Andrei Vagin [Wed, 7 Apr 2021 06:40:51 +0000 (23:40 -0700)]
net: remove the new_ifindex argument from dev_change_net_namespace
Here is only one place where we want to specify new_ifindex. In all
other cases, callers pass 0 as new_ifindex. It looks reasonable to add a
low-level function with new_ifindex and to convert
dev_change_net_namespace to a static inline wrapper.
Fixes:
eeb85a14ee34 ("net: Allow to specify ifindex when device is moved to another namespace")
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Andrei Vagin <avagin@gmail.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrei Vagin [Wed, 7 Apr 2021 06:40:03 +0000 (23:40 -0700)]
net: introduce nla_policy for IFLA_NEW_IFINDEX
In this case, we don't need to check that new_ifindex is positive in
validate_linkmsg.
Fixes:
eeb85a14ee34 ("net: Allow to specify ifindex when device is moved to another namespace")
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Andrei Vagin <avagin@gmail.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 7 Apr 2021 21:38:24 +0000 (14:38 -0700)]
Merge tag 'mlx5-updates-2021-04-06' of git://git./linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5-updates-2021-04-06
Introduce TC sample offload
Background
----------
The tc sample action allows user to sample traffic matched by tc
classifier. The sampling consists of choosing packets randomly and
sampling them using psample module.
The tc sample parameters include group id, sampling rate and packet's
truncation (to save kernel-user traffic).
Sample in TC SW
---------------
User must specify rate and group id for sample action, truncate is
optional.
tc filter add dev enp4s0f0_0 ingress protocol ip prio 1 flower \
src_mac 02:25:d0:14:01:02 dst_mac 02:25:d0:14:01:03 \
action sample rate 10 group 5 trunc 60 \
action mirred egress redirect dev enp4s0f0_1
The tc sample action kernel module 'act_sample' will call another
kernel module 'psample' to send sampled packets to userspace.
MLX5 sample HW offload - MLX5 driver patches
--------------------------------------------
The sample action is translated to a goto flow table object
destination which samples packets according to the provided
sample ratio. Sampled packets are duplicated. One copy is
processed by a termination table, named the sample table,
which sends the packet to the eswitch manager port (that will
be processed by software).
The second copy is processed by the default table which executes
the subsequent actions. The default table is created per <vport,
chain, prio> tuple as rules with different prios and chains may
overlap.
For example, for the following typical flow table:
+-------------------------------+
+ original flow table +
+-------------------------------+
+ original match +
+-------------------------------+
+ sample action + other actions +
+-------------------------------+
We translate the tc filter with sample action to the following HW model:
+---------------------+
+ original flow table +
+---------------------+
+ original match +
+---------------------+
|
v
+------------------------------------------------+
+ Flow Sampler Object +
+------------------------------------------------+
+ sample ratio +
+------------------------------------------------+
+ sample table id | default table id +
+------------------------------------------------+
| |
v v
+-----------------------------+ +----------------------------------------+
+ sample table + + default table per <vport, chain, prio> +
+-----------------------------+ +----------------------------------------+
+ forward to management vport + + original match +
+-----------------------------+ +----------------------------------------+
+ other actions +
+----------------------------------------+
Flow sampler object
-------------------
Hardware introduces flow sampler object to do sample. It is a new
destination type. Driver needs to specify two flow table ids in it.
One is sample table id. The other one is the default table id.
Sample table samples the packets according to the sample rate and
forward the sampled packets to eswitch manager port. Default table
finishes the subsequent actions.
Group id and reg_c0
-------------------
Userspace program will take different actions for sampled packets
according to tc sample action group id. So hardware must pass group
id to software for each sampled packets. In Paul Blakey's "Introduce
connection tracking offload" patch set, reg_c0 lower 16 bits are used
for miss packet chain id restore. We convert reg_c0 lower 16 bits to
a common object pool, so other features can also use it.
Since sample group id is 32 bits, create a 16 bits object id to map
the group id and write the object id to reg_c0 lower 16 bits. reg_c0
can only be used for matching. Write reg_c0 to flow_tag, so software
can get the object id via flow_tag and find group id via the common
object pool.
Sampler restore handle
----------------------
Use common object pool to create an object id to map sample parameters.
Allocate a modify header action to write the object id to reg_c0 lower
16 bits. Create a restore rule to pass the object id to software. So
software can identify sampled packets via the object id and send it to
userspace.
Aggregate the modify header action, restore rule and object id to a
sample restore handle. Re-use identical sample restore handle for
the same object id.
Send sampled packets to userspace
---------------------------------
The destination for sampled packets is eswitch manager port, so
representors can receive sampled packets together with the group id.
Driver will send sampled packets and group id to userspace via psample.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
wengjianfeng [Wed, 7 Apr 2021 03:16:38 +0000 (11:16 +0800)]
nfc/fdp: remove unnecessary assignment and label
In function fdp_nci_patch_otp and fdp_nci_patch_ram,many goto
out statements are used, and out label just return variable r.
in some places,just jump to the out label, and in other places,
assign a value to the variable r,then jump to the out label.
It is unnecessary, we just use return sentences to replace goto
sentences and delete out label.
Signed-off-by: wengjianfeng <wengjianfeng@yulong.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vadim Pasternak [Tue, 6 Apr 2021 12:27:33 +0000 (15:27 +0300)]
mlxsw: core: Remove critical trip points from thermal zones
Disable software thermal protection by removing critical trip points
from all thermal zones.
The software thermal protection is redundant given there are two layers
of protection below it in firmware and hardware. The first layer is
performed by firmware, the second, in case firmware was not able to
perform protection, by hardware.
The temperature threshold set for hardware protection is always higher
than for firmware.
Signed-off-by: Vadim Pasternak <vadimp@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Voon Weifeng [Tue, 6 Apr 2021 01:32:50 +0000 (09:32 +0800)]
stmmac: intel: Enable SERDES PHY rx clk for PSE
EHL PSE SGMII mode requires to ungate the SERDES PHY rx clk for power up
sequence and vice versa.
Signed-off-by: Voon Weifeng <weifeng.voon@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski [Wed, 7 Apr 2021 00:23:59 +0000 (17:23 -0700)]
ethtool: document PHY tunable callbacks
Add missing kdoc for phy tunable callbacks.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Wed, 7 Apr 2021 21:09:40 +0000 (14:09 -0700)]
Merge branch 'mptcp-next'
Mat Martineau says:
====================
mptcp: Cleanup, a new test case, and header trimming
Some more patches to include from the MPTCP tree:
Patches 1-6 refactor an address-related data structure and reduce some
duplicate code that handles IPv4 and IPv6 addresses.
Patch 7 adds a test case for the MPTCP netlink interface, passing a
specific ifindex to the kernel.
Patch 8 drops extra header options from IPv4 address echo packets,
improving consistency and testability between IPv4 and IPv6.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Davide Caratti [Wed, 7 Apr 2021 00:16:04 +0000 (17:16 -0700)]
mptcp: drop all sub-options except ADD_ADDR when the echo bit is set
Current Linux carries echo-ed ADD_ADDR over pure TCP ACKs, so there is no
need to add a DSS element that would fit only ADD_ADDR with IPv4 address.
Drop the DSS from echo-ed ADD_ADDR, regardless of the IP version.
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geliang Tang [Wed, 7 Apr 2021 00:16:03 +0000 (17:16 -0700)]
selftests: mptcp: add the net device name testcase
This patch added a new testcase for setting the net device name. In it,
pass the net device name to pm_nl_ctl to set the ifindex field of struct
mptcp_pm_addr_entry.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geliang Tang [Wed, 7 Apr 2021 00:16:02 +0000 (17:16 -0700)]
mptcp: unify add_addr(6)_generate_hmac
The length of the IPv4 address is 4 octets and IPv6 is 16. That's the only
difference between add_addr_generate_hmac and add_addr6_generate_hmac.
This patch dropped the duplicate code and unify them into one.
Co-developed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geliang Tang [Wed, 7 Apr 2021 00:16:01 +0000 (17:16 -0700)]
mptcp: drop MPTCP_ADDR_IPVERSION_4/6
Since the type of the address family in struct mptcp_options_received
became sa_family_t, we should set AF_INET/AF_INET6 to it, instead of
using MPTCP_ADDR_IPVERSION_4/6.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geliang Tang [Wed, 7 Apr 2021 00:16:00 +0000 (17:16 -0700)]
mptcp: use mptcp_addr_info in mptcp_options_received
This patch added a new struct mptcp_addr_info member addr in struct
mptcp_options_received, and dropped the original family, addr_id, addr,
addr6 and port fields in it. Then we can pass the parameter mp_opt.addr
directly to mptcp_pm_add_addr_received and mptcp_pm_add_addr_echoed.
Since the port number became big-endian now, use htons to convert the
incoming port number to it. Also use ntohs to convert it when passing
it to add_addr_generate_hmac or printing it out.
Co-developed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geliang Tang [Wed, 7 Apr 2021 00:15:59 +0000 (17:15 -0700)]
mptcp: drop OPTION_MPTCP_ADD_ADDR6
Since the family field was added in struct mptcp_out_options, no need to
use OPTION_MPTCP_ADD_ADDR6 to identify the IPv6 address. Drop it.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geliang Tang [Wed, 7 Apr 2021 00:15:58 +0000 (17:15 -0700)]
mptcp: use mptcp_addr_info in mptcp_out_options
This patch moved the mptcp_addr_info struct from protocol.h to mptcp.h,
added a new struct mptcp_addr_info member addr in struct mptcp_out_options,
and dropped the original addr, addr6, addr_id and port fields in it. Then
we can use opts->addr to get the adding address from PM directly using
mptcp_pm_add_addr_signal.
Since the port number became big-endian now, use ntohs to convert it
before sending it out with the ADD_ADDR suboption. Also convert it
when passing it to add_addr_generate_hmac or printing it out.
Co-developed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geliang Tang [Wed, 7 Apr 2021 00:15:57 +0000 (17:15 -0700)]
mptcp: move flags and ifindex out of mptcp_addr_info
This patch moved the flags and ifindex fields from struct mptcp_addr_info
to struct mptcp_pm_addr_entry. Add the flags and ifindex values as two new
parameters to __mptcp_subflow_connect.
In mptcp_pm_create_subflow_or_signal_addr, pass the local address entry's
flags and ifindex fields to __mptcp_subflow_connect.
In mptcp_pm_nl_add_addr_received, just pass two zeros to it.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Chris Mi [Mon, 21 Sep 2020 08:45:07 +0000 (16:45 +0800)]
net/mlx5e: TC, Add support to offload sample action
The following diagram illustrates the hardware model for tc sample action:
+---------------------+
+ original flow table +
+---------------------+
+ original match +
+---------------------+
|
v
+------------------------------------------------+
+ Flow Sampler Object +
+------------------------------------------------+
+ sample ratio +
+------------------------------------------------+
+ sample table id | default table id +
+------------------------------------------------+
| |
v v
+-----------------------------+ +----------------------------------------+
+ sample table + + default table per <vport, chain, prio> +
+-----------------------------+ +----------------------------------------+
+ forward to management vport + + original match +
+-----------------------------+ +----------------------------------------+
+ other actions +
+----------------------------------------+
The sample action is translated to a goto flow table object
destination which samples packets according to the provided
sample ratio. Sampled packets are duplicated. One copy is
processed by a termination table, named the sample table,
which sends the packet to the eswitch manager port (that will
be processed by software).
The second copy is processed by the default table which executes
the subsequent actions. The default table is created per <vport,
chain, prio> tuple as rules with different prios and chains may
overlap.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Tue, 26 Jan 2021 03:15:46 +0000 (11:15 +0800)]
net/mlx5e: TC, Handle sampled packets
Mark the sampled packets with a sample restore object. Send sampled
packets using the psample api.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Fri, 15 Jan 2021 08:03:47 +0000 (16:03 +0800)]
net/mlx5e: TC, Refactor tc update skb function
As a pre-step to process sampled packet in this function.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Tue, 26 Jan 2021 03:08:28 +0000 (11:08 +0800)]
net/mlx5e: TC, Add sampler restore handle API
Use common object pool to create an object ID to map sample parameters.
Allocate a modify header action to write the object ID to reg_c0 lower
16 bits. Create a restore rule to pass the object ID to software. So
software can identify sampled packets via the object ID and send it to
userspace.
Aggregate the modify header action, restore rule and object ID to a
sample restore handle. Re-use identical sample restore handle for
the same object ID.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Mon, 21 Sep 2020 05:25:54 +0000 (13:25 +0800)]
net/mlx5e: TC, Add sampler object API
In order to offload sample action, HW introduces sampler object. The
sampler object samples packets according to the provided sample ratio.
Sampled packets are duplicated. One copy is processed by a termination
table, named the sample table, which sends the packet up to software.
The second copy is processed by the default table.
Instantiate sampler object. Re-use identical sampler object for
the same sample ratio, sample table and default table as a prestep for
offloading tc sample actions.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Fri, 18 Sep 2020 09:46:57 +0000 (17:46 +0800)]
net/mlx5e: TC, Add sampler termination table API
Sampled packets are sent to software using termination tables. There
is only one rule in that table that is to forward sampled packets to
the e-switch management vport.
Create a sampler termination table and rule for each eswitch.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Mon, 31 Aug 2020 05:28:35 +0000 (13:28 +0800)]
net/mlx5e: TC, Parse sample action
Parse TC sample action and save sample parameters in flow attribute
data structure.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Mon, 31 Aug 2020 05:27:53 +0000 (13:27 +0800)]
net/mlx5: Instantiate separate mapping objects for FDB and NIC tables
Currently, the u32 chain id is mapped to u16 value which is stored on
the lower 16 bits of reg_c0 for FDB and reg_b for NIC tables. The
mapping is internally maintained by the chains object. However, with
the introduction of reg_c0 objects the fdb may store more than just
the chain id on reg_c0. This is not relevant for NIC tables.
Separate the chains mapping instantiation for FDB and NIC tables.
Remove the mapping from the chains object. For FDB tables, create
the mapping per eswitch. For NIC tables, create the mapping per tc
table. Pass the corresponding mapping pointer when creating the
chains object.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Thu, 10 Sep 2020 07:28:02 +0000 (15:28 +0800)]
net/mlx5: Map register values to restore objects
Currently reg_c0 lower 16 bits and reg_b are used to store the chain
id that missed in FDB and NIC tables accordingly. However, the
registers' values may index a restore object, rather than a single u32
value. Different object types can be used to restore mutually exclusive
contexts such as chain id and sample group id.
Use the mapping object to associate an index with a restore object
as a prestep for supporting additional restore types.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Fri, 9 Oct 2020 03:06:33 +0000 (11:06 +0800)]
net/mlx5: E-switch, Set per vport table default group number
Different per voprt table is created using a different per vport table
namespace. Because we can't use variable to set the namespace member
value. If max group number is 0 in the namespace, use the eswitch
default max group number.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Mon, 31 Aug 2020 05:23:25 +0000 (13:23 +0800)]
net/mlx5: E-switch, Generalize per vport table API
Currently, per vport table was used only for port mirroring actions.
However, sample action will also require a per vport table instance.
Generalize the vport table API to work with multiple namespaces where
each namespace manages its own vport table instance.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Thu, 14 Jan 2021 07:12:36 +0000 (15:12 +0800)]
net/mlx5: E-switch, Rename functions to follow naming convention.
Public api starts with mlx5 and remove mlx5 for non-public api.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Chris Mi [Mon, 31 Aug 2020 05:22:20 +0000 (13:22 +0800)]
net/mlx5: E-switch, Move vport table functions to a new file
Currently, the vport table functions are in common eswitch offload
file. This file is too big. Move the vport table create, delete and
lookup functions to a separate file. Put the file in esw directory.
Pre-step for generalizing its functionality for serving both the
mirroring and the sample features.
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Jakub Kicinski [Tue, 6 Apr 2021 22:59:31 +0000 (15:59 -0700)]
docs: ethtool: correct quotes
Quotes to backticks. All commands use backticks since the names
are constants.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Peng Zhang [Tue, 6 Apr 2021 15:54:52 +0000 (17:54 +0200)]
nfp: flower: add support for packet-per-second policing
Allow hardware offload of a policer action attached to a matchall filter
which enforces a packets-per-second rate-limit.
e.g.
tc filter add dev tap1 parent ffff: u32 match \
u32 0 0 police pkts_rate 3000 pkts_burst 1000
Signed-off-by: Peng Zhang <peng.zhang@corigine.com>
Signed-off-by: Baowen Zheng <baowen.zheng@corigine.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Louis Peens <louis.peens@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 6 Apr 2021 23:36:41 +0000 (16:36 -0700)]
Merge git://git./linux/kernel/git/pablo/nf-next
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following batch contains Netfilter/IPVS updates for your net-next tree:
1) Simplify log infrastructure modularity: Merge ipv4, ipv6, bridge,
netdev and ARP families to nf_log_syslog.c. Add module softdeps.
This fixes a rare deadlock condition that might occur when log
module autoload is required. From Florian Westphal.
2) Moves part of netfilter related pernet data from struct net to
net_generic() infrastructure. All of these users can be modules,
so if they are not loaded there is no need to waste space. Size
reduction is 7 cachelines on x86_64, also from Florian.
2) Update nftables audit support to report events once per table,
to get it aligned with iptables. From Richard Guy Briggs.
3) Check for stale routes from the flowtable garbage collector path.
This is fixing IPv6 which breaks due missing check for the dst_cookie.
4) Add a nfnl_fill_hdr() function to simplify netlink + nfnetlink
headers setup.
5) Remove documentation on several statified functions.
6) Remove printk on netns creation for the FTP IPVS tracker,
from Florian Westphal.
7) Remove unnecessary nf_tables_destroy_list_lock spinlock
initialization, from Yang Yingliang.
7) Remove a duplicated forward declaration in ipset,
from Wan Jiabing.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Andy Shevchenko [Tue, 6 Apr 2021 10:22:51 +0000 (13:22 +0300)]
time64.h: Consolidated PSEC_PER_SEC definition
We have currently three users of the PSEC_PER_SEC each of them defining it
individually. Instead, move it to time64.h to be available for everyone.
There is a new user coming with the same constant in use. It will also
make its life easier.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andy Shevchenko [Tue, 6 Apr 2021 10:13:06 +0000 (13:13 +0300)]
stmmac: intel: Drop duplicate ID in the list of PCI device IDs
The PCI device IDs are defined with a prefix PCI_DEVICE_ID.
There is no need to repeat the ID part at the end of each definition.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Wong Vee Khee <vee.khee.wong@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Qiheng Lin [Tue, 6 Apr 2021 03:18:13 +0000 (11:18 +0800)]
netdevsim: remove unneeded semicolon
Eliminate the following coccicheck warning:
drivers/net/netdevsim/fib.c:569:2-3: Unneeded semicolon
Signed-off-by: Qiheng Lin <linqiheng@huawei.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Qiheng Lin [Tue, 6 Apr 2021 03:04:33 +0000 (11:04 +0800)]
net: ethernet: mtk_eth_soc: remove unneeded semicolon
Eliminate the following coccicheck warning:
drivers/net/ethernet/mediatek/mtk_ppe.c:270:2-3: Unneeded semicolon
Signed-off-by: Qiheng Lin <linqiheng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
wengjianfeng [Tue, 6 Apr 2021 01:59:54 +0000 (09:59 +0800)]
nfc: s3fwrn5: remove unnecessary label
In function s3fwrn5_nci_post_setup, the variable ret is assigned then
goto out label, which just return ret, so we use return to replace it.
Other goto sentences are similar, we use return sentences to replace
goto sentences and delete out label.
Signed-off-by: wengjianfeng <wengjianfeng@yulong.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 6 Apr 2021 23:22:37 +0000 (16:22 -0700)]
Merge branch 'usbnet-speed'
Grant Grundler says:
====================
usbnet: speed reporting for devices without MDIO
This series introduces support for USB network devices that report
speed as a part of their protocol, not emulating an MII to be accessed
over MDIO.
v2: rebased on recent upstream changes
v3: incorporated hints on naming and comments
v4: fix misplaced hunks; reword some commit messages;
add same change for cdc_ether
v4-repost: added "net-next" to subject and Andrew Lunn's Reviewed-by
I'm reposting Oliver Neukum's <oneukum@suse.com> patch series with
fix ups for "misplaced hunks" (landed in the wrong patches).
Please fixup the "author" if "git am" fails to attribute the
patches 1-3 (of 4) to Oliver.
I've tested v4 series with "5.12-rc3+" kernel on Intel NUC6i5SYB
and + Sabrent NT-S25G. Google Pixelbook Go (chromeos-4.4 kernel)
+ Alpha Network AUE2500C were connected directly to the NT-S25G
to get 2.5Gbps link rate:
Settings for enx002427880815:
Supported ports: [ ]
Supported link modes: Not reported
Supported pause frame use: No
Supports auto-negotiation: No
Supported FEC modes: Not reported
Advertised link modes: Not reported
Advertised pause frame use: No
Advertised auto-negotiation: No
Advertised FEC modes: Not reported
Speed: 2500Mb/s
Duplex: Half
Auto-negotiation: off
Port: Twisted Pair
PHYAD: 0
Transceiver: internal
MDI-X: Unknown
Current message level: 0x00000007 (7)
drv probe link
Link detected: yes
"Duplex" is a lie since we get no information about it.
I expect "Auto-Negotiation" is always true for cdc_ncm and
cdc_ether devices and perhaps someone knows offhand how
to have ethtool report "true" instead.
But this is good step in the right direction.
base-commit:
1c273e10bc0cc7efb933e0ca10e260cdfc9f0b8c
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Grant Grundler [Mon, 5 Apr 2021 23:13:44 +0000 (16:13 -0700)]
net: cdc_ether: record speed in status method
Until very recently, the usbnet framework only had support functions
for devices which reported the link speed by explicitly querying the
PHY over a MDIO interface. However, the cdc_ether devices send
notifications when the link state or link speeds change and do not
expose the PHY (or modem) directly.
Support funtions (e.g. usbnet_get_link_ksettings_internal()) to directly
query state recorded by the cdc_ether driver were added in a previous patch.
Instead of cdc_ether spewing the link speed into the dmesg buffer,
record the link speed encoded in these notifications and tell the
usbnet framework to use the new functions to get link speed/state.
User space can now get the most recent link speed/state using ethtool.
v4: added to series since cdc_ether uses same notifications
as cdc_ncm driver.
Signed-off-by: Grant Grundler <grundler@chromium.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Oliver Neukum [Mon, 5 Apr 2021 23:13:43 +0000 (16:13 -0700)]
net: cdc_ncm: record speed in status method
Until very recently, the usbnet framework only had support functions
for devices which reported the link speed by explicitly querying the
PHY over a MDIO interface. However, the cdc_ncm devices send
notifications when the link state or link speeds change and do not
expose the PHY (or modem) directly.
Support funtions (e.g. usbnet_get_link_ksettings_internal()) to directly
query state recorded by the cdc_ncm driver were added in a previous patch.
So instead of cdc_ncm spewing the link speed into the dmesg buffer,
record the link speed encoded in these notifications and tell the
usbnet framework to use the new functions to get link speed/state.
Link speed/state is now available via ethtool.
This is especially useful given all current RTL8156 devices emit
a connection/speed status notification every 32ms and this would
fill the dmesg buffer. This implementation replaces the one
recently submitted in
de658a195ee23ca6aaffe197d1d2ea040beea0a2 :
"net: usb: cdc_ncm: don't spew notifications"
v2: rebased on upstream
v3: changed variable names
v4: rewrote commit message
Signed-off-by: Oliver Neukum <oneukum@suse.com>
Tested-by: Roland Dreier <roland@kernel.org>
Signed-off-by: Grant Grundler <grundler@chromium.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Oliver Neukum [Mon, 5 Apr 2021 23:13:42 +0000 (16:13 -0700)]
usbnet: add method for reporting speed without MII
The old method for reporting link speed assumed a driver uses the
generic phy (mii) MDIO read/write functions. CDC devices don't
expose the phy.
Add a primitive internal version reporting back directly what
the CDC notification/status operations recorded.
v2: rebased on upstream
v3: changed names and made clear which units are used
v4: moved hunks to correct patch; rewrote commmit messages
Signed-off-by: Oliver Neukum <oneukum@suse.com>
Tested-by: Roland Dreier <roland@kernel.org>
Reviewed-by: Grant Grundler <grundler@chromium.org>
Tested-by: Grant Grundler <grundler@chromium.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Oliver Neukum [Mon, 5 Apr 2021 23:13:41 +0000 (16:13 -0700)]
usbnet: add _mii suffix to usbnet_set/get_link_ksettings
The generic functions assumed devices provided an MDIO interface (accessed
via older mii code, not phylib). This is true only for genuine ethernet.
Devices with a higher level of abstraction or based on different
technologies do not have MDIO. To support this case, first rename
the existing functions with _mii suffix.
v2: rebased on changed upstream
v3: changed names to clearly say that this does NOT use phylib
v4: moved hunks to correct patch; reworded commmit messages
Signed-off-by : Oliver Neukum <oneukum@suse.com>
Tested-by: Roland Dreier <roland@kernel.org>
Reviewed-by: Grant Grundler <grundler@chromium.org>
Tested-by: Grant Grundler <grundler@chromium.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Manoj Basapathi [Mon, 5 Apr 2021 17:02:42 +0000 (22:32 +0530)]
tcp: Reset tcp connections in SYN-SENT state
Userspace sends tcp connection (sock) destroy on network switch
i.e switching the default network of the device between multiple
networks(Cellular/Wifi/Ethernet).
Kernel though doesn't send reset for the connections in SYN-SENT state
and these connections continue to remain.
Even as per RFC 793, there is no hard rule to not send RST on ABORT in
this state.
Modify tcp_abort and tcp_disconnect behavior to send RST for connections
in syn-sent state to avoid lingering connections on network switch.
Signed-off-by: Manoj Basapathi <manojbm@codeaurora.org>
Signed-off-by: Sauvik Saha <ssaha@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Florian Westphal [Thu, 1 Apr 2021 14:11:14 +0000 (16:11 +0200)]
net: remove obsolete members from struct net
all have been moved to generic_net infra. On x86_64, this reduces
struct net size from 70 to 63 cache lines (4480 to 4032 byte).
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:13 +0000 (16:11 +0200)]
netfilter: conntrack: move ecache dwork to net_generic infra
dwork struct is large (>128 byte) and not needed when conntrack module
is not loaded.
Place it in net_generic data instead. The struct net dwork member is now
obsolete and will be removed in a followup patch.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:12 +0000 (16:11 +0200)]
netfilter: conntrack: move sysctl pointer to net_generic infra
No need to keep this in struct net, place it in the net_generic data.
The sysctl pointer is removed from struct net in a followup patch.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:11 +0000 (16:11 +0200)]
netfilter: x_tables: move known table lists to net_generic infra
Will reduce struct net size by 208 bytes.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:10 +0000 (16:11 +0200)]
netfilter: nf_tables: use net_generic infra for transaction data
This moves all nf_tables pernet data from struct net to a net_generic
extension, with the exception of the gencursor.
The latter is used in the data path and also outside of the nf_tables
core. All others are only used from the configuration plane.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:09 +0000 (16:11 +0200)]
netfilter: ebtables: use net_generic infra
ebtables currently uses net->xt.tables[BRIDGE], but upcoming
patch will move net->xt.tables away from struct net.
To avoid exposing x_tables internals to ebtables, use a private list
instead.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:08 +0000 (16:11 +0200)]
netfilter: nf_defrag_ipv4: use net_generic infra
This allows followup patch to remove the defrag_ipv4 member from struct
net. It also allows to auto-remove the hooks later on by adding a
_disable() function. This will be done later in a follow patch series.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:07 +0000 (16:11 +0200)]
netfilter: nf_defrag_ipv6: use net_generic infra
This allows followup patch to remove these members from struct net.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:06 +0000 (16:11 +0200)]
netfilter: cttimeout: use net_generic infra
reduce size of struct net and make this self-contained.
The member in struct net is kept to minimize changes to struct net
layout, it will be removed in a separate patch.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:05 +0000 (16:11 +0200)]
netfilter: nfnetlink: use net_generic infra
No need to place it in struct net, nfnetlink is a module and usage
doesn't occur in fastpath.
Also remove rcu usage:
Not a single reader of net->nfnl uses rcu accessors.
When exit_batch callbacks are executed the net namespace is already dead
so no calls to these functions are possible anymore (else we'd get NULL
deref crash too).
If the module is removed, then modules that call any of those functions
have been removed too so no calls to nfnl functions are possible either.
The nfnl and nfl_stash pointers in struct net are no longer used, they
will be removed in a followup patch to minimize changes to struct net
(causes rebuild for entire network stack).
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Thu, 1 Apr 2021 14:11:04 +0000 (16:11 +0200)]
netfilter: nfnetlink: add and use nfnetlink_broadcast
This removes the only reference of net->nfnl outside of the nfnetlink
module. This allows to move net->nfnl to net_generic infra.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Krzysztof Kozlowski [Mon, 5 Apr 2021 18:15:48 +0000 (20:15 +0200)]
net: smsc911x: skip acpi_device_id table when !CONFIG_ACPI
The driver can match via multiple methods. Its acpi_device_id table is
referenced via ACPI_PTR() so it will be unused for !CONFIG_ACPI builds:
drivers/net/ethernet/smsc/smsc911x.c:2652:36: warning:
‘smsc911x_acpi_match’ defined but not used [-Wunused-const-variable=]
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Salil Mehta [Mon, 5 Apr 2021 17:28:25 +0000 (18:28 +0100)]
net: hns3: Limiting the scope of vector_ring_chain variable
Limiting the scope of the variable vector_ring_chain to the block where it
is used.
Fixes:
424eb834a9be ("net: hns3: Unified HNS3 {VF|PF} Ethernet Driver for hip08 SoC")
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrei Vagin [Mon, 5 Apr 2021 07:12:23 +0000 (00:12 -0700)]
net: Allow to specify ifindex when device is moved to another namespace
Currently, we can specify ifindex on link creation. This change allows
to specify ifindex when a device is moved to another network namespace.
Even now, a device ifindex can be changed if there is another device
with the same ifindex in the target namespace. So this change doesn't
introduce completely new behavior, it adds more control to the process.
CRIU users want to restore containers with pre-created network devices.
A user will provide network devices and instructions where they have to
be restored, then CRIU will restore network namespaces and move devices
into them. The problem is that devices have to be restored with the same
indexes that they have before C/R.
Cc: Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>
Suggested-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Andrei Vagin <avagin@gmail.com>
Reviewed-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Zheng Yongjun [Mon, 5 Apr 2021 10:54:35 +0000 (18:54 +0800)]
net: nfc: Fix spelling errors in net/nfc module
These patches fix a series of spelling errors in net/nfc module.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Wu XiangCheng [Sun, 4 Apr 2021 14:23:15 +0000 (22:23 +0800)]
tipc: Fix a kernel-doc warning in name_table.c
Fix kernel-doc warning:
Documentation/networking/tipc:66: /home/sfr/next/next/net/tipc/name_table.c
:558: WARNING: Unexpected indentation.
Documentation/networking/tipc:66: /home/sfr/next/next/net/tipc/name_table.c
:559: WARNING: Block quote ends without a blank line; unexpected unindent.
Due to blank line missing.
Fixes:
908148bc5046 ("tipc: refactor tipc_sendmsg() and tipc_lookup_anycast()")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Link: https://lore.kernel.org/netdev/20210318172255.63185609@canb.auug.org.au/
Signed-off-by: Wu XiangCheng <bobwxc@email.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Sun, 4 Apr 2021 13:38:23 +0000 (13:38 +0000)]
mld: change lockdep annotation for ip6_sf_socklist and ipv6_mc_socklist
struct ip6_sf_socklist and ipv6_mc_socklist are per-socket MLD data.
These data are protected by rtnl lock, socket lock, and RCU.
So, when these are used, it verifies whether rtnl lock is acquired or not.
ip6_mc_msfget() is called by do_ipv6_getsockopt().
But caller doesn't acquire rtnl lock.
So, when these data are used in the ip6_mc_msfget() lockdep warns about it.
But accessing these is actually safe because socket lock was acquired by
do_ipv6_getsockopt().
So, it changes lockdep annotation from rtnl lock to socket lock.
(rtnl_dereference -> sock_dereference)
Locking graph for mld data is like below:
When writing mld data:
do_ipv6_setsockopt()
rtnl_lock
lock_sock
(mld functions)
idev->mc_lock(if per-interface mld data is modified)
When reading mld data:
do_ipv6_getsockopt()
lock_sock
ip6_mc_msfget()
Splat looks like:
=============================
WARNING: suspicious RCU usage
5.12.0-rc4+ #503 Not tainted
-----------------------------
net/ipv6/mcast.c:610 suspicious rcu_dereference_protected() usage!
other info that might help us debug this:
rcu_scheduler_active = 2, debug_locks = 1
1 lock held by mcast-listener-/923:
#0:
ffff888007958a70 (sk_lock-AF_INET6){+.+.}-{0:0}, at:
ipv6_get_msfilter+0xaf/0x190
stack backtrace:
CPU: 1 PID: 923 Comm: mcast-listener- Not tainted 5.12.0-rc4+ #503
Call Trace:
dump_stack+0xa4/0xe5
ip6_mc_msfget+0x553/0x6c0
? ipv6_sock_mc_join_ssm+0x10/0x10
? lockdep_hardirqs_on_prepare+0x3e0/0x3e0
? mark_held_locks+0xb7/0x120
? lockdep_hardirqs_on_prepare+0x27c/0x3e0
? __local_bh_enable_ip+0xa5/0xf0
? lock_sock_nested+0x82/0xf0
ipv6_get_msfilter+0xc3/0x190
? compat_ipv6_get_msfilter+0x300/0x300
? lock_downgrade+0x690/0x690
do_ipv6_getsockopt.isra.6.constprop.13+0x1809/0x29e0
? do_ipv6_mcast_group_source+0x150/0x150
? register_lock_class+0x1750/0x1750
? kvm_sched_clock_read+0x14/0x30
? sched_clock+0x5/0x10
? sched_clock_cpu+0x18/0x170
? find_held_lock+0x3a/0x1c0
? lock_downgrade+0x690/0x690
? ipv6_getsockopt+0xdb/0x1b0
ipv6_getsockopt+0xdb/0x1b0
[ ... ]
Fixes:
88e2ca308094 ("mld: convert ifmcaddr6 to RCU")
Reported-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Christophe JAILLET [Sun, 4 Apr 2021 12:42:57 +0000 (14:42 +0200)]
qede: Use 'skb_add_rx_frag()' instead of hand coding it
Some lines of code can be merged into an equivalent 'skb_add_rx_frag()'
call which is less verbose.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Christophe JAILLET [Sun, 4 Apr 2021 12:42:44 +0000 (14:42 +0200)]
qede: Remove a erroneous ++ in 'qede_rx_build_jumbo()'
This ++ is confusing. It looks duplicated with the one already performed in
'skb_fill_page_desc()'.
In fact, it is harmless. 'nr_frags' is written twice with the same value.
Once, because of the nr_frags++, and once because of the 'nr_frags = i + 1'
in 'skb_fill_page_desc()'.
So axe this post-increment to avoid confusion.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Christophe JAILLET [Sun, 4 Apr 2021 09:45:11 +0000 (11:45 +0200)]
sfc: Use 'skb_add_rx_frag()' instead of hand coding it
Some lines of code can be merged into an equivalent 'skb_add_rx_frag()'
call which is less verbose.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
Christophe JAILLET [Sun, 4 Apr 2021 08:54:37 +0000 (10:54 +0200)]
ibmvnic: Use 'skb_frag_address()' instead of hand coding it
'page_address(skb_frag_page()) + skb_frag_off()' can be replaced by an
equivalent 'skb_frag_address()' which is less verbose.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
Christophe JAILLET [Sun, 4 Apr 2021 06:33:44 +0000 (08:33 +0200)]
net: ag71xx: Slightly simplify 'ag71xx_rx_packets()'
There is no need to use 'list_for_each_entry_safe' here, as nothing is
removed from the list in the 'for' loop.
Use 'list_for_each_entry' instead, it is slightly less verbose.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
Xie He [Fri, 2 Apr 2021 09:30:00 +0000 (02:30 -0700)]
net: x25: Queue received packets in the drivers instead of per-CPU queues
X.25 Layer 3 (the Packet Layer) expects layer 2 to provide a reliable
datalink service such that no packets are reordered or dropped. And
X.25 Layer 2 (the LAPB layer) is indeed designed to provide such service.
However, this reliability is not preserved when a driver calls "netif_rx"
to deliver the received packets to layer 3, because "netif_rx" will put
the packets into per-CPU queues before they are delivered to layer 3.
If there are multiple CPUs, the order of the packets may not be preserved.
The per-CPU queues may also drop packets if there are too many.
Therefore, we should not call "netif_rx" to let it queue the packets.
Instead, we should use our own queue that won't reorder or drop packets.
This patch changes all X.25 drivers to use their own queues instead of
calling "netif_rx". The patch also documents this requirement in the
"x25-iface" documentation.
Cc: Martin Schiller <ms@dev.tdt.de>
Signed-off-by: Xie He <xie.he.0141@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Christophe JAILLET [Sun, 4 Apr 2021 07:11:03 +0000 (09:11 +0200)]
net: openvswitch: Use 'skb_push_rcsum()' instead of hand coding it
'skb_push()'/'skb_postpush_rcsum()' can be replaced by an equivalent
'skb_push_rcsum()' which is less verbose.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 4 Apr 2021 08:41:08 +0000 (01:41 -0700)]
Merge tag 'mlx5-updates-2021-04-02' of git://git./linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5-updates-2021-04-02
This series provides trivial updates and cleanup to mlx5 driver
1) Support for matching on ct_state inv and rel flag in connection tracking
2) Reject TC rules that redirect from a VF to itself
3) Parav provided some E-Switch cleanups that could be summarized to:
3.1) Packing and Reduce structure sizes
3.2) Dynamic allocation of rate limit tables and structures
4) Vu Makes the netdev arfs and vlan tables allocation dynamic.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 3 Apr 2021 21:13:40 +0000 (14:13 -0700)]
Merge branch 'stmmac-xdp'
Ong Boon Leong says:
====================
stmmac: Add XDP support
This is the v4 patch series for adding XDP native support to stmmac.
Changes in v4:
5/6: Move TX clean timer setup to the end of NAPI RX process and
group it under stmmac_finalize_xdp_rx().
Also, fixed stmmac_xdp_xmit_back() returns STMMAC_XDP_CONSUMED
if XDP buffer conversion to XDP frame fails.
6/6: Move xdp_do_flush(0 into stmmac_finalize_xdp_rx() and combine
the XDP verdict of XDP TX and XDP REDIRECT together.
I retested the patch series on the 'xdp2' and 'xdp_redirect' related to
changes above and found the result to be satisfactory.
History of previous patch series:
v3: https://patchwork.kernel.org/project/netdevbpf/cover/
20210331154135.8507-1-boon.leong.ong@intel.com/
v2: https://patchwork.kernel.org/project/netdevbpf/list/?series=457757
v1: https://patchwork.kernel.org/project/netdevbpf/list/?series=457139
It will be great if community can help to test or review the v4 series
and provide me any input if any.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Ong Boon Leong [Thu, 1 Apr 2021 02:11:17 +0000 (10:11 +0800)]
net: stmmac: Add support for XDP_REDIRECT action
This patch adds the support of XDP_REDIRECT to another remote cpu for
further action. It also implements ndo_xdp_xmit ops, enabling the driver
to transmit packets forwarded to it by XDP program running on another
interface.
This patch has been tested using "xdp_redirect_cpu" for XDP_REDIRECT
+ drop testing. It also been tested with "xdp_redirect" sample app
which can be used to exercise ndo_xdp_xmit ops. The burst traffics are
generated using pktgen_sample03_burst_single_flow.sh in samples/pktgen
directory.
v4: Move xdp_do_flush() processing into stmmac_finalize_xdp_rx() and
combined the XDP verdict of XDP TX and REDIRECT together.
v3: Added 'nq->trans_start = jiffies' to avoid TX time-out as we are
sharing TX queue between slow path and XDP. Thanks to Jakub Kicinski
for point out.
Signed-off-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ong Boon Leong [Thu, 1 Apr 2021 02:11:16 +0000 (10:11 +0800)]
net: stmmac: Add support for XDP_TX action
This patch adds support for XDP_TX action which enables XDP program to
transmit back received frames.
This patch has been tested with the "xdp2" app located in samples/bpf
dir. The DUT receives burst traffic packet generated using pktgen script
'pktgen_sample03_burst_single_flow.sh'.
v4: Moved stmmac_tx_timer_arm() to be done once at the end of NAPI RX.
Fixed stmmac_xdp_xmit_back() to return STMMAC_XDP_CONSUMED if
XDP buffer to frame conversion fails. Thanks to Jakub's input.
v3: Added 'nq->trans_start = jiffies' to avoid TX time-out as we are
sharing TX queue between slow path and XDP. Thanks to Jakub Kicinski
for pointing out.
Signed-off-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ong Boon Leong [Thu, 1 Apr 2021 02:11:15 +0000 (10:11 +0800)]
net: stmmac: Add initial XDP support
This patch adds the initial XDP support to stmmac driver. It supports
XDP_PASS, XDP_DROP and XDP_ABORTED actions. Upcoming patches will add
support for XDP_TX and XDP_REDIRECT.
To support XDP headroom, this patch adds page_offset into RX buffer and
change the dma_sync_single_for_device|cpu(). The DMA address used for
RX operation are changed to take into page_offset too. As page_pool
can handle dma_sync_single_for_device() on behalf of driver with
PP_FLAG_DMA_SYNC_DEV flag, we skip doing that in stmmac driver.
Current stmmac driver supports split header support (SPH) in RX but
the flexibility of splitting header and payload at different position
makes it very complex to be supported for XDP processing. In addition,
jumbo frame is not supported in XDP to keep the initial codes simple.
This patch has been tested with the sample app "xdp1" located in
samples/bpf directory for both SKB and Native (XDP) mode. The burst
traffic generated using pktgen_sample03_burst_single_flow.sh in
samples/pktgen directory.
Changes in v3:
- factor in xdp header and tail adjustment done by XDP program.
Thanks to Jakub Kicinski for pointing out the gap in v2.
Changes in v2:
- fix for "warning: variable 'len' set but not used" reported by lkp.
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ong Boon Leong [Thu, 1 Apr 2021 02:11:14 +0000 (10:11 +0800)]
net: stmmac: arrange Tx tail pointer update to stmmac_flush_tx_descriptors
This patch organizes TX tail pointer update into a new function called
stmmac_flush_tx_descriptors() so that we can reuse it in stmmac_xmit(),
stmmac_tso_xmit() and up-coming XDP implementation.
Changes to v2:
- Fix for warning: unused variable ‘desc_size’
https://patchwork.hopto.org/static/nipa/457321/
12170149/build_32bit/stderr
Signed-off-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ong Boon Leong [Thu, 1 Apr 2021 02:11:13 +0000 (10:11 +0800)]
net: stmmac: make SPH enable/disable to be configurable
SPH functionality splits header and payload according to split mode and
offsef fields (SPLM and SPLOFST). It is beneficials for Linux network
stack RX processing however it adds a lot of complexity in XDP
processing.
So, this patch makes the split-header (SPH) capability of the controller
is stored in "priv->sph_cap" and the enabling/disabling of SPH is decided
by "priv->sph".
This is to prepare initial XDP enabling for stmmac to disable the use of
SPH whenever XDP is enabled.
Signed-off-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ong Boon Leong [Thu, 1 Apr 2021 02:11:12 +0000 (10:11 +0800)]
net: stmmac: set IRQ affinity hint for multi MSI vectors
Certain platform likes Intel mGBE has independent hardware IRQ resources
for TX and RX DMA operation. In preparation to support XDP TX, we add IRQ
affinity hint to group both RX and TX queue of the same queue ID to the
same CPU.
Changes in v2:
- IRQ affinity hint need to set to null before IRQ is released.
Thanks to issue reported by Song, Yoong Siang.
Reported-by: Song, Yoong Siang <yoong.siang.song@intel.com>
Signed-off-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso [Wed, 31 Mar 2021 21:10:15 +0000 (23:10 +0200)]
netfilter: nftables: remove documentation on static functions
Since
4f16d25c68ec ("netfilter: nftables: add nft_parse_register_load()
and use it") and
345023b0db31 ("netfilter: nftables: add
nft_parse_register_store() and use it"), the following functions are not
exported symbols anymore:
- nft_parse_register()
- nft_validate_register_load()
- nft_validate_register_store()
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Dan Carpenter [Fri, 2 Apr 2021 11:45:44 +0000 (14:45 +0300)]
netfilter: nftables: fix a warning message in nf_tables_commit_audit_collect()
The first argument of a WARN_ONCE() is a condition. This WARN_ONCE()
will only print the table name, and is potentially problematic if the
table name has a %s in it.
Fixes:
c520292f29b8 ("audit: log nftables configuration change events once per table")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Florian Westphal [Tue, 30 Mar 2021 06:42:32 +0000 (08:42 +0200)]
netfilter: ipvs: do not printk on netns creation
This causes dmesg spew during normal operation, so remove this.
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Julian Anastasov <ja@ssi.bg>
Reviewed-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Vu Pham [Sat, 27 Feb 2021 05:17:12 +0000 (21:17 -0800)]
net/mlx5e: Dynamic alloc vlan table for netdev when needed
Dynamic allocate vlan table in mlx5e_priv for EN netdev
when needed. Don't allocate it for representor netdev.
Signed-off-by: Vu Pham <vuhuong@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Vu Pham [Fri, 26 Feb 2021 23:51:12 +0000 (15:51 -0800)]
net/mlx5e: Dynamic alloc arfs table for netdev when needed
Dynamic allocate arfs table in mlx5e_priv for EN netdev
when needed. Don't allocate it for representor netdev.
Signed-off-by: Vu Pham <vuhuong@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Ariel Levkovich [Thu, 11 Mar 2021 18:37:21 +0000 (20:37 +0200)]
net/mlx5e: Reject tc rules which redirect from a VF to itself
Since there are self loopback prevention mechanisms at the
VF level, offloading such rules which redirect from a VF
to itself in the eswitch will break the datapath since the
packets will be dropped once they go back to the vport they
came from.
Therefore, offloading such rules will be rejected and left to
be handled by SW.
Signed-off-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Roi Dayan [Mon, 15 Mar 2021 11:31:40 +0000 (13:31 +0200)]
net/mlx5: Use ida_alloc_range() instead of ida_simple_alloc()
ida_simple_alloc() and remove functions are deprecated.
Related change:
commit
3264ceec8f17 ("lib/idr.c: document that ida_simple_{get,remove}() are deprecated")
Signed-off-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Wed, 3 Feb 2021 13:32:50 +0000 (15:32 +0200)]
net/mlx5: E-Switch, move QoS specific fields to existing qos struct
Function QoS related fields are already defined in qos related struct.
min and max rate are left out to mlx5_vport_info struct.
Move them to existing qos struct.
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Wed, 3 Feb 2021 04:59:32 +0000 (06:59 +0200)]
net/mlx5: E-Switch, cut down mlx5_vport_info structure size by 8 bytes
Structure mlx5_vport_info consumes 40 bytes of space due to a hole
in it. After packing it reduces to 32 bytes.
Currently:
pahole -C mlx5_vport_info drivers/net/ethernet/mellanox/mlx5/core/eswitch.o
struct mlx5_vport_info {
u8 mac[6]; /* 0 6 */
u16 vlan; /* 6 2 */
u8 qos; /* 8 1 */
/* XXX 7 bytes hole, try to pack */
u64 node_guid; /* 16 8 */
int link_state; /* 24 4 */
u32 min_rate; /* 28 4 */
u32 max_rate; /* 32 4 */
bool spoofchk; /* 36 1 */
bool trusted; /* 37 1 */
/* size: 40, cachelines: 1, members: 9 */
/* sum members: 31, holes: 1, sum holes: 7 */
/* padding: 2 */
/* last cacheline: 40 bytes */
};
After packing:
$ pahole -C mlx5_vport_info drivers/net/ethernet/mellanox/mlx5/core/eswitch.o
struct mlx5_vport_info {
u8 mac[6]; /* 0 6 */
u16 vlan; /* 6 2 */
u64 node_guid; /* 8 8 */
int link_state; /* 16 4 */
u32 min_rate; /* 20 4 */
u32 max_rate; /* 24 4 */
u8 qos; /* 28 1 */
u8 spoofchk:1; /* 29: 0 1 */
u8 trusted:1; /* 29: 1 1 */
/* size: 32, cachelines: 1, members: 9 */
/* padding: 2 */
/* bit_padding: 6 bits */
/* last cacheline: 32 bytes */
};
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Fri, 19 Feb 2021 07:36:54 +0000 (09:36 +0200)]
net/mlx5: Pair mutex_destory with mutex_init for rate limit table
Add missing mutex_destroy() to pair with mutex_init().
This should be done only when table is initialized, hence perform
mutex_init() only when table is initialized.
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Fri, 19 Feb 2021 10:06:54 +0000 (12:06 +0200)]
net/mlx5: Allocate rate limit table when rate is configured
A device supports 128 rate limiters. A static table allocation consumes
8KB of memory even when rate is not configured.
Instead, allocate the table when at least one rate is configured.
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Thu, 25 Feb 2021 11:38:07 +0000 (13:38 +0200)]
net/mlx5: Use helper to increment, decrement rate entry refcount
Rate limit entry refcount can be incremented uniformly when it is newly
allocated or reused.
So simplify the code to increment refcount at one place.
Use decrement refcount helper in two routines.
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Fri, 19 Feb 2021 07:29:56 +0000 (09:29 +0200)]
net/mlx5: Use helpers to allocate and free rl table entries
User helper routines to allocate and free rate limit table entries.
Subsequent patch extends use of these helpers to do allocation
during rate entry allocation callback.
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Wed, 24 Feb 2021 09:04:19 +0000 (11:04 +0200)]
net/mlx5: Do not hold mutex while reading table constants
Table max_size, min and max rate are constants initialized while table
is created. Reading it doesn't need to hold a table mutex. Hence, read
them without holding table mutex.
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Fri, 19 Feb 2021 06:18:12 +0000 (08:18 +0200)]
net/mlx5: Pack mlx5_rl_entry structure
mlx5_rl_entry structure is not properly packed as shown below. Due to this
an array of size 9144 bytes allocated which is aligned to 16Kbytes.
Hence, pack the structure and avoid the wastage.
This offers 8Kbytes of saving per mlx5_core_dev struct.
pahole -C mlx5_rl_entry drivers/net/ethernet/mellanox/mlx5/core/en_main.o
Existing layout:
struct mlx5_rl_entry {
u8 rl_raw[48]; /* 0 48 */
u16 index; /* 48 2 */
/* XXX 6 bytes hole, try to pack */
u64 refcount; /* 56 8 */
/* --- cacheline 1 boundary (64 bytes) --- */
u16 uid; /* 64 2 */
u8 dedicated:1; /* 66: 0 1 */
/* size: 72, cachelines: 2, members: 5 */
/* sum members: 60, holes: 1, sum holes: 6 */
/* sum bitfield members: 1 bits (0 bytes) */
/* padding: 5 */
/* bit_padding: 7 bits */
/* last cacheline: 8 bytes */
};
After alignment:
struct mlx5_rl_entry {
u8 rl_raw[48]; /* 0 48 */
u64 refcount; /* 48 8 */
u16 index; /* 56 2 */
u16 uid; /* 58 2 */
u8 dedicated:1; /* 60: 0 1 */
/* size: 64, cachelines: 1, members: 5 */
/* padding: 3 */
/* bit_padding: 7 bits */
};
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Tue, 9 Feb 2021 07:44:27 +0000 (09:44 +0200)]
net/mlx5: Use unsigned int for free_count
Fix the warning due to missing int.
WARNING: Prefer 'unsigned int' to bare use of 'unsigned'
+ unsigned free_count;
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Wed, 3 Feb 2021 13:32:50 +0000 (15:32 +0200)]
net/mlx5: E-Switch, move QoS specific fields to existing qos struct
Function QoS related fields are already defined in qos related struct.
min and max rate are left out to mlx5_vport_info struct.
Move them to existing qos struct.
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Parav Pandit [Wed, 3 Feb 2021 04:59:32 +0000 (06:59 +0200)]
net/mlx5: E-Switch, cut down mlx5_vport_info structure size by 8 bytes
Structure mlx5_vport_info consumes 40 bytes of space due to a hole
in it. After packing it reduces to 32 bytes.
Currently:
pahole -C mlx5_vport_info drivers/net/ethernet/mellanox/mlx5/core/eswitch.o
struct mlx5_vport_info {
u8 mac[6]; /* 0 6 */
u16 vlan; /* 6 2 */
u8 qos; /* 8 1 */
/* XXX 7 bytes hole, try to pack */
u64 node_guid; /* 16 8 */
int link_state; /* 24 4 */
u32 min_rate; /* 28 4 */
u32 max_rate; /* 32 4 */
bool spoofchk; /* 36 1 */
bool trusted; /* 37 1 */
/* size: 40, cachelines: 1, members: 9 */
/* sum members: 31, holes: 1, sum holes: 7 */
/* padding: 2 */
/* last cacheline: 40 bytes */
};
After packing:
$ pahole -C mlx5_vport_info drivers/net/ethernet/mellanox/mlx5/core/eswitch.o
struct mlx5_vport_info {
u8 mac[6]; /* 0 6 */
u16 vlan; /* 6 2 */
u64 node_guid; /* 8 8 */
int link_state; /* 16 4 */
u32 min_rate; /* 20 4 */
u32 max_rate; /* 24 4 */
u8 qos; /* 28 1 */
u8 spoofchk:1; /* 29: 0 1 */
u8 trusted:1; /* 29: 1 1 */
/* size: 32, cachelines: 1, members: 9 */
/* padding: 2 */
/* bit_padding: 6 bits */
/* last cacheline: 32 bytes */
};
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Ariel Levkovich [Mon, 15 Mar 2021 17:59:52 +0000 (19:59 +0200)]
net/mlx5: CT: Add support for matching on ct_state inv and rel flags
Add support for matching on ct_state inv and rel flags.
Currently the support is only for match on -inv and -rel.
Matching on +inv and +rel will be rejected.
Example:
$ tc filter add dev ens1f0_0 ingress prio 1 chain 1 proto ip flower \
ct_state -est-rel+trk \
action mirred egress redirect dev ens1f0_1
$ tc filter add dev ens1f0_1 ingress prio 1 chain 1 proto ip flower \
ct_state +trk+est-inv \
action mirred egress redirect dev ens1f0_0
Signed-off-by: Ariel Levkovich <lariel@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Eric Dumazet [Fri, 2 Apr 2021 18:10:37 +0000 (11:10 -0700)]
tcp: reorder tcp_congestion_ops for better cache locality
Group all the often used fields in the first cache line,
to reduce cache line misses.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Fri, 2 Apr 2021 18:07:46 +0000 (11:07 -0700)]
net: reorganize fields in netns_mib
Order fields to increase locality for most used protocols.
udplite and icmp are moved at the end.
Same for proc_net_devsnmp6 which is not used in fast path.
This potentially saves one cache line miss for typical TCP/UDP over IPv4/IPv6.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Dan Carpenter [Fri, 2 Apr 2021 11:44:42 +0000 (14:44 +0300)]
nfc: pn533: prevent potential memory corruption
If the "type_a->nfcid_len" is too large then it would lead to memory
corruption in pn533_target_found_type_a() when we do:
memcpy(nfc_tgt->nfcid1, tgt_type_a->nfcid_data, nfc_tgt->nfcid1_len);
Fixes:
c3b1e1e8a76f ("NFC: Export NFCID1 from pn533")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 2 Apr 2021 21:25:47 +0000 (14:25 -0700)]
Merge branch 'dpaa2-rx-copybreak'
Ioana Ciornei says:
====================
dpaa2-eth: add rx copybreak support
DMA unmapping, allocating a new buffer and DMA mapping it back on the
refill path is really not that efficient. Proper buffer recycling (page
pool, flipping the page and using the other half) cannot be done for
DPAA2 since it's not a ring based controller but it rather deals with
multiple queues which all get their buffers from the same buffer pool on
Rx.
To circumvent these limitations, add support for Rx copybreak in
dpaa2-eth.
Below you can find a summary of the tests that were run to end up
with the default rx copybreak value of 512.
A bit about the setup - a LS2088A SoC, 8 x Cortex A72 @ 1.8GHz, IPfwd
zero loss test @ 20Gbit/s throughput. I tested multiple frame sizes to
get an idea where is the break even point.
Here are 2 sets of results, (1) is the baseline and (2) is just
allocating a new skb for all frames sizes received (as if the copybreak
was even to the MTU). All numbers are in Mpps.
64 128 256 512 640 768 896
(1) 3.23 3.23 3.24 3.21 3.1 2.76 2.71
(2) 3.95 3.88 3.79 3.62 3.3 3.02 2.65
It seems that even for 512 bytes frame sizes it's comfortably better when
allocating a new skb. After that, we see diminishing rewards or even worse.
Changes in v2:
- properly marked dpaa2_eth_copybreak as static
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Ioana Ciornei [Fri, 2 Apr 2021 09:55:32 +0000 (12:55 +0300)]
dpaa2-eth: export the rx copybreak value as an ethtool tunable
It's useful, especially for debugging purposes, to have the Rx copybreak
value changeable at runtime. Export it as an ethtool tunable.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ioana Ciornei [Fri, 2 Apr 2021 09:55:31 +0000 (12:55 +0300)]
dpaa2-eth: add rx copybreak support
DMA unmapping, allocating a new buffer and DMA mapping it back on the
refill path is really not that efficient. Proper buffer recycling (page
pool, flipping the page and using the other half) cannot be done for
DPAA2 since it's not a ring based controller but it rather deals with
multiple queues which all get their buffers from the same buffer pool on
Rx.
To circumvent these limitations, add support for Rx copybreak. For small
sized packets instead of creating a skb around the buffer in which the
frame was received, allocate a new sk buffer altogether, copy the
contents of the frame and release the initial page back into the buffer
pool.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>