| Age | Commit message (Collapse) | Author |
|---|
|
Merge mlx4 bug fixes from Amir Vadai.
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Under heavy CPU load, changing, ring size/mtu/etc. could result in transmit
timeout, since stop-start port might take more than 10 seconds.
Calling netif_detach_device to prevent tx queue transmit timeout.
netif_detach_device() is not called under ndo_stop, because netif_carrier_off
will prevent the timeout, and device should not be marked as not present, or
else user won't be able to start it later on.
CC: Ben Hutchings <bhutchings@solarflare.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>
|
|
Mac reassignments should only be done when not supported by the firmware. To
accomplish that, checking firmware capability bit to know whether we should
reassign macs in the driver.
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>
|
|
The Firmware dynamically changes flow steering hash configuration from covering
L2 only to "full" L2/L3/L4 mode needed. The dynamic change allows the driver
to set hard coded hash configuration which is changed by the firmware from L2
to L2/L3/L4 when attaching the first L3/L4 flow steering rule and back to L2
when there are no more such rules.
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>
|
|
As part of the driver unload flow, all steering rules must be deleted,
make sure to remove the rules that were set through ethtool.
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>
|
|
is down
Attaching steering rules while the interface is down is an invalid operation, block it.
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>
|
|
The vlan mask field should be validated and assigned according to the field
size which is 12 bits. Also replace the numeric 0xfff mask with existing kernel
macro.
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>
|
|
When attaching flow steering rules via Ethtool accept only valid vlans IDs e.g
in the range: [0,4095].
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>
|
|
Destination mac is a mandatory specification for ip/udp steering rules.
When attaching multicast steering rules via ethtool the unicast mac of the
interface was added to the rule specification instead of the multicast mac.
The following commit sets the corresponding multicast mac for the rule multicast ip.
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>
|
|
The allow_loopback flag was wrongly set using arithmetic bit operation, change
the code to use logical bit operation.
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>
|
|
Some of the fields for struct mlx4_net_trans_rule_hw_ctrl were packed into u32
and accessed through bit field operations. Expose and access them directly as
u8.
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>
|
|
Implement ethtool get_drvinfo.
Signed-off-by: Yan Burman <yanb@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Since all users are write-lock, it does not make sense to use
rwlock here. Use simple spinlock.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Supporting access to skb->pkt_type is a bit tricky if we want
to have a generic code, allowing pkt_type to be moved in struct sk_buff
pkt_type is a bit field, so compiler cannot really help us to find
its offset. Let's use a helper for this : It will throw a one time
message if pkt_type no longer starts at a byte boundary or is
no longer a 3bit field.
Reported-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Manish chopra <manish.chopra@qlogic.com>
Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
o ipv4 address was not getting programmed properly because of
improper byte order conversion
Signed-off-by: Manish chopra <manish.chopra@qlogic.com>
Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Manish chopra <manish.chopra@qlogic.com>
Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Manish chopra <manish.chopra@qlogic.com>
Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Makes more sense to have randomly generated address by default than to
have all zeroes. It also allows user to for example put the bond into
bridge without need to have any slaves in it.
Also note that this changes only behaviour of bonds with no slaves. Once
the first slave device is enslaved, its address will be used (no change
here).
Also, fix dev_assign_type values on the way.
Reported-by: Pavel Šimerda <psimerda@redhat.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Instead of jumping aroung bugs that are easily fixed just don't let them in:
affected drivers should be either fixed or have NETIF_F_HW_VLAN_FILTER
removed from advertised features.
Quick grep in drivers/net shows two drivers that have NETIF_F_HW_VLAN_FILTER
but not ndo_vlan_rx_add/kill_vid(), but those are false-positives (features
are commented out).
OTOH two drivers have ndo_vlan_rx_add/kill_vid() implemented but don't
advertise NETIF_F_HW_VLAN_FILTER. Those are:
+ethernet/cisco/enic/enic_main.c
+ethernet/qlogic/qlcnic/qlcnic_main.c
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
All users of xfrm_addr_cmp() use its result as boolean.
Introduce xfrm_addr_equal() (which is equal to !xfrm_addr_cmp())
and convert all users.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Included changes:
- fix recently introduced output behaviour
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Bring in the 'net' tree so that we can get some ipv4/ipv6 bug
fixes that some net-next work will build upon.
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This patch adds anti-spoofing checks in sit.c as specified in RFC3964
section 5.2 for 6to4 and RFC5969 section 12 for 6rd. I left out the
checks which could easily be implemented with netfilter.
Specifically this patch adds following logic (based loosely on the
pseudocode in RFC3964 section 5.2):
if prefix (inner_src_v6) == rd6_prefix (2002::/16 is the default)
and outer_src_v4 != embedded_ipv4 (inner_src_v6)
drop
if prefix (inner_dst_v6) == rd6_prefix (or 2002::/16 is the default)
and outer_dst_v4 != embedded_ipv4 (inner_dst_v6)
drop
accept
To accomplish the specified security checks proposed by above RFCs,
it is still necessary to employ uRPF filters with netfilter. These new
checks only kick in if the employed addresses are within the 2002::/16 or
another range specified by the 6rd-prefix (which defaults to 2002::/16).
Cc: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
* remove unused members(!): imask, ievent
* move space consuming interrupt name strings (int_name_* members) to
external structures, unessential for the driver's hot path
* keep high priority hot path data within the first 2 cache lines
This reduces struct gfar_priv_grp from 6 to 3 cache lines.
(Also fixed checkpatch warnings for the old code, in the process.)
Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Factor out redundant code (improve readability, source code size).
Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Resize and regroup structure members to eliminate memory holes and
to pack the structure into 2 cache lines (from 3).
tx_ring_size was resized from 4 to 2 bytes and few members were re-grouped
in order to eliminate byte holes and achieve compactness.
Where possible, few members were grouped according to their usage and access
order (i.e. start_xmit vs. clean_tx_ring members), less important members
were pushed at the end.
Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When attempting to build linux-next with user namespaces enabled I ran
into this fun build error.
CC net/ipv6/inet6_connection_sock.o
.../net/ipv6/inet6_connection_sock.c: In function ‘inet6_csk_bind_conflict’:
.../net/ipv6/inet6_connection_sock.c:37:12: error: incompatible types when initializing type ‘int’ using
type ‘kuid_t’
.../net/ipv6/inet6_connection_sock.c:54:30: error: incompatible type for argument 1 of ‘uid_eq’
.../include/linux/uidgid.h:48:20: note: expected ‘kuid_t’ but argument is of type ‘int’
make[3]: *** [net/ipv6/inet6_connection_sock.o] Error 1
make[2]: *** [net/ipv6] Error 2
make[2]: *** Waiting for unfinished jobs....
Using kuid_t instead of int to hold the uid fixes this.
Cc: Tom Herbert <therbert@google.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
v3: make pktgen_threads list per-namespace
v2: remove a useless check
This patch add net namespace to pktgen, so that
we can use pktgen in different namespaces.
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Add napi support
Before this patch
iperf -s -i 1
------------------------------------------------------------
Server listening on TCP port 5001
TCP window size: 85.3 KByte (default)
------------------------------------------------------------
[ 4] local 10.192.242.153 port 5001 connected with 10.192.242.138 port 50004
[ ID] Interval Transfer Bandwidth
[ 4] 0.0- 1.0 sec 41.2 MBytes 345 Mbits/sec
[ 4] 1.0- 2.0 sec 43.7 MBytes 367 Mbits/sec
[ 4] 2.0- 3.0 sec 42.8 MBytes 359 Mbits/sec
[ 4] 3.0- 4.0 sec 43.7 MBytes 367 Mbits/sec
[ 4] 4.0- 5.0 sec 42.7 MBytes 359 Mbits/sec
[ 4] 5.0- 6.0 sec 43.8 MBytes 367 Mbits/sec
[ 4] 6.0- 7.0 sec 43.0 MBytes 361 Mbits/sec
After this patch
[ 4] 2.0- 3.0 sec 51.6 MBytes 433 Mbits/sec
[ 4] 3.0- 4.0 sec 51.8 MBytes 435 Mbits/sec
[ 4] 4.0- 5.0 sec 52.2 MBytes 438 Mbits/sec
[ 4] 5.0- 6.0 sec 52.1 MBytes 437 Mbits/sec
[ 4] 6.0- 7.0 sec 52.1 MBytes 437 Mbits/sec
[ 4] 7.0- 8.0 sec 52.3 MBytes 439 Mbits/sec
Signed-off-by: Frank Li <Frank.Li@freescale.com>
Signed-off-by: Fugang Duan <B38611@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Cleanup the format of ethoc.c to meet network driver style as
per checkpatch.pl.
Signed-off-by: Barry Grussling <barry@grussling.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
A GRE tunnel can be configured so that outgoing tunnel packets inherit
the value of the TOS field from the inner IP header. In doing so, when
a non-IP packet is transmitted through the tunnel, the TOS field will
always be set to 0.
Instead, the user should be able to configure a different TOS value as
the fallback to use for non-IP packets. This is helpful when the non-IP
packets are all control packets and should be handled by routers outside
the tunnel as having Internet Control precedence. One example of this is
the NHRP packets that control a DMVPN-compatible mGRE tunnel; they are
encapsulated directly by GRE and do not contain an inner IP header.
Under the existing behavior, the IFLA_GRE_TOS parameter must be set to
'1' for the TOS value to be inherited. Now, only the least significant
bit of this parameter must be set to '1', and when a non-IP packet is
sent through the tunnel, the upper 6 bits of this same parameter will be
copied into the TOS field. (The ECN bits get masked off as before.)
This behavior is backwards-compatible with existing configurations and
iproute2 versions.
Signed-off-by: David Ward <david.ward@ll.mit.edu>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
There are some usecase when lifetime of ipv4 addresses might be helpful.
For example:
1) initramfs networkmanager uses a DHCP daemon to learn network
configuration parameters
2) initramfs networkmanager addresses, routes and DNS configuration
3) initramfs networkmanager is requested to stop
4) initramfs networkmanager stops all daemons including dhclient
5) there are addresses and routes configured but no daemon running. If
the system doesn't start networkmanager for some reason, addresses and
routes will be used forever, which violates RFC 2131.
This patch is essentially a backport of ivp6 address lifetime mechanism
for ipv4 addresses.
Current "ip" tool supports this without any patch (since it does not
distinguish between ipv4 and ipv6 addresses in this perspective.
Also, this should be back-compatible with all current netlink users.
Reported-by: Pavel Šimerda <psimerda@redhat.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Jesper Dangaard Brouer says:
====================
This patchset is V2, with some trivial code fixes, which were noticed
by DaveM. It is still a partly respin of my fragmentation optimization
patches: http://thread.gmane.org/gmane.linux.network/250914
This is not the complete patchset, from the gmane link above. In this
patchset, I primarily focus on adjusting cacheline for better SMP/NUMA
performance.
Once this patchset have been agreed upon, I will continue and respin
the rest of my patches.
This time around, I have created a frag DoS generator, via the tool
trafgen (http://netsniff-ng.org/). To create a stable DoS scenario
(no longer relying on frame dropping due to disabled flow-control).
Two 10G interfaces are under-test, and uses Ethernet flow-control. A
third interface is used for generating the DoS attack (this interface
is also 10G, but it does not need to be, as 500Kpps DoS is enough).
Test types summary (netperf):
Test-20G64K == 2x10G with 65K fragments
Test-20G3F == 2x10G with 3x fragments (3*1472 bytes)
Test-20G64K+DoS == Same as 20G64K with frag DoS
Test-20G3F+DoS == Same as 20G3F with frag DoS
Patch list:
Patch-01 - net: cacheline adjust struct netns_frags for better frag performance
Patch-02 - net: cacheline adjust struct inet_frags for better frag performance
Patch-03 - net: cacheline adjust struct inet_frag_queue
Patch-04 - net: frag helper functions for mem limit tracking
Patch-05 - net: use lib/percpu_counter API for fragmentation mem accounting
Patch-06 - net: frag, move LRU list maintenance outside of rwlock
Performance table summary:
Test-type: Test-20G64K Test-20G3F 20G64K+DoS 20G3F+DoS
---------- ----------- ---------- ---------- ---------
net-next: 15114.5 Mbit/s 8954.21 2444.28 3918.01 Mbit/s
Patch-01: 16075.8 Mbit/s 8976.18 2621.49 4072.79 Mbit/s
Patch-02: 17806.9 Mbit/s 9280.32 2478.62 4274.59 Mbit/s
Patch-03: 17317.4 Mbit/s 9308.62 2546.05 4336.59 Mbit/s
Patch-04: 17635.9 Mbit/s 9256.16 2535.25 4327.63 Mbit/s
Patch-05: 18027.0 Mbit/s 9918.99 2492.62 3621.68 Mbit/s
Patch-06: 18486.7 Mbit/s 10723.20 3657.85 4560.64 Mbit/s
I cannot explain the under-DoS regression that patch-05/percpu_counter
introduces. But patch-06/LRU-lock corrects the situation again.
Below is a testlab setup description, with links to the trafgen DoS
packet config used.
Testlab
=======
Server setup
------------
The machine acting as a server:
- 2x CPU (E5-2630)
- Thus a NUMA arch/machine
- 4x 10Gbit/s ports
- NICs 2x Intel Dual port 82599 based (driver ixgbe)
Setup:
- Interfaces uses Ethernet flow control
- Flush all iptables
- Remove all iptables related module.
- Kill irqbalance
- Pin each 10G NIC port to a *single* CPU each
Pinning can easily be done by command hacks::
for x in /proc/irq/*/eth8*/../smp_affinity_list ; do echo 1 > $x; done
for x in /proc/irq/*/eth9*/../smp_affinity_list ; do echo 3 > $x; done
for x in /proc/irq/*/eth31*/../smp_affinity_list; do echo 6 > $x; done
for x in /proc/irq/*/eth32*/../smp_affinity_list; do echo 8 > $x; done
Notice NUMA setting: The CPU to NIC tying is carefully choosen
according to the NUMA node setup. Thus, NICs connected to a PCI-e
slot that is connected to a physical CPU socket are tied together.
Choosing only a single CPU per NIC (port) is just to ease provoking
and debugging this performance issue. (In real setups, you can choose
more CPU, just remember the NUMA node in the equation).
Tools
-----
Netperf is used, with option -T to ensure CPU binding.
The netserver processes, are NAPI pinned::
numactl -m0 -c0 netserver
numactl -m1 -c 1 netserver -p 1337
I now have a frag DoS generator, created via the tool:
trafgen (see: http://netsniff-ng.org/)
Trafgen packet config file:
http://people.netfilter.org/hawk/frag_work/trafgen/frag_packet03_small_frag.txf
Notice, I'm using features of trafgen, recently developed by Daniel
Borkmann, thus you need the latest git tree to use my trafgen packet
config.
git://github.com/borkmann/netsniff-ng.git
Command line:
trafgen --dev eth51 --conf frag_packet03_small_frag.txf -V -k 100 --cpus 2
Tests types
-----------
Test(20G64K) UDP-64K 2x 10Gbit/s with no DoS traffic:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
export SIZE=$((65507)); export TIME=$((20)); export LOG=/tmp/netperf.log ;\
netperf -p 1337 -H 192.168.31.2 -T7,7 -t UDP_STREAM -l $TIME -- -m $SIZE >> ${LOG}.31 &\
netperf -H 192.168.81.2 -T2,2 -t UDP_STREAM -l $TIME -- -m $SIZE >> ${LOG}.81 && \
wait $! && tail -n3 ${LOG}.* && \
tail -n3 ${LOG}.{31,81} | awk 'BEGIN{sum=0;} /212992 / {sum+=$4; print " +"$4} /==/ {print " file:"$2} END{print "sum:"sum" Mbit/s"}'
Test(20G3F) UDP-3xfrags 2x 10Gbit/s with no DoS traffic:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
export SIZE=$((3*1472)); export TIME=$((20)); export LOG=/tmp/netperf.log ;\
netperf -p 1337 -H 192.168.31.2 -T7,7 -t UDP_STREAM -l $TIME -- -m $SIZE >> ${LOG}.31 &\
netperf -H 192.168.81.2 -T2,2 -t UDP_STREAM -l $TIME -- -m $SIZE >> ${LOG}.81 && \
wait $! && tail -n3 ${LOG}.* && \
tail -n3 ${LOG}.{31,81} | awk 'BEGIN{sum=0;} /212992 / {sum+=$4; print " +"$4} /==/ {print " file:"$2} END{print "sum:"sum" Mbit/s"}'
Awk script for summming results:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
tail -n3 ${LOG}.{31,81} | awk 'BEGIN{sum=0;} /212992 / {sum+=$4; print " +"$4} /==/ {print " file:"$2} END{print "sum:"sum" Mbit/s"}'
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Updating the fragmentation queues LRU (Least-Recently-Used) list,
required taking the hash writer lock. However, the LRU list isn't
tied to the hash at all, so we can use a separate lock for it.
Original-idea-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Replace the per network namespace shared atomic "mem" accounting
variable, in the fragmentation code, with a lib/percpu_counter.
Getting percpu_counter to scale to the fragmentation code usage
requires some tweaks.
At first view, percpu_counter looks superfast, but it does not
scale on multi-CPU/NUMA machines, because the default batch size
is too small, for frag code usage. Thus, I have adjusted the
batch size by using __percpu_counter_add() directly, instead of
percpu_counter_sub() and percpu_counter_add().
The batch size is increased to 130.000, based on the largest 64K
fragment memory usage. This does introduce some imprecise
memory accounting, but its does not need to be strict for this
use-case.
It is also essential, that the percpu_counter, does not
share cacheline with other writers, to make this scale.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This change is primarily a preparation to ease the extension of memory
limit tracking.
The change does reduce the number atomic operation, during freeing of
a frag queue. This does introduce a some performance improvement, as
these atomic operations are at the core of the performance problems
seen on NUMA systems.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Fragmentation code cacheline adjusting of struct inet_frag_queue.
Take advantage of the size of struct timer_list, and move all but
spinlock_t lock, below the timer struct. On 64-bit 'lru_list',
'list' and 'refcnt', fits exactly into the next cacheline, and a
new cacheline starts at 'fragments'.
The netns_frags *net pointer is moved to the end of the struct,
because its used in a compare, with "next/close-by" elements of
which this struct is embedded into.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The globally shared rwlock, of struct inet_frags, shares
cacheline with the 'rnd' number, which is used by the hash
calculations. Fix this, as this obviously is a bad idea, as
unnecessary cache-misses will occur when accessing the 'rnd'
number.
Also small note that, moving function ptr (*match) up in struct,
is to avoid it lands on the next cacheline (on 64-bit).
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
