]> git.karo-electronics.de Git - linux-beck.git/log
linux-beck.git
12 years agoixgbe: ptp code cleanup
Jacob Keller [Tue, 22 May 2012 06:08:32 +0000 (06:08 +0000)]
ixgbe: ptp code cleanup

This patch fixes two minor nits from Richard Cochran. The first is a case of
ambitious line wrapping that wasn't necessary. The second is to re-order the
flag checks for PPS support. Previously, the hardware test was done first, and
the interrupt flag test was done second. Now, test the interrupt flag and use
the unlikely macro.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
12 years agoixgbe: do not compile ixgbe_sysfs.c when CONFIG_IXGBE_HWMON is not set
Emil Tantilov [Wed, 16 May 2012 07:06:38 +0000 (07:06 +0000)]
ixgbe: do not compile ixgbe_sysfs.c when CONFIG_IXGBE_HWMON is not set

ixgbe_sysfs.c is only needed when CONFIG_IXGBE_HWMON is configured in the
kernel.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Acked-by: Don Skidmore <Donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
12 years agoixgbe: align flow control DV macros with datasheet
John Fastabend [Wed, 28 Mar 2012 11:42:45 +0000 (11:42 +0000)]
ixgbe: align flow control DV macros with datasheet

The flow control DV macros are used to calculate the flow control
high and low thresholds. This patch annotates these macros slightly
better and fixes the issues below.

The macro variables are renamed LINK to _max_frame_link and TC to
_max_frame_tc. This was to avoid confusion and make them more
readable. It was found that people auditing the code read TC to be
'traffic class' in the 802.1Q definition instead of the max frame
size of the tc. Hopefully it is clear now.

This audit also found the following real deviations from the
theoretical values. Fixed in this patch.

  * I multiplied the DV calculations by (36/25) which always
    evaluates to 1. This does not match the intended theoretical
    value of 1.44.

  * IXGBE_BT2KB added 1023 to account for rounding however this
    really should be 8 * 1023 - 1 to account for division by 8k.

  * x2 multiplication of max frame in DV calculations to account
    for updated hardware recommendations.

With this patch the DV values are inline with the recommendations
in the 82599 and 82598 data sheets. Its worth noting I did not
see any dropped frames with flow control on in my experiments without
this patch. However aligning with the hardware specs and
recommendations seems like a good idea here to account for worst
case scenarios.

Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
12 years agoe1000e: use more informative logging macros when netdev not yet registered
Bruce Allan [Thu, 7 Jun 2012 02:23:37 +0000 (02:23 +0000)]
e1000e: use more informative logging macros when netdev not yet registered

Based on a report from Ethan Zhao, before calling register_netdev() the
driver should be using logging macros that do not display the potentially
confusing "(unregistered net_device)" yet still display the useful driver
name and PCI bus/device/function.

Reported-by: Ethan Zhao <ethan.kernel@gmail.com>
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
12 years agobonding: drop_monitor aware
Eric Dumazet [Wed, 13 Jun 2012 05:30:07 +0000 (05:30 +0000)]
bonding: drop_monitor aware

When packets are dropped in TX path, its better to use kfree_skb()
instead of dev_kfree_skb() to give proper drop_monitor events.

Also move the kfree_skb() call after read_unlock() in bond_alb_xmit()
and bond_xmit_activebackup()

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodcbnl: Use type safe nlmsg_data()
Thomas Graf [Wed, 13 Jun 2012 02:55:01 +0000 (02:55 +0000)]
dcbnl: Use type safe nlmsg_data()

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodcbnl: Move dcb app allocation into dcb_app_add()
Thomas Graf [Wed, 13 Jun 2012 02:55:00 +0000 (02:55 +0000)]
dcbnl: Move dcb app allocation into dcb_app_add()

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodcbnl: Move dcb app lookup code into dcb_app_lookup()
Thomas Graf [Wed, 13 Jun 2012 02:54:59 +0000 (02:54 +0000)]
dcbnl: Move dcb app lookup code into dcb_app_lookup()

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodcbnl: Return consistent error codes
Thomas Graf [Wed, 13 Jun 2012 02:54:58 +0000 (02:54 +0000)]
dcbnl: Return consistent error codes

EMSGSIZE - ran out of space while constructing message
EOPNOTSUPP - driver/hardware does not support operation
ENODEV - network device not found
EINVAL - invalid message

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodcbnl: Use dcbnl_newmsg() where possible
Thomas Graf [Wed, 13 Jun 2012 02:54:57 +0000 (02:54 +0000)]
dcbnl: Use dcbnl_newmsg() where possible

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodcbnl: Remove now unused dcbnl_reply()
Thomas Graf [Wed, 13 Jun 2012 02:54:56 +0000 (02:54 +0000)]
dcbnl: Remove now unused dcbnl_reply()

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodcbnl: Shorten all command handling functions
Thomas Graf [Wed, 13 Jun 2012 02:54:55 +0000 (02:54 +0000)]
dcbnl: Shorten all command handling functions

Allocating and sending the skb in dcb_doit() allows for much
shorter and cleaner command handling functions.

The huge switch statement is replaced with an array based definition
of the handling function and reply message type.

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodcbnl: Prepare framework to shorten handling functions
Thomas Graf [Wed, 13 Jun 2012 02:54:54 +0000 (02:54 +0000)]
dcbnl: Prepare framework to shorten handling functions

There is no need to allocate and send the reply message in each
handling function separately. Instead, the reply skb can be allocated
and sent in dcb_doit() directly.

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
David S. Miller [Wed, 13 Jun 2012 04:59:18 +0000 (21:59 -0700)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net

Conflicts:
MAINTAINERS
drivers/net/wireless/iwlwifi/pcie/trans.c

The iwlwifi conflict was resolved by keeping the code added
in 'net' that turns off the buggy chip feature.

The MAINTAINERS conflict was merely overlapping changes, one
change updated all the wireless web site URLs and the other
changed some GIT trees to be Johannes's instead of John's.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoethtool: Make more commands available to unprivileged processes
Ben Hutchings [Tue, 12 Jun 2012 13:05:41 +0000 (13:05 +0000)]
ethtool: Make more commands available to unprivileged processes

'Get' commands should generally not require CAP_NET_ADMIN, with
the exception of those that expose internal state.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet-next: add dev_loopback_xmit() to avoid duplicate code
Michel Machado [Tue, 12 Jun 2012 10:16:35 +0000 (10:16 +0000)]
net-next: add dev_loopback_xmit() to avoid duplicate code

Add dev_loopback_xmit() in order to deduplicate functions
ip_dev_loopback_xmit() (in net/ipv4/ip_output.c) and
ip6_dev_loopback_xmit() (in net/ipv6/ip6_output.c).

I was about to reinvent the wheel when I noticed that
ip_dev_loopback_xmit() and ip6_dev_loopback_xmit() do exactly what I
need and are not IP-only functions, but they were not available to reuse
elsewhere.

ip6_dev_loopback_xmit() does not have line "skb_dst_force(skb);", but I
understand that this is harmless, and should be in dev_loopback_xmit().

Signed-off-by: Michel Machado <michel@digirati.com.br>
CC: "David S. Miller" <davem@davemloft.net>
CC: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
CC: James Morris <jmorris@namei.org>
CC: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
CC: Patrick McHardy <kaber@trash.net>
CC: Eric Dumazet <edumazet@google.com>
CC: Jiri Pirko <jpirko@redhat.com>
CC: "Michał Mirosław" <mirq-linux@rere.qmqm.pl>
CC: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobonding: remove packet cloning in recv_probe()
Eric Dumazet [Mon, 11 Jun 2012 19:23:07 +0000 (19:23 +0000)]
bonding: remove packet cloning in recv_probe()

Cloning all packets in input path have a significant cost.

Use skb_header_pointer()/skb_copy_bits() instead of pskb_may_pull() so
that recv_probe handlers (bond_3ad_lacpdu_recv / bond_arp_rcv /
rlb_arp_recv ) dont touch input skb.

bond_handle_frame() can avoid the skb_clone()/dev_kfree_skb()

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jay Vosburgh <fubar@us.ibm.com>
Cc: Andy Gospodarek <andy@greyhouse.net>
Cc: Jiri Bohac <jbohac@suse.cz>
Cc: Nicolas de Pesloüan <nicolas.2p.debian@free.fr>
Cc: Maciej Żenczykowski <maze@google.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agousbnet: don't initialize transfer buffer before submit status URB
tom.leiming@gmail.com [Mon, 11 Jun 2012 15:19:45 +0000 (15:19 +0000)]
usbnet: don't initialize transfer buffer before submit status URB

The line below in intr_complete isn't needed,

memset(urb->transfer_buffer, 0, urb->transfer_buffer_length);

so just remove it.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agousbnet: remove declaration for intr_complete
tom.leiming@gmail.com [Mon, 11 Jun 2012 15:19:44 +0000 (15:19 +0000)]
usbnet: remove declaration for intr_complete

Remove declaration for intr_complete so that ctags may be happy to
decrease duplicated symbols, also decrease one line code.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agousbnet: remove flag of EVENT_DEV_WAKING
tom.leiming@gmail.com [Mon, 11 Jun 2012 15:19:43 +0000 (15:19 +0000)]
usbnet: remove flag of EVENT_DEV_WAKING

The flag of EVENT_DEV_WAKING is not used any more, so just remove it.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agousbnet:cdc-phonet: remove usb_get/put_dev in .probe and .disconnect
tom.leiming@gmail.com [Mon, 11 Jun 2012 15:19:41 +0000 (15:19 +0000)]
usbnet:cdc-phonet: remove usb_get/put_dev in .probe and .disconnect

usb_device is parent device of usb_interface in the view of driver
model, so its reference count is always held during .probe/.disconnect
of usb_interface instance.

This patch just removes the unnecessay usb_get/put_dev.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agousbnet:pegasus: remove usb_get/put_dev in .probe and .disconnect
tom.leiming@gmail.com [Mon, 11 Jun 2012 15:19:40 +0000 (15:19 +0000)]
usbnet:pegasus: remove usb_get/put_dev in .probe and .disconnect

usb_device is parent device of usb_interface in the view of driver
model, so its reference count is always held during .probe/.disconnect
of usb_interface instance.

This patch just removes the unnecessay usb_get/put_dev.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agousbnet: remove usb_get/put_dev in .probe and .disconnect
tom.leiming@gmail.com [Mon, 11 Jun 2012 15:19:39 +0000 (15:19 +0000)]
usbnet: remove usb_get/put_dev in .probe and .disconnect

usb_device is parent device of usb_interface in the view of driver
model, so its reference count is always held during .probe/.disconnect
of usb_interface instance.

This patch just removes the unnecessay usb_get/put_dev.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobonding: Fix corrupted queue_mapping
Eric Dumazet [Tue, 12 Jun 2012 06:03:51 +0000 (06:03 +0000)]
bonding: Fix corrupted queue_mapping

In the transmit path of the bonding driver, skb->cb is used to
stash the skb->queue_mapping so that the bonding device can set its
own queue mapping.  This value becomes corrupted since the skb->cb is
also used in __dev_xmit_skb.

When transmitting through bonding driver, bond_select_queue is
called from dev_queue_xmit.  In bond_select_queue the original
skb->queue_mapping is copied into skb->cb (via bond_queue_mapping)
and skb->queue_mapping is overwritten with the bond driver queue.

Subsequently in dev_queue_xmit, __dev_xmit_skb is called which writes
the packet length into skb->cb, thereby overwriting the stashed
queue mappping.  In bond_dev_queue_xmit (called from hard_start_xmit),
the queue mapping for the skb is set to the stashed value which is now
the skb length and hence is an invalid queue for the slave device.

If we want to save skb->queue_mapping into skb->cb[], best place is to
add a field in struct qdisc_skb_cb, to make sure it wont conflict with
other layers (eg : Qdiscc, Infiniband...)

This patchs also makes sure (struct qdisc_skb_cb)->data is aligned on 8
bytes :

netem qdisc for example assumes it can store an u64 in it, without
misalignment penalty.

Note : we only have 20 bytes left in (struct qdisc_skb_cb)->data[].
The largest user is CHOKe and it fills it.

Based on a previous patch from Tom Herbert.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Tom Herbert <therbert@google.com>
Cc: John Fastabend <john.r.fastabend@intel.com>
Cc: Roland Dreier <roland@kernel.org>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoipv4: Add interface option to enable routing of 127.0.0.0/8
Thomas Graf [Tue, 12 Jun 2012 00:44:01 +0000 (00:44 +0000)]
ipv4: Add interface option to enable routing of 127.0.0.0/8

Routing of 127/8 is tradtionally forbidden, we consider
packets from that address block martian when routing and do
not process corresponding ARP requests.

This is a sane default but renders a huge address space
practically unuseable.

The RFC states that no address within the 127/8 block should
ever appear on any network anywhere but it does not forbid
the use of such addresses outside of the loopback device in
particular. For example to address a pool of virtual guests
behind a load balancer.

This patch adds a new interface option 'route_localnet'
enabling routing of the 127/8 address block and processing
of ARP requests on a specific interface.

Note that for the feature to work, the default local route
covering 127/8 dev lo needs to be removed.

Example:
  $ sysctl -w net.ipv4.conf.eth0.route_localnet=1
  $ ip route del 127.0.0.0/8 dev lo table local
  $ ip addr add 127.1.0.1/16 dev eth0
  $ ip route flush cache

V2: Fix invalid check to auto flush cache (thanks davem)

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobonding:record primary when modify it via sysfs
Weiping Pan [Sun, 10 Jun 2012 23:00:20 +0000 (23:00 +0000)]
bonding:record primary when modify it via sysfs

If we modify primary via sysfs and it is not a valid slave,
we should record it for future use, and this behavior is the same with
bond_check_params().

Signed-off-by: Weiping Pan <wpan@redhat.com>
Acked-by: Nicolas de Pesloüan <nicolas.2p.debian@free.fr>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge branch 'master' of git://1984.lsi.us.es/net
David S. Miller [Tue, 12 Jun 2012 22:18:15 +0000 (15:18 -0700)]
Merge branch 'master' of git://1984.lsi.us.es/net

12 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
John W. Linville [Tue, 12 Jun 2012 18:25:04 +0000 (14:25 -0400)]
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-next into for-davem

12 years agotilegx network driver: initial support
Chris Metcalf [Thu, 7 Jun 2012 10:45:02 +0000 (10:45 +0000)]
tilegx network driver: initial support

This change adds support for the tilegx network driver based on the
GXIO IORPC support in the tilegx software stack, using the on-chip
mPIPE packet processing engine.

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agophy: Use pr_<level>
Joe Perches [Sat, 9 Jun 2012 07:49:07 +0000 (07:49 +0000)]
phy: Use pr_<level>

Use a more current logging style.

Add pr_fmt and missing newlines.
Remove embedded prefixes.
Neaten phy_print_status to avoid using KERN_CONT.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agotg3: Apply short DMA frag workaround to 5906
Matt Carlson [Thu, 7 Jun 2012 12:56:54 +0000 (12:56 +0000)]
tg3: Apply short DMA frag workaround to 5906

5906 devices also need the short DMA fragment workaround.  This patch
makes the necessary change.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Tested-by: Christian Kujau <lists@nerdbynature.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoaf_packet: use sizeof instead of constant in spkt_device
danborkmann@iogearbox.net [Sun, 10 Jun 2012 08:59:28 +0000 (08:59 +0000)]
af_packet: use sizeof instead of constant in spkt_device

This small patch removes access to the last element of the spkt_device
array through a constant. Instead, it is accessed by sizeof() to respect
possible changes in if_packet.h.

Signed-off-by: Daniel Borkmann <daniel.borkmann@tik.ee.ethz.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet: stmmac: Fix clock en-/disable calls
Stefan Roese [Thu, 7 Jun 2012 23:21:05 +0000 (23:21 +0000)]
net: stmmac: Fix clock en-/disable calls

clk_{un}prepare is mandatory for platforms using common clock framework.
Since these drivers are used by SPEAr platform, which supports common
clock framework, add clk_{un}prepare() support for them. Otherwise
the clocks are not correctly en-/disabled and ethernet support doesn't
work.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Viresh Kumar <viresh.linux@gmail.com>
Cc: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Fix BUG triggered by __rt{,6}_get_peer().
David S. Miller [Mon, 11 Jun 2012 22:52:29 +0000 (15:52 -0700)]
inet: Fix BUG triggered by __rt{,6}_get_peer().

If no peer actually gets attached (either because create is zero or
the peer allocation fails) we'll trigger a BUG because we
unconditionally do an rt{,6}_peer_ptr() afterwards.

Fix this by guarding it with the proper check.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonetfilter: nf_ct_tcp, udp: fix compilation with sysctl disabled
Pablo Neira Ayuso [Mon, 11 Jun 2012 21:58:01 +0000 (23:58 +0200)]
netfilter: nf_ct_tcp, udp: fix compilation with sysctl disabled

This patch fixes the compilation of the TCP and UDP trackers with sysctl
compilation disabled:

net/netfilter/nf_conntrack_proto_udp.c: In function ‘udp_init_net_data’:
net/netfilter/nf_conntrack_proto_udp.c:279:13: error: ‘struct nf_proto_net’ has no member named
 ‘user’
net/netfilter/nf_conntrack_proto_tcp.c:1606:9: error: ‘struct nf_proto_net’ has no member named
 ‘user’
net/netfilter/nf_conntrack_proto_tcp.c:1643:9: error: ‘struct nf_proto_net’ has no member named
 ‘user’

Reported-by: Fengguang Wu <wfg@linux.intel.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agolpc_eth: fix tx completion
Eric Dumazet [Mon, 11 Jun 2012 07:21:36 +0000 (07:21 +0000)]
lpc_eth: fix tx completion

__lpc_handle_xmit() has two bugs :

1) It can leak skbs in case TXSTATUS_ERROR is set

2) It can wake up txqueue while no slot was freed.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Roland Stigge <stigge@antcom.de>
Tested-by: Roland Stigge <stigge@antcom.de>
Cc: Kevin Wells <kevin.wells@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agolpc_eth: add missing ndo_change_mtu()
Eric Dumazet [Sun, 10 Jun 2012 23:24:00 +0000 (23:24 +0000)]
lpc_eth: add missing ndo_change_mtu()

lpc_eth does a copy of transmitted skbs to DMA area, without checking
skb lengths, so can trigger buffer overflows :

memcpy(pldat->tx_buff_v + txidx * ENET_MAXF_SIZE, skb->data, len);

One way to get bigger skbs is to allow MTU changes above the 1500 limit.

Calling eth_change_mtu() in ndo_change_mtu() makes sure this cannot
happen.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Roland Stigge <stigge@antcom.de>
Cc: Kevin Wells <kevin.wells@nxp.com>
Acked-by: Roland Stigge <stigge@antcom.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet: keep name_hlist close to name
Eric Dumazet [Mon, 11 Jun 2012 06:36:13 +0000 (06:36 +0000)]
net: keep name_hlist close to name

__dev_get_by_name() is slow because pm_qos_req has been inserted between
name[] and name_hlist, adding cache misses.

pm_qos_req has nothing to do at the beginning of struct net_device

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge branch 'master' of git://1984.lsi.us.es/net-next
David S. Miller [Mon, 11 Jun 2012 19:56:14 +0000 (12:56 -0700)]
Merge branch 'master' of git://1984.lsi.us.es/net-next

12 years agossb: add missing PCI ID for b/g/n single band BCM4322
Jonas Gorski [Sun, 10 Jun 2012 20:11:56 +0000 (22:11 +0200)]
ssb: add missing PCI ID for b/g/n single band BCM4322

14e4:432c is found on some bcm63xx devices. The device is working fine
with b43.

Reported-by: Álvaro Fernández Rojas <noltari@gmail.com>
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agoath9k_hw: Initvals update for AR9462
Sujith Manoharan [Fri, 8 Jun 2012 07:54:55 +0000 (13:24 +0530)]
ath9k_hw: Initvals update for AR9462

MSI is enabled by default for most of the 4th generation
chips. Add this for AR9462 - this fixes PowerSave operation,
the chip was not entering Network-Sleep mode earlier.
With proper powering down of the MAC now, power consumption
in associated state is reduced considerably.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agomwifiex: scan less channels per scan command to improve Tx traffic
Amitkumar Karwar [Thu, 7 Jun 2012 04:12:42 +0000 (21:12 -0700)]
mwifiex: scan less channels per scan command to improve Tx traffic

Currently 4 channels are scanned per scan command. if scan request
is issued by user during Tx traffic, radio will be out of channel
for "4 * per_chan_scan_time" for each scan command and will not be
able to receive Rx packets. This adds delay in data traffic. We can
minimize it by reducing number of channels scanned per scan command
in this scenario.

We can not always scan 1 channel per scan command due to limitation
of number of command buffers. So we add code to decide number of
channels scanned per scan command in associated state.

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agomwifiex: fix simultaneous scan and Tx traffic problem
Amitkumar Karwar [Thu, 7 Jun 2012 04:12:41 +0000 (21:12 -0700)]
mwifiex: fix simultaneous scan and Tx traffic problem

If scan operation is started when Tx traffic is already running,
driver locks Tx queue until it gets completed. With this logic
there is a delay for Tx packets.

This patch implements new approach to give Tx path higher priority
in this case. Driver internally sends multiple synchronous scan
commands to firmware when scan is requested by user. Now we will
make sure that Tx queue is empty everytime before sending next scan
command. If Tx queue isn't empty scan command will be postponsed by
20msec. This rule will be followed until Tx queue becomes empty or
timeout of 1 second happens. In case of timeout scan operation will
be aborted.

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agomwifiex: shorten per channel scan time
Bing Zhao [Thu, 7 Jun 2012 04:12:40 +0000 (21:12 -0700)]
mwifiex: shorten per channel scan time

Currently the scan time per channel for active scanning is set to
200ms. It takes quite a while to finsh scanning on all channels,
especially with a dual band configuration.

Change the per channel scan time settings to the following values:

passive scan: 110ms
active scan: 30ms
specific scan: 30ms

Above settings have been tested on x86 and arm platforms.

Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agoMerge branch 'for-john' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi...
John W. Linville [Mon, 11 Jun 2012 18:50:59 +0000 (14:50 -0400)]
Merge branch 'for-john' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-next

Conflicts:
drivers/net/wireless/iwlwifi/iwl-eeprom.c

12 years agoMerge tag 'nfc-next-3.6-1' of git://git.kernel.org/pub/scm/linux/kernel/git/sameo...
John W. Linville [Mon, 11 Jun 2012 18:46:04 +0000 (14:46 -0400)]
Merge tag 'nfc-next-3.6-1' of git://git.kernel.org/pub/scm/linux/kernel/git/sameo/nfc-3.0

12 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
John W. Linville [Mon, 11 Jun 2012 18:36:48 +0000 (14:36 -0400)]
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless into for-davem

12 years agoinet: Avoid potential NULL peer dereference.
David S. Miller [Mon, 11 Jun 2012 11:13:57 +0000 (04:13 -0700)]
inet: Avoid potential NULL peer dereference.

We handle NULL in rt{,6}_set_peer but then our caller will try to pass
that NULL pointer into inet_putpeer() which isn't ready for it.

Fix this by moving the NULL check one level up, and then remove the
now unnecessary NULL check from inetpeer_ptr_set_peer().

Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Use FIB table peer roots in routes.
David S. Miller [Mon, 11 Jun 2012 09:01:56 +0000 (02:01 -0700)]
inet: Use FIB table peer roots in routes.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Add inetpeer tree roots to the FIB tables.
David S. Miller [Mon, 11 Jun 2012 07:01:52 +0000 (00:01 -0700)]
inet: Add inetpeer tree roots to the FIB tables.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Add family scope inetpeer flushes.
David S. Miller [Sun, 10 Jun 2012 07:24:21 +0000 (00:24 -0700)]
inet: Add family scope inetpeer flushes.

This implementation can deal with having many inetpeer roots, which is
a necessary prerequisite for per-FIB table rooted peer tables.

Each family (AF_INET, AF_INET6) has a sequence number which we bump
when we get a family invalidation request.

Each peer lookup cheaply checks whether the flush sequence of the
root we are using is out of date, and if so flushes it and updates
the sequence number.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoipv4: Kill ip_rt_frag_needed().
David S. Miller [Sun, 10 Jun 2012 07:04:12 +0000 (00:04 -0700)]
ipv4: Kill ip_rt_frag_needed().

There is zero point to this function.

It's only real substance is to perform an extremely outdated BSD4.2
ICMP check, which we can safely remove.  If you really have a MTU
limited link being routed by a BSD4.2 derived system, here's a nickel
go buy yourself a real router.

The other actions of ip_rt_frag_needed(), checking and conditionally
updating the peer, are done by the per-protocol handlers of the ICMP
event.

TCP, UDP, et al. have a handler which will receive this event and
transmit it back into the associated route via dst_ops->update_pmtu().

This simplification is important, because it eliminates the one place
where we do not have a proper route context in which to make an
inetpeer lookup.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Hide route peer accesses behind helpers.
David S. Miller [Sun, 10 Jun 2012 05:36:36 +0000 (22:36 -0700)]
inet: Hide route peer accesses behind helpers.

We encode the pointer(s) into an unsigned long with one state bit.

The state bit is used so we can store the inetpeer tree root to use
when resolving the peer later.

Later the peer roots will be per-FIB table, and this change works to
facilitate that.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodummy: fix rcu_sched self-detected stalls
Eric Dumazet [Sun, 10 Jun 2012 21:11:57 +0000 (21:11 +0000)]
dummy: fix rcu_sched self-detected stalls

Trying to "modprobe dummy numdummies=30000" triggers :

INFO: rcu_sched self-detected stall on CPU { 8} (t=60000 jiffies)

After this splat, RTNL is locked and reboot is needed.

We must call cond_resched() to avoid this, even holding RTNL.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet: Reorder initialization in ip_route_output to fix gcc warning
Roland Dreier [Sun, 10 Jun 2012 20:05:24 +0000 (20:05 +0000)]
net: Reorder initialization in ip_route_output to fix gcc warning

If I build with W=1, for every file that includes <net/route.h>, I get the warning

    include/net/route.h: In function 'ip_route_output':
    include/net/route.h:135:3: warning: initialized field overwritten [-Woverride-init]
    include/net/route.h:135:3: warning: (near initialization for 'fl4') [-Woverride-init]

(This is with "gcc (Debian 4.6.3-1) 4.6.3")

A fix seems pretty trivial: move the initialization of .flowi4_tos
earlier.  As far as I can tell, this has no effect on code generation.

Signed-off-by: Roland Dreier <roland@purestorage.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agovirtio-net: fix a race on 32bit arches
Eric Dumazet [Tue, 5 Jun 2012 22:35:24 +0000 (22:35 +0000)]
virtio-net: fix a race on 32bit arches

commit 3fa2a1df909 (virtio-net: per cpu 64 bit stats (v2)) added a race
on 32bit arches.

We must use separate syncp for rx and tx path as they can be run at the
same time on different cpus. Thus one sequence increment can be lost and
readers spin forever.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Stephen Hemminger <shemminger@vyatta.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agor8169: avoid NAPI scheduling delay.
françois romieu [Sat, 9 Jun 2012 10:53:16 +0000 (10:53 +0000)]
r8169: avoid NAPI scheduling delay.

While reworking the r8169 driver a few months ago to perform the
smallest amount of work in the irq handler, I took care of avoiding
any irq mask register operation in the slow work dedicated user
context thread. The slow work thread scheduled an extra round of NAPI
work which would ultimately set the irq mask register as required,
thus keeping such irq mask operations in the NAPI handler.
It would eventually race with the irq handler and delay NAPI execution
for - assuming no further irq - a whole ksoftirqd period. Mildly a
problem for rare link changes or corner case PCI events.

The race was always lost after the last bh disabling lock had been
removed from the work thread and people started wondering where those
pesky "NOHZ: local_softirq_pending 08" messages came from.

Actually the irq mask register _can_ be set up directly in the slow
work thread.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Reported-by: Dave Jones <davej@redhat.com>
Tested-by: Marc Dionne <marc.c.dionne@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet: Make linux/tcp.h C++ friendly (trivial)
Paul Pluzhnikov [Sat, 9 Jun 2012 14:53:03 +0000 (07:53 -0700)]
net: Make linux/tcp.h C++ friendly (trivial)

I originally sent this patch to <trivial@kernel.org>, but Jiri Kosina did
not feel that this is fully appropriate for the trivial tree.

Using linux/tcp.h from C++ results in:

cat t.cc
#include <linux/tcp.h>
int main() { }

g++ -c t.cc

In file included from t.cc:1:
/usr/include/linux/tcp.h:72: error: '__u32 __fswab32(__u32)' cannot appear in a constant-expression
/usr/include/linux/tcp.h:72: error: a function call cannot appear in a constant-expression
...

Attached trivial patch fixes this problem.

Tested:
- the t.cc above compiles with g++ and
- the following program generates the same output before/after
  the patch:

#include <linux/tcp.h>
#include <stdio.h>

int main ()
{
#define P(a) printf("%s: %08x\n", #a, (int)a)
 P(TCP_FLAG_CWR);
 P(TCP_FLAG_ECE);
 P(TCP_FLAG_URG);
 P(TCP_FLAG_ACK);
 P(TCP_FLAG_PSH);
 P(TCP_FLAG_RST);
 P(TCP_FLAG_SYN);
 P(TCP_FLAG_FIN);
 P(TCP_RESERVED_BITS);
 P(TCP_DATA_OFFSET);
#undef P
 return 0;
}

Signed-off-by: Paul Pluzhnikov <ppluzhnikov@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Pass inetpeer root into inet_getpeer*() interfaces.
David S. Miller [Sun, 10 Jun 2012 02:12:36 +0000 (19:12 -0700)]
inet: Pass inetpeer root into inet_getpeer*() interfaces.

Otherwise we reference potentially non-existing members when
ipv6 is disabled.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoaf_unix: remove unix_iter_state
Eric Dumazet [Fri, 8 Jun 2012 22:10:20 +0000 (22:10 +0000)]
af_unix: remove unix_iter_state

As pointed out by Michael Tokarev , struct unix_iter_state is no longer
needed.

Suggested-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoipv6: Do not mark ipv6_inetpeer_ops as __net_initdata.
David S. Miller [Sun, 10 Jun 2012 02:00:16 +0000 (19:00 -0700)]
ipv6: Do not mark ipv6_inetpeer_ops as __net_initdata.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Consolidate inetpeer_invalidate_tree() interfaces.
David S. Miller [Sat, 9 Jun 2012 23:32:41 +0000 (16:32 -0700)]
inet: Consolidate inetpeer_invalidate_tree() interfaces.

We only need one interface for this operation, since we always know
which inetpeer root we want to flush.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Initialize per-netns inetpeer roots in net/ipv{4,6}/route.c
David S. Miller [Sat, 9 Jun 2012 23:27:05 +0000 (16:27 -0700)]
inet: Initialize per-netns inetpeer roots in net/ipv{4,6}/route.c

Instead of net/ipv4/inetpeer.c

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years ago[PATCH] tcp: Cache inetpeer in timewait socket, and only when necessary.
David S. Miller [Sat, 9 Jun 2012 21:56:12 +0000 (14:56 -0700)]
[PATCH] tcp: Cache inetpeer in timewait socket, and only when necessary.

Since it's guarenteed that we will access the inetpeer if we're trying
to do timewait recycling and TCP options were enabled on the
connection, just cache the peer in the timewait socket.

In the future, inetpeer lookups will be context dependent (per routing
realm), and this helps facilitate that as well.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agotcp: Get rid of inetpeer special cases.
David S. Miller [Sat, 9 Jun 2012 08:25:47 +0000 (01:25 -0700)]
tcp: Get rid of inetpeer special cases.

The get_peer method TCP uses is full of special cases that make no
sense accommodating, and it also gets in the way of doing more
reasonable things here.

First of all, if the socket doesn't have a usable cached route, there
is no sense in trying to optimize timewait recycling.

Likewise for the case where we have IP options, such as SRR enabled,
that make the IP header destination address (and thus the destination
address of the route key) differ from that of the connection's
destination address.

Just return a NULL peer in these cases, and thus we're also able to
get rid of the clumsy inetpeer release logic.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinet: Create and use rt{,6}_get_peer_create().
David S. Miller [Sat, 9 Jun 2012 06:24:18 +0000 (23:24 -0700)]
inet: Create and use rt{,6}_get_peer_create().

There's a lot of places that open-code rt{,6}_get_peer() only because
they want to set 'create' to one.  So add an rt{,6}_get_peer_create()
for their sake.

There were also a few spots open-coding plain rt{,6}_get_peer() and
those are transformed here as well.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonetdev: fix drivers/net/phy/ kernel-doc warnings
Randy Dunlap [Fri, 8 Jun 2012 14:07:19 +0000 (14:07 +0000)]
netdev: fix drivers/net/phy/ kernel-doc warnings

Fix kernel-doc warnings in drivers/net/phy:

Warning(drivers/net/phy/mdio_bus.c:109): No description found for parameter 'mdio_bus_np'
Warning(drivers/net/phy/mdio_bus.c:109): Excess function parameter 'mdio_np' description in 'of_mdio_find_bus'

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet/core: fix kernel-doc warnings
Randy Dunlap [Fri, 8 Jun 2012 14:01:44 +0000 (14:01 +0000)]
net/core: fix kernel-doc warnings

Fix kernel-doc warnings in net/core:

Warning(net/core/skbuff.c:3368): No description found for parameter 'delta_truesize'
Warning(net/core/filter.c:628): No description found for parameter 'pfp'
Warning(net/core/filter.c:628): Excess function parameter 'sk' description in 'sk_unattached_filter_create'

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobe2net: fix a race in be_xmit()
Eric Dumazet [Thu, 7 Jun 2012 22:59:59 +0000 (22:59 +0000)]
be2net: fix a race in be_xmit()

As soon as hardware is notified of a transmit, we no longer can assume
skb can be dereferenced, as TX completion might have freed the packet.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agol2tp: fix a race in l2tp_ip_sendmsg()
Eric Dumazet [Fri, 8 Jun 2012 06:25:00 +0000 (06:25 +0000)]
l2tp: fix a race in l2tp_ip_sendmsg()

Commit 081b1b1bb27f (l2tp: fix l2tp_ip_sendmsg() route handling) added
a race, in case IP route cache is disabled.

In this case, we should not do the dst_release(&rt->dst), since it'll
free the dst immediately, instead of waiting a RCU grace period.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: James Chapman <jchapman@katalix.com>
Cc: Denys Fedoryshchenko <denys@visp.net.lb>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoaf_unix: speedup /proc/net/unix
Eric Dumazet [Fri, 8 Jun 2012 05:03:21 +0000 (05:03 +0000)]
af_unix: speedup /proc/net/unix

/proc/net/unix has quadratic behavior, and can hold unix_table_lock for
a while if high number of unix sockets are alive. (90 ms for 200k
sockets...)

We already have a hash table, so its quite easy to use it.

Problem is unbound sockets are still hashed in a single hash slot
(unix_socket_table[UNIX_HASH_TABLE])

This patch also spreads unbound sockets to 256 hash slots, to speedup
both /proc/net/unix and unix_diag.

Time to read /proc/net/unix with 200k unix sockets :
(time dd if=/proc/net/unix of=/dev/null bs=4k)

before : 520 secs
after : 2 secs

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinetpeer: add parameter net for inet_getpeer_v4,v6
Gao feng [Fri, 8 Jun 2012 01:21:40 +0000 (01:21 +0000)]
inetpeer: add parameter net for inet_getpeer_v4,v6

add struct net as a parameter of inet_getpeer_v[4,6],
use net to replace &init_net.

and modify some places to provide net for inet_getpeer_v[4,6]

Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoinetpeer: add namespace support for inetpeer
Gao feng [Fri, 8 Jun 2012 01:20:41 +0000 (01:20 +0000)]
inetpeer: add namespace support for inetpeer

now inetpeer doesn't support namespace,the information will
be leaking across namespace.

this patch move the global vars v4_peers and v6_peers to
netns_ipv4 and netns_ipv6 as a field peers.

add struct pernet_operations inetpeer_ops to initial pernet
inetpeer data.

and change family_to_base and inet_getpeer to support namespace.

Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agowl18xx: avoid some -Wformat warnings
John W. Linville [Fri, 8 Jun 2012 18:24:53 +0000 (14:24 -0400)]
wl18xx: avoid some -Wformat warnings

  CC      drivers/net/wireless/ti/wl18xx/main.o
drivers/net/wireless/ti/wl18xx/main.c: In function ‘wl18xx_conf_init’:
drivers/net/wireless/ti/wl18xx/main.c:1024:3: warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has type ‘long unsigned int’ [-Wformat]
drivers/net/wireless/ti/wl18xx/main.c:1024:3: warning: format ‘%d’ expects argument of type ‘int’, but argument 3 has type ‘size_t’ [-Wformat]

Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agomac80211: add back channel change flag
Stanislaw Gruszka [Thu, 7 Jun 2012 12:47:21 +0000 (14:47 +0200)]
mac80211: add back channel change flag

commit 24398e39c8ee4a9d9123eed322b859ece4d16cac
Author: Johannes Berg <johannes.berg@intel.com>
Date:   Wed Mar 28 10:58:36 2012 +0200

    mac80211: set HT channel before association

removed IEEE80211_CONF_CHANGE_CHANNEL argument from ieee80211_hw_config,
which is required by iwl4965 driver, otherwise that driver does not
configure channel properly and is not able to associate.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agoNFC: Fix possible NULL ptr deref when getting the name of a socket
Sasha Levin [Wed, 6 Jun 2012 21:02:55 +0000 (23:02 +0200)]
NFC: Fix possible NULL ptr deref when getting the name of a socket

llcp_sock_getname() might get called before the LLCP socket was created.
This condition isn't checked, and llcp_sock_getname will simply deref a
NULL ptr in that case.

This exists starting with d646960 ("NFC: Initial LLCP support").

Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agoiwlwifi: disable the buggy chain extension feature in HW
Emmanuel Grumbach [Wed, 6 Jun 2012 11:55:02 +0000 (13:55 +0200)]
iwlwifi: disable the buggy chain extension feature in HW

This feature has been reported to be buggy and enabled by
default. We therefore need to disable it manually.

Cc: stable@vger.kernel.org
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agoiwlwifi: don't mess up the SCD when removing a key
Emmanuel Grumbach [Wed, 6 Jun 2012 07:13:36 +0000 (09:13 +0200)]
iwlwifi: don't mess up the SCD when removing a key

When we remove a key, we put a key index which was supposed
to tell the fw that we are actually removing the key. But
instead the fw took that index as a valid index and messed
up the SRAM of the device.

This memory corruption on the device mangled the data of
the SCD. The impact on the user is that SCD queue 2 got
stuck after having removed keys.
The message is the log that was printed is:

Queue 2 stuck for 10000ms

This doesn't seem to fix the higher queues that get stuck
from time to time.

Cc: stable@vger.kernel.org [2.6.27+]
Reviewed-by: Meenakshi Venkataraman <meenakshi.venkataraman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agobcma: fix null pointer in bcma_core_pci_irq_ctl
Hauke Mehrtens [Tue, 5 Jun 2012 18:58:20 +0000 (20:58 +0200)]
bcma: fix null pointer in bcma_core_pci_irq_ctl

pc could be null if hosttype != BCMA_HOSTTYPE_PCI.
If we are on a device without a pci core this function is called with
pc = null by b43 and brcmsmac. If the host type is PCI we have a pci
core as well and pc can not be null.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agob43: do not call ieee80211_unregister_hw if we are not registred
Oleksij Rempel [Tue, 5 Jun 2012 18:39:32 +0000 (20:39 +0200)]
b43: do not call ieee80211_unregister_hw if we are not registred

this patch fixes kernel Oops on "rmmod b43" if firmware was not loaded:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000088
IP: [<ffffffff8104e988>] drain_workqueue+0x25/0x142
PGD 153ac6067 PUD 153b82067 PMD 0
Oops: 0000 [#1] SMP

Signed-off-by: Oleksij Rempel <bug-track@fisher-privat.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agoiwlwifi: use correct supported firmware for 6035 and 6000g2
Meenakshi Venkataraman [Tue, 5 Jun 2012 18:24:37 +0000 (20:24 +0200)]
iwlwifi: use correct supported firmware for 6035 and 6000g2

My patch

   iwlwifi: use correct released ucode version

did not correctly report supported firmware
for the 6035 device. This patch fixes it. The
minimum supported firmware version for 6035
is v6.

Also correct the minimum supported firmware
version for the 6000g2 series of devices.

Cc: stable@kernel.org
Signed-off-by: Meenakshi Venkataraman <meenakshi.venkataraman@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agowireless: add my new trees to MAINTAINERS
Johannes Berg [Tue, 5 Jun 2012 13:42:55 +0000 (15:42 +0200)]
wireless: add my new trees to MAINTAINERS

Add my new trees to the MAINTAINERS file
for the components that I maintain in the
new trees.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agomac80211_hwsim: Set IEEE80211_STAT_ACK flag when userspace indicates that the frame...
Qasim Javed [Tue, 5 Jun 2012 06:25:44 +0000 (01:25 -0500)]
mac80211_hwsim: Set IEEE80211_STAT_ACK flag when userspace indicates that the frame has been acknowledged.

The station fail average is not updated correctly since the
IEEE80211_STAT_ACK flag is not set when using wmediumd with
mac80211_hwsim. Set this flag when wmediumd indicates that the frame
was successfully transmitted (eventually).

Signed-off-by: Qasim Javed <qasimj@gmail.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agortl8187: ->brightness_set can not sleep
Stanislaw Gruszka [Wed, 16 May 2012 09:06:21 +0000 (11:06 +0200)]
rtl8187: ->brightness_set can not sleep

Fix:

BUG: sleeping function called from invalid context at kernel/workqueue.c:2547
in_atomic(): 1, irqs_disabled(): 0, pid: 629, name: wpa_supplicant
2 locks held by wpa_supplicant/629:
 #0:  (rtnl_mutex){+.+.+.}, at: [<c08b2b84>] rtnl_lock+0x14/0x20
 #1:  (&trigger->leddev_list_lock){.+.?..}, at: [<c0867f41>] led_trigger_event+0x21/0x80
Pid: 629, comm: wpa_supplicant Not tainted 3.3.0-0.rc3.git5.1.fc17.i686
Call Trace:
 [<c046a9f6>] __might_sleep+0x126/0x1d0
 [<c0457d6c>] wait_on_work+0x2c/0x1d0
 [<c045a09a>] __cancel_work_timer+0x6a/0x120
 [<c045a160>] cancel_delayed_work_sync+0x10/0x20
 [<f7dd3c22>] rtl8187_led_brightness_set+0x82/0xf0 [rtl8187]
 [<c0867f7c>] led_trigger_event+0x5c/0x80
 [<f7ff5e6d>] ieee80211_led_radio+0x1d/0x40 [mac80211]
 [<f7ff3583>] ieee80211_stop_device+0x13/0x230 [mac80211]

Removing _sync is ok, because if led_on work is currently running
it will be finished before led_off work start to perform, since
they are always queued on the same mac80211 local->workqueue.

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=795176

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Acked-by: Hin-Tak Leung <htl10@users.sourceforge.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agonet/wireless: ipw2100: Fix WARN_ON occurring in wiphy_register called by ipw2100_pci_...
Stanislav Yakovlev [Mon, 14 May 2012 23:06:19 +0000 (19:06 -0400)]
net/wireless: ipw2100: Fix WARN_ON occurring in wiphy_register called by ipw2100_pci_init_one

The problem was found by Larry Finger:
http://marc.info/?l=linux-wireless&m=133702401700614&w=2

The problem is identical to the one for ipw2200 which is already fixed:
http://marc.info/?l=linux-wireless&m=133457257407196&w=2

[   17.766431] ------------[ cut here ]------------
[   17.766467] WARNING: at net/wireless/core.c:562 wiphy_register+0x34c/0x3c0 [cfg80211]()
[   17.766471] Hardware name: Latitude D600
[   17.766474] Modules linked in: ipw2100(+) libipw pcmcia cfg80211 ppdev parport_pc yenta_socket sr_mod pcmcia_rsrc parport iTCO_wdt cdrom sg rfkill pcmcia_
core lib80211 tg3 video button battery ac iTCO_vendor_support joydev shpchp pcspkr pciehp pci_hotplug autofs4 radeon ttm drm_kms_helper uhci_hcd ehci_hcd rtc
_cmos thermal drm hwmon i2c_algo_bit i2c_core processor usbcore usb_common ata_generic ata_piix ahci libahci libata
[   17.766525] Pid: 474, comm: modprobe Not tainted 3.4.0-rc7-wl+ #6
[   17.766528] Call Trace:
[   17.766541]  [<c066ad08>] ? printk+0x28/0x2a
[   17.766552]  [<c0230edd>] warn_slowpath_common+0x6d/0xa0
[   17.766563]  [<e0b253bc>] ? wiphy_register+0x34c/0x3c0 [cfg80211]
[   17.766573]  [<e0b253bc>] ? wiphy_register+0x34c/0x3c0 [cfg80211]
[   17.766578]  [<c0230f2d>] warn_slowpath_null+0x1d/0x20
[   17.766588]  [<e0b253bc>] wiphy_register+0x34c/0x3c0 [cfg80211]
[   17.766605]  [<e0b5b0d6>] ipw2100_wdev_init+0x196/0x1c0 [ipw2100]
[   17.766616]  [<e0b5d962>] ipw2100_pci_init_one+0x2b2/0x694 [ipw2100]
[   17.766632]  [<c047ce52>] local_pci_probe+0x42/0xb0
[   17.766637]  [<c047e2b0>] pci_device_probe+0x60/0x90
[   17.766645]  [<c0376de2>] ? sysfs_create_link+0x12/0x20
[   17.766654]  [<c050f1f6>] really_probe+0x56/0x2e0
[   17.766659]  [<c037636d>] ? create_dir+0x5d/0xa0
[   17.766667]  [<c0518c6b>] ? pm_runtime_barrier+0x3b/0xa0
[   17.766672]  [<c050f5e4>] driver_probe_device+0x44/0xa0
[   17.766677]  [<c047e227>] ? pci_match_device+0x97/0xa0
[   17.766681]  [<c050f6c9>] __driver_attach+0x89/0x90
[   17.766686]  [<c050f640>] ? driver_probe_device+0xa0/0xa0
[   17.766691]  [<c050da2a>] bus_for_each_dev+0x3a/0x70
[   17.766695]  [<c050ee6c>] driver_attach+0x1c/0x30
[   17.766699]  [<c050f640>] ? driver_probe_device+0xa0/0xa0
[   17.766704]  [<c050ea77>] bus_add_driver+0x187/0x280
[   17.766710]  [<c045b9cd>] ? kset_find_obj+0x2d/0x60
[   17.766715]  [<c047e2e0>] ? pci_device_probe+0x90/0x90
[   17.766719]  [<c047e2e0>] ? pci_device_probe+0x90/0x90
[   17.766724]  [<c050fb85>] driver_register+0x65/0x110
[   17.766729]  [<c047e09d>] __pci_register_driver+0x3d/0xa0
[   17.766738]  [<e09f705c>] ipw2100_init+0x5c/0x1000 [ipw2100]
[   17.766743]  [<c020110f>] do_one_initcall+0x2f/0x170
[   17.766749]  [<e09f7000>] ? 0xe09f6fff
[   17.766757]  [<c0287ce8>] sys_init_module+0xa8/0x210
[   17.766766]  [<c067a075>] syscall_call+0x7/0xb
[   17.766769] ---[ end trace 559898c6bb0d1c75 ]---
[   17.767093] ipw2100: probe of 0000:02:03.0 failed with error -5

This warning appears only if we apply Ben Hutchings' fix
http://marc.info/?l=linux-wireless&m=132720204412667&w=2
for the bug reported by Cesare Leonardi
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=656813
with cfg80211 warning during device registration
("cfg80211: failed to add phy80211 symlink to netdev!").

We separate device bring up and registration with network stack
to avoid the problem.

Signed-off-by: Stanislav Yakovlev <stas.yakovlev@gmail.com>
Tested-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
12 years agoMerge branch 'wl12xx-next' into for-linville
Luciano Coelho [Fri, 8 Jun 2012 13:43:19 +0000 (16:43 +0300)]
Merge branch 'wl12xx-next' into for-linville

12 years agoRevert "niu: Add support for byte queue limits."
David S. Miller [Fri, 8 Jun 2012 07:28:16 +0000 (00:28 -0700)]
Revert "niu: Add support for byte queue limits."

This reverts commit efa230f2c68abab817f13473077f8d0cc74f43f3.

BQL doesn't work with how this driver currently only takes TX
interrupts every 1/4 of the TX ring.  That behavior needs to be fixed,
but that's a larger non-trivial task and for now we have to revert
BQL support as this makes the device currently completely unusable.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agowlcore/wl12xx/wl18xx: make NVS file optional for wl18xx
Arik Nemtsov [Tue, 29 May 2012 15:38:05 +0000 (18:38 +0300)]
wlcore/wl12xx/wl18xx: make NVS file optional for wl18xx

Don't spew errors when we can't find the NVS file in wlcore. Instead
fail the wl12xx boot HW op if the NVS isn't found.

Take this opportunity to remove some dead code from register_hw()
which looks for the NVS again needlessly.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowl18xx: clean up phy module parameters
Arik Nemtsov [Tue, 29 May 2012 09:44:12 +0000 (12:44 +0300)]
wl18xx: clean up phy module parameters

Give all wl18xx phy module paramters -1 as a default value, indicating
the paramter was not set. Add previous default values to the default
18xx priv conf structure.

Remove the board_type field from wl18xx priv. The field with the same
name inside the phy conf is good enough for our purposes.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowl18xx: align wl18xx_conf_phy with FW variant and remove it
Arik Nemtsov [Tue, 29 May 2012 09:40:50 +0000 (12:40 +0300)]
wl18xx: align wl18xx_conf_phy with FW variant and remove it

wl18xx_conf_phy represents part of the FW native wl18xx_mac_and_phy_params
structure. Remove it and replace the phy part of the wl18xx conf with the
FW bound structure. This allows us to set/override all members.

Increment the wlconf version to ensure compatibility with the new
structure

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowl18xx: update fw statistics
Ido Reis [Tue, 22 May 2012 09:34:10 +0000 (12:34 +0300)]
wl18xx: update fw statistics

Aligned to the struct in FW 8.2.0.0.91 and updated the debugfs entries
accordingly.

Signed-off-by: Ido Reis <idor@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowl18xx: allow FW-log by default for PG2.0
Arik Nemtsov [Wed, 23 May 2012 05:39:43 +0000 (08:39 +0300)]
wl18xx: allow FW-log by default for PG2.0

This is supported by new FW versions (.88+).

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowl18xx: read configuration structure from a binary file
Luciano Coelho [Thu, 7 Jun 2012 20:39:28 +0000 (23:39 +0300)]
wl18xx: read configuration structure from a binary file

Instead of using the hardcoded configuration structure, try to read it
from a "firmware" file called wl18xx-conf.bin.  If the file doesn't
exist, fall back to the hardcoded version.  If the file exists but is
illegal, bail out.

Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowlcore/wl18xx: export conf struct in a debugfs file
Luciano Coelho [Thu, 7 Jun 2012 20:39:27 +0000 (23:39 +0300)]
wlcore/wl18xx: export conf struct in a debugfs file

Add conf file header structure, magic and version values and export
the entire conf struct in debugfs.

Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowlcore/wl18xx: the conf structs must be packed so they can be exported
Luciano Coelho [Thu, 7 Jun 2012 20:39:26 +0000 (23:39 +0300)]
wlcore/wl18xx: the conf structs must be packed so they can be exported

Since we are now going to export the conf structure and read it from a
file, it should be packed to avoid surprises with padding bytes.

Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowlcore: use u8 instead of enum for bcn_filt_mode
Luciano Coelho [Thu, 7 Jun 2012 20:39:25 +0000 (23:39 +0300)]
wlcore: use u8 instead of enum for bcn_filt_mode

Since we will export the conf structure as a file, we need to use well
defined types.  Instead of using enum, whose size may vary, use u8 for
bcn_filt_mode instead.

Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowlcore/wl18xx/wl12xx: use u8 instead of bool for host_fast_wakeup_support
Luciano Coelho [Thu, 7 Jun 2012 20:39:24 +0000 (23:39 +0300)]
wlcore/wl18xx/wl12xx: use u8 instead of bool for host_fast_wakeup_support

The conf structure is going to be exported to a file, so we should use
only well defined types.  bool is not well defined and may vary from
platform to platform, so change the host_fast_wakeup_support type to
u8 instead.

Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowlcore: export raw binary with the FW statistics in debugfs
Luciano Coelho [Thu, 7 Jun 2012 20:39:23 +0000 (23:39 +0300)]
wlcore: export raw binary with the FW statistics in debugfs

Instead of parsing all the binary data returned by the firmware, we
should simply export the binary and let the userspace do the parsing.

This commit adds a new file to debugfs to do that.

Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowl18xx: add support to clear FW statistics
Luciano Coelho [Thu, 7 Jun 2012 20:39:22 +0000 (23:39 +0300)]
wl18xx: add support to clear FW statistics

This patch calls ACX_CLEAR_STATISTICS to clear the firmware
statistics.  The trigger is a new debugfs file called
clear_fw_statistics in the fw_stats directory.

Signed-off-by: Luciano Coelho <coelho@ti.com>
12 years agowlcore: add debugfs control over rx interrupt pacing
Eyal Shapira [Thu, 7 Jun 2012 20:39:21 +0000 (23:39 +0300)]
wlcore: add debugfs control over rx interrupt pacing

Add control over several conf fields which combined
control the rx interrupt pacing mechanism, that is avoiding
getting an interrupt following a single frame rx but instead
have the FW trigger the interrupt only after a certain
amount of frames received or a timeout.

Signed-off-by: Eyal Shapira <eyal@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>