Shradha Shah [Thu, 23 Jan 2014 13:19:55 +0000 (13:19 +0000)]
Add Shradha Shah as the sfc driver maintainer.
I will be taking over the work from Ben Hutchings.
Signed-off-by: Shradha Shah <sshah@solarflare.com> Acked-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Or Gerlitz [Thu, 23 Jan 2014 09:28:13 +0000 (11:28 +0200)]
net/vxlan: Share RX skb de-marking and checksum checks with ovs
Make sure the practice set by commit 0afb166 "vxlan: Add capability
of Rx checksum offload for inner packet" is applied when the skb
goes through the portion of the RX code which is shared between
vxlan netdevices and ovs vxlan port instances.
Cc: Joseph Gasparakis <joseph.gasparakis@intel.com> Cc: Pravin B Shelar <pshelar@nicira.com> Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Dan Carpenter [Thu, 23 Jan 2014 08:26:37 +0000 (11:26 +0300)]
tulip: cleanup by using ARRAY_SIZE()
In this situation then ARRAY_SIZE() and sizeof() are the same, but we're
really dealing with array indexes and not byte offsets so ARRAY_SIZE()
is cleaner.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Acked-by: Grant Grundler <grundler@parisc-linux.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Duan Jiong [Thu, 23 Jan 2014 06:00:25 +0000 (14:00 +0800)]
ip_tunnel: clear IPCB in ip_tunnel_xmit() in case dst_link_failure() is called
commit a622260254ee48("ip_tunnel: fix kernel panic with icmp_dest_unreach")
clear IPCB in ip_tunnel_xmit() , or else skb->cb[] may contain garbage from
GSO segmentation layer.
But commit 0e6fbc5b6c621("ip_tunnels: extend iptunnel_xmit()") refactor codes,
and it clear IPCB behind the dst_link_failure().
So clear IPCB in ip_tunnel_xmit() just like commti a622260254ee48("ip_tunnel:
fix kernel panic with icmp_dest_unreach").
Signed-off-by: Duan Jiong <duanj.fnst@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Gavin Shan [Thu, 23 Jan 2014 04:27:35 +0000 (12:27 +0800)]
net/cxgb4: Don't retrieve stats during recovery
We possibly retrieve the adapter's statistics during EEH recovery
and that should be disallowed. Otherwise, it would possibly incur
replicate EEH error and EEH recovery is going to fail eventually.
The patch reuses statistics lock and checks net_device is attached
before going to retrieve statistics, so that the problem can be
avoided.
Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Gavin Shan [Thu, 23 Jan 2014 04:27:34 +0000 (12:27 +0800)]
net/cxgb4: Avoid disabling PCI device for towice
If we have EEH error happens to the adapter and we have to remove
it from the system for some reasons (e.g. more than 5 EEH errors
detected from the device in last hour), the adapter will be disabled
for towice separately by eeh_err_detected() and remove_one(), which
will incur following unexpected backtrace. The patch tries to avoid
it.
Vlad Yasevich [Wed, 22 Jan 2014 17:54:15 +0000 (12:54 -0500)]
net: Correctly sync addresses from multiple sources to single device
When we have multiple devices attempting to sync the same address
to a single destination, each device should be permitted to sync
it once. To accomplish this, pass the 'sync_cnt' of the source
address when adding the addresss to the lower device. 'sync_cnt'
tracks how many time a given address has been succefully synced.
This way, we know that if the 'sync_cnt' passed in is 0, we should
sync this address.
Also, turn 'synced' member back into the counter as was originally
done in
commit 4543fbefe6e06a9e40d9f2b28d688393a299f079.
net: count hw_addr syncs so that unsync works properly.
It tracks how many time a given address has been added via a
'sync' operation. For every successfull 'sync' the counter is
incremented, and for ever 'unsync', the counter is decremented.
This makes sure that the address will be properly removed from
the the lower device when all the upper devices have removed it.
Reported-by: Andrey Dmitrov <andrey.dmitrov@oktetlabs.ru> CC: Andrey Dmitrov <andrey.dmitrov@oktetlabs.ru> CC: Alexandra N. Kossovsky <Alexandra.Kossovsky@oktetlabs.ru> CC: Konstantin Ushakov <Konstantin.Ushakov@oktetlabs.ru> Signed-off-by: Vlad Yasevich <vyasevic@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Shlomo Pongratz [Wed, 22 Jan 2014 13:23:29 +0000 (15:23 +0200)]
net/udp_offload: Handle static checker complaints
Fixed few issues around using __rcu prefix and rcu_assign_pointer, also
fixed a warning print to use ntohs(port) and not htons(port).
net/ipv4/udp_offload.c:112:9: error: incompatible types in comparison expression (different address spaces)
net/ipv4/udp_offload.c:113:9: error: incompatible types in comparison expression (different address spaces)
net/ipv4/udp_offload.c:176:19: error: incompatible types in comparison expression (different address spaces)
Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com> Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Christoph Paasch [Wed, 22 Jan 2014 12:58:44 +0000 (13:58 +0100)]
tcp: metrics: Handle v6/v4-mapped sockets in tcp-metrics
A socket may be v6/v4-mapped. In that case sk->sk_family is AF_INET6,
but the IP being used is actually an IPv4-address.
Current's tcp-metrics will thus represent it as an IPv6-address:
root@server:~# ip tcp_metrics
::ffff:10.1.1.2 age 22.920sec rtt 18750us rttvar 15000us cwnd 10
10.1.1.2 age 47.970sec rtt 16250us rttvar 10000us cwnd 10
This patch modifies the tcp-metrics so that they are able to handle the
v6/v4-mapped sockets correctly.
Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Yann Droneaud [Wed, 22 Jan 2014 19:25:24 +0000 (20:25 +0100)]
6lowpan: add a license to 6lowpan_iphc module
Since commit 8df8c56a5abc, 6lowpan_iphc is a module of its own.
Unfortunately, it lacks some infrastructure to behave like a
good kernel citizen:
kernel: 6lowpan_iphc: module license 'unspecified' taints kernel.
kernel: Disabling lock debugging due to kernel taint
This patch adds the basic MODULE_LICENSE(); with GPL license:
the code was copied from net/ieee802154/6lowpan.c which is GPL
and the module exports symbol with EXPORT_SYMBOL_GPL();.
Cc: Jukka Rissanen <jukka.rissanen@linux.intel.com> Cc: Alexander Aring <alex.aring@gmail.com> Cc: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Yann Droneaud <ydroneaud@opteya.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 22 Jan 2014 08:05:55 +0000 (09:05 +0100)]
rtnetlink: provide api for getting and setting slave info
Recent patch
bonding: add netlink attributes to slave link dev (1d3ee88ae0d6)
Introduced yet another device specific way to access slave information
over rtnetlink. There is one already there for bridge.
This patch introduces generic way to do this, for getting and setting
info as well by extending link_ops. Later on, this new interface will
be used for bridge ports as well.
Signed-off-by: Jiri Pirko <jiri@resnulli.us> Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 22 Jan 2014 08:05:54 +0000 (09:05 +0100)]
rtnetlink: put "BOND" into nl attribute names which are related to bonding
Signed-off-by: Jiri Pirko <jiri@resnulli.us> Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Wed, 22 Jan 2014 08:05:53 +0000 (09:05 +0100)]
bonding: change name of sysfs dir for bonding slaves
Allow user to identify easily what the attributes are related to. Change
the name of the group to "bonding_slave" to be similar to master which
is named "bonding".
Signed-off-by: Jiri Pirko <jiri@resnulli.us> Signed-off-by: David S. Miller <davem@davemloft.net>
viresh kumar [Wed, 22 Jan 2014 06:53:33 +0000 (12:23 +0530)]
net/neighbour: queue work on power efficient wq
Workqueue used in neighbour layer have no real dependency of scheduling these on
the cpu which scheduled them.
On a idle system, it is observed that an idle cpu wakes up many times just to
service this work. It would be better if we can schedule it on a cpu which the
scheduler believes to be the most appropriate one.
This patch replaces normal workqueues with power efficient versions. This
doesn't change existing behavior of code unless CONFIG_WQ_POWER_EFFICIENT is
enabled.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
viresh kumar [Wed, 22 Jan 2014 06:53:32 +0000 (12:23 +0530)]
net/ipv4: queue work on power efficient wq
Workqueue used in ipv4 layer have no real dependency of scheduling these on the
cpu which scheduled them.
On a idle system, it is observed that an idle cpu wakes up many times just to
service this work. It would be better if we can schedule it on a cpu which the
scheduler believes to be the most appropriate one.
This patch replaces normal workqueues with power efficient versions. This
doesn't change existing behavior of code unless CONFIG_WQ_POWER_EFFICIENT is
enabled.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net>
FX Le Bail [Wed, 22 Jan 2014 06:42:37 +0000 (07:42 +0100)]
ipv6: enable anycast addresses as source addresses for datagrams
This change allows to consider an anycast address valid as source address
when given via an IPV6_PKTINFO or IPV6_2292PKTINFO ancillary data item.
So, when sending a datagram with ancillary data, the unicast and anycast
addresses are handled in the same way.
- Adds ipv6_chk_acast_addr_src() to check if an anycast address is link-local
on given interface or is global.
- Uses it in ip6_datagram_send_ctl().
Signed-off-by: Francois-Xavier Le Bail <fx.lebail@yahoo.com> Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Dominic Curran [Wed, 22 Jan 2014 03:03:23 +0000 (03:03 +0000)]
tuntap: Fix for a race in accessing numqueues
A patch for fixing a race between queue selection and changing queues
was introduced in commit 92bb73ea2("tuntap: fix a possible race between
queue selection and changing queues").
The fix was to prevent the driver from re-reading the tun->numqueues
more than once within tun_select_queue() using ACCESS_ONCE().
We have been experiancing 'Divide-by-zero' errors in tun_net_xmit()
since we moved from 3.6 to 3.10, and believe that they come from a
simular source where the value of tun->numqueues changes to zero
between the first and a subsequent read of tun->numqueues.
The fix is a simular use of ACCESS_ONCE(), as well as a multiply
instead of a divide in the if statement.
Signed-off-by: Dominic Curran <dominic.curran@citrix.com> Cc: Jason Wang <jasowang@redhat.com> Cc: Maxim Krasnyansky <maxk@qti.qualcomm.com> Acked-by: Jason Wang <jasowang@redhat.com> Acked-by: Max Krasnyansky <maxk@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Toshiaki Makita [Wed, 22 Jan 2014 00:29:37 +0000 (09:29 +0900)]
bridge: Remove unnecessary vlan_put_tag in br_handle_vlan
br_handle_vlan() pushes HW accelerated vlan tag into skbuff when outgoing
port is the bridge device.
This is unnecessary because __netif_receive_skb_core() can handle skbs
with HW accelerated vlan tag. In current implementation,
__netif_receive_skb_core() needs to extract the vlan tag embedded in skb
data. This could cause low network performance especially when receiving
frames at a high frame rate on the bridge device.
Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp> Acked-by: Vlad Yasevich <vyasevic@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Christoph Paasch [Tue, 21 Jan 2014 12:30:26 +0000 (13:30 +0100)]
tcp: metrics: Fix rcu-race when deleting multiple entries
In bbf852b96ebdc6d1 I introduced the tmlist, which allows to delete
multiple entries from the cache that match a specified destination if no
source-IP is specified.
However, as the cache is an RCU-list, we should not create this tmlist, as
it will change the tcpm_next pointer of the element that will be deleted
and so a thread iterating over the cache's entries while holding the
RCU-lock might get "redirected" to this tmlist.
This patch fixes this, by reverting back to the old behavior prior to bbf852b96ebdc6d1, which means that we simply change the tcpm_next
pointer of the previous element (pp) to jump over the one we are
deleting.
The difference is that we call kfree_rcu() directly on the cache entry,
which allows us to delete multiple entries from the list.
Fixes: bbf852b96ebdc6d1 (tcp: metrics: Delete all entries matching a certain destination) Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 23 Jan 2014 01:45:39 +0000 (17:45 -0800)]
Merge branch 'bonding'
Ding Tianhong says:
====================
bonding: fix primary problem for bonding
If the slave's name changed, and the bond params primary is exist,
the bond should deal with the situation in two ways:
1) If the slave was the primary slave yet, clean the primary slave
and reselect active slave.
2) If the slave's new name is as same as bond primary, set the slave
as primary slave and reselect active slave.
If the new primary is not matching any slave in the bond, the bond should
record it to params, clean the primary slave and select a new active slave.
Update bonding.txt for primary description.
v2.1->v1: Because there are too many indentions and useless verification, so rewrite
the logic for updating the primary slave.
Modify some comments for to clean the typos.
v3->v2.1: Veaceslav disagree the first patch and modify the logic for it
(bonding: update the primary slave when changing slave's name)
and resend it himself (bonding: handle slave's name change with primary_slave logic),
so remove the first patch and send the last two patches.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Wed, 22 Jan 2014 20:07:53 +0000 (21:07 +0100)]
net: vxlan: convert to act as a pernet subsystem
As per suggestion from Eric W. Biederman, vxlan should be using
{un,}register_pernet_subsys() instead of {un,}register_pernet_device()
to ensure the vxlan_net structure is initialized before and cleaned
up after all network devices in a given network namespace i.e. when
dealing with network notifiers. This is similarly handeled already in
commit 91e2ff3528ac ("net: Teach vlans to cleanup as a pernet subsystem")
and, thus, improves upon fd27e0d44a89 ("net: vxlan: do not use vxlan_net
before checking event type"). Just as in 91e2ff3528ac, we do not need
to explicitly handle deletion of vxlan devices as network namespace
exit calls dellink on all remaining virtual devices, and
rtnl_link_unregister() calls dellink on all outstanding devices in that
network namespace, so we can entirely drop the pernet exit operation
as well. Moreover, on vxlan module exit, rcu_barrier() is called by
netns since commit 3a765edadb28 ("netns: Add an explicit rcu_barrier
to unregister_pernet_{device|subsys}"), so this may be omitted. Tested
with various scenarios and works well on my side.
Suggested-by: Eric W. Biederman <ebiederm@xmission.com> Cc: Jesse Brandeburg <jesse.brandeburg@intel.com> Cc: Eric W. Biederman <ebiederm@xmission.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Harry Mason [Fri, 17 Jan 2014 13:22:32 +0000 (13:22 +0000)]
sch_htb: let skb->priority refer to non-leaf class
If the class in skb->priority is not a leaf, apply filters from the
selected class, not the qdisc. This lets netfilter or user space
partially classify the packet.
Signed-off-by: Harry Mason <harry.mason@smoothwall.net> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Neil Horman [Wed, 22 Jan 2014 21:01:44 +0000 (16:01 -0500)]
af_packet: Add Queue mapping mode to af_packet fanout operation
This patch adds a queue mapping mode to the fanout operation of af_packet
sockets. This allows user space af_packet users to better filter on flows
ingressing and egressing via a specific hardware queue, and avoids the potential
packet reordering that can occur when FANOUT_CPU is being used and irq affinity
varies.
Tested successfully by myself. applies to net-next
Signed-off-by: Neil Horman <nhorman@tuxdriver.com> CC: "David S. Miller" <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 22 Jan 2014 23:39:18 +0000 (15:39 -0800)]
Merge branch 'bonding_option_api'
Nikolay Aleksandrov says:
====================
bonding: introduce new option API
This patchset's goal is to introduce a new option API which should be used
to properly describe the bonding options with their mode dependcies and
requirements. With this patchset applied we get centralized option
manipulation, automatic RTNL acquire per option setting, automatic option
range checking, mode dependcy checking and other various flags which are
described in detail in patch 01's commit message and comments.
Also the parameter passing is changed to use a specialized structure which
is initialized to a value depending on the needs.
The main exported functions are:
__bond_opt_set() - set an option (RTNL should be acquired prior)
bond_opt_init(val|str) - init a bond_opt_value struct for value or string
parameter passing
bond_opt_tryset_rtnl() - function which tries to acquire rtnl, mainly used
for sysfs
bond_opt_parse - used to parse or check for valid values
bond_opt_get - retrieve a pointer to bond_option struct for some option
bond_opt_get_val - retrieve a pointer to a bond_opt_value struct for
some value
The same functions are used to set an option via sysfs and netlink, just
the parameter that's passed is usually initialized in a different way.
The converted options have multiple style fixes, there're some longer
lines but they looked either ugly or were strings/pr_warnings, if you
think some line would be better broken just let me know :-) there're
also a few sscanf false-positive warnings.
I decided to keep the "unsuppmodes" way of mode dep checking since it's
straight forward, if we make a more general way for checking dependencies
it'll be easy to change it.
Future plans for this work include:
- Automatic sysfs generation from the bond_opts[].
- Use of the API in bond_check_params() and thus cleaning it up (this has
actually started, I'll take care of the rest in a separate patch)
- Clean up all option-unrelated files of option definitions and functions
I've tried to leave as much documentation as possible, if there's anything
unclear please let me know. One more thing, I haven't moved all
option-related functions from bonding.h to the new bond_options.h, this
will be done in a separate patch, it's in my todo list.
This patchset has been tested by setting each converted option via sysfs
and netlink to a couple of wrong values, a couple of correct values and
some random values, also for the opts that have flags they have been
tested as well.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds the necessary changes so slaves would use
the new bonding option API. Also move the option to its own set function
in bond_options.c and fix some style errors.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: convert queue_id to use the new option API
This patch adds the necessary changes so queue_id would use
the new bonding option API. Also move it to its own set function in
bond_options.c and fix some style errors.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: convert num_peer_notif to use the new option API
This patch adds the necessary changes so num_peer_notif would use
the new bonding option API.
When the auto-sysfs generation is done an alias should be added for
this option as there're currently 2 entries in sysfs for it.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: convert arp_ip_target to use the new option API
This patch adds the necessary changes so arp_ip_target would use
the new bonding option API. This option is an exception because of
the way it's currently implemented that's why its netlink code is
a bit different from the other options to keep the functionality as
before and at the same time to have a single set function.
This patch also fixes a few stylistic errors.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: convert fail_over_mac to use the new option API
This patch adds the necessary changes so fail_over_mac would use
the new bonding option API. Also fixes a trivial copy/paste error in
bond_check_params where the wrong variable was used for the error msg.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: convert mode setting to use the new option API
This patch makes the bond's mode setting use the new option API and
adds support for dependency printing which relies on having an entry for
the mode option in the bond_opts[] array.
Also add the ability to print the mode name when mode dependency fails
and fix some trivial/style errors.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds the necessary basic infrastructure to support
centralized and unified option manipulation API for the bonding. The new
structure bond_option will be used to describe each option with its
dependencies on modes which will be checked automatically thus removing a
lot of duplicated code. Also automatic range checking is added for
some options. Currently the option setting function requires RTNL to
be acquired prior to calling it, since many options already rely on RTNL
it seemed like the best choice to protect all against common race
conditions.
In order to add an option the following steps need to be done:
1. Add an entry BOND_OPT_<option> to bond_options.h so it gets a unique id
and a bit corresponding to the id
2. Add a bond_option entry to the bond_opts[] array in bond_options.c which
describes the option, its dependencies and its manipulation function
3. Add code to export the option through sysfs and/or as a module parameter
(the sysfs export will be made automatically in the future)
The options can have different flags set, currently the following are
supported:
BOND_OPTFLAG_NOSLAVES - require that the bond device has no slaves prior
to setting the option
BOND_OPTFLAG_IFDOWN - require that the bond device is down prior to
setting the option
BOND_OPTFLAG_RAWVAL - don't parse the value but return it raw for the
option to parse
There's a new value structure to describe different types of values
which can have the following flags:
BOND_VALFLAG_DEFAULT - marks the default option (permanent string alias
to this option is "default")
BOND_VALFLAG_MIN - the minimum value that this option can have
BOND_VALFLAG_MAX - the maximum value that this option can have
An example would be nice here, so if we have an option which can have
the values "off"(2), "special"(4, default) and supports a range, say
16 - 32, it should be defined as follows:
"off", 2,
"special", 4, BOND_VALFLAG_DEFAULT,
"rangemin", 16, BOND_VALFLAG_MIN,
"rangemax", 32, BOND_VALFLAG_MAX
So we have the valid intervals: [2, 2], [4, 4], [16, 32]
Also the valid strings: "off" = 2, "special" and "default" = 4
"rangemin" = 16, "rangemax" = 32
BOND_VALFLAG_(MIN|MAX) can be used to specify a valid range for an
option, if MIN is omitted then 0 is considered as a minimum. If an
exact match is found in the values[] table it will be returned,
otherwise the range is tried (if available).
The option parameter passing is done by using a special structure called
bond_opt_value which can take either a string or a value to parse. One
of the bond_opt_init(val|str) macros should be used depending on which
one does the user want to parse (string or value). Then a call to
__bond_opt_set should be done under RTNL.
Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 22 Jan 2014 07:19:26 +0000 (23:19 -0800)]
Merge branch 'reciprocal'
Hannes Frederic Sowa says:
====================
reciprocal_divide update
This patch is on top of aee636c4809fa5 ("bpf: do not use reciprocal
divide") from Eric that sits in net tree. It will not create a merge
conflict, but it depends on this one, so we suggest, if possible, to
merge net into net-next.
We are proposing this change with only small modifications from the
v2 version, namely updating the name of trim to reciprocal_scale
(as commented on by Ben Hutchings and Eric Dumazet, thanks!).
We thought about introducing the reciprocal_divide algorithm in
parallel to the one already used by the kernel but faced organizational
issues, leading us to the conclusion that it is best to just replace
the old one: We could not come up with names for the different
implementations and also with a way to describe the differences to
guide developers which one to choose in which situation. This is
because we cannot specify the correct semantics for the version
which is currently used by the kernel. Altough it seems to not be
causing problems in the kernel, we cannot surely say so in the
case of flex_array for the future. Current usage seems ok, but
future users could run into problems.
Changelog:
v1->v2:
- changed name to prandom_u32_max in p1
- changed name to trim in p2
- reworked code in p3
v2->v3:
- p1 and p3 stays unchanged, only small update in commit
message in p3
- changed name to reciprocal_scale in p2
- fixed kernel doc format
v3->v4:
- pseduo -> pseudo (thanks to Tilman Schmidt)
v4->v5:
- fix pseduo -> pseudo for real now, sorry for the noise
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
reciprocal_divide: update/correction of the algorithm
Jakub Zawadzki noticed that some divisions by reciprocal_divide()
were not correct [1][2], which he could also show with BPF code
after divisions are transformed into reciprocal_value() for runtime
invariance which can be passed to reciprocal_divide() later on;
reverse in BPF dump ended up with a different, off-by-one K in
some situations.
This has been fixed by Eric Dumazet in commit aee636c4809fa5
("bpf: do not use reciprocal divide"). This follow-up patch
improves reciprocal_value() and reciprocal_divide() to work in
all cases by using Granlund and Montgomery method, so that also
future use is safe and without any non-obvious side-effects.
Known problems with the old implementation were that division by 1
always returned 0 and some off-by-ones when the dividend and divisor
where very large. This seemed to not be problematic with its
current users, as far as we can tell. Eric Dumazet checked for
the slab usage, we cannot surely say so in the case of flex_array.
Still, in order to fix that, we propose an extension from the
original implementation from commit 6a2d7a955d8d resp. [3][4],
by using the algorithm proposed in "Division by Invariant Integers
Using Multiplication" [5], Torbjörn Granlund and Peter L.
Montgomery, that is, pseudocode for q = n/d where q, n, d is in
u32 universe:
1) Initialization:
int l = ceil(log_2 d)
uword m' = floor((1<<32)*((1<<l)-d)/d)+1
int sh_1 = min(l,1)
int sh_2 = max(l-1,0)
2) For q = n/d, all uword:
uword t = (n*m')>>32
q = (t+((n-t)>>sh_1))>>sh_2
The assembler implementation from Agner Fog [6] also helped a lot
while implementing. We have tested the implementation on x86_64,
ppc64, i686, s390x; on x86_64/haswell we're still half the latency
compared to normal divide.
Reported-by: Jakub Zawadzki <darkjames-ws@darkjames.pl> Cc: Eric Dumazet <eric.dumazet@gmail.com> Cc: Austin S Hemmelgarn <ahferroin7@gmail.com> Cc: linux-kernel@vger.kernel.org Cc: Jesse Gross <jesse@nicira.com> Cc: Jamal Hadi Salim <jhs@mojatatu.com> Cc: Stephen Hemminger <stephen@networkplumber.org> Cc: Matt Mackall <mpm@selenic.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Christoph Lameter <cl@linux-foundation.org> Cc: Andy Gospodarek <andy@greyhouse.net> Cc: Veaceslav Falico <vfalico@redhat.com> Cc: Jay Vosburgh <fubar@us.ibm.com> Cc: Jakub Zawadzki <darkjames-ws@darkjames.pl> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Wed, 22 Jan 2014 01:29:40 +0000 (02:29 +0100)]
net: introduce reciprocal_scale helper and convert users
As David Laight suggests, we shouldn't necessarily call this
reciprocal_divide() when users didn't requested a reciprocal_value();
lets keep the basic idea and call it reciprocal_scale(). More
background information on this topic can be found in [1].
Suggested-by: David Laight <david.laight@aculab.com> Cc: Jakub Zawadzki <darkjames-ws@darkjames.pl> Cc: Eric Dumazet <eric.dumazet@gmail.com> Cc: linux-kernel@vger.kernel.org Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Wed, 22 Jan 2014 01:29:39 +0000 (02:29 +0100)]
random32: add prandom_u32_max and convert open coded users
Many functions have open coded a function that returns a random
number in range [0,N-1]. Under the assumption that we have a PRNG
such as taus113 with being well distributed in [0, ~0U] space,
we can implement such a function as uword t = (n*m')>>32, where
m' is a random number obtained from PRNG, n the right open interval
border and t our resulting random number, with n,m',t in u32 universe.
Lets go with Joe and simply call it prandom_u32_max(), although
technically we have an right open interval endpoint, but that we
have documented. Other users can further be migrated to the new
prandom_u32_max() function later on; for now, we need to make sure
to migrate reciprocal_divide() users for the reciprocal_divide()
follow-up fixup since their function signatures are going to change.
Joint work with Hannes Frederic Sowa.
Cc: Jakub Zawadzki <darkjames-ws@darkjames.pl> Cc: Eric Dumazet <eric.dumazet@gmail.com> Cc: linux-kernel@vger.kernel.org Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Moni Shoua [Tue, 21 Jan 2014 08:19:37 +0000 (10:19 +0200)]
net/mlx4_core: Remove unnecessary validation for port number
This is a fix to a regression introduced by commit:
"982290a net/mlx4_core: Check port number for validity
before accessing data"
IPoIB could not attach to multicast group and we get this in dmesg:
[144214.145008] ib0: failed to attach to multicast group, ret = -22
[144214.145016] ib0: couldn't attach QP to multicast group ff12:401b:ffff:0000:0000:0000:ffff:ffff
[144214.145019] ib0: multicast join failed for ff12:401b:ffff:0000:0000:0000:ffff:ffff, status -22
The cause to the problem is because port is extracted from gid[5].
Which is only valid for Ethernet.
Removed this validation in mlx4_qp_attach_common(), which is accessed
from both Ethernet and IB flows.
Error flow for bad port value in Ethernet is already exists in that
function.
Signed-off-by: Moni Shoua <monis@mellanox.co.il> Signed-off-by: Matan Barak <matanb@mellanox.com> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Somnath Kotur [Tue, 21 Jan 2014 10:20:55 +0000 (15:50 +0530)]
be2net: Fix be_vlan_add/rem_vid() routines
The current logic to put interface into VLAN Promiscous mode is not correct.
We should increment "adapter->vlans_added" before calling be_vid_config().
Also removed some unwanted log messages.
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>
Ariel Elior [Tue, 21 Jan 2014 08:31:20 +0000 (10:31 +0200)]
bnx2x: Fix VF flr flow
When a VF originating from a given PF is flr-ed, that PF gets an interrupt
from the chip management and takes a part in the flr process.
This patch fixes several corner cases in which the driver performs its part
of the flr flow out-of-order, causing the FW to assert due to badly timed
messages received from the driver.
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com> Signed-off-by: Ariel Elior <ariele@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Mon, 20 Jan 2014 23:19:37 +0000 (00:19 +0100)]
net: filter: let bpf_tell_extensions return SKF_AD_MAX
Michal Sekletar added in commit ea02f9411d9f ("net: introduce
SO_BPF_EXTENSIONS") a facility where user space can enquire
the BPF ancillary instruction set, which is imho a step into
the right direction for letting user space high-level to BPF
optimizers make an informed decision for possibly using these
extensions.
The original rationale was to return through a getsockopt(2)
a bitfield of which instructions are supported and which
are not, as of right now, we just return 0 to indicate a
base support for SKF_AD_PROTOCOL up to SKF_AD_PAY_OFFSET.
Limitations of this approach are that this API which we need
to maintain for a long time can only support a maximum of 32
extensions, and needs to be additionally maintained/updated
when each new extension that comes in.
I thought about this a bit more and what we can do here to
overcome this is to just return SKF_AD_MAX. Since we never
remove any extension since we cannot break user space and
always linearly increase SKF_AD_MAX on each newly added
extension, user space can make a decision on what extensions
are supported in the whole set of extensions and which aren't,
by just checking which of them from the whole set have an
offset < SKF_AD_MAX of the underlying kernel.
Since SKF_AD_MAX must be updated each time we add new ones,
we don't need to introduce an additional enum and got
maintenance for free. At some point in time when
SO_BPF_EXTENSIONS becomes ubiquitous for most kernels, then
an application can simply make use of this and easily be run
on newer or older underlying kernels without needing to be
recompiled, of course. Since that is for 3.14, it's not too
late to do this change.
Cc: Michal Sekletar <msekleta@redhat.com> Cc: Eric Dumazet <edumazet@google.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Acked-by: Michal Sekletar <msekleta@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 22 Jan 2014 02:55:22 +0000 (18:55 -0800)]
net: Fix some fallout from the etner_addr_copy() changes.
net/appletalk/aarp.c: In function ‘__aarp_send_query’:
net/appletalk/aarp.c:137:2: error: implicit declaration of function ‘ether_addr_copy’ [-Werror=implicit-function-declaration]
...
net/atm/lec.c: In function ‘send_to_lecd’:
net/atm/lec.c:524:3: warning: passing argument 1 of ‘ether_addr_copy’ from incompatible pointer type [enabled by default]
In file included from net/atm/lec.c:17:0:
include/linux/etherdevice.h:227:20: note: expected ‘u8 *’ but argument is of type ‘unsigned char (*)[6]’
...
net/caif/caif_usb.c: In function ‘cfusbl_create’:
net/caif/caif_usb.c:108:2: error: implicit declaration of function ‘ether_addr_copy’ [-Werror=implicit-function-declaration]
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 22 Jan 2014 02:41:46 +0000 (18:41 -0800)]
Merge branch 'sctp'
Wang Weidong says:
====================
sctp: remove some macro locking wrappers
In sctp.h we can find some macro locking wrappers. As Neil point out that:
"Its because in the origional implementation of the sctp protocol, there was a
user space test harness which built the kernel module for userspace execution to
cary our some unit testing on the code. It did so by redefining some of those
locking macros to user space friendly code. IIRC we haven't use those unit
tests in years, and so should be removing them, not adding them to other
locations."
So I remove them.
====================
Acked-by: Neil Horman <nhorman@tuxdriver.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
wangweidong [Tue, 21 Jan 2014 07:44:10 +0000 (15:44 +0800)]
sctp: remove macros sctp_read_[un]lock
Redefined read_[un]lock to sctp_read_[un]lock for user space
friendly code which we haven't use in years, and the macros
we never used, so removing them.
Signed-off-by: Wang Weidong <wangweidong1@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 22 Jan 2014 02:05:12 +0000 (18:05 -0800)]
Merge branch 'gro_udp_encap'
Or Gerlitz says:
====================
net: Add GRO support for UDP encapsulating protocols
This series adds GRO handlers for protocols that do UDP encapsulation, with the
intent of being able to coalesce packets which encapsulate packets belonging to
the same TCP session.
For GRO purposes, the destination UDP port takes the role of the ether type
field in the ethernet header or the next protocol in the IP header.
The UDP GRO handler will only attempt to coalesce packets whose destination
port is registered to have gro handler.
The patches done against net-next 75e4364f67 "net: stmmac: fix NULL pointer
dereference in stmmac_get_tx_hwtstamp"
Or.
v4 --> v5 changes:
- followed Eric's directives to avoid using atomic get/put ops on the
udp gro receive and complete callbacks and instead keep the rcu_read_lock
when calling the next handler on the chain.
v3 --> v4 changes:
- applied feedback from Tom on some micro-optimizations that save
branches and goto directives in the udp gro logic
- applied feedback from Eric on correct RCU programming for the
add/remove flow of the upper protocols udp gro handlers
v2 --> v3 changes:
- moved to use linked list to store the udp gro handlers, this solves the
problem of consuming 512KB of memory for the handlers.
- use a mark on the skb GRO CB data to disallow running the udp gro_receive twice
on a packet, this solves the problem of udp encapsulated packets whose inner VM
packet is udp and happen to carry a port which has registered offloads - and flush it.
- invoke the udp offload protocol registration and de-registration from the vxlan driver
in a sleepable context
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com> Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Or Gerlitz [Mon, 20 Jan 2014 11:59:19 +0000 (13:59 +0200)]
net: Add GRO support for UDP encapsulating protocols
Add GRO handlers for protocols that do UDP encapsulation, with the intent of
being able to coalesce packets which encapsulate packets belonging to
the same TCP session.
For GRO purposes, the destination UDP port takes the role of the ether type
field in the ethernet header or the next protocol in the IP header.
The UDP GRO handler will only attempt to coalesce packets whose destination
port is registered to have gro handler.
Use a mark on the skb GRO CB data to disallow (flush) running the udp gro receive
code twice on a packet. This solves the problem of udp encapsulated packets whose
inner VM packet is udp and happen to carry a port which has registered offloads.
Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com> Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vince Bridgers [Mon, 20 Jan 2014 11:39:01 +0000 (05:39 -0600)]
stmmac: Fix kernel crashes for jumbo frames
These changes correct the following issues with jumbo frames on the
stmmac driver:
1) The Synopsys EMAC can be configured to support different FIFO
sizes at core configuration time. There's no way to query the
controller and know the FIFO size, so the driver needs to get this
information from the device tree in order to know how to correctly
handle MTU changes and setting up dma buffers. The default
max-frame-size is as currently used, which is the size of a jumbo
frame.
2) The driver was enabling Jumbo frames by default, but was not allocating
dma buffers of sufficient size to handle the maximum possible packet
size that could be received. This led to memory corruption since DMAs were
occurring beyond the extent of the allocated receive buffers for certain types
of network traffic.
Vince Bridgers [Mon, 20 Jan 2014 11:39:00 +0000 (05:39 -0600)]
dts: Add a binding for Synopsys emac max-frame-size
This change adds a parameter for the Synopsys 10/100/1000
stmmac Ethernet driver to configure the maximum frame
size supported by the EMAC driver. Synopsys allows the FIFO
sizes to be configured when the cores are built for a particular
device, but do not provide a way for the driver to read
information from the device about the maximum MTU size
supported as limited by the device's FIFO size.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Dan Carpenter [Mon, 20 Jan 2014 10:28:59 +0000 (13:28 +0300)]
rxrpc: out of bound read in debug code
Smatch complains because we are using an untrusted index into the
rxrpc_acks[] array. It's just a read and it's only in the debug code,
but it's simple enough to add a check and fix it.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Yegor Yefremov [Mon, 20 Jan 2014 09:49:39 +0000 (10:49 +0100)]
8021q: update description
Replace deprecated 'vconfig' tool with 'ip' from 'iproute2'. Add
some beautifications like replacing 'ethernet' with 'Ethernet' and
removing unneeded spaces.
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
ipv6: protect protocols not handling ipv4 from v4 connection/bind attempts
Some ipv6 protocols cannot handle ipv4 addresses, so we must not allow
connecting and binding to them. sendmsg logic does already check msg->name
for this but must trust already connected sockets which could be set up
for connection to ipv4 address family.
Per-socket flag ipv6only is of no use here, as it is under users control
by setsockopt.
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: David S. Miller <davem@davemloft.net>
FX Le Bail [Sun, 19 Jan 2014 16:00:36 +0000 (17:00 +0100)]
ipv6: enable anycast addresses as source addresses in ICMPv6 error messages
- Uses ipv6_anycast_destination() in icmp6_send().
Suggested-by: Bill Fink <billfink@mindspring.com> Signed-off-by: Francois-Xavier Le Bail <fx.lebail@yahoo.com> Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Peter Pan(潘卫平) [Sun, 19 Jan 2014 12:44:46 +0000 (20:44 +0800)]
tcp: delete redundant calls of tcp_mtup_init()
As tcp_rcv_state_process() has already calls tcp_mtup_init() for non-fastopen
sock, we can delete the redundant calls of tcp_mtup_init() in
tcp_{v4,v6}_syn_recv_sock().
Signed-off-by: Weiping Pan <panweiping3@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jesse Brandeburg [Fri, 17 Jan 2014 19:00:33 +0000 (11:00 -0800)]
net: add vxlan description
Add a description to the vxlan module, helping save the world
from the minions of destruction and confusion.
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com> CC: Stephen Hemminger <stephen@networkplumber.org> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 21 Jan 2014 22:46:56 +0000 (14:46 -0800)]
Merge branch 'sfc'
Shradha Shah says:
====================
Bug Fixes for SFC driver
I am taking over the upstream patch submission work for the
sfc driver from Ben Hutchings.
These patches are bug fixes to the sfc driver involving
replacement of the PORT RESET MC command and fixing transposed
ptp_{under,over}size_sync_window_statistics
The PORT_RESET bug fix is needed for all versions supporting EF10
i.e all versions including and after 3.12.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Somehow I transposed these two while bringing the original statistics
support upstream.
Fixes: 99691c4ac112 ('sfc: Add PTP counters to ethtool stats') Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: Shradha Shah <sshah@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / linux-beck.git / log