ipath_verbs.c

来自「linux 内核源代码」· C语言 代码 · 共 1,887 行 · 第 1/4 页

 * @hdr: the packet header
 * @hdrwords: the number of words in the header
 * @ss: the SGE to send
 * @len: the length of the packet in bytes
 */
int ipath_verbs_send(struct ipath_qp *qp, struct ipath_ib_header *hdr,
		     u32 hdrwords, struct ipath_sge_state *ss, u32 len)
{
	struct ipath_devdata *dd = to_idev(qp->ibqp.device)->dd;
	u32 plen;
	int ret;
	u32 dwords = (len + 3) >> 2;

	/* +1 is for the qword padding of pbc */
	plen = hdrwords + dwords + 1;

	/* Drop non-VL15 packets if we are not in the active state */
	if (!(dd->ipath_flags & IPATH_LINKACTIVE) &&
	    qp->ibqp.qp_type != IB_QPT_SMI) {
		if (qp->s_wqe)
			ipath_send_complete(qp, qp->s_wqe, IB_WC_SUCCESS);
		ret = 0;
	} else
		ret = ipath_verbs_send_pio(qp, (u32 *) hdr, hdrwords,
					   ss, len, plen, dwords);

	return ret;
}

int ipath_snapshot_counters(struct ipath_devdata *dd, u64 *swords,
			    u64 *rwords, u64 *spkts, u64 *rpkts,
			    u64 *xmit_wait)
{
	int ret;

	if (!(dd->ipath_flags & IPATH_INITTED)) {
		/* no hardware, freeze, etc. */
		ret = -EINVAL;
		goto bail;
	}
	*swords = ipath_snap_cntr(dd, dd->ipath_cregs->cr_wordsendcnt);
	*rwords = ipath_snap_cntr(dd, dd->ipath_cregs->cr_wordrcvcnt);
	*spkts = ipath_snap_cntr(dd, dd->ipath_cregs->cr_pktsendcnt);
	*rpkts = ipath_snap_cntr(dd, dd->ipath_cregs->cr_pktrcvcnt);
	*xmit_wait = ipath_snap_cntr(dd, dd->ipath_cregs->cr_sendstallcnt);

	ret = 0;

bail:
	return ret;
}

/**
 * ipath_get_counters - get various chip counters
 * @dd: the infinipath device
 * @cntrs: counters are placed here
 *
 * Return the counters needed by recv_pma_get_portcounters().
 */
int ipath_get_counters(struct ipath_devdata *dd,
		       struct ipath_verbs_counters *cntrs)
{
	struct ipath_cregs const *crp = dd->ipath_cregs;
	int ret;

	if (!(dd->ipath_flags & IPATH_INITTED)) {
		/* no hardware, freeze, etc. */
		ret = -EINVAL;
		goto bail;
	}
	cntrs->symbol_error_counter =
		ipath_snap_cntr(dd, crp->cr_ibsymbolerrcnt);
	cntrs->link_error_recovery_counter =
		ipath_snap_cntr(dd, crp->cr_iblinkerrrecovcnt);
	/*
	 * The link downed counter counts when the other side downs the
	 * connection.  We add in the number of times we downed the link
	 * due to local link integrity errors to compensate.
	 */
	cntrs->link_downed_counter =
		ipath_snap_cntr(dd, crp->cr_iblinkdowncnt);
	cntrs->port_rcv_errors =
		ipath_snap_cntr(dd, crp->cr_rxdroppktcnt) +
		ipath_snap_cntr(dd, crp->cr_rcvovflcnt) +
		ipath_snap_cntr(dd, crp->cr_portovflcnt) +
		ipath_snap_cntr(dd, crp->cr_err_rlencnt) +
		ipath_snap_cntr(dd, crp->cr_invalidrlencnt) +
		ipath_snap_cntr(dd, crp->cr_errlinkcnt) +
		ipath_snap_cntr(dd, crp->cr_erricrccnt) +
		ipath_snap_cntr(dd, crp->cr_errvcrccnt) +
		ipath_snap_cntr(dd, crp->cr_errlpcrccnt) +
		ipath_snap_cntr(dd, crp->cr_badformatcnt) +
		dd->ipath_rxfc_unsupvl_errs;
	cntrs->port_rcv_remphys_errors =
		ipath_snap_cntr(dd, crp->cr_rcvebpcnt);
	cntrs->port_xmit_discards = ipath_snap_cntr(dd, crp->cr_unsupvlcnt);
	cntrs->port_xmit_data = ipath_snap_cntr(dd, crp->cr_wordsendcnt);
	cntrs->port_rcv_data = ipath_snap_cntr(dd, crp->cr_wordrcvcnt);
	cntrs->port_xmit_packets = ipath_snap_cntr(dd, crp->cr_pktsendcnt);
	cntrs->port_rcv_packets = ipath_snap_cntr(dd, crp->cr_pktrcvcnt);
	cntrs->local_link_integrity_errors =
		(dd->ipath_flags & IPATH_GPIO_ERRINTRS) ?
		dd->ipath_lli_errs : dd->ipath_lli_errors;
	cntrs->excessive_buffer_overrun_errors = dd->ipath_overrun_thresh_errs;

	ret = 0;

bail:
	return ret;
}

/**
 * ipath_ib_piobufavail - callback when a PIO buffer is available
 * @arg: the device pointer
 *
 * This is called from ipath_intr() at interrupt level when a PIO buffer is
 * available after ipath_verbs_send() returned an error that no buffers were
 * available.  Return 1 if we consumed all the PIO buffers and we still have
 * QPs waiting for buffers (for now, just do a tasklet_hi_schedule and
 * return zero).
 */
int ipath_ib_piobufavail(struct ipath_ibdev *dev)
{
	struct ipath_qp *qp;
	unsigned long flags;

	if (dev == NULL)
		goto bail;

	spin_lock_irqsave(&dev->pending_lock, flags);
	while (!list_empty(&dev->piowait)) {
		qp = list_entry(dev->piowait.next, struct ipath_qp,
				piowait);
		list_del_init(&qp->piowait);
		clear_bit(IPATH_S_BUSY, &qp->s_busy);
		tasklet_hi_schedule(&qp->s_task);
	}
	spin_unlock_irqrestore(&dev->pending_lock, flags);

bail:
	return 0;
}

static int ipath_query_device(struct ib_device *ibdev,
			      struct ib_device_attr *props)
{
	struct ipath_ibdev *dev = to_idev(ibdev);

	memset(props, 0, sizeof(*props));

	props->device_cap_flags = IB_DEVICE_BAD_PKEY_CNTR |
		IB_DEVICE_BAD_QKEY_CNTR | IB_DEVICE_SHUTDOWN_PORT |
		IB_DEVICE_SYS_IMAGE_GUID;
	props->page_size_cap = PAGE_SIZE;
	props->vendor_id = dev->dd->ipath_vendorid;
	props->vendor_part_id = dev->dd->ipath_deviceid;
	props->hw_ver = dev->dd->ipath_pcirev;

	props->sys_image_guid = dev->sys_image_guid;

	props->max_mr_size = ~0ull;
	props->max_qp = ib_ipath_max_qps;
	props->max_qp_wr = ib_ipath_max_qp_wrs;
	props->max_sge = ib_ipath_max_sges;
	props->max_cq = ib_ipath_max_cqs;
	props->max_ah = ib_ipath_max_ahs;
	props->max_cqe = ib_ipath_max_cqes;
	props->max_mr = dev->lk_table.max;
	props->max_fmr = dev->lk_table.max;
	props->max_map_per_fmr = 32767;
	props->max_pd = ib_ipath_max_pds;
	props->max_qp_rd_atom = IPATH_MAX_RDMA_ATOMIC;
	props->max_qp_init_rd_atom = 255;
	/* props->max_res_rd_atom */
	props->max_srq = ib_ipath_max_srqs;
	props->max_srq_wr = ib_ipath_max_srq_wrs;
	props->max_srq_sge = ib_ipath_max_srq_sges;
	/* props->local_ca_ack_delay */
	props->atomic_cap = IB_ATOMIC_GLOB;
	props->max_pkeys = ipath_get_npkeys(dev->dd);
	props->max_mcast_grp = ib_ipath_max_mcast_grps;
	props->max_mcast_qp_attach = ib_ipath_max_mcast_qp_attached;
	props->max_total_mcast_qp_attach = props->max_mcast_qp_attach *
		props->max_mcast_grp;

	return 0;
}

const u8 ipath_cvt_physportstate[16] = {
	[INFINIPATH_IBCS_LT_STATE_DISABLED] = 3,
	[INFINIPATH_IBCS_LT_STATE_LINKUP] = 5,
	[INFINIPATH_IBCS_LT_STATE_POLLACTIVE] = 2,
	[INFINIPATH_IBCS_LT_STATE_POLLQUIET] = 2,
	[INFINIPATH_IBCS_LT_STATE_SLEEPDELAY] = 1,
	[INFINIPATH_IBCS_LT_STATE_SLEEPQUIET] = 1,
	[INFINIPATH_IBCS_LT_STATE_CFGDEBOUNCE] = 4,
	[INFINIPATH_IBCS_LT_STATE_CFGRCVFCFG] = 4,
	[INFINIPATH_IBCS_LT_STATE_CFGWAITRMT] = 4,
	[INFINIPATH_IBCS_LT_STATE_CFGIDLE] = 4,
	[INFINIPATH_IBCS_LT_STATE_RECOVERRETRAIN] = 6,
	[INFINIPATH_IBCS_LT_STATE_RECOVERWAITRMT] = 6,
	[INFINIPATH_IBCS_LT_STATE_RECOVERIDLE] = 6,
};

u32 ipath_get_cr_errpkey(struct ipath_devdata *dd)
{
	return ipath_read_creg32(dd, dd->ipath_cregs->cr_errpkey);
}

static int ipath_query_port(struct ib_device *ibdev,
			    u8 port, struct ib_port_attr *props)
{
	struct ipath_ibdev *dev = to_idev(ibdev);
	struct ipath_devdata *dd = dev->dd;
	enum ib_mtu mtu;
	u16 lid = dd->ipath_lid;
	u64 ibcstat;

	memset(props, 0, sizeof(*props));
	props->lid = lid ? lid : __constant_be16_to_cpu(IB_LID_PERMISSIVE);
	props->lmc = dd->ipath_lmc;
	props->sm_lid = dev->sm_lid;
	props->sm_sl = dev->sm_sl;
	ibcstat = dd->ipath_lastibcstat;
	props->state = ((ibcstat >> 4) & 0x3) + 1;
	/* See phys_state_show() */
	props->phys_state = ipath_cvt_physportstate[
		dd->ipath_lastibcstat & 0xf];
	props->port_cap_flags = dev->port_cap_flags;
	props->gid_tbl_len = 1;
	props->max_msg_sz = 0x80000000;
	props->pkey_tbl_len = ipath_get_npkeys(dd);
	props->bad_pkey_cntr = ipath_get_cr_errpkey(dd) -
		dev->z_pkey_violations;
	props->qkey_viol_cntr = dev->qkey_violations;
	props->active_width = IB_WIDTH_4X;
	/* See rate_show() */
	props->active_speed = 1;	/* Regular 10Mbs speed. */
	props->max_vl_num = 1;		/* VLCap = VL0 */
	props->init_type_reply = 0;

	/*
	 * Note: the chip supports a maximum MTU of 4096, but the driver
	 * hasn't implemented this feature yet, so set the maximum value
	 * to 2048.
	 */
	props->max_mtu = IB_MTU_2048;
	switch (dd->ipath_ibmtu) {
	case 4096:
		mtu = IB_MTU_4096;
		break;
	case 2048:
		mtu = IB_MTU_2048;
		break;
	case 1024:
		mtu = IB_MTU_1024;
		break;
	case 512:
		mtu = IB_MTU_512;
		break;
	case 256:
		mtu = IB_MTU_256;
		break;
	default:
		mtu = IB_MTU_2048;
	}
	props->active_mtu = mtu;
	props->subnet_timeout = dev->subnet_timeout;

	return 0;
}

static int ipath_modify_device(struct ib_device *device,
			       int device_modify_mask,
			       struct ib_device_modify *device_modify)
{
	int ret;

	if (device_modify_mask & ~(IB_DEVICE_MODIFY_SYS_IMAGE_GUID |
				   IB_DEVICE_MODIFY_NODE_DESC)) {
		ret = -EOPNOTSUPP;
		goto bail;
	}

	if (device_modify_mask & IB_DEVICE_MODIFY_NODE_DESC)
		memcpy(device->node_desc, device_modify->node_desc, 64);

	if (device_modify_mask & IB_DEVICE_MODIFY_SYS_IMAGE_GUID)
		to_idev(device)->sys_image_guid =
			cpu_to_be64(device_modify->sys_image_guid);

	ret = 0;

bail:
	return ret;
}

static int ipath_modify_port(struct ib_device *ibdev,
			     u8 port, int port_modify_mask,
			     struct ib_port_modify *props)
{
	struct ipath_ibdev *dev = to_idev(ibdev);

	dev->port_cap_flags |= props->set_port_cap_mask;
	dev->port_cap_flags &= ~props->clr_port_cap_mask;
	if (port_modify_mask & IB_PORT_SHUTDOWN)
		ipath_set_linkstate(dev->dd, IPATH_IB_LINKDOWN);
	if (port_modify_mask & IB_PORT_RESET_QKEY_CNTR)
		dev->qkey_violations = 0;
	return 0;
}

static int ipath_query_gid(struct ib_device *ibdev, u8 port,
			   int index, union ib_gid *gid)
{
	struct ipath_ibdev *dev = to_idev(ibdev);
	int ret;

	if (index >= 1) {
		ret = -EINVAL;
		goto bail;
	}
	gid->global.subnet_prefix = dev->gid_prefix;
	gid->global.interface_id = dev->dd->ipath_guid;

	ret = 0;

bail:
	return ret;
}

static struct ib_pd *ipath_alloc_pd(struct ib_device *ibdev,
				    struct ib_ucontext *context,
				    struct ib_udata *udata)
{
	struct ipath_ibdev *dev = to_idev(ibdev);
	struct ipath_pd *pd;
	struct ib_pd *ret;

	/*
	 * This is actually totally arbitrary.	Some correctness tests
	 * assume there's a maximum number of PDs that can be allocated.
	 * We don't actually have this limit, but we fail the test if
	 * we allow allocations of more than we report for this value.
	 */

	pd = kmalloc(sizeof *pd, GFP_KERNEL);
	if (!pd) {
		ret = ERR_PTR(-ENOMEM);
		goto bail;
	}

	spin_lock(&dev->n_pds_lock);
	if (dev->n_pds_allocated == ib_ipath_max_pds) {
		spin_unlock(&dev->n_pds_lock);
		kfree(pd);
		ret = ERR_PTR(-ENOMEM);
		goto bail;
	}

	dev->n_pds_allocated++;
	spin_unlock(&dev->n_pds_lock);

	/* ib_alloc_pd() will initialize pd->ibpd. */
	pd->user = udata != NULL;

	ret = &pd->ibpd;

bail:
	return ret;
}

static int ipath_dealloc_pd(struct ib_pd *ibpd)
{
	struct ipath_pd *pd = to_ipd(ibpd);
	struct ipath_ibdev *dev = to_idev(ibpd->device);

	spin_lock(&dev->n_pds_lock);
	dev->n_pds_allocated--;
	spin_unlock(&dev->n_pds_lock);

	kfree(pd);

	return 0;
}

/**
 * ipath_create_ah - create an address handle
 * @pd: the protection domain
 * @ah_attr: the attributes of the AH
 *
 * This may be called from interrupt context.
 */
static struct ib_ah *ipath_create_ah(struct ib_pd *pd,
				     struct ib_ah_attr *ah_attr)
{
	struct ipath_ah *ah;
	struct ib_ah *ret;
	struct ipath_ibdev *dev = to_idev(pd->device);
	unsigned long flags;

	/* A multicast address requires a GRH (see ch. 8.4.1). */
	if (ah_attr->dlid >= IPATH_MULTICAST_LID_BASE &&
	    ah_attr->dlid != IPATH_PERMISSIVE_LID &&
	    !(ah_attr->ah_flags & IB_AH_GRH)) {
		ret = ERR_PTR(-EINVAL);
		goto bail;
	}

	if (ah_attr->dlid == 0) {
		ret = ERR_PTR(-EINVAL);
		goto bail;
	}

	if (ah_attr->port_num < 1 ||
	    ah_attr->port_num > pd->device->phys_port_cnt) {
		ret = ERR_PTR(-EINVAL);
		goto bail;
	}

	ah = kmalloc(sizeof *ah, GFP_ATOMIC);
	if (!ah) {
		ret = ERR_PTR(-ENOMEM);
		goto bail;
	}

	spin_lock_irqsave(&dev->n_ahs_lock, flags);
	if (dev->n_ahs_allocated == ib_ipath_max_ahs) {
		spin_unlock_irqrestore(&dev->n_ahs_lock, flags);
		kfree(ah);
		ret = ERR_PTR(-ENOMEM);
		goto bail;
	}

	dev->n_ahs_allocated++;
	spin_unlock_irqrestore(&dev->n_ahs_lock, flags);

	/* ib_create_ah() will initialize ah->ibah. */
	ah->attr = *ah_attr;

	ret = &ah->ibah;

bail:
	return ret;
}

/**
 * ipath_destroy_ah - destroy an address handle
 * @ibah: the AH to destroy
 *
 * This may be called from interrupt context.
 */
static int ipath_destroy_ah(struct ib_ah *ibah)
{
	struct ipath_ibdev *dev = to_idev(ibah->device);
	struct ipath_ah *ah = to_iah(ibah);
	unsigned long flags;

	spin_lock_irqsave(&dev->n_ahs_lock, flags);
	dev->n_ahs_allocated--;
	spin_unlock_irqrestore(&dev->n_ahs_lock, flags);

	kfree(ah);

	return 0;
}

static int ipath_query_ah(struct ib_ah *ibah, struct ib_ah_attr *ah_attr)
{
	struct ipath_ah *ah = to_iah(ibah);

	*ah_attr = ah->attr;