platform/kernel/kernel-mfld-blackbay.git
13 years agophy: add the IC+ IP1001 driver
Giuseppe CAVALLARO [Wed, 8 Dec 2010 23:05:13 +0000 (23:05 +0000)]
phy: add the IC+ IP1001 driver

This patch adds the IC+ IP1001 (Gigabit Ethernet Transceiver) driver.
I've had to add an additional delay (2ns) to adjust RX clock phase at
GMII/ RGMII interface (according to the PHY data-sheet). This helps to
have the RGMII working on some ST platforms.

Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoatm: correct sysfs 'device' link creation and parent relationships
Dan Williams [Wed, 8 Dec 2010 19:40:47 +0000 (19:40 +0000)]
atm: correct sysfs 'device' link creation and parent relationships

The ATM subsystem was incorrectly creating the 'device' link for ATM
nodes in sysfs.  This led to incorrect device/parent relationships
exposed by sysfs and udev.  Instead of rolling the 'device' link by hand
in the generic ATM code, pass each ATM driver's bus device down to the
sysfs code and let sysfs do this stuff correctly.

Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMAINTAINERS: remove me from tulip
Kyle McMartin [Thu, 9 Dec 2010 09:50:08 +0000 (09:50 +0000)]
MAINTAINERS: remove me from tulip

It was a nice idea, but -ENOTIME and -ENOHW. I never got around to doing
a lot of the clean up that I intended to.

Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
Acked-by: Grant Grundler <grundler@parisc-linux.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoSCTP: Fix SCTP_SET_PEER_PRIMARY_ADDR to accpet v4mapped address
Wei Yongjun [Tue, 7 Dec 2010 17:11:09 +0000 (17:11 +0000)]
SCTP: Fix SCTP_SET_PEER_PRIMARY_ADDR to accpet v4mapped address

SCTP_SET_PEER_PRIMARY_ADDR does not accpet v4mapped address, using
v4mapped address in SCTP_SET_PEER_PRIMARY_ADDR socket option will
get -EADDRNOTAVAIL error if v4map is enabled. This patch try to
fix it by mapping v4mapped address to v4 address if allowed.

Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Acked-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoenic: Bug Fix: Pass napi reference to the isr that services receive queue
Vasanthy Kolluri [Wed, 8 Dec 2010 13:05:45 +0000 (13:05 +0000)]
enic: Bug Fix: Pass napi reference to the isr that services receive queue

Pass reference to napi instead of enic device to the isr that services receive queue.

Signed-off-by: Vasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com>
Signed-off-by: David Wang <dwang2@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Fri, 10 Dec 2010 20:51:02 +0000 (12:51 -0800)]
Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-2.6

13 years agoipv6: fix nl group when advertising a new link
Nicolas Dichtel [Tue, 7 Dec 2010 23:38:31 +0000 (23:38 +0000)]
ipv6: fix nl group when advertising a new link

New idev are advertised with NL group RTNLGRP_IPV6_IFADDR, but
should use RTNLGRP_IPV6_IFINFO.
Bug was introduced by commit 8d7a76c9.

Signed-off-by: Wang Xuefu <xuefu.wang@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Thomas Graf <tgraf@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoconnector: add module alias
Stephen Hemminger [Fri, 10 Dec 2010 20:27:49 +0000 (12:27 -0800)]
connector: add module alias

Since connector can be built as a module and uses netlink socket
to communicate. The module should have an alias to autoload when socket
of NETLINK_CONNECTOR type is requested.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Acked-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet: Document the kernel_recvmsg() function
Martin Lucina [Fri, 10 Dec 2010 00:04:05 +0000 (00:04 +0000)]
net: Document the kernel_recvmsg() function

Signed-off-by: Martin Lucina <mato@kotelna.sk>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agor8169: Fix runtime power management
Rafael J. Wysocki [Wed, 8 Dec 2010 15:32:14 +0000 (15:32 +0000)]
r8169: Fix runtime power management

I noticed that one of the post-2.6.36 patches broke runtime PM of the
r8169 on my MSI Wind test machine in such a way that the link was not
brought up after reconnecting the network cable.

In the process of debugging the issue I realized that we only should
invoke the runtime PM functions in rtl8169_check_link_status() when
link change is reported and if we do so, the problem goes away.
Moreover, this allows rtl8169_runtime_idle() to be simplified quite
a bit.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agohso: IP checksuming doesn't work on GE0301 option cards
Thomas Bogendoerfer [Tue, 7 Dec 2010 22:21:52 +0000 (22:21 +0000)]
hso: IP checksuming doesn't work on GE0301 option cards

There is definitly a problem, that some option cards send up broken
IP pakets leading to corrupted IP packets. These corruptions aren't
detected, because the driver claims that the packets are already
checksummed. This change removes the CHECKSUM_UNNECESSARY option
and let IP detect broken data.

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoxfrm: Fix xfrm_state_migrate leak
Thomas Egerer [Mon, 6 Dec 2010 23:28:56 +0000 (23:28 +0000)]
xfrm: Fix xfrm_state_migrate leak

xfrm_state_migrate calls kfree instead of xfrm_state_put to free
a failed state. According to git commit 553f9118 this can cause
memory leaks.

Signed-off-by: Thomas Egerer <thomas.egerer@secunet.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet: Convert netpoll blocking api in bonding driver to be a counter
Neil Horman [Mon, 6 Dec 2010 09:05:50 +0000 (09:05 +0000)]
net: Convert netpoll blocking api in bonding driver to be a counter

A while back I made some changes to enable netpoll in the bonding driver.  Among
them was a per-cpu flag that indicated we were in a path that held locks which
could cause the netpoll path to block in during tx, and as such the tx path
should queue the frame for later use.  This appears to have given rise to a
regression.  If one of those paths on which we hold the per-cpu flag yields the
cpu, its possible for us to come back on a different cpu, leading to us clearing
a different flag than we set.  This results in odd netpoll drops, and BUG
backtraces appearing in the log, as we check to make sure that we only clear set
bits, and only set clear bits.  I had though briefly about changing the
offending paths so that they wouldn't sleep, but looking at my origional work
more closely, it doesn't appear that a per-cpu flag is warranted.  We alrady
gate the checking of this flag on IFF_IN_NETPOLL, so we don't hit this in the
normal tx case anyway.  And practically speaking, the normal use case for
netpoll is to only have one client anyway, so we're not going to erroneously
queue netpoll frames when its actually safe to do so.  As such, lets just
convert that per-cpu flag to an atomic counter.  It fixes the rescheduling bugs,
is equivalent from a performance perspective and actually eliminates some code
in the process.

Tested by the reporter and myself, successfully

Reported-by: Liang Zheng <lzheng@redhat.com>
CC: Jay Vosburgh <fubar@us.ibm.com>
CC: Andy Gospodarek <andy@greyhouse.net>
CC: David S. Miller <davem@davemloft.net>
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoeconet: Fix crash in aun_incoming().
David S. Miller [Thu, 9 Dec 2010 02:42:23 +0000 (18:42 -0800)]
econet: Fix crash in aun_incoming().

Unconditional use of skb->dev won't work here,
try to fetch the econet device via skb_dst()->dev
instead.

Suggested by Eric Dumazet.

Reported-by: Nelson Elhage <nelhage@ksplice.com>
Tested-by: Nelson Elhage <nelhage@ksplice.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agotcp: protect sysctl_tcp_cookie_size reads
Eric Dumazet [Tue, 7 Dec 2010 12:20:47 +0000 (12:20 +0000)]
tcp: protect sysctl_tcp_cookie_size reads

Make sure sysctl_tcp_cookie_size is read once in
tcp_cookie_size_check(), or we might return an illegal value to caller
if sysctl_tcp_cookie_size is changed by another cpu.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: William Allen Simpson <william.allen.simpson@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agotcp: avoid a possible divide by zero
Eric Dumazet [Tue, 7 Dec 2010 12:03:55 +0000 (12:03 +0000)]
tcp: avoid a possible divide by zero

sysctl_tcp_tso_win_divisor might be set to zero while one cpu runs in
tcp_tso_should_defer(). Make sure we dont allow a divide by zero by
reading sysctl_tcp_tso_win_divisor exactly once.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoorinoco: fix TKIP countermeasure behaviour
David Kilroy [Sun, 5 Dec 2010 15:43:55 +0000 (15:43 +0000)]
orinoco: fix TKIP countermeasure behaviour

Enable the port when disabling countermeasures, and disable it on
enabling countermeasures.

This bug causes the response of the system to certain attacks to be
ineffective.

It also prevents wpa_supplicant from getting scan results, as
wpa_supplicant disables countermeasures on startup - preventing the
hardware from scanning.

wpa_supplicant works with ap_mode=2 despite this bug because the commit
handler re-enables the port.

The log tends to look like:

State: DISCONNECTED -> SCANNING
Starting AP scan for wildcard SSID
Scan requested (ret=0) - scan timeout 5 seconds
EAPOL: disable timer tick
EAPOL: Supplicant port status: Unauthorized
Scan timeout - try to get results
Failed to get scan results
Failed to get scan results - try scanning again
Setting scan request: 1 sec 0 usec
Starting AP scan for wildcard SSID
Scan requested (ret=-1) - scan timeout 5 seconds
Failed to initiate AP scan.

Reported by: Giacomo Comes <comes@naic.edu>
Signed-off by: David Kilroy <kilroyd@googlemail.com>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoorinoco: clear countermeasure setting on commit
David Kilroy [Sun, 5 Dec 2010 15:45:58 +0000 (15:45 +0000)]
orinoco: clear countermeasure setting on commit

... and interface up.

In these situations, you are usually trying to connect to a new AP, so
keeping TKIP countermeasures active is confusing. This is already how
the driver behaves (inadvertently). However, querying SIOCGIWAUTH may
tell userspace that countermeasures are active when they aren't.

Clear the setting so that the reporting matches what the driver has
done..

Signed-off by: David Kilroy <kilroyd@googlemail.com>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Fix BUG in pskb_expand_head when transmitting shared skbs
Helmut Schaa [Thu, 2 Dec 2010 17:44:09 +0000 (18:44 +0100)]
mac80211: Fix BUG in pskb_expand_head when transmitting shared skbs

mac80211 doesn't handle shared skbs correctly at the moment. As a result
a possible resize can trigger a BUG in pskb_expand_head.

[  676.030000] Kernel bug detected[#1]:
[  676.030000] Cpu 0
[  676.030000] $ 0   : 00000000 00000000 819662ff 00000002
[  676.030000] $ 4   : 81966200 00000020 00000000 00000020
[  676.030000] $ 8   : 819662e0 800043c0 00000002 00020000
[  676.030000] $12   : 3b9aca00 00000000 00000000 00470000
[  676.030000] $16   : 80ea2000 00000000 00000000 00000000
[  676.030000] $20   : 818aa200 80ea2018 80ea2000 00000008
[  676.030000] $24   : 00000002 800ace5c
[  676.030000] $28   : 8199a000 8199bd20 81938f88 80f180d4
[  676.030000] Hi    : 0000026e
[  676.030000] Lo    : 0000757e
[  676.030000] epc   : 801245e4 pskb_expand_head+0x44/0x1d8
[  676.030000]     Not tainted
[  676.030000] ra    : 80f180d4 ieee80211_skb_resize+0xb0/0x114 [mac80211]
[  676.030000] Status: 1000a403    KERNEL EXL IE
[  676.030000] Cause : 10800024
[  676.030000] PrId  : 0001964c (MIPS 24Kc)
[  676.030000] Modules linked in: mac80211_hwsim rt2800lib rt2x00soc rt2x00pci rt2x00lib mac80211 crc_itu_t crc_ccitt cfg80211 compat arc4 aes_generic deflate ecb cbc [last unloaded: rt2800pci]
[  676.030000] Process kpktgend_0 (pid: 97, threadinfo=8199a000, task=81879f48, tls=00000000)
[  676.030000] Stack : ffffffff 00000000 00000000 00000014 00000004 80ea2000 00000000 00000000
[  676.030000]         818aa200 80f180d4 ffffffff 0000000a 81879f78 81879f48 81879f48 00000018
[  676.030000]         81966246 80ea2000 818432e0 80f1a420 80203050 81814d98 00000001 81879f48
[  676.030000]         81879f48 00000018 81966246 818432e0 0000001a 8199bdd4 0000001c 80f1b72c
[  676.030000]         80203020 8001292c 80ef4aa2 7f10b55d 801ab5b8 81879f48 00000188 80005c90
[  676.030000]         ...
[  676.030000] Call Trace:
[  676.030000] [<801245e4>] pskb_expand_head+0x44/0x1d8
[  676.030000] [<80f180d4>] ieee80211_skb_resize+0xb0/0x114 [mac80211]
[  676.030000] [<80f1a420>] ieee80211_xmit+0x150/0x22c [mac80211]
[  676.030000] [<80f1b72c>] ieee80211_subif_start_xmit+0x6f4/0x73c [mac80211]
[  676.030000] [<8014361c>] pktgen_thread_worker+0xfac/0x16f8
[  676.030000] [<8002ebe8>] kthread+0x7c/0x88
[  676.030000] [<80008e0c>] kernel_thread_helper+0x10/0x18
[  676.030000]
[  676.030000]
[  676.030000] Code: 24020001  10620005  2502001f <0200000d0804917a  00000000  2502001f  00441023  00531021

Fix this by making a local copy of shared skbs prior to mangeling them.
To avoid copying the skb unnecessarily move the skb_copy call below the
checks that don't need write access to the skb.

Also, move the assignment of nh_pos and h_pos below the skb_copy to point
to the correct skb.

It would be possible to avoid another resize of the copied skb by using
skb_copy_expand instead of skb_copy but that would make the patch more
complex. Also, shared skbs are a corner case right now, so the resize
shouldn't matter much.

Cc: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_htc: Fix suspend/resume
Sujith Manoharan [Wed, 1 Dec 2010 07:00:27 +0000 (12:30 +0530)]
ath9k_htc: Fix suspend/resume

The HW has to be set to FULLSLEEP mode during suspend,
when no interface has been brought up. Not doing this would
break resume, as the chip won't be powered up at all.

Signed-off-by: Sujith Manoharan <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: Put the right tsf value in mesh beacons
Javier Cardona [Tue, 7 Dec 2010 21:37:56 +0000 (13:37 -0800)]
ath5k: Put the right tsf value in mesh beacons

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: Prevent mesh interfaces from being counted as ad-hoc
Javier Cardona [Tue, 7 Dec 2010 21:36:55 +0000 (13:36 -0800)]
ath5k: Prevent mesh interfaces from being counted as ad-hoc

This results in an erroneus num_adhoc_vifs count, as the this counter
was incremented but not decremented for mesh interfaces.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: Fix beaconing in mesh mode
Javier Cardona [Tue, 7 Dec 2010 21:35:55 +0000 (13:35 -0800)]
ath5k: Fix beaconing in mesh mode

This patch fixes the oops below when attempting to bring up a mesh
interface on ath5k hardware.

[  128.933099] kernel BUG at drivers/net/wireless/ath/ath5k/base.c:197!
[  128.933099] invalid opcode: 0000 [#1]
(...)
[  128.933099] Call Trace:
[  128.933099]  [<c83b77fa>] ? ath5k_beacon_update+0x57/0x1f8 [ath5k]
[  128.933099]  [<c02d9a40>] ? __sysfs_add_one+0x28/0x76
[  128.933099]  [<c83b830e>] ? ath5k_bss_info_changed+0x13f/0x173
[ath5k]
[  128.933099]  [<c82ff629>] ? ieee80211_config_beacon+0xc0/0x17e
[mac80211]
[  128.933099]  [<c82f073e>] ?
ieee80211_bss_info_change_notify+0x182/0x18b [mac80211]
[  128.933099]  [<c83b81cf>] ? ath5k_bss_info_changed+0x0/0x173 [ath5k]
[  128.933099]  [<c82ff6d6>] ? ieee80211_config_beacon+0x16d/0x17e
[mac80211]
[  128.933099]  [<c82ff753>] ? ieee80211_add_beacon+0x34/0x39 [mac80211]
[  128.933099]  [<c830a4ed>] ? ieee80211s_init+0xf8/0x10f [mac80211]
[  128.933099]  [<c830a5df>] ? ieee80211_mesh_init_sdata+0xdb/0x154 [mac80211]

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoorinoco: initialise priv->hw before assigning the interrupt
David Kilroy [Tue, 7 Dec 2010 18:50:42 +0000 (18:50 +0000)]
orinoco: initialise priv->hw before assigning the interrupt

The interrupt handler takes a lock - but since commit bcad6e80f3f this
lock goes through an indirection specified in the hermes_t structure.
We must therefore initialise the structure before setting up the
interrupt handler.

Fix orinoco_cs and spectrum_cs

<https://bugzilla.kernel.org/show_bug.cgi?id=23932>

Bisected by: Matt Domsch <Matt_Domsch@dell.com>
Signed-off by: David Kilroy <kilroyd@googlemail.com>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoehea: Fixing LRO configuration
Breno Leitao [Wed, 8 Dec 2010 20:19:14 +0000 (12:19 -0800)]
ehea: Fixing LRO configuration

In order to set LRO on ehea, the user must set a module parameter, which
is not the standard way to do so. This patch adds a way to set LRO using
the ethtool tool.

Signed-off-by: Breno Leitao <leitao@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agotcp: Replace time wait bucket msg by counter
Tom Herbert [Wed, 8 Dec 2010 20:16:33 +0000 (12:16 -0800)]
tcp: Replace time wait bucket msg by counter

Rather than printing the message to the log, use a mib counter to keep
track of the count of occurences of time wait bucket overflow.  Reduces
spam in logs.

Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agox25: decrement netdev reference counts on unload
Apollon Oikonomopoulos [Tue, 7 Dec 2010 09:43:30 +0000 (09:43 +0000)]
x25: decrement netdev reference counts on unload

x25 does not decrement the network device reference counts on module unload.
Thus unregistering any pre-existing interface after unloading the x25 module
hangs and results in

 unregister_netdevice: waiting for tap0 to become free. Usage count = 1

This patch decrements the reference counts of all interfaces in x25_link_free,
the way it is already done in x25_link_device_down for NETDEV_DOWN events.

Signed-off-by: Apollon Oikonomopoulos <apollon@noc.grnet.gr>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agodriver/net/benet: fix be_cmd_multicast_set() memcpy bug
Joe Jin [Mon, 6 Dec 2010 03:00:59 +0000 (03:00 +0000)]
driver/net/benet: fix be_cmd_multicast_set() memcpy bug

Regarding  benet be_cmd_multicast_set() function, now using
netdev_for_each_mc_addr() helper for mac address copy, but
when copying to req->mac[] did not increase of the index.

Cc: Sathya Perla <sathyap@serverengines.com>
Cc: Subbu Seetharaman <subbus@serverengines.com>
Cc: Sarveshwar Bandi <sarveshwarb@serverengines.com>
Cc: Ajit Khaparde <ajitk@serverengines.com>
Signed-off-by: Joe Jin <joe.jin@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agol2tp: Fix modalias of l2tp_ip
Michal Marek [Mon, 6 Dec 2010 02:39:12 +0000 (02:39 +0000)]
l2tp: Fix modalias of l2tp_ip

Using the SOCK_DGRAM enum results in
"net-pf-2-proto-SOCK_DGRAM-type-115", so use the numeric value like it
is done in net/dccp.

Signed-off-by: Michal Marek <mmarek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoeconet: Do the correct cleanup after an unprivileged SIOCSIFADDR.
Nelson Elhage [Wed, 8 Dec 2010 18:13:55 +0000 (10:13 -0800)]
econet: Do the correct cleanup after an unprivileged SIOCSIFADDR.

We need to drop the mutex and do a dev_put, so set an error code and break like
the other paths, instead of returning directly.

Signed-off-by: Nelson Elhage <nelhage@ksplice.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMerge branch 'sfc-2.6.37' of git://git.kernel.org/pub/scm/linux/kernel/git/bwh/sfc-2.6
David S. Miller [Wed, 8 Dec 2010 20:13:23 +0000 (12:13 -0800)]
Merge branch 'sfc-2.6.37' of git://git./linux/kernel/git/bwh/sfc-2.6

13 years agollc: fix a device refcount imbalance
Eric Dumazet [Sun, 5 Dec 2010 02:03:26 +0000 (02:03 +0000)]
llc: fix a device refcount imbalance

Le dimanche 05 décembre 2010 Ã  12:23 +0100, Eric Dumazet a Ã©crit :
> Le dimanche 05 décembre 2010 Ã  09:19 +0100, Eric Dumazet a Ã©crit :
>
> > Hmm..
> >
> > If somebody can explain why RTNL is held in arp_ioctl() (and therefore
> > in arp_req_delete()), we might first remove RTNL use in arp_ioctl() so
> > that your patch can be applied.
> >
> > Right now it is not good, because RTNL wont be necessarly held when you
> > are going to call arp_invalidate() ?
>
> While doing this analysis, I found a refcount bug in llc, I'll send a
> patch for net-2.6

Oh well, of course I must first fix the bug in net-2.6, and wait David
pull the fix in net-next-2.6 before sending this rcu conversion.

Note: this patch should be sent to stable teams (2.6.34 and up)

[PATCH net-2.6] llc: fix a device refcount imbalance

commit abf9d537fea225 (llc: add support for SO_BINDTODEVICE) added one
refcount imbalance in llc_ui_bind(), because dev_getbyhwaddr() doesnt
take a reference on device, while dev_get_by_index() does.

Fix this using RCU locking. And since an RCU conversion will be done for
2.6.38 for dev_getbyhwaddr(), put the rcu_read_lock/unlock exactly at
their final place.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: stable@kernel.org
Cc: Octavian Purdila <opurdila@ixiacom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoifb: goto resched directly if error happens and dp->tq isn't empty
Changli Gao [Sat, 4 Dec 2010 14:09:08 +0000 (14:09 +0000)]
ifb: goto resched directly if error happens and dp->tq isn't empty

If we break the loop when there are still skbs in tq and no skb in
rq, the skbs will be left in txq until new skbs are enqueued into rq.
In rare cases, no new skb is queued, then these skbs will stay in rq
forever.

After this patch, if tq isn't empty when we break the loop, we goto
resched directly.

Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agotcp: Bug fix in initialization of receive window.
Nandita Dukkipati [Fri, 3 Dec 2010 13:33:44 +0000 (13:33 +0000)]
tcp: Bug fix in initialization of receive window.

The bug has to do with boundary checks on the initial receive window.
If the initial receive window falls between init_cwnd and the
receive window specified by the user, the initial window is incorrectly
brought down to init_cwnd. The correct behavior is to allow it to
remain unchanged.

Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agocxgb4: fix MAC address hash filter
Dimitris Michailidis [Fri, 3 Dec 2010 10:39:04 +0000 (10:39 +0000)]
cxgb4: fix MAC address hash filter

Fix the calculation of the inexact hash-based MAC address filter.
It's 64 bits but current code is missing a ULL.  Results in filtering out
some legitimate packets.

Signed-off-by: Dimitris Michailidis <dm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoCAIF: Fix U5500 compile error for shared memory driver
Kim Lilliestierna XX [Tue, 30 Nov 2010 09:11:22 +0000 (09:11 +0000)]
CAIF: Fix U5500 compile error for shared memory driver

Rearrange pr_fmt so it compiles.

Signed-off-by: Sjur Braendeland <sjur.brandeland@stericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Wed, 8 Dec 2010 16:13:01 +0000 (08:13 -0800)]
Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-2.6

13 years agosfc: Fix NAPI list corruption during ring reallocation
Ben Hutchings [Tue, 7 Dec 2010 19:47:34 +0000 (19:47 +0000)]
sfc: Fix NAPI list corruption during ring reallocation

Call netif_napi_{add,del}() on the NAPI contexts in the new and
old channels, respectively.

Since efx_init_napi() cannot fail, make its return type void.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
13 years agoath9k: fix beacon resource related race condition
Rajkumar Manoharan [Tue, 7 Dec 2010 12:12:18 +0000 (17:42 +0530)]
ath9k: fix beacon resource related race condition

The beacon tasklet is accesssing the bslot info for beacon generation.
Meanwhile the same slot can be freed on interface deletion.
Current the remove_interface disables the beacon alert after freeing the slot.
This may leads to null pointer access.

This patch disables SWBA and kills the beacon tasklet to prevent access
to the slot to be freed. After releasing the slot, swba will be enabled again
upon the availablity of beaconing interfaces.

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Fix STA disconnect issue due to received MIC failed bcast frames
Senthil Balasubramanian [Mon, 6 Dec 2010 13:39:27 +0000 (19:09 +0530)]
ath9k: Fix STA disconnect issue due to received MIC failed bcast frames

AR_RxKeyIdxValid will not be set for bcast/mcast frames and so relying
this status for MIC failed frames is buggy.

Due to this, MIC failure events for broadcast frames are not sent to
supplicant resulted in AP disconnecting the STA.

Able to pass Wifi Test case 5.2.18 with this fix.

Cc: Stable <stable@kernel.org> (2.6.36+)
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: fix a DMA related race condition on reset
Felix Fietkau [Sun, 5 Dec 2010 19:17:53 +0000 (20:17 +0100)]
ath9k: fix a DMA related race condition on reset

When ath_drain_all_txq fails to stop DMA, it issues a hw reset. This reset
happens at a very problematic point in time, when the hardware rx path has
not been stopped yet. This could lead to memory corruption, hardware hangs
or other issues.
To fix these issues, simply remove the reset entirely and check the tx DMA
stop status to prevent problems with fast channel changes.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: fix bug in tx power
Matteo Croce [Fri, 3 Dec 2010 01:25:08 +0000 (02:25 +0100)]
ath9k: fix bug in tx power

The ath9k driver subtracts 3 dBm to the txpower as with two radios the
signal power is doubled.
The resulting value is assigned in an u16 which overflows and makes
the card work at full power.

Cc: stable@kernel.org
Signed-off-by: Matteo Croce <matteo@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agosfc: Fix crash in legacy onterrupt handler during ring reallocation
Ben Hutchings [Tue, 7 Dec 2010 19:24:45 +0000 (19:24 +0000)]
sfc: Fix crash in legacy onterrupt handler during ring reallocation

If we are using a legacy interrupt, our IRQ may be shared and our
interrupt handler may be called even though interrupts are disabled on
the NIC. When we change ring sizes, we reallocate the event queue and
the interrupt handler may use an invalid pointer when called for
another device's interrupt.

Maintain a legacy_irq_enabled flag and test that at the top of the
interrupt handler.  Note that this problem results from the need to
work around broken INT_ISR0 reads, and does not affect the legacy
interrupt handler for Falcon A1.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
13 years agoixgbe: fix possible NULL pointer deference in shutdown path
Don Skidmore [Wed, 1 Dec 2010 20:54:53 +0000 (20:54 +0000)]
ixgbe: fix possible NULL pointer deference in shutdown path

After freeing the rings we were not zeroing out the ring count values.
This patch now clears these counts correctly.

Reported-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agofilter: fix sk_filter rcu handling
Eric Dumazet [Mon, 6 Dec 2010 17:29:43 +0000 (09:29 -0800)]
filter: fix sk_filter rcu handling

Pavel Emelyanov tried to fix a race between sk_filter_(de|at)tach and
sk_clone() in commit 47e958eac280c263397

Problem is we can have several clones sharing a common sk_filter, and
these clones might want to sk_filter_attach() their own filters at the
same time, and can overwrite old_filter->rcu, corrupting RCU queues.

We can not use filter->rcu without being sure no other thread could do
the same thing.

Switch code to a more conventional ref-counting technique : Do the
atomic decrement immediately and queue one rcu call back when last
reference is released.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agotulip: fix hang in dmfe driver on sending of big packet
Alexander V. Lukyanov [Tue, 30 Nov 2010 03:57:39 +0000 (03:57 +0000)]
tulip: fix hang in dmfe driver on sending of big packet

This patch fixes hang in dmfe driver on attempt of sending a big packet.
Without this patch the code stops the queue and never wakes it again.

Signed-off-by: Alexander V. Lukyanov <lav@netis.ru>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agostmmac: priv->lock can be used uninitialized
Vlad Lungu [Mon, 29 Nov 2010 22:52:52 +0000 (22:52 +0000)]
stmmac: priv->lock can be used uninitialized

To reproduce: if connman (http://connman.net/) is started,
inserting the stmmac module triggers a "BUG: spinlock bad magic on CPU#0".

Registering the device in stmmac_probe() sends a notification to connman
which brings the interface up before the lock is initialized.

Signed-off-by: Vlad Lungu <vlad.lungu@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoRevert "ath9k: Fix STA disconnect issue due to received MIC failed bcast frames"
John W. Linville [Thu, 2 Dec 2010 19:10:58 +0000 (14:10 -0500)]
Revert "ath9k: Fix STA disconnect issue due to received MIC failed bcast frames"

This reverts commit 916448e77f6bcaaa7f13c3de0c3851783ae2bfd0.

"As far as I can tell, either of these patches breaks multiple VIF
scenarios.  I'm not sure exactly why, but I had to revert this to
get any of my interfaces to associate."
  -- Ben Greear <greearb@candelatech.com>

http://marc.info/?l=linux-wireless&m=129123368719339&w=2

Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: fix more bitfield related endian issues
Felix Fietkau [Wed, 1 Dec 2010 18:07:47 +0000 (19:07 +0100)]
ath9k_hw: fix more bitfield related endian issues

A few LNA control related flags were also specified as a bitfields, however
for some strange reason they were written in big-endian order this time.
Fix this by using flags instead.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: fix endian issues with CTLs on AR9003
Felix Fietkau [Wed, 1 Dec 2010 18:07:46 +0000 (19:07 +0100)]
ath9k_hw: fix endian issues with CTLs on AR9003

Parsing data using bitfields is messy, because it makes endian handling
much harder. AR9002 and earlier got it right, AR9003 got it wrong.
This might lead to either using too high or too low tx power values,
depending on frequency and eeprom settings.
Fix it by getting rid of the CTL related bitfields entirely and use
masks instead.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Fix bug in reading input gpio state for ar9003
Vasanthakumar Thiagarajan [Wed, 1 Dec 2010 07:24:09 +0000 (23:24 -0800)]
ath9k: Fix bug in reading input gpio state for ar9003

The register which gives input gpio state is 0x404c for ar9003,
currently 0x4048 is wrongly used. This will disable RF and make
it unusable on some of AR9003.

Cc:stable@kernel.org
Signed-off-by: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/padovan/blueto...
John W. Linville [Thu, 2 Dec 2010 19:00:51 +0000 (14:00 -0500)]
Merge branch 'master' of git://git./linux/kernel/git/padovan/bluetooth-2.6

13 years agonet/ipv6/sit.c: return unhandled skb to tunnel4_rcv
David McCullough [Mon, 29 Nov 2010 19:32:34 +0000 (19:32 +0000)]
net/ipv6/sit.c: return unhandled skb to tunnel4_rcv

I found a problem using an IPv6 over IPv4 tunnel.  When CONFIG_IPV6_SIT
was enabled, the packets would be rejected as net/ipv6/sit.c was catching
all IPPROTO_IPV6 packets and returning an ICMP port unreachable error.

I think this patch fixes the problem cleanly.  I believe the code in
net/ipv4/tunnel4.c:tunnel4_rcv takes care of it properly if none of the
handlers claim the skb.

Signed-off-by: David McCullough <david_mccullough@mcafee.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agobonding: check for assigned mac before adopting the slaves mac address
David Strand [Wed, 1 Dec 2010 19:43:08 +0000 (11:43 -0800)]
bonding: check for assigned mac before adopting the slaves mac address

Restore the check for an unassigned mac address before adopting the
first slaves as it's own. The change in behavior was introduced by:

commit c20811a79e671a6a1fe86a8c1afe04aca8a7f085
Author: Jiri Pirko <jpirko@redhat.com>

    bonding: move dev_addr cpy to bond_enslave

Signed-off-by: David Strand <dpstrand@gmail.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agob44: fix workarround for wap54g10
Hauke Mehrtens [Sat, 27 Nov 2010 06:47:43 +0000 (06:47 +0000)]
b44: fix workarround for wap54g10

The code for the b44_wap54g10_workaround was never included, because
the config option was wrong. The nvram_get function was never in
mainline kernel, only in external OpenWrt patches.

The code should be compiled in when CONFIG_BCM47XX is selected and not
when CONFIG_SSB_DRIVER_MIPS is selected, because nvram_getenv is only
available on bcm47xx platforms and now in the mainline kernel code.
Using an include is better than a second function declaration, to fix
this when the function signature changes.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMake the ip6_tunnel reflect the true mtu.
Anders Franzen [Wed, 24 Nov 2010 05:47:18 +0000 (05:47 +0000)]
Make the ip6_tunnel reflect the true mtu.

The ip6_tunnel always assumes it consumes 40 bytes (ip6 hdr) of the mtu of the
underlaying device. So for a normal ethernet bearer, the mtu of the ip6_tunnel is
1460.
However, when creating a tunnel the encap limit option is enabled by default, and it
consumes 8 bytes more, so the true mtu shall be 1452.

I dont really know if this breaks some statement in some RFC, so this is a request for
comments.

Signed-off-by: Anders Franzen <anders.franzen@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoBluetooth: Add new PID for Atheros 3011
Bala Shanmugam [Fri, 26 Nov 2010 12:05:46 +0000 (17:35 +0530)]
Bluetooth: Add new PID for Atheros 3011

Atheros 3011 has small sflash firmware and needs to be
blacklisted in transport driver to load actual firmware
in DFU driver.

Signed-off-by: Bala Shanmugam <sbalashanmugam@atheros.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>
13 years agoBluetooth: Fix log spamming in btusb due to autosuspend
Stefan Seyfried [Tue, 30 Nov 2010 20:49:08 +0000 (21:49 +0100)]
Bluetooth: Fix log spamming in btusb due to autosuspend

If a device is autosuspended an inability to resubmit URBs is
to be expected. Check the error code and only log real errors.
(Now that autosuspend is default enabled for btusb, those log
messages were happening all the time e.g. with a BT mouse)

Signed-off-by: Stefan Seyfried <seife+kernel@b1-systems.com>
Signed-off-by: Oliver Neukum <oneukum@suse.de>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>
13 years agoath9k: Fix STA disconnect issue due to received MIC failed bcast frames
Senthil Balasubramanian [Tue, 30 Nov 2010 14:45:39 +0000 (20:15 +0530)]
ath9k: Fix STA disconnect issue due to received MIC failed bcast frames

AR_RxKeyIdxValid will not be set for bcast/mcast frames and so relying
this status for MIC failed frames is buggy.

Due to this, MIC failure events for broadcast frames are not sent to
supplicant resulted in AP disconnecting the STA.

Able to pass Wifi Test case 5.2.18 with this fix.

Cc: Stable <stable@kernel.org> (2.6.36+)
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Fix STA disconnect due to MIC failure
Senthil Balasubramanian [Tue, 30 Nov 2010 14:45:38 +0000 (20:15 +0530)]
mac80211: Fix STA disconnect due to MIC failure

Th commit titled "mac80211: clean up rx handling wrt. found_sta"
removed found_sta variable which caused a MIC failure event
to be reported twice for a single failure to supplicant resulted
in STA disconnect.

This should fix WPA specific countermeasures WiFi test case (5.2.17)
issues with mac80211 based drivers which report MIC failure events in
rx status.

Cc: Stable <stable@kernel.org> (2.6.37)
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: ignore non-bcast mcast deauth/disassoc franes
Christian Lamparter [Mon, 29 Nov 2010 19:53:23 +0000 (20:53 +0100)]
mac80211: ignore non-bcast mcast deauth/disassoc franes

This patch fixes an curious issue due to insufficient
rx frame filtering.

Saqeb Akhter reported frequent disconnects while streaming
videos over samba: <http://marc.info/?m=128600031109136>
> [ 1166.512087] wlan1: deauthenticated from 30:46:9a:10:49:f7 (Reason: 7)
> [ 1526.059997] wlan1: deauthenticated from 30:46:9a:10:49:f7 (Reason: 7)
> [ 2125.324356] wlan1: deauthenticated from 30:46:9a:10:49:f7 (Reason: 7)
> [...]

The reason is that the device generates frames with slightly
bogus SA/TA addresses.

e.g.:
 [ 2314.402316] Ignore 9f:1f:31:f8:64:ff
 [ 2314.402321] Ignore 9f:1f:31:f8:64:ff
 [ 2352.453804] Ignore 0d:1f:31:f8:64:ff
 [ 2352.453808] Ignore 0d:1f:31:f8:64:ff
     ^^ the group-address flag is set!
 (the correct SA/TA would be: 00:1f:31:f8:64:ff)

Since the AP does not know from where the frames come, it
generates a DEAUTH response for the (invalid) mcast address.
This mcast deauth frame then passes through all filters and
tricks the stack into thinking that the AP brutally kicked
us!

This patch fixes the problem by simply ignoring
non-broadcast, group-addressed deauth/disassoc frames.

Cc: Jouni Malinen <j@w1.fi>
Cc: Johannes Berg <johannes@sipsolutions.net>
Reported-by: Saqeb Akhter <saqeb.akhter@gmail.com>
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoorinoco: abort scan on interface down
David Kilroy [Wed, 24 Nov 2010 20:33:02 +0000 (20:33 +0000)]
orinoco: abort scan on interface down

This fixes the problem causing the following trace:

------------[ cut here ]------------
WARNING: at linux-2.6.34/net/wireless/core.c:633 wdev_cleanup_work+0xb7/0xe0 [cfg80211]()
Hardware name: Latitude C840
Pid: 707, comm: cfg80211 Not tainted 2.6.34.7-0.5-desktop #1
Call Trace:
 [<c02065c3>] try_stack_unwind+0x173/0x190
 [<c02051cf>] dump_trace+0x3f/0xe0
 [<c020662b>] show_trace_log_lvl+0x4b/0x60
 [<c0206658>] show_trace+0x18/0x20
 [<c064e0b3>] dump_stack+0x6d/0x72
 [<c02443ae>] warn_slowpath_common+0x6e/0xb0
 [<c0244403>] warn_slowpath_null+0x13/0x20
 [<e2db5497>] wdev_cleanup_work+0xb7/0xe0 [cfg80211]
 [<c025cfa9>] run_workqueue+0x79/0x170
 [<c025d123>] worker_thread+0x83/0xe0
 [<c025fef4>] kthread+0x74/0x80
 [<c0203826>] kernel_thread_helper+0x6/0x10
---[ end trace 3f0348b3b0c6f4ff ]---

Reported by: Giacomo Comes <comes@naic.edu>
Signed-off-by: David Kilroy <kilroyd@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocarl9170: fix carl9170_tx_prepare typo
Christian Lamparter [Fri, 26 Nov 2010 22:29:23 +0000 (23:29 +0100)]
carl9170: fix carl9170_tx_prepare typo

commit: "carl9170: revamp carl9170_tx_prepare"
introduced a peculiar bug that would only show
up if the the module parameter noht is set to 1.

Then all outbound voice, video and background
frames would each invoke a (bogus) RTS/CTS
handshake.

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Fix frame injection using non-AP vif
Jouni Malinen [Fri, 26 Nov 2010 18:41:55 +0000 (20:41 +0200)]
mac80211: Fix frame injection using non-AP vif

In order for frame injection to work properly for some use cases
(e.g., finding the station entry and keys for encryption), mac80211
needs to find the correct sdata entry. This works when the main vif
is in AP mode, but commit a2c1e3dad516618cb0fbfb1a62c36d0b0744573a
broke this particular use case for station main vif. While this type of
injection is quite unusual operation, it has some uses and we should fix
it. Do this by changing the monitor vif sdata selection to allow station
vif to be selected instead of limiting it to just AP vifs. We still need
to skip some iftypes to avoid selecting unsuitable vif for injection.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Disable SWBA interrupt on remove_interface
Rajkumar Manoharan [Fri, 26 Nov 2010 17:54:31 +0000 (23:24 +0530)]
ath9k: Disable SWBA interrupt on remove_interface

while removing beaconing mode interface, SWBA interrupt
was never disabled when there are no other beaconing interfaces.

Cc: stable@kernel.org
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k/carl9170: advertise P2P
Johannes Berg [Fri, 26 Nov 2010 10:38:04 +0000 (11:38 +0100)]
ath9k/carl9170: advertise P2P

With some upcoming changes we'd like to use
the interface types for P2P capability tests.
Enable them now so that when we add those
tests in wpa_supplicant, nothing will break.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: use per-device struct for pm_qos_* operations
Gabor Juhos [Thu, 25 Nov 2010 17:26:07 +0000 (18:26 +0100)]
ath9k: use per-device struct for pm_qos_* operations

The ath9k driver uses a shared pm_qos_request_list structure for all
devices. This causes the following warning if more than one device is
present in the system:

WARNING: at kernel/pm_qos_params.c:234 ath9k_init_device+0x5e8/0x6b0()
pm_qos_add_request() called for already added request
Modules linked in:
Call Trace:
[<802b1cdc>] dump_stack+0x8/0x34
[<8007dd90>] warn_slowpath_common+0x78/0xa4
[<8007de44>] warn_slowpath_fmt+0x2c/0x38
[<801b0828>] ath9k_init_device+0x5e8/0x6b0
[<801bc508>] ath_pci_probe+0x2dc/0x39c
[<80176254>] pci_device_probe+0x64/0xa4
[<8019471c>] driver_probe_device+0xbc/0x188
[<80194854>] __driver_attach+0x6c/0xa4
[<80193e20>] bus_for_each_dev+0x60/0xb0
[<80193580>] bus_add_driver+0xcc/0x268
[<80194c08>] driver_register+0xe0/0x198
[<801764e0>] __pci_register_driver+0x50/0xe0
[<80365f48>] ath9k_init+0x3c/0x6c
[<8006050c>] do_one_initcall+0xfc/0x1d8
[<80355340>] kernel_init+0xd4/0x174
[<800639a4>] kernel_thread_helper+0x10/0x18

---[ end trace 5345fc6f870564a6 ]---

This patch fixes that warning by using a separate pm_qos_request_list
sructure for each device.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agolibertas: fix invalid access
Sven Neumann [Wed, 24 Nov 2010 16:02:00 +0000 (16:02 +0000)]
libertas: fix invalid access

card->priv must not be accessed after lbs_remove_card() was called
as lbs_remove_card() frees card->priv via free_netdev().

For libertas_sdio this is a regression introduced by 23b149c1890f9.
The correct fix to the issue described there is simply to remove the
assignment. This flag is set at the appropriate time inside
lbs_remove_card anyway.

Reported-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: Sven Neumann <s.neumann@raumfeld.com>
Signed-off-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agolibertas: fix memory corruption in lbs_remove_card()
Daniel Drake [Wed, 24 Nov 2010 16:00:49 +0000 (16:00 +0000)]
libertas: fix memory corruption in lbs_remove_card()

"priv" is stored at the end of the wiphy structure, which is freed
during the call to lbs_cfg_free(). It must not be touched afterwards.

Remove the unnecessary NULL assignment causing this memory corruption.

Signed-off-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoaf_unix: limit recursion level
Eric Dumazet [Thu, 25 Nov 2010 04:11:39 +0000 (04:11 +0000)]
af_unix: limit recursion level

Its easy to eat all kernel memory and trigger NMI watchdog, using an
exploit program that queues unix sockets on top of others.

lkml ref : http://lkml.org/lkml/2010/11/25/8

This mechanism is used in applications, one choice we have is to have a
recursion limit.

Other limits might be needed as well (if we queue other types of files),
since the passfd mechanism is currently limited by socket receive queue
sizes only.

Add a recursion_level to unix socket, allowing up to 4 levels.

Each time we send an unix socket through sendfd mechanism, we copy its
recursion level (plus one) to receiver. This recursion level is cleared
when socket receive queue is emptied.

Reported-by: Марк Коренберг <socketpair@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agopch_gbe driver: The wrong of initializer entry
Toshiharu Okada [Mon, 29 Nov 2010 06:18:07 +0000 (06:18 +0000)]
pch_gbe driver: The wrong of initializer entry

The wrong of initializer entry was modified.

Signed-off-by: Toshiharu Okada <toshiharu-linux@dsn.okisemi.com>
Reported-by: Dr. David Alan Gilbert <linux@treblig.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agopch_gbe dreiver: chang author
Toshiharu Okada [Sun, 21 Nov 2010 19:58:37 +0000 (19:58 +0000)]
pch_gbe dreiver: chang author

This driver's AUTHOR was changed to "Toshiharu Okada" from "Masayuki Ohtake".
I update the Kconfig, renamed "Topcliff" to "EG20T".

Signed-off-by: Toshiharu Okada <toshiharu-linux@dsn.okisemi.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoucc_geth: fix ucc halt problem in half duplex mode
Yang Li [Thu, 25 Nov 2010 23:29:58 +0000 (23:29 +0000)]
ucc_geth: fix ucc halt problem in half duplex mode

In commit 58933c64(ucc_geth: Fix the wrong the Rx/Tx FIFO size),
the UCC_GETH_UTFTT_INIT is set to 512 based on the recommendation
of the QE Reference Manual.  But that will sometimes cause tx halt
while working in half duplex mode.

According to errata draft QE_GENERAL-A003(High Tx Virtual FIFO
threshold size can cause UCC to halt), setting UTFTT less than
[(UTFS x (M - 8)/M) - 128] will prevent this from happening
(M is the minimum buffer size).

The patch changes UTFTT back to 256.

Signed-off-by: Li Yang <leoli@freescale.com>
Cc: Jean-Denis Boyer <jdboyer@media5corp.com>
Cc: Andreas Schmitz <Andreas.Schmitz@riedel.net>
Cc: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoinet: Fix __inet_inherit_port() to correctly increment bsockets and num_owners
Nagendra Tomar [Fri, 26 Nov 2010 14:26:27 +0000 (14:26 +0000)]
inet: Fix __inet_inherit_port() to correctly increment bsockets and num_owners

inet sockets corresponding to passive connections are added to the bind hash
using ___inet_inherit_port(). These sockets are later removed from the bind
hash using __inet_put_port(). These two functions are not exactly symmetrical.
__inet_put_port() decrements hashinfo->bsockets and tb->num_owners, whereas
___inet_inherit_port() does not increment them. This results in both of these
going to -ve values.

This patch fixes this by calling inet_bind_hash() from ___inet_inherit_port(),
which does the right thing.

'bsockets' and 'num_owners' were introduced by commit a9d8f9110d7e953c
(inet: Allowing more than 64k connections and heavily optimize bind(0))

Signed-off-by: Nagendra Singh Tomar <tomer_iisc@yahoo.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoehea: Add some info messages and fix an issue
Breno Leitao [Fri, 26 Nov 2010 07:26:27 +0000 (07:26 +0000)]
ehea: Add some info messages and fix an issue

This patch adds some debug information about ehea not being able to
allocate enough spaces. Also it correctly updates the amount of available
skb.

Signed-off-by: Breno Leitao <leitao@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agohso: fix disable_net
Filip Aben [Thu, 25 Nov 2010 03:40:50 +0000 (03:40 +0000)]
hso: fix disable_net

The HSO driver incorrectly creates a serial device instead of a net
device when disable_net is set. It shouldn't create anything for the
network interface.

Signed-off-by: Filip Aben <f.aben@option.com>
Reported-by: Piotr Isajew <pki@ex.com.pl>
Reported-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoNET: wan/x25_asy, move lapb_unregister to x25_asy_close_tty
Jiri Slaby [Wed, 24 Nov 2010 13:54:54 +0000 (13:54 +0000)]
NET: wan/x25_asy, move lapb_unregister to x25_asy_close_tty

We register lapb when tty is created, but unregister it only when the
device is UP. So move the lapb_unregister to x25_asy_close_tty after
the device is down.

The old behaviour causes ldisc switching to fail each second attempt,
because we noted for us that the device is unused, so we use it the
second time, but labp layer still have it registered, so it fails
obviously.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Reported-by: Sergey Lapin <slapin@ossfans.org>
Cc: Andrew Hendry <andrew.hendry@gmail.com>
Tested-by: Sergey Lapin <slapin@ossfans.org>
Tested-by: Mikhail Ulyanov <ulyanov.mikhail@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agocxgb4vf: fix setting unicast/multicast addresses ...
Casey Leedom [Wed, 24 Nov 2010 12:23:57 +0000 (12:23 +0000)]
cxgb4vf: fix setting unicast/multicast addresses ...

We were truncating the number of unicast and multicast MAC addresses
supported.  Additionally, we were incorrectly computing the MAC Address
hash (a "1 << N" where we needed a "1ULL << N").

Signed-off-by: Casey Leedom <leedom@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet, ppp: Report correct error code if unit allocation failed
Cyrill Gorcunov [Tue, 23 Nov 2010 11:43:44 +0000 (11:43 +0000)]
net, ppp: Report correct error code if unit allocation failed

Allocating unit from ird might return several error codes
not only -EAGAIN, so it should not be changed and returned
precisely. Same time unit release procedure should be invoked
only if device is unregistering.

Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
CC: Paul Mackerras <paulus@samba.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoDECnet: don't leak uninitialized stack byte
Dan Rosenberg [Tue, 23 Nov 2010 11:02:13 +0000 (11:02 +0000)]
DECnet: don't leak uninitialized stack byte

A single uninitialized padding byte is leaked to userspace.

Signed-off-by: Dan Rosenberg <drosenberg@vsecurity.com>
CC: stable <stable@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoau1000_eth: fix invalid address accessing the MAC enable register
Wolfgang Grandegger [Tue, 23 Nov 2010 06:40:25 +0000 (06:40 +0000)]
au1000_eth: fix invalid address accessing the MAC enable register

"aup->enable" holds already the address pointing to the MAC enable
register. The bug was introduced by commit d0e7cb:

"au1000-eth: remove volatiles, switch to I/O accessors".

CC: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Wolfgang Grandegger <wg@denx.de>
Acked-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agodccp: fix error in updating the GAR
Gerrit Renker [Tue, 23 Nov 2010 02:36:56 +0000 (02:36 +0000)]
dccp: fix error in updating the GAR

This fixes a bug in updating the Greatest Acknowledgment number Received (GAR):
the current implementation does not track the greatest received value -
lower values in the range AWL..AWH (RFC 4340, 7.5.1) erase higher ones.

Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMerge branch 'vhost-net' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
David S. Miller [Sun, 28 Nov 2010 19:27:44 +0000 (11:27 -0800)]
Merge branch 'vhost-net' of git://git./linux/kernel/git/mst/vhost

13 years agotcp: restrict net.ipv4.tcp_adv_win_scale (#20312)
Alexey Dobriyan [Mon, 22 Nov 2010 12:54:21 +0000 (12:54 +0000)]
tcp: restrict net.ipv4.tcp_adv_win_scale (#20312)

tcp_win_from_space() does the following:

      if (sysctl_tcp_adv_win_scale <= 0)
              return space >> (-sysctl_tcp_adv_win_scale);
      else
              return space - (space >> sysctl_tcp_adv_win_scale);

"space" is int.

As per C99 6.5.7 (3) shifting int for 32 or more bits is
undefined behaviour.

Indeed, if sysctl_tcp_adv_win_scale is exactly 32,
space >> 32 equals space and function returns 0.

Which means we busyloop in tcp_fixup_rcvbuf().

Restrict net.ipv4.tcp_adv_win_scale to [-31, 31].

Fix https://bugzilla.kernel.org/show_bug.cgi?id=20312

Steps to reproduce:

      echo 32 >/proc/sys/net/ipv4/tcp_adv_win_scale
      wget www.kernel.org
      [softlockup]

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonetns: Don't leak others' openreq-s in proc
Pavel Emelyanov [Mon, 22 Nov 2010 03:26:12 +0000 (03:26 +0000)]
netns: Don't leak others' openreq-s in proc

The /proc/net/tcp leaks openreq sockets from other namespaces.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoNet: ceph: Makefile: Remove unnessary code
Tracey Dent [Sun, 21 Nov 2010 15:23:50 +0000 (15:23 +0000)]
Net: ceph: Makefile: Remove unnessary code

Remove the if and else conditional because the code is in mainline and there
is no need in it being there.

Signed-off-by: Tracey Dent <tdent48227@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agovhost/net: fix rcu check usage
Michael S. Tsirkin [Sun, 14 Nov 2010 15:31:52 +0000 (17:31 +0200)]
vhost/net: fix rcu check usage

Incorrect rcu check was used as rcu isn't done
under mutex here. Force check to 1 for now,
to stop it from complaining.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
13 years agoeconet: fix CVE-2010-3848
Phil Blundell [Wed, 24 Nov 2010 19:51:47 +0000 (11:51 -0800)]
econet: fix CVE-2010-3848

Don't declare variable sized array of iovecs on the stack since this
could cause stack overflow if msg->msgiovlen is large.  Instead, coalesce
the user-supplied data into a new buffer and use a single iovec for it.

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoeconet: fix CVE-2010-3850
Phil Blundell [Wed, 24 Nov 2010 19:49:53 +0000 (11:49 -0800)]
econet: fix CVE-2010-3850

Add missing check for capable(CAP_NET_ADMIN) in SIOCSIFADDR operation.

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoeconet: disallow NULL remote addr for sendmsg(), fixes CVE-2010-3849
Phil Blundell [Wed, 24 Nov 2010 19:49:19 +0000 (11:49 -0800)]
econet: disallow NULL remote addr for sendmsg(), fixes CVE-2010-3849

Later parts of econet_sendmsg() rely on saddr != NULL, so return early
with EINVAL if NULL was passed otherwise an oops may occur.

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agotcp: Make TCP_MAXSEG minimum more correct.
David S. Miller [Wed, 24 Nov 2010 19:47:22 +0000 (11:47 -0800)]
tcp: Make TCP_MAXSEG minimum more correct.

Use TCP_MIN_MSS instead of constant 64.

Reported-by: Min Zhang <mzhang@mvista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoisdn: icn: Fix stack corruption bug.
Steven Rostedt [Wed, 24 Nov 2010 19:19:05 +0000 (11:19 -0800)]
isdn: icn: Fix stack corruption bug.

Running randconfig with ktest.pl I hit this bug:

[   16.101158] ICN-ISDN-driver Rev 1.65.6.8 mem=0x000d0000
[   16.106376] icn: (line0) ICN-2B, port 0x320 added
[   16.111064] Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: c1642880
[   16.111066]
[   16.121214] Pid: 1, comm: swapper Not tainted 2.6.37-rc2-test-00124-g6656b3f #8
[   16.128499] Call Trace:
[   16.130942]  [<c0f51662>] ? printk+0x1d/0x23
[   16.135200]  [<c0f5153f>] panic+0x5c/0x162
[   16.139286]  [<c0d62a9a>] ? icn_addcard+0x6d/0xbe
[   16.143975]  [<c0445783>] print_tainted+0x0/0x8c
[   16.148582]  [<c1642880>] ? icn_init+0xd8/0xdf
[   16.153012]  [<c1642880>] icn_init+0xd8/0xdf
[   16.157271]  [<c04012e5>] do_one_initcall+0x8c/0x143
[   16.162222]  [<c16427a8>] ? icn_init+0x0/0xdf
[   16.166566]  [<c15f1a05>] kernel_init+0x13f/0x1da
[   16.171256]  [<c15f18c6>] ? kernel_init+0x0/0x1da
[   16.175945]  [<c0403bfe>] kernel_thread_helper+0x6/0x10
[   16.181181] panic occurred, switching back to text console

Looking into it I found that the stack was corrupted by the assignment
of the Rev #. The variable rev is given 10 bytes, and in this output the
characters that were copied was: " 1.65.6.8 $". Which was 11 characters
plus the null ending character for a total of 12 bytes, thus corrupting
the stack.

This patch ups the variable size to 20 bytes as well as changes the
strcpy to strncpy. I also added a check to make sure '$' is found.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Wed, 24 Nov 2010 17:16:14 +0000 (09:16 -0800)]
Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-2.6

13 years agoaf_unix: limit unix_tot_inflight
Eric Dumazet [Wed, 24 Nov 2010 17:15:27 +0000 (09:15 -0800)]
af_unix: limit unix_tot_inflight

Vegard Nossum found a unix socket OOM was possible, posting an exploit
program.

My analysis is we can eat all LOWMEM memory before unix_gc() being
called from unix_release_sock(). Moreover, the thread blocked in
unix_gc() can consume huge amount of time to perform cleanup because of
huge working set.

One way to handle this is to have a sensible limit on unix_tot_inflight,
tested from wait_for_unix_gc() and to force a call to unix_gc() if this
limit is hit.

This solves the OOM and also reduce overall latencies, and should not
slowdown normal workloads.

Reported-by: Vegard Nossum <vegard.nossum@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agowireless: b43: fix error path in SDIO
Guennadi Liakhovetski [Tue, 23 Nov 2010 16:10:24 +0000 (17:10 +0100)]
wireless: b43: fix error path in SDIO

Fix unbalanced call to sdio_release_host() on the error path.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoBluetooth: Fix not returning proper error in SCO
Gustavo F. Padovan [Mon, 1 Nov 2010 19:08:50 +0000 (19:08 +0000)]
Bluetooth: Fix not returning proper error in SCO

Return 0 in that situation could lead to errors in the caller.

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>
13 years agocarl9170: fix virtual interface setup crash
Christian Lamparter [Sat, 20 Nov 2010 12:15:27 +0000 (13:15 +0100)]
carl9170: fix virtual interface setup crash

This patch fixes a faulty bound check which caused a
crash when too many virtual interface were brought up.

BUG: unable to handle kernel NULL pointer dereference at 00000004
IP: [<f8125f67>] carl9170_op_add_interface+0x1d7/0x2c0 [carl9170]
*pde = 00000000
Oops: 0002 [#1] PREEMPT
Modules linked in: carl9170 [...]
Pid: 4720, comm: wpa_supplicant Not tainted 2.6.37-rc2-wl+
EIP: 0060:[<f8125f67>] EFLAGS: 00210206 CPU: 0
EIP is at carl9170_op_add_interface+0x1d7/0x2c0 [carl9170]
EAX: 00000000 ...
Process wpa_supplicant
Stack:
 f4f88f34 fffffff4 ..
Call Trace:
 [<f8f4e666>] ? ieee80211_do_open+0x406/0x5c0 [mac80211]
 [...]
Code: <89> 42 04 ...
EIP: [<f8125f67>] carl9170_op_add_interface+0x1d7/0x2c0 [carl9170]
CR2: 0000000000000004

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agossb: b43-pci-bridge: Add new vendor for BCM4318
Daniel Klaffenbach [Sat, 20 Nov 2010 03:25:21 +0000 (21:25 -0600)]
ssb: b43-pci-bridge: Add new vendor for BCM4318

Add new vendor for Broadcom 4318.

Signed-off-by: Daniel Klaffenbach <danielklaffenbach@gmail.com>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@kernel.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: fix timeout on stopping rx dma
Felix Fietkau [Sat, 20 Nov 2010 02:08:47 +0000 (03:08 +0100)]
ath9k: fix timeout on stopping rx dma

It seems that using ath9k_hw_stoppcurecv to stop rx dma is not enough.
When it's time to stop DMA, the PCU is still busy, so the rx enable
bit never clears.
Using ath9k_hw_abortpcurecv helps with getting rx stopped much faster,
with this change, I cannot reproduce the rx stop related WARN_ON anymore.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoof/phylib: Use device tree properties to initialize Marvell PHYs.
David Daney [Fri, 19 Nov 2010 12:13:18 +0000 (12:13 +0000)]
of/phylib: Use device tree properties to initialize Marvell PHYs.

Some aspects of PHY initialization are board dependent, things like
indicator LED connections and some clocking modes cannot be determined
by probing.  The dev_flags element of struct phy_device can be used to
control these things if an appropriate value can be passed from the
Ethernet driver.  We run into problems however if the PHY connections
are specified by the device tree.  There is no way for the Ethernet
driver to know what flags it should pass.

If we are using the device tree, the struct phy_device will be
populated with the device tree node corresponding to the PHY, and we
can extract extra configuration information from there.

The next question is what should the format of that information be?
It is highly device specific, and the device tree representation
should not be tied to any arbitrary kernel defined constants.  A
straight forward representation is just to specify the exact bits that
should be set using the "marvell,reg-init" property:

      phy5: ethernet-phy@5 {
        reg = <5>;
        compatible = "marvell,88e1149r";
        marvell,reg-init =
                /* led[0]:1000, led[1]:100, led[2]:10, led[3]:tx */
                <3 0x10 0 0x5777>, /* Reg 3,16 <- 0x5777 */
                /* mix %:0, led[0123]:drive low off hiZ */
                <3 0x11 0 0x00aa>, /* Reg 3,17 <- 0x00aa */
                /* default blink periods. */
                <3 0x12 0 0x4105>, /* Reg 3,18 <- 0x4105 */
                /* led[4]:rx, led[5]:dplx, led[45]:drive low off hiZ */
                <3 0x13 0 0x0a60>; /* Reg 3,19 <- 0x0a60 */
      };

      phy6: ethernet-phy@6 {
        reg = <6>;
        compatible = "marvell,88e1118";
        marvell,reg-init =
                /* Fix rx and tx clock transition timing */
                <2 0x15 0xffcf 0>, /* Reg 2,21 Clear bits 4, 5 */
                /* Adjust LED drive. */
                <3 0x11 0 0x442a>, /* Reg 3,17 <- 0442a */
                /* irq, blink-activity, blink-link */
                <3 0x10 0 0x0242>; /* Reg 3,16 <- 0x0242 */
      };

The Marvell PHYs have a page select register at register 22 (0x16), we
can specify any register by its page and register number.  These are
the first and second word.  The third word contains a mask to be ANDed
with the existing register value, and the fourth word is ORed with the
result to yield the new register value.  The new marvell_of_reg_init
function leaves the page select register unchanged, so a call to it
can be dropped into the .config_init functions without unduly
affecting the state of the PHY.

If CONFIG_OF_MDIO is not set, there is no of_node, or no
"marvell,reg-init" property, the PHY initialization is unchanged.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Cyril Chemparathy <cyril@ti.com>
Cc: David Daney <ddaney@caviumnetworks.com>
Cc: Arnaud Patard <arnaud.patard@rtp-net.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>