Remove early buffer posting logic from burst receive loop to address
several issues:
- Posting receive descriptors without first posting completion
entries risks overflowing the completion queue.
- Posting receive descriptors without updating rx_raw_prod
creates the possibility that the receive descriptor doorbell
can be written twice with the same value.
- Having this logic in the inner descriptor processing loop
can impact performance.
Fixes: 637e34befd ("net/bnxt: optimize Rx processing")
Fixes: 04067844a3 ("net/bnxt: reduce CQ queue size without aggregation ring")
Cc: stable@dpdk.org
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
When host SW issues a HWRM_RING_FREE for Tx/Rx/AGG ring in HW,
the FW flushes the BDs associated with the ring and performs other
cleanup in the HW. The host software should ideally check for an
indication from the FW indicating this step has been completed
successfully to avoid unexpected errors during cleanup.
The FW issues a HWRM_DONE response to the RING_FREE request on
the corresponding CQ ring. Poll the CQs during cleanup and
ensure the HWRM_FREE command is completed not just based on the
value of valid bit but also the HWRM_DONE response for the ring.
If the HWRM_DONE response is not seen, force the cleanup to
complete just based on the valid bit.
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
bnxt_rx_queue_count_op() incorrectly returns the number of
filled but unprocessed completion queue entries instead of
the number of filled but unprocessed received packet
completions. Fix by properly accounting for the number of
completion ring entries used by the various received packet
completion types.
Fixes: 34c0ba839b ("net/bnxt: fix Rx queue count")
Cc: stable@dpdk.org
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Update copyright year in various files.
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
1. On the new 58818 chips, the RX completion is largely the same except
for the new completion opcode and the stripped VLAN format and
checksum status. Added bnxt_parse_csum_v2(), bnxt_parse_pkt_type_v2()
and bnxt_rx_vlan_v2() to support the new RX completion logic.
2. Disable vector mode RX/TX for 58818 chips for now.
3. The cfa_code format on 58818 chips is different than legacy chips.
So skip cfa_code parsing logic on 58818 chips for now.
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Add the new chip specific TPA v2 logic to bnxt_tpa_start() to fully
support TPA on the new chip.
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
The setting of the mbuf ol_flags field for tunneled packets
should be different depending upon whether DEV_RX_OFFLOAD_OUTER_*
offloads are enabled. Initialize ol_flags mappings based on
the receive offload configuration when the receive ring is
initialized.
Cc: stable@dpdk.org
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Refactor offload flags mapping table to be dynamic and
per-ring instead of static and global.
Cc: stable@dpdk.org
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Change the ring logic so that the index increments
unbounded and mask it only when needed.
Modify the existing macros so that the index is not masked.
Add a new macro RING_IDX() to mask it only when needed.
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Lance Richardson <lance.richardson@broadcom.com>
Refactor code to remove references to Thor.
Instead use P5 as in phase 5 of development cycle since it is applicable
to boards other than Thor as well.
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Lance Richardson <lance.richardson@broadcom.com>
The spinlock was trying to protect scenarios where rx_queue stop/start
could be initiated dynamically. Assigning bnxt_dummy_recv_pkts and
bnxt_dummy_xmit_pkts immediately to avoid concurrent access of mbuf in Rx
and cleanup path should help achieve the same result.
Fixes: 14255b3515 ("net/bnxt: fix queue start/stop operations")
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Rahul Gupta <rahul.gupta@broadcom.com>
The CFA code from mark was stored in the deprecated mbuf field udata64.
It is moved to a dynamic field in order to allow removal of udata64.
Note: the new field has 32 bits, smaller than udata64.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Use table to translate receive descriptor status flags to
rte_mbuf ol_flags values.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Change the type of the software receive mbuf ring from an array
of structures containing an mbuf pointer to an array of pointers
to struct rte_mbuf for consistency with how this ring is currently
used by the vector mode receive function.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Add bnxt vector PMD support using NEON SIMD instructions.
Also update the 20.08 release notes with this information.
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Added code to support Tx/Rx from a VF representor port.
The VF-reps use the RX/TX rings of the Trusted VF/PF.
For each VF-rep, the Trusted VF/PF driver issues a VFR_ALLOC FW cmd that
returns "cfa_code" and "cfa_action" values.
The FW sets up the filter tables in such a way that VF traffic by
default (in absence of other rules) gets punted to the parent function
i.e. either the Trusted VF or the PF.
The cfa_code value in the RX-compl informs the driver of the source VF.
For traffic being transmitted from the VF-rep, the TX BD is tagged with
a cfa_action value that informs the HW to punt it to the corresponding
VF.
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
The current mark handling uses the meta data field of the rxcmp as the
first level check for determining gfid vs lfid. When the meta data is
zero due to only the lowest 16bits of the gfid being set, the cfa code
is incorrectly interpreted as being an lfid. Changing code to look at
meta fmt instead of the meta data directly for the determination.
Fixes: b87abb2e55 ("net/bnxt: support marking packet")
Signed-off-by: Mike Baucom <michael.baucom@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
Certain applications(Ex: OVS-DPDK) can issue rte_flow_create with mark
id set to 0. The mark table entry creation in the driver was assuming
non-zero mark ids. Fix it to have a valid flag for each entry.
Also, it is possible that both RSS flags and cfa_code/ mark id are set
as part of the Rx packet completion descriptor if the 'action' for a
flow is MARK + RSS.
Fix Rx completion processing to look for both fields as opposed to only
either.
Fixes: 94eb699bc8 ("net/bnxt: support flow mark action")
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Add support for RTE_FLOW_ACTION_TYPE_MARK.
Use the flow_id provided by FW during flow creation to lookup the
mark id provided by the application.
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
Add support for LRO for adapters based on Thor (BCM57508).
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Refactor init and uninit functions so that the driver can fail
the eth_dev_ops callbacks and accessing Tx and Rx queues
when device is in reset or in error state.
Transmit and receive queues are freed during reset cleanup and
reallocated during recovery. So we block all data path handling
in this state. The eth_dev dev_started field is updated depending
on the status of the device.
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reduce code duplication and prepare for newer controllers that
use different doorbell protocols by refactoring doorbell handling
code.
Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
1) Use nb_rx_pkts instead of checking producer indices of Rx and
aggregator rings to decide if any Rx completions were processed.
2) Post Rx buffers early in Rx processing instead of waiting for
the budgeted burst quota.
3) Ring Rx CQ DB after Rx buffers are posted.
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
For frames where the hardware is not able to calculate checksum
we are indicating such frames to be bad. And that is incorrect.
Indicate PKT_RX_IP_CKSUM_UNKNOWN or PKT_RX_L4_CKSUM_UNKNOWN
for such frames.
Fixes: 7ec39d8c52 ("net/bnxt: update status of Rx IP/L4 CKSUM")
Cc: stable@dpdk.org
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Fix Rx checksum status for tunnel frames as seen by hardware.
Current code does not handle cases for tunnel frames correctly.
Fixes: 7ec39d8c52 ("net/bnxt: update status of Rx IP/L4 CKSUM")
Cc: stable@dpdk.org
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Update the license header on bnxt files to be the standard
BSD-3-Clause license used for the rest of DPDK,
bring the files in compliance with the DPDK licensing policy.
Signed-off-by: Scott Branden <scott.branden@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Currently this is implemented entirely in the PMD as there is no
explicit support in the HW. Re-program the RSS Table without this queue
on stop and add it back to the table on start.
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Renamed data type from phys_addr_t to rte_iova_t.
Signed-off-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
update ol_flags with the appropriate status of IP/L4 cksum in Rx path.
Fixes: 2eb53b134a ("net/bnxt: add initial Rx code")
Cc: stable@dpdk.org
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
This patch adds support to enable and disable LRO
To support this feature, the driver creates an aggregator ring.
When the hardware starts doing LRO, it sends a tpa_start completion.
When the driver receives a tpa_end completion, it indicates that the
LRO chaining is complete.
Signed-off-by: Steeven Li <steeven.li@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
This patch adds support to modify MTU using the set_mtu dev_op.
To support frames > 2k, the PMD creates an aggregator ring.
When a frame greater than 2k is received, it is fragmented
and the resulting fragments are DMA'ed to the aggregator ring.
Now the driver can support jumbo frames upto 9500 bytes.
Signed-off-by: Steeven Li <steeven.li@broadcom.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Perform allocation and free()ing of ring and information structures
for the TX, RX, and completion rings. The previous patches had
so far provided top level stubs and generic ring support, while this
patch does the real allocation and freeing of the memory specific to
each different type of generic ring.
For example bnxt_init_tx_ring_struct() or bnxt_init_rx_ring_struct() is
now allocating memory based on the socked_id being provided.
bnxt_tx_queue_setup_op() or bnxt_rx_queue_setup_op() have gone through
some reformatting to perform a graceful cleanup in case memory
allocation fails.
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com>
Reviewed-by: David Christensen <david.christensen@broadcom.com>
This patch adds initial implementation of rx_pkt_burst() function for Rx.
bnxt_recv_pkts() is the top level function for doing Rx.
This patch also adds code to allocate rings in the ASIC.
For each Rx queue allocated in the PMD driver, a corresponding ring
in hardware will be created. Every time a frame is received a Rx ring
is selected based on the hardware configuration like RSS, MAC or VLAN,
COS and such. The hardware uses a completion ring to indicate the
availability of a packet.
This patch also brings in functions like bnxt_init_one_rx_ring()
bnxt_init_rx_ring_struct() which initializes various structures before
a Rx can begin.
bnxt_init_rxbds() initializes the Rx Buffer Descriptors while
bnxt_alloc_rx_data() allocates a buffer in the host to receive the
incoming packet.
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com>
Reviewed-by: David Christensen <david.christensen@broadcom.com>
