diff options
Diffstat (limited to 'drivers/infiniband/hw/hfi1/hfi.h')
-rw-r--r-- | drivers/infiniband/hw/hfi1/hfi.h | 1950 |
1 files changed, 1950 insertions, 0 deletions
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h new file mode 100644 index 000000000000..4417a0fd3ef9 --- /dev/null +++ b/drivers/infiniband/hw/hfi1/hfi.h @@ -0,0 +1,1950 @@ +#ifndef _HFI1_KERNEL_H +#define _HFI1_KERNEL_H +/* + * Copyright(c) 2015, 2016 Intel Corporation. + * + * This file is provided under a dual BSD/GPLv2 license. When using or + * redistributing this file, you may do so under either license. + * + * GPL LICENSE SUMMARY + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of version 2 of the GNU General Public License as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * BSD LICENSE + * + * 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. + * - Neither the name of Intel Corporation nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * 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 + * OWNER 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. + * + */ + +#include <linux/interrupt.h> +#include <linux/pci.h> +#include <linux/dma-mapping.h> +#include <linux/mutex.h> +#include <linux/list.h> +#include <linux/scatterlist.h> +#include <linux/slab.h> +#include <linux/io.h> +#include <linux/fs.h> +#include <linux/completion.h> +#include <linux/kref.h> +#include <linux/sched.h> +#include <linux/cdev.h> +#include <linux/delay.h> +#include <linux/kthread.h> +#include <rdma/rdma_vt.h> + +#include "chip_registers.h" +#include "common.h" +#include "verbs.h" +#include "pio.h" +#include "chip.h" +#include "mad.h" +#include "qsfp.h" +#include "platform.h" +#include "affinity.h" + +/* bumped 1 from s/w major version of TrueScale */ +#define HFI1_CHIP_VERS_MAJ 3U + +/* don't care about this except printing */ +#define HFI1_CHIP_VERS_MIN 0U + +/* The Organization Unique Identifier (Mfg code), and its position in GUID */ +#define HFI1_OUI 0x001175 +#define HFI1_OUI_LSB 40 + +#define DROP_PACKET_OFF 0 +#define DROP_PACKET_ON 1 + +extern unsigned long hfi1_cap_mask; +#define HFI1_CAP_KGET_MASK(mask, cap) ((mask) & HFI1_CAP_##cap) +#define HFI1_CAP_UGET_MASK(mask, cap) \ + (((mask) >> HFI1_CAP_USER_SHIFT) & HFI1_CAP_##cap) +#define HFI1_CAP_KGET(cap) (HFI1_CAP_KGET_MASK(hfi1_cap_mask, cap)) +#define HFI1_CAP_UGET(cap) (HFI1_CAP_UGET_MASK(hfi1_cap_mask, cap)) +#define HFI1_CAP_IS_KSET(cap) (!!HFI1_CAP_KGET(cap)) +#define HFI1_CAP_IS_USET(cap) (!!HFI1_CAP_UGET(cap)) +#define HFI1_MISC_GET() ((hfi1_cap_mask >> HFI1_CAP_MISC_SHIFT) & \ + HFI1_CAP_MISC_MASK) +/* Offline Disabled Reason is 4-bits */ +#define HFI1_ODR_MASK(rsn) ((rsn) & OPA_PI_MASK_OFFLINE_REASON) + +/* + * Control context is always 0 and handles the error packets. + * It also handles the VL15 and multicast packets. + */ +#define HFI1_CTRL_CTXT 0 + +/* + * Driver context will store software counters for each of the events + * associated with these status registers + */ +#define NUM_CCE_ERR_STATUS_COUNTERS 41 +#define NUM_RCV_ERR_STATUS_COUNTERS 64 +#define NUM_MISC_ERR_STATUS_COUNTERS 13 +#define NUM_SEND_PIO_ERR_STATUS_COUNTERS 36 +#define NUM_SEND_DMA_ERR_STATUS_COUNTERS 4 +#define NUM_SEND_EGRESS_ERR_STATUS_COUNTERS 64 +#define NUM_SEND_ERR_STATUS_COUNTERS 3 +#define NUM_SEND_CTXT_ERR_STATUS_COUNTERS 5 +#define NUM_SEND_DMA_ENG_ERR_STATUS_COUNTERS 24 + +/* + * per driver stats, either not device nor port-specific, or + * summed over all of the devices and ports. + * They are described by name via ipathfs filesystem, so layout + * and number of elements can change without breaking compatibility. + * If members are added or deleted hfi1_statnames[] in debugfs.c must + * change to match. + */ +struct hfi1_ib_stats { + __u64 sps_ints; /* number of interrupts handled */ + __u64 sps_errints; /* number of error interrupts */ + __u64 sps_txerrs; /* tx-related packet errors */ + __u64 sps_rcverrs; /* non-crc rcv packet errors */ + __u64 sps_hwerrs; /* hardware errors reported (parity, etc.) */ + __u64 sps_nopiobufs; /* no pio bufs avail from kernel */ + __u64 sps_ctxts; /* number of contexts currently open */ + __u64 sps_lenerrs; /* number of kernel packets where RHF != LRH len */ + __u64 sps_buffull; + __u64 sps_hdrfull; +}; + +extern struct hfi1_ib_stats hfi1_stats; +extern const struct pci_error_handlers hfi1_pci_err_handler; + +/* + * First-cut criterion for "device is active" is + * two thousand dwords combined Tx, Rx traffic per + * 5-second interval. SMA packets are 64 dwords, + * and occur "a few per second", presumably each way. + */ +#define HFI1_TRAFFIC_ACTIVE_THRESHOLD (2000) + +/* + * Below contains all data related to a single context (formerly called port). + */ + +#ifdef CONFIG_DEBUG_FS +struct hfi1_opcode_stats_perctx; +#endif + +struct ctxt_eager_bufs { + ssize_t size; /* total size of eager buffers */ + u32 count; /* size of buffers array */ + u32 numbufs; /* number of buffers allocated */ + u32 alloced; /* number of rcvarray entries used */ + u32 rcvtid_size; /* size of each eager rcv tid */ + u32 threshold; /* head update threshold */ + struct eager_buffer { + void *addr; + dma_addr_t phys; + ssize_t len; + } *buffers; + struct { + void *addr; + dma_addr_t phys; + } *rcvtids; +}; + +struct exp_tid_set { + struct list_head list; + u32 count; +}; + +struct hfi1_ctxtdata { + /* shadow the ctxt's RcvCtrl register */ + u64 rcvctrl; + /* rcvhdrq base, needs mmap before useful */ + void *rcvhdrq; + /* kernel virtual address where hdrqtail is updated */ + volatile __le64 *rcvhdrtail_kvaddr; + /* + * Shared page for kernel to signal user processes that send buffers + * need disarming. The process should call HFI1_CMD_DISARM_BUFS + * or HFI1_CMD_ACK_EVENT with IPATH_EVENT_DISARM_BUFS set. + */ + unsigned long *user_event_mask; + /* when waiting for rcv or pioavail */ + wait_queue_head_t wait; + /* rcvhdrq size (for freeing) */ + size_t rcvhdrq_size; + /* number of rcvhdrq entries */ + u16 rcvhdrq_cnt; + /* size of each of the rcvhdrq entries */ + u16 rcvhdrqentsize; + /* mmap of hdrq, must fit in 44 bits */ + dma_addr_t rcvhdrq_phys; + dma_addr_t rcvhdrqtailaddr_phys; + struct ctxt_eager_bufs egrbufs; + /* this receive context's assigned PIO ACK send context */ + struct send_context *sc; + + /* dynamic receive available interrupt timeout */ + u32 rcvavail_timeout; + /* + * number of opens (including slave sub-contexts) on this instance + * (ignoring forks, dup, etc. for now) + */ + int cnt; + /* + * how much space to leave at start of eager TID entries for + * protocol use, on each TID + */ + /* instead of calculating it */ + unsigned ctxt; + /* non-zero if ctxt is being shared. */ + u16 subctxt_cnt; + /* non-zero if ctxt is being shared. */ + u16 subctxt_id; + u8 uuid[16]; + /* job key */ + u16 jkey; + /* number of RcvArray groups for this context. */ + u32 rcv_array_groups; + /* index of first eager TID entry. */ + u32 eager_base; + /* number of expected TID entries */ + u32 expected_count; + /* index of first expected TID entry. */ + u32 expected_base; + + struct exp_tid_set tid_group_list; + struct exp_tid_set tid_used_list; + struct exp_tid_set tid_full_list; + + /* lock protecting all Expected TID data */ + struct mutex exp_lock; + /* number of pio bufs for this ctxt (all procs, if shared) */ + u32 piocnt; + /* first pio buffer for this ctxt */ + u32 pio_base; + /* chip offset of PIO buffers for this ctxt */ + u32 piobufs; + /* per-context configuration flags */ + u32 flags; + /* per-context event flags for fileops/intr communication */ + unsigned long event_flags; + /* WAIT_RCV that timed out, no interrupt */ + u32 rcvwait_to; + /* WAIT_PIO that timed out, no interrupt */ + u32 piowait_to; + /* WAIT_RCV already happened, no wait */ + u32 rcvnowait; + /* WAIT_PIO already happened, no wait */ + u32 pionowait; + /* total number of polled urgent packets */ + u32 urgent; + /* saved total number of polled urgent packets for poll edge trigger */ + u32 urgent_poll; + /* pid of process using this ctxt */ + pid_t pid; + pid_t subpid[HFI1_MAX_SHARED_CTXTS]; + /* same size as task_struct .comm[], command that opened context */ + char comm[TASK_COMM_LEN]; + /* so file ops can get at unit */ + struct hfi1_devdata *dd; + /* so functions that need physical port can get it easily */ + struct hfi1_pportdata *ppd; + /* A page of memory for rcvhdrhead, rcvegrhead, rcvegrtail * N */ + void *subctxt_uregbase; + /* An array of pages for the eager receive buffers * N */ + void *subctxt_rcvegrbuf; + /* An array of pages for the eager header queue entries * N */ + void *subctxt_rcvhdr_base; + /* The version of the library which opened this ctxt */ + u32 userversion; + /* Bitmask of active slaves */ + u32 active_slaves; + /* Type of packets or conditions we want to poll for */ + u16 poll_type; + /* receive packet sequence counter */ + u8 seq_cnt; + u8 redirect_seq_cnt; + /* ctxt rcvhdrq head offset */ + u32 head; + u32 pkt_count; + /* QPs waiting for context processing */ + struct list_head qp_wait_list; + /* interrupt handling */ + u64 imask; /* clear interrupt mask */ + int ireg; /* clear interrupt register */ + unsigned numa_id; /* numa node of this context */ + /* verbs stats per CTX */ + struct hfi1_opcode_stats_perctx *opstats; + /* + * This is the kernel thread that will keep making + * progress on the user sdma requests behind the scenes. + * There is one per context (shared contexts use the master's). + */ + struct task_struct *progress; + struct list_head sdma_queues; + /* protect sdma queues */ + spinlock_t sdma_qlock; + + /* Is ASPM interrupt supported for this context */ + bool aspm_intr_supported; + /* ASPM state (enabled/disabled) for this context */ + bool aspm_enabled; + /* Timer for re-enabling ASPM if interrupt activity quietens down */ + struct timer_list aspm_timer; + /* Lock to serialize between intr, timer intr and user threads */ + spinlock_t aspm_lock; + /* Is ASPM processing enabled for this context (in intr context) */ + bool aspm_intr_enable; + /* Last interrupt timestamp */ + ktime_t aspm_ts_last_intr; + /* Last timestamp at which we scheduled a timer for this context */ + ktime_t aspm_ts_timer_sched; + + /* + * The interrupt handler for a particular receive context can vary + * throughout it's lifetime. This is not a lock protected data member so + * it must be updated atomically and the prev and new value must always + * be valid. Worst case is we process an extra interrupt and up to 64 + * packets with the wrong interrupt handler. + */ + int (*do_interrupt)(struct hfi1_ctxtdata *rcd, int threaded); +}; + +/* + * Represents a single packet at a high level. Put commonly computed things in + * here so we do not have to keep doing them over and over. The rule of thumb is + * if something is used one time to derive some value, store that something in + * here. If it is used multiple times, then store the result of that derivation + * in here. + */ +struct hfi1_packet { + void *ebuf; + void *hdr; + struct hfi1_ctxtdata *rcd; + __le32 *rhf_addr; + struct rvt_qp *qp; + struct hfi1_other_headers *ohdr; + u64 rhf; + u32 maxcnt; + u32 rhqoff; + u32 hdrqtail; + int numpkt; + u16 tlen; + u16 hlen; + s16 etail; + u16 rsize; + u8 updegr; + u8 rcv_flags; + u8 etype; +}; + +static inline bool has_sc4_bit(struct hfi1_packet *p) +{ + return !!rhf_dc_info(p->rhf); +} + +/* + * Private data for snoop/capture support. + */ +struct hfi1_snoop_data { + int mode_flag; + struct cdev cdev; + struct device *class_dev; + /* protect snoop data */ + spinlock_t snoop_lock; + struct list_head queue; + wait_queue_head_t waitq; + void *filter_value; + int (*filter_callback)(void *hdr, void *data, void *value); + u64 dcc_cfg; /* saved value of DCC Cfg register */ +}; + +/* snoop mode_flag values */ +#define HFI1_PORT_SNOOP_MODE 1U +#define HFI1_PORT_CAPTURE_MODE 2U + +struct rvt_sge_state; + +/* + * Get/Set IB link-level config parameters for f_get/set_ib_cfg() + * Mostly for MADs that set or query link parameters, also ipath + * config interfaces + */ +#define HFI1_IB_CFG_LIDLMC 0 /* LID (LS16b) and Mask (MS16b) */ +#define HFI1_IB_CFG_LWID_DG_ENB 1 /* allowed Link-width downgrade */ +#define HFI1_IB_CFG_LWID_ENB 2 /* allowed Link-width */ +#define HFI1_IB_CFG_LWID 3 /* currently active Link-width */ +#define HFI1_IB_CFG_SPD_ENB 4 /* allowed Link speeds */ +#define HFI1_IB_CFG_SPD 5 /* current Link spd */ +#define HFI1_IB_CFG_RXPOL_ENB 6 /* Auto-RX-polarity enable */ +#define HFI1_IB_CFG_LREV_ENB 7 /* Auto-Lane-reversal enable */ +#define HFI1_IB_CFG_LINKLATENCY 8 /* Link Latency (IB1.2 only) */ +#define HFI1_IB_CFG_HRTBT 9 /* IB heartbeat off/enable/auto; DDR/QDR only */ +#define HFI1_IB_CFG_OP_VLS 10 /* operational VLs */ +#define HFI1_IB_CFG_VL_HIGH_CAP 11 /* num of VL high priority weights */ +#define HFI1_IB_CFG_VL_LOW_CAP 12 /* num of VL low priority weights */ +#define HFI1_IB_CFG_OVERRUN_THRESH 13 /* IB overrun threshold */ +#define HFI1_IB_CFG_PHYERR_THRESH 14 /* IB PHY error threshold */ +#define HFI1_IB_CFG_LINKDEFAULT 15 /* IB link default (sleep/poll) */ +#define HFI1_IB_CFG_PKEYS 16 /* update partition keys */ +#define HFI1_IB_CFG_MTU 17 /* update MTU in IBC */ +#define HFI1_IB_CFG_VL_HIGH_LIMIT 19 +#define HFI1_IB_CFG_PMA_TICKS 20 /* PMA sample tick resolution */ +#define HFI1_IB_CFG_PORT 21 /* switch port we are connected to */ + +/* + * HFI or Host Link States + * + * These describe the states the driver thinks the logical and physical + * states are in. Used as an argument to set_link_state(). Implemented + * as bits for easy multi-state checking. The actual state can only be + * one. + */ +#define __HLS_UP_INIT_BP 0 +#define __HLS_UP_ARMED_BP 1 +#define __HLS_UP_ACTIVE_BP 2 +#define __HLS_DN_DOWNDEF_BP 3 /* link down default */ +#define __HLS_DN_POLL_BP 4 +#define __HLS_DN_DISABLE_BP 5 +#define __HLS_DN_OFFLINE_BP 6 +#define __HLS_VERIFY_CAP_BP 7 +#define __HLS_GOING_UP_BP 8 +#define __HLS_GOING_OFFLINE_BP 9 +#define __HLS_LINK_COOLDOWN_BP 10 + +#define HLS_UP_INIT BIT(__HLS_UP_INIT_BP) +#define HLS_UP_ARMED BIT(__HLS_UP_ARMED_BP) +#define HLS_UP_ACTIVE BIT(__HLS_UP_ACTIVE_BP) +#define HLS_DN_DOWNDEF BIT(__HLS_DN_DOWNDEF_BP) /* link down default */ +#define HLS_DN_POLL BIT(__HLS_DN_POLL_BP) +#define HLS_DN_DISABLE BIT(__HLS_DN_DISABLE_BP) +#define HLS_DN_OFFLINE BIT(__HLS_DN_OFFLINE_BP) +#define HLS_VERIFY_CAP BIT(__HLS_VERIFY_CAP_BP) +#define HLS_GOING_UP BIT(__HLS_GOING_UP_BP) +#define HLS_GOING_OFFLINE BIT(__HLS_GOING_OFFLINE_BP) +#define HLS_LINK_COOLDOWN BIT(__HLS_LINK_COOLDOWN_BP) + +#define HLS_UP (HLS_UP_INIT | HLS_UP_ARMED | HLS_UP_ACTIVE) +#define HLS_DOWN ~(HLS_UP) + +/* use this MTU size if none other is given */ +#define HFI1_DEFAULT_ACTIVE_MTU 10240 +/* use this MTU size as the default maximum */ +#define HFI1_DEFAULT_MAX_MTU 10240 +/* default partition key */ +#define DEFAULT_PKEY 0xffff + +/* + * Possible fabric manager config parameters for fm_{get,set}_table() + */ +#define FM_TBL_VL_HIGH_ARB 1 /* Get/set VL high prio weights */ +#define FM_TBL_VL_LOW_ARB 2 /* Get/set VL low prio weights */ +#define FM_TBL_BUFFER_CONTROL 3 /* Get/set Buffer Control */ +#define FM_TBL_SC2VLNT 4 /* Get/set SC->VLnt */ +#define FM_TBL_VL_PREEMPT_ELEMS 5 /* Get (no set) VL preempt elems */ +#define FM_TBL_VL_PREEMPT_MATRIX 6 /* Get (no set) VL preempt matrix */ + +/* + * Possible "operations" for f_rcvctrl(ppd, op, ctxt) + * these are bits so they can be combined, e.g. + * HFI1_RCVCTRL_INTRAVAIL_ENB | HFI1_RCVCTRL_CTXT_ENB + */ +#define HFI1_RCVCTRL_TAILUPD_ENB 0x01 +#define HFI1_RCVCTRL_TAILUPD_DIS 0x02 +#define HFI1_RCVCTRL_CTXT_ENB 0x04 +#define HFI1_RCVCTRL_CTXT_DIS 0x08 +#define HFI1_RCVCTRL_INTRAVAIL_ENB 0x10 +#define HFI1_RCVCTRL_INTRAVAIL_DIS 0x20 +#define HFI1_RCVCTRL_PKEY_ENB 0x40 /* Note, default is enabled */ +#define HFI1_RCVCTRL_PKEY_DIS 0x80 +#define HFI1_RCVCTRL_TIDFLOW_ENB 0x0400 +#define HFI1_RCVCTRL_TIDFLOW_DIS 0x0800 +#define HFI1_RCVCTRL_ONE_PKT_EGR_ENB 0x1000 +#define HFI1_RCVCTRL_ONE_PKT_EGR_DIS 0x2000 +#define HFI1_RCVCTRL_NO_RHQ_DROP_ENB 0x4000 +#define HFI1_RCVCTRL_NO_RHQ_DROP_DIS 0x8000 +#define HFI1_RCVCTRL_NO_EGR_DROP_ENB 0x10000 +#define HFI1_RCVCTRL_NO_EGR_DROP_DIS 0x20000 + +/* partition enforcement flags */ +#define HFI1_PART_ENFORCE_IN 0x1 +#define HFI1_PART_ENFORCE_OUT 0x2 + +/* how often we check for synthetic counter wrap around */ +#define SYNTH_CNT_TIME 2 + +/* Counter flags */ +#define CNTR_NORMAL 0x0 /* Normal counters, just read register */ +#define CNTR_SYNTH 0x1 /* Synthetic counters, saturate at all 1s */ +#define CNTR_DISABLED 0x2 /* Disable this counter */ +#define CNTR_32BIT 0x4 /* Simulate 64 bits for this counter */ +#define CNTR_VL 0x8 /* Per VL counter */ +#define CNTR_SDMA 0x10 +#define CNTR_INVALID_VL -1 /* Specifies invalid VL */ +#define CNTR_MODE_W 0x0 +#define CNTR_MODE_R 0x1 + +/* VLs Supported/Operational */ +#define HFI1_MIN_VLS_SUPPORTED 1 +#define HFI1_MAX_VLS_SUPPORTED 8 + +static inline void incr_cntr64(u64 *cntr) +{ + if (*cntr < (u64)-1LL) + (*cntr)++; +} + +static inline void incr_cntr32(u32 *cntr) +{ + if (*cntr < (u32)-1LL) + (*cntr)++; +} + +#define MAX_NAME_SIZE 64 +struct hfi1_msix_entry { + enum irq_type type; + struct msix_entry msix; + void *arg; + char name[MAX_NAME_SIZE]; + cpumask_t mask; +}; + +/* per-SL CCA information */ +struct cca_timer { + struct hrtimer hrtimer; + struct hfi1_pportdata *ppd; /* read-only */ + int sl; /* read-only */ + u16 ccti; /* read/write - current value of CCTI */ +}; + +struct link_down_reason { + /* + * SMA-facing value. Should be set from .latest when + * HLS_UP_* -> HLS_DN_* transition actually occurs. + */ + u8 sma; + u8 latest; +}; + +enum { + LO_PRIO_TABLE, + HI_PRIO_TABLE, + MAX_PRIO_TABLE +}; + +struct vl_arb_cache { + /* protect vl arb cache */ + spinlock_t lock; + struct ib_vl_weight_elem table[VL_ARB_TABLE_SIZE]; +}; + +/* + * The structure below encapsulates data relevant to a physical IB Port. + * Current chips support only one such port, but the separation + * clarifies things a bit. Note that to conform to IB conventions, + * port-numbers are one-based. The first or only port is port1. + */ +struct hfi1_pportdata { + struct hfi1_ibport ibport_data; + + struct hfi1_devdata *dd; + struct kobject pport_cc_kobj; + struct kobject sc2vl_kobj; + struct kobject sl2sc_kobj; + struct kobject vl2mtu_kobj; + + /* PHY support */ + u32 port_type; + struct qsfp_data qsfp_info; + + /* GUID for this interface, in host order */ + u64 guid; + /* GUID for peer interface, in host order */ + u64 neighbor_guid; + + /* up or down physical link state */ + u32 linkup; + + /* + * this address is mapped read-only into user processes so they can + * get status cheaply, whenever they want. One qword of status per port + */ + u64 *statusp; + + /* SendDMA related entries */ + + struct workqueue_struct *hfi1_wq; + + /* move out of interrupt context */ + struct work_struct link_vc_work; + struct work_struct link_up_work; + struct work_struct link_down_work; + struct work_struct sma_message_work; + struct work_struct freeze_work; + struct work_struct link_downgrade_work; + struct work_struct link_bounce_work; + /* host link state variables */ + struct mutex hls_lock; + u32 host_link_state; + + spinlock_t sdma_alllock ____cacheline_aligned_in_smp; + + u32 lstate; /* logical link state */ + + /* these are the "32 bit" regs */ + + u32 ibmtu; /* The MTU programmed for this unit */ + /* + * Current max size IB packet (in bytes) including IB headers, that + * we can send. Changes when ibmtu changes. + */ + u32 ibmaxlen; + u32 current_egress_rate; /* units [10^6 bits/sec] */ + /* LID programmed for this instance */ + u16 lid; + /* list of pkeys programmed; 0 if not set */ + u16 pkeys[MAX_PKEY_VALUES]; + u16 link_width_supported; + u16 link_width_downgrade_supported; + u16 link_speed_supported; + u16 link_width_enabled; + u16 link_width_downgrade_enabled; + u16 link_speed_enabled; + u16 link_width_active; + u16 link_width_downgrade_tx_active; + u16 link_width_downgrade_rx_active; + u16 link_speed_active; + u8 vls_supported; + u8 vls_operational; + u8 actual_vls_operational; + /* LID mask control */ + u8 lmc; + /* Rx Polarity inversion (compensate for ~tx on partner) */ + u8 rx_pol_inv; + + u8 hw_pidx; /* physical port index */ + u8 port; /* IB port number and index into dd->pports - 1 */ + /* type of neighbor node */ + u8 neighbor_type; + u8 neighbor_normal; + u8 neighbor_fm_security; /* 1 if firmware checking is disabled */ + u8 neighbor_port_number; + u8 is_sm_config_started; + u8 offline_disabled_reason; + u8 is_active_optimize_enabled; + u8 driver_link_ready; /* driver ready for active link */ + u8 link_enabled; /* link enabled? */ + u8 linkinit_reason; + u8 local_tx_rate; /* rate given to 8051 firmware */ + u8 last_pstate; /* info only */ + + /* placeholders for IB MAD packet settings */ + u8 overrun_threshold; + u8 phy_error_threshold; + + /* Used to override LED behavior for things like maintenance beaconing*/ + /* + * Alternates per phase of blink + * [0] holds LED off duration, [1] holds LED on duration + */ + unsigned long led_override_vals[2]; + u8 led_override_phase; /* LSB picks from vals[] */ + atomic_t led_override_timer_active; + /* Used to flash LEDs in override mode */ + struct timer_list led_override_timer; + + u32 sm_trap_qp; + u32 sa_qp; + + /* + * cca_timer_lock protects access to the per-SL cca_timer + * structures (specifically the ccti member). + */ + spinlock_t cca_timer_lock ____cacheline_aligned_in_smp; + struct cca_timer cca_timer[OPA_MAX_SLS]; + + /* List of congestion control table entries */ + struct ib_cc_table_entry_shadow ccti_entries[CC_TABLE_SHADOW_MAX]; + + /* congestion entries, each entry corresponding to a SL */ + struct opa_congestion_setting_entry_shadow + congestion_entries[OPA_MAX_SLS]; + + /* + * cc_state_lock protects (write) access to the per-port + * struct cc_state. + */ + spinlock_t cc_state_lock ____cacheline_aligned_in_smp; + + struct cc_state __rcu *cc_state; + + /* Total number of congestion control table entries */ + u16 total_cct_entry; + + /* Bit map identifying service level */ + u32 cc_sl_control_map; + + /* CA's max number of 64 entry units in the congestion control table */ + u8 cc_max_table_entries; + + /* + * begin congestion log related entries + * cc_log_lock protects all congestion log related data + */ + spinlock_t cc_log_lock ____cacheline_aligned_in_smp; + u8 threshold_cong_event_map[OPA_MAX_SLS / 8]; + u16 threshold_event_counter; + struct opa_hfi1_cong_log_event_internal cc_events[OPA_CONG_LOG_ELEMS]; + int cc_log_idx; /* index for logging events */ + int cc_mad_idx; /* index for reporting events */ + /* end congestion log related entries */ + + struct vl_arb_cache vl_arb_cache[MAX_PRIO_TABLE]; + + /* port relative counter buffer */ + u64 *cntrs; + /* port relative synthetic counter buffer */ + u64 *scntrs; + /* port_xmit_discards are synthesized from different egress errors */ + u64 port_xmit_discards; + u64 port_xmit_discards_vl[C_VL_COUNT]; + u64 port_xmit_constraint_errors; + u64 port_rcv_constraint_errors; + /* count of 'link_err' interrupts from DC */ + u64 link_downed; + /* number of times link retrained successfully */ + u64 link_up; + /* number of times a link unknown frame was reported */ + u64 unknown_frame_count; + /* port_ltp_crc_mode is returned in 'portinfo' MADs */ + u16 port_ltp_crc_mode; + /* port_crc_mode_enabled is the crc we support */ + u8 port_crc_mode_enabled; + /* mgmt_allowed is also returned in 'portinfo' MADs */ + u8 mgmt_allowed; + u8 part_enforce; /* partition enforcement flags */ + struct link_down_reason local_link_down_reason; + struct link_down_reason neigh_link_down_reason; + /* Value to be sent to link peer on LinkDown .*/ + u8 remote_link_down_reason; + /* Error events that will cause a port bounce. */ + u32 port_error_action; + struct work_struct linkstate_active_work; + /* Does this port need to prescan for FECNs */ + bool cc_prescan; +}; + +typedef int (*rhf_rcv_function_ptr)(struct hfi1_packet *packet); + +typedef void (*opcode_handler)(struct hfi1_packet *packet); + +/* return values for the RHF receive functions */ +#define RHF_RCV_CONTINUE 0 /* keep going */ +#define RHF_RCV_DONE 1 /* stop, this packet processed */ +#define RHF_RCV_REPROCESS 2 /* stop. retain this packet */ + +struct rcv_array_data { + u8 group_size; + u16 ngroups; + u16 nctxt_extra; +}; + +struct per_vl_data { + u16 mtu; + struct send_context *sc; +}; + +/* 16 to directly index */ +#define PER_VL_SEND_CONTEXTS 16 + +struct err_info_rcvport { + u8 status_and_code; + u64 packet_flit1; + u64 packet_flit2; +}; + +struct err_info_constraint { + u8 status; + u16 pkey; + u32 slid; +}; + +struct hfi1_temp { + unsigned int curr; /* current temperature */ + unsigned int lo_lim; /* low temperature limit */ + unsigned int hi_lim; /* high temperature limit */ + unsigned int crit_lim; /* critical temperature limit */ + u8 triggers; /* temperature triggers */ +}; + +/* common data between shared ASIC HFIs */ +struct hfi1_asic_data { + struct hfi1_devdata *dds[2]; /* back pointers */ + struct mutex asic_resource_mutex; +}; + +/* device data struct now contains only "general per-device" info. + * fields related to a physical IB port are in a hfi1_pportdata struct. + */ +struct sdma_engine; +struct sdma_vl_map; + +#define BOARD_VERS_MAX 96 /* how long the version string can be */ +#define SERIAL_MAX 16 /* length of the serial number */ + +typedef int (*send_routine)(struct rvt_qp *, struct hfi1_pkt_state *, u64); +struct hfi1_devdata { + struct hfi1_ibdev verbs_dev; /* must be first */ + struct list_head list; + /* pointers to related structs for this device */ + /* pci access data structure */ + struct pci_dev *pcidev; + struct cdev user_cdev; + struct cdev diag_cdev; + struct cdev ui_cdev; + struct device *user_device; + struct device *diag_device; + struct device *ui_device; + + /* mem-mapped pointer to base of chip regs */ + u8 __iomem *kregbase; + /* end of mem-mapped chip space excluding sendbuf and user regs */ + u8 __iomem *kregend; + /* physical address of chip for io_remap, etc. */ + resource_size_t physaddr; + /* receive context data */ + struct hfi1_ctxtdata **rcd; + /* send context data */ + struct send_context_info *send_contexts; + /* map hardware send contexts to software index */ + u8 *hw_to_sw; + /* spinlock for allocating and releasing send context resources */ + spinlock_t sc_lock; + /* Per VL data. Enough for all VLs but not all elements are set/used. */ + struct per_vl_data vld[PER_VL_SEND_CONTEXTS]; + /* lock for pio_map */ + spinlock_t pio_map_lock; + /* array of kernel send contexts */ + struct send_context **kernel_send_context; + /* array of vl maps */ + struct pio_vl_map __rcu *pio_map; + /* seqlock for sc2vl */ + seqlock_t sc2vl_lock; + u64 sc2vl[4]; + /* Send Context initialization lock. */ + spinlock_t sc_init_lock; + + /* fields common to all SDMA engines */ + + /* default flags to last descriptor */ + u64 default_desc1; + volatile __le64 *sdma_heads_dma; /* DMA'ed by chip */ + dma_addr_t sdma_heads_phys; + void *sdma_pad_dma; /* DMA'ed by chip */ + dma_addr_t sdma_pad_phys; + /* for deallocation */ + size_t sdma_heads_size; + /* number from the chip */ + u32 chip_sdma_engines; + /* num used */ + u32 num_sdma; + /* lock for sdma_map */ + spinlock_t sde_map_lock; + /* array of engines sized by num_sdma */ + struct sdma_engine *per_sdma; + /* array of vl maps */ + struct sdma_vl_map __rcu *sdma_map; + /* SPC freeze waitqueue and variable */ + wait_queue_head_t sdma_unfreeze_wq; + atomic_t sdma_unfreeze_count; + + /* common data between shared ASIC HFIs in this OS */ + struct hfi1_asic_data *asic_data; + + /* hfi1_pportdata, points to array of (physical) port-specific + * data structs, indexed by pidx (0..n-1) + */ + struct hfi1_pportdata *pport; + + /* mem-mapped pointer to base of PIO buffers */ + void __iomem *piobase; + /* + * write-combining mem-mapped pointer to base of RcvArray + * memory. + */ + void __iomem *rcvarray_wc; + /* + * credit return base - a per-NUMA range of DMA address that + * the chip will use to update the per-context free counter + */ + struct credit_return_base *cr_base; + + /* send context numbers and sizes for each type */ + struct sc_config_sizes sc_sizes[SC_MAX]; + + u32 lcb_access_count; /* count of LCB users */ + + char *boardname; /* human readable board info */ + + /* device (not port) flags, basically device capabilities */ + u32 flags; + + /* reset value */ + u64 z_int_counter; + u64 z_rcv_limit; + u64 z_send_schedule; + /* percpu int_counter */ + u64 __percpu *int_counter; + u64 __percpu *rcv_limit; + u64 __percpu *send_schedule; + /* number of receive contexts in use by the driver */ + u32 num_rcv_contexts; + /* number of pio send contexts in use by the driver */ + u32 num_send_contexts; + /* + * number of ctxts available for PSM open + */ + u32 freectxts; + /* total number of available user/PSM contexts */ + u32 num_user_contexts; + /* base receive interrupt timeout, in CSR units */ + u32 rcv_intr_timeout_csr; + + u64 __iomem *egrtidbase; + spinlock_t sendctrl_lock; /* protect changes to SendCtrl */ + spinlock_t rcvctrl_lock; /* protect changes to RcvCtrl */ + /* around rcd and (user ctxts) ctxt_cnt use (intr vs free) */ + spinlock_t uctxt_lock; /* rcd and user context changes */ + /* exclusive access to 8051 */ + spinlock_t dc8051_lock; + /* exclusive access to 8051 memory */ + spinlock_t dc8051_memlock; + int dc8051_timed_out; /* remember if the 8051 timed out */ + /* + * A page that will hold event notification bitmaps for all + * contexts. This page will be mapped into all processes. + */ + unsigned long *events; + /* + * per unit status, see also portdata statusp + * mapped read-only into user processes so they can get unit and + * IB link status cheaply + */ + struct hfi1_status *status; + u32 freezelen; /* max length of freezemsg */ + + /* revision register shadow */ + u64 revision; + /* Base GUID for device (network order) */ + u64 base_guid; + + /* these are the "32 bit" regs */ + + /* value we put in kr_rcvhdrsize */ + u32 rcvhdrsize; + /* number of receive contexts the chip supports */ + u32 chip_rcv_contexts; + /* number of receive array entries */ + u32 chip_rcv_array_count; + /* number of PIO send contexts the chip supports */ + u32 chip_send_contexts; + /* number of bytes in the PIO memory buffer */ + u32 chip_pio_mem_size; + /* number of bytes in the SDMA memory buffer */ + u32 chip_sdma_mem_size; + + /* size of each rcvegrbuffer */ + u32 rcvegrbufsize; + /* log2 of above */ + u16 rcvegrbufsize_shift; + /* both sides of the PCIe link are gen3 capable */ + u8 link_gen3_capable; + /* localbus width (1, 2,4,8,16,32) from config space */ + u32 lbus_width; + /* localbus speed in MHz */ + u32 lbus_speed; + int unit; /* unit # of this chip */ + int node; /* home node of this chip */ + + /* save these PCI fields to restore after a reset */ + u32 pcibar0; + u32 pcibar1; + u32 pci_rom; + u16 pci_command; + u16 pcie_devctl; + u16 pcie_lnkctl; + u16 pcie_devctl2; + u32 pci_msix0; + u32 pci_lnkctl3; + u32 pci_tph2; + + /* + * ASCII serial number, from flash, large enough for original + * all digit strings, and longer serial number format + */ + u8 serial[SERIAL_MAX]; + /* human readable board version */ + u8 boardversion[BOARD_VERS_MAX]; + u8 lbus_info[32]; /* human readable localbus info */ + /* chip major rev, from CceRevision */ + u8 majrev; + /* chip minor rev, from CceRevision */ + u8 minrev; + /* hardware ID */ + u8 hfi1_id; + /* implementation code */ + u8 icode; + /* default link down value (poll/sleep) */ + u8 link_default; + /* vAU of this device */ + u8 vau; + /* vCU of this device */ + u8 vcu; + /* link credits of this device */ + u16 link_credits; + /* initial vl15 credits to use */ + u16 vl15_init; + + /* Misc small ints */ + /* Number of physical ports available */ + u8 num_pports; + /* Lowest context number which can be used by user processes */ + u8 first_user_ctxt; + u8 n_krcv_queues; + u8 qos_shift; + u8 qpn_mask; + + u16 rhf_offset; /* offset of RHF within receive header entry */ + u16 irev; /* implementation revision */ + u16 dc8051_ver; /* 8051 firmware version */ + + struct platform_config platform_config; + struct platform_config_cache pcfg_cache; + + struct diag_client *diag_client; + spinlock_t hfi1_diag_trans_lock; /* protect diag observer ops */ + + u8 psxmitwait_supported; + /* cycle length of PS* counters in HW (in picoseconds) */ + u16 psxmitwait_check_rate; + /* high volume overflow errors deferred to tasklet */ + struct tasklet_struct error_tasklet; + + /* MSI-X information */ + struct hfi1_msix_entry *msix_entries; + u32 num_msix_entries; + + /* INTx information */ + u32 requested_intx_irq; /* did we request one? */ + char intx_name[MAX_NAME_SIZE]; /* INTx name */ + + /* general interrupt: mask of handled interrupts */ + u64 gi_mask[CCE_NUM_INT_CSRS]; + + struct rcv_array_data rcv_entries; + + /* + * 64 bit synthetic counters + */ + struct timer_list synth_stats_timer; + + /* + * device counters + */ + char *cntrnames; + size_t cntrnameslen; + size_t ndevcntrs; + u64 *cntrs; + u64 *scntrs; + + /* + * remembered values for synthetic counters + */ + u64 last_tx; + u64 last_rx; + + /* + * per-port counters + */ + size_t nportcntrs; + char *portcntrnames; + size_t portcntrnameslen; + + struct hfi1_snoop_data hfi1_snoop; + + struct err_info_rcvport err_info_rcvport; + struct err_info_constraint err_info_rcv_constraint; + struct er |