linux/drivers/infiniband/hw/ipath/ipath_ud.c

580 lines
16 KiB
C
Raw Normal View History

/*
* Copyright (c) 2006, 2007, 2008 QLogic Corporation. All rights reserved.
* Copyright (c) 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.
*/
#include <rdma/ib_smi.h>
#include "ipath_verbs.h"
#include "ipath_kernel.h"
/**
* ipath_ud_loopback - handle send on loopback QPs
* @sqp: the sending QP
* @swqe: the send work request
*
* This is called from ipath_make_ud_req() to forward a WQE addressed
* to the same HCA.
* Note that the receive interrupt handler may be calling ipath_ud_rcv()
* while this is being called.
*/
static void ipath_ud_loopback(struct ipath_qp *sqp, struct ipath_swqe *swqe)
{
struct ipath_ibdev *dev = to_idev(sqp->ibqp.device);
struct ipath_qp *qp;
struct ib_ah_attr *ah_attr;
unsigned long flags;
struct ipath_rq *rq;
struct ipath_srq *srq;
struct ipath_sge_state rsge;
struct ipath_sge *sge;
struct ipath_rwq *wq;
struct ipath_rwqe *wqe;
void (*handler)(struct ib_event *, void *);
struct ib_wc wc;
u32 tail;
u32 rlen;
u32 length;
qp = ipath_lookup_qpn(&dev->qp_table, swqe->wr.wr.ud.remote_qpn);
if (!qp || !(ib_ipath_state_ops[qp->state] & IPATH_PROCESS_RECV_OK)) {
dev->n_pkt_drops++;
goto done;
}
/*
* Check that the qkey matches (except for QP0, see 9.6.1.4.1).
* Qkeys with the high order bit set mean use the
* qkey from the QP context instead of the WR (see 10.2.5).
*/
if (unlikely(qp->ibqp.qp_num &&
((int) swqe->wr.wr.ud.remote_qkey < 0 ?
sqp->qkey : swqe->wr.wr.ud.remote_qkey) != qp->qkey)) {
/* XXX OK to lose a count once in a while. */
dev->qkey_violations++;
dev->n_pkt_drops++;
goto drop;
}
/*
* A GRH is expected to preceed the data even if not
* present on the wire.
*/
length = swqe->length;
memset(&wc, 0, sizeof wc);
wc.byte_len = length + sizeof(struct ib_grh);
if (swqe->wr.opcode == IB_WR_SEND_WITH_IMM) {
wc.wc_flags = IB_WC_WITH_IMM;
RDMA/core: Add memory management extensions support This patch adds support for the IB "base memory management extension" (BMME) and the equivalent iWARP operations (which the iWARP verbs mandates all devices must implement). The new operations are: - Allocate an ib_mr for use in fast register work requests. - Allocate/free a physical buffer lists for use in fast register work requests. This allows device drivers to allocate this memory as needed for use in posting send requests (eg via dma_alloc_coherent). - New send queue work requests: * send with remote invalidate * fast register memory region * local invalidate memory region * RDMA read with invalidate local memory region (iWARP only) Consumer interface details: - A new device capability flag IB_DEVICE_MEM_MGT_EXTENSIONS is added to indicate device support for these features. - New send work request opcodes IB_WR_FAST_REG_MR, IB_WR_LOCAL_INV, IB_WR_RDMA_READ_WITH_INV are added. - A new consumer API function, ib_alloc_mr() is added to allocate fast register memory regions. - New consumer API functions, ib_alloc_fast_reg_page_list() and ib_free_fast_reg_page_list() are added to allocate and free device-specific memory for fast registration page lists. - A new consumer API function, ib_update_fast_reg_key(), is added to allow the key portion of the R_Key and L_Key of a fast registration MR to be updated. Consumers call this if desired before posting a IB_WR_FAST_REG_MR work request. Consumers can use this as follows: - MR is allocated with ib_alloc_mr(). - Page list memory is allocated with ib_alloc_fast_reg_page_list(). - MR R_Key/L_Key "key" field is updated with ib_update_fast_reg_key(). - MR made VALID and bound to a specific page list via ib_post_send(IB_WR_FAST_REG_MR) - MR made INVALID via ib_post_send(IB_WR_LOCAL_INV), ib_post_send(IB_WR_RDMA_READ_WITH_INV) or an incoming send with invalidate operation. - MR is deallocated with ib_dereg_mr() - page lists dealloced via ib_free_fast_reg_page_list(). Applications can allocate a fast register MR once, and then can repeatedly bind the MR to different physical block lists (PBLs) via posting work requests to a send queue (SQ). For each outstanding MR-to-PBL binding in the SQ pipe, a fast_reg_page_list needs to be allocated (the fast_reg_page_list is owned by the low-level driver from the consumer posting a work request until the request completes). Thus pipelining can be achieved while still allowing device-specific page_list processing. The 32-bit fast register memory key/STag is composed of a 24-bit index and an 8-bit key. The application can change the key each time it fast registers thus allowing more control over the peer's use of the key/STag (ie it can effectively be changed each time the rkey is rebound to a page list). Signed-off-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2008-07-15 14:48:45 +08:00
wc.ex.imm_data = swqe->wr.ex.imm_data;
}
/*
* This would be a lot simpler if we could call ipath_get_rwqe()
* but that uses state that the receive interrupt handler uses
* so we would need to lock out receive interrupts while doing
* local loopback.
*/
if (qp->ibqp.srq) {
srq = to_isrq(qp->ibqp.srq);
handler = srq->ibsrq.event_handler;
rq = &srq->rq;
} else {
srq = NULL;
handler = NULL;
rq = &qp->r_rq;
}
/*
* Get the next work request entry to find where to put the data.
* Note that it is safe to drop the lock after changing rq->tail
* since ipath_post_receive() won't fill the empty slot.
*/
spin_lock_irqsave(&rq->lock, flags);
wq = rq->wq;
tail = wq->tail;
/* Validate tail before using it since it is user writable. */
if (tail >= rq->size)
tail = 0;
if (unlikely(tail == wq->head)) {
spin_unlock_irqrestore(&rq->lock, flags);
dev->n_pkt_drops++;
goto drop;
}
wqe = get_rwqe_ptr(rq, tail);
rsge.sg_list = qp->r_ud_sg_list;
if (!ipath_init_sge(qp, wqe, &rlen, &rsge)) {
spin_unlock_irqrestore(&rq->lock, flags);
dev->n_pkt_drops++;
goto drop;
}
/* Silently drop packets which are too big. */
if (wc.byte_len > rlen) {
spin_unlock_irqrestore(&rq->lock, flags);
dev->n_pkt_drops++;
goto drop;
}
if (++tail >= rq->size)
tail = 0;
wq->tail = tail;
wc.wr_id = wqe->wr_id;
if (handler) {
u32 n;
/*
* validate head pointer value and compute
* the number of remaining WQEs.
*/
n = wq->head;
if (n >= rq->size)
n = 0;
if (n < tail)
n += rq->size - tail;
else
n -= tail;
if (n < srq->limit) {
struct ib_event ev;
srq->limit = 0;
spin_unlock_irqrestore(&rq->lock, flags);
ev.device = qp->ibqp.device;
ev.element.srq = qp->ibqp.srq;
ev.event = IB_EVENT_SRQ_LIMIT_REACHED;
handler(&ev, srq->ibsrq.srq_context);
} else
spin_unlock_irqrestore(&rq->lock, flags);
} else
spin_unlock_irqrestore(&rq->lock, flags);
ah_attr = &to_iah(swqe->wr.wr.ud.ah)->attr;
if (ah_attr->ah_flags & IB_AH_GRH) {
ipath_copy_sge(&rsge, &ah_attr->grh, sizeof(struct ib_grh));
wc.wc_flags |= IB_WC_GRH;
} else
ipath_skip_sge(&rsge, sizeof(struct ib_grh));
sge = swqe->sg_list;
while (length) {
u32 len = sge->length;
if (len > length)
len = length;
if (len > sge->sge_length)
len = sge->sge_length;
BUG_ON(len == 0);
ipath_copy_sge(&rsge, sge->vaddr, len);
sge->vaddr += len;
sge->length -= len;
sge->sge_length -= len;
if (sge->sge_length == 0) {
if (--swqe->wr.num_sge)
sge++;
} else if (sge->length == 0 && sge->mr != NULL) {
if (++sge->n >= IPATH_SEGSZ) {
if (++sge->m >= sge->mr->mapsz)
break;
sge->n = 0;
}
sge->vaddr =
sge->mr->map[sge->m]->segs[sge->n].vaddr;
sge->length =
sge->mr->map[sge->m]->segs[sge->n].length;
}
length -= len;
}
wc.status = IB_WC_SUCCESS;
wc.opcode = IB_WC_RECV;
wc.qp = &qp->ibqp;
wc.src_qp = sqp->ibqp.qp_num;
/* XXX do we know which pkey matched? Only needed for GSI. */
wc.pkey_index = 0;
wc.slid = dev->dd->ipath_lid |
(ah_attr->src_path_bits &
((1 << dev->dd->ipath_lmc) - 1));
wc.sl = ah_attr->sl;
wc.dlid_path_bits =
ah_attr->dlid & ((1 << dev->dd->ipath_lmc) - 1);
wc.port_num = 1;
/* Signal completion event if the solicited bit is set. */
ipath_cq_enter(to_icq(qp->ibqp.recv_cq), &wc,
swqe->wr.send_flags & IB_SEND_SOLICITED);
drop:
if (atomic_dec_and_test(&qp->refcount))
wake_up(&qp->wait);
done:;
}
/**
* ipath_make_ud_req - construct a UD request packet
* @qp: the QP
*
* Return 1 if constructed; otherwise, return 0.
*/
int ipath_make_ud_req(struct ipath_qp *qp)
{
struct ipath_ibdev *dev = to_idev(qp->ibqp.device);
struct ipath_other_headers *ohdr;
struct ib_ah_attr *ah_attr;
struct ipath_swqe *wqe;
unsigned long flags;
u32 nwords;
u32 extra_bytes;
u32 bth0;
u16 lrh0;
u16 lid;
int ret = 0;
int next_cur;
spin_lock_irqsave(&qp->s_lock, flags);
if (!(ib_ipath_state_ops[qp->state] & IPATH_PROCESS_NEXT_SEND_OK)) {
if (!(ib_ipath_state_ops[qp->state] & IPATH_FLUSH_SEND))
goto bail;
/* We are in the error state, flush the work request. */
if (qp->s_last == qp->s_head)
goto bail;
/* If DMAs are in progress, we can't flush immediately. */
if (atomic_read(&qp->s_dma_busy)) {
qp->s_flags |= IPATH_S_WAIT_DMA;
goto bail;
}
wqe = get_swqe_ptr(qp, qp->s_last);
ipath_send_complete(qp, wqe, IB_WC_WR_FLUSH_ERR);
goto done;
}
if (qp->s_cur == qp->s_head)
goto bail;
wqe = get_swqe_ptr(qp, qp->s_cur);
next_cur = qp->s_cur + 1;
if (next_cur >= qp->s_size)
next_cur = 0;
/* Construct the header. */
ah_attr = &to_iah(wqe->wr.wr.ud.ah)->attr;
if (ah_attr->dlid >= IPATH_MULTICAST_LID_BASE) {
if (ah_attr->dlid != IPATH_PERMISSIVE_LID)
dev->n_multicast_xmit++;
else
dev->n_unicast_xmit++;
} else {
dev->n_unicast_xmit++;
lid = ah_attr->dlid & ~((1 << dev->dd->ipath_lmc) - 1);
if (unlikely(lid == dev->dd->ipath_lid)) {
/*
* If DMAs are in progress, we can't generate
* a completion for the loopback packet since
* it would be out of order.
* XXX Instead of waiting, we could queue a
* zero length descriptor so we get a callback.
*/
if (atomic_read(&qp->s_dma_busy)) {
qp->s_flags |= IPATH_S_WAIT_DMA;
goto bail;
}
qp->s_cur = next_cur;
spin_unlock_irqrestore(&qp->s_lock, flags);
ipath_ud_loopback(qp, wqe);
spin_lock_irqsave(&qp->s_lock, flags);
ipath_send_complete(qp, wqe, IB_WC_SUCCESS);
goto done;
}
}
qp->s_cur = next_cur;
extra_bytes = -wqe->length & 3;
nwords = (wqe->length + extra_bytes) >> 2;
/* header size in 32-bit words LRH+BTH+DETH = (8+12+8)/4. */
qp->s_hdrwords = 7;
qp->s_cur_size = wqe->length;
qp->s_cur_sge = &qp->s_sge;
qp->s_dmult = ah_attr->static_rate;
qp->s_wqe = wqe;
qp->s_sge.sge = wqe->sg_list[0];
qp->s_sge.sg_list = wqe->sg_list + 1;
qp->s_sge.num_sge = wqe->wr.num_sge;
if (ah_attr->ah_flags & IB_AH_GRH) {
/* Header size in 32-bit words. */
qp->s_hdrwords += ipath_make_grh(dev, &qp->s_hdr.u.l.grh,
&ah_attr->grh,
qp->s_hdrwords, nwords);
lrh0 = IPATH_LRH_GRH;
ohdr = &qp->s_hdr.u.l.oth;
/*
* Don't worry about sending to locally attached multicast
* QPs. It is unspecified by the spec. what happens.
*/
} else {
/* Header size in 32-bit words. */
lrh0 = IPATH_LRH_BTH;
ohdr = &qp->s_hdr.u.oth;
}
if (wqe->wr.opcode == IB_WR_SEND_WITH_IMM) {
qp->s_hdrwords++;
2008-04-17 12:09:32 +08:00
ohdr->u.ud.imm_data = wqe->wr.ex.imm_data;
bth0 = IB_OPCODE_UD_SEND_ONLY_WITH_IMMEDIATE << 24;
} else
bth0 = IB_OPCODE_UD_SEND_ONLY << 24;
lrh0 |= ah_attr->sl << 4;
if (qp->ibqp.qp_type == IB_QPT_SMI)
lrh0 |= 0xF000; /* Set VL (see ch. 13.5.3.1) */
qp->s_hdr.lrh[0] = cpu_to_be16(lrh0);
qp->s_hdr.lrh[1] = cpu_to_be16(ah_attr->dlid); /* DEST LID */
qp->s_hdr.lrh[2] = cpu_to_be16(qp->s_hdrwords + nwords +
SIZE_OF_CRC);
lid = dev->dd->ipath_lid;
if (lid) {
lid |= ah_attr->src_path_bits &
((1 << dev->dd->ipath_lmc) - 1);
qp->s_hdr.lrh[3] = cpu_to_be16(lid);
} else
qp->s_hdr.lrh[3] = IB_LID_PERMISSIVE;
if (wqe->wr.send_flags & IB_SEND_SOLICITED)
bth0 |= 1 << 23;
bth0 |= extra_bytes << 20;
bth0 |= qp->ibqp.qp_type == IB_QPT_SMI ? IPATH_DEFAULT_P_KEY :
ipath_get_pkey(dev->dd, qp->s_pkey_index);
ohdr->bth[0] = cpu_to_be32(bth0);
/*
* Use the multicast QP if the destination LID is a multicast LID.
*/
ohdr->bth[1] = ah_attr->dlid >= IPATH_MULTICAST_LID_BASE &&
ah_attr->dlid != IPATH_PERMISSIVE_LID ?
__constant_cpu_to_be32(IPATH_MULTICAST_QPN) :
cpu_to_be32(wqe->wr.wr.ud.remote_qpn);
ohdr->bth[2] = cpu_to_be32(qp->s_next_psn++ & IPATH_PSN_MASK);
/*
* Qkeys with the high order bit set mean use the
* qkey from the QP context instead of the WR (see 10.2.5).
*/
ohdr->u.ud.deth[0] = cpu_to_be32((int)wqe->wr.wr.ud.remote_qkey < 0 ?
qp->qkey : wqe->wr.wr.ud.remote_qkey);
ohdr->u.ud.deth[1] = cpu_to_be32(qp->ibqp.qp_num);
done:
ret = 1;
goto unlock;
bail:
qp->s_flags &= ~IPATH_S_BUSY;
unlock:
spin_unlock_irqrestore(&qp->s_lock, flags);
return ret;
}
/**
* ipath_ud_rcv - receive an incoming UD packet
* @dev: the device the packet came in on
* @hdr: the packet header
* @has_grh: true if the packet has a GRH
* @data: the packet data
* @tlen: the packet length
* @qp: the QP the packet came on
*
* This is called from ipath_qp_rcv() to process an incoming UD packet
* for the given QP.
* Called at interrupt level.
*/
void ipath_ud_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr,
int has_grh, void *data, u32 tlen, struct ipath_qp *qp)
{
struct ipath_other_headers *ohdr;
int opcode;
u32 hdrsize;
u32 pad;
struct ib_wc wc;
u32 qkey;
u32 src_qp;
u16 dlid;
int header_in_data;
/* Check for GRH */
if (!has_grh) {
ohdr = &hdr->u.oth;
hdrsize = 8 + 12 + 8; /* LRH + BTH + DETH */
qkey = be32_to_cpu(ohdr->u.ud.deth[0]);
src_qp = be32_to_cpu(ohdr->u.ud.deth[1]);
header_in_data = 0;
} else {
ohdr = &hdr->u.l.oth;
hdrsize = 8 + 40 + 12 + 8; /* LRH + GRH + BTH + DETH */
/*
* The header with GRH is 68 bytes and the core driver sets
* the eager header buffer size to 56 bytes so the last 12
* bytes of the IB header is in the data buffer.
*/
header_in_data = dev->dd->ipath_rcvhdrentsize == 16;
if (header_in_data) {
qkey = be32_to_cpu(((__be32 *) data)[1]);
src_qp = be32_to_cpu(((__be32 *) data)[2]);
data += 12;
} else {
qkey = be32_to_cpu(ohdr->u.ud.deth[0]);
src_qp = be32_to_cpu(ohdr->u.ud.deth[1]);
}
}
src_qp &= IPATH_QPN_MASK;
/*
* Check that the permissive LID is only used on QP0
* and the QKEY matches (see 9.6.1.4.1 and 9.6.1.5.1).
*/
if (qp->ibqp.qp_num) {
if (unlikely(hdr->lrh[1] == IB_LID_PERMISSIVE ||
hdr->lrh[3] == IB_LID_PERMISSIVE)) {
dev->n_pkt_drops++;
goto bail;
}
if (unlikely(qkey != qp->qkey)) {
/* XXX OK to lose a count once in a while. */
dev->qkey_violations++;
dev->n_pkt_drops++;
goto bail;
}
} else if (hdr->lrh[1] == IB_LID_PERMISSIVE ||
hdr->lrh[3] == IB_LID_PERMISSIVE) {
struct ib_smp *smp = (struct ib_smp *) data;
if (smp->mgmt_class != IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE) {
dev->n_pkt_drops++;
goto bail;
}
}
/*
* The opcode is in the low byte when its in network order
* (top byte when in host order).
*/
opcode = be32_to_cpu(ohdr->bth[0]) >> 24;
if (qp->ibqp.qp_num > 1 &&
opcode == IB_OPCODE_UD_SEND_ONLY_WITH_IMMEDIATE) {
if (header_in_data) {
RDMA/core: Add memory management extensions support This patch adds support for the IB "base memory management extension" (BMME) and the equivalent iWARP operations (which the iWARP verbs mandates all devices must implement). The new operations are: - Allocate an ib_mr for use in fast register work requests. - Allocate/free a physical buffer lists for use in fast register work requests. This allows device drivers to allocate this memory as needed for use in posting send requests (eg via dma_alloc_coherent). - New send queue work requests: * send with remote invalidate * fast register memory region * local invalidate memory region * RDMA read with invalidate local memory region (iWARP only) Consumer interface details: - A new device capability flag IB_DEVICE_MEM_MGT_EXTENSIONS is added to indicate device support for these features. - New send work request opcodes IB_WR_FAST_REG_MR, IB_WR_LOCAL_INV, IB_WR_RDMA_READ_WITH_INV are added. - A new consumer API function, ib_alloc_mr() is added to allocate fast register memory regions. - New consumer API functions, ib_alloc_fast_reg_page_list() and ib_free_fast_reg_page_list() are added to allocate and free device-specific memory for fast registration page lists. - A new consumer API function, ib_update_fast_reg_key(), is added to allow the key portion of the R_Key and L_Key of a fast registration MR to be updated. Consumers call this if desired before posting a IB_WR_FAST_REG_MR work request. Consumers can use this as follows: - MR is allocated with ib_alloc_mr(). - Page list memory is allocated with ib_alloc_fast_reg_page_list(). - MR R_Key/L_Key "key" field is updated with ib_update_fast_reg_key(). - MR made VALID and bound to a specific page list via ib_post_send(IB_WR_FAST_REG_MR) - MR made INVALID via ib_post_send(IB_WR_LOCAL_INV), ib_post_send(IB_WR_RDMA_READ_WITH_INV) or an incoming send with invalidate operation. - MR is deallocated with ib_dereg_mr() - page lists dealloced via ib_free_fast_reg_page_list(). Applications can allocate a fast register MR once, and then can repeatedly bind the MR to different physical block lists (PBLs) via posting work requests to a send queue (SQ). For each outstanding MR-to-PBL binding in the SQ pipe, a fast_reg_page_list needs to be allocated (the fast_reg_page_list is owned by the low-level driver from the consumer posting a work request until the request completes). Thus pipelining can be achieved while still allowing device-specific page_list processing. The 32-bit fast register memory key/STag is composed of a 24-bit index and an 8-bit key. The application can change the key each time it fast registers thus allowing more control over the peer's use of the key/STag (ie it can effectively be changed each time the rkey is rebound to a page list). Signed-off-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2008-07-15 14:48:45 +08:00
wc.ex.imm_data = *(__be32 *) data;
data += sizeof(__be32);
} else
RDMA/core: Add memory management extensions support This patch adds support for the IB "base memory management extension" (BMME) and the equivalent iWARP operations (which the iWARP verbs mandates all devices must implement). The new operations are: - Allocate an ib_mr for use in fast register work requests. - Allocate/free a physical buffer lists for use in fast register work requests. This allows device drivers to allocate this memory as needed for use in posting send requests (eg via dma_alloc_coherent). - New send queue work requests: * send with remote invalidate * fast register memory region * local invalidate memory region * RDMA read with invalidate local memory region (iWARP only) Consumer interface details: - A new device capability flag IB_DEVICE_MEM_MGT_EXTENSIONS is added to indicate device support for these features. - New send work request opcodes IB_WR_FAST_REG_MR, IB_WR_LOCAL_INV, IB_WR_RDMA_READ_WITH_INV are added. - A new consumer API function, ib_alloc_mr() is added to allocate fast register memory regions. - New consumer API functions, ib_alloc_fast_reg_page_list() and ib_free_fast_reg_page_list() are added to allocate and free device-specific memory for fast registration page lists. - A new consumer API function, ib_update_fast_reg_key(), is added to allow the key portion of the R_Key and L_Key of a fast registration MR to be updated. Consumers call this if desired before posting a IB_WR_FAST_REG_MR work request. Consumers can use this as follows: - MR is allocated with ib_alloc_mr(). - Page list memory is allocated with ib_alloc_fast_reg_page_list(). - MR R_Key/L_Key "key" field is updated with ib_update_fast_reg_key(). - MR made VALID and bound to a specific page list via ib_post_send(IB_WR_FAST_REG_MR) - MR made INVALID via ib_post_send(IB_WR_LOCAL_INV), ib_post_send(IB_WR_RDMA_READ_WITH_INV) or an incoming send with invalidate operation. - MR is deallocated with ib_dereg_mr() - page lists dealloced via ib_free_fast_reg_page_list(). Applications can allocate a fast register MR once, and then can repeatedly bind the MR to different physical block lists (PBLs) via posting work requests to a send queue (SQ). For each outstanding MR-to-PBL binding in the SQ pipe, a fast_reg_page_list needs to be allocated (the fast_reg_page_list is owned by the low-level driver from the consumer posting a work request until the request completes). Thus pipelining can be achieved while still allowing device-specific page_list processing. The 32-bit fast register memory key/STag is composed of a 24-bit index and an 8-bit key. The application can change the key each time it fast registers thus allowing more control over the peer's use of the key/STag (ie it can effectively be changed each time the rkey is rebound to a page list). Signed-off-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2008-07-15 14:48:45 +08:00
wc.ex.imm_data = ohdr->u.ud.imm_data;
wc.wc_flags = IB_WC_WITH_IMM;
hdrsize += sizeof(u32);
} else if (opcode == IB_OPCODE_UD_SEND_ONLY) {
RDMA/core: Add memory management extensions support This patch adds support for the IB "base memory management extension" (BMME) and the equivalent iWARP operations (which the iWARP verbs mandates all devices must implement). The new operations are: - Allocate an ib_mr for use in fast register work requests. - Allocate/free a physical buffer lists for use in fast register work requests. This allows device drivers to allocate this memory as needed for use in posting send requests (eg via dma_alloc_coherent). - New send queue work requests: * send with remote invalidate * fast register memory region * local invalidate memory region * RDMA read with invalidate local memory region (iWARP only) Consumer interface details: - A new device capability flag IB_DEVICE_MEM_MGT_EXTENSIONS is added to indicate device support for these features. - New send work request opcodes IB_WR_FAST_REG_MR, IB_WR_LOCAL_INV, IB_WR_RDMA_READ_WITH_INV are added. - A new consumer API function, ib_alloc_mr() is added to allocate fast register memory regions. - New consumer API functions, ib_alloc_fast_reg_page_list() and ib_free_fast_reg_page_list() are added to allocate and free device-specific memory for fast registration page lists. - A new consumer API function, ib_update_fast_reg_key(), is added to allow the key portion of the R_Key and L_Key of a fast registration MR to be updated. Consumers call this if desired before posting a IB_WR_FAST_REG_MR work request. Consumers can use this as follows: - MR is allocated with ib_alloc_mr(). - Page list memory is allocated with ib_alloc_fast_reg_page_list(). - MR R_Key/L_Key "key" field is updated with ib_update_fast_reg_key(). - MR made VALID and bound to a specific page list via ib_post_send(IB_WR_FAST_REG_MR) - MR made INVALID via ib_post_send(IB_WR_LOCAL_INV), ib_post_send(IB_WR_RDMA_READ_WITH_INV) or an incoming send with invalidate operation. - MR is deallocated with ib_dereg_mr() - page lists dealloced via ib_free_fast_reg_page_list(). Applications can allocate a fast register MR once, and then can repeatedly bind the MR to different physical block lists (PBLs) via posting work requests to a send queue (SQ). For each outstanding MR-to-PBL binding in the SQ pipe, a fast_reg_page_list needs to be allocated (the fast_reg_page_list is owned by the low-level driver from the consumer posting a work request until the request completes). Thus pipelining can be achieved while still allowing device-specific page_list processing. The 32-bit fast register memory key/STag is composed of a 24-bit index and an 8-bit key. The application can change the key each time it fast registers thus allowing more control over the peer's use of the key/STag (ie it can effectively be changed each time the rkey is rebound to a page list). Signed-off-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Roland Dreier <rolandd@cisco.com>
2008-07-15 14:48:45 +08:00
wc.ex.imm_data = 0;
wc.wc_flags = 0;
} else {
dev->n_pkt_drops++;
goto bail;
}
/* Get the number of bytes the message was padded by. */
pad = (be32_to_cpu(ohdr->bth[0]) >> 20) & 3;
if (unlikely(tlen < (hdrsize + pad + 4))) {
/* Drop incomplete packets. */
dev->n_pkt_drops++;
goto bail;
}
tlen -= hdrsize + pad + 4;
/* Drop invalid MAD packets (see 13.5.3.1). */
if (unlikely((qp->ibqp.qp_num == 0 &&
(tlen != 256 ||
(be16_to_cpu(hdr->lrh[0]) >> 12) != 15)) ||
(qp->ibqp.qp_num == 1 &&
(tlen != 256 ||
(be16_to_cpu(hdr->lrh[0]) >> 12) == 15)))) {
dev->n_pkt_drops++;
goto bail;
}
/*
* A GRH is expected to preceed the data even if not
* present on the wire.
*/
wc.byte_len = tlen + sizeof(struct ib_grh);
/*
* Get the next work request entry to find where to put the data.
*/
if (qp->r_flags & IPATH_R_REUSE_SGE)
qp->r_flags &= ~IPATH_R_REUSE_SGE;
else if (!ipath_get_rwqe(qp, 0)) {
/*
* Count VL15 packets dropped due to no receive buffer.
* Otherwise, count them as buffer overruns since usually,
* the HW will be able to receive packets even if there are
* no QPs with posted receive buffers.
*/
if (qp->ibqp.qp_num == 0)
dev->n_vl15_dropped++;
else
dev->rcv_errors++;
goto bail;
}
/* Silently drop packets which are too big. */
if (wc.byte_len > qp->r_len) {
qp->r_flags |= IPATH_R_REUSE_SGE;
dev->n_pkt_drops++;
goto bail;
}
if (has_grh) {
ipath_copy_sge(&qp->r_sge, &hdr->u.l.grh,
sizeof(struct ib_grh));
wc.wc_flags |= IB_WC_GRH;
} else
ipath_skip_sge(&qp->r_sge, sizeof(struct ib_grh));
ipath_copy_sge(&qp->r_sge, data,
wc.byte_len - sizeof(struct ib_grh));
if (!test_and_clear_bit(IPATH_R_WRID_VALID, &qp->r_aflags))
goto bail;
wc.wr_id = qp->r_wr_id;
wc.status = IB_WC_SUCCESS;
wc.opcode = IB_WC_RECV;
wc.vendor_err = 0;
wc.qp = &qp->ibqp;
wc.src_qp = src_qp;
/* XXX do we know which pkey matched? Only needed for GSI. */
wc.pkey_index = 0;
wc.slid = be16_to_cpu(hdr->lrh[3]);
wc.sl = (be16_to_cpu(hdr->lrh[0]) >> 4) & 0xF;
dlid = be16_to_cpu(hdr->lrh[1]);
/*
* Save the LMC lower bits if the destination LID is a unicast LID.
*/
wc.dlid_path_bits = dlid >= IPATH_MULTICAST_LID_BASE ? 0 :
dlid & ((1 << dev->dd->ipath_lmc) - 1);
wc.port_num = 1;
/* Signal completion event if the solicited bit is set. */
ipath_cq_enter(to_icq(qp->ibqp.recv_cq), &wc,
(ohdr->bth[0] &
__constant_cpu_to_be32(1 << 23)) != 0);
bail:;
}