/*
* Copyright (c) 2016, 2017 Oracle. All rights reserved.
* 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
* 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 BSD-type
* 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.
*
* Neither the name of the Network Appliance, Inc. nor the names of
* its contributors may be used to endorse or promote products
* derived from this software without specific prior written
* permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Author: Tom Tucker <tom@opengridcomputing.com>
*/
/* Operation
*
* The main entry point is svc_rdma_recvfrom. This is called from
* svc_recv when the transport indicates there is incoming data to
* be read. "Data Ready" is signaled when an RDMA Receive completes,
* or when a set of RDMA Reads complete.
*
* An svc_rqst is passed in. This structure contains an array of
* free pages (rq_pages) that will contain the incoming RPC message.
*
* Short messages are moved directly into svc_rqst::rq_arg, and
* the RPC Call is ready to be processed by the Upper Layer.
* svc_rdma_recvfrom returns the length of the RPC Call message,
* completing the reception of the RPC Call.
*
* However, when an incoming message has Read chunks,
* svc_rdma_recvfrom must post RDMA Reads to pull the RPC Call's
* data payload from the client. svc_rdma_recvfrom sets up the
* RDMA Reads using pages in svc_rqst::rq_pages, which are
* transferred to an svc_rdma_op_ctxt for the duration of the
* I/O. svc_rdma_recvfrom then returns zero, since the RPC message
* is still not yet ready.
*
* When the Read chunk payloads have become available on the
* server, "Data Ready" is raised again, and svc_recv calls
* svc_rdma_recvfrom again. This second call may use a different
* svc_rqst than the first one, thus any information that needs
* to be preserved across these two calls is kept in an
* svc_rdma_op_ctxt.
*
* The second call to svc_rdma_recvfrom performs final assembly
* of the RPC Call message, using the RDMA Read sink pages kept in
* the svc_rdma_op_ctxt. The xdr_buf is copied from the
* svc_rdma_op_ctxt to the second svc_rqst. The second call returns
* the length of the completed RPC Call message.
*
* Page Management
*
* Pages under I/O must be transferred from the first svc_rqst to an
* svc_rdma_op_ctxt before the first svc_rdma_recvfrom call returns.
*
* The first svc_rqst supplies pages for RDMA Reads. These are moved
* from rqstp::rq_pages into ctxt::pages. The consumed elements of
* the rq_pages array are set to NULL and refilled with the first
* svc_rdma_recvfrom call returns.
*
* During the second svc_rdma_recvfrom call, RDMA Read sink pages
* are transferred from the svc_rdma_op_ctxt to the second svc_rqst
* (see rdma_read_complete() below).
*/
#include <asm/unaligned.h>
#include <rdma/ib_verbs.h>
#include <rdma/rdma_cm.h>
#include <linux/spinlock.h>
#include <linux/sunrpc/xdr.h>
#include <linux/sunrpc/debug.h>
#include <linux/sunrpc/rpc_rdma.h>
#include <linux/sunrpc/svc_rdma.h>
#define RPCDBG_FACILITY RPCDBG_SVCXPRT
/*
* Replace the pages in the rq_argpages array with the pages from the SGE in
* the RDMA_RECV completion. The SGL should contain full pages up until the
* last one.
*/
static void rdma_build_arg_xdr(struct svc_rqst *rqstp,
struct svc_rdma_op_ctxt *ctxt,
u32 byte_count)
{
struct page *page;
u32 bc;
int sge_no;
/* Swap the page in the SGE with the page in argpages */
page = ctxt->pages[0];
put_page(rqstp->rq_pages[0]);
rqstp->rq_pages[0] = page;
/* Set up the XDR head */
rqstp->rq_arg.head[0].iov_base = page_address(page);
rqstp->rq_arg.head[0].iov_len =
min_t(size_t, byte_count, ctxt->sge[0].length);
rqstp->rq_arg.len = byte_count;
rqstp->rq_arg.buflen = byte_count;
/* Compute bytes past head in the SGL */
bc = byte_count - rqstp->rq_arg.head[0].iov_len;
/* If data remains, store it in the pagelist */
rqstp->rq_arg.page_len = bc;
rqstp->rq_arg.page_base = 0;
sge_no = 1;
while (bc && sge_no < ctxt->count) {
page = ctxt->pages[sge_no];
put_page(rqstp->rq_pages[sge_no]);
rqstp->rq_pages[sge_no] = page;
bc -= min_t(u32, bc, ctxt->sge[sge_no].length);
sge_no++;
}
rqstp->rq_respages = &rqstp->rq_pages[sge_no];
rqstp->rq_next_page = rqstp->rq_respages + 1;
/* If not all pages were used from the SGL, free the remaining ones */
bc = sge_no;
while (sge_no < ctxt->count) {
page = ctxt->pages[sge_no++];
put_page(page);
}
ctxt->count = bc;
/* Set up tail */
rqstp->rq_arg.tail[0].iov_base = NULL;
rqstp->rq_arg.tail[0].iov_len = 0;
}
/* This accommodates the largest possible Write chunk,
* in one segment.
*/
#define MAX_BYTES_WRITE_SEG ((u32)(RPCSVC_MAXPAGES << PAGE_SHIFT))
/* This accommodates the largest possible Position-Zero
* Read chunk or Reply chunk, in one segment.
*/
#define MAX_BYTES_SPECIAL_SEG ((u32)((RPCSVC_MAXPAGES + 2) << PAGE_SHIFT))
/* Sanity check the Read list.
*
* Implementation limits:
* - This implementation supports only one Read chunk.
*
* Sanity checks:
* - Read list does not overflow buffer.
* - Segment size limited by largest NFS data payload.
*
* The segment count is limited to how many segments can
* fit in the transport header without overflowing the
* buffer. That's about 40 Read segments for a 1KB inline
* threshold.
*
* Returns pointer to the following Write list.
*/
static __be32 *xdr_check_read_list(__be32 *p, const __be32 *end)
{
u32 position;
bool first;
first = true;
while (*p++ != xdr_zero) {
if (first) {
position = be32_to_cpup(p++);
first = false;
} else if (be32_to_cpup(p++) != position) {
return NULL;
}
p++; /* handle */
if (be32_to_cpup(p++) > MAX_BYTES_SPECIAL_SEG)
return NULL;
p += 2; /* offset */
if (p > end)
return NULL;
}
return p;
}
/* The segment count is limited to how many segments can
* fit in the transport header without overflowing the
* buffer. That's about 60 Write segments for a 1KB inline
* threshold.
*/
static __be32 *xdr_check_write_chunk(__be32 *p, const __be32 *end,
u32 maxlen)
{
u32 i, segcount;
segcount = be32_to_cpup(p++);
for (i = 0; i < segcount; i++) {
p++; /* handle */
if (be32_to_cpup(p++) > maxlen)
return NULL;
p += 2; /* offset */
if (p > end)
return NULL;
}
return p;
}
/* Sanity check the Write list.
*
* Implementation limits:
* - This implementation supports only one Write chunk.
*
* Sanity checks:
* - Write list does not overflow buffer.
* - Segment size limited by largest NFS data payload.
*
* Returns pointer to the following Reply chunk.
*/
static __be32 *xdr_check_write_list(__be32 *p, const __be32 *end)
{
u32 chcount;
chcount = 0;
while (*p++ != xdr_zero) {
p = xdr_check_write_chunk(p, end, MAX_BYTES_WRITE_SEG);
if (!p)
return NULL;
if (chcount++ > 1)
return NULL;
}
return p;
}
/* Sanity check the Reply chunk.
*
* Sanity checks:
* - Reply chunk does not overflow buffer.
* - Segment size limited by largest NFS data payload.
*
* Returns pointer to the following RPC header.
*/
static __be32 *xdr_check_reply_chunk(__be32 *p, const __be32 *end)
{
if (*p++ != xdr_zero) {
p = xdr_check_write_chunk(p, end, MAX_BYTES_SPECIAL_SEG);
if (!p)
return NULL;
}
return p;
}
/* On entry, xdr->head[0].iov_base points to first byte in the
* RPC-over-RDMA header.
*
* On successful exit, head[0] points to first byte past the
* RPC-over-RDMA header. For RDMA_MSG, this is the RPC message.
* The length of the RPC-over-RDMA header is returned.
*
* Assumptions:
* - The transport header is entirely contained in the head iovec.
*/
static int svc_rdma_xdr_decode_req(struct xdr_buf *rq_arg)
{
__be32 *p, *end, *rdma_argp;
unsigned int hdr_len;
char *proc;
/* Verify that there's enough bytes for header + something */
if (rq_arg->len <= RPCRDMA_HDRLEN_ERR)
goto out_short;
rdma_argp = rq_arg->head[0].iov_base;
if (*(rdma_argp + 1) != rpcrdma_version)
goto out_version;
switch (*(rdma_argp + 3)) {
case rdma_msg:
proc = "RDMA_MSG";
break;
case rdma_nomsg:
proc = "RDMA_NOMSG";
break;
case rdma_done:
goto out_drop;
case rdma_error:
goto out_drop;
default:
goto out_proc;
}
end = (__be32 *)((unsigned long)rdma_argp + rq_arg->len);
p = xdr_check_read_list(rdma_argp + 4, end);
if (!p)
goto out_inval;
p = xdr_check_write_list(p, end);
if (!p)
goto out_inval;
p = xdr_check_reply_chunk(p, end);
if (!p)
goto out_inval;
if (p > end)
goto out_inval;
rq_arg->head[0].iov_base = p;
hdr_len = (unsigned long)p - (unsigned long)rdma_argp;
rq_arg->head[0].iov_len -= hdr_len;
rq_arg->len -= hdr_len;
dprintk("svcrdma: received %s request for XID 0x%08x, hdr_len=%u\n",
proc, be32_to_cpup(rdma_argp), hdr_len);
return hdr_len;
out_short:
dprintk("svcrdma: header too short = %d\n", rq_arg->len);
return -EINVAL;
out_version:
dprintk("svcrdma: bad xprt version: %u\n",
be32_to_cpup(rdma_argp + 1));
return -EPROTONOSUPPORT;
out_drop:
dprintk("svcrdma: dropping RDMA_DONE/ERROR message\n");
return 0;
out_proc:
dprintk("svcrdma: bad rdma procedure (%u)\n",
be32_to_cpup(rdma_argp + 3));
return -EINVAL;
out_inval:
dprintk("svcrdma: failed to parse transport header\n");
return -EINVAL;
}
static void rdma_read_complete(struct svc_rqst *rqstp,
struct svc_rdma_op_ctxt *head)
{
int page_no;
/* Copy RPC pages */
for (page_no = 0; page_no < head->count; page_no++) {
put_page(rqstp->rq_pages[page_no]);
rqstp->rq_pages[page_no] = head->pages[page_no];
}
/* Point rq_arg.pages past header */
rqstp->rq_arg.pages = &rqstp->rq_pages[head->hdr_count];
rqstp->rq_arg.page_len = head->arg.page_len;
/* rq_respages starts after the last arg page */
rqstp->rq_respages = &rqstp->rq_pages[page_no];
rqstp->rq_next_page = rqstp->rq_respages + 1;
/* Rebuild rq_arg head and tail. */
rqstp->rq_arg.head[0] = head->arg.head[0];
rqstp->rq_arg.tail[0] = head->arg.tail[0];
rqstp->rq_arg.len = head->arg.len;
rqstp->rq_arg.buflen = head->arg.buflen;
}
static void svc_rdma_send_error(struct svcxprt_rdma *xprt,
__be32 *rdma_argp, int status)
{
struct svc_rdma_op_ctxt *ctxt;
__be32 *p, *err_msgp;
unsigned int length;
struct page *page;
int ret;
ret = svc_rdma_repost_recv(xprt, GFP_KERNEL);
if (ret)
return;
page = alloc_page(GFP_KERNEL);
if (!page)
return;
err_msgp = page_address(page);
p = err_msgp;
*p++ = *rdma_argp;
*p++ = *(rdma_argp + 1);
*p++ = xprt->sc_fc_credits;
*p++ = rdma_error;
if (status == -EPROTONOSUPPORT) {
*p++ = err_vers;
*p++ = rpcrdma_version;
*p++ = rpcrdma_version;
} else {
*p++ = err_chunk;
}
length = (unsigned long)p - (unsigned long)err_msgp;
/* Map transport header; no RPC message payload */
ctxt = svc_rdma_get_context(xprt);
ret = svc_rdma_map_reply_hdr(xprt, ctxt, err_msgp, length);
if (ret) {
dprintk("svcrdma: Error %d mapping send for protocol error\n",
ret);
return;
}
ret = svc_rdma_post_send_wr(xprt, ctxt, 1, 0);
if (ret) {
dprintk("svcrdma: Error %d posting send for protocol error\n",
ret);
svc_rdma_unmap_dma(ctxt);
svc_rdma_put_context(ctxt, 1);
}
}
/* By convention, backchannel calls arrive via rdma_msg type
* messages, and never populate the chunk lists. This makes
* the RPC/RDMA header small and fixed in size, so it is
* straightforward to check the RPC header's direction field.
*/
static bool svc_rdma_is_backchannel_reply(struct svc_xprt *xprt,
__be32 *rdma_resp)
{
__be32 *p;
if (!xprt->xpt_bc_xprt)
return false;
p = rdma_resp + 3;
if (*p++ != rdma_msg)
return false;
if (*p++ != xdr_zero)
return false;
if (*p++ != xdr_zero)
return false;
if (*p++ != xdr_zero)
return false;
/* XID sanity */
if (*p++ != *rdma_resp)
return false;
/* call direction */
if (*p == cpu_to_be32(RPC_CALL))
return false;
return true;
}
/**
* svc_rdma_recvfrom - Receive an RPC call
* @rqstp: request structure into which to receive an RPC Call
*
* Returns:
* The positive number of bytes in the RPC Call message,
* %0 if there were no Calls ready to return,
* %-EINVAL if the Read chunk data is too large,
* %-ENOMEM if rdma_rw context pool was exhausted,
* %-ENOTCONN if posting failed (connection is lost),
* %-EIO if rdma_rw initialization failed (DMA mapping, etc).
*
* Called in a loop when XPT_DATA is set. XPT_DATA is cleared only
* when there are no remaining ctxt's to process.
*
* The next ctxt is removed from the "receive" lists.
*
* - If the ctxt completes a Read, then finish assembling the Call
* message and return the number of bytes in the message.
*
* - If the ctxt completes a Receive, then construct the Call
* message from the contents of the Receive buffer.
*
* - If there are no Read chunks in this message, then finish
* assembling the Call message and return the number of bytes
* in the message.
*
* - If there are Read chunks in this message, post Read WRs to
* pull that payload and return 0.
*/
int svc_rdma_recvfrom(struct svc_rqst *rqstp)
{
struct svc_xprt *xprt = rqstp->rq_xprt;
struct svcxprt_rdma *rdma_xprt =
container_of(xprt, struct svcxprt_rdma, sc_xprt);
struct svc_rdma_op_ctxt *ctxt;
__be32 *p;
int ret;
spin_lock(&rdma_xprt->sc_rq_dto_lock);
if (!list_empty(&rdma_xprt->sc_read_complete_q)) {
ctxt = list_first_entry(&rdma_xprt->sc_read_complete_q,
struct svc_rdma_op_ctxt, list);
list_del(&ctxt->list);
spin_unlock(&rdma_xprt->sc_rq_dto_lock);
rdma_read_complete(rqstp, ctxt);
goto complete;
} else if (!list_empty(&rdma_xprt->sc_rq_dto_q)) {
ctxt = list_first_entry(&rdma_xprt->sc_rq_dto_q,
struct svc_rdma_op_ctxt, list);
list_del(&ctxt->list);
} else {
/* No new incoming requests, terminate the loop */
clear_bit(XPT_DATA, &xprt->xpt_flags);
spin_unlock(&rdma_xprt->sc_rq_dto_lock);
return 0;
}
spin_unlock(&rdma_xprt->sc_rq_dto_lock);
dprintk("svcrdma: recvfrom: ctxt=%p on xprt=%p, rqstp=%p\n",
ctxt, rdma_xprt, rqstp);
atomic_inc(&rdma_stat_recv);
/* Build up the XDR from the receive buffers. */
rdma_build_arg_xdr(rqstp, ctxt, ctxt->byte_len);
/* Decode the RDMA header. */
p = (__be32 *)rqstp->rq_arg.head[0].iov_base;
ret = svc_rdma_xdr_decode_req(&rqstp->rq_arg);
if (ret < 0)
goto out_err;
if (ret == 0)
goto out_drop;
rqstp->rq_xprt_hlen = ret;
if (svc_rdma_is_backchannel_reply(xprt, p)) {
ret = svc_rdma_handle_bc_reply(xprt->xpt_bc_xprt, p,
&rqstp->rq_arg);
svc_rdma_put_context(ctxt, 0);
if (ret)
goto repost;
return ret;
}
p += rpcrdma_fixed_maxsz;
if (*p != xdr_zero)
goto out_readchunk;
complete:
svc_rdma_put_context(ctxt, 0);
dprintk("svcrdma: recvfrom: xprt=%p, rqstp=%p, rq_arg.len=%u\n",
rdma_xprt, rqstp, rqstp->rq_arg.len);
rqstp->rq_prot = IPPROTO_MAX;
svc_xprt_copy_addrs(rqstp, xprt);
return rqstp->rq_arg.len;
out_readchunk:
ret = svc_rdma_recv_read_chunk(rdma_xprt, rqstp, ctxt, p);
if (ret < 0)
goto out_postfail;
return 0;
out_err:
svc_rdma_send_error(rdma_xprt, p, ret);
svc_rdma_put_context(ctxt, 0);
return 0;
out_postfail:
if (ret == -EINVAL)
svc_rdma_send_error(rdma_xprt, p, ret);
svc_rdma_put_context(ctxt, 1);
return ret;
out_drop:
svc_rdma_put_context(ctxt, 1);
repost:
return svc_rdma_repost_recv(rdma_xprt, GFP_KERNEL);
}
