the attributes of bridge or br_port by netlink,
a relevant netlink notification will be sent, but if we change them
by ioctl or sysfs, no notification will be sent.
We should ensure that whenever those attributes change internally or from
sysfs/ioctl, that a netlink notification is sent out to listeners.
Also, NetworkManager will use this in the future to listen for out-of-band
bridge master attribute updates and incorporate them into the runtime
configuration.
This patch is used for ioctl.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now when we change the attributes of bridge or br_port by netlink,
a relevant netlink notification will be sent, but if we change them
by ioctl or sysfs, no notification will be sent.
We should ensure that whenever those attributes change internally or from
sysfs/ioctl, that a netlink notification is sent out to listeners.
Also, NetworkManager will use this in the future to listen for out-of-band
bridge master attribute updates and incorporate them into the runtime
configuration.
This patch is used for br_sysfs_if, and we also move br_ifinfo_notify out
of store_flag.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now when we change the attributes of bridge or br_port by netlink,
a relevant netlink notification will be sent, but if we change them
by ioctl or sysfs, no notification will be sent.
We should ensure that whenever those attributes change internally or from
sysfs/ioctl, that a netlink notification is sent out to listeners.
Also, NetworkManager will use this in the future to listen for out-of-band
bridge master attribute updates and incorporate them into the runtime
configuration.
This patch is used for br_sysfs_br. and we also need to remove some
rtnl_trylock in old functions so that we can call it in a common one.
For group_addr_store, we cannot make it use store_bridge_parm, because
it's not a string-to-long convert, we will add notification on it
individually.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are some repetitive codes in stp_state_store, we can remove
them by calling store_bridge_parm.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are some repetitive codes in forward_delay_store, we can remove
them by calling store_bridge_parm.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are some repetitive codes in flush_store, we can remove
them by calling store_bridge_parm, also, it would send rtnl notification
after we add it in store_bridge_parm in the following patches.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Reviewed-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fix typos in Documentation/networking/dsa.
Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The original tokenized iid support implemented via f53adae4ea ("net: ipv6:
add tokenized interface identifier support") didn't allow for clearing a
device token as it was intended that this addressing mode was the only one
active for globally scoped IPv6 addresses. Later we relaxed that restriction
via 617fe29d45 ("net: ipv6: only invalidate previously tokenized addresses"),
and we should also allow for clearing tokens as there's no good reason why
it shouldn't be allowed.
Fixes: 617fe29d45 ("net: ipv6: only invalidate previously tokenized addresses")
Reported-by: Robin H. Johnson <robbat2@gentoo.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
sock_owned_by_user should not be used without socket lock held. It seems
to be a common practice to check .owned before lock reclassification, so
provide a little help to abstract this check away.
Cc: linux-cifs@vger.kernel.org
Cc: linux-bluetooth@vger.kernel.org
Cc: linux-nfs@vger.kernel.org
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
On GMAC4.xx each descriptor contains 2 buffers of 16KB (each).
Initially, those 2 buffers was filled in dwmac4_rd_prepare_tx_desc but
it is actually not needed. Indeed, stmmac driver supports frame up to
9000 bytes (jumbo). So only one buffer is needed.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alexandre TORGUE <alexandre.torgue@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Willem de Bruijn says:
====================
fix two more udp pull header issues
Follow up patches to the fixes to RxRPC and SunRPC. A scan of the
code showed two other interfaces that expect UDP packets to have
a udphdr when queued: read packet length with ioctl SIOCINQ and
receive payload checksum with socket option IP_CHECKSUM.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
On udp sockets, recv cmsg IP_CMSG_CHECKSUM returns a checksum over
the packet payload. Since commit e6afc8ace6 pulled the headers,
taking skb->data as the start of transport header is incorrect. Use
the transport header pointer.
Also, when peeking at an offset from the start of the packet, only
return a checksum from the start of the peeked data. Note that the
cmsg does not subtract a tail checkum when reading truncated data.
Fixes: e6afc8ace6 ("udp: remove headers from UDP packets before queueing")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On udp sockets, ioctl SIOCINQ returns the payload size of the first
packet. Since commit e6afc8ace6 pulled the headers, the result is
incorrect when subtracting header length. Remove that operation.
Fixes: e6afc8ace6 ("udp: remove headers from UDP packets before queueing")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn says:
====================
DSA refactoring: set 1
There has been a long running effort to refractor DSA probing to make
the switches true linux devices. Here are a small collection of
patches moving in this direction. Most have been seen before.
We take a little step forward by passing the dsa device point to the
driver, thus allowing it to perform resource allocations using the
normal mechanisms. This device structure will later be replaced by the
devices own device structure.
Future patches will add a true driver probe function, so we rename the
current probe function, cleaning up the namespace.
phys_port_mask continually confuses me, thinking it is about PHYs. But
it is actually about ports enabled to the outside world. So rename it to
enabled_port_mask.
Lots more patches yet to follow, this is just doing some ground work.
v2:
enabled_port_mask instread of user_port_masks
Added Tested-by's and Reviewed-by.
====================
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
mv88e6xxx_lookup_name() returns the model name of a switch at a given
address on an MII bus. Using mii_bus to identify the bus rather than
the host device is more logical, so change the parameter.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The phys in phys_port_mask suggests this mask is about PHYs. In fact,
it means physical ports. Rename to enabled_port_mask, indicating
external enabled ports of the switch, which is hopefully less
confusing.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rename the function called from the DSA to perform a probe for the
switch. This makes the normal _probe() name available for a standard
Linux device driver probe function.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rather than looking up the mii bus and address every time, do it once
at probe, and keep it in the private structure. Centralise this probe
code in mv88e6xxx.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The drivers now allocate their own memory for private usage. Remove
the allocation from the core code.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now the switch devices have a dev pointer, make use of it for allocating
the drivers private data structures using a devm_kzalloc().
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
By passing a device structure to the switch devices, it allows them
to use devm_* methods for resource management.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
all drivers - removing the duplicated enum ieee80211_band and
replacing it by enum nl80211_band. On top of that, just a small
documentation update.
-----BEGIN PGP SIGNATURE-----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=2xfe
-----END PGP SIGNATURE-----
Merge tag 'mac80211-next-for-davem-2016-04-13' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next
Johannes Berg says:
====================
To synchronize with Kalle, here's just a big change that affects
all drivers - removing the duplicated enum ieee80211_band and
replacing it by enum nl80211_band. On top of that, just a small
documentation update.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
We remove a couple of leftover fields in struct tipc_bearer. Those
were used by the old broadcast implementation, and are not needed
any longer. There is no functional changes in this commit.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
John Crispin says:
====================
net: mediatek: make the driver pass stress tests
While testing the driver we managed to get the TX path to stall and fail
to recover. When dual MAC support was added to the driver, the whole queue
stop/wake code was not properly adapted. There was also a regression in the
locking of the xmit function. The fact that watchdog_timeo was not set and
that the tx_timeout code failed to properly reset the dma, irq and queue
just made the mess complete.
This series make the driver pass stress testing. With this series applied
the testbed has been running for several days and still has not locked up.
We have a second setup that has a small hack patch applied to randomly stop
irqs and/or one of the queues and successfully manages to recover from these
simulated tx stalls.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The QID field gets set to the mac id. This made the DMA linked list queue
the traffic of each MAC on a different internal queue. However during long
term testing we found that this will cause traffic stalls as the multi
queue setup requires a more complete initialisation which is not part of
the upstream driver yet.
This patch removes the code setting the QID field, resulting in all
traffic ending up in queue 0 which works without any special setup.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The worker always touches both netdevs. It is ethernet core and not MAC
specific. We only need one worker, which belongs into the ethernets core
struct.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The driver supports 2 MACs. Both run on the same DMA ring. If we hit a TX
timeout we need to stop both netdevs before restarting them again. If we
don't do this, mtk_stop() wont shutdown DMA and the consecutive call to
mtk_open() wont restart DMA and enable IRQs.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Inside the TX path there is a lock inside the tx_map function. This is
however too late. The patch moves the lock to the start of the xmit
function right before the free count check of the DMA ring happens.
If we do not do this, the code becomes racy leading to TX stalls and
dropped packets. This happens as there are 2 netdevs running on the
same physical DMA ring.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The driver supports 2 MACs. Both run on the same DMA ring. If we go
above/below the TX rings threshold value, we always need to wake/stop
the queue of both devices. Not doing to can cause TX stalls and packet
drops on one of the devices.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
HW reset is triggered in the mtk_hw_init() function. There is no need to
also reset the core during probe.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The code used to also support the PDMA engine, which had 2 packet pointers
per descriptor. Because of this we had to divide the result by 2 and round
it up. This is no longer needed as the code only supports QDMA.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The original commit failed to set watchdog_timeo. This patch sets
watchdog_timeo to HZ.
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following patchset contains the first batch of Netfilter updates for
your net-next tree.
1) Define pr_fmt() in nf_conntrack, from Weongyo Jeong.
2) Define and register netfilter's afinfo for the bridge family,
this comes in preparation for native nfqueue's bridge for nft,
from Stephane Bryant.
3) Add new attributes to store layer 2 and VLAN headers to nfqueue,
also from Stephane Bryant.
4) Parse new NFQA_VLAN and NFQA_L2HDR nfqueue netlink attributes
coming from userspace, from Stephane Bryant.
5) Use net->ipv6.devconf_all->hop_limit instead of hardcoded hop_limit
in IPv6 SYNPROXY, from Liping Zhang.
6) Remove unnecessary check for dst == NULL in nf_reject_ipv6,
from Haishuang Yan.
7) Deinline ctnetlink event report functions, from Florian Westphal.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Not performance critical, it is only invoked when an expectation is
added/destroyed.
While at it, kill unused nf_ct_expect_event() wrapper.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Way too large; move it to nf_conntrack_ecache.c.
Reduces total object size by 1216 byte on my machine.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Johan Hedberg says:
====================
pull request: bluetooth-next 2016-04-12
Here's a set of Bluetooth & 802.15.4 patches intended for the 4.7 kernel:
- Fix for race condition in vhci driver
- Memory leak fix for ieee802154/adf7242 driver
- Improvements to deal with single-mode (LE-only) Bluetooth controllers
- Fix for allowing the BT_SECURITY_FIPS security level
- New BCM2E71 ACPI ID
- NULL pointer dereference fix fox hci_ldisc driver
Let me know if there are any issues pulling. Thanks.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This enum is already perfectly aliased to enum nl80211_band, and
the only reason for it is that we get IEEE80211_NUM_BANDS out of
it. There's no really good reason to not declare the number of
bands in nl80211 though, so do that and remove the cfg80211 one.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The roaming cases for the Connect command were not fully covered and
neither Connect nor Associate command uses of the prev_bssid parameter
were very clear. Add details to describe how the prev_bssid argument is
supposed to be used and when the driver should use association or
reassociation.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
MDIO devices can be stacked upon each other. The current code supports
two levels, which until recently has been enough for a DSA mdio bus on
top of another bus. Now we have hardware which has an MDIO mux in the
middle.
Define an MDIO MUTEX class with three levels.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
David Howells says:
====================
RxRPC: 2nd rewrite part 1
Okay, I'm in the process of rewriting the RxRPC rewrite. The primary aim of
this second rewrite is to strictly control the number of active connections we
know about and to get rid of connections we don't need much more quickly.
On top of this, there are fixes to the protocol handling which will all occur
in later parts.
Here's the first set of patches from the second go, aimed at net-next. These
are all fixes and cleanups preparatory to the main event.
Notable parts of this set include:
(1) A fix for the AFS filesystem to wait for outstanding calls to complete
before closing the RxRPC socket.
(2) Differentiation of local and remote abort codes. At a future point
userspace will get to see this via control message data on recvmsg().
(3) Absorb the rxkad module into the af_rxrpc module to prevent a dependency
loop.
(4) Create a null security module and unconditionalise calls into the
security module that's in force (there will always be a security module
applied to a connection, even if it's just the null one).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
