These are my last couple of days at Microsoft.
Remove the old email from MAINTAINERS.
Will no longer have free access to Azure to work on Netvsc.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Add a new structure for indirect AGE update.
This new structure enables:
1. Update timeout value.
2. Stop AGE checking.
3. Start AGE checking.
4. restart AGE checking.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Ori Kam <orika@nvidia.com>
Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
When application use queue-based flow rule management and operate the
same flow rule on the same queue, e.g create/destroy/query, API of
querying aged flow rules should also have queue id parameter just like
other queue-based flow APIs.
By this way, PMD can work in more optimized way since resources are
isolated by queue and needn't synchronize.
If application do use queue-based flow management but configure port
without RTE_FLOW_PORT_FLAG_STRICT_QUEUE, which means application operate
a given flow rule on different queues, the queue id parameter will
be ignored.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Ori Kam <orika@nvidia.com>
Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
The data-path focused flow rule management can manage flow rules in more
optimized way than traditional one by using hints provided by
application in initialization phase.
In addition to the current hints we have in port attr, more hints could
be provided by application about its behaviour.
One example is how the application do with the same flow rule ?
A. create/destroy flow on same queue but query flow on different queue
or queue-less way (i.e, counter query)
B. All flow operations will be exactly on the same queue, by which PMD
could be in more optimized way then A because resource could be
isolated and access based on queue, without lock, for example.
This patch add flag about above situation and could be extended to cover
more situations.
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Ori Kam <orika@nvidia.com>
Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Based on the hard limits configured in the SA context,
PMD passes corresponding event subtype to the application
to notify hard expiry event
Signed-off-by: Vamsi Attunuru <vattunuru@marvell.com>
Removed unnecessary datapointer(DPTR) update and remove ESN update
from microcode command word 0 based on the latest microcode.
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Signed-off-by: Vidya Sagar Velumuri <vvelumuri@marvell.com>
As per recent change in Linux-5.4.x AF driver, mailbox is updated to
configure mapping between channel and BPID.
Due to mbox mismatch, PFC was broken. Patch syncs mailbox definition
for the same. Also fixes the PFC configuration issues.
Fixes: ff1400aa9d ("net/cnxk: add receive channel backpressure for SDP")
Cc: stable@dpdk.org
Signed-off-by: Sunil Kumar Kori <skori@marvell.com>
Sync nix_hw_info structure with kernel.
Maintain default RR_QUANTUM for VF TL2 same as kernel to make
equal distribution among all VFs.
Signed-off-by: Satha Rao <skoteshwar@marvell.com>
Each TX schedule config mail box supports a maximum 20 register updates.
This patch will send node weight updates in multiple mailboxes when
TM is created with more than 20 scheduler nodes.
Fixes: 464c9f9193 ("common/cnxk: support NIX TM dynamic update")
Cc: stable@dpdk.org
Signed-off-by: Satha Rao <skoteshwar@marvell.com>
RQ mask config needs to enable SPB_ENA in order for zero for
being able to override it with meta AURA.
Also fix flow control config to catch invalid rxchan config
errors.
Fixes: ddf955d391 ("common/cnxk: support CPT second pass")
Fixes: da57d4589a ("common/cnxk: support NIX flow control")
Cc: stable@dpdk.org
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
In outbound inline case, use NIX Tx offset instead of
NIX Tx address for cn10kb as per new instruction format.
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Fix later skip to include mbuf priv data as mbuf->buf_addr
is populated based on calculation including per-mbuf priv area.
Fixes: 706eeae607 ("net/cnxk: add multi-segment Rx for CN10K")
Cc: stable@dpdk.org
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Add support to override soft expiry poll frequency via devargs.
Also provide helper API to indicate reassembly support on a chip
and documentation for devargs that are already present.
Fixes: 780b9c8924 ("net/cnxk: support zero AURA for inline meta")
Signed-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>
Set maximum frame size on SDP NIX side to 16KB for CN93 A0 and B0,
CNF95N A0 and CNF95O A0 SOC type. Rest of the SoCs SDP NIX to 64KB.
Signed-off-by: Sathesh Edara <sedara@marvell.com>
Verify that the message length is non zero and keep processing
virtual channel messages after the event is received.
Fixes: 6bc987ecb8 ("net/iavf: support IPsec inline crypto")
Cc: stable@dpdk.org
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
PMD tx path does not support VXLAN_GPE tunnel offload. Because it does not
process RTE_MBUF_F_TX_TUNNEL_VXLAN_GPE flag in mbuf, and then the "L4TUNT"
field will not be set in Tx context descriptor.
This patch is to add the RTE_MBUF_F_TX_TUNNEL_VXLAN_GPE flag to
support Tx VXLAN_GPE offload under the scenario if the offload tso
and VXLAN_GPE tunnel are both required, so that it would avoid
tx queue overflowing.
Fixes: daa02b5cdd ("mbuf: add namespace to offload flags")
Cc: stable@dpdk.org
Signed-off-by: Mingjin Ye <mingjinx.ye@intel.com>
Tested-by: Ke Xu <ke1.xu@intel.com>
Use large VF queue stop request when large VF is enabled
Fixes: 9cf9c02bf6 ("net/iavf: add enable/disable queues for large VF")
Cc: stable@dpdk.org
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
All callbacks registered for ethdev events are called in
eal-intr-thread, and some of them execute virtchnl commands.
Because interrupts are disabled in the intr thread, no response
will be received for these commands. So all callbacks should
be called in a new context.
When the device is bonded, the bond pmd registers a callback for
the LSC event to execute virtchnl commands to reinitialize the
device, and it would also raise the above issue.
This commit adds a new thread to call all event callbacks.
Fixes: 48de41ca11 ("net/avf: enable link status update")
Fixes: 8410842505 ("net/iavf: support asynchronous virtual channel message")
Cc: stable@dpdk.org
Signed-off-by: Yiding Zhou <yidingx.zhou@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
proto_hdr defines a bit mask of the protocol sequence as RTE_PTYPE_*,
The last RTE_PTYPE* in the mask indicates the split position.
To get the split position from proto_hdr, the order of judgement should
be from inner to outer layer, so for tunneling packets the tunnel header
should be placed at the end of the judgement condition.
Fixes: 629dad3ef3 ("net/ice: support buffer split in scalar Rx")
Signed-off-by: Yuan Wang <yuanx.wang@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Replace mempool_put with memset 0, the internal session memory block
is no longer allocated from a mempool
Fixes: 3f3fc3308b ("security: remove private mempool usage")
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Akhil Goyal <gakhil@marvell.com>
Replace mempool_put with memset 0, the internal session memory block
is no longer allocated from a mempool
Fixes: 3f3fc3308b ("security: remove private mempool usage")
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Akhil Goyal <gakhil@marvell.com>
Fix issue with final word being dropped when retrieving module EEPROM
data for the ice driver.
Take for simplicity the case when `info->offset` is zero and `info->len`
is equal to `SFF_READ_BLOCK_SIZE`. In this case, memcpy would not be
called despite there presumably being room in the buffer
(as we have requested 8 bytes of data and the memcpy would write
precisely 8 bytes). The same edge case will be hit on the final
iteration of the for loop whenever a multiple of 8 bytes is requested,
as the final word will not be written to the data buffer.
Fixes: 5256925658 ("net/ice: support module EEPROM")
Cc: stable@dpdk.org
Signed-off-by: David MacDougal <david.macdougal@gmail.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Return error if SPI from the flow spec doesn't match
the one from the crypto session.
Fixes: 6bc987ecb8 ("net/iavf: support IPsec inline crypto")
Cc: stable@dpdk.org
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
This patch adds per queue offload check to turn off Rx timestamp
offload by default.
Fixes: b5cd735132 ("net/iavf: enable Rx timestamp on flex descriptor")
Cc: stable@dpdk.org
Signed-off-by: Simei Su <simei.su@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Add Protocol Agnostic Flow (raw flow) support for flow subscription
in AVF.
For example, testpmd creates a flow subscription raw packet rule:
rule: eth + ipv4 src is 1.1.1.1 dst is 2.2.2.2
cmd: flow create 0 ingress pattern raw pattern spec \
00000000000000000000000008004500001400000000000000000101010102020202 \
pattern mask \
0000000000000000000000000000000000000000000000000000FFFFFFFFFFFFFFFF \
/ end actions port_representor port_id 0 / end
Signed-off-by: Jie Wang <jie1x.wang@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Fix an error check where the return code was assigned to a
unsigned integer which can hide negative error codes.
Fixes: 6bc987ecb8 ("net/iavf: support IPsec inline crypto")
Cc: stable@dpdk.org
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
In case rte_eth_dma_zone_reserve fails in ice_tx_queue_setup
ice_tx_queue_release is called on 0 allocated but not initialized
txq struct.
This may happen on ENOMEM condition, size exhaustion of
memconfig->memzones array as well as some others.
Fixes: edec6dd838 ("net/ice: remove redundant functions")
Cc: stable@dpdk.org
Signed-off-by: Tomasz Jonak <tomasz@graphiant.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Passing tainted expression "rss_meta->proto_hdrs.count" to
"iavf_refine_proto_hdrs", which uses it as a loop boundary.
Replace tainted expression with a temp variable to avoid the
tainted scalar coverity warning.
Coverity issue: 381131
Fixes: f30157d988 ("net/iavf: support PPPoL2TPv2oUDP RSS Hash")
Cc: stable@dpdk.org
Signed-off-by: Steve Yang <stevex.yang@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
When the promiscuous mode is enabled on a VF, the IXGBE_VMOLR_VPE
bit (VLAN Promiscuous Enable) is set. This means that the VF will
receive packets whose VLAN is not the same as the VLAN of the VF.
For instance, in this situation:
┌────────┐ ┌────────┐ ┌────────┐
│ │ │ │ │ │
│ │ │ │ │ │
│ VF0├────┤VF1 VF2├────┤VF3 │
│ │ │ │ │ │
└────────┘ └────────┘ └────────┘
VM1 VM2 VM3
vf 0: vlan 1000
vf 1: vlan 1000
vf 2: vlan 1001
vf 3: vlan 1001
If we tcpdump on VF3, we see all the packets, even those transmitted
on vlan 1000.
This behavior prevents to bridge VF1 and VF2 in VM2, because it will
create a loop: packets transmitted on VF1 will be received by VF2 and
vice-versa, and bridged again through the software bridge.
This patch remove the activation of VLAN Promiscuous when a VF enables
the promiscuous mode. However, the IXGBE_VMOLR_UPE bit (Unicast
Promiscuous) is kept, so that a VF receives all packets that has the
same VLAN, whatever the destination MAC address.
A similar patch was accepted in Linux kernel (see link).
Link: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=7bb0fb7c63df
Fixes: 0355c379b7 ("net/ixgbe: support VF promiscuous by PF driver")
Cc: stable@dpdk.org
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Wenjun Wu <wenjun1.wu@intel.com>
After a VF requested to remove the promiscuous flag on an interface, the
broadcast packets are not received anymore. This breaks some protocols
like ARP.
In ixgbe_update_vf_xcast_mode(), we should keep the IXGBE_VMOLR_BAM
bit (Broadcast Accept) on promiscuous removal. This flag is already set
by default in ixgbe_vf_reset_event() on VF reset.
A similar patch was accepted in Linux kernel (see link).
Link: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=803e9895ea2b
Fixes: 0355c379b7 ("net/ixgbe: support VF promiscuous by PF driver")
Cc: stable@dpdk.org
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Wenjun Wu <wenjun1.wu@intel.com>
Config load functions updated to support 100G rates for subport and pipes.
Added new parse function to convert string to unsigned long long.
Added error checks.
Fixed format warnings.
Signed-off-by: Megha Ajmera <megha.ajmera@intel.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
This library has no maintainer and, for now, nobody expressed interest
in taking over.
Mark this experimental library as deprecated and announce plan for
removal in v23.11.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Ferruh Yigit <ferruh.yigit@amd.com>
Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
Bernard is no longer with Intel and is no longer involved in the DPDK
community, so remove him from the MAINTAINERS file.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>
When adding buffer split feature to mlx in DPDK 20.11,
it has been forgotten to fill the feature matrix.
Fixes: 6c8f7f1c18 ("net/mlx5: report Rx buffer split capabilities")
Cc: stable@dpdk.org
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Raslan Darawsheh <rasland@nvidia.com>
The "mlx5_os_parse_eth_devargs()" function parses the ETH devargs into a
specific structure called "eth_da".
It gets structure called "devargs" as a member of EAL device containing
the relevant information.
When "devargs" structure is invalid, the function avoids parsing it.
However, when it valid but its field "args" is invalid, the function
tries to parse it and dereference to NULL pointer.
This patch adds check to avoid this NULL dereferencing.
Fixes: 919488fbfa ("net/mlx5: support Sub-Function")
Cc: stable@dpdk.org
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
The common MLX5 probe function parses first the devargs and save them in
a dictionary.
It gets structure called "devargs" as a member of EAL device containing
all needed information.
When "devargs" structure is invalid, the function avoids parsing it.
However, when it valid but its field "args" is invalid, the function
tries to parse it and dereference to NULL pointer.
This patch adds check to avoid this NULL dereferencing.
Fixes: a729d2f093 ("common/mlx5: refactor devargs management")
Cc: stable@dpdk.org
Signed-off-by: Michael Baum <michaelba@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>
For the flows with multiple tunnel layers and containing
tunnel decap and modify actions, for example:
... / vxlan / eth / ipv4 proto is 4 / end
actions raw_decap / modify_field / ...
(note: proto 4 means we have the IP-over-IP tunnel in VXLAN payload)
We have added the multiple tunnel layers validation rejecting
the flows like above mentioned one.
The hardware supports the above match combination till the inner
IP-over-IP header (not including the last one), both for IP-over-IPv4
and IP-over-IPv6, so we should not blindly reject. Also, for the modify
actions following the decap we should set the layer attributes correctly.
This patch reverts the below code changes to support the match, and
adjusts the layers update in case of decap with outer tunnel header.
Fixes: fa06906a48 ("net/mlx5: fix IPIP multi-tunnel validation")
Cc: stable@dpdk.org
Signed-off-by: Jiawei Wang <jiaweiw@nvidia.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
Add support for port_representor item, it will match on traffic
originated from representor port specified in the pattern. This item
is supported in FDB steering domain only (in the flow with transfer
attribute).
For example, below flow will redirect the destination of traffic from
ethdev 1 to ethdev 2.
testpmd> ... pattern eth / port_representor port_id is 1 / end actions
represented_port ethdev_port_id 2 / ...
To handle abovementioned item, Tx queue matching is added in the driver,
and the flow will be expanded to number of the Tx queues. If the spec of
port_representor is NULL, the flow will not be expanded and match on
traffic from any representor port.
Signed-off-by: Sean Zhang <xiazhang@nvidia.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
This patch adds the creation of control flow rules required to receive
default traffic (based on port configuration) with HWS.
Control flow rules are created on port start and destroyed on port stop.
Handling of destroying these rules was already implemented before that
patch.
Control flow rules are created if and only if flow isolation mode is
disabled and the creation process goes as follows:
- Port configuration is collected into a set of flags. Each flag
corresponds to a certain Ethernet pattern type, defined by
mlx5_flow_ctrl_rx_eth_pattern_type enumeration. There is a separate
flag for VLAN filtering.
- For each possible Ethernet pattern type and:
- For each possible RSS action configuration:
- If configuration flags do not match this combination, it is
omitted.
- A template table is created using this combination of pattern
and actions template (templates are fetched from hw_ctrl_rx
struct stored in the port's private data).
- Flow rules are created in this table.
Signed-off-by: Dariusz Sosnowski <dsosnowski@nvidia.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
