Vivien Didelot says:
====================
net: dsa: mv88e6xxx: rework ATU support
The purpose of this patch series is to rework the code related to the
Address Translation Unit (ATU), and bring support for it to the 88E6390
family of switch chips.
All Global (1) ATU related code have been reworked and moved to its own
file. Some port related bits used for ATU configuration (such as the
Learn2All and MessagePort feature) have also been taken care of.
The ports' mode and egress flooding mode have been refactored to fix the
egress of frames with unknown unicast or multicast destination address,
and write all these bits regardless the port mode (Normal, DSA, etc.)
Finally remove the eth_addr_greater which was only used by mv88e6xxx.
Changes in v2:
- add Reviewed-by tags
- split mv88e6xxx_g1_atu_set_age_time and mv88e6xxx_atu_setup addition
- remove DSA_TAG_PROTO_TRAILER check
- split Message Port and Learn2All addition
- remove unused MV88E6XXX_FLAG_G1_ATU_FID flag
- add dsa_is_normal_port helper
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
eth_addr_greater() was introduced for the mv88e6xxx driver, but is not
used anymore. There is no other user, thus remove this function.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a new operation to disable the DA, SA and VTU priority override.
Setting such limit is not likely to be used soon, so provide a
port_disable_pri_override operation directly. This can be changed later
for port_set_pri_override when we'll need it.
Also remove the now obsolete mv88e6xxx_6320_family helper.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a new operation to disable the limiting of learnt MAC addresses.
Setting such limit is not likely to be used soon, so provide a
port_disable_learn_limit operation directly. This can be changed later
for port_set_learn_limit when we'll need it.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Marvell switch ports can be configured to allow or prevent egress of
frames with an unknown unicast or multicast destination address.
Some switch chips such as 88E6095 and 88E6185 have two disjoint bits in
Port Control Register (0x04) bit 2 "Forward Unknown" (for unicast) and
Port Control 2 Register (0x08) bit 6 "Default Forward" (for multicast).
Other chips such as 88E6085, 88E6123, 88E6352, and 88E6390 have a 2-bit
value in Port Control Register (0x04) bits 3:2 "EgressFloods".
The current code does not fully implement the disjoint bits variant and
assigns incorrect ones to some chip models. Fix that with two
implementation references (6185 and 6352 that I currently have) of a
port_set_egress_floods operation (as named in datasheets).
Old chips such as 88E6060 don't have egress flooding mode, so don't
error out if the operation is not provided.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A switch port mode is defined by the association of its egress mode, its
frame mode and if supported or required, the ether type value.
Pack all this in a mv88e6xxx_set_port_mode function and provide helpers
for the Normal Network mode, the DSA mode, and the Ether Type DSA mode,
as well as an helper to setup a port's mode depending on its nature.
Define PORT_ETH_TYPE_DEFAULT for the 0x9100 reset value of port E Type.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduce a dsa_is_normal_port helper to check if a given port is a
normal user port as opposed to a CPU port or DSA link.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Not all Marvell switch chips support port trunking, which is embedded in
the port vector data for ATU operations.
Rename the portv_trunkid member of the mv88e6xxx_atu_entry structure to
portvec to be more concise and consistent with the different chips.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rename the _mv88e6xxx_fid_new helper to mv88e6xxx_atu_new to get rid of
the old underscore prefix naming convention and be consistent with the
rest of the chip-wide ATU API.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a fresh documented implementation of the ATU Move operation, and use
it to replace the current ATU Remove operation.
Note that not all Marvell switch chip support the ATU Move operation.
For those supporting it, the number of bits used to mask the destination
port may vary. 6352 and such use 4-bit, while 6390 use 5-bit. Thus add a
new atu_move_port_mask member in the info structure to describe the
presence and variant of ATU Move operation.
Note that the ATU Move operation is not documented in the 6185 datasheet
but the chip does support the operation.
All remaining _mv88e6xxx_atu_* functions are now unused as well as the
MV88E6XXX_FLAG_G1_ATU_FID flag, thus remove them.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a fresh documented implementation of the ATU Flush/Move operation.
Use it to replace the current ATU Flush operation.
_mv88e6xxx_atu_flush_move is still used by the Move operation so keep it
until the Move operation is refactored in a next commit.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add and use a fresh documented implementation of the ATU GetNext.
Since it is not necessary to write the MAC address to iterate from, only
do it once directly in the ATU GetNext operation, if the provided ATU
entry structure is not valid. This makes the user code simpler.
Also, there is no need to loop when getting a single ATU entry. So
remove the mv88e6xxx_atu_get helper and add a simpler snippet in
mv88e6xxx_port_db_load_purge to lookup a given MAC address.
The _mv88e6xxx_atu_mac_{read,write} are not used anymore thus remove
them. _mv88e6xxx_atu_data_{read,write} are still used so keep them.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All Marvell switch chips have an ATU accessed using the same Global (1)
register layout. Only the handling of the FID differs as more bits were
necessary to support more and more databases.
Add and use a fresh documented implementation of the ATU Load/Purge.
The static mv88e6xxx_g1_atu_{fid_write,op_wait,op,data_write,mac_write}
functions won't need to be exposed in the end so for the moment keep
their counterparts _mv88e6xxx_atu_{wait,cmd,data_write,mac_write} as is,
since they are still used by other ATU operations.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The ATU Learn2All feature allows newly learnt addresses to be spanned on
ports marked as "Message Port", currently all DSA ports.
This commit enables this feature which is necessary and quite convenient
for multi-chip switch fabrics.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All interconnectable Marvell switch chips have an ATU Learn2All feature
which allows newly learnt addresses to be spanned on ports marked as
"Message Port".
This commit configures the DSA ports as Message Port. Note that this has
no effect until the Learn2All feature is enabled.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move the configuration of the default ageing time in a new
mv88e6xxx_atu_setup function.
That function will be extended later to contain all ATU related
configuration bits.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move the ATU ageing time setter code in a new global1_atu.c file, which
will be extended in future patches to contains all consequent Global (1)
ATU support code.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a mv88e6xxx_port_mask() helper to get the bitmask of ports in a
switch chip, that will be used in several features.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Improve bpf_{prog,jit_binary}_{un,}lock_ro() by throwing a
one-time warning in case of an error when the image couldn't
be set read-only, and also mark struct bpf_prog as locked when
bpf_prog_lock_ro() was called.
Reason for the latter is that bpf_prog_unlock_ro() is called from
various places including error paths, and we shouldn't mess with
page attributes when really not needed.
For bpf_jit_binary_unlock_ro() this is not needed as jited flag
implicitly indicates this, thus for archs with ARCH_HAS_SET_MEMORY
we're guaranteed to have a previously locked image. Overall, this
should also help us to identify any further potential issues with
set_memory_*() helpers.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko says:
====================
mlxsw: Add support for TC flower offload statistics
Arkadi says:
This patchset adds support for retrieving TC flower statistics for
offloaded rules, which includes packets count, bytes count and last used
time stamp. Currently The statistics are gathered on a per-rule basis.
This patchset also includes generic allocator for counters.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for TC flower offload statistics including number of packets,
bytes and last use timestamp. Currently the statistics are gathered on a
per-rule basis.
Signed-off-by: Arkadi Sharshvesky <arkadis@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for packets and byte statistics on TCAM entries. The counters
are allocated from the generic flow counters pool.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for Policing and Counting action block. This action block
will be used to bind counter to TCAM entries.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduce periodic task for dumping the activity status for the ACL
rule TCAM entries. This is done in order to emulate last use statistics.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.comi>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently the ACL rules can be accessed only by hashing. In order to
dump the activity the rules are also placed in a list.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for retrieving TCAM entry activity. In order to support ACL
rule activity corresponding TCAM entry should be queried.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for allocating generic flow counter. Generic flow counter
can count packets or packets and bytes and can be assigned to different
hardware processes. First use will be for counting packets and bytes of
ACL rules, and will be introduced in the following patches.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Reviewed-by: Ido schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The MGPC register retrieves generic flow counter value. It will be
used to query ACL counters.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add implementation for counter allocator. The ASIC has special memory
pool for various counting purposes. Counter memory is distributed between
equal size banks.
The static sub-pool configuration should specify the following parameters
for each sub-pool:
- Number of required banks.
- Maximum entry size.
Each module can add dedicated sub-pool or use existing one.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Recent merge of 'linux-kselftest-4.11-rc1' tree broke bpf test build.
None of the tests were building and test_verifier.c had tons of compiler errors.
Fix it and add #ifdef CAP_IS_SUPPORTED to support old versions of libcap.
Tested on centos 6.8 and 7
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use setup_timer() instead of init_timer() to simplify the code.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use setup_timer() instead of init_timer() to simplify the code.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Multipath routes can be rendered usesless when a device in one of the
paths is deleted. For example:
$ ip -f mpls ro ls
100
nexthop as to 200 via inet 172.16.2.2 dev virt12
nexthop as to 300 via inet 172.16.3.2 dev br0
101
nexthop as to 201 via inet6 2000:2::2 dev virt12
nexthop as to 301 via inet6 2000:3::2 dev br0
$ ip li del br0
When br0 is deleted the other hop is not considered in
mpls_select_multipath because of the alive check -- rt_nhn_alive
is 0.
rt_nhn_alive is decremented once in mpls_ifdown when the device is taken
down (NETDEV_DOWN) and again when it is deleted (NETDEV_UNREGISTER). For
a 2 hop route, deleting one device drops the alive count to 0. Since
devices are taken down before unregistering, the decrement on
NETDEV_UNREGISTER is redundant.
Fixes: c89359a42e ("mpls: support for dead routes")
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In some cases during XenBus disconnect event handling and subsequent
queue resource release there may be some TX handlers active on
other processors. Use RCU in order to synchronize with them.
Signed-off-by: Igor Druzhinin <igor.druzhinin@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski says:
====================
nfp: XDP adjust head support
This series adds support for XDP adjust head. Bulk of the code
is actually just paying technical debt. On reconfiguration request
nfp was allocating new resources separately leaving device running
with the existing set of rings. We used to manage the new resources
in special ring set structures. This set is simply separating the
datapath part of the device structure from the control information
allowing the new datapath structure to be allocated with all new
memory and rings. The swap operation is now greatly simplified.
We also save a lot of parameter passing this way. Hopefully the churn
is worth the negative diffstat.
Support for XDP adjust head is done in a pretty standard way. NFP
is a bit special because it prepends metadata before packet data
so we have to do a bit of memcpying in case XDP will run. We also
luck out a little bit because the fact that we already have prepend
space allocated means that one byte is enough to store the extra XDP
space (256 of standard prepend space is a bit inconvenient since
it would normally require 16bits or boolean with additional shifts).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Support prepending data from XDP. We are already always allocating
some headroom because FW may prepend metadata to packets.
xdp_adjust_head() can be supported by making sure that headroom is
big enough for XDP. In case FW had prepended metadata to the packet,
however, we have to move it out of the way before we call XDP.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
XDP may require us to move metadata to make room for pushing
headers. Track meta data location with a pointer and pass
it explicitly to functions.
While at it validate that meta_len from the descriptor is not
bogus.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rename pkt_off variable to dma_off, it should hold data offset
counting from beginning of DMA mapping. Compute the value only
in XDP context.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
NFP_NET_CFG_RX_OFFSET is 32bit wide, make sure what we read from
there is reasonable for packet headroom. This allows us to store
the rx_offset in a 8bit variable.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of testing if xdp_prog is present store the dma direction
in data path structure.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of passing around sets of rings and their parameters just
store all information in the data path structure.
We will no longer user xchg() on XDP programs when we swap programs
while the traffic is guaranteed not to be flowing. This allows us
to simply assign the entire data path structures instead of copying
field by field.
The optimization to reallocate only the rings on the side (RX/TX)
which has been changed is also removed since it seems like it's not
worth the code complexity.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use xdp_prog member of data path struct to carry the xdp_prog to
alloc/free free functions.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move the mtu member from ring set to data path struct.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use fl_bufsz member of data path struct to carry desired size of
free list entries.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of passing variables around use dp to store number of tx rings
for the stack and number of IRQ vectors.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Make callers of nfp_net_ring_reconfig() pass newly allocated data
path structure. We will gradually make use of that structure
instead of passing parameters around to all the allocation functions.
This commit adds allocation and propagation of new data path struct,
no parameters are converted, yet.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Control BAR pointer is used to unmask interrupts so it should be
in the first cacheline of adapter structure.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move all data path information into a separate structure. This way
we will be able to allocate new data path with all new rings etc.
and swap it in easily.
No functional changes.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Joao Pinto says:
====================
prepare mac operations for multiple queues
As agreed with David Miller, this patch-set is the first of 3 to enable
multiple queues in stmmac.
This first one concentrates on mac operations adding functionalities as:
a) Configuration through DT
b) RX and TX scheduling algorithms programming
b) TX queues weight programming (essential in weightes algorithms)
c) RX enable as DCB or AVB (preparing for future AVB support)
d) Mapping RX queue to DMA channel
e) IRQ treatment prepared for multiple queues
f) Debug dump prepared for multiple queues
g) CBS configuration
In v3 patch-set version I included a new patch to enable CBS configuration
(Patch 9).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds the configuration of the AVB Credit-Based Shaper.
Signed-off-by: Joao Pinto <jpinto@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
