pandora-kernel.git
12 years agotcp: avoid order-1 allocations on wifi and tx path
Eric Dumazet [Tue, 10 Apr 2012 20:30:48 +0000 (20:30 +0000)]
tcp: avoid order-1 allocations on wifi and tx path

Marc Merlin reported many order-1 allocations failures in TX path on its
wireless setup, that dont make any sense with MTU=1500 network, and non
SG capable hardware.

After investigation, it turns out TCP uses sk_stream_alloc_skb() and
used as a convention skb_tailroom(skb) to know how many bytes of data
payload could be put in this skb (for non SG capable devices)

Note : these skb used kmalloc-4096 (MTU=1500 + MAX_HEADER +
sizeof(struct skb_shared_info) being above 2048)

Later, mac80211 layer need to add some bytes at the tail of skb
(IEEE80211_ENCRYPT_TAILROOM = 18 bytes) and since no more tailroom is
available has to call pskb_expand_head() and request order-1
allocations.

This patch changes sk_stream_alloc_skb() so that only
sk->sk_prot->max_header bytes of headroom are reserved, and use a new
skb field, avail_size to hold the data payload limit.

This way, order-0 allocations done by TCP stack can leave more than 2 KB
of tailroom and no more allocation is performed in mac80211 layer (or
any layer needing some tailroom)

avail_size is unioned with mark/dropcount, since mark will be set later
in IP stack for output packets. Therefore, skb size is unchanged.

Reported-by: Marc MERLIN <marc@merlins.org>
Tested-by: Marc MERLIN <marc@merlins.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet: allow pskb_expand_head() to get maximum tailroom
Eric Dumazet [Tue, 10 Apr 2012 20:08:39 +0000 (20:08 +0000)]
net: allow pskb_expand_head() to get maximum tailroom

Marc Merlin reported many order-1 allocations failures in TX path on its
wireless setup, that dont make any sense with MTU=1500 network, and non
SG capable hardware.

Turns out part of the problem comes from pskb_expand_head() not using
ksize() to get exact head size given by kmalloc(). Doing the same thing
than __alloc_skb() allows more tailroom in skb and can prevent future
reallocations.

As a bonus, struct skb_shared_info becomes cache line aligned.

Reported-by: Marc MERLIN <marc@merlins.org>
Tested-by: Marc MERLIN <marc@merlins.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobridge: Do not send queries on multicast group leaves
Herbert Xu [Wed, 4 Apr 2012 01:01:20 +0000 (01:01 +0000)]
bridge: Do not send queries on multicast group leaves

As it stands the bridge IGMP snooping system will respond to
group leave messages with queries for remaining membership.
This is both unnecessary and undesirable.  First of all any
multicast routers present should be doing this rather than us.
What's more the queries that we send may end up upsetting other
multicast snooping swithces in the system that are buggy.

In fact, we can simply remove the code that send these queries
because the existing membership expiry mechanism doesn't rely
on them anyway.

So this patch simply removes all code associated with group
queries in response to group leave messages.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMAINTAINERS: Mark NATSEMI driver as orphan'd.
David S. Miller [Wed, 11 Apr 2012 01:10:43 +0000 (21:10 -0400)]
MAINTAINERS: Mark NATSEMI driver as orphan'd.

After discussion with Tim Hockin.

Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agotcp: fix tcp_rcv_rtt_update() use of an unscaled RTT sample
Neal Cardwell [Tue, 10 Apr 2012 07:59:20 +0000 (07:59 +0000)]
tcp: fix tcp_rcv_rtt_update() use of an unscaled RTT sample

Fix a code path in tcp_rcv_rtt_update() that was comparing scaled and
unscaled RTT samples.

The intent in the code was to only use the 'm' measurement if it was a
new minimum.  However, since 'm' had not yet been shifted left 3 bits
but 'new_sample' had, this comparison would nearly always succeed,
leading us to erroneously set our receive-side RTT estimate to the 'm'
sample when that sample could be nearly 8x too high to use.

The overall effect is to often cause the receive-side RTT estimate to
be significantly too large (up to 40% too large for brief periods in
my tests).

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agotcp: restore correct limit
Eric Dumazet [Tue, 10 Apr 2012 00:56:42 +0000 (00:56 +0000)]
tcp: restore correct limit

Commit c43b874d5d714f (tcp: properly initialize tcp memory limits) tried
to fix a regression added in commits 4acb4190 & 3dc43e3,
but still get it wrong.

Result is machines with low amount of memory have too small tcp_rmem[2]
value and slow tcp receives : Per socket limit being 1/1024 of memory
instead of 1/128 in old kernels, so rcv window is capped to small
values.

Fix this to match comment and previous behavior.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Glauber Costa <glommer@parallels.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, 10 Apr 2012 18:38:31 +0000 (14:38 -0400)]
Merge branch 'master' of git://1984.lsi.us.es/net

12 years agonetfilter: nf_conntrack: fix incorrect logic in nf_conntrack_init_net
Gao feng [Sat, 7 Apr 2012 16:08:28 +0000 (16:08 +0000)]
netfilter: nf_conntrack: fix incorrect logic in nf_conntrack_init_net

in function nf_conntrack_init_net,when nf_conntrack_timeout_init falied,
we should call nf_conntrack_ecache_fini to do rollback.
but the current code calls nf_conntrack_timeout_fini.

Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
12 years agonetfilter: nf_ct_ipv4: packets with wrong ihl are invalid
Jozsef Kadlecsik [Tue, 3 Apr 2012 20:02:01 +0000 (22:02 +0200)]
netfilter: nf_ct_ipv4: packets with wrong ihl are invalid

It was reported that the Linux kernel sometimes logs:

klogd: [2629147.402413] kernel BUG at net / netfilter /
nf_conntrack_proto_tcp.c: 447!
klogd: [1072212.887368] kernel BUG at net / netfilter /
nf_conntrack_proto_tcp.c: 392

ipv4_get_l4proto() in nf_conntrack_l3proto_ipv4.c and tcp_error() in
nf_conntrack_proto_tcp.c should catch malformed packets, so the errors
at the indicated lines - TCP options parsing - should not happen.
However, tcp_error() relies on the "dataoff" offset to the TCP header,
calculated by ipv4_get_l4proto().  But ipv4_get_l4proto() does not check
bogus ihl values in IPv4 packets, which then can slip through tcp_error()
and get caught at the TCP options parsing routines.

The patch fixes ipv4_get_l4proto() by invalidating packets with bogus
ihl value.

The patch closes netfilter bugzilla id 771.

Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
12 years agonetfilter: nf_ct_ipv4: handle invalid IPv4 and IPv6 packets consistently
Jozsef Kadlecsik [Mon, 9 Apr 2012 14:32:16 +0000 (16:32 +0200)]
netfilter: nf_ct_ipv4: handle invalid IPv4 and IPv6 packets consistently

IPv6 conntrack marked invalid packets as INVALID and let the user
drop those by an explicit rule, while IPv4 conntrack dropped such
packets itself.

IPv4 conntrack is changed so that it marks INVALID packets and let
the user to drop them.

Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
12 years agonetfilter: ip6_tables: ip6t_ext_hdr is now static inline
Pablo Neira Ayuso [Fri, 6 Apr 2012 16:12:54 +0000 (18:12 +0200)]
netfilter: ip6_tables: ip6t_ext_hdr is now static inline

We may hit this in xt_LOG:

net/built-in.o:xt_LOG.c:function dump_ipv6_packet:
error: undefined reference to 'ip6t_ext_hdr'

happens with these config options:

CONFIG_NETFILTER_XT_TARGET_LOG=y
CONFIG_IP6_NF_IPTABLES=m

ip6t_ext_hdr is fairly small and it is called in the packet path.
Make it static inline.

Reported-by: Simon Kirby <sim@netnation.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
12 years agonetfilter: nf_ct_tcp: don't scale the size of the window up twice
Changli Gao [Sun, 1 Apr 2012 17:25:06 +0000 (17:25 +0000)]
netfilter: nf_ct_tcp: don't scale the size of the window up twice

For a picked up connection, the window win is scaled twice: one is by the
initialization code, and the other is by the sender updating code.

I use the temporary variable swin instead of modifying the variable win.

Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Acked-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
12 years agoMake the "word-at-a-time" helper functions more commonly usable
Linus Torvalds [Fri, 6 Apr 2012 20:54:56 +0000 (13:54 -0700)]
Make the "word-at-a-time" helper functions more commonly usable

I have a new optimized x86 "strncpy_from_user()" that will use these
same helper functions for all the same reasons the name lookup code uses
them.  This is preparation for that.

This moves them into an architecture-specific header file.  It's
architecture-specific for two reasons:

 - some of the functions are likely to want architecture-specific
   implementations.  Even if the current code happens to be "generic" in
   the sense that it should work on any little-endian machine, it's
   likely that the "multiply by a big constant and shift" implementation
   is less than optimal for an architecture that has a guaranteed fast
   bit count instruction, for example.

 - I expect that if architectures like sparc want to start playing
   around with this, we'll need to abstract out a few more details (in
   particular the actual unaligned accesses).  So we're likely to have
   more architecture-specific stuff if non-x86 architectures start using
   this.

   (and if it turns out that non-x86 architectures don't start using
   this, then having it in an architecture-specific header is still the
   right thing to do, of course)

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 6 Apr 2012 17:37:38 +0000 (10:37 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking updates from David Miller:

 1) Fix inaccuracies in network driver interface documentation, from Ben
    Hutchings.

 2) Fix handling of negative offsets in BPF JITs, from Jan Seiffert.

 3) Compile warning, locking, and refcounting fixes in netfilter's
    xt_CT, from Pablo Neira Ayuso.

 4) phonet sendmsg needs to validate user length just like any other
    datagram protocol, fix from Sasha Levin.

 5) Ipv6 multicast code uses wrong loop index, from RongQing Li.

 6) Link handling and firmware fixes in bnx2x driver from Yaniv Rosner
    and Yuval Mintz.

 7) mlx4 erroneously allocates 4 pages at a time, regardless of page
    size, fix from Thadeu Lima de Souza Cascardo.

 8) SCTP socket option wasn't extended in a backwards compatible way,
    fix from Thomas Graf.

 9) Add missing address change event emissions to bonding, from Shlomo
    Pongratz.

10) /proc/net/dev regressed because it uses a private offset to track
    where we are in the hash table, but this doesn't track the offset
    pullback that the seq_file code does resulting in some entries being
    missed in large dumps.

    Fix from Eric Dumazet.

11) do_tcp_sendpage() unloads the send queue way too fast, because it
    invokes tcp_push() when it shouldn't.  Let the natural sequence
    generated by the splice paths, and the assosciated MSG_MORE
    settings, guide the tcp_push() calls.

    Otherwise what goes out of TCP is spaghetti and doesn't batch
    effectively into GSO/TSO clusters.

    From Eric Dumazet.

12) Once we put a SKB into either the netlink receiver's queue or a
    socket error queue, it can be consumed and freed up, therefore we
    cannot touch it after queueing it like that.

    Fixes from Eric Dumazet.

13) PPP has this annoying behavior in that for every transmit call it
    immediately stops the TX queue, then calls down into the next layer
    to transmit the PPP frame.

    But if that next layer can take it immediately, it just un-stops the
    TX queue right before returning from the transmit method.

    Besides being useless work, it makes several facilities unusable, in
    particular things like the equalizers.  Well behaved devices should
    only stop the TX queue when they really are full, and in PPP's case
    when it gets backlogged to the downstream device.

    David Woodhouse therefore fixed PPP to not stop the TX queue until
    it's downstream can't take data any more.

14) IFF_UNICAST_FLT got accidently lost in some recent stmmac driver
    changes, re-add.  From Marc Kleine-Budde.

15) Fix link flaps in ixgbe, from Eric W. Multanen.

16) Descriptor writeback fixes in e1000e from Matthew Vick.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (47 commits)
  net: fix a race in sock_queue_err_skb()
  netlink: fix races after skb queueing
  doc, net: Update ndo_start_xmit return type and values
  doc, net: Remove instruction to set net_device::trans_start
  doc, net: Update netdev operation names
  doc, net: Update documentation of synchronisation for TX multiqueue
  doc, net: Remove obsolete reference to dev->poll
  ethtool: Remove exception to the requirement of holding RTNL lock
  MAINTAINERS: update for Marvell Ethernet drivers
  bonding: properly unset current_arp_slave on slave link up
  phonet: Check input from user before allocating
  tcp: tcp_sendpages() should call tcp_push() once
  ipv6: fix array index in ip6_mc_add_src()
  mlx4: allocate just enough pages instead of always 4 pages
  stmmac: re-add IFF_UNICAST_FLT for dwmac1000
  bnx2x: Clear MDC/MDIO warning message
  bnx2x: Fix BCM57711+BCM84823 link issue
  bnx2x: Clear BCM84833 LED after fan failure
  bnx2x: Fix BCM84833 PHY FW version presentation
  bnx2x: Fix link issue for BCM8727 boards.
  ...

12 years agonet: fix a race in sock_queue_err_skb()
Eric Dumazet [Fri, 6 Apr 2012 08:49:10 +0000 (10:49 +0200)]
net: fix a race in sock_queue_err_skb()

As soon as an skb is queued into socket error queue, another thread
can consume it, so we are not allowed to reference skb anymore, or risk
use after free.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonetlink: fix races after skb queueing
Eric Dumazet [Thu, 5 Apr 2012 22:17:46 +0000 (22:17 +0000)]
netlink: fix races after skb queueing

As soon as an skb is queued into socket receive_queue, another thread
can consume it, so we are not allowed to reference skb anymore, or risk
use after free.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodoc, net: Update ndo_start_xmit return type and values
Ben Hutchings [Thu, 5 Apr 2012 14:40:25 +0000 (14:40 +0000)]
doc, net: Update ndo_start_xmit return type and values

Commit dc1f8bf68b311b1537cb65893430b6796118498a ('netdev: change
transmit to limited range type') changed the required return type and
9a1654ba0b50402a6bd03c7b0fe9b0200a5ea7b1 ('net: Optimize
hard_start_xmit() return checking') changed the valid numerical
return values.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodoc, net: Remove instruction to set net_device::trans_start
Ben Hutchings [Thu, 5 Apr 2012 14:40:06 +0000 (14:40 +0000)]
doc, net: Remove instruction to set net_device::trans_start

Commit 08baf561083bc27a953aa087dd8a664bb2b88e8e ('net:
txq_trans_update() helper') made it unnecessary for most drivers to
set net_device::trans_start (or netdev_queue::trans_start).

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodoc, net: Update netdev operation names
Ben Hutchings [Thu, 5 Apr 2012 14:39:47 +0000 (14:39 +0000)]
doc, net: Update netdev operation names

Commits d314774cf2cd5dfeb39a00d37deee65d4c627927 ('netdev: network
device operations infrastructure') and
008298231abbeb91bc7be9e8b078607b816d1a4a ('netdev: add more functions
to netdevice ops') moved and renamed net device operation pointers.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodoc, net: Update documentation of synchronisation for TX multiqueue
Ben Hutchings [Thu, 5 Apr 2012 14:39:30 +0000 (14:39 +0000)]
doc, net: Update documentation of synchronisation for TX multiqueue

Commits e308a5d806c852f56590ffdd3834d0df0cbed8d7 ('netdev: Add
netdev->addr_list_lock protection.') and
e8a0464cc950972824e2e128028ae3db666ec1ed ('netdev: Allocate multiple
queues for TX.') introduced more fine-grained locks.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodoc, net: Remove obsolete reference to dev->poll
Ben Hutchings [Thu, 5 Apr 2012 14:39:10 +0000 (14:39 +0000)]
doc, net: Remove obsolete reference to dev->poll

Commit bea3348eef27e6044b6161fd04c3152215f96411 ('[NET]: Make NAPI
polling independent of struct net_device objects.') removed the
automatic disabling of NAPI polling by dev_close(), and drivers
must now do this themselves.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoethtool: Remove exception to the requirement of holding RTNL lock
Ben Hutchings [Thu, 5 Apr 2012 14:38:49 +0000 (14:38 +0000)]
ethtool: Remove exception to the requirement of holding RTNL lock

Commit e52ac3398c3d772d372b9b62ab408fd5eec96840 ('net: Use device
model to get driver name in skb_gso_segment()') removed the only
in-tree caller of ethtool ops that doesn't hold the RTNL lock.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
Linus Torvalds [Fri, 6 Apr 2012 05:13:39 +0000 (22:13 -0700)]
Merge tag 'fixes-for-linus' of git://git./linux/kernel/git/arm/arm-soc

Pull "ARM: SoC fixes: from Olof Johansson:
 "A bunch of fixes for regressions (and a few other problems) in
  3.4-rc1:

 - Fix for regression of mach/io.h cleanup on platforms with PCI or
   PCMCIA (adding back the include file on those for now)
 - AT91 fixes for usb and spi
 - smsc911x ethernet fixes for i.MX
 - smsc911x fixes for OMAP
 - gpio fixes for Tegra
 - A handful of build error and warning fixes for various platforms
 - cpufreq kconfig dependencies, build and lowlevel debug fixes for
   Samsung platforms

  In other words, more or less the regular collection of -rc1/2 type
  material.  A few of them, in particular the smsc911x for OMAP series,
  aren't technically regressions for 3.4, but they're valid fixes and
  we're still relatively early in the rc cycle so it seems appropriate
  to include them."

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (60 commits)
  ARM: fix __io macro for PCMCIA
  ARM: EXYNOS: Fix compiler warning in dma.c file
  ARM: EXYNOS: fix ISO C90 warning
  ARM: OMAP2+: hwmod: Fix wrong SYSC_TYPE1_XXX_MASK bit definitions
  ARM: OMAP2+: hwmod: Make omap_hwmod_softreset wait for reset status
  ARM: OMAP2+: hwmod: Restore sysc after a reset
  ARM: OMAP2+: omap_hwmod: Allow io_ring wakeup configuration for all modules
  ARM: OMAP3: clock data: fill in some missing clockdomains
  ARM: OMAP4: clock data: Force a DPLL clkdm/pwrdm ON before a relock
  ARM: OMAP4: clock data: fix mult and div mask for USB_DPLL
  ARM: OMAP2+: powerdomain: Wait for powerdomain transition in pwrdm_state_switch()
  gpio: tegra: Iterate over the correct number of banks
  gpio: tegra: fix register address calculations for Tegra30
  EXYNOS: fix dependency for EXYNOS_CPUFREQ
  ARM: at91: dt: remove unit-address part for memory nodes
  ARM: at91: fix check of valid GPIO for SPI and USB
  USB: ehci-atmel: add needed of.h header file
  ARM: at91/NAND DT bindings: add comments
  ARM: at91/at91sam9x5.dtsi: fix NAND ale/cle in DT file
  USB: ohci-at91: trivial return code name change
  ...

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lliubbo...
Linus Torvalds [Fri, 6 Apr 2012 04:38:51 +0000 (21:38 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/lliubbo/blackfin

Pull a few blackfin compile fixes from Bob Liu.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lliubbo/blackfin:
  blackfin: update defconfig for bf527-ezkit
  blackfin: gpio: fix compile error if !CONFIG_GPIOLIB
  blackfin: fix L1 data A overflow link issue

12 years agoblackfin: update defconfig for bf527-ezkit
Bob Liu [Thu, 5 Apr 2012 02:40:35 +0000 (10:40 +0800)]
blackfin: update defconfig for bf527-ezkit

To fix compile error:
drivers/usb/musb/blackfin.h:51:3: error: #error "Please use PIO mode in MUSB
driver on bf52x chip v0.0 and v0.1"
make[4]: *** [drivers/usb/musb/blackfin.o] Error 1

Signed-off-by: Bob Liu <lliubbo@gmail.com>
12 years agoblackfin: gpio: fix compile error if !CONFIG_GPIOLIB
Bob Liu [Tue, 27 Mar 2012 03:27:15 +0000 (11:27 +0800)]
blackfin: gpio: fix compile error if !CONFIG_GPIOLIB

Add __gpio_get_value()/__gpio_set_value() to fix compile error if
CONFIG_GPIOLIB = n.

Signed-off-by: Bob Liu <lliubbo@gmail.com>
12 years agoblackfin: fix L1 data A overflow link issue
Mike Frysinger [Wed, 28 Mar 2012 03:43:02 +0000 (11:43 +0800)]
blackfin: fix L1 data A overflow link issue

This patch fix below compile error:
"bfin-uclinux-ld: L1 data A overflow!"

It is due to the recent lib/gen_crc32table.c change:
46c5801eaf86e83cb3a4142ad35188db5011fff0
crc32: bolt on crc32c

it added 8KiB more data to __cacheline_aligned which cause blackfin L1 data
cache overflow.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Bob Liu <lliubbo@gmail.com>
12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/apm
Linus Torvalds [Fri, 6 Apr 2012 00:34:30 +0000 (17:34 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/apm

Pull an APM fix from Jiri Kosina:
 "One deadlock/race fix from Niel that got introduced when we were
  moving away from freezer_*_count() to wait_event_freezable()."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/apm:
  APM: fix deadlock in APM_IOC_SUSPEND ioctl

12 years agoMerge tag 'omap-fixes-a2-for-3.4rc' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Fri, 6 Apr 2012 00:09:45 +0000 (17:09 -0700)]
Merge tag 'omap-fixes-a2-for-3.4rc' of git://git./linux/kernel/git/pjw/omap-pending into fixes

From Paul Walmsley:

OMAP clock, powerdomain, clockdomain, and hwmod fixes intended for the
early v3.4-rc series.  Also contains an HSMMC integration refinement
of an earlier hardware bug workaround.

* tag 'omap-fixes-a2-for-3.4rc' of git://git.kernel.org/pub/scm/linux/kernel/git/pjw/omap-pending:
  ARM: OMAP2+: hwmod: Fix wrong SYSC_TYPE1_XXX_MASK bit definitions
  ARM: OMAP2+: hwmod: Make omap_hwmod_softreset wait for reset status
  ARM: OMAP2+: hwmod: Restore sysc after a reset
  ARM: OMAP2+: omap_hwmod: Allow io_ring wakeup configuration for all modules
  ARM: OMAP3: clock data: fill in some missing clockdomains
  ARM: OMAP4: clock data: Force a DPLL clkdm/pwrdm ON before a relock
  ARM: OMAP4: clock data: fix mult and div mask for USB_DPLL
  ARM: OMAP2+: powerdomain: Wait for powerdomain transition in pwrdm_state_switch()
  ARM: OMAP AM3517/3505: clock data: change EMAC clocks aliases
  ARM: OMAP: clock: fix race in disable all clocks
  ARM: OMAP4: hwmod data: Add aliases for McBSP fclk clocks
  ARM: OMAP3xxx: clock data: fix DPLL4 CLKSEL masks
  ARM: OMAP3xxx: HSMMC: avoid erratum workaround when transceiver is attached
  ARM: OMAP44xx: clockdomain data: correct the emu_sys_clkdm CLKTRCTRL data

12 years agoMAINTAINERS: update for Marvell Ethernet drivers
stephen hemminger [Mon, 2 Apr 2012 12:59:47 +0000 (12:59 +0000)]
MAINTAINERS: update for Marvell Ethernet drivers

Marvell has agreed to do maintenance on the sky2 driver.
   * Add the developer to the maintainers file
   * Remove the old reference to the long gone (sk98lin) driver
   * Rearrange to fit current topic organization

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobonding: properly unset current_arp_slave on slave link up
Veaceslav Falico [Thu, 5 Apr 2012 03:47:43 +0000 (03:47 +0000)]
bonding: properly unset current_arp_slave on slave link up

When a slave comes up, we're unsetting the current_arp_slave without
removing active flags from it, which can lead to situations where we have
more than one slave with active flags in active-backup mode.

To avoid this situation we must remove the active flags from a slave before
removing it as a current_arp_slave.

Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Marcelo Ricardo Leitner <mleitner@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agophonet: Check input from user before allocating
Sasha Levin [Thu, 5 Apr 2012 12:07:45 +0000 (12:07 +0000)]
phonet: Check input from user before allocating

A phonet packet is limited to USHRT_MAX bytes, this is never checked during
tx which means that the user can specify any size he wishes, and the kernel
will attempt to allocate that size.

In the good case, it'll lead to the following warning, but it may also cause
the kernel to kick in the OOM and kill a random task on the server.

[ 8921.744094] WARNING: at mm/page_alloc.c:2255 __alloc_pages_slowpath+0x65/0x730()
[ 8921.749770] Pid: 5081, comm: trinity Tainted: G        W    3.4.0-rc1-next-20120402-sasha #46
[ 8921.756672] Call Trace:
[ 8921.758185]  [<ffffffff810b2ba7>] warn_slowpath_common+0x87/0xb0
[ 8921.762868]  [<ffffffff810b2be5>] warn_slowpath_null+0x15/0x20
[ 8921.765399]  [<ffffffff8117eae5>] __alloc_pages_slowpath+0x65/0x730
[ 8921.769226]  [<ffffffff81179c8a>] ? zone_watermark_ok+0x1a/0x20
[ 8921.771686]  [<ffffffff8117d045>] ? get_page_from_freelist+0x625/0x660
[ 8921.773919]  [<ffffffff8117f3a8>] __alloc_pages_nodemask+0x1f8/0x240
[ 8921.776248]  [<ffffffff811c03e0>] kmalloc_large_node+0x70/0xc0
[ 8921.778294]  [<ffffffff811c4bd4>] __kmalloc_node_track_caller+0x34/0x1c0
[ 8921.780847]  [<ffffffff821b0e3c>] ? sock_alloc_send_pskb+0xbc/0x260
[ 8921.783179]  [<ffffffff821b3c65>] __alloc_skb+0x75/0x170
[ 8921.784971]  [<ffffffff821b0e3c>] sock_alloc_send_pskb+0xbc/0x260
[ 8921.787111]  [<ffffffff821b002e>] ? release_sock+0x7e/0x90
[ 8921.788973]  [<ffffffff821b0ff0>] sock_alloc_send_skb+0x10/0x20
[ 8921.791052]  [<ffffffff824cfc20>] pep_sendmsg+0x60/0x380
[ 8921.792931]  [<ffffffff824cb4a6>] ? pn_socket_bind+0x156/0x180
[ 8921.794917]  [<ffffffff824cb50f>] ? pn_socket_autobind+0x3f/0x90
[ 8921.797053]  [<ffffffff824cb63f>] pn_socket_sendmsg+0x4f/0x70
[ 8921.798992]  [<ffffffff821ab8e7>] sock_aio_write+0x187/0x1b0
[ 8921.801395]  [<ffffffff810e325e>] ? sub_preempt_count+0xae/0xf0
[ 8921.803501]  [<ffffffff8111842c>] ? __lock_acquire+0x42c/0x4b0
[ 8921.805505]  [<ffffffff821ab760>] ? __sock_recv_ts_and_drops+0x140/0x140
[ 8921.807860]  [<ffffffff811e07cc>] do_sync_readv_writev+0xbc/0x110
[ 8921.809986]  [<ffffffff811958e7>] ? might_fault+0x97/0xa0
[ 8921.811998]  [<ffffffff817bd99e>] ? security_file_permission+0x1e/0x90
[ 8921.814595]  [<ffffffff811e17e2>] do_readv_writev+0xe2/0x1e0
[ 8921.816702]  [<ffffffff810b8dac>] ? do_setitimer+0x1ac/0x200
[ 8921.818819]  [<ffffffff810e2ec1>] ? get_parent_ip+0x11/0x50
[ 8921.820863]  [<ffffffff810e325e>] ? sub_preempt_count+0xae/0xf0
[ 8921.823318]  [<ffffffff811e1926>] vfs_writev+0x46/0x60
[ 8921.825219]  [<ffffffff811e1a3f>] sys_writev+0x4f/0xb0
[ 8921.827127]  [<ffffffff82658039>] system_call_fastpath+0x16/0x1b
[ 8921.829384] ---[ end trace dffe390f30db9eb7 ]---

Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Acked-by: Rémi Denis-Courmont <remi.denis-courmont@nokia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agotcp: tcp_sendpages() should call tcp_push() once
Eric Dumazet [Thu, 5 Apr 2012 03:05:35 +0000 (03:05 +0000)]
tcp: tcp_sendpages() should call tcp_push() once

commit 2f533844242 (tcp: allow splice() to build full TSO packets) added
a regression for splice() calls using SPLICE_F_MORE.

We need to call tcp_flush() at the end of the last page processed in
tcp_sendpages(), or else transmits can be deferred and future sends
stall.

Add a new internal flag, MSG_SENDPAGE_NOTLAST, acting like MSG_MORE, but
with different semantic.

For all sendpage() providers, its a transparent change. Only
sock_sendpage() and tcp_sendpages() can differentiate the two different
flags provided by pipe_to_sendpage()

Reported-by: Tom Herbert <therbert@google.com>
Cc: Nandita Dukkipati <nanditad@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: H.K. Jerry Chu <hkchu@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Cc: Mahesh Bandewar <maheshb@google.com>
Cc: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail>com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge branch 'akpm' (Andrew's patch-bomb)
Linus Torvalds [Thu, 5 Apr 2012 22:30:34 +0000 (15:30 -0700)]
Merge branch 'akpm' (Andrew's patch-bomb)

Merge batch of fixes from Andrew Morton:
 "The simple_open() cleanup was held back while I wanted for laggards to
  merge things.

  I still need to send a few checkpoint/restore patches.  I've been
  wobbly about merging them because I'm wobbly about the overall
  prospects for success of the project.  But after speaking with Pavel
  at the LSF conference, it sounds like they're further toward
  completion than I feared - apparently davem is at the "has stopped
  complaining" stage regarding the net changes.  So I need to go back
  and re-review those patchs and their (lengthy) discussion."

* emailed from Andrew Morton <akpm@linux-foundation.org>: (16 patches)
  memcg swap: use mem_cgroup_uncharge_swap fix
  backlight: add driver for DA9052/53 PMIC v1
  C6X: use set_current_blocked() and block_sigmask()
  MAINTAINERS: add entry for sparse checker
  MAINTAINERS: fix REMOTEPROC F: typo
  alpha: use set_current_blocked() and block_sigmask()
  simple_open: automatically convert to simple_open()
  scripts/coccinelle/api/simple_open.cocci: semantic patch for simple_open()
  libfs: add simple_open()
  hugetlbfs: remove unregister_filesystem() when initializing module
  drivers/rtc/rtc-88pm860x.c: fix rtc irq enable callback
  fs/xattr.c:setxattr(): improve handling of allocation failures
  fs/xattr.c:listxattr(): fall back to vmalloc() if kmalloc() failed
  fs/xattr.c: suppress page allocation failure warnings from sys_listxattr()
  sysrq: use SEND_SIG_FORCED instead of force_sig()
  proc: fix mount -t proc -o AAA

12 years agomemcg swap: use mem_cgroup_uncharge_swap fix
Michal Hocko [Thu, 5 Apr 2012 21:25:16 +0000 (14:25 -0700)]
memcg swap: use mem_cgroup_uncharge_swap fix

Although mem_cgroup_uncharge_swap has an empty placeholder for
!CONFIG_CGROUP_MEM_RES_CTLR_SWAP the definition is placed in the
CONFIG_SWAP ifdef block so we are missing the same definition for
!CONFIG_SWAP which implies !CONFIG_CGROUP_MEM_RES_CTLR_SWAP.

This has not been an issue before, because mem_cgroup_uncharge_swap was
not called from !CONFIG_SWAP context.  But Hugh Dickins has a cleanup
patch to call __mem_cgroup_commit_charge_swapin which is defined also
for !CONFIG_SWAP.

Let's move both the empty definition and declaration outside of the
CONFIG_SWAP block to avoid the following compilation error:

  mm/memcontrol.c: In function '__mem_cgroup_commit_charge_swapin':
  mm/memcontrol.c:2837: error: implicit declaration of function 'mem_cgroup_uncharge_swap'

if CONFIG_SWAP is disabled.

Reported-by: David Rientjes <rientjes@google.com>
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agobacklight: add driver for DA9052/53 PMIC v1
Ashish Jangam [Thu, 5 Apr 2012 21:25:15 +0000 (14:25 -0700)]
backlight: add driver for DA9052/53 PMIC v1

DA9052/53 PMIC has capability to supply power for upto 3 banks of 6
white serial LEDS.  It can also control intensity of independent banks
and to drive these banks boost converter will provide up to 24V and
forward current of max 50mA.

This patch allows to control intensity of the individual WLEDs bank
through DA9052/53 PMIC.

This patch is functionally tested on Samsung SMDKV6410.

Signed-off-by: David Dajun Chen <dchen@diasemi.com>
Signed-off-by: Ashish Jangam <ashish.jangam@kpitcummins.com>
Cc: Richard Purdie <rpurdie@rpsys.net>
Cc: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoC6X: use set_current_blocked() and block_sigmask()
Matt Fleming [Thu, 5 Apr 2012 21:25:14 +0000 (14:25 -0700)]
C6X: use set_current_blocked() and block_sigmask()

As described in e6fa16ab9c1e ("signal: sigprocmask() should do
retarget_shared_pending()") the modification of current->blocked is
incorrect as we need to check whether the signal we're about to block is
pending in the shared queue.

Also, use the new helper function introduced in commit 5e6292c0f28f
("signal: add block_sigmask() for adding sigmask to current->blocked")
which centralises the code for updating current->blocked after
successfully delivering a signal and reduces the amount of duplicate
code across architectures.  In the past some architectures got this code
wrong, so using this helper function should stop that from happening
again.

Acked-by: Mark Salter <msalter@redhat.com>
Cc: Aurelien Jacquiot <a-jacquiot@ti.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMAINTAINERS: add entry for sparse checker
Christopher Li [Thu, 5 Apr 2012 21:25:14 +0000 (14:25 -0700)]
MAINTAINERS: add entry for sparse checker

Signed-off-by: Christopher Li <sparse@chrisli.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMAINTAINERS: fix REMOTEPROC F: typo
Joe Perches [Thu, 5 Apr 2012 21:25:13 +0000 (14:25 -0700)]
MAINTAINERS: fix REMOTEPROC F: typo

remoteproc.txt should have been .h

Signed-off-by: Joe Perches <joe@perches.com>
Cc: Ohad Ben-Cohen <ohad@wizery.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoalpha: use set_current_blocked() and block_sigmask()
Matt Fleming [Thu, 5 Apr 2012 21:25:12 +0000 (14:25 -0700)]
alpha: use set_current_blocked() and block_sigmask()

As described in e6fa16ab9c1e ("signal: sigprocmask() should do
retarget_shared_pending()") the modification of current->blocked is
incorrect as we need to check for shared signals we're about to block.

Also, use the new helper function introduced in commit 5e6292c0f28f
("signal: add block_sigmask() for adding sigmask to current->blocked")
which centralises the code for updating current->blocked after
successfully delivering a signal and reduces the amount of duplicate
code across architectures.  In the past some architectures got this code
wrong, so using this helper function should stop that from happening
again.

Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agosimple_open: automatically convert to simple_open()
Stephen Boyd [Thu, 5 Apr 2012 21:25:11 +0000 (14:25 -0700)]
simple_open: automatically convert to simple_open()

Many users of debugfs copy the implementation of default_open() when
they want to support a custom read/write function op.  This leads to a
proliferation of the default_open() implementation across the entire
tree.

Now that the common implementation has been consolidated into libfs we
can replace all the users of this function with simple_open().

This replacement was done with the following semantic patch:

<smpl>
@ open @
identifier open_f != simple_open;
identifier i, f;
@@
-int open_f(struct inode *i, struct file *f)
-{
(
-if (i->i_private)
-f->private_data = i->i_private;
|
-f->private_data = i->i_private;
)
-return 0;
-}

@ has_open depends on open @
identifier fops;
identifier open.open_f;
@@
struct file_operations fops = {
...
-.open = open_f,
+.open = simple_open,
...
};
</smpl>

[akpm@linux-foundation.org: checkpatch fixes]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Julia Lawall <Julia.Lawall@lip6.fr>
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoscripts/coccinelle/api/simple_open.cocci: semantic patch for simple_open()
Julia Lawall [Thu, 5 Apr 2012 21:25:10 +0000 (14:25 -0700)]
scripts/coccinelle/api/simple_open.cocci: semantic patch for simple_open()

Find instances of an open-coded simple_open() and replace them with
calls to simple_open().

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Reported-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agolibfs: add simple_open()
Stephen Boyd [Thu, 5 Apr 2012 21:25:09 +0000 (14:25 -0700)]
libfs: add simple_open()

debugfs and a few other drivers use an open-coded version of
simple_open() to pass a pointer from the file to the read/write file
ops.  Add support for this simple case to libfs so that we can remove
the many duplicate copies of this simple function.

Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agohugetlbfs: remove unregister_filesystem() when initializing module
Hillf Danton [Thu, 5 Apr 2012 21:25:09 +0000 (14:25 -0700)]
hugetlbfs: remove unregister_filesystem() when initializing module

It was introduced by d1d5e05ffdc1 ("hugetlbfs: return error code when
initializing module") but as Al pointed out, is a bad idea.

Quoted comments from Al:
 "Note that unregister_filesystem() in module init is *always* wrong;
  it's not an issue here (it's done too early to care about and
  realistically the box is not going anywhere - it'll panic when attempt
  to exec /sbin/init fails, if not earlier), but it's a damn bad
  example.

  Consider a normal fs module.  Somebody loads it and in parallel with
  that we get a mount attempt on that fs type.  It comes between
  register and failure exits that causes unregister; at that point we
  are screwed since grabbing a reference to module as done by mount is
  enough to prevent exit, but not to prevent the failure of init.  As
  the result, module will get freed when init fails, mounted fs of that
  type be damned."

So remove it.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/rtc/rtc-88pm860x.c: fix rtc irq enable callback
Jett.Zhou [Thu, 5 Apr 2012 21:25:08 +0000 (14:25 -0700)]
drivers/rtc/rtc-88pm860x.c: fix rtc irq enable callback

According to 88pm860x spec, rtc alarm irq enable control is bit3 for
RTC_ALARM_EN, so fix it.

Signed-off-by: Jett.Zhou <jtzhou@marvell.com>
Cc: Axel Lin <axel.lin@gmail.com>
Cc: Haojian Zhuang <haojian.zhuang@marvell.com>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agofs/xattr.c:setxattr(): improve handling of allocation failures
Andrew Morton [Thu, 5 Apr 2012 21:25:07 +0000 (14:25 -0700)]
fs/xattr.c:setxattr(): improve handling of allocation failures

This allocation can be as large as 64k.

 - Add __GFP_NOWARN so the a falied kmalloc() is silent

 - Fall back to vmalloc() if the kmalloc() failed

Cc: Dave Chinner <david@fromorbit.com>
Cc: Dave Jones <davej@codemonkey.org.uk>
Cc: David Rientjes <rientjes@google.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agofs/xattr.c:listxattr(): fall back to vmalloc() if kmalloc() failed
Andrew Morton [Thu, 5 Apr 2012 21:25:07 +0000 (14:25 -0700)]
fs/xattr.c:listxattr(): fall back to vmalloc() if kmalloc() failed

This allocation can be as large as 64k.  As David points out, "falling
back to vmalloc here is much better solution than failing to retreive
the attribute - it will work no matter how fragmented memory gets.  That
means we don't get incomplete backups occurring after days or months of
uptime and successful backups".

Cc: Dave Chinner <david@fromorbit.com>
Cc: Dave Jones <davej@codemonkey.org.uk>
Cc: David Rientjes <rientjes@google.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agofs/xattr.c: suppress page allocation failure warnings from sys_listxattr()
Dave Jones [Thu, 5 Apr 2012 21:25:06 +0000 (14:25 -0700)]
fs/xattr.c: suppress page allocation failure warnings from sys_listxattr()

This size is user controllable, up to a maximum of XATTR_LIST_MAX (64k).
So it's trivial for someone to trigger a stream of order:4 page
allocation errors.

Signed-off-by: Dave Jones <davej@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Dave Chinner <david@fromorbit.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agosysrq: use SEND_SIG_FORCED instead of force_sig()
Anton Vorontsov [Thu, 5 Apr 2012 21:25:05 +0000 (14:25 -0700)]
sysrq: use SEND_SIG_FORCED instead of force_sig()

Change send_sig_all() to use do_send_sig_info(SEND_SIG_FORCED) instead
of force_sig(SIGKILL).  With the recent changes we do not need force_ to
kill the CLONE_NEWPID tasks.

And this is more correct.  force_sig() can race with the exiting thread,
while do_send_sig_info(group => true) kill the whole process.

Some more notes from Oleg Nesterov:

> Just one note. This change makes no difference for sysrq_handle_kill().
> But it obviously changes the behaviour sysrq_handle_term(). I think
> this is fine, if you want to really kill the task which blocks/ignores
> SIGTERM you can use sysrq_handle_kill().
>
> Even ignoring the reasons why force_sig() is simply wrong here,
> force_sig(SIGTERM) looks strange. The task won't be killed if it has
> a handler, but SIG_IGN can't help. However if it has the handler
> but blocks SIGTERM temporary (this is very common) it will be killed.

Also,

> force_sig() can't kill the process if the main thread has already
> exited. IOW, it is trivial to create the process which can't be
> killed by sysrq.

So, this patch fixes the issue.

Suggested-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org>
Cc: Alan Cox <alan@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoproc: fix mount -t proc -o AAA
Vasiliy Kulikov [Thu, 5 Apr 2012 21:25:04 +0000 (14:25 -0700)]
proc: fix mount -t proc -o AAA

The proc_parse_options() call from proc_mount() runs only once at boot
time.  So on any later mount attempt, any mount options are ignored
because ->s_root is already initialized.

As a consequence, "mount -o <options>" will ignore the options.  The
only way to change mount options is "mount -o remount,<options>".

To fix this, parse the mount options unconditionally.

Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>
Reported-by: Arkadiusz Miskiewicz <a.miskiewicz@gmail.com>
Tested-by: Arkadiusz Miskiewicz <a.miskiewicz@gmail.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoARM: fix __io macro for PCMCIA
Rob Herring [Wed, 4 Apr 2012 22:48:04 +0000 (17:48 -0500)]
ARM: fix __io macro for PCMCIA

With commit c334bc1 (ARM: make mach/io.h include optional), PCMCIA was
broken. PCMCIA depends on __io() returning a valid i/o address, and most
ARM platforms require IO_SPACE_LIMIT be set to 0xffffffff for PCMCIA. This
needs a better fix with a fixed i/o address mapping, but for now we just
restore things to the previous behavior.

This fixes at91, omap1, pxa and sa11xx. pxa needs io.h if PCI is enabled,
but PCMCIA is not. sa11xx already has IO_SPACE_LIMIT set to 0xffffffff,
so it doesn't need an io.h.

Signed-off-by: Rob Herring <rob.herring@calxeda.com>
Cc: Joachim Eastwood <joachim.eastwood@jotron.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Andrew Victor <linux@maxim.org.za>
Cc: Nicolas Ferre <nicolas.ferre@atmel.com>
Cc: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com>
Tested-by: Paul Parsons <lost.distance@yahoo.com> (pxa270)
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Olof Johansson <olof@lixom.net>
12 years agoMerge branch 'v3.4-samsung-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Thu, 5 Apr 2012 17:49:45 +0000 (10:49 -0700)]
Merge branch 'v3.4-samsung-fixes-1' of git://git./linux/kernel/git/kgene/linux-samsung into fixes

* 'v3.4-samsung-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: EXYNOS: Fix compiler warning in dma.c file
  ARM: EXYNOS: fix ISO C90 warning
  ARM: EXYNOS: use chip_id reg in uncompress to select uart base phys
  ARM: EXYNOS: fix CONFIG_DEBUG_LL
  ARM: S3C24XX: fix missing common.h in mach-s3c24xx/

12 years agoMerge branch 'v3.4-cpufreq-exynos-fixes' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Thu, 5 Apr 2012 17:48:56 +0000 (10:48 -0700)]
Merge branch 'v3.4-cpufreq-exynos-fixes' of git://git./linux/kernel/git/kgene/linux-samsung into fixes

* 'v3.4-cpufreq-exynos-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  EXYNOS: fix dependency for EXYNOS_CPUFREQ

12 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux...
Olof Johansson [Thu, 5 Apr 2012 17:44:59 +0000 (10:44 -0700)]
Merge branch 'fixes' of git://git./linux/kernel/git/tmlind/linux-omap into fixes

From Tony Lindgren:

Note that this also contains a set of fixes that are not regressions
or oopses to properly deal with the smsc911x regulator issue.

Basically the regulators must be per board file as the regulators
can also come from drivers, such as twl4030. So it's best to dumb
down gpmc-smsc911x.c to not even care about the regulators.

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
  ARM: OMAP: fix section mismatches in usb-host.c
  ARM: OMAP2+: Fix omap2+ build error
  ARM: OMAP2+: smsc911x: Add fixed board regulators
  ARM: OMAP2+: smsc911x: Remove regulator support from gmpc-smsc911x
  ARM: OMAP2+: smsc911x: Remove unused rate calculation
  ARM: OMAP2+ smsc911x: Fix possible stale smsc911x flags
  ARM: OMAP2+: smsc911x: Remove odd gpmc_cfg/board_data redirection
  ARM: OMAP3+: fix oops triggered in omap_prcm_register_chain_handler(v1)
  ARM: OMAP2+: OPP: allow OPP enumeration to continue if device is not present
  arm: omap3: pm34xx.c: Replace printk() with appropriate pr_*()
  arm: omap3: pm34xx.c: Fix omap3_pm_init() error out paths
  ARM: OMAP4: Workaround the OCP synchronisation issue with 32K synctimer.
  ARM: OMAP4: prm: fix interrupt register offsets
  ARM: OMAP: hwmod: Use sysc_fields->srst_shift and get rid of hardcoded SYSC_TYPE2_SOFTRESET_MASK

12 years agoMerge branch 'at91-fixes' of git://github.com/at91linux/linux-at91 into fixes
Olof Johansson [Thu, 5 Apr 2012 17:43:06 +0000 (10:43 -0700)]
Merge branch 'at91-fixes' of git://github.com/at91linux/linux-at91 into fixes

* 'at91-fixes' of git://github.com/at91linux/linux-at91:
  ARM: at91: dt: remove unit-address part for memory nodes
  ARM: at91: fix check of valid GPIO for SPI and USB
  USB: ehci-atmel: add needed of.h header file
  ARM: at91/NAND DT bindings: add comments
  ARM: at91/at91sam9x5.dtsi: fix NAND ale/cle in DT file
  USB: ohci-at91: trivial return code name change
  USB: ohci-at91: change maximum number of ports
  USB: ohci-at91: rework and fix initialization
  ARM: at91/dts: USB host vbus is active low
  ARM: at91/USB host: specify and handle properly vbus_pin_active_low
  USB: ohci-at91: fix vbus_pin_active_low handling
  ARM: at91/at91sam9x5: add clkdev entries for DMA controllers

12 years agoMerge branch 'for-3.4/fixes-for-rc2' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Thu, 5 Apr 2012 17:41:49 +0000 (10:41 -0700)]
Merge branch 'for-3.4/fixes-for-rc2' of git://git./linux/kernel/git/swarren/linux-tegra into fixes

* 'for-3.4/fixes-for-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/swarren/linux-tegra:
  gpio: tegra: Iterate over the correct number of banks
  gpio: tegra: fix register address calculations for Tegra30

12 years agoARM: EXYNOS: Fix compiler warning in dma.c file
Sachin Kamat [Thu, 5 Apr 2012 15:01:13 +0000 (08:01 -0700)]
ARM: EXYNOS: Fix compiler warning in dma.c file

Fixes the following warning:
warning: 'dma_dmamask' defined but not used [-Wunused-variable]

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
12 years agoARM: EXYNOS: fix ISO C90 warning
Il Han [Thu, 5 Apr 2012 14:59:36 +0000 (07:59 -0700)]
ARM: EXYNOS: fix ISO C90 warning

ISO C90 forbids mixed declarations and code.
Fix it.

Signed-off-by: Il Han <corone.il.han@gmail.com>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
12 years agoMerge branches 'clock_fixes_3.4rc', 'clockdomain_fixes_3.4rc', 'hsmmc_erratum_2_1_1_1...
Paul Walmsley [Thu, 5 Apr 2012 09:00:22 +0000 (03:00 -0600)]
Merge branches 'clock_fixes_3.4rc', 'clockdomain_fixes_3.4rc', 'hsmmc_erratum_2_1_1_128_refine_3.4rc1', 'hwmod_data_fixes_a_3.4rc', 'hwmod_fixes_a2_3.4rc' and 'powerdomain_fixes_a_3.4rc' into omap-fixes-a2-for-3.4rc-branch

12 years agoARM: OMAP2+: hwmod: Fix wrong SYSC_TYPE1_XXX_MASK bit definitions
Vaibhav Hiremath [Fri, 17 Feb 2012 11:26:01 +0000 (16:56 +0530)]
ARM: OMAP2+: hwmod: Fix wrong SYSC_TYPE1_XXX_MASK bit definitions

In the SYSC_TYPE1_XXX_MASK configuration, SYSC_XXX_SHIFT macro
is used which is not defined anywhere in the kernel.
Until now the build was going through successfully, since it
is not being used anywhere in kernel.

This bug got introduced by the commit
358f0e630d5409ab3837b86db3595560eae773b6 ("OMAP3: hwmod: support
to specify the offset position of various SYSCONFIG register bits.")

Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Acked-by: Benoit Cousson <b-cousson@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
12 years agoARM: OMAP2+: hwmod: Make omap_hwmod_softreset wait for reset status
Rajendra Nayak [Tue, 13 Mar 2012 17:25:24 +0000 (22:55 +0530)]
ARM: OMAP2+: hwmod: Make omap_hwmod_softreset wait for reset status

omap_hwmod_softreset() does not seem to wait for reset status
after doing a softreset. Make it use _ocp_softreset() instead
which does this correctly.

Signed-off-by: Rajendra Nayak <rnayak@ti.com>
Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Anand Gadiyar <gadiyar@ti.com>
Cc: Shubhrajyoti D <shubhrajyoti@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
12 years agoARM: OMAP2+: hwmod: Restore sysc after a reset
Rajendra Nayak [Tue, 13 Mar 2012 17:25:23 +0000 (22:55 +0530)]
ARM: OMAP2+: hwmod: Restore sysc after a reset

After a softreset, make sure the sysc settings are correctly
restored.

Reported-by: Anand Gadiyar <gadiyar@ti.com>
Signed-off-by: Rajendra Nayak <rnayak@ti.com>
Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Shubhrajyoti D <shubhrajyoti@ti.com>
[paul@pwsan.com: combined post-reset SYSCONFIG reload code into the
 _reset() function to avoid duplication and future mistakes]
Signed-off-by: Paul Walmsley <paul@pwsan.com>
12 years agoARM: OMAP2+: omap_hwmod: Allow io_ring wakeup configuration for all modules
Govindraj.R [Thu, 5 Apr 2012 08:59:32 +0000 (02:59 -0600)]
ARM: OMAP2+: omap_hwmod: Allow io_ring wakeup configuration for all modules

Some modules doesn't have SYSC_HAS_ENAWAKEUP bit available (ex: usb
host uhh module) in absence of this flag
omap_hwmod_enable/disable_wakeup avoids configuring pad mux wakeup
capability.

Configure sysc if SYSC_HAS_ENAWAKEUP is available and for other cases
try enabling/disabling wakeup from mux_pad pins.

Cc: Paul Walmsley <paul@pwsan.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Rajendra Nayak <rnayak@ti.com>
Signed-off-by: Govindraj.R <govindraj.raja@ti.com>
[paul@pwsan.com: updated function kerneldoc documentation]
Signed-off-by: Paul Walmsley <paul@pwsan.com>
12 years agoipv6: fix array index in ip6_mc_add_src()
RongQing.Li [Wed, 4 Apr 2012 16:47:04 +0000 (16:47 +0000)]
ipv6: fix array index in ip6_mc_add_src()

Convert array index from the loop bound to the loop index.

And remove the void type conversion to ip6_mc_del1_src() return
code, seem it is unnecessary, since ip6_mc_del1_src() does not
use __must_check similar attribute, no compiler will report the
warning when it is removed.

v2: enrich the commit header

Signed-off-by: RongQing.Li <roy.qing.li@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Thu, 5 Apr 2012 01:37:09 +0000 (18:37 -0700)]
Merge git://git.samba.org/sfrench/cifs-2.6

Pull CIFS fixes from Steve French.

* git://git.samba.org/sfrench/cifs-2.6:
  Fix UNC parsing on mount
  Remove unnecessary check for NULL in password parser
  CIFS: Fix VFS lock usage for oplocked files
  Revert "CIFS: Fix VFS lock usage for oplocked files"
  cifs: writing past end of struct in cifs_convert_address()
  cifs: silence compiler warnings showing up with gcc-4.7.0
  CIFS: Fix VFS lock usage for oplocked files

12 years agomlx4: allocate just enough pages instead of always 4 pages
Thadeu Lima de Souza Cascardo [Wed, 4 Apr 2012 09:40:40 +0000 (09:40 +0000)]
mlx4: allocate just enough pages instead of always 4 pages

The driver uses a 2-order allocation, which is too much on architectures
like ppc64, which has a 64KiB page. This particular allocation is used
for large packet fragments that may have a size of 512, 1024, 4096 or
fill the whole allocation. So, a minimum size of 16384 is good enough
and will be the same size that is used in architectures of 4KiB sized
pages.

This will avoid allocation failures that we see when the system is under
stress, but still has plenty of memory, like the one below.

This will also allow us to set the interface MTU to higher values like
9000, which was not possible on ppc64 without this patch.

Node 1 DMA: 737*64kB 37*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB 0*8192kB 0*16384kB = 51904kB
83137 total pagecache pages
0 pages in swap cache
Swap cache stats: add 0, delete 0, find 0/0
Free swap  = 10420096kB
Total swap = 10420096kB
107776 pages RAM
1184 pages reserved
147343 pages shared
28152 pages non-shared
netstat: page allocation failure. order:2, mode:0x4020
Call Trace:
[c0000001a4fa3770] [c000000000012f04] .show_stack+0x74/0x1c0 (unreliable)
[c0000001a4fa3820] [c00000000016af38] .__alloc_pages_nodemask+0x618/0x930
[c0000001a4fa39a0] [c0000000001a71a0] .alloc_pages_current+0xb0/0x170
[c0000001a4fa3a40] [d00000000dcc3e00] .mlx4_en_alloc_frag+0x200/0x240 [mlx4_en]
[c0000001a4fa3b10] [d00000000dcc3f8c] .mlx4_en_complete_rx_desc+0x14c/0x250 [mlx4_en]
[c0000001a4fa3be0] [d00000000dcc4eec] .mlx4_en_process_rx_cq+0x62c/0x850 [mlx4_en]
[c0000001a4fa3d20] [d00000000dcc5150] .mlx4_en_poll_rx_cq+0x40/0x90 [mlx4_en]
[c0000001a4fa3dc0] [c0000000004e2bb8] .net_rx_action+0x178/0x450
[c0000001a4fa3eb0] [c00000000009c9b8] .__do_softirq+0x118/0x290
[c0000001a4fa3f90] [c000000000031df8] .call_do_softirq+0x14/0x24
[c000000184c3b520] [c00000000000e700] .do_softirq+0xf0/0x110
[c000000184c3b5c0] [c00000000009c6d4] .irq_exit+0xb4/0xc0
[c000000184c3b640] [c00000000000e964] .do_IRQ+0x144/0x230

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Tested-by: Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge tag 'for_linus-3.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/jwesse...
Linus Torvalds [Thu, 5 Apr 2012 00:26:08 +0000 (17:26 -0700)]
Merge tag 'for_linus-3.4-rc2' of git://git./linux/kernel/git/jwessel/kgdb

Pull KGDB/KDB regression fixes from Jason Wessel:
 - Fix a Smatch warning that appeared in the 3.4 merge window
 - Fix kgdb test suite with SMP for all archs without HW single stepping
 - Fix kgdb sw breakpoints with CONFIG_DEBUG_RODATA=y limitations on x86
 - Fix oops on kgdb test suite with CONFIG_DEBUG_RODATA
 - Fix kgdb test suite with SMP for all archs with HW single stepping

* tag 'for_linus-3.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/jwessel/kgdb:
  x86,kgdb: Fix DEBUG_RODATA limitation using text_poke()
  kgdb,debug_core: pass the breakpoint struct instead of address and memory
  kgdbts: (2 of 2) fix single step awareness to work correctly with SMP
  kgdbts: (1 of 2) fix single step awareness to work correctly with SMP
  kgdbts: Fix kernel oops with CONFIG_DEBUG_RODATA
  kdb: Fix smatch warning on dbg_io_ops->is_console

12 years agoMerge branch 'for-linus' of git://git.linaro.org/people/mszyprowski/linux-dma-mapping
Linus Torvalds [Thu, 5 Apr 2012 00:13:43 +0000 (17:13 -0700)]
Merge branch 'for-linus' of git://git.linaro.org/people/mszyprowski/linux-dma-mapping

Pull DMA mapping branch from Marek Szyprowski:
 "Short summary for the whole series:

  A few limitations have been identified in the current dma-mapping
  design and its implementations for various architectures.  There exist
  more than one function for allocating and freeing the buffers:
  currently these 3 are used dma_{alloc, free}_coherent,
  dma_{alloc,free}_writecombine, dma_{alloc,free}_noncoherent.

  For most of the systems these calls are almost equivalent and can be
  interchanged.  For others, especially the truly non-coherent ones
  (like ARM), the difference can be easily noticed in overall driver
  performance.  Sadly not all architectures provide implementations for
  all of them, so the drivers might need to be adapted and cannot be
  easily shared between different architectures.  The provided patches
  unify all these functions and hide the differences under the already
  existing dma attributes concept.  The thread with more references is
  available here:

    http://www.spinics.net/lists/linux-sh/msg09777.html

  These patches are also a prerequisite for unifying DMA-mapping
  implementation on ARM architecture with the common one provided by
  dma_map_ops structure and extending it with IOMMU support.  More
  information is available in the following thread:

    http://thread.gmane.org/gmane.linux.kernel.cross-arch/12819

  More works on dma-mapping framework are planned, especially in the
  area of buffer sharing and managing the shared mappings (together with
  the recently introduced dma_buf interface: commit d15bd7ee445d
  "dma-buf: Introduce dma buffer sharing mechanism").

  The patches in the current set introduce a new alloc/free methods
  (with support for memory attributes) in dma_map_ops structure, which
  will later replace dma_alloc_coherent and dma_alloc_writecombine
  functions."

People finally started piping up with support for merging this, so I'm
merging it as the last of the pending stuff from the merge window.
Looks like pohmelfs is going to wait for 3.5 and more external support
for merging.

* 'for-linus' of git://git.linaro.org/people/mszyprowski/linux-dma-mapping:
  common: DMA-mapping: add NON-CONSISTENT attribute
  common: DMA-mapping: add WRITE_COMBINE attribute
  common: dma-mapping: introduce mmap method
  common: dma-mapping: remove old alloc_coherent and free_coherent methods
  Hexagon: adapt for dma_map_ops changes
  Unicore32: adapt for dma_map_ops changes
  Microblaze: adapt for dma_map_ops changes
  SH: adapt for dma_map_ops changes
  Alpha: adapt for dma_map_ops changes
  SPARC: adapt for dma_map_ops changes
  PowerPC: adapt for dma_map_ops changes
  MIPS: adapt for dma_map_ops changes
  X86 & IA64: adapt for dma_map_ops changes
  common: dma-mapping: introduce generic alloc() and free() methods

12 years agostmmac: re-add IFF_UNICAST_FLT for dwmac1000
Marc Kleine-Budde [Tue, 3 Apr 2012 22:13:01 +0000 (22:13 +0000)]
stmmac: re-add IFF_UNICAST_FLT for dwmac1000

In commit (bfab27a stmmac: add the experimental PCI support) the
IFF_UNICAST_FLT flag has been removed from the stmmac_mac_device_setup()
function. This patch re-adds the flag.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Clear MDC/MDIO warning message
Yaniv Rosner [Wed, 4 Apr 2012 01:40:02 +0000 (01:40 +0000)]
bnx2x: Clear MDC/MDIO warning message

This patch clears a warning message of "MDC/MDIO access timeout" which may
appear when interface is loaded due to missing clock setting before resetting
the LED, and starting periodic function too early.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Fix BCM57711+BCM84823 link issue
Yaniv Rosner [Wed, 4 Apr 2012 01:29:02 +0000 (01:29 +0000)]
bnx2x: Fix BCM57711+BCM84823 link issue

Fix a link problem on the second port of BCM57711 + BCM84823 boards due to
incorrect macro usage.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Clear BCM84833 LED after fan failure
Yaniv Rosner [Wed, 4 Apr 2012 01:29:01 +0000 (01:29 +0000)]
bnx2x: Clear BCM84833 LED after fan failure

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Fix BCM84833 PHY FW version presentation
Yaniv Rosner [Wed, 4 Apr 2012 01:29:00 +0000 (01:29 +0000)]
bnx2x: Fix BCM84833 PHY FW version presentation

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Fix link issue for BCM8727 boards.
Yaniv Rosner [Wed, 4 Apr 2012 01:28:59 +0000 (01:28 +0000)]
bnx2x: Fix link issue for BCM8727 boards.

This patch fixes a link problem on BCM57712 + BCM8727 designs in which the TX
laser is controller by GPIO, after 1.60.xx drivers were previously loaded.
On these designs the TX_LASER is enabled by logic AND between the PHY
(through MDIO), and the GPIO. When an old driver is used, it disables the
MDIO part, hence the GPIO control had no affect de facto.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Restore 1G LED on BCM57712+BCM8727 designs.
Yaniv Rosner [Wed, 4 Apr 2012 01:28:58 +0000 (01:28 +0000)]
bnx2x: Restore 1G LED on BCM57712+BCM8727 designs.

Fix no-LED problem when link speed is 1G on BCM57712 + BCM8727 designs, by
removing a logic error checking for a different PHY.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Fix BCM578x0-SFI pre-emphasis settings
Yaniv Rosner [Wed, 4 Apr 2012 01:28:57 +0000 (01:28 +0000)]
bnx2x: Fix BCM578x0-SFI pre-emphasis settings

Fix 578x0-SFI pre-emphasis settings per HW recommendations to achieve better
link strength.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Fix BCM57810-KR AN speed transition
Yaniv Rosner [Wed, 4 Apr 2012 01:28:56 +0000 (01:28 +0000)]
bnx2x: Fix BCM57810-KR AN speed transition

BCM57810-KR link may not come up in 1G after running loopback test, so set
the relevant registers to their default values before starting KR autoneg.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: Fix BCM57810-KR FC
Yaniv Rosner [Wed, 4 Apr 2012 01:28:55 +0000 (01:28 +0000)]
bnx2x: Fix BCM57810-KR FC

Fix 57810-KR flow-control handling link is achieved via CL37 AN.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobnx2x: PFC fix
Yaniv Rosner [Wed, 4 Apr 2012 01:28:54 +0000 (01:28 +0000)]
bnx2x: PFC fix

Fix a problem in which PFC frames are not honored, due to incorrect link
attributes synchronization following PMF migration, and verify PFC XON is not
stuck from previous link change.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agosky2: copy received packets on inefficient unaligned architecture
stephen hemminger [Wed, 4 Apr 2012 12:10:27 +0000 (12:10 +0000)]
sky2: copy received packets on inefficient unaligned architecture

Modified from original patch from Chris.

The sky2 driver has to have 8 byte alignment of receive buffer
on some chip versions. On architectures which don't support efficient
unaligned access this doesn't work very well. The solution is to
just copy all received packets which is what the driver already
does for small packets.

This allows the driver to be used on the Tilera TILEmpower-Gx, since
the tile architecture doesn't currently handle kernel unaligned accesses,
just userspace.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Acked-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet/bonding: correctly proxy slave neigh param setup ndo function
Shlomo Pongratz [Tue, 3 Apr 2012 22:56:20 +0000 (22:56 +0000)]
net/bonding: correctly proxy slave neigh param setup ndo function

The current implemenation was buggy for slaves who use ndo_neigh_setup,
since the networking stack invokes the bonding device ndo entry (from
neigh_params_alloc) before any devices are enslaved, and the bonding
driver can't further delegate the call at that point in time. As a
result when bonding IPoIB devices, the neigh_cleanup hasn't been called.

Fix that by deferring the actual call into the slave ndo_neigh_setup
from the time the bonding neigh_setup is called.

Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet/bonding: emit address change event also in bond_release
Shlomo Pongratz [Tue, 3 Apr 2012 22:56:19 +0000 (22:56 +0000)]
net/bonding: emit address change event also in bond_release

commit 7d26bb103c4 "bonding: emit event when bonding changes MAC" didn't
take care to emit the NETDEV_CHANGEADDR event in bond_release, where bonding
actually changes the mac address (to all zeroes). As a result the neighbours
aren't deleted by the core networking code (which does so upon getting that
event).

Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agosctp: Allow struct sctp_event_subscribe to grow without breaking binaries
Thomas Graf [Tue, 3 Apr 2012 22:17:53 +0000 (22:17 +0000)]
sctp: Allow struct sctp_event_subscribe to grow without breaking binaries

getsockopt(..., SCTP_EVENTS, ...) performs a length check and returns
an error if the user provides less bytes than the size of struct
sctp_event_subscribe.

Struct sctp_event_subscribe needs to be extended by an u8 for every
new event or notification type that is added.

This obviously makes getsockopt fail for binaries that are compiled
against an older versions of <net/sctp/user.h> which do not contain
all event types.

This patch changes getsockopt behaviour to no longer return an error
if not enough bytes are being provided by the user. Instead, it
returns as much of sctp_event_subscribe as fits into the provided buffer.

This leads to the new behavior that users see what they have been aware
of at compile time.

The setsockopt(..., SCTP_EVENTS, ...) API is already behaving like this.

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge tag 'pm-for-3.4-part-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Wed, 4 Apr 2012 21:26:40 +0000 (14:26 -0700)]
Merge tag 'pm-for-3.4-part-2' of git://git./linux/kernel/git/rafael/linux-pm

Pull more power management updates from Rafael Wysocki:
 - Patch series that hopefully fixes races between the freezer and
   request_firmware() and request_firmware_nowait() for good, with two
   cleanups from Stephen Boyd on top.
 - Runtime PM fix from Alan Stern preventing tasks from getting stuck
   indefinitely in the runtime PM wait queue.
 - Device PM QoS update from MyungJoo Ham introducing a new variant of
   pm_qos_update_request() allowing the callers to specify a timeout.

* tag 'pm-for-3.4-part-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  PM / QoS: add pm_qos_update_request_timeout() API
  firmware_class: Move request_firmware_nowait() to workqueues
  firmware_class: Reorganize fw_create_instance()
  PM / Sleep: Mitigate race between the freezer and request_firmware()
  PM / Sleep: Move disabling of usermode helpers to the freezer
  PM / Hibernate: Disable usermode helpers right before freezing tasks
  firmware_class: Do not warn that system is not ready from async loads
  firmware_class: Split _request_firmware() into three functions, v2
  firmware_class: Rework usermodehelper check
  PM / Runtime: don't forget to wake up waitqueue on failure

12 years agoARM: OMAP3: clock data: fill in some missing clockdomains
Paul Walmsley [Wed, 4 Apr 2012 16:20:15 +0000 (10:20 -0600)]
ARM: OMAP3: clock data: fill in some missing clockdomains

Several clocks are missing clockdomains.  This can cause problems with
the hwmod and power management code.  Fill these in.

Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Matt Porter <mporter@ti.com>
Cc: Vaibhav Hiremath <hvaibhav@ti.com>
12 years agoARM: OMAP4: clock data: Force a DPLL clkdm/pwrdm ON before a relock
Rajendra Nayak [Wed, 4 Apr 2012 16:20:01 +0000 (10:20 -0600)]
ARM: OMAP4: clock data: Force a DPLL clkdm/pwrdm ON before a relock

All DPLLs except USB are in ALWON powerdomain. Make sure the
clkdm/pwrdm for USB DPLL (l3init) is turned on before attempting
a DPLL relock. So, mark the database accordingly.

Without this fix, it was seen that DPLL relock fails while testing
relock in a loop of USB DPLL.

Cc: Nishanth Menon <nm@ti.com>
Tested-by: Ameya Palande <ameya.palande@ti.com>
Signed-off-by: Rajendra Nayak <rnayak@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
12 years agoARM: OMAP4: clock data: fix mult and div mask for USB_DPLL
Ameya Palande [Wed, 4 Apr 2012 16:19:31 +0000 (10:19 -0600)]
ARM: OMAP4: clock data: fix mult and div mask for USB_DPLL

According to OMAP4 TRM Table 3-1183, CM_CLKSEL_DPLL_USB register defines
following fields for multiplication and division factors:

DPLL_MULT (bits 19:8) DPLL multiplier factor (2 to 4095)
DPLL_DIV (bits 7:0) DPLL divider factor (0 to 255)

Acked-by: Benoit Cousson <b-cousson@ti.com>
Signed-off-by: Ameya Palande <ameya.palande@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
12 years agoARM: OMAP2+: powerdomain: Wait for powerdomain transition in pwrdm_state_switch()
Santosh Shilimkar [Mon, 12 Mar 2012 14:34:32 +0000 (20:04 +0530)]
ARM: OMAP2+: powerdomain: Wait for powerdomain transition in pwrdm_state_switch()

Commit b1cbdb00d ("OMAP: clockdomain: Wait for powerdomain to be ON
when using clockdomain force wakeup") was assuming that
pwrdm_state_switch() does wait for the powerdomain transition which is
not the case.  The missing wait for the powerdomain transition
violates the sequence which the hardware expects, which causes power
management failures on some devices.

Fix this API by adding the pwrdm_wait_transition().

Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Cc: Rajendra Nayak <rnayak@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
[paul@pwsan.com: added some more details in the commit log]
Signed-off-by: Paul Walmsley <paul@pwsan.com>
12 years agogpio: tegra: Iterate over the correct number of banks
Stephen Warren [Fri, 16 Mar 2012 23:37:24 +0000 (17:37 -0600)]
gpio: tegra: Iterate over the correct number of banks

When Tegra30 support was added to the Tegra GPIO driver, a few places
which iterated over all banks were not converted to use the variable
tegra_gpio_bank_count rather than hard-coding the bank count. Fix this.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Olof Johansson <olof@lixom.net>
12 years agogpio: tegra: fix register address calculations for Tegra30
Stephen Warren [Fri, 16 Mar 2012 23:35:08 +0000 (17:35 -0600)]
gpio: tegra: fix register address calculations for Tegra30

Tegra20 and Tegra30 share the same register layout within registers, but
the addresses of the registers is a little different. Fix the driver to
cope with this.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Olof Johansson <olof@lixom.net>
12 years agoMerge branch 'selinux' ("struct common_audit_data" sanitizer)
Linus Torvalds [Wed, 4 Apr 2012 17:11:24 +0000 (10:11 -0700)]
Merge branch 'selinux' ("struct common_audit_data" sanitizer)

Merge common_audit_data cleanup patches from Eric Paris.

This is really too late, but it's a long-overdue cleanup of the costly
wrapper functions for the security layer.

The "struct common_audit_data" is used all over in critical paths,
allocated and initialized on the stack.  And used to be much too large,
causing not only unnecessarily big stack frames but the clearing of the
(mostly useless) data was also very visible in profiles.

As a particular example, in one microbenchmark for just doing "stat()"
over files a lot, selinux_inode_permission() used 7% of the CPU time.
That's despite the fact that it doesn't actually *do* anything: it is
just a helper wrapper function in the selinux security layer.

This patch-series shrinks "struct common_audit_data" sufficiently that
code generation for these kinds of wrapper functions is improved
noticeably, and we spend much less time just initializing data that we
will never use.

The functions still get called all the time, and it still shows up at
3.5+% in my microbenchmark, but it's quite a bit lower down the list,
and much less noticeable.

* Emailed patches from Eric Paris <eparis@redhat.com>:
  lsm_audit: don't specify the audit pre/post callbacks in 'struct common_audit_data'
  SELinux: do not allocate stack space for AVC data unless needed
  SELinux: remove avd from slow_avc_audit()
  SELinux: remove avd from selinux_audit_data
  LSM: shrink the common_audit_data data union
  LSM: shrink sizeof LSM specific portion of common_audit_data

12 years agoMerge tag 'regmap-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap
Linus Torvalds [Wed, 4 Apr 2012 17:10:14 +0000 (10:10 -0700)]
Merge tag 'regmap-3.4' of git://git./linux/kernel/git/broonie/regmap

Pull a single regmap fix from Mark Brown:
 "A simple bug that's been lurking for a while but not terribly visible
  since a high proportion of chips have no register 0 so the normal
  failure is that we end up doing a bit of extra I/O."

* tag 'regmap-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:
  regmap: rbtree: Fix register default look-up in sync

12 years agoMerge tag 'regulator-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie...
Linus Torvalds [Wed, 4 Apr 2012 17:09:30 +0000 (10:09 -0700)]
Merge tag 'regulator-3.4' of git://git./linux/kernel/git/broonie/regulator

Pull regulator fixes from Mark Brown:
 "A bunch of smallish fixes that came up during the merge window as
  things got more testing - even more fixes from Axel, a fix for error
  handling in more complex systems using -EPROBE_DEFER and a couple of
  small fixes for the new dummy regulators."

* tag 'regulator-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: Remove non-existent parameter from fixed-helper.c kernel doc
  regulator: Fix setting new voltage in s5m8767_set_voltage
  regulator: fix sysfs name collision between dummy and fixed dummy regulator
  regulator: Fix deadlock on removal of regulators with supplies
  regulator: Fix comments in include/linux/regulator/machine.h
  regulator: Only update [LDOx|DCx]_HIB_MODE bits in wm8350_[ldo|dcdc]_set_suspend_disable
  regulator: Fix setting low power mode for wm831x aldo
  regulator: Return microamps in wm8350_isink_get_current
  regulator: wm8350: Fix the logic to choose best current limit setting
  regulator: wm831x-isink: Fix the logic to choose best current limit setting
  regulator: wm831x-dcdc: Fix the logic to choose best current limit setting
  regulator: anatop: patching to device-tree property "reg".
  regulator: Do proper shift to set correct bit for DC[2|5]_HIB_MODE setting
  regulator: Fix restoring pmic.dcdcx_hib_mode settings in wm8350_dcdc_set_suspend_enable
  regulator: Fix unbalanced lock/unlock in mc13892_regulator_probe error path
  regulator: Fix set and get current limit for wm831x_buckv
  regulator: tps6586x: Fix list minimal voltage setting for LDO0

12 years agoEXYNOS: fix dependency for EXYNOS_CPUFREQ
Kukjin Kim [Wed, 4 Apr 2012 17:09:03 +0000 (10:09 -0700)]
EXYNOS: fix dependency for EXYNOS_CPUFREQ

This fixes the CPUFREQ dependency for regarding EXYNOS SoCs
such as EXYNOS4210, EXYNOS4X12 and EXYNOS5250. Its cpufreq
driver should be built with selection of SoC arch part.

Reported-by: Russell King <rmk+kernel@arm.linux.org.uk>
Acked-by: Dave Jones <davej@redhat.com>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
12 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 4 Apr 2012 17:04:42 +0000 (10:04 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar.

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/p4: Add format attributes
  tracing, sched, vfs: Fix 'old_pid' usage in trace_sched_process_exec()

12 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 4 Apr 2012 17:04:01 +0000 (10:04 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar.

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86, kvm: Call restore_sched_clock_state() only after %gs is initialized
  x86: Use -mno-avx when available
  x86: Remove the ancient and deprecated disable_hlt() and enable_hlt() facility
  x86: Preserve lazy irq disable semantics in fixup_irqs()

12 years agoARM: at91: dt: remove unit-address part for memory nodes
Ludovic Desroches [Mon, 2 Apr 2012 18:44:20 +0000 (20:44 +0200)]
ARM: at91: dt: remove unit-address part for memory nodes

Because of the inclusion of skeleton.dtsi, the memory node is
named "memory" we where not modifying the already included one
but creating a new one. It caused bad memory node detection during
early_init_dt_scan_memory() so we modify them.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Cc: devicetree-discuss@lists.ozlabs.org
12 years agoARM: at91: fix check of valid GPIO for SPI and USB
Nicolas Ferre [Wed, 28 Mar 2012 09:58:58 +0000 (11:58 +0200)]
ARM: at91: fix check of valid GPIO for SPI and USB

SPI chip select pins have to be checked by gpio_is_valid().
The USB host overcurrent_pin checking was missing.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
12 years agoUSB: ehci-atmel: add needed of.h header file
Nicolas Ferre [Tue, 27 Mar 2012 16:23:31 +0000 (18:23 +0200)]
USB: ehci-atmel: add needed of.h header file

Compilation error in case of non-DT configuration without this
of.h header file.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
12 years agoARM: at91/NAND DT bindings: add comments
Nicolas Ferre [Thu, 22 Mar 2012 13:48:47 +0000 (14:48 +0100)]
ARM: at91/NAND DT bindings: add comments

Add comments to NAND "gpios" property to make it clearer.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>