linux_dsm_epyc7002/drivers/infiniband/hw/ocrdma/ocrdma.h

610 lines
13 KiB
C
Raw Normal View History

/* This file is part of the Emulex RoCE Device Driver for
* RoCE (RDMA over Converged Ethernet) adapters.
* Copyright (C) 2012-2015 Emulex. All rights reserved.
* EMULEX and SLI are trademarks of Emulex.
* www.emulex.com
*
* 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 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.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
* WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
* OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
* ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Contact Information:
* linux-drivers@emulex.com
*
* Emulex
* 3333 Susan Street
* Costa Mesa, CA 92626
*/
#ifndef __OCRDMA_H__
#define __OCRDMA_H__
#include <linux/mutex.h>
#include <linux/list.h>
#include <linux/spinlock.h>
#include <linux/pci.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_user_verbs.h>
#include <rdma/ib_addr.h>
#include <be_roce.h>
#include "ocrdma_sli.h"
#define OCRDMA_ROCE_DRV_VERSION "11.0.0.0"
#define OCRDMA_ROCE_DRV_DESC "Emulex OneConnect RoCE Driver"
#define OCRDMA_NODE_DESC "Emulex OneConnect RoCE HCA"
#define OC_NAME_SH OCRDMA_NODE_DESC "(Skyhawk)"
#define OC_NAME_UNKNOWN OCRDMA_NODE_DESC "(Unknown)"
#define OC_SKH_DEVICE_PF 0x720
#define OC_SKH_DEVICE_VF 0x728
#define OCRDMA_MAX_AH 512
#define OCRDMA_UVERBS(CMD_NAME) (1ull << IB_USER_VERBS_CMD_##CMD_NAME)
#define convert_to_64bit(lo, hi) ((u64)hi << 32 | (u64)lo)
#define EQ_INTR_PER_SEC_THRSH_HI 150000
#define EQ_INTR_PER_SEC_THRSH_LOW 100000
#define EQ_AIC_MAX_EQD 20
#define EQ_AIC_MIN_EQD 0
void ocrdma_eqd_set_task(struct work_struct *work);
struct ocrdma_dev_attr {
u8 fw_ver[32];
u32 vendor_id;
u32 device_id;
u16 max_pd;
u16 max_dpp_pds;
u16 max_cq;
u16 max_cqe;
u16 max_qp;
u16 max_wqe;
u16 max_rqe;
u16 max_srq;
u32 max_inline_data;
int max_send_sge;
int max_recv_sge;
int max_srq_sge;
int max_rdma_sge;
int max_mr;
u64 max_mr_size;
u32 max_num_mr_pbl;
int max_mw;
int max_fmr;
int max_map_per_fmr;
int max_pages_per_frmr;
u16 max_ord_per_qp;
u16 max_ird_per_qp;
int device_cap_flags;
u8 cq_overflow_detect;
u8 srq_supported;
u32 wqe_size;
u32 rqe_size;
u32 ird_page_size;
u8 local_ca_ack_delay;
u8 ird;
u8 num_ird_pages;
u8 udp_encap;
};
struct ocrdma_dma_mem {
void *va;
dma_addr_t pa;
u32 size;
};
struct ocrdma_pbl {
void *va;
dma_addr_t pa;
};
struct ocrdma_queue_info {
void *va;
dma_addr_t dma;
u32 size;
u16 len;
u16 entry_size; /* Size of an element in the queue */
u16 id; /* qid, where to ring the doorbell. */
u16 head, tail;
bool created;
};
struct ocrdma_aic_obj { /* Adaptive interrupt coalescing (AIC) info */
u32 prev_eqd;
u64 eq_intr_cnt;
u64 prev_eq_intr_cnt;
};
struct ocrdma_eq {
struct ocrdma_queue_info q;
u32 vector;
int cq_cnt;
struct ocrdma_dev *dev;
char irq_name[32];
struct ocrdma_aic_obj aic_obj;
};
struct ocrdma_mq {
struct ocrdma_queue_info sq;
struct ocrdma_queue_info cq;
bool rearm_cq;
};
struct mqe_ctx {
struct mutex lock; /* for serializing mailbox commands on MQ */
wait_queue_head_t cmd_wait;
u32 tag;
u16 cqe_status;
u16 ext_status;
bool cmd_done;
bool fw_error_state;
};
struct ocrdma_hw_mr {
u32 lkey;
u8 fr_mr;
u8 remote_atomic;
u8 remote_rd;
u8 remote_wr;
u8 local_rd;
u8 local_wr;
u8 mw_bind;
u8 rsvd;
u64 len;
struct ocrdma_pbl *pbl_table;
u32 num_pbls;
u32 num_pbes;
u32 pbl_size;
u32 pbe_size;
u64 fbo;
u64 va;
};
struct ocrdma_mr {
struct ib_mr ibmr;
struct ib_umem *umem;
struct ocrdma_hw_mr hwmr;
u64 *pages;
u32 npages;
};
struct ocrdma_stats {
u8 type;
struct ocrdma_dev *dev;
};
struct ocrdma_pd_resource_mgr {
u32 pd_norm_start;
u16 pd_norm_count;
u16 pd_norm_thrsh;
u16 max_normal_pd;
u32 pd_dpp_start;
u16 pd_dpp_count;
u16 pd_dpp_thrsh;
u16 max_dpp_pd;
u16 dpp_page_index;
unsigned long *pd_norm_bitmap;
unsigned long *pd_dpp_bitmap;
bool pd_prealloc_valid;
};
struct stats_mem {
struct ocrdma_mqe mqe;
void *va;
dma_addr_t pa;
u32 size;
char *debugfs_mem;
};
struct phy_info {
u16 auto_speeds_supported;
u16 fixed_speeds_supported;
u16 phy_type;
u16 interface_type;
};
RDMA/ocrdma: Depend on async link events from CNA Recently Dough Ledford reported a deadlock happening between ocrdma-load sequence and NetworkManager service issuing "open" on be2net interface. The deadlock happens when any be2net hook (e.g. open/close) is called in parallel to insmod ocrdma.ko. A. be2net is sending administrative open/close event to ocrdma holding device_list_mutex. It does this from ndo_open/ndo_stop hooks of be2net. So sequence of locks is rtnl_lock---> device_list lock B. When new ocrdma roce device gets registered, infiniband stack now takes rtnl_lock in ib_register_device() in GID initialization routines. So sequence of locks in this path is device_list lock ---> rtnl_lock. This improper locking sequence causes deadlock. With this patch we stop using administrative open and close events injected by be2net driver. These events were used to dispatch PORT_ACTIVE and PORT_ERROR events to the IB-stack. This patch implements a logic to receive async-link-events generated from CNA whenever link-state-change is detected. Now on, these async-events will be used to dispatch PORT_ACTIVE and PORT_ERROR events to IB-stack. Depending on async-events from CNA removes the need to hold device-list-mutex and thus breaks the busy-wait scenario. Reported-by: Doug Ledford <dledford@redhat.com> CC: Sathya Perla <sathya.perla@avagotech.com> Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@avagotech.com> Signed-off-by: Selvin Xavier <selvin.xavier@avagotech.com> Signed-off-by: Devesh Sharma <devesh.sharma@avagotech.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-12-25 01:14:07 +07:00
enum ocrdma_flags {
OCRDMA_FLAGS_LINK_STATUS_INIT = 0x01
};
struct ocrdma_dev {
struct ib_device ibdev;
struct ocrdma_dev_attr attr;
struct mutex dev_lock; /* provides syncronise access to device data */
spinlock_t flush_q_lock ____cacheline_aligned;
struct ocrdma_cq **cq_tbl;
struct ocrdma_qp **qp_tbl;
struct ocrdma_eq *eq_tbl;
int eq_cnt;
struct delayed_work eqd_work;
u16 base_eqid;
u16 max_eq;
/* provided synchronization to sgid table for
* updating gid entries triggered by notifier.
*/
spinlock_t sgid_lock;
int gsi_qp_created;
struct ocrdma_cq *gsi_sqcq;
struct ocrdma_cq *gsi_rqcq;
struct {
struct ocrdma_av *va;
dma_addr_t pa;
u32 size;
u32 num_ah;
/* provide synchronization for av
* entry allocations.
*/
spinlock_t lock;
u32 ahid;
struct ocrdma_pbl pbl;
} av_tbl;
void *mbx_cmd;
struct ocrdma_mq mq;
struct mqe_ctx mqe_ctx;
struct be_dev_info nic_info;
struct phy_info phy;
char model_number[32];
u32 hba_port_num;
struct list_head entry;
int id;
u64 *stag_arr;
u8 sl; /* service level */
bool pfc_state;
atomic_t update_sl;
u16 pvid;
u32 asic_id;
RDMA/ocrdma: Depend on async link events from CNA Recently Dough Ledford reported a deadlock happening between ocrdma-load sequence and NetworkManager service issuing "open" on be2net interface. The deadlock happens when any be2net hook (e.g. open/close) is called in parallel to insmod ocrdma.ko. A. be2net is sending administrative open/close event to ocrdma holding device_list_mutex. It does this from ndo_open/ndo_stop hooks of be2net. So sequence of locks is rtnl_lock---> device_list lock B. When new ocrdma roce device gets registered, infiniband stack now takes rtnl_lock in ib_register_device() in GID initialization routines. So sequence of locks in this path is device_list lock ---> rtnl_lock. This improper locking sequence causes deadlock. With this patch we stop using administrative open and close events injected by be2net driver. These events were used to dispatch PORT_ACTIVE and PORT_ERROR events to the IB-stack. This patch implements a logic to receive async-link-events generated from CNA whenever link-state-change is detected. Now on, these async-events will be used to dispatch PORT_ACTIVE and PORT_ERROR events to IB-stack. Depending on async-events from CNA removes the need to hold device-list-mutex and thus breaks the busy-wait scenario. Reported-by: Doug Ledford <dledford@redhat.com> CC: Sathya Perla <sathya.perla@avagotech.com> Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@avagotech.com> Signed-off-by: Selvin Xavier <selvin.xavier@avagotech.com> Signed-off-by: Devesh Sharma <devesh.sharma@avagotech.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-12-25 01:14:07 +07:00
u32 flags;
ulong last_stats_time;
struct mutex stats_lock; /* provide synch for debugfs operations */
struct stats_mem stats_mem;
struct ocrdma_stats rsrc_stats;
struct ocrdma_stats rx_stats;
struct ocrdma_stats wqe_stats;
struct ocrdma_stats tx_stats;
struct ocrdma_stats db_err_stats;
struct ocrdma_stats tx_qp_err_stats;
struct ocrdma_stats rx_qp_err_stats;
struct ocrdma_stats tx_dbg_stats;
struct ocrdma_stats rx_dbg_stats;
struct ocrdma_stats driver_stats;
struct ocrdma_stats reset_stats;
struct dentry *dir;
atomic_t async_err_stats[OCRDMA_MAX_ASYNC_ERRORS];
atomic_t cqe_err_stats[OCRDMA_MAX_CQE_ERR];
struct ocrdma_pd_resource_mgr *pd_mgr;
};
struct ocrdma_cq {
struct ib_cq ibcq;
struct ocrdma_cqe *va;
u32 phase;
u32 getp; /* pointer to pending wrs to
* return to stack, wrap arounds
* at max_hw_cqe
*/
u32 max_hw_cqe;
bool phase_change;
spinlock_t cq_lock ____cacheline_aligned; /* provide synchronization
* to cq polling
*/
/* syncronizes cq completion handler invoked from multiple context */
spinlock_t comp_handler_lock ____cacheline_aligned;
u16 id;
u16 eqn;
struct ocrdma_ucontext *ucontext;
dma_addr_t pa;
u32 len;
u32 cqe_cnt;
/* head of all qp's sq and rq for which cqes need to be flushed
* by the software.
*/
struct list_head sq_head, rq_head;
};
struct ocrdma_pd {
struct ib_pd ibpd;
struct ocrdma_ucontext *uctx;
u32 id;
int num_dpp_qp;
u32 dpp_page;
bool dpp_enabled;
};
struct ocrdma_ah {
struct ib_ah ibah;
struct ocrdma_av *av;
u16 sgid_index;
u32 id;
u8 hdr_type;
};
struct ocrdma_qp_hwq_info {
u8 *va; /* virtual address */
u32 max_sges;
u32 head, tail;
u32 entry_size;
u32 max_cnt;
u32 max_wqe_idx;
u16 dbid; /* qid, where to ring the doorbell. */
u32 len;
dma_addr_t pa;
};
struct ocrdma_srq {
struct ib_srq ibsrq;
u8 __iomem *db;
struct ocrdma_qp_hwq_info rq;
u64 *rqe_wr_id_tbl;
u32 *idx_bit_fields;
u32 bit_fields_len;
/* provide synchronization to multiple context(s) posting rqe */
spinlock_t q_lock ____cacheline_aligned;
struct ocrdma_pd *pd;
u32 id;
};
struct ocrdma_qp {
struct ib_qp ibqp;
u8 __iomem *sq_db;
struct ocrdma_qp_hwq_info sq;
struct {
uint64_t wrid;
uint16_t dpp_wqe_idx;
uint16_t dpp_wqe;
uint8_t signaled;
uint8_t rsvd[3];
} *wqe_wr_id_tbl;
u32 max_inline_data;
/* provide synchronization to multiple context(s) posting wqe, rqe */
spinlock_t q_lock ____cacheline_aligned;
struct ocrdma_cq *sq_cq;
/* list maintained per CQ to flush SQ errors */
struct list_head sq_entry;
u8 __iomem *rq_db;
struct ocrdma_qp_hwq_info rq;
u64 *rqe_wr_id_tbl;
struct ocrdma_cq *rq_cq;
struct ocrdma_srq *srq;
/* list maintained per CQ to flush RQ errors */
struct list_head rq_entry;
enum ocrdma_qp_state state; /* QP state */
int cap_flags;
u32 max_ord, max_ird;
u32 id;
struct ocrdma_pd *pd;
enum ib_qp_type qp_type;
int sgid_idx;
u32 qkey;
bool dpp_enabled;
u8 *ird_q_va;
bool signaled;
};
struct ocrdma_ucontext {
struct ib_ucontext ibucontext;
struct list_head mm_head;
struct mutex mm_list_lock; /* protects list entries of mm type */
struct ocrdma_pd *cntxt_pd;
int pd_in_use;
struct {
u32 *va;
dma_addr_t pa;
u32 len;
} ah_tbl;
};
struct ocrdma_mm {
struct {
u64 phy_addr;
unsigned long len;
} key;
struct list_head entry;
};
static inline struct ocrdma_dev *get_ocrdma_dev(struct ib_device *ibdev)
{
return container_of(ibdev, struct ocrdma_dev, ibdev);
}
static inline struct ocrdma_ucontext *get_ocrdma_ucontext(struct ib_ucontext
*ibucontext)
{
return container_of(ibucontext, struct ocrdma_ucontext, ibucontext);
}
static inline struct ocrdma_pd *get_ocrdma_pd(struct ib_pd *ibpd)
{
return container_of(ibpd, struct ocrdma_pd, ibpd);
}
static inline struct ocrdma_cq *get_ocrdma_cq(struct ib_cq *ibcq)
{
return container_of(ibcq, struct ocrdma_cq, ibcq);
}
static inline struct ocrdma_qp *get_ocrdma_qp(struct ib_qp *ibqp)
{
return container_of(ibqp, struct ocrdma_qp, ibqp);
}
static inline struct ocrdma_mr *get_ocrdma_mr(struct ib_mr *ibmr)
{
return container_of(ibmr, struct ocrdma_mr, ibmr);
}
static inline struct ocrdma_ah *get_ocrdma_ah(struct ib_ah *ibah)
{
return container_of(ibah, struct ocrdma_ah, ibah);
}
static inline struct ocrdma_srq *get_ocrdma_srq(struct ib_srq *ibsrq)
{
return container_of(ibsrq, struct ocrdma_srq, ibsrq);
}
static inline int is_cqe_valid(struct ocrdma_cq *cq, struct ocrdma_cqe *cqe)
{
int cqe_valid;
cqe_valid = le32_to_cpu(cqe->flags_status_srcqpn) & OCRDMA_CQE_VALID;
return (cqe_valid == cq->phase);
}
static inline int is_cqe_for_sq(struct ocrdma_cqe *cqe)
{
return (le32_to_cpu(cqe->flags_status_srcqpn) &
OCRDMA_CQE_QTYPE) ? 0 : 1;
}
static inline int is_cqe_invalidated(struct ocrdma_cqe *cqe)
{
return (le32_to_cpu(cqe->flags_status_srcqpn) &
OCRDMA_CQE_INVALIDATE) ? 1 : 0;
}
static inline int is_cqe_imm(struct ocrdma_cqe *cqe)
{
return (le32_to_cpu(cqe->flags_status_srcqpn) &
OCRDMA_CQE_IMM) ? 1 : 0;
}
static inline int is_cqe_wr_imm(struct ocrdma_cqe *cqe)
{
return (le32_to_cpu(cqe->flags_status_srcqpn) &
OCRDMA_CQE_WRITE_IMM) ? 1 : 0;
}
static inline int ocrdma_resolve_dmac(struct ocrdma_dev *dev,
struct ib_ah_attr *ah_attr, u8 *mac_addr)
{
struct in6_addr in6;
memcpy(&in6, ah_attr->grh.dgid.raw, sizeof(in6));
if (rdma_is_multicast_addr(&in6))
rdma_get_mcast_mac(&in6, mac_addr);
else if (rdma_link_local_addr(&in6))
rdma_get_ll_mac(&in6, mac_addr);
else
memcpy(mac_addr, ah_attr->dmac, ETH_ALEN);
return 0;
}
static inline char *hca_name(struct ocrdma_dev *dev)
{
switch (dev->nic_info.pdev->device) {
case OC_SKH_DEVICE_PF:
case OC_SKH_DEVICE_VF:
return OC_NAME_SH;
default:
return OC_NAME_UNKNOWN;
}
}
static inline int ocrdma_get_eq_table_index(struct ocrdma_dev *dev,
int eqid)
{
int indx;
for (indx = 0; indx < dev->eq_cnt; indx++) {
if (dev->eq_tbl[indx].q.id == eqid)
return indx;
}
return -EINVAL;
}
static inline u8 ocrdma_get_asic_type(struct ocrdma_dev *dev)
{
if (dev->nic_info.dev_family == 0xF && !dev->asic_id) {
pci_read_config_dword(
dev->nic_info.pdev,
OCRDMA_SLI_ASIC_ID_OFFSET, &dev->asic_id);
}
return (dev->asic_id & OCRDMA_SLI_ASIC_GEN_NUM_MASK) >>
OCRDMA_SLI_ASIC_GEN_NUM_SHIFT;
}
static inline u8 ocrdma_get_pfc_prio(u8 *pfc, u8 prio)
{
return *(pfc + prio);
}
static inline u8 ocrdma_get_app_prio(u8 *app_prio, u8 prio)
{
return *(app_prio + prio);
}
static inline u8 ocrdma_is_enabled_and_synced(u32 state)
{ /* May also be used to interpret TC-state, QCN-state
* Appl-state and Logical-link-state in future.
*/
return (state & OCRDMA_STATE_FLAG_ENABLED) &&
(state & OCRDMA_STATE_FLAG_SYNC);
}
RDMA/ocrdma: Depend on async link events from CNA Recently Dough Ledford reported a deadlock happening between ocrdma-load sequence and NetworkManager service issuing "open" on be2net interface. The deadlock happens when any be2net hook (e.g. open/close) is called in parallel to insmod ocrdma.ko. A. be2net is sending administrative open/close event to ocrdma holding device_list_mutex. It does this from ndo_open/ndo_stop hooks of be2net. So sequence of locks is rtnl_lock---> device_list lock B. When new ocrdma roce device gets registered, infiniband stack now takes rtnl_lock in ib_register_device() in GID initialization routines. So sequence of locks in this path is device_list lock ---> rtnl_lock. This improper locking sequence causes deadlock. With this patch we stop using administrative open and close events injected by be2net driver. These events were used to dispatch PORT_ACTIVE and PORT_ERROR events to the IB-stack. This patch implements a logic to receive async-link-events generated from CNA whenever link-state-change is detected. Now on, these async-events will be used to dispatch PORT_ACTIVE and PORT_ERROR events to IB-stack. Depending on async-events from CNA removes the need to hold device-list-mutex and thus breaks the busy-wait scenario. Reported-by: Doug Ledford <dledford@redhat.com> CC: Sathya Perla <sathya.perla@avagotech.com> Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@avagotech.com> Signed-off-by: Selvin Xavier <selvin.xavier@avagotech.com> Signed-off-by: Devesh Sharma <devesh.sharma@avagotech.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-12-25 01:14:07 +07:00
static inline u8 ocrdma_get_ae_link_state(u32 ae_state)
{
return ((ae_state & OCRDMA_AE_LSC_LS_MASK) >> OCRDMA_AE_LSC_LS_SHIFT);
}
static inline bool ocrdma_is_udp_encap_supported(struct ocrdma_dev *dev)
{
return (dev->attr.udp_encap & OCRDMA_L3_TYPE_IPV4) ||
(dev->attr.udp_encap & OCRDMA_L3_TYPE_IPV6);
}
#endif