aboutsummaryrefslogtreecommitdiff
path: root/net/sunrpc
diff options
context:
space:
mode:
Diffstat (limited to 'net/sunrpc')
-rw-r--r--net/sunrpc/Makefile3
-rw-r--r--net/sunrpc/auth_gss/auth_gss.c6
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_crypto.c350
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_keys.c12
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_mech.c89
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_seqnum.c22
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_wrap.c24
-rw-r--r--net/sunrpc/clnt.c328
-rw-r--r--net/sunrpc/rpcb_clnt.c10
-rw-r--r--net/sunrpc/sched.c2
-rw-r--r--net/sunrpc/xprt.c42
-rw-r--r--net/sunrpc/xprtmultipath.c475
-rw-r--r--net/sunrpc/xprtrdma/backchannel.c2
-rw-r--r--net/sunrpc/xprtrdma/fmr_ops.c28
-rw-r--r--net/sunrpc/xprtrdma/frwr_ops.c143
-rw-r--r--net/sunrpc/xprtrdma/physical_ops.c1
-rw-r--r--net/sunrpc/xprtrdma/rpc_rdma.c108
-rw-r--r--net/sunrpc/xprtrdma/verbs.c204
-rw-r--r--net/sunrpc/xprtrdma/xprt_rdma.h14
-rw-r--r--net/sunrpc/xprtsock.c4
20 files changed, 1313 insertions, 554 deletions
diff --git a/net/sunrpc/Makefile b/net/sunrpc/Makefile
index b512fbd9d79a..ea7ffa12e0f9 100644
--- a/net/sunrpc/Makefile
+++ b/net/sunrpc/Makefile
@@ -12,7 +12,8 @@ sunrpc-y := clnt.o xprt.o socklib.o xprtsock.o sched.o \
svc.o svcsock.o svcauth.o svcauth_unix.o \
addr.o rpcb_clnt.o timer.o xdr.o \
sunrpc_syms.o cache.o rpc_pipe.o \
- svc_xprt.o
+ svc_xprt.o \
+ xprtmultipath.o
sunrpc-$(CONFIG_SUNRPC_DEBUG) += debugfs.o
sunrpc-$(CONFIG_SUNRPC_BACKCHANNEL) += backchannel_rqst.o
sunrpc-$(CONFIG_PROC_FS) += stats.o
diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c
index 799e65b944b9..8c6bc795f060 100644
--- a/net/sunrpc/auth_gss/auth_gss.c
+++ b/net/sunrpc/auth_gss/auth_gss.c
@@ -740,7 +740,7 @@ gss_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
default:
printk(KERN_CRIT "%s: bad return from "
"gss_fill_context: %zd\n", __func__, err);
- BUG();
+ gss_msg->msg.errno = -EIO;
}
goto err_release_msg;
}
@@ -1181,12 +1181,12 @@ static struct rpc_auth *
gss_create(struct rpc_auth_create_args *args, struct rpc_clnt *clnt)
{
struct gss_auth *gss_auth;
- struct rpc_xprt *xprt = rcu_access_pointer(clnt->cl_xprt);
+ struct rpc_xprt_switch *xps = rcu_access_pointer(clnt->cl_xpi.xpi_xpswitch);
while (clnt != clnt->cl_parent) {
struct rpc_clnt *parent = clnt->cl_parent;
/* Find the original parent for this transport */
- if (rcu_access_pointer(parent->cl_xprt) != xprt)
+ if (rcu_access_pointer(parent->cl_xpi.xpi_xpswitch) != xps)
break;
clnt = parent;
}
diff --git a/net/sunrpc/auth_gss/gss_krb5_crypto.c b/net/sunrpc/auth_gss/gss_krb5_crypto.c
index fee3c15a4b52..d94a8e1e9f05 100644
--- a/net/sunrpc/auth_gss/gss_krb5_crypto.c
+++ b/net/sunrpc/auth_gss/gss_krb5_crypto.c
@@ -34,11 +34,12 @@
* WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
*/
+#include <crypto/hash.h>
+#include <crypto/skcipher.h>
#include <linux/err.h>
#include <linux/types.h>
#include <linux/mm.h>
#include <linux/scatterlist.h>
-#include <linux/crypto.h>
#include <linux/highmem.h>
#include <linux/pagemap.h>
#include <linux/random.h>
@@ -51,7 +52,7 @@
u32
krb5_encrypt(
- struct crypto_blkcipher *tfm,
+ struct crypto_skcipher *tfm,
void * iv,
void * in,
void * out,
@@ -60,24 +61,28 @@ krb5_encrypt(
u32 ret = -EINVAL;
struct scatterlist sg[1];
u8 local_iv[GSS_KRB5_MAX_BLOCKSIZE] = {0};
- struct blkcipher_desc desc = { .tfm = tfm, .info = local_iv };
+ SKCIPHER_REQUEST_ON_STACK(req, tfm);
- if (length % crypto_blkcipher_blocksize(tfm) != 0)
+ if (length % crypto_skcipher_blocksize(tfm) != 0)
goto out;
- if (crypto_blkcipher_ivsize(tfm) > GSS_KRB5_MAX_BLOCKSIZE) {
+ if (crypto_skcipher_ivsize(tfm) > GSS_KRB5_MAX_BLOCKSIZE) {
dprintk("RPC: gss_k5encrypt: tfm iv size too large %d\n",
- crypto_blkcipher_ivsize(tfm));
+ crypto_skcipher_ivsize(tfm));
goto out;
}
if (iv)
- memcpy(local_iv, iv, crypto_blkcipher_ivsize(tfm));
+ memcpy(local_iv, iv, crypto_skcipher_ivsize(tfm));
memcpy(out, in, length);
sg_init_one(sg, out, length);
- ret = crypto_blkcipher_encrypt_iv(&desc, sg, sg, length);
+ skcipher_request_set_callback(req, 0, NULL, NULL);
+ skcipher_request_set_crypt(req, sg, sg, length, local_iv);
+
+ ret = crypto_skcipher_encrypt(req);
+ skcipher_request_zero(req);
out:
dprintk("RPC: krb5_encrypt returns %d\n", ret);
return ret;
@@ -85,7 +90,7 @@ out:
u32
krb5_decrypt(
- struct crypto_blkcipher *tfm,
+ struct crypto_skcipher *tfm,
void * iv,
void * in,
void * out,
@@ -94,23 +99,27 @@ krb5_decrypt(
u32 ret = -EINVAL;
struct scatterlist sg[1];
u8 local_iv[GSS_KRB5_MAX_BLOCKSIZE] = {0};
- struct blkcipher_desc desc = { .tfm = tfm, .info = local_iv };
+ SKCIPHER_REQUEST_ON_STACK(req, tfm);
- if (length % crypto_blkcipher_blocksize(tfm) != 0)
+ if (length % crypto_skcipher_blocksize(tfm) != 0)
goto out;
- if (crypto_blkcipher_ivsize(tfm) > GSS_KRB5_MAX_BLOCKSIZE) {
+ if (crypto_skcipher_ivsize(tfm) > GSS_KRB5_MAX_BLOCKSIZE) {
dprintk("RPC: gss_k5decrypt: tfm iv size too large %d\n",
- crypto_blkcipher_ivsize(tfm));
+ crypto_skcipher_ivsize(tfm));
goto out;
}
if (iv)
- memcpy(local_iv,iv, crypto_blkcipher_ivsize(tfm));
+ memcpy(local_iv,iv, crypto_skcipher_ivsize(tfm));
memcpy(out, in, length);
sg_init_one(sg, out, length);
- ret = crypto_blkcipher_decrypt_iv(&desc, sg, sg, length);
+ skcipher_request_set_callback(req, 0, NULL, NULL);
+ skcipher_request_set_crypt(req, sg, sg, length, local_iv);
+
+ ret = crypto_skcipher_decrypt(req);
+ skcipher_request_zero(req);
out:
dprintk("RPC: gss_k5decrypt returns %d\n",ret);
return ret;
@@ -119,9 +128,11 @@ out:
static int
checksummer(struct scatterlist *sg, void *data)
{
- struct hash_desc *desc = data;
+ struct ahash_request *req = data;
+
+ ahash_request_set_crypt(req, sg, NULL, sg->length);
- return crypto_hash_update(desc, sg, sg->length);
+ return crypto_ahash_update(req);
}
static int
@@ -152,13 +163,13 @@ make_checksum_hmac_md5(struct krb5_ctx *kctx, char *header, int hdrlen,
struct xdr_buf *body, int body_offset, u8 *cksumkey,
unsigned int usage, struct xdr_netobj *cksumout)
{
- struct hash_desc desc;
struct scatterlist sg[1];
int err;
u8 checksumdata[GSS_KRB5_MAX_CKSUM_LEN];
u8 rc4salt[4];
- struct crypto_hash *md5;
- struct crypto_hash *hmac_md5;
+ struct crypto_ahash *md5;
+ struct crypto_ahash *hmac_md5;
+ struct ahash_request *req;
if (cksumkey == NULL)
return GSS_S_FAILURE;
@@ -174,61 +185,79 @@ make_checksum_hmac_md5(struct krb5_ctx *kctx, char *header, int hdrlen,
return GSS_S_FAILURE;
}
- md5 = crypto_alloc_hash("md5", 0, CRYPTO_ALG_ASYNC);
+ md5 = crypto_alloc_ahash("md5", 0, CRYPTO_ALG_ASYNC);
if (IS_ERR(md5))
return GSS_S_FAILURE;
- hmac_md5 = crypto_alloc_hash(kctx->gk5e->cksum_name, 0,
- CRYPTO_ALG_ASYNC);
+ hmac_md5 = crypto_alloc_ahash(kctx->gk5e->cksum_name, 0,
+ CRYPTO_ALG_ASYNC);
if (IS_ERR(hmac_md5)) {
- crypto_free_hash(md5);
+ crypto_free_ahash(md5);
+ return GSS_S_FAILURE;
+ }
+
+ req = ahash_request_alloc(md5, GFP_KERNEL);
+ if (!req) {
+ crypto_free_ahash(hmac_md5);
+ crypto_free_ahash(md5);
return GSS_S_FAILURE;
}
- desc.tfm = md5;
- desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
+ ahash_request_set_callback(req, CRYPTO_TFM_REQ_MAY_SLEEP, NULL, NULL);
- err = crypto_hash_init(&desc);
+ err = crypto_ahash_init(req);
if (err)
goto out;
sg_init_one(sg, rc4salt, 4);
- err = crypto_hash_update(&desc, sg, 4);
+ ahash_request_set_crypt(req, sg, NULL, 4);
+ err = crypto_ahash_update(req);
if (err)
goto out;
sg_init_one(sg, header, hdrlen);
- err = crypto_hash_update(&desc, sg, hdrlen);
+ ahash_request_set_crypt(req, sg, NULL, hdrlen);
+ err = crypto_ahash_update(req);
if (err)
goto out;
err = xdr_process_buf(body, body_offset, body->len - body_offset,
- checksummer, &desc);
+ checksummer, req);
if (err)
goto out;
- err = crypto_hash_final(&desc, checksumdata);
+ ahash_request_set_crypt(req, NULL, checksumdata, 0);
+ err = crypto_ahash_final(req);
if (err)
goto out;
- desc.tfm = hmac_md5;
- desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
+ ahash_request_free(req);
+ req = ahash_request_alloc(hmac_md5, GFP_KERNEL);
+ if (!req) {
+ crypto_free_ahash(hmac_md5);
+ crypto_free_ahash(md5);
+ return GSS_S_FAILURE;
+ }
+
+ ahash_request_set_callback(req, CRYPTO_TFM_REQ_MAY_SLEEP, NULL, NULL);
- err = crypto_hash_init(&desc);
+ err = crypto_ahash_init(req);
if (err)
goto out;
- err = crypto_hash_setkey(hmac_md5, cksumkey, kctx->gk5e->keylength);
+ err = crypto_ahash_setkey(hmac_md5, cksumkey, kctx->gk5e->keylength);
if (err)
goto out;
- sg_init_one(sg, checksumdata, crypto_hash_digestsize(md5));
- err = crypto_hash_digest(&desc, sg, crypto_hash_digestsize(md5),
- checksumdata);
+ sg_init_one(sg, checksumdata, crypto_ahash_digestsize(md5));
+ ahash_request_set_crypt(req, sg, checksumdata,
+ crypto_ahash_digestsize(md5));
+ err = crypto_ahash_digest(req);
if (err)
goto out;
memcpy(cksumout->data, checksumdata, kctx->gk5e->cksumlength);
cksumout->len = kctx->gk5e->cksumlength;
out:
- crypto_free_hash(md5);
- crypto_free_hash(hmac_md5);
+ ahash_request_free(req);
+ crypto_free_ahash(md5);
+ crypto_free_ahash(hmac_md5);
return err ? GSS_S_FAILURE : 0;
}
@@ -242,7 +271,8 @@ make_checksum(struct krb5_ctx *kctx, char *header, int hdrlen,
struct xdr_buf *body, int body_offset, u8 *cksumkey,
unsigned int usage, struct xdr_netobj *cksumout)
{
- struct hash_desc desc;
+ struct crypto_ahash *tfm;
+ struct ahash_request *req;
struct scatterlist sg[1];
int err;
u8 checksumdata[GSS_KRB5_MAX_CKSUM_LEN];
@@ -259,32 +289,41 @@ make_checksum(struct krb5_ctx *kctx, char *header, int hdrlen,
return GSS_S_FAILURE;
}
- desc.tfm = crypto_alloc_hash(kctx->gk5e->cksum_name, 0, CRYPTO_ALG_ASYNC);
- if (IS_ERR(desc.tfm))
+ tfm = crypto_alloc_ahash(kctx->gk5e->cksum_name, 0, CRYPTO_ALG_ASYNC);
+ if (IS_ERR(tfm))
return GSS_S_FAILURE;
- desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
- checksumlen = crypto_hash_digestsize(desc.tfm);
+ req = ahash_request_alloc(tfm, GFP_KERNEL);
+ if (!req) {
+ crypto_free_ahash(tfm);
+ return GSS_S_FAILURE;
+ }
+
+ ahash_request_set_callback(req, CRYPTO_TFM_REQ_MAY_SLEEP, NULL, NULL);
+
+ checksumlen = crypto_ahash_digestsize(tfm);
if (cksumkey != NULL) {
- err = crypto_hash_setkey(desc.tfm, cksumkey,
- kctx->gk5e->keylength);
+ err = crypto_ahash_setkey(tfm, cksumkey,
+ kctx->gk5e->keylength);
if (err)
goto out;
}
- err = crypto_hash_init(&desc);
+ err = crypto_ahash_init(req);
if (err)
goto out;
sg_init_one(sg, header, hdrlen);
- err = crypto_hash_update(&desc, sg, hdrlen);
+ ahash_request_set_crypt(req, sg, NULL, hdrlen);
+ err = crypto_ahash_update(req);
if (err)
goto out;
err = xdr_process_buf(body, body_offset, body->len - body_offset,
- checksummer, &desc);
+ checksummer, req);
if (err)
goto out;
- err = crypto_hash_final(&desc, checksumdata);
+ ahash_request_set_crypt(req, NULL, checksumdata, 0);
+ err = crypto_ahash_final(req);
if (err)
goto out;
@@ -307,7 +346,8 @@ make_checksum(struct krb5_ctx *kctx, char *header, int hdrlen,
}
cksumout->len = kctx->gk5e->cksumlength;
out:
- crypto_free_hash(desc.tfm);
+ ahash_request_free(req);
+ crypto_free_ahash(tfm);
return err ? GSS_S_FAILURE : 0;
}
@@ -323,7 +363,8 @@ make_checksum_v2(struct krb5_ctx *kctx, char *header, int hdrlen,
struct xdr_buf *body, int body_offset, u8 *cksumkey,
unsigned int usage, struct xdr_netobj *cksumout)
{
- struct hash_desc desc;
+ struct crypto_ahash *tfm;
+ struct ahash_request *req;
struct scatterlist sg[1];
int err;
u8 checksumdata[GSS_KRB5_MAX_CKSUM_LEN];
@@ -340,31 +381,39 @@ make_checksum_v2(struct krb5_ctx *kctx, char *header, int hdrlen,
return GSS_S_FAILURE;
}
- desc.tfm = crypto_alloc_hash(kctx->gk5e->cksum_name, 0,
- CRYPTO_ALG_ASYNC);
- if (IS_ERR(desc.tfm))
+ tfm = crypto_alloc_ahash(kctx->gk5e->cksum_name, 0, CRYPTO_ALG_ASYNC);
+ if (IS_ERR(tfm))
return GSS_S_FAILURE;
- checksumlen = crypto_hash_digestsize(desc.tfm);
- desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
+ checksumlen = crypto_ahash_digestsize(tfm);
+
+ req = ahash_request_alloc(tfm, GFP_KERNEL);
+ if (!req) {
+ crypto_free_ahash(tfm);
+ return GSS_S_FAILURE;
+ }
- err = crypto_hash_setkey(desc.tfm, cksumkey, kctx->gk5e->keylength);
+ ahash_request_set_callback(req, CRYPTO_TFM_REQ_MAY_SLEEP, NULL, NULL);
+
+ err = crypto_ahash_setkey(tfm, cksumkey, kctx->gk5e->keylength);
if (err)
goto out;
- err = crypto_hash_init(&desc);
+ err = crypto_ahash_init(req);
if (err)
goto out;
err = xdr_process_buf(body, body_offset, body->len - body_offset,
- checksummer, &desc);
+ checksummer, req);
if (err)
goto out;
if (header != NULL) {
sg_init_one(sg, header, hdrlen);
- err = crypto_hash_update(&desc, sg, hdrlen);
+ ahash_request_set_crypt(req, sg, NULL, hdrlen);
+ err = crypto_ahash_update(req);
if (err)
goto out;
}
- err = crypto_hash_final(&desc, checksumdata);
+ ahash_request_set_crypt(req, NULL, checksumdata, 0);
+ err = crypto_ahash_final(req);
if (err)
goto out;
@@ -381,13 +430,14 @@ make_checksum_v2(struct krb5_ctx *kctx, char *header, int hdrlen,
break;
}
out:
- crypto_free_hash(desc.tfm);
+ ahash_request_free(req);
+ crypto_free_ahash(tfm);
return err ? GSS_S_FAILURE : 0;
}
struct encryptor_desc {
u8 iv[GSS_KRB5_MAX_BLOCKSIZE];
- struct blkcipher_desc desc;
+ struct skcipher_request *req;
int pos;
struct xdr_buf *outbuf;
struct page **pages;
@@ -402,6 +452,7 @@ encryptor(struct scatterlist *sg, void *data)
{
struct encryptor_desc *desc = data;
struct xdr_buf *outbuf = desc->outbuf;
+ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(desc->req);
struct page *in_page;
int thislen = desc->fraglen + sg->length;
int fraglen, ret;
@@ -427,7 +478,7 @@ encryptor(struct scatterlist *sg, void *data)
desc->fraglen += sg->length;
desc->pos += sg->length;
- fraglen = thislen & (crypto_blkcipher_blocksize(desc->desc.tfm) - 1);
+ fraglen = thislen & (crypto_skcipher_blocksize(tfm) - 1);
thislen -= fraglen;
if (thislen == 0)
@@ -436,8 +487,10 @@ encryptor(struct scatterlist *sg, void *data)
sg_mark_end(&desc->infrags[desc->fragno - 1]);
sg_mark_end(&desc->outfrags[desc->fragno - 1]);
- ret = crypto_blkcipher_encrypt_iv(&desc->desc, desc->outfrags,
- desc->infrags, thislen);
+ skcipher_request_set_crypt(desc->req, desc->infrags, desc->outfrags,
+ thislen, desc->iv);
+
+ ret = crypto_skcipher_encrypt(desc->req);
if (ret)
return ret;
@@ -459,18 +512,20 @@ encryptor(struct scatterlist *sg, void *data)
}
int
-gss_encrypt_xdr_buf(struct crypto_blkcipher *tfm, struct xdr_buf *buf,
+gss_encrypt_xdr_buf(struct crypto_skcipher *tfm, struct xdr_buf *buf,
int offset, struct page **pages)
{
int ret;
struct encryptor_desc desc;
+ SKCIPHER_REQUEST_ON_STACK(req, tfm);
+
+ BUG_ON((buf->len - offset) % crypto_skcipher_blocksize(tfm) != 0);
- BUG_ON((buf->len - offset) % crypto_blkcipher_blocksize(tfm) != 0);
+ skcipher_request_set_tfm(req, tfm);
+ skcipher_request_set_callback(req, 0, NULL, NULL);
memset(desc.iv, 0, sizeof(desc.iv));
- desc.desc.tfm = tfm;
- desc.desc.info = desc.iv;
- desc.desc.flags = 0;
+ desc.req = req;
desc.pos = offset;
desc.outbuf = buf;
desc.pages = pages;
@@ -481,12 +536,13 @@ gss_encrypt_xdr_buf(struct crypto_blkcipher *tfm, struct xdr_buf *buf,
sg_init_table(desc.outfrags, 4);
ret = xdr_process_buf(buf, offset, buf->len - offset, encryptor, &desc);
+ skcipher_request_zero(req);
return ret;
}
struct decryptor_desc {
u8 iv[GSS_KRB5_MAX_BLOCKSIZE];
- struct blkcipher_desc desc;
+ struct skcipher_request *req;
struct scatterlist frags[4];
int fragno;
int fraglen;
@@ -497,6 +553,7 @@ decryptor(struct scatterlist *sg, void *data)
{
struct decryptor_desc *desc = data;
int thislen = desc->fraglen + sg->length;
+ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(desc->req);
int fraglen, ret;
/* Worst case is 4 fragments: head, end of page 1, start
@@ -507,7 +564,7 @@ decryptor(struct scatterlist *sg, void *data)
desc->fragno++;
desc->fraglen += sg->length;
- fraglen = thislen & (crypto_blkcipher_blocksize(desc->desc.tfm) - 1);
+ fraglen = thislen & (crypto_skcipher_blocksize(tfm) - 1);
thislen -= fraglen;
if (thislen == 0)
@@ -515,8 +572,10 @@ decryptor(struct scatterlist *sg, void *data)
sg_mark_end(&desc->frags[desc->fragno - 1]);
- ret = crypto_blkcipher_decrypt_iv(&desc->desc, desc->frags,
- desc->frags, thislen);
+ skcipher_request_set_crypt(desc->req, desc->frags, desc->frags,
+ thislen, desc->iv);
+
+ ret = crypto_skcipher_decrypt(desc->req);
if (ret)
return ret;
@@ -535,24 +594,29 @@ decryptor(struct scatterlist *sg, void *data)
}
int
-gss_decrypt_xdr_buf(struct crypto_blkcipher *tfm, struct xdr_buf *buf,
+gss_decrypt_xdr_buf(struct crypto_skcipher *tfm, struct xdr_buf *buf,
int offset)
{
+ int ret;
struct decryptor_desc desc;
+ SKCIPHER_REQUEST_ON_STACK(req, tfm);
/* XXXJBF: */
- BUG_ON((buf->len - offset) % crypto_blkcipher_blocksize(tfm) != 0);
+ BUG_ON((buf->len - offset) % crypto_skcipher_blocksize(tfm) != 0);
+
+ skcipher_request_set_tfm(req, tfm);
+ skcipher_request_set_callback(req, 0, NULL, NULL);
memset(desc.iv, 0, sizeof(desc.iv));
- desc.desc.tfm = tfm;
- desc.desc.info = desc.iv;
- desc.desc.flags = 0;
+ desc.req = req;
desc.fragno = 0;
desc.fraglen = 0;
sg_init_table(desc.frags, 4);
- return xdr_process_buf(buf, offset, buf->len - offset, decryptor, &desc);
+ ret = xdr_process_buf(buf, offset, buf->len - offset, decryptor, &desc);
+ skcipher_request_zero(req);
+ return ret;
}
/*
@@ -594,12 +658,12 @@ xdr_extend_head(struct xdr_buf *buf, unsigned int base, unsigned int shiftlen)
}
static u32
-gss_krb5_cts_crypt(struct crypto_blkcipher *cipher, struct xdr_buf *buf,
+gss_krb5_cts_crypt(struct crypto_skcipher *cipher, struct xdr_buf *buf,
u32 offset, u8 *iv, struct page **pages, int encrypt)
{
u32 ret;
struct scatterlist sg[1];
- struct blkcipher_desc desc = { .tfm = cipher, .info = iv };
+ SKCIPHER_REQUEST_ON_STACK(req, cipher);
u8 data[GSS_KRB5_MAX_BLOCKSIZE * 2];
struct page **save_pages;
u32 len = buf->len - offset;
@@ -625,10 +689,16 @@ gss_krb5_cts_crypt(struct crypto_blkcipher *cipher, struct xdr_buf *buf,
sg_init_one(sg, data, len);
+ skcipher_request_set_tfm(req, cipher);
+ skcipher_request_set_callback(req, 0, NULL, NULL);
+ skcipher_request_set_crypt(req, sg, sg, len, iv);
+
if (encrypt)
- ret = crypto_blkcipher_encrypt_iv(&desc, sg, sg, len);
+ ret = crypto_skcipher_encrypt(req);
else
- ret = crypto_blkcipher_decrypt_iv(&desc, sg, sg, len);
+ ret = crypto_skcipher_decrypt(req);
+
+ skcipher_request_zero(req);
if (ret)
goto out;
@@ -647,7 +717,7 @@ gss_krb5_aes_encrypt(struct krb5_ctx *kctx, u32 offset,
struct xdr_netobj hmac;
u8 *cksumkey;
u8 *ecptr;
- struct crypto_blkcipher *cipher, *aux_cipher;
+ struct crypto_skcipher *cipher, *aux_cipher;
int blocksize;
struct page **save_pages;
int nblocks, nbytes;
@@ -666,7 +736,7 @@ gss_krb5_aes_encrypt(struct krb5_ctx *kctx, u32 offset,
cksumkey = kctx->acceptor_integ;
usage = KG_USAGE_ACCEPTOR_SEAL;
}
- blocksize = crypto_blkcipher_blocksize(cipher);
+ blocksize = crypto_skcipher_blocksize(cipher);
/* hide the gss token header and insert the confounder */
offset += GSS_KRB5_TOK_HDR_LEN;
@@ -719,20 +789,24 @@ gss_krb5_aes_encrypt(struct krb5_ctx *kctx, u32 offset,
memset(desc.iv, 0, sizeof(desc.iv));
if (cbcbytes) {
+ SKCIPHER_REQUEST_ON_STACK(req, aux_cipher);
+
desc.pos = offset + GSS_KRB5_TOK_HDR_LEN;
desc.fragno = 0;
desc.fraglen = 0;
desc.pages = pages;
desc.outbuf = buf;
- desc.desc.info = desc.iv;
- desc.desc.flags = 0;
- desc.desc.tfm = aux_cipher;
+ desc.req = req;
+
+ skcipher_request_set_tfm(req, aux_cipher);
+ skcipher_request_set_callback(req, 0, NULL, NULL);
sg_init_table(desc.infrags, 4);
sg_init_table(desc.outfrags, 4);
err = xdr_process_buf(buf, offset + GSS_KRB5_TOK_HDR_LEN,
cbcbytes, encryptor, &desc);
+ skcipher_request_zero(req);
if (err)
goto out_err;
}
@@ -763,7 +837,7 @@ gss_krb5_aes_decrypt(struct krb5_ctx *kctx, u32 offset, struct xdr_buf *buf,
struct xdr_buf subbuf;
u32 ret = 0;
u8 *cksum_key;
- struct crypto_blkcipher *cipher, *aux_cipher;
+ struct crypto_skcipher *cipher, *aux_cipher;
struct xdr_netobj our_hmac_obj;
u8 our_hmac[GSS_KRB5_MAX_CKSUM_LEN];
u8 pkt_hmac[GSS_KRB5_MAX_CKSUM_LEN];
@@ -782,7 +856,7 @@ gss_krb5_aes_decrypt(struct krb5_ctx *kctx, u32 offset, struct xdr_buf *buf,
cksum_key = kctx->initiator_integ;
usage = KG_USAGE_INITIATOR_SEAL;
}
- blocksize = crypto_blkcipher_blocksize(cipher);
+ blocksize = crypto_skcipher_blocksize(cipher);
/* create a segment skipping the header and leaving out the checksum */
@@ -799,15 +873,19 @@ gss_krb5_aes_decrypt(struct krb5_ctx *kctx, u32 offset, struct xdr_buf *buf,
memset(desc.iv, 0, sizeof(desc.iv));
if (cbcbytes) {
+ SKCIPHER_REQUEST_ON_STACK(req, aux_cipher);
+
desc.fragno = 0;
desc.fraglen = 0;
- desc.desc.info = desc.iv;
- desc.desc.flags = 0;
- desc.desc.tfm = aux_cipher;
+ desc.req = req;
+
+ skcipher_request_set_tfm(req, aux_cipher);
+ skcipher_request_set_callback(req, 0, NULL, NULL);
sg_init_table(desc.frags, 4);
ret = xdr_process_buf(&subbuf, 0, cbcbytes, decryptor, &desc);
+ skcipher_request_zero(req);
if (ret)
goto out_err;
}
@@ -850,61 +928,62 @@ out_err:
* Set the key of the given cipher.
*/
int
-krb5_rc4_setup_seq_key(struct krb5_ctx *kctx, struct crypto_blkcipher *cipher,
+krb5_rc4_setup_seq_key(struct krb5_ctx *kctx, struct crypto_skcipher *cipher,
unsigned char *cksum)
{
- struct crypto_hash *hmac;
- struct hash_desc desc;
- struct scatterlist sg[1];
+ struct crypto_shash *hmac;
+ struct shash_desc *desc;
u8 Kseq[GSS_KRB5_MAX_KEYLEN];
u32 zeroconstant = 0;
int err;
dprintk("%s: entered\n", __func__);
- hmac = crypto_alloc_hash(kctx->gk5e->cksum_name, 0, CRYPTO_ALG_ASYNC);
+ hmac = crypto_alloc_shash(kctx->gk5e->cksum_name, 0, 0);
if (IS_ERR(hmac)) {
dprintk("%s: error %ld, allocating hash '%s'\n",
__func__, PTR_ERR(hmac), kctx->gk5e->cksum_name);
return PTR_ERR(hmac);
}
- desc.tfm = hmac;
- desc.flags = 0;
+ desc = kmalloc(sizeof(*desc), GFP_KERNEL);
+ if (!desc) {
+ dprintk("%s: failed to allocate shash descriptor for '%s'\n",
+ __func__, kctx->gk5e->cksum_name);
+ crypto_free_shash(hmac);
+ return -ENOMEM;
+ }
- err = crypto_hash_init(&desc);
- if (err)
- goto out_err;
+ desc->tfm = hmac;
+ desc->flags = 0;
/* Compute intermediate Kseq from session key */
- err = crypto_hash_setkey(hmac, kctx->Ksess, kctx->gk5e->keylength);
+ err = crypto_shash_setkey(hmac, kctx->Ksess, kctx->gk5e->keylength);
if (err)
goto out_err;
- sg_init_one(sg, &zeroconstant, 4);
- err = crypto_hash_digest(&desc, sg, 4, Kseq);
+ err = crypto_shash_digest(desc, (u8 *)&zeroconstant, 4, Kseq);
if (err)
goto out_err;
/* Compute final Kseq from the checksum and intermediate Kseq */
- err = crypto_hash_setkey(hmac, Kseq, kctx->gk5e->keylength);
+ err = crypto_shash_setkey(hmac, Kseq, kctx->gk5e->keylength);
if (err)
goto out_err;
- sg_set_buf(sg, cksum, 8);
-
- err = crypto_hash_digest(&desc, sg, 8, Kseq);
+ err = crypto_shash_digest(desc, cksum, 8, Kseq);
if (err)
goto out_err;
- err = crypto_blkcipher_setkey(cipher, Kseq, kctx->gk5e->keylength);
+ err = crypto_skcipher_setkey(cipher, Kseq, kctx->gk5e->keylength);
if (err)
goto out_err;
err = 0;
out_err:
- crypto_free_hash(hmac);
+ kzfree(desc);
+ crypto_free_shash(hmac);
dprintk("%s: returning %d\n", __func__, err);
return err;
}
@@ -914,12 +993,11 @@ out_err:
* Set the key of cipher kctx->enc.
*/
int
-krb5_rc4_setup_enc_key(struct krb5_ctx *kctx, struct crypto_blkcipher *cipher,
+krb5_rc4_setup_enc_key(struct krb5_ctx *kctx, struct crypto_skcipher *cipher,
s32 seqnum)
{
- struct crypto_hash *hmac;
- struct hash_desc desc;
- struct scatterlist sg[1];
+ struct crypto_shash *hmac;
+ struct shash_desc *desc;
u8 Kcrypt[GSS_KRB5_MAX_KEYLEN];
u8 zeroconstant[4] = {0};
u8 seqnumarray[4];
@@ -927,35 +1005,38 @@ krb5_rc4_setup_enc_key(struct krb5_ctx *kctx, struct crypto_blkcipher *cipher,
dprintk("%s: entered, seqnum %u\n", __func__, seqnum);
- hmac = crypto_alloc_hash(kctx->gk5e->cksum_name, 0, CRYPTO_ALG_ASYNC);
+ hmac = crypto_alloc_shash(kctx->gk5e->cksum_name, 0, 0);
if (IS_ERR(hmac)) {
dprintk("%s: error %ld, allocating hash '%s'\n",
__func__, PTR_ERR(hmac), kctx->gk5e->cksum_name);
return PTR_ERR(hmac);
}
- desc.tfm = hmac;
- desc.flags = 0;
+ desc = kmalloc(sizeof(*desc), GFP_KERNEL);
+ if (!desc) {
+ dprintk("%s: failed to allocate shash descriptor for '%s'\n",
+ __func__, kctx->gk5e->cksum_name);
+ crypto_free_shash(hmac);
+ return -ENOMEM;
+ }
- err = crypto_hash_init(&desc);
- if (err)
- goto out_err;
+ desc->tfm = hmac;
+ desc->flags = 0;
/* Compute intermediate Kcrypt from session key */
for (i = 0; i < kctx->gk5e->keylength; i++)
Kcrypt[i] = kctx->Ksess[i] ^ 0xf0;
- err = crypto_hash_setkey(hmac, Kcrypt, kctx->gk5e->keylength);
+ err = crypto_shash_setkey(hmac, Kcrypt, kctx->gk5e->keylength);
if (err)
goto out_err;
- sg_init_one(sg, zeroconstant, 4);
- err = crypto_hash_digest(&desc, sg, 4, Kcrypt);
+ err = crypto_shash_digest(desc, zeroconstant, 4, Kcrypt);
if (err)
goto out_err;
/* Compute final Kcrypt from the seqnum and intermediate Kcrypt */
- err = crypto_hash_setkey(hmac, Kcrypt, kctx->gk5e->keylength);
+ err = crypto_shash_setkey(hmac, Kcrypt, kctx->gk5e->keylength);
if (err)
goto out_err;
@@ -964,20 +1045,19 @@ krb5_rc4_setup_enc_key(struct krb5_ctx *kctx, struct crypto_blkcipher *cipher,
seqnumarray[2] = (unsigned char) ((seqnum >> 8) & 0xff);
seqnumarray[3] = (unsigned char) ((seqnum >> 0) & 0xff);
- sg_set_buf(sg, seqnumarray, 4);
-
- err = crypto_hash_digest(&desc, sg, 4, Kcrypt);
+ err = crypto_shash_digest(desc, seqnumarray, 4, Kcrypt);
if (err)
goto out_err;
- err = crypto_blkcipher_setkey(cipher, Kcrypt, kctx->gk5e->keylength);
+ err = crypto_skcipher_setkey(cipher, Kcrypt, kctx->gk5e->keylength);
if (err)
goto out_err;
err = 0;
out_err:
- crypto_free_hash(hmac);
+ kzfree(desc);
+ crypto_free_shash(hmac);
dprintk("%s: returning %d\n", __func__, err);
return err;
}
diff --git a/net/sunrpc/auth_gss/gss_krb5_keys.c b/net/sunrpc/auth_gss/gss_krb5_keys.c
index 234fa8d0fd9b..870133146026 100644
--- a/net/sunrpc/auth_gss/gss_krb5_keys.c
+++ b/net/sunrpc/auth_gss/gss_krb5_keys.c
@@ -54,9 +54,9 @@
* WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
*/
+#include <crypto/skcipher.h>
#include <linux/err.h>
#include <linux/types.h>
-#include <linux/crypto.h>
#include <linux/sunrpc/gss_krb5.h>
#include <linux/sunrpc/xdr.h>
#include <linux/lcm.h>
@@ -147,7 +147,7 @@ u32 krb5_derive_key(const struct gss_krb5_enctype *gk5e,
size_t blocksize, keybytes, keylength, n;
unsigned char *inblockdata, *outblockdata, *rawkey;
struct xdr_netobj inblock, outblock;
- struct crypto_blkcipher *cipher;
+ struct crypto_skcipher *cipher;
u32 ret = EINVAL;
blocksize = gk5e->blocksize;
@@ -157,11 +157,11 @@ u32 krb5_derive_key(const struct gss_krb5_enctype *gk5e,
if ((inkey->len != keylength) || (outkey->len != keylength))
goto err_return;
- cipher = crypto_alloc_blkcipher(gk5e->encrypt_name, 0,
- CRYPTO_ALG_ASYNC);
+ cipher = crypto_alloc_skcipher(gk5e->encrypt_name, 0,
+ CRYPTO_ALG_ASYNC);
if (IS_ERR(cipher))
goto err_return;
- if (crypto_blkcipher_setkey(cipher, inkey->data, inkey->len))
+ if (crypto_skcipher_setkey(cipher, inkey->data, inkey->len))
goto err_return;
/* allocate and set up buffers */
@@ -238,7 +238,7 @@ err_free_in:
memset(inblockdata, 0, blocksize);
kfree(inblockdata);
err_free_cipher:
- crypto_free_blkcipher(cipher);
+ crypto_free_skcipher(cipher);
err_return:
return ret;
}
diff --git a/net/sunrpc/auth_gss/gss_krb5_mech.c b/net/sunrpc/auth_gss/gss_krb5_mech.c
index 28db442a0034..71341ccb9890 100644
--- a/net/sunrpc/auth_gss/gss_krb5_mech.c
+++ b/net/sunrpc/auth_gss/gss_krb5_mech.c
@@ -34,6 +34,8 @@
*
*/
+#include <crypto/hash.h>
+#include <crypto/skcipher.h>
#include <linux/err.h>
#include <linux/module.h>
#include <linux/init.h>
@@ -42,7 +44,6 @@
#include <linux/sunrpc/auth.h>
#include <linux/sunrpc/gss_krb5.h>
#include <linux/sunrpc/xdr.h>
-#include <linux/crypto.h>
#include <linux/sunrpc/gss_krb5_enctypes.h>
#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
@@ -217,7 +218,7 @@ simple_get_netobj(const void *p, const void *end, struct xdr_netobj *res)
static inline const void *
get_key(const void *p, const void *end,
- struct krb5_ctx *ctx, struct crypto_blkcipher **res)
+ struct krb5_ctx *ctx, struct crypto_skcipher **res)
{
struct xdr_netobj key;
int alg;
@@ -245,7 +246,7 @@ get_key(const void *p, const void *end,
if (IS_ERR(p))
goto out_err;
- *res = crypto_alloc_blkcipher(ctx->gk5e->encrypt_name, 0,
+ *res = crypto_alloc_skcipher(ctx->gk5e->encrypt_name, 0,
CRYPTO_ALG_ASYNC);
if (IS_ERR(*res)) {
printk(KERN_WARNING "gss_kerberos_mech: unable to initialize "
@@ -253,7 +254,7 @@ get_key(const void *p, const void *end,
*res = NULL;
goto out_err_free_key;
}
- if (crypto_blkcipher_setkey(*res, key.data, key.len)) {
+ if (crypto_skcipher_setkey(*res, key.data, key.len)) {
printk(KERN_WARNING "gss_kerberos_mech: error setting key for "
"crypto algorithm %s\n", ctx->gk5e->encrypt_name);
goto out_err_free_tfm;
@@ -263,7 +264,7 @@ get_key(const void *p, const void *end,
return p;
out_err_free_tfm:
- crypto_free_blkcipher(*res);
+ crypto_free_skcipher(*res);
out_err_free_key:
kfree(key.data);
p = ERR_PTR(-EINVAL);
@@ -335,30 +336,30 @@ gss_import_v1_context(const void *p, const void *end, struct krb5_ctx *ctx)
return 0;
out_err_free_key2:
- crypto_free_blkcipher(ctx->seq);
+ crypto_free_skcipher(ctx->seq);
out_err_free_key1:
- crypto_free_blkcipher(ctx->enc);
+ crypto_free_skcipher(ctx->enc);
out_err_free_mech:
kfree(ctx->mech_used.data);
out_err:
return PTR_ERR(p);
}
-static struct crypto_blkcipher *
+static struct crypto_skcipher *
context_v2_alloc_cipher(struct krb5_ctx *ctx, const char *cname, u8 *key)
{
- struct crypto_blkcipher *cp;
+ struct crypto_skcipher *cp;
- cp = crypto_alloc_blkcipher(cname, 0, CRYPTO_ALG_ASYNC);
+ cp = crypto_alloc_skcipher(cname, 0, CRYPTO_ALG_ASYNC);
if (IS_ERR(cp)) {
dprintk("gss_kerberos_mech: unable to initialize "
"crypto algorithm %s\n", cname);
return NULL;
}
- if (crypto_blkcipher_setkey(cp, key, ctx->gk5e->keylength)) {
+ if (crypto_skcipher_setkey(cp, key, ctx->gk5e->keylength)) {
dprintk("gss_kerberos_mech: error setting key for "
"crypto algorithm %s\n", cname);
- crypto_free_blkcipher(cp);
+ crypto_free_skcipher(cp);
return NULL;
}
return cp;
@@ -412,9 +413,9 @@ context_derive_keys_des3(struct krb5_ctx *ctx, gfp_t gfp_mask)
return 0;
out_free_enc:
- crypto_free_blkcipher(ctx->enc);
+ crypto_free_skcipher(ctx->enc);
out_free_seq:
- crypto_free_blkcipher(ctx->seq);
+ crypto_free_skcipher(ctx->seq);
out_err:
return -EINVAL;
}
@@ -427,18 +428,17 @@ out_err:
static int
context_derive_keys_rc4(struct krb5_ctx *ctx)
{
- struct crypto_hash *hmac;
+ struct crypto_shash *hmac;
char sigkeyconstant[] = "signaturekey";
int slen = strlen(sigkeyconstant) + 1; /* include null terminator */
- struct hash_desc desc;
- struct scatterlist sg[1];
+ struct shash_desc *desc;
int err;
dprintk("RPC: %s: entered\n", __func__);
/*
* derive cksum (aka Ksign) key
*/
- hmac = crypto_alloc_hash(ctx->gk5e->cksum_name, 0, CRYPTO_ALG_ASYNC);
+ hmac = crypto_alloc_shash(ctx->gk5e->cksum_name, 0, 0);
if (IS_ERR(hmac)) {
dprintk("%s: error %ld allocating hash '%s'\n",
__func__, PTR_ERR(hmac), ctx->gk5e->cksum_name);
@@ -446,37 +446,40 @@ context_derive_keys_rc4(struct krb5_ctx *ctx)
goto out_err;
}
- err = crypto_hash_setkey(hmac, ctx->Ksess, ctx->gk5e->keylength);
+ err = crypto_shash_setkey(hmac, ctx->Ksess, ctx->gk5e->keylength);
if (err)
goto out_err_free_hmac;
- sg_init_table(sg, 1);
- sg_set_buf(sg, sigkeyconstant, slen);
- desc.tfm = hmac;
- desc.flags = 0;
-
- err = crypto_hash_init(&desc);
- if (err)
+ desc = kmalloc(sizeof(*desc), GFP_KERNEL);
+ if (!desc) {
+ dprintk("%s: failed to allocate hash descriptor for '%s'\n",
+ __func__, ctx->gk5e->cksum_name);
+ err = -ENOMEM;
goto out_err_free_hmac;
+ }
+
+ desc->tfm = hmac;
+ desc->flags = 0;
- err = crypto_hash_digest(&desc, sg, slen, ctx->cksum);
+ err = crypto_shash_digest(desc, sigkeyconstant, slen, ctx->cksum);
+ kzfree(desc);
if (err)
goto out_err_free_hmac;
/*
- * allocate hash, and blkciphers for data and seqnum encryption
+ * allocate hash, and skciphers for data and seqnum encryption
*/
- ctx->enc = crypto_alloc_blkcipher(ctx->gk5e->encrypt_name, 0,
- CRYPTO_ALG_ASYNC);
+ ctx->enc = crypto_alloc_skcipher(ctx->gk5e->encrypt_name, 0,
+ CRYPTO_ALG_ASYNC);
if (IS_ERR(ctx->enc)) {
err = PTR_ERR(ctx->enc);
goto out_err_free_hmac;
}
- ctx->seq = crypto_alloc_blkcipher(ctx->gk5e->encrypt_name, 0,
- CRYPTO_ALG_ASYNC);
+ ctx->seq = crypto_alloc_skcipher(ctx->gk5e->encrypt_name, 0,
+ CRYPTO_ALG_ASYNC);
if (IS_ERR(ctx->seq)) {
- crypto_free_blkcipher(ctx->enc);
+ crypto_free_skcipher(ctx->enc);
err = PTR_ERR(ctx->seq);
goto out_err_free_hmac;
}
@@ -486,7 +489,7 @@ context_derive_keys_rc4(struct krb5_ctx *ctx)
err = 0;
out_err_free_hmac:
- crypto_free_hash(hmac);
+ crypto_free_shash(hmac);
out_err:
dprintk("RPC: %s: returning %d\n", __func__, err);
return err;
@@ -588,7 +591,7 @@ context_derive_keys_new(struct krb5_ctx *ctx, gfp_t gfp_mask)
context_v2_alloc_cipher(ctx, "cbc(aes)",
ctx->acceptor_seal);
if (ctx->acceptor_enc_aux == NULL) {
- crypto_free_blkcipher(ctx->initiator_enc_aux);
+ crypto_free_skcipher(ctx->initiator_enc_aux);
goto out_free_acceptor_enc;
}
}
@@ -596,9 +599,9 @@ context_derive_keys_new(struct krb5_ctx *ctx, gfp_t gfp_mask)
return 0;
out_free_acceptor_enc:
- crypto_free_blkcipher(ctx->acceptor_enc);
+ crypto_free_skcipher(ctx->acceptor_enc);
out_free_initiator_enc:
- crypto_free_blkcipher(ctx->initiator_enc);
+ crypto_free_skcipher(ctx->initiator_enc);
out_err:
return -EINVAL;
}
@@ -710,12 +713,12 @@ static void
gss_delete_sec_context_kerberos(void *internal_ctx) {
struct krb5_ctx *kctx = internal_ctx;
- crypto_free_blkcipher(kctx->seq);
- crypto_free_blkcipher(kctx->enc);
- crypto_free_blkcipher(kctx->acceptor_enc);
- crypto_free_blkcipher(kctx->initiator_enc);
- crypto_free_blkcipher(kctx->acceptor_enc_aux);
- crypto_free_blkcipher(kctx->initiator_enc_aux);
+ crypto_free_skcipher(kctx->seq);
+ crypto_free_skcipher(kctx->enc);
+ crypto_free_skcipher(kctx->acceptor_enc);
+ crypto_free_skcipher(kctx->initiator_enc);
+ crypto_free_skcipher(kctx->acceptor_enc_aux);
+ crypto_free_skcipher(kctx->initiator_enc_aux);
kfree(kctx->mech_used.data);
kfree(kctx);
}
diff --git a/net/sunrpc/auth_gss/gss_krb5_seqnum.c b/net/sunrpc/auth_gss/gss_krb5_seqnum.c
index 20d55c793eb6..c8b9082f4a9d 100644
--- a/net/sunrpc/auth_gss/gss_krb5_seqnum.c
+++ b/net/sunrpc/auth_gss/gss_krb5_seqnum.c
@@ -31,9 +31,9 @@
* PERFORMANCE OF THIS SOFTWARE.
*/
+#include <crypto/skcipher.h>
#include <linux/types.h>
#include <linux/sunrpc/gss_krb5.h>
-#include <linux/crypto.h>
#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
# define RPCDBG_FACILITY RPCDBG_AUTH
@@ -43,13 +43,13 @@ static s32
krb5_make_rc4_seq_num(struct krb5_ctx *kctx, int direction, s32 seqnum,
unsigned char *cksum, unsigned char *buf)
{
- struct crypto_blkcipher *cipher;
+ struct crypto_skcipher *cipher;
unsigned char plain[8];
s32 code;
dprintk("RPC: %s:\n", __func__);
- cipher = crypto_alloc_blkcipher(kctx->gk5e->encrypt_name, 0,
- CRYPTO_ALG_ASYNC);
+ cipher = crypto_alloc_skcipher(kctx->gk5e->encrypt_name, 0,
+ CRYPTO_ALG_ASYNC);
if (IS_ERR(cipher))
return PTR_ERR(cipher);
@@ -68,12 +68,12 @@ krb5_make_rc4_seq_num(struct krb5_ctx *kctx, int direction, s32 seqnum,
code = krb5_encrypt(cipher, cksum, plain, buf, 8);
out:
- crypto_free_blkcipher(cipher);
+ crypto_free_skcipher(cipher);
return code;
}
s32
krb5_make_seq_num(struct krb5_ctx *kctx,
- struct crypto_blkcipher *key,
+ struct crypto_skcipher *key,
int direction,
u32 seqnum,
unsigned char *cksum, unsigned char *buf)
@@ -101,13 +101,13 @@ static s32
krb5_get_rc4_seq_num(struct krb5_ctx *kctx, unsigned char *cksum,
unsigned char *buf, int *direction, s32 *seqnum)
{
- struct crypto_blkcipher *cipher;
+ struct crypto_skcipher *cipher;
unsigned char plain[8];
s32 code;
dprintk("RPC: %s:\n", __func__);
- cipher = crypto_alloc_blkcipher(kctx->gk5e->encrypt_name, 0,
- CRYPTO_ALG_ASYNC);
+ cipher = crypto_alloc_skcipher(kctx->gk5e->encrypt_name, 0,
+ CRYPTO_ALG_ASYNC);
if (IS_ERR(cipher))
return PTR_ERR(cipher);
@@ -130,7 +130,7 @@ krb5_get_rc4_seq_num(struct krb5_ctx *kctx, unsigned char *cksum,
*seqnum = ((plain[0] << 24) | (plain[1] << 16) |
(plain[2] << 8) | (plain[3]));
out:
- crypto_free_blkcipher(cipher);
+ crypto_free_skcipher(cipher);
return code;
}
@@ -142,7 +142,7 @@ krb5_get_seq_num(struct krb5_ctx *kctx,
{
s32 code;
unsigned char plain[8];
- struct crypto_blkcipher *key = kctx->seq;
+ struct crypto_skcipher *key = kctx->seq;
dprintk("RPC: krb5_get_seq_num:\n");
diff --git a/net/sunrpc/auth_gss/gss_krb5_wrap.c b/net/sunrpc/auth_gss/gss_krb5_wrap.c
index ca7e92a32f84..765088e4ad84 100644
--- a/net/sunrpc/auth_gss/gss_krb5_wrap.c
+++ b/net/sunrpc/auth_gss/gss_krb5_wrap.c
@@ -28,12 +28,12 @@
* SUCH DAMAGES.
*/
+#include <crypto/skcipher.h>
#include <linux/types.h>
#include <linux/jiffies.h>
#include <linux/sunrpc/gss_krb5.h>
#include <linux/random.h>
#include <linux/pagemap.h>
-#include <linux/crypto.h>
#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
# define RPCDBG_FACILITY RPCDBG_AUTH
@@ -174,7 +174,7 @@ gss_wrap_kerberos_v1(struct krb5_ctx *kctx, int offset,
now = get_seconds();
- blocksize = crypto_blkcipher_blocksize(kctx->enc);
+ blocksize = crypto_skcipher_blocksize(kctx->enc);
gss_krb5_add_padding(buf, offset, blocksize);
BUG_ON((buf->len - offset) % blocksize);
plainlen = conflen + buf->len - offset;
@@ -239,10 +239,10 @@ gss_wrap_kerberos_v1(struct krb5_ctx *kctx, int offset,
return GSS_S_FAILURE;
if (kctx->enctype == ENCTYPE_ARCFOUR_HMAC) {
- struct crypto_blkcipher *cipher;
+ struct crypto_skcipher *cipher;
int err;
- cipher = crypto_alloc_blkcipher(kctx->gk5e->encrypt_name, 0,
- CRYPTO_ALG_ASYNC);
+ cipher = crypto_alloc_skcipher(kctx->gk5e->encrypt_name, 0,
+ CRYPTO_ALG_ASYNC);
if (IS_ERR(cipher))
return GSS_S_FAILURE;
@@ -250,7 +250,7 @@ gss_wrap_kerberos_v1(struct krb5_ctx *kctx, int offset,
err = gss_encrypt_xdr_buf(cipher, buf,
offset + headlen - conflen, pages);
- crypto_free_blkcipher(cipher);
+ crypto_free_skcipher(cipher);
if (err)
return GSS_S_FAILURE;
} else {
@@ -327,18 +327,18 @@ gss_unwrap_kerberos_v1(struct krb5_ctx *kctx, int offset, struct xdr_buf *buf)
return GSS_S_BAD_SIG;
if (kctx->enctype == ENCTYPE_ARCFOUR_HMAC) {
- struct crypto_blkcipher *cipher;
+ struct crypto_skcipher *cipher;
int err;
- cipher = crypto_alloc_blkcipher(kctx->gk5e->encrypt_name, 0,
- CRYPTO_ALG_ASYNC);
+ cipher = crypto_alloc_skcipher(kctx->gk5e->encrypt_name, 0,
+ CRYPTO_ALG_ASYNC);
if (IS_ERR(cipher))
return GSS_S_FAILURE;
krb5_rc4_setup_enc_key(kctx, cipher, seqnum);
err = gss_decrypt_xdr_buf(cipher, buf, crypt_offset);
- crypto_free_blkcipher(cipher);
+ crypto_free_skcipher(cipher);
if (err)
return GSS_S_DEFECTIVE_TOKEN;
} else {
@@ -371,7 +371,7 @@ gss_unwrap_kerberos_v1(struct krb5_ctx *kctx, int offset, struct xdr_buf *buf)
/* Copy the data back to the right position. XXX: Would probably be
* better to copy and encrypt at the same time. */
- blocksize = crypto_blkcipher_blocksize(kctx->enc);
+ blocksize = crypto_skcipher_blocksize(kctx->enc);
data_start = ptr + (GSS_KRB5_TOK_HDR_LEN + kctx->gk5e->cksumlength) +
conflen;
orig_start = buf->head[0].iov_base + offset;
@@ -473,7 +473,7 @@ gss_wrap_kerberos_v2(struct krb5_ctx *kctx, u32 offset,
*ptr++ = 0xff;
be16ptr = (__be16 *)ptr;
- blocksize = crypto_blkcipher_blocksize(kctx->acceptor_enc);
+ blocksize = crypto_skcipher_blocksize(kctx->acceptor_enc);
*be16ptr++ = 0;
/* "inner" token header always uses 0 for RRC */
*be16ptr++ = 0;
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index b7f21044f4d8..7e0c9bf22df8 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -354,6 +354,7 @@ static void rpc_free_clid(struct rpc_clnt *clnt)
}
static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args,
+ struct rpc_xprt_switch *xps,
struct rpc_xprt *xprt,
struct rpc_clnt *parent)
{
@@ -411,6 +412,8 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args,
}
rpc_clnt_set_transport(clnt, xprt, timeout);
+ xprt_iter_init(&clnt->cl_xpi, xps);
+ xprt_switch_put(xps);
clnt->cl_rtt = &clnt->cl_rtt_default;
rpc_init_rtt(&clnt->cl_rtt_default, clnt->cl_timeout->to_initval);
@@ -438,6 +441,7 @@ out_no_clid:
out_err:
rpciod_down();
out_no_rpciod:
+ xprt_switch_put(xps);
xprt_put(xprt);
return ERR_PTR(err);
}
@@ -446,8 +450,13 @@ struct rpc_clnt *rpc_create_xprt(struct rpc_create_args *args,
struct rpc_xprt *xprt)
{
struct rpc_clnt *clnt = NULL;
+ struct rpc_xprt_switch *xps;
- clnt = rpc_new_client(args, xprt, NULL);
+ xps = xprt_switch_alloc(xprt, GFP_KERNEL);
+ if (xps == NULL)
+ return ERR_PTR(-ENOMEM);
+
+ clnt = rpc_new_client(args, xps, xprt, NULL);
if (IS_ERR(clnt))
return clnt;
@@ -564,6 +573,7 @@ EXPORT_SYMBOL_GPL(rpc_create);
static struct rpc_clnt *__rpc_clone_client(struct rpc_create_args *args,
struct rpc_clnt *clnt)
{
+ struct rpc_xprt_switch *xps;
struct rpc_xprt *xprt;
struct rpc_clnt *new;
int err;
@@ -571,13 +581,17 @@ static struct rpc_clnt *__rpc_clone_client(struct rpc_create_args *args,
err = -ENOMEM;
rcu_read_lock();
xprt = xprt_get(rcu_dereference(clnt->cl_xprt));
+ xps = xprt_switch_get(rcu_dereference(clnt->cl_xpi.xpi_xpswitch));
rcu_read_unlock();
- if (xprt == NULL)
+ if (xprt == NULL || xps == NULL) {
+ xprt_put(xprt);
+ xprt_switch_put(xps);
goto out_err;
+ }
args->servername = xprt->servername;
args->nodename = clnt->cl_nodename;
- new = rpc_new_client(args, xprt, clnt);
+ new = rpc_new_client(args, xps, xprt, clnt);
if (IS_ERR(new)) {
err = PTR_ERR(new);
goto out_err;
@@ -657,6 +671,7 @@ int rpc_switch_client_transport(struct rpc_clnt *clnt,
{
const struct rpc_timeout *old_timeo;
rpc_authflavor_t pseudoflavor;
+ struct rpc_xprt_switch *xps, *oldxps;
struct rpc_xprt *xprt, *old;
struct rpc_clnt *parent;
int err;
@@ -668,10 +683,17 @@ int rpc_switch_client_transport(struct rpc_clnt *clnt,
return PTR_ERR(xprt);
}
+ xps = xprt_switch_alloc(xprt, GFP_KERNEL);
+ if (xps == NULL) {
+ xprt_put(xprt);
+ return -ENOMEM;
+ }
+
pseudoflavor = clnt->cl_auth->au_flavor;
old_timeo = clnt->cl_timeout;
old = rpc_clnt_set_transport(clnt, xprt, timeout);
+ oldxps = xprt_iter_xchg_switch(&clnt->cl_xpi, xps);
rpc_unregister_client(clnt);
__rpc_clnt_remove_pipedir(clnt);
@@ -697,20 +719,74 @@ int rpc_switch_client_transport(struct rpc_clnt *clnt,
synchronize_rcu();
if (parent != clnt)
rpc_release_client(parent);
+ xprt_switch_put(oldxps);
xprt_put(old);
dprintk("RPC: replaced xprt for clnt %p\n", clnt);
return 0;
out_revert:
+ xps = xprt_iter_xchg_switch(&clnt->cl_xpi, oldxps);
rpc_clnt_set_transport(clnt, old, old_timeo);
clnt->cl_parent = parent;
rpc_client_register(clnt, pseudoflavor, NULL);
+ xprt_switch_put(xps);
xprt_put(xprt);
dprintk("RPC: failed to switch xprt for clnt %p\n", clnt);
return err;
}
EXPORT_SYMBOL_GPL(rpc_switch_client_transport);
+static
+int rpc_clnt_xprt_iter_init(struct rpc_clnt *clnt, struct rpc_xprt_iter *xpi)
+{
+ struct rpc_xprt_switch *xps;
+
+ rcu_read_lock();
+ xps = xprt_switch_get(rcu_dereference(clnt->cl_xpi.xpi_xpswitch));
+ rcu_read_unlock();
+ if (xps == NULL)
+ return -EAGAIN;
+ xprt_iter_init_listall(xpi, xps);
+ xprt_switch_put(xps);
+ return 0;
+}
+
+/**
+ * rpc_clnt_iterate_for_each_xprt - Apply a function to all transports
+ * @clnt: pointer to client
+ * @fn: function to apply
+ * @data: void pointer to function data
+ *
+ * Iterates through the list of RPC transports currently attached to the
+ * client and applies the function fn(clnt, xprt, data).
+ *
+ * On error, the iteration stops, and the function returns the error value.
+ */
+int rpc_clnt_iterate_for_each_xprt(struct rpc_clnt *clnt,
+ int (*fn)(struct rpc_clnt *, struct rpc_xprt *, void *),
+ void *data)
+{
+ struct rpc_xprt_iter xpi;
+ int ret;
+
+ ret = rpc_clnt_xprt_iter_init(clnt, &xpi);
+ if (ret)
+ return ret;
+ for (;;) {
+ struct rpc_xprt *xprt = xprt_iter_get_next(&xpi);
+
+ if (!xprt)
+ break;
+ ret = fn(clnt, xprt, data);
+ xprt_put(xprt);
+ if (ret < 0)
+ break;
+ }
+ xprt_iter_destroy(&xpi);
+ return ret;
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_iterate_for_each_xprt);
+
/*
* Kill all tasks for the given client.
* XXX: kill their descendants as well?
@@ -783,6 +859,7 @@ rpc_free_client(struct rpc_clnt *clnt)
rpc_free_iostats(clnt->cl_metrics);
clnt->cl_metrics = NULL;
xprt_put(rcu_dereference_raw(clnt->cl_xprt));
+ xprt_iter_destroy(&clnt->cl_xpi);
rpciod_down();
rpc_free_clid(clnt);
kfree(clnt);
@@ -868,6 +945,7 @@ EXPORT_SYMBOL_GPL(rpc_bind_new_program);
void rpc_task_release_client(struct rpc_task *task)
{
struct rpc_clnt *clnt = task->tk_client;
+ struct rpc_xprt *xprt = task->tk_xprt;
if (clnt != NULL) {
/* Remove from client task list */
@@ -878,13 +956,22 @@ void rpc_task_release_client(struct rpc_task *task)
rpc_release_client(clnt);
}
+
+ if (xprt != NULL) {
+ task->tk_xprt = NULL;
+
+ xprt_put(xprt);
+ }
}
static
void rpc_task_set_client(struct rpc_task *task, struct rpc_clnt *clnt)
{
+
if (clnt != NULL) {
rpc_task_release_client(task);
+ if (task->tk_xprt == NULL)
+ task->tk_xprt = xprt_iter_get_next(&clnt->cl_xpi);
task->tk_client = clnt;
atomic_inc(&clnt->cl_count);
if (clnt->cl_softrtry)
@@ -900,14 +987,6 @@ void rpc_task_set_client(struct rpc_task *task, struct rpc_clnt *clnt)
}
}
-void rpc_task_reset_client(struct rpc_task *task, struct rpc_clnt *clnt)
-{
- rpc_task_release_client(task);
- rpc_task_set_client(task, clnt);
-}
-EXPORT_SYMBOL_GPL(rpc_task_reset_client);
-
-
static void
rpc_task_set_rpc_message(struct rpc_task *task, const struct rpc_message *msg)
{
@@ -2104,11 +2183,9 @@ call_timeout(struct rpc_task *task)
}
if (RPC_IS_SOFT(task)) {
if (clnt->cl_chatty) {
- rcu_read_lock();
printk(KERN_NOTICE "%s: server %s not responding, timed out\n",
clnt->cl_program->name,
- rcu_dereference(clnt->cl_xprt)->servername);
- rcu_read_unlock();
+ task->tk_xprt->servername);
}
if (task->tk_flags & RPC_TASK_TIMEOUT)
rpc_exit(task, -ETIMEDOUT);
@@ -2120,11 +2197,9 @@ call_timeout(struct rpc_task *task)
if (!(task->tk_flags & RPC_CALL_MAJORSEEN)) {
task->tk_flags |= RPC_CALL_MAJORSEEN;
if (clnt->cl_chatty) {
- rcu_read_lock();
printk(KERN_NOTICE "%s: server %s not responding, still trying\n",
clnt->cl_program->name,
- rcu_dereference(clnt->cl_xprt)->servername);
- rcu_read_unlock();
+ task->tk_xprt->servername);
}
}
rpc_force_rebind(clnt);
@@ -2154,11 +2229,9 @@ call_decode(struct rpc_task *task)
if (task->tk_flags & RPC_CALL_MAJORSEEN) {
if (clnt->cl_chatty) {
- rcu_read_lock();
printk(KERN_NOTICE "%s: server %s OK\n",
clnt->cl_program->name,
- rcu_dereference(clnt->cl_xprt)->servername);
- rcu_read_unlock();
+ task->tk_xprt->servername);
}
task->tk_flags &= ~RPC_CALL_MAJORSEEN;
}
@@ -2312,11 +2385,9 @@ rpc_verify_header(struct rpc_task *task)
task->tk_action = call_bind;
goto out_retry;
case RPC_AUTH_TOOWEAK:
- rcu_read_lock();
printk(KERN_NOTICE "RPC: server %s requires stronger "
"authentication.\n",
- rcu_dereference(clnt->cl_xprt)->servername);
- rcu_read_unlock();
+ task->tk_xprt->servername);
break;
default:
dprintk("RPC: %5u %s: unknown auth error: %x\n",
@@ -2341,27 +2412,27 @@ rpc_verify_header(struct rpc_task *task)
case RPC_SUCCESS:
return p;
case RPC_PROG_UNAVAIL:
- dprintk_rcu("RPC: %5u %s: program %u is unsupported "
+ dprintk("RPC: %5u %s: program %u is unsupported "
"by server %s\n", task->tk_pid, __func__,
(unsigned int)clnt->cl_prog,
- rcu_dereference(clnt->cl_xprt)->servername);
+ task->tk_xprt->servername);
error = -EPFNOSUPPORT;
goto out_err;
case RPC_PROG_MISMATCH:
- dprintk_rcu("RPC: %5u %s: program %u, version %u unsupported "
+ dprintk("RPC: %5u %s: program %u, version %u unsupported "
"by server %s\n", task->tk_pid, __func__,
(unsigned int)clnt->cl_prog,
(unsigned int)clnt->cl_vers,
- rcu_dereference(clnt->cl_xprt)->servername);
+ task->tk_xprt->servername);
error = -EPROTONOSUPPORT;
goto out_err;
case RPC_PROC_UNAVAIL:
- dprintk_rcu("RPC: %5u %s: proc %s unsupported by program %u, "
+ dprintk("RPC: %5u %s: proc %s unsupported by program %u, "
"version %u on server %s\n",
task->tk_pid, __func__,
rpc_proc_name(task),
clnt->cl_prog, clnt->cl_vers,
- rcu_dereference(clnt->cl_xprt)->servername);
+ task->tk_xprt->servername);
error = -EOPNOTSUPP;
goto out_err;
case RPC_GARBAGE_ARGS:
@@ -2421,7 +2492,10 @@ static int rpc_ping(struct rpc_clnt *clnt)
return err;
}
-struct rpc_task *rpc_call_null(struct rpc_clnt *clnt, struct rpc_cred *cred, int flags)
+static
+struct rpc_task *rpc_call_null_helper(struct rpc_clnt *clnt,
+ struct rpc_xprt *xprt, struct rpc_cred *cred, int flags,
+ const struct rpc_call_ops *ops, void *data)
{
struct rpc_message msg = {
.rpc_proc = &rpcproc_null,
@@ -2429,14 +2503,140 @@ struct rpc_task *rpc_call_null(struct rpc_clnt *clnt, struct rpc_cred *cred, int
};
struct rpc_task_setup task_setup_data = {
.rpc_client = clnt,
+ .rpc_xprt = xprt,
.rpc_message = &msg,
- .callback_ops = &rpc_default_ops,
+ .callback_ops = (ops != NULL) ? ops : &rpc_default_ops,
+ .callback_data = data,
.flags = flags,
};
+
return rpc_run_task(&task_setup_data);
}
+
+struct rpc_task *rpc_call_null(struct rpc_clnt *clnt, struct rpc_cred *cred, int flags)
+{
+ return rpc_call_null_helper(clnt, NULL, cred, flags, NULL, NULL);
+}
EXPORT_SYMBOL_GPL(rpc_call_null);
+struct rpc_cb_add_xprt_calldata {
+ struct rpc_xprt_switch *xps;
+ struct rpc_xprt *xprt;
+};
+
+static void rpc_cb_add_xprt_done(struct rpc_task *task, void *calldata)
+{
+ struct rpc_cb_add_xprt_calldata *data = calldata;
+
+ if (task->tk_status == 0)
+ rpc_xprt_switch_add_xprt(data->xps, data->xprt);
+}
+
+static void rpc_cb_add_xprt_release(void *calldata)
+{
+ struct rpc_cb_add_xprt_calldata *data = calldata;
+
+ xprt_put(data->xprt);
+ xprt_switch_put(data->xps);
+ kfree(data);
+}
+
+const static struct rpc_call_ops rpc_cb_add_xprt_call_ops = {
+ .rpc_call_done = rpc_cb_add_xprt_done,
+ .rpc_release = rpc_cb_add_xprt_release,
+};
+
+/**
+ * rpc_clnt_test_and_add_xprt - Test and add a new transport to a rpc_clnt
+ * @clnt: pointer to struct rpc_clnt
+ * @xps: pointer to struct rpc_xprt_switch,
+ * @xprt: pointer struct rpc_xprt
+ * @dummy: unused
+ */
+int rpc_clnt_test_and_add_xprt(struct rpc_clnt *clnt,
+ struct rpc_xprt_switch *xps, struct rpc_xprt *xprt,
+ void *dummy)
+{
+ struct rpc_cb_add_xprt_calldata *data;
+ struct rpc_cred *cred;
+ struct rpc_task *task;
+
+ data = kmalloc(sizeof(*data), GFP_NOFS);
+ if (!data)
+ return -ENOMEM;
+ data->xps = xprt_switch_get(xps);
+ data->xprt = xprt_get(xprt);
+
+ cred = authnull_ops.lookup_cred(NULL, NULL, 0);
+ task = rpc_call_null_helper(clnt, xprt, cred,
+ RPC_TASK_SOFT|RPC_TASK_SOFTCONN|RPC_TASK_ASYNC,
+ &rpc_cb_add_xprt_call_ops, data);
+ put_rpccred(cred);
+ if (IS_ERR(task))
+ return PTR_ERR(task);
+ rpc_put_task(task);
+ return 1;
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_test_and_add_xprt);
+
+/**
+ * rpc_clnt_add_xprt - Add a new transport to a rpc_clnt
+ * @clnt: pointer to struct rpc_clnt
+ * @xprtargs: pointer to struct xprt_create
+ * @setup: callback to test and/or set up the connection
+ * @data: pointer to setup function data
+ *
+ * Creates a new transport using the parameters set in args and
+ * adds it to clnt.
+ * If ping is set, then test that connectivity succeeds before
+ * adding the new transport.
+ *
+ */
+int rpc_clnt_add_xprt(struct rpc_clnt *clnt,
+ struct xprt_create *xprtargs,
+ int (*setup)(struct rpc_clnt *,
+ struct rpc_xprt_switch *,
+ struct rpc_xprt *,
+ void *),
+ void *data)
+{
+ struct rpc_xprt_switch *xps;
+ struct rpc_xprt *xprt;
+ unsigned char resvport;
+ int ret = 0;
+
+ rcu_read_lock();
+ xps = xprt_switch_get(rcu_dereference(clnt->cl_xpi.xpi_xpswitch));
+ xprt = xprt_iter_xprt(&clnt->cl_xpi);
+ if (xps == NULL || xprt == NULL) {
+ rcu_read_unlock();
+ return -EAGAIN;
+ }
+ resvport = xprt->resvport;
+ rcu_read_unlock();
+
+ xprt = xprt_create_transport(xprtargs);
+ if (IS_ERR(xprt)) {
+ ret = PTR_ERR(xprt);
+ goto out_put_switch;
+ }
+ xprt->resvport = resvport;
+
+ rpc_xprt_switch_set_roundrobin(xps);
+ if (setup) {
+ ret = setup(clnt, xps, xprt, data);
+ if (ret != 0)
+ goto out_put_xprt;
+ }
+ rpc_xprt_switch_add_xprt(xps, xprt);
+out_put_xprt:
+ xprt_put(xprt);
+out_put_switch:
+ xprt_switch_put(xps);
+ return ret;
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_add_xprt);
+
#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
static void rpc_show_header(void)
{
@@ -2483,57 +2683,39 @@ void rpc_show_tasks(struct net *net)
#endif
#if IS_ENABLED(CONFIG_SUNRPC_SWAP)
+static int
+rpc_clnt_swap_activate_callback(struct rpc_clnt *clnt,
+ struct rpc_xprt *xprt,
+ void *dummy)
+{
+ return xprt_enable_swap(xprt);
+}
+
int
rpc_clnt_swap_activate(struct rpc_clnt *clnt)
{
- int ret = 0;
- struct rpc_xprt *xprt;
-
- if (atomic_inc_return(&clnt->cl_swapper) == 1) {
-retry:
- rcu_read_lock();
- xprt = xprt_get(rcu_dereference(clnt->cl_xprt));
- rcu_read_unlock();
- if (!xprt) {
- /*
- * If we didn't get a reference, then we likely are
- * racing with a migration event. Wait for a grace
- * period and try again.
- */
- synchronize_rcu();
- goto retry;
- }
-
- ret = xprt_enable_swap(xprt);
- xprt_put(xprt);
- }
- return ret;
+ if (atomic_inc_return(&clnt->cl_swapper) == 1)
+ return rpc_clnt_iterate_for_each_xprt(clnt,
+ rpc_clnt_swap_activate_callback, NULL);
+ return 0;
}
EXPORT_SYMBOL_GPL(rpc_clnt_swap_activate);
+static int
+rpc_clnt_swap_deactivate_callback(struct rpc_clnt *clnt,
+ struct rpc_xprt *xprt,
+ void *dummy)
+{
+ xprt_disable_swap(xprt);
+ return 0;
+}
+
void
rpc_clnt_swap_deactivate(struct rpc_clnt *clnt)
{
- struct rpc_xprt *xprt;
-
- if (atomic_dec_if_positive(&clnt->cl_swapper) == 0) {
-retry:
- rcu_read_lock();
- xprt = xprt_get(rcu_dereference(clnt->cl_xprt));
- rcu_read_unlock();
- if (!xprt) {
- /*
- * If we didn't get a reference, then we likely are
- * racing with a migration event. Wait for a grace
- * period and try again.
- */
- synchronize_rcu();
- goto retry;
- }
-
- xprt_disable_swap(xprt);
- xprt_put(xprt);
- }
+ if (atomic_dec_if_positive(&clnt->cl_swapper) == 0)
+ rpc_clnt_iterate_for_each_xprt(clnt,
+ rpc_clnt_swap_deactivate_callback, NULL);
}
EXPORT_SYMBOL_GPL(rpc_clnt_swap_deactivate);
#endif /* CONFIG_SUNRPC_SWAP */
diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c
index cf5770d8f49a..5b30603596d0 100644
--- a/net/sunrpc/rpcb_clnt.c
+++ b/net/sunrpc/rpcb_clnt.c
@@ -648,10 +648,10 @@ static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbi
static struct rpc_clnt *rpcb_find_transport_owner(struct rpc_clnt *clnt)
{
struct rpc_clnt *parent = clnt->cl_parent;
- struct rpc_xprt *xprt = rcu_dereference(clnt->cl_xprt);
+ struct rpc_xprt_switch *xps = rcu_access_pointer(clnt->cl_xpi.xpi_xpswitch);
while (parent != clnt) {
- if (rcu_dereference(parent->cl_xprt) != xprt)
+ if (rcu_access_pointer(parent->cl_xpi.xpi_xpswitch) != xps)
break;
if (clnt->cl_autobind)
break;
@@ -683,11 +683,9 @@ void rpcb_getport_async(struct rpc_task *task)
int status;
rcu_read_lock();
- do {
- clnt = rpcb_find_transport_owner(task->tk_client);
- xprt = xprt_get(rcu_dereference(clnt->cl_xprt));
- } while (xprt == NULL);
+ clnt = rpcb_find_transport_owner(task->tk_client);
rcu_read_unlock();
+ xprt = xprt_get(task->tk_xprt);
dprintk("RPC: %5u %s(%s, %u, %u, %d)\n",
task->tk_pid, __func__,
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 73ad57a59989..fcfd48d263f6 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -909,6 +909,8 @@ static void rpc_init_task(struct rpc_task *task, const struct rpc_task_setup *ta
/* Initialize workqueue for async tasks */
task->tk_workqueue = task_setup_data->workqueue;
+ task->tk_xprt = xprt_get(task_setup_data->rpc_xprt);
+
if (task->tk_ops->rpc_call_prepare != NULL)
task->tk_action = rpc_prepare_task;
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 37edea6fa92d..216a1385718a 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -48,6 +48,7 @@
#include <linux/sunrpc/clnt.h>
#include <linux/sunrpc/metrics.h>
#include <linux/sunrpc/bc_xprt.h>
+#include <linux/rcupdate.h>
#include <trace/events/sunrpc.h>
@@ -1166,7 +1167,7 @@ void xprt_free(struct rpc_xprt *xprt)
{
put_net(xprt->xprt_net);
xprt_free_all_slots(xprt);
- kfree(xprt);
+ kfree_rcu(xprt, rcu);
}
EXPORT_SYMBOL_GPL(xprt_free);
@@ -1180,7 +1181,7 @@ EXPORT_SYMBOL_GPL(xprt_free);
*/
void xprt_reserve(struct rpc_task *task)
{
- struct rpc_xprt *xprt;
+ struct rpc_xprt *xprt = task->tk_xprt;
task->tk_status = 0;
if (task->tk_rqstp != NULL)
@@ -1188,11 +1189,8 @@ void xprt_reserve(struct rpc_task *task)
task->tk_timeout = 0;
task->tk_status = -EAGAIN;
- rcu_read_lock();
- xprt = rcu_dereference(task->tk_client->cl_xprt);
if (!xprt_throttle_congested(xprt, task))
xprt->ops->alloc_slot(xprt, task);
- rcu_read_unlock();
}
/**
@@ -1206,7 +1204,7 @@ void xprt_reserve(struct rpc_task *task)
*/
void xprt_retry_reserve(struct rpc_task *task)
{
- struct rpc_xprt *xprt;
+ struct rpc_xprt *xprt = task->tk_xprt;
task->tk_status = 0;
if (task->tk_rqstp != NULL)
@@ -1214,10 +1212,7 @@ void xprt_retry_reserve(struct rpc_task *task)
task->tk_timeout = 0;
task->tk_status = -EAGAIN;
- rcu_read_lock();
- xprt = rcu_dereference(task->tk_client->cl_xprt);
xprt->ops->alloc_slot(xprt, task);
- rcu_read_unlock();
}
static inline __be32 xprt_alloc_xid(struct rpc_xprt *xprt)
@@ -1264,11 +1259,9 @@ void xprt_release(struct rpc_task *task)
if (req == NULL) {
if (task->tk_client) {
- rcu_read_lock();
- xprt = rcu_dereference(task->tk_client->cl_xprt);
+ xprt = task->tk_xprt;
if (xprt->snd_task == task)
xprt_release_write(xprt, task);
- rcu_read_unlock();
}
return;
}
@@ -1307,7 +1300,7 @@ void xprt_release(struct rpc_task *task)
static void xprt_init(struct rpc_xprt *xprt, struct net *net)
{
- atomic_set(&xprt->count, 1);
+ kref_init(&xprt->kref);
spin_lock_init(&xprt->transport_lock);
spin_lock_init(&xprt->reserve_lock);
@@ -1318,6 +1311,7 @@ static void xprt_init(struct rpc_xprt *xprt, struct net *net)
spin_lock_init(&xprt->bc_pa_lock);
INIT_LIST_HEAD(&xprt->bc_pa_list);
#endif /* CONFIG_SUNRPC_BACKCHANNEL */
+ INIT_LIST_HEAD(&xprt->xprt_switch);
xprt->last_used = jiffies;
xprt->cwnd = RPC_INITCWND;
@@ -1415,6 +1409,24 @@ static void xprt_destroy(struct rpc_xprt *xprt)
xprt->ops->destroy(xprt);
}
+static void xprt_destroy_kref(struct kref *kref)
+{
+ xprt_destroy(container_of(kref, struct rpc_xprt, kref));
+}
+
+/**
+ * xprt_get - return a reference to an RPC transport.
+ * @xprt: pointer to the transport
+ *
+ */
+struct rpc_xprt *xprt_get(struct rpc_xprt *xprt)
+{
+ if (xprt != NULL && kref_get_unless_zero(&xprt->kref))
+ return xprt;
+ return NULL;
+}
+EXPORT_SYMBOL_GPL(xprt_get);
+
/**
* xprt_put - release a reference to an RPC transport.
* @xprt: pointer to the transport
@@ -1422,7 +1434,7 @@ static void xprt_destroy(struct rpc_xprt *xprt)
*/
void xprt_put(struct rpc_xprt *xprt)
{
- if (atomic_dec_and_test(&xprt->count))
- xprt_destroy(xprt);
+ if (xprt != NULL)
+ kref_put(&xprt->kref, xprt_destroy_kref);
}
EXPORT_SYMBOL_GPL(xprt_put);
diff --git a/net/sunrpc/xprtmultipath.c b/net/sunrpc/xprtmultipath.c
new file mode 100644
index 000000000000..e7fd76975d86
--- /dev/null
+++ b/net/sunrpc/xprtmultipath.c
@@ -0,0 +1,475 @@
+/*
+ * Multipath support for RPC
+ *
+ * Copyright (c) 2015, 2016, Primary Data, Inc. All rights reserved.
+ *
+ * Trond Myklebust <trond.myklebust@primarydata.com>
+ *
+ */
+#include <linux/types.h>
+#include <linux/kref.h>
+#include <linux/list.h>
+#include <linux/rcupdate.h>
+#include <linux/rculist.h>
+#include <linux/slab.h>
+#include <asm/cmpxchg.h>
+#include <linux/spinlock.h>
+#include <linux/sunrpc/xprt.h>
+#include <linux/sunrpc/xprtmultipath.h>
+
+typedef struct rpc_xprt *(*xprt_switch_find_xprt_t)(struct list_head *head,
+ const struct rpc_xprt *cur);
+
+static const struct rpc_xprt_iter_ops rpc_xprt_iter_singular;
+static const struct rpc_xprt_iter_ops rpc_xprt_iter_roundrobin;
+static const struct rpc_xprt_iter_ops rpc_xprt_iter_listall;
+
+static void xprt_switch_add_xprt_locked(struct rpc_xprt_switch *xps,
+ struct rpc_xprt *xprt)
+{
+ if (unlikely(xprt_get(xprt) == NULL))
+ return;
+ list_add_tail_rcu(&xprt->xprt_switch, &xps->xps_xprt_list);
+ smp_wmb();
+ if (xps->xps_nxprts == 0)
+ xps->xps_net = xprt->xprt_net;
+ xps->xps_nxprts++;
+}
+
+/**
+ * rpc_xprt_switch_add_xprt - Add a new rpc_xprt to an rpc_xprt_switch
+ * @xps: pointer to struct rpc_xprt_switch
+ * @xprt: pointer to struct rpc_xprt
+ *
+ * Adds xprt to the end of the list of struct rpc_xprt in xps.
+ */
+void rpc_xprt_switch_add_xprt(struct rpc_xprt_switch *xps,
+ struct rpc_xprt *xprt)
+{
+ if (xprt == NULL)
+ return;
+ spin_lock(&xps->xps_lock);
+ if (xps->xps_net == xprt->xprt_net || xps->xps_net == NULL)
+ xprt_switch_add_xprt_locked(xps, xprt);
+ spin_unlock(&xps->xps_lock);
+}
+
+static void xprt_switch_remove_xprt_locked(struct rpc_xprt_switch *xps,
+ struct rpc_xprt *xprt)
+{
+ if (unlikely(xprt == NULL))
+ return;
+ xps->xps_nxprts--;
+ if (xps->xps_nxprts == 0)
+ xps->xps_net = NULL;
+ smp_wmb();
+ list_del_rcu(&xprt->xprt_switch);
+}
+
+/**
+ * rpc_xprt_switch_remove_xprt - Removes an rpc_xprt from a rpc_xprt_switch
+ * @xps: pointer to struct rpc_xprt_switch
+ * @xprt: pointer to struct rpc_xprt
+ *
+ * Removes xprt from the list of struct rpc_xprt in xps.
+ */
+void rpc_xprt_switch_remove_xprt(struct rpc_xprt_switch *xps,
+ struct rpc_xprt *xprt)
+{
+ spin_lock(&xps->xps_lock);
+ xprt_switch_remove_xprt_locked(xps, xprt);
+ spin_unlock(&xps->xps_lock);
+ xprt_put(xprt);
+}
+
+/**
+ * xprt_switch_alloc - Allocate a new struct rpc_xprt_switch
+ * @xprt: pointer to struct rpc_xprt
+ * @gfp_flags: allocation flags
+ *
+ * On success, returns an initialised struct rpc_xprt_switch, containing
+ * the entry xprt. Returns NULL on failure.
+ */
+struct rpc_xprt_switch *xprt_switch_alloc(struct rpc_xprt *xprt,
+ gfp_t gfp_flags)
+{
+ struct rpc_xprt_switch *xps;
+
+ xps = kmalloc(sizeof(*xps), gfp_flags);
+ if (xps != NULL) {
+ spin_lock_init(&xps->xps_lock);
+ kref_init(&xps->xps_kref);
+ xps->xps_nxprts = 0;
+ INIT_LIST_HEAD(&xps->xps_xprt_list);
+ xps->xps_iter_ops = &rpc_xprt_iter_singular;
+ xprt_switch_add_xprt_locked(xps, xprt);
+ }
+
+ return xps;
+}
+
+static void xprt_switch_free_entries(struct rpc_xprt_switch *xps)
+{
+ spin_lock(&xps->xps_lock);
+ while (!list_empty(&xps->xps_xprt_list)) {
+ struct rpc_xprt *xprt;
+
+ xprt = list_first_entry(&xps->xps_xprt_list,
+ struct rpc_xprt, xprt_switch);
+ xprt_switch_remove_xprt_locked(xps, xprt);
+ spin_unlock(&xps->xps_lock);
+ xprt_put(xprt);
+ spin_lock(&xps->xps_lock);
+ }
+ spin_unlock(&xps->xps_lock);
+}
+
+static void xprt_switch_free(struct kref *kref)
+{
+ struct rpc_xprt_switch *xps = container_of(kref,
+ struct rpc_xprt_switch, xps_kref);
+
+ xprt_switch_free_entries(xps);
+ kfree_rcu(xps, xps_rcu);
+}
+
+/**
+ * xprt_switch_get - Return a reference to a rpc_xprt_switch
+ * @xps: pointer to struct rpc_xprt_switch
+ *
+ * Returns a reference to xps unless the refcount is already zero.
+ */
+struct rpc_xprt_switch *xprt_switch_get(struct rpc_xprt_switch *xps)
+{
+ if (xps != NULL && kref_get_unless_zero(&xps->xps_kref))
+ return xps;
+ return NULL;
+}
+
+/**
+ * xprt_switch_put - Release a reference to a rpc_xprt_switch
+ * @xps: pointer to struct rpc_xprt_switch
+ *
+ * Release the reference to xps, and free it once the refcount is zero.
+ */
+void xprt_switch_put(struct rpc_xprt_switch *xps)
+{
+ if (xps != NULL)
+ kref_put(&xps->xps_kref, xprt_switch_free);
+}
+
+/**
+ * rpc_xprt_switch_set_roundrobin - Set a round-robin policy on rpc_xprt_switch
+ * @xps: pointer to struct rpc_xprt_switch
+ *
+ * Sets a round-robin default policy for iterators acting on xps.
+ */
+void rpc_xprt_switch_set_roundrobin(struct rpc_xprt_switch *xps)
+{
+ if (READ_ONCE(xps->xps_iter_ops) != &rpc_xprt_iter_roundrobin)
+ WRITE_ONCE(xps->xps_iter_ops, &rpc_xprt_iter_roundrobin);
+}
+
+static
+const struct rpc_xprt_iter_ops *xprt_iter_ops(const struct rpc_xprt_iter *xpi)
+{
+ if (xpi->xpi_ops != NULL)
+ return xpi->xpi_ops;
+ return rcu_dereference(xpi->xpi_xpswitch)->xps_iter_ops;
+}
+
+static
+void xprt_iter_no_rewind(struct rpc_xprt_iter *xpi)
+{
+}
+
+static
+void xprt_iter_default_rewind(struct rpc_xprt_iter *xpi)
+{
+ WRITE_ONCE(xpi->xpi_cursor, NULL);
+}
+
+static
+struct rpc_xprt *xprt_switch_find_first_entry(struct list_head *head)
+{
+ return list_first_or_null_rcu(head, struct rpc_xprt, xprt_switch);
+}
+
+static
+struct rpc_xprt *xprt_iter_first_entry(struct rpc_xprt_iter *xpi)
+{
+ struct rpc_xprt_switch *xps = rcu_dereference(xpi->xpi_xpswitch);
+
+ if (xps == NULL)
+ return NULL;
+ return xprt_switch_find_first_entry(&xps->xps_xprt_list);
+}
+
+static
+struct rpc_xprt *xprt_switch_find_current_entry(struct list_head *head,
+ const struct rpc_xprt *cur)
+{
+ struct rpc_xprt *pos;
+
+ list_for_each_entry_rcu(pos, head, xprt_switch) {
+ if (cur == pos)
+ return pos;
+ }
+ return NULL;
+}
+
+static
+struct rpc_xprt *xprt_iter_current_entry(struct rpc_xprt_iter *xpi)
+{
+ struct rpc_xprt_switch *xps = rcu_dereference(xpi->xpi_xpswitch);
+ struct list_head *head;
+
+ if (xps == NULL)
+ return NULL;
+ head = &xps->xps_xprt_list;
+ if (xpi->xpi_cursor == NULL || xps->xps_nxprts < 2)
+ return xprt_switch_find_first_entry(head);
+ return xprt_switch_find_current_entry(head, xpi->xpi_cursor);
+}
+
+static
+struct rpc_xprt *xprt_switch_find_next_entry(struct list_head *head,
+ const struct rpc_xprt *cur)
+{
+ struct rpc_xprt *pos, *prev = NULL;
+
+ list_for_each_entry_rcu(pos, head, xprt_switch) {
+ if (cur == prev)
+ return pos;
+ prev = pos;
+ }
+ return NULL;
+}
+
+static
+struct rpc_xprt *xprt_switch_set_next_cursor(struct list_head *head,
+ struct rpc_xprt **cursor,
+ xprt_switch_find_xprt_t find_next)
+{
+ struct rpc_xprt *cur, *pos, *old;
+
+ cur = READ_ONCE(*cursor);
+ for (;;) {
+ old = cur;
+ pos = find_next(head, old);
+ if (pos == NULL)
+ break;
+ cur = cmpxchg_relaxed(cursor, old, pos);
+ if (cur == old)
+ break;
+ }
+ return pos;
+}
+
+static
+struct rpc_xprt *xprt_iter_next_entry_multiple(struct rpc_xprt_iter *xpi,
+ xprt_switch_find_xprt_t find_next)
+{
+ struct rpc_xprt_switch *xps = rcu_dereference(xpi->xpi_xpswitch);
+ struct list_head *head;
+
+ if (xps == NULL)
+ return NULL;
+ head = &xps->xps_xprt_list;
+ if (xps->xps_nxprts < 2)
+ return xprt_switch_find_first_entry(head);
+ return xprt_switch_set_next_cursor(head, &xpi->xpi_cursor, find_next);
+}
+
+static
+struct rpc_xprt *xprt_switch_find_next_entry_roundrobin(struct list_head *head,
+ const struct rpc_xprt *cur)
+{
+ struct rpc_xprt *ret;
+
+ ret = xprt_switch_find_next_entry(head, cur);
+ if (ret != NULL)
+ return ret;
+ return xprt_switch_find_first_entry(head);
+}
+
+static
+struct rpc_xprt *xprt_iter_next_entry_roundrobin(struct rpc_xprt_iter *xpi)
+{
+ return xprt_iter_next_entry_multiple(xpi,
+ xprt_switch_find_next_entry_roundrobin);
+}
+
+static
+struct rpc_xprt *xprt_iter_next_entry_all(struct rpc_xprt_iter *xpi)
+{
+ return xprt_iter_next_entry_multiple(xpi, xprt_switch_find_next_entry);
+}
+
+/*
+ * xprt_iter_rewind - Resets the xprt iterator
+ * @xpi: pointer to rpc_xprt_iter
+ *
+ * Resets xpi to ensure that it points to the first entry in the list
+ * of transports.
+ */
+static
+void xprt_iter_rewind(struct rpc_xprt_iter *xpi)
+{
+ rcu_read_lock();
+ xprt_iter_ops(xpi)->xpi_rewind(xpi);
+ rcu_read_unlock();
+}
+
+static void __xprt_iter_init(struct rpc_xprt_iter *xpi,
+ struct rpc_xprt_switch *xps,
+ const struct rpc_xprt_iter_ops *ops)
+{
+ rcu_assign_pointer(xpi->xpi_xpswitch, xprt_switch_get(xps));
+ xpi->xpi_cursor = NULL;
+ xpi->xpi_ops = ops;
+}
+
+/**
+ * xprt_iter_init - Initialise an xprt iterator
+ * @xpi: pointer to rpc_xprt_iter
+ * @xps: pointer to rpc_xprt_switch
+ *
+ * Initialises the iterator to use the default iterator ops
+ * as set in xps. This function is mainly intended for internal
+ * use in the rpc_client.
+ */
+void xprt_iter_init(struct rpc_xprt_iter *xpi,
+ struct rpc_xprt_switch *xps)
+{
+ __xprt_iter_init(xpi, xps, NULL);
+}
+
+/**
+ * xprt_iter_init_listall - Initialise an xprt iterator
+ * @xpi: pointer to rpc_xprt_iter
+ * @xps: pointer to rpc_xprt_switch
+ *
+ * Initialises the iterator to iterate once through the entire list
+ * of entries in xps.
+ */
+void xprt_iter_init_listall(struct rpc_xprt_iter *xpi,
+ struct rpc_xprt_switch *xps)
+{
+ __xprt_iter_init(xpi, xps, &rpc_xprt_iter_listall);
+}
+
+/**
+ * xprt_iter_xchg_switch - Atomically swap out the rpc_xprt_switch
+ * @xpi: pointer to rpc_xprt_iter
+ * @xps: pointer to a new rpc_xprt_switch or NULL
+ *
+ * Swaps out the existing xpi->xpi_xpswitch with a new value.
+ */
+struct rpc_xprt_switch *xprt_iter_xchg_switch(struct rpc_xprt_iter *xpi,
+ struct rpc_xprt_switch *newswitch)
+{
+ struct rpc_xprt_switch __rcu *oldswitch;
+
+ /* Atomically swap out the old xpswitch */
+ oldswitch = xchg(&xpi->xpi_xpswitch, RCU_INITIALIZER(newswitch));
+ if (newswitch != NULL)
+ xprt_iter_rewind(xpi);
+ return rcu_dereference_protected(oldswitch, true);
+}
+
+/**
+ * xprt_iter_destroy - Destroys the xprt iterator
+ * @xpi pointer to rpc_xprt_iter
+ */
+void xprt_iter_destroy(struct rpc_xprt_iter *xpi)
+{
+ xprt_switch_put(xprt_iter_xchg_switch(xpi, NULL));
+}
+
+/**
+ * xprt_iter_xprt - Returns the rpc_xprt pointed to by the cursor
+ * @xpi: pointer to rpc_xprt_iter
+ *
+ * Returns a pointer to the struct rpc_xprt that is currently
+ * pointed to by the cursor.
+ * Caller must be holding rcu_read_lock().
+ */
+struct rpc_xprt *xprt_iter_xprt(struct rpc_xprt_iter *xpi)
+{
+ WARN_ON_ONCE(!rcu_read_lock_held());
+ return xprt_iter_ops(xpi)->xpi_xprt(xpi);
+}
+
+static
+struct rpc_xprt *xprt_iter_get_helper(struct rpc_xprt_iter *xpi,
+ struct rpc_xprt *(*fn)(struct rpc_xprt_iter *))
+{
+ struct rpc_xprt *ret;
+
+ do {
+ ret = fn(xpi);
+ if (ret == NULL)
+ break;
+ ret = xprt_get(ret);
+ } while (ret == NULL);
+ return ret;
+}
+
+/**
+ * xprt_iter_get_xprt - Returns the rpc_xprt pointed to by the cursor
+ * @xpi: pointer to rpc_xprt_iter
+ *
+ * Returns a reference to the struct rpc_xprt that is currently
+ * pointed to by the cursor.
+ */
+struct rpc_xprt *xprt_iter_get_xprt(struct rpc_xprt_iter *xpi)
+{
+ struct rpc_xprt *xprt;
+
+ rcu_read_lock();
+ xprt = xprt_iter_get_helper(xpi, xprt_iter_ops(xpi)->xpi_xprt);
+ rcu_read_unlock();
+ return xprt;
+}
+
+/**
+ * xprt_iter_get_next - Returns the next rpc_xprt following the cursor
+ * @xpi: pointer to rpc_xprt_iter
+ *
+ * Returns a reference to the struct rpc_xprt that immediately follows the
+ * entry pointed to by the cursor.
+ */
+struct rpc_xprt *xprt_iter_get_next(struct rpc_xprt_iter *xpi)
+{
+ struct rpc_xprt *xprt;
+
+ rcu_read_lock();
+ xprt = xprt_iter_get_helper(xpi, xprt_iter_ops(xpi)->xpi_next);
+ rcu_read_unlock();
+ return xprt;
+}
+
+/* Policy for always returning the first entry in the rpc_xprt_switch */
+static
+const struct rpc_xprt_iter_ops rpc_xprt_iter_singular = {
+ .xpi_rewind = xprt_iter_no_rewind,
+ .xpi_xprt = xprt_iter_first_entry,
+ .xpi_next = xprt_iter_first_entry,
+};
+
+/* Policy for round-robin iteration of entries in the rpc_xprt_switch */
+static
+const struct rpc_xprt_iter_ops rpc_xprt_iter_roundrobin = {
+ .xpi_rewind = xprt_iter_default_rewind,
+ .xpi_xprt = xprt_iter_current_entry,
+ .xpi_next = xprt_iter_next_entry_roundrobin,
+};
+
+/* Policy for once-through iteration of entries in the rpc_xprt_switch */
+static
+const struct rpc_xprt_iter_ops rpc_xprt_iter_listall = {
+ .xpi_rewind = xprt_iter_default_rewind,
+ .xpi_xprt = xprt_iter_current_entry,
+ .xpi_next = xprt_iter_next_entry_all,
+};
diff --git a/net/sunrpc/xprtrdma/backchannel.c b/net/sunrpc/xprtrdma/backchannel.c
index cc1251d07297..2dcd7640eeb5 100644
--- a/net/sunrpc/xprtrdma/backchannel.c
+++ b/net/sunrpc/xprtrdma/backchannel.c
@@ -341,6 +341,8 @@ void rpcrdma_bc_receive_call(struct rpcrdma_xprt *r_xprt,
rqst->rq_reply_bytes_recvd = 0;
rqst->rq_bytes_sent = 0;
rqst->rq_xid = headerp->rm_xid;
+
+ rqst->rq_private_buf.len = size;
set_bit(RPC_BC_PA_IN_USE, &rqst->rq_bc_pa_state);
buf = &rqst->rq_rcv_buf;
diff --git a/net/sunrpc/xprtrdma/fmr_ops.c b/net/sunrpc/xprtrdma/fmr_ops.c
index c14f3a4bff68..b289e106540b 100644
--- a/net/sunrpc/xprtrdma/fmr_ops.c
+++ b/net/sunrpc/xprtrdma/fmr_ops.c
@@ -80,13 +80,13 @@ fmr_op_init(struct rpcrdma_xprt *r_xprt)
if (!r)
goto out;
- r->r.fmr.physaddrs = kmalloc(RPCRDMA_MAX_FMR_SGES *
- sizeof(u64), GFP_KERNEL);
- if (!r->r.fmr.physaddrs)
+ r->fmr.physaddrs = kmalloc(RPCRDMA_MAX_FMR_SGES *
+ sizeof(u64), GFP_KERNEL);
+ if (!r->fmr.physaddrs)
goto out_free;
- r->r.fmr.fmr = ib_alloc_fmr(pd, mr_access_flags, &fmr_attr);
- if (IS_ERR(r->r.fmr.fmr))
+ r->fmr.fmr = ib_alloc_fmr(pd, mr_access_flags, &fmr_attr);
+ if (IS_ERR(r->fmr.fmr))
goto out_fmr_err;
list_add(&r->mw_list, &buf->rb_mws);
@@ -95,9 +95,9 @@ fmr_op_init(struct rpcrdma_xprt *r_xprt)
return 0;
out_fmr_err:
- rc = PTR_ERR(r->r.fmr.fmr);
+ rc = PTR_ERR(r->fmr.fmr);
dprintk("RPC: %s: ib_alloc_fmr status %i\n", __func__, rc);
- kfree(r->r.fmr.physaddrs);
+ kfree(r->fmr.physaddrs);
out_free:
kfree(r);
out:
@@ -109,7 +109,7 @@ __fmr_unmap(struct rpcrdma_mw *r)
{
LIST_HEAD(l);
- list_add(&r->r.fmr.fmr->list, &l);
+ list_add(&r->fmr.fmr->list, &l);
return ib_unmap_fmr(&l);
}
@@ -148,7 +148,7 @@ fmr_op_map(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg,
nsegs = RPCRDMA_MAX_FMR_SGES;
for (i = 0; i < nsegs;) {
rpcrdma_map_one(device, seg, direction);
- mw->r.fmr.physaddrs[i] = seg->mr_dma;
+ mw->fmr.physaddrs[i] = seg->mr_dma;
len += seg->mr_len;
++seg;
++i;
@@ -158,13 +158,13 @@ fmr_op_map(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg,
break;
}
- rc = ib_map_phys_fmr(mw->r.fmr.fmr, mw->r.fmr.physaddrs,
+ rc = ib_map_phys_fmr(mw->fmr.fmr, mw->fmr.physaddrs,
i, seg1->mr_dma);
if (rc)
goto out_maperr;
seg1->rl_mw = mw;
- seg1->mr_rkey = mw->r.fmr.fmr->rkey;
+ seg1->mr_rkey = mw->fmr.fmr->rkey;
seg1->mr_base = seg1->mr_dma + pageoff;
seg1->mr_nsegs = i;
seg1->mr_len = len;
@@ -219,7 +219,7 @@ fmr_op_unmap_sync(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
seg = &req->rl_segments[i];
mw = seg->rl_mw;
- list_add(&mw->r.fmr.fmr->list, &unmap_list);
+ list_add(&mw->fmr.fmr->list, &unmap_list);
i += seg->mr_nsegs;
}
@@ -281,9 +281,9 @@ fmr_op_destroy(struct rpcrdma_buffer *buf)
while (!list_empty(&buf->rb_all)) {
r = list_entry(buf->rb_all.next, struct rpcrdma_mw, mw_all);
list_del(&r->mw_all);
- kfree(r->r.fmr.physaddrs);
+ kfree(r->fmr.physaddrs);
- rc = ib_dealloc_fmr(r->r.fmr.fmr);
+ rc = ib_dealloc_fmr(r->fmr.fmr);
if (rc)
dprintk("RPC: %s: ib_dealloc_fmr failed %i\n",
__func__, rc);
diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index e16567389e28..c250924a9fd3 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -109,20 +109,20 @@ static void
__frwr_recovery_worker(struct work_struct *work)
{
struct rpcrdma_mw *r = container_of(work, struct rpcrdma_mw,
- r.frmr.fr_work);
- struct rpcrdma_xprt *r_xprt = r->r.frmr.fr_xprt;
+ frmr.fr_work);
+ struct rpcrdma_xprt *r_xprt = r->frmr.fr_xprt;
unsigned int depth = r_xprt->rx_ia.ri_max_frmr_depth;
struct ib_pd *pd = r_xprt->rx_ia.ri_pd;
- if (ib_dereg_mr(r->r.frmr.fr_mr))
+ if (ib_dereg_mr(r->frmr.fr_mr))
goto out_fail;
- r->r.frmr.fr_mr = ib_alloc_mr(pd, IB_MR_TYPE_MEM_REG, depth);
- if (IS_ERR(r->r.frmr.fr_mr))
+ r->frmr.fr_mr = ib_alloc_mr(pd, IB_MR_TYPE_MEM_REG, depth);
+ if (IS_ERR(r->frmr.fr_mr))
goto out_fail;
dprintk("RPC: %s: recovered FRMR %p\n", __func__, r);
- r->r.frmr.fr_state = FRMR_IS_INVALID;
+ r->frmr.fr_state = FRMR_IS_INVALID;
rpcrdma_put_mw(r_xprt, r);
return;
@@ -137,15 +137,15 @@ out_fail:
static void
__frwr_queue_recovery(struct rpcrdma_mw *r)
{
- INIT_WORK(&r->r.frmr.fr_work, __frwr_recovery_worker);
- queue_work(frwr_recovery_wq, &r->r.frmr.fr_work);
+ INIT_WORK(&r->frmr.fr_work, __frwr_recovery_worker);
+ queue_work(frwr_recovery_wq, &r->frmr.fr_work);
}
static int
__frwr_init(struct rpcrdma_mw *r, struct ib_pd *pd, struct ib_device *device,
unsigned int depth)
{
- struct rpcrdma_frmr *f = &r->r.frmr;
+ struct rpcrdma_frmr *f = &r->frmr;
int rc;
f->fr_mr = ib_alloc_mr(pd, IB_MR_TYPE_MEM_REG, depth);
@@ -158,6 +158,8 @@ __frwr_init(struct rpcrdma_mw *r, struct ib_pd *pd, struct ib_device *device,
sg_init_table(f->sg, depth);
+ init_completion(&f->fr_linv_done);
+
return 0;
out_mr_err:
@@ -179,11 +181,11 @@ __frwr_release(struct rpcrdma_mw *r)
{
int rc;
- rc = ib_dereg_mr(r->r.frmr.fr_mr);
+ rc = ib_dereg_mr(r->frmr.fr_mr);
if (rc)
dprintk("RPC: %s: ib_dereg_mr status %i\n",
__func__, rc);
- kfree(r->r.frmr.sg);
+ kfree(r->frmr.sg);
}
static int
@@ -244,39 +246,76 @@ frwr_op_maxpages(struct rpcrdma_xprt *r_xprt)
rpcrdma_max_segments(r_xprt) * ia->ri_max_frmr_depth);
}
-/* If FAST_REG or LOCAL_INV failed, indicate the frmr needs
- * to be reset.
+static void
+__frwr_sendcompletion_flush(struct ib_wc *wc, struct rpcrdma_frmr *frmr,
+ const char *wr)
+{
+ frmr->fr_state = FRMR_IS_STALE;
+ if (wc->status != IB_WC_WR_FLUSH_ERR)
+ pr_err("rpcrdma: %s: %s (%u/0x%x)\n",
+ wr, ib_wc_status_msg(wc->status),
+ wc->status, wc->vendor_err);
+}
+
+/**
+ * frwr_wc_fastreg - Invoked by RDMA provider for each polled FastReg WC
+ * @cq: completion queue (ignored)
+ * @wc: completed WR
*
- * WARNING: Only wr_id and status are reliable at this point
*/
static void
-__frwr_sendcompletion_flush(struct ib_wc *wc, struct rpcrdma_mw *r)
+frwr_wc_fastreg(struct ib_cq *cq, struct ib_wc *wc)
{
- if (likely(wc->status == IB_WC_SUCCESS))
- return;
-
- /* WARNING: Only wr_id and status are reliable at this point */
- r = (struct rpcrdma_mw *)(unsigned long)wc->wr_id;
- if (wc->status == IB_WC_WR_FLUSH_ERR)
- dprintk("RPC: %s: frmr %p flushed\n", __func__, r);
- else
- pr_warn("RPC: %s: frmr %p error, status %s (%d)\n",
- __func__, r, ib_wc_status_msg(wc->status), wc->status);
+ struct rpcrdma_frmr *frmr;
+ struct ib_cqe *cqe;
- r->r.frmr.fr_state = FRMR_IS_STALE;
+ /* WARNING: Only wr_cqe and status are reliable at this point */
+ if (wc->status != IB_WC_SUCCESS) {
+ cqe = wc->wr_cqe;
+ frmr = container_of(cqe, struct rpcrdma_frmr, fr_cqe);
+ __frwr_sendcompletion_flush(wc, frmr, "fastreg");
+ }
}
+/**
+ * frwr_wc_localinv - Invoked by RDMA provider for each polled LocalInv WC
+ * @cq: completion queue (ignored)
+ * @wc: completed WR
+ *
+ */
static void
-frwr_sendcompletion(struct ib_wc *wc)
+frwr_wc_localinv(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rpcrdma_mw *r = (struct rpcrdma_mw *)(unsigned long)wc->wr_id;
- struct rpcrdma_frmr *f = &r->r.frmr;
+ struct rpcrdma_frmr *frmr;
+ struct ib_cqe *cqe;
- if (unlikely(wc->status != IB_WC_SUCCESS))
- __frwr_sendcompletion_flush(wc, r);
+ /* WARNING: Only wr_cqe and status are reliable at this point */
+ if (wc->status != IB_WC_SUCCESS) {
+ cqe = wc->wr_cqe;
+ frmr = container_of(cqe, struct rpcrdma_frmr, fr_cqe);
+ __frwr_sendcompletion_flush(wc, frmr, "localinv");
+ }
+}
- if (f->fr_waiter)
- complete(&f->fr_linv_done);
+/**
+ * frwr_wc_localinv - Invoked by RDMA provider for each polled LocalInv WC
+ * @cq: completion queue (ignored)
+ * @wc: completed WR
+ *
+ * Awaken anyone waiting for an MR to finish being fenced.
+ */
+static void
+frwr_wc_localinv_wake(struct ib_cq *cq, struct ib_wc *wc)
+{
+ struct rpcrdma_frmr *frmr;
+ struct ib_cqe *cqe;
+
+ /* WARNING: Only wr_cqe and status are reliable at this point */
+ cqe = wc->wr_cqe;
+ frmr = container_of(cqe, struct rpcrdma_frmr, fr_cqe);
+ if (wc->status != IB_WC_SUCCESS)
+ __frwr_sendcompletion_flush(wc, frmr, "localinv");
+ complete_all(&frmr->fr_linv_done);
}
static int
@@ -313,8 +352,7 @@ frwr_op_init(struct rpcrdma_xprt *r_xprt)
list_add(&r->mw_list, &buf->rb_mws);
list_add(&r->mw_all, &buf->rb_all);
- r->mw_sendcompletion = frwr_sendcompletion;
- r->r.frmr.fr_xprt = r_xprt;
+ r->frmr.fr_xprt = r_xprt;
}
return 0;
@@ -347,10 +385,9 @@ frwr_op_map(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg,
mw = rpcrdma_get_mw(r_xprt);
if (!mw)
return -ENOMEM;
- } while (mw->r.frmr.fr_state != FRMR_IS_INVALID);
- frmr = &mw->r.frmr;
+ } while (mw->frmr.fr_state != FRMR_IS_INVALID);
+ frmr = &mw->frmr;
frmr->fr_state = FRMR_IS_VALID;
- frmr->fr_waiter = false;
mr = frmr->fr_mr;
reg_wr = &frmr->fr_regwr;
@@ -400,7 +437,8 @@ frwr_op_map(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg,
reg_wr->wr.next = NULL;
reg_wr->wr.opcode = IB_WR_REG_MR;
- reg_wr->wr.wr_id = (uintptr_t)mw;
+ frmr->fr_cqe.done = frwr_wc_fastreg;
+ reg_wr->wr.wr_cqe = &frmr->fr_cqe;
reg_wr->wr.num_sge = 0;
reg_wr->wr.send_flags = 0;
reg_wr->mr = mr;
@@ -434,15 +472,15 @@ static struct ib_send_wr *
__frwr_prepare_linv_wr(struct rpcrdma_mr_seg *seg)
{
struct rpcrdma_mw *mw = seg->rl_mw;
- struct rpcrdma_frmr *f = &mw->r.frmr;
+ struct rpcrdma_frmr *f = &mw->frmr;
struct ib_send_wr *invalidate_wr;
- f->fr_waiter = false;
f->fr_state = FRMR_IS_INVALID;
invalidate_wr = &f->fr_invwr;
memset(invalidate_wr, 0, sizeof(*invalidate_wr));
- invalidate_wr->wr_id = (unsigned long)(void *)mw;
+ f->fr_cqe.done = frwr_wc_localinv;
+ invalidate_wr->wr_cqe = &f->fr_cqe;
invalidate_wr->opcode = IB_WR_LOCAL_INV;
invalidate_wr->ex.invalidate_rkey = f->fr_mr->rkey;
@@ -455,7 +493,7 @@ __frwr_dma_unmap(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg,
{
struct ib_device *device = r_xprt->rx_ia.ri_device;
struct rpcrdma_mw *mw = seg->rl_mw;
- struct rpcrdma_frmr *f = &mw->r.frmr;
+ struct rpcrdma_frmr *f = &mw->frmr;
seg->rl_mw = NULL;
@@ -504,15 +542,15 @@ frwr_op_unmap_sync(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
i += seg->mr_nsegs;
}
- f = &seg->rl_mw->r.frmr;
+ f = &seg->rl_mw->frmr;
/* Strong send queue ordering guarantees that when the
* last WR in the chain completes, all WRs in the chain
* are complete.
*/
f->fr_invwr.send_flags = IB_SEND_SIGNALED;
- f->fr_waiter = true;
- init_completion(&f->fr_linv_done);
+ f->fr_cqe.done = frwr_wc_localinv_wake;
+ reinit_completion(&f->fr_linv_done);
INIT_CQCOUNT(&r_xprt->rx_ep);
/* Transport disconnect drains the receive CQ before it
@@ -520,14 +558,18 @@ frwr_op_unmap_sync(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
* unless ri_id->qp is a valid pointer.
*/
rc = ib_post_send(ia->ri_id->qp, invalidate_wrs, &bad_wr);
- if (rc)
+ if (rc) {
pr_warn("%s: ib_post_send failed %i\n", __func__, rc);
+ rdma_disconnect(ia->ri_id);
+ goto unmap;
+ }
wait_for_completion(&f->fr_linv_done);
/* ORDER: Now DMA unmap all of the req's MRs, and return
* them to the free MW list.
*/
+unmap:
for (i = 0, nchunks = req->rl_nchunks; nchunks; nchunks--) {
seg = &req->rl_segments[i];
@@ -549,7 +591,7 @@ frwr_op_unmap(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg)
struct rpcrdma_mr_seg *seg1 = seg;
struct rpcrdma_ia *ia = &r_xprt->rx_ia;
struct rpcrdma_mw *mw = seg1->rl_mw;
- struct rpcrdma_frmr *frmr = &mw->r.frmr;
+ struct rpcrdma_frmr *frmr = &mw->frmr;
struct ib_send_wr *invalidate_wr, *bad_wr;
int rc, nsegs = seg->mr_nsegs;
@@ -557,10 +599,11 @@ frwr_op_unmap(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg)
seg1->rl_mw = NULL;
frmr->fr_state = FRMR_IS_INVALID;
- invalidate_wr = &mw->r.frmr.fr_invwr;
+ invalidate_wr = &mw->frmr.fr_invwr;
memset(invalidate_wr, 0, sizeof(*invalidate_wr));
- invalidate_wr->wr_id = (uintptr_t)mw;
+ frmr->fr_cqe.done = frwr_wc_localinv;
+ invalidate_wr->wr_cqe = &frmr->fr_cqe;
invalidate_wr->opcode = IB_WR_LOCAL_INV;
invalidate_wr->ex.invalidate_rkey = frmr->fr_mr->rkey;
DECR_CQCOUNT(&r_xprt->rx_ep);
diff --git a/net/sunrpc/xprtrdma/physical_ops.c b/net/sunrpc/xprtrdma/physical_ops.c
index dbb302ecf590..481b9b6f4a15 100644
--- a/net/sunrpc/xprtrdma/physical_ops.c
+++ b/net/sunrpc/xprtrdma/physical_ops.c
@@ -68,7 +68,6 @@ physical_op_map(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr_seg *seg,
rpcrdma_map_one(ia->ri_device, seg, rpcrdma_data_dir(writing));
seg->mr_rkey = ia->ri_dma_mr->rkey;
seg->mr_base = seg->mr_dma;
- seg->mr_nsegs = 1;
return 1;
}
diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c
index 0f28f2d743ed..888823bb6dae 100644
--- a/net/sunrpc/xprtrdma/rpc_rdma.c
+++ b/net/sunrpc/xprtrdma/rpc_rdma.c
@@ -132,6 +132,33 @@ rpcrdma_tail_pullup(struct xdr_buf *buf)
return tlen;
}
+/* Split "vec" on page boundaries into segments. FMR registers pages,
+ * not a byte range. Other modes coalesce these segments into a single
+ * MR when they can.
+ */
+static int
+rpcrdma_convert_kvec(struct kvec *vec, struct rpcrdma_mr_seg *seg,
+ int n, int nsegs)
+{
+ size_t page_offset;
+ u32 remaining;
+ char *base;
+
+ base = vec->iov_base;
+ page_offset = offset_in_page(base);
+ remaining = vec->iov_len;
+ while (remaining && n < nsegs) {
+ seg[n].mr_page = NULL;
+ seg[n].mr_offset = base;
+ seg[n].mr_len = min_t(u32, PAGE_SIZE - page_offset, remaining);
+ remaining -= seg[n].mr_len;
+ base += seg[n].mr_len;
+ ++n;
+ page_offset = 0;
+ }
+ return n;
+}
+
/*
* Chunk assembly from upper layer xdr_buf.
*
@@ -150,11 +177,10 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos,
int page_base;
struct page **ppages;
- if (pos == 0 && xdrbuf->head[0].iov_len) {
- seg[n].mr_page = NULL;
- seg[n].mr_offset = xdrbuf->head[0].iov_base;
- seg[n].mr_len = xdrbuf->head[0].iov_len;
- ++n;
+ if (pos == 0) {
+ n = rpcrdma_convert_kvec(&xdrbuf->head[0], seg, n, nsegs);
+ if (n == nsegs)
+ return -EIO;
}
len = xdrbuf->page_len;
@@ -192,13 +218,9 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos,
* xdr pad bytes, saving the server an RDMA operation. */
if (xdrbuf->tail[0].iov_len < 4 && xprt_rdma_pad_optimize)
return n;
+ n = rpcrdma_convert_kvec(&xdrbuf->tail[0], seg, n, nsegs);
if (n == nsegs)
- /* Tail remains, but we're out of segments */
return -EIO;
- seg[n].mr_page = NULL;
- seg[n].mr_offset = xdrbuf->tail[0].iov_base;
- seg[n].mr_len = xdrbuf->tail[0].iov_len;
- ++n;
}
return n;
@@ -773,20 +795,17 @@ rpcrdma_reply_handler(struct rpcrdma_rep *rep)
struct rpcrdma_xprt *r_xprt = rep->rr_rxprt;
struct rpc_xprt *xprt = &r_xprt->rx_xprt;
__be32 *iptr;
- int rdmalen, status;
+ int rdmalen, status, rmerr;
unsigned long cwnd;
- u32 credits;
dprintk("RPC: %s: incoming rep %p\n", __func__, rep);
if (rep->rr_len == RPCRDMA_BAD_LEN)
goto out_badstatus;
- if (rep->rr_len < RPCRDMA_HDRLEN_MIN)
+ if (rep->rr_len < RPCRDMA_HDRLEN_ERR)
goto out_shortreply;
headerp = rdmab_to_msg(rep->rr_rdmabuf);
- if (headerp->rm_vers != rpcrdma_version)
- goto out_badversion;
#if defined(CONFIG_SUNRPC_BACKCHANNEL)
if (rpcrdma_is_bcall(headerp))
goto out_bcall;
@@ -809,15 +828,16 @@ rpcrdma_reply_handler(struct rpcrdma_rep *rep)
*/
list_del_init(&rqst->rq_list);
spin_unlock_bh(&xprt->transport_lock);
- dprintk("RPC: %s: reply 0x%p completes request 0x%p\n"
- " RPC request 0x%p xid 0x%08x\n",
- __func__, rep, req, rqst,
- be32_to_cpu(headerp->rm_xid));
+ dprintk("RPC: %s: reply %p completes request %p (xid 0x%08x)\n",
+ __func__, rep, req, be32_to_cpu(headerp->rm_xid));
/* from here on, the reply is no longer an orphan */
req->rl_reply = rep;
xprt->reestablish_timeout = 0;
+ if (headerp->rm_vers != rpcrdma_version)
+ goto out_badversion;
+
/* check for expected message types */
/* The order of some of these tests is important. */
switch (headerp->rm_type) {
@@ -878,6 +898,9 @@ rpcrdma_reply_handler(struct rpcrdma_rep *rep)
status = rdmalen;
break;
+ case rdma_error:
+ goto out_rdmaerr;
+
badheader:
default:
dprintk("%s: invalid rpcrdma reply header (type %d):"
@@ -893,6 +916,7 @@ badheader:
break;
}
+out:
/* Invalidate and flush the data payloads before waking the
* waiting application. This guarantees the memory region is
* properly fenced from the server before the application
@@ -903,15 +927,9 @@ badheader:
if (req->rl_nchunks)
r_xprt->rx_ia.ri_ops->ro_unmap_sync(r_xprt, req);
- credits = be32_to_cpu(headerp->rm_credit);
- if (credits == 0)
- credits = 1; /* don't deadlock */
- else if (credits > r_xprt->rx_buf.rb_max_requests)
- credits = r_xprt->rx_buf.rb_max_requests;
-
spin_lock_bh(&xprt->transport_lock);
cwnd = xprt->cwnd;
- xprt->cwnd = credits << RPC_CWNDSHIFT;
+ xprt->cwnd = atomic_read(&r_xprt->rx_buf.rb_credits) << RPC_CWNDSHIFT;
if (xprt->cwnd > cwnd)
xprt_release_rqst_cong(rqst->rq_task);
@@ -935,13 +953,43 @@ out_bcall:
return;
#endif
-out_shortreply:
- dprintk("RPC: %s: short/invalid reply\n", __func__);
- goto repost;
-
+/* If the incoming reply terminated a pending RPC, the next
+ * RPC call will post a replacement receive buffer as it is
+ * being marshaled.
+ */
out_badversion:
dprintk("RPC: %s: invalid version %d\n",
__func__, be32_to_cpu(headerp->rm_vers));
+ status = -EIO;
+ r_xprt->rx_stats.bad_reply_count++;
+ goto out;
+
+out_rdmaerr:
+ rmerr = be32_to_cpu(headerp->rm_body.rm_error.rm_err);
+ switch (rmerr) {
+ case ERR_VERS:
+ pr_err("%s: server reports header version error (%u-%u)\n",
+ __func__,
+ be32_to_cpu(headerp->rm_body.rm_error.rm_vers_low),
+ be32_to_cpu(headerp->rm_body.rm_error.rm_vers_high));
+ break;
+ case ERR_CHUNK:
+ pr_err("%s: server reports header decoding error\n",
+ __func__);
+ break;
+ default:
+ pr_err("%s: server reports unknown error %d\n",
+ __func__, rmerr);
+ }
+ status = -EREMOTEIO;
+ r_xprt->rx_stats.bad_reply_count++;
+ goto out;
+
+/* If no pending RPC transaction was matched, post a replacement
+ * receive buffer before returning.
+ */
+out_shortreply:
+ dprintk("RPC: %s: short/invalid reply\n", __func__);
goto repost;
out_nomatch:
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 878f1bfb1db9..f5ed9f982cd7 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -112,89 +112,65 @@ rpcrdma_qp_async_error_upcall(struct ib_event *event, void *context)
}
}
+/**
+ * rpcrdma_wc_send - Invoked by RDMA provider for each polled Send WC
+ * @cq: completion queue (ignored)
+ * @wc: completed WR
+ *
+ */
static void
-rpcrdma_cq_async_error_upcall(struct ib_event *event, void *context)
+rpcrdma_wc_send(struct ib_cq *cq, struct ib_wc *wc)
{
- struct rpcrdma_ep *ep = context;
-
- pr_err("RPC: %s: %s on device %s ep %p\n",
- __func__, ib_event_msg(event->event),
- event->device->name, context);
- if (ep->rep_connected == 1) {
- ep->rep_connected = -EIO;
- rpcrdma_conn_func(ep);
- wake_up_all(&ep->rep_connect_wait);
- }
+ /* WARNING: Only wr_cqe and status are reliable at this point */
+ if (wc->status != IB_WC_SUCCESS && wc->status != IB_WC_WR_FLUSH_ERR)
+ pr_err("rpcrdma: Send: %s (%u/0x%x)\n",
+ ib_wc_status_msg(wc->status),
+ wc->status, wc->vendor_err);
}
static void
-rpcrdma_sendcq_process_wc(struct ib_wc *wc)
+rpcrdma_receive_worker(struct work_struct *work)
{
- /* WARNING: Only wr_id and status are reliable at this point */
- if (wc->wr_id == RPCRDMA_IGNORE_COMPLETION) {
- if (wc->status != IB_WC_SUCCESS &&
- wc->status != IB_WC_WR_FLUSH_ERR)
- pr_err("RPC: %s: SEND: %s\n",
- __func__, ib_wc_status_msg(wc->status));
- } else {
- struct rpcrdma_mw *r;
+ struct rpcrdma_rep *rep =
+ container_of(work, struct rpcrdma_rep, rr_work);
- r = (struct rpcrdma_mw *)(unsigned long)wc->wr_id;
- r->mw_sendcompletion(wc);
- }
+ rpcrdma_reply_handler(rep);
}
-/* The common case is a single send completion is waiting. By
- * passing two WC entries to ib_poll_cq, a return code of 1
- * means there is exactly one WC waiting and no more. We don't
- * have to invoke ib_poll_cq again to know that the CQ has been
- * properly drained.
+/* Perform basic sanity checking to avoid using garbage
+ * to update the credit grant value.
*/
static void
-rpcrdma_sendcq_poll(struct ib_cq *cq)
+rpcrdma_update_granted_credits(struct rpcrdma_rep *rep)
{
- struct ib_wc *pos, wcs[2];
- int count, rc;
+ struct rpcrdma_msg *rmsgp = rdmab_to_msg(rep->rr_rdmabuf);
+ struct rpcrdma_buffer *buffer = &rep->rr_rxprt->rx_buf;
+ u32 credits;
- do {
- pos = wcs;
+ if (rep->rr_len < RPCRDMA_HDRLEN_ERR)
+ return;
- rc = ib_poll_cq(cq, ARRAY_SIZE(wcs), pos);
- if (rc < 0)
- break;
+ credits = be32_to_cpu(rmsgp->rm_credit);
+ if (credits == 0)
+ credits = 1; /* don't deadlock */
+ else if (credits > buffer->rb_max_requests)
+ credits = buffer->rb_max_requests;
- count = rc;
- while (count-- > 0)
- rpcrdma_sendcq_process_wc(pos++);
- } while (rc == ARRAY_SIZE(wcs));
- return;
+ atomic_set(&buffer->rb_credits, credits);
}
-/* Handle provider send completion upcalls.
+/**
+ * rpcrdma_receive_wc - Invoked by RDMA provider for each polled Receive WC
+ * @cq: completion queue (ignored)
+ * @wc: completed WR
+ *
*/
static void
-rpcrdma_sendcq_upcall(struct ib_cq *cq, void *cq_context)
+rpcrdma_receive_wc(struct ib_cq *cq, struct ib_wc *wc)
{
- do {
- rpcrdma_sendcq_poll(cq);
- } while (ib_req_notify_cq(cq, IB_CQ_NEXT_COMP |
- IB_CQ_REPORT_MISSED_EVENTS) > 0);
-}
-
-static void
-rpcrdma_receive_worker(struct work_struct *work)
-{
- struct rpcrdma_rep *rep =
- container_of(work, struct rpcrdma_rep, rr_work);
-
- rpcrdma_reply_handler(rep);
-}
-
-static void
-rpcrdma_recvcq_process_wc(struct ib_wc *wc)
-{
- struct rpcrdma_rep *rep =
- (struct rpcrdma_rep *)(unsigned long)wc->wr_id;
+ struct ib_cqe *cqe = wc->wr_cqe;
+ struct rpcrdma_rep *rep = container_of(cqe, struct rpcrdma_rep,
+ rr_cqe);
/* WARNING: Only wr_id and status are reliable at this point */
if (wc->status != IB_WC_SUCCESS)
@@ -211,7 +187,8 @@ rpcrdma_recvcq_process_wc(struct ib_wc *wc)
ib_dma_sync_single_for_cpu(rep->rr_device,
rdmab_addr(rep->rr_rdmabuf),
rep->rr_len, DMA_FROM_DEVICE);
- prefetch(rdmab_to_msg(rep->rr_rdmabuf));
+
+ rpcrdma_update_granted_credits(rep);
out_schedule:
queue_work(rpcrdma_receive_wq, &rep->rr_work);
@@ -219,57 +196,20 @@ out_schedule:
out_fail:
if (wc->status != IB_WC_WR_FLUSH_ERR)
- pr_err("RPC: %s: rep %p: %s\n",
- __func__, rep, ib_wc_status_msg(wc->status));
+ pr_err("rpcrdma: Recv: %s (%u/0x%x)\n",
+ ib_wc_status_msg(wc->status),
+ wc->status, wc->vendor_err);
rep->rr_len = RPCRDMA_BAD_LEN;
goto out_schedule;
}
-/* The wc array is on stack: automatic memory is always CPU-local.
- *
- * struct ib_wc is 64 bytes, making the poll array potentially
- * large. But this is at the bottom of the call chain. Further
- * substantial work is done in another thread.
- */
-static void
-rpcrdma_recvcq_poll(struct ib_cq *cq)
-{
- struct ib_wc *pos, wcs[4];
- int count, rc;
-
- do {
- pos = wcs;
-
- rc = ib_poll_cq(cq, ARRAY_SIZE(wcs), pos);
- if (rc < 0)
- break;
-
- count = rc;
- while (count-- > 0)
- rpcrdma_recvcq_process_wc(pos++);
- } while (rc == ARRAY_SIZE(wcs));
-}
-
-/* Handle provider receive completion upcalls.
- */
-static void
-rpcrdma_recvcq_upcall(struct ib_cq *cq, void *cq_context)
-{
- do {
- rpcrdma_recvcq_poll(cq);
- } while (ib_req_notify_cq(cq, IB_CQ_NEXT_COMP |
- IB_CQ_REPORT_MISSED_EVENTS) > 0);
-}
-
static void
rpcrdma_flush_cqs(struct rpcrdma_ep *ep)
{
struct ib_wc wc;
while (ib_poll_cq(ep->rep_attr.recv_cq, 1, &wc) > 0)
- rpcrdma_recvcq_process_wc(&wc);
- while (ib_poll_cq(ep->rep_attr.send_cq, 1, &wc) > 0)
- rpcrdma_sendcq_process_wc(&wc);
+ rpcrdma_receive_wc(NULL, &wc);
}
static int
@@ -330,6 +270,7 @@ rpcrdma_conn_upcall(struct rdma_cm_id *id, struct rdma_cm_event *event)
connected:
dprintk("RPC: %s: %sconnected\n",
__func__, connstate > 0 ? "" : "dis");
+ atomic_set(&xprt->rx_buf.rb_credits, 1);
ep->rep_connected = connstate;
rpcrdma_conn_func(ep);
wake_up_all(&ep->rep_connect_wait);
@@ -560,9 +501,8 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia,
struct rpcrdma_create_data_internal *cdata)
{
struct ib_cq *sendcq, *recvcq;
- struct ib_cq_init_attr cq_attr = {};
unsigned int max_qp_wr;
- int rc, err;
+ int rc;
if (ia->ri_device->attrs.max_sge < RPCRDMA_MAX_IOVS) {
dprintk("RPC: %s: insufficient sge's available\n",
@@ -614,9 +554,9 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia,
init_waitqueue_head(&ep->rep_connect_wait);
INIT_DELAYED_WORK(&ep->rep_connect_worker, rpcrdma_connect_worker);
- cq_attr.cqe = ep->rep_attr.cap.max_send_wr + 1;
- sendcq = ib_create_cq(ia->ri_device, rpcrdma_sendcq_upcall,
- rpcrdma_cq_async_error_upcall, NULL, &cq_attr);
+ sendcq = ib_alloc_cq(ia->ri_device, NULL,
+ ep->rep_attr.cap.max_send_wr + 1,
+ 0, IB_POLL_SOFTIRQ);
if (IS_ERR(sendcq)) {
rc = PTR_ERR(sendcq);
dprintk("RPC: %s: failed to create send CQ: %i\n",
@@ -624,16 +564,9 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia,
goto out1;
}
- rc = ib_req_notify_cq(sendcq, IB_CQ_NEXT_COMP);
- if (rc) {
- dprintk("RPC: %s: ib_req_notify_cq failed: %i\n",
- __func__, rc);
- goto out2;
- }
-
- cq_attr.cqe = ep->rep_attr.cap.max_recv_wr + 1;
- recvcq = ib_create_cq(ia->ri_device, rpcrdma_recvcq_upcall,
- rpcrdma_cq_async_error_upcall, NULL, &cq_attr);
+ recvcq = ib_alloc_cq(ia->ri_device, NULL,
+ ep->rep_attr.cap.max_recv_wr + 1,
+ 0, IB_POLL_SOFTIRQ);
if (IS_ERR(recvcq)) {
rc = PTR_ERR(recvcq);
dprintk("RPC: %s: failed to create recv CQ: %i\n",
@@ -641,14 +574,6 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia,
goto out2;
}
- rc = ib_req_notify_cq(recvcq, IB_CQ_NEXT_COMP);
- if (rc) {
- dprintk("RPC: %s: ib_req_notify_cq failed: %i\n",
- __func__, rc);
- ib_destroy_cq(recvcq);
- goto out2;
- }
-
ep->rep_attr.send_cq = sendcq;
ep->rep_attr.recv_cq = recvcq;
@@ -673,10 +598,7 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia,
return 0;
out2:
- err = ib_destroy_cq(sendcq);
- if (err)
- dprintk("RPC: %s: ib_destroy_cq returned %i\n",
- __func__, err);
+ ib_free_cq(sendcq);
out1:
if (ia->ri_dma_mr)
ib_dereg_mr(ia->ri_dma_mr);
@@ -711,15 +633,8 @@ rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia)
ia->ri_id->qp = NULL;
}
- rc = ib_destroy_cq(ep->rep_attr.recv_cq);
- if (rc)
- dprintk("RPC: %s: ib_destroy_cq returned %i\n",
- __func__, rc);
-
- rc = ib_destroy_cq(ep->rep_attr.send_cq);
- if (rc)
- dprintk("RPC: %s: ib_destroy_cq returned %i\n",
- __func__, rc);
+ ib_free_cq(ep->rep_attr.recv_cq);
+ ib_free_cq(ep->rep_attr.send_cq);
if (ia->ri_dma_mr) {
rc = ib_dereg_mr(ia->ri_dma_mr);
@@ -898,6 +813,7 @@ rpcrdma_create_req(struct rpcrdma_xprt *r_xprt)
spin_lock(&buffer->rb_reqslock);
list_add(&req->rl_all, &buffer->rb_allreqs);
spin_unlock(&buffer->rb_reqslock);
+ req->rl_cqe.done = rpcrdma_wc_send;
req->rl_buffer = &r_xprt->rx_buf;
return req;
}
@@ -923,6 +839,7 @@ rpcrdma_create_rep(struct rpcrdma_xprt *r_xprt)
}
rep->rr_device = ia->ri_device;
+ rep->rr_cqe.done = rpcrdma_receive_wc;
rep->rr_rxprt = r_xprt;
INIT_WORK(&rep->rr_work, rpcrdma_receive_worker);
return rep;
@@ -943,6 +860,7 @@ rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt)
buf->rb_max_requests = r_xprt->rx_data.max_requests;
buf->rb_bc_srv_max_requests = 0;
spin_lock_init(&buf->rb_lock);
+ atomic_set(&buf->rb_credits, 1);
rc = ia->ri_ops->ro_init(r_xprt);
if (rc)
@@ -1259,7 +1177,7 @@ rpcrdma_ep_post(struct rpcrdma_ia *ia,
}
send_wr.next = NULL;
- send_wr.wr_id = RPCRDMA_IGNORE_COMPLETION;
+ send_wr.wr_cqe = &req->rl_cqe;
send_wr.sg_list = iov;
send_wr.num_sge = req->rl_niovs;
send_wr.opcode = IB_WR_SEND;
@@ -1297,7 +1215,7 @@ rpcrdma_ep_post_recv(struct rpcrdma_ia *ia,
int rc;
recv_wr.next = NULL;
- recv_wr.wr_id = (u64) (unsigned long) rep;
+ recv_wr.wr_cqe = &rep->rr_cqe;
recv_wr.sg_list = &rep->rr_rdmabuf->rg_iov;
recv_wr.num_sge = 1;
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index 38fe11b09875..2ebc743cb96f 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -95,10 +95,6 @@ struct rpcrdma_ep {
#define INIT_CQCOUNT(ep) atomic_set(&(ep)->rep_cqcount, (ep)->rep_cqinit)
#define DECR_CQCOUNT(ep) atomic_sub_return(1, &(ep)->rep_cqcount)
-/* Force completion handler to ignore the signal
- */
-#define RPCRDMA_IGNORE_COMPLETION (0ULL)
-
/* Pre-allocate extra Work Requests for handling backward receives
* and sends. This is a fixed value because the Work Queues are
* allocated when the forward channel is set up.
@@ -171,6 +167,7 @@ rdmab_to_msg(struct rpcrdma_regbuf *rb)
struct rpcrdma_buffer;
struct rpcrdma_rep {
+ struct ib_cqe rr_cqe;
unsigned int rr_len;
struct ib_device *rr_device;
struct rpcrdma_xprt *rr_rxprt;
@@ -204,11 +201,11 @@ struct rpcrdma_frmr {
struct scatterlist *sg;
int sg_nents;
struct ib_mr *fr_mr;
+ struct ib_cqe fr_cqe;
enum rpcrdma_frmr_state fr_state;
+ struct completion fr_linv_done;
struct work_struct fr_work;
struct rpcrdma_xprt *fr_xprt;
- bool fr_waiter;
- struct completion fr_linv_done;;
union {
struct ib_reg_wr fr_regwr;
struct ib_send_wr fr_invwr;
@@ -224,8 +221,7 @@ struct rpcrdma_mw {
union {
struct rpcrdma_fmr fmr;
struct rpcrdma_frmr frmr;
- } r;
- void (*mw_sendcompletion)(struct ib_wc *);
+ };
struct list_head mw_list;
struct list_head mw_all;
};
@@ -281,6 +277,7 @@ struct rpcrdma_req {
struct rpcrdma_regbuf *rl_sendbuf;
struct rpcrdma_mr_seg rl_segments[RPCRDMA_MAX_SEGS];
+ struct ib_cqe rl_cqe;
struct list_head rl_all;
bool rl_backchannel;
};
@@ -311,6 +308,7 @@ struct rpcrdma_buffer {
struct list_head rb_send_bufs;
struct list_head rb_recv_bufs;
u32 rb_max_requests;
+ atomic_t rb_credits; /* most recent credit grant */
u32 rb_bc_srv_max_requests;
spinlock_t rb_reqslock; /* protect rb_allreqs */
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index fde2138b81e7..65e759569e48 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -1844,9 +1844,7 @@ static int xs_bind(struct sock_xprt *transport, struct socket *sock)
*/
static void xs_local_rpcbind(struct rpc_task *task)
{
- rcu_read_lock();
- xprt_set_bound(rcu_dereference(task->tk_client->cl_xprt));
- rcu_read_unlock();
+ xprt_set_bound(task->tk_xprt);
}
static void xs_local_set_port(struct rpc_xprt *xprt, unsigned short port)
generated by cgit v1.2.3 (git 2.39.1) at 2024-09-28 23:26:44 +0000