Jeff Kirsher says:
====================
Intel Wired LAN Driver Updates 2015-05-28
This series contains updates to ethtool, ixgbe, i40e and i40evf.
John adds helper routines for ethtool to pass VF to rx_flow_spec. Since
the ring_cookie is 64 bits wide which is much larger than what could be
used for actual queue index values, provide helper routines to pack a VF
index into the cookie. Then John provides a ixgbe patch to allow flow
director to use the entire queue space.
Neerav provides a i40e patch to collect XOFF Rx stats, where it was not
being collected before.
Anjali provides ATR support for tunneled packets, as well as stats to
count tunnel ATR hits. Cleaned up PF struct members which are
unnecessary, since we can use the stat index macro directly. Cleaned
up flow director ATR/SB messages to a higher debug level since they
are not useful unless silicon validation is happening.
Greg provides a patch to disable offline diagnostics if VFs are enabled
since ethtool offline diagnostic tests are not designed (out of scope)
to disable VF functions for testing and re-enable afterward. Also cleans
up TODO comment that is no longer needed.
Vasu provides a fix an FCoE EOF case where i40e_fcoe_ctxt_eof() maybe
called before i40e_fcoe_eof_is_supported() is called.
Jesse adds skb->xmit_more support for i40evf. Then provides a performance
enhancement for i40evf by inlining some functions which provides a 15%
gain in small packet performance. Also cleans up the use of time_stamp
since it is no longer used to determine if there is a tx_hang and was
a part of a previous tx_hang design which is no longer used.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
As sock refcnt is taken when sock timer is started in
sk_reset_timer(), the sock refcnt should be put when sock timer
to be deleted is in pending state no matter what "probing_state"
value of tipc sock is.
Reviewed-by: Erik Hugne <erik.hugne@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes "vlaue" for "value" in include/linux/if_vlan.h.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
classic BPF already exposes skb->dev->ifindex via SKF_AD_IFINDEX extension.
Allow eBPF program to access it as well. Note that classic aborts execution
of the program if 'skb->dev == NULL' (which is inconvenient for program
writers), whereas eBPF returns zero in such case.
Also expose the 'skb_iif' field, since programs triggered by redirected
packet need to known the original interface index.
Summary:
__skb->ifindex -> skb->dev->ifindex
__skb->ingress_ifindex -> skb->skb_iif
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The current algorithm for deciding on the number of VRSS channels is
not optimal since we open up the min of number of CPUs online and the
number of VRSS channels the host is offering. So on a 32 VCPU guest
we could potentially open 32 VRSS subchannels. Experimentation has
shown that it is best to limit the number of VRSS channels to the number
of CPUs within a NUMA node.
Here is the new algorithm for deciding on the number of sub-channels we
would open up:
1) Pick the minimum of what the host is offering and what the driver
in the guest is specifying as the default value.
2) Pick the minimum of (1) and the numbers of CPUs in the NUMA
node the primary channel is bound to.
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is similar to b1cb59cf2efe(net: sysctl_net_core: check SNDBUF
and RCVBUF for min length). I don't think too small values can cause
crashes in the case of udp and tcp, but I've seen this set to too
small values which triggered awful performance. It also makes the
setting consistent across all the wmem/rmem sysctls.
Signed-off-by: Sorin Dumitru <sdumitru@ixiacom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Mathieu Olivari says:
====================
Add support for QCA IPQ806x Ethernet GMAC controller
This patch set adds support for the integrated Ethernet GMAC controller
on QCA IPQ806x SoC. This controller is based on a Gigabit Synopsys
DesignWare IP, already supported in the stmmac driver located in
drivers/net/ethernet/stmicro/stmmac.
This change is done as a follow-up to the following thread:
*http://www.spinics.net/lists/netdev/msg311265.html
While previous attempt was creating a new driver to drive this controller,
this new post leverages the existing stmmac driver by implementing the
SoC specific glue to it.
Aside from the pure stmmac glue layer, we have a couple of related
patches:
*IPQ806x NSS clock addition is cherry-picked and refreshed from the
following thread: https://lkml.org/lkml/2014/8/6/390
*phy-handle and fixed-link support are also added in this change set so the
driver can be fully functional on platforms using device-trees as well as
ethernet switches.
V2:
*Fix MODULE_LICENSE to "Dual BSD/GPL" as the dwmac-ipq806x.c is using
ISC license.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the device tree bindings documentation for the QCA IPQ806x
variant of the Synopsys DesignWare MAC.
Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The ethernet controller available in IPQ806x is a Synopsys DesignWare
Gigabit MAC IP core, already supported by the stmmac driver.
This glue layer implements some platform specific settings required to
get the controller working on an IPQ806x based platform.
Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case DT is used, this change adds the ability to the stmmac driver to
detect a fixed-link PHY, instanciate it, and use it during
phy_connect().
Fixed link PHYs DT usage is described in:
Documentation/devicetree/bindings/net/fixed-link.txt
Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
On stmmac driver, PHY specification in device-tree was done using the
non-standard property "snps,phy-addr". Specifying a PHY on a different
MDIO bus that the one within the stmmac controller doesn't seem to be
possible when device-tree is used.
This change adds support for the phy-handle property, as specified in
Documentation/devicetree/bindings/net/ethernet.txt.
Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the NSS/GMAC clocks and the TCM clock and NSS resets.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Mathieu Olivari <mathieu@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Steffen Klassert says:
====================
pull request (net-next): ipsec-next 2015-05-28
1) Remove xfrm_queue_purge as this is the same as skb_queue_purge.
2) Optimize policy and state walk.
3) Use a sane return code if afinfo registration fails.
4) Only check fori a acquire state if the state is not valid.
5) Remove a unnecessary NULL check before xfrm_pol_hold
as it checks the input for NULL.
6) Return directly if the xfrm hold queue is empty, avoid
to take a lock as it is nothing to do in this case.
7) Optimize the inexact policy search and allow for matching
of policies with priority ~0U.
All from Li RongQing.
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace confusing QL_STATUS_INVALID_PARAM == -1 == -EPERM with -EINVAL
and QLC_STATUS_UNSUPPORTED_CMD == -2 == -ENOENT with -EOPNOTSUPP, the
latter error code is arguable, but it is already used in the driver,
so let it be here as well.
Also remove always false (!buf) check on read(), the driver should
not care if userspace gets its EFAULT or not.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Acked-by: Rajesh Borundia <rajesh.borundia@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All of these files were only building on non-x86 because of
the indirect of inclusion of vmalloc.h by, of all things,
"net/inet_hashtables.h"
None of this got caught during build testing, because on x86
there is an implicit vmalloc.h include via on of the arch asm/
headers.
This fixes all of these
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Bump.
Change-ID: I54ec2787a9fead5e18447078f26e5dd27f01da44
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The driver doesn't use the time_stamp member to determine if there is a
tx_hang any more. There really isn't any point to the variable at all
so just remove it. It was left over from a previous tx_hang design.
Change-ID: I4c814827e1bcb46e45118fe37acdcfa814fb62a0
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Inlining these functions gives us about 15% more 64 byte packets per
second when using pktgen. 13.3 million to 15 million with a single
queue.
Also fix the function names in i40evf to i40evf not i40e while we are
touching the function header.
Change-ID: I3294ae9b085cf438672b6db5f9af122490ead9d0
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Eric added support for skb->xmit_more in i40e, this ports that into
i40evf as well.
Support skb->xmit_more in i40evf is straightforward; we need to move
around i40e_maybe_stop_tx() call to correctly test netif_xmit_stopped()
before taking the decision to not kick the NIC.
Change-ID: Idddda6a2e4a7ab335631c91ced51f55b25eb8468
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
These are not useful unless SV is happening as there is a FD flush counter
that tracks this.
Change-ID: If2655b5a29687247d03a51d35f69854bbeb711ce
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Because i40e_fcoe_ctxt_eof should never be called without
i40e_fcoe_eof_is_supported being called first, the EOF in fcoe_ctxt_eof
should always be valid and therefore we do not need to print an error
if it is not valid.
However, a WARN ON to easily catch any calls to i40e_fcoe_ctxt_eof that
aren't preceded with a call to i40e_fcoe_eof_is_supported is helpful.
Change-ID: I3b536b1981ec0bce80576a74440b7dea3908bdb9
Signed-off-by: Vasu Dev <vasu.dev@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
There's no need for a counter so remove the TODO comment.
Change-ID: I3321dda04934c4f5fda9b279ab666192bda44214
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
We can use the stat index macro directly, a variable is not required.
Change-ID: I19f08ac16353dc0cd87a1a8248d714e15a54aa8a
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Add a 3rd dynamic filter counter to track Tunneled ATR hits separately.
Ethtool port stat "fdir_atr_tunnel_match"
Change-ID: Idd978b6db2a462b5722397cd2ffd04ef055f8655
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Without this, RSS would have done inner header load balancing. Now we can
get the benefits of ATR for tunneled packets to better align TX and RX
queues with the right core/interrupt.
Change-ID: I07d0e0a192faf28fdd33b2f04c32b2a82ff97ddd
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Require the user to disable virtual functions before running the device
offline diagnostics. The offline diagnostics are intended to ensure
basic operation of the device - it is beyond the scope of the diagnostic
test to handle the additional complexity of bringing all the virtual
functions offline and then back online for each test run.
Change-ID: Ic0b854851a09fc85df0c9e82c220e45885457c30
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
When PFC is enabled for any UP in single TC configuration the driver didn't
collect the PFC XOFF RX stats. Though a single TC with PFC enabled is not a
common scenario do not prevent the driver from collecting stats if firmware
indicates that PFC is enabled.
Change-ID: Ie20bd58b07608b528f3c6d95894c9ae56b00077a
Signed-off-by: Neerav Parikh <neerav.parikh@intel.com>
Tested-by: Jim Young <james.m.young@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Flow director is exported to user space using the ethtool ntuple
support. However, currently it only supports steering traffic to a
subset of the queues in use by the hardware. This change allows
flow director to specify queues that have been assigned to virtual
functions by partitioning the ring_cookie into a 8bit VF specifier
followed by 32bit queue index. At the moment we don't have any
ethernet drivers with more than 2^32 queues on a single function
as best I can tell and nor do I expect this to happen anytime
soon. This way the ring_cookie's normal use for specifying a queue
on a specific PCI function continues to work as expected.
CC: Alex Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The ring_cookie is 64 bits wide which is much larger than can be used
for actual queue index values. So provide some helper routines to
pack a VF index into the cookie. This is useful to steer packets to
a VF ring without having to know the queue layout of the device.
CC: Alex Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
After commit 07f4c90062 ("tcp/dccp: try to not exhaust
ip_local_port_range in connect()") it is advised to have an even number
of ports described in /proc/sys/net/ipv4/ip_local_port_range
This means start/end values should have a different parity.
Let's warn sysadmins of this, so that they can update their settings
if they want to.
Suggested-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
__inet_hash_connect() does not use its third argument (port_offset)
if socket was already bound to a source port.
No need to perform useless but expensive md5 computations.
Reported-by: Crestez Dan Leonard <cdleonard@gmail.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Hariprasad Shenai says:
====================
cxgb4/cxgb4vf: Adds FL starvation support and cleanup
This patch series adds the following.
Adds debugfs entry to inject freelist starvation and some function and
argument cleanup
This patch series has been created against net-next tree and includes
patches on cxgb4 and cxgb4vf driver.
We have included all the maintainers of respective drivers. Kindly review
the change and let us know in case of any review comments.
Thanks
V2:
Skipping patch "cxgb4: Add support for loopback between VI of same port".
This needs some major code change, since module param is not recommended.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch changes variable name 'fn' to 'pf' of structure adapter.
A 'fn' usually stands for PCI function which could be a PF or a VF.
However, the use of this particular variable is explicitly limited to PF
only. So, be specific about it in the variable name.
Also corrects arguments passed for fn t4_ofld_eq_free, t4_ctrl_eq_free,
t4_eth_eq_free, t4_iq_free, t4_alloc_vi, t4_fw_hello, t4_wr_mbox and
t4_cfg_pfvf function.
Also renames cxgb4_t4_bar2_sge_qregs to t4_bar2_sge_qregs and renames
the latter function name in cxgb4vf driver to t4vf_bar2_sge_qregs to
avoid conflicts. Also fixes alignment for these function.
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add debugfs entry to inject Freelist starvation, used only for debugging
purpose.
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
vmalloc.h used to be included from include/net/inet_hashtables.h
but it is no longer the case.
Fixes: 095dc8e0c3 ("tcp: fix/cleanup inet_ehash_locks_alloc()")
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Eric Dumazet <edumzet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Aleksey Makarov says:
====================
Adding support for Cavium ThunderX network controller
This patchset adds support for the Cavium ThunderX network controller.
changes in v6:
* unused preprocessor symbols were removed
* reduce no of atomic operations in SQ maintenance
* support for TCP segmentation at driver level
* reset RBDR if fifo state is FAIL
* fixed an issue with link state mailbox message
changes in v5:
* __packed were removed. now we rely on C language ABI
* nic_dbg() -> netdev_dbg()
* fixes for a typo, constant spelling and using BIT_ULL
* use print_hex_dump()
* unnecessary conditions in a long if() chain were removed
changes in v4:
* the patch "pci: Add Cavium PCI vendor id" was attributed correctly
* a note that Cavium id is used in many drivers was added
* the license comments now match MODULE_LICENSE
* a comment explaining usage of writeq_relaxed()/readq_relaxed() was added
changes in v3:
* code cleanup
* issues discovered by reviewers were addressed
changes in v2:
* non-generic module parameters removed
* ethtool support added (nicvf_set_rxnfc())
v5: https://lkml.kernel.org/g/<1432344498-17131-1-git-send-email-aleksey.makarov@caviumnetworks.com>
v4: https://lkml.kernel.org/g/<1432000757-28700-1-git-send-email-aleksey.makarov@auriga.com>
v3: https://lkml.kernel.org/g/<1431747401-20847-1-git-send-email-aleksey.makarov@auriga.com>
v2: https://lkml.kernel.org/g/<1415596445-10061-1-git-send-email-rric@kernel.org>
v1: https://lkml.kernel.org/g/<20141030165434.GW20170@rric.localhost>
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds support for the Cavium ThunderX network controller.
The driver is on the pci bus and thus requires the Thunder PCIe host
controller driver to be enabled.
Signed-off-by: Maciej Czekaj <mjc@semihalf.com>
Signed-off-by: David Daney <david.daney@cavium.com>
Signed-off-by: Sunil Goutham <sgoutham@cavium.com>
Signed-off-by: Ganapatrao Kulkarni <ganapatrao.kulkarni@caviumnetworks.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: Tomasz Nowicki <tomasz.nowicki@linaro.org>
Signed-off-by: Robert Richter <rrichter@cavium.com>
Signed-off-by: Kamil Rytarowski <kamil@semihalf.com>
Signed-off-by: Thanneeru Srinivasulu <tsrinivasulu@caviumnetworks.com>
Signed-off-by: Sruthi Vangala <svangala@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This vendor id will be used by network (vNIC), USB (xHCI),
SATA (AHCI), GPIO, I2C, MMC and maybe other drivers
for ThunderX SoC.
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sunil Goutham <sgoutham@cavium.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While 3b52960266 ("test_bpf: add more eBPF jump torture cases")
added the int3 bug test case only for eBPF, which needs exactly 11
passes to converge, here's a version for classic BPF with 11 passes,
and one that would need 70 passes on x86_64 to actually converge for
being successfully JITed. Effectively, all jumps are being optimized
out resulting in a JIT image of just 89 bytes (from originally max
BPF insns), only returning K.
Might be useful as a receipe for folks wanting to craft a test case
when backporting the fix in commit 3f7352bf21 ("x86: bpf_jit: fix
compilation of large bpf programs") while not having eBPF. The 2nd
one is delegated to the interpreter as the last pass still results
in shrinking, in other words, this one won't be JITed on x86_64.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Edward Cree says:
====================
sfc: add MCDI tracing
This patchset adds support for logging MCDI (Management-Controller-to-
Driver Interface) interactions between the sfc driver and a bound device,
to aid in debugging.
Solarflare has a tool to decode the resulting traces and will look to
open-source this if there is any external interest, but the protocol is
already detailed in drivers/net/ethernet/sfc/mcdi_pcol.h.
The logging buffer we allocate per MCDI context is a work area for
constructing each individual message before logging it with netif_info.
The reason the buffer is long-lived is simply to avoid the overhead of
allocating and freeing it every MCDI call, since MCDIs are already known
to be serialised for other reasons.
--
v4: remove patch #4, which has already been applied via sshah
v3: add some explanations to cover letter and patch #4
v2: avoid long lines in cover letter; fix multiline comment style
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
As many issues are encountered at probe time, where MCDI logging can't be
enabled through the sysfs node, this change adds a module parameter
'mcdi_logging_default', which defaults to false. When set to true, newly-
probed functions will have MCDI logging enabled. The setting can
subsequently be changed as normal through the sysfs node.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
MCDI tracing is enabled per-function with a sysfs file
/sys/class/net/<NET_DEV>/device/mcdi_logging
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
MCDI tracing is conditional on CONFIG_SFC_MCDI_LOGGING, which is enabled
by default.
Each MCDI command will produce a console line like
sfc dom🚌dev:fn ifname: MCDI RPC REQ: xxxxxxxx [yyyyyyyy...]
where xxxxxxxx etc. are the raw MCDI payload in 32-bit hex chunks.
The response will then produce a similar line with "RESP" instead of "REQ",
and containing the MCDI response payload (if any).
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We're seeing some softlockups from this function when there
are a lot fdb entries on a vxlan device. Taking the lock for
each bucket instead of the whole table is enough to fix that.
Signed-off-by: Sorin Dumitru <sdumitru@ixiacom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A long standing problem on busy servers is the tiny available TCP port
range (/proc/sys/net/ipv4/ip_local_port_range) and the default
sequential allocation of source ports in connect() system call.
If a host is having a lot of active TCP sessions, chances are
very high that all ports are in use by at least one flow,
and subsequent bind(0) attempts fail, or have to scan a big portion of
space to find a slot.
In this patch, I changed the starting point in __inet_hash_connect()
so that we try to favor even [1] ports, leaving odd ports for bind()
users.
We still perform a sequential search, so there is no guarantee, but
if connect() targets are very different, end result is we leave
more ports available to bind(), and we spread them all over the range,
lowering time for both connect() and bind() to find a slot.
This strategy only works well if /proc/sys/net/ipv4/ip_local_port_range
is even, ie if start/end values have different parity.
Therefore, default /proc/sys/net/ipv4/ip_local_port_range was changed to
32768 - 60999 (instead of 32768 - 61000)
There is no change on security aspects here, only some poor hashing
schemes could be eventually impacted by this change.
[1] : The odd/even property depends on ip_local_port_range values parity
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Florian Westphal says:
====================
net: force refragmentation for DF reassembed skbs
output path tests:
if (skb->len > mtu) ip_fragment()
This breaks connectivity in one corner case:
If the skb was reassembled, but has the DF bit set and ..
.. its reassembled size is <= outdev mtu ..
.. we will forward a DF packet larger than what the sender
transmitted on wire.
If a router later in the path can't forward this packet, it will send an
icmp error in response to an mtu that the original sender never exceeded.
This changes ipv4 defrag/output path to
a) force refragmentation for DF reassembled skbs and
b) set DF bit on all fragments when refragmenting if it was set on original
frags.
tested via:
from scapy.all import *
dip="10.23.42.2"
payload="A"*1400
packet=IP(dst=dip,id=12345,flags='DF')/UDP(sport=42,dport=42)/payload
frags=fragment(packet,fragsize=1200)
for fragment in frags:
send(fragment)
Without this patch, we generate fragments without df bit set based
on the outgoing device mtu when fragmenting after forwarding, ie.
IP (ttl 64, id 12345, offset 0, flags [+, DF], proto UDP (17), length 1204)
192.168.7.1.42 > 10.23.42.2.42: UDP, length 1400
IP (ttl 64, id 12345, offset 1184, flags [DF], proto UDP (17), length 244)
192.168.7.1 > 10.23.42.2: ip-proto-17
on ingress will either turn into
IP (ttl 63, id 12345, offset 0, flags [+], proto UDP (17), length 1396)
192.168.7.1.42 > 10.23.42.2.42: UDP, length 1400
IP (ttl 63, id 12345, offset 1376, flags [none], proto UDP (17), length 52)
(mtu 1400: We strip df and send larger fragment), or
IP (ttl 63, id 12345, offset 0, flags [DF], proto UDP (17), length 1428)
192.168.7.1.42 > 10.23.42.2.42: [udp sum ok] UDP, length 1400
if mtu is 1500. And in this case things break; router with a smaller mtu
will send icmp error, but original sender only sent packets <= 1204 byte.
With patch, we keep intent of such fragments and will emit DF-fragments
that won't exceed 1204 byte in size.
Joint work with Hannes Frederic Sowa.
Changes since v2:
- split unrelated patches from series
- rework changelog of patch #2 to better illustrate breakage
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
We currently always send fragments without DF bit set.
Thus, given following setup:
mtu1500 - mtu1500:1400 - mtu1400:1280 - mtu1280
A R1 R2 B
Where R1 and R2 run linux with netfilter defragmentation/conntrack
enabled, then if Host A sent a fragmented packet _with_ DF set to B, R1
will respond with icmp too big error if one of these fragments exceeded
1400 bytes.
However, if R1 receives fragment sizes 1200 and 100, it would
forward the reassembled packet without refragmenting, i.e.
R2 will send an icmp error in response to a packet that was never sent,
citing mtu that the original sender never exceeded.
The other minor issue is that a refragmentation on R1 will conceal the
MTU of R2-B since refragmentation does not set DF bit on the fragments.
This modifies ip_fragment so that we track largest fragment size seen
both for DF and non-DF packets, and set frag_max_size to the largest
value.
If the DF fragment size is larger or equal to the non-df one, we will
consider the packet a path mtu probe:
We set DF bit on the reassembled skb and also tag it with a new IPCB flag
to force refragmentation even if skb fits outdev mtu.
We will also set DF bit on each fragment in this case.
Joint work with Hannes Frederic Sowa.
Reported-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
ip_skb_dst_mtu is small inline helper, but its called in several places.
before: 17061 44 0 17105 42d1 net/ipv4/ip_output.o
after: 16805 44 0 16849 41d1 net/ipv4/ip_output.o
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Florian Fainelli says:
====================
net: phy: phy_interface_is_rgmii helper
As you suggested, here is the helper function to avoid missing some RGMII
interface checks. Had to wait for net to be merged in net-next to avoid
submitting the same patch/commit.
Dan, you might want to rebase your dp83867 submission to use that helper
when you this patchset gets merged into net-next, thanks!
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Update all open-coded tests for all 4 PHY_INTERFACE_MODE_RGMII* values
to use the newly introduced helper: phy_interface_is_rgmii.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
