d/numam-dpdk
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
4f28d9a1a6
numam-dpdk/drivers/net/bnxt/bnxt_ring.c

843 lines
22 KiB
C
Raw Normal View History

net/bnxt: convert to SPDX license tag 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>
2018-04-02 22:34:32 +00:00
/* SPDX-License-Identifier: BSD-3-Clause
net/bnxt: update copyright year 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>
2021-01-27 16:09:43 +00:00
* Copyright(c) 2014-2021 Broadcom
net/bnxt: convert to SPDX license tag 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>
2018-04-02 22:34:32 +00:00
* All rights reserved.
net/bnxt: declare generic ring structs Declare generic ring structures and a free() function. These are generic ring management functions which will be used to create Tx, Rx and Completion rings in the subsequent patches, and tie them to the HWRM managed ring resources. This generic ring structure is shared all the ring types and tracks the the host Buffer Descriptors (BDs) and the HWRM assigned ID. 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>
2016-06-15 21:23:07 +00:00
*/
net/bnxt: support LRO 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>
2017-06-01 17:07:10 +00:00
#include <rte_bitmap.h>
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
#include <rte_memzone.h>
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
#include <rte_malloc.h>
net/bnxt: support lack of huge pages rte_malloc_virt2phy() does not return a physical address if huge pages aren't in use. Further, rte_memzone->phys_addr is not a physical address. Use rte_mem_virt2phy() and manually lock pages to support lack of huge pages. Also check the return value of rte_mem_virt2phy() Verify the function returns an address. Otherwise return an error and log a message. Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2017-06-01 17:07:01 +00:00
#include <unistd.h>
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
net/bnxt: declare generic ring structs Declare generic ring structures and a free() function. These are generic ring management functions which will be used to create Tx, Rx and Completion rings in the subsequent patches, and tie them to the HWRM managed ring resources. This generic ring structure is shared all the ring types and tracks the the host Buffer Descriptors (BDs) and the HWRM assigned ID. 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>
2016-06-15 21:23:07 +00:00
#include "bnxt.h"
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
#include "bnxt_hwrm.h"
net/bnxt: declare generic ring structs Declare generic ring structures and a free() function. These are generic ring management functions which will be used to create Tx, Rx and Completion rings in the subsequent patches, and tie them to the HWRM managed ring resources. This generic ring structure is shared all the ring types and tracks the the host Buffer Descriptors (BDs) and the HWRM assigned ID. 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>
2016-06-15 21:23:07 +00:00
#include "bnxt_ring.h"
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
#include "bnxt_rxq.h"
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
#include "bnxt_rxr.h"
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
#include "bnxt_txq.h"
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
#include "bnxt_txr.h"
#include "hsi_struct_def_dpdk.h"
net/bnxt: declare generic ring structs Declare generic ring structures and a free() function. These are generic ring management functions which will be used to create Tx, Rx and Completion rings in the subsequent patches, and tie them to the HWRM managed ring resources. This generic ring structure is shared all the ring types and tracks the the host Buffer Descriptors (BDs) and the HWRM assigned ID. 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>
2016-06-15 21:23:07 +00:00
/*
* Generic ring handling
*/
void bnxt_free_ring(struct bnxt_ring *ring)
{
net/bnxt: validate structs and pointers before use Validate pointers aren't pointing to uninitialized areas including txq and rxq before using them to avoid bnxt driver from crashing. Signed-off-by: Rahul Gupta <rahul.gupta@broadcom.com> Signed-off-by: Jay Ding <jay.ding@broadcom.com> Signed-off-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Tested-by: Randy Schacher <stuart.schacher@broadcom.com>
2018-05-01 01:06:09 +00:00
if (!ring)
return;
net/bnxt: declare generic ring structs Declare generic ring structures and a free() function. These are generic ring management functions which will be used to create Tx, Rx and Completion rings in the subsequent patches, and tie them to the HWRM managed ring resources. This generic ring structure is shared all the ring types and tracks the the host Buffer Descriptors (BDs) and the HWRM assigned ID. 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>
2016-06-15 21:23:07 +00:00
if (ring->vmem_size && *ring->vmem) {
memset((char *)*ring->vmem, 0, ring->vmem_size);
*ring->vmem = NULL;
}
net/bnxt: avoid freeing memzone multiple times Since we are storing the mem_zone address for each ring created, we are freeing the same address multiple times. For example the memory zone created for Rx is being freed during Rx ring cleanup, AGG ring cleanup and CQ cleanup. Avoid this by storing the memory zone address in RXQ instead and free it as a part of queue_release dev_op. In the same way do the same for TX queues as well. Fixes: 51c87ebafc7d ("net/bnxt: add Tx queue create/destroy") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-04-17 01:11:25 +00:00
ring->mem_zone = NULL;
net/bnxt: declare generic ring structs Declare generic ring structures and a free() function. These are generic ring management functions which will be used to create Tx, Rx and Completion rings in the subsequent patches, and tie them to the HWRM managed ring resources. This generic ring structure is shared all the ring types and tracks the the host Buffer Descriptors (BDs) and the HWRM assigned ID. 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>
2016-06-15 21:23:07 +00:00
}
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
/*
* Ring groups
*/
net/bnxt: fix initialization During initialization sequence in bnxt_alloc_mem() if bnxt_alloc_async_cp_ring() fails, driver invokes bnxt_free_mem() which in turn call bnxt_free_vnic_mem() which causes the error logs "bnxt_free_vnic_mem(): VNIC is not freed yet!". This is because vnic memory is allocated but not initialized yet. Fix this by moving bnxt_init_vnics() from bnxt_init_nic() to bnxt_alloc_vnic_mem(). Also get rid of bnxt_init_nic() as the initialization is done in respective functions now. Fixes: 1bf01f5135f8 ("net/bnxt: prevent device access when device is in reset") Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-11-04 10:02:39 +00:00
static void bnxt_init_ring_grps(struct bnxt *bp)
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
{
unsigned int i;
for (i = 0; i < bp->max_ring_grps; i++)
memset(&bp->grp_info[i], (uint8_t)HWRM_NA_SIGNATURE,
sizeof(struct bnxt_ring_grp_info));
}
net/bnxt: prevent device access when device is in reset 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>
2019-10-02 01:23:22 +00:00
int bnxt_alloc_ring_grps(struct bnxt *bp)
{
if (bp->max_tx_rings == 0) {
PMD_DRV_LOG(ERR, "No TX rings available!\n");
return -EBUSY;
}
/* THOR does not support ring groups.
* But we will use the array to save RSS context IDs.
*/
net/bnxt: remove references to Thor 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>
2020-12-01 19:15:23 +00:00
if (BNXT_CHIP_P5(bp)) {
bp->max_ring_grps = BNXT_MAX_RSS_CTXTS_P5;
net/bnxt: prevent device access when device is in reset 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>
2019-10-02 01:23:22 +00:00
} else if (bp->max_ring_grps < bp->rx_cp_nr_rings) {
/* 1 ring is for default completion ring */
PMD_DRV_LOG(ERR, "Insufficient resource: Ring Group\n");
return -ENOSPC;
}
if (BNXT_HAS_RING_GRPS(bp)) {
bp->grp_info = rte_zmalloc("bnxt_grp_info",
sizeof(*bp->grp_info) *
bp->max_ring_grps, 0);
if (!bp->grp_info) {
PMD_DRV_LOG(ERR,
"Failed to alloc grp info tbl.\n");
return -ENOMEM;
}
net/bnxt: fix initialization During initialization sequence in bnxt_alloc_mem() if bnxt_alloc_async_cp_ring() fails, driver invokes bnxt_free_mem() which in turn call bnxt_free_vnic_mem() which causes the error logs "bnxt_free_vnic_mem(): VNIC is not freed yet!". This is because vnic memory is allocated but not initialized yet. Fix this by moving bnxt_init_vnics() from bnxt_init_nic() to bnxt_alloc_vnic_mem(). Also get rid of bnxt_init_nic() as the initialization is done in respective functions now. Fixes: 1bf01f5135f8 ("net/bnxt: prevent device access when device is in reset") Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-11-04 10:02:39 +00:00
bnxt_init_ring_grps(bp);
net/bnxt: prevent device access when device is in reset 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>
2019-10-02 01:23:22 +00:00
}
return 0;
}
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
/*
* Allocates a completion ring with vmem and stats optionally also allocating
* a TX and/or RX ring. Passing NULL as tx_ring_info and/or rx_ring_info
* to not allocate them.
*
* Order in the allocation is:
* stats - Always non-zero length
* cp vmem - Always zero-length, supported for the bnxt_ring abstraction
* tx vmem - Only non-zero length if tx_ring_info is not NULL
* rx vmem - Only non-zero length if rx_ring_info is not NULL
* cp bd ring - Always non-zero length
* tx bd ring - Only non-zero length if tx_ring_info is not NULL
* rx bd ring - Only non-zero length if rx_ring_info is not NULL
*/
int bnxt_alloc_rings(struct bnxt *bp, uint16_t qidx,
net/bnxt: avoid freeing memzone multiple times Since we are storing the mem_zone address for each ring created, we are freeing the same address multiple times. For example the memory zone created for Rx is being freed during Rx ring cleanup, AGG ring cleanup and CQ cleanup. Avoid this by storing the memory zone address in RXQ instead and free it as a part of queue_release dev_op. In the same way do the same for TX queues as well. Fixes: 51c87ebafc7d ("net/bnxt: add Tx queue create/destroy") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-04-17 01:11:25 +00:00
struct bnxt_tx_queue *txq,
struct bnxt_rx_queue *rxq,
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
struct bnxt_cp_ring_info *cp_ring_info,
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
struct bnxt_cp_ring_info *nq_ring_info,
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
const char *suffix)
{
struct bnxt_ring *cp_ring = cp_ring_info->cp_ring_struct;
net/bnxt: avoid freeing memzone multiple times Since we are storing the mem_zone address for each ring created, we are freeing the same address multiple times. For example the memory zone created for Rx is being freed during Rx ring cleanup, AGG ring cleanup and CQ cleanup. Avoid this by storing the memory zone address in RXQ instead and free it as a part of queue_release dev_op. In the same way do the same for TX queues as well. Fixes: 51c87ebafc7d ("net/bnxt: add Tx queue create/destroy") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-04-17 01:11:25 +00:00
struct bnxt_rx_ring_info *rx_ring_info = rxq ? rxq->rx_ring : NULL;
struct bnxt_tx_ring_info *tx_ring_info = txq ? txq->tx_ring : NULL;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
struct bnxt_ring *tx_ring;
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
struct bnxt_ring *rx_ring;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
struct rte_pci_device *pdev = bp->pdev;
net/bnxt: switch to the new offload API Update bnxt PMD to new ethdev offloads API. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-04-07 17:40:55 +00:00
uint64_t rx_offloads = bp->eth_dev->data->dev_conf.rxmode.offloads;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
const struct rte_memzone *mz = NULL;
char mz_name[RTE_MEMZONE_NAMESIZE];
drivers/net: rename physical address type to IOVA 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>
2017-10-20 12:31:31 +00:00
rte_iova_t mz_phys_addr;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
int stats_len = (tx_ring_info || rx_ring_info) ?
net/bnxt: update HWRM to version 1.9.2 Update HWRM structures to version 1.9.2 Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-05-03 17:23:55 +00:00
RTE_CACHE_LINE_ROUNDUP(sizeof(struct hwrm_stat_ctx_query_output) -
sizeof (struct hwrm_resp_hdr)) : 0;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
stats_len = RTE_ALIGN(stats_len, 128);
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
int cp_vmem_start = stats_len;
int cp_vmem_len = RTE_CACHE_LINE_ROUNDUP(cp_ring->vmem_size);
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
cp_vmem_len = RTE_ALIGN(cp_vmem_len, 128);
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
int nq_vmem_len = nq_ring_info ?
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
RTE_CACHE_LINE_ROUNDUP(cp_ring->vmem_size) : 0;
nq_vmem_len = RTE_ALIGN(nq_vmem_len, 128);
int nq_vmem_start = cp_vmem_start + cp_vmem_len;
int tx_vmem_start = nq_vmem_start + nq_vmem_len;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
int tx_vmem_len =
tx_ring_info ? RTE_CACHE_LINE_ROUNDUP(tx_ring_info->
tx_ring_struct->vmem_size) : 0;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
tx_vmem_len = RTE_ALIGN(tx_vmem_len, 128);
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
int rx_vmem_start = tx_vmem_start + tx_vmem_len;
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
int rx_vmem_len = rx_ring_info ?
RTE_CACHE_LINE_ROUNDUP(rx_ring_info->
rx_ring_struct->vmem_size) : 0;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
rx_vmem_len = RTE_ALIGN(rx_vmem_len, 128);
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
int ag_vmem_start = 0;
int ag_vmem_len = 0;
int cp_ring_start = 0;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
int nq_ring_start = 0;
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
ag_vmem_start = rx_vmem_start + rx_vmem_len;
ag_vmem_len = rx_ring_info ? RTE_CACHE_LINE_ROUNDUP(
rx_ring_info->ag_ring_struct->vmem_size) : 0;
cp_ring_start = ag_vmem_start + ag_vmem_len;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
cp_ring_start = RTE_ALIGN(cp_ring_start, 4096);
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
int cp_ring_len = RTE_CACHE_LINE_ROUNDUP(cp_ring->ring_size *
sizeof(struct cmpl_base));
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
cp_ring_len = RTE_ALIGN(cp_ring_len, 128);
nq_ring_start = cp_ring_start + cp_ring_len;
nq_ring_start = RTE_ALIGN(nq_ring_start, 4096);
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
int nq_ring_len = nq_ring_info ? cp_ring_len : 0;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
int tx_ring_start = nq_ring_start + nq_ring_len;
net/bnxt: fix ring alignment for Thor-based adapters When using transmit/receive queue sizes smaller than 256, alignment requirements are not being met for Thor-based adapters. Fix by forcing memory addresses used for transmit/receive/aggregation ring allocations to be on 4K boundaries. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Cc: stable@dpdk.org Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:56 +00:00
tx_ring_start = RTE_ALIGN(tx_ring_start, 4096);
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
int tx_ring_len = tx_ring_info ?
RTE_CACHE_LINE_ROUNDUP(tx_ring_info->tx_ring_struct->ring_size *
sizeof(struct tx_bd_long)) : 0;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
tx_ring_len = RTE_ALIGN(tx_ring_len, 4096);
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
int rx_ring_start = tx_ring_start + tx_ring_len;
net/bnxt: fix ring alignment for Thor-based adapters When using transmit/receive queue sizes smaller than 256, alignment requirements are not being met for Thor-based adapters. Fix by forcing memory addresses used for transmit/receive/aggregation ring allocations to be on 4K boundaries. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Cc: stable@dpdk.org Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:56 +00:00
rx_ring_start = RTE_ALIGN(rx_ring_start, 4096);
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
int rx_ring_len = rx_ring_info ?
RTE_CACHE_LINE_ROUNDUP(rx_ring_info->rx_ring_struct->ring_size *
sizeof(struct rx_prod_pkt_bd)) : 0;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
rx_ring_len = RTE_ALIGN(rx_ring_len, 4096);
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
net/bnxt: support LRO 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>
2017-06-01 17:07:10 +00:00
int ag_ring_start = rx_ring_start + rx_ring_len;
net/bnxt: fix ring alignment for Thor-based adapters When using transmit/receive queue sizes smaller than 256, alignment requirements are not being met for Thor-based adapters. Fix by forcing memory addresses used for transmit/receive/aggregation ring allocations to be on 4K boundaries. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Cc: stable@dpdk.org Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:56 +00:00
ag_ring_start = RTE_ALIGN(ag_ring_start, 4096);
net/bnxt: support LRO 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>
2017-06-01 17:07:10 +00:00
int ag_ring_len = rx_ring_len * AGG_RING_SIZE_FACTOR;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
ag_ring_len = RTE_ALIGN(ag_ring_len, 4096);
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
net/bnxt: support LRO 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>
2017-06-01 17:07:10 +00:00
int ag_bitmap_start = ag_ring_start + ag_ring_len;
int ag_bitmap_len = rx_ring_info ?
RTE_CACHE_LINE_ROUNDUP(rte_bitmap_get_memory_footprint(
rx_ring_info->rx_ring_struct->ring_size *
AGG_RING_SIZE_FACTOR)) : 0;
int tpa_info_start = ag_bitmap_start + ag_bitmap_len;
net/bnxt: support LRO on Thor adapters 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>
2019-10-04 03:48:57 +00:00
int tpa_info_len = 0;
if (rx_ring_info && (rx_offloads & DEV_RX_OFFLOAD_TCP_LRO)) {
int tpa_max = BNXT_TPA_MAX_AGGS(bp);
tpa_info_len = tpa_max * sizeof(struct bnxt_tpa_info);
tpa_info_len = RTE_CACHE_LINE_ROUNDUP(tpa_info_len);
}
net/bnxt: support LRO 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>
2017-06-01 17:07:10 +00:00
int total_alloc_len = tpa_info_start;
net/bnxt: support LRO on Thor adapters 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>
2019-10-04 03:48:57 +00:00
total_alloc_len += tpa_info_len;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
snprintf(mz_name, RTE_MEMZONE_NAMESIZE,
net/bnxt: use macro for PCI log format Use PCI_PRI_FMT instead of "%04x:%02x:%02x:%02x" print format. Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2020-04-21 21:33:44 +00:00
"bnxt_" PCI_PRI_FMT "-%04x_%s", pdev->addr.domain,
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
pdev->addr.bus, pdev->addr.devid, pdev->addr.function, qidx,
suffix);
mz_name[RTE_MEMZONE_NAMESIZE - 1] = 0;
mz = rte_memzone_lookup(mz_name);
if (!mz) {
net/bnxt: support lack of huge pages rte_malloc_virt2phy() does not return a physical address if huge pages aren't in use. Further, rte_memzone->phys_addr is not a physical address. Use rte_mem_virt2phy() and manually lock pages to support lack of huge pages. Also check the return value of rte_mem_virt2phy() Verify the function returns an address. Otherwise return an error and log a message. Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2017-06-01 17:07:01 +00:00
mz = rte_memzone_reserve_aligned(mz_name, total_alloc_len,
net/bnxt: use contiguous allocation for DMA memory All hardware drivers should allocate IOVA-contiguous memzones for their hardware resources. Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com> Tested-by: Santosh Shukla <santosh.shukla@caviumnetworks.com> Tested-by: Hemant Agrawal <hemant.agrawal@nxp.com> Tested-by: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
2018-04-11 12:29:51 +00:00
SOCKET_ID_ANY,
RTE_MEMZONE_2MB |
RTE_MEMZONE_SIZE_HINT_ONLY |
RTE_MEMZONE_IOVA_CONTIG,
getpagesize());
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
if (mz == NULL)
return -ENOMEM;
}
memset(mz->addr, 0, mz->len);
memzone: rename address from physical to IOVA The struct rte_memzone field .phys_addr is renamed to .iova. The deprecated name is kept in an anonymous union to avoid breaking the API. Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
2017-11-04 01:22:28 +00:00
mz_phys_addr = mz->iova;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
if (tx_ring_info) {
net/bnxt: avoid freeing memzone multiple times Since we are storing the mem_zone address for each ring created, we are freeing the same address multiple times. For example the memory zone created for Rx is being freed during Rx ring cleanup, AGG ring cleanup and CQ cleanup. Avoid this by storing the memory zone address in RXQ instead and free it as a part of queue_release dev_op. In the same way do the same for TX queues as well. Fixes: 51c87ebafc7d ("net/bnxt: add Tx queue create/destroy") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-04-17 01:11:25 +00:00
txq->mz = mz;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
tx_ring = tx_ring_info->tx_ring_struct;
tx_ring->bd = ((char *)mz->addr + tx_ring_start);
tx_ring_info->tx_desc_ring = (struct tx_bd_long *)tx_ring->bd;
net/bnxt: support lack of huge pages rte_malloc_virt2phy() does not return a physical address if huge pages aren't in use. Further, rte_memzone->phys_addr is not a physical address. Use rte_mem_virt2phy() and manually lock pages to support lack of huge pages. Also check the return value of rte_mem_virt2phy() Verify the function returns an address. Otherwise return an error and log a message. Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2017-06-01 17:07:01 +00:00
tx_ring->bd_dma = mz_phys_addr + tx_ring_start;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
tx_ring_info->tx_desc_mapping = tx_ring->bd_dma;
net/bnxt: allocate Rx/Tx and completion rings 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>
2016-06-15 21:23:15 +00:00
tx_ring->mem_zone = (const void *)mz;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
if (!tx_ring->bd)
return -ENOMEM;
if (tx_ring->vmem_size) {
tx_ring->vmem =
(void **)((char *)mz->addr + tx_vmem_start);
tx_ring_info->tx_buf_ring =
net/bnxt: optimize Tx completion handling Avoid copying mbuf pointers to separate array for bulk mbuf free when handling transmit completions for vector mode transmit. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2021-03-06 15:19:11 +00:00
(struct rte_mbuf **)tx_ring->vmem;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
}
}
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
if (rx_ring_info) {
net/bnxt: avoid freeing memzone multiple times Since we are storing the mem_zone address for each ring created, we are freeing the same address multiple times. For example the memory zone created for Rx is being freed during Rx ring cleanup, AGG ring cleanup and CQ cleanup. Avoid this by storing the memory zone address in RXQ instead and free it as a part of queue_release dev_op. In the same way do the same for TX queues as well. Fixes: 51c87ebafc7d ("net/bnxt: add Tx queue create/destroy") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-04-17 01:11:25 +00:00
rxq->mz = mz;
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
rx_ring = rx_ring_info->rx_ring_struct;
rx_ring->bd = ((char *)mz->addr + rx_ring_start);
rx_ring_info->rx_desc_ring =
(struct rx_prod_pkt_bd *)rx_ring->bd;
net/bnxt: support lack of huge pages rte_malloc_virt2phy() does not return a physical address if huge pages aren't in use. Further, rte_memzone->phys_addr is not a physical address. Use rte_mem_virt2phy() and manually lock pages to support lack of huge pages. Also check the return value of rte_mem_virt2phy() Verify the function returns an address. Otherwise return an error and log a message. Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2017-06-01 17:07:01 +00:00
rx_ring->bd_dma = mz_phys_addr + rx_ring_start;
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
rx_ring_info->rx_desc_mapping = rx_ring->bd_dma;
net/bnxt: allocate Rx/Tx and completion rings 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>
2016-06-15 21:23:15 +00:00
rx_ring->mem_zone = (const void *)mz;
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
if (!rx_ring->bd)
return -ENOMEM;
if (rx_ring->vmem_size) {
rx_ring->vmem =
(void **)((char *)mz->addr + rx_vmem_start);
rx_ring_info->rx_buf_ring =
net/bnxt: use appropriate type for Rx ring 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>
2020-09-09 15:52:56 +00:00
(struct rte_mbuf **)rx_ring->vmem;
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
}
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
rx_ring = rx_ring_info->ag_ring_struct;
rx_ring->bd = ((char *)mz->addr + ag_ring_start);
rx_ring_info->ag_desc_ring =
(struct rx_prod_pkt_bd *)rx_ring->bd;
memzone: rename address from physical to IOVA The struct rte_memzone field .phys_addr is renamed to .iova. The deprecated name is kept in an anonymous union to avoid breaking the API. Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
2017-11-04 01:22:28 +00:00
rx_ring->bd_dma = mz->iova + ag_ring_start;
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
rx_ring_info->ag_desc_mapping = rx_ring->bd_dma;
rx_ring->mem_zone = (const void *)mz;
if (!rx_ring->bd)
return -ENOMEM;
if (rx_ring->vmem_size) {
rx_ring->vmem =
(void **)((char *)mz->addr + ag_vmem_start);
rx_ring_info->ag_buf_ring =
net/bnxt: use appropriate type for Rx ring 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>
2020-09-09 15:52:56 +00:00
(struct rte_mbuf **)rx_ring->vmem;
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
}
net/bnxt: support LRO 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>
2017-06-01 17:07:10 +00:00
rx_ring_info->ag_bitmap =
rte_bitmap_init(rx_ring_info->rx_ring_struct->ring_size *
AGG_RING_SIZE_FACTOR, (uint8_t *)mz->addr +
ag_bitmap_start, ag_bitmap_len);
/* TPA info */
net/bnxt: switch to the new offload API Update bnxt PMD to new ethdev offloads API. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-04-07 17:40:55 +00:00
if (rx_offloads & DEV_RX_OFFLOAD_TCP_LRO)
net/bnxt: support LRO 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>
2017-06-01 17:07:10 +00:00
rx_ring_info->tpa_info =
((struct bnxt_tpa_info *)((char *)mz->addr +
tpa_info_start));
net/bnxt: add initial Rx code 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>
2016-06-15 21:23:14 +00:00
}
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
cp_ring->bd = ((char *)mz->addr + cp_ring_start);
net/bnxt: support lack of huge pages rte_malloc_virt2phy() does not return a physical address if huge pages aren't in use. Further, rte_memzone->phys_addr is not a physical address. Use rte_mem_virt2phy() and manually lock pages to support lack of huge pages. Also check the return value of rte_mem_virt2phy() Verify the function returns an address. Otherwise return an error and log a message. Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2017-06-01 17:07:01 +00:00
cp_ring->bd_dma = mz_phys_addr + cp_ring_start;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
cp_ring_info->cp_desc_ring = cp_ring->bd;
cp_ring_info->cp_desc_mapping = cp_ring->bd_dma;
net/bnxt: allocate Rx/Tx and completion rings 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>
2016-06-15 21:23:15 +00:00
cp_ring->mem_zone = (const void *)mz;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
if (!cp_ring->bd)
return -ENOMEM;
if (cp_ring->vmem_size)
*cp_ring->vmem = ((char *)mz->addr + stats_len);
if (stats_len) {
cp_ring_info->hw_stats = mz->addr;
net/bnxt: support lack of huge pages rte_malloc_virt2phy() does not return a physical address if huge pages aren't in use. Further, rte_memzone->phys_addr is not a physical address. Use rte_mem_virt2phy() and manually lock pages to support lack of huge pages. Also check the return value of rte_mem_virt2phy() Verify the function returns an address. Otherwise return an error and log a message. Signed-off-by: Stephen Hurd <stephen.hurd@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2017-06-01 17:07:01 +00:00
cp_ring_info->hw_stats_map = mz_phys_addr;
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
}
cp_ring_info->hw_stats_ctx_id = HWRM_NA_SIGNATURE;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
net/bnxt: avoid null pointer dereference Avoid null pointer dereference when allocating an insulated completion ring by basing nq ring allocation on whether an nq ring was requested instead of whether the device supports nq rings. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-17 10:41:34 +00:00
if (nq_ring_info) {
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
struct bnxt_ring *nq_ring = nq_ring_info->cp_ring_struct;
nq_ring->bd = (char *)mz->addr + nq_ring_start;
nq_ring->bd_dma = mz_phys_addr + nq_ring_start;
nq_ring_info->cp_desc_ring = nq_ring->bd;
nq_ring_info->cp_desc_mapping = nq_ring->bd_dma;
nq_ring->mem_zone = (const void *)mz;
if (!nq_ring->bd)
return -ENOMEM;
if (nq_ring->vmem_size)
*nq_ring->vmem = (char *)mz->addr + nq_vmem_start;
nq_ring_info->hw_stats_ctx_id = HWRM_NA_SIGNATURE;
}
net/bnxt: add initial Tx code Initial implementation of tx_pkt_burst for transmit. bnxt_xmit_pkts() is the top level function that is called during Tx. bnxt_handle_tx_cp() is used to check and process the Tx completions generated for the Tx Buffer Descriptors sent by the hardware. This patch also adds code to allocate rings in the hardware. For each Tx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a Tx request is initiated via the bnxt_xmit_pkts() call, a Buffer Descriptor is created and is sent to the hardware via the associated Tx ring. On completing the Tx operation, the hardware will generates the status in the form of a completion. This completion is processed by the bnxt_handle_tx_cp() function. Functions like bnxt_init_tx_ring_struct() and bnxt_init_one_tx_ring() are used to initialize various members of the structure before starting Tx operations. 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>
2016-06-15 21:23:13 +00:00
return 0;
}
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
net/bnxt: set ring coalesce parameters for Stratus NIC Set ring coalesce parameters for Stratus NIC. Other skews don't necessarily need this. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:32 +00:00
static void bnxt_init_dflt_coal(struct bnxt_coal *coal)
{
/* Tick values in micro seconds.
* 1 coal_buf x bufs_per_record = 1 completion record.
*/
coal->num_cmpl_aggr_int = BNXT_NUM_CMPL_AGGR_INT;
/* This is a 6-bit value and must not be 0, or we'll get non stop IRQ */
coal->num_cmpl_dma_aggr = BNXT_NUM_CMPL_DMA_AGGR;
/* This is a 6-bit value and must not be 0, or we'll get non stop IRQ */
coal->num_cmpl_dma_aggr_during_int = BNXT_NUM_CMPL_DMA_AGGR_DURING_INT;
coal->int_lat_tmr_max = BNXT_INT_LAT_TMR_MAX;
/* min timer set to 1/2 of interrupt timer */
coal->int_lat_tmr_min = BNXT_INT_LAT_TMR_MIN;
/* buf timer set to 1/4 of interrupt timer */
coal->cmpl_aggr_dma_tmr = BNXT_CMPL_AGGR_DMA_TMR;
coal->cmpl_aggr_dma_tmr_during_int = BNXT_CMPL_AGGR_DMA_TMR_DURING_INT;
}
net/bnxt: refactor doorbell handling 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>
2019-06-02 17:42:42 +00:00
static void bnxt_set_db(struct bnxt *bp,
struct bnxt_db_info *db,
uint32_t ring_type,
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
uint32_t map_idx,
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
uint32_t fid,
uint32_t ring_mask)
net/bnxt: refactor doorbell handling 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>
2019-06-02 17:42:42 +00:00
{
net/bnxt: remove references to Thor 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>
2020-12-01 19:15:23 +00:00
if (BNXT_CHIP_P5(bp)) {
net/bnxt: support 58818 chip family The new chip (Stingray 2) is part of the P5 chip family with a number of changes: 1. Implement the epoch doorbell bit for 58818 chip. With the new doorbell infrastructure and the unbounded index logic, now set the epoch doorbell bit to support proper doorbell operation on the new chip. Toggle epoch bit of all rings when it's wrapped to support doorbell overflow checking. 2. Get the legacy doorbell size from firmware. Legacy doorbell support has been removed in Stingray 2. So, the fast path doorbell pages start from the base of the BAR. Drivers need to use legacy_l2_db_space_size_kb field in the hwrm_func_qcfg_output response to get the legacy doorbell page offset from the BAR. 3. Set VALID doorbell bit on 58818 chip family. This class of chip has a valid doorbell bit added and it needs to be set. 4. Use "chip_num" returned by firmware. The "chip_num" field in the HWRM_VER_GET output returns the chip number. Use this value to identify chip category for 58818 chip family. 5. Added device ids for Stingray2 PF/VF devices. Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2020-12-20 05:24:25 +00:00
int db_offset = DB_PF_OFFSET;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
switch (ring_type) {
case HWRM_RING_ALLOC_INPUT_RING_TYPE_TX:
db->db_key64 = DBR_PATH_L2 | DBR_TYPE_SQ;
break;
case HWRM_RING_ALLOC_INPUT_RING_TYPE_RX:
case HWRM_RING_ALLOC_INPUT_RING_TYPE_RX_AGG:
db->db_key64 = DBR_PATH_L2 | DBR_TYPE_SRQ;
break;
case HWRM_RING_ALLOC_INPUT_RING_TYPE_L2_CMPL:
db->db_key64 = DBR_PATH_L2 | DBR_TYPE_CQ;
break;
case HWRM_RING_ALLOC_INPUT_RING_TYPE_NQ:
net/bnxt: cleanup NQ doorbell Simplify nq doorbell handling code by removing redundant db parameter and consolidating NQ doorbell macro into the inline function that uses it. Add "enable interrupt" variant of nq write. This will be used in a subsequent commit. When initializing nq doorbell, don't assume that only the "disable interrupt" form will be used. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-17 10:41:33 +00:00
db->db_key64 = DBR_PATH_L2;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
break;
}
net/bnxt: support 58818 chip family The new chip (Stingray 2) is part of the P5 chip family with a number of changes: 1. Implement the epoch doorbell bit for 58818 chip. With the new doorbell infrastructure and the unbounded index logic, now set the epoch doorbell bit to support proper doorbell operation on the new chip. Toggle epoch bit of all rings when it's wrapped to support doorbell overflow checking. 2. Get the legacy doorbell size from firmware. Legacy doorbell support has been removed in Stingray 2. So, the fast path doorbell pages start from the base of the BAR. Drivers need to use legacy_l2_db_space_size_kb field in the hwrm_func_qcfg_output response to get the legacy doorbell page offset from the BAR. 3. Set VALID doorbell bit on 58818 chip family. This class of chip has a valid doorbell bit added and it needs to be set. 4. Use "chip_num" returned by firmware. The "chip_num" field in the HWRM_VER_GET output returns the chip number. Use this value to identify chip category for 58818 chip family. 5. Added device ids for Stingray2 PF/VF devices. Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2020-12-20 05:24:25 +00:00
if (BNXT_CHIP_SR2(bp)) {
db->db_key64 |= DBR_VALID;
db_offset = bp->legacy_db_size;
} else if (BNXT_VF(bp)) {
db_offset = DB_VF_OFFSET;
}
db->doorbell = (char *)bp->doorbell_base + db_offset;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
db->db_key64 |= (uint64_t)fid << DBR_XID_SFT;
db->db_64 = true;
} else {
db->doorbell = (char *)bp->doorbell_base + map_idx * 0x80;
switch (ring_type) {
case HWRM_RING_ALLOC_INPUT_RING_TYPE_TX:
db->db_key32 = DB_KEY_TX;
break;
case HWRM_RING_ALLOC_INPUT_RING_TYPE_RX:
db->db_key32 = DB_KEY_RX;
break;
case HWRM_RING_ALLOC_INPUT_RING_TYPE_L2_CMPL:
db->db_key32 = DB_KEY_CP;
break;
}
db->db_64 = false;
net/bnxt: refactor doorbell handling 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>
2019-06-02 17:42:42 +00:00
}
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
db->db_ring_mask = ring_mask;
net/bnxt: support 58818 chip family The new chip (Stingray 2) is part of the P5 chip family with a number of changes: 1. Implement the epoch doorbell bit for 58818 chip. With the new doorbell infrastructure and the unbounded index logic, now set the epoch doorbell bit to support proper doorbell operation on the new chip. Toggle epoch bit of all rings when it's wrapped to support doorbell overflow checking. 2. Get the legacy doorbell size from firmware. Legacy doorbell support has been removed in Stingray 2. So, the fast path doorbell pages start from the base of the BAR. Drivers need to use legacy_l2_db_space_size_kb field in the hwrm_func_qcfg_output response to get the legacy doorbell page offset from the BAR. 3. Set VALID doorbell bit on 58818 chip family. This class of chip has a valid doorbell bit added and it needs to be set. 4. Use "chip_num" returned by firmware. The "chip_num" field in the HWRM_VER_GET output returns the chip number. Use this value to identify chip category for 58818 chip family. 5. Added device ids for Stingray2 PF/VF devices. Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2020-12-20 05:24:25 +00:00
if (BNXT_CHIP_SR2(bp)) {
db->db_epoch_mask = db->db_ring_mask + 1;
db->db_epoch_shift = DBR_EPOCH_SFT -
rte_log2_u32(db->db_epoch_mask);
}
net/bnxt: refactor doorbell handling 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>
2019-06-02 17:42:42 +00:00
}
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
static int bnxt_alloc_cmpl_ring(struct bnxt *bp, int queue_index,
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
struct bnxt_cp_ring_info *cpr)
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
{
struct bnxt_ring *cp_ring = cpr->cp_ring_struct;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
uint32_t nq_ring_id = HWRM_NA_SIGNATURE;
net/bnxt: change MSI-X vector to queue mapping DPDK PCIe-VFIO framework configures base MSIX vector for interrupts which is supported by other h/w. In case of bnxt, base MSIX vector starts with the RX completion queue 0. To comply with the DPDK framework We need to increase the map index by 1 so that RXTX completion queues events can be delivered to appropriate event listeners by kernel VFIO. Fixes: bd0a14c99f65 ("net/bnxt: use dedicated CPR for async events") Cc: stable@dpdk.org Signed-off-by: Rahul Gupta <rahul.gupta@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-10 01:41:49 +00:00
int cp_ring_index = queue_index + BNXT_RX_VEC_START;
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
struct bnxt_cp_ring_info *nqr = bp->rxtx_nq_ring;
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
uint8_t ring_type;
int rc = 0;
ring_type = HWRM_RING_ALLOC_INPUT_RING_TYPE_L2_CMPL;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
if (BNXT_HAS_NQ(bp)) {
if (nqr) {
nq_ring_id = nqr->cp_ring_struct->fw_ring_id;
} else {
PMD_DRV_LOG(ERR, "NQ ring is NULL\n");
return -EINVAL;
}
}
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
rc = bnxt_hwrm_ring_alloc(bp, cp_ring, ring_type, cp_ring_index,
net/bnxt: support CoS classification Class of Service (CoS) is a way to manage multiple types of traffic over a network to offer different types of services to applications. CoS classification (priority to cosqueue) is determined by the user and configured through the PF driver. DPDK driver queries this configuration and maps the cos queue ids to different VNICs. This patch adds this support. Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:58 +00:00
HWRM_NA_SIGNATURE, nq_ring_id, 0);
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
if (rc)
return rc;
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
cpr->cp_raw_cons = 0;
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
bnxt_set_db(bp, &cpr->cp_db, ring_type, cp_ring_index,
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
cp_ring->fw_ring_id, cp_ring->ring_mask);
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
bnxt_db_cq(cpr);
return 0;
}
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
int bnxt_alloc_rxtx_nq_ring(struct bnxt *bp)
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
{
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
struct bnxt_cp_ring_info *nqr;
struct bnxt_ring *ring;
int ring_index = BNXT_NUM_ASYNC_CPR(bp);
unsigned int socket_id;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
uint8_t ring_type;
int rc = 0;
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
if (!BNXT_HAS_NQ(bp) || bp->rxtx_nq_ring)
return 0;
eal: rename lcore master and slave Replace master lcore with main lcore and replace slave lcore with worker lcore. Keep the old functions and macros but mark them as deprecated for this release. The "--master-lcore" command line option is also deprecated and any usage will print a warning and use "--main-lcore" as replacement. Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
2020-10-15 22:57:19 +00:00
socket_id = rte_lcore_to_socket_id(rte_get_main_lcore());
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
nqr = rte_zmalloc_socket("nqr",
sizeof(struct bnxt_cp_ring_info),
RTE_CACHE_LINE_SIZE, socket_id);
if (nqr == NULL)
return -ENOMEM;
ring = rte_zmalloc_socket("bnxt_cp_ring_struct",
sizeof(struct bnxt_ring),
RTE_CACHE_LINE_SIZE, socket_id);
if (ring == NULL) {
rte_free(nqr);
return -ENOMEM;
}
ring->bd = (void *)nqr->cp_desc_ring;
ring->bd_dma = nqr->cp_desc_mapping;
ring->ring_size = rte_align32pow2(DEFAULT_CP_RING_SIZE);
ring->ring_mask = ring->ring_size - 1;
ring->vmem_size = 0;
ring->vmem = NULL;
net/bnxt: fix structure variable initialization During port start if bnxt_alloc_all_hwrm_stat_ctxs() fails, in the cleanup path we do see errors like below: bnxt_hwrm_ring_free(): hwrm_ring_free cp failed. rc:2 bnxt_hwrm_ring_free(): hwrm_ring_free rx failed. rc:2 The reason for this is in bnxt_free_all_hwrm_rings(), the check is made against "ring->fw_ring_id != INVALID_HW_RING_ID" which always return true as ring->fw_ring_id is not set to INVALID_HW_RING_ID while initialising the ring structs. Fixes: 6eb3cc2294fd ("net/bnxt: add initial Tx code") Fixes: 2eb53b134aae ("net/bnxt: add initial Rx code") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2020-08-28 05:01:09 +00:00
ring->fw_ring_id = INVALID_HW_RING_ID;
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
nqr->cp_ring_struct = ring;
rc = bnxt_alloc_rings(bp, 0, NULL, NULL, nqr, NULL, "l2_nqr");
if (rc) {
rte_free(ring);
rte_free(nqr);
return -ENOMEM;
}
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
ring_type = HWRM_RING_ALLOC_INPUT_RING_TYPE_NQ;
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
rc = bnxt_hwrm_ring_alloc(bp, ring, ring_type, ring_index,
net/bnxt: support CoS classification Class of Service (CoS) is a way to manage multiple types of traffic over a network to offer different types of services to applications. CoS classification (priority to cosqueue) is determined by the user and configured through the PF driver. DPDK driver queries this configuration and maps the cos queue ids to different VNICs. This patch adds this support. Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:58 +00:00
HWRM_NA_SIGNATURE, HWRM_NA_SIGNATURE, 0);
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
if (rc) {
rte_free(ring);
rte_free(nqr);
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
return rc;
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
}
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
bnxt_set_db(bp, &nqr->cp_db, ring_type, ring_index,
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
ring->fw_ring_id, ring->ring_mask);
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
bnxt_db_nq(nqr);
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
bp->rxtx_nq_ring = nqr;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
return 0;
}
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
/* Free RX/TX NQ ring. */
void bnxt_free_rxtx_nq_ring(struct bnxt *bp)
{
struct bnxt_cp_ring_info *nqr = bp->rxtx_nq_ring;
if (!nqr)
return;
bnxt_free_nq_ring(bp, nqr);
bnxt_free_ring(nqr->cp_ring_struct);
rte_free(nqr->cp_ring_struct);
nqr->cp_ring_struct = NULL;
rte_free(nqr);
bp->rxtx_nq_ring = NULL;
}
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
static int bnxt_alloc_rx_ring(struct bnxt *bp, int queue_index)
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
{
struct bnxt_rx_queue *rxq = bp->rx_queues[queue_index];
struct bnxt_cp_ring_info *cpr = rxq->cp_ring;
struct bnxt_ring *cp_ring = cpr->cp_ring_struct;
struct bnxt_rx_ring_info *rxr = rxq->rx_ring;
struct bnxt_ring *ring = rxr->rx_ring_struct;
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
uint8_t ring_type;
int rc = 0;
ring_type = HWRM_RING_ALLOC_INPUT_RING_TYPE_RX;
rc = bnxt_hwrm_ring_alloc(bp, ring, ring_type,
queue_index, cpr->hw_stats_ctx_id,
net/bnxt: support CoS classification Class of Service (CoS) is a way to manage multiple types of traffic over a network to offer different types of services to applications. CoS classification (priority to cosqueue) is determined by the user and configured through the PF driver. DPDK driver queries this configuration and maps the cos queue ids to different VNICs. This patch adds this support. Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:58 +00:00
cp_ring->fw_ring_id, 0);
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
if (rc)
return rc;
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
rxr->rx_raw_prod = 0;
net/bnxt: create ring group array only when needed Fix an overrun of the ring group array with BCM5750X-based adapters by ensuring that the ring group array is not allocated or accessed for adapters that do not support ring groups. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-10 17:11:19 +00:00
if (BNXT_HAS_RING_GRPS(bp))
bp->grp_info[queue_index].rx_fw_ring_id = ring->fw_ring_id;
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
bnxt_set_db(bp, &rxr->rx_db, ring_type, queue_index, ring->fw_ring_id,
ring->ring_mask);
bnxt_db_write(&rxr->rx_db, rxr->rx_raw_prod);
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
return 0;
}
static int bnxt_alloc_rx_agg_ring(struct bnxt *bp, int queue_index)
{
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
unsigned int map_idx = queue_index + bp->rx_cp_nr_rings;
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
struct bnxt_rx_queue *rxq = bp->rx_queues[queue_index];
struct bnxt_cp_ring_info *cpr = rxq->cp_ring;
struct bnxt_ring *cp_ring = cpr->cp_ring_struct;
struct bnxt_rx_ring_info *rxr = rxq->rx_ring;
struct bnxt_ring *ring = rxr->ag_ring_struct;
uint32_t hw_stats_ctx_id = HWRM_NA_SIGNATURE;
net/bnxt: refactor doorbell handling 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>
2019-06-02 17:42:42 +00:00
uint8_t ring_type;
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
int rc = 0;
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
ring->fw_rx_ring_id = rxr->rx_ring_struct->fw_ring_id;
net/bnxt: remove references to Thor 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>
2020-12-01 19:15:23 +00:00
if (BNXT_CHIP_P5(bp)) {
net/bnxt: support thor controller This commit adds support to the bnxt PMD for devices based on the BCM57508 "thor" Ethernet controller. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com>
2019-06-02 17:42:44 +00:00
ring_type = HWRM_RING_ALLOC_INPUT_RING_TYPE_RX_AGG;
hw_stats_ctx_id = cpr->hw_stats_ctx_id;
} else {
ring_type = HWRM_RING_ALLOC_INPUT_RING_TYPE_RX;
}
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
rc = bnxt_hwrm_ring_alloc(bp, ring, ring_type, map_idx,
net/bnxt: support CoS classification Class of Service (CoS) is a way to manage multiple types of traffic over a network to offer different types of services to applications. CoS classification (priority to cosqueue) is determined by the user and configured through the PF driver. DPDK driver queries this configuration and maps the cos queue ids to different VNICs. This patch adds this support. Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:58 +00:00
hw_stats_ctx_id, cp_ring->fw_ring_id, 0);
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
if (rc)
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
return rc;
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
rxr->ag_raw_prod = 0;
net/bnxt: create ring group array only when needed Fix an overrun of the ring group array with BCM5750X-based adapters by ensuring that the ring group array is not allocated or accessed for adapters that do not support ring groups. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-10 17:11:19 +00:00
if (BNXT_HAS_RING_GRPS(bp))
bp->grp_info[queue_index].ag_fw_ring_id = ring->fw_ring_id;
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
bnxt_set_db(bp, &rxr->ag_db, ring_type, map_idx, ring->fw_ring_id,
ring->ring_mask);
bnxt_db_write(&rxr->ag_db, rxr->ag_raw_prod);
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
return 0;
}
int bnxt_alloc_hwrm_rx_ring(struct bnxt *bp, int queue_index)
{
struct bnxt_rx_queue *rxq = bp->rx_queues[queue_index];
struct bnxt_cp_ring_info *cpr = rxq->cp_ring;
struct bnxt_ring *cp_ring = cpr->cp_ring_struct;
struct bnxt_rx_ring_info *rxr = rxq->rx_ring;
net/bnxt: refactor HWRM ring allocation routine bnxt_alloc_hwrm_rings() was duplicating a snippet of code that was already present in bnxt_alloc_hwrm_rx_ring() with just one line/function invocation missing. Add that line and replace the snippet of code with a call to bnxt_hwrm_alloc_rx_ring Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2021-02-19 09:28:03 +00:00
struct bnxt_coal coal;
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
int rc;
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
net/bnxt: fix Rx completion ring size calculation The size of the receive completion ring should be recalculated when MTU is increased to a size that requires scattered receive or when LRO is enabled. Move logic for this calculation from the ring configuration path to the device start path. - Made size calculation dependent only on scattered_rx status. - Moved calculation of scattered_rx up in the initialization sequence. - Made LRO offload status part of scattered_rx calculation. When the completion ring size is too small, completion overflows can occur causing the ring to be disabled in hardware. Fixes: 04067844a3e9 ("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>
2021-01-29 18:07:09 +00:00
/*
* Storage for the cp ring is allocated based on worst-case
* usage, the actual size to be used by hw is computed here.
*/
cp_ring->ring_size = rxr->rx_ring_struct->ring_size * 2;
if (bp->eth_dev->data->scattered_rx)
cp_ring->ring_size *= AGG_RING_SIZE_FACTOR;
cp_ring->ring_mask = cp_ring->ring_size - 1;
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
rc = bnxt_alloc_cmpl_ring(bp, queue_index, cpr);
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
if (rc)
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
goto err_out;
net/bnxt: create ring group array only when needed Fix an overrun of the ring group array with BCM5750X-based adapters by ensuring that the ring group array is not allocated or accessed for adapters that do not support ring groups. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-10 17:11:19 +00:00
if (BNXT_HAS_RING_GRPS(bp)) {
bp->grp_info[queue_index].fw_stats_ctx = cpr->hw_stats_ctx_id;
bp->grp_info[queue_index].cp_fw_ring_id = cp_ring->fw_ring_id;
}
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
net/bnxt: refactor HWRM ring allocation routine bnxt_alloc_hwrm_rings() was duplicating a snippet of code that was already present in bnxt_alloc_hwrm_rx_ring() with just one line/function invocation missing. Add that line and replace the snippet of code with a call to bnxt_hwrm_alloc_rx_ring Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2021-02-19 09:28:03 +00:00
bnxt_init_dflt_coal(&coal);
bnxt_hwrm_set_ring_coal(bp, &coal, cp_ring->fw_ring_id);
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
if (!BNXT_NUM_ASYNC_CPR(bp) && !queue_index) {
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
/*
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
* If a dedicated async event completion ring is not enabled,
* use the first completion ring from PF or VF as the default
* completion ring for async event handling.
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
*/
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
bp->async_cp_ring = cpr;
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
rc = bnxt_hwrm_set_async_event_cr(bp);
if (rc)
goto err_out;
}
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
rc = bnxt_alloc_rx_ring(bp, queue_index);
if (rc)
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
goto err_out;
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
rc = bnxt_alloc_rx_agg_ring(bp, queue_index);
if (rc)
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
goto err_out;
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
net/bnxt: fix Rx queue start/stop Driver should not change "deferred_start" state of the rx queues. It should get the state in queue_setup_op() and use that value. Since the deferred start state was being used in the packet receive functions to determine whether a stopped rx ring should be polled, introduced a per-rxq flag to track queue stopped/started state. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-02 17:17:32 +00:00
if (rxq->rx_started) {
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
if (bnxt_init_one_rx_ring(rxq)) {
net/bnxt: update indentation and coding style Indentation and coding style fixes. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Acked-by: Lance Richardson <lance.richardson@broadcom.com>
2020-12-05 00:15:05 +00:00
PMD_DRV_LOG(ERR, "bnxt_init_one_rx_ring failed!\n");
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
bnxt_rx_queue_release_op(rxq);
rc = -ENOMEM;
goto err_out;
}
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
bnxt_db_write(&rxr->rx_db, rxr->rx_raw_prod);
bnxt_db_write(&rxr->ag_db, rxr->ag_raw_prod);
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
}
rxq->index = queue_index;
net/bnxt: support NEON 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>
2020-07-07 22:22:25 +00:00
#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
net/bnxt: reset Rx allocation state on Rx queue restart Call bnxt_rxq_vec_setup() to ensure that rxq->rxrearm_nb and rxq->rxrearm_start are initialized correctly when a receive queue is restarted. Fixes: bc4a000f2f53 ("net/bnxt: implement SSE vector mode") Cc: stable@dpdk.org Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Xiaobing Zhang <xiaobing.zhang@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-09 17:43:34 +00:00
bnxt_rxq_vec_setup(rxq);
#endif
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
return 0;
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
err_out:
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
PMD_DRV_LOG(ERR,
"Failed to allocate receive queue %d, rc %d.\n",
queue_index, rc);
net/bnxt: fix queue start/stop operations Packets destined to the to-be-stopped queue should not be dropped (neither in HW nor in the driver), so re-program the RSS Table without this queue on stop and add it back to the table on start unless it is a Representor VF. Since 0th entry is used for default ring, use fw_grp_id + 1 to change the RSS table population logic by programming valid IDs instead of the default zeroth entry in case of an invalid fw_grp_id. Destroy and recreate the trio of Rx rings(compl, Rx, AG) every time in start so that HW is in sync with software. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ray Jui <ray.jui@broadcom.com> Reviewed-by: Scott Branden <scott.branden@broadcom.com> Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:35 +00:00
return rc;
}
net/bnxt: refactor doorbell handling 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>
2019-06-02 17:42:42 +00:00
net/bnxt: handle ring cleanup in case of error Initialize all rings to INVALID_HW_RING_ID. This can be used to determine the rings to free if allocation fails. Signed-off-by: Rahul Gupta <rahul.gupta@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-02 23:25:52 +00:00
/* Initialise all rings to -1, its used to free rings later if allocation
* of few rings fails.
*/
static void bnxt_init_all_rings(struct bnxt *bp)
{
unsigned int i = 0;
struct bnxt_rx_queue *rxq;
struct bnxt_ring *cp_ring;
struct bnxt_ring *ring;
struct bnxt_rx_ring_info *rxr;
struct bnxt_tx_queue *txq;
for (i = 0; i < bp->rx_cp_nr_rings; i++) {
rxq = bp->rx_queues[i];
/* Rx-compl */
cp_ring = rxq->cp_ring->cp_ring_struct;
cp_ring->fw_ring_id = INVALID_HW_RING_ID;
/* Rx-Reg */
rxr = rxq->rx_ring;
ring = rxr->rx_ring_struct;
ring->fw_ring_id = INVALID_HW_RING_ID;
/* Rx-AGG */
ring = rxr->ag_ring_struct;
ring->fw_ring_id = INVALID_HW_RING_ID;
}
for (i = 0; i < bp->tx_cp_nr_rings; i++) {
txq = bp->tx_queues[i];
/* Tx cmpl */
cp_ring = txq->cp_ring->cp_ring_struct;
cp_ring->fw_ring_id = INVALID_HW_RING_ID;
/*Tx Ring */
ring = txq->tx_ring->tx_ring_struct;
ring->fw_ring_id = INVALID_HW_RING_ID;
}
}
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
/* ring_grp usage:
* [0] = default completion ring
* [1 -> +rx_cp_nr_rings] = rx_cp, rx rings
* [1+rx_cp_nr_rings + 1 -> +tx_cp_nr_rings] = tx_cp, tx rings
*/
int bnxt_alloc_hwrm_rings(struct bnxt *bp)
{
net/bnxt: set ring coalesce parameters for Stratus NIC Set ring coalesce parameters for Stratus NIC. Other skews don't necessarily need this. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:32 +00:00
struct bnxt_coal coal;
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
unsigned int i;
net/bnxt: refactor doorbell handling 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>
2019-06-02 17:42:42 +00:00
uint8_t ring_type;
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
int rc = 0;
net/bnxt: set ring coalesce parameters for Stratus NIC Set ring coalesce parameters for Stratus NIC. Other skews don't necessarily need this. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:32 +00:00
bnxt_init_dflt_coal(&coal);
net/bnxt: handle ring cleanup in case of error Initialize all rings to INVALID_HW_RING_ID. This can be used to determine the rings to free if allocation fails. Signed-off-by: Rahul Gupta <rahul.gupta@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-02 23:25:52 +00:00
bnxt_init_all_rings(bp);
net/bnxt: set ring coalesce parameters for Stratus NIC Set ring coalesce parameters for Stratus NIC. Other skews don't necessarily need this. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:32 +00:00
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
for (i = 0; i < bp->rx_cp_nr_rings; i++) {
net/bnxt: refactor HWRM ring allocation routine bnxt_alloc_hwrm_rings() was duplicating a snippet of code that was already present in bnxt_alloc_hwrm_rx_ring() with just one line/function invocation missing. Add that line and replace the snippet of code with a call to bnxt_hwrm_alloc_rx_ring Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2021-02-19 09:28:03 +00:00
rc = bnxt_alloc_hwrm_rx_ring(bp, i);
if (rc)
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
goto err_out;
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
}
for (i = 0; i < bp->tx_cp_nr_rings; i++) {
struct bnxt_tx_queue *txq = bp->tx_queues[i];
struct bnxt_cp_ring_info *cpr = txq->cp_ring;
struct bnxt_ring *cp_ring = cpr->cp_ring_struct;
struct bnxt_tx_ring_info *txr = txq->tx_ring;
struct bnxt_ring *ring = txr->tx_ring_struct;
net/bnxt: use first completion ring for fwd and async event In order to save completion resource, use the first completion ring from PF or VF as the default completion ring for async event & HWRM forward response handling. Add bnxt_hwrm_set_async_event_cr() to set async_event_cr for either PF or VF. Fixes: 7bc8e9a227cc ("net/bnxt: support async link notification") Signed-off-by: Qingmin Liu <qingmin.liu@broadcom.com> Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-05-22 18:13:44 +00:00
unsigned int idx = i + bp->rx_cp_nr_rings;
net/bnxt: support CoS classification Class of Service (CoS) is a way to manage multiple types of traffic over a network to offer different types of services to applications. CoS classification (priority to cosqueue) is determined by the user and configured through the PF driver. DPDK driver queries this configuration and maps the cos queue ids to different VNICs. This patch adds this support. Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:58 +00:00
uint16_t tx_cosq_id = 0;
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
net/bnxt: use common NQ ring Thor queue scaling is currently limited by the number of NQs that can be allocated. Fix by using a common NQ for all receive/transmit rings instead of allocating a separate NQ for each ring. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") 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> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
2019-10-04 03:48:59 +00:00
if (bnxt_alloc_cmpl_ring(bp, idx, cpr))
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
goto err_out;
net/bnxt: support CoS classification Class of Service (CoS) is a way to manage multiple types of traffic over a network to offer different types of services to applications. CoS classification (priority to cosqueue) is determined by the user and configured through the PF driver. DPDK driver queries this configuration and maps the cos queue ids to different VNICs. This patch adds this support. Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:58 +00:00
if (bp->vnic_cap_flags & BNXT_VNIC_CAP_COS_CLASSIFY)
tx_cosq_id = bp->tx_cosq_id[i < bp->max_lltc ? i : 0];
else
tx_cosq_id = bp->tx_cosq_id[0];
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
/* Tx ring */
net/bnxt: refactor doorbell handling 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>
2019-06-02 17:42:42 +00:00
ring_type = HWRM_RING_ALLOC_INPUT_RING_TYPE_TX;
net/bnxt: refactor ring allocation Reduce code duplication and prepare for supporting hardware with different ring allocation requirements by refactoring ring allocation code. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-06-02 17:42:43 +00:00
rc = bnxt_hwrm_ring_alloc(bp, ring,
ring_type,
net/bnxt: fix doorbell register offset for Tx ring For Tx-ring # 104 and higher, the doorbell register was incorrectly configured due to which FW was not able to receive the notification of packet to transmit. With this fix, user can run traffic upto 256 rings. Fixes: 6eb3cc2294fd ("net/bnxt: add initial Tx code") Cc: stable@dpdk.org Signed-off-by: Rahul Gupta <rahul.gupta@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-17 10:41:31 +00:00
i, cpr->hw_stats_ctx_id,
net/bnxt: support CoS classification Class of Service (CoS) is a way to manage multiple types of traffic over a network to offer different types of services to applications. CoS classification (priority to cosqueue) is determined by the user and configured through the PF driver. DPDK driver queries this configuration and maps the cos queue ids to different VNICs. This patch adds this support. Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:58 +00:00
cp_ring->fw_ring_id,
tx_cosq_id);
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
if (rc)
goto err_out;
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
bnxt_set_db(bp, &txr->tx_db, ring_type, i, ring->fw_ring_id,
ring->ring_mask);
net/bnxt: support set MTU 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>
2017-06-01 17:07:09 +00:00
txq->index = idx;
net/bnxt: set ring coalesce parameters for Stratus NIC Set ring coalesce parameters for Stratus NIC. Other skews don't necessarily need this. Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2018-06-28 20:15:32 +00:00
bnxt_hwrm_set_ring_coal(bp, &coal, cp_ring->fw_ring_id);
net/bnxt: allocate rings and groups Add a top level functions to initialize ring groups, and functions to allocate and free all the rings via HWRM. A ring group is identified by an index. It consists of Rx or Tx ring id, completion ring id and a statistics context. Once a ring group is initialized, use this group index while creating the rings in the ASIC using the appropriate HWRM API added via earlier patches. Functions added: bnxt_free_cp_ring Calls the HWRM function generic ring free with arguments specific to a completion ring and sanitizes the host completion structure bnxt_free_all_hwrm_rings Frees all the HWRM allocated hardware rings bnxt_free_all_hwrm_resources Frees all the resources allocated via the HRM in the hardware bnxt_alloc_hwrm_rings Allocates all the HWRM rings needed in the current configuration This should be the last functionality needed to add start/stop device operations. 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>
2016-06-15 21:23:28 +00:00
}
err_out:
return rc;
}
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
/* Allocate dedicated async completion ring. */
int bnxt_alloc_async_cp_ring(struct bnxt *bp)
{
struct bnxt_cp_ring_info *cpr = bp->async_cp_ring;
net/bnxt: avoid null pointer dereference Commit "bd0a14c99f65" enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring on non Stingray Platforms. This causes a segfault due to NULL pointer dereference in bnxt_alloc_async_cp_ring() on stingray. Fix this by checking the pointer validity before accessing it. Fixes: bd0a14c99f65 ("net/bnxt: use dedicated CPR for async events") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by: Rahul Gupta <rahul.gupta@broadcom.com> Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
2019-10-02 01:23:33 +00:00
struct bnxt_ring *cp_ring;
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
uint8_t ring_type;
int rc;
net/bnxt: avoid null pointer dereference Commit "bd0a14c99f65" enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring on non Stingray Platforms. This causes a segfault due to NULL pointer dereference in bnxt_alloc_async_cp_ring() on stingray. Fix this by checking the pointer validity before accessing it. Fixes: bd0a14c99f65 ("net/bnxt: use dedicated CPR for async events") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by: Rahul Gupta <rahul.gupta@broadcom.com> Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
2019-10-02 01:23:33 +00:00
if (BNXT_NUM_ASYNC_CPR(bp) == 0 || cpr == NULL)
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
return 0;
net/bnxt: avoid null pointer dereference Commit "bd0a14c99f65" enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring on non Stingray Platforms. This causes a segfault due to NULL pointer dereference in bnxt_alloc_async_cp_ring() on stingray. Fix this by checking the pointer validity before accessing it. Fixes: bd0a14c99f65 ("net/bnxt: use dedicated CPR for async events") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by: Rahul Gupta <rahul.gupta@broadcom.com> Reviewed-by: Lance Richardson <lance.richardson@broadcom.com>
2019-10-02 01:23:33 +00:00
cp_ring = cpr->cp_ring_struct;
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
if (BNXT_HAS_NQ(bp))
ring_type = HWRM_RING_ALLOC_INPUT_RING_TYPE_NQ;
else
ring_type = HWRM_RING_ALLOC_INPUT_RING_TYPE_L2_CMPL;
rc = bnxt_hwrm_ring_alloc(bp, cp_ring, ring_type, 0,
net/bnxt: support CoS classification Class of Service (CoS) is a way to manage multiple types of traffic over a network to offer different types of services to applications. CoS classification (priority to cosqueue) is determined by the user and configured through the PF driver. DPDK driver queries this configuration and maps the cos queue ids to different VNICs. This patch adds this support. Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Santoshkumar Karanappa Rastapur <santosh.rastapur@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-10-04 03:48:58 +00:00
HWRM_NA_SIGNATURE, HWRM_NA_SIGNATURE, 0);
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
if (rc)
return rc;
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
cpr->cp_raw_cons = 0;
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
cpr->valid = 0;
bnxt_set_db(bp, &cpr->cp_db, ring_type, 0,
net/bnxt: modify ring index logic 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>
2020-12-07 17:54:30 +00:00
cp_ring->fw_ring_id, cp_ring->ring_mask);
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
if (BNXT_HAS_NQ(bp))
bnxt_db_nq(cpr);
else
bnxt_db_cq(cpr);
return bnxt_hwrm_set_async_event_cr(bp);
}
/* Free dedicated async completion ring. */
void bnxt_free_async_cp_ring(struct bnxt *bp)
{
struct bnxt_cp_ring_info *cpr = bp->async_cp_ring;
if (BNXT_NUM_ASYNC_CPR(bp) == 0 || cpr == NULL)
return;
if (BNXT_HAS_NQ(bp))
bnxt_free_nq_ring(bp, cpr);
else
bnxt_free_cp_ring(bp, cpr);
bnxt_free_ring(cpr->cp_ring_struct);
rte_free(cpr->cp_ring_struct);
cpr->cp_ring_struct = NULL;
rte_free(cpr);
bp->async_cp_ring = NULL;
}
int bnxt_alloc_async_ring_struct(struct bnxt *bp)
{
struct bnxt_cp_ring_info *cpr = NULL;
struct bnxt_ring *ring = NULL;
unsigned int socket_id;
if (BNXT_NUM_ASYNC_CPR(bp) == 0)
return 0;
eal: rename lcore master and slave Replace master lcore with main lcore and replace slave lcore with worker lcore. Keep the old functions and macros but mark them as deprecated for this release. The "--master-lcore" command line option is also deprecated and any usage will print a warning and use "--main-lcore" as replacement. Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
2020-10-15 22:57:19 +00:00
socket_id = rte_lcore_to_socket_id(rte_get_main_lcore());
net/bnxt: use dedicated CPR for async events This commit enables the creation of a dedicated completion ring for asynchronous event handling instead of handling these events on a receive completion ring. For the stingray platform and other platforms needing tighter control of resource utilization, we retain the ability to process async events on a receive completion ring. For Thor-based adapters, we use a dedicated NQ (notification queue) ring for async events (async events can't currently be received on a completion ring due to a firmware limitation). Rename "def_cp_ring" to "async_cp_ring" to better reflect its purpose (async event notifications) and to avoid confusion with VNIC default receive completion rings. Allow rxq 0 to be stopped when not being used for async events. Signed-off-by: Lance Richardson <lance.richardson@broadcom.com> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
2019-07-24 16:49:32 +00:00
cpr = rte_zmalloc_socket("cpr",
sizeof(struct bnxt_cp_ring_info),
RTE_CACHE_LINE_SIZE, socket_id);
if (cpr == NULL)
return -ENOMEM;
ring = rte_zmalloc_socket("bnxt_cp_ring_struct",
sizeof(struct bnxt_ring),
RTE_CACHE_LINE_SIZE, socket_id);
if (ring == NULL) {
rte_free(cpr);
return -ENOMEM;
}
ring->bd = (void *)cpr->cp_desc_ring;
ring->bd_dma = cpr->cp_desc_mapping;
ring->ring_size = rte_align32pow2(DEFAULT_CP_RING_SIZE);
ring->ring_mask = ring->ring_size - 1;
ring->vmem_size = 0;
ring->vmem = NULL;
bp->async_cp_ring = cpr;
cpr->cp_ring_struct = ring;
return bnxt_alloc_rings(bp, 0, NULL, NULL,
bp->async_cp_ring, NULL,
"def_cp");
}
Reference in New Issue Copy Permalink