Skip to content

C
clone-Linux

gsplhtlxg / clone-Linux

通知 2
Star 0
Fork 1
C
clone-Linux

体验新版 GitCode，发现更多精彩内容 >>

切换分支/标签
  1. 14 1月, 2015 24 次提交
  3. 13 1月, 2015 16 次提交
    • D
      Merge branch 'tuntap_queues' · d2c60b13
      由 David S. Miller 提交于 
      Pankaj Gupta says:

====================
Increase the limit of tuntap queues

Networking under KVM works best if we allocate a per-vCPU rx and tx
queue in a virtual NIC. This requires a per-vCPU queue on the host side.
Modern physical NICs have multiqueue support for large number of queues.
To scale vNIC to run multiple queues parallel to maximum number of vCPU's
we need to increase number of queues support in tuntap.

Changes from v4:
PATCH2: Michael.S.Tsirkin - Updated change comment message.

Changes from v3:
PATCH1: Michael.S.Tsirkin - Some cleanups and updated commit message.
                            Perf numbers on 10 Gbs NIC
Changes from v2:
PATCH 3: David Miller     - flex array adds extra level of indirection
                            for preallocated array.(dropped, as flow array
			    is allocated using kzalloc with failover to zalloc).
Changes from v1:
PATCH 2: David Miller     - sysctl changes to limit number of queues
                            not required for unprivileged users(dropped).

Changes from RFC
PATCH 1: Sergei Shtylyov  - Add an empty line after declarations.
PATCH 2: Jiri Pirko -       Do not introduce new module paramaters.
	 Michael.S.Tsirkin- We can use sysctl for limiting max number
                            of queues.

This series is to increase the number of tuntap queues. Original work is being
done by 'jasowang@redhat.com'. I am taking this 'https://lkml.org/lkml/2013/6/19/29'
patch series as a reference. As per discussion in the patch series:

There were two reasons which prevented us from increasing number of tun queues:

- The netdev_queue array in netdevice were allocated through kmalloc, which may
  cause a high order memory allocation too when we have several queues.
  E.g. sizeof(netdev_queue) is 320, which means a high order allocation would
  happens when the device has more than 16 queues.

- We store the hash buckets in tun_struct which results a very large size of
  tun_struct, this high order memory allocation fail easily when the memory is
  fragmented.

The patch 60877a32 increases the number of tx
queues. Memory allocation fallback to vzalloc() when kmalloc() fails.

This series tries to address following issues:

- Increase the number of netdev_queue queues for rx similarly its done for tx
  queues by falling back to vzalloc() when memory allocation with kmalloc() fails.

- Increase number of queues to 256, maximum number is equal to maximum number
  of vCPUS allowed in a guest.

I have also done testing with multiple parallel Netperf sessions for different
combination of queues and CPU's. It seems to be working fine without much increase
in cpu load with increase in number of queues. I also see good increase in throughput
with increase in number of queues. Though i had limitation of 8 physical CPU's.

For this test: Two Hosts(Host1 & Host2) are directly connected with cable
Host1 is running Guest1. Data is sent from Host2 to Guest1 via Host1.

Host kernel: 3.19.0-rc2+, AMD Opteron(tm) Processor 6320
NIC : Emulex Corporation OneConnect 10Gb NIC (be3)

Patch Applied  %usr   %nice    %sys %iowait    %irq   %soft  %steal  %guest  %gnice   %idle  throughput
Single Queue, 2 vCPU's
-------------
Before Patch :all    0.19    0.00    0.16    0.07    0.04    0.10    0.00    0.18    0.00   99.26  57864.18
After  Patch :all    0.99    0.00    0.64    0.69    0.07    0.26    0.00    1.58    0.00   95.77  57735.77

With 2 Queues, 2 vCPU's
---------------
Before Patch :all    0.19    0.00    0.19    0.10    0.04    0.11    0.00    0.28    0.00   99.08  63083.09
After  Patch :all    0.87    0.00    0.73    0.78    0.09    0.35    0.00    2.04    0.00   95.14  62917.03

With 4 Queues, 4 vCPU's
--------------
Before Patch :all    0.20    0.00    0.21    0.11    0.04    0.12    0.00    0.32    0.00   99.00  80865.06
After  Patch :all    0.71    0.00    0.93    0.85    0.11    0.51    0.00    2.62    0.00   94.27  86463.19

With 8 Queues, 8 vCPU's
--------------
Before Patch :all    0.19    0.00    0.18    0.09    0.04    0.11    0.00    0.23    0.00   99.17  86795.31
After  Patch :all    0.65    0.00    1.18    0.93    0.13    0.68    0.00    3.38    0.00   93.05  89459.93

With 16 Queues, 8 vCPU's
--------------
After  Patch :all    0.61    0.00    1.59    0.97    0.18    0.92    0.00    4.32    0.00   91.41  120951.60
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      d2c60b13
    • P
      tuntap: Increase the number of queues in tun. · baf71c5c
      由 Pankaj Gupta 提交于 
      Networking under kvm works best if we allocate a per-vCPU RX and TX
queue in a virtual NIC. This requires a per-vCPU queue on the host side.

It is now safe to increase the maximum number of queues.
Preceding patch: 'net: allow large number of rx queues'
made sure this won't cause failures due to high order memory
allocations. Increase it to 256: this is the max number of vCPUs
KVM supports.

Size of tun_struct changes from 8512 to 10496 after this patch. This keeps
pages allocated for tun_struct before and after the patch to 3.
Signed-off-by: NPankaj Gupta <pagupta@redhat.com>
Reviewed-by: NDavid Gibson <dgibson@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      baf71c5c
    • P
      net: allow large number of rx queues · 10595902
      由 Pankaj Gupta 提交于 
      netif_alloc_rx_queues() uses kcalloc() to allocate memory
for "struct netdev_queue *_rx" array.
If we are doing large rx queue allocation kcalloc() might
fail, so this patch does a fallback to vzalloc().
Similar implementation is done for tx queue allocation in
netif_alloc_netdev_queues().

We avoid failure of high order memory allocation
with the help of vzalloc(), this allows us to do large
rx and tx queue allocation which in turn helps us to
increase the number of queues in tun.

As vmalloc() adds overhead on a critical network path,
__GFP_REPEAT flag is used with kzalloc() to do this fallback
only when really needed.
Signed-off-by: NPankaj Gupta <pagupta@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NDavid Gibson <dgibson@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      10595902
    • K
      team: Remove dead code · e350a96e
      由 Kenneth Williams 提交于 
      The deleted lines are called from a function which is called:
1) Only through __team_options_register via team_options_register and
2) Only during initialization / mode initialization when there are no
ports attached.
Therefore the ports list is guarenteed to be empty and this code will
never be executed.
Signed-off-by: NKenneth Williams <ken@williamsclan.us>
Acked-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      e350a96e
    • D
      net: bnx2x: avoid macro redefinition · b1e8bc61
      由 David Decotigny 提交于 
      Signed-off-by: NDavid Decotigny <decot@googlers.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      b1e8bc61
    • R
      net: sched: sch_teql: Remove unused function · ddcde70c
      由 Rickard Strandqvist 提交于 
      Remove the function teql_neigh_release() that is not used anywhere.

This was partially found by using a static code analysis program called cppcheck.
Signed-off-by: NRickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      ddcde70c
    • R
      net: xfrm: xfrm_algo: Remove unused function · 83400b99
      由 Rickard Strandqvist 提交于 
      Remove the function aead_entries() that is not used anywhere.

This was partially found by using a static code analysis program called cppcheck.
Signed-off-by: NRickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      83400b99
    • D
      Merge branch 'bridge_vlan_ranges' · d0d2cc53
      由 David S. Miller 提交于 
      Roopa Prabhu says:

====================
bridge: support for vlan range in setlink/dellink

This series adds new flags in IFLA_BRIDGE_VLAN_INFO to indicate
vlan range.

Will post corresponding iproute2 patches if these get accepted.

v1-> v2
    - changed patches to use a nested list attribute
    IFLA_BRIDGE_VLAN_INFO_LIST as suggested by scott feldman
    - dropped notification changes from the series. Will post them
    separately after this range message is accepted.

v2 -> v3
    - incorporated some review feedback
    - include patches to fill vlan ranges during getlink
    - Dropped IFLA_BRIDGE_VLAN_INFO_LIST. I think it may get
    confusing to userspace if we introduce yet another way to
    send lists. With getlink already sending nested
    IFLA_BRIDGE_VLAN_INFO in IFLA_AF_SPEC, It seems better to
    use the existing format for lists and just use the flags from v2
    to mark vlan ranges
====================
Signed-off-by: NRoopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: NWilson Kok <wkok@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      d0d2cc53
    • R
      bridge: new function to pack vlans into ranges during gets · 36cd0ffb
      由 Roopa Prabhu 提交于 
      This patch adds new function to pack vlans into ranges
whereever applicable using the flags BRIDGE_VLAN_INFO_RANGE_BEGIN
and BRIDGE VLAN_INFO_RANGE_END

Old vlan packing code is moved to a new function and continues to be
called when filter_mask is RTEXT_FILTER_BRVLAN.
Signed-off-by: NRoopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      36cd0ffb
    • R
      rtnetlink: new filter RTEXT_FILTER_BRVLAN_COMPRESSED · 35a27cee
      由 Roopa Prabhu 提交于 
      This filter is same as RTEXT_FILTER_BRVLAN except that it tries
to compress the consecutive vlans into ranges.

This helps on systems with large number of configured vlans.
Signed-off-by: NRoopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      35a27cee
    • R
      bridge: support for multiple vlans and vlan ranges in setlink and dellink requests · bdced7ef
      由 Roopa Prabhu 提交于 
      This patch changes bridge IFLA_AF_SPEC netlink attribute parser to
look for more than one IFLA_BRIDGE_VLAN_INFO attribute. This allows
userspace to pack more than one vlan in the setlink msg.

The dumps were already sending more than one vlan info in the getlink msg.

This patch also adds bridge_vlan_info flags BRIDGE_VLAN_INFO_RANGE_BEGIN and
BRIDGE_VLAN_INFO_RANGE_END to indicate start and end of vlan range

This patch also deletes unused ifla_br_policy.
Signed-off-by: NRoopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      bdced7ef
    • V
      drivers: net: xen-netfront: remove residual dead code · dd2e8bf5
      由 Vincenzo Maffione 提交于 
      This patch removes some unused arrays from the netfront private
data structures. These arrays were used in "flip" receive mode.
Signed-off-by: NVincenzo Maffione <v.maffione@gmail.com>
Reviewed-by: NDavid Vrabel <david.vrabel@citrix.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      dd2e8bf5
    • S
      Driver: Vmxnet3: Reinitialize vmxnet3 backend on wakeup from hibernate · 5ec82c1e
      由 Shrikrishna Khare 提交于 
      Failing to reinitialize on wakeup results in loss of network connectivity for
vmxnet3 interface.
Signed-off-by: NSrividya Murali <smurali@vmware.com>
Signed-off-by: NShrikrishna Khare <skhare@vmware.com>
Reviewed-by: NShreyas N Bhatewara <sbhatewara@vmware.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      5ec82c1e
    • J
      bonding: cleanup bond_opts array · 7bfa0145
      由 Jonathan Toppins 提交于 
      Remove the empty array element initializer and size the array with
BOND_OPT_LAST so the compiler will complain if more elements are in
there than should be.

An interesting unwanted side effect of this initializer is that if one
inserts new options into the middle of the array then this initializer
will zero out the option that equals BOND_OPT_TLB_DYNAMIC_LB+1.

Example:
Extend the OPTS enum:
enum {
   ...
   BOND_OPT_TLB_DYNAMIC_LB,
   BOND_OPT_LACP_NEW1,
   BOND_OPT_LAST
};

Now insert into bond_opts array:
static const struct bond_option bond_opts[] = {
      ...
      [BOND_OPT_LACP_RATE] = { .... unchanged stuff .... },
      [BOND_OPT_LACP_NEW1] = { ... new stuff ... },
      ...
      [BOND_OPT_TLB_DYNAMIC_LB] = { .... unchanged stuff ....},
      { } // MARK A
};

Since BOND_OPT_LACP_NEW1 = BOND_OPT_TLB_DYNAMIC_LB+1, the last
initializer (MARK A) will overwrite the contents of BOND_OPT_LACP_NEW1
and can be easily viewed with the crash utility.
Signed-off-by: NJonathan Toppins <jtoppins@cumulusnetworks.com>
Cc: Andy Gospodarek <gospo@cumulusnetworks.com>
Cc: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: NAndy Gospodarek <gospo@cumulusnetworks.com>
Acked-by: NNikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      7bfa0145
    • D
      Merge branch 'tipc-namespaces' · d9fbfb94
      由 David S. Miller 提交于 
      Ying Xue says:

====================
tipc: make tipc support namespace

This patchset aims to add net namespace support for TIPC stack.

Currently TIPC module declares the following global resources:
- TIPC network idenfication number
- TIPC node table
- TIPC bearer list table
- TIPC broadcast link
- TIPC socket reference table
- TIPC name service table
- TIPC node address
- TIPC service subscriber server
- TIPC random value
- TIPC netlink

In order that TIPC is aware of namespace, above each resource must be
allocated, initialized and destroyed inside per namespace. Therefore,
the major works of this patchset are to isolate these global resources
and make them private for each namespace. However, before these changes
come true, some necessary preparation works must be first done: convert
socket reference table with generic rhashtable, cleanup core.c and
core.h files, remove unnecessary wrapper functions for kernel timer
interfaces and so on.

It should be noted that commit ##1 ("tipc: fix bug in broadcast
retransmit code") was already submitted to 'net' tree, so please see
below link:

http://patchwork.ozlabs.org/patch/426717/

Since it is prerequisite for the rest of the series to apply, I
prepend them to the series.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      d9fbfb94
    • Y
      tipc: make netlink support net namespace · d49e2041
      由 Ying Xue 提交于 
      Currently tipc module only allows users sitting on "init_net" namespace
to configure it through netlink interface. But now almost each tipc
component is able to be aware of net namespace, so it's time to open
the permission for users residing in other namespaces, allowing them
to configure their own tipc stack instance through netlink interface.
Signed-off-by: NYing Xue <ying.xue@windriver.com>
Tested-by: NTero Aho <Tero.Aho@coriant.com>
Reviewed-by: NJon Maloy <jon.maloy@ericsson.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      d49e2041