linux/drivers/infiniband/hw/qib/qib.h
Mike Marciniszyn e67306a380 IB/qib: Defer HCA error events to tasklet
With ib_qib options:

    options ib_qib krcvqs=1 pcie_caps=0x51 rcvhdrcnt=4096 singleport=1 ibmtu=4

a run of ib_write_bw -a yields the following:

    ------------------------------------------------------------------
     #bytes     #iterations    BW peak[MB/sec]    BW average[MB/sec]
     1048576   5000           2910.64            229.80
    ------------------------------------------------------------------

The top cpu use in a profile is:

    CPU: Intel Architectural Perfmon, speed 2400.15 MHz (estimated)
    Counted CPU_CLK_UNHALTED events (Clock cycles when not halted) with a unit mask
    of 0x00 (No unit mask) count 1002300
    Counted LLC_MISSES events (Last level cache demand requests from this core that
    missed the LLC) with a unit mask of 0x41 (No unit mask) count 10000
    samples  %        samples  %        app name                 symbol name
    15237    29.2642  964      17.1195  ib_qib.ko                qib_7322intr
    12320    23.6618  1040     18.4692  ib_qib.ko                handle_7322_errors
    4106      7.8860  0              0  vmlinux                  vsnprintf


Analysis of the stats, profile, the code, and the annotated profile indicate:
 - All of the overflow interrupts (one per packet overflow) are
   serviced on CPU0 with no mitigation on the frequency.
 - All of the receive interrupts are being serviced by CPU0.  (That is
   the way truescale.cmds statically allocates the kctx IRQs to CPU)
 - The code is spending all of its time servicing QIB_I_C_ERROR
   RcvEgrFullErr interrupts on CPU0, starving the packet receive
   processing.
 - The decode_err routine is very inefficient, using a printf variant
   to format a "%s" and continues to loop when the errs mask has been
   cleared.
 - Both qib_7322intr and handle_7322_errors read pci registers, which
   is very inefficient.

The fix does the following:
 - Adds a tasklet to service QIB_I_C_ERROR
 - Replaces the very inefficient scnprintf() with a memcpy().  A field
   is added to qib_hwerror_msgs to save the sizeof("string") at
   compile time so that a strlen is not needed during err_decode().
 - The most frequent errors (Overflows) are serviced first to exit the
   loop as early as possible.
 - The loop now exits as soon as the errs mask is clear rather than
   fruitlessly looping through the msp array.

With this fix the performance changes to:

    ------------------------------------------------------------------
     #bytes     #iterations    BW peak[MB/sec]    BW average[MB/sec]
     1048576   5000           2990.64            2941.35
    ------------------------------------------------------------------

During testing of the error handling overflow patch, it was determined
that some CPU's were slower when servicing both overflow and receive
interrupts on CPU0 with different MSI interrupt vectors.

This patch adds an option (krcvq01_no_msi) to not use a dedicated MSI
interrupt for kctx's < 2 and to service them on the default interrupt.
For some CPUs, the cost of the interrupt enter/exit is more costly
than then the additional PCI read in the default handler.

Signed-off-by: Mike Marciniszyn <mike.marciniszyn@qlogic.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
2011-07-22 11:56:05 -07:00

1448 lines
48 KiB
C

#ifndef _QIB_KERNEL_H
#define _QIB_KERNEL_H
/*
* Copyright (c) 2006, 2007, 2008, 2009, 2010 QLogic Corporation.
* All rights reserved.
* Copyright (c) 2003, 2004, 2005, 2006 PathScale, Inc. 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.
*/
/*
* This header file is the base header file for qlogic_ib kernel code
* qib_user.h serves a similar purpose for user code.
*/
#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 "qib_common.h"
#include "qib_verbs.h"
/* only s/w major version of QLogic_IB we can handle */
#define QIB_CHIP_VERS_MAJ 2U
/* don't care about this except printing */
#define QIB_CHIP_VERS_MIN 0U
/* The Organization Unique Identifier (Mfg code), and its position in GUID */
#define QIB_OUI 0x001175
#define QIB_OUI_LSB 40
/*
* 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 qib_statnames[] in qib_fs.c must
* change to match.
*/
struct qlogic_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 qlogic_ib_stats qib_stats;
extern struct pci_error_handlers qib_pci_err_handler;
extern struct pci_driver qib_driver;
#define QIB_CHIP_SWVERSION QIB_CHIP_VERS_MAJ
/*
* First-cut critierion 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 QIB_TRAFFIC_ACTIVE_THRESHOLD (2000)
/*
* Struct used to indicate which errors are logged in each of the
* error-counters that are logged to EEPROM. A counter is incremented
* _once_ (saturating at 255) for each event with any bits set in
* the error or hwerror register masks below.
*/
#define QIB_EEP_LOG_CNT (4)
struct qib_eep_log_mask {
u64 errs_to_log;
u64 hwerrs_to_log;
};
/*
* Below contains all data related to a single context (formerly called port).
*/
struct qib_ctxtdata {
void **rcvegrbuf;
dma_addr_t *rcvegrbuf_phys;
/* rcvhdrq base, needs mmap before useful */
void *rcvhdrq;
/* kernel virtual address where hdrqtail is updated */
void *rcvhdrtail_kvaddr;
/*
* temp buffer for expected send setup, allocated at open, instead
* of each setup call
*/
void *tid_pg_list;
/*
* Shared page for kernel to signal user processes that send buffers
* need disarming. The process should call QIB_CMD_DISARM_BUFS
* or QIB_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;
/*
* rcvegr bufs base, physical, must fit
* in 44 bits so 32 bit programs mmap64 44 bit works)
*/
dma_addr_t rcvegr_phys;
/* mmap of hdrq, must fit in 44 bits */
dma_addr_t rcvhdrq_phys;
dma_addr_t rcvhdrqtailaddr_phys;
/*
* 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;
/* number of eager TID entries. */
u16 rcvegrcnt;
/* index of first eager TID entry. */
u16 rcvegr_tid_base;
/* 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;
/* how many alloc_pages() chunks in rcvegrbuf_pages */
u32 rcvegrbuf_chunks;
/* how many egrbufs per chunk */
u32 rcvegrbufs_perchunk;
/* order for rcvegrbuf_pages */
size_t rcvegrbuf_size;
/* rcvhdrq size (for freeing) */
size_t rcvhdrq_size;
/* per-context flags for fileops/intr communication */
unsigned long flag;
/* next expected TID to check when looking for free */
u32 tidcursor;
/* 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[QLOGIC_IB_MAX_SUBCTXT];
/* same size as task_struct .comm[], command that opened context */
char comm[16];
/* pkeys set by this use of this ctxt */
u16 pkeys[4];
/* so file ops can get at unit */
struct qib_devdata *dd;
/* so funcs that need physical port can get it easily */
struct qib_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;
};
struct qib_sge_state;
struct qib_sdma_txreq {
int flags;
int sg_count;
dma_addr_t addr;
void (*callback)(struct qib_sdma_txreq *, int);
u16 start_idx; /* sdma private */
u16 next_descq_idx; /* sdma private */
struct list_head list; /* sdma private */
};
struct qib_sdma_desc {
__le64 qw[2];
};
struct qib_verbs_txreq {
struct qib_sdma_txreq txreq;
struct qib_qp *qp;
struct qib_swqe *wqe;
u32 dwords;
u16 hdr_dwords;
u16 hdr_inx;
struct qib_pio_header *align_buf;
struct qib_mregion *mr;
struct qib_sge_state *ss;
};
#define QIB_SDMA_TXREQ_F_USELARGEBUF 0x1
#define QIB_SDMA_TXREQ_F_HEADTOHOST 0x2
#define QIB_SDMA_TXREQ_F_INTREQ 0x4
#define QIB_SDMA_TXREQ_F_FREEBUF 0x8
#define QIB_SDMA_TXREQ_F_FREEDESC 0x10
#define QIB_SDMA_TXREQ_S_OK 0
#define QIB_SDMA_TXREQ_S_SENDERROR 1
#define QIB_SDMA_TXREQ_S_ABORTED 2
#define QIB_SDMA_TXREQ_S_SHUTDOWN 3
/*
* 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 QIB_IB_CFG_LIDLMC 0 /* LID (LS16b) and Mask (MS16b) */
#define QIB_IB_CFG_LWID_ENB 2 /* allowed Link-width */
#define QIB_IB_CFG_LWID 3 /* currently active Link-width */
#define QIB_IB_CFG_SPD_ENB 4 /* allowed Link speeds */
#define QIB_IB_CFG_SPD 5 /* current Link spd */
#define QIB_IB_CFG_RXPOL_ENB 6 /* Auto-RX-polarity enable */
#define QIB_IB_CFG_LREV_ENB 7 /* Auto-Lane-reversal enable */
#define QIB_IB_CFG_LINKLATENCY 8 /* Link Latency (IB1.2 only) */
#define QIB_IB_CFG_HRTBT 9 /* IB heartbeat off/enable/auto; DDR/QDR only */
#define QIB_IB_CFG_OP_VLS 10 /* operational VLs */
#define QIB_IB_CFG_VL_HIGH_CAP 11 /* num of VL high priority weights */
#define QIB_IB_CFG_VL_LOW_CAP 12 /* num of VL low priority weights */
#define QIB_IB_CFG_OVERRUN_THRESH 13 /* IB overrun threshold */
#define QIB_IB_CFG_PHYERR_THRESH 14 /* IB PHY error threshold */
#define QIB_IB_CFG_LINKDEFAULT 15 /* IB link default (sleep/poll) */
#define QIB_IB_CFG_PKEYS 16 /* update partition keys */
#define QIB_IB_CFG_MTU 17 /* update MTU in IBC */
#define QIB_IB_CFG_LSTATE 18 /* update linkcmd and linkinitcmd in IBC */
#define QIB_IB_CFG_VL_HIGH_LIMIT 19
#define QIB_IB_CFG_PMA_TICKS 20 /* PMA sample tick resolution */
#define QIB_IB_CFG_PORT 21 /* switch port we are connected to */
/*
* for CFG_LSTATE: LINKCMD in upper 16 bits, LINKINITCMD in lower 16
* IB_LINKINITCMD_POLL and SLEEP are also used as set/get values for
* QIB_IB_CFG_LINKDEFAULT cmd
*/
#define IB_LINKCMD_DOWN (0 << 16)
#define IB_LINKCMD_ARMED (1 << 16)
#define IB_LINKCMD_ACTIVE (2 << 16)
#define IB_LINKINITCMD_NOP 0
#define IB_LINKINITCMD_POLL 1
#define IB_LINKINITCMD_SLEEP 2
#define IB_LINKINITCMD_DISABLE 3
/*
* valid states passed to qib_set_linkstate() user call
*/
#define QIB_IB_LINKDOWN 0
#define QIB_IB_LINKARM 1
#define QIB_IB_LINKACTIVE 2
#define QIB_IB_LINKDOWN_ONLY 3
#define QIB_IB_LINKDOWN_SLEEP 4
#define QIB_IB_LINKDOWN_DISABLE 5
/*
* These 7 values (SDR, DDR, and QDR may be ORed for auto-speed
* negotiation) are used for the 3rd argument to path_f_set_ib_cfg
* with cmd QIB_IB_CFG_SPD_ENB, by direct calls or via sysfs. They
* are also the the possible values for qib_link_speed_enabled and active
* The values were chosen to match values used within the IB spec.
*/
#define QIB_IB_SDR 1
#define QIB_IB_DDR 2
#define QIB_IB_QDR 4
#define QIB_DEFAULT_MTU 4096
/* max number of IB ports supported per HCA */
#define QIB_MAX_IB_PORTS 2
/*
* Possible IB config parameters for f_get/set_ib_table()
*/
#define QIB_IB_TBL_VL_HIGH_ARB 1 /* Get/set VL high priority weights */
#define QIB_IB_TBL_VL_LOW_ARB 2 /* Get/set VL low priority weights */
/*
* Possible "operations" for f_rcvctrl(ppd, op, ctxt)
* these are bits so they can be combined, e.g.
* QIB_RCVCTRL_INTRAVAIL_ENB | QIB_RCVCTRL_CTXT_ENB
*/
#define QIB_RCVCTRL_TAILUPD_ENB 0x01
#define QIB_RCVCTRL_TAILUPD_DIS 0x02
#define QIB_RCVCTRL_CTXT_ENB 0x04
#define QIB_RCVCTRL_CTXT_DIS 0x08
#define QIB_RCVCTRL_INTRAVAIL_ENB 0x10
#define QIB_RCVCTRL_INTRAVAIL_DIS 0x20
#define QIB_RCVCTRL_PKEY_ENB 0x40 /* Note, default is enabled */
#define QIB_RCVCTRL_PKEY_DIS 0x80
#define QIB_RCVCTRL_BP_ENB 0x0100
#define QIB_RCVCTRL_BP_DIS 0x0200
#define QIB_RCVCTRL_TIDFLOW_ENB 0x0400
#define QIB_RCVCTRL_TIDFLOW_DIS 0x0800
/*
* Possible "operations" for f_sendctrl(ppd, op, var)
* these are bits so they can be combined, e.g.
* QIB_SENDCTRL_BUFAVAIL_ENB | QIB_SENDCTRL_ENB
* Some operations (e.g. DISARM, ABORT) are known to
* be "one-shot", so do not modify shadow.
*/
#define QIB_SENDCTRL_DISARM (0x1000)
#define QIB_SENDCTRL_DISARM_BUF(bufn) ((bufn) | QIB_SENDCTRL_DISARM)
/* available (0x2000) */
#define QIB_SENDCTRL_AVAIL_DIS (0x4000)
#define QIB_SENDCTRL_AVAIL_ENB (0x8000)
#define QIB_SENDCTRL_AVAIL_BLIP (0x10000)
#define QIB_SENDCTRL_SEND_DIS (0x20000)
#define QIB_SENDCTRL_SEND_ENB (0x40000)
#define QIB_SENDCTRL_FLUSH (0x80000)
#define QIB_SENDCTRL_CLEAR (0x100000)
#define QIB_SENDCTRL_DISARM_ALL (0x200000)
/*
* These are the generic indices for requesting per-port
* counter values via the f_portcntr function. They
* are always returned as 64 bit values, although most
* are 32 bit counters.
*/
/* send-related counters */
#define QIBPORTCNTR_PKTSEND 0U
#define QIBPORTCNTR_WORDSEND 1U
#define QIBPORTCNTR_PSXMITDATA 2U
#define QIBPORTCNTR_PSXMITPKTS 3U
#define QIBPORTCNTR_PSXMITWAIT 4U
#define QIBPORTCNTR_SENDSTALL 5U
/* receive-related counters */
#define QIBPORTCNTR_PKTRCV 6U
#define QIBPORTCNTR_PSRCVDATA 7U
#define QIBPORTCNTR_PSRCVPKTS 8U
#define QIBPORTCNTR_RCVEBP 9U
#define QIBPORTCNTR_RCVOVFL 10U
#define QIBPORTCNTR_WORDRCV 11U
/* IB link related error counters */
#define QIBPORTCNTR_RXLOCALPHYERR 12U
#define QIBPORTCNTR_RXVLERR 13U
#define QIBPORTCNTR_ERRICRC 14U
#define QIBPORTCNTR_ERRVCRC 15U
#define QIBPORTCNTR_ERRLPCRC 16U
#define QIBPORTCNTR_BADFORMAT 17U
#define QIBPORTCNTR_ERR_RLEN 18U
#define QIBPORTCNTR_IBSYMBOLERR 19U
#define QIBPORTCNTR_INVALIDRLEN 20U
#define QIBPORTCNTR_UNSUPVL 21U
#define QIBPORTCNTR_EXCESSBUFOVFL 22U
#define QIBPORTCNTR_ERRLINK 23U
#define QIBPORTCNTR_IBLINKDOWN 24U
#define QIBPORTCNTR_IBLINKERRRECOV 25U
#define QIBPORTCNTR_LLI 26U
/* other error counters */
#define QIBPORTCNTR_RXDROPPKT 27U
#define QIBPORTCNTR_VL15PKTDROP 28U
#define QIBPORTCNTR_ERRPKEY 29U
#define QIBPORTCNTR_KHDROVFL 30U
/* sampling counters (these are actually control registers) */
#define QIBPORTCNTR_PSINTERVAL 31U
#define QIBPORTCNTR_PSSTART 32U
#define QIBPORTCNTR_PSSTAT 33U
/* how often we check for packet activity for "power on hours (in seconds) */
#define ACTIVITY_TIMER 5
/* Below is an opaque struct. Each chip (device) can maintain
* private data needed for its operation, but not germane to the
* rest of the driver. For convenience, we define another that
* is chip-specific, per-port
*/
struct qib_chip_specific;
struct qib_chipport_specific;
enum qib_sdma_states {
qib_sdma_state_s00_hw_down,
qib_sdma_state_s10_hw_start_up_wait,
qib_sdma_state_s20_idle,
qib_sdma_state_s30_sw_clean_up_wait,
qib_sdma_state_s40_hw_clean_up_wait,
qib_sdma_state_s50_hw_halt_wait,
qib_sdma_state_s99_running,
};
enum qib_sdma_events {
qib_sdma_event_e00_go_hw_down,
qib_sdma_event_e10_go_hw_start,
qib_sdma_event_e20_hw_started,
qib_sdma_event_e30_go_running,
qib_sdma_event_e40_sw_cleaned,
qib_sdma_event_e50_hw_cleaned,
qib_sdma_event_e60_hw_halted,
qib_sdma_event_e70_go_idle,
qib_sdma_event_e7220_err_halted,
qib_sdma_event_e7322_err_halted,
qib_sdma_event_e90_timer_tick,
};
extern char *qib_sdma_state_names[];
extern char *qib_sdma_event_names[];
struct sdma_set_state_action {
unsigned op_enable:1;
unsigned op_intenable:1;
unsigned op_halt:1;
unsigned op_drain:1;
unsigned go_s99_running_tofalse:1;
unsigned go_s99_running_totrue:1;
};
struct qib_sdma_state {
struct kref kref;
struct completion comp;
enum qib_sdma_states current_state;
struct sdma_set_state_action *set_state_action;
unsigned current_op;
unsigned go_s99_running;
unsigned first_sendbuf;
unsigned last_sendbuf; /* really last +1 */
/* debugging/devel */
enum qib_sdma_states previous_state;
unsigned previous_op;
enum qib_sdma_events last_event;
};
struct xmit_wait {
struct timer_list timer;
u64 counter;
u8 flags;
struct cache {
u64 psxmitdata;
u64 psrcvdata;
u64 psxmitpkts;
u64 psrcvpkts;
u64 psxmitwait;
} counter_cache;
};
/*
* 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 qib_pportdata {
struct qib_ibport ibport_data;
struct qib_devdata *dd;
struct qib_chippport_specific *cpspec; /* chip-specific per-port */
struct kobject pport_kobj;
struct kobject sl2vl_kobj;
struct kobject diagc_kobj;
/* GUID for this interface, in network order */
__be64 guid;
/* QIB_POLL, etc. link-state specific flags, per port */
u32 lflags;
/* qib_lflags driver is waiting for */
u32 state_wanted;
spinlock_t lflags_lock;
/* number of (port-specific) interrupts for this port -- saturates... */
u32 int_counter;
/* ref count for each pkey */
atomic_t pkeyrefs[4];
/*
* this address is mapped readonly into user processes so they can
* get status cheaply, whenever they want. One qword of status per port
*/
u64 *statusp;
/* SendDMA related entries */
spinlock_t sdma_lock;
struct qib_sdma_state sdma_state;
unsigned long sdma_buf_jiffies;
struct qib_sdma_desc *sdma_descq;
u64 sdma_descq_added;
u64 sdma_descq_removed;
u16 sdma_descq_cnt;
u16 sdma_descq_tail;
u16 sdma_descq_head;
u16 sdma_next_intr;
u16 sdma_reset_wait;
u8 sdma_generation;
struct tasklet_struct sdma_sw_clean_up_task;
struct list_head sdma_activelist;
dma_addr_t sdma_descq_phys;
volatile __le64 *sdma_head_dma; /* DMA'ed by chip */
dma_addr_t sdma_head_phys;
wait_queue_head_t state_wait; /* for state_wanted */
/* HoL blocking for SMP replies */
unsigned hol_state;
struct timer_list hol_timer;
/*
* Shadow copies of registers; size indicates read access size.
* Most of them are readonly, but some are write-only register,
* where we manipulate the bits in the shadow copy, and then write
* the shadow copy to qlogic_ib.
*
* We deliberately make most of these 32 bits, since they have
* restricted range. For any that we read, we won't to generate 32
* bit accesses, since Opteron will generate 2 separate 32 bit HT
* transactions for a 64 bit read, and we want to avoid unnecessary
* bus transactions.
*/
/* This is the 64 bit group */
/* last ibcstatus. opaque outside chip-specific code */
u64 lastibcstat;
/* these are the "32 bit" regs */
/*
* the following two are 32-bit bitmasks, but {test,clear,set}_bit
* all expect bit fields to be "unsigned long"
*/
unsigned long p_rcvctrl; /* shadow per-port rcvctrl */
unsigned long p_sendctrl; /* shadow per-port sendctrl */
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;
/*
* ibmaxlen at init time, limited by chip and by receive buffer
* size. Not changed after init.
*/
u32 init_ibmaxlen;
/* LID programmed for this instance */
u16 lid;
/* list of pkeys programmed; 0 if not set */
u16 pkeys[4];
/* LID mask control */
u8 lmc;
u8 link_width_supported;
u8 link_speed_supported;
u8 link_width_enabled;
u8 link_speed_enabled;
u8 link_width_active;
u8 link_speed_active;
u8 vls_supported;
u8 vls_operational;
/* 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 */
u8 delay_mult;
/* used to override LED behavior */
u8 led_override; /* Substituted for normal value, if non-zero */
u16 led_override_timeoff; /* delta to next timer event */
u8 led_override_vals[2]; /* Alternates per blink-frame */
u8 led_override_phase; /* Just counts, LSB picks from vals[] */
atomic_t led_override_timer_active;
/* Used to flash LEDs in override mode */
struct timer_list led_override_timer;
struct xmit_wait cong_stats;
struct timer_list symerr_clear_timer;
};
/* Observers. Not to be taken lightly, possibly not to ship. */
/*
* If a diag read or write is to (bottom <= offset <= top),
* the "hoook" is called, allowing, e.g. shadows to be
* updated in sync with the driver. struct diag_observer
* is the "visible" part.
*/
struct diag_observer;
typedef int (*diag_hook) (struct qib_devdata *dd,
const struct diag_observer *op,
u32 offs, u64 *data, u64 mask, int only_32);
struct diag_observer {
diag_hook hook;
u32 bottom;
u32 top;
};
extern int qib_register_observer(struct qib_devdata *dd,
const struct diag_observer *op);
/* Only declared here, not defined. Private to diags */
struct diag_observer_list_elt;
/* device data struct now contains only "general per-device" info.
* fields related to a physical IB port are in a qib_pportdata struct,
* described above) while fields only used by a particular chip-type are in
* a qib_chipdata struct, whose contents are opaque to this file.
*/
struct qib_devdata {
struct qib_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 device *user_device;
struct device *diag_device;
/* mem-mapped pointer to base of chip regs */
u64 __iomem *kregbase;
/* end of mem-mapped chip space excluding sendbuf and user regs */
u64 __iomem *kregend;
/* physical address of chip for io_remap, etc. */
resource_size_t physaddr;
/* qib_cfgctxts pointers */
struct qib_ctxtdata **rcd; /* Receive Context Data */
/* qib_pportdata, points to array of (physical) port-specific
* data structs, indexed by pidx (0..n-1)
*/
struct qib_pportdata *pport;
struct qib_chip_specific *cspec; /* chip-specific */
/* kvirt address of 1st 2k pio buffer */
void __iomem *pio2kbase;
/* kvirt address of 1st 4k pio buffer */
void __iomem *pio4kbase;
/* mem-mapped pointer to base of PIO buffers (if using WC PAT) */
void __iomem *piobase;
/* mem-mapped pointer to base of user chip regs (if using WC PAT) */
u64 __iomem *userbase;
void __iomem *piovl15base; /* base of VL15 buffers, if not WC */
/*
* points to area where PIOavail registers will be DMA'ed.
* Has to be on a page of it's own, because the page will be
* mapped into user program space. This copy is *ONLY* ever
* written by DMA, not by the driver! Need a copy per device
* when we get to multiple devices
*/
volatile __le64 *pioavailregs_dma; /* DMA'ed by chip */
/* physical address where updates occur */
dma_addr_t pioavailregs_phys;
/* device-specific implementations of functions needed by
* common code. Contrary to previous consensus, we can't
* really just point to a device-specific table, because we
* may need to "bend", e.g. *_f_put_tid
*/
/* fallback to alternate interrupt type if possible */
int (*f_intr_fallback)(struct qib_devdata *);
/* hard reset chip */
int (*f_reset)(struct qib_devdata *);
void (*f_quiet_serdes)(struct qib_pportdata *);
int (*f_bringup_serdes)(struct qib_pportdata *);
int (*f_early_init)(struct qib_devdata *);
void (*f_clear_tids)(struct qib_devdata *, struct qib_ctxtdata *);
void (*f_put_tid)(struct qib_devdata *, u64 __iomem*,
u32, unsigned long);
void (*f_cleanup)(struct qib_devdata *);
void (*f_setextled)(struct qib_pportdata *, u32);
/* fill out chip-specific fields */
int (*f_get_base_info)(struct qib_ctxtdata *, struct qib_base_info *);
/* free irq */
void (*f_free_irq)(struct qib_devdata *);
struct qib_message_header *(*f_get_msgheader)
(struct qib_devdata *, __le32 *);
void (*f_config_ctxts)(struct qib_devdata *);
int (*f_get_ib_cfg)(struct qib_pportdata *, int);
int (*f_set_ib_cfg)(struct qib_pportdata *, int, u32);
int (*f_set_ib_loopback)(struct qib_pportdata *, const char *);
int (*f_get_ib_table)(struct qib_pportdata *, int, void *);
int (*f_set_ib_table)(struct qib_pportdata *, int, void *);
u32 (*f_iblink_state)(u64);
u8 (*f_ibphys_portstate)(u64);
void (*f_xgxs_reset)(struct qib_pportdata *);
/* per chip actions needed for IB Link up/down changes */
int (*f_ib_updown)(struct qib_pportdata *, int, u64);
u32 __iomem *(*f_getsendbuf)(struct qib_pportdata *, u64, u32 *);
/* Read/modify/write of GPIO pins (potentially chip-specific */
int (*f_gpio_mod)(struct qib_devdata *dd, u32 out, u32 dir,
u32 mask);
/* Enable writes to config EEPROM (if supported) */
int (*f_eeprom_wen)(struct qib_devdata *dd, int wen);
/*
* modify rcvctrl shadow[s] and write to appropriate chip-regs.
* see above QIB_RCVCTRL_xxx_ENB/DIS for operations.
* (ctxt == -1) means "all contexts", only meaningful for
* clearing. Could remove if chip_spec shutdown properly done.
*/
void (*f_rcvctrl)(struct qib_pportdata *, unsigned int op,
int ctxt);
/* Read/modify/write sendctrl appropriately for op and port. */
void (*f_sendctrl)(struct qib_pportdata *, u32 op);
void (*f_set_intr_state)(struct qib_devdata *, u32);
void (*f_set_armlaunch)(struct qib_devdata *, u32);
void (*f_wantpiobuf_intr)(struct qib_devdata *, u32);
int (*f_late_initreg)(struct qib_devdata *);
int (*f_init_sdma_regs)(struct qib_pportdata *);
u16 (*f_sdma_gethead)(struct qib_pportdata *);
int (*f_sdma_busy)(struct qib_pportdata *);
void (*f_sdma_update_tail)(struct qib_pportdata *, u16);
void (*f_sdma_set_desc_cnt)(struct qib_pportdata *, unsigned);
void (*f_sdma_sendctrl)(struct qib_pportdata *, unsigned);
void (*f_sdma_hw_clean_up)(struct qib_pportdata *);
void (*f_sdma_hw_start_up)(struct qib_pportdata *);
void (*f_sdma_init_early)(struct qib_pportdata *);
void (*f_set_cntr_sample)(struct qib_pportdata *, u32, u32);
void (*f_update_usrhead)(struct qib_ctxtdata *, u64, u32, u32, u32);
u32 (*f_hdrqempty)(struct qib_ctxtdata *);
u64 (*f_portcntr)(struct qib_pportdata *, u32);
u32 (*f_read_cntrs)(struct qib_devdata *, loff_t, char **,
u64 **);
u32 (*f_read_portcntrs)(struct qib_devdata *, loff_t, u32,
char **, u64 **);
u32 (*f_setpbc_control)(struct qib_pportdata *, u32, u8, u8);
void (*f_initvl15_bufs)(struct qib_devdata *);
void (*f_init_ctxt)(struct qib_ctxtdata *);
void (*f_txchk_change)(struct qib_devdata *, u32, u32, u32,
struct qib_ctxtdata *);
void (*f_writescratch)(struct qib_devdata *, u32);
int (*f_tempsense_rd)(struct qib_devdata *, int regnum);
char *boardname; /* human readable board info */
/* template for writing TIDs */
u64 tidtemplate;
/* value to write to free TIDs */
u64 tidinvalid;
/* number of registers used for pioavail */
u32 pioavregs;
/* device (not port) flags, basically device capabilities */
u32 flags;
/* last buffer for user use */
u32 lastctxt_piobuf;
/* saturating counter of (non-port-specific) device interrupts */
u32 int_counter;
/* pio bufs allocated per ctxt */
u32 pbufsctxt;
/* if remainder on bufs/ctxt, ctxts < extrabuf get 1 extra */
u32 ctxts_extrabuf;
/*
* number of ctxts configured as max; zero is set to number chip
* supports, less gives more pio bufs/ctxt, etc.
*/
u32 cfgctxts;
/*
* hint that we should update pioavailshadow before
* looking for a PIO buffer
*/
u32 upd_pio_shadow;
/* internal debugging stats */
u32 maxpkts_call;
u32 avgpkts_call;
u64 nopiobufs;
/* PCI Vendor ID (here for NodeInfo) */
u16 vendorid;
/* PCI Device ID (here for NodeInfo) */
u16 deviceid;
/* for write combining settings */
unsigned long wc_cookie;
unsigned long wc_base;
unsigned long wc_len;
/* shadow copy of struct page *'s for exp tid pages */
struct page **pageshadow;
/* shadow copy of dma handles for exp tid pages */
dma_addr_t *physshadow;
u64 __iomem *egrtidbase;
spinlock_t sendctrl_lock; /* protect changes to sendctrl shadow */
/* around rcd and (user ctxts) ctxt_cnt use (intr vs free) */
spinlock_t uctxt_lock; /* rcd and user context changes */
/*
* per unit status, see also portdata statusp
* mapped readonly into user processes so they can get unit and
* IB link status cheaply
*/
u64 *devstatusp;
char *freezemsg; /* freeze msg if hw error put chip in freeze */
u32 freezelen; /* max length of freezemsg */
/* timer used to prevent stats overflow, error throttling, etc. */
struct timer_list stats_timer;
/* timer to verify interrupts work, and fallback if possible */
struct timer_list intrchk_timer;
unsigned long ureg_align; /* user register alignment */
/*
* Protects pioavailshadow, pioavailkernel, pio_need_disarm, and
* pio_writing.
*/
spinlock_t pioavail_lock;
/*
* Shadow copies of registers; size indicates read access size.
* Most of them are readonly, but some are write-only register,
* where we manipulate the bits in the shadow copy, and then write
* the shadow copy to qlogic_ib.
*
* We deliberately make most of these 32 bits, since they have
* restricted range. For any that we read, we won't to generate 32
* bit accesses, since Opteron will generate 2 separate 32 bit HT
* transactions for a 64 bit read, and we want to avoid unnecessary
* bus transactions.
*/
/* This is the 64 bit group */
unsigned long pioavailshadow[6];
/* bitmap of send buffers available for the kernel to use with PIO. */
unsigned long pioavailkernel[6];
/* bitmap of send buffers which need to be disarmed. */
unsigned long pio_need_disarm[3];
/* bitmap of send buffers which are being written to. */
unsigned long pio_writing[3];
/* kr_revision shadow */
u64 revision;
/* Base GUID for device (from eeprom, network order) */
__be64 base_guid;
/*
* kr_sendpiobufbase value (chip offset of pio buffers), and the
* base of the 2KB buffer s(user processes only use 2K)
*/
u64 piobufbase;
u32 pio2k_bufbase;
/* these are the "32 bit" regs */
/* number of GUIDs in the flash for this interface */
u32 nguid;
/*
* the following two are 32-bit bitmasks, but {test,clear,set}_bit
* all expect bit fields to be "unsigned long"
*/
unsigned long rcvctrl; /* shadow per device rcvctrl */
unsigned long sendctrl; /* shadow per device sendctrl */
/* value we put in kr_rcvhdrcnt */
u32 rcvhdrcnt;
/* value we put in kr_rcvhdrsize */
u32 rcvhdrsize;
/* value we put in kr_rcvhdrentsize */
u32 rcvhdrentsize;
/* kr_ctxtcnt value */
u32 ctxtcnt;
/* kr_pagealign value */
u32 palign;
/* number of "2KB" PIO buffers */
u32 piobcnt2k;
/* size in bytes of "2KB" PIO buffers */
u32 piosize2k;
/* max usable size in dwords of a "2KB" PIO buffer before going "4KB" */
u32 piosize2kmax_dwords;
/* number of "4KB" PIO buffers */
u32 piobcnt4k;
/* size in bytes of "4KB" PIO buffers */
u32 piosize4k;
/* kr_rcvegrbase value */
u32 rcvegrbase;
/* kr_rcvtidbase value */
u32 rcvtidbase;
/* kr_rcvtidcnt value */
u32 rcvtidcnt;
/* kr_userregbase */
u32 uregbase;
/* shadow the control register contents */
u32 control;
/* chip address space used by 4k pio buffers */
u32 align4k;
/* size of each rcvegrbuffer */
u32 rcvegrbufsize;
/* 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 */
/* start of CHIP_SPEC move to chipspec, but need code changes */
/* low and high portions of MSI capability/vector */
u32 msi_lo;
/* saved after PCIe init for restore after reset */
u32 msi_hi;
/* MSI data (vector) saved for restore */
u16 msi_data;
/* so we can rewrite it after a chip reset */
u32 pcibar0;
/* so we can rewrite it after a chip reset */
u32 pcibar1;
u64 rhdrhead_intr_off;
/*
* ASCII serial number, from flash, large enough for original
* all digit strings, and longer QLogic serial number format
*/
u8 serial[16];
/* human readable board version */
u8 boardversion[96];
u8 lbus_info[32]; /* human readable localbus info */
/* chip major rev, from qib_revision */
u8 majrev;
/* chip minor rev, from qib_revision */
u8 minrev;
/* 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 qpn_mask;
u8 skip_kctxt_mask;
u16 rhf_offset; /* offset of RHF within receive header entry */
/*
* GPIO pins for twsi-connected devices, and device code for eeprom
*/
u8 gpio_sda_num;
u8 gpio_scl_num;
u8 twsi_eeprom_dev;
u8 board_atten;
/* Support (including locks) for EEPROM logging of errors and time */
/* control access to actual counters, timer */
spinlock_t eep_st_lock;
/* control high-level access to EEPROM */
struct mutex eep_lock;
uint64_t traffic_wds;
/* active time is kept in seconds, but logged in hours */
atomic_t active_time;
/* Below are nominal shadow of EEPROM, new since last EEPROM update */
uint8_t eep_st_errs[QIB_EEP_LOG_CNT];
uint8_t eep_st_new_errs[QIB_EEP_LOG_CNT];
uint16_t eep_hrs;
/*
* masks for which bits of errs, hwerrs that cause
* each of the counters to increment.
*/
struct qib_eep_log_mask eep_st_masks[QIB_EEP_LOG_CNT];
struct qib_diag_client *diag_client;
spinlock_t qib_diag_trans_lock; /* protect diag observer ops */
struct diag_observer_list_elt *diag_observer_list;
u8 psxmitwait_supported;
/* cycle length of PS* counters in HW (in picoseconds) */
u16 psxmitwait_check_rate;
/* high volume overflow errors defered to tasklet */
struct tasklet_struct error_tasklet;
};
/* hol_state values */
#define QIB_HOL_UP 0
#define QIB_HOL_INIT 1
#define QIB_SDMA_SENDCTRL_OP_ENABLE (1U << 0)
#define QIB_SDMA_SENDCTRL_OP_INTENABLE (1U << 1)
#define QIB_SDMA_SENDCTRL_OP_HALT (1U << 2)
#define QIB_SDMA_SENDCTRL_OP_CLEANUP (1U << 3)
#define QIB_SDMA_SENDCTRL_OP_DRAIN (1U << 4)
/* operation types for f_txchk_change() */
#define TXCHK_CHG_TYPE_DIS1 3
#define TXCHK_CHG_TYPE_ENAB1 2
#define TXCHK_CHG_TYPE_KERN 1
#define TXCHK_CHG_TYPE_USER 0
#define QIB_CHASE_TIME msecs_to_jiffies(145)
#define QIB_CHASE_DIS_TIME msecs_to_jiffies(160)
/* Private data for file operations */
struct qib_filedata {
struct qib_ctxtdata *rcd;
unsigned subctxt;
unsigned tidcursor;
struct qib_user_sdma_queue *pq;
int rec_cpu_num; /* for cpu affinity; -1 if none */
};
extern struct list_head qib_dev_list;
extern spinlock_t qib_devs_lock;
extern struct qib_devdata *qib_lookup(int unit);
extern u32 qib_cpulist_count;
extern unsigned long *qib_cpulist;
extern unsigned qib_wc_pat;
int qib_init(struct qib_devdata *, int);
int init_chip_wc_pat(struct qib_devdata *dd, u32);
int qib_enable_wc(struct qib_devdata *dd);
void qib_disable_wc(struct qib_devdata *dd);
int qib_count_units(int *npresentp, int *nupp);
int qib_count_active_units(void);
int qib_cdev_init(int minor, const char *name,
const struct file_operations *fops,
struct cdev **cdevp, struct device **devp);
void qib_cdev_cleanup(struct cdev **cdevp, struct device **devp);
int qib_dev_init(void);
void qib_dev_cleanup(void);
int qib_diag_add(struct qib_devdata *);
void qib_diag_remove(struct qib_devdata *);
void qib_handle_e_ibstatuschanged(struct qib_pportdata *, u64);
void qib_sdma_update_tail(struct qib_pportdata *, u16); /* hold sdma_lock */
int qib_decode_err(struct qib_devdata *dd, char *buf, size_t blen, u64 err);
void qib_bad_intrstatus(struct qib_devdata *);
void qib_handle_urcv(struct qib_devdata *, u64);
/* clean up any per-chip chip-specific stuff */
void qib_chip_cleanup(struct qib_devdata *);
/* clean up any chip type-specific stuff */
void qib_chip_done(void);
/* check to see if we have to force ordering for write combining */
int qib_unordered_wc(void);
void qib_pio_copy(void __iomem *to, const void *from, size_t count);
void qib_disarm_piobufs(struct qib_devdata *, unsigned, unsigned);
int qib_disarm_piobufs_ifneeded(struct qib_ctxtdata *);
void qib_disarm_piobufs_set(struct qib_devdata *, unsigned long *, unsigned);
void qib_cancel_sends(struct qib_pportdata *);
int qib_create_rcvhdrq(struct qib_devdata *, struct qib_ctxtdata *);
int qib_setup_eagerbufs(struct qib_ctxtdata *);
void qib_set_ctxtcnt(struct qib_devdata *);
int qib_create_ctxts(struct qib_devdata *dd);
struct qib_ctxtdata *qib_create_ctxtdata(struct qib_pportdata *, u32);
void qib_init_pportdata(struct qib_pportdata *, struct qib_devdata *, u8, u8);
void qib_free_ctxtdata(struct qib_devdata *, struct qib_ctxtdata *);
u32 qib_kreceive(struct qib_ctxtdata *, u32 *, u32 *);
int qib_reset_device(int);
int qib_wait_linkstate(struct qib_pportdata *, u32, int);
int qib_set_linkstate(struct qib_pportdata *, u8);
int qib_set_mtu(struct qib_pportdata *, u16);
int qib_set_lid(struct qib_pportdata *, u32, u8);
void qib_hol_down(struct qib_pportdata *);
void qib_hol_init(struct qib_pportdata *);
void qib_hol_up(struct qib_pportdata *);
void qib_hol_event(unsigned long);
void qib_disable_after_error(struct qib_devdata *);
int qib_set_uevent_bits(struct qib_pportdata *, const int);
/* for use in system calls, where we want to know device type, etc. */
#define ctxt_fp(fp) \
(((struct qib_filedata *)(fp)->private_data)->rcd)
#define subctxt_fp(fp) \
(((struct qib_filedata *)(fp)->private_data)->subctxt)
#define tidcursor_fp(fp) \
(((struct qib_filedata *)(fp)->private_data)->tidcursor)
#define user_sdma_queue_fp(fp) \
(((struct qib_filedata *)(fp)->private_data)->pq)
static inline struct qib_devdata *dd_from_ppd(struct qib_pportdata *ppd)
{
return ppd->dd;
}
static inline struct qib_devdata *dd_from_dev(struct qib_ibdev *dev)
{
return container_of(dev, struct qib_devdata, verbs_dev);
}
static inline struct qib_devdata *dd_from_ibdev(struct ib_device *ibdev)
{
return dd_from_dev(to_idev(ibdev));
}
static inline struct qib_pportdata *ppd_from_ibp(struct qib_ibport *ibp)
{
return container_of(ibp, struct qib_pportdata, ibport_data);
}
static inline struct qib_ibport *to_iport(struct ib_device *ibdev, u8 port)
{
struct qib_devdata *dd = dd_from_ibdev(ibdev);
unsigned pidx = port - 1; /* IB number port from 1, hdw from 0 */
WARN_ON(pidx >= dd->num_pports);
return &dd->pport[pidx].ibport_data;
}
/*
* values for dd->flags (_device_ related flags) and
*/
#define QIB_HAS_LINK_LATENCY 0x1 /* supports link latency (IB 1.2) */
#define QIB_INITTED 0x2 /* chip and driver up and initted */
#define QIB_DOING_RESET 0x4 /* in the middle of doing chip reset */
#define QIB_PRESENT 0x8 /* chip accesses can be done */
#define QIB_PIO_FLUSH_WC 0x10 /* Needs Write combining flush for PIO */
#define QIB_HAS_THRESH_UPDATE 0x40
#define QIB_HAS_SDMA_TIMEOUT 0x80
#define QIB_USE_SPCL_TRIG 0x100 /* SpecialTrigger launch enabled */
#define QIB_NODMA_RTAIL 0x200 /* rcvhdrtail register DMA enabled */
#define QIB_HAS_INTX 0x800 /* Supports INTx interrupts */
#define QIB_HAS_SEND_DMA 0x1000 /* Supports Send DMA */
#define QIB_HAS_VLSUPP 0x2000 /* Supports multiple VLs; PBC different */
#define QIB_HAS_HDRSUPP 0x4000 /* Supports header suppression */
#define QIB_BADINTR 0x8000 /* severe interrupt problems */
#define QIB_DCA_ENABLED 0x10000 /* Direct Cache Access enabled */
#define QIB_HAS_QSFP 0x20000 /* device (card instance) has QSFP */
/*
* values for ppd->lflags (_ib_port_ related flags)
*/
#define QIBL_LINKV 0x1 /* IB link state valid */
#define QIBL_LINKDOWN 0x8 /* IB link is down */
#define QIBL_LINKINIT 0x10 /* IB link level is up */
#define QIBL_LINKARMED 0x20 /* IB link is ARMED */
#define QIBL_LINKACTIVE 0x40 /* IB link is ACTIVE */
/* leave a gap for more IB-link state */
#define QIBL_IB_AUTONEG_INPROG 0x1000 /* non-IBTA DDR/QDR neg active */
#define QIBL_IB_AUTONEG_FAILED 0x2000 /* non-IBTA DDR/QDR neg failed */
#define QIBL_IB_LINK_DISABLED 0x4000 /* Linkdown-disable forced,
* Do not try to bring up */
#define QIBL_IB_FORCE_NOTIFY 0x8000 /* force notify on next ib change */
/* IB dword length mask in PBC (lower 11 bits); same for all chips */
#define QIB_PBC_LENGTH_MASK ((1 << 11) - 1)
/* ctxt_flag bit offsets */
/* waiting for a packet to arrive */
#define QIB_CTXT_WAITING_RCV 2
/* master has not finished initializing */
#define QIB_CTXT_MASTER_UNINIT 4
/* waiting for an urgent packet to arrive */
#define QIB_CTXT_WAITING_URG 5
/* free up any allocated data at closes */
void qib_free_data(struct qib_ctxtdata *dd);
void qib_chg_pioavailkernel(struct qib_devdata *, unsigned, unsigned,
u32, struct qib_ctxtdata *);
struct qib_devdata *qib_init_iba7322_funcs(struct pci_dev *,
const struct pci_device_id *);
struct qib_devdata *qib_init_iba7220_funcs(struct pci_dev *,
const struct pci_device_id *);
struct qib_devdata *qib_init_iba6120_funcs(struct pci_dev *,
const struct pci_device_id *);
void qib_free_devdata(struct qib_devdata *);
struct qib_devdata *qib_alloc_devdata(struct pci_dev *pdev, size_t extra);
#define QIB_TWSI_NO_DEV 0xFF
/* Below qib_twsi_ functions must be called with eep_lock held */
int qib_twsi_reset(struct qib_devdata *dd);
int qib_twsi_blk_rd(struct qib_devdata *dd, int dev, int addr, void *buffer,
int len);
int qib_twsi_blk_wr(struct qib_devdata *dd, int dev, int addr,
const void *buffer, int len);
void qib_get_eeprom_info(struct qib_devdata *);
int qib_update_eeprom_log(struct qib_devdata *dd);
void qib_inc_eeprom_err(struct qib_devdata *dd, u32 eidx, u32 incr);
void qib_dump_lookup_output_queue(struct qib_devdata *);
void qib_force_pio_avail_update(struct qib_devdata *);
void qib_clear_symerror_on_linkup(unsigned long opaque);
/*
* Set LED override, only the two LSBs have "public" meaning, but
* any non-zero value substitutes them for the Link and LinkTrain
* LED states.
*/
#define QIB_LED_PHYS 1 /* Physical (linktraining) GREEN LED */
#define QIB_LED_LOG 2 /* Logical (link) YELLOW LED */
void qib_set_led_override(struct qib_pportdata *ppd, unsigned int val);
/* send dma routines */
int qib_setup_sdma(struct qib_pportdata *);
void qib_teardown_sdma(struct qib_pportdata *);
void __qib_sdma_intr(struct qib_pportdata *);
void qib_sdma_intr(struct qib_pportdata *);
int qib_sdma_verbs_send(struct qib_pportdata *, struct qib_sge_state *,
u32, struct qib_verbs_txreq *);
/* ppd->sdma_lock should be locked before calling this. */
int qib_sdma_make_progress(struct qib_pportdata *dd);
/* must be called under qib_sdma_lock */
static inline u16 qib_sdma_descq_freecnt(const struct qib_pportdata *ppd)
{
return ppd->sdma_descq_cnt -
(ppd->sdma_descq_added - ppd->sdma_descq_removed) - 1;
}
static inline int __qib_sdma_running(struct qib_pportdata *ppd)
{
return ppd->sdma_state.current_state == qib_sdma_state_s99_running;
}
int qib_sdma_running(struct qib_pportdata *);
void __qib_sdma_process_event(struct qib_pportdata *, enum qib_sdma_events);
void qib_sdma_process_event(struct qib_pportdata *, enum qib_sdma_events);
/*
* number of words used for protocol header if not set by qib_userinit();
*/
#define QIB_DFLT_RCVHDRSIZE 9
/*
* We need to be able to handle an IB header of at least 24 dwords.
* We need the rcvhdrq large enough to handle largest IB header, but
* still have room for a 2KB MTU standard IB packet.
* Additionally, some processor/memory controller combinations
* benefit quite strongly from having the DMA'ed data be cacheline
* aligned and a cacheline multiple, so we set the size to 32 dwords
* (2 64-byte primary cachelines for pretty much all processors of
* interest). The alignment hurts nothing, other than using somewhat
* more memory.
*/
#define QIB_RCVHDR_ENTSIZE 32
int qib_get_user_pages(unsigned long, size_t, struct page **);
void qib_release_user_pages(struct page **, size_t);
int qib_eeprom_read(struct qib_devdata *, u8, void *, int);
int qib_eeprom_write(struct qib_devdata *, u8, const void *, int);
u32 __iomem *qib_getsendbuf_range(struct qib_devdata *, u32 *, u32, u32);
void qib_sendbuf_done(struct qib_devdata *, unsigned);
static inline void qib_clear_rcvhdrtail(const struct qib_ctxtdata *rcd)
{
*((u64 *) rcd->rcvhdrtail_kvaddr) = 0ULL;
}
static inline u32 qib_get_rcvhdrtail(const struct qib_ctxtdata *rcd)
{
/*
* volatile because it's a DMA target from the chip, routine is
* inlined, and don't want register caching or reordering.
*/
return (u32) le64_to_cpu(
*((volatile __le64 *)rcd->rcvhdrtail_kvaddr)); /* DMA'ed */
}
static inline u32 qib_get_hdrqtail(const struct qib_ctxtdata *rcd)
{
const struct qib_devdata *dd = rcd->dd;
u32 hdrqtail;
if (dd->flags & QIB_NODMA_RTAIL) {
__le32 *rhf_addr;
u32 seq;
rhf_addr = (__le32 *) rcd->rcvhdrq +
rcd->head + dd->rhf_offset;
seq = qib_hdrget_seq(rhf_addr);
hdrqtail = rcd->head;
if (seq == rcd->seq_cnt)
hdrqtail++;
} else
hdrqtail = qib_get_rcvhdrtail(rcd);
return hdrqtail;
}
/*
* sysfs interface.
*/
extern const char ib_qib_version[];
int qib_device_create(struct qib_devdata *);
void qib_device_remove(struct qib_devdata *);
int qib_create_port_files(struct ib_device *ibdev, u8 port_num,
struct kobject *kobj);
int qib_verbs_register_sysfs(struct qib_devdata *);
void qib_verbs_unregister_sysfs(struct qib_devdata *);
/* Hook for sysfs read of QSFP */
extern int qib_qsfp_dump(struct qib_pportdata *ppd, char *buf, int len);
int __init qib_init_qibfs(void);
int __exit qib_exit_qibfs(void);
int qibfs_add(struct qib_devdata *);
int qibfs_remove(struct qib_devdata *);
int qib_pcie_init(struct pci_dev *, const struct pci_device_id *);
int qib_pcie_ddinit(struct qib_devdata *, struct pci_dev *,
const struct pci_device_id *);
void qib_pcie_ddcleanup(struct qib_devdata *);
int qib_pcie_params(struct qib_devdata *, u32, u32 *, struct msix_entry *);
int qib_reinit_intr(struct qib_devdata *);
void qib_enable_intx(struct pci_dev *);
void qib_nomsi(struct qib_devdata *);
void qib_nomsix(struct qib_devdata *);
void qib_pcie_getcmd(struct qib_devdata *, u16 *, u8 *, u8 *);
void qib_pcie_reenable(struct qib_devdata *, u16, u8, u8);
/*
* dma_addr wrappers - all 0's invalid for hw
*/
dma_addr_t qib_map_page(struct pci_dev *, struct page *, unsigned long,
size_t, int);
const char *qib_get_unit_name(int unit);
/*
* Flush write combining store buffers (if present) and perform a write
* barrier.
*/
#if defined(CONFIG_X86_64)
#define qib_flush_wc() asm volatile("sfence" : : : "memory")
#else
#define qib_flush_wc() wmb() /* no reorder around wc flush */
#endif
/* global module parameter variables */
extern unsigned qib_ibmtu;
extern ushort qib_cfgctxts;
extern ushort qib_num_cfg_vls;
extern ushort qib_mini_init; /* If set, do few (ideally 0) writes to chip */
extern unsigned qib_n_krcv_queues;
extern unsigned qib_sdma_fetch_arb;
extern unsigned qib_compat_ddr_negotiate;
extern int qib_special_trigger;
extern struct mutex qib_mutex;
/* Number of seconds before our card status check... */
#define STATUS_TIMEOUT 60
#define QIB_DRV_NAME "ib_qib"
#define QIB_USER_MINOR_BASE 0
#define QIB_TRACE_MINOR 127
#define QIB_DIAGPKT_MINOR 128
#define QIB_DIAG_MINOR_BASE 129
#define QIB_NMINORS 255
#define PCI_VENDOR_ID_PATHSCALE 0x1fc1
#define PCI_VENDOR_ID_QLOGIC 0x1077
#define PCI_DEVICE_ID_QLOGIC_IB_6120 0x10
#define PCI_DEVICE_ID_QLOGIC_IB_7220 0x7220
#define PCI_DEVICE_ID_QLOGIC_IB_7322 0x7322
/*
* qib_early_err is used (only!) to print early errors before devdata is
* allocated, or when dd->pcidev may not be valid, and at the tail end of
* cleanup when devdata may have been freed, etc. qib_dev_porterr is
* the same as qib_dev_err, but is used when the message really needs
* the IB port# to be definitive as to what's happening..
* All of these go to the trace log, and the trace log entry is done
* first to avoid possible serial port delays from printk.
*/
#define qib_early_err(dev, fmt, ...) \
do { \
dev_err(dev, fmt, ##__VA_ARGS__); \
} while (0)
#define qib_dev_err(dd, fmt, ...) \
do { \
dev_err(&(dd)->pcidev->dev, "%s: " fmt, \
qib_get_unit_name((dd)->unit), ##__VA_ARGS__); \
} while (0)
#define qib_dev_porterr(dd, port, fmt, ...) \
do { \
dev_err(&(dd)->pcidev->dev, "%s: IB%u:%u " fmt, \
qib_get_unit_name((dd)->unit), (dd)->unit, (port), \
##__VA_ARGS__); \
} while (0)
#define qib_devinfo(pcidev, fmt, ...) \
do { \
dev_info(&(pcidev)->dev, fmt, ##__VA_ARGS__); \
} while (0)
/*
* this is used for formatting hw error messages...
*/
struct qib_hwerror_msgs {
u64 mask;
const char *msg;
size_t sz;
};
#define QLOGIC_IB_HWE_MSG(a, b) { .mask = a, .msg = b }
/* in qib_intr.c... */
void qib_format_hwerrors(u64 hwerrs,
const struct qib_hwerror_msgs *hwerrmsgs,
size_t nhwerrmsgs, char *msg, size_t lmsg);
#endif /* _QIB_KERNEL_H */