kernel/kernel-generic.git
11 years agoMerge tag 'batman-adv-for-davem' of git://git.open-mesh.org/linux-merge
David S. Miller [Thu, 30 May 2013 22:53:18 +0000 (15:53 -0700)]
Merge tag 'batman-adv-for-davem' of git://git.open-mesh.org/linux-merge

Included changes:
- reduce broadcast overhead on non-lossy wired links
- fix typos in kernel doc
- use eth_hdr() when possible
- use netdev_allock_skb_ip_align() and don't deal with NET_IP_ALIGN
- change VID semantic in the BLA component
- other minor cleanups and code refactoring

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet/usb/kalmia: use %*phC to dump small buffers
Andy Shevchenko [Wed, 29 May 2013 07:22:30 +0000 (10:22 +0300)]
net/usb/kalmia: use %*phC to dump small buffers

Instead of dereferencing pointer and put values on stack we could use nice
%*phC specifier.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobonding: trivial: update the comments to reflect the reality
Veaceslav Falico [Tue, 28 May 2013 01:26:13 +0000 (01:26 +0000)]
bonding: trivial: update the comments to reflect the reality

Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobonding: trivial: remove unused parameter from alb_swap_mac_addr()
Veaceslav Falico [Mon, 27 May 2013 23:14:51 +0000 (23:14 +0000)]
bonding: trivial: remove unused parameter from alb_swap_mac_addr()

After b924551 ("bonding: fix enslaving in alb mode when link down") we
don't need the bond parameter in alb_swap_mac_addr(), so remove it.

Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agovxlan: defer vxlan init as late as possible
Cong Wang [Mon, 27 May 2013 22:35:53 +0000 (22:35 +0000)]
vxlan: defer vxlan init as late as possible

When vxlan is compiled as builtin, its init code
runs before IPv6 init, this could cause problems
if we create IPv6 socket in the latter patch.

Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agovxlan: use unsigned int instead of unsigned
Cong Wang [Mon, 27 May 2013 22:35:52 +0000 (22:35 +0000)]
vxlan: use unsigned int instead of unsigned

'unsigned int' is slightly better.

Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agovxlan: remove the unused rcu head from struct vxlan_rdst
Cong Wang [Mon, 27 May 2013 22:35:51 +0000 (22:35 +0000)]
vxlan: remove the unused rcu head from struct vxlan_rdst

Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet, ipv4, ipv6: Correct assignment of skb->network_header to skb->tail
Simon Horman [Tue, 28 May 2013 20:34:29 +0000 (20:34 +0000)]
net, ipv4, ipv6: Correct assignment of skb->network_header to skb->tail

This corrects an regression introduced by "net: Use 16bits for *_headers
fields of struct skbuff" when NET_SKBUFF_DATA_USES_OFFSET is not set. In
that case skb->tail will be a pointer however skb->network_header is now
an offset.

This patch corrects the problem by adding a wrapper to return skb tail as
an offset regardless of the value of NET_SKBUFF_DATA_USES_OFFSET. It seems
that skb->tail that this offset may be more than 64k and some care has been
taken to treat such cases as an error.

Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agosctp: Correct access to skb->{network, transport}_header
Simon Horman [Tue, 28 May 2013 20:34:28 +0000 (20:34 +0000)]
sctp: Correct access to skb->{network, transport}_header

This corrects an regression introduced by "net: Use 16bits for *_headers
fields of struct skbuff" when NET_SKBUFF_DATA_USES_OFFSET is not set. In
that case sk_buff_data_t will be a pointer, however,
skb->{network,transport}_header is now __u16.

Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoipv4: Correct comparisons and calculations using skb->tail and skb-transport_header
Simon Horman [Tue, 28 May 2013 20:34:27 +0000 (20:34 +0000)]
ipv4: Correct comparisons and calculations using skb->tail and skb-transport_header

This corrects an regression introduced by "net: Use 16bits for *_headers
fields of struct skbuff" when NET_SKBUFF_DATA_USES_OFFSET is not set. In
that case skb->tail will be a pointer whereas skb->transport_header
will be an offset from head. This is corrected by using wrappers that
ensure that comparisons and calculations are always made using pointers.

Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoipv6: Correct comparisons and calculations using skb->tail and skb-transport_header
Simon Horman [Tue, 28 May 2013 20:34:26 +0000 (20:34 +0000)]
ipv6: Correct comparisons and calculations using skb->tail and skb-transport_header

This corrects an regression introduced by "net: Use 16bits for *_headers
fields of struct skbuff" when NET_SKBUFF_DATA_USES_OFFSET is not set. In
that case skb->tail will be a pointer whereas skb->transport_header
will be an offset from head. This is corrected by using wrappers that
ensure that comparisons and calculations are always made using pointers.

Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: Correct comparisons and calculations using skb->tail and skb-transport_header
Simon Horman [Tue, 28 May 2013 20:34:25 +0000 (20:34 +0000)]
net: Correct comparisons and calculations using skb->tail and skb-transport_header

This corrects an regression introduced by "net: Use 16bits for *_headers
fields of struct skbuff" when NET_SKBUFF_DATA_USES_OFFSET is not set. In
that case skb->tail will be a pointer whereas skb->transport_header
will be an offset from head. This is corrected by using wrappers that
ensure that comparisons and calculations are always made using pointers.

Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agocxgb3: Correct comparisons and calculations using skb->tail and skb-transport_header
Simon Horman [Tue, 28 May 2013 20:34:24 +0000 (20:34 +0000)]
cxgb3: Correct comparisons and calculations using skb->tail and skb-transport_header

This corrects an regression introduced by "net: Use 16bits for *_headers
fields of struct skbuff" when NET_SKBUFF_DATA_USES_OFFSET is not set. In
that case skb->tail will be a pointer whereas skb->transport_header
will be an offset from head. This is corrected by using wrappers that
ensure that comparisons and calculations are always made using pointers.

Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoisdn: Correct comparison of skb->tail and skb-transport_header
Simon Horman [Tue, 28 May 2013 20:34:23 +0000 (20:34 +0000)]
isdn: Correct comparison of skb->tail and skb-transport_header

This corrects an regression introduced by "net: Use 16bits for *_headers
fields of struct skbuff" when NET_SKBUFF_DATA_USES_OFFSET is not set. In
that case skb->tail will be a pointer whereas skb->transport_header
will be an offset from head. This is corrected by using wrappers that
ensure that the comparison is always between pointers.

Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: always pass struct netdev_notifier_info to netdevice notifiers
Cong Wang [Wed, 29 May 2013 03:30:50 +0000 (11:30 +0800)]
net: always pass struct netdev_notifier_info to netdevice notifiers

commit 351638e7deeed2ec8ce451b53d3 (net: pass info struct via netdevice notifier)
breaks booting of my KVM guest, this is due to we still forget to pass
struct netdev_notifier_info in several places. This patch completes it.

Cc: Jiri Pirko <jiri@resnulli.us>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobatman-adv: Start new development cycle
Simon Wunderlich [Sat, 20 Apr 2013 15:15:09 +0000 (17:15 +0200)]
batman-adv: Start new development cycle

Signed-off-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
11 years agobatman-adv: Move call to batadv_nc_skb_forward() from routing.c to send.c
Martin Hundebøll [Sat, 20 Apr 2013 11:54:39 +0000 (13:54 +0200)]
batman-adv: Move call to batadv_nc_skb_forward() from routing.c to send.c

The call to batadv_nc_skb_forward() fits better in
batadv_send_skb_to_orig(), as this is where the actual next hop is
looked up.

To let the caller of batadv_send_skb_to_orig() know wether the skb is
transmitted, buffered or failed, the return value is changed from
boolean to int.

Signed-off-by: Martin Hundebøll <martin@hundeboll.net>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
11 years agobatman-adv: print the VID properly
Antonio Quartulli [Fri, 19 Apr 2013 16:07:01 +0000 (18:07 +0200)]
batman-adv: print the VID properly

Since the MSB bits of any vid variable are now used for
storing flags, print the vid properly by taking the flags
away and printing -1 in case of VID representing no real
VLAN.

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: change VID semantic in the BLA code
Antonio Quartulli [Fri, 19 Apr 2013 16:07:00 +0000 (18:07 +0200)]
batman-adv: change VID semantic in the BLA code

In order to make batman-adv fully vlan aware later, the
semantic used for variables storing the VLAN ID values has
to be changed in order to be adapted to the new one which
will be used batman-adv wide.

In particular, the VID has to be an "_unsigned_ short int"
and its 4 MSB will be used as a flag bitfield, while the
remaining 12 bits are used to store the real VID value

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Acked-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de>
11 years agobatman-adv: do not print orig nodes without nc neighbors on nc table print
Marek Lindner [Wed, 17 Apr 2013 20:56:03 +0000 (04:56 +0800)]
batman-adv: do not print orig nodes without nc neighbors on nc table print

Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
11 years agobatman-adv: Remove unnecessary INIT_HLIST_NODE() calls
Linus Lüssing [Fri, 19 Apr 2013 10:06:56 +0000 (12:06 +0200)]
batman-adv: Remove unnecessary INIT_HLIST_NODE() calls

There's no need to for an explicit hlist_node initialization if it is
added to a list right away, like it's the case with the
hlist_add_head()s here.

Signed-off-by: Linus Lüssing <linus.luessing@web.de>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
11 years agobatman-adv: pass a 16bit long flag argument to tt_global_add()
Antonio Quartulli [Thu, 18 Apr 2013 13:13:01 +0000 (15:13 +0200)]
batman-adv: pass a 16bit long flag argument to tt_global_add()

it may be the case that we want to store some local TT client flags in a global
entry, therefore the tt_global_add needs to get a proper argument for this

Signed-off-by: Antonio Quartulli <ordex@autistici.org>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: don't deal with NET_IP_ALIGN manually
Antonio Quartulli [Tue, 2 Apr 2013 20:28:44 +0000 (22:28 +0200)]
batman-adv: don't deal with NET_IP_ALIGN manually

Instead of dealing with NET_IP_ALIGN during allocation and
headroom reservation, it is possible to use
netdev_alloc_skb_ip_align() which transparently allocate
and reserve the correct amount of data

Signed-off-by: Antonio Quartulli <ordex@autistici.org>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: refactor batadv_tt_local_event()
Antonio Quartulli [Wed, 3 Apr 2013 09:15:33 +0000 (11:15 +0200)]
batman-adv: refactor batadv_tt_local_event()

Instead of passing a generic combination of flags as
argument, it is easier to pass the entire tt_common
structure (containing the flags already set) plus a
bitfield of event flags that will be unified with
the already existing ones before inserting the client
in the event queue.
In this way invocations of the modified function can be
simplified.

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: move batadv_slide_own_bcast_window to bat_iv_ogm.c
Antonio Quartulli [Wed, 17 Apr 2013 15:44:43 +0000 (17:44 +0200)]
batman-adv: move batadv_slide_own_bcast_window to bat_iv_ogm.c

batadv_slide_own_bcast_window() is used only in bat_iv_ogm.c
and it is currently touching only batman_iv specific
attributes.

Move it into bat_iv_ogm.c and make it static.

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: move ring_buffer helper functions in bat_iv_ogm
Antonio Quartulli [Mon, 8 Apr 2013 07:38:12 +0000 (09:38 +0200)]
batman-adv: move ring_buffer helper functions in bat_iv_ogm

the two lonely ring_buffer helper functions are used by the
bat_iv_ogm module only and therefore they can be moved
inside it.

Reported-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: use eth_hdr() when it makes sense
Antonio Quartulli [Mon, 8 Apr 2013 13:08:18 +0000 (15:08 +0200)]
batman-adv: use eth_hdr() when it makes sense

Instead of casting the result of skb_mac_header() to
"struct ethhdr *" every time, the eth_hdr inline function
can be use to beautify the code and improve its readability.

Signed-off-by: Antonio Quartulli <ordex@autistici.org>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: don't initialise batman_iv private members in hard-interface.c
Antonio Quartulli [Tue, 2 Apr 2013 10:16:53 +0000 (12:16 +0200)]
batman-adv: don't initialise batman_iv private members in hard-interface.c

hard-interface.c has to do not contain any routing algorithm
specific code.

Allocate the hard-interface with kzalloc() and remove any
useless and algorithm specific member initialisation

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: do not silently ignore wrong condition
Antonio Quartulli [Sat, 30 Mar 2013 16:22:00 +0000 (17:22 +0100)]
batman-adv: do not silently ignore wrong condition

Only one neigh_node per orig_node should match a given
neighbor address, therefore, if more than one matching
neigh_node is found, a WARNING has to be triggered to let
the user know that something is wrong in the originator
state instead of silently skipping the error.

Signed-off-by: Antonio Quartulli <ordex@autistici.org>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: don't check compat version twice
Antonio Quartulli [Thu, 28 Mar 2013 13:21:12 +0000 (14:21 +0100)]
batman-adv: don't check compat version twice

Compatibility version is checked upon packet reception
before calling any handler. For this reason it does need to
be checked once more in the handler itself.

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: don't check the source address twice
Antonio Quartulli [Mon, 25 Mar 2013 21:27:00 +0000 (22:27 +0100)]
batman-adv: don't check the source address twice

The source address has already been checked in
batadv_check_management_packet() upon packet reception and
therefore it does not need to be checked again in
ogm_process()

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: slightly improve neighbor creation debug message
Antonio Quartulli [Mon, 25 Mar 2013 12:54:45 +0000 (13:54 +0100)]
batman-adv: slightly improve neighbor creation debug message

print the interface along with the new neighbor mac address

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: drop useless argument seqno in neighbor creation
Antonio Quartulli [Mon, 25 Mar 2013 12:49:46 +0000 (13:49 +0100)]
batman-adv: drop useless argument seqno in neighbor creation

the sequence number is not stored in struct neigh_node,
therefore there is no need to pass such value to the
neigh_node creation procedure.

At the moment the value is only used by a debug message, but
given the fact that the seqno is not related to the neighbor
object, it is better to print it elsewhere.

Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
11 years agobatman-adv: fix typos in kernel doc & comments
Marek Lindner [Sun, 10 Mar 2013 11:29:15 +0000 (19:29 +0800)]
batman-adv: fix typos in kernel doc & comments

Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
11 years agobatman-adv: send each broadcast only once on non-wireless interfaces
Matthias Schiffer [Sat, 9 Mar 2013 22:14:23 +0000 (23:14 +0100)]
batman-adv: send each broadcast only once on non-wireless interfaces

While it makes sense to send each broadcast thrice on 802.11 (WLAN) interfaces
as broadcasts are often unreliable on these, there is no reason to do so on
other interface types.

The increased the overhead can be harmful on low-bandwidth links like VPN
connections over slow internet lines, therefore it is better to reduce the
number of broadcast packets sent on non-wireless links to one.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
11 years agobatman-adv: split batadv_is_wifi_iface() into two functions
Matthias Schiffer [Sat, 9 Mar 2013 22:14:22 +0000 (23:14 +0100)]
batman-adv: split batadv_is_wifi_iface() into two functions

Previously batadv_is_wifi_iface() did two things at once: looking up a
net_device from an interface index, and determining if it is a wifi device.

The second part is useful itself when the caller already has a net_device
reference.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
11 years agonet: Fix build warnings after mac_header and transport_header became __u16.
David S. Miller [Tue, 28 May 2013 20:15:50 +0000 (13:15 -0700)]
net: Fix build warnings after mac_header and transport_header became __u16.

net/core/skbuff.c: In function ‘__alloc_skb_head’:
net/core/skbuff.c:203:2: warning: large integer implicitly truncated to unsigned type [-Woverflow]
net/core/skbuff.c: In function ‘__alloc_skb’:
net/core/skbuff.c:279:2: warning: large integer implicitly truncated to unsigned type [-Woverflow]
net/core/skbuff.c:280:2: warning: large integer implicitly truncated to unsigned type [-Woverflow]
net/core/skbuff.c: In function ‘build_skb’:
net/core/skbuff.c:348:2: warning: large integer implicitly truncated to unsigned type [-Woverflow]
net/core/skbuff.c:349:2: warning: large integer implicitly truncated to unsigned type [-Woverflow]

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoarp: flush arp cache on IFF_NOARP change
Timo Teräs [Tue, 28 May 2013 01:30:23 +0000 (01:30 +0000)]
arp: flush arp cache on IFF_NOARP change

IFF_NOARP affects what kind of neighbor entries are created
(nud NOARP or nud INCOMPLETE). If the flag changes, flush the arp
cache to refresh all entries.

Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
v2->v3: shortened notifier_info struct name
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: pass changed flags along with NETDEV_CHANGE event
Jiri Pirko [Tue, 28 May 2013 01:30:22 +0000 (01:30 +0000)]
net: pass changed flags along with NETDEV_CHANGE event

Use new netdevice notifier infrastructure to pass along changed flags.

Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
v2->v3: shortened notifier_info struct name
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: pass info struct via netdevice notifier
Jiri Pirko [Tue, 28 May 2013 01:30:21 +0000 (01:30 +0000)]
net: pass info struct via netdevice notifier

So far, only net_device * could be passed along with netdevice notifier
event. This patch provides a possibility to pass custom structure
able to provide info that event listener needs to know.

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
v2->v3: fix typo on simeth
shortened dev_getter
shortened notifier_info struct name
v1->v2: fix notifier_call parameter in call_netdevice_notifier()
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobonding: remove ifenslave.c from kernel source
Cong Wang [Mon, 27 May 2013 15:49:16 +0000 (15:49 +0000)]
bonding: remove ifenslave.c from kernel source

As Stephen proposed:
Since bonding supports configuration via iproute (netlink) and sysfs, I think
it is time to purge the old ifenslave code out of Documentation/networking
and update the documentation.

Suggested-by: Stephen Hemminger <stephen@networkplumber.org>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Jay Vosburgh <fubar@us.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agodoc:networking: Fix typo in documentation/networking
Masanari Iida [Fri, 24 May 2013 07:05:59 +0000 (07:05 +0000)]
doc:networking: Fix typo in documentation/networking

Correct spelling typo

Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobonding: allow xmit hash policy change while bond dev is up
Nikolay Aleksandrov [Fri, 24 May 2013 00:59:47 +0000 (00:59 +0000)]
bonding: allow xmit hash policy change while bond dev is up

Since the xmit_hash_policy pointer is always valid and not dependent on
anything, we can change it while the bond device is up and running. The
only downside would be the out of order packets but that is a small price
to pay.

Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet/phy: Use module_spi_driver in spi_ks8995.c
Sachin Kamat [Mon, 27 May 2013 01:51:39 +0000 (01:51 +0000)]
net/phy: Use module_spi_driver in spi_ks8995.c

module_spi_driver() removes some boilerplate and makes the code
simpler.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonetpoll: remove return value from netpoll_rx_disable()
dingtianhong [Mon, 27 May 2013 19:53:31 +0000 (19:53 +0000)]
netpoll: remove return value from netpoll_rx_disable()

The netpoll_rx_disable() will always return 0, it is no use and looks wordy,
so remove the unnecessary code and get rid of it in _dev_open and _dev_close.

Signed-off-by: Ding Tianhong <dingtianhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: micrel : ks8851-ml: add dt support
Jean-Christophe PLAGNIOL-VILLARD [Thu, 23 May 2013 23:01:22 +0000 (23:01 +0000)]
net: micrel : ks8851-ml: add dt support

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'mpls_gso'
David S. Miller [Tue, 28 May 2013 05:51:06 +0000 (22:51 -0700)]
Merge branch 'mpls_gso'

Simon Horman says:

====================
In the case where a non-MPLS packet is received and an MPLS stack is
added it may well be the case that the original skb is GSO but the
NIC used for transmit does not support GSO of MPLS packets.

The aim of this short series is to provide GSO in software for MPLS packets
whose skbs are GSO.

Change since v4:

Update first patch of the series to use 16 bits for all *_headers
rather than just inner_*_headers

Simon Horman (2):
  net: Use 16bits for *_headers fields of struct skbuff
  MPLS: Add limited GSO support
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMPLS: Add limited GSO support
Simon Horman [Thu, 23 May 2013 21:02:52 +0000 (21:02 +0000)]
MPLS: Add limited GSO support

In the case where a non-MPLS packet is received and an MPLS stack is
added it may well be the case that the original skb is GSO but the
NIC used for transmit does not support GSO of MPLS packets.

The aim of this code is to provide GSO in software for MPLS packets
whose skbs are GSO.

SKB Usage:

When an implementation adds an MPLS stack to a non-MPLS packet it should do
the following to skb metadata:

* Set skb->inner_protocol to the old non-MPLS ethertype of the packet.
  skb->inner_protocol is added by this patch.

* Set skb->protocol to the new MPLS ethertype of the packet.

* Set skb->network_header to correspond to the
  end of the L3 header, including the MPLS label stack.

I have posted a patch, "[PATCH v3.29] datapath: Add basic MPLS support to
kernel" which adds MPLS support to the kernel datapath of Open vSwtich.
That patch sets the above requirements in datapath/actions.c:push_mpls()
and was used to exercise this code.  The datapath patch is against the Open
vSwtich tree but it is intended that it be added to the Open vSwtich code
present in the mainline Linux kernel at some point.

Features:

I believe that the approach that I have taken is at least partially
consistent with the handling of other protocols.  Jesse, I understand that
you have some ideas here.  I am more than happy to change my implementation.

This patch adds dev->mpls_features which may be used by devices
to advertise features supported for MPLS packets.

A new NETIF_F_MPLS_GSO feature is added for devices which support
hardware MPLS GSO offload.  Currently no devices support this
and MPLS GSO always falls back to software.

Alternate Implementation:

One possible alternate implementation is to teach netif_skb_features()
and skb_network_protocol() about MPLS, in a similar way to their
understanding of VLANs. I believe this would avoid the need
for net/mpls/mpls_gso.c and in particular the calls to
__skb_push() and __skb_push() in mpls_gso_segment().

I have decided on the implementation in this patch as it should
not introduce any overhead in the case where mpls_gso is not compiled
into the kernel or inserted as a module.

MPLS GSO suggested by Jesse Gross.
Based in part on "v4 GRE: Add TCP segmentation offload for GRE"
by Pravin B Shelar.

Cc: Jesse Gross <jesse@nicira.com>
Cc: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: Use 16bits for *_headers fields of struct skbuff
Simon Horman [Thu, 23 May 2013 21:02:51 +0000 (21:02 +0000)]
net: Use 16bits for *_headers fields of struct skbuff

In order to mitigate ongoing incresase in the size of struct skbuff
use 16 bit integer offsets rather than pointers for inner_*_headers.

This appears to reduce the size of struct skbuff from 0xd0 to 0xc0
bytes on x86_64 with the following all unset.

CONFIG_XFRM
CONFIG_NF_CONNTRACK
CONFIG_NF_CONNTRACK_MODULE
NET_SKBUFF_NF_DEFRAG_NEEDED
CONFIG_BRIDGE_NETFILTER
CONFIG_NET_SCHED
CONFIG_IPV6_NDISC_NODETYPE
CONFIG_NET_DMA
CONFIG_NETWORK_SECMARK

Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agophy: bcm63xx: report Broadcom BCM63xx PHYs as internal
Florian Fainelli [Thu, 23 May 2013 01:11:13 +0000 (01:11 +0000)]
phy: bcm63xx: report Broadcom BCM63xx PHYs as internal

The Broadcom BCM63xx PHY driver is for the SoC internal PHYs, flag these
as internal PHY devices.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agophy: allow drivers to flag a PHY device as internal
Florian Fainelli [Thu, 23 May 2013 01:11:12 +0000 (01:11 +0000)]
phy: allow drivers to flag a PHY device as internal

libphy currently always reports a PHY as an external transceiver from
the ethtool output. This is inaccurate, because some drivers should be
able to tell that a PHY device is an internal transceiver of an Ethernet
MAC. Add a new flag (PHY_IS_INTERNAL) which can be set by PHY drivers
just like other flags, and a corresponding helper: phy_is_internal()
which can be used by networking drivers to query if a given
PHY device is internal.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: ethtool: disambiguate XCVR_* meaning
Florian Fainelli [Thu, 23 May 2013 01:11:11 +0000 (01:11 +0000)]
net: ethtool: disambiguate XCVR_* meaning

Add a comment which explains the real meaning of XCVR_INTERNAL (PHY and
Ethernet MAC in the same package/die) and XCVR_EXTERNAL (PHY and
Ethernet MAC in a different package/die). Most if not all of the drivers
setting their transceiver type already do it the way the comment
describes it.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: wan: remove unnecessary platform_set_drvdata()
Jingoo Han [Mon, 27 May 2013 19:08:17 +0000 (19:08 +0000)]
net: wan: remove unnecessary platform_set_drvdata()

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d0631001288a5974afc0b2a5f568bcdecb4d
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: wireless: remove unnecessary platform_set_drvdata()
Jingoo Han [Mon, 27 May 2013 19:07:11 +0000 (19:07 +0000)]
net: wireless: remove unnecessary platform_set_drvdata()

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d0631001288a5974afc0b2a5f568bcdecb4d
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: irda: remove unnecessary platform_set_drvdata()
Jingoo Han [Mon, 27 May 2013 19:03:49 +0000 (19:03 +0000)]
net: irda: remove unnecessary platform_set_drvdata()

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d0631001288a5974afc0b2a5f568bcdecb4d
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: ethernet: remove unnecessary platform_set_drvdata()
Jingoo Han [Mon, 27 May 2013 19:01:12 +0000 (19:01 +0000)]
net: ethernet: remove unnecessary platform_set_drvdata()

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d0631001288a5974afc0b2a5f568bcdecb4d
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Rob Herring <rob.herring@calxeda.com>
Acked-by: Roland Stigge <stigge@antcom.de>
Acked-by: Mugunthan V N <mugunthanvnm@ti.com>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Tested-by: Roland Stigge <stigge@antcom.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'fec_regulator_suspend_resume'
David S. Miller [Tue, 28 May 2013 05:30:23 +0000 (22:30 -0700)]
Merge branch 'fec_regulator_suspend_resume'

Fabio Estevam says:

====================
This series provides support for controlling the PHY regulator in suspend/resume.

Fabio Estevam (5):
  fec: Place the phy regulator in the private structure
  fec: Invert the order of error path sequence
  fec: Disable the PHY regulator on error and removal
  fec: Remove irqs first
  fec: Handle the regulator in suspend/resume
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agofec: Handle the regulator in suspend/resume
Fabio Estevam [Mon, 27 May 2013 03:48:33 +0000 (03:48 +0000)]
fec: Handle the regulator in suspend/resume

In order to save power, let's disable the regulator in the suspend function and
enable it in resume.

Tested on a mx28evk board.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agofec: Remove irqs first
Fabio Estevam [Mon, 27 May 2013 03:48:32 +0000 (03:48 +0000)]
fec: Remove irqs first

During probe the clocks are enabled prior than the acquiring the interrupts.

In the remove function we need to do the opposite: first remove the interrupts
and then disable the clocks.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agofec: Disable the PHY regulator on error and removal
Fabio Estevam [Mon, 27 May 2013 03:48:31 +0000 (03:48 +0000)]
fec: Disable the PHY regulator on error and removal

In the case of error during probe, disable the PHY regulator.

Do the same in fec_drv_remove().

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agofec: Invert the order of error path sequence
Fabio Estevam [Mon, 27 May 2013 03:48:30 +0000 (03:48 +0000)]
fec: Invert the order of error path sequence

Currently when fec_enet_init fails it jumps to 'failed_init' error path, which
will attemp to free the interrupts.

This is wrong because at this point the interrupts have not even been acquired.

Swap failed_init/failed_irq to fix the error path.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agofec: Place the phy regulator in the private structure
Fabio Estevam [Mon, 27 May 2013 03:48:29 +0000 (03:48 +0000)]
fec: Place the phy regulator in the private structure

Instead of using a local reg_phy structure, let's put it inside the private
structure, so that we are able to have access to the regulator structure even
when we are outside fec_probe().

This is in preparation for controlling the FEC PHY regulator in the suspend and
resume functions.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobe2net: Trim padded packets for Lancer
Somnath Kotur [Sun, 26 May 2013 21:09:06 +0000 (21:09 +0000)]
be2net: Trim padded packets for Lancer

For padded packets, Lancer computes incorrect checksum. The workaround is
to trim such packets. This workaround is mainly for IPv4 packets.

Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Somnath Kotur <somnath.kotur@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobe2net: Pad skb to meet min Tx pkt size in lancer
Somnath Kotur [Sun, 26 May 2013 21:08:47 +0000 (21:08 +0000)]
be2net: Pad skb to meet min Tx pkt size in lancer

In Lancer, packets that are 32 bytes or less may cause a transmit stall.
The work-around is to pad such packets to a 36 byte length.

Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Somnath Kotur <somnath.kotur@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobe2net: cleanup be_get_drvinfo()
Somnath Kotur [Sun, 26 May 2013 21:08:36 +0000 (21:08 +0000)]
be2net: cleanup be_get_drvinfo()

Removing the be_cmd_get_fw_ver() query from be_get_drvinfo() and invoking
the same from be_setup() and/or post firmware download.

Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Somnath Kotur <somnath.kotur@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: Change to D3hot only on removal
Yuval Mintz [Mon, 27 May 2013 04:08:31 +0000 (04:08 +0000)]
bnx2x: Change to D3hot only on removal

This changes the PCI power management scheme of the bnx2x driver to be similar
to those of most network drivers - the driver will now changes the power state
into D3hot whenever the driver will be removed, instead of whenever an
interface is unloaded.

This change enables the driver to access its eeprom via ethtool callbacks
even when interfaces are unloaded (such access requires the function to be
in D0active).

Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: Implement PCI shutdown
Yuval Mintz [Mon, 27 May 2013 04:08:30 +0000 (04:08 +0000)]
bnx2x: Implement PCI shutdown

Implement the PCI shutdown callback to support un-orderly shutdown, i.e.,
shutdowns in which the remove callback will not be called.

Due to the lack of this functionality, when an un-orderly shutdown occurred,
it was possible for wake-on-lan to remain disabled.

This is now fixed as the callback introduces the correct place in which
`system_state' can be queried and wake-on-lan be configured accordingly.

Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: Count number of possible FCoE interfaces
Dmitry Kravkov [Mon, 27 May 2013 04:08:29 +0000 (04:08 +0000)]
bnx2x: Count number of possible FCoE interfaces

Commit 0eb43b4bb081a1a22574daab9c05286a600dd7fe
"bnx2x, bnx2fc: Use per port max exchange resources" has changed the number
of available FCoE exchanges, even in scenarios when some of the functions has
no FCoE support; This needlessly degraded the available resources.

Remedy this by calculating the maximal number of functions that may actually
utilize said connection.

Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
CC: Bhanu Prakash Gollapudi <bprakash@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: Ack unknown VF messages
Ariel Elior [Mon, 27 May 2013 04:08:28 +0000 (04:08 +0000)]
bnx2x: Ack unknown VF messages

A PF should ack the firmware when receiving unknown messages through the VF's
mailbox. This prevents the VF mailbox from being stuck in case of a VF sending
a message unknown to the PF (e.g. VF with more advanced version).

Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: Add and correct PCI link speed prints
Dmitry Kravkov [Mon, 27 May 2013 04:08:27 +0000 (04:08 +0000)]
bnx2x: Add and correct PCI link speed prints

This adds the print of the PCI gen3 link speed (8GHz), as well as correcting
the same print for 57712 boards (the print erroneously showed a 2.5GHz speed).

Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: Zero VFs starting MACs
Ariel Elior [Mon, 27 May 2013 04:08:26 +0000 (04:08 +0000)]
bnx2x: Zero VFs starting MACs

Hypervisor/Supervisor should set the VF's MAC prior to its load;
Using a randomly generated MAC as a default is a bad practice.

Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotcp: Remove 2 indentation levels in tcp_rcv_state_process
Joe Perches [Fri, 24 May 2013 18:06:58 +0000 (18:06 +0000)]
tcp: Remove 2 indentation levels in tcp_rcv_state_process

case TCP_FIN_WAIT1 can also be simplified by reversing tests
and adding breaks;

Add braces after case and move automatic definitions.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotcp: Remove another indentation level in tcp_rcv_state_process
Joe Perches [Fri, 24 May 2013 18:36:13 +0000 (18:36 +0000)]
tcp: Remove another indentation level in tcp_rcv_state_process

case TCP_SYN_RECV: can have another indentation level removed
by converting

if (acceptable) {
...;
} else {
return 1;
}

to
if (!acceptable)
return 1;
...;

Reflow code and comments to fit 80 columns.

Another pure cleanup patch.

Signed-off-by: Joe Perches <joe@perches.com>
Improved-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotcp: remove one indentation level in tcp_rcv_state_process()
Eric Dumazet [Fri, 24 May 2013 15:03:54 +0000 (15:03 +0000)]
tcp: remove one indentation level in tcp_rcv_state_process()

Remove one level of indentation 'introduced' in commit
c3ae62af8e75 (tcp: should drop incoming frames without ACK flag set)

if (true) {
        ...
}

@acceptable variable is a boolean.

This patch is a pure cleanup.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: add netnotifier event for upper device change
Jiri Pirko [Sat, 25 May 2013 04:12:10 +0000 (04:12 +0000)]
net: add netnotifier event for upper device change

Now when upper device is changed, event is not propagated via RT Netlink
to userspace. Userspace might never now about the change. Fix this by
adding upper-device-change notifier event.

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: ethernet: use platform_{get,set}_drvdata()
Jingoo Han [Thu, 23 May 2013 00:52:31 +0000 (00:52 +0000)]
net: ethernet: use platform_{get,set}_drvdata()

Use the wrapper functions for getting and setting the driver data using
platform_device instead of using dev_{get,set}_drvdata() with &pdev->dev,
so we can directly pass a struct platform_device.

Also, unnecessary dev_set_drvdata() is removed, because the driver core
clears the driver data to NULL after device_release or on probe failure.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: ipv6: Add IPv6 support to the ping socket.
Lorenzo Colitti [Wed, 22 May 2013 20:17:31 +0000 (20:17 +0000)]
net: ipv6: Add IPv6 support to the ping socket.

This adds the ability to send ICMPv6 echo requests without a
raw socket. The equivalent ability for ICMPv4 was added in
2011.

Instead of having separate code paths for IPv4 and IPv6, make
most of the code in net/ipv4/ping.c dual-stack and only add a
few IPv6-specific bits (like the protocol definition) to a new
net/ipv6/ping.c. Hopefully this will reduce divergence and/or
duplication of bugs in the future.

Caveats:

- Setting options via ancillary data (e.g., using IPV6_PKTINFO
  to specify the outgoing interface) is not yet supported.
- There are no separate security settings for IPv4 and IPv6;
  everything is controlled by /proc/net/ipv4/ping_group_range.
- The proc interface does not yet display IPv6 ping sockets
  properly.

Tested with a patched copy of ping6 and using raw socket calls.
Compiles and works with all of CONFIG_IPV6={n,m,y}.

Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'qlcnic'
David S. Miller [Sat, 25 May 2013 06:05:11 +0000 (23:05 -0700)]
Merge branch 'qlcnic'

Shahed Shaikh says:

====================
This patch series includes following changes:
* Support for spoof check configuration per VF using
  iproute2 tool.
* Set HW mask for 8300 adapter in INT-x mode to stop generating
  interrupts.
* Updated IRQ name for 8200 and 8300 Series adapter as per
  format used by other multiqueue drivers.
* Remove qlcnic_config_npars module parameter.
* Initialize trans_work and idc_aen_work at VF probe.
* Convert netsted if-else into switch-case statement.
* Change in diagnostics routine.
* Modify reset recovery path in diag mode by providing diagnostics
  routines enough time to unwind before proceeding with reset recovery.
* Implement GET_LED_STATUS command for 82xx adapter and refactor
  qlcnic_store_beacon() to split 8200 and 8300 specific calls.
* Some cleanup and log enhancements
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Update version to 5.2.43
Shahed Shaikh [Thu, 23 May 2013 21:04:36 +0000 (21:04 +0000)]
qlcnic: Update version to 5.2.43

Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Enhance virtual NIC logging
Manish Chopra [Thu, 23 May 2013 21:04:35 +0000 (21:04 +0000)]
qlcnic: Enhance virtual NIC logging

Signed-off-by: Manish Chopra <manish.chopra@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: qlcnic_get_board_name() function cleanup
Manish Chopra [Thu, 23 May 2013 21:04:34 +0000 (21:04 +0000)]
qlcnic: qlcnic_get_board_name() function cleanup

Signed-off-by: Manish Chopra <manish.chopra@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Implement GET_LED_STATUS command for 82xx adapter.
Himanshu Madhani [Thu, 23 May 2013 21:04:33 +0000 (21:04 +0000)]
qlcnic: Implement GET_LED_STATUS command for 82xx adapter.

o GET_LED_STATUS command will be used by driver to get
  current beacon state from 82xx adapter.
o Refactored qlcnic_store_beacon() to split 8200 and 8300
  specific calls.

Signed-off-by: Himanshu Madhani <himanshu.madhani@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: modify reset recovery path in diag mode
Sony Chacko [Thu, 23 May 2013 21:04:32 +0000 (21:04 +0000)]
qlcnic: modify reset recovery path in diag mode

Provide diagnostics routines enough time to unwind before
proceeding with reset recovery.

Signed-off-by: Sony Chacko <sony.chacko@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: diagnostics routine changes
Sony Chacko [Thu, 23 May 2013 21:04:31 +0000 (21:04 +0000)]
qlcnic: diagnostics routine changes

Test and set diagnostics mode bit before starting
diagnostics tests. Stop diagnostics tests if adapter is resetting.

Signed-off-by: Sony Chacko <sony.chacko@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Convert nested if-else to switch-case
Jitendra Kalsaria [Thu, 23 May 2013 21:04:30 +0000 (21:04 +0000)]
qlcnic: Convert nested if-else to switch-case

Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Initialize trans_work and idc_aen_work at VF probe.
Sucheta Chakraborty [Thu, 23 May 2013 21:04:29 +0000 (21:04 +0000)]
qlcnic: Initialize trans_work and idc_aen_work at VF probe.

o work_struct should be initialized before cancel_delayed_work
  call to destroy it.

Signed-off-by: Sucheta Chakraborty <sucheta.chakraborty@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Remove qlcnic_config_npars module parameter
Shahed Shaikh [Thu, 23 May 2013 21:04:28 +0000 (21:04 +0000)]
qlcnic: Remove qlcnic_config_npars module parameter

  qlcnic_config_npars module parameter is used to configure NPAR
  operational modes where NPAR function is used for PCI passthrough.
  Removing this paramter as PCI passthrough is not supported for
  NPAR functions.

Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Update IRQ name for 8200 and 8300 Series adapter.
Himanshu Madhani [Thu, 23 May 2013 21:04:27 +0000 (21:04 +0000)]
qlcnic: Update IRQ name for 8200 and 8300 Series adapter.

o Updated IRQ name for 8200 and 8300 Series adapter as per
  format used by other multiqueue drivers.

Signed-off-by: Himanshu Madhani <himanshu.madhani@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Disable INT-x interrupt for 83xx on driver unload
Himanshu Madhani [Thu, 23 May 2013 21:04:26 +0000 (21:04 +0000)]
qlcnic: Disable INT-x interrupt for 83xx on driver unload

o Set HW mask for 8300 Series adapter, in INT-x mode, to
  stop generating interrupt during driver unload.

Signed-off-by: Himanshu Madhani <himanshu.madhani@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Support spoof check config.
Rajesh Borundia [Thu, 23 May 2013 21:04:25 +0000 (21:04 +0000)]
qlcnic: Support spoof check config.

o Add support for spoof check configuration per VF using
  iproute2 tool.

Signed-off-by: Rajesh Borundia <rajesh.borundia@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Remove unnecessary lock around tg3_flag_set
Nithin Sujir [Thu, 23 May 2013 11:11:27 +0000 (11:11 +0000)]
tg3: Remove unnecessary lock around tg3_flag_set

The spinlock was needed when flags used to be a u32 and set/cleared
using bit operations. Now that we use the atomic set_bit, this lock
isn't needed.

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Fix misplaced empty line
Nithin Sujir [Thu, 23 May 2013 11:11:26 +0000 (11:11 +0000)]
tg3: Fix misplaced empty line

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Use descriptive label names in tg3_start
Nithin Sujir [Thu, 23 May 2013 11:11:25 +0000 (11:11 +0000)]
tg3: Use descriptive label names in tg3_start

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Make tg3_rings_reset() more concise
Nithin Sujir [Thu, 23 May 2013 11:11:24 +0000 (11:11 +0000)]
tg3: Make tg3_rings_reset() more concise

Simplify the rings reset function and increase readability by moving the
control block disable code into separate functions.

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Simplify ring control block setup
Nithin Sujir [Thu, 23 May 2013 11:11:23 +0000 (11:11 +0000)]
tg3: Simplify ring control block setup

The current code calls tg3_set_bdinfo() separately on napi0, followed by
a loop that does napi1+. Simplify it by setting bdinfo in the loop for
all napi contexts.

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Split APE driver state change out of boot reset signature update
Nithin Sujir [Thu, 23 May 2013 11:11:22 +0000 (11:11 +0000)]
tg3: Split APE driver state change out of boot reset signature update

Unlike the boot signature that needs to be set before every reset, the
ape state only needs to be updated to tell the firmware that the driver
is now taking/releasing control of the hardware. Move the calls to
tg3_ape_driver_state_change() to better, more appropriate places.

Also, the firmware does not distinguish between SUSPEND and START states
anymore. Remove the SUSPEND case in the switch.

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
David S. Miller [Fri, 24 May 2013 23:48:28 +0000 (16:48 -0700)]
Merge git://git./linux/kernel/git/davem/net

Merge net into net-next because some upcoming net-next changes
build on top of bug fixes that went into net.

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes
Linus Torvalds [Fri, 24 May 2013 15:58:53 +0000 (08:58 -0700)]
Merge git://git./linux/kernel/git/steve/gfs2-3.0-fixes

Pull gfs2 fixes from Steven Whitehouse:
 "This time there are just four fixes.  There are a couple of minor
  updates to the quota code, a fix for KConfig to ensure that only valid
  combinations including GFS2 can be built, and a fix for a typo
  affecting end i/o processing when writing the journal.

  Also, there is a temporary fix for a performance regression relating
  to block reservations and directories.  A longer fix will be applied
  in due course, but this deals with the most immediate problem for now"

* git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes:
  GFS2: Fix typo in gfs2_log_end_write loop
  GFS2: fix DLM depends to fix build errors
  GFS2: Use single-block reservations for directories
  GFS2: two minor quota fixups

11 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Fri, 24 May 2013 15:57:38 +0000 (08:57 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc

Pull powerpc fixes from Benjamin Herrenschmidt:
 "Some more P8 related bits, a bunch of fixes for our P7+/P8 HW crypto
  drivers, some added workarounds for those radeons that don't do proper
  64-bit MSIs and a couple of other trivialities by myself."

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/pseries: Make 32-bit MSI quirk work on systems lacking firmware support
  powerpc/powernv: Build a zImage.epapr
  powerpc: Make radeon 32-bit MSI quirk work on powernv
  powerpc: Context switch more PMU related SPRs
  powerpc/powernv: Fix condition for when to invalidate the TCE cache
  powerpc/pci: Fix bogus message at boot about empty memory resources
  powerpc: Fix TLB cleanup at boot on POWER8
  drivers/crypto/nx: Fixes for multiple races and issues

11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 24 May 2013 15:27:32 +0000 (08:27 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:
 "It's been a while since my last pull request so quite a few fixes have
  piled up."

Indeed.

 1) Fix nf_{log,queue} compilation with PROC_FS disabled, from Pablo
    Neira Ayuso.

 2) Fix data corruption on some tg3 chips with TSO enabled, from Michael
    Chan.

 3) Fix double insertion of VLAN tags in be2net driver, from Sarveshwar
    Bandi.

 4) Don't have TCP's MD5 support pass > PAGE_SIZE page offsets in
    scatter-gather entries into the crypto layer, the crypto layer can't
    handle that.  From Eric Dumazet.

 5) Fix lockdep splat in 802.1Q MRP code, also from Eric Dumazet.

 6) Fix OOPS in netfilter log module when called from conntrack, from
    Hans Schillstrom.

 7) FEC driver needs to use netif_tx_{lock,unlock}_bh() rather than the
    non-BH disabling variants.  From Fabio Estevam.

 8) TCP GSO can generate out-of-order packets, fix from Eric Dumazet.

 9) vxlan driver doesn't update 'used' field of fdb entries when it
    should, from Sridhar Samudrala.

10) ipv6 should use kzalloc() to allocate inet6 socket cork options,
    otherwise we can OOPS in ip6_cork_release().  From Eric Dumazet.

11) Fix races in bonding set mode, from Nikolay Aleksandrov.

12) Fix checksum generation regression added by "r8169: fix 8168evl
    frame padding.", from Francois Romieu.

13) ip_gre can look at stale SKB data pointer, fix from Eric Dumazet.

14) Fix checksum handling when GSO is enabled in bnx2x driver with
    certain chips, from Yuval Mintz.

15) Fix double free in batman-adv, from Martin Hundebøll.

16) Fix device startup synchronization with firmware in tg3 driver, from
    Nithin Sujit.

17) perf networking dropmonitor doesn't work at all due to mixed up
    trace parameter ordering, from Ben Hutchings.

18) Fix proportional rate reduction handling in tcp_ack(), from Nandita
    Dukkipati.

19) IPSEC layer doesn't return an error when a valid state is detected,
    causing an OOPS.  Fix from Timo Teräs.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (85 commits)
  be2net: bug fix on returning an invalid nic descriptor
  tcp: xps: fix reordering issues
  net: Revert unused variable changes.
  xfrm: properly handle invalid states as an error
  virtio_net: enable napi for all possible queues during open
  tcp: bug fix in proportional rate reduction.
  net: ethernet: sun: drop unused variable
  net: ethernet: korina: drop unused variable
  net: ethernet: apple: drop unused variable
  qmi_wwan: Added support for Cinterion's PLxx WWAN Interface
  perf: net_dropmonitor: Remove progress indicator
  perf: net_dropmonitor: Use bisection in symbol lookup
  perf: net_dropmonitor: Do not assume ordering of dictionaries
  perf: net_dropmonitor: Fix symbol-relative addresses
  perf: net_dropmonitor: Fix trace parameter order
  net: fec: use a more proper compatible string for MVF type device
  qlcnic: Fix updating netdev->features
  qlcnic: remove netdev->trans_start updates within the driver
  qlcnic: Return proper error codes from probe failure paths
  tg3: Update version to 3.132
  ...