Hauke Mehrtens [Fri, 20 Dec 2013 01:16:07 +0000 (02:16 +0100)]
b44: rename B44_PHY_ADDR_NO_PHY to B44_PHY_ADDR_NO_LOCAL_PHY
The PHY address 30 means there is no local PHY, but there could be an
external PHY like a switch connected via MII. This is the case on most
embedded home routers where this driver is used.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by: David S. Miller <davem@davemloft.net>
Hauke Mehrtens [Fri, 20 Dec 2013 01:16:06 +0000 (02:16 +0100)]
b44: check register instead of PHY address to detect external PHY
The Ethernet core supported by b44 supports an internal PHY integrated
into the mac core, which is supported by the b44 driver and an external
PHY to which the mac core is connected. This external PHY could be a
switch connected through MII, which is often the case when this core is
used on home routers. The usage of an external PHY was assumed when the
PHY address 30 was used and an internal PHY was assumed when the PHY
address was different. To verify that b44_phy_reset() was called and
checked if it worked, otherwise PHY address 30 was assumed, an external
PHY. It is better to check the register which says which PHY is
connected to the MAC instead of checking the PHY address.
The interface to an external PHY was only activated when this register
was set.
This also changes B44_FLAG_INTERNAL_PHY to B44_FLAG_EXTERNAL_PHY, it is
easier to check.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 17 Dec 2013 17:58:30 +0000 (09:58 -0800)]
tcp: autocork should not hold first packet in write queue
Willem noticed a TCP_RR regression caused by TCP autocorking
on a Mellanox test bed. MLX4_EN_TX_COAL_TIME is 16 us, which can be
right above RTT between hosts.
We can receive a ACK for a packet still in NIC TX ring buffer or in a
softnet completion queue.
Fix this by always pushing the skb if it is at the head of write queue.
Also, as TX completion is lockless, it's safer to perform sk_wmem_alloc
test after setting TSQ_THROTTLED.
Reported-by: Willem de Bruijn <willemb@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Fixes: f54b311142a9 ("tcp: auto corking") Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Fri, 20 Dec 2013 20:32:32 +0000 (12:32 -0800)]
net_sched: fix regression in tc_action_ops
list_for_each_entry(a, &act_base, head) doesn't
exit with a = NULL if we reached the end of the list.
tcf_unregister_action(), tc_lookup_action_n() and tc_lookup_action()
need fixes.
Remove tc_lookup_action_id() as its unused and not worth 'fixing'
Signed-off-by: Eric Dumazet <edumazet@google.com> Fixes: 1f747c26c48b ("net_sched: convert tc_action_ops to use struct list_head") Cc: Cong Wang <xiyou.wangcong@gmail.com> Reviewed-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Fri, 20 Dec 2013 18:04:18 +0000 (10:04 -0800)]
net_sched: fix a regression in tcf_proto_lookup_ops()
list_for_each_entry(t, &tcf_proto_base, head) doesn't
exit with t = NULL if we reached the end of the list.
Signed-off-by: Eric Dumazet <edumazet@google.com> Fixes: 3627287463b4 ("net_sched: convert tcf_proto_ops to use struct
list_head") Cc: Cong Wang <xiyou.wangcong@gmail.com> Reviewed-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
WANG Cong [Fri, 20 Dec 2013 08:08:51 +0000 (00:08 -0800)]
net_sched: fix a regression in tc actions
This patch fixes:
1) pass mask rather than size to tcf_hashinfo_init()
2) the cleanup should be in reversed order in mirred_cleanup_module()
Reported-by: Eric Dumazet <eric.dumazet@gmail.com> Fixes: 369ba56787d7469c0afd ("net_sched: init struct tcf_hashinfo at register time") Cc: Eric Dumazet <eric.dumazet@gmail.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 20 Dec 2013 00:10:14 +0000 (19:10 -0500)]
Merge branch 'sh_eth'
Sergei Shtylyov says:
====================
Add PHY IRQ support to the 'sh_eth' driver
This patchset adds support for passing PHY's IRQ to the 'sh_eth' driver from
the platform code and removes now useless code that hinders the PHY IRQ handing
by philib.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Sergei Shtylyov [Thu, 19 Dec 2013 22:41:12 +0000 (01:41 +0300)]
sh_eth: do not reset PHY needlessly
There's no need anymore to call phy_init_hw() to reset/resume the PHY from the
driver, as the call chain in phylib already has reached it, and so reset/resumed
the PHY (even resuming it twice). This duplicate reset is not only needless, it
e.g. clears the PHY's interrupt enables just setup by phylib and so prevents the
expected IRQs from the PHY.
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This patchset contains:
1. Support in ndo_get_phys_port_id added by Hadar.
2. Change the driver to use by default CQE/EQE size of 64 bytes done by Eyal.
This doubles the packet-rate of the NIC.
3. Configure the XPS queue mapping on driver load - added by Ido.
4. Fixes for some small bugs done by Jenny and Matan
Patchset was applied and tested against commit: "2372175 lib: hash: follow-up
fixups for arch hash"
Changes from V1:
- Removed Patch 10 "net/mlx4_en: Fix Supported/Advertised link mode reported by
ethtool". This patch is needed to be rewritten from scratch and I wouldn't
like it to block the rest of the patches in this set.
Also, fix to Kconfig suggested by Ben will be sent in next patchset.
Changes from V0:
- Found some issues in "Reuse memory in RX flow" patch from V0.
Removing this patch from the patchset till analyzed and fixed.
- Fix some coding style issues in patch 6 "Configure the XPS queue mapping on
driver load"
- Changed patch 9 "Add NAPI support for transmit side" to use NAPI_POLL_WEIGHT
instead of MLX4_EN_TX_BUDGET
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Matan Barak [Thu, 19 Dec 2013 19:20:18 +0000 (21:20 +0200)]
net/mlx4_core: Check port number for validity before accessing data
Need to validate port number at mlx4_promisc_qp() before use.
Since port number is extracted from gid, as a cooked or corrupted gid
could lead to a crash.
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>
Add NAPI for TX side,
implement its support and provide NAPI callback.
Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.com> Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
MLX4_DEV_EVENT_SLAVE_INIT and MLX4_DEV_EVENT_SLAVE_SHUTDOWN
events used by Hypervisor to inform the PPF IB driver that
IB para-virtualization must be initialized/destroyed for a slave.
If this event is catched by ETH VF annoying but harmless error message
is printed into dmesg. Remove dmesg prints for these events.
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.co.il> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eyal Perry [Thu, 19 Dec 2013 19:20:15 +0000 (21:20 +0200)]
net/mlx4_core: Set CQE/EQE size to 64B by default
To achieve out of the box performance default is to use 64 byte CQE/EQE.
In tests that we conduct in our labs, we achieved a performance
improvement of twice the message rate. For older VF/libmlx4 support,
enable_64b_cqe_eqe must be set to 0 (disabled).
Signed-off-by: Eyal Perry <eyalpe@mellanox.com> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Hadar Hen Zion [Thu, 19 Dec 2013 19:20:13 +0000 (21:20 +0200)]
net/mlx4_en: Implement ndo_get_phys_port_id
Use the port GUID read from the firmware to identify the physical port.
This port identifier is available via ndo_get_phys_port_id for both PF
and VF net-devices.
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Hadar Hen Zion [Thu, 19 Dec 2013 19:20:12 +0000 (21:20 +0200)]
net/mlx4_core: Expose physical port id as PF/VF capability
Add the infrastructure needed to support ndo_get_phys_port_id which
allows users to identify to which physical port a net-device is connected
to by reading a unique port id.
This will work for VFs and PFs.
The driver uses a new device capability - phys_port_id, The PF driver
reads the port phys_port_id from Firmware and stores it. The VF driver
reads the port phys_port_id from the PF using QUERY_FUNC_CAP command.
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Hadar Hen Zion [Thu, 19 Dec 2013 19:20:11 +0000 (21:20 +0200)]
net/mlx4_core: Introduce nic_info new flag in QUERY_FUNC_CAP
Set nic_info field in QUERY_FUNC_CAP, which designates
supplementary NIC information is provided by the hypervisor.
When set, the following fields are valid: nic_num_rings,
nic_indirection_tbl_sz, cur_mac and phys_port_id.
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Hadar Hen Zion [Thu, 19 Dec 2013 19:20:10 +0000 (21:20 +0200)]
net/mlx4_core: Rename QUERY_FUNC_CAP fields
Use correct names for QUERY_FUNC_CAP fields: flags0 and flags1.
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Hadar Hen Zion [Thu, 19 Dec 2013 19:20:09 +0000 (21:20 +0200)]
net/mlx4_core: Remove zeroed out of explicit QUERY_FUNC_CAP fields
All mailboxes are already zeroed by commit: 571b8b9 net/mlx4_core: Initialize all mailbox buffers to zero before use
Remove explicit zero set for force mac and force vlan fields in
mlx4_QUERY_FUNC_CAP_wrapper
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com> Signed-off-by: Amir Vadai <amirv@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 18 Dec 2013 14:46:23 +0000 (06:46 -0800)]
rps: NUMA flow limit allocations
Given we allocate memory for each cpu, we can do this
using NUMA affinities, instead of using NUMA policies
of the process changing flow_limit_cpu_bitmap value.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Rashika Kheria [Thu, 19 Dec 2013 08:55:34 +0000 (14:25 +0530)]
drivers: net: Mark functions as static in debug.c
This patch marks the function il_clear_traffic_stats() in debug.c as
static because they are not used outside this file.
Thus, it also removes the following warnings in
wireless/iwlegacy/debug.c:
drivers/net/wireless/iwlegacy/debug.c:35:1: warning: no previous prototype for ‘il_clear_traffic_stats’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Rashika Kheria [Thu, 19 Dec 2013 08:54:38 +0000 (14:24 +0530)]
drivers: net: Mark functions as static in 4965-debug.c
This patch marks the function il4965_ucode_rx_stats_read(),
il4965_ucode_tx_stats_read() and il4965_ucode_general_stats_read() in
4965-debug.c as static because they are not used outside this file.
Thus, it also removes the following warnings in
wireless/iwlegacy/4965-debug.c:
drivers/net/wireless/iwlegacy/4965-debug.c:59:1: warning: no previous prototype for ‘il4965_ucode_rx_stats_read’ [-Wmissing-prototypes]
drivers/net/wireless/iwlegacy/4965-debug.c:471:1: warning: no previous prototype for ‘il4965_ucode_tx_stats_read’ [-Wmissing-prototypes]
drivers/net/wireless/iwlegacy/4965-debug.c:637:1: warning: no previous prototype for ‘il4965_ucode_general_stats_read’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Rashika Kheria [Thu, 19 Dec 2013 08:53:49 +0000 (14:23 +0530)]
drivers: net: Mark functions as static in 3945-debug.c
This patch marks the function il3945_ucode_rx_stats_read(),
il3945_ucode_tx_stats_read() and il3945_ucode_general_stats_read() in
3945-debug.c as static because they are not used outside this file.
Thus, it also removes the following warnings in
wireless/iwlegacy/3945-debug.c:
drivers/net/wireless/iwlegacy/3945-debug.c:52:1: warning: no previous prototype for ‘il3945_ucode_rx_stats_read’ [-Wmissing-prototypes]
drivers/net/wireless/iwlegacy/3945-debug.c:317:1: warning: no previous prototype for ‘il3945_ucode_tx_stats_read’ [-Wmissing-prototypes]
drivers/net/wireless/iwlegacy/3945-debug.c:407:1: warning: no previous prototype for ‘il3945_ucode_general_stats_read’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Rashika Kheria [Thu, 19 Dec 2013 08:52:37 +0000 (14:22 +0530)]
drivers: net: Include new header file in sbni.c
Create a new header file include/net/Space.h which contains
prototype declaration of sbni_probe().
Include the new header file in drivers/net/Space.c and
drivers/net/wan/sbni.c because they use this function.
This eliminates the following warning in wan/sbni.c:
drivers/net/wan/sbni.c:224:12: warning: no previous prototype for ‘sbni_probe’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Rashika Kheria [Thu, 19 Dec 2013 08:49:44 +0000 (14:19 +0530)]
drivers: net: Mark functions as static in stmmac_platform.c
This patch marks the function stmmac_pltfr_freeze() and
stmmac_pltfr_restore() in stmmac_platform.c as static because they are
not used outside this file.
Thus, it also removes the following warnings in
ethernet/stmicro/stmmac/stmmac_platform.c:
drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c:222:5: warning: no previous prototype for ‘stmmac_pltfr_freeze’ [-Wmissing-prototypes]
drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c:236:5: warning: no previous prototype for ‘stmmac_pltfr_restore’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Rashika Kheria [Thu, 19 Dec 2013 08:48:11 +0000 (14:18 +0530)]
drivers: net: Mark functions as static in qlcnic_83xx_hw.c
This patch marks the function qlcnic_83xx_clear_legacy_intr_mask(),
qlcnic_83xx_set_legacy_intr_mask() and
qlcnic_83xx_enable_legacy_msix_mbx_intr() in qlcnic_83xx_hw.c as static
because they are not used outside this file.
Thus, it also removes the following warnings
in ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c:
drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c:318:13: warning: no previous prototype for ‘qlcnic_83xx_clear_legacy_intr_mask’ [-Wmissing-prototypes]
drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c:323:13: warning: no previous prototype for ‘qlcnic_83xx_set_legacy_intr_mask’ [-Wmissing-prototypes]
drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c:343:13: warning: no previous prototype for ‘qlcnic_83xx_enable_legacy_msix_mbx_intr’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Rashika Kheria [Thu, 19 Dec 2013 08:47:27 +0000 (14:17 +0530)]
drivers: net: Mark functions as static in qlcnic_io.c
This patch marks the function qlcnic_enable_tx_intr(),
qlcnic_83xx_enable_tx_intr() and qlcnic_83xx_disable_tx_intr() in
qlcnic_io.c as static because they are not used outside this file.
Thus, it also removes the following warnings in
ethernet/qlogic/qlcnic/qlcnic_io.c:
drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c:130:13: warning: no previous prototype for ‘qlcnic_enable_tx_intr’ [-Wmissing-prototypes]
drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c:147:13: warning: no previous prototype for ‘qlcnic_83xx_enable_tx_intr’ [-Wmissing-prototypes]
drivers/net/ethernet/qlogic/qlcnic/qlcnic_io.c:153:13: warning: no previous prototype for ‘qlcnic_83xx_disable_tx_intr’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org> Signed-off-by: David S. Miller <davem@davemloft.net>
1) Use the user supplied policy index instead of a generated one
if present. From Fan Du.
2) Make xfrm migration namespace aware. From Fan Du.
3) Make the xfrm state and policy locks namespace aware. From Fan Du.
4) Remove ancient sleeping when the SA is in acquire state,
we now queue packets to the policy instead. This replaces the
sleeping code.
5) Remove FLOWI_FLAG_CAN_SLEEP. This was used to notify xfrm about the
posibility to sleep. The sleeping code is gone, so remove it.
6) Check user specified spi for IPComp. Thr spi for IPcomp is only
16 bit wide, so check for a valid value. From Fan Du.
7) Export verify_userspi_info to check for valid user supplied spi ranges
with pfkey and netlink. From Fan Du.
8) RFC3173 states that if the total size of a compressed payload and the IPComp
header is not smaller than the size of the original payload, the IP datagram
must be sent in the original non-compressed form. These packets are dropped
by the inbound policy check because they are not transformed. Document the need
to set 'level use' for IPcomp to receive such packets anyway. From Fan Du.
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Tested with modified iproute2 to verify attributes can be set at bond creation
time or set later. Verified sysfs interface to attributes continues to work.
I have one more set after this one for all of the 802.3ad attributes. Whew.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: add packets_per_slave attribute netlink support
Add IFLA_BOND_PACKETS_PER_SLAVE to allow get/set of bonding parameter
packets_per_slave via netlink.
Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com> Signed-off-by: Jiri Pirko <jiri@resnulli.us> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: add lp_interval attribute netlink support
Add IFLA_BOND_LP_INTERVAL to allow get/set of bonding parameter
lp_interval via netlink.
Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com> Signed-off-by: Jiri Pirko <jiri@resnulli.us> Signed-off-by: David S. Miller <davem@davemloft.net>
Add IFLA_BOND_MIN_LINKS to allow get/set of bonding parameter
min_links via netlink.
Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com> Signed-off-by: Jiri Pirko <jiri@resnulli.us> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: add all_slaves_active attribute netlink support
Add IFLA_BOND_ALL_SLAVES_ACTIVE to allow get/set of bonding parameter
all_slaves_active via netlink.
Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com> Signed-off-by: Jiri Pirko <jiri@resnulli.us> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: add num_grat_arp attribute netlink support
Add IFLA_BOND_NUM_PEER_NOTIF to allow get/set of bonding parameter
num_grat_arp via netlink. Bonding parameter num_unsol_na is
synonymous with num_grat_arp, so add only one netlink attribute
to represent both bonding parameters.
Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com> Signed-off-by: Jiri Pirko <jiri@resnulli.us> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 19 Dec 2013 20:13:14 +0000 (15:13 -0500)]
Merge branch 'for-davem' of git://gitorious.org/linux-can/linux-can-next
Marc Kleine-Budde says:
====================
this is a pull request of four patches for net-next/master.
There is one patch by Markus Pargmann, which speeds up the c_can
driver, a patch by John Whitmore which updates the in tree
documentation. A patch by Jeff Kirsher which replaces the FSF's address
by a link and a patch by Alexander Shiyan which converts the mcp251x
driver to make use of managed resources.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Paul Durrant [Tue, 17 Dec 2013 11:44:35 +0000 (11:44 +0000)]
xen-netback: add gso_segs calculation
netback already has code which parses IPv4 and v6 headers to set up checksum
offsets and these are always applied to GSO packets being sent from
frontends. It's therefore suboptimal that GSOs are being marked
SKB_GSO_DODGY to defer the gso_segs calculation when netback already has all
necessary information to hand to do the calculation. This patch adds that
calculation.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com> Cc: Wei Liu <wei.liu2@citrix.com> Cc: Ian Campbell <ian.campbell@citrix.com> Cc: David Vrabel <david.vrabel@citrix.com> Cc: Eric Dumazet <eric.dumazet@gmail.com> Acked-by: Eric Dumazet <edumazet@google.com> Acked-by: Wei Liu <wei.liu2@citrix.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Terry Lam [Sun, 15 Dec 2013 08:30:21 +0000 (00:30 -0800)]
net-qdisc-hhf: Heavy-Hitter Filter (HHF) qdisc
This patch implements the first size-based qdisc that attempts to
differentiate between small flows and heavy-hitters. The goal is to
catch the heavy-hitters and move them to a separate queue with less
priority so that bulk traffic does not affect the latency of critical
traffic. Currently "less priority" means less weight (2:1 in
particular) in a Weighted Deficit Round Robin (WDRR) scheduler.
In essence, this patch addresses the "delay-bloat" problem due to
bloated buffers. In some systems, large queues may be necessary for
obtaining CPU efficiency, or due to the presence of unresponsive
traffic like UDP, or just a large number of connections with each
having a small amount of outstanding traffic. In these circumstances,
HHF aims to reduce the HoL blocking for latency sensitive traffic,
while not impacting the queues built up by bulk traffic. HHF can also
be used in conjunction with other AQM mechanisms such as CoDel.
To capture heavy-hitters, we implement the "multi-stage filter" design
in the following paper:
C. Estan and G. Varghese, "New Directions in Traffic Measurement and
Accounting", in ACM SIGCOMM, 2002.
Some configurable qdisc settings through 'tc':
- hhf_reset_timeout: period to reset counter values in the multi-stage
filter (default 40ms)
- hhf_admit_bytes: threshold to classify heavy-hitters
(default 128KB)
- hhf_evict_timeout: threshold to evict idle heavy-hitters
(default 1s)
- hhf_non_hh_weight: Weighted Deficit Round Robin (WDRR) weight for
non-heavy-hitters (default 2)
- hh_flows_limit: max number of heavy-hitter flow entries
(default 2048)
Note that the ratio between hhf_admit_bytes and hhf_reset_timeout
reflects the bandwidth of heavy-hitters that we attempt to capture
(25Mbps with the above default settings).
The false negative rate (heavy-hitter flows getting away unclassified)
is zero by the design of the multi-stage filter algorithm.
With 100 heavy-hitter flows, using four hashes and 4000 counters yields
a false positive rate (non-heavy-hitters mistakenly classified as
heavy-hitters) of less than 1e-4.
Signed-off-by: Terry Lam <vtlam@google.com> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Francesco Fusco [Wed, 18 Dec 2013 15:05:48 +0000 (16:05 +0100)]
lib: hash: follow-up fixups for arch hash
This patch adds the include file to pull in __read_mostly on some
architectures e.g. ppc and also fixes up signatures in generic
asm.
Reported-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Francesco Fusco <ffusco@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
tanxiaojun [Mon, 16 Dec 2013 13:32:46 +0000 (21:32 +0800)]
bridge: spelling fixes
Fix spelling errors in bridge driver.
Signed-off-by: Tan Xiaojun <tanxiaojun@huawei.com> Acked-by: Stephen Hemminger <stephen@networkplumber.org> Signed-off-by: David S. Miller <davem@davemloft.net>
IPV6_PMTU_INTERFACE is the same as IPV6_PMTU_PROBE for ipv6. Add it
nontheless for symmetry with IPv4 sockets. Also drop incoming MTU
information if this mode is enabled.
The additional bit in ipv6_pinfo just eats in the padding behind the
bitfield. There are no changes to the layout of the struct at all.
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: David S. Miller <davem@davemloft.net>
ipv4: new ip_no_pmtu_disc mode to always discard incoming frag needed msgs
This new mode discards all incoming fragmentation-needed notifications
as I guess was originally intended with this knob. To not break backward
compatibility too much, I only added a special case for mode 2 in the
receiving path.
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 18 Dec 2013 19:58:41 +0000 (14:58 -0500)]
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next
Jeff Kirsher says:
====================
Intel Wired LAN Driver Updates
This series contains updates to i40e, ixgbevf, ixgbe and igb.
Don provides an ixgbevf patch to add DCB configuration into the queue
setup so that we won't have to allocate queues in a separate place when
enabling DCB.
Guenter Roeck provides 2 patches for ixgbe to simplify the code by
attaching hwmon sysfs attributes to hwmon device instead of PCI device.
Also fix an issues where the temperature sensor attribute index was
being started with the value 0 and not 1 as per the hwmon API.
Carolyn provides igb patches to fix queue allocation method to
accommodate changes during runtime. This includes changing how the
driver initializes MSIx and checks for MSIx configuration to make it
easier to reconfigure the device when queue changes happen at runtime.
Neerav and Shannon fixes i40e debugfs commands that dump hex information
by using print_hex_dump().
Shannon provides several i40e fixes which include the prevention of
null pointer exception in the dump descriptor by checking that rings
were allocated before trying to reference them. Fixed up a couple of
scanfs to accept various base numbers instead of silently requiring hex.
Anjali fixes up i40e where the incorrect defines were being used for
misc interrupts.
Alan Cox provides a fix for i40e where we assume that the resulting
buffer is zero terminated when we then re-use it. The sscanf is limited
to 512 bytes but needs to be 511 to allow for a terminator.
Stephen Hemminger fixes i40e by making local functions static and removes
unused code (i40e_aq_add/remove_vlan() functions).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
WANG Cong [Mon, 16 Dec 2013 04:15:11 +0000 (20:15 -0800)]
net_sched: convert tcf_proto_ops to use struct list_head
We don't need to maintain our own singly linked list code.
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
WANG Cong [Mon, 16 Dec 2013 04:15:10 +0000 (20:15 -0800)]
net_sched: convert tc_action_ops to use struct list_head
We don't need to maintain our own singly linked list code.
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
WANG Cong [Mon, 16 Dec 2013 04:15:09 +0000 (20:15 -0800)]
net_sched: convert tcf_hashinfo to hlist and use spinlock
So that we don't need to play with singly linked list,
and since the code is not on hot path, we can use spinlock
instead of rwlock.
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
WANG Cong [Mon, 16 Dec 2013 04:15:08 +0000 (20:15 -0800)]
net_sched: init struct tcf_hashinfo at register time
It looks weird to store the lock out of the struct but
still points to a static variable. Just move them into the struct.
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
WANG Cong [Mon, 16 Dec 2013 04:15:07 +0000 (20:15 -0800)]
net_sched: cls: refactor out struct tcf_ext_map
These information can be saved in tcf_exts, and this will
simplify the code.
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
WANG Cong [Mon, 16 Dec 2013 04:15:05 +0000 (20:15 -0800)]
net_sched: act: use standard struct list_head
Currently actions are chained by a singly linked list,
therefore it is a bit hard to add and remove a specific
entry. Convert it to struct list_head so that in the
latter patch we can remove an action without finding
its head.
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
WANG Cong [Mon, 16 Dec 2013 04:15:04 +0000 (20:15 -0800)]
net_sched: remove get_stats from tc_action_ops
It is not used.
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Carolyn Wyborny [Tue, 10 Dec 2013 07:58:34 +0000 (07:58 +0000)]
igb: Change to use statically allocated array for MSIx entries
This patch changes how the driver initializes MSIx and checks
for MSIx configuration. This change makes it easier to reconfigure the
device when queue changes happen at runtime using ethtool's set_channels
feature.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com> Tested-by: Aaron Brown <aaron.f.brown@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Carolyn Wyborny [Tue, 10 Dec 2013 07:58:29 +0000 (07:58 +0000)]
igb: Fix queue allocation method to accommodate changing during runtime
When changing number of queues using ethtool's set_channels during runtime,
a queue allocation could fail, which can leave the device in a down state.
In order to preserve the usability of the device in this scenario, this patch
changes the driver to allocate the number of queues only if they have not
been allocated already. The first allocation is then done for the max number
of queues, which is the default queues for this driver. With this change,
queue quantity changes are not subject to queue allocation failures.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com> Tested-by: Aaron Brown <aaron.f.brown@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Guenter Roeck [Tue, 26 Nov 2013 07:15:28 +0000 (07:15 +0000)]
ixgbe: Convert to use devm_hwmon_device_register_with_groups
Simplify the code. Attach hwmon sysfs attributes to hwmon device
instead of pci device. Avoid race conditions caused by attributes
being created after hwmon device registration. Implicitly
(through hwmon API) add mandatory 'name' sysfs attribute.
Other cleanup:
Instead of allocating memory for hwmon attributes, move attributes
and all other hwmon related data into struct hwmon_buff and allocate
the entire structure using devm_kzalloc.
Check return value from calls to igb_add_hwmon_attr() one by one instead
of logically combining them all together.
Signed-off-by: Guenter Roeck <linux@roeck-us.net> Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Don Skidmore [Sat, 21 Sep 2013 01:40:49 +0000 (01:40 +0000)]
ixgbevf: add DCB configuration into queue setup
This patch takes the DCB config checks and adds them to the normal setting
up of the queues. This way we won't have to allocation queues in a separate
place for enabling DCB.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com> Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com> Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com> Tested-By: Jack Morgan<jack.morgan@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Alan Cox [Thu, 12 Dec 2013 02:44:24 +0000 (02:44 +0000)]
i40e: Fix off by one in i40e_dbg_command_write
We assume that the resulting buffer is zero terminated when we then
re-use it. The sscanf is limited to 512 bytes but needs to be 511
to allow for a terminator.
One of a set of problems noted by Jackie Chang
Signed-off-by: Alan Cox <alan@linux.intel.com> Acked-by: Shannon Nelson <Shannon.nelson@intel.com> Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
David S. Miller [Wed, 18 Dec 2013 05:36:51 +0000 (00:36 -0500)]
Merge branch 'vlan_tpid'
Atzm Watanabe says:
====================
packet: deliver VLAN TPID to userspace
This patchset enables userspace to get VLAN TPID as well as the VLAN TCI.
After the 802.1AD support, userspace packet receivers (packet dumper,
software switch, and the like) need how to know VLAN TPID in order to
reconstruct original tagged frame.
v4: Simply use sizeof(tp_padding) for zeroing the padding bytes,
commented by David Laight.
Use __u16 for tp_vlan_tpid in tpacket_hdr_variant1,
commented by Daniel Borkmann.
v3: Add a definition which indicates whether tp_vlan_tpid is valid.
Explicitly define pad bytes for tpacket{2,3}_hdr and pick the area
for tp_vlan_tpid from the definition. Commented by David Laight.
v2: Add BUILD_BUG_ON() to make current aligned size of
struct tpacket{2,3}_hdr clear. Commented by Ben Hutchings.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
projects / linux-beck.git / log