Skip to content

L
linux
Commit 0bf48289 authored by Steve Wise's avatar Steve Wise Committed by J. Bruce Fields
Browse files
Options

svcrdma: refactor marshalling logic 

This patch refactors the NFSRDMA server marshalling logic to
remove the intermediary map structures.  It also fixes an existing bug
where the NFSRDMA server was not minding the device fast register page
list length limitations.
Signed-off-by: default avatarTom Tucker <tom@opengridcomputing.com>
Signed-off-by: default avatarSteve Wise <swise@opengridcomputing.com>
parent 1b19453d
Expand all Show whitespace changes
Inline Side-by-side
Showing with 332 additions and 606 deletions
include/linux/sunrpc/svc_rdma.h
View file @ 0bf48289
......@@ -115,14 +115,13 @@ struct svc_rdma_fastreg_mr {
struct list_head frmr_list;
};
struct svc_rdma_req_map {
struct svc_rdma_fastreg_mr *frmr;
unsigned long count;
union {
struct kvec sge[RPCSVC_MAXPAGES];
struct svc_rdma_chunk_sge ch[RPCSVC_MAXPAGES];
unsigned long lkey[RPCSVC_MAXPAGES];
};
};
#define RDMACTXT_F_FAST_UNREG 1
#define RDMACTXT_F_LAST_CTXT 2
#define SVCRDMA_DEVCAP_FAST_REG 1 /* fast mr registration */
......
net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
View file @ 0bf48289
This diff is collapsed.
net/sunrpc/xprtrdma/svc_rdma_sendto.c
View file @ 0bf48289
/*
* Copyright (c) 2014 Open Grid Computing, Inc. All rights reserved.
* Copyright (c) 2005-2006 Network Appliance, Inc. All rights reserved.
*
* This software is available to you under a choice of one of two
......@@ -49,152 +50,6 @@
#define RPCDBG_FACILITY RPCDBG_SVCXPRT
/* Encode an XDR as an array of IB SGE
*
* Assumptions:
* - head[0] is physically contiguous.
* - tail[0] is physically contiguous.
* - pages[] is not physically or virtually contiguous and consists of
* PAGE_SIZE elements.
*
* Output:
* SGE[0] reserved for RCPRDMA header
* SGE[1] data from xdr->head[]
* SGE[2..sge_count-2] data from xdr->pages[]
* SGE[sge_count-1] data from xdr->tail.
*
* The max SGE we need is the length of the XDR / pagesize + one for
* head + one for tail + one for RPCRDMA header. Since RPCSVC_MAXPAGES
* reserves a page for both the request and the reply header, and this
* array is only concerned with the reply we are assured that we have
* on extra page for the RPCRMDA header.
*/
static int fast_reg_xdr(struct svcxprt_rdma *xprt,
struct xdr_buf *xdr,
struct svc_rdma_req_map *vec)
{
int sge_no;
u32 sge_bytes;
u32 page_bytes;
u32 page_off;
int page_no = 0;
u8 *frva;
struct svc_rdma_fastreg_mr *frmr;
frmr = svc_rdma_get_frmr(xprt);
if (IS_ERR(frmr))
return -ENOMEM;
vec->frmr = frmr;
/* Skip the RPCRDMA header */
sge_no = 1;
/* Map the head. */
frva = (void *)((unsigned long)(xdr->head[0].iov_base) & PAGE_MASK);
vec->sge[sge_no].iov_base = xdr->head[0].iov_base;
vec->sge[sge_no].iov_len = xdr->head[0].iov_len;
vec->count = 2;
sge_no++;
/* Map the XDR head */
frmr->kva = frva;
frmr->direction = DMA_TO_DEVICE;
frmr->access_flags = 0;
frmr->map_len = PAGE_SIZE;
frmr->page_list_len = 1;
page_off = (unsigned long)xdr->head[0].iov_base & ~PAGE_MASK;
frmr->page_list->page_list[page_no] =
ib_dma_map_page(xprt->sc_cm_id->device,
virt_to_page(xdr->head[0].iov_base),
page_off,
PAGE_SIZE - page_off,
DMA_TO_DEVICE);
if (ib_dma_mapping_error(xprt->sc_cm_id->device,
frmr->page_list->page_list[page_no]))
goto fatal_err;
atomic_inc(&xprt->sc_dma_used);
/* Map the XDR page list */
page_off = xdr->page_base;
page_bytes = xdr->page_len + page_off;
if (!page_bytes)
goto encode_tail;
/* Map the pages */
vec->sge[sge_no].iov_base = frva + frmr->map_len + page_off;
vec->sge[sge_no].iov_len = page_bytes;
sge_no++;
while (page_bytes) {
struct page *page;
page = xdr->pages[page_no++];
sge_bytes = min_t(u32, page_bytes, (PAGE_SIZE - page_off));
page_bytes -= sge_bytes;
frmr->page_list->page_list[page_no] =
ib_dma_map_page(xprt->sc_cm_id->device,
page, page_off,
sge_bytes, DMA_TO_DEVICE);
if (ib_dma_mapping_error(xprt->sc_cm_id->device,
frmr->page_list->page_list[page_no]))
goto fatal_err;
atomic_inc(&xprt->sc_dma_used);
page_off = 0; /* reset for next time through loop */
frmr->map_len += PAGE_SIZE;
frmr->page_list_len++;
}
vec->count++;
encode_tail:
/* Map tail */
if (0 == xdr->tail[0].iov_len)
goto done;
vec->count++;
vec->sge[sge_no].iov_len = xdr->tail[0].iov_len;
if (((unsigned long)xdr->tail[0].iov_base & PAGE_MASK) ==
((unsigned long)xdr->head[0].iov_base & PAGE_MASK)) {
/*
* If head and tail use the same page, we don't need
* to map it again.
*/
vec->sge[sge_no].iov_base = xdr->tail[0].iov_base;
} else {
void *va;
/* Map another page for the tail */
page_off = (unsigned long)xdr->tail[0].iov_base & ~PAGE_MASK;
va = (void *)((unsigned long)xdr->tail[0].iov_base & PAGE_MASK);
vec->sge[sge_no].iov_base = frva + frmr->map_len + page_off;
frmr->page_list->page_list[page_no] =
ib_dma_map_page(xprt->sc_cm_id->device, virt_to_page(va),
page_off,
PAGE_SIZE,
DMA_TO_DEVICE);
if (ib_dma_mapping_error(xprt->sc_cm_id->device,
frmr->page_list->page_list[page_no]))
goto fatal_err;
atomic_inc(&xprt->sc_dma_used);
frmr->map_len += PAGE_SIZE;
frmr->page_list_len++;
}
done:
if (svc_rdma_fastreg(xprt, frmr))
goto fatal_err;
return 0;
fatal_err:
printk("svcrdma: Error fast registering memory for xprt %p\n", xprt);
vec->frmr = NULL;
svc_rdma_put_frmr(xprt, frmr);
return -EIO;
}
static int map_xdr(struct svcxprt_rdma *xprt,
struct xdr_buf *xdr,
struct svc_rdma_req_map *vec)
......@@ -208,9 +63,6 @@ static int map_xdr(struct svcxprt_rdma *xprt,
BUG_ON(xdr->len !=
(xdr->head[0].iov_len + xdr->page_len + xdr->tail[0].iov_len));
if (xprt->sc_frmr_pg_list_len)
return fast_reg_xdr(xprt, xdr, vec);
/* Skip the first sge, this is for the RPCRDMA header */
sge_no = 1;
......@@ -282,8 +134,6 @@ static dma_addr_t dma_map_xdr(struct svcxprt_rdma *xprt,
}
/* Assumptions:
* - We are using FRMR
* - or -
* - The specified write_len can be represented in sc_max_sge * PAGE_SIZE
*/
static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp,
......@@ -327,7 +177,6 @@ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp,
sge_bytes = min_t(size_t,
bc, vec->sge[xdr_sge_no].iov_len-sge_off);
sge[sge_no].length = sge_bytes;
if (!vec->frmr) {
sge[sge_no].addr =
dma_map_xdr(xprt, &rqstp->rq_res, xdr_off,
sge_bytes, DMA_TO_DEVICE);
......@@ -337,13 +186,7 @@ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp,
goto err;
atomic_inc(&xprt->sc_dma_used);
sge[sge_no].lkey = xprt->sc_dma_lkey;
} else {
sge[sge_no].addr = (unsigned long)
vec->sge[xdr_sge_no].iov_base + sge_off;
sge[sge_no].lkey = vec->frmr->mr->lkey;
}
ctxt->count++;
ctxt->frmr = vec->frmr;
sge_off = 0;
sge_no++;
xdr_sge_no++;
......@@ -369,7 +212,6 @@ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp,
return 0;
err:
svc_rdma_unmap_dma(ctxt);
svc_rdma_put_frmr(xprt, vec->frmr);
svc_rdma_put_context(ctxt, 0);
/* Fatal error, close transport */
return -EIO;
......@@ -397,9 +239,6 @@ static int send_write_chunks(struct svcxprt_rdma *xprt,
res_ary = (struct rpcrdma_write_array *)
&rdma_resp->rm_body.rm_chunks[1];
if (vec->frmr)
max_write = vec->frmr->map_len;
else
max_write = xprt->sc_max_sge * PAGE_SIZE;
/* Write chunks start at the pagelist */
......@@ -472,9 +311,6 @@ static int send_reply_chunks(struct svcxprt_rdma *xprt,
res_ary = (struct rpcrdma_write_array *)
&rdma_resp->rm_body.rm_chunks[2];
if (vec->frmr)
max_write = vec->frmr->map_len;
else
max_write = xprt->sc_max_sge * PAGE_SIZE;
/* xdr offset starts at RPC message */
......@@ -545,7 +381,6 @@ static int send_reply(struct svcxprt_rdma *rdma,
int byte_count)
{
struct ib_send_wr send_wr;
struct ib_send_wr inv_wr;
int sge_no;
int sge_bytes;
int page_no;
......@@ -559,7 +394,6 @@ static int send_reply(struct svcxprt_rdma *rdma,
"svcrdma: could not post a receive buffer, err=%d."
"Closing transport %p.\n", ret, rdma);
set_bit(XPT_CLOSE, &rdma->sc_xprt.xpt_flags);
svc_rdma_put_frmr(rdma, vec->frmr);
svc_rdma_put_context(ctxt, 0);
return -ENOTCONN;
}
......@@ -567,11 +401,6 @@ static int send_reply(struct svcxprt_rdma *rdma,
/* Prepare the context */
ctxt->pages[0] = page;
ctxt->count = 1;
ctxt->frmr = vec->frmr;
if (vec->frmr)
set_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags);
else
clear_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags);
/* Prepare the SGE for the RPCRDMA Header */
ctxt->sge[0].lkey = rdma->sc_dma_lkey;
......@@ -590,7 +419,6 @@ static int send_reply(struct svcxprt_rdma *rdma,
int xdr_off = 0;
sge_bytes = min_t(size_t, vec->sge[sge_no].iov_len, byte_count);
byte_count -= sge_bytes;
if (!vec->frmr) {
ctxt->sge[sge_no].addr =
dma_map_xdr(rdma, &rqstp->rq_res, xdr_off,
sge_bytes, DMA_TO_DEVICE);
......@@ -600,11 +428,6 @@ static int send_reply(struct svcxprt_rdma *rdma,
goto err;
atomic_inc(&rdma->sc_dma_used);
ctxt->sge[sge_no].lkey = rdma->sc_dma_lkey;
} else {
ctxt->sge[sge_no].addr = (unsigned long)
vec->sge[sge_no].iov_base;
ctxt->sge[sge_no].lkey = vec->frmr->mr->lkey;
}
ctxt->sge[sge_no].length = sge_bytes;
}
BUG_ON(byte_count != 0);
......@@ -627,6 +450,7 @@ static int send_reply(struct svcxprt_rdma *rdma,
ctxt->sge[page_no+1].length = 0;
}
rqstp->rq_next_page = rqstp->rq_respages + 1;
BUG_ON(sge_no > rdma->sc_max_sge);
memset(&send_wr, 0, sizeof send_wr);
ctxt->wr_op = IB_WR_SEND;
......@@ -635,15 +459,6 @@ static int send_reply(struct svcxprt_rdma *rdma,
send_wr.num_sge = sge_no;
send_wr.opcode = IB_WR_SEND;
send_wr.send_flags = IB_SEND_SIGNALED;
if (vec->frmr) {
/* Prepare INVALIDATE WR */
memset(&inv_wr, 0, sizeof inv_wr);
inv_wr.opcode = IB_WR_LOCAL_INV;
inv_wr.send_flags = IB_SEND_SIGNALED;
inv_wr.ex.invalidate_rkey =
vec->frmr->mr->lkey;
send_wr.next = &inv_wr;
}
ret = svc_rdma_send(rdma, &send_wr);
if (ret)
......@@ -653,7 +468,6 @@ static int send_reply(struct svcxprt_rdma *rdma,
err:
svc_rdma_unmap_dma(ctxt);
svc_rdma_put_frmr(rdma, vec->frmr);
svc_rdma_put_context(ctxt, 1);
return -EIO;
}
......
net/sunrpc/xprtrdma/svc_rdma_transport.c
View file @ 0bf48289
/*
* Copyright (c) 2014 Open Grid Computing, Inc. All rights reserved.
* Copyright (c) 2005-2007 Network Appliance, Inc. All rights reserved.
*
* This software is available to you under a choice of one of two
......@@ -162,7 +163,6 @@ struct svc_rdma_req_map *svc_rdma_get_req_map(void)
schedule_timeout_uninterruptible(msecs_to_jiffies(500));
}
map->count = 0;
map->frmr = NULL;
return map;
}
......@@ -338,22 +338,21 @@ static void process_context(struct svcxprt_rdma *xprt,
switch (ctxt->wr_op) {
case IB_WR_SEND:
if (test_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags))
svc_rdma_put_frmr(xprt, ctxt->frmr);
BUG_ON(ctxt->frmr);
svc_rdma_put_context(ctxt, 1);
break;
case IB_WR_RDMA_WRITE:
BUG_ON(ctxt->frmr);
svc_rdma_put_context(ctxt, 0);
break;
case IB_WR_RDMA_READ:
case IB_WR_RDMA_READ_WITH_INV:
svc_rdma_put_frmr(xprt, ctxt->frmr);
if (test_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags)) {
struct svc_rdma_op_ctxt *read_hdr = ctxt->read_hdr;
BUG_ON(!read_hdr);
if (test_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags))
svc_rdma_put_frmr(xprt, ctxt->frmr);
spin_lock_bh(&xprt->sc_rq_dto_lock);
set_bit(XPT_DATA, &xprt->sc_xprt.xpt_flags);
list_add_tail(&read_hdr->dto_q,
......@@ -365,6 +364,7 @@ static void process_context(struct svcxprt_rdma *xprt,
break;
default:
BUG_ON(1);
printk(KERN_ERR "svcrdma: unexpected completion type, "
"opcode=%d\n",
ctxt->wr_op);
......@@ -380,30 +380,43 @@ static void process_context(struct svcxprt_rdma *xprt,
static void sq_cq_reap(struct svcxprt_rdma *xprt)
{
struct svc_rdma_op_ctxt *ctxt = NULL;
struct ib_wc wc;
struct ib_wc wc_a[6];
struct ib_wc *wc;
struct ib_cq *cq = xprt->sc_sq_cq;
int ret;
memset(wc_a, 0, sizeof(wc_a));
if (!test_and_clear_bit(RDMAXPRT_SQ_PENDING, &xprt->sc_flags))
return;
ib_req_notify_cq(xprt->sc_sq_cq, IB_CQ_NEXT_COMP);
atomic_inc(&rdma_stat_sq_poll);
while ((ret = ib_poll_cq(cq, 1, &wc)) > 0) {
if (wc.status != IB_WC_SUCCESS)
while ((ret = ib_poll_cq(cq, ARRAY_SIZE(wc_a), wc_a)) > 0) {
int i;
for (i = 0; i < ret; i++) {
wc = &wc_a[i];
if (wc->status != IB_WC_SUCCESS) {
dprintk("svcrdma: sq wc err status %d\n",
wc->status);
/* Close the transport */
set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags);
}
/* Decrement used SQ WR count */
atomic_dec(&xprt->sc_sq_count);
wake_up(&xprt->sc_send_wait);
ctxt = (struct svc_rdma_op_ctxt *)(unsigned long)wc.wr_id;
ctxt = (struct svc_rdma_op_ctxt *)
(unsigned long)wc->wr_id;
if (ctxt)
process_context(xprt, ctxt);
svc_xprt_put(&xprt->sc_xprt);
}
}
if (ctxt)
atomic_inc(&rdma_stat_sq_prod);
......@@ -995,7 +1008,11 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt)
need_dma_mr = 0;
break;
case RDMA_TRANSPORT_IB:
if (!(devattr.device_cap_flags & IB_DEVICE_LOCAL_DMA_LKEY)) {
if (!(newxprt->sc_dev_caps & SVCRDMA_DEVCAP_FAST_REG)) {
need_dma_mr = 1;
dma_mr_acc = IB_ACCESS_LOCAL_WRITE;
} else if (!(devattr.device_cap_flags &
IB_DEVICE_LOCAL_DMA_LKEY)) {
need_dma_mr = 1;
dma_mr_acc = IB_ACCESS_LOCAL_WRITE;
} else
......@@ -1192,14 +1209,7 @@ static int svc_rdma_has_wspace(struct svc_xprt *xprt)
container_of(xprt, struct svcxprt_rdma, sc_xprt);
/*
* If there are fewer SQ WR available than required to send a
* simple response, return false.
*/
if ((rdma->sc_sq_depth - atomic_read(&rdma->sc_sq_count) < 3))
return 0;
/*
* ...or there are already waiters on the SQ,
* If there are already waiters on the SQ,
* return false.
*/
if (waitqueue_active(&rdma->sc_send_wait))
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7