diff options
Diffstat (limited to 'drivers/net/ethernet/mellanox/mlx5/core')
71 files changed, 6401 insertions, 646 deletions
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig index fc52d742b7f7..5aee05992f27 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig +++ b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig @@ -11,9 +11,23 @@ config MLX5_CORE Core driver for low level functionality of the ConnectX-4 and Connect-IB cards by Mellanox Technologies. +config MLX5_ACCEL + bool + +config MLX5_FPGA + bool "Mellanox Technologies Innova support" + depends on MLX5_CORE + select MLX5_ACCEL + ---help--- + Build support for the Innova family of network cards by Mellanox + Technologies. Innova network cards are comprised of a ConnectX chip + and an FPGA chip on one board. If you select this option, the + mlx5_core driver will include the Innova FPGA core and allow building + sandbox-specific client drivers. + config MLX5_CORE_EN bool "Mellanox Technologies ConnectX-4 Ethernet support" - depends on NETDEVICES && ETHERNET && PCI && MLX5_CORE + depends on NETDEVICES && ETHERNET && INET && PCI && MLX5_CORE depends on IPV6=y || IPV6=n || MLX5_CORE=m imply PTP_1588_CLOCK default n @@ -38,3 +52,15 @@ config MLX5_CORE_IPOIB default n ---help--- MLX5 IPoIB offloads & acceleration support. + +config MLX5_EN_IPSEC + bool "IPSec XFRM cryptography-offload accelaration" + depends on MLX5_ACCEL + depends on MLX5_CORE_EN + depends on XFRM_OFFLOAD + depends on INET_ESP_OFFLOAD || INET6_ESP_OFFLOAD + default n + ---help--- + Build support for IPsec cryptography-offload accelaration in the NIC. + Note: Support for hardware with this capability needs to be selected + for this option to become available. diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile index 9e644615f07a..9d17e4e76d3a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile +++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile @@ -1,15 +1,24 @@ obj-$(CONFIG_MLX5_CORE) += mlx5_core.o +subdir-ccflags-y += -I$(src) mlx5_core-y := main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \ health.o mcg.o cq.o srq.o alloc.o qp.o port.o mr.o pd.o \ mad.o transobj.o vport.o sriov.o fs_cmd.o fs_core.o \ - fs_counters.o rl.o lag.o dev.o + fs_counters.o rl.o lag.o dev.o wq.o lib/gid.o -mlx5_core-$(CONFIG_MLX5_CORE_EN) += wq.o eswitch.o eswitch_offloads.o \ +mlx5_core-$(CONFIG_MLX5_ACCEL) += accel/ipsec.o + +mlx5_core-$(CONFIG_MLX5_FPGA) += fpga/cmd.o fpga/core.o fpga/conn.o fpga/sdk.o \ + fpga/ipsec.o + +mlx5_core-$(CONFIG_MLX5_CORE_EN) += eswitch.o eswitch_offloads.o \ en_main.o en_common.o en_fs.o en_ethtool.o en_tx.o \ en_rx.o en_rx_am.o en_txrx.o en_clock.o vxlan.o \ en_tc.o en_arfs.o en_rep.o en_fs_ethtool.o en_selftest.o mlx5_core-$(CONFIG_MLX5_CORE_EN_DCB) += en_dcbnl.o -mlx5_core-$(CONFIG_MLX5_CORE_IPOIB) += ipoib.o +mlx5_core-$(CONFIG_MLX5_CORE_IPOIB) += ipoib/ipoib.o ipoib/ethtool.o + +mlx5_core-$(CONFIG_MLX5_EN_IPSEC) += en_accel/ipsec.o en_accel/ipsec_rxtx.o \ + en_accel/ipsec_stats.o diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/accel/Makefile new file mode 100644 index 000000000000..d8e17110f25d --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/Makefile @@ -0,0 +1 @@ +subdir-ccflags-y += -I$(src)/.. diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c new file mode 100644 index 000000000000..53e69edaedde --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c @@ -0,0 +1,78 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/mlx5/device.h> + +#include "accel/ipsec.h" +#include "mlx5_core.h" +#include "fpga/ipsec.h" + +void *mlx5_accel_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, + struct mlx5_accel_ipsec_sa *cmd) +{ + if (!MLX5_IPSEC_DEV(mdev)) + return ERR_PTR(-EOPNOTSUPP); + + return mlx5_fpga_ipsec_sa_cmd_exec(mdev, cmd); +} + +int mlx5_accel_ipsec_sa_cmd_wait(void *ctx) +{ + return mlx5_fpga_ipsec_sa_cmd_wait(ctx); +} + +u32 mlx5_accel_ipsec_device_caps(struct mlx5_core_dev *mdev) +{ + return mlx5_fpga_ipsec_device_caps(mdev); +} + +unsigned int mlx5_accel_ipsec_counters_count(struct mlx5_core_dev *mdev) +{ + return mlx5_fpga_ipsec_counters_count(mdev); +} + +int mlx5_accel_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters, + unsigned int count) +{ + return mlx5_fpga_ipsec_counters_read(mdev, counters, count); +} + +int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev) +{ + return mlx5_fpga_ipsec_init(mdev); +} + +void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev) +{ + mlx5_fpga_ipsec_cleanup(mdev); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h new file mode 100644 index 000000000000..d6e20fea9554 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h @@ -0,0 +1,138 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef __MLX5_ACCEL_IPSEC_H__ +#define __MLX5_ACCEL_IPSEC_H__ + +#ifdef CONFIG_MLX5_ACCEL + +#include <linux/mlx5/driver.h> + +enum { + MLX5_ACCEL_IPSEC_DEVICE = BIT(1), + MLX5_ACCEL_IPSEC_IPV6 = BIT(2), + MLX5_ACCEL_IPSEC_ESP = BIT(3), + MLX5_ACCEL_IPSEC_LSO = BIT(4), +}; + +#define MLX5_IPSEC_SADB_IP_AH BIT(7) +#define MLX5_IPSEC_SADB_IP_ESP BIT(6) +#define MLX5_IPSEC_SADB_SA_VALID BIT(5) +#define MLX5_IPSEC_SADB_SPI_EN BIT(4) +#define MLX5_IPSEC_SADB_DIR_SX BIT(3) +#define MLX5_IPSEC_SADB_IPV6 BIT(2) + +enum { + MLX5_IPSEC_CMD_ADD_SA = 0, + MLX5_IPSEC_CMD_DEL_SA = 1, +}; + +enum mlx5_accel_ipsec_enc_mode { + MLX5_IPSEC_SADB_MODE_NONE = 0, + MLX5_IPSEC_SADB_MODE_AES_GCM_128_AUTH_128 = 1, + MLX5_IPSEC_SADB_MODE_AES_GCM_256_AUTH_128 = 3, +}; + +#define MLX5_IPSEC_DEV(mdev) (mlx5_accel_ipsec_device_caps(mdev) & \ + MLX5_ACCEL_IPSEC_DEVICE) + +struct mlx5_accel_ipsec_sa { + __be32 cmd; + u8 key_enc[32]; + u8 key_auth[32]; + __be32 sip[4]; + __be32 dip[4]; + union { + struct { + __be32 reserved; + u8 salt_iv[8]; + __be32 salt; + } __packed gcm; + struct { + u8 salt[16]; + } __packed cbc; + }; + __be32 spi; + __be32 sw_sa_handle; + __be16 tfclen; + u8 enc_mode; + u8 sip_masklen; + u8 dip_masklen; + u8 flags; + u8 reserved[2]; +} __packed; + +/** + * mlx5_accel_ipsec_sa_cmd_exec - Execute an IPSec SADB command + * @mdev: mlx5 device + * @cmd: command to execute + * May be called from atomic context. Returns context pointer, or error + * Caller must eventually call mlx5_accel_ipsec_sa_cmd_wait from non-atomic + * context, to cleanup the context pointer + */ +void *mlx5_accel_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, + struct mlx5_accel_ipsec_sa *cmd); + +/** + * mlx5_accel_ipsec_sa_cmd_wait - Wait for command execution completion + * @context: Context pointer returned from call to mlx5_accel_ipsec_sa_cmd_exec + * Sleeps (killable) until command execution is complete. + * Returns the command result, or -EINTR if killed + */ +int mlx5_accel_ipsec_sa_cmd_wait(void *context); + +u32 mlx5_accel_ipsec_device_caps(struct mlx5_core_dev *mdev); + +unsigned int mlx5_accel_ipsec_counters_count(struct mlx5_core_dev *mdev); +int mlx5_accel_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters, + unsigned int count); + +int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev); +void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev); + +#else + +#define MLX5_IPSEC_DEV(mdev) false + +static inline int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev) +{ + return 0; +} + +static inline void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev) +{ +} + +#endif + +#endif /* __MLX5_ACCEL_IPSEC_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/alloc.c b/drivers/net/ethernet/mellanox/mlx5/core/alloc.c index 66bd213f35ce..3c95f7f53802 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/alloc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/alloc.c @@ -274,7 +274,6 @@ void mlx5_db_free(struct mlx5_core_dev *dev, struct mlx5_db *db) } EXPORT_SYMBOL_GPL(mlx5_db_free); - void mlx5_fill_page_array(struct mlx5_buf *buf, __be64 *pas) { u64 addr; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c index 5bdaf3d545b2..f5a2c605749f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c @@ -217,7 +217,6 @@ static void free_cmd(struct mlx5_cmd_work_ent *ent) kfree(ent); } - static int verify_signature(struct mlx5_cmd_work_ent *ent) { struct mlx5_cmd_mailbox *next = ent->out->next; @@ -308,6 +307,7 @@ static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op, case MLX5_CMD_OP_SET_FLOW_TABLE_ROOT: case MLX5_CMD_OP_DEALLOC_ENCAP_HEADER: case MLX5_CMD_OP_DEALLOC_MODIFY_HEADER_CONTEXT: + case MLX5_CMD_OP_FPGA_DESTROY_QP: return MLX5_CMD_STAT_OK; case MLX5_CMD_OP_QUERY_HCA_CAP: @@ -420,6 +420,10 @@ static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op, case MLX5_CMD_OP_QUERY_FLOW_COUNTER: case MLX5_CMD_OP_ALLOC_ENCAP_HEADER: case MLX5_CMD_OP_ALLOC_MODIFY_HEADER_CONTEXT: + case MLX5_CMD_OP_FPGA_CREATE_QP: + case MLX5_CMD_OP_FPGA_MODIFY_QP: + case MLX5_CMD_OP_FPGA_QUERY_QP: + case MLX5_CMD_OP_FPGA_QUERY_QP_COUNTERS: *status = MLX5_DRIVER_STATUS_ABORTED; *synd = MLX5_DRIVER_SYND; return -EIO; @@ -586,6 +590,11 @@ const char *mlx5_command_str(int command) MLX5_COMMAND_STR_CASE(DEALLOC_ENCAP_HEADER); MLX5_COMMAND_STR_CASE(ALLOC_MODIFY_HEADER_CONTEXT); MLX5_COMMAND_STR_CASE(DEALLOC_MODIFY_HEADER_CONTEXT); + MLX5_COMMAND_STR_CASE(FPGA_CREATE_QP); + MLX5_COMMAND_STR_CASE(FPGA_MODIFY_QP); + MLX5_COMMAND_STR_CASE(FPGA_QUERY_QP); + MLX5_COMMAND_STR_CASE(FPGA_QUERY_QP_COUNTERS); + MLX5_COMMAND_STR_CASE(FPGA_DESTROY_QP); default: return "unknown command opcode"; } } @@ -774,7 +783,7 @@ static void cb_timeout_handler(struct work_struct *work) mlx5_core_warn(dev, "%s(0x%x) timeout. Will cause a leak of a command resource\n", mlx5_command_str(msg_to_opcode(ent->in)), msg_to_opcode(ent->in)); - mlx5_cmd_comp_handler(dev, 1UL << ent->idx); + mlx5_cmd_comp_handler(dev, 1UL << ent->idx, true); } static void cmd_work_handler(struct work_struct *work) @@ -786,6 +795,8 @@ static void cmd_work_handler(struct work_struct *work) struct mlx5_cmd_layout *lay; struct semaphore *sem; unsigned long flags; + bool poll_cmd = ent->polling; + sem = ent->page_queue ? &cmd->pages_sem : &cmd->sem; down(sem); @@ -804,6 +815,7 @@ static void cmd_work_handler(struct work_struct *work) } cmd->ent_arr[ent->idx] = ent; + set_bit(MLX5_CMD_ENT_STATE_PENDING_COMP, &ent->state); lay = get_inst(cmd, ent->idx); ent->lay = lay; memset(lay, 0, sizeof(*lay)); @@ -825,17 +837,31 @@ static void cmd_work_handler(struct work_struct *work) if (ent->callback) schedule_delayed_work(&ent->cb_timeout_work, cb_timeout); + /* Skip sending command to fw if internal error */ + if (pci_channel_offline(dev->pdev) || + dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) { + u8 status = 0; + u32 drv_synd; + + ent->ret = mlx5_internal_err_ret_value(dev, msg_to_opcode(ent->in), &drv_synd, &status); + MLX5_SET(mbox_out, ent->out, status, status); + MLX5_SET(mbox_out, ent->out, syndrome, drv_synd); + + mlx5_cmd_comp_handler(dev, 1UL << ent->idx, true); + return; + } + /* ring doorbell after the descriptor is valid */ mlx5_core_dbg(dev, "writing 0x%x to command doorbell\n", 1 << ent->idx); wmb(); iowrite32be(1 << ent->idx, &dev->iseg->cmd_dbell); mmiowb(); /* if not in polling don't use ent after this point */ - if (cmd->mode == CMD_MODE_POLLING) { + if (cmd->mode == CMD_MODE_POLLING || poll_cmd) { poll_timeout(ent); /* make sure we read the descriptor after ownership is SW */ rmb(); - mlx5_cmd_comp_handler(dev, 1UL << ent->idx); + mlx5_cmd_comp_handler(dev, 1UL << ent->idx, (ent->ret == -ETIMEDOUT)); } } @@ -859,7 +885,7 @@ static const char *deliv_status_to_str(u8 status) case MLX5_CMD_DELIVERY_STAT_IN_LENGTH_ERR: return "command input length error"; case MLX5_CMD_DELIVERY_STAT_OUT_LENGTH_ERR: - return "command ouput length error"; + return "command output length error"; case MLX5_CMD_DELIVERY_STAT_RES_FLD_NOT_CLR_ERR: return "reserved fields not cleared"; case MLX5_CMD_DELIVERY_STAT_CMD_DESCR_ERR: @@ -875,11 +901,11 @@ static int wait_func(struct mlx5_core_dev *dev, struct mlx5_cmd_work_ent *ent) struct mlx5_cmd *cmd = &dev->cmd; int err; - if (cmd->mode == CMD_MODE_POLLING) { + if (cmd->mode == CMD_MODE_POLLING || ent->polling) { wait_for_completion(&ent->done); } else if (!wait_for_completion_timeout(&ent->done, timeout)) { ent->ret = -ETIMEDOUT; - mlx5_cmd_comp_handler(dev, 1UL << ent->idx); + mlx5_cmd_comp_handler(dev, 1UL << ent->idx, true); } err = ent->ret; @@ -903,7 +929,7 @@ static int mlx5_cmd_invoke(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *in, struct mlx5_cmd_msg *out, void *uout, int uout_size, mlx5_cmd_cbk_t callback, void *context, int page_queue, u8 *status, - u8 token) + u8 token, bool force_polling) { struct mlx5_cmd *cmd = &dev->cmd; struct mlx5_cmd_work_ent *ent; @@ -921,6 +947,7 @@ static int mlx5_cmd_invoke(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *in, return PTR_ERR(ent); ent->token = token; + ent->polling = force_polling; if (!callback) init_completion(&ent->done); @@ -986,7 +1013,6 @@ static ssize_t dbg_write(struct file *filp, const char __user *buf, return err ? err : count; } - static const struct file_operations fops = { .owner = THIS_MODULE, .open = simple_open, @@ -1138,7 +1164,7 @@ err_alloc: } static void mlx5_free_cmd_msg(struct mlx5_core_dev *dev, - struct mlx5_cmd_msg *msg) + struct mlx5_cmd_msg *msg) { struct mlx5_cmd_mailbox *head = msg->next; struct mlx5_cmd_mailbox *next; @@ -1375,7 +1401,7 @@ static void free_msg(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *msg) } } -void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec) +void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool forced) { struct mlx5_cmd *cmd = &dev->cmd; struct mlx5_cmd_work_ent *ent; @@ -1395,6 +1421,19 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec) struct semaphore *sem; ent = cmd->ent_arr[i]; + + /* if we already completed the command, ignore it */ + if (!test_and_clear_bit(MLX5_CMD_ENT_STATE_PENDING_COMP, + &ent->state)) { + /* only real completion can free the cmd slot */ + if (!forced) { + mlx5_core_err(dev, "Command completion arrived after timeout (entry idx = %d).\n", + ent->idx); + free_ent(cmd, ent->idx); + } + continue; + } + if (ent->callback) cancel_delayed_work(&ent->cb_timeout_work); if (ent->page_queue) @@ -1417,7 +1456,10 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec) mlx5_core_dbg(dev, "command completed. ret 0x%x, delivery status %s(0x%x)\n", ent->ret, deliv_status_to_str(ent->status), ent->status); } - free_ent(cmd, ent->idx); + + /* only real completion will free the entry slot */ + if (!forced) + free_ent(cmd, ent->idx); if (ent->callback) { ds = ent->ts2 - ent->ts1; @@ -1506,7 +1548,8 @@ static int is_manage_pages(void *in) } static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out, - int out_size, mlx5_cmd_cbk_t callback, void *context) + int out_size, mlx5_cmd_cbk_t callback, void *context, + bool force_polling) { struct mlx5_cmd_msg *inb; struct mlx5_cmd_msg *outb; @@ -1551,7 +1594,7 @@ static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out, } err = mlx5_cmd_invoke(dev, inb, outb, out, out_size, callback, context, - pages_queue, &status, token); + pages_queue, &status, token, force_polling); if (err) goto out_out; @@ -1579,7 +1622,7 @@ int mlx5_cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out, { int err; - err = cmd_exec(dev, in, in_size, out, out_size, NULL, NULL); + err = cmd_exec(dev, in, in_size, out, out_size, NULL, NULL, false); return err ? : mlx5_cmd_check(dev, in, out); } EXPORT_SYMBOL(mlx5_cmd_exec); @@ -1588,10 +1631,22 @@ int mlx5_cmd_exec_cb(struct mlx5_core_dev *dev, void *in, int in_size, void *out, int out_size, mlx5_cmd_cbk_t callback, void *context) { - return cmd_exec(dev, in, in_size, out, out_size, callback, context); + return cmd_exec(dev, in, in_size, out, out_size, callback, context, + false); } EXPORT_SYMBOL(mlx5_cmd_exec_cb); +int mlx5_cmd_exec_polling(struct mlx5_core_dev *dev, void *in, int in_size, + void *out, int out_size) +{ + int err; + + err = cmd_exec(dev, in, in_size, out, out_size, NULL, NULL, true); + + return err ? : mlx5_cmd_check(dev, in, out); +} +EXPORT_SYMBOL(mlx5_cmd_exec_polling); + static void destroy_msg_cache(struct mlx5_core_dev *dev) { struct cmd_msg_cache *ch; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c index e94a9532e218..7ecadb501743 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/debugfs.c @@ -168,7 +168,6 @@ static ssize_t average_read(struct file *filp, char __user *buf, size_t count, return ret; } - static ssize_t average_write(struct file *filp, const char __user *buf, size_t count, loff_t *pos) { @@ -405,7 +404,7 @@ static u64 cq_read_field(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, u32 *out; int err; - out = mlx5_vzalloc(outlen); + out = kvzalloc(outlen, GFP_KERNEL); if (!out) return param; @@ -466,7 +465,6 @@ static ssize_t dbg_read(struct file *filp, char __user *buf, size_t count, return -EINVAL; } - if (is_str) ret = snprintf(tbuf, sizeof(tbuf), "%s\n", (const char *)(unsigned long)field); else @@ -562,7 +560,6 @@ void mlx5_debug_qp_remove(struct mlx5_core_dev *dev, struct mlx5_core_qp *qp) rem_res_tree(qp->dbg); } - int mlx5_debug_eq_add(struct mlx5_core_dev *dev, struct mlx5_eq *eq) { int err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 0099a3e397bc..e1b7ddfecd01 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -52,8 +52,10 @@ #define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v) -#define MLX5E_HW2SW_MTU(hwmtu) ((hwmtu) - (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN)) -#define MLX5E_SW2HW_MTU(swmtu) ((swmtu) + (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN)) +#define MLX5E_ETH_HARD_MTU (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN) + +#define MLX5E_HW2SW_MTU(priv, hwmtu) ((hwmtu) - ((priv)->hard_mtu)) +#define MLX5E_SW2HW_MTU(priv, swmtu) ((swmtu) + ((priv)->hard_mtu)) #define MLX5E_MAX_NUM_TC 8 @@ -70,6 +72,8 @@ #define MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE_MPW 0x6 #define MLX5_RX_HEADROOM NET_SKB_PAD +#define MLX5_SKB_FRAG_SZ(len) (SKB_DATA_ALIGN(len) + \ + SKB_DATA_ALIGN(sizeof(struct skb_shared_info))) #define MLX5_MPWRQ_MIN_LOG_STRIDE_SZ(mdev) \ (6 + MLX5_CAP_GEN(mdev, cache_line_128byte)) /* HW restriction */ @@ -213,6 +217,7 @@ struct mlx5e_cq_moder { struct mlx5e_params { u8 log_sq_size; u8 rq_wq_type; + u16 rq_headroom; u8 mpwqe_log_stride_sz; u8 mpwqe_log_num_strides; u8 log_rq_size; @@ -323,6 +328,7 @@ struct mlx5e_sq_dma { enum { MLX5E_SQ_STATE_ENABLED, + MLX5E_SQ_STATE_IPSEC, }; struct mlx5e_sq_wqe_info { @@ -443,6 +449,11 @@ struct mlx5e_dma_info { dma_addr_t addr; }; +struct mlx5e_wqe_frag_info { + struct mlx5e_dma_info di; + u32 offset; +}; + struct mlx5e_umr_dma_info { __be64 *mtt; dma_addr_t mtt_addr; @@ -458,13 +469,15 @@ struct mlx5e_mpw_info { struct mlx5e_rx_am_stats { int ppms; /* packets per msec */ + int bpms; /* bytes per msec */ int epms; /* events per msec */ }; struct mlx5e_rx_am_sample { - ktime_t time; - unsigned int pkt_ctr; - u16 event_ctr; + ktime_t time; + u32 pkt_ctr; + u32 byte_ctr; + u16 event_ctr; }; struct mlx5e_rx_am { /* Adaptive Moderation */ @@ -502,7 +515,12 @@ struct mlx5e_rq { struct mlx5_wq_ll wq; union { - struct mlx5e_dma_info *dma_info; + struct { + struct mlx5e_wqe_frag_info *frag_info; + u32 frag_sz; /* max possible skb frag_sz */ + bool page_reuse; + bool xdp_xmit; + } wqe; struct { struct mlx5e_mpw_info *info; void *mtt_no_align; @@ -623,6 +641,8 @@ struct mlx5e_tc_table { struct rhashtable_params ht_params; struct rhashtable ht; + + DECLARE_HASHTABLE(mod_hdr_tbl, 8); }; struct mlx5e_vlan_table { @@ -743,6 +763,7 @@ struct mlx5e_priv { struct mlx5e_tir indir_tir[MLX5E_NUM_INDIR_TIRS]; struct mlx5e_tir direct_tir[MLX5E_MAX_NUM_CHANNELS]; u32 tx_rates[MLX5E_MAX_NUM_SQS]; + int hard_mtu; struct mlx5e_flow_steering fs; struct mlx5e_vxlan_db vxlan; @@ -764,6 +785,9 @@ struct mlx5e_priv { const struct mlx5e_profile *profile; void *ppriv; +#ifdef CONFIG_MLX5_EN_IPSEC + struct mlx5e_ipsec *ipsec; +#endif }; struct mlx5e_profile { @@ -778,6 +802,7 @@ struct mlx5e_profile { void (*enable)(struct mlx5e_priv *priv); void (*disable)(struct mlx5e_priv *priv); void (*update_stats)(struct mlx5e_priv *priv); + void (*update_carrier)(struct mlx5e_priv *priv); int (*max_nch)(struct mlx5_core_dev *mdev); struct { mlx5e_fp_handle_rx_cqe handle_rx_cqe; @@ -812,13 +837,12 @@ void mlx5e_dealloc_rx_wqe(struct mlx5e_rq *rq, u16 ix); void mlx5e_dealloc_rx_mpwqe(struct mlx5e_rq *rq, u16 ix); void mlx5e_post_rx_mpwqe(struct mlx5e_rq *rq); void mlx5e_free_rx_mpwqe(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi); -struct mlx5_cqe64 *mlx5e_get_cqe(struct mlx5e_cq *cq); void mlx5e_rx_am(struct mlx5e_rq *rq); void mlx5e_rx_am_work(struct work_struct *work); struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode); -void mlx5e_update_stats(struct mlx5e_priv *priv); +void mlx5e_update_stats(struct mlx5e_priv *priv, bool full); int mlx5e_create_flow_steering(struct mlx5e_priv *priv); void mlx5e_destroy_flow_steering(struct mlx5e_priv *priv); @@ -845,8 +869,8 @@ void mlx5e_timestamp_init(struct mlx5e_priv *priv); void mlx5e_timestamp_cleanup(struct mlx5e_priv *priv); void mlx5e_pps_event_handler(struct mlx5e_priv *priv, struct ptp_clock_event *event); -int mlx5e_hwstamp_set(struct net_device *dev, struct ifreq *ifr); -int mlx5e_hwstamp_get(struct net_device *dev, struct ifreq *ifr); +int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr); +int mlx5e_hwstamp_get(struct mlx5e_priv *priv, struct ifreq *ifr); int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool val); int mlx5e_vlan_rx_add_vid(struct net_device *dev, __always_unused __be16 proto, @@ -1003,7 +1027,7 @@ int mlx5e_create_direct_tirs(struct mlx5e_priv *priv); void mlx5e_destroy_direct_tirs(struct mlx5e_priv *priv); void mlx5e_destroy_rqt(struct mlx5e_priv *priv, struct mlx5e_rqt *rqt); -int mlx5e_create_ttc_table(struct mlx5e_priv *priv, u32 underlay_qpn); +int mlx5e_create_ttc_table(struct mlx5e_priv *priv); void mlx5e_destroy_ttc_table(struct mlx5e_priv *priv); int mlx5e_create_tis(struct mlx5_core_dev *mdev, int tc, @@ -1017,6 +1041,31 @@ int mlx5e_open(struct net_device *netdev); void mlx5e_update_stats_work(struct work_struct *work); u32 mlx5e_choose_lro_timeout(struct mlx5_core_dev *mdev, u32 wanted_timeout); +/* ethtool helpers */ +void mlx5e_ethtool_get_drvinfo(struct mlx5e_priv *priv, + struct ethtool_drvinfo *drvinfo); +void mlx5e_ethtool_get_strings(struct mlx5e_priv *priv, + uint32_t stringset, uint8_t *data); +int mlx5e_ethtool_get_sset_count(struct mlx5e_priv *priv, int sset); +void mlx5e_ethtool_get_ethtool_stats(struct mlx5e_priv *priv, + struct ethtool_stats *stats, u64 *data); +void mlx5e_ethtool_get_ringparam(struct mlx5e_priv *priv, + struct ethtool_ringparam *param); +int mlx5e_ethtool_set_ringparam(struct mlx5e_priv *priv, + struct ethtool_ringparam *param); +void mlx5e_ethtool_get_channels(struct mlx5e_priv *priv, + struct ethtool_channels *ch); +int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv, + struct ethtool_channels *ch); +int mlx5e_ethtool_get_coalesce(struct mlx5e_priv *priv, + struct ethtool_coalesce *coal); +int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv, + struct ethtool_coalesce *coal); +int mlx5e_ethtool_get_ts_info(struct mlx5e_priv *priv, + struct ethtool_ts_info *info); +int mlx5e_ethtool_flash_device(struct mlx5e_priv *priv, + struct ethtool_flash *flash); + /* mlx5e generic netdev management API */ struct net_device* mlx5e_create_netdev(struct mlx5_core_dev *mdev, const struct mlx5e_profile *profile, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/Makefile new file mode 100644 index 000000000000..d8e17110f25d --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/Makefile @@ -0,0 +1 @@ +subdir-ccflags-y += -I$(src)/.. diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c new file mode 100644 index 000000000000..bac5103efad3 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c @@ -0,0 +1,461 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <crypto/internal/geniv.h> +#include <crypto/aead.h> +#include <linux/inetdevice.h> +#include <linux/netdevice.h> +#include <linux/module.h> + +#include "en.h" +#include "accel/ipsec.h" +#include "en_accel/ipsec.h" +#include "en_accel/ipsec_rxtx.h" + +struct mlx5e_ipsec_sa_entry { + struct hlist_node hlist; /* Item in SADB_RX hashtable */ + unsigned int handle; /* Handle in SADB_RX */ + struct xfrm_state *x; + struct mlx5e_ipsec *ipsec; + void *context; +}; + +struct xfrm_state *mlx5e_ipsec_sadb_rx_lookup(struct mlx5e_ipsec *ipsec, + unsigned int handle) +{ + struct mlx5e_ipsec_sa_entry *sa_entry; + struct xfrm_state *ret = NULL; + + rcu_read_lock(); + hash_for_each_possible_rcu(ipsec->sadb_rx, sa_entry, hlist, handle) + if (sa_entry->handle == handle) { + ret = sa_entry->x; + xfrm_state_hold(ret); + break; + } + rcu_read_unlock(); + + return ret; +} + +static int mlx5e_ipsec_sadb_rx_add(struct mlx5e_ipsec_sa_entry *sa_entry) +{ + struct mlx5e_ipsec *ipsec = sa_entry->ipsec; + unsigned long flags; + int ret; + + spin_lock_irqsave(&ipsec->sadb_rx_lock, flags); + ret = ida_simple_get(&ipsec->halloc, 1, 0, GFP_KERNEL); + if (ret < 0) + goto out; + + sa_entry->handle = ret; + hash_add_rcu(ipsec->sadb_rx, &sa_entry->hlist, sa_entry->handle); + ret = 0; + +out: + spin_unlock_irqrestore(&ipsec->sadb_rx_lock, flags); + return ret; +} + +static void mlx5e_ipsec_sadb_rx_del(struct mlx5e_ipsec_sa_entry *sa_entry) +{ + struct mlx5e_ipsec *ipsec = sa_entry->ipsec; + unsigned long flags; + + spin_lock_irqsave(&ipsec->sadb_rx_lock, flags); + hash_del_rcu(&sa_entry->hlist); + spin_unlock_irqrestore(&ipsec->sadb_rx_lock, flags); +} + +static void mlx5e_ipsec_sadb_rx_free(struct mlx5e_ipsec_sa_entry *sa_entry) +{ + struct mlx5e_ipsec *ipsec = sa_entry->ipsec; + unsigned long flags; + + /* Wait for the hash_del_rcu call in sadb_rx_del to affect data path */ + synchronize_rcu(); + spin_lock_irqsave(&ipsec->sadb_rx_lock, flags); + ida_simple_remove(&ipsec->halloc, sa_entry->handle); + spin_unlock_irqrestore(&ipsec->sadb_rx_lock, flags); +} + +static enum mlx5_accel_ipsec_enc_mode mlx5e_ipsec_enc_mode(struct xfrm_state *x) +{ + unsigned int key_len = (x->aead->alg_key_len + 7) / 8 - 4; + + switch (key_len) { + case 16: + return MLX5_IPSEC_SADB_MODE_AES_GCM_128_AUTH_128; + case 32: + return MLX5_IPSEC_SADB_MODE_AES_GCM_256_AUTH_128; + default: + netdev_warn(x->xso.dev, "Bad key len: %d for alg %s\n", + key_len, x->aead->alg_name); + return -1; + } +} + +static void mlx5e_ipsec_build_hw_sa(u32 op, struct mlx5e_ipsec_sa_entry *sa_entry, + struct mlx5_accel_ipsec_sa *hw_sa) +{ + struct xfrm_state *x = sa_entry->x; + struct aead_geniv_ctx *geniv_ctx; + unsigned int crypto_data_len; + struct crypto_aead *aead; + unsigned int key_len; + int ivsize; + + memset(hw_sa, 0, sizeof(*hw_sa)); + + if (op == MLX5_IPSEC_CMD_ADD_SA) { + crypto_data_len = (x->aead->alg_key_len + 7) / 8; + key_len = crypto_data_len - 4; /* 4 bytes salt at end */ + aead = x->data; + geniv_ctx = crypto_aead_ctx(aead); + ivsize = crypto_aead_ivsize(aead); + + memcpy(&hw_sa->key_enc, x->aead->alg_key, key_len); + /* Duplicate 128 bit key twice according to HW layout */ + if (key_len == 16) + memcpy(&hw_sa->key_enc[16], x->aead->alg_key, key_len); + memcpy(&hw_sa->gcm.salt_iv, geniv_ctx->salt, ivsize); + hw_sa->gcm.salt = *((__be32 *)(x->aead->alg_key + key_len)); + } + + hw_sa->cmd = htonl(op); + hw_sa->flags |= MLX5_IPSEC_SADB_SA_VALID | MLX5_IPSEC_SADB_SPI_EN; + if (x->props.family == AF_INET) { + hw_sa->sip[3] = x->props.saddr.a4; + hw_sa->dip[3] = x->id.daddr.a4; + hw_sa->sip_masklen = 32; + hw_sa->dip_masklen = 32; + } else { + memcpy(hw_sa->sip, x->props.saddr.a6, sizeof(hw_sa->sip)); + memcpy(hw_sa->dip, x->id.daddr.a6, sizeof(hw_sa->dip)); + hw_sa->sip_masklen = 128; + hw_sa->dip_masklen = 128; + hw_sa->flags |= MLX5_IPSEC_SADB_IPV6; + } + hw_sa->spi = x->id.spi; + hw_sa->sw_sa_handle = htonl(sa_entry->handle); + switch (x->id.proto) { + case IPPROTO_ESP: + hw_sa->flags |= MLX5_IPSEC_SADB_IP_ESP; + break; + case IPPROTO_AH: + hw_sa->flags |= MLX5_IPSEC_SADB_IP_AH; + break; + default: + break; + } + hw_sa->enc_mode = mlx5e_ipsec_enc_mode(x); + if (!(x->xso.flags & XFRM_OFFLOAD_INBOUND)) + hw_sa->flags |= MLX5_IPSEC_SADB_DIR_SX; +} + +static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x) +{ + struct net_device *netdev = x->xso.dev; + struct mlx5e_priv *priv; + + priv = netdev_priv(netdev); + + if (x->props.aalgo != SADB_AALG_NONE) { + netdev_info(netdev, "Cannot offload authenticated xfrm states\n"); + return -EINVAL; + } + if (x->props.ealgo != SADB_X_EALG_AES_GCM_ICV16) { + netdev_info(netdev, "Only AES-GCM-ICV16 xfrm state may be offloaded\n"); + return -EINVAL; + } + if (x->props.calgo != SADB_X_CALG_NONE) { + netdev_info(netdev, "Cannot offload compressed xfrm states\n"); + return -EINVAL; + } + if (x->props.flags & XFRM_STATE_ESN) { + netdev_info(netdev, "Cannot offload ESN xfrm states\n"); + return -EINVAL; + } + if (x->props.family != AF_INET && + x->props.family != AF_INET6) { + netdev_info(netdev, "Only IPv4/6 xfrm states may be offloaded\n"); + return -EINVAL; + } + if (x->props.mode != XFRM_MODE_TRANSPORT && + x->props.mode != XFRM_MODE_TUNNEL) { + dev_info(&netdev->dev, "Only transport and tunnel xfrm states may be offloaded\n"); + return -EINVAL; + } + if (x->id.proto != IPPROTO_ESP) { + netdev_info(netdev, "Only ESP xfrm state may be offloaded\n"); + return -EINVAL; + } + if (x->encap) { + netdev_info(netdev, "Encapsulated xfrm state may not be offloaded\n"); + return -EINVAL; + } + if (!x->aead) { + netdev_info(netdev, "Cannot offload xfrm states without aead\n"); + return -EINVAL; + } + if (x->aead->alg_icv_len != 128) { + netdev_info(netdev, "Cannot offload xfrm states with AEAD ICV length other than 128bit\n"); + return -EINVAL; + } + if ((x->aead->alg_key_len != 128 + 32) && + (x->aead->alg_key_len != 256 + 32)) { + netdev_info(netdev, "Cannot offload xfrm states with AEAD key length other than 128/256 bit\n"); + return -EINVAL; + } + if (x->tfcpad) { + netdev_info(netdev, "Cannot offload xfrm states with tfc padding\n"); + return -EINVAL; + } + if (!x->geniv) { + netdev_info(netdev, "Cannot offload xfrm states without geniv\n"); + return -EINVAL; + } + if (strcmp(x->geniv, "seqiv")) { + netdev_info(netdev, "Cannot offload xfrm states with geniv other than seqiv\n"); + return -EINVAL; + } + if (x->props.family == AF_INET6 && + !(mlx5_accel_ipsec_device_caps(priv->mdev) & MLX5_ACCEL_IPSEC_IPV6)) { + netdev_info(netdev, "IPv6 xfrm state offload is not supported by this device\n"); + return -EINVAL; + } + return 0; +} + +static int mlx5e_xfrm_add_state(struct xfrm_state *x) +{ + struct mlx5e_ipsec_sa_entry *sa_entry = NULL; + struct net_device *netdev = x->xso.dev; + struct mlx5_accel_ipsec_sa hw_sa; + struct mlx5e_priv *priv; + void *context; + int err; + + priv = netdev_priv(netdev); + + err = mlx5e_xfrm_validate_state(x); + if (err) + return err; + + sa_entry = kzalloc(sizeof(*sa_entry), GFP_KERNEL); + if (!sa_entry) { + err = -ENOMEM; + goto out; + } + + sa_entry->x = x; + sa_entry->ipsec = priv->ipsec; + + /* Add the SA to handle processed incoming packets before the add SA + * completion was received + */ + if (x->xso.flags & XFRM_OFFLOAD_INBOUND) { + err = mlx5e_ipsec_sadb_rx_add(sa_entry); + if (err) { + netdev_info(netdev, "Failed adding to SADB_RX: %d\n", err); + goto err_entry; + } + } + + mlx5e_ipsec_build_hw_sa(MLX5_IPSEC_CMD_ADD_SA, sa_entry, &hw_sa); + context = mlx5_accel_ipsec_sa_cmd_exec(sa_entry->ipsec->en_priv->mdev, &hw_sa); + if (IS_ERR(context)) { + err = PTR_ERR(context); + goto err_sadb_rx; + } + + err = mlx5_accel_ipsec_sa_cmd_wait(context); + if (err) + goto err_sadb_rx; + + x->xso.offload_handle = (unsigned long)sa_entry; + goto out; + +err_sadb_rx: + if (x->xso.flags & XFRM_OFFLOAD_INBOUND) { + mlx5e_ipsec_sadb_rx_del(sa_entry); + mlx5e_ipsec_sadb_rx_free(sa_entry); + } +err_entry: + kfree(sa_entry); +out: + return err; +} + +static void mlx5e_xfrm_del_state(struct xfrm_state *x) +{ + struct mlx5e_ipsec_sa_entry *sa_entry; + struct mlx5_accel_ipsec_sa hw_sa; + void *context; + + if (!x->xso.offload_handle) + return; + + sa_entry = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle; + WARN_ON(sa_entry->x != x); + + if (x->xso.flags & XFRM_OFFLOAD_INBOUND) + mlx5e_ipsec_sadb_rx_del(sa_entry); + + mlx5e_ipsec_build_hw_sa(MLX5_IPSEC_CMD_DEL_SA, sa_entry, &hw_sa); + context = mlx5_accel_ipsec_sa_cmd_exec(sa_entry->ipsec->en_priv->mdev, &hw_sa); + if (IS_ERR(context)) + return; + + sa_entry->context = context; +} + +static void mlx5e_xfrm_free_state(struct xfrm_state *x) +{ + struct mlx5e_ipsec_sa_entry *sa_entry; + int res; + + if (!x->xso.offload_handle) + return; + + sa_entry = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle; + WARN_ON(sa_entry->x != x); + + res = mlx5_accel_ipsec_sa_cmd_wait(sa_entry->context); + sa_entry->context = NULL; + if (res) { + /* Leftover object will leak */ + return; + } + + if (x->xso.flags & XFRM_OFFLOAD_INBOUND) + mlx5e_ipsec_sadb_rx_free(sa_entry); + + kfree(sa_entry); +} + +int mlx5e_ipsec_init(struct mlx5e_priv *priv) +{ + struct mlx5e_ipsec *ipsec = NULL; + + if (!MLX5_IPSEC_DEV(priv->mdev)) { + netdev_dbg(priv->netdev, "Not an IPSec offload device\n"); + return 0; + } + + ipsec = kzalloc(sizeof(*ipsec), GFP_KERNEL); + if (!ipsec) + return -ENOMEM; + + hash_init(ipsec->sadb_rx); + spin_lock_init(&ipsec->sadb_rx_lock); + ida_init(&ipsec->halloc); + ipsec->en_priv = priv; + ipsec->en_priv->ipsec = ipsec; + netdev_dbg(priv->netdev, "IPSec attached to netdevice\n"); + return 0; +} + +void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv) +{ + struct mlx5e_ipsec *ipsec = priv->ipsec; + + if (!ipsec) + return; + + ida_destroy(&ipsec->halloc); + kfree(ipsec); + priv->ipsec = NULL; +} + +static bool mlx5e_ipsec_offload_ok(struct sk_buff *skb, struct xfrm_state *x) +{ + if (x->props.family == AF_INET) { + /* Offload with IPv4 options is not supported yet */ + if (ip_hdr(skb)->ihl > 5) + return false; + } else { + /* Offload with IPv6 extension headers is not support yet */ + if (ipv6_ext_hdr(ipv6_hdr(skb)->nexthdr)) + return false; + } + + return true; +} + +static const struct xfrmdev_ops mlx5e_ipsec_xfrmdev_ops = { + .xdo_dev_state_add = mlx5e_xfrm_add_state, + .xdo_dev_state_delete = mlx5e_xfrm_del_state, + .xdo_dev_state_free = mlx5e_xfrm_free_state, + .xdo_dev_offload_ok = mlx5e_ipsec_offload_ok, +}; + +void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv) +{ + struct mlx5_core_dev *mdev = priv->mdev; + struct net_device *netdev = priv->netdev; + + if (!priv->ipsec) + return; + + if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_ESP) || + !MLX5_CAP_ETH(mdev, swp)) { + mlx5_core_dbg(mdev, "mlx5e: ESP and SWP offload not supported\n"); + return; + } + + mlx5_core_info(mdev, "mlx5e: IPSec ESP acceleration enabled\n"); + netdev->xfrmdev_ops = &mlx5e_ipsec_xfrmdev_ops; + netdev->features |= NETIF_F_HW_ESP; + netdev->hw_enc_features |= NETIF_F_HW_ESP; + + if (!MLX5_CAP_ETH(mdev, swp_csum)) { + mlx5_core_dbg(mdev, "mlx5e: SWP checksum not supported\n"); + return; + } + + netdev->features |= NETIF_F_HW_ESP_TX_CSUM; + netdev->hw_enc_features |= NETIF_F_HW_ESP_TX_CSUM; + + if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_LSO) || + !MLX5_CAP_ETH(mdev, swp_lso)) { + mlx5_core_dbg(mdev, "mlx5e: ESP LSO not supported\n"); + return; + } + + mlx5_core_dbg(mdev, "mlx5e: ESP GSO capability turned on\n"); + netdev->features |= NETIF_F_GSO_ESP; + netdev->hw_features |= NETIF_F_GSO_ESP; + netdev->hw_enc_features |= NETIF_F_GSO_ESP; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h new file mode 100644 index 000000000000..56e00baf16cc --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h @@ -0,0 +1,140 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef __MLX5E_IPSEC_H__ +#define __MLX5E_IPSEC_H__ + +#ifdef CONFIG_MLX5_EN_IPSEC + +#include <linux/mlx5/device.h> +#include <net/xfrm.h> +#include <linux/idr.h> + +#define MLX5E_IPSEC_SADB_RX_BITS 10 +#define MLX5E_METADATA_ETHER_TYPE (0x8CE4) +#define MLX5E_METADATA_ETHER_LEN 8 + +struct mlx5e_priv; + +struct mlx5e_ipsec_sw_stats { + atomic64_t ipsec_rx_drop_sp_alloc; + atomic64_t ipsec_rx_drop_sadb_miss; + atomic64_t ipsec_rx_drop_syndrome; + atomic64_t ipsec_tx_drop_bundle; + atomic64_t ipsec_tx_drop_no_state; + atomic64_t ipsec_tx_drop_not_ip; + atomic64_t ipsec_tx_drop_trailer; + atomic64_t ipsec_tx_drop_metadata; +}; + +struct mlx5e_ipsec_stats { + u64 ipsec_dec_in_packets; + u64 ipsec_dec_out_packets; + u64 ipsec_dec_bypass_packets; + u64 ipsec_enc_in_packets; + u64 ipsec_enc_out_packets; + u64 ipsec_enc_bypass_packets; + u64 ipsec_dec_drop_packets; + u64 ipsec_dec_auth_fail_packets; + u64 ipsec_enc_drop_packets; + u64 ipsec_add_sa_success; + u64 ipsec_add_sa_fail; + u64 ipsec_del_sa_success; + u64 ipsec_del_sa_fail; + u64 ipsec_cmd_drop; +}; + +struct mlx5e_ipsec { + struct mlx5e_priv *en_priv; + DECLARE_HASHTABLE(sadb_rx, MLX5E_IPSEC_SADB_RX_BITS); + spinlock_t sadb_rx_lock; /* Protects sadb_rx and halloc */ + struct ida halloc; + struct mlx5e_ipsec_sw_stats sw_stats; + struct mlx5e_ipsec_stats stats; +}; + +void mlx5e_ipsec_build_inverse_table(void); +int mlx5e_ipsec_init(struct mlx5e_priv *priv); +void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv); +void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv); + +int mlx5e_ipsec_get_count(struct mlx5e_priv *priv); +int mlx5e_ipsec_get_strings(struct mlx5e_priv *priv, uint8_t *data); +void mlx5e_ipsec_update_stats(struct mlx5e_priv *priv); +int mlx5e_ipsec_get_stats(struct mlx5e_priv *priv, u64 *data); + +struct xfrm_state *mlx5e_ipsec_sadb_rx_lookup(struct mlx5e_ipsec *dev, + unsigned int handle); + +#else + +static inline void mlx5e_ipsec_build_inverse_table(void) +{ +} + +static inline int mlx5e_ipsec_init(struct mlx5e_priv *priv) +{ + return 0; +} + +static inline void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv) +{ +} + +static inline void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv) +{ +} + +static inline int mlx5e_ipsec_get_count(struct mlx5e_priv *priv) +{ + return 0; +} + +static inline int mlx5e_ipsec_get_strings(struct mlx5e_priv *priv, + uint8_t *data) +{ + return 0; +} + +static inline void mlx5e_ipsec_update_stats(struct mlx5e_priv *priv) +{ +} + +static inline int mlx5e_ipsec_get_stats(struct mlx5e_priv *priv, u64 *data) +{ + return 0; +} + +#endif + +#endif /* __MLX5E_IPSEC_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c new file mode 100644 index 000000000000..4614ddfa91eb --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c @@ -0,0 +1,378 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <crypto/aead.h> +#include <net/xfrm.h> +#include <net/esp.h> + +#include "en_accel/ipsec_rxtx.h" +#include "en_accel/ipsec.h" +#include "en.h" + +enum { + MLX5E_IPSEC_RX_SYNDROME_DECRYPTED = 0x11, + MLX5E_IPSEC_RX_SYNDROME_AUTH_FAILED = 0x12, +}; + +struct mlx5e_ipsec_rx_metadata { + unsigned char reserved; + __be32 sa_handle; +} __packed; + +enum { + MLX5E_IPSEC_TX_SYNDROME_OFFLOAD = 0x8, + MLX5E_IPSEC_TX_SYNDROME_OFFLOAD_WITH_LSO_TCP = 0x9, +}; + +struct mlx5e_ipsec_tx_metadata { + __be16 mss_inv; /* 1/MSS in 16bit fixed point, only for LSO */ + __be16 seq; /* LSBs of the first TCP seq, only for LSO */ + u8 esp_next_proto; /* Next protocol of ESP */ +} __packed; + +struct mlx5e_ipsec_metadata { + unsigned char syndrome; + union { + unsigned char raw[5]; + /* from FPGA to host, on successful decrypt */ + struct mlx5e_ipsec_rx_metadata rx; + /* from host to FPGA */ + struct mlx5e_ipsec_tx_metadata tx; + } __packed content; + /* packet type ID field */ + __be16 ethertype; +} __packed; + +#define MAX_LSO_MSS 2048 + +/* Pre-calculated (Q0.16) fixed-point inverse 1/x function */ +static __be16 mlx5e_ipsec_inverse_table[MAX_LSO_MSS]; + +static inline __be16 mlx5e_ipsec_mss_inv(struct sk_buff *skb) +{ + return mlx5e_ipsec_inverse_table[skb_shinfo(skb)->gso_size]; +} + +static struct mlx5e_ipsec_metadata *mlx5e_ipsec_add_metadata(struct sk_buff *skb) +{ + struct mlx5e_ipsec_metadata *mdata; + struct ethhdr *eth; + + if (unlikely(skb_cow_head(skb, sizeof(*mdata)))) + return ERR_PTR(-ENOMEM); + + eth = (struct ethhdr *)skb_push(skb, sizeof(*mdata)); + skb->mac_header -= sizeof(*mdata); + mdata = (struct mlx5e_ipsec_metadata *)(eth + 1); + + memmove(skb->data, skb->data + sizeof(*mdata), + 2 * ETH_ALEN); + + eth->h_proto = cpu_to_be16(MLX5E_METADATA_ETHER_TYPE); + + memset(mdata->content.raw, 0, sizeof(mdata->content.raw)); + return mdata; +} + +static int mlx5e_ipsec_remove_trailer(struct sk_buff *skb, struct xfrm_state *x) +{ + unsigned int alen = crypto_aead_authsize(x->data); + struct ipv6hdr *ipv6hdr = ipv6_hdr(skb); + struct iphdr *ipv4hdr = ip_hdr(skb); + unsigned int trailer_len; + u8 plen; + int ret; + + ret = skb_copy_bits(skb, skb->len - alen - 2, &plen, 1); + if (unlikely(ret)) + return ret; + + trailer_len = alen + plen + 2; + + pskb_trim(skb, skb->len - trailer_len); + if (skb->protocol == htons(ETH_P_IP)) { + ipv4hdr->tot_len = htons(ntohs(ipv4hdr->tot_len) - trailer_len); + ip_send_check(ipv4hdr); + } else { + ipv6hdr->payload_len = htons(ntohs(ipv6hdr->payload_len) - + trailer_len); + } + return 0; +} + +static void mlx5e_ipsec_set_swp(struct sk_buff *skb, + struct mlx5_wqe_eth_seg *eseg, u8 mode, + struct xfrm_offload *xo) +{ + u8 proto; + + /* Tunnel Mode: + * SWP: OutL3 InL3 InL4 + * Pkt: MAC IP ESP IP L4 + * + * Transport Mode: + * SWP: OutL3 InL4 + * InL3 + * Pkt: MAC IP ESP L4 + * + * Offsets are in 2-byte words, counting from start of frame + */ + eseg->swp_outer_l3_offset = skb_network_offset(skb) / 2; + if (skb->protocol == htons(ETH_P_IPV6)) + eseg->swp_flags |= MLX5_ETH_WQE_SWP_OUTER_L3_IPV6; + + if (mode == XFRM_MODE_TUNNEL) { + eseg->swp_inner_l3_offset = skb_inner_network_offset(skb) / 2; + if (xo->proto == IPPROTO_IPV6) { + eseg->swp_flags |= MLX5_ETH_WQE_SWP_INNER_L3_IPV6; + proto = inner_ipv6_hdr(skb)->nexthdr; + } else { + proto = inner_ip_hdr(skb)->protocol; + } + } else { + eseg->swp_inner_l3_offset = skb_network_offset(skb) / 2; + if (skb->protocol == htons(ETH_P_IPV6)) + eseg->swp_flags |= MLX5_ETH_WQE_SWP_INNER_L3_IPV6; + proto = xo->proto; + } + switch (proto) { + case IPPROTO_UDP: + eseg->swp_flags |= MLX5_ETH_WQE_SWP_INNER_L4_UDP; + /* Fall through */ + case IPPROTO_TCP: + eseg->swp_inner_l4_offset = skb_inner_transport_offset(skb) / 2; + break; + } +} + +static void mlx5e_ipsec_set_iv(struct sk_buff *skb, struct xfrm_offload *xo) +{ + int iv_offset; + __be64 seqno; + + /* Place the SN in the IV field */ + seqno = cpu_to_be64(xo->seq.low + ((u64)xo->seq.hi << 32)); + iv_offset = skb_transport_offset(skb) + sizeof(struct ip_esp_hdr); + skb_store_bits(skb, iv_offset, &seqno, 8); +} + +static void mlx5e_ipsec_set_metadata(struct sk_buff *skb, + struct mlx5e_ipsec_metadata *mdata, + struct xfrm_offload *xo) +{ + struct ip_esp_hdr *esph; + struct tcphdr *tcph; + + if (skb_is_gso(skb)) { + /* Add LSO metadata indication */ + esph = ip_esp_hdr(skb); + tcph = inner_tcp_hdr(skb); + netdev_dbg(skb->dev, " Offloading GSO packet outer L3 %u; L4 %u; Inner L3 %u; L4 %u\n", + skb->network_header, + skb->transport_header, + skb->inner_network_header, + skb->inner_transport_header); + netdev_dbg(skb->dev, " Offloading GSO packet of len %u; mss %u; TCP sp %u dp %u seq 0x%x ESP seq 0x%x\n", + skb->len, skb_shinfo(skb)->gso_size, + ntohs(tcph->source), ntohs(tcph->dest), + ntohl(tcph->seq), ntohl(esph->seq_no)); + mdata->syndrome = MLX5E_IPSEC_TX_SYNDROME_OFFLOAD_WITH_LSO_TCP; + mdata->content.tx.mss_inv = mlx5e_ipsec_mss_inv(skb); + mdata->content.tx.seq = htons(ntohl(tcph->seq) & 0xFFFF); + } else { + mdata->syndrome = MLX5E_IPSEC_TX_SYNDROME_OFFLOAD; + } + mdata->content.tx.esp_next_proto = xo->proto; + + netdev_dbg(skb->dev, " TX metadata syndrome %u proto %u mss_inv %04x seq %04x\n", + mdata->syndrome, mdata->content.tx.esp_next_proto, + ntohs(mdata->content.tx.mss_inv), + ntohs(mdata->content.tx.seq)); +} + +struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev, + struct mlx5e_tx_wqe *wqe, + struct sk_buff *skb) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + struct xfrm_offload *xo = xfrm_offload(skb); + struct mlx5e_ipsec_metadata *mdata; + struct xfrm_state *x; + + if (!xo) + return skb; + + if (unlikely(skb->sp->len != 1)) { + atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_bundle); + goto drop; + } + + x = xfrm_input_state(skb); + if (unlikely(!x)) { + atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_no_state); + goto drop; + } + + if (unlikely(!x->xso.offload_handle || + (skb->protocol != htons(ETH_P_IP) && + skb->protocol != htons(ETH_P_IPV6)))) { + atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_not_ip); + goto drop; + } + + if (!skb_is_gso(skb)) + if (unlikely(mlx5e_ipsec_remove_trailer(skb, x))) { + atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_trailer); + goto drop; + } + mdata = mlx5e_ipsec_add_metadata(skb); + if (unlikely(IS_ERR(mdata))) { + atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_metadata); + goto drop; + } + mlx5e_ipsec_set_swp(skb, &wqe->eth, x->props.mode, xo); + mlx5e_ipsec_set_iv(skb, xo); + mlx5e_ipsec_set_metadata(skb, mdata, xo); + + return skb; + +drop: + kfree_skb(skb); + return NULL; +} + +static inline struct xfrm_state * +mlx5e_ipsec_build_sp(struct net_device *netdev, struct sk_buff *skb, + struct mlx5e_ipsec_metadata *mdata) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + struct xfrm_offload *xo; + struct xfrm_state *xs; + u32 sa_handle; + + skb->sp = secpath_dup(skb->sp); + if (unlikely(!skb->sp)) { + atomic64_inc(&priv->ipsec->sw_stats.ipsec_rx_drop_sp_alloc); + return NULL; + } + + sa_handle = be32_to_cpu(mdata->content.rx.sa_handle); + xs = mlx5e_ipsec_sadb_rx_lookup(priv->ipsec, sa_handle); + if (unlikely(!xs)) { + atomic64_inc(&priv->ipsec->sw_stats.ipsec_rx_drop_sadb_miss); + return NULL; + } + + skb->sp->xvec[skb->sp->len++] = xs; + skb->sp->olen++; + + xo = xfrm_offload(skb); + xo->flags = CRYPTO_DONE; + switch (mdata->syndrome) { + case MLX5E_IPSEC_RX_SYNDROME_DECRYPTED: + xo->status = CRYPTO_SUCCESS; + break; + case MLX5E_IPSEC_RX_SYNDROME_AUTH_FAILED: + xo->status = CRYPTO_TUNNEL_ESP_AUTH_FAILED; + break; + default: + atomic64_inc(&priv->ipsec->sw_stats.ipsec_rx_drop_syndrome); + return NULL; + } + return xs; +} + +struct sk_buff *mlx5e_ipsec_handle_rx_skb(struct net_device *netdev, + struct sk_buff *skb) +{ + struct mlx5e_ipsec_metadata *mdata; + struct ethhdr *old_eth; + struct ethhdr *new_eth; + struct xfrm_state *xs; + __be16 *ethtype; + + /* Detect inline metadata */ + if (skb->len < ETH_HLEN + MLX5E_METADATA_ETHER_LEN) + return skb; + ethtype = (__be16 *)(skb->data + ETH_ALEN * 2); + if (*ethtype != cpu_to_be16(MLX5E_METADATA_ETHER_TYPE)) + return skb; + + /* Use the metadata */ + mdata = (struct mlx5e_ipsec_metadata *)(skb->data + ETH_HLEN); + xs = mlx5e_ipsec_build_sp(netdev, skb, mdata); + if (unlikely(!xs)) { + kfree_skb(skb); + return NULL; + } + + /* Remove the metadata from the buffer */ + old_eth = (struct ethhdr *)skb->data; + new_eth = (struct ethhdr *)(skb->data + MLX5E_METADATA_ETHER_LEN); + memmove(new_eth, old_eth, 2 * ETH_ALEN); + /* Ethertype is already in its new place */ + skb_pull_inline(skb, MLX5E_METADATA_ETHER_LEN); + + return skb; +} + +bool mlx5e_ipsec_feature_check(struct sk_buff *skb, struct net_device *netdev, + netdev_features_t features) +{ + struct xfrm_state *x; + + if (skb->sp && skb->sp->len) { + x = skb->sp->xvec[0]; + if (x && x->xso.offload_handle) + return true; + } + return false; +} + +void mlx5e_ipsec_build_inverse_table(void) +{ + u16 mss_inv; + u32 mss; + + /* Calculate 1/x inverse table for use in GSO data path. + * Using this table, we provide the IPSec accelerator with the value of + * 1/gso_size so that it can infer the position of each segment inside + * the GSO, and increment the ESP sequence number, and generate the IV. + * The HW needs this value in Q0.16 fixed-point number format + */ + mlx5e_ipsec_inverse_table[1] = htons(0xFFFF); + for (mss = 2; mss < MAX_LSO_MSS; mss++) { + mss_inv = div_u64(1ULL << 32, mss) >> 16; + mlx5e_ipsec_inverse_table[mss] = htons(mss_inv); + } +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h new file mode 100644 index 000000000000..e37ae2598dbb --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h @@ -0,0 +1,55 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef __MLX5E_IPSEC_RXTX_H__ +#define __MLX5E_IPSEC_RXTX_H__ + +#ifdef CONFIG_MLX5_EN_IPSEC + +#include <linux/skbuff.h> +#include "en.h" + +struct sk_buff *mlx5e_ipsec_handle_rx_skb(struct net_device *netdev, + struct sk_buff *skb); +void mlx5e_ipsec_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe); + +void mlx5e_ipsec_inverse_table_init(void); +bool mlx5e_ipsec_feature_check(struct sk_buff *skb, struct net_device *netdev, + netdev_features_t features); +struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev, + struct mlx5e_tx_wqe *wqe, + struct sk_buff *skb); + +#endif /* CONFIG_MLX5_EN_IPSEC */ + +#endif /* __MLX5E_IPSEC_RXTX_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_stats.c new file mode 100644 index 000000000000..6fea59223dc4 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_stats.c @@ -0,0 +1,133 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/ethtool.h> +#include <net/sock.h> + +#include "en.h" +#include "accel/ipsec.h" +#include "fpga/sdk.h" +#include "en_accel/ipsec.h" + +static const struct counter_desc mlx5e_ipsec_hw_stats_desc[] = { + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_in_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_out_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_bypass_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_enc_in_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_enc_out_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_enc_bypass_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_drop_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_dec_auth_fail_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_enc_drop_packets) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_add_sa_success) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_add_sa_fail) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_del_sa_success) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_del_sa_fail) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_stats, ipsec_cmd_drop) }, +}; + +static const struct counter_desc mlx5e_ipsec_sw_stats_desc[] = { + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_rx_drop_sp_alloc) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_rx_drop_sadb_miss) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_rx_drop_syndrome) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_bundle) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_no_state) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_not_ip) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_trailer) }, + { MLX5E_DECLARE_STAT(struct mlx5e_ipsec_sw_stats, ipsec_tx_drop_metadata) }, +}; + +#define MLX5E_READ_CTR_ATOMIC64(ptr, dsc, i) \ + atomic64_read((atomic64_t *)((char *)(ptr) + (dsc)[i].offset)) + +#define NUM_IPSEC_HW_COUNTERS ARRAY_SIZE(mlx5e_ipsec_hw_stats_desc) +#define NUM_IPSEC_SW_COUNTERS ARRAY_SIZE(mlx5e_ipsec_sw_stats_desc) + +#define NUM_IPSEC_COUNTERS (NUM_IPSEC_HW_COUNTERS + NUM_IPSEC_SW_COUNTERS) + +int mlx5e_ipsec_get_count(struct mlx5e_priv *priv) +{ + if (!priv->ipsec) + return 0; + + return NUM_IPSEC_COUNTERS; +} + +int mlx5e_ipsec_get_strings(struct mlx5e_priv *priv, uint8_t *data) +{ + unsigned int i, idx = 0; + + if (!priv->ipsec) + return 0; + + for (i = 0; i < NUM_IPSEC_HW_COUNTERS; i++) + strcpy(data + (idx++) * ETH_GSTRING_LEN, + mlx5e_ipsec_hw_stats_desc[i].format); + + for (i = 0; i < NUM_IPSEC_SW_COUNTERS; i++) + strcpy(data + (idx++) * ETH_GSTRING_LEN, + mlx5e_ipsec_sw_stats_desc[i].format); + + return NUM_IPSEC_COUNTERS; +} + +void mlx5e_ipsec_update_stats(struct mlx5e_priv *priv) +{ + int ret; + + if (!priv->ipsec) + return; + + ret = mlx5_accel_ipsec_counters_read(priv->mdev, (u64 *)&priv->ipsec->stats, + NUM_IPSEC_HW_COUNTERS); + if (ret) + memset(&priv->ipsec->stats, 0, sizeof(priv->ipsec->stats)); +} + +int mlx5e_ipsec_get_stats(struct mlx5e_priv *priv, u64 *data) +{ + int i, idx = 0; + + if (!priv->ipsec) + return 0; + + for (i = 0; i < NUM_IPSEC_HW_COUNTERS; i++) + data[idx++] = MLX5E_READ_CTR64_CPU(&priv->ipsec->stats, + mlx5e_ipsec_hw_stats_desc, i); + + for (i = 0; i < NUM_IPSEC_SW_COUNTERS; i++) + data[idx++] = MLX5E_READ_CTR_ATOMIC64(&priv->ipsec->sw_stats, + mlx5e_ipsec_sw_stats_desc, i); + + return NUM_IPSEC_COUNTERS; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c index c8a005326e30..12d3ced61114 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c @@ -178,34 +178,26 @@ static int arfs_add_default_rule(struct mlx5e_priv *priv, struct mlx5_flow_destination dest; MLX5_DECLARE_FLOW_ACT(flow_act); struct mlx5_flow_spec *spec; + enum mlx5e_traffic_types tt; int err = 0; - spec = mlx5_vzalloc(sizeof(*spec)); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); if (!spec) { - netdev_err(priv->netdev, "%s: alloc failed\n", __func__); err = -ENOMEM; goto out; } dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR; - switch (type) { - case ARFS_IPV4_TCP: - dest.tir_num = tir[MLX5E_TT_IPV4_TCP].tirn; - break; - case ARFS_IPV4_UDP: - dest.tir_num = tir[MLX5E_TT_IPV4_UDP].tirn; - break; - case ARFS_IPV6_TCP: - dest.tir_num = tir[MLX5E_TT_IPV6_TCP].tirn; - break; - case ARFS_IPV6_UDP: - dest.tir_num = tir[MLX5E_TT_IPV6_UDP].tirn; - break; - default: + tt = arfs_get_tt(type); + if (tt == -EINVAL) { + netdev_err(priv->netdev, "%s: bad arfs_type: %d\n", + __func__, type); err = -EINVAL; goto out; } + dest.tir_num = tir[tt].tirn; + arfs_t->default_rule = mlx5_add_flow_rules(arfs_t->ft.t, spec, &flow_act, &dest, 1); @@ -237,7 +229,7 @@ static int arfs_create_groups(struct mlx5e_flow_table *ft, ft->g = kcalloc(MLX5E_ARFS_NUM_GROUPS, sizeof(*ft->g), GFP_KERNEL); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in || !ft->g) { kvfree(ft->g); kvfree(in); @@ -481,9 +473,8 @@ static struct mlx5_flow_handle *arfs_add_rule(struct mlx5e_priv *priv, struct mlx5_flow_table *ft; int err = 0; - spec = mlx5_vzalloc(sizeof(*spec)); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); if (!spec) { - netdev_err(priv->netdev, "%s: alloc failed\n", __func__); err = -ENOMEM; goto out; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c index e706a87fc8b2..66f432385dbb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_clock.c @@ -86,9 +86,8 @@ static void mlx5e_timestamp_overflow(struct work_struct *work) schedule_delayed_work(&tstamp->overflow_work, tstamp->overflow_period); } -int mlx5e_hwstamp_set(struct net_device *dev, struct ifreq *ifr) +int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr) { - struct mlx5e_priv *priv = netdev_priv(dev); struct hwtstamp_config config; int err; @@ -128,11 +127,12 @@ int mlx5e_hwstamp_set(struct net_device *dev, struct ifreq *ifr) case HWTSTAMP_FILTER_PTP_V2_EVENT: case HWTSTAMP_FILTER_PTP_V2_SYNC: case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ: + case HWTSTAMP_FILTER_NTP_ALL: /* Disable CQE compression */ - netdev_warn(dev, "Disabling cqe compression"); + netdev_warn(priv->netdev, "Disabling cqe compression"); err = mlx5e_modify_rx_cqe_compression_locked(priv, false); if (err) { - netdev_err(dev, "Failed disabling cqe compression err=%d\n", err); + netdev_err(priv->netdev, "Failed disabling cqe compression err=%d\n", err); mutex_unlock(&priv->state_lock); return err; } @@ -150,9 +150,8 @@ int mlx5e_hwstamp_set(struct net_device *dev, struct ifreq *ifr) sizeof(config)) ? -EFAULT : 0; } -int mlx5e_hwstamp_get(struct net_device *dev, struct ifreq *ifr) +int mlx5e_hwstamp_get(struct mlx5e_priv *priv, struct ifreq *ifr) { - struct mlx5e_priv *priv = netdev_priv(dev); struct hwtstamp_config *cfg = &priv->tstamp.hwtstamp_config; if (!MLX5_CAP_GEN(priv->mdev, device_frequency_khz)) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_common.c b/drivers/net/ethernet/mellanox/mlx5/core/en_common.c index f1f17f7a3cd0..ece3fb147e3e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_common.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_common.c @@ -65,7 +65,7 @@ static int mlx5e_create_mkey(struct mlx5_core_dev *mdev, u32 pdn, u32 *in; int err; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -145,9 +145,8 @@ int mlx5e_refresh_tirs(struct mlx5e_priv *priv, bool enable_uc_lb) int inlen; void *in; - inlen = MLX5_ST_SZ_BYTES(modify_tir_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) goto out; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c index 8fa23f6a1f67..2eb54d36e16e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c @@ -464,6 +464,8 @@ static void mlx5e_dcbnl_getpermhwaddr(struct net_device *netdev, if (!perm_addr) return; + memset(perm_addr, 0xff, MAX_ADDR_LEN); + mlx5_query_nic_vport_mac_address(priv->mdev, 0, perm_addr); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c index ce7b09d72ff6..917fade5f5d5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c @@ -31,15 +31,15 @@ */ #include "en.h" +#include "en_accel/ipsec.h" -static void mlx5e_get_drvinfo(struct net_device *dev, - struct ethtool_drvinfo *drvinfo) +void mlx5e_ethtool_get_drvinfo(struct mlx5e_priv *priv, + struct ethtool_drvinfo *drvinfo) { - struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5_core_dev *mdev = priv->mdev; strlcpy(drvinfo->driver, DRIVER_NAME, sizeof(drvinfo->driver)); - strlcpy(drvinfo->version, DRIVER_VERSION " (" DRIVER_RELDATE ")", + strlcpy(drvinfo->version, DRIVER_VERSION, sizeof(drvinfo->version)); snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version), "%d.%d.%04d (%.16s)", @@ -49,6 +49,14 @@ static void mlx5e_get_drvinfo(struct net_device *dev, sizeof(drvinfo->bus_info)); } +static void mlx5e_get_drvinfo(struct net_device *dev, + struct ethtool_drvinfo *drvinfo) +{ + struct mlx5e_priv *priv = netdev_priv(dev); + + mlx5e_ethtool_get_drvinfo(priv, drvinfo); +} + struct ptys2ethtool_config { __ETHTOOL_DECLARE_LINK_MODE_MASK(supported); __ETHTOOL_DECLARE_LINK_MODE_MASK(advertised); @@ -135,6 +143,9 @@ static unsigned long mlx5e_query_pfc_combined(struct mlx5e_priv *priv) u8 pfc_en_rx; int err; + if (MLX5_CAP_GEN(mdev, port_type) != MLX5_CAP_PORT_TYPE_ETH) + return 0; + err = mlx5_query_port_pfc(mdev, &pfc_en_tx, &pfc_en_rx); return err ? 0 : pfc_en_tx | pfc_en_rx; @@ -147,6 +158,9 @@ static bool mlx5e_query_global_pause_combined(struct mlx5e_priv *priv) u32 tx_pause; int err; + if (MLX5_CAP_GEN(mdev, port_type) != MLX5_CAP_PORT_TYPE_ETH) + return false; + err = mlx5_query_port_pause(mdev, &rx_pause, &tx_pause); return err ? false : rx_pause | tx_pause; @@ -160,9 +174,8 @@ static bool mlx5e_query_global_pause_combined(struct mlx5e_priv *priv) ((mlx5e_query_global_pause_combined(priv) + hweight8(mlx5e_query_pfc_combined(priv))) * \ NUM_PPORT_PER_PRIO_PFC_COUNTERS) -static int mlx5e_get_sset_count(struct net_device *dev, int sset) +int mlx5e_ethtool_get_sset_count(struct mlx5e_priv *priv, int sset) { - struct mlx5e_priv *priv = netdev_priv(dev); switch (sset) { case ETH_SS_STATS: @@ -174,7 +187,8 @@ static int mlx5e_get_sset_count(struct net_device *dev, int sset) MLX5E_NUM_SQ_STATS(priv) + MLX5E_NUM_PFC_COUNTERS(priv) + ARRAY_SIZE(mlx5e_pme_status_desc) + - ARRAY_SIZE(mlx5e_pme_error_desc); + ARRAY_SIZE(mlx5e_pme_error_desc) + + mlx5e_ipsec_get_count(priv); case ETH_SS_PRIV_FLAGS: return ARRAY_SIZE(mlx5e_priv_flags); @@ -186,6 +200,13 @@ static int mlx5e_get_sset_count(struct net_device *dev, int sset) } } +static int mlx5e_get_sset_count(struct net_device *dev, int sset) +{ + struct mlx5e_priv *priv = netdev_priv(dev); + + return mlx5e_ethtool_get_sset_count(priv, sset); +} + static void mlx5e_fill_stats_strings(struct mlx5e_priv *priv, uint8_t *data) { int i, j, tc, prio, idx = 0; @@ -256,6 +277,9 @@ static void mlx5e_fill_stats_strings(struct mlx5e_priv *priv, uint8_t *data) for (i = 0; i < ARRAY_SIZE(mlx5e_pme_error_desc); i++) strcpy(data + (idx++) * ETH_GSTRING_LEN, mlx5e_pme_error_desc[i].format); + /* IPSec counters */ + idx += mlx5e_ipsec_get_strings(priv, data + idx * ETH_GSTRING_LEN); + if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) return; @@ -273,10 +297,9 @@ static void mlx5e_fill_stats_strings(struct mlx5e_priv *priv, uint8_t *data) priv->channel_tc2txq[i][tc]); } -static void mlx5e_get_strings(struct net_device *dev, - uint32_t stringset, uint8_t *data) +void mlx5e_ethtool_get_strings(struct mlx5e_priv *priv, + uint32_t stringset, uint8_t *data) { - struct mlx5e_priv *priv = netdev_priv(dev); int i; switch (stringset) { @@ -297,10 +320,17 @@ static void mlx5e_get_strings(struct net_device *dev, } } -static void mlx5e_get_ethtool_stats(struct net_device *dev, - struct ethtool_stats *stats, u64 *data) +static void mlx5e_get_strings(struct net_device *dev, + uint32_t stringset, uint8_t *data) { struct mlx5e_priv *priv = netdev_priv(dev); + + mlx5e_ethtool_get_strings(priv, stringset, data); +} + +void mlx5e_ethtool_get_ethtool_stats(struct mlx5e_priv *priv, + struct ethtool_stats *stats, u64 *data) +{ struct mlx5e_channels *channels; struct mlx5_priv *mlx5_priv; int i, j, tc, prio, idx = 0; @@ -311,7 +341,7 @@ static void mlx5e_get_ethtool_stats(struct net_device *dev, mutex_lock(&priv->state_lock); if (test_bit(MLX5E_STATE_OPENED, &priv->state)) - mlx5e_update_stats(priv); + mlx5e_update_stats(priv, true); channels = &priv->channels; mutex_unlock(&priv->state_lock); @@ -378,6 +408,9 @@ static void mlx5e_get_ethtool_stats(struct net_device *dev, data[idx++] = MLX5E_READ_CTR64_CPU(mlx5_priv->pme_stats.error_counters, mlx5e_pme_error_desc, i); + /* IPSec counters */ + idx += mlx5e_ipsec_get_stats(priv, data + idx); + if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) return; @@ -395,6 +428,15 @@ static void mlx5e_get_ethtool_stats(struct net_device *dev, sq_stats_desc, j); } +static void mlx5e_get_ethtool_stats(struct net_device *dev, + struct ethtool_stats *stats, + u64 *data) +{ + struct mlx5e_priv *priv = netdev_priv(dev); + + mlx5e_ethtool_get_ethtool_stats(priv, stats, data); +} + static u32 mlx5e_rx_wqes_to_packets(struct mlx5e_priv *priv, int rq_wq_type, int num_wqe) { @@ -439,10 +481,9 @@ static u32 mlx5e_packets_to_rx_wqes(struct mlx5e_priv *priv, int rq_wq_type, return 1 << (order_base_2(num_wqes)); } -static void mlx5e_get_ringparam(struct net_device *dev, - struct ethtool_ringparam *param) +void mlx5e_ethtool_get_ringparam(struct mlx5e_priv *priv, + struct ethtool_ringparam *param) { - struct mlx5e_priv *priv = netdev_priv(dev); int rq_wq_type = priv->channels.params.rq_wq_type; param->rx_max_pending = mlx5e_rx_wqes_to_packets(priv, rq_wq_type, @@ -453,10 +494,17 @@ static void mlx5e_get_ringparam(struct net_device *dev, param->tx_pending = 1 << priv->channels.params.log_sq_size; } -static int mlx5e_set_ringparam(struct net_device *dev, - struct ethtool_ringparam *param) +static void mlx5e_get_ringparam(struct net_device *dev, + struct ethtool_ringparam *param) { struct mlx5e_priv *priv = netdev_priv(dev); + + mlx5e_ethtool_get_ringparam(priv, param); +} + +int mlx5e_ethtool_set_ringparam(struct mlx5e_priv *priv, + struct ethtool_ringparam *param) +{ int rq_wq_type = priv->channels.params.rq_wq_type; struct mlx5e_channels new_channels = {}; u32 rx_pending_wqes; @@ -468,12 +516,12 @@ static int mlx5e_set_ringparam(struct net_device *dev, int err = 0; if (param->rx_jumbo_pending) { - netdev_info(dev, "%s: rx_jumbo_pending not supported\n", + netdev_info(priv->netdev, "%s: rx_jumbo_pending not supported\n", __func__); return -EINVAL; } if (param->rx_mini_pending) { - netdev_info(dev, "%s: rx_mini_pending not supported\n", + netdev_info(priv->netdev, "%s: rx_mini_pending not supported\n", __func__); return -EINVAL; } @@ -486,13 +534,13 @@ static int mlx5e_set_ringparam(struct net_device *dev, param->rx_pending); if (param->rx_pending < min_rq_size) { - netdev_info(dev, "%s: rx_pending (%d) < min (%d)\n", + netdev_info(priv->netdev, "%s: rx_pending (%d) < min (%d)\n", __func__, param->rx_pending, min_rq_size); return -EINVAL; } if (param->rx_pending > max_rq_size) { - netdev_info(dev, "%s: rx_pending (%d) > max (%d)\n", + netdev_info(priv->netdev, "%s: rx_pending (%d) > max (%d)\n", __func__, param->rx_pending, max_rq_size); return -EINVAL; @@ -501,19 +549,19 @@ static int mlx5e_set_ringparam(struct net_device *dev, num_mtts = MLX5E_REQUIRED_MTTS(rx_pending_wqes); if (priv->channels.params.rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ && !MLX5E_VALID_NUM_MTTS(num_mtts)) { - netdev_info(dev, "%s: rx_pending (%d) request can't be satisfied, try to reduce.\n", + netdev_info(priv->netdev, "%s: rx_pending (%d) request can't be satisfied, try to reduce.\n", __func__, param->rx_pending); return -EINVAL; } if (param->tx_pending < (1 << MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE)) { - netdev_info(dev, "%s: tx_pending (%d) < min (%d)\n", + netdev_info(priv->netdev, "%s: tx_pending (%d) < min (%d)\n", __func__, param->tx_pending, 1 << MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE); return -EINVAL; } if (param->tx_pending > (1 << MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE)) { - netdev_info(dev, "%s: tx_pending (%d) > max (%d)\n", + netdev_info(priv->netdev, "%s: tx_pending (%d) > max (%d)\n", __func__, param->tx_pending, 1 << MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE); return -EINVAL; @@ -549,26 +597,39 @@ unlock: return err; } -static void mlx5e_get_channels(struct net_device *dev, - struct ethtool_channels *ch) +static int mlx5e_set_ringparam(struct net_device *dev, + struct ethtool_ringparam *param) { struct mlx5e_priv *priv = netdev_priv(dev); + return mlx5e_ethtool_set_ringparam(priv, param); +} + +void mlx5e_ethtool_get_channels(struct mlx5e_priv *priv, + struct ethtool_channels *ch) +{ ch->max_combined = priv->profile->max_nch(priv->mdev); ch->combined_count = priv->channels.params.num_channels; } -static int mlx5e_set_channels(struct net_device *dev, - struct ethtool_channels *ch) +static void mlx5e_get_channels(struct net_device *dev, + struct ethtool_channels *ch) { struct mlx5e_priv *priv = netdev_priv(dev); + + mlx5e_ethtool_get_channels(priv, ch); +} + +int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv, + struct ethtool_channels *ch) +{ unsigned int count = ch->combined_count; struct mlx5e_channels new_channels = {}; bool arfs_enabled; int err = 0; if (!count) { - netdev_info(dev, "%s: combined_count=0 not supported\n", + netdev_info(priv->netdev, "%s: combined_count=0 not supported\n", __func__); return -EINVAL; } @@ -593,7 +654,7 @@ static int mlx5e_set_channels(struct net_device *dev, if (err) goto out; - arfs_enabled = dev->features & NETIF_F_NTUPLE; + arfs_enabled = priv->netdev->features & NETIF_F_NTUPLE; if (arfs_enabled) mlx5e_arfs_disable(priv); @@ -603,7 +664,7 @@ static int mlx5e_set_channels(struct net_device *dev, if (arfs_enabled) { err = mlx5e_arfs_enable(priv); if (err) - netdev_err(dev, "%s: mlx5e_arfs_enable failed: %d\n", + netdev_err(priv->netdev, "%s: mlx5e_arfs_enable failed: %d\n", __func__, err); } @@ -613,11 +674,17 @@ out: return err; } -static int mlx5e_get_coalesce(struct net_device *netdev, - struct ethtool_coalesce *coal) +static int mlx5e_set_channels(struct net_device *dev, + struct ethtool_channels *ch) { - struct mlx5e_priv *priv = netdev_priv(netdev); + struct mlx5e_priv *priv = netdev_priv(dev); + + return mlx5e_ethtool_set_channels(priv, ch); +} +int mlx5e_ethtool_get_coalesce(struct mlx5e_priv *priv, + struct ethtool_coalesce *coal) +{ if (!MLX5_CAP_GEN(priv->mdev, cq_moderation)) return -EOPNOTSUPP; @@ -630,6 +697,14 @@ static int mlx5e_get_coalesce(struct net_device *netdev, return 0; } +static int mlx5e_get_coalesce(struct net_device *netdev, + struct ethtool_coalesce *coal) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_get_coalesce(priv, coal); +} + static void mlx5e_set_priv_channels_coalesce(struct mlx5e_priv *priv, struct ethtool_coalesce *coal) { @@ -653,10 +728,9 @@ mlx5e_set_priv_channels_coalesce(struct mlx5e_priv *priv, struct ethtool_coalesc } } -static int mlx5e_set_coalesce(struct net_device *netdev, - struct ethtool_coalesce *coal) +int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv, + struct ethtool_coalesce *coal) { - struct mlx5e_priv *priv = netdev_priv(netdev); struct mlx5_core_dev *mdev = priv->mdev; struct mlx5e_channels new_channels = {}; int err = 0; @@ -699,6 +773,14 @@ out: return err; } +static int mlx5e_set_coalesce(struct net_device *netdev, + struct ethtool_coalesce *coal) +{ + struct mlx5e_priv *priv = netdev_priv(netdev); + + return mlx5e_ethtool_set_coalesce(priv, coal); +} + static void ptys2ethtool_supported_link(unsigned long *supported_modes, u32 eth_proto_cap) { @@ -723,24 +805,81 @@ static void ptys2ethtool_adver_link(unsigned long *advertising_modes, __ETHTOOL_LINK_MODE_MASK_NBITS); } -static void ptys2ethtool_supported_port(struct ethtool_link_ksettings *link_ksettings, - u32 eth_proto_cap) +static void ptys2ethtool_supported_advertised_port(struct ethtool_link_ksettings *link_ksettings, + u32 eth_proto_cap, + u8 connector_type) { - if (eth_proto_cap & (MLX5E_PROT_MASK(MLX5E_10GBASE_CR) - | MLX5E_PROT_MASK(MLX5E_10GBASE_SR) - | MLX5E_PROT_MASK(MLX5E_40GBASE_CR4) - | MLX5E_PROT_MASK(MLX5E_40GBASE_SR4) - | MLX5E_PROT_MASK(MLX5E_100GBASE_SR4) - | MLX5E_PROT_MASK(MLX5E_1000BASE_CX_SGMII))) { - ethtool_link_ksettings_add_link_mode(link_ksettings, supported, FIBRE); + if (!connector_type || connector_type >= MLX5E_CONNECTOR_TYPE_NUMBER) { + if (eth_proto_cap & (MLX5E_PROT_MASK(MLX5E_10GBASE_CR) + | MLX5E_PROT_MASK(MLX5E_10GBASE_SR) + | MLX5E_PROT_MASK(MLX5E_40GBASE_CR4) + | MLX5E_PROT_MASK(MLX5E_40GBASE_SR4) + | MLX5E_PROT_MASK(MLX5E_100GBASE_SR4) + | MLX5E_PROT_MASK(MLX5E_1000BASE_CX_SGMII))) { + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, + FIBRE); + ethtool_link_ksettings_add_link_mode(link_ksettings, + advertising, + FIBRE); + } + + if (eth_proto_cap & (MLX5E_PROT_MASK(MLX5E_100GBASE_KR4) + | MLX5E_PROT_MASK(MLX5E_40GBASE_KR4) + | MLX5E_PROT_MASK(MLX5E_10GBASE_KR) + | MLX5E_PROT_MASK(MLX5E_10GBASE_KX4) + | MLX5E_PROT_MASK(MLX5E_1000BASE_KX))) { + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, + Backplane); + ethtool_link_ksettings_add_link_mode(link_ksettings, + advertising, + Backplane); + } + return; } - if (eth_proto_cap & (MLX5E_PROT_MASK(MLX5E_100GBASE_KR4) - | MLX5E_PROT_MASK(MLX5E_40GBASE_KR4) - | MLX5E_PROT_MASK(MLX5E_10GBASE_KR) - | MLX5E_PROT_MASK(MLX5E_10GBASE_KX4) - | MLX5E_PROT_MASK(MLX5E_1000BASE_KX))) { - ethtool_link_ksettings_add_link_mode(link_ksettings, supported, Backplane); + switch (connector_type) { + case MLX5E_PORT_TP: + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, TP); + ethtool_link_ksettings_add_link_mode(link_ksettings, + advertising, TP); + break; + case MLX5E_PORT_AUI: + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, AUI); + ethtool_link_ksettings_add_link_mode(link_ksettings, + advertising, AUI); + break; + case MLX5E_PORT_BNC: + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, BNC); + ethtool_link_ksettings_add_link_mode(link_ksettings, + advertising, BNC); + break; + case MLX5E_PORT_MII: + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, MII); + ethtool_link_ksettings_add_link_mode(link_ksettings, + advertising, MII); + break; + case MLX5E_PORT_FIBRE: + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, FIBRE); + ethtool_link_ksettings_add_link_mode(link_ksettings, + advertising, FIBRE); + break; + case MLX5E_PORT_DA: + ethtool_link_ksettings_add_link_mode(link_ksettings, + supported, Backplane); + ethtool_link_ksettings_add_link_mode(link_ksettings, + advertising, Backplane); + break; + case MLX5E_PORT_NONE: + case MLX5E_PORT_OTHER: + default: + break; } } @@ -791,10 +930,8 @@ static void get_supported(u32 eth_proto_cap, { unsigned long *supported = link_ksettings->link_modes.supported; - ptys2ethtool_supported_port(link_ksettings, eth_proto_cap); ptys2ethtool_supported_link(supported, eth_proto_cap); ethtool_link_ksettings_add_link_mode(link_ksettings, supported, Pause); - ethtool_link_ksettings_add_link_mode(link_ksettings, supported, Asym_Pause); } static void get_advertising(u32 eth_proto_cap, u8 tx_pause, @@ -804,14 +941,29 @@ static void get_advertising(u32 eth_proto_cap, u8 tx_pause, unsigned long *advertising = link_ksettings->link_modes.advertising; ptys2ethtool_adver_link(advertising, eth_proto_cap); - if (tx_pause) + if (rx_pause) ethtool_link_ksettings_add_link_mode(link_ksettings, advertising, Pause); if (tx_pause ^ rx_pause) ethtool_link_ksettings_add_link_mode(link_ksettings, advertising, Asym_Pause); } -static u8 get_connector_port(u32 eth_proto) +static int ptys2connector_type[MLX5E_CONNECTOR_TYPE_NUMBER] = { + [MLX5E_PORT_UNKNOWN] = PORT_OTHER, + [MLX5E_PORT_NONE] = PORT_NONE, + [MLX5E_PORT_TP] = PORT_TP, + [MLX5E_PORT_AUI] = PORT_AUI, + [MLX5E_PORT_BNC] = PORT_BNC, + [MLX5E_PORT_MII] = PORT_MII, + [MLX5E_PORT_FIBRE] = PORT_FIBRE, + [MLX5E_PORT_DA] = PORT_DA, + [MLX5E_PORT_OTHER] = PORT_OTHER, + }; + +static u8 get_connector_port(u32 eth_proto, u8 connector_type) { + if (connector_type && connector_type < MLX5E_CONNECTOR_TYPE_NUMBER) + return ptys2connector_type[connector_type]; + if (eth_proto & (MLX5E_PROT_MASK(MLX5E_10GBASE_SR) | MLX5E_PROT_MASK(MLX5E_40GBASE_SR4) | MLX5E_PROT_MASK(MLX5E_100GBASE_SR4) @@ -849,12 +1001,15 @@ static int mlx5e_get_link_ksettings(struct net_device *netdev, struct mlx5e_priv *priv = netdev_priv(netdev); struct mlx5_core_dev *mdev = priv->mdev; u32 out[MLX5_ST_SZ_DW(ptys_reg)] = {0}; + u32 rx_pause = 0; + u32 tx_pause = 0; u32 eth_proto_cap; u32 eth_proto_admin; u32 eth_proto_lp; u32 eth_proto_oper; u8 an_disable_admin; u8 an_status; + u8 connector_type; int err; err = mlx5_query_port_ptys(mdev, out, sizeof(out), MLX5_PTYS_EN, 1); @@ -870,17 +1025,23 @@ static int mlx5e_get_link_ksettings(struct net_device *netdev, eth_proto_lp = MLX5_GET(ptys_reg, out, eth_proto_lp_advertise); an_disable_admin = MLX5_GET(ptys_reg, out, an_disable_admin); an_status = MLX5_GET(ptys_reg, out, an_status); + connector_type = MLX5_GET(ptys_reg, out, connector_type); + + mlx5_query_port_pause(mdev, &rx_pause, &tx_pause); ethtool_link_ksettings_zero_link_mode(link_ksettings, supported); ethtool_link_ksettings_zero_link_mode(link_ksettings, advertising); get_supported(eth_proto_cap, link_ksettings); - get_advertising(eth_proto_admin, 0, 0, link_ksettings); + get_advertising(eth_proto_admin, tx_pause, rx_pause, link_ksettings); get_speed_duplex(netdev, eth_proto_oper, link_ksettings); eth_proto_oper = eth_proto_oper ? eth_proto_oper : eth_proto_cap; - link_ksettings->base.port = get_connector_port(eth_proto_oper); + link_ksettings->base.port = get_connector_port(eth_proto_oper, + connector_type); + ptys2ethtool_supported_advertised_port(link_ksettings, eth_proto_admin, + connector_type); get_lp_advertising(eth_proto_lp, link_ksettings); if (an_status == MLX5_AN_COMPLETE) @@ -1045,7 +1206,7 @@ static int mlx5e_set_rxfh(struct net_device *dev, const u32 *indir, (hfunc != ETH_RSS_HASH_TOP)) return -EINVAL; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -1219,13 +1380,12 @@ static int mlx5e_set_pauseparam(struct net_device *netdev, return err; } -static int mlx5e_get_ts_info(struct net_device *dev, - struct ethtool_ts_info *info) +int mlx5e_ethtool_get_ts_info(struct mlx5e_priv *priv, + struct ethtool_ts_info *info) { - struct mlx5e_priv *priv = netdev_priv(dev); int ret; - ret = ethtool_op_get_ts_info(dev, info); + ret = ethtool_op_get_ts_info(priv->netdev, info); if (ret) return ret; @@ -1239,15 +1399,23 @@ static int mlx5e_get_ts_info(struct net_device *dev, SOF_TIMESTAMPING_RX_HARDWARE | SOF_TIMESTAMPING_RAW_HARDWARE; - info->tx_types = (BIT(1) << HWTSTAMP_TX_OFF) | - (BIT(1) << HWTSTAMP_TX_ON); + info->tx_types = BIT(HWTSTAMP_TX_OFF) | + BIT(HWTSTAMP_TX_ON); - info->rx_filters = (BIT(1) << HWTSTAMP_FILTER_NONE) | - (BIT(1) << HWTSTAMP_FILTER_ALL); + info->rx_filters = BIT(HWTSTAMP_FILTER_NONE) | + BIT(HWTSTAMP_FILTER_ALL); return 0; } +static int mlx5e_get_ts_info(struct net_device *dev, + struct ethtool_ts_info *info) +{ + struct mlx5e_priv *priv = netdev_priv(dev); + + return mlx5e_ethtool_get_ts_info(priv, info); +} + static __u32 mlx5e_get_wol_supported(struct mlx5_core_dev *mdev) { __u32 ret = 0; @@ -1635,6 +1803,40 @@ static int mlx5e_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd) return err; } +int mlx5e_ethtool_flash_device(struct mlx5e_priv *priv, + struct ethtool_flash *flash) +{ + struct mlx5_core_dev *mdev = priv->mdev; + struct net_device *dev = priv->netdev; + const struct firmware *fw; + int err; + + if (flash->region != ETHTOOL_FLASH_ALL_REGIONS) + return -EOPNOTSUPP; + + err = request_firmware_direct(&fw, flash->data, &dev->dev); + if (err) + return err; + + dev_hold(dev); + rtnl_unlock(); + + err = mlx5_firmware_flash(mdev, fw); + release_firmware(fw); + + rtnl_lock(); + dev_put(dev); + return err; +} + +static int mlx5e_flash_device(struct net_device *dev, + struct ethtool_flash *flash) +{ + struct mlx5e_priv *priv = netdev_priv(dev); + + return mlx5e_ethtool_flash_device(priv, flash); +} + const struct ethtool_ops mlx5e_ethtool_ops = { .get_drvinfo = mlx5e_get_drvinfo, .get_link = ethtool_op_get_link, @@ -1655,6 +1857,7 @@ const struct ethtool_ops mlx5e_ethtool_ops = { .set_rxfh = mlx5e_set_rxfh, .get_rxnfc = mlx5e_get_rxnfc, .set_rxnfc = mlx5e_set_rxnfc, + .flash_device = mlx5e_flash_device, .get_tunable = mlx5e_get_tunable, .set_tunable = mlx5e_set_tunable, .get_pauseparam = mlx5e_get_pauseparam, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c index 576d6787b484..dfccb5305e9c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c @@ -170,7 +170,6 @@ static int __mlx5e_add_vlan_rule(struct mlx5e_priv *priv, spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS; - switch (rule_type) { case MLX5E_VLAN_RULE_TYPE_UNTAGGED: rule_p = &priv->fs.vlan.untagged_rule; @@ -218,11 +217,9 @@ static int mlx5e_add_vlan_rule(struct mlx5e_priv *priv, struct mlx5_flow_spec *spec; int err = 0; - spec = mlx5_vzalloc(sizeof(*spec)); - if (!spec) { - netdev_err(priv->netdev, "%s: alloc failed\n", __func__); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); + if (!spec) return -ENOMEM; - } if (rule_type == MLX5E_VLAN_RULE_TYPE_MATCH_VID) mlx5e_vport_context_update_vlans(priv); @@ -660,11 +657,9 @@ mlx5e_generate_ttc_rule(struct mlx5e_priv *priv, struct mlx5_flow_spec *spec; int err = 0; - spec = mlx5_vzalloc(sizeof(*spec)); - if (!spec) { - netdev_err(priv->netdev, "%s: alloc failed\n", __func__); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); + if (!spec) return ERR_PTR(-ENOMEM); - } if (proto) { spec->match_criteria_enable = MLX5_MATCH_OUTER_HEADERS; @@ -742,7 +737,7 @@ static int mlx5e_create_ttc_table_groups(struct mlx5e_ttc_table *ttc) sizeof(*ft->g), GFP_KERNEL); if (!ft->g) return -ENOMEM; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) { kfree(ft->g); return -ENOMEM; @@ -800,7 +795,7 @@ void mlx5e_destroy_ttc_table(struct mlx5e_priv *priv) mlx5e_destroy_flow_table(&ttc->ft); } -int mlx5e_create_ttc_table(struct mlx5e_priv *priv, u32 underlay_qpn) +int mlx5e_create_ttc_table(struct mlx5e_priv *priv) { struct mlx5e_ttc_table *ttc = &priv->fs.ttc; struct mlx5_flow_table_attr ft_attr = {}; @@ -810,7 +805,6 @@ int mlx5e_create_ttc_table(struct mlx5e_priv *priv, u32 underlay_qpn) ft_attr.max_fte = MLX5E_TTC_TABLE_SIZE; ft_attr.level = MLX5E_TTC_FT_LEVEL; ft_attr.prio = MLX5E_NIC_PRIO; - ft_attr.underlay_qpn = underlay_qpn; ft->t = mlx5_create_flow_table(priv->fs.ns, &ft_attr); if (IS_ERR(ft->t)) { @@ -853,11 +847,9 @@ static int mlx5e_add_l2_flow_rule(struct mlx5e_priv *priv, u8 *mc_dmac; u8 *mv_dmac; - spec = mlx5_vzalloc(sizeof(*spec)); - if (!spec) { - netdev_err(priv->netdev, "%s: alloc failed\n", __func__); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); + if (!spec) return -ENOMEM; - } mc_dmac = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, outer_headers.dmac_47_16); @@ -917,7 +909,7 @@ static int mlx5e_create_l2_table_groups(struct mlx5e_l2_table *l2_table) ft->g = kcalloc(MLX5E_NUM_L2_GROUPS, sizeof(*ft->g), GFP_KERNEL); if (!ft->g) return -ENOMEM; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) { kfree(ft->g); return -ENOMEM; @@ -1072,7 +1064,7 @@ static int mlx5e_create_vlan_table_groups(struct mlx5e_flow_table *ft) int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in); int err; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -1147,7 +1139,7 @@ int mlx5e_create_flow_steering(struct mlx5e_priv *priv) priv->netdev->hw_features &= ~NETIF_F_NTUPLE; } - err = mlx5e_create_ttc_table(priv, 0); + err = mlx5e_create_ttc_table(priv); if (err) { netdev_err(priv->netdev, "Failed to create ttc table, err=%d\n", err); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c index 85bf4a389295..bdd82c9b3992 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c @@ -296,7 +296,7 @@ add_ethtool_flow_rule(struct mlx5e_priv *priv, struct mlx5_flow_handle *rule; int err = 0; - spec = mlx5_vzalloc(sizeof(*spec)); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); if (!spec) return ERR_PTR(-ENOMEM); err = set_flow_attrs(spec->match_criteria, spec->match_value, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index a61b71b6fff3..1eac5003084f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -39,6 +39,9 @@ #include "en.h" #include "en_tc.h" #include "en_rep.h" +#include "en_accel/ipsec.h" +#include "en_accel/ipsec_rxtx.h" +#include "accel/ipsec.h" #include "vxlan.h" struct mlx5e_rq_param { @@ -96,9 +99,12 @@ void mlx5e_set_rq_type_params(struct mlx5_core_dev *mdev, params->log_rq_size = is_kdump_kernel() ? MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE : MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE; + params->rq_headroom = params->xdp_prog ? + XDP_PACKET_HEADROOM : MLX5_RX_HEADROOM; + params->rq_headroom += NET_IP_ALIGN; /* Extra room needed for build_skb */ - params->lro_wqe_sz -= MLX5_RX_HEADROOM + + params->lro_wqe_sz -= params->rq_headroom + SKB_DATA_ALIGN(sizeof(struct skb_shared_info)); } @@ -112,7 +118,7 @@ void mlx5e_set_rq_type_params(struct mlx5_core_dev *mdev, static void mlx5e_set_rq_params(struct mlx5_core_dev *mdev, struct mlx5e_params *params) { u8 rq_type = mlx5e_check_fragmented_striding_rq_cap(mdev) && - !params->xdp_prog ? + !params->xdp_prog && !MLX5_IPSEC_DEV(mdev) ? MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ : MLX5_WQ_TYPE_LINKED_LIST; mlx5e_set_rq_type_params(mdev, params, rq_type); @@ -124,7 +130,8 @@ static void mlx5e_update_carrier(struct mlx5e_priv *priv) u8 port_state; port_state = mlx5_query_vport_state(mdev, - MLX5_QUERY_VPORT_STATE_IN_OP_MOD_VNIC_VPORT, 0); + MLX5_QUERY_VPORT_STATE_IN_OP_MOD_VNIC_VPORT, + 0); if (port_state == VPORT_STATE_UP) { netdev_info(priv->netdev, "Link up\n"); @@ -142,7 +149,8 @@ static void mlx5e_update_carrier_work(struct work_struct *work) mutex_lock(&priv->state_lock); if (test_bit(MLX5E_STATE_OPENED, &priv->state)) - mlx5e_update_carrier(priv); + if (priv->profile->update_carrier) + priv->profile->update_carrier(priv); mutex_unlock(&priv->state_lock); } @@ -195,6 +203,7 @@ static void mlx5e_update_sw_counters(struct mlx5e_priv *priv) s->rx_buff_alloc_err += rq_stats->buff_alloc_err; s->rx_cqe_compress_blks += rq_stats->cqe_compress_blks; s->rx_cqe_compress_pkts += rq_stats->cqe_compress_pkts; + s->rx_page_reuse += rq_stats->page_reuse; s->rx_cache_reuse += rq_stats->cache_reuse; s->rx_cache_full += rq_stats->cache_full; s->rx_cache_empty += rq_stats->cache_empty; @@ -243,18 +252,14 @@ static void mlx5e_update_vport_counters(struct mlx5e_priv *priv) mlx5_cmd_exec(mdev, in, sizeof(in), out, outlen); } -static void mlx5e_update_pport_counters(struct mlx5e_priv *priv) +static void mlx5e_update_pport_counters(struct mlx5e_priv *priv, bool full) { struct mlx5e_pport_stats *pstats = &priv->stats.pport; struct mlx5_core_dev *mdev = priv->mdev; + u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {0}; int sz = MLX5_ST_SZ_BYTES(ppcnt_reg); int prio; void *out; - u32 *in; - - in = mlx5_vzalloc(sz); - if (!in) - goto free_out; MLX5_SET(ppcnt_reg, in, local_port, 1); @@ -262,6 +267,9 @@ static void mlx5e_update_pport_counters(struct mlx5e_priv *priv) MLX5_SET(ppcnt_reg, in, grp, MLX5_IEEE_802_3_COUNTERS_GROUP); mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0); + if (!full) + return; + out = pstats->RFC_2863_counters; MLX5_SET(ppcnt_reg, in, grp, MLX5_RFC_2863_COUNTERS_GROUP); mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0); @@ -287,53 +295,57 @@ static void mlx5e_update_pport_counters(struct mlx5e_priv *priv) mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0); } - -free_out: - kvfree(in); } static void mlx5e_update_q_counter(struct mlx5e_priv *priv) { struct mlx5e_qcounter_stats *qcnt = &priv->stats.qcnt; + u32 out[MLX5_ST_SZ_DW(query_q_counter_out)]; + int err; if (!priv->q_counter) return; - mlx5_core_query_out_of_buffer(priv->mdev, priv->q_counter, - &qcnt->rx_out_of_buffer); + err = mlx5_core_query_q_counter(priv->mdev, priv->q_counter, 0, out, sizeof(out)); + if (err) + return; + + qcnt->rx_out_of_buffer = MLX5_GET(query_q_counter_out, out, out_of_buffer); } static void mlx5e_update_pcie_counters(struct mlx5e_priv *priv) { struct mlx5e_pcie_stats *pcie_stats = &priv->stats.pcie; struct mlx5_core_dev *mdev = priv->mdev; + u32 in[MLX5_ST_SZ_DW(mpcnt_reg)] = {0}; int sz = MLX5_ST_SZ_BYTES(mpcnt_reg); void *out; - u32 *in; if (!MLX5_CAP_MCAM_FEATURE(mdev, pcie_performance_group)) return; - in = mlx5_vzalloc(sz); - if (!in) - return; - out = pcie_stats->pcie_perf_counters; MLX5_SET(mpcnt_reg, in, grp, MLX5_PCIE_PERFORMANCE_COUNTERS_GROUP); mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_MPCNT, 0, 0); - - kvfree(in); } -void mlx5e_update_stats(struct mlx5e_priv *priv) +void mlx5e_update_stats(struct mlx5e_priv *priv, bool full) { - mlx5e_update_pcie_counters(priv); - mlx5e_update_pport_counters(priv); + if (full) { + mlx5e_update_pcie_counters(priv); + mlx5e_ipsec_update_stats(priv); + } + mlx5e_update_pport_counters(priv, full); mlx5e_update_vport_counters(priv); mlx5e_update_q_counter(priv); mlx5e_update_sw_counters(priv); } +static void mlx5e_update_ndo_stats(struct mlx5e_priv *priv) +{ + mlx5e_update_stats(priv, false); +} + void mlx5e_update_stats_work(struct work_struct *work) { struct delayed_work *dwork = to_delayed_work(work); @@ -503,7 +515,7 @@ static int mlx5e_create_umr_mkey(struct mlx5_core_dev *mdev, if (!MLX5E_VALID_NUM_MTTS(npages)) return -EINVAL; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -544,7 +556,6 @@ static int mlx5e_alloc_rq(struct mlx5e_channel *c, void *rqc = rqp->rqc; void *rqc_wq = MLX5_ADDR_OF(rqc, rqc, wq); u32 byte_count; - u32 frag_sz; int npages; int wq_sz; int err; @@ -576,13 +587,8 @@ static int mlx5e_alloc_rq(struct mlx5e_channel *c, goto err_rq_wq_destroy; } - if (rq->xdp_prog) { - rq->buff.map_dir = DMA_BIDIRECTIONAL; - rq->rx_headroom = XDP_PACKET_HEADROOM; - } else { - rq->buff.map_dir = DMA_FROM_DEVICE; - rq->rx_headroom = MLX5_RX_HEADROOM; - } + rq->buff.map_dir = rq->xdp_prog ? DMA_BIDIRECTIONAL : DMA_FROM_DEVICE; + rq->rx_headroom = params->rq_headroom; switch (rq->wq_type) { case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ: @@ -591,6 +597,13 @@ static int mlx5e_alloc_rq(struct mlx5e_channel *c, rq->dealloc_wqe = mlx5e_dealloc_rx_mpwqe; rq->handle_rx_cqe = c->priv->profile->rx_handlers.handle_rx_cqe_mpwqe; +#ifdef CONFIG_MLX5_EN_IPSEC + if (MLX5_IPSEC_DEV(mdev)) { + err = -EINVAL; + netdev_err(c->netdev, "MPWQE RQ with IPSec offload not supported\n"); + goto err_rq_wq_destroy; + } +#endif if (!rq->handle_rx_cqe) { err = -EINVAL; netdev_err(c->netdev, "RX handler of MPWQE RQ is not set, err %d\n", err); @@ -613,18 +626,24 @@ static int mlx5e_alloc_rq(struct mlx5e_channel *c, goto err_destroy_umr_mkey; break; default: /* MLX5_WQ_TYPE_LINKED_LIST */ - rq->dma_info = kzalloc_node(wq_sz * sizeof(*rq->dma_info), - GFP_KERNEL, cpu_to_node(c->cpu)); - if (!rq->dma_info) { + rq->wqe.frag_info = + kzalloc_node(wq_sz * sizeof(*rq->wqe.frag_info), + GFP_KERNEL, cpu_to_node(c->cpu)); + if (!rq->wqe.frag_info) { err = -ENOMEM; goto err_rq_wq_destroy; } rq->alloc_wqe = mlx5e_alloc_rx_wqe; rq->dealloc_wqe = mlx5e_dealloc_rx_wqe; - rq->handle_rx_cqe = c->priv->profile->rx_handlers.handle_rx_cqe; +#ifdef CONFIG_MLX5_EN_IPSEC + if (c->priv->ipsec) + rq->handle_rx_cqe = mlx5e_ipsec_handle_rx_cqe; + else +#endif + rq->handle_rx_cqe = c->priv->profile->rx_handlers.handle_rx_cqe; if (!rq->handle_rx_cqe) { - kfree(rq->dma_info); + kfree(rq->wqe.frag_info); err = -EINVAL; netdev_err(c->netdev, "RX handler of RQ is not set, err %d\n", err); goto err_rq_wq_destroy; @@ -632,16 +651,17 @@ static int mlx5e_alloc_rq(struct mlx5e_channel *c, rq->buff.wqe_sz = params->lro_en ? params->lro_wqe_sz : - MLX5E_SW2HW_MTU(c->netdev->mtu); + MLX5E_SW2HW_MTU(c->priv, c->netdev->mtu); +#ifdef CONFIG_MLX5_EN_IPSEC + if (MLX5_IPSEC_DEV(mdev)) + rq->buff.wqe_sz += MLX5E_METADATA_ETHER_LEN; +#endif + rq->wqe.page_reuse = !params->xdp_prog && !params->lro_en; byte_count = rq->buff.wqe_sz; /* calc the required page order */ - frag_sz = rq->rx_headroom + - byte_count /* packet data */ + - SKB_DATA_ALIGN(sizeof(struct skb_shared_info)); - frag_sz = SKB_DATA_ALIGN(frag_sz); - - npages = DIV_ROUND_UP(frag_sz, PAGE_SIZE); + rq->wqe.frag_sz = MLX5_SKB_FRAG_SZ(rq->rx_headroom + byte_count); + npages = DIV_ROUND_UP(rq->wqe.frag_sz, PAGE_SIZE); rq->buff.page_order = order_base_2(npages); byte_count |= MLX5_HW_START_PADDING; @@ -686,7 +706,7 @@ static void mlx5e_free_rq(struct mlx5e_rq *rq) mlx5_core_destroy_mkey(rq->mdev, &rq->umr_mkey); break; default: /* MLX5_WQ_TYPE_LINKED_LIST */ - kfree(rq->dma_info); + kfree(rq->wqe.frag_info); } for (i = rq->page_cache.head; i != rq->page_cache.tail; @@ -711,7 +731,7 @@ static int mlx5e_create_rq(struct mlx5e_rq *rq, inlen = MLX5_ST_SZ_BYTES(create_rq_in) + sizeof(u64) * rq->wq_ctrl.buf.npages; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -748,7 +768,7 @@ static int mlx5e_modify_rq_state(struct mlx5e_rq *rq, int curr_state, int err; inlen = MLX5_ST_SZ_BYTES(modify_rq_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -776,7 +796,7 @@ static int mlx5e_modify_rq_scatter_fcs(struct mlx5e_rq *rq, bool enable) int err; inlen = MLX5_ST_SZ_BYTES(modify_rq_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -805,7 +825,7 @@ static int mlx5e_modify_rq_vsd(struct mlx5e_rq *rq, bool vsd) int err; inlen = MLX5_ST_SZ_BYTES(modify_rq_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -868,6 +888,16 @@ static void mlx5e_free_rx_descs(struct mlx5e_rq *rq) mlx5_wq_ll_pop(&rq->wq, wqe_ix_be, &wqe->next.next_wqe_index); } + + if (rq->wq_type == MLX5_WQ_TYPE_LINKED_LIST && rq->wqe.page_reuse) { + /* Clean outstanding pages on handled WQEs that decided to do page-reuse, + * but yet to be re-posted. + */ + int wq_sz = mlx5_wq_ll_get_size(&rq->wq); + + for (wqe_ix = 0; wqe_ix < wq_sz; wqe_ix++) + rq->dealloc_wqe(rq, wqe_ix); + } } static int mlx5e_open_rq(struct mlx5e_channel *c, @@ -1086,6 +1116,8 @@ static int mlx5e_alloc_txqsq(struct mlx5e_channel *c, sq->uar_map = mdev->mlx5e_res.bfreg.map; sq->max_inline = params->tx_max_inline; sq->min_inline_mode = params->tx_min_inline_mode; + if (MLX5_IPSEC_DEV(c->priv->mdev)) + set_bit(MLX5E_SQ_STATE_IPSEC, &sq->state); param->wq.db_numa_node = cpu_to_node(c->cpu); err = mlx5_wq_cyc_create(mdev, ¶m->wq, sqc_wq, &sq->wq, &sq->wq_ctrl); @@ -1134,7 +1166,7 @@ static int mlx5e_create_sq(struct mlx5_core_dev *mdev, inlen = MLX5_ST_SZ_BYTES(create_sq_in) + sizeof(u64) * csp->wq_ctrl->buf.npages; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -1182,7 +1214,7 @@ static int mlx5e_modify_sq(struct mlx5_core_dev *mdev, u32 sqn, int err; inlen = MLX5_ST_SZ_BYTES(modify_sq_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -1496,7 +1528,7 @@ static int mlx5e_create_cq(struct mlx5e_cq *cq, struct mlx5e_cq_param *param) inlen = MLX5_ST_SZ_BYTES(create_cq_in) + sizeof(u64) * cq->wq_ctrl.frag_buf.npages; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -1905,6 +1937,7 @@ static void mlx5e_build_sq_param(struct mlx5e_priv *priv, mlx5e_build_sq_param_common(priv, param); MLX5_SET(wq, wq, log_wq_sz, params->log_sq_size); + MLX5_SET(sqc, sqc, allow_swp, !!MLX5_IPSEC_DEV(priv->mdev)); } static void mlx5e_build_common_cq_param(struct mlx5e_priv *priv, @@ -2091,7 +2124,7 @@ mlx5e_create_rqt(struct mlx5e_priv *priv, int sz, struct mlx5e_rqt *rqt) int i; inlen = MLX5_ST_SZ_BYTES(create_rqt_in) + sizeof(u32) * sz; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -2210,7 +2243,7 @@ int mlx5e_redirect_rqt(struct mlx5e_priv *priv, u32 rqtn, int sz, int err; inlen = MLX5_ST_SZ_BYTES(modify_rqt_in) + sizeof(u32) * sz; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -2433,7 +2466,7 @@ static int mlx5e_modify_tirs_lro(struct mlx5e_priv *priv) int ix; inlen = MLX5_ST_SZ_BYTES(modify_tir_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -2465,7 +2498,7 @@ free_in: static int mlx5e_set_mtu(struct mlx5e_priv *priv, u16 mtu) { struct mlx5_core_dev *mdev = priv->mdev; - u16 hw_mtu = MLX5E_SW2HW_MTU(mtu); + u16 hw_mtu = MLX5E_SW2HW_MTU(priv, mtu); int err; err = mlx5_set_port_mtu(mdev, hw_mtu, 1); @@ -2487,7 +2520,7 @@ static void mlx5e_query_mtu(struct mlx5e_priv *priv, u16 *mtu) if (err || !hw_mtu) /* fallback to port oper mtu */ mlx5_query_port_oper_mtu(mdev, &hw_mtu, 1); - *mtu = MLX5E_HW2SW_MTU(hw_mtu); + *mtu = MLX5E_HW2SW_MTU(priv, hw_mtu); } static int mlx5e_set_dev_port_mtu(struct mlx5e_priv *priv) @@ -2596,9 +2629,10 @@ void mlx5e_switch_priv_channels(struct mlx5e_priv *priv, { struct net_device *netdev = priv->netdev; int new_num_txqs; - + int carrier_ok; new_num_txqs = new_chs->num * new_chs->params.num_tc; + carrier_ok = netif_carrier_ok(netdev); netif_carrier_off(netdev); if (new_num_txqs < netdev->real_num_tx_queues) @@ -2616,7 +2650,9 @@ void mlx5e_switch_priv_channels(struct mlx5e_priv *priv, mlx5e_refresh_tirs(priv, false); mlx5e_activate_priv_channels(priv); - mlx5e_update_carrier(priv); + /* return carrier back if needed */ + if (carrier_ok) + netif_carrier_on(netdev); } int mlx5e_open_locked(struct net_device *netdev) @@ -2632,7 +2668,8 @@ int mlx5e_open_locked(struct net_device *netdev) mlx5e_refresh_tirs(priv, false); mlx5e_activate_priv_channels(priv); - mlx5e_update_carrier(priv); + if (priv->profile->update_carrier) + priv->profile->update_carrier(priv); mlx5e_timestamp_init(priv); if (priv->profile->update_stats) @@ -2850,7 +2887,7 @@ int mlx5e_create_indirect_tirs(struct mlx5e_priv *priv) int tt; inlen = MLX5_ST_SZ_BYTES(create_tir_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -2889,7 +2926,7 @@ int mlx5e_create_direct_tirs(struct mlx5e_priv *priv) int ix; inlen = MLX5_ST_SZ_BYTES(create_tir_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -2976,7 +3013,7 @@ static int mlx5e_setup_tc(struct net_device *netdev, u8 tc) new_channels.params = priv->channels.params; new_channels.params.num_tc = tc ? tc : 1; - if (test_bit(MLX5E_STATE_OPENED, &priv->state)) { + if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) { priv->channels.params = new_channels.params; goto out; } @@ -2992,13 +3029,17 @@ out: } static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle, - __be16 proto, struct tc_to_netdev *tc) + u32 chain_index, __be16 proto, + struct tc_to_netdev *tc) { struct mlx5e_priv *priv = netdev_priv(dev); if (TC_H_MAJ(handle) != TC_H_MAJ(TC_H_INGRESS)) goto mqprio; + if (chain_index) + return -EOPNOTSUPP; + switch (tc->type) { case TC_SETUP_CLSFLOWER: switch (tc->cls_flower->command) { @@ -3053,8 +3094,6 @@ mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats) PPORT_802_3_GET(pstats, a_frame_check_sequence_errors); stats->rx_frame_errors = PPORT_802_3_GET(pstats, a_alignment_errors); stats->tx_aborted_errors = PPORT_2863_GET(pstats, if_out_discards); - stats->tx_carrier_errors = - PPORT_802_3_GET(pstats, a_symbol_error_during_carrier); stats->rx_errors = stats->rx_length_errors + stats->rx_crc_errors + stats->rx_frame_errors; stats->tx_errors = stats->tx_aborted_errors + stats->tx_carrier_errors; @@ -3064,7 +3103,6 @@ mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats) */ stats->multicast = VPORT_COUNTER_GET(vstats, received_eth_multicast.packets); - } static void mlx5e_set_rx_mode(struct net_device *dev) @@ -3307,11 +3345,13 @@ out: static int mlx5e_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) { + struct mlx5e_priv *priv = netdev_priv(dev); + switch (cmd) { case SIOCSHWTSTAMP: - return mlx5e_hwstamp_set(dev, ifr); + return mlx5e_hwstamp_set(priv, ifr); case SIOCGHWTSTAMP: - return mlx5e_hwstamp_get(dev, ifr); + return mlx5e_hwstamp_get(priv, ifr); default: return -EOPNOTSUPP; } @@ -3490,6 +3530,11 @@ static netdev_features_t mlx5e_features_check(struct sk_buff *skb, features = vlan_features_check(skb, features); features = vxlan_features_check(skb, features); +#ifdef CONFIG_MLX5_EN_IPSEC + if (mlx5e_ipsec_feature_check(skb, netdev, features)) + return features; +#endif + /* Validate if the tunneled packet is being offloaded by HW */ if (skb->encapsulation && (features & NETIF_F_CSUM_MASK || features & NETIF_F_GSO_MASK)) @@ -3537,6 +3582,12 @@ static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog) goto unlock; } + if ((netdev->features & NETIF_F_HW_ESP) && prog) { + netdev_warn(netdev, "can't set XDP with IPSec offload\n"); + err = -EINVAL; + goto unlock; + } + was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state); /* no need for full reset when exchanging programs */ reset = (!priv->channels.params.xdp_prog || !prog); @@ -3596,11 +3647,19 @@ unlock: return err; } -static bool mlx5e_xdp_attached(struct net_device *dev) +static u32 mlx5e_xdp_query(struct net_device *dev) { struct mlx5e_priv *priv = netdev_priv(dev); + const struct bpf_prog *xdp_prog; + u32 prog_id = 0; + + mutex_lock(&priv->state_lock); + xdp_prog = priv->channels.params.xdp_prog; + if (xdp_prog) + prog_id = xdp_prog->aux->id; + mutex_unlock(&priv->state_lock); - return !!priv->channels.params.xdp_prog; + return prog_id; } static int mlx5e_xdp(struct net_device *dev, struct netdev_xdp *xdp) @@ -3609,7 +3668,8 @@ static int mlx5e_xdp(struct net_device *dev, struct netdev_xdp *xdp) case XDP_SETUP_PROG: return mlx5e_xdp_set(dev, xdp->prog); case XDP_QUERY_PROG: - xdp->prog_attached = mlx5e_xdp_attached(dev); + xdp->prog_id = mlx5e_xdp_query(dev); + xdp->prog_attached = !!xdp->prog_id; return 0; default: return -EINVAL; @@ -3715,7 +3775,7 @@ static int mlx5e_check_required_hca_cap(struct mlx5_core_dev *mdev) if (!MLX5_CAP_ETH(mdev, self_lb_en_modifiable)) mlx5_core_warn(mdev, "Self loop back prevention is not supported\n"); if (!MLX5_CAP_GEN(mdev, cq_moderation)) - mlx5_core_warn(mdev, "CQ modiration is not supported\n"); + mlx5_core_warn(mdev, "CQ moderation is not supported\n"); return 0; } @@ -3848,7 +3908,7 @@ void mlx5e_build_nic_params(struct mlx5_core_dev *mdev, /* set CQE compression */ params->rx_cqe_compress_def = false; if (MLX5_CAP_GEN(mdev, cqe_compression) && - MLX5_CAP_GEN(mdev, vport_group_manager)) + MLX5_CAP_GEN(mdev, vport_group_manager)) params->rx_cqe_compress_def = cqe_compress_heuristic(link_speed, pci_bw); MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_CQE_COMPRESS, params->rx_cqe_compress_def); @@ -3857,6 +3917,7 @@ void mlx5e_build_nic_params(struct mlx5_core_dev *mdev, mlx5e_set_rq_params(mdev, params); /* HW LRO */ + /* TODO: && MLX5_CAP_ETH(mdev, lro_cap) */ if (params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) params->lro_en = hw_lro_heuristic(link_speed, pci_bw); @@ -3897,6 +3958,7 @@ static void mlx5e_build_nic_netdev_priv(struct mlx5_core_dev *mdev, priv->netdev = netdev; priv->profile = profile; priv->ppriv = ppriv; + priv->hard_mtu = MLX5E_ETH_HARD_MTU; mlx5e_build_nic_params(mdev, &priv->channels.params, profile->max_nch(mdev)); @@ -4017,6 +4079,8 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev) if (MLX5_CAP_GEN(mdev, vport_group_manager)) netdev->switchdev_ops = &mlx5e_switchdev_ops; #endif + + mlx5e_ipsec_build_netdev(priv); } static void mlx5e_create_q_counter(struct mlx5e_priv *priv) @@ -4045,14 +4109,19 @@ static void mlx5e_nic_init(struct mlx5_core_dev *mdev, void *ppriv) { struct mlx5e_priv *priv = netdev_priv(netdev); + int err; mlx5e_build_nic_netdev_priv(mdev, netdev, profile, ppriv); + err = mlx5e_ipsec_init(priv); + if (err) + mlx5_core_err(mdev, "IPSec initialization failed, %d\n", err); mlx5e_build_nic_netdev(netdev); mlx5e_vxlan_init(priv); } static void mlx5e_nic_cleanup(struct mlx5e_priv *priv) { + mlx5e_ipsec_cleanup(priv); mlx5e_vxlan_cleanup(priv); if (priv->channels.params.xdp_prog) @@ -4142,7 +4211,7 @@ static void mlx5e_nic_enable(struct mlx5e_priv *priv) /* MTU range: 68 - hw-specific max */ netdev->min_mtu = ETH_MIN_MTU; mlx5_query_port_max_mtu(priv->mdev, &max_mtu, 1); - netdev->max_mtu = MLX5E_HW2SW_MTU(max_mtu); + netdev->max_mtu = MLX5E_HW2SW_MTU(priv, max_mtu); mlx5e_set_dev_port_mtu(priv); mlx5_lag_add(mdev, netdev); @@ -4199,8 +4268,9 @@ static const struct mlx5e_profile mlx5e_nic_profile = { .cleanup_tx = mlx5e_cleanup_nic_tx, .enable = mlx5e_nic_enable, .disable = mlx5e_nic_disable, - .update_stats = mlx5e_update_stats, + .update_stats = mlx5e_update_ndo_stats, .max_nch = mlx5e_get_max_num_channels, + .update_carrier = mlx5e_update_carrier, .rx_handlers.handle_rx_cqe = mlx5e_handle_rx_cqe, .rx_handlers.handle_rx_cqe_mpwqe = mlx5e_handle_rx_cqe_mpwrq, .max_tc = MLX5E_MAX_NUM_TC, @@ -4241,7 +4311,8 @@ struct net_device *mlx5e_create_netdev(struct mlx5_core_dev *mdev, return netdev; err_cleanup_nic: - profile->cleanup(priv); + if (profile->cleanup) + profile->cleanup(priv); free_netdev(netdev); return NULL; @@ -4442,6 +4513,7 @@ static struct mlx5_interface mlx5e_interface = { void mlx5e_init(void) { + mlx5e_ipsec_build_inverse_table(); mlx5e_build_ptys2ethtool_map(); mlx5_register_interface(&mlx5e_interface); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 79462c0368a0..45e60be9c277 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -652,7 +652,8 @@ static int mlx5e_rep_get_phys_port_name(struct net_device *dev, } static int mlx5e_rep_ndo_setup_tc(struct net_device *dev, u32 handle, - __be16 proto, struct tc_to_netdev *tc) + u32 chain_index, __be16 proto, + struct tc_to_netdev *tc) { struct mlx5e_priv *priv = netdev_priv(dev); @@ -664,9 +665,13 @@ static int mlx5e_rep_ndo_setup_tc(struct net_device *dev, u32 handle, struct net_device *uplink_dev = mlx5_eswitch_get_uplink_netdev(esw); return uplink_dev->netdev_ops->ndo_setup_tc(uplink_dev, handle, + chain_index, proto, tc); } + if (chain_index) + return -EOPNOTSUPP; + switch (tc->type) { case TC_SETUP_CLSFLOWER: switch (tc->cls_flower->command) { @@ -791,6 +796,8 @@ static void mlx5e_build_rep_params(struct mlx5_core_dev *mdev, params->tx_max_inline = mlx5e_get_max_inline_cap(mdev); params->num_tc = 1; params->lro_wqe_sz = MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ; + + mlx5_query_min_inline(mdev, ¶ms->tx_min_inline_mode); } static void mlx5e_build_rep_netdev(struct net_device *netdev) @@ -828,6 +835,9 @@ static void mlx5e_init_rep(struct mlx5_core_dev *mdev, INIT_DELAYED_WORK(&priv->update_stats_work, mlx5e_update_stats_work); priv->channels.params.num_channels = profile->max_nch(mdev); + + priv->hard_mtu = MLX5E_ETH_HARD_MTU; + mlx5e_build_rep_params(mdev, &priv->channels.params); mlx5e_build_rep_netdev(netdev); } @@ -911,6 +921,7 @@ static struct mlx5e_profile mlx5e_rep_profile = { .cleanup_tx = mlx5e_cleanup_nic_tx, .update_stats = mlx5e_rep_update_stats, .max_nch = mlx5e_get_rep_max_num_channels, + .update_carrier = NULL, .rx_handlers.handle_rx_cqe = mlx5e_handle_rx_cqe_rep, .rx_handlers.handle_rx_cqe_mpwqe = NULL /* Not supported */, .max_tc = 1, @@ -1014,7 +1025,6 @@ err_destroy_netdev: mlx5e_destroy_netdev(netdev_priv(netdev)); kfree(rpriv); return err; - } static void diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c index 7b1566f0ae58..325b2c8c1c6d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c @@ -40,7 +40,8 @@ #include "en_tc.h" #include "eswitch.h" #include "en_rep.h" -#include "ipoib.h" +#include "ipoib/ipoib.h" +#include "en_accel/ipsec_rxtx.h" static inline bool mlx5e_rx_hw_stamp(struct mlx5e_tstamp *tstamp) { @@ -160,6 +161,11 @@ static inline u32 mlx5e_decompress_cqes_start(struct mlx5e_rq *rq, #define RQ_PAGE_SIZE(rq) ((1 << rq->buff.page_order) << PAGE_SHIFT) +static inline bool mlx5e_page_is_reserved(struct page *page) +{ + return page_is_pfmemalloc(page) || page_to_nid(page) != numa_node_id(); +} + static inline bool mlx5e_rx_cache_put(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info) { @@ -238,22 +244,54 @@ void mlx5e_page_release(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info, put_page(dma_info->page); } +static inline bool mlx5e_page_reuse(struct mlx5e_rq *rq, + struct mlx5e_wqe_frag_info *wi) +{ + return rq->wqe.page_reuse && wi->di.page && + (wi->offset + rq->wqe.frag_sz <= RQ_PAGE_SIZE(rq)) && + !mlx5e_page_is_reserved(wi->di.page); +} + int mlx5e_alloc_rx_wqe(struct mlx5e_rq *rq, struct mlx5e_rx_wqe *wqe, u16 ix) { - struct mlx5e_dma_info *di = &rq->dma_info[ix]; + struct mlx5e_wqe_frag_info *wi = &rq->wqe.frag_info[ix]; - if (unlikely(mlx5e_page_alloc_mapped(rq, di))) - return -ENOMEM; + /* check if page exists, hence can be reused */ + if (!wi->di.page) { + if (unlikely(mlx5e_page_alloc_mapped(rq, &wi->di))) + return -ENOMEM; + wi->offset = 0; + } - wqe->data.addr = cpu_to_be64(di->addr + rq->rx_headroom); + wqe->data.addr = cpu_to_be64(wi->di.addr + wi->offset + + rq->rx_headroom); return 0; } +static inline void mlx5e_free_rx_wqe(struct mlx5e_rq *rq, + struct mlx5e_wqe_frag_info *wi) +{ + mlx5e_page_release(rq, &wi->di, true); + wi->di.page = NULL; +} + +static inline void mlx5e_free_rx_wqe_reuse(struct mlx5e_rq *rq, + struct mlx5e_wqe_frag_info *wi) +{ + if (mlx5e_page_reuse(rq, wi)) { + rq->stats.page_reuse++; + return; + } + + mlx5e_free_rx_wqe(rq, wi); +} + void mlx5e_dealloc_rx_wqe(struct mlx5e_rq *rq, u16 ix) { - struct mlx5e_dma_info *di = &rq->dma_info[ix]; + struct mlx5e_wqe_frag_info *wi = &rq->wqe.frag_info[ix]; - mlx5e_page_release(rq, di, true); + if (wi->di.page) + mlx5e_free_rx_wqe(rq, wi); } static inline int mlx5e_mpwqe_strides_per_page(struct mlx5e_rq *rq) @@ -648,9 +686,8 @@ static inline bool mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq, prefetchw(wqe); if (unlikely(dma_len < MLX5E_XDP_MIN_INLINE || - MLX5E_SW2HW_MTU(rq->netdev->mtu) < dma_len)) { + MLX5E_SW2HW_MTU(rq->channel->priv, rq->netdev->mtu) < dma_len)) { rq->stats.xdp_drop++; - mlx5e_page_release(rq, di, true); return false; } @@ -661,7 +698,6 @@ static inline bool mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq, sq->db.doorbell = false; } rq->stats.xdp_tx_full++; - mlx5e_page_release(rq, di, true); return false; } @@ -686,10 +722,15 @@ static inline bool mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq, cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | MLX5_OPCODE_SEND); + /* move page to reference to sq responsibility, + * and mark so it's not put back in page-cache. + */ + rq->wqe.xdp_xmit = true; sq->db.di[pi] = *di; sq->pc++; sq->db.doorbell = true; + rq->stats.xdp_tx++; return true; } @@ -726,35 +767,34 @@ static inline int mlx5e_xdp_handle(struct mlx5e_rq *rq, trace_xdp_exception(rq->netdev, prog, act); case XDP_DROP: rq->stats.xdp_drop++; - mlx5e_page_release(rq, di, true); return true; } } static inline struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe, - u16 wqe_counter, u32 cqe_bcnt) + struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt) { - struct mlx5e_dma_info *di; + struct mlx5e_dma_info *di = &wi->di; struct sk_buff *skb; void *va, *data; u16 rx_headroom = rq->rx_headroom; bool consumed; + u32 frag_size; - di = &rq->dma_info[wqe_counter]; - va = page_address(di->page); + va = page_address(di->page) + wi->offset; data = va + rx_headroom; + frag_size = MLX5_SKB_FRAG_SZ(rx_headroom + cqe_bcnt); dma_sync_single_range_for_cpu(rq->pdev, - di->addr, - rx_headroom, - rq->buff.wqe_sz, + di->addr + wi->offset, + 0, frag_size, DMA_FROM_DEVICE); prefetch(data); + wi->offset += frag_size; if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) { rq->stats.wqe_err++; - mlx5e_page_release(rq, di, true); return NULL; } @@ -764,16 +804,14 @@ struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe, if (consumed) return NULL; /* page/packet was consumed by XDP */ - skb = build_skb(va, RQ_PAGE_SIZE(rq)); + skb = build_skb(va, frag_size); if (unlikely(!skb)) { rq->stats.buff_alloc_err++; - mlx5e_page_release(rq, di, true); return NULL; } - /* queue up for recycling ..*/ + /* queue up for recycling/reuse */ page_ref_inc(di->page); - mlx5e_page_release(rq, di, true); skb_reserve(skb, rx_headroom); skb_put(skb, cqe_bcnt); @@ -783,6 +821,7 @@ struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe, void mlx5e_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) { + struct mlx5e_wqe_frag_info *wi; struct mlx5e_rx_wqe *wqe; __be16 wqe_counter_be; struct sk_buff *skb; @@ -792,15 +831,27 @@ void mlx5e_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) wqe_counter_be = cqe->wqe_counter; wqe_counter = be16_to_cpu(wqe_counter_be); wqe = mlx5_wq_ll_get_wqe(&rq->wq, wqe_counter); + wi = &rq->wqe.frag_info[wqe_counter]; cqe_bcnt = be32_to_cpu(cqe->byte_cnt); - skb = skb_from_cqe(rq, cqe, wqe_counter, cqe_bcnt); - if (!skb) + skb = skb_from_cqe(rq, cqe, wi, cqe_bcnt); + if (!skb) { + /* probably for XDP */ + if (rq->wqe.xdp_xmit) { + wi->di.page = NULL; + rq->wqe.xdp_xmit = false; + /* do not return page to cache, it will be returned on XDP_TX completion */ + goto wq_ll_pop; + } + /* probably an XDP_DROP, save the page-reuse checks */ + mlx5e_free_rx_wqe(rq, wi); goto wq_ll_pop; + } mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb); napi_gro_receive(rq->cq.napi, skb); + mlx5e_free_rx_wqe_reuse(rq, wi); wq_ll_pop: mlx5_wq_ll_pop(&rq->wq, wqe_counter_be, &wqe->next.next_wqe_index); @@ -812,6 +863,7 @@ void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) struct mlx5e_priv *priv = netdev_priv(netdev); struct mlx5e_rep_priv *rpriv = priv->ppriv; struct mlx5_eswitch_rep *rep = rpriv->rep; + struct mlx5e_wqe_frag_info *wi; struct mlx5e_rx_wqe *wqe; struct sk_buff *skb; __be16 wqe_counter_be; @@ -821,11 +873,21 @@ void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) wqe_counter_be = cqe->wqe_counter; wqe_counter = be16_to_cpu(wqe_counter_be); wqe = mlx5_wq_ll_get_wqe(&rq->wq, wqe_counter); + wi = &rq->wqe.frag_info[wqe_counter]; cqe_bcnt = be32_to_cpu(cqe->byte_cnt); - skb = skb_from_cqe(rq, cqe, wqe_counter, cqe_bcnt); - if (!skb) + skb = skb_from_cqe(rq, cqe, wi, cqe_bcnt); + if (!skb) { + if (rq->wqe.xdp_xmit) { + wi->di.page = NULL; + rq->wqe.xdp_xmit = false; + /* do not return page to cache, it will be returned on XDP_TX completion */ + goto wq_ll_pop; + } + /* probably an XDP_DROP, save the page-reuse checks */ + mlx5e_free_rx_wqe(rq, wi); goto wq_ll_pop; + } mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb); @@ -834,6 +896,7 @@ void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) napi_gro_receive(rq->cq.napi, skb); + mlx5e_free_rx_wqe_reuse(rq, wi); wq_ll_pop: mlx5_wq_ll_pop(&rq->wq, wqe_counter_be, &wqe->next.next_wqe_index); @@ -934,7 +997,7 @@ int mlx5e_poll_rx_cq(struct mlx5e_cq *cq, int budget) work_done += mlx5e_decompress_cqes_cont(rq, cq, 0, budget); for (; work_done < budget; work_done++) { - struct mlx5_cqe64 *cqe = mlx5e_get_cqe(cq); + struct mlx5_cqe64 *cqe = mlx5_cqwq_get_cqe(&cq->wq); if (!cqe) break; @@ -988,7 +1051,7 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq) u16 wqe_counter; bool last_wqe; - cqe = mlx5e_get_cqe(cq); + cqe = mlx5_cqwq_get_cqe(&cq->wq); if (!cqe) break; @@ -1038,8 +1101,6 @@ void mlx5e_free_xdpsq_descs(struct mlx5e_xdpsq *sq) #ifdef CONFIG_MLX5_CORE_IPOIB #define MLX5_IB_GRH_DGID_OFFSET 24 -#define MLX5_IB_GRH_BYTES 40 -#define MLX5_IPOIB_ENCAP_LEN 4 #define MLX5_GID_SIZE 16 static inline void mlx5i_complete_rx_cqe(struct mlx5e_rq *rq, @@ -1048,6 +1109,8 @@ static inline void mlx5i_complete_rx_cqe(struct mlx5e_rq *rq, struct sk_buff *skb) { struct net_device *netdev = rq->netdev; + struct mlx5e_tstamp *tstamp = rq->tstamp; + char *pseudo_header; u8 *dgid; u8 g; @@ -1071,13 +1134,19 @@ static inline void mlx5i_complete_rx_cqe(struct mlx5e_rq *rq, skb->ip_summed = CHECKSUM_COMPLETE; skb->csum = csum_unfold((__force __sum16)cqe->check_sum); + if (unlikely(mlx5e_rx_hw_stamp(tstamp))) + mlx5e_fill_hwstamp(tstamp, get_cqe_ts(cqe), skb_hwtstamps(skb)); + skb_record_rx_queue(skb, rq->ix); if (likely(netdev->features & NETIF_F_RXHASH)) mlx5e_skb_set_hash(cqe, skb); + /* 20 bytes of ipoib header and 4 for encap existing */ + pseudo_header = skb_push(skb, MLX5_IPOIB_PSEUDO_LEN); + memset(pseudo_header, 0, MLX5_IPOIB_PSEUDO_LEN); skb_reset_mac_header(skb); - skb_pull(skb, MLX5_IPOIB_ENCAP_LEN); + skb_pull(skb, MLX5_IPOIB_HARD_LEN); skb->dev = netdev; @@ -1088,6 +1157,7 @@ static inline void mlx5i_complete_rx_cqe(struct mlx5e_rq *rq, void mlx5i_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) { + struct mlx5e_wqe_frag_info *wi; struct mlx5e_rx_wqe *wqe; __be16 wqe_counter_be; struct sk_buff *skb; @@ -1097,18 +1167,60 @@ void mlx5i_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) wqe_counter_be = cqe->wqe_counter; wqe_counter = be16_to_cpu(wqe_counter_be); wqe = mlx5_wq_ll_get_wqe(&rq->wq, wqe_counter); + wi = &rq->wqe.frag_info[wqe_counter]; cqe_bcnt = be32_to_cpu(cqe->byte_cnt); - skb = skb_from_cqe(rq, cqe, wqe_counter, cqe_bcnt); + skb = skb_from_cqe(rq, cqe, wi, cqe_bcnt); if (!skb) - goto wq_ll_pop; + goto wq_free_wqe; mlx5i_complete_rx_cqe(rq, cqe, cqe_bcnt, skb); napi_gro_receive(rq->cq.napi, skb); -wq_ll_pop: +wq_free_wqe: + mlx5e_free_rx_wqe_reuse(rq, wi); mlx5_wq_ll_pop(&rq->wq, wqe_counter_be, &wqe->next.next_wqe_index); } #endif /* CONFIG_MLX5_CORE_IPOIB */ + +#ifdef CONFIG_MLX5_EN_IPSEC + +void mlx5e_ipsec_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe) +{ + struct mlx5e_wqe_frag_info *wi; + struct mlx5e_rx_wqe *wqe; + __be16 wqe_counter_be; + struct sk_buff *skb; + u16 wqe_counter; + u32 cqe_bcnt; + + wqe_counter_be = cqe->wqe_counter; + wqe_counter = be16_to_cpu(wqe_counter_be); + wqe = mlx5_wq_ll_get_wqe(&rq->wq, wqe_counter); + wi = &rq->wqe.frag_info[wqe_counter]; + cqe_bcnt = be32_to_cpu(cqe->byte_cnt); + + skb = skb_from_cqe(rq, cqe, wi, cqe_bcnt); + if (unlikely(!skb)) { + /* a DROP, save the page-reuse checks */ + mlx5e_free_rx_wqe(rq, wi); + goto wq_ll_pop; + } + skb = mlx5e_ipsec_handle_rx_skb(rq->netdev, skb); + if (unlikely(!skb)) { + mlx5e_free_rx_wqe(rq, wi); + goto wq_ll_pop; + } + + mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb); + napi_gro_receive(rq->cq.napi, skb); + + mlx5e_free_rx_wqe_reuse(rq, wi); +wq_ll_pop: + mlx5_wq_ll_pop(&rq->wq, wqe_counter_be, + &wqe->next.next_wqe_index); +} + +#endif /* CONFIG_MLX5_EN_IPSEC */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c index 02dd3a95ed8f..acf32fe952cd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c @@ -183,28 +183,27 @@ static void mlx5e_am_exit_parking(struct mlx5e_rx_am *am) mlx5e_am_step(am); } +#define IS_SIGNIFICANT_DIFF(val, ref) \ + (((100 * abs((val) - (ref))) / (ref)) > 10) /* more than 10% difference */ + static int mlx5e_am_stats_compare(struct mlx5e_rx_am_stats *curr, struct mlx5e_rx_am_stats *prev) { - int diff; - - if (!prev->ppms) - return curr->ppms ? MLX5E_AM_STATS_BETTER : + if (!prev->bpms) + return curr->bpms ? MLX5E_AM_STATS_BETTER : MLX5E_AM_STATS_SAME; - diff = curr->ppms - prev->ppms; - if (((100 * abs(diff)) / prev->ppms) > 10) /* more than 10% diff */ - return (diff > 0) ? MLX5E_AM_STATS_BETTER : - MLX5E_AM_STATS_WORSE; + if (IS_SIGNIFICANT_DIFF(curr->bpms, prev->bpms)) + return (curr->bpms > prev->bpms) ? MLX5E_AM_STATS_BETTER : + MLX5E_AM_STATS_WORSE; - if (!prev->epms) - return curr->epms ? MLX5E_AM_STATS_WORSE : - MLX5E_AM_STATS_SAME; + if (IS_SIGNIFICANT_DIFF(curr->ppms, prev->ppms)) + return (curr->ppms > prev->ppms) ? MLX5E_AM_STATS_BETTER : + MLX5E_AM_STATS_WORSE; - diff = curr->epms - prev->epms; - if (((100 * abs(diff)) / prev->epms) > 10) /* more than 10% diff */ - return (diff < 0) ? MLX5E_AM_STATS_BETTER : - MLX5E_AM_STATS_WORSE; + if (IS_SIGNIFICANT_DIFF(curr->epms, prev->epms)) + return (curr->epms < prev->epms) ? MLX5E_AM_STATS_BETTER : + MLX5E_AM_STATS_WORSE; return MLX5E_AM_STATS_SAME; } @@ -266,10 +265,13 @@ static void mlx5e_am_sample(struct mlx5e_rq *rq, { s->time = ktime_get(); s->pkt_ctr = rq->stats.packets; + s->byte_ctr = rq->stats.bytes; s->event_ctr = rq->cq.event_ctr; } #define MLX5E_AM_NEVENTS 64 +#define BITS_PER_TYPE(type) (sizeof(type) * BITS_PER_BYTE) +#define BIT_GAP(bits, end, start) ((((end) - (start)) + BIT_ULL(bits)) & (BIT_ULL(bits) - 1)) static void mlx5e_am_calc_stats(struct mlx5e_rx_am_sample *start, struct mlx5e_rx_am_sample *end, @@ -277,13 +279,17 @@ static void mlx5e_am_calc_stats(struct mlx5e_rx_am_sample *start, { /* u32 holds up to 71 minutes, should be enough */ u32 delta_us = ktime_us_delta(end->time, start->time); - unsigned int npkts = end->pkt_ctr - start->pkt_ctr; + u32 npkts = BIT_GAP(BITS_PER_TYPE(u32), end->pkt_ctr, start->pkt_ctr); + u32 nbytes = BIT_GAP(BITS_PER_TYPE(u32), end->byte_ctr, + start->byte_ctr); if (!delta_us) return; - curr_stats->ppms = (npkts * USEC_PER_MSEC) / delta_us; - curr_stats->epms = (MLX5E_AM_NEVENTS * USEC_PER_MSEC) / delta_us; + curr_stats->ppms = DIV_ROUND_UP(npkts * USEC_PER_MSEC, delta_us); + curr_stats->bpms = DIV_ROUND_UP(nbytes * USEC_PER_MSEC, delta_us); + curr_stats->epms = DIV_ROUND_UP(MLX5E_AM_NEVENTS * USEC_PER_MSEC, + delta_us); } void mlx5e_rx_am_work(struct work_struct *work) @@ -308,7 +314,8 @@ void mlx5e_rx_am(struct mlx5e_rq *rq) switch (am->state) { case MLX5E_AM_MEASURE_IN_PROGRESS: - nevents = rq->cq.event_ctr - am->start_sample.event_ctr; + nevents = BIT_GAP(BITS_PER_TYPE(u16), rq->cq.event_ctr, + am->start_sample.event_ctr); if (nevents < MLX5E_AM_NEVENTS) break; mlx5e_am_sample(rq, &end_sample); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c b/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c index 5225f2226a67..898759fcf9ec 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_selftest.c @@ -132,14 +132,14 @@ static struct sk_buff *mlx5e_test_get_udp_skb(struct mlx5e_priv *priv) skb_reserve(skb, NET_IP_ALIGN); /* Reserve for ethernet and IP header */ - ethh = (struct ethhdr *)skb_push(skb, ETH_HLEN); + ethh = skb_push(skb, ETH_HLEN); skb_reset_mac_header(skb); skb_set_network_header(skb, skb->len); - iph = (struct iphdr *)skb_put(skb, sizeof(struct iphdr)); + iph = skb_put(skb, sizeof(struct iphdr)); skb_set_transport_header(skb, skb->len); - udph = (struct udphdr *)skb_put(skb, sizeof(struct udphdr)); + udph = skb_put(skb, sizeof(struct udphdr)); /* Fill ETH header */ ether_addr_copy(ethh->h_dest, priv->netdev->dev_addr); @@ -167,12 +167,12 @@ static struct sk_buff *mlx5e_test_get_udp_skb(struct mlx5e_priv *priv) ip_send_check(iph); /* Fill test header and data */ - mlxh = (struct mlx5ehdr *)skb_put(skb, sizeof(*mlxh)); + mlxh = skb_put(skb, sizeof(*mlxh)); mlxh->version = 0; mlxh->magic = cpu_to_be64(MLX5E_TEST_MAGIC); strlcpy(mlxh->text, mlx5e_test_text, sizeof(mlxh->text)); datalen -= sizeof(*mlxh); - memset(skb_put(skb, datalen), 0, datalen); + skb_put_zero(skb, datalen); skb->csum = 0; skb->ip_summed = CHECKSUM_PARTIAL; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h index 53e4992d6511..e65517eafc58 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h @@ -79,6 +79,7 @@ struct mlx5e_sw_stats { u64 rx_buff_alloc_err; u64 rx_cqe_compress_blks; u64 rx_cqe_compress_pkts; + u64 rx_page_reuse; u64 rx_cache_reuse; u64 rx_cache_full; u64 rx_cache_empty; @@ -117,6 +118,7 @@ static const struct counter_desc sw_stats_desc[] = { { MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_buff_alloc_err) }, { MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cqe_compress_blks) }, { MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cqe_compress_pkts) }, + { MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_page_reuse) }, { MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cache_reuse) }, { MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cache_full) }, { MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_cache_empty) }, @@ -268,7 +270,7 @@ static const struct counter_desc pport_2819_stats_desc[] = { }; static const struct counter_desc pport_phy_statistical_stats_desc[] = { - { "rx_symbol_errors_phy", PPORT_PHY_STATISTICAL_OFF(phy_symbol_errors) }, + { "rx_pcs_symbol_err_phy", PPORT_PHY_STATISTICAL_OFF(phy_symbol_errors) }, { "rx_corrected_bits_phy", PPORT_PHY_STATISTICAL_OFF(phy_corrected_bits) }, }; @@ -319,6 +321,7 @@ struct mlx5e_rq_stats { u64 buff_alloc_err; u64 cqe_compress_blks; u64 cqe_compress_pkts; + u64 page_reuse; u64 cache_reuse; u64 cache_full; u64 cache_empty; @@ -341,6 +344,7 @@ static const struct counter_desc rq_stats_desc[] = { { MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, buff_alloc_err) }, { MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cqe_compress_blks) }, { MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cqe_compress_pkts) }, + { MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, page_reuse) }, { MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cache_reuse) }, { MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cache_full) }, { MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, cache_empty) }, @@ -417,20 +421,13 @@ struct mlx5e_stats { }; static const struct counter_desc mlx5e_pme_status_desc[] = { - { "module_plug", 0 }, { "module_unplug", 8 }, }; static const struct counter_desc mlx5e_pme_error_desc[] = { - { "module_pwr_budget_exd", 0 }, /* power budget exceed */ - { "module_long_range", 8 }, /* long range for non MLNX cable */ - { "module_bus_stuck", 16 }, /* bus stuck (I2C or data shorted) */ - { "module_no_eeprom", 24 }, /* no eeprom/retry time out */ - { "module_enforce_part", 32 }, /* enforce part number list */ - { "module_unknown_id", 40 }, /* unknown identifier */ - { "module_high_temp", 48 }, /* high temperature */ + { "module_bus_stuck", 16 }, /* bus stuck (I2C or data shorted) */ + { "module_high_temp", 48 }, /* high temperature */ { "module_bad_shorted", 56 }, /* bad or shorted cable/module */ - { "module_unknown_status", 64 }, }; #endif /* __MLX5_EN_STATS_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c index 11c27e4fadf6..3c536f560dd2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -43,6 +43,7 @@ #include <net/tc_act/tc_vlan.h> #include <net/tc_act/tc_tunnel_key.h> #include <net/tc_act/tc_pedit.h> +#include <net/tc_act/tc_csum.h> #include <net/vxlan.h> #include <net/arp.h> #include "en.h" @@ -68,7 +69,8 @@ struct mlx5e_tc_flow { u64 cookie; u8 flags; struct mlx5_flow_handle *rule; - struct list_head encap; /* flows sharing the same encap */ + struct list_head encap; /* flows sharing the same encap ID */ + struct list_head mod_hdr; /* flows sharing the same mod hdr ID */ union { struct mlx5_esw_flow_attr esw_attr[0]; struct mlx5_nic_flow_attr nic_attr[0]; @@ -89,6 +91,135 @@ enum { #define MLX5E_TC_TABLE_NUM_ENTRIES 1024 #define MLX5E_TC_TABLE_NUM_GROUPS 4 +struct mod_hdr_key { + int num_actions; + void *actions; +}; + +struct mlx5e_mod_hdr_entry { + /* a node of a hash table which keeps all the mod_hdr entries */ + struct hlist_node mod_hdr_hlist; + + /* flows sharing the same mod_hdr entry */ + struct list_head flows; + + struct mod_hdr_key key; + + u32 mod_hdr_id; +}; + +#define MLX5_MH_ACT_SZ MLX5_UN_SZ_BYTES(set_action_in_add_action_in_auto) + +static inline u32 hash_mod_hdr_info(struct mod_hdr_key *key) +{ + return jhash(key->actions, + key->num_actions * MLX5_MH_ACT_SZ, 0); +} + +static inline int cmp_mod_hdr_info(struct mod_hdr_key *a, + struct mod_hdr_key *b) +{ + if (a->num_actions != b->num_actions) + return 1; + + return memcmp(a->actions, b->actions, a->num_actions * MLX5_MH_ACT_SZ); +} + +static int mlx5e_attach_mod_hdr(struct mlx5e_priv *priv, + struct mlx5e_tc_flow *flow, + struct mlx5e_tc_flow_parse_attr *parse_attr) +{ + struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; + int num_actions, actions_size, namespace, err; + struct mlx5e_mod_hdr_entry *mh; + struct mod_hdr_key key; + bool found = false; + u32 hash_key; + + num_actions = parse_attr->num_mod_hdr_actions; + actions_size = MLX5_MH_ACT_SZ * num_actions; + + key.actions = parse_attr->mod_hdr_actions; + key.num_actions = num_actions; + + hash_key = hash_mod_hdr_info(&key); + + if (flow->flags & MLX5E_TC_FLOW_ESWITCH) { + namespace = MLX5_FLOW_NAMESPACE_FDB; + hash_for_each_possible(esw->offloads.mod_hdr_tbl, mh, + mod_hdr_hlist, hash_key) { + if (!cmp_mod_hdr_info(&mh->key, &key)) { + found = true; + break; + } + } + } else { + namespace = MLX5_FLOW_NAMESPACE_KERNEL; + hash_for_each_possible(priv->fs.tc.mod_hdr_tbl, mh, + mod_hdr_hlist, hash_key) { + if (!cmp_mod_hdr_info(&mh->key, &key)) { + found = true; + break; + } + } + } + + if (found) + goto attach_flow; + + mh = kzalloc(sizeof(*mh) + actions_size, GFP_KERNEL); + if (!mh) + return -ENOMEM; + + mh->key.actions = (void *)mh + sizeof(*mh); + memcpy(mh->key.actions, key.actions, actions_size); + mh->key.num_actions = num_actions; + INIT_LIST_HEAD(&mh->flows); + + err = mlx5_modify_header_alloc(priv->mdev, namespace, + mh->key.num_actions, + mh->key.actions, + &mh->mod_hdr_id); + if (err) + goto out_err; + + if (flow->flags & MLX5E_TC_FLOW_ESWITCH) + hash_add(esw->offloads.mod_hdr_tbl, &mh->mod_hdr_hlist, hash_key); + else + hash_add(priv->fs.tc.mod_hdr_tbl, &mh->mod_hdr_hlist, hash_key); + +attach_flow: + list_add(&flow->mod_hdr, &mh->flows); + if (flow->flags & MLX5E_TC_FLOW_ESWITCH) + flow->esw_attr->mod_hdr_id = mh->mod_hdr_id; + else + flow->nic_attr->mod_hdr_id = mh->mod_hdr_id; + + return 0; + +out_err: + kfree(mh); + return err; +} + +static void mlx5e_detach_mod_hdr(struct mlx5e_priv *priv, + struct mlx5e_tc_flow *flow) +{ + struct list_head *next = flow->mod_hdr.next; + + list_del(&flow->mod_hdr); + + if (list_empty(next)) { + struct mlx5e_mod_hdr_entry *mh; + + mh = list_entry(next, struct mlx5e_mod_hdr_entry, flows); + + mlx5_modify_header_dealloc(priv->mdev, mh->mod_hdr_id); + hash_del(&mh->mod_hdr_hlist); + kfree(mh); + } +} + static struct mlx5_flow_handle * mlx5e_tc_add_nic_flow(struct mlx5e_priv *priv, struct mlx5e_tc_flow_parse_attr *parse_attr, @@ -120,10 +251,7 @@ mlx5e_tc_add_nic_flow(struct mlx5e_priv *priv, } if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) { - err = mlx5_modify_header_alloc(dev, MLX5_FLOW_NAMESPACE_KERNEL, - parse_attr->num_mod_hdr_actions, - parse_attr->mod_hdr_actions, - &attr->mod_hdr_id); + err = mlx5e_attach_mod_hdr(priv, flow, parse_attr); flow_act.modify_id = attr->mod_hdr_id; kfree(parse_attr->mod_hdr_actions); if (err) { @@ -165,8 +293,7 @@ err_add_rule: } err_create_ft: if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) - mlx5_modify_header_dealloc(priv->mdev, - attr->mod_hdr_id); + mlx5e_detach_mod_hdr(priv, flow); err_create_mod_hdr_id: mlx5_fc_destroy(dev, counter); @@ -176,6 +303,7 @@ err_create_mod_hdr_id: static void mlx5e_tc_del_nic_flow(struct mlx5e_priv *priv, struct mlx5e_tc_flow *flow) { + struct mlx5_nic_flow_attr *attr = flow->nic_attr; struct mlx5_fc *counter = NULL; counter = mlx5_flow_rule_counter(flow->rule); @@ -187,9 +315,8 @@ static void mlx5e_tc_del_nic_flow(struct mlx5e_priv *priv, priv->fs.tc.t = NULL; } - if (flow->nic_attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) - mlx5_modify_header_dealloc(priv->mdev, - flow->nic_attr->mod_hdr_id); + if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) + mlx5e_detach_mod_hdr(priv, flow); } static void mlx5e_detach_encap(struct mlx5e_priv *priv, @@ -212,10 +339,7 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, } if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) { - err = mlx5_modify_header_alloc(priv->mdev, MLX5_FLOW_NAMESPACE_FDB, - parse_attr->num_mod_hdr_actions, - parse_attr->mod_hdr_actions, - &attr->mod_hdr_id); + err = mlx5e_attach_mod_hdr(priv, flow, parse_attr); kfree(parse_attr->mod_hdr_actions); if (err) { rule = ERR_PTR(err); @@ -230,9 +354,8 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv, return rule; err_add_rule: - if (flow->esw_attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) - mlx5_modify_header_dealloc(priv->mdev, - attr->mod_hdr_id); + if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) + mlx5e_detach_mod_hdr(priv, flow); err_mod_hdr: mlx5_eswitch_del_vlan_action(esw, attr); err_add_vlan: @@ -249,19 +372,18 @@ static void mlx5e_tc_del_fdb_flow(struct mlx5e_priv *priv, if (flow->flags & MLX5E_TC_FLOW_OFFLOADED) { flow->flags &= ~MLX5E_TC_FLOW_OFFLOADED; - mlx5_eswitch_del_offloaded_rule(esw, flow->rule, flow->esw_attr); + mlx5_eswitch_del_offloaded_rule(esw, flow->rule, attr); } - mlx5_eswitch_del_vlan_action(esw, flow->esw_attr); + mlx5_eswitch_del_vlan_action(esw, attr); - if (flow->esw_attr->action & MLX5_FLOW_CONTEXT_ACTION_ENCAP) { + if (attr->action & MLX5_FLOW_CONTEXT_ACTION_ENCAP) { mlx5e_detach_encap(priv, flow); - kvfree(flow->esw_attr->parse_attr); + kvfree(attr->parse_attr); } - if (flow->esw_attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) - mlx5_modify_header_dealloc(priv->mdev, - attr->mod_hdr_id); + if (attr->action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR) + mlx5e_detach_mod_hdr(priv, flow); } void mlx5e_tc_encap_flows_add(struct mlx5e_priv *priv, @@ -384,7 +506,7 @@ static void mlx5e_detach_encap(struct mlx5e_priv *priv, if (e->flags & MLX5_ENCAP_ENTRY_VALID) mlx5_encap_dealloc(priv->mdev, e->encap_id); - hlist_del_rcu(&e->encap_hlist); + hash_del_rcu(&e->encap_hlist); kfree(e->encap_header); kfree(e); } @@ -580,7 +702,9 @@ static int __parse_cls_flower(struct mlx5e_priv *priv, BIT(FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS) | BIT(FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS) | BIT(FLOW_DISSECTOR_KEY_ENC_PORTS) | - BIT(FLOW_DISSECTOR_KEY_ENC_CONTROL))) { + BIT(FLOW_DISSECTOR_KEY_ENC_CONTROL) | + BIT(FLOW_DISSECTOR_KEY_TCP) | + BIT(FLOW_DISSECTOR_KEY_IP))) { netdev_warn(priv->netdev, "Unsupported key used: 0x%x\n", f->dissector->used_keys); return -EOPNOTSUPP; @@ -764,6 +888,34 @@ static int __parse_cls_flower(struct mlx5e_priv *priv, *min_inline = MLX5_INLINE_MODE_IP; } + if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_IP)) { + struct flow_dissector_key_ip *key = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_IP, + f->key); + struct flow_dissector_key_ip *mask = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_IP, + f->mask); + + MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_ecn, mask->tos & 0x3); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_ecn, key->tos & 0x3); + + MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_dscp, mask->tos >> 2); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_dscp, key->tos >> 2); + + MLX5_SET(fte_match_set_lyr_2_4, headers_c, ttl_hoplimit, mask->ttl); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, ttl_hoplimit, key->ttl); + + if (mask->ttl && + !MLX5_CAP_ESW_FLOWTABLE_FDB(priv->mdev, + ft_field_support.outer_ipv4_ttl)) + return -EOPNOTSUPP; + + if (mask->tos || mask->ttl) + *min_inline = MLX5_INLINE_MODE_IP; + } + if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_PORTS)) { struct flow_dissector_key_ports *key = skb_flow_dissector_target(f->dissector, @@ -807,6 +959,25 @@ static int __parse_cls_flower(struct mlx5e_priv *priv, *min_inline = MLX5_INLINE_MODE_TCP_UDP; } + if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_TCP)) { + struct flow_dissector_key_tcp *key = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_TCP, + f->key); + struct flow_dissector_key_tcp *mask = + skb_flow_dissector_target(f->dissector, + FLOW_DISSECTOR_KEY_TCP, + f->mask); + + MLX5_SET(fte_match_set_lyr_2_4, headers_c, tcp_flags, + ntohs(mask->flags)); + MLX5_SET(fte_match_set_lyr_2_4, headers_v, tcp_flags, + ntohs(key->flags)); + + if (mask->flags) + *min_inline = MLX5_INLINE_MODE_TCP_UDP; + } + return 0; } @@ -887,33 +1058,37 @@ struct mlx5_fields { u32 offset; }; +#define OFFLOAD(fw_field, size, field, off) \ + {MLX5_ACTION_IN_FIELD_OUT_ ## fw_field, size, offsetof(struct pedit_headers, field) + (off)} + static struct mlx5_fields fields[] = { - {MLX5_ACTION_IN_FIELD_OUT_DMAC_47_16, 4, offsetof(struct pedit_headers, eth.h_dest[0])}, - {MLX5_ACTION_IN_FIELD_OUT_DMAC_15_0, 2, offsetof(struct pedit_headers, eth.h_dest[4])}, - {MLX5_ACTION_IN_FIELD_OUT_SMAC_47_16, 4, offsetof(struct pedit_headers, eth.h_source[0])}, - {MLX5_ACTION_IN_FIELD_OUT_SMAC_15_0, 2, offsetof(struct pedit_headers, eth.h_source[4])}, - {MLX5_ACTION_IN_FIELD_OUT_ETHERTYPE, 2, offsetof(struct pedit_headers, eth.h_proto)}, - - {MLX5_ACTION_IN_FIELD_OUT_IP_DSCP, 1, offsetof(struct pedit_headers, ip4.tos)}, - {MLX5_ACTION_IN_FIELD_OUT_IP_TTL, 1, offsetof(struct pedit_headers, ip4.ttl)}, - {MLX5_ACTION_IN_FIELD_OUT_SIPV4, 4, offsetof(struct pedit_headers, ip4.saddr)}, - {MLX5_ACTION_IN_FIELD_OUT_DIPV4, 4, offsetof(struct pedit_headers, ip4.daddr)}, - - {MLX5_ACTION_IN_FIELD_OUT_SIPV6_127_96, 4, offsetof(struct pedit_headers, ip6.saddr.s6_addr32[0])}, - {MLX5_ACTION_IN_FIELD_OUT_SIPV6_95_64, 4, offsetof(struct pedit_headers, ip6.saddr.s6_addr32[1])}, - {MLX5_ACTION_IN_FIELD_OUT_SIPV6_63_32, 4, offsetof(struct pedit_headers, ip6.saddr.s6_addr32[2])}, - {MLX5_ACTION_IN_FIELD_OUT_SIPV6_31_0, 4, offsetof(struct pedit_headers, ip6.saddr.s6_addr32[3])}, - {MLX5_ACTION_IN_FIELD_OUT_DIPV6_127_96, 4, offsetof(struct pedit_headers, ip6.daddr.s6_addr32[0])}, - {MLX5_ACTION_IN_FIELD_OUT_DIPV6_95_64, 4, offsetof(struct pedit_headers, ip6.daddr.s6_addr32[1])}, - {MLX5_ACTION_IN_FIELD_OUT_DIPV6_63_32, 4, offsetof(struct pedit_headers, ip6.daddr.s6_addr32[2])}, - {MLX5_ACTION_IN_FIELD_OUT_DIPV6_31_0, 4, offsetof(struct pedit_headers, ip6.daddr.s6_addr32[3])}, - - {MLX5_ACTION_IN_FIELD_OUT_TCP_SPORT, 2, offsetof(struct pedit_headers, tcp.source)}, - {MLX5_ACTION_IN_FIELD_OUT_TCP_DPORT, 2, offsetof(struct pedit_headers, tcp.dest)}, - {MLX5_ACTION_IN_FIELD_OUT_TCP_FLAGS, 1, offsetof(struct pedit_headers, tcp.ack_seq) + 5}, - - {MLX5_ACTION_IN_FIELD_OUT_UDP_SPORT, 2, offsetof(struct pedit_headers, udp.source)}, - {MLX5_ACTION_IN_FIELD_OUT_UDP_DPORT, 2, offsetof(struct pedit_headers, udp.dest)}, + OFFLOAD(DMAC_47_16, 4, eth.h_dest[0], 0), + OFFLOAD(DMAC_47_16, 4, eth.h_dest[0], 0), + OFFLOAD(DMAC_15_0, 2, eth.h_dest[4], 0), + OFFLOAD(SMAC_47_16, 4, eth.h_source[0], 0), + OFFLOAD(SMAC_15_0, 2, eth.h_source[4], 0), + OFFLOAD(ETHERTYPE, 2, eth.h_proto, 0), + + OFFLOAD(IP_TTL, 1, ip4.ttl, 0), + OFFLOAD(SIPV4, 4, ip4.saddr, 0), + OFFLOAD(DIPV4, 4, ip4.daddr, 0), + + OFFLOAD(SIPV6_127_96, 4, ip6.saddr.s6_addr32[0], 0), + OFFLOAD(SIPV6_95_64, 4, ip6.saddr.s6_addr32[1], 0), + OFFLOAD(SIPV6_63_32, 4, ip6.saddr.s6_addr32[2], 0), + OFFLOAD(SIPV6_31_0, 4, ip6.saddr.s6_addr32[3], 0), + OFFLOAD(DIPV6_127_96, 4, ip6.daddr.s6_addr32[0], 0), + OFFLOAD(DIPV6_95_64, 4, ip6.daddr.s6_addr32[1], 0), + OFFLOAD(DIPV6_63_32, 4, ip6.daddr.s6_addr32[2], 0), + OFFLOAD(DIPV6_31_0, 4, ip6.daddr.s6_addr32[3], 0), + OFFLOAD(IPV6_HOPLIMIT, 1, ip6.hop_limit, 0), + + OFFLOAD(TCP_SPORT, 2, tcp.source, 0), + OFFLOAD(TCP_DPORT, 2, tcp.dest, 0), + OFFLOAD(TCP_FLAGS, 1, tcp.ack_seq, 5), + + OFFLOAD(UDP_SPORT, 2, udp.source, 0), + OFFLOAD(UDP_DPORT, 2, udp.dest, 0), }; /* On input attr->num_mod_hdr_actions tells how many HW actions can be parsed at @@ -925,12 +1100,14 @@ static int offload_pedit_fields(struct pedit_headers *masks, struct mlx5e_tc_flow_parse_attr *parse_attr) { struct pedit_headers *set_masks, *add_masks, *set_vals, *add_vals; - int i, action_size, nactions, max_actions, first, last; + int i, action_size, nactions, max_actions, first, last, next_z; void *s_masks_p, *a_masks_p, *vals_p; - u32 s_mask, a_mask, val; struct mlx5_fields *f; u8 cmd, field_bsize; + u32 s_mask, a_mask; unsigned long mask; + __be32 mask_be32; + __be16 mask_be16; void *action; set_masks = &masks[TCA_PEDIT_KEY_EX_CMD_SET]; @@ -946,7 +1123,8 @@ static int offload_pedit_fields(struct pedit_headers *masks, for (i = 0; i < ARRAY_SIZE(fields); i++) { f = &fields[i]; /* avoid seeing bits set from previous iterations */ - s_mask = a_mask = mask = val = 0; + s_mask = 0; + a_mask = 0; s_masks_p = (void *)set_masks + f->offset; a_masks_p = (void *)add_masks + f->offset; @@ -981,13 +1159,21 @@ static int offload_pedit_fields(struct pedit_headers *masks, memset(a_masks_p, 0, f->size); } - memcpy(&val, vals_p, f->size); - field_bsize = f->size * BITS_PER_BYTE; + + if (field_bsize == 32) { + mask_be32 = *(__be32 *)&mask; + mask = (__force unsigned long)cpu_to_le32(be32_to_cpu(mask_be32)); + } else if (field_bsize == 16) { + mask_be16 = *(__be16 *)&mask; + mask = (__force unsigned long)cpu_to_le16(be16_to_cpu(mask_be16)); + } + first = find_first_bit(&mask, field_bsize); + next_z = find_next_zero_bit(&mask, field_bsize, first); last = find_last_bit(&mask, field_bsize); - if (first > 0 || last != (field_bsize - 1)) { - printk(KERN_WARNING "mlx5: partial rewrite (mask %lx) is currently not offloaded\n", + if (first < next_z && next_z < last) { + printk(KERN_WARNING "mlx5: rewrite of few sub-fields (mask %lx) isn't offloaded\n", mask); return -EOPNOTSUPP; } @@ -996,17 +1182,17 @@ static int offload_pedit_fields(struct pedit_headers *masks, MLX5_SET(set_action_in, action, field, f->field); if (cmd == MLX5_ACTION_TYPE_SET) { - MLX5_SET(set_action_in, action, offset, 0); + MLX5_SET(set_action_in, action, offset, first); /* length is num of bits to be written, zero means length of 32 */ - MLX5_SET(set_action_in, action, length, field_bsize); + MLX5_SET(set_action_in, action, length, (last - first + 1)); } if (field_bsize == 32) - MLX5_SET(set_action_in, action, data, ntohl(val)); + MLX5_SET(set_action_in, action, data, ntohl(*(__be32 *)vals_p) >> first); else if (field_bsize == 16) - MLX5_SET(set_action_in, action, data, ntohs(val)); + MLX5_SET(set_action_in, action, data, ntohs(*(__be16 *)vals_p) >> first); else if (field_bsize == 8) - MLX5_SET(set_action_in, action, data, val); + MLX5_SET(set_action_in, action, data, *(u8 *)vals_p >> first); action += action_size; nactions++; @@ -1109,6 +1295,28 @@ out_err: return err; } +static bool csum_offload_supported(struct mlx5e_priv *priv, u32 action, u32 update_flags) +{ + u32 prot_flags = TCA_CSUM_UPDATE_FLAG_IPV4HDR | TCA_CSUM_UPDATE_FLAG_TCP | + TCA_CSUM_UPDATE_FLAG_UDP; + + /* The HW recalcs checksums only if re-writing headers */ + if (!(action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)) { + netdev_warn(priv->netdev, + "TC csum action is only offloaded with pedit\n"); + return false; + } + + if (update_flags & ~prot_flags) { + netdev_warn(priv->netdev, + "can't offload TC csum action for some header/s - flags %#x\n", + update_flags); + return false; + } + + return true; +} + static int parse_tc_nic_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, struct mlx5e_tc_flow_parse_attr *parse_attr, struct mlx5e_tc_flow *flow) @@ -1126,10 +1334,6 @@ static int parse_tc_nic_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, tcf_exts_to_list(exts, &actions); list_for_each_entry(a, &actions, list) { - /* Only support a single action per rule */ - if (attr->action) - return -EINVAL; - if (is_tcf_gact_shot(a)) { attr->action |= MLX5_FLOW_CONTEXT_ACTION_DROP; if (MLX5_CAP_FLOWTABLE(priv->mdev, @@ -1149,6 +1353,14 @@ static int parse_tc_nic_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, continue; } + if (is_tcf_csum(a)) { + if (csum_offload_supported(priv, attr->action, + tcf_csum_update_flags(a))) + continue; + + return -EOPNOTSUPP; + } + if (is_tcf_skbedit_mark(a)) { u32 mark = tcf_skbedit_mark(a); @@ -1404,8 +1616,8 @@ static int mlx5e_create_encap_header_ipv4(struct mlx5e_priv *priv, if (!(nud_state & NUD_VALID)) { neigh_event_send(n, NULL); - neigh_release(n); - return -EAGAIN; + err = -EAGAIN; + goto out; } err = mlx5_encap_alloc(priv->mdev, e->tunnel_type, @@ -1510,8 +1722,8 @@ static int mlx5e_create_encap_header_ipv6(struct mlx5e_priv *priv, if (!(nud_state & NUD_VALID)) { neigh_event_send(n, NULL); - neigh_release(n); - return -EAGAIN; + err = -EAGAIN; + goto out; } err = mlx5_encap_alloc(priv->mdev, e->tunnel_type, @@ -1651,6 +1863,14 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv, struct tcf_exts *exts, continue; } + if (is_tcf_csum(a)) { + if (csum_offload_supported(priv, attr->action, + tcf_csum_update_flags(a))) + continue; + + return -EOPNOTSUPP; + } + if (is_tcf_mirred_egress_redirect(a)) { int ifindex = tcf_mirred_ifindex(a); struct net_device *out_dev, *encap_dev = NULL; @@ -1738,7 +1958,7 @@ int mlx5e_configure_flower(struct mlx5e_priv *priv, __be16 protocol, } flow = kzalloc(sizeof(*flow) + attr_size, GFP_KERNEL); - parse_attr = mlx5_vzalloc(sizeof(*parse_attr)); + parse_attr = kvzalloc(sizeof(*parse_attr), GFP_KERNEL); if (!parse_attr || !flow) { err = -ENOMEM; goto err_free; @@ -1823,9 +2043,7 @@ int mlx5e_stats_flower(struct mlx5e_priv *priv, { struct mlx5e_tc_table *tc = &priv->fs.tc; struct mlx5e_tc_flow *flow; - struct tc_action *a; struct mlx5_fc *counter; - LIST_HEAD(actions); u64 bytes; u64 packets; u64 lastuse; @@ -1844,13 +2062,7 @@ int mlx5e_stats_flower(struct mlx5e_priv *priv, mlx5_fc_query_cached(counter, &bytes, &packets, &lastuse); - preempt_disable(); - - tcf_exts_to_list(f->exts, &actions); - list_for_each_entry(a, &actions, list) - tcf_action_stats_update(a, bytes, packets, lastuse); - - preempt_enable(); + tcf_exts_stats_update(f->exts, bytes, packets, lastuse); return 0; } @@ -1866,6 +2078,8 @@ int mlx5e_tc_init(struct mlx5e_priv *priv) { struct mlx5e_tc_table *tc = &priv->fs.tc; + hash_init(tc->mod_hdr_tbl); + tc->ht_params = mlx5e_tc_flow_ht_params; return rhashtable_init(&tc->ht, &tc->ht_params); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c index ab3bb026ff9e..aaa0f4ebba9a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c @@ -33,7 +33,8 @@ #include <linux/tcp.h> #include <linux/if_vlan.h> #include "en.h" -#include "ipoib.h" +#include "ipoib/ipoib.h" +#include "en_accel/ipsec_rxtx.h" #define MLX5E_SQ_NOPS_ROOM MLX5_SEND_WQE_MAX_WQEBBS #define MLX5E_SQ_STOP_ROOM (MLX5_SEND_WQE_MAX_WQEBBS +\ @@ -245,7 +246,7 @@ mlx5e_txwqe_build_dsegs(struct mlx5e_txqsq *sq, struct sk_buff *skb, int fsz = skb_frag_size(frag); dma_addr = skb_frag_dma_map(sq->pdev, frag, 0, fsz, - DMA_TO_DEVICE); + DMA_TO_DEVICE); if (unlikely(dma_mapping_error(sq->pdev, dma_addr))) return -ENOMEM; @@ -299,12 +300,9 @@ mlx5e_txwqe_complete(struct mlx5e_txqsq *sq, struct sk_buff *skb, } } -static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb) +static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb, + struct mlx5e_tx_wqe *wqe, u16 pi) { - struct mlx5_wq_cyc *wq = &sq->wq; - - u16 pi = sq->pc & wq->sz_m1; - struct mlx5e_tx_wqe *wqe = mlx5_wq_cyc_get_wqe(wq, pi); struct mlx5e_tx_wqe_info *wi = &sq->db.wqe_info[pi]; struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl; @@ -319,8 +317,6 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb) u16 ds_cnt; u16 ihs; - memset(wqe, 0, sizeof(*wqe)); - mlx5e_txwqe_build_eseg_csum(sq, skb, eseg); if (skb_is_gso(skb)) { @@ -375,8 +371,21 @@ netdev_tx_t mlx5e_xmit(struct sk_buff *skb, struct net_device *dev) { struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5e_txqsq *sq = priv->txq2sq[skb_get_queue_mapping(skb)]; + struct mlx5_wq_cyc *wq = &sq->wq; + u16 pi = sq->pc & wq->sz_m1; + struct mlx5e_tx_wqe *wqe = mlx5_wq_cyc_get_wqe(wq, pi); + + memset(wqe, 0, sizeof(*wqe)); - return mlx5e_sq_xmit(sq, skb); +#ifdef CONFIG_MLX5_EN_IPSEC + if (sq->state & BIT(MLX5E_SQ_STATE_IPSEC)) { + skb = mlx5e_ipsec_handle_tx_skb(dev, wqe, skb); + if (unlikely(!skb)) + return NETDEV_TX_OK; + } +#endif + + return mlx5e_sq_xmit(sq, skb, wqe, pi); } bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq, int napi_budget) @@ -409,7 +418,7 @@ bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq, int napi_budget) u16 wqe_counter; bool last_wqe; - cqe = mlx5e_get_cqe(cq); + cqe = mlx5_cqwq_get_cqe(&cq->wq); if (!cqe) break; @@ -557,11 +566,16 @@ netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb, if (skb_is_gso(skb)) { opcode = MLX5_OPCODE_LSO; ihs = mlx5e_txwqe_build_eseg_gso(sq, skb, eseg, &num_bytes); + sq->stats.packets += skb_shinfo(skb)->gso_segs; } else { ihs = mlx5e_calc_min_inline(sq->min_inline_mode, skb); num_bytes = max_t(unsigned int, skb->len, ETH_ZLEN); + sq->stats.packets++; } + sq->stats.bytes += num_bytes; + sq->stats.xmit_more += skb->xmit_more; + ds_cnt = sizeof(*wqe) / MLX5_SEND_WQE_DS; if (ihs) { memcpy(eseg->inline_hdr.start, skb_data, ihs); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c index 5ca6714e3e02..92db28a9ed43 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c @@ -32,23 +32,6 @@ #include "en.h" -struct mlx5_cqe64 *mlx5e_get_cqe(struct mlx5e_cq *cq) -{ - struct mlx5_cqwq *wq = &cq->wq; - u32 ci = mlx5_cqwq_get_ci(wq); - struct mlx5_cqe64 *cqe = mlx5_cqwq_get_wqe(wq, ci); - u8 cqe_ownership_bit = cqe->op_own & MLX5_CQE_OWNER_MASK; - u8 sw_ownership_val = mlx5_cqwq_get_wrap_cnt(wq) & 1; - - if (cqe_ownership_bit != sw_ownership_val) - return NULL; - - /* ensure cqe content is read after cqe ownership bit */ - dma_rmb(); - - return cqe; -} - static inline void mlx5e_poll_ico_single_cqe(struct mlx5e_cq *cq, struct mlx5e_icosq *sq, struct mlx5_cqe64 *cqe, @@ -89,7 +72,7 @@ static void mlx5e_poll_ico_cq(struct mlx5e_cq *cq) if (unlikely(!test_bit(MLX5E_SQ_STATE_ENABLED, &sq->state))) return; - cqe = mlx5e_get_cqe(cq); + cqe = mlx5_cqwq_get_cqe(&cq->wq); if (likely(!cqe)) return; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c index ea5d8d37a75c..af51a5d2b912 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c @@ -35,6 +35,7 @@ #include <linux/mlx5/driver.h> #include <linux/mlx5/cmd.h> #include "mlx5_core.h" +#include "fpga/core.h" #ifdef CONFIG_MLX5_CORE_EN #include "eswitch.h" #endif @@ -156,6 +157,10 @@ static const char *eqe_type_str(u8 type) return "MLX5_EVENT_TYPE_PAGE_FAULT"; case MLX5_EVENT_TYPE_PPS_EVENT: return "MLX5_EVENT_TYPE_PPS_EVENT"; + case MLX5_EVENT_TYPE_NIC_VPORT_CHANGE: + return "MLX5_EVENT_TYPE_NIC_VPORT_CHANGE"; + case MLX5_EVENT_TYPE_FPGA_ERROR: + return "MLX5_EVENT_TYPE_FPGA_ERROR"; default: return "Unrecognized event"; } @@ -186,7 +191,7 @@ static void eq_update_ci(struct mlx5_eq *eq, int arm) { __be32 __iomem *addr = eq->doorbell + (arm ? 0 : 2); u32 val = (eq->cons_index & 0xffffff) | (eq->eqn << 24); - __raw_writel((__force u32) cpu_to_be32(val), addr); + __raw_writel((__force u32)cpu_to_be32(val), addr); /* We still want ordering, just not swabbing, so add a barrier */ mb(); } @@ -422,7 +427,7 @@ static irqreturn_t mlx5_eq_int(int irq, void *eq_ptr) break; case MLX5_EVENT_TYPE_CMD: - mlx5_cmd_comp_handler(dev, be32_to_cpu(eqe->data.cmd.vector)); + mlx5_cmd_comp_handler(dev, be32_to_cpu(eqe->data.cmd.vector), false); break; case MLX5_EVENT_TYPE_PORT_CHANGE: @@ -476,6 +481,11 @@ static irqreturn_t mlx5_eq_int(int irq, void *eq_ptr) if (dev->event) dev->event(dev, MLX5_DEV_EVENT_PPS, (unsigned long)eqe); break; + + case MLX5_EVENT_TYPE_FPGA_ERROR: + mlx5_fpga_event(dev, eqe->type, &eqe->data.raw); + break; + default: mlx5_core_warn(dev, "Unhandled event 0x%x on EQ 0x%x\n", eqe->type, eq->eqn); @@ -548,7 +558,7 @@ int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx, inlen = MLX5_ST_SZ_BYTES(create_eq_in) + MLX5_FLD_SZ_BYTES(create_eq_in, pas[0]) * eq->buf.npages; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) { err = -ENOMEM; goto err_buf; @@ -667,7 +677,6 @@ int mlx5_eq_init(struct mlx5_core_dev *dev) return err; } - void mlx5_eq_cleanup(struct mlx5_core_dev *dev) { mlx5_eq_debugfs_cleanup(dev); @@ -679,7 +688,6 @@ int mlx5_start_eqs(struct mlx5_core_dev *dev) u64 async_event_mask = MLX5_ASYNC_EVENT_MASK; int err; - if (MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_ETH && MLX5_CAP_GEN(dev, vport_group_manager) && mlx5_core_is_pf(dev)) @@ -693,6 +701,9 @@ int mlx5_start_eqs(struct mlx5_core_dev *dev) if (MLX5_CAP_GEN(dev, pps)) async_event_mask |= (1ull << MLX5_EVENT_TYPE_PPS_EVENT); + if (MLX5_CAP_GEN(dev, fpga)) + async_event_mask |= (1ull << MLX5_EVENT_TYPE_FPGA_ERROR); + err = mlx5_create_map_eq(dev, &table->cmd_eq, MLX5_EQ_VEC_CMD, MLX5_NUM_CMD_EQE, 1ull << MLX5_EVENT_TYPE_CMD, "mlx5_cmd_eq", MLX5_EQ_TYPE_ASYNC); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index 2e34d95ea776..89bfda419efe 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -248,11 +248,10 @@ __esw_fdb_set_vport_rule(struct mlx5_eswitch *esw, u32 vport, bool rx_rule, if (rx_rule) match_header |= MLX5_MATCH_MISC_PARAMETERS; - spec = mlx5_vzalloc(sizeof(*spec)); - if (!spec) { - esw_warn(esw->dev, "FDB: Failed to alloc match parameters\n"); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); + if (!spec) return NULL; - } + dmac_v = MLX5_ADDR_OF(fte_match_param, spec->match_value, outer_headers.dmac_47_16); dmac_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, @@ -350,10 +349,9 @@ static int esw_create_legacy_fdb_table(struct mlx5_eswitch *esw, int nvports) return -EOPNOTSUPP; } - flow_group_in = mlx5_vzalloc(inlen); + flow_group_in = kvzalloc(inlen, GFP_KERNEL); if (!flow_group_in) return -ENOMEM; - memset(flow_group_in, 0, inlen); table_size = BIT(MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size)); @@ -961,7 +959,7 @@ static int esw_vport_enable_egress_acl(struct mlx5_eswitch *esw, return -EOPNOTSUPP; } - flow_group_in = mlx5_vzalloc(inlen); + flow_group_in = kvzalloc(inlen, GFP_KERNEL); if (!flow_group_in) return -ENOMEM; @@ -1078,7 +1076,7 @@ static int esw_vport_enable_ingress_acl(struct mlx5_eswitch *esw, return -EOPNOTSUPP; } - flow_group_in = mlx5_vzalloc(inlen); + flow_group_in = kvzalloc(inlen, GFP_KERNEL); if (!flow_group_in) return -ENOMEM; @@ -1219,7 +1217,6 @@ static int esw_vport_ingress_config(struct mlx5_eswitch *esw, "vport[%d] configure ingress rules failed, illegal mac with spoofchk\n", vport->vport); return -EPERM; - } esw_vport_cleanup_ingress_rules(esw, vport); @@ -1241,11 +1238,9 @@ static int esw_vport_ingress_config(struct mlx5_eswitch *esw, "vport[%d] configure ingress rules, vlan(%d) qos(%d)\n", vport->vport, vport->info.vlan, vport->info.qos); - spec = mlx5_vzalloc(sizeof(*spec)); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); if (!spec) { err = -ENOMEM; - esw_warn(esw->dev, "vport[%d] configure ingress rules failed, err(%d)\n", - vport->vport, err); goto out; } @@ -1322,11 +1317,9 @@ static int esw_vport_egress_config(struct mlx5_eswitch *esw, "vport[%d] configure egress rules, vlan(%d) qos(%d)\n", vport->vport, vport->info.vlan, vport->info.qos); - spec = mlx5_vzalloc(sizeof(*spec)); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); if (!spec) { err = -ENOMEM; - esw_warn(esw->dev, "vport[%d] configure egress rules failed, err(%d)\n", - vport->vport, err); goto out; } @@ -1775,6 +1768,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev) } hash_init(esw->offloads.encap_tbl); + hash_init(esw->offloads.mod_hdr_tbl); mutex_init(&esw->state_lock); for (vport_num = 0; vport_num < total_vports; vport_num++) { @@ -2158,7 +2152,7 @@ int mlx5_eswitch_get_vport_stats(struct mlx5_eswitch *esw, if (!LEGAL_VPORT(esw, vport)) return -EINVAL; - out = mlx5_vzalloc(outlen); + out = kvzalloc(outlen, GFP_KERNEL); if (!out) return -ENOMEM; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index b746f62c8c79..834a33050969 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -207,6 +207,7 @@ struct mlx5_esw_offload { struct mlx5_flow_group *vport_rx_group; struct mlx5_eswitch_rep *vport_reps; DECLARE_HASHTABLE(encap_tbl, 8); + DECLARE_HASHTABLE(mod_hdr_tbl, 8); u8 inline_mode; u64 num_flows; u8 encap; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index f991f669047e..95b64025ce36 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -311,9 +311,8 @@ mlx5_eswitch_add_send_to_vport_rule(struct mlx5_eswitch *esw, int vport, u32 sqn struct mlx5_flow_spec *spec; void *misc; - spec = mlx5_vzalloc(sizeof(*spec)); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); if (!spec) { - esw_warn(esw->dev, "FDB: Failed to alloc match parameters\n"); flow_rule = ERR_PTR(-ENOMEM); goto out; } @@ -401,9 +400,8 @@ static int esw_add_fdb_miss_rule(struct mlx5_eswitch *esw) struct mlx5_flow_spec *spec; int err = 0; - spec = mlx5_vzalloc(sizeof(*spec)); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); if (!spec) { - esw_warn(esw->dev, "FDB: Failed to alloc match parameters\n"); err = -ENOMEM; goto out; } @@ -488,7 +486,7 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw, int nvports) u32 *flow_group_in; esw_debug(esw->dev, "Create offloads FDB Tables\n"); - flow_group_in = mlx5_vzalloc(inlen); + flow_group_in = kvzalloc(inlen, GFP_KERNEL); if (!flow_group_in) return -ENOMEM; @@ -631,7 +629,7 @@ static int esw_create_vport_rx_group(struct mlx5_eswitch *esw) int err = 0; int nvports = priv->sriov.num_vfs + 2; - flow_group_in = mlx5_vzalloc(inlen); + flow_group_in = kvzalloc(inlen, GFP_KERNEL); if (!flow_group_in) return -ENOMEM; @@ -675,9 +673,8 @@ mlx5_eswitch_create_vport_rx_rule(struct mlx5_eswitch *esw, int vport, u32 tirn) struct mlx5_flow_spec *spec; void *misc; - spec = mlx5_vzalloc(sizeof(*spec)); + spec = kvzalloc(sizeof(*spec), GFP_KERNEL); if (!spec) { - esw_warn(esw->dev, "Failed to alloc match parameters\n"); flow_rule = ERR_PTR(-ENOMEM); goto out; } @@ -694,7 +691,7 @@ mlx5_eswitch_create_vport_rx_rule(struct mlx5_eswitch *esw, int vport, u32 tirn) flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST; flow_rule = mlx5_add_flow_rules(esw->offloads.ft_offloads, spec, - &flow_act, &dest, 1); + &flow_act, &dest, 1); if (IS_ERR(flow_rule)) { esw_warn(esw->dev, "fs offloads: Failed to add vport rx rule err %ld\n", PTR_ERR(flow_rule)); goto out; @@ -906,21 +903,34 @@ static int esw_inline_mode_to_devlink(u8 mlx5_mode, u8 *mode) return 0; } -int mlx5_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode) +static int mlx5_devlink_eswitch_check(struct devlink *devlink) { - struct mlx5_core_dev *dev; - u16 cur_mlx5_mode, mlx5_mode = 0; + struct mlx5_core_dev *dev = devlink_priv(devlink); - dev = devlink_priv(devlink); + if (MLX5_CAP_GEN(dev, port_type) != MLX5_CAP_PORT_TYPE_ETH) + return -EOPNOTSUPP; if (!MLX5_CAP_GEN(dev, vport_group_manager)) return -EOPNOTSUPP; - cur_mlx5_mode = dev->priv.eswitch->mode; - - if (cur_mlx5_mode == SRIOV_NONE) + if (dev->priv.eswitch->mode == SRIOV_NONE) return -EOPNOTSUPP; + return 0; +} + +int mlx5_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode) +{ + struct mlx5_core_dev *dev = devlink_priv(devlink); + u16 cur_mlx5_mode, mlx5_mode = 0; + int err; + + err = mlx5_devlink_eswitch_check(devlink); + if (err) + return err; + + cur_mlx5_mode = dev->priv.eswitch->mode; + if (esw_mode_from_devlink(mode, &mlx5_mode)) return -EINVAL; @@ -937,15 +947,12 @@ int mlx5_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode) int mlx5_devlink_eswitch_mode_get(struct devlink *devlink, u16 *mode) { - struct mlx5_core_dev *dev; - - dev = devlink_priv(devlink); - - if (!MLX5_CAP_GEN(dev, vport_group_manager)) - return -EOPNOTSUPP; + struct mlx5_core_dev *dev = devlink_priv(devlink); + int err; - if (dev->priv.eswitch->mode == SRIOV_NONE) - return -EOPNOTSUPP; + err = mlx5_devlink_eswitch_check(devlink); + if (err) + return err; return esw_mode_to_devlink(dev->priv.eswitch->mode, mode); } @@ -954,15 +961,12 @@ int mlx5_devlink_eswitch_inline_mode_set(struct devlink *devlink, u8 mode) { struct mlx5_core_dev *dev = devlink_priv(devlink); struct mlx5_eswitch *esw = dev->priv.eswitch; - int num_vports = esw->enabled_vports; int err, vport; u8 mlx5_mode; - if (!MLX5_CAP_GEN(dev, vport_group_manager)) - return -EOPNOTSUPP; - - if (esw->mode == SRIOV_NONE) - return -EOPNOTSUPP; + err = mlx5_devlink_eswitch_check(devlink); + if (err) + return err; switch (MLX5_CAP_ETH(dev, wqe_inline_mode)) { case MLX5_CAP_INLINE_MODE_NOT_REQUIRED: @@ -985,7 +989,7 @@ int mlx5_devlink_eswitch_inline_mode_set(struct devlink *devlink, u8 mode) if (err) goto out; - for (vport = 1; vport < num_vports; vport++) { + for (vport = 1; vport < esw->enabled_vports; vport++) { err = mlx5_modify_nic_vport_min_inline(dev, vport, mlx5_mode); if (err) { esw_warn(dev, "Failed to set min inline on vport %d\n", @@ -1010,12 +1014,11 @@ int mlx5_devlink_eswitch_inline_mode_get(struct devlink *devlink, u8 *mode) { struct mlx5_core_dev *dev = devlink_priv(devlink); struct mlx5_eswitch *esw = dev->priv.eswitch; + int err; - if (!MLX5_CAP_GEN(dev, vport_group_manager)) - return -EOPNOTSUPP; - - if (esw->mode == SRIOV_NONE) - return -EOPNOTSUPP; + err = mlx5_devlink_eswitch_check(devlink); + if (err) + return err; return esw_inline_mode_to_devlink(esw->offloads.inline_mode, mode); } @@ -1062,11 +1065,9 @@ int mlx5_devlink_eswitch_encap_mode_set(struct devlink *devlink, u8 encap) struct mlx5_eswitch *esw = dev->priv.eswitch; int err; - if (!MLX5_CAP_GEN(dev, vport_group_manager)) - return -EOPNOTSUPP; - - if (esw->mode == SRIOV_NONE) - return -EOPNOTSUPP; + err = mlx5_devlink_eswitch_check(devlink); + if (err) + return err; if (encap != DEVLINK_ESWITCH_ENCAP_MODE_NONE && (!MLX5_CAP_ESW_FLOWTABLE_FDB(dev, encap) || @@ -1096,7 +1097,7 @@ int mlx5_devlink_eswitch_encap_mode_set(struct devlink *devlink, u8 encap) if (err) { esw_warn(esw->dev, "Failed re-creating fast FDB table, err %d\n", err); esw->offloads.encap = !encap; - (void) esw_create_offloads_fast_fdb_table(esw); + (void)esw_create_offloads_fast_fdb_table(esw); } return err; } @@ -1105,12 +1106,11 @@ int mlx5_devlink_eswitch_encap_mode_get(struct devlink *devlink, u8 *encap) { struct mlx5_core_dev *dev = devlink_priv(devlink); struct mlx5_eswitch *esw = dev->priv.eswitch; + int err; - if (!MLX5_CAP_GEN(dev, vport_group_manager)) - return -EOPNOTSUPP; - - if (esw->mode == SRIOV_NONE) - return -EOPNOTSUPP; + err = mlx5_devlink_eswitch_check(devlink); + if (err) + return err; *encap = esw->offloads.encap; return 0; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/fpga/Makefile new file mode 100644 index 000000000000..d8e17110f25d --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/Makefile @@ -0,0 +1 @@ +subdir-ccflags-y += -I$(src)/.. diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.c new file mode 100644 index 000000000000..e37453d838db --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.c @@ -0,0 +1,238 @@ +/* + * Copyright (c) 2017, Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/etherdevice.h> +#include <linux/mlx5/cmd.h> +#include <linux/mlx5/driver.h> +#include <linux/mlx5/device.h> + +#include "mlx5_core.h" +#include "fpga/cmd.h" + +#define MLX5_FPGA_ACCESS_REG_SZ (MLX5_ST_SZ_DW(fpga_access_reg) + \ + MLX5_FPGA_ACCESS_REG_SIZE_MAX) + +int mlx5_fpga_access_reg(struct mlx5_core_dev *dev, u8 size, u64 addr, + void *buf, bool write) +{ + u32 in[MLX5_FPGA_ACCESS_REG_SZ] = {0}; + u32 out[MLX5_FPGA_ACCESS_REG_SZ]; + int err; + + if (size & 3) + return -EINVAL; + if (addr & 3) + return -EINVAL; + if (size > MLX5_FPGA_ACCESS_REG_SIZE_MAX) + return -EINVAL; + + MLX5_SET(fpga_access_reg, in, size, size); + MLX5_SET64(fpga_access_reg, in, address, addr); + if (write) + memcpy(MLX5_ADDR_OF(fpga_access_reg, in, data), buf, size); + + err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out), + MLX5_REG_FPGA_ACCESS_REG, 0, write); + if (err) + return err; + + if (!write) + memcpy(buf, MLX5_ADDR_OF(fpga_access_reg, out, data), size); + + return 0; +} + +int mlx5_fpga_caps(struct mlx5_core_dev *dev, u32 *caps) +{ + u32 in[MLX5_ST_SZ_DW(fpga_cap)] = {0}; + + return mlx5_core_access_reg(dev, in, sizeof(in), caps, + MLX5_ST_SZ_BYTES(fpga_cap), + MLX5_REG_FPGA_CAP, 0, 0); +} + +int mlx5_fpga_ctrl_op(struct mlx5_core_dev *dev, u8 op) +{ + u32 in[MLX5_ST_SZ_DW(fpga_ctrl)] = {0}; + u32 out[MLX5_ST_SZ_DW(fpga_ctrl)]; + + MLX5_SET(fpga_ctrl, in, operation, op); + + return mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out), + MLX5_REG_FPGA_CTRL, 0, true); +} + +int mlx5_fpga_sbu_caps(struct mlx5_core_dev *dev, void *caps, int size) +{ + unsigned int cap_size = MLX5_CAP_FPGA(dev, sandbox_extended_caps_len); + u64 addr = MLX5_CAP64_FPGA(dev, sandbox_extended_caps_addr); + unsigned int read; + int ret = 0; + + if (cap_size > size) { + mlx5_core_warn(dev, "Not enough buffer %u for FPGA SBU caps %u", + size, cap_size); + return -EINVAL; + } + + while (cap_size > 0) { + read = min_t(unsigned int, cap_size, + MLX5_FPGA_ACCESS_REG_SIZE_MAX); + + ret = mlx5_fpga_access_reg(dev, read, addr, caps, false); + if (ret) { + mlx5_core_warn(dev, "Error reading FPGA SBU caps %u bytes at address 0x%llx: %d", + read, addr, ret); + return ret; + } + + cap_size -= read; + addr += read; + caps += read; + } + + return ret; +} + +int mlx5_fpga_query(struct mlx5_core_dev *dev, struct mlx5_fpga_query *query) +{ + u32 in[MLX5_ST_SZ_DW(fpga_ctrl)] = {0}; + u32 out[MLX5_ST_SZ_DW(fpga_ctrl)]; + int err; + + err = mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out), + MLX5_REG_FPGA_CTRL, 0, false); + if (err) + return err; + + query->status = MLX5_GET(fpga_ctrl, out, status); + query->admin_image = MLX5_GET(fpga_ctrl, out, flash_select_admin); + query->oper_image = MLX5_GET(fpga_ctrl, out, flash_select_oper); + return 0; +} + +int mlx5_fpga_create_qp(struct mlx5_core_dev *dev, void *fpga_qpc, + u32 *fpga_qpn) +{ + u32 in[MLX5_ST_SZ_DW(fpga_create_qp_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(fpga_create_qp_out)]; + int ret; + + MLX5_SET(fpga_create_qp_in, in, opcode, MLX5_CMD_OP_FPGA_CREATE_QP); + memcpy(MLX5_ADDR_OF(fpga_create_qp_in, in, fpga_qpc), fpga_qpc, + MLX5_FLD_SZ_BYTES(fpga_create_qp_in, fpga_qpc)); + + ret = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); + if (ret) + return ret; + + memcpy(fpga_qpc, MLX5_ADDR_OF(fpga_create_qp_out, out, fpga_qpc), + MLX5_FLD_SZ_BYTES(fpga_create_qp_out, fpga_qpc)); + *fpga_qpn = MLX5_GET(fpga_create_qp_out, out, fpga_qpn); + return ret; +} + +int mlx5_fpga_modify_qp(struct mlx5_core_dev *dev, u32 fpga_qpn, + enum mlx5_fpga_qpc_field_select fields, + void *fpga_qpc) +{ + u32 in[MLX5_ST_SZ_DW(fpga_modify_qp_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(fpga_modify_qp_out)]; + + MLX5_SET(fpga_modify_qp_in, in, opcode, MLX5_CMD_OP_FPGA_MODIFY_QP); + MLX5_SET(fpga_modify_qp_in, in, field_select, fields); + MLX5_SET(fpga_modify_qp_in, in, fpga_qpn, fpga_qpn); + memcpy(MLX5_ADDR_OF(fpga_modify_qp_in, in, fpga_qpc), fpga_qpc, + MLX5_FLD_SZ_BYTES(fpga_modify_qp_in, fpga_qpc)); + + return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); +} + +int mlx5_fpga_query_qp(struct mlx5_core_dev *dev, + u32 fpga_qpn, void *fpga_qpc) +{ + u32 in[MLX5_ST_SZ_DW(fpga_query_qp_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(fpga_query_qp_out)]; + int ret; + + MLX5_SET(fpga_query_qp_in, in, opcode, MLX5_CMD_OP_FPGA_QUERY_QP); + MLX5_SET(fpga_query_qp_in, in, fpga_qpn, fpga_qpn); + + ret = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); + if (ret) + return ret; + + memcpy(fpga_qpc, MLX5_ADDR_OF(fpga_query_qp_out, out, fpga_qpc), + MLX5_FLD_SZ_BYTES(fpga_query_qp_out, fpga_qpc)); + return ret; +} + +int mlx5_fpga_destroy_qp(struct mlx5_core_dev *dev, u32 fpga_qpn) +{ + u32 in[MLX5_ST_SZ_DW(fpga_destroy_qp_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(fpga_destroy_qp_out)]; + + MLX5_SET(fpga_destroy_qp_in, in, opcode, MLX5_CMD_OP_FPGA_DESTROY_QP); + MLX5_SET(fpga_destroy_qp_in, in, fpga_qpn, fpga_qpn); + + return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); +} + +int mlx5_fpga_query_qp_counters(struct mlx5_core_dev *dev, u32 fpga_qpn, + bool clear, struct mlx5_fpga_qp_counters *data) +{ + u32 in[MLX5_ST_SZ_DW(fpga_query_qp_counters_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(fpga_query_qp_counters_out)]; + int ret; + + MLX5_SET(fpga_query_qp_counters_in, in, opcode, + MLX5_CMD_OP_FPGA_QUERY_QP_COUNTERS); + MLX5_SET(fpga_query_qp_counters_in, in, clear, clear); + MLX5_SET(fpga_query_qp_counters_in, in, fpga_qpn, fpga_qpn); + + ret = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); + if (ret) + return ret; + + data->rx_ack_packets = MLX5_GET64(fpga_query_qp_counters_out, out, + rx_ack_packets); + data->rx_send_packets = MLX5_GET64(fpga_query_qp_counters_out, out, + rx_send_packets); + data->tx_ack_packets = MLX5_GET64(fpga_query_qp_counters_out, out, + tx_ack_packets); + data->tx_send_packets = MLX5_GET64(fpga_query_qp_counters_out, out, + tx_send_packets); + data->rx_total_drop = MLX5_GET64(fpga_query_qp_counters_out, out, + rx_total_drop); + + return ret; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.h new file mode 100644 index 000000000000..94bdfd47c3f0 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/cmd.h @@ -0,0 +1,84 @@ +/* + * Copyright (c) 2017, Mellanox Technologies, Ltd. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#ifndef __MLX5_FPGA_H__ +#define __MLX5_FPGA_H__ + +#include <linux/mlx5/driver.h> + +enum mlx5_fpga_image { + MLX5_FPGA_IMAGE_USER = 0, + MLX5_FPGA_IMAGE_FACTORY, +}; + +enum mlx5_fpga_status { + MLX5_FPGA_STATUS_SUCCESS = 0, + MLX5_FPGA_STATUS_FAILURE = 1, + MLX5_FPGA_STATUS_IN_PROGRESS = 2, + MLX5_FPGA_STATUS_NONE = 0xFFFF, +}; + +struct mlx5_fpga_query { + enum mlx5_fpga_image admin_image; + enum mlx5_fpga_image oper_image; + enum mlx5_fpga_status status; +}; + +enum mlx5_fpga_qpc_field_select { + MLX5_FPGA_QPC_STATE = BIT(0), +}; + +struct mlx5_fpga_qp_counters { + u64 rx_ack_packets; + u64 rx_send_packets; + u64 tx_ack_packets; + u64 tx_send_packets; + u64 rx_total_drop; +}; + +int mlx5_fpga_caps(struct mlx5_core_dev *dev, u32 *caps); +int mlx5_fpga_query(struct mlx5_core_dev *dev, struct mlx5_fpga_query *query); +int mlx5_fpga_ctrl_op(struct mlx5_core_dev *dev, u8 op); +int mlx5_fpga_access_reg(struct mlx5_core_dev *dev, u8 size, u64 addr, + void *buf, bool write); +int mlx5_fpga_sbu_caps(struct mlx5_core_dev *dev, void *caps, int size); + +int mlx5_fpga_create_qp(struct mlx5_core_dev *dev, void *fpga_qpc, + u32 *fpga_qpn); +int mlx5_fpga_modify_qp(struct mlx5_core_dev *dev, u32 fpga_qpn, + enum mlx5_fpga_qpc_field_select fields, void *fpga_qpc); +int mlx5_fpga_query_qp(struct mlx5_core_dev *dev, u32 fpga_qpn, void *fpga_qpc); +int mlx5_fpga_query_qp_counters(struct mlx5_core_dev *dev, u32 fpga_qpn, + bool clear, struct mlx5_fpga_qp_counters *data); +int mlx5_fpga_destroy_qp(struct mlx5_core_dev *dev, u32 fpga_qpn); + +#endif /* __MLX5_FPGA_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c new file mode 100644 index 000000000000..c4392f741c5f --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.c @@ -0,0 +1,1042 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <net/addrconf.h> +#include <linux/etherdevice.h> +#include <linux/mlx5/vport.h> + +#include "mlx5_core.h" +#include "lib/mlx5.h" +#include "fpga/conn.h" + +#define MLX5_FPGA_PKEY 0xFFFF +#define MLX5_FPGA_PKEY_INDEX 0 /* RoCE PKEY 0xFFFF is always at index 0 */ +#define MLX5_FPGA_RECV_SIZE 2048 +#define MLX5_FPGA_PORT_NUM 1 +#define MLX5_FPGA_CQ_BUDGET 64 + +static int mlx5_fpga_conn_map_buf(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_dma_buf *buf) +{ + struct device *dma_device; + int err = 0; + + if (unlikely(!buf->sg[0].data)) + goto out; + + dma_device = &conn->fdev->mdev->pdev->dev; + buf->sg[0].dma_addr = dma_map_single(dma_device, buf->sg[0].data, + buf->sg[0].size, buf->dma_dir); + err = dma_mapping_error(dma_device, buf->sg[0].dma_addr); + if (unlikely(err)) { + mlx5_fpga_warn(conn->fdev, "DMA error on sg 0: %d\n", err); + err = -ENOMEM; + goto out; + } + + if (!buf->sg[1].data) + goto out; + + buf->sg[1].dma_addr = dma_map_single(dma_device, buf->sg[1].data, + buf->sg[1].size, buf->dma_dir); + err = dma_mapping_error(dma_device, buf->sg[1].dma_addr); + if (unlikely(err)) { + mlx5_fpga_warn(conn->fdev, "DMA error on sg 1: %d\n", err); + dma_unmap_single(dma_device, buf->sg[0].dma_addr, + buf->sg[0].size, buf->dma_dir); + err = -ENOMEM; + } + +out: + return err; +} + +static void mlx5_fpga_conn_unmap_buf(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_dma_buf *buf) +{ + struct device *dma_device; + + dma_device = &conn->fdev->mdev->pdev->dev; + if (buf->sg[1].data) + dma_unmap_single(dma_device, buf->sg[1].dma_addr, + buf->sg[1].size, buf->dma_dir); + + if (likely(buf->sg[0].data)) + dma_unmap_single(dma_device, buf->sg[0].dma_addr, + buf->sg[0].size, buf->dma_dir); +} + +static int mlx5_fpga_conn_post_recv(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_dma_buf *buf) +{ + struct mlx5_wqe_data_seg *data; + unsigned int ix; + int err = 0; + + err = mlx5_fpga_conn_map_buf(conn, buf); + if (unlikely(err)) + goto out; + + if (unlikely(conn->qp.rq.pc - conn->qp.rq.cc >= conn->qp.rq.size)) { + mlx5_fpga_conn_unmap_buf(conn, buf); + return -EBUSY; + } + + ix = conn->qp.rq.pc & (conn->qp.rq.size - 1); + data = mlx5_wq_cyc_get_wqe(&conn->qp.wq.rq, ix); + data->byte_count = cpu_to_be32(buf->sg[0].size); + data->lkey = cpu_to_be32(conn->fdev->conn_res.mkey.key); + data->addr = cpu_to_be64(buf->sg[0].dma_addr); + + conn->qp.rq.pc++; + conn->qp.rq.bufs[ix] = buf; + + /* Make sure that descriptors are written before doorbell record. */ + dma_wmb(); + *conn->qp.wq.rq.db = cpu_to_be32(conn->qp.rq.pc & 0xffff); +out: + return err; +} + +static void mlx5_fpga_conn_notify_hw(struct mlx5_fpga_conn *conn, void *wqe) +{ + /* ensure wqe is visible to device before updating doorbell record */ + dma_wmb(); + *conn->qp.wq.sq.db = cpu_to_be32(conn->qp.sq.pc); + /* Make sure that doorbell record is visible before ringing */ + wmb(); + mlx5_write64(wqe, conn->fdev->conn_res.uar->map + MLX5_BF_OFFSET, NULL); +} + +static void mlx5_fpga_conn_post_send(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_dma_buf *buf) +{ + struct mlx5_wqe_ctrl_seg *ctrl; + struct mlx5_wqe_data_seg *data; + unsigned int ix, sgi; + int size = 1; + + ix = conn->qp.sq.pc & (conn->qp.sq.size - 1); + + ctrl = mlx5_wq_cyc_get_wqe(&conn->qp.wq.sq, ix); + data = (void *)(ctrl + 1); + + for (sgi = 0; sgi < ARRAY_SIZE(buf->sg); sgi++) { + if (!buf->sg[sgi].data) + break; + data->byte_count = cpu_to_be32(buf->sg[sgi].size); + data->lkey = cpu_to_be32(conn->fdev->conn_res.mkey.key); + data->addr = cpu_to_be64(buf->sg[sgi].dma_addr); + data++; + size++; + } + + ctrl->imm = 0; + ctrl->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE; + ctrl->opmod_idx_opcode = cpu_to_be32(((conn->qp.sq.pc & 0xffff) << 8) | + MLX5_OPCODE_SEND); + ctrl->qpn_ds = cpu_to_be32(size | (conn->qp.mqp.qpn << 8)); + + conn->qp.sq.pc++; + conn->qp.sq.bufs[ix] = buf; + mlx5_fpga_conn_notify_hw(conn, ctrl); +} + +int mlx5_fpga_conn_send(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_dma_buf *buf) +{ + unsigned long flags; + int err; + + if (!conn->qp.active) + return -ENOTCONN; + + err = mlx5_fpga_conn_map_buf(conn, buf); + if (err) + return err; + + spin_lock_irqsave(&conn->qp.sq.lock, flags); + + if (conn->qp.sq.pc - conn->qp.sq.cc >= conn->qp.sq.size) { + list_add_tail(&buf->list, &conn->qp.sq.backlog); + goto out_unlock; + } + + mlx5_fpga_conn_post_send(conn, buf); + +out_unlock: + spin_unlock_irqrestore(&conn->qp.sq.lock, flags); + return err; +} + +static int mlx5_fpga_conn_post_recv_buf(struct mlx5_fpga_conn *conn) +{ + struct mlx5_fpga_dma_buf *buf; + int err; + + buf = kzalloc(sizeof(*buf) + MLX5_FPGA_RECV_SIZE, 0); + if (!buf) + return -ENOMEM; + + buf->sg[0].data = (void *)(buf + 1); + buf->sg[0].size = MLX5_FPGA_RECV_SIZE; + buf->dma_dir = DMA_FROM_DEVICE; + + err = mlx5_fpga_conn_post_recv(conn, buf); + if (err) + kfree(buf); + + return err; +} + +static int mlx5_fpga_conn_create_mkey(struct mlx5_core_dev *mdev, u32 pdn, + struct mlx5_core_mkey *mkey) +{ + int inlen = MLX5_ST_SZ_BYTES(create_mkey_in); + void *mkc; + u32 *in; + int err; + + in = kvzalloc(inlen, GFP_KERNEL); + if (!in) + return -ENOMEM; + + mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry); + MLX5_SET(mkc, mkc, access_mode, MLX5_MKC_ACCESS_MODE_PA); + MLX5_SET(mkc, mkc, lw, 1); + MLX5_SET(mkc, mkc, lr, 1); + + MLX5_SET(mkc, mkc, pd, pdn); + MLX5_SET(mkc, mkc, length64, 1); + MLX5_SET(mkc, mkc, qpn, 0xffffff); + + err = mlx5_core_create_mkey(mdev, mkey, in, inlen); + + kvfree(in); + return err; +} + +static void mlx5_fpga_conn_rq_cqe(struct mlx5_fpga_conn *conn, + struct mlx5_cqe64 *cqe, u8 status) +{ + struct mlx5_fpga_dma_buf *buf; + int ix, err; + + ix = be16_to_cpu(cqe->wqe_counter) & (conn->qp.rq.size - 1); + buf = conn->qp.rq.bufs[ix]; + conn->qp.rq.bufs[ix] = NULL; + if (!status) + buf->sg[0].size = be32_to_cpu(cqe->byte_cnt); + conn->qp.rq.cc++; + + if (unlikely(status && (status != MLX5_CQE_SYNDROME_WR_FLUSH_ERR))) + mlx5_fpga_warn(conn->fdev, "RQ buf %p on FPGA QP %u completion status %d\n", + buf, conn->fpga_qpn, status); + else + mlx5_fpga_dbg(conn->fdev, "RQ buf %p on FPGA QP %u completion status %d\n", + buf, conn->fpga_qpn, status); + + mlx5_fpga_conn_unmap_buf(conn, buf); + + if (unlikely(status || !conn->qp.active)) { + conn->qp.active = false; + kfree(buf); + return; + } + + mlx5_fpga_dbg(conn->fdev, "Message with %u bytes received successfully\n", + buf->sg[0].size); + conn->recv_cb(conn->cb_arg, buf); + + buf->sg[0].size = MLX5_FPGA_RECV_SIZE; + err = mlx5_fpga_conn_post_recv(conn, buf); + if (unlikely(err)) { + mlx5_fpga_warn(conn->fdev, + "Failed to re-post recv buf: %d\n", err); + kfree(buf); + } +} + +static void mlx5_fpga_conn_sq_cqe(struct mlx5_fpga_conn *conn, + struct mlx5_cqe64 *cqe, u8 status) +{ + struct mlx5_fpga_dma_buf *buf, *nextbuf; + unsigned long flags; + int ix; + + spin_lock_irqsave(&conn->qp.sq.lock, flags); + + ix = be16_to_cpu(cqe->wqe_counter) & (conn->qp.sq.size - 1); + buf = conn->qp.sq.bufs[ix]; + conn->qp.sq.bufs[ix] = NULL; + conn->qp.sq.cc++; + + /* Handle backlog still under the spinlock to ensure message post order */ + if (unlikely(!list_empty(&conn->qp.sq.backlog))) { + if (likely(conn->qp.active)) { + nextbuf = list_first_entry(&conn->qp.sq.backlog, + struct mlx5_fpga_dma_buf, list); + list_del(&nextbuf->list); + mlx5_fpga_conn_post_send(conn, nextbuf); + } + } + + spin_unlock_irqrestore(&conn->qp.sq.lock, flags); + + if (unlikely(status && (status != MLX5_CQE_SYNDROME_WR_FLUSH_ERR))) + mlx5_fpga_warn(conn->fdev, "SQ buf %p on FPGA QP %u completion status %d\n", + buf, conn->fpga_qpn, status); + else + mlx5_fpga_dbg(conn->fdev, "SQ buf %p on FPGA QP %u completion status %d\n", + buf, conn->fpga_qpn, status); + + mlx5_fpga_conn_unmap_buf(conn, buf); + + if (likely(buf->complete)) + buf->complete(conn, conn->fdev, buf, status); + + if (unlikely(status)) + conn->qp.active = false; +} + +static void mlx5_fpga_conn_handle_cqe(struct mlx5_fpga_conn *conn, + struct mlx5_cqe64 *cqe) +{ + u8 opcode, status = 0; + + opcode = cqe->op_own >> 4; + + switch (opcode) { + case MLX5_CQE_REQ_ERR: + status = ((struct mlx5_err_cqe *)cqe)->syndrome; + /* Fall through */ + case MLX5_CQE_REQ: + mlx5_fpga_conn_sq_cqe(conn, cqe, status); + break; + + case MLX5_CQE_RESP_ERR: + status = ((struct mlx5_err_cqe *)cqe)->syndrome; + /* Fall through */ + case MLX5_CQE_RESP_SEND: + mlx5_fpga_conn_rq_cqe(conn, cqe, status); + break; + default: + mlx5_fpga_warn(conn->fdev, "Unexpected cqe opcode %u\n", + opcode); + } +} + +static void mlx5_fpga_conn_arm_cq(struct mlx5_fpga_conn *conn) +{ + mlx5_cq_arm(&conn->cq.mcq, MLX5_CQ_DB_REQ_NOT, + conn->fdev->conn_res.uar->map, conn->cq.wq.cc); +} + +static void mlx5_fpga_conn_cq_event(struct mlx5_core_cq *mcq, + enum mlx5_event event) +{ + struct mlx5_fpga_conn *conn; + + conn = container_of(mcq, struct mlx5_fpga_conn, cq.mcq); + mlx5_fpga_warn(conn->fdev, "CQ event %u on CQ #%u\n", event, mcq->cqn); +} + +static void mlx5_fpga_conn_event(struct mlx5_core_qp *mqp, int event) +{ + struct mlx5_fpga_conn *conn; + + conn = container_of(mqp, struct mlx5_fpga_conn, qp.mqp); + mlx5_fpga_warn(conn->fdev, "QP event %u on QP #%u\n", event, mqp->qpn); +} + +static inline void mlx5_fpga_conn_cqes(struct mlx5_fpga_conn *conn, + unsigned int budget) +{ + struct mlx5_cqe64 *cqe; + + while (budget) { + cqe = mlx5_cqwq_get_cqe(&conn->cq.wq); + if (!cqe) + break; + + budget--; + mlx5_cqwq_pop(&conn->cq.wq); + mlx5_fpga_conn_handle_cqe(conn, cqe); + mlx5_cqwq_update_db_record(&conn->cq.wq); + } + if (!budget) { + tasklet_schedule(&conn->cq.tasklet); + return; + } + + mlx5_fpga_dbg(conn->fdev, "Re-arming CQ with cc# %u\n", conn->cq.wq.cc); + /* ensure cq space is freed before enabling more cqes */ + wmb(); + mlx5_fpga_conn_arm_cq(conn); +} + +static void mlx5_fpga_conn_cq_tasklet(unsigned long data) +{ + struct mlx5_fpga_conn *conn = (void *)data; + + if (unlikely(!conn->qp.active)) + return; + mlx5_fpga_conn_cqes(conn, MLX5_FPGA_CQ_BUDGET); +} + +static void mlx5_fpga_conn_cq_complete(struct mlx5_core_cq *mcq) +{ + struct mlx5_fpga_conn *conn; + + conn = container_of(mcq, struct mlx5_fpga_conn, cq.mcq); + if (unlikely(!conn->qp.active)) + return; + mlx5_fpga_conn_cqes(conn, MLX5_FPGA_CQ_BUDGET); +} + +static int mlx5_fpga_conn_create_cq(struct mlx5_fpga_conn *conn, int cq_size) +{ + struct mlx5_fpga_device *fdev = conn->fdev; + struct mlx5_core_dev *mdev = fdev->mdev; + u32 temp_cqc[MLX5_ST_SZ_DW(cqc)] = {0}; + struct mlx5_wq_param wqp; + struct mlx5_cqe64 *cqe; + int inlen, err, eqn; + unsigned int irqn; + void *cqc, *in; + __be64 *pas; + u32 i; + + cq_size = roundup_pow_of_two(cq_size); + MLX5_SET(cqc, temp_cqc, log_cq_size, ilog2(cq_size)); + + wqp.buf_numa_node = mdev->priv.numa_node; + wqp.db_numa_node = mdev->priv.numa_node; + + err = mlx5_cqwq_create(mdev, &wqp, temp_cqc, &conn->cq.wq, + &conn->cq.wq_ctrl); + if (err) + return err; + + for (i = 0; i < mlx5_cqwq_get_size(&conn->cq.wq); i++) { + cqe = mlx5_cqwq_get_wqe(&conn->cq.wq, i); + cqe->op_own = MLX5_CQE_INVALID << 4 | MLX5_CQE_OWNER_MASK; + } + + inlen = MLX5_ST_SZ_BYTES(create_cq_in) + + sizeof(u64) * conn->cq.wq_ctrl.frag_buf.npages; + in = kvzalloc(inlen, GFP_KERNEL); + if (!in) { + err = -ENOMEM; + goto err_cqwq; + } + + err = mlx5_vector2eqn(mdev, smp_processor_id(), &eqn, &irqn); + if (err) + goto err_cqwq; + + cqc = MLX5_ADDR_OF(create_cq_in, in, cq_context); + MLX5_SET(cqc, cqc, log_cq_size, ilog2(cq_size)); + MLX5_SET(cqc, cqc, c_eqn, eqn); + MLX5_SET(cqc, cqc, uar_page, fdev->conn_res.uar->index); + MLX5_SET(cqc, cqc, log_page_size, conn->cq.wq_ctrl.frag_buf.page_shift - + MLX5_ADAPTER_PAGE_SHIFT); + MLX5_SET64(cqc, cqc, dbr_addr, conn->cq.wq_ctrl.db.dma); + + pas = (__be64 *)MLX5_ADDR_OF(create_cq_in, in, pas); + mlx5_fill_page_frag_array(&conn->cq.wq_ctrl.frag_buf, pas); + + err = mlx5_core_create_cq(mdev, &conn->cq.mcq, in, inlen); + kvfree(in); + + if (err) + goto err_cqwq; + + conn->cq.mcq.cqe_sz = 64; + conn->cq.mcq.set_ci_db = conn->cq.wq_ctrl.db.db; + conn->cq.mcq.arm_db = conn->cq.wq_ctrl.db.db + 1; + *conn->cq.mcq.set_ci_db = 0; + *conn->cq.mcq.arm_db = 0; + conn->cq.mcq.vector = 0; + conn->cq.mcq.comp = mlx5_fpga_conn_cq_complete; + conn->cq.mcq.event = mlx5_fpga_conn_cq_event; + conn->cq.mcq.irqn = irqn; + conn->cq.mcq.uar = fdev->conn_res.uar; + tasklet_init(&conn->cq.tasklet, mlx5_fpga_conn_cq_tasklet, + (unsigned long)conn); + + mlx5_fpga_dbg(fdev, "Created CQ #0x%x\n", conn->cq.mcq.cqn); + + goto out; + +err_cqwq: + mlx5_cqwq_destroy(&conn->cq.wq_ctrl); +out: + return err; +} + +static void mlx5_fpga_conn_destroy_cq(struct mlx5_fpga_conn *conn) +{ + tasklet_disable(&conn->cq.tasklet); + tasklet_kill(&conn->cq.tasklet); + mlx5_core_destroy_cq(conn->fdev->mdev, &conn->cq.mcq); + mlx5_cqwq_destroy(&conn->cq.wq_ctrl); +} + +static int mlx5_fpga_conn_create_wq(struct mlx5_fpga_conn *conn, void *qpc) +{ + struct mlx5_fpga_device *fdev = conn->fdev; + struct mlx5_core_dev *mdev = fdev->mdev; + struct mlx5_wq_param wqp; + + wqp.buf_numa_node = mdev->priv.numa_node; + wqp.db_numa_node = mdev->priv.numa_node; + + return mlx5_wq_qp_create(mdev, &wqp, qpc, &conn->qp.wq, + &conn->qp.wq_ctrl); +} + +static int mlx5_fpga_conn_create_qp(struct mlx5_fpga_conn *conn, + unsigned int tx_size, unsigned int rx_size) +{ + struct mlx5_fpga_device *fdev = conn->fdev; + struct mlx5_core_dev *mdev = fdev->mdev; + u32 temp_qpc[MLX5_ST_SZ_DW(qpc)] = {0}; + void *in = NULL, *qpc; + int err, inlen; + + conn->qp.rq.pc = 0; + conn->qp.rq.cc = 0; + conn->qp.rq.size = roundup_pow_of_two(rx_size); + conn->qp.sq.pc = 0; + conn->qp.sq.cc = 0; + conn->qp.sq.size = roundup_pow_of_two(tx_size); + + MLX5_SET(qpc, temp_qpc, log_rq_stride, ilog2(MLX5_SEND_WQE_DS) - 4); + MLX5_SET(qpc, temp_qpc, log_rq_size, ilog2(conn->qp.rq.size)); + MLX5_SET(qpc, temp_qpc, log_sq_size, ilog2(conn->qp.sq.size)); + err = mlx5_fpga_conn_create_wq(conn, temp_qpc); + if (err) + goto out; + + conn->qp.rq.bufs = kvzalloc(sizeof(conn->qp.rq.bufs[0]) * + conn->qp.rq.size, GFP_KERNEL); + if (!conn->qp.rq.bufs) { + err = -ENOMEM; + goto err_wq; + } + + conn->qp.sq.bufs = kvzalloc(sizeof(conn->qp.sq.bufs[0]) * + conn->qp.sq.size, GFP_KERNEL); + if (!conn->qp.sq.bufs) { + err = -ENOMEM; + goto err_rq_bufs; + } + + inlen = MLX5_ST_SZ_BYTES(create_qp_in) + + MLX5_FLD_SZ_BYTES(create_qp_in, pas[0]) * + conn->qp.wq_ctrl.buf.npages; + in = kvzalloc(inlen, GFP_KERNEL); + if (!in) { + err = -ENOMEM; + goto err_sq_bufs; + } + + qpc = MLX5_ADDR_OF(create_qp_in, in, qpc); + MLX5_SET(qpc, qpc, uar_page, fdev->conn_res.uar->index); + MLX5_SET(qpc, qpc, log_page_size, + conn->qp.wq_ctrl.buf.page_shift - MLX5_ADAPTER_PAGE_SHIFT); + MLX5_SET(qpc, qpc, fre, 1); + MLX5_SET(qpc, qpc, rlky, 1); + MLX5_SET(qpc, qpc, st, MLX5_QP_ST_RC); + MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED); + MLX5_SET(qpc, qpc, pd, fdev->conn_res.pdn); + MLX5_SET(qpc, qpc, log_rq_stride, ilog2(MLX5_SEND_WQE_DS) - 4); + MLX5_SET(qpc, qpc, log_rq_size, ilog2(conn->qp.rq.size)); + MLX5_SET(qpc, qpc, rq_type, MLX5_NON_ZERO_RQ); + MLX5_SET(qpc, qpc, log_sq_size, ilog2(conn->qp.sq.size)); + MLX5_SET(qpc, qpc, cqn_snd, conn->cq.mcq.cqn); + MLX5_SET(qpc, qpc, cqn_rcv, conn->cq.mcq.cqn); + MLX5_SET64(qpc, qpc, dbr_addr, conn->qp.wq_ctrl.db.dma); + if (MLX5_CAP_GEN(mdev, cqe_version) == 1) + MLX5_SET(qpc, qpc, user_index, 0xFFFFFF); + + mlx5_fill_page_array(&conn->qp.wq_ctrl.buf, + (__be64 *)MLX5_ADDR_OF(create_qp_in, in, pas)); + + err = mlx5_core_create_qp(mdev, &conn->qp.mqp, in, inlen); + if (err) + goto err_sq_bufs; + + conn->qp.mqp.event = mlx5_fpga_conn_event; + mlx5_fpga_dbg(fdev, "Created QP #0x%x\n", conn->qp.mqp.qpn); + + goto out; + +err_sq_bufs: + kvfree(conn->qp.sq.bufs); +err_rq_bufs: + kvfree(conn->qp.rq.bufs); +err_wq: + mlx5_wq_destroy(&conn->qp.wq_ctrl); +out: + kvfree(in); + return err; +} + +static void mlx5_fpga_conn_free_recv_bufs(struct mlx5_fpga_conn *conn) +{ + int ix; + + for (ix = 0; ix < conn->qp.rq.size; ix++) { + if (!conn->qp.rq.bufs[ix]) + continue; + mlx5_fpga_conn_unmap_buf(conn, conn->qp.rq.bufs[ix]); + kfree(conn->qp.rq.bufs[ix]); + conn->qp.rq.bufs[ix] = NULL; + } +} + +static void mlx5_fpga_conn_flush_send_bufs(struct mlx5_fpga_conn *conn) +{ + struct mlx5_fpga_dma_buf *buf, *temp; + int ix; + + for (ix = 0; ix < conn->qp.sq.size; ix++) { + buf = conn->qp.sq.bufs[ix]; + if (!buf) + continue; + conn->qp.sq.bufs[ix] = NULL; + mlx5_fpga_conn_unmap_buf(conn, buf); + if (!buf->complete) + continue; + buf->complete(conn, conn->fdev, buf, MLX5_CQE_SYNDROME_WR_FLUSH_ERR); + } + list_for_each_entry_safe(buf, temp, &conn->qp.sq.backlog, list) { + mlx5_fpga_conn_unmap_buf(conn, buf); + if (!buf->complete) + continue; + buf->complete(conn, conn->fdev, buf, MLX5_CQE_SYNDROME_WR_FLUSH_ERR); + } +} + +static void mlx5_fpga_conn_destroy_qp(struct mlx5_fpga_conn *conn) +{ + mlx5_core_destroy_qp(conn->fdev->mdev, &conn->qp.mqp); + mlx5_fpga_conn_free_recv_bufs(conn); + mlx5_fpga_conn_flush_send_bufs(conn); + kvfree(conn->qp.sq.bufs); + kvfree(conn->qp.rq.bufs); + mlx5_wq_destroy(&conn->qp.wq_ctrl); +} + +static inline int mlx5_fpga_conn_reset_qp(struct mlx5_fpga_conn *conn) +{ + struct mlx5_core_dev *mdev = conn->fdev->mdev; + + mlx5_fpga_dbg(conn->fdev, "Modifying QP %u to RST\n", conn->qp.mqp.qpn); + + return mlx5_core_qp_modify(mdev, MLX5_CMD_OP_2RST_QP, 0, NULL, + &conn->qp.mqp); +} + +static inline int mlx5_fpga_conn_init_qp(struct mlx5_fpga_conn *conn) +{ + struct mlx5_fpga_device *fdev = conn->fdev; + struct mlx5_core_dev *mdev = fdev->mdev; + u32 *qpc = NULL; + int err; + + mlx5_fpga_dbg(conn->fdev, "Modifying QP %u to INIT\n", conn->qp.mqp.qpn); + + qpc = kzalloc(MLX5_ST_SZ_BYTES(qpc), GFP_KERNEL); + if (!qpc) { + err = -ENOMEM; + goto out; + } + + MLX5_SET(qpc, qpc, st, MLX5_QP_ST_RC); + MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED); + MLX5_SET(qpc, qpc, primary_address_path.pkey_index, MLX5_FPGA_PKEY_INDEX); + MLX5_SET(qpc, qpc, primary_address_path.port, MLX5_FPGA_PORT_NUM); + MLX5_SET(qpc, qpc, pd, conn->fdev->conn_res.pdn); + MLX5_SET(qpc, qpc, cqn_snd, conn->cq.mcq.cqn); + MLX5_SET(qpc, qpc, cqn_rcv, conn->cq.mcq.cqn); + MLX5_SET64(qpc, qpc, dbr_addr, conn->qp.wq_ctrl.db.dma); + + err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_RST2INIT_QP, 0, qpc, + &conn->qp.mqp); + if (err) { + mlx5_fpga_warn(fdev, "qp_modify RST2INIT failed: %d\n", err); + goto out; + } + +out: + kfree(qpc); + return err; +} + +static inline int mlx5_fpga_conn_rtr_qp(struct mlx5_fpga_conn *conn) +{ + struct mlx5_fpga_device *fdev = conn->fdev; + struct mlx5_core_dev *mdev = fdev->mdev; + u32 *qpc = NULL; + int err; + + mlx5_fpga_dbg(conn->fdev, "QP RTR\n"); + + qpc = kzalloc(MLX5_ST_SZ_BYTES(qpc), GFP_KERNEL); + if (!qpc) { + err = -ENOMEM; + goto out; + } + + MLX5_SET(qpc, qpc, mtu, MLX5_QPC_MTU_1K_BYTES); + MLX5_SET(qpc, qpc, log_msg_max, (u8)MLX5_CAP_GEN(mdev, log_max_msg)); + MLX5_SET(qpc, qpc, remote_qpn, conn->fpga_qpn); + MLX5_SET(qpc, qpc, next_rcv_psn, + MLX5_GET(fpga_qpc, conn->fpga_qpc, next_send_psn)); + MLX5_SET(qpc, qpc, primary_address_path.pkey_index, MLX5_FPGA_PKEY_INDEX); + MLX5_SET(qpc, qpc, primary_address_path.port, MLX5_FPGA_PORT_NUM); + ether_addr_copy(MLX5_ADDR_OF(qpc, qpc, primary_address_path.rmac_47_32), + MLX5_ADDR_OF(fpga_qpc, conn->fpga_qpc, fpga_mac_47_32)); + MLX5_SET(qpc, qpc, primary_address_path.udp_sport, + MLX5_CAP_ROCE(mdev, r_roce_min_src_udp_port)); + MLX5_SET(qpc, qpc, primary_address_path.src_addr_index, + conn->qp.sgid_index); + MLX5_SET(qpc, qpc, primary_address_path.hop_limit, 0); + memcpy(MLX5_ADDR_OF(qpc, qpc, primary_address_path.rgid_rip), + MLX5_ADDR_OF(fpga_qpc, conn->fpga_qpc, fpga_ip), + MLX5_FLD_SZ_BYTES(qpc, primary_address_path.rgid_rip)); + + err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_INIT2RTR_QP, 0, qpc, + &conn->qp.mqp); + if (err) { + mlx5_fpga_warn(fdev, "qp_modify RST2INIT failed: %d\n", err); + goto out; + } + +out: + kfree(qpc); + return err; +} + +static inline int mlx5_fpga_conn_rts_qp(struct mlx5_fpga_conn *conn) +{ + struct mlx5_fpga_device *fdev = conn->fdev; + struct mlx5_core_dev *mdev = fdev->mdev; + u32 *qpc = NULL; + u32 opt_mask; + int err; + + mlx5_fpga_dbg(conn->fdev, "QP RTS\n"); + + qpc = kzalloc(MLX5_ST_SZ_BYTES(qpc), GFP_KERNEL); + if (!qpc) { + err = -ENOMEM; + goto out; + } + + MLX5_SET(qpc, qpc, log_ack_req_freq, 8); + MLX5_SET(qpc, qpc, min_rnr_nak, 0x12); + MLX5_SET(qpc, qpc, primary_address_path.ack_timeout, 0x12); /* ~1.07s */ + MLX5_SET(qpc, qpc, next_send_psn, + MLX5_GET(fpga_qpc, conn->fpga_qpc, next_rcv_psn)); + MLX5_SET(qpc, qpc, retry_count, 7); + MLX5_SET(qpc, qpc, rnr_retry, 7); /* Infinite retry if RNR NACK */ + + opt_mask = MLX5_QP_OPTPAR_RNR_TIMEOUT; + err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_RTR2RTS_QP, opt_mask, qpc, + &conn->qp.mqp); + if (err) { + mlx5_fpga_warn(fdev, "qp_modify RST2INIT failed: %d\n", err); + goto out; + } + +out: + kfree(qpc); + return err; +} + +static int mlx5_fpga_conn_connect(struct mlx5_fpga_conn *conn) +{ + struct mlx5_fpga_device *fdev = conn->fdev; + int err; + + MLX5_SET(fpga_qpc, conn->fpga_qpc, state, MLX5_FPGA_QPC_STATE_ACTIVE); + err = mlx5_fpga_modify_qp(conn->fdev->mdev, conn->fpga_qpn, + MLX5_FPGA_QPC_STATE, &conn->fpga_qpc); + if (err) { + mlx5_fpga_err(fdev, "Failed to activate FPGA RC QP: %d\n", err); + goto out; + } + + err = mlx5_fpga_conn_reset_qp(conn); + if (err) { + mlx5_fpga_err(fdev, "Failed to change QP state to reset\n"); + goto err_fpga_qp; + } + + err = mlx5_fpga_conn_init_qp(conn); + if (err) { + mlx5_fpga_err(fdev, "Failed to modify QP from RESET to INIT\n"); + goto err_fpga_qp; + } + conn->qp.active = true; + + while (!mlx5_fpga_conn_post_recv_buf(conn)) + ; + + err = mlx5_fpga_conn_rtr_qp(conn); + if (err) { + mlx5_fpga_err(fdev, "Failed to change QP state from INIT to RTR\n"); + goto err_recv_bufs; + } + + err = mlx5_fpga_conn_rts_qp(conn); + if (err) { + mlx5_fpga_err(fdev, "Failed to change QP state from RTR to RTS\n"); + goto err_recv_bufs; + } + goto out; + +err_recv_bufs: + mlx5_fpga_conn_free_recv_bufs(conn); +err_fpga_qp: + MLX5_SET(fpga_qpc, conn->fpga_qpc, state, MLX5_FPGA_QPC_STATE_INIT); + if (mlx5_fpga_modify_qp(conn->fdev->mdev, conn->fpga_qpn, + MLX5_FPGA_QPC_STATE, &conn->fpga_qpc)) + mlx5_fpga_err(fdev, "Failed to revert FPGA QP to INIT\n"); +out: + return err; +} + +struct mlx5_fpga_conn *mlx5_fpga_conn_create(struct mlx5_fpga_device *fdev, + struct mlx5_fpga_conn_attr *attr, + enum mlx5_ifc_fpga_qp_type qp_type) +{ + struct mlx5_fpga_conn *ret, *conn; + u8 *remote_mac, *remote_ip; + int err; + + if (!attr->recv_cb) + return ERR_PTR(-EINVAL); + + conn = kzalloc(sizeof(*conn), GFP_KERNEL); + if (!conn) + return ERR_PTR(-ENOMEM); + + conn->fdev = fdev; + INIT_LIST_HEAD(&conn->qp.sq.backlog); + + spin_lock_init(&conn->qp.sq.lock); + + conn->recv_cb = attr->recv_cb; + conn->cb_arg = attr->cb_arg; + + remote_mac = MLX5_ADDR_OF(fpga_qpc, conn->fpga_qpc, remote_mac_47_32); + err = mlx5_query_nic_vport_mac_address(fdev->mdev, 0, remote_mac); + if (err) { + mlx5_fpga_err(fdev, "Failed to query local MAC: %d\n", err); + ret = ERR_PTR(err); + goto err; + } + + /* Build Modified EUI-64 IPv6 address from the MAC address */ + remote_ip = MLX5_ADDR_OF(fpga_qpc, conn->fpga_qpc, remote_ip); + remote_ip[0] = 0xfe; + remote_ip[1] = 0x80; + addrconf_addr_eui48(&remote_ip[8], remote_mac); + + err = mlx5_core_reserved_gid_alloc(fdev->mdev, &conn->qp.sgid_index); + if (err) { + mlx5_fpga_err(fdev, "Failed to allocate SGID: %d\n", err); + ret = ERR_PTR(err); + goto err; + } + + err = mlx5_core_roce_gid_set(fdev->mdev, conn->qp.sgid_index, + MLX5_ROCE_VERSION_2, + MLX5_ROCE_L3_TYPE_IPV6, + remote_ip, remote_mac, true, 0); + if (err) { + mlx5_fpga_err(fdev, "Failed to set SGID: %d\n", err); + ret = ERR_PTR(err); + goto err_rsvd_gid; + } + mlx5_fpga_dbg(fdev, "Reserved SGID index %u\n", conn->qp.sgid_index); + + /* Allow for one cqe per rx/tx wqe, plus one cqe for the next wqe, + * created during processing of the cqe + */ + err = mlx5_fpga_conn_create_cq(conn, + (attr->tx_size + attr->rx_size) * 2); + if (err) { + mlx5_fpga_err(fdev, "Failed to create CQ: %d\n", err); + ret = ERR_PTR(err); + goto err_gid; + } + + mlx5_fpga_conn_arm_cq(conn); + + err = mlx5_fpga_conn_create_qp(conn, attr->tx_size, attr->rx_size); + if (err) { + mlx5_fpga_err(fdev, "Failed to create QP: %d\n", err); + ret = ERR_PTR(err); + goto err_cq; + } + + MLX5_SET(fpga_qpc, conn->fpga_qpc, state, MLX5_FPGA_QPC_STATE_INIT); + MLX5_SET(fpga_qpc, conn->fpga_qpc, qp_type, qp_type); + MLX5_SET(fpga_qpc, conn->fpga_qpc, st, MLX5_FPGA_QPC_ST_RC); + MLX5_SET(fpga_qpc, conn->fpga_qpc, ether_type, ETH_P_8021Q); + MLX5_SET(fpga_qpc, conn->fpga_qpc, vid, 0); + MLX5_SET(fpga_qpc, conn->fpga_qpc, next_rcv_psn, 1); + MLX5_SET(fpga_qpc, conn->fpga_qpc, next_send_psn, 0); + MLX5_SET(fpga_qpc, conn->fpga_qpc, pkey, MLX5_FPGA_PKEY); + MLX5_SET(fpga_qpc, conn->fpga_qpc, remote_qpn, conn->qp.mqp.qpn); + MLX5_SET(fpga_qpc, conn->fpga_qpc, rnr_retry, 7); + MLX5_SET(fpga_qpc, conn->fpga_qpc, retry_count, 7); + + err = mlx5_fpga_create_qp(fdev->mdev, &conn->fpga_qpc, + &conn->fpga_qpn); + if (err) { + mlx5_fpga_err(fdev, "Failed to create FPGA RC QP: %d\n", err); + ret = ERR_PTR(err); + goto err_qp; + } + + err = mlx5_fpga_conn_connect(conn); + if (err) { + ret = ERR_PTR(err); + goto err_conn; + } + + mlx5_fpga_dbg(fdev, "FPGA QPN is %u\n", conn->fpga_qpn); + ret = conn; + goto out; + +err_conn: + mlx5_fpga_destroy_qp(conn->fdev->mdev, conn->fpga_qpn); +err_qp: + mlx5_fpga_conn_destroy_qp(conn); +err_cq: + mlx5_fpga_conn_destroy_cq(conn); +err_gid: + mlx5_core_roce_gid_set(fdev->mdev, conn->qp.sgid_index, 0, 0, NULL, + NULL, false, 0); +err_rsvd_gid: + mlx5_core_reserved_gid_free(fdev->mdev, conn->qp.sgid_index); +err: + kfree(conn); +out: + return ret; +} + +void mlx5_fpga_conn_destroy(struct mlx5_fpga_conn *conn) +{ + struct mlx5_fpga_device *fdev = conn->fdev; + struct mlx5_core_dev *mdev = fdev->mdev; + int err = 0; + + conn->qp.active = false; + tasklet_disable(&conn->cq.tasklet); + synchronize_irq(conn->cq.mcq.irqn); + + mlx5_fpga_destroy_qp(conn->fdev->mdev, conn->fpga_qpn); + err = mlx5_core_qp_modify(mdev, MLX5_CMD_OP_2ERR_QP, 0, NULL, + &conn->qp.mqp); + if (err) + mlx5_fpga_warn(fdev, "qp_modify 2ERR failed: %d\n", err); + mlx5_fpga_conn_destroy_qp(conn); + mlx5_fpga_conn_destroy_cq(conn); + + mlx5_core_roce_gid_set(conn->fdev->mdev, conn->qp.sgid_index, 0, 0, + NULL, NULL, false, 0); + mlx5_core_reserved_gid_free(conn->fdev->mdev, conn->qp.sgid_index); + kfree(conn); +} + +int mlx5_fpga_conn_device_init(struct mlx5_fpga_device *fdev) +{ + int err; + + err = mlx5_nic_vport_enable_roce(fdev->mdev); + if (err) { + mlx5_fpga_err(fdev, "Failed to enable RoCE: %d\n", err); + goto out; + } + + fdev->conn_res.uar = mlx5_get_uars_page(fdev->mdev); + if (IS_ERR(fdev->conn_res.uar)) { + err = PTR_ERR(fdev->conn_res.uar); + mlx5_fpga_err(fdev, "get_uars_page failed, %d\n", err); + goto err_roce; + } + mlx5_fpga_dbg(fdev, "Allocated UAR index %u\n", + fdev->conn_res.uar->index); + + err = mlx5_core_alloc_pd(fdev->mdev, &fdev->conn_res.pdn); + if (err) { + mlx5_fpga_err(fdev, "alloc pd failed, %d\n", err); + goto err_uar; + } + mlx5_fpga_dbg(fdev, "Allocated PD %u\n", fdev->conn_res.pdn); + + err = mlx5_fpga_conn_create_mkey(fdev->mdev, fdev->conn_res.pdn, + &fdev->conn_res.mkey); + if (err) { + mlx5_fpga_err(fdev, "create mkey failed, %d\n", err); + goto err_dealloc_pd; + } + mlx5_fpga_dbg(fdev, "Created mkey 0x%x\n", fdev->conn_res.mkey.key); + + return 0; + +err_dealloc_pd: + mlx5_core_dealloc_pd(fdev->mdev, fdev->conn_res.pdn); +err_uar: + mlx5_put_uars_page(fdev->mdev, fdev->conn_res.uar); +err_roce: + mlx5_nic_vport_disable_roce(fdev->mdev); +out: + return err; +} + +void mlx5_fpga_conn_device_cleanup(struct mlx5_fpga_device *fdev) +{ + mlx5_core_destroy_mkey(fdev->mdev, &fdev->conn_res.mkey); + mlx5_core_dealloc_pd(fdev->mdev, fdev->conn_res.pdn); + mlx5_put_uars_page(fdev->mdev, fdev->conn_res.uar); + mlx5_nic_vport_disable_roce(fdev->mdev); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.h new file mode 100644 index 000000000000..44bd9eccc711 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/conn.h @@ -0,0 +1,96 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef __MLX5_FPGA_CONN_H__ +#define __MLX5_FPGA_CONN_H__ + +#include <linux/mlx5/cq.h> +#include <linux/mlx5/qp.h> + +#include "fpga/core.h" +#include "fpga/sdk.h" +#include "wq.h" + +struct mlx5_fpga_conn { + struct mlx5_fpga_device *fdev; + + void (*recv_cb)(void *cb_arg, struct mlx5_fpga_dma_buf *buf); + void *cb_arg; + + /* FPGA QP */ + u32 fpga_qpc[MLX5_ST_SZ_DW(fpga_qpc)]; + u32 fpga_qpn; + + /* CQ */ + struct { + struct mlx5_cqwq wq; + struct mlx5_frag_wq_ctrl wq_ctrl; + struct mlx5_core_cq mcq; + struct tasklet_struct tasklet; + } cq; + + /* QP */ + struct { + bool active; + int sgid_index; + struct mlx5_wq_qp wq; + struct mlx5_wq_ctrl wq_ctrl; + struct mlx5_core_qp mqp; + struct { + spinlock_t lock; /* Protects all SQ state */ + unsigned int pc; + unsigned int cc; + unsigned int size; + struct mlx5_fpga_dma_buf **bufs; + struct list_head backlog; + } sq; + struct { + unsigned int pc; + unsigned int cc; + unsigned int size; + struct mlx5_fpga_dma_buf **bufs; + } rq; + } qp; +}; + +int mlx5_fpga_conn_device_init(struct mlx5_fpga_device *fdev); +void mlx5_fpga_conn_device_cleanup(struct mlx5_fpga_device *fdev); +struct mlx5_fpga_conn * +mlx5_fpga_conn_create(struct mlx5_fpga_device *fdev, + struct mlx5_fpga_conn_attr *attr, + enum mlx5_ifc_fpga_qp_type qp_type); +void mlx5_fpga_conn_destroy(struct mlx5_fpga_conn *conn); +int mlx5_fpga_conn_send(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_dma_buf *buf); + +#endif /* __MLX5_FPGA_CONN_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c new file mode 100644 index 000000000000..9034e9960a76 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c @@ -0,0 +1,283 @@ +/* + * Copyright (c) 2017, Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/module.h> +#include <linux/etherdevice.h> +#include <linux/mlx5/driver.h> + +#include "mlx5_core.h" +#include "lib/mlx5.h" +#include "fpga/core.h" +#include "fpga/conn.h" + +static const char *const mlx5_fpga_error_strings[] = { + "Null Syndrome", + "Corrupted DDR", + "Flash Timeout", + "Internal Link Error", + "Watchdog HW Failure", + "I2C Failure", + "Image Changed", + "Temperature Critical", +}; + +static struct mlx5_fpga_device *mlx5_fpga_device_alloc(void) +{ + struct mlx5_fpga_device *fdev = NULL; + + fdev = kzalloc(sizeof(*fdev), GFP_KERNEL); + if (!fdev) + return NULL; + + spin_lock_init(&fdev->state_lock); + fdev->state = MLX5_FPGA_STATUS_NONE; + return fdev; +} + +static const char *mlx5_fpga_image_name(enum mlx5_fpga_image image) +{ + switch (image) { + case MLX5_FPGA_IMAGE_USER: + return "user"; + case MLX5_FPGA_IMAGE_FACTORY: + return "factory"; + default: + return "unknown"; + } +} + +static int mlx5_fpga_device_load_check(struct mlx5_fpga_device *fdev) +{ + struct mlx5_fpga_query query; + int err; + + err = mlx5_fpga_query(fdev->mdev, &query); + if (err) { + mlx5_fpga_err(fdev, "Failed to query status: %d\n", err); + return err; + } + + fdev->last_admin_image = query.admin_image; + fdev->last_oper_image = query.oper_image; + + mlx5_fpga_dbg(fdev, "Status %u; Admin image %u; Oper image %u\n", + query.status, query.admin_image, query.oper_image); + + if (query.status != MLX5_FPGA_STATUS_SUCCESS) { + mlx5_fpga_err(fdev, "%s image failed to load; status %u\n", + mlx5_fpga_image_name(fdev->last_oper_image), + query.status); + return -EIO; + } + + return 0; +} + +static int mlx5_fpga_device_brb(struct mlx5_fpga_device *fdev) +{ + int err; + struct mlx5_core_dev *mdev = fdev->mdev; + + err = mlx5_fpga_ctrl_op(mdev, MLX5_FPGA_CTRL_OPERATION_SANDBOX_BYPASS_ON); + if (err) { + mlx5_fpga_err(fdev, "Failed to set bypass on: %d\n", err); + return err; + } + err = mlx5_fpga_ctrl_op(mdev, MLX5_FPGA_CTRL_OPERATION_RESET_SANDBOX); + if (err) { + mlx5_fpga_err(fdev, "Failed to reset SBU: %d\n", err); + return err; + } + err = mlx5_fpga_ctrl_op(mdev, MLX5_FPGA_CTRL_OPERATION_SANDBOX_BYPASS_OFF); + if (err) { + mlx5_fpga_err(fdev, "Failed to set bypass off: %d\n", err); + return err; + } + return 0; +} + +int mlx5_fpga_device_start(struct mlx5_core_dev *mdev) +{ + struct mlx5_fpga_device *fdev = mdev->fpga; + unsigned long flags; + unsigned int max_num_qps; + int err; + + if (!fdev) + return 0; + + err = mlx5_fpga_device_load_check(fdev); + if (err) + goto out; + + err = mlx5_fpga_caps(fdev->mdev, + fdev->mdev->caps.hca_cur[MLX5_CAP_FPGA]); + if (err) + goto out; + + mlx5_fpga_info(fdev, "device %u; %s image, version %u\n", + MLX5_CAP_FPGA(fdev->mdev, fpga_device), + mlx5_fpga_image_name(fdev->last_oper_image), + MLX5_CAP_FPGA(fdev->mdev, image_version)); + + max_num_qps = MLX5_CAP_FPGA(mdev, shell_caps.max_num_qps); + err = mlx5_core_reserve_gids(mdev, max_num_qps); + if (err) + goto out; + + err = mlx5_fpga_conn_device_init(fdev); + if (err) + goto err_rsvd_gid; + + if (fdev->last_oper_image == MLX5_FPGA_IMAGE_USER) { + err = mlx5_fpga_device_brb(fdev); + if (err) + goto err_conn_init; + } + + goto out; + +err_conn_init: + mlx5_fpga_conn_device_cleanup(fdev); + +err_rsvd_gid: + mlx5_core_unreserve_gids(mdev, max_num_qps); +out: + spin_lock_irqsave(&fdev->state_lock, flags); + fdev->state = err ? MLX5_FPGA_STATUS_FAILURE : MLX5_FPGA_STATUS_SUCCESS; + spin_unlock_irqrestore(&fdev->state_lock, flags); + return err; +} + +int mlx5_fpga_init(struct mlx5_core_dev *mdev) +{ + struct mlx5_fpga_device *fdev = NULL; + + if (!MLX5_CAP_GEN(mdev, fpga)) { + mlx5_core_dbg(mdev, "FPGA capability not present\n"); + return 0; + } + + mlx5_core_dbg(mdev, "Initializing FPGA\n"); + + fdev = mlx5_fpga_device_alloc(); + if (!fdev) + return -ENOMEM; + + fdev->mdev = mdev; + mdev->fpga = fdev; + + return 0; +} + +void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev) +{ + struct mlx5_fpga_device *fdev = mdev->fpga; + unsigned int max_num_qps; + unsigned long flags; + int err; + + if (!fdev) + return; + + spin_lock_irqsave(&fdev->state_lock, flags); + if (fdev->state != MLX5_FPGA_STATUS_SUCCESS) { + spin_unlock_irqrestore(&fdev->state_lock, flags); + return; + } + fdev->state = MLX5_FPGA_STATUS_NONE; + spin_unlock_irqrestore(&fdev->state_lock, flags); + + if (fdev->last_oper_image == MLX5_FPGA_IMAGE_USER) { + err = mlx5_fpga_ctrl_op(mdev, MLX5_FPGA_CTRL_OPERATION_SANDBOX_BYPASS_ON); + if (err) + mlx5_fpga_err(fdev, "Failed to re-set SBU bypass on: %d\n", + err); + } + + mlx5_fpga_conn_device_cleanup(fdev); + max_num_qps = MLX5_CAP_FPGA(mdev, shell_caps.max_num_qps); + mlx5_core_unreserve_gids(mdev, max_num_qps); +} + +void mlx5_fpga_cleanup(struct mlx5_core_dev *mdev) +{ + struct mlx5_fpga_device *fdev = mdev->fpga; + + mlx5_fpga_device_stop(mdev); + kfree(fdev); + mdev->fpga = NULL; +} + +static const char *mlx5_fpga_syndrome_to_string(u8 syndrome) +{ + if (syndrome < ARRAY_SIZE(mlx5_fpga_error_strings)) + return mlx5_fpga_error_strings[syndrome]; + return "Unknown"; +} + +void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, void *data) +{ + struct mlx5_fpga_device *fdev = mdev->fpga; + const char *event_name; + bool teardown = false; + unsigned long flags; + u8 syndrome; + + if (event != MLX5_EVENT_TYPE_FPGA_ERROR) { + mlx5_fpga_warn_ratelimited(fdev, "Unexpected event %u\n", + event); + return; + } + + syndrome = MLX5_GET(fpga_error_event, data, syndrome); + event_name = mlx5_fpga_syndrome_to_string(syndrome); + + spin_lock_irqsave(&fdev->state_lock, flags); + switch (fdev->state) { + case MLX5_FPGA_STATUS_SUCCESS: + mlx5_fpga_warn(fdev, "Error %u: %s\n", syndrome, event_name); + teardown = true; + break; + default: + mlx5_fpga_warn_ratelimited(fdev, "Unexpected error event %u: %s\n", + syndrome, event_name); + } + spin_unlock_irqrestore(&fdev->state_lock, flags); + /* We tear-down the card's interfaces and functionality because + * the FPGA bump-on-the-wire is misbehaving and we lose ability + * to communicate with the network. User may still be able to + * recover by re-programming or debugging the FPGA + */ + if (teardown) + mlx5_trigger_health_work(fdev->mdev); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h new file mode 100644 index 000000000000..82405ed84725 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/core.h @@ -0,0 +1,113 @@ +/* + * Copyright (c) 2017, Mellanox Technologies, Ltd. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#ifndef __MLX5_FPGA_CORE_H__ +#define __MLX5_FPGA_CORE_H__ + +#ifdef CONFIG_MLX5_FPGA + +#include "fpga/cmd.h" + +/* Represents an Innova device */ +struct mlx5_fpga_device { + struct mlx5_core_dev *mdev; + spinlock_t state_lock; /* Protects state transitions */ + enum mlx5_fpga_status state; + enum mlx5_fpga_image last_admin_image; + enum mlx5_fpga_image last_oper_image; + + /* QP Connection resources */ + struct { + u32 pdn; + struct mlx5_core_mkey mkey; + struct mlx5_uars_page *uar; + } conn_res; + + struct mlx5_fpga_ipsec *ipsec; +}; + +#define mlx5_fpga_dbg(__adev, format, ...) \ + dev_dbg(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d:(pid %d): " format, \ + __func__, __LINE__, current->pid, ##__VA_ARGS__) + +#define mlx5_fpga_err(__adev, format, ...) \ + dev_err(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d:(pid %d): " format, \ + __func__, __LINE__, current->pid, ##__VA_ARGS__) + +#define mlx5_fpga_warn(__adev, format, ...) \ + dev_warn(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d:(pid %d): " format, \ + __func__, __LINE__, current->pid, ##__VA_ARGS__) + +#define mlx5_fpga_warn_ratelimited(__adev, format, ...) \ + dev_warn_ratelimited(&(__adev)->mdev->pdev->dev, "FPGA: %s:%d: " \ + format, __func__, __LINE__, ##__VA_ARGS__) + +#define mlx5_fpga_notice(__adev, format, ...) \ + dev_notice(&(__adev)->mdev->pdev->dev, "FPGA: " format, ##__VA_ARGS__) + +#define mlx5_fpga_info(__adev, format, ...) \ + dev_info(&(__adev)->mdev->pdev->dev, "FPGA: " format, ##__VA_ARGS__) + +int mlx5_fpga_init(struct mlx5_core_dev *mdev); +void mlx5_fpga_cleanup(struct mlx5_core_dev *mdev); +int mlx5_fpga_device_start(struct mlx5_core_dev *mdev); +void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev); +void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, void *data); + +#else + +static inline int mlx5_fpga_init(struct mlx5_core_dev *mdev) +{ + return 0; +} + +static inline void mlx5_fpga_cleanup(struct mlx5_core_dev *mdev) +{ +} + +static inline int mlx5_fpga_device_start(struct mlx5_core_dev *mdev) +{ + return 0; +} + +static inline void mlx5_fpga_device_stop(struct mlx5_core_dev *mdev) +{ +} + +static inline void mlx5_fpga_event(struct mlx5_core_dev *mdev, u8 event, + void *data) +{ +} + +#endif + +#endif /* __MLX5_FPGA_CORE_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c new file mode 100644 index 000000000000..35d0e33381ca --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c @@ -0,0 +1,376 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/mlx5/driver.h> + +#include "mlx5_core.h" +#include "fpga/ipsec.h" +#include "fpga/sdk.h" +#include "fpga/core.h" + +#define SBU_QP_QUEUE_SIZE 8 + +enum mlx5_ipsec_response_syndrome { + MLX5_IPSEC_RESPONSE_SUCCESS = 0, + MLX5_IPSEC_RESPONSE_ILLEGAL_REQUEST = 1, + MLX5_IPSEC_RESPONSE_SADB_ISSUE = 2, + MLX5_IPSEC_RESPONSE_WRITE_RESPONSE_ISSUE = 3, +}; + +enum mlx5_fpga_ipsec_sacmd_status { + MLX5_FPGA_IPSEC_SACMD_PENDING, + MLX5_FPGA_IPSEC_SACMD_SEND_FAIL, + MLX5_FPGA_IPSEC_SACMD_COMPLETE, +}; + +struct mlx5_ipsec_command_context { + struct mlx5_fpga_dma_buf buf; + struct mlx5_accel_ipsec_sa sa; + enum mlx5_fpga_ipsec_sacmd_status status; + int status_code; + struct completion complete; + struct mlx5_fpga_device *dev; + struct list_head list; /* Item in pending_cmds */ +}; + +struct mlx5_ipsec_sadb_resp { + __be32 syndrome; + __be32 sw_sa_handle; + u8 reserved[24]; +} __packed; + +struct mlx5_fpga_ipsec { + struct list_head pending_cmds; + spinlock_t pending_cmds_lock; /* Protects pending_cmds */ + u32 caps[MLX5_ST_SZ_DW(ipsec_extended_cap)]; + struct mlx5_fpga_conn *conn; +}; + +static bool mlx5_fpga_is_ipsec_device(struct mlx5_core_dev *mdev) +{ + if (!mdev->fpga || !MLX5_CAP_GEN(mdev, fpga)) + return false; + + if (MLX5_CAP_FPGA(mdev, ieee_vendor_id) != + MLX5_FPGA_CAP_SANDBOX_VENDOR_ID_MLNX) + return false; + + if (MLX5_CAP_FPGA(mdev, sandbox_product_id) != + MLX5_FPGA_CAP_SANDBOX_PRODUCT_ID_IPSEC) + return false; + + return true; +} + +static void mlx5_fpga_ipsec_send_complete(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_device *fdev, + struct mlx5_fpga_dma_buf *buf, + u8 status) +{ + struct mlx5_ipsec_command_context *context; + + if (status) { + context = container_of(buf, struct mlx5_ipsec_command_context, + buf); + mlx5_fpga_warn(fdev, "IPSec command send failed with status %u\n", + status); + context->status = MLX5_FPGA_IPSEC_SACMD_SEND_FAIL; + complete(&context->complete); + } +} + +static inline int syndrome_to_errno(enum mlx5_ipsec_response_syndrome syndrome) +{ + switch (syndrome) { + case MLX5_IPSEC_RESPONSE_SUCCESS: + return 0; + case MLX5_IPSEC_RESPONSE_SADB_ISSUE: + return -EEXIST; + case MLX5_IPSEC_RESPONSE_ILLEGAL_REQUEST: + return -EINVAL; + case MLX5_IPSEC_RESPONSE_WRITE_RESPONSE_ISSUE: + return -EIO; + } + return -EIO; +} + +static void mlx5_fpga_ipsec_recv(void *cb_arg, struct mlx5_fpga_dma_buf *buf) +{ + struct mlx5_ipsec_sadb_resp *resp = buf->sg[0].data; + struct mlx5_ipsec_command_context *context; + enum mlx5_ipsec_response_syndrome syndrome; + struct mlx5_fpga_device *fdev = cb_arg; + unsigned long flags; + + if (buf->sg[0].size < sizeof(*resp)) { + mlx5_fpga_warn(fdev, "Short receive from FPGA IPSec: %u < %zu bytes\n", + buf->sg[0].size, sizeof(*resp)); + return; + } + + mlx5_fpga_dbg(fdev, "mlx5_ipsec recv_cb syndrome %08x sa_id %x\n", + ntohl(resp->syndrome), ntohl(resp->sw_sa_handle)); + + spin_lock_irqsave(&fdev->ipsec->pending_cmds_lock, flags); + context = list_first_entry_or_null(&fdev->ipsec->pending_cmds, + struct mlx5_ipsec_command_context, + list); + if (context) + list_del(&context->list); + spin_unlock_irqrestore(&fdev->ipsec->pending_cmds_lock, flags); + + if (!context) { + mlx5_fpga_warn(fdev, "Received IPSec offload response without pending command request\n"); + return; + } + mlx5_fpga_dbg(fdev, "Handling response for %p\n", context); + + if (context->sa.sw_sa_handle != resp->sw_sa_handle) { + mlx5_fpga_err(fdev, "mismatch SA handle. cmd 0x%08x vs resp 0x%08x\n", + ntohl(context->sa.sw_sa_handle), + ntohl(resp->sw_sa_handle)); + return; + } + + syndrome = ntohl(resp->syndrome); + context->status_code = syndrome_to_errno(syndrome); + context->status = MLX5_FPGA_IPSEC_SACMD_COMPLETE; + + if (context->status_code) + mlx5_fpga_warn(fdev, "IPSec SADB command failed with syndrome %08x\n", + syndrome); + complete(&context->complete); +} + +void *mlx5_fpga_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, + struct mlx5_accel_ipsec_sa *cmd) +{ + struct mlx5_ipsec_command_context *context; + struct mlx5_fpga_device *fdev = mdev->fpga; + unsigned long flags; + int res = 0; + + BUILD_BUG_ON((sizeof(struct mlx5_accel_ipsec_sa) & 3) != 0); + if (!fdev || !fdev->ipsec) + return ERR_PTR(-EOPNOTSUPP); + + context = kzalloc(sizeof(*context), GFP_ATOMIC); + if (!context) + return ERR_PTR(-ENOMEM); + + memcpy(&context->sa, cmd, sizeof(*cmd)); + context->buf.complete = mlx5_fpga_ipsec_send_complete; + context->buf.sg[0].size = sizeof(context->sa); + context->buf.sg[0].data = &context->sa; + init_completion(&context->complete); + context->dev = fdev; + spin_lock_irqsave(&fdev->ipsec->pending_cmds_lock, flags); + list_add_tail(&context->list, &fdev->ipsec->pending_cmds); + spin_unlock_irqrestore(&fdev->ipsec->pending_cmds_lock, flags); + + context->status = MLX5_FPGA_IPSEC_SACMD_PENDING; + + res = mlx5_fpga_sbu_conn_sendmsg(fdev->ipsec->conn, &context->buf); + if (res) { + mlx5_fpga_warn(fdev, "Failure sending IPSec command: %d\n", + res); + spin_lock_irqsave(&fdev->ipsec->pending_cmds_lock, flags); + list_del(&context->list); + spin_unlock_irqrestore(&fdev->ipsec->pending_cmds_lock, flags); + kfree(context); + return ERR_PTR(res); + } + /* Context will be freed by wait func after completion */ + return context; +} + +int mlx5_fpga_ipsec_sa_cmd_wait(void *ctx) +{ + struct mlx5_ipsec_command_context *context = ctx; + int res; + + res = wait_for_completion_killable(&context->complete); + if (res) { + mlx5_fpga_warn(context->dev, "Failure waiting for IPSec command response\n"); + return -EINTR; + } + + if (context->status == MLX5_FPGA_IPSEC_SACMD_COMPLETE) + res = context->status_code; + else + res = -EIO; + + kfree(context); + return res; +} + +u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev) +{ + struct mlx5_fpga_device *fdev = mdev->fpga; + u32 ret = 0; + + if (mlx5_fpga_is_ipsec_device(mdev)) + ret |= MLX5_ACCEL_IPSEC_DEVICE; + else + return ret; + + if (!fdev->ipsec) + return ret; + + if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, esp)) + ret |= MLX5_ACCEL_IPSEC_ESP; + + if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, ipv6)) + ret |= MLX5_ACCEL_IPSEC_IPV6; + + if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, lso)) + ret |= MLX5_ACCEL_IPSEC_LSO; + + return ret; +} + +unsigned int mlx5_fpga_ipsec_counters_count(struct mlx5_core_dev *mdev) +{ + struct mlx5_fpga_device *fdev = mdev->fpga; + + if (!fdev || !fdev->ipsec) + return 0; + + return MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, + number_of_ipsec_counters); +} + +int mlx5_fpga_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters, + unsigned int counters_count) +{ + struct mlx5_fpga_device *fdev = mdev->fpga; + unsigned int i; + __be32 *data; + u32 count; + u64 addr; + int ret; + + if (!fdev || !fdev->ipsec) + return 0; + + addr = (u64)MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, + ipsec_counters_addr_low) + + ((u64)MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, + ipsec_counters_addr_high) << 32); + + count = mlx5_fpga_ipsec_counters_count(mdev); + + data = kzalloc(sizeof(*data) * count * 2, GFP_KERNEL); + if (!data) { + ret = -ENOMEM; + goto out; + } + + ret = mlx5_fpga_mem_read(fdev, count * sizeof(u64), addr, data, + MLX5_FPGA_ACCESS_TYPE_DONTCARE); + if (ret < 0) { + mlx5_fpga_err(fdev, "Failed to read IPSec counters from HW: %d\n", + ret); + goto out; + } + ret = 0; + + if (count > counters_count) + count = counters_count; + + /* Each counter is low word, then high. But each word is big-endian */ + for (i = 0; i < count; i++) + counters[i] = (u64)ntohl(data[i * 2]) | + ((u64)ntohl(data[i * 2 + 1]) << 32); + +out: + kfree(data); + return ret; +} + +int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev) +{ + struct mlx5_fpga_conn_attr init_attr = {0}; + struct mlx5_fpga_device *fdev = mdev->fpga; + struct mlx5_fpga_conn *conn; + int err; + + if (!mlx5_fpga_is_ipsec_device(mdev)) + return 0; + + fdev->ipsec = kzalloc(sizeof(*fdev->ipsec), GFP_KERNEL); + if (!fdev->ipsec) + return -ENOMEM; + + err = mlx5_fpga_get_sbu_caps(fdev, sizeof(fdev->ipsec->caps), + fdev->ipsec->caps); + if (err) { + mlx5_fpga_err(fdev, "Failed to retrieve IPSec extended capabilities: %d\n", + err); + goto error; + } + + INIT_LIST_HEAD(&fdev->ipsec->pending_cmds); + spin_lock_init(&fdev->ipsec->pending_cmds_lock); + + init_attr.rx_size = SBU_QP_QUEUE_SIZE; + init_attr.tx_size = SBU_QP_QUEUE_SIZE; + init_attr.recv_cb = mlx5_fpga_ipsec_recv; + init_attr.cb_arg = fdev; + conn = mlx5_fpga_sbu_conn_create(fdev, &init_attr); + if (IS_ERR(conn)) { + err = PTR_ERR(conn); + mlx5_fpga_err(fdev, "Error creating IPSec command connection %d\n", + err); + goto error; + } + fdev->ipsec->conn = conn; + return 0; + +error: + kfree(fdev->ipsec); + fdev->ipsec = NULL; + return err; +} + +void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev) +{ + struct mlx5_fpga_device *fdev = mdev->fpga; + + if (!mlx5_fpga_is_ipsec_device(mdev)) + return; + + mlx5_fpga_sbu_conn_destroy(fdev->ipsec->conn); + kfree(fdev->ipsec); + fdev->ipsec = NULL; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h new file mode 100644 index 000000000000..26a3e4b56972 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h @@ -0,0 +1,94 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef __MLX5_FPGA_IPSEC_H__ +#define __MLX5_FPGA_IPSEC_H__ + +#include "accel/ipsec.h" + +#ifdef CONFIG_MLX5_FPGA + +void *mlx5_fpga_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, + struct mlx5_accel_ipsec_sa *cmd); +int mlx5_fpga_ipsec_sa_cmd_wait(void *context); + +u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev); +unsigned int mlx5_fpga_ipsec_counters_count(struct mlx5_core_dev *mdev); +int mlx5_fpga_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters, + unsigned int counters_count); + +int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev); +void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev); + +#else + +static inline void *mlx5_fpga_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, + struct mlx5_accel_ipsec_sa *cmd) +{ + return ERR_PTR(-EOPNOTSUPP); +} + +static inline int mlx5_fpga_ipsec_sa_cmd_wait(void *context) +{ + return -EOPNOTSUPP; +} + +static inline u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev) +{ + return 0; +} + +static inline unsigned int +mlx5_fpga_ipsec_counters_count(struct mlx5_core_dev *mdev) +{ + return 0; +} + +static inline int mlx5_fpga_ipsec_counters_read(struct mlx5_core_dev *mdev, + u64 *counters) +{ + return 0; +} + +static inline int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev) +{ + return 0; +} + +static inline void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev) +{ +} + +#endif /* CONFIG_MLX5_FPGA */ + +#endif /* __MLX5_FPGA_SADB_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.c new file mode 100644 index 000000000000..3c11d6e2160a --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.c @@ -0,0 +1,164 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#include <linux/mlx5/device.h> + +#include "fpga/core.h" +#include "fpga/conn.h" +#include "fpga/sdk.h" + +struct mlx5_fpga_conn * +mlx5_fpga_sbu_conn_create(struct mlx5_fpga_device *fdev, + struct mlx5_fpga_conn_attr *attr) +{ + return mlx5_fpga_conn_create(fdev, attr, MLX5_FPGA_QPC_QP_TYPE_SANDBOX_QP); +} +EXPORT_SYMBOL(mlx5_fpga_sbu_conn_create); + +void mlx5_fpga_sbu_conn_destroy(struct mlx5_fpga_conn *conn) +{ + mlx5_fpga_conn_destroy(conn); +} +EXPORT_SYMBOL(mlx5_fpga_sbu_conn_destroy); + +int mlx5_fpga_sbu_conn_sendmsg(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_dma_buf *buf) +{ + return mlx5_fpga_conn_send(conn, buf); +} +EXPORT_SYMBOL(mlx5_fpga_sbu_conn_sendmsg); + +static int mlx5_fpga_mem_read_i2c(struct mlx5_fpga_device *fdev, size_t size, + u64 addr, u8 *buf) +{ + size_t max_size = MLX5_FPGA_ACCESS_REG_SIZE_MAX; + size_t bytes_done = 0; + u8 actual_size; + int err; + + if (!fdev->mdev) + return -ENOTCONN; + + while (bytes_done < size) { + actual_size = min(max_size, (size - bytes_done)); + + err = mlx5_fpga_access_reg(fdev->mdev, actual_size, + addr + bytes_done, + buf + bytes_done, false); + if (err) { + mlx5_fpga_err(fdev, "Failed to read over I2C: %d\n", + err); + break; + } + + bytes_done += actual_size; + } + + return err; +} + +static int mlx5_fpga_mem_write_i2c(struct mlx5_fpga_device *fdev, size_t size, + u64 addr, u8 *buf) +{ + size_t max_size = MLX5_FPGA_ACCESS_REG_SIZE_MAX; + size_t bytes_done = 0; + u8 actual_size; + int err; + + if (!fdev->mdev) + return -ENOTCONN; + + while (bytes_done < size) { + actual_size = min(max_size, (size - bytes_done)); + + err = mlx5_fpga_access_reg(fdev->mdev, actual_size, + addr + bytes_done, + buf + bytes_done, true); + if (err) { + mlx5_fpga_err(fdev, "Failed to write FPGA crspace\n"); + break; + } + + bytes_done += actual_size; + } + + return err; +} + +int mlx5_fpga_mem_read(struct mlx5_fpga_device *fdev, size_t size, u64 addr, + void *buf, enum mlx5_fpga_access_type access_type) +{ + int ret; + + switch (access_type) { + case MLX5_FPGA_ACCESS_TYPE_I2C: + ret = mlx5_fpga_mem_read_i2c(fdev, size, addr, buf); + if (ret) + return ret; + break; + default: + mlx5_fpga_warn(fdev, "Unexpected read access_type %u\n", + access_type); + return -EACCES; + } + + return size; +} +EXPORT_SYMBOL(mlx5_fpga_mem_read); + +int mlx5_fpga_mem_write(struct mlx5_fpga_device *fdev, size_t size, u64 addr, + void *buf, enum mlx5_fpga_access_type access_type) +{ + int ret; + + switch (access_type) { + case MLX5_FPGA_ACCESS_TYPE_I2C: + ret = mlx5_fpga_mem_write_i2c(fdev, size, addr, buf); + if (ret) + return ret; + break; + default: + mlx5_fpga_warn(fdev, "Unexpected write access_type %u\n", + access_type); + return -EACCES; + } + + return size; +} +EXPORT_SYMBOL(mlx5_fpga_mem_write); + +int mlx5_fpga_get_sbu_caps(struct mlx5_fpga_device *fdev, int size, void *buf) +{ + return mlx5_fpga_sbu_caps(fdev->mdev, buf, size); +} +EXPORT_SYMBOL(mlx5_fpga_get_sbu_caps); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.h new file mode 100644 index 000000000000..baa537e54a49 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/sdk.h @@ -0,0 +1,204 @@ +/* + * Copyright (c) 2017 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef MLX5_FPGA_SDK_H +#define MLX5_FPGA_SDK_H + +#include <linux/types.h> +#include <linux/dma-direction.h> + +/** + * DOC: Innova SDK + * This header defines the in-kernel API for Innova FPGA client drivers. + */ + +enum mlx5_fpga_access_type { + MLX5_FPGA_ACCESS_TYPE_I2C = 0x0, + MLX5_FPGA_ACCESS_TYPE_DONTCARE = 0x0, +}; + +struct mlx5_fpga_conn; +struct mlx5_fpga_device; + +/** + * struct mlx5_fpga_dma_entry - A scatter-gather DMA entry + */ +struct mlx5_fpga_dma_entry { + /** @data: Virtual address pointer to the data */ + void *data; + /** @size: Size in bytes of the data */ + unsigned int size; + /** @dma_addr: Private member. Physical DMA-mapped address of the data */ + dma_addr_t dma_addr; +}; + +/** + * struct mlx5_fpga_dma_buf - A packet buffer + * May contain up to 2 scatter-gather data entries + */ +struct mlx5_fpga_dma_buf { + /** @dma_dir: DMA direction */ + enum dma_data_direction dma_dir; + /** @sg: Scatter-gather entries pointing to the data in memory */ + struct mlx5_fpga_dma_entry sg[2]; + /** @list: Item in SQ backlog, for TX packets */ + struct list_head list; + /** + * @complete: Completion routine, for TX packets + * @conn: FPGA Connection this packet was sent to + * @fdev: FPGA device this packet was sent to + * @buf: The packet buffer + * @status: 0 if successful, or an error code otherwise + */ + void (*complete)(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_device *fdev, + struct mlx5_fpga_dma_buf *buf, u8 status); +}; + +/** + * struct mlx5_fpga_conn_attr - FPGA connection attributes + * Describes the attributes of a connection + */ +struct mlx5_fpga_conn_attr { + /** @tx_size: Size of connection TX queue, in packets */ + unsigned int tx_size; + /** @rx_size: Size of connection RX queue, in packets */ + unsigned int rx_size; + /** + * @recv_cb: Callback function which is called for received packets + * @cb_arg: The value provided in mlx5_fpga_conn_attr.cb_arg + * @buf: A buffer containing a received packet + * + * buf is guaranteed to only contain a single scatter-gather entry. + * The size of the actual packet received is specified in buf.sg[0].size + * When this callback returns, the packet buffer may be re-used for + * subsequent receives. + */ + void (*recv_cb)(void *cb_arg, struct mlx5_fpga_dma_buf *buf); + void *cb_arg; +}; + +/** + * mlx5_fpga_sbu_conn_create() - Initialize a new FPGA SBU connection + * @fdev: The FPGA device + * @attr: Attributes of the new connection + * + * Sets up a new FPGA SBU connection with the specified attributes. + * The receive callback function may be called for incoming messages even + * before this function returns. + * + * The caller must eventually destroy the connection by calling + * mlx5_fpga_sbu_conn_destroy. + * + * Return: A new connection, or ERR_PTR() error value otherwise. + */ +struct mlx5_fpga_conn * +mlx5_fpga_sbu_conn_create(struct mlx5_fpga_device *fdev, + struct mlx5_fpga_conn_attr *attr); + +/** + * mlx5_fpga_sbu_conn_destroy() - Destroy an FPGA SBU connection + * @conn: The FPGA SBU connection to destroy + * + * Cleans up an FPGA SBU connection which was previously created with + * mlx5_fpga_sbu_conn_create. + */ +void mlx5_fpga_sbu_conn_destroy(struct mlx5_fpga_conn *conn); + +/** + * mlx5_fpga_sbu_conn_sendmsg() - Queue the transmission of a packet + * @fdev: An FPGA SBU connection + * @buf: The packet buffer + * + * Queues a packet for transmission over an FPGA SBU connection. + * The buffer should not be modified or freed until completion. + * Upon completion, the buf's complete() callback is invoked, indicating the + * success or error status of the transmission. + * + * Return: 0 if successful, or an error value otherwise. + */ +int mlx5_fpga_sbu_conn_sendmsg(struct mlx5_fpga_conn *conn, + struct mlx5_fpga_dma_buf *buf); + +/** + * mlx5_fpga_mem_read() - Read from FPGA memory address space + * @fdev: The FPGA device + * @size: Size of chunk to read, in bytes + * @addr: Starting address to read from, in FPGA address space + * @buf: Buffer to read into + * @access_type: Method for reading + * + * Reads from the specified address into the specified buffer. + * The address may point to configuration space or to DDR. + * Large reads may be performed internally as several non-atomic operations. + * This function may sleep, so should not be called from atomic contexts. + * + * Return: 0 if successful, or an error value otherwise. + */ +int mlx5_fpga_mem_read(struct mlx5_fpga_device *fdev, size_t size, u64 addr, + void *buf, enum mlx5_fpga_access_type access_type); + +/** + * mlx5_fpga_mem_write() - Write to FPGA memory address space + * @fdev: The FPGA device + * @size: Size of chunk to write, in bytes + * @addr: Starting address to write to, in FPGA address space + * @buf: Buffer which contains data to write + * @access_type: Method for writing + * + * Writes the specified buffer data to FPGA memory at the specified address. + * The address may point to configuration space or to DDR. + * Large writes may be performed internally as several non-atomic operations. + * This function may sleep, so should not be called from atomic contexts. + * + * Return: 0 if successful, or an error value otherwise. + */ +int mlx5_fpga_mem_write(struct mlx5_fpga_device *fdev, size_t size, u64 addr, + void *buf, enum mlx5_fpga_access_type access_type); + +/** + * mlx5_fpga_get_sbu_caps() - Read the SBU capabilities + * @fdev: The FPGA device + * @size: Size of the buffer to read into + * @buf: Buffer to read the capabilities into + * + * Reads the FPGA SBU capabilities into the specified buffer. + * The format of the capabilities buffer is SBU-dependent. + * + * Return: 0 if successful + * -EINVAL if the buffer is not large enough to contain SBU caps + * or any other error value otherwise. + */ +int mlx5_fpga_get_sbu_caps(struct mlx5_fpga_device *fdev, int size, void *buf); + +#endif /* MLX5_FPGA_SDK_H */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c index 19e3d2fc2099..e750f07793b8 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c @@ -40,28 +40,25 @@ #include "eswitch.h" int mlx5_cmd_update_root_ft(struct mlx5_core_dev *dev, - struct mlx5_flow_table *ft) + struct mlx5_flow_table *ft, u32 underlay_qpn) { u32 in[MLX5_ST_SZ_DW(set_flow_table_root_in)] = {0}; u32 out[MLX5_ST_SZ_DW(set_flow_table_root_out)] = {0}; if ((MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_IB) && - ft->underlay_qpn == 0) + underlay_qpn == 0) return 0; MLX5_SET(set_flow_table_root_in, in, opcode, MLX5_CMD_OP_SET_FLOW_TABLE_ROOT); MLX5_SET(set_flow_table_root_in, in, table_type, ft->type); MLX5_SET(set_flow_table_root_in, in, table_id, ft->id); + MLX5_SET(set_flow_table_root_in, in, underlay_qpn, underlay_qpn); if (ft->vport) { MLX5_SET(set_flow_table_root_in, in, vport_number, ft->vport); MLX5_SET(set_flow_table_root_in, in, other_vport, 1); } - if ((MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_IB) && - ft->underlay_qpn != 0) - MLX5_SET(set_flow_table_root_in, in, underlay_qpn, ft->underlay_qpn); - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); } @@ -81,28 +78,33 @@ int mlx5_cmd_create_flow_table(struct mlx5_core_dev *dev, MLX5_CMD_OP_CREATE_FLOW_TABLE); MLX5_SET(create_flow_table_in, in, table_type, type); - MLX5_SET(create_flow_table_in, in, level, level); - MLX5_SET(create_flow_table_in, in, log_size, log_size); + MLX5_SET(create_flow_table_in, in, flow_table_context.level, level); + MLX5_SET(create_flow_table_in, in, flow_table_context.log_size, log_size); if (vport) { MLX5_SET(create_flow_table_in, in, vport_number, vport); MLX5_SET(create_flow_table_in, in, other_vport, 1); } - MLX5_SET(create_flow_table_in, in, decap_en, en_encap_decap); - MLX5_SET(create_flow_table_in, in, encap_en, en_encap_decap); + MLX5_SET(create_flow_table_in, in, flow_table_context.decap_en, + en_encap_decap); + MLX5_SET(create_flow_table_in, in, flow_table_context.encap_en, + en_encap_decap); switch (op_mod) { case FS_FT_OP_MOD_NORMAL: if (next_ft) { - MLX5_SET(create_flow_table_in, in, table_miss_mode, 1); - MLX5_SET(create_flow_table_in, in, table_miss_id, next_ft->id); + MLX5_SET(create_flow_table_in, in, + flow_table_context.table_miss_action, 1); + MLX5_SET(create_flow_table_in, in, + flow_table_context.table_miss_id, next_ft->id); } break; case FS_FT_OP_MOD_LAG_DEMUX: MLX5_SET(create_flow_table_in, in, op_mod, 0x1); if (next_ft) - MLX5_SET(create_flow_table_in, in, lag_master_next_table_id, + MLX5_SET(create_flow_table_in, in, + flow_table_context.lag_master_next_table_id, next_ft->id); break; } @@ -149,10 +151,10 @@ int mlx5_cmd_modify_flow_table(struct mlx5_core_dev *dev, MLX5_MODIFY_FLOW_TABLE_LAG_NEXT_TABLE_ID); if (next_ft) { MLX5_SET(modify_flow_table_in, in, - lag_master_next_table_id, next_ft->id); + flow_table_context.lag_master_next_table_id, next_ft->id); } else { MLX5_SET(modify_flow_table_in, in, - lag_master_next_table_id, 0); + flow_table_context.lag_master_next_table_id, 0); } } else { if (ft->vport) { @@ -163,11 +165,14 @@ int mlx5_cmd_modify_flow_table(struct mlx5_core_dev *dev, MLX5_SET(modify_flow_table_in, in, modify_field_select, MLX5_MODIFY_FLOW_TABLE_MISS_TABLE_ID); if (next_ft) { - MLX5_SET(modify_flow_table_in, in, table_miss_mode, 1); - MLX5_SET(modify_flow_table_in, in, table_miss_id, + MLX5_SET(modify_flow_table_in, in, + flow_table_context.table_miss_action, 1); + MLX5_SET(modify_flow_table_in, in, + flow_table_context.table_miss_id, next_ft->id); } else { - MLX5_SET(modify_flow_table_in, in, table_miss_mode, 0); + MLX5_SET(modify_flow_table_in, in, + flow_table_context.table_miss_action, 0); } } @@ -235,11 +240,9 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev, u32 *in; int err; - in = mlx5_vzalloc(inlen); - if (!in) { - mlx5_core_warn(dev, "failed to allocate inbox\n"); + in = kvzalloc(inlen, GFP_KERNEL); + if (!in) return -ENOMEM; - } MLX5_SET(set_fte_in, in, opcode, MLX5_CMD_OP_SET_FLOW_TABLE_ENTRY); MLX5_SET(set_fte_in, in, op_mod, opmod); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h index 8fad80688536..0f98a7cf4877 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h @@ -71,7 +71,8 @@ int mlx5_cmd_delete_fte(struct mlx5_core_dev *dev, unsigned int index); int mlx5_cmd_update_root_ft(struct mlx5_core_dev *dev, - struct mlx5_flow_table *ft); + struct mlx5_flow_table *ft, + u32 underlay_qpn); int mlx5_cmd_fc_alloc(struct mlx5_core_dev *dev, u16 *id); int mlx5_cmd_fc_free(struct mlx5_core_dev *dev, u16 id); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c index b8a176503d38..e8690fe46bf2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c @@ -104,6 +104,7 @@ struct node_caps { size_t arr_sz; long *caps; }; + static struct init_tree_node { enum fs_node_type type; struct init_tree_node *children; @@ -376,11 +377,9 @@ static void del_rule(struct fs_node *node) int err; bool update_fte = false; - match_value = mlx5_vzalloc(match_len); - if (!match_value) { - mlx5_core_warn(dev, "failed to allocate inbox\n"); + match_value = kvzalloc(match_len, GFP_KERNEL); + if (!match_value) return; - } fs_get_obj(rule, node); fs_get_obj(fte, rule->node.parent); @@ -650,7 +649,7 @@ static int update_root_ft_create(struct mlx5_flow_table *ft, struct fs_prio if (ft->level >= min_level) return 0; - err = mlx5_cmd_update_root_ft(root->dev, ft); + err = mlx5_cmd_update_root_ft(root->dev, ft, root->underlay_qpn); if (err) mlx5_core_warn(root->dev, "Update root flow table of id=%u failed\n", ft->id); @@ -818,8 +817,6 @@ static struct mlx5_flow_table *__mlx5_create_flow_table(struct mlx5_flow_namespa goto unlock_root; } - ft->underlay_qpn = ft_attr->underlay_qpn; - tree_init_node(&ft->node, 1, del_flow_table); log_table_sz = ft->max_fte ? ilog2(ft->max_fte) : 0; next_ft = find_next_chained_ft(fs_prio); @@ -864,7 +861,7 @@ struct mlx5_flow_table *mlx5_create_vport_flow_table(struct mlx5_flow_namespace ft_attr.level = level; ft_attr.prio = prio; - return __mlx5_create_flow_table(ns, &ft_attr, FS_FT_OP_MOD_NORMAL, 0); + return __mlx5_create_flow_table(ns, &ft_attr, FS_FT_OP_MOD_NORMAL, vport); } struct mlx5_flow_table* @@ -1159,7 +1156,7 @@ static struct mlx5_flow_group *create_autogroup(struct mlx5_flow_table *ft, if (!ft->autogroup.active) return ERR_PTR(-ENOENT); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return ERR_PTR(-ENOMEM); @@ -1489,7 +1486,8 @@ static int update_root_ft_destroy(struct mlx5_flow_table *ft) new_root_ft = find_next_ft(ft); if (new_root_ft) { - int err = mlx5_cmd_update_root_ft(root->dev, new_root_ft); + int err = mlx5_cmd_update_root_ft(root->dev, new_root_ft, + root->underlay_qpn); if (err) { mlx5_core_warn(root->dev, "Update root flow table of id=%u failed\n", @@ -1778,7 +1776,7 @@ static struct mlx5_flow_root_namespace *create_root_ns(struct mlx5_flow_steering struct mlx5_flow_namespace *ns; /* Create the root namespace */ - root_ns = mlx5_vzalloc(sizeof(*root_ns)); + root_ns = kvzalloc(sizeof(*root_ns), GFP_KERNEL); if (!root_ns) return NULL; @@ -1861,7 +1859,6 @@ static int create_anchor_flow_table(struct mlx5_flow_steering *steering) static int init_root_ns(struct mlx5_flow_steering *steering) { - steering->root_ns = create_root_ns(steering, FS_FT_NIC_RX); if (!steering->root_ns) goto cleanup; @@ -2062,3 +2059,21 @@ err: mlx5_cleanup_fs(dev); return err; } + +int mlx5_fs_add_rx_underlay_qpn(struct mlx5_core_dev *dev, u32 underlay_qpn) +{ + struct mlx5_flow_root_namespace *root = dev->priv.steering->root_ns; + + root->underlay_qpn = underlay_qpn; + return 0; +} +EXPORT_SYMBOL(mlx5_fs_add_rx_underlay_qpn); + +int mlx5_fs_remove_rx_underlay_qpn(struct mlx5_core_dev *dev, u32 underlay_qpn) +{ + struct mlx5_flow_root_namespace *root = dev->priv.steering->root_ns; + + root->underlay_qpn = 0; + return 0; +} +EXPORT_SYMBOL(mlx5_fs_remove_rx_underlay_qpn); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h index 81eafc7b9dd9..990acee6fb09 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h @@ -118,7 +118,6 @@ struct mlx5_flow_table { /* FWD rules that point on this flow table */ struct list_head fwd_rules; u32 flags; - u32 underlay_qpn; }; struct mlx5_fc_cache { @@ -195,6 +194,7 @@ struct mlx5_flow_root_namespace { struct mlx5_flow_table *root_ft; /* Should be held when chaining flow tables */ struct mutex chain_lock; + u32 underlay_qpn; }; int mlx5_init_fc_stats(struct mlx5_core_dev *dev); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fw.c b/drivers/net/ethernet/mellanox/mlx5/core/fw.c index 1bc14d0fded8..fa33d59ab485 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fw.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fw.c @@ -34,6 +34,7 @@ #include <linux/mlx5/cmd.h> #include <linux/module.h> #include "mlx5_core.h" +#include "../../mlxfw/mlxfw.h" static int mlx5_cmd_query_adapter(struct mlx5_core_dev *dev, u32 *out, int outlen) @@ -195,3 +196,298 @@ int mlx5_cmd_teardown_hca(struct mlx5_core_dev *dev) MLX5_SET(teardown_hca_in, in, opcode, MLX5_CMD_OP_TEARDOWN_HCA); return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); } + +int mlx5_cmd_force_teardown_hca(struct mlx5_core_dev *dev) +{ + u32 out[MLX5_ST_SZ_DW(teardown_hca_out)] = {0}; + u32 in[MLX5_ST_SZ_DW(teardown_hca_in)] = {0}; + int force_state; + int ret; + + if (!MLX5_CAP_GEN(dev, force_teardown)) { + mlx5_core_dbg(dev, "force teardown is not supported in the firmware\n"); + return -EOPNOTSUPP; + } + + MLX5_SET(teardown_hca_in, in, opcode, MLX5_CMD_OP_TEARDOWN_HCA); + MLX5_SET(teardown_hca_in, in, profile, MLX5_TEARDOWN_HCA_IN_PROFILE_FORCE_CLOSE); + + ret = mlx5_cmd_exec_polling(dev, in, sizeof(in), out, sizeof(out)); + if (ret) + return ret; + + force_state = MLX5_GET(teardown_hca_out, out, force_state); + if (force_state == MLX5_TEARDOWN_HCA_OUT_FORCE_STATE_FAIL) { + mlx5_core_err(dev, "teardown with force mode failed\n"); + return -EIO; + } + + return 0; +} + +enum mlxsw_reg_mcc_instruction { + MLX5_REG_MCC_INSTRUCTION_LOCK_UPDATE_HANDLE = 0x01, + MLX5_REG_MCC_INSTRUCTION_RELEASE_UPDATE_HANDLE = 0x02, + MLX5_REG_MCC_INSTRUCTION_UPDATE_COMPONENT = 0x03, + MLX5_REG_MCC_INSTRUCTION_VERIFY_COMPONENT = 0x04, + MLX5_REG_MCC_INSTRUCTION_ACTIVATE = 0x06, + MLX5_REG_MCC_INSTRUCTION_CANCEL = 0x08, +}; + +static int mlx5_reg_mcc_set(struct mlx5_core_dev *dev, + enum mlxsw_reg_mcc_instruction instr, + u16 component_index, u32 update_handle, + u32 component_size) +{ + u32 out[MLX5_ST_SZ_DW(mcc_reg)]; + u32 in[MLX5_ST_SZ_DW(mcc_reg)]; + + memset(in, 0, sizeof(in)); + + MLX5_SET(mcc_reg, in, instruction, instr); + MLX5_SET(mcc_reg, in, component_index, component_index); + MLX5_SET(mcc_reg, in, update_handle, update_handle); + MLX5_SET(mcc_reg, in, component_size, component_size); + + return mlx5_core_access_reg(dev, in, sizeof(in), out, + sizeof(out), MLX5_REG_MCC, 0, 1); +} + +static int mlx5_reg_mcc_query(struct mlx5_core_dev *dev, + u32 *update_handle, u8 *error_code, + u8 *control_state) +{ + u32 out[MLX5_ST_SZ_DW(mcc_reg)]; + u32 in[MLX5_ST_SZ_DW(mcc_reg)]; + int err; + + memset(in, 0, sizeof(in)); + memset(out, 0, sizeof(out)); + MLX5_SET(mcc_reg, in, update_handle, *update_handle); + + err = mlx5_core_access_reg(dev, in, sizeof(in), out, + sizeof(out), MLX5_REG_MCC, 0, 0); + if (err) + goto out; + + *update_handle = MLX5_GET(mcc_reg, out, update_handle); + *error_code = MLX5_GET(mcc_reg, out, error_code); + *control_state = MLX5_GET(mcc_reg, out, control_state); + +out: + return err; +} + +static int mlx5_reg_mcda_set(struct mlx5_core_dev *dev, + u32 update_handle, + u32 offset, u16 size, + u8 *data) +{ + int err, in_size = MLX5_ST_SZ_BYTES(mcda_reg) + size; + u32 out[MLX5_ST_SZ_DW(mcda_reg)]; + int i, j, dw_size = size >> 2; + __be32 data_element; + u32 *in; + + in = kzalloc(in_size, GFP_KERNEL); + if (!in) + return -ENOMEM; + + MLX5_SET(mcda_reg, in, update_handle, update_handle); + MLX5_SET(mcda_reg, in, offset, offset); + MLX5_SET(mcda_reg, in, size, size); + + for (i = 0; i < dw_size; i++) { + j = i * 4; + data_element = htonl(*(u32 *)&data[j]); + memcpy(MLX5_ADDR_OF(mcda_reg, in, data) + j, &data_element, 4); + } + + err = mlx5_core_access_reg(dev, in, in_size, out, + sizeof(out), MLX5_REG_MCDA, 0, 1); + kfree(in); + return err; +} + +static int mlx5_reg_mcqi_query(struct mlx5_core_dev *dev, + u16 component_index, + u32 *max_component_size, + u8 *log_mcda_word_size, + u16 *mcda_max_write_size) +{ + u32 out[MLX5_ST_SZ_DW(mcqi_reg) + MLX5_ST_SZ_DW(mcqi_cap)]; + int offset = MLX5_ST_SZ_DW(mcqi_reg); + u32 in[MLX5_ST_SZ_DW(mcqi_reg)]; + int err; + + memset(in, 0, sizeof(in)); + memset(out, 0, sizeof(out)); + + MLX5_SET(mcqi_reg, in, component_index, component_index); + MLX5_SET(mcqi_reg, in, data_size, MLX5_ST_SZ_BYTES(mcqi_cap)); + + err = mlx5_core_access_reg(dev, in, sizeof(in), out, + sizeof(out), MLX5_REG_MCQI, 0, 0); + if (err) + goto out; + + *max_component_size = MLX5_GET(mcqi_cap, out + offset, max_component_size); + *log_mcda_word_size = MLX5_GET(mcqi_cap, out + offset, log_mcda_word_size); + *mcda_max_write_size = MLX5_GET(mcqi_cap, out + offset, mcda_max_write_size); + +out: + return err; +} + +struct mlx5_mlxfw_dev { + struct mlxfw_dev mlxfw_dev; + struct mlx5_core_dev *mlx5_core_dev; +}; + +static int mlx5_component_query(struct mlxfw_dev *mlxfw_dev, + u16 component_index, u32 *p_max_size, + u8 *p_align_bits, u16 *p_max_write_size) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + + return mlx5_reg_mcqi_query(dev, component_index, p_max_size, + p_align_bits, p_max_write_size); +} + +static int mlx5_fsm_lock(struct mlxfw_dev *mlxfw_dev, u32 *fwhandle) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + u8 control_state, error_code; + int err; + + *fwhandle = 0; + err = mlx5_reg_mcc_query(dev, fwhandle, &error_code, &control_state); + if (err) + return err; + + if (control_state != MLXFW_FSM_STATE_IDLE) + return -EBUSY; + + return mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_LOCK_UPDATE_HANDLE, + 0, *fwhandle, 0); +} + +static int mlx5_fsm_component_update(struct mlxfw_dev *mlxfw_dev, u32 fwhandle, + u16 component_index, u32 component_size) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + + return mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_UPDATE_COMPONENT, + component_index, fwhandle, component_size); +} + +static int mlx5_fsm_block_download(struct mlxfw_dev *mlxfw_dev, u32 fwhandle, + u8 *data, u16 size, u32 offset) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + + return mlx5_reg_mcda_set(dev, fwhandle, offset, size, data); +} + +static int mlx5_fsm_component_verify(struct mlxfw_dev *mlxfw_dev, u32 fwhandle, + u16 component_index) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + + return mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_VERIFY_COMPONENT, + component_index, fwhandle, 0); +} + +static int mlx5_fsm_activate(struct mlxfw_dev *mlxfw_dev, u32 fwhandle) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + + return mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_ACTIVATE, 0, + fwhandle, 0); +} + +static int mlx5_fsm_query_state(struct mlxfw_dev *mlxfw_dev, u32 fwhandle, + enum mlxfw_fsm_state *fsm_state, + enum mlxfw_fsm_state_err *fsm_state_err) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + u8 control_state, error_code; + int err; + + err = mlx5_reg_mcc_query(dev, &fwhandle, &error_code, &control_state); + if (err) + return err; + + *fsm_state = control_state; + *fsm_state_err = min_t(enum mlxfw_fsm_state_err, error_code, + MLXFW_FSM_STATE_ERR_MAX); + return 0; +} + +static void mlx5_fsm_cancel(struct mlxfw_dev *mlxfw_dev, u32 fwhandle) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + + mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_CANCEL, 0, fwhandle, 0); +} + +static void mlx5_fsm_release(struct mlxfw_dev *mlxfw_dev, u32 fwhandle) +{ + struct mlx5_mlxfw_dev *mlx5_mlxfw_dev = + container_of(mlxfw_dev, struct mlx5_mlxfw_dev, mlxfw_dev); + struct mlx5_core_dev *dev = mlx5_mlxfw_dev->mlx5_core_dev; + + mlx5_reg_mcc_set(dev, MLX5_REG_MCC_INSTRUCTION_RELEASE_UPDATE_HANDLE, 0, + fwhandle, 0); +} + +static const struct mlxfw_dev_ops mlx5_mlxfw_dev_ops = { + .component_query = mlx5_component_query, + .fsm_lock = mlx5_fsm_lock, + .fsm_component_update = mlx5_fsm_component_update, + .fsm_block_download = mlx5_fsm_block_download, + .fsm_component_verify = mlx5_fsm_component_verify, + .fsm_activate = mlx5_fsm_activate, + .fsm_query_state = mlx5_fsm_query_state, + .fsm_cancel = mlx5_fsm_cancel, + .fsm_release = mlx5_fsm_release +}; + +int mlx5_firmware_flash(struct mlx5_core_dev *dev, + const struct firmware *firmware) +{ + struct mlx5_mlxfw_dev mlx5_mlxfw_dev = { + .mlxfw_dev = { + .ops = &mlx5_mlxfw_dev_ops, + .psid = dev->board_id, + .psid_size = strlen(dev->board_id), + }, + .mlx5_core_dev = dev + }; + + if (!MLX5_CAP_GEN(dev, mcam_reg) || + !MLX5_CAP_MCAM_REG(dev, mcqi) || + !MLX5_CAP_MCAM_REG(dev, mcc) || + !MLX5_CAP_MCAM_REG(dev, mcda)) { + pr_info("%s flashing isn't supported by the running FW\n", __func__); + return -EOPNOTSUPP; + } + + return mlxfw_firmware_flash(&mlx5_mlxfw_dev.mlxfw_dev, firmware); +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c index d0515391d33b..4b6b03d6297f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/health.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c @@ -67,6 +67,7 @@ enum { enum { MLX5_DROP_NEW_HEALTH_WORK, + MLX5_DROP_NEW_RECOVERY_WORK, }; static u8 get_nic_state(struct mlx5_core_dev *dev) @@ -90,7 +91,7 @@ static void trigger_cmd_completions(struct mlx5_core_dev *dev) spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags); mlx5_core_dbg(dev, "vector 0x%llx\n", vector); - mlx5_cmd_comp_handler(dev, vector); + mlx5_cmd_comp_handler(dev, vector, true); return; no_trig: @@ -111,14 +112,14 @@ static int in_fatal(struct mlx5_core_dev *dev) return 0; } -void mlx5_enter_error_state(struct mlx5_core_dev *dev) +void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force) { mutex_lock(&dev->intf_state_mutex); if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) goto unlock; mlx5_core_err(dev, "start\n"); - if (pci_channel_offline(dev->pdev) || in_fatal(dev)) { + if (pci_channel_offline(dev->pdev) || in_fatal(dev) || force) { dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR; trigger_cmd_completions(dev); } @@ -185,6 +186,7 @@ static void health_care(struct work_struct *work) struct mlx5_core_health *health; struct mlx5_core_dev *dev; struct mlx5_priv *priv; + unsigned long flags; health = container_of(work, struct mlx5_core_health, work); priv = container_of(health, struct mlx5_priv, health); @@ -192,13 +194,13 @@ static void health_care(struct work_struct *work) mlx5_core_warn(dev, "handling bad device here\n"); mlx5_handle_bad_state(dev); - spin_lock(&health->wq_lock); - if (!test_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags)) + spin_lock_irqsave(&health->wq_lock, flags); + if (!test_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags)) schedule_delayed_work(&health->recover_work, recover_delay); else dev_err(&dev->pdev->dev, "new health works are not permitted at this stage\n"); - spin_unlock(&health->wq_lock); + spin_unlock_irqrestore(&health->wq_lock, flags); } static const char *hsynd_str(u8 synd) @@ -269,16 +271,28 @@ static unsigned long get_next_poll_jiffies(void) return next; } +void mlx5_trigger_health_work(struct mlx5_core_dev *dev) +{ + struct mlx5_core_health *health = &dev->priv.health; + unsigned long flags; + + spin_lock_irqsave(&health->wq_lock, flags); + if (!test_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags)) + queue_work(health->wq, &health->work); + else + dev_err(&dev->pdev->dev, + "new health works are not permitted at this stage\n"); + spin_unlock_irqrestore(&health->wq_lock, flags); +} + static void poll_health(unsigned long data) { struct mlx5_core_dev *dev = (struct mlx5_core_dev *)data; struct mlx5_core_health *health = &dev->priv.health; u32 count; - if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) { - mod_timer(&health->timer, get_next_poll_jiffies()); - return; - } + if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) + goto out; count = ioread32be(health->health_counter); if (count == health->prev) @@ -290,21 +304,16 @@ static void poll_health(unsigned long data) if (health->miss_counter == MAX_MISSES) { dev_err(&dev->pdev->dev, "device's health compromised - reached miss count\n"); print_health_info(dev); - } else { - mod_timer(&health->timer, get_next_poll_jiffies()); } if (in_fatal(dev) && !health->sick) { health->sick = true; print_health_info(dev); - spin_lock(&health->wq_lock); - if (!test_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags)) - queue_work(health->wq, &health->work); - else - dev_err(&dev->pdev->dev, - "new health works are not permitted at this stage\n"); - spin_unlock(&health->wq_lock); + mlx5_trigger_health_work(dev); } + +out: + mod_timer(&health->timer, get_next_poll_jiffies()); } void mlx5_start_health_poll(struct mlx5_core_dev *dev) @@ -314,6 +323,7 @@ void mlx5_start_health_poll(struct mlx5_core_dev *dev) init_timer(&health->timer); health->sick = 0; clear_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags); + clear_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags); health->health = &dev->iseg->health; health->health_counter = &dev->iseg->health_counter; @@ -333,14 +343,26 @@ void mlx5_stop_health_poll(struct mlx5_core_dev *dev) void mlx5_drain_health_wq(struct mlx5_core_dev *dev) { struct mlx5_core_health *health = &dev->priv.health; + unsigned long flags; - spin_lock(&health->wq_lock); + spin_lock_irqsave(&health->wq_lock, flags); set_bit(MLX5_DROP_NEW_HEALTH_WORK, &health->flags); - spin_unlock(&health->wq_lock); + set_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags); + spin_unlock_irqrestore(&health->wq_lock, flags); cancel_delayed_work_sync(&health->recover_work); cancel_work_sync(&health->work); } +void mlx5_drain_health_recovery(struct mlx5_core_dev *dev) +{ + struct mlx5_core_health *health = &dev->priv.health; + + spin_lock(&health->wq_lock); + set_bit(MLX5_DROP_NEW_RECOVERY_WORK, &health->flags); + spin_unlock(&health->wq_lock); + cancel_delayed_work_sync(&dev->priv.health.recover_work); +} + void mlx5_health_cleanup(struct mlx5_core_dev *dev) { struct mlx5_core_health *health = &dev->priv.health; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/Makefile new file mode 100644 index 000000000000..d8e17110f25d --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/Makefile @@ -0,0 +1 @@ +subdir-ccflags-y += -I$(src)/.. diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c new file mode 100644 index 000000000000..eb04e97d8765 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ethtool.c @@ -0,0 +1,145 @@ +/* + * Copyright (c) 2017, Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include "en.h" +#include "ipoib.h" + +static void mlx5i_get_drvinfo(struct net_device *dev, + struct ethtool_drvinfo *drvinfo) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + mlx5e_ethtool_get_drvinfo(priv, drvinfo); +} + +static void mlx5i_get_strings(struct net_device *dev, + uint32_t stringset, uint8_t *data) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + mlx5e_ethtool_get_strings(priv, stringset, data); +} + +static int mlx5i_get_sset_count(struct net_device *dev, int sset) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + return mlx5e_ethtool_get_sset_count(priv, sset); +} + +static void mlx5i_get_ethtool_stats(struct net_device *dev, + struct ethtool_stats *stats, + u64 *data) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + mlx5e_ethtool_get_ethtool_stats(priv, stats, data); +} + +static int mlx5i_set_ringparam(struct net_device *dev, + struct ethtool_ringparam *param) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + return mlx5e_ethtool_set_ringparam(priv, param); +} + +static void mlx5i_get_ringparam(struct net_device *dev, + struct ethtool_ringparam *param) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + mlx5e_ethtool_get_ringparam(priv, param); +} + +static int mlx5i_set_channels(struct net_device *dev, + struct ethtool_channels *ch) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + return mlx5e_ethtool_set_channels(priv, ch); +} + +static void mlx5i_get_channels(struct net_device *dev, + struct ethtool_channels *ch) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + mlx5e_ethtool_get_channels(priv, ch); +} + +static int mlx5i_set_coalesce(struct net_device *netdev, + struct ethtool_coalesce *coal) +{ + struct mlx5e_priv *priv = mlx5i_epriv(netdev); + + return mlx5e_ethtool_set_coalesce(priv, coal); +} + +static int mlx5i_get_coalesce(struct net_device *netdev, + struct ethtool_coalesce *coal) +{ + struct mlx5e_priv *priv = mlx5i_epriv(netdev); + + return mlx5e_ethtool_get_coalesce(priv, coal); +} + +static int mlx5i_get_ts_info(struct net_device *netdev, + struct ethtool_ts_info *info) +{ + struct mlx5e_priv *priv = mlx5i_epriv(netdev); + + return mlx5e_ethtool_get_ts_info(priv, info); +} + +static int mlx5i_flash_device(struct net_device *netdev, + struct ethtool_flash *flash) +{ + struct mlx5e_priv *priv = mlx5i_epriv(netdev); + + return mlx5e_ethtool_flash_device(priv, flash); +} + +const struct ethtool_ops mlx5i_ethtool_ops = { + .get_drvinfo = mlx5i_get_drvinfo, + .get_strings = mlx5i_get_strings, + .get_sset_count = mlx5i_get_sset_count, + .get_ethtool_stats = mlx5i_get_ethtool_stats, + .get_ringparam = mlx5i_get_ringparam, + .set_ringparam = mlx5i_set_ringparam, + .flash_device = mlx5i_flash_device, + .get_channels = mlx5i_get_channels, + .set_channels = mlx5i_set_channels, + .get_coalesce = mlx5i_get_coalesce, + .set_coalesce = mlx5i_set_coalesce, + .get_ts_info = mlx5i_get_ts_info, +}; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c index 019c230da498..1ee5bce85901 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c @@ -36,20 +36,38 @@ #include "ipoib.h" #define IB_DEFAULT_Q_KEY 0xb1b +#define MLX5I_PARAMS_DEFAULT_LOG_RQ_SIZE 9 static int mlx5i_open(struct net_device *netdev); static int mlx5i_close(struct net_device *netdev); static int mlx5i_dev_init(struct net_device *dev); static void mlx5i_dev_cleanup(struct net_device *dev); +static int mlx5i_change_mtu(struct net_device *netdev, int new_mtu); +static int mlx5i_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd); static const struct net_device_ops mlx5i_netdev_ops = { .ndo_open = mlx5i_open, .ndo_stop = mlx5i_close, .ndo_init = mlx5i_dev_init, .ndo_uninit = mlx5i_dev_cleanup, + .ndo_change_mtu = mlx5i_change_mtu, + .ndo_do_ioctl = mlx5i_ioctl, }; /* IPoIB mlx5 netdev profile */ +static void mlx5i_build_nic_params(struct mlx5_core_dev *mdev, + struct mlx5e_params *params) +{ + /* Override RQ params as IPoIB supports only LINKED LIST RQ for now */ + mlx5e_set_rq_type_params(mdev, params, MLX5_WQ_TYPE_LINKED_LIST); + + /* RQ size in ipoib by default is 512 */ + params->log_rq_size = is_kdump_kernel() ? + MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE : + MLX5I_PARAMS_DEFAULT_LOG_RQ_SIZE; + + params->lro_en = false; +} /* Called directly after IPoIB netdevice was created to initialize SW structs */ static void mlx5i_init(struct mlx5_core_dev *mdev, @@ -59,15 +77,18 @@ static void mlx5i_init(struct mlx5_core_dev *mdev, { struct mlx5e_priv *priv = mlx5i_epriv(netdev); + /* priv init */ priv->mdev = mdev; priv->netdev = netdev; priv->profile = profile; priv->ppriv = ppriv; + priv->hard_mtu = MLX5_IB_GRH_BYTES + MLX5_IPOIB_HARD_LEN; + mutex_init(&priv->state_lock); mlx5e_build_nic_params(mdev, &priv->channels.params, profile->max_nch(mdev)); + mlx5i_build_nic_params(mdev, &priv->channels.params); - mutex_init(&priv->state_lock); - + /* netdev init */ netdev->hw_features |= NETIF_F_SG; netdev->hw_features |= NETIF_F_IP_CSUM; netdev->hw_features |= NETIF_F_IPV6_CSUM; @@ -78,6 +99,7 @@ static void mlx5i_init(struct mlx5_core_dev *mdev, netdev->hw_features |= NETIF_F_RXHASH; netdev->netdev_ops = &mlx5i_netdev_ops; + netdev->ethtool_ops = &mlx5i_ethtool_ops; } /* Called directly before IPoIB netdevice is destroyed to cleanup SW structs */ @@ -98,7 +120,7 @@ static int mlx5i_create_underlay_qp(struct mlx5_core_dev *mdev, struct mlx5_core void *qpc; inlen = MLX5_ST_SZ_BYTES(create_qp_in); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -156,6 +178,8 @@ out: static void mlx5i_destroy_underlay_qp(struct mlx5_core_dev *mdev, struct mlx5_core_qp *qp) { + mlx5_fs_remove_rx_underlay_qpn(mdev, qp->qpn); + mlx5_core_destroy_qp(mdev, qp); } @@ -170,6 +194,8 @@ static int mlx5i_init_tx(struct mlx5e_priv *priv) return err; } + mlx5_fs_add_rx_underlay_qpn(priv->mdev, ipriv->qp.qpn); + err = mlx5e_create_tis(priv->mdev, 0 /* tc */, ipriv->qp.qpn, &priv->tisn[0]); if (err) { mlx5_core_warn(priv->mdev, "create tis failed, %d\n", err); @@ -189,7 +215,6 @@ static void mlx5i_cleanup_tx(struct mlx5e_priv *priv) static int mlx5i_create_flow_steering(struct mlx5e_priv *priv) { - struct mlx5i_priv *ipriv = priv->ppriv; int err; priv->fs.ns = mlx5_get_flow_namespace(priv->mdev, @@ -205,7 +230,7 @@ static int mlx5i_create_flow_steering(struct mlx5e_priv *priv) priv->netdev->hw_features &= ~NETIF_F_NTUPLE; } - err = mlx5e_create_ttc_table(priv, ipriv->qp.qpn); + err = mlx5e_create_ttc_table(priv); if (err) { netdev_err(priv->netdev, "Failed to create ttc table, err=%d\n", err); @@ -283,6 +308,7 @@ static const struct mlx5e_profile mlx5i_nic_profile = { .disable = NULL, /* mlx5i_disable */ .update_stats = NULL, /* mlx5i_update_stats */ .max_nch = mlx5e_get_max_num_channels, + .update_carrier = NULL, /* no HW update in IB link */ .rx_handlers.handle_rx_cqe = mlx5i_handle_rx_cqe, .rx_handlers.handle_rx_cqe_mpwqe = NULL, /* Not supported */ .max_tc = MLX5I_MAX_NUM_TC, @@ -290,6 +316,35 @@ static const struct mlx5e_profile mlx5i_nic_profile = { /* mlx5i netdev NDos */ +static int mlx5i_change_mtu(struct net_device *netdev, int new_mtu) +{ + struct mlx5e_priv *priv = mlx5i_epriv(netdev); + struct mlx5e_channels new_channels = {}; + int curr_mtu; + int err = 0; + + mutex_lock(&priv->state_lock); + + curr_mtu = netdev->mtu; + netdev->mtu = new_mtu; + + if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) + goto out; + + new_channels.params = priv->channels.params; + err = mlx5e_open_channels(priv, &new_channels); + if (err) { + netdev->mtu = curr_mtu; + goto out; + } + + mlx5e_switch_priv_channels(priv, &new_channels, NULL); + +out: + mutex_unlock(&priv->state_lock); + return err; +} + static int mlx5i_dev_init(struct net_device *dev) { struct mlx5e_priv *priv = mlx5i_epriv(dev); @@ -303,6 +358,20 @@ static int mlx5i_dev_init(struct net_device *dev) return 0; } +static int mlx5i_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) +{ + struct mlx5e_priv *priv = mlx5i_epriv(dev); + + switch (cmd) { + case SIOCSHWTSTAMP: + return mlx5e_hwstamp_set(priv, ifr); + case SIOCGHWTSTAMP: + return mlx5e_hwstamp_get(priv, ifr); + default: + return -EOPNOTSUPP; + } +} + static void mlx5i_dev_cleanup(struct net_device *dev) { struct mlx5e_priv *priv = mlx5i_epriv(dev); @@ -329,6 +398,8 @@ static int mlx5i_open(struct net_device *netdev) mlx5e_refresh_tirs(priv, false); mlx5e_activate_priv_channels(priv); + mlx5e_timestamp_init(priv); + mutex_unlock(&priv->state_lock); return 0; @@ -352,6 +423,7 @@ static int mlx5i_close(struct net_device *netdev) clear_bit(MLX5E_STATE_OPENED, &priv->state); + mlx5e_timestamp_cleanup(priv); netif_carrier_off(priv->netdev); mlx5e_deactivate_priv_channels(priv); mlx5e_close_channels(&priv->channels); @@ -503,4 +575,3 @@ void mlx5_rdma_netdev_free(struct net_device *netdev) mlx5e_destroy_mdev_resources(priv->mdev); } EXPORT_SYMBOL(mlx5_rdma_netdev_free); - diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib.h b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h index 213191a78464..a0f405f520f7 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h @@ -38,6 +38,13 @@ #define MLX5I_MAX_NUM_TC 1 +extern const struct ethtool_ops mlx5i_ethtool_ops; + +#define MLX5_IB_GRH_BYTES 40 +#define MLX5_IPOIB_ENCAP_LEN 4 +#define MLX5_IPOIB_PSEUDO_LEN 20 +#define MLX5_IPOIB_HARD_LEN (MLX5_IPOIB_PSEUDO_LEN + MLX5_IPOIB_ENCAP_LEN) + /* ipoib rdma netdev's private data structure */ struct mlx5i_priv { struct rdma_netdev rn; /* keep this first */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index b5d5519542e8..a3a836bdcfd2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -61,6 +61,11 @@ struct mlx5_lag { struct lag_tracker tracker; struct delayed_work bond_work; struct notifier_block nb; + + /* Admin state. Allow lag only if allowed is true + * even if network conditions for lag were met + */ + bool allowed; }; /* General purpose, use for short periods of time. @@ -214,6 +219,7 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) struct lag_tracker tracker; u8 v2p_port1, v2p_port2; int i, err; + bool do_bond; if (!dev0 || !dev1) return; @@ -222,13 +228,9 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) tracker = ldev->tracker; mutex_unlock(&lag_mutex); - if (tracker.is_bonded && !mlx5_lag_is_bonded(ldev)) { - if (mlx5_sriov_is_enabled(dev0) || - mlx5_sriov_is_enabled(dev1)) { - mlx5_core_warn(dev0, "LAG is not supported with SRIOV"); - return; - } + do_bond = tracker.is_bonded && ldev->allowed; + if (do_bond && !mlx5_lag_is_bonded(ldev)) { for (i = 0; i < MLX5_MAX_PORTS; i++) mlx5_remove_dev_by_protocol(ldev->pf[i].dev, MLX5_INTERFACE_PROTOCOL_IB); @@ -237,7 +239,7 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) mlx5_add_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); mlx5_nic_vport_enable_roce(dev1); - } else if (tracker.is_bonded && mlx5_lag_is_bonded(ldev)) { + } else if (do_bond && mlx5_lag_is_bonded(ldev)) { mlx5_infer_tx_affinity_mapping(&tracker, &v2p_port1, &v2p_port2); @@ -252,7 +254,7 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) "Failed to modify LAG (%d)\n", err); } - } else if (!tracker.is_bonded && mlx5_lag_is_bonded(ldev)) { + } else if (!do_bond && mlx5_lag_is_bonded(ldev)) { mlx5_remove_dev_by_protocol(dev0, MLX5_INTERFACE_PROTOCOL_IB); mlx5_nic_vport_disable_roce(dev1); @@ -411,6 +413,15 @@ static int mlx5_lag_netdev_event(struct notifier_block *this, return NOTIFY_DONE; } +static bool mlx5_lag_check_prereq(struct mlx5_lag *ldev) +{ + if ((ldev->pf[0].dev && mlx5_sriov_is_enabled(ldev->pf[0].dev)) || + (ldev->pf[1].dev && mlx5_sriov_is_enabled(ldev->pf[1].dev))) + return false; + else + return true; +} + static struct mlx5_lag *mlx5_lag_dev_alloc(void) { struct mlx5_lag *ldev; @@ -420,6 +431,7 @@ static struct mlx5_lag *mlx5_lag_dev_alloc(void) return NULL; INIT_DELAYED_WORK(&ldev->bond_work, mlx5_do_bond_work); + ldev->allowed = mlx5_lag_check_prereq(ldev); return ldev; } @@ -444,7 +456,9 @@ static void mlx5_lag_dev_add_pf(struct mlx5_lag *ldev, ldev->tracker.netdev_state[fn].link_up = 0; ldev->tracker.netdev_state[fn].tx_enabled = 0; + ldev->allowed = mlx5_lag_check_prereq(ldev); dev->priv.lag = ldev; + mutex_unlock(&lag_mutex); } @@ -464,10 +478,10 @@ static void mlx5_lag_dev_remove_pf(struct mlx5_lag *ldev, memset(&ldev->pf[i], 0, sizeof(*ldev->pf)); dev->priv.lag = NULL; + ldev->allowed = mlx5_lag_check_prereq(ldev); mutex_unlock(&lag_mutex); } - /* Must be called with intf_mutex held */ void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev) { @@ -543,6 +557,44 @@ bool mlx5_lag_is_active(struct mlx5_core_dev *dev) } EXPORT_SYMBOL(mlx5_lag_is_active); +static int mlx5_lag_set_state(struct mlx5_core_dev *dev, bool allow) +{ + struct mlx5_lag *ldev; + int ret = 0; + bool lag_active; + + mlx5_dev_list_lock(); + + ldev = mlx5_lag_dev_get(dev); + if (!ldev) { + ret = -ENODEV; + goto unlock; + } + lag_active = mlx5_lag_is_bonded(ldev); + if (!mlx5_lag_check_prereq(ldev) && allow) { + ret = -EINVAL; + goto unlock; + } + if (ldev->allowed == allow) + goto unlock; + ldev->allowed = allow; + if ((lag_active && !allow) || allow) + mlx5_do_bond(ldev); +unlock: + mlx5_dev_list_unlock(); + return ret; +} + +int mlx5_lag_forbid(struct mlx5_core_dev *dev) +{ + return mlx5_lag_set_state(dev, false); +} + +int mlx5_lag_allow(struct mlx5_core_dev *dev) +{ + return mlx5_lag_set_state(dev, true); +} + struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev) { struct net_device *ndev = NULL; @@ -586,4 +638,3 @@ bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv) /* If bonded, we do not add an IB device for PF1. */ return false; } - diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/lib/Makefile new file mode 100644 index 000000000000..d8e17110f25d --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/Makefile @@ -0,0 +1 @@ +subdir-ccflags-y += -I$(src)/.. diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/gid.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/gid.c new file mode 100644 index 000000000000..573f59f46d41 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/gid.c @@ -0,0 +1,155 @@ +/* + * Copyright (c) 2017, Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/mlx5/driver.h> +#include <linux/etherdevice.h> +#include <linux/idr.h> +#include "mlx5_core.h" +#include "lib/mlx5.h" + +void mlx5_init_reserved_gids(struct mlx5_core_dev *dev) +{ + unsigned int tblsz = MLX5_CAP_ROCE(dev, roce_address_table_size); + + ida_init(&dev->roce.reserved_gids.ida); + dev->roce.reserved_gids.start = tblsz; + dev->roce.reserved_gids.count = 0; +} + +void mlx5_cleanup_reserved_gids(struct mlx5_core_dev *dev) +{ + WARN_ON(!ida_is_empty(&dev->roce.reserved_gids.ida)); + dev->roce.reserved_gids.start = 0; + dev->roce.reserved_gids.count = 0; + ida_destroy(&dev->roce.reserved_gids.ida); +} + +int mlx5_core_reserve_gids(struct mlx5_core_dev *dev, unsigned int count) +{ + if (test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state)) { + mlx5_core_err(dev, "Cannot reserve GIDs when interfaces are up\n"); + return -EPERM; + } + if (dev->roce.reserved_gids.start < count) { + mlx5_core_warn(dev, "GID table exhausted attempting to reserve %d more GIDs\n", + count); + return -ENOMEM; + } + if (dev->roce.reserved_gids.count + count > MLX5_MAX_RESERVED_GIDS) { + mlx5_core_warn(dev, "Unable to reserve %d more GIDs\n", count); + return -ENOMEM; + } + + dev->roce.reserved_gids.start -= count; + dev->roce.reserved_gids.count += count; + mlx5_core_dbg(dev, "Reserved %u GIDs starting at %u\n", + dev->roce.reserved_gids.count, + dev->roce.reserved_gids.start); + return 0; +} + +void mlx5_core_unreserve_gids(struct mlx5_core_dev *dev, unsigned int count) +{ + WARN(test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state), "Unreserving GIDs when interfaces are up"); + WARN(count > dev->roce.reserved_gids.count, "Unreserving %u GIDs when only %u reserved", + count, dev->roce.reserved_gids.count); + + dev->roce.reserved_gids.start += count; + dev->roce.reserved_gids.count -= count; + mlx5_core_dbg(dev, "%u GIDs starting at %u left reserved\n", + dev->roce.reserved_gids.count, + dev->roce.reserved_gids.start); +} + +int mlx5_core_reserved_gid_alloc(struct mlx5_core_dev *dev, int *gid_index) +{ + int end = dev->roce.reserved_gids.start + + dev->roce.reserved_gids.count; + int index = 0; + + index = ida_simple_get(&dev->roce.reserved_gids.ida, + dev->roce.reserved_gids.start, end, + GFP_KERNEL); + if (index < 0) + return index; + + mlx5_core_dbg(dev, "Allocating reserved GID %u\n", index); + *gid_index = index; + return 0; +} + +void mlx5_core_reserved_gid_free(struct mlx5_core_dev *dev, int gid_index) +{ + mlx5_core_dbg(dev, "Freeing reserved GID %u\n", gid_index); + ida_simple_remove(&dev->roce.reserved_gids.ida, gid_index); +} + +unsigned int mlx5_core_reserved_gids_count(struct mlx5_core_dev *dev) +{ + return dev->roce.reserved_gids.count; +} +EXPORT_SYMBOL_GPL(mlx5_core_reserved_gids_count); + +int mlx5_core_roce_gid_set(struct mlx5_core_dev *dev, unsigned int index, + u8 roce_version, u8 roce_l3_type, const u8 *gid, + const u8 *mac, bool vlan, u16 vlan_id) +{ +#define MLX5_SET_RA(p, f, v) MLX5_SET(roce_addr_layout, p, f, v) + u32 in[MLX5_ST_SZ_DW(set_roce_address_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(set_roce_address_out)] = {0}; + void *in_addr = MLX5_ADDR_OF(set_roce_address_in, in, roce_address); + char *addr_l3_addr = MLX5_ADDR_OF(roce_addr_layout, in_addr, + source_l3_address); + void *addr_mac = MLX5_ADDR_OF(roce_addr_layout, in_addr, + source_mac_47_32); + int gidsz = MLX5_FLD_SZ_BYTES(roce_addr_layout, source_l3_address); + + if (MLX5_CAP_GEN(dev, port_type) != MLX5_CAP_PORT_TYPE_ETH) + return -EINVAL; + + if (gid) { + if (vlan) { + MLX5_SET_RA(in_addr, vlan_valid, 1); + MLX5_SET_RA(in_addr, vlan_id, vlan_id); + } + + ether_addr_copy(addr_mac, mac); + MLX5_SET_RA(in_addr, roce_version, roce_version); + MLX5_SET_RA(in_addr, roce_l3_type, roce_l3_type); + memcpy(addr_l3_addr, gid, gidsz); + } + + MLX5_SET(set_roce_address_in, in, roce_address_index, index); + MLX5_SET(set_roce_address_in, in, opcode, MLX5_CMD_OP_SET_ROCE_ADDRESS); + return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); +} +EXPORT_SYMBOL(mlx5_core_roce_gid_set); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h new file mode 100644 index 000000000000..7550b1cc8c6a --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mlx5.h @@ -0,0 +1,43 @@ +/* + * Copyright (c) 2017, Mellanox Technologies, Ltd. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#ifndef __LIB_MLX5_H__ +#define __LIB_MLX5_H__ + +void mlx5_init_reserved_gids(struct mlx5_core_dev *dev); +void mlx5_cleanup_reserved_gids(struct mlx5_core_dev *dev); +int mlx5_core_reserve_gids(struct mlx5_core_dev *dev, unsigned int count); +void mlx5_core_unreserve_gids(struct mlx5_core_dev *dev, unsigned int count); +int mlx5_core_reserved_gid_alloc(struct mlx5_core_dev *dev, int *gid_index); +void mlx5_core_reserved_gid_free(struct mlx5_core_dev *dev, int gid_index); + +#endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 0c123d571b4c..c065132b956d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -56,6 +56,9 @@ #ifdef CONFIG_MLX5_CORE_EN #include "eswitch.h" #endif +#include "lib/mlx5.h" +#include "fpga/core.h" +#include "accel/ipsec.h" MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>"); MODULE_DESCRIPTION("Mellanox Connect-IB, ConnectX-4 core driver"); @@ -175,8 +178,9 @@ static struct mlx5_profile profile[] = { }, }; -#define FW_INIT_TIMEOUT_MILI 2000 -#define FW_INIT_WAIT_MS 2 +#define FW_INIT_TIMEOUT_MILI 2000 +#define FW_INIT_WAIT_MS 2 +#define FW_PRE_INIT_TIMEOUT_MILI 10000 static int wait_fw_init(struct mlx5_core_dev *dev, u32 max_wait_mili) { @@ -355,12 +359,11 @@ static void mlx5_disable_msix(struct mlx5_core_dev *dev) kfree(priv->msix_arr); } -struct mlx5_reg_host_endianess { +struct mlx5_reg_host_endianness { u8 he; u8 rsvd[15]; }; - #define CAP_MASK(pos, size) ((u64)((1 << (size)) - 1) << (pos)) enum { @@ -474,7 +477,7 @@ static int handle_hca_cap_atomic(struct mlx5_core_dev *dev) req_endianness = MLX5_CAP_ATOMIC(dev, - supported_atomic_req_8B_endianess_mode_1); + supported_atomic_req_8B_endianness_mode_1); if (req_endianness != MLX5_ATOMIC_REQ_MODE_HOST_ENDIANNESS) return 0; @@ -486,7 +489,7 @@ static int handle_hca_cap_atomic(struct mlx5_core_dev *dev) set_hca_cap = MLX5_ADDR_OF(set_hca_cap_in, set_ctx, capability); /* Set requestor to host endianness */ - MLX5_SET(atomic_caps, set_hca_cap, atomic_req_8B_endianess_mode, + MLX5_SET(atomic_caps, set_hca_cap, atomic_req_8B_endianness_mode, MLX5_ATOMIC_REQ_MODE_HOST_ENDIANNESS); err = set_caps(dev, set_ctx, set_sz, MLX5_SET_HCA_CAP_OP_MOD_ATOMIC); @@ -537,8 +540,10 @@ static int handle_hca_cap(struct mlx5_core_dev *dev) /* disable cmdif checksum */ MLX5_SET(cmd_hca_cap, set_hca_cap, cmdif_checksum, 0); - /* If the HCA supports 4K UARs use it */ - if (MLX5_CAP_GEN_MAX(dev, uar_4k)) + /* Enable 4K UAR only when HCA supports it and page size is bigger + * than 4K. + */ + if (MLX5_CAP_GEN_MAX(dev, uar_4k) && PAGE_SIZE > 4096) MLX5_SET(cmd_hca_cap, set_hca_cap, uar_4k, 1); MLX5_SET(cmd_hca_cap, set_hca_cap, log_uar_page_sz, PAGE_SHIFT - 12); @@ -559,8 +564,8 @@ query_ex: static int set_hca_ctrl(struct mlx5_core_dev *dev) { - struct mlx5_reg_host_endianess he_in; - struct mlx5_reg_host_endianess he_out; + struct mlx5_reg_host_endianness he_in; + struct mlx5_reg_host_endianness he_out; int err; if (!mlx5_core_is_pf(dev)) @@ -612,7 +617,6 @@ static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i) struct mlx5_priv *priv = &mdev->priv; struct msix_entry *msix = priv->msix_arr; int irq = msix[i + MLX5_EQ_VEC_COMP_BASE].vector; - int err; if (!zalloc_cpumask_var(&priv->irq_info[i].mask, GFP_KERNEL)) { mlx5_core_warn(mdev, "zalloc_cpumask_var failed"); @@ -622,18 +626,11 @@ static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i) cpumask_set_cpu(cpumask_local_spread(i, priv->numa_node), priv->irq_info[i].mask); - err = irq_set_affinity_hint(irq, priv->irq_info[i].mask); - if (err) { - mlx5_core_warn(mdev, "irq_set_affinity_hint failed,irq 0x%.4x", - irq); - goto err_clear_mask; - } + if (IS_ENABLED(CONFIG_SMP) && + irq_set_affinity_hint(irq, priv->irq_info[i].mask)) + mlx5_core_warn(mdev, "irq_set_affinity_hint failed, irq 0x%.4x", irq); return 0; - -err_clear_mask: - free_cpumask_var(priv->irq_info[i].mask); - return err; } static void mlx5_irq_clear_affinity_hint(struct mlx5_core_dev *mdev, int i) @@ -941,6 +938,8 @@ static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv) mlx5_init_mkey_table(dev); + mlx5_init_reserved_gids(dev); + err = mlx5_init_rl_table(dev); if (err) { dev_err(&pdev->dev, "Failed to init rate limiting\n"); @@ -961,8 +960,16 @@ static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv) goto err_eswitch_cleanup; } + err = mlx5_fpga_init(dev); + if (err) { + dev_err(&pdev->dev, "Failed to init fpga device %d\n", err); + goto err_sriov_cleanup; + } + return 0; +err_sriov_cleanup: + mlx5_sriov_cleanup(dev); err_eswitch_cleanup: #ifdef CONFIG_MLX5_CORE_EN mlx5_eswitch_cleanup(dev->priv.eswitch); @@ -986,11 +993,13 @@ out: static void mlx5_cleanup_once(struct mlx5_core_dev *dev) { + mlx5_fpga_cleanup(dev); mlx5_sriov_cleanup(dev); #ifdef CONFIG_MLX5_CORE_EN mlx5_eswitch_cleanup(dev->priv.eswitch); #endif mlx5_cleanup_rl_table(dev); + mlx5_cleanup_reserved_gids(dev); mlx5_cleanup_mkey_table(dev); mlx5_cleanup_srq_table(dev); mlx5_cleanup_qp_table(dev); @@ -1019,6 +1028,15 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, */ dev->state = MLX5_DEVICE_STATE_UP; + /* wait for firmware to accept initialization segments configurations + */ + err = wait_fw_init(dev, FW_PRE_INIT_TIMEOUT_MILI); + if (err) { + dev_err(&dev->pdev->dev, "Firmware over %d MS in pre-initializing state, aborting\n", + FW_PRE_INIT_TIMEOUT_MILI); + goto out_err; + } + err = mlx5_cmd_init(dev); if (err) { dev_err(&pdev->dev, "Failed initializing command interface, aborting\n"); @@ -1147,6 +1165,17 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, goto err_sriov; } + err = mlx5_fpga_device_start(dev); + if (err) { + dev_err(&pdev->dev, "fpga device start failed %d\n", err); + goto err_fpga_start; + } + err = mlx5_accel_ipsec_init(dev); + if (err) { + dev_err(&pdev->dev, "IPSec device start failed %d\n", err); + goto err_ipsec_start; + } + if (mlx5_device_registered(dev)) { mlx5_attach_device(dev); } else { @@ -1165,6 +1194,11 @@ out: return 0; err_reg_dev: + mlx5_accel_ipsec_cleanup(dev); +err_ipsec_start: + mlx5_fpga_device_stop(dev); + +err_fpga_start: mlx5_sriov_detach(dev); err_sriov: @@ -1224,7 +1258,7 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, int err = 0; if (cleanup) - mlx5_drain_health_wq(dev); + mlx5_drain_health_recovery(dev); mutex_lock(&dev->intf_state_mutex); if (test_bit(MLX5_INTERFACE_STATE_DOWN, &dev->intf_state)) { @@ -1235,9 +1269,15 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, goto out; } + clear_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state); + set_bit(MLX5_INTERFACE_STATE_DOWN, &dev->intf_state); + if (mlx5_device_registered(dev)) mlx5_detach_device(dev); + mlx5_accel_ipsec_cleanup(dev); + mlx5_fpga_device_stop(dev); + mlx5_sriov_detach(dev); #ifdef CONFIG_MLX5_CORE_EN mlx5_eswitch_detach(dev->priv.eswitch); @@ -1262,8 +1302,6 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, mlx5_cmd_cleanup(dev); out: - clear_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state); - set_bit(MLX5_INTERFACE_STATE_DOWN, &dev->intf_state); mutex_unlock(&dev->intf_state_mutex); return err; } @@ -1408,7 +1446,7 @@ static pci_ers_result_t mlx5_pci_err_detected(struct pci_dev *pdev, dev_info(&pdev->dev, "%s was called\n", __func__); - mlx5_enter_error_state(dev); + mlx5_enter_error_state(dev, false); mlx5_unload_one(dev, priv, false); /* In case of kernel call drain the health wq */ if (state) { @@ -1495,24 +1533,52 @@ static const struct pci_error_handlers mlx5_err_handler = { .resume = mlx5_pci_resume }; +static int mlx5_try_fast_unload(struct mlx5_core_dev *dev) +{ + int ret; + + if (!MLX5_CAP_GEN(dev, force_teardown)) { + mlx5_core_dbg(dev, "force teardown is not supported in the firmware\n"); + return -EOPNOTSUPP; + } + + if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) { + mlx5_core_dbg(dev, "Device in internal error state, giving up\n"); + return -EAGAIN; + } + + ret = mlx5_cmd_force_teardown_hca(dev); + if (ret) { + mlx5_core_dbg(dev, "Firmware couldn't do fast unload error: %d\n", ret); + return ret; + } + + mlx5_enter_error_state(dev, true); + + return 0; +} + static void shutdown(struct pci_dev *pdev) { struct mlx5_core_dev *dev = pci_get_drvdata(pdev); struct mlx5_priv *priv = &dev->priv; + int err; dev_info(&pdev->dev, "Shutdown was called\n"); /* Notify mlx5 clients that the kernel is being shut down */ set_bit(MLX5_INTERFACE_STATE_SHUTDOWN, &dev->intf_state); - mlx5_unload_one(dev, priv, false); + err = mlx5_try_fast_unload(dev); + if (err) + mlx5_unload_one(dev, priv, false); mlx5_pci_disable_device(dev); } static const struct pci_device_id mlx5_core_pci_table[] = { - { PCI_VDEVICE(MELLANOX, 0x1011) }, /* Connect-IB */ + { PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_CONNECTIB) }, { PCI_VDEVICE(MELLANOX, 0x1012), MLX5_PCI_DEV_IS_VF}, /* Connect-IB VF */ - { PCI_VDEVICE(MELLANOX, 0x1013) }, /* ConnectX-4 */ + { PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_CONNECTX4) }, { PCI_VDEVICE(MELLANOX, 0x1014), MLX5_PCI_DEV_IS_VF}, /* ConnectX-4 VF */ - { PCI_VDEVICE(MELLANOX, 0x1015) }, /* ConnectX-4LX */ + { PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_CONNECTX4_LX) }, { PCI_VDEVICE(MELLANOX, 0x1016), MLX5_PCI_DEV_IS_VF}, /* ConnectX-4LX VF */ { PCI_VDEVICE(MELLANOX, 0x1017) }, /* ConnectX-5, PCIe 3.0 */ { PCI_VDEVICE(MELLANOX, 0x1018), MLX5_PCI_DEV_IS_VF}, /* ConnectX-5 VF */ @@ -1520,6 +1586,8 @@ static const struct pci_device_id mlx5_core_pci_table[] = { { PCI_VDEVICE(MELLANOX, 0x101a), MLX5_PCI_DEV_IS_VF}, /* ConnectX-5 Ex VF */ { PCI_VDEVICE(MELLANOX, 0x101b) }, /* ConnectX-6 */ { PCI_VDEVICE(MELLANOX, 0x101c), MLX5_PCI_DEV_IS_VF}, /* ConnectX-6 VF */ + { PCI_VDEVICE(MELLANOX, 0xa2d2) }, /* BlueField integrated ConnectX-5 network controller */ + { PCI_VDEVICE(MELLANOX, 0xa2d3), MLX5_PCI_DEV_IS_VF}, /* BlueField integrated ConnectX-5 network controller VF */ { 0, } }; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index fbc6e9e9e305..6a3d6bef7dd4 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -37,10 +37,10 @@ #include <linux/kernel.h> #include <linux/sched.h> #include <linux/if_link.h> +#include <linux/firmware.h> #define DRIVER_NAME "mlx5_core" -#define DRIVER_VERSION "3.0-1" -#define DRIVER_RELDATE "January 2015" +#define DRIVER_VERSION "5.0-0" #define MLX5_TOTAL_VPORTS(mdev) (1 + pci_sriov_get_totalvfs(mdev->pdev)) @@ -84,12 +84,13 @@ int mlx5_query_hca_caps(struct mlx5_core_dev *dev); int mlx5_query_board_id(struct mlx5_core_dev *dev); int mlx5_cmd_init_hca(struct mlx5_core_dev *dev); int mlx5_cmd_teardown_hca(struct mlx5_core_dev *dev); +int mlx5_cmd_force_teardown_hca(struct mlx5_core_dev *dev); void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event, unsigned long param); void mlx5_core_page_fault(struct mlx5_core_dev *dev, struct mlx5_pagefault *pfault); void mlx5_port_module_event(struct mlx5_core_dev *dev, struct mlx5_eqe *eqe); -void mlx5_enter_error_state(struct mlx5_core_dev *dev); +void mlx5_enter_error_state(struct mlx5_core_dev *dev, bool force); void mlx5_disable_device(struct mlx5_core_dev *dev); void mlx5_recover_device(struct mlx5_core_dev *dev); int mlx5_sriov_init(struct mlx5_core_dev *dev); @@ -153,6 +154,8 @@ int mlx5_set_mtpps(struct mlx5_core_dev *mdev, u32 *mtpps, u32 mtpps_size); int mlx5_query_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 *arm, u8 *mode); int mlx5_set_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 arm, u8 mode); +int mlx5_firmware_flash(struct mlx5_core_dev *dev, const struct firmware *fw); + void mlx5e_init(void); void mlx5e_cleanup(void); @@ -168,4 +171,7 @@ static inline int mlx5_lag_is_lacp_owner(struct mlx5_core_dev *dev) MLX5_CAP_GEN(dev, lag_master); } +int mlx5_lag_allow(struct mlx5_core_dev *dev); +int mlx5_lag_forbid(struct mlx5_core_dev *dev); + #endif /* __MLX5_CORE_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c index a57d5a81eb05..e36d3e3675f9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c @@ -279,7 +279,7 @@ static int give_pages(struct mlx5_core_dev *dev, u16 func_id, int npages, int i; inlen += npages * MLX5_FLD_SZ_BYTES(manage_pages_in, pas[0]); - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) { err = -ENOMEM; mlx5_core_warn(dev, "vzalloc failed %d\n", inlen); @@ -376,7 +376,7 @@ static int reclaim_pages(struct mlx5_core_dev *dev, u32 func_id, int npages, *nclaimed = 0; outlen += npages * MLX5_FLD_SZ_BYTES(manage_pages_out, pas[0]); - out = mlx5_vzalloc(outlen); + out = kvzalloc(outlen, GFP_KERNEL); if (!out) return -ENOMEM; @@ -403,7 +403,6 @@ static int reclaim_pages(struct mlx5_core_dev *dev, u32 func_id, int npages, for (i = 0; i < num_claimed; i++) free_4k(dev, MLX5_GET64(manage_pages_out, out, pas[i])); - if (nclaimed) *nclaimed = num_claimed; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/port.c b/drivers/net/ethernet/mellanox/mlx5/core/port.c index 141583daf5a2..1975d4388d4f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/port.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/port.c @@ -47,8 +47,8 @@ int mlx5_core_access_reg(struct mlx5_core_dev *dev, void *data_in, u32 *in = NULL; void *data; - in = mlx5_vzalloc(inlen); - out = mlx5_vzalloc(outlen); + in = kvzalloc(inlen, GFP_KERNEL); + out = kvzalloc(outlen, GFP_KERNEL); if (!in || !out) goto out; @@ -454,7 +454,7 @@ int mlx5_core_query_ib_ppcnt(struct mlx5_core_dev *dev, u32 *in; int err; - in = mlx5_vzalloc(sz); + in = kvzalloc(sz, GFP_KERNEL); if (!in) { err = -ENOMEM; return err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/qp.c b/drivers/net/ethernet/mellanox/mlx5/core/qp.c index cbbcef2884be..340f281c9801 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/qp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/qp.c @@ -30,7 +30,6 @@ * SOFTWARE. */ - #include <linux/gfp.h> #include <linux/export.h> #include <linux/mlx5/cmd.h> @@ -519,23 +518,3 @@ int mlx5_core_query_q_counter(struct mlx5_core_dev *dev, u16 counter_id, return mlx5_cmd_exec(dev, in, sizeof(in), out, out_size); } EXPORT_SYMBOL_GPL(mlx5_core_query_q_counter); - -int mlx5_core_query_out_of_buffer(struct mlx5_core_dev *dev, u16 counter_id, - u32 *out_of_buffer) -{ - int outlen = MLX5_ST_SZ_BYTES(query_q_counter_out); - void *out; - int err; - - out = mlx5_vzalloc(outlen); - if (!out) - return -ENOMEM; - - err = mlx5_core_query_q_counter(dev, counter_id, 0, out, outlen); - if (!err) - *out_of_buffer = MLX5_GET(query_q_counter_out, out, - out_of_buffer); - - kfree(out); - return err; -} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c index e08627785590..bcdf7779c48d 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/sriov.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/sriov.c @@ -175,15 +175,20 @@ int mlx5_core_sriov_configure(struct pci_dev *pdev, int num_vfs) if (!mlx5_core_is_pf(dev)) return -EPERM; - if (num_vfs && mlx5_lag_is_active(dev)) { - mlx5_core_warn(dev, "can't turn sriov on while LAG is active"); - return -EINVAL; + if (num_vfs) { + int ret; + + ret = mlx5_lag_forbid(dev); + if (ret && (ret != -ENODEV)) + return ret; } - if (num_vfs) + if (num_vfs) { err = mlx5_sriov_enable(pdev, num_vfs); - else + } else { mlx5_sriov_disable(pdev); + mlx5_lag_allow(dev); + } return err ? err : num_vfs; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c index 3099630015d7..f774de6f5fcb 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c @@ -162,7 +162,7 @@ static int create_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, pas_size = get_pas_size(in); inlen = MLX5_ST_SZ_BYTES(create_srq_in) + pas_size; - create_in = mlx5_vzalloc(inlen); + create_in = kvzalloc(inlen, GFP_KERNEL); if (!create_in) return -ENOMEM; @@ -221,7 +221,7 @@ static int query_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, void *srqc; int err; - srq_out = mlx5_vzalloc(MLX5_ST_SZ_BYTES(query_srq_out)); + srq_out = kvzalloc(MLX5_ST_SZ_BYTES(query_srq_out), GFP_KERNEL); if (!srq_out) return -ENOMEM; @@ -256,7 +256,7 @@ static int create_xrc_srq_cmd(struct mlx5_core_dev *dev, pas_size = get_pas_size(in); inlen = MLX5_ST_SZ_BYTES(create_xrc_srq_in) + pas_size; - create_in = mlx5_vzalloc(inlen); + create_in = kvzalloc(inlen, GFP_KERNEL); if (!create_in) return -ENOMEM; @@ -320,7 +320,7 @@ static int query_xrc_srq_cmd(struct mlx5_core_dev *dev, void *xrc_srqc; int err; - xrcsrq_out = mlx5_vzalloc(MLX5_ST_SZ_BYTES(query_xrc_srq_out)); + xrcsrq_out = kvzalloc(MLX5_ST_SZ_BYTES(query_xrc_srq_out), GFP_KERNEL); if (!xrcsrq_out) return -ENOMEM; memset(xrcsrq_in, 0, sizeof(xrcsrq_in)); @@ -357,7 +357,7 @@ static int create_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, pas_size = get_pas_size(in); inlen = MLX5_ST_SZ_BYTES(create_rmp_in) + pas_size; - create_in = mlx5_vzalloc(inlen); + create_in = kvzalloc(inlen, GFP_KERNEL); if (!create_in) return -ENOMEM; @@ -390,7 +390,7 @@ static int arm_rmp_cmd(struct mlx5_core_dev *dev, void *bitmask; int err; - in = mlx5_vzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in)); + in = kvzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in), GFP_KERNEL); if (!in) return -ENOMEM; @@ -417,7 +417,7 @@ static int query_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq, void *rmpc; int err; - rmp_out = mlx5_vzalloc(MLX5_ST_SZ_BYTES(query_rmp_out)); + rmp_out = kvzalloc(MLX5_ST_SZ_BYTES(query_rmp_out), GFP_KERNEL); if (!rmp_out) return -ENOMEM; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/transobj.c b/drivers/net/ethernet/mellanox/mlx5/core/transobj.c index a00ff49eec18..5e128d7a9ffd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/transobj.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/transobj.c @@ -284,7 +284,7 @@ int mlx5_core_arm_rmp(struct mlx5_core_dev *dev, u32 rmpn, u16 lwm) void *bitmask; int err; - in = mlx5_vzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in)); + in = kvzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in), GFP_KERNEL); if (!in) return -ENOMEM; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/vport.c b/drivers/net/ethernet/mellanox/mlx5/core/vport.c index 15c2294dd2b4..5abfec1c3399 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/vport.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/vport.c @@ -172,7 +172,7 @@ int mlx5_query_nic_vport_mac_address(struct mlx5_core_dev *mdev, u8 *out_addr; int err; - out = mlx5_vzalloc(outlen); + out = kvzalloc(outlen, GFP_KERNEL); if (!out) return -ENOMEM; @@ -197,11 +197,9 @@ int mlx5_modify_nic_vport_mac_address(struct mlx5_core_dev *mdev, void *nic_vport_ctx; u8 *perm_mac; - in = mlx5_vzalloc(inlen); - if (!in) { - mlx5_core_warn(mdev, "failed to allocate inbox\n"); + in = kvzalloc(inlen, GFP_KERNEL); + if (!in) return -ENOMEM; - } MLX5_SET(modify_nic_vport_context_in, in, field_select.permanent_address, 1); @@ -231,7 +229,7 @@ int mlx5_query_nic_vport_mtu(struct mlx5_core_dev *mdev, u16 *mtu) u32 *out; int err; - out = mlx5_vzalloc(outlen); + out = kvzalloc(outlen, GFP_KERNEL); if (!out) return -ENOMEM; @@ -251,7 +249,7 @@ int mlx5_modify_nic_vport_mtu(struct mlx5_core_dev *mdev, u16 mtu) void *in; int err; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -501,7 +499,7 @@ int mlx5_query_nic_vport_system_image_guid(struct mlx5_core_dev *mdev, u32 *out; int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out); - out = mlx5_vzalloc(outlen); + out = kvzalloc(outlen, GFP_KERNEL); if (!out) return -ENOMEM; @@ -521,7 +519,7 @@ int mlx5_query_nic_vport_node_guid(struct mlx5_core_dev *mdev, u64 *node_guid) u32 *out; int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out); - out = mlx5_vzalloc(outlen); + out = kvzalloc(outlen, GFP_KERNEL); if (!out) return -ENOMEM; @@ -551,7 +549,7 @@ int mlx5_modify_nic_vport_node_guid(struct mlx5_core_dev *mdev, if (!MLX5_CAP_ESW(mdev, nic_vport_node_guid_modify)) return -EOPNOTSUPP; - in = mlx5_vzalloc(inlen); + in = kvzalloc(inlen, GFP_KERNEL); if (!in) return -ENOMEM; @@ -577,7 +575,7 @@ int mlx5_query_nic_vport_qkey_viol_cntr(struct mlx5_core_dev *mdev, u32 *out; int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out); - out = mlx5_vzalloc(outlen); + out = kvzalloc(outlen, GFP_KERNEL); if (!out) return -ENOMEM; @@ -879,11 +877,9 @@ int mlx5_modify_nic_vport_promisc(struct mlx5_core_dev *mdev, int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in); int err; - in = mlx5_vzalloc(inlen); - if (!in) { - mlx5_core_err(mdev, "failed to allocate inbox\n"); + in = kvzalloc(inlen, GFP_KERNEL); + if (!in) return -ENOMEM; - } MLX5_SET(modify_nic_vport_context_in, in, field_select.promisc, 1); MLX5_SET(modify_nic_vport_context_in, in, @@ -913,11 +909,9 @@ static int mlx5_nic_vport_update_roce_state(struct mlx5_core_dev *mdev, int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in); int err; - in = mlx5_vzalloc(inlen); - if (!in) { - mlx5_core_warn(mdev, "failed to allocate inbox\n"); + in = kvzalloc(inlen, GFP_KERNEL); + if (!in) return -ENOMEM; - } MLX5_SET(modify_nic_vport_context_in, in, field_select.roce_en, 1); MLX5_SET(modify_nic_vport_context_in, in, nic_vport_context.roce_en, @@ -932,12 +926,16 @@ static int mlx5_nic_vport_update_roce_state(struct mlx5_core_dev *mdev, int mlx5_nic_vport_enable_roce(struct mlx5_core_dev *mdev) { + if (atomic_inc_return(&mdev->roce.roce_en) != 1) + return 0; return mlx5_nic_vport_update_roce_state(mdev, MLX5_VPORT_ROCE_ENABLED); } EXPORT_SYMBOL_GPL(mlx5_nic_vport_enable_roce); int mlx5_nic_vport_disable_roce(struct mlx5_core_dev *mdev) { + if (atomic_dec_return(&mdev->roce.roce_en) != 0) + return 0; return mlx5_nic_vport_update_roce_state(mdev, MLX5_VPORT_ROCE_DISABLED); } EXPORT_SYMBOL_GPL(mlx5_nic_vport_disable_roce); @@ -952,7 +950,7 @@ int mlx5_core_query_vport_counter(struct mlx5_core_dev *dev, u8 other_vport, int err; is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager); - in = mlx5_vzalloc(in_sz); + in = kvzalloc(in_sz, GFP_KERNEL); if (!in) { err = -ENOMEM; return err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.c b/drivers/net/ethernet/mellanox/mlx5/core/wq.c index 921673c42bc9..6bcfc25350f5 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/wq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.c @@ -54,6 +54,12 @@ static u32 mlx5_wq_cyc_get_byte_size(struct mlx5_wq_cyc *wq) return mlx5_wq_cyc_get_size(wq) << wq->log_stride; } +static u32 mlx5_wq_qp_get_byte_size(struct mlx5_wq_qp *wq) +{ + return mlx5_wq_cyc_get_byte_size(&wq->rq) + + mlx5_wq_cyc_get_byte_size(&wq->sq); +} + static u32 mlx5_cqwq_get_byte_size(struct mlx5_cqwq *wq) { return mlx5_cqwq_get_size(wq) << wq->log_stride; @@ -99,6 +105,46 @@ err_db_free: return err; } +int mlx5_wq_qp_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, + void *qpc, struct mlx5_wq_qp *wq, + struct mlx5_wq_ctrl *wq_ctrl) +{ + int err; + + wq->rq.log_stride = MLX5_GET(qpc, qpc, log_rq_stride) + 4; + wq->rq.sz_m1 = (1 << MLX5_GET(qpc, qpc, log_rq_size)) - 1; + + wq->sq.log_stride = ilog2(MLX5_SEND_WQE_BB); + wq->sq.sz_m1 = (1 << MLX5_GET(qpc, qpc, log_sq_size)) - 1; + + err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node); + if (err) { + mlx5_core_warn(mdev, "mlx5_db_alloc_node() failed, %d\n", err); + return err; + } + + err = mlx5_buf_alloc_node(mdev, mlx5_wq_qp_get_byte_size(wq), + &wq_ctrl->buf, param->buf_numa_node); + if (err) { + mlx5_core_warn(mdev, "mlx5_buf_alloc_node() failed, %d\n", err); + goto err_db_free; + } + + wq->rq.buf = wq_ctrl->buf.direct.buf; + wq->sq.buf = wq->rq.buf + mlx5_wq_cyc_get_byte_size(&wq->rq); + wq->rq.db = &wq_ctrl->db.db[MLX5_RCV_DBR]; + wq->sq.db = &wq_ctrl->db.db[MLX5_SND_DBR]; + + wq_ctrl->mdev = mdev; + + return 0; + +err_db_free: + mlx5_db_free(mdev, &wq_ctrl->db); + + return err; +} + int mlx5_cqwq_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, void *cqc, struct mlx5_cqwq *wq, struct mlx5_frag_wq_ctrl *wq_ctrl) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.h b/drivers/net/ethernet/mellanox/mlx5/core/wq.h index d8afed898c31..718589d0cec2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/wq.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.h @@ -34,6 +34,8 @@ #define __MLX5_WQ_H__ #include <linux/mlx5/mlx5_ifc.h> +#include <linux/mlx5/cq.h> +#include <linux/mlx5/qp.h> struct mlx5_wq_param { int linear; @@ -60,6 +62,11 @@ struct mlx5_wq_cyc { u8 log_stride; }; +struct mlx5_wq_qp { + struct mlx5_wq_cyc rq; + struct mlx5_wq_cyc sq; +}; + struct mlx5_cqwq { struct mlx5_frag_buf frag_buf; __be32 *db; @@ -87,6 +94,10 @@ int mlx5_wq_cyc_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, struct mlx5_wq_ctrl *wq_ctrl); u32 mlx5_wq_cyc_get_size(struct mlx5_wq_cyc *wq); +int mlx5_wq_qp_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, + void *qpc, struct mlx5_wq_qp *wq, + struct mlx5_wq_ctrl *wq_ctrl); + int mlx5_cqwq_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param, void *cqc, struct mlx5_cqwq *wq, struct mlx5_frag_wq_ctrl *wq_ctrl); @@ -146,6 +157,22 @@ static inline void mlx5_cqwq_update_db_record(struct mlx5_cqwq *wq) *wq->db = cpu_to_be32(wq->cc & 0xffffff); } +static inline struct mlx5_cqe64 *mlx5_cqwq_get_cqe(struct mlx5_cqwq *wq) +{ + u32 ci = mlx5_cqwq_get_ci(wq); + struct mlx5_cqe64 *cqe = mlx5_cqwq_get_wqe(wq, ci); + u8 cqe_ownership_bit = cqe->op_own & MLX5_CQE_OWNER_MASK; + u8 sw_ownership_val = mlx5_cqwq_get_wrap_cnt(wq) & 1; + + if (cqe_ownership_bit != sw_ownership_val) + return NULL; + + /* ensure cqe content is read after cqe ownership bit */ + dma_rmb(); + + return cqe; +} + static inline int mlx5_wq_ll_is_full(struct mlx5_wq_ll *wq) { return wq->cur_sz == wq->sz_m1; |