summaryrefslogtreecommitdiff
path: root/net/sunrpc
diff options
context:
space:
mode:
Diffstat (limited to 'net/sunrpc')
-rw-r--r--net/sunrpc/auth_gss/Makefile4
-rw-r--r--net/sunrpc/auth_gss/auth_gss.c29
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_crypto.c10
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_mech.c4
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_seal.c26
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_unseal.c16
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_wrap.c72
-rw-r--r--net/sunrpc/auth_gss/gss_spkm3_mech.c4
-rw-r--r--net/sunrpc/auth_gss/gss_spkm3_token.c2
-rw-r--r--net/sunrpc/auth_gss/svcauth_gss.c12
-rw-r--r--net/sunrpc/auth_unix.c2
-rw-r--r--net/sunrpc/clnt.c161
-rw-r--r--net/sunrpc/rpc_pipe.c2
-rw-r--r--net/sunrpc/rpcb_clnt.c377
-rw-r--r--net/sunrpc/sched.c23
-rw-r--r--net/sunrpc/svc.c110
-rw-r--r--net/sunrpc/xprt.c9
-rw-r--r--net/sunrpc/xprtrdma/svc_rdma.c35
-rw-r--r--net/sunrpc/xprtrdma/svc_rdma_recvfrom.c84
-rw-r--r--net/sunrpc/xprtrdma/svc_rdma_sendto.c167
-rw-r--r--net/sunrpc/xprtrdma/svc_rdma_transport.c195
-rw-r--r--net/sunrpc/xprtsock.c2
22 files changed, 748 insertions, 598 deletions
diff --git a/net/sunrpc/auth_gss/Makefile b/net/sunrpc/auth_gss/Makefile
index f3431a7e33da..4de8bcf26fa7 100644
--- a/net/sunrpc/auth_gss/Makefile
+++ b/net/sunrpc/auth_gss/Makefile
@@ -5,12 +5,12 @@
obj-$(CONFIG_SUNRPC_GSS) += auth_rpcgss.o
auth_rpcgss-objs := auth_gss.o gss_generic_token.o \
- gss_mech_switch.o svcauth_gss.o gss_krb5_crypto.o
+ gss_mech_switch.o svcauth_gss.o
obj-$(CONFIG_RPCSEC_GSS_KRB5) += rpcsec_gss_krb5.o
rpcsec_gss_krb5-objs := gss_krb5_mech.o gss_krb5_seal.o gss_krb5_unseal.o \
- gss_krb5_seqnum.o gss_krb5_wrap.o
+ gss_krb5_seqnum.o gss_krb5_wrap.o gss_krb5_crypto.o
obj-$(CONFIG_RPCSEC_GSS_SPKM3) += rpcsec_gss_spkm3.o
diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c
index cc12d5f5d5da..853a4142cea1 100644
--- a/net/sunrpc/auth_gss/auth_gss.c
+++ b/net/sunrpc/auth_gss/auth_gss.c
@@ -33,8 +33,6 @@
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
- * $Id$
*/
@@ -63,22 +61,11 @@ static const struct rpc_credops gss_nullops;
# define RPCDBG_FACILITY RPCDBG_AUTH
#endif
-#define NFS_NGROUPS 16
-
-#define GSS_CRED_SLACK 1024 /* XXX: unused */
+#define GSS_CRED_SLACK 1024
/* length of a krb5 verifier (48), plus data added before arguments when
* using integrity (two 4-byte integers): */
#define GSS_VERF_SLACK 100
-/* XXX this define must match the gssd define
-* as it is passed to gssd to signal the use of
-* machine creds should be part of the shared rpc interface */
-
-#define CA_RUN_AS_MACHINE 0x00000200
-
-/* dump the buffer in `emacs-hexl' style */
-#define isprint(c) ((c > 0x1f) && (c < 0x7f))
-
struct gss_auth {
struct kref kref;
struct rpc_auth rpc_auth;
@@ -146,7 +133,7 @@ simple_get_netobj(const void *p, const void *end, struct xdr_netobj *dest)
q = (const void *)((const char *)p + len);
if (unlikely(q > end || q < p))
return ERR_PTR(-EFAULT);
- dest->data = kmemdup(p, len, GFP_KERNEL);
+ dest->data = kmemdup(p, len, GFP_NOFS);
if (unlikely(dest->data == NULL))
return ERR_PTR(-ENOMEM);
dest->len = len;
@@ -171,7 +158,7 @@ gss_alloc_context(void)
{
struct gss_cl_ctx *ctx;
- ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+ ctx = kzalloc(sizeof(*ctx), GFP_NOFS);
if (ctx != NULL) {
ctx->gc_proc = RPC_GSS_PROC_DATA;
ctx->gc_seq = 1; /* NetApp 6.4R1 doesn't accept seq. no. 0 */
@@ -272,7 +259,7 @@ __gss_find_upcall(struct rpc_inode *rpci, uid_t uid)
return NULL;
}
-/* Try to add a upcall to the pipefs queue.
+/* Try to add an upcall to the pipefs queue.
* If an upcall owned by our uid already exists, then we return a reference
* to that upcall instead of adding the new upcall.
*/
@@ -341,7 +328,7 @@ gss_alloc_msg(struct gss_auth *gss_auth, uid_t uid)
{
struct gss_upcall_msg *gss_msg;
- gss_msg = kzalloc(sizeof(*gss_msg), GFP_KERNEL);
+ gss_msg = kzalloc(sizeof(*gss_msg), GFP_NOFS);
if (gss_msg != NULL) {
INIT_LIST_HEAD(&gss_msg->list);
rpc_init_wait_queue(&gss_msg->rpc_waitqueue, "RPCSEC_GSS upcall waitq");
@@ -493,7 +480,6 @@ gss_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
{
const void *p, *end;
void *buf;
- struct rpc_clnt *clnt;
struct gss_upcall_msg *gss_msg;
struct inode *inode = filp->f_path.dentry->d_inode;
struct gss_cl_ctx *ctx;
@@ -503,11 +489,10 @@ gss_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
if (mlen > MSG_BUF_MAXSIZE)
goto out;
err = -ENOMEM;
- buf = kmalloc(mlen, GFP_KERNEL);
+ buf = kmalloc(mlen, GFP_NOFS);
if (!buf)
goto out;
- clnt = RPC_I(inode)->private;
err = -EFAULT;
if (copy_from_user(buf, src, mlen))
goto err;
@@ -806,7 +791,7 @@ gss_create_cred(struct rpc_auth *auth, struct auth_cred *acred, int flags)
dprintk("RPC: gss_create_cred for uid %d, flavor %d\n",
acred->uid, auth->au_flavor);
- if (!(cred = kzalloc(sizeof(*cred), GFP_KERNEL)))
+ if (!(cred = kzalloc(sizeof(*cred), GFP_NOFS)))
goto out_err;
rpcauth_init_cred(&cred->gc_base, acred, auth, &gss_credops);
diff --git a/net/sunrpc/auth_gss/gss_krb5_crypto.c b/net/sunrpc/auth_gss/gss_krb5_crypto.c
index 1d52308ca324..c93fca204558 100644
--- a/net/sunrpc/auth_gss/gss_krb5_crypto.c
+++ b/net/sunrpc/auth_gss/gss_krb5_crypto.c
@@ -83,8 +83,6 @@ out:
return ret;
}
-EXPORT_SYMBOL(krb5_encrypt);
-
u32
krb5_decrypt(
struct crypto_blkcipher *tfm,
@@ -118,8 +116,6 @@ out:
return ret;
}
-EXPORT_SYMBOL(krb5_decrypt);
-
static int
checksummer(struct scatterlist *sg, void *data)
{
@@ -161,8 +157,6 @@ out:
return err ? GSS_S_FAILURE : 0;
}
-EXPORT_SYMBOL(make_checksum);
-
struct encryptor_desc {
u8 iv[8]; /* XXX hard-coded blocksize */
struct blkcipher_desc desc;
@@ -262,8 +256,6 @@ gss_encrypt_xdr_buf(struct crypto_blkcipher *tfm, struct xdr_buf *buf,
return ret;
}
-EXPORT_SYMBOL(gss_encrypt_xdr_buf);
-
struct decryptor_desc {
u8 iv[8]; /* XXX hard-coded blocksize */
struct blkcipher_desc desc;
@@ -334,5 +326,3 @@ gss_decrypt_xdr_buf(struct crypto_blkcipher *tfm, struct xdr_buf *buf,
return xdr_process_buf(buf, offset, buf->len - offset, decryptor, &desc);
}
-
-EXPORT_SYMBOL(gss_decrypt_xdr_buf);
diff --git a/net/sunrpc/auth_gss/gss_krb5_mech.c b/net/sunrpc/auth_gss/gss_krb5_mech.c
index 60c3dba545d7..ef45eba22485 100644
--- a/net/sunrpc/auth_gss/gss_krb5_mech.c
+++ b/net/sunrpc/auth_gss/gss_krb5_mech.c
@@ -70,7 +70,7 @@ simple_get_netobj(const void *p, const void *end, struct xdr_netobj *res)
q = (const void *)((const char *)p + len);
if (unlikely(q > end || q < p))
return ERR_PTR(-EFAULT);
- res->data = kmemdup(p, len, GFP_KERNEL);
+ res->data = kmemdup(p, len, GFP_NOFS);
if (unlikely(res->data == NULL))
return ERR_PTR(-ENOMEM);
res->len = len;
@@ -131,7 +131,7 @@ gss_import_sec_context_kerberos(const void *p,
struct krb5_ctx *ctx;
int tmp;
- if (!(ctx = kzalloc(sizeof(*ctx), GFP_KERNEL)))
+ if (!(ctx = kzalloc(sizeof(*ctx), GFP_NOFS)))
goto out_err;
p = simple_get_bytes(p, end, &ctx->initiate, sizeof(ctx->initiate));
diff --git a/net/sunrpc/auth_gss/gss_krb5_seal.c b/net/sunrpc/auth_gss/gss_krb5_seal.c
index 5f1d36dfbcf7..b8f42ef7178e 100644
--- a/net/sunrpc/auth_gss/gss_krb5_seal.c
+++ b/net/sunrpc/auth_gss/gss_krb5_seal.c
@@ -78,7 +78,7 @@ gss_get_mic_kerberos(struct gss_ctx *gss_ctx, struct xdr_buf *text,
struct krb5_ctx *ctx = gss_ctx->internal_ctx_id;
char cksumdata[16];
struct xdr_netobj md5cksum = {.len = 0, .data = cksumdata};
- unsigned char *ptr, *krb5_hdr, *msg_start;
+ unsigned char *ptr, *msg_start;
s32 now;
u32 seq_send;
@@ -87,36 +87,36 @@ gss_get_mic_kerberos(struct gss_ctx *gss_ctx, struct xdr_buf *text,
now = get_seconds();
- token->len = g_token_size(&ctx->mech_used, 24);
+ token->len = g_token_size(&ctx->mech_used, GSS_KRB5_TOK_HDR_LEN + 8);
ptr = token->data;
- g_make_token_header(&ctx->mech_used, 24, &ptr);
+ g_make_token_header(&ctx->mech_used, GSS_KRB5_TOK_HDR_LEN + 8, &ptr);
- *ptr++ = (unsigned char) ((KG_TOK_MIC_MSG>>8)&0xff);
- *ptr++ = (unsigned char) (KG_TOK_MIC_MSG&0xff);
+ /* ptr now at header described in rfc 1964, section 1.2.1: */
+ ptr[0] = (unsigned char) ((KG_TOK_MIC_MSG >> 8) & 0xff);
+ ptr[1] = (unsigned char) (KG_TOK_MIC_MSG & 0xff);
- /* ptr now at byte 2 of header described in rfc 1964, section 1.2.1: */
- krb5_hdr = ptr - 2;
- msg_start = krb5_hdr + 24;
+ msg_start = ptr + GSS_KRB5_TOK_HDR_LEN + 8;
- *(__be16 *)(krb5_hdr + 2) = htons(SGN_ALG_DES_MAC_MD5);
- memset(krb5_hdr + 4, 0xff, 4);
+ *(__be16 *)(ptr + 2) = htons(SGN_ALG_DES_MAC_MD5);
+ memset(ptr + 4, 0xff, 4);
- if (make_checksum("md5", krb5_hdr, 8, text, 0, &md5cksum))
+ if (make_checksum("md5", ptr, 8, text, 0, &md5cksum))
return GSS_S_FAILURE;
if (krb5_encrypt(ctx->seq, NULL, md5cksum.data,
md5cksum.data, md5cksum.len))
return GSS_S_FAILURE;
- memcpy(krb5_hdr + 16, md5cksum.data + md5cksum.len - 8, 8);
+ memcpy(ptr + GSS_KRB5_TOK_HDR_LEN, md5cksum.data + md5cksum.len - 8, 8);
spin_lock(&krb5_seq_lock);
seq_send = ctx->seq_send++;
spin_unlock(&krb5_seq_lock);
if (krb5_make_seq_num(ctx->seq, ctx->initiate ? 0 : 0xff,
- seq_send, krb5_hdr + 16, krb5_hdr + 8))
+ seq_send, ptr + GSS_KRB5_TOK_HDR_LEN,
+ ptr + 8))
return GSS_S_FAILURE;
return (ctx->endtime < now) ? GSS_S_CONTEXT_EXPIRED : GSS_S_COMPLETE;
diff --git a/net/sunrpc/auth_gss/gss_krb5_unseal.c b/net/sunrpc/auth_gss/gss_krb5_unseal.c
index d91a5d004803..066ec73c84d6 100644
--- a/net/sunrpc/auth_gss/gss_krb5_unseal.c
+++ b/net/sunrpc/auth_gss/gss_krb5_unseal.c
@@ -92,30 +92,30 @@ gss_verify_mic_kerberos(struct gss_ctx *gss_ctx,
read_token->len))
return GSS_S_DEFECTIVE_TOKEN;
- if ((*ptr++ != ((KG_TOK_MIC_MSG>>8)&0xff)) ||
- (*ptr++ != ( KG_TOK_MIC_MSG &0xff)) )
+ if ((ptr[0] != ((KG_TOK_MIC_MSG >> 8) & 0xff)) ||
+ (ptr[1] != (KG_TOK_MIC_MSG & 0xff)))
return GSS_S_DEFECTIVE_TOKEN;
/* XXX sanity-check bodysize?? */
- signalg = ptr[0] + (ptr[1] << 8);
+ signalg = ptr[2] + (ptr[3] << 8);
if (signalg != SGN_ALG_DES_MAC_MD5)
return GSS_S_DEFECTIVE_TOKEN;
- sealalg = ptr[2] + (ptr[3] << 8);
+ sealalg = ptr[4] + (ptr[5] << 8);
if (sealalg != SEAL_ALG_NONE)
return GSS_S_DEFECTIVE_TOKEN;
- if ((ptr[4] != 0xff) || (ptr[5] != 0xff))
+ if ((ptr[6] != 0xff) || (ptr[7] != 0xff))
return GSS_S_DEFECTIVE_TOKEN;
- if (make_checksum("md5", ptr - 2, 8, message_buffer, 0, &md5cksum))
+ if (make_checksum("md5", ptr, 8, message_buffer, 0, &md5cksum))
return GSS_S_FAILURE;
if (krb5_encrypt(ctx->seq, NULL, md5cksum.data, md5cksum.data, 16))
return GSS_S_FAILURE;
- if (memcmp(md5cksum.data + 8, ptr + 14, 8))
+ if (memcmp(md5cksum.data + 8, ptr + GSS_KRB5_TOK_HDR_LEN, 8))
return GSS_S_BAD_SIG;
/* it got through unscathed. Make sure the context is unexpired */
@@ -127,7 +127,7 @@ gss_verify_mic_kerberos(struct gss_ctx *gss_ctx,
/* do sequencing checks */
- if (krb5_get_seq_num(ctx->seq, ptr + 14, ptr + 6, &direction, &seqnum))
+ if (krb5_get_seq_num(ctx->seq, ptr + GSS_KRB5_TOK_HDR_LEN, ptr + 8, &direction, &seqnum))
return GSS_S_FAILURE;
if ((ctx->initiate && direction != 0xff) ||
diff --git a/net/sunrpc/auth_gss/gss_krb5_wrap.c b/net/sunrpc/auth_gss/gss_krb5_wrap.c
index b00b1b426301..ae8e69b59c4c 100644
--- a/net/sunrpc/auth_gss/gss_krb5_wrap.c
+++ b/net/sunrpc/auth_gss/gss_krb5_wrap.c
@@ -87,8 +87,8 @@ out:
return 0;
}
-static inline void
-make_confounder(char *p, int blocksize)
+static void
+make_confounder(char *p, u32 conflen)
{
static u64 i = 0;
u64 *q = (u64 *)p;
@@ -102,8 +102,22 @@ make_confounder(char *p, int blocksize)
* uniqueness would mean worrying about atomicity and rollover, and I
* don't care enough. */
- BUG_ON(blocksize != 8);
- *q = i++;
+ /* initialize to random value */
+ if (i == 0) {
+ i = random32();
+ i = (i << 32) | random32();
+ }
+
+ switch (conflen) {
+ case 16:
+ *q++ = i++;
+ /* fall through */
+ case 8:
+ *q++ = i++;
+ break;
+ default:
+ BUG();
+ }
}
/* Assumptions: the head and tail of inbuf are ours to play with.
@@ -122,7 +136,7 @@ gss_wrap_kerberos(struct gss_ctx *ctx, int offset,
char cksumdata[16];
struct xdr_netobj md5cksum = {.len = 0, .data = cksumdata};
int blocksize = 0, plainlen;
- unsigned char *ptr, *krb5_hdr, *msg_start;
+ unsigned char *ptr, *msg_start;
s32 now;
int headlen;
struct page **tmp_pages;
@@ -149,26 +163,26 @@ gss_wrap_kerberos(struct gss_ctx *ctx, int offset,
buf->len += headlen;
BUG_ON((buf->len - offset - headlen) % blocksize);
- g_make_token_header(&kctx->mech_used, 24 + plainlen, &ptr);
+ g_make_token_header(&kctx->mech_used,
+ GSS_KRB5_TOK_HDR_LEN + 8 + plainlen, &ptr);
- *ptr++ = (unsigned char) ((KG_TOK_WRAP_MSG>>8)&0xff);
- *ptr++ = (unsigned char) (KG_TOK_WRAP_MSG&0xff);
+ /* ptr now at header described in rfc 1964, section 1.2.1: */
+ ptr[0] = (unsigned char) ((KG_TOK_WRAP_MSG >> 8) & 0xff);
+ ptr[1] = (unsigned char) (KG_TOK_WRAP_MSG & 0xff);
- /* ptr now at byte 2 of header described in rfc 1964, section 1.2.1: */
- krb5_hdr = ptr - 2;
- msg_start = krb5_hdr + 24;
+ msg_start = ptr + 24;
- *(__be16 *)(krb5_hdr + 2) = htons(SGN_ALG_DES_MAC_MD5);
- memset(krb5_hdr + 4, 0xff, 4);
- *(__be16 *)(krb5_hdr + 4) = htons(SEAL_ALG_DES);
+ *(__be16 *)(ptr + 2) = htons(SGN_ALG_DES_MAC_MD5);
+ memset(ptr + 4, 0xff, 4);
+ *(__be16 *)(ptr + 4) = htons(SEAL_ALG_DES);
make_confounder(msg_start, blocksize);
/* XXXJBF: UGH!: */
tmp_pages = buf->pages;
buf->pages = pages;
- if (make_checksum("md5", krb5_hdr, 8, buf,
+ if (make_checksum("md5", ptr, 8, buf,
offset + headlen - blocksize, &md5cksum))
return GSS_S_FAILURE;
buf->pages = tmp_pages;
@@ -176,7 +190,7 @@ gss_wrap_kerberos(struct gss_ctx *ctx, int offset,
if (krb5_encrypt(kctx->seq, NULL, md5cksum.data,
md5cksum.data, md5cksum.len))
return GSS_S_FAILURE;
- memcpy(krb5_hdr + 16, md5cksum.data + md5cksum.len - 8, 8);
+ memcpy(ptr + GSS_KRB5_TOK_HDR_LEN, md5cksum.data + md5cksum.len - 8, 8);
spin_lock(&krb5_seq_lock);
seq_send = kctx->seq_send++;
@@ -185,7 +199,7 @@ gss_wrap_kerberos(struct gss_ctx *ctx, int offset,
/* XXX would probably be more efficient to compute checksum
* and encrypt at the same time: */
if ((krb5_make_seq_num(kctx->seq, kctx->initiate ? 0 : 0xff,
- seq_send, krb5_hdr + 16, krb5_hdr + 8)))
+ seq_send, ptr + GSS_KRB5_TOK_HDR_LEN, ptr + 8)))
return GSS_S_FAILURE;
if (gss_encrypt_xdr_buf(kctx->enc, buf, offset + headlen - blocksize,
@@ -219,38 +233,38 @@ gss_unwrap_kerberos(struct gss_ctx *ctx, int offset, struct xdr_buf *buf)
buf->len - offset))
return GSS_S_DEFECTIVE_TOKEN;
- if ((*ptr++ != ((KG_TOK_WRAP_MSG>>8)&0xff)) ||
- (*ptr++ != (KG_TOK_WRAP_MSG &0xff)) )
+ if ((ptr[0] != ((KG_TOK_WRAP_MSG >> 8) & 0xff)) ||
+ (ptr[1] != (KG_TOK_WRAP_MSG & 0xff)))
return GSS_S_DEFECTIVE_TOKEN;
/* XXX sanity-check bodysize?? */
/* get the sign and seal algorithms */
- signalg = ptr[0] + (ptr[1] << 8);
+ signalg = ptr[2] + (ptr[3] << 8);
if (signalg != SGN_ALG_DES_MAC_MD5)
return GSS_S_DEFECTIVE_TOKEN;
- sealalg = ptr[2] + (ptr[3] << 8);
+ sealalg = ptr[4] + (ptr[5] << 8);
if (sealalg != SEAL_ALG_DES)
return GSS_S_DEFECTIVE_TOKEN;
- if ((ptr[4] != 0xff) || (ptr[5] != 0xff))
+ if ((ptr[6] != 0xff) || (ptr[7] != 0xff))
return GSS_S_DEFECTIVE_TOKEN;
if (gss_decrypt_xdr_buf(kctx->enc, buf,
- ptr + 22 - (unsigned char *)buf->head[0].iov_base))
+ ptr + GSS_KRB5_TOK_HDR_LEN + 8 - (unsigned char *)buf->head[0].iov_base))
return GSS_S_DEFECTIVE_TOKEN;
- if (make_checksum("md5", ptr - 2, 8, buf,
- ptr + 22 - (unsigned char *)buf->head[0].iov_base, &md5cksum))
+ if (make_checksum("md5", ptr, 8, buf,
+ ptr + GSS_KRB5_TOK_HDR_LEN + 8 - (unsigned char *)buf->head[0].iov_base, &md5cksum))
return GSS_S_FAILURE;
if (krb5_encrypt(kctx->seq, NULL, md5cksum.data,
md5cksum.data, md5cksum.len))
return GSS_S_FAILURE;
- if (memcmp(md5cksum.data + 8, ptr + 14, 8))
+ if (memcmp(md5cksum.data + 8, ptr + GSS_KRB5_TOK_HDR_LEN, 8))
return GSS_S_BAD_SIG;
/* it got through unscathed. Make sure the context is unexpired */
@@ -262,8 +276,8 @@ gss_unwrap_kerberos(struct gss_ctx *ctx, int offset, struct xdr_buf *buf)
/* do sequencing checks */
- if (krb5_get_seq_num(kctx->seq, ptr + 14, ptr + 6, &direction,
- &seqnum))
+ if (krb5_get_seq_num(kctx->seq, ptr + GSS_KRB5_TOK_HDR_LEN, ptr + 8,
+ &direction, &seqnum))
return GSS_S_BAD_SIG;
if ((kctx->initiate && direction != 0xff) ||
@@ -274,7 +288,7 @@ gss_unwrap_kerberos(struct gss_ctx *ctx, int offset, struct xdr_buf *buf)
* better to copy and encrypt at the same time. */
blocksize = crypto_blkcipher_blocksize(kctx->enc);
- data_start = ptr + 22 + blocksize;
+ data_start = ptr + GSS_KRB5_TOK_HDR_LEN + 8 + blocksize;
orig_start = buf->head[0].iov_base + offset;
data_len = (buf->head[0].iov_base + buf->head[0].iov_len) - data_start;
memmove(orig_start, data_start, data_len);
diff --git a/net/sunrpc/auth_gss/gss_spkm3_mech.c b/net/sunrpc/auth_gss/gss_spkm3_mech.c
index 5deb4b6e4514..035e1dd6af1b 100644
--- a/net/sunrpc/auth_gss/gss_spkm3_mech.c
+++ b/net/sunrpc/auth_gss/gss_spkm3_mech.c
@@ -76,7 +76,7 @@ simple_get_netobj(const void *p, const void *end, struct xdr_netobj *res)
q = (const void *)((const char *)p + len);
if (unlikely(q > end || q < p))
return ERR_PTR(-EFAULT);
- res->data = kmemdup(p, len, GFP_KERNEL);
+ res->data = kmemdup(p, len, GFP_NOFS);
if (unlikely(res->data == NULL))
return ERR_PTR(-ENOMEM);
return q;
@@ -90,7 +90,7 @@ gss_import_sec_context_spkm3(const void *p, size_t len,
struct spkm3_ctx *ctx;
int version;
- if (!(ctx = kzalloc(sizeof(*ctx), GFP_KERNEL)))
+ if (!(ctx = kzalloc(sizeof(*ctx), GFP_NOFS)))
goto out_err;
p = simple_get_bytes(p, end, &version, sizeof(version));
diff --git a/net/sunrpc/auth_gss/gss_spkm3_token.c b/net/sunrpc/auth_gss/gss_spkm3_token.c
index 6cdd241ad267..3308157436d2 100644
--- a/net/sunrpc/auth_gss/gss_spkm3_token.c
+++ b/net/sunrpc/auth_gss/gss_spkm3_token.c
@@ -90,7 +90,7 @@ asn1_bitstring_len(struct xdr_netobj *in, int *enclen, int *zerobits)
int
decode_asn1_bitstring(struct xdr_netobj *out, char *in, int enclen, int explen)
{
- if (!(out->data = kzalloc(explen,GFP_KERNEL)))
+ if (!(out->data = kzalloc(explen,GFP_NOFS)))
return 0;
out->len = explen;
memcpy(out->data, in, enclen);
diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c
index 5905d56737d6..81ae3d62a0cc 100644
--- a/net/sunrpc/auth_gss/svcauth_gss.c
+++ b/net/sunrpc/auth_gss/svcauth_gss.c
@@ -1144,20 +1144,20 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp)
case RPC_GSS_SVC_NONE:
break;
case RPC_GSS_SVC_INTEGRITY:
+ /* placeholders for length and seq. number: */
+ svc_putnl(resv, 0);
+ svc_putnl(resv, 0);
if (unwrap_integ_data(&rqstp->rq_arg,
gc->gc_seq, rsci->mechctx))
goto garbage_args;
+ break;
+ case RPC_GSS_SVC_PRIVACY:
/* placeholders for length and seq. number: */
svc_putnl(resv, 0);
svc_putnl(resv, 0);
- break;
- case RPC_GSS_SVC_PRIVACY:
if (unwrap_priv_data(rqstp, &rqstp->rq_arg,
gc->gc_seq, rsci->mechctx))
goto garbage_args;
- /* placeholders for length and seq. number: */
- svc_putnl(resv, 0);
- svc_putnl(resv, 0);
break;
default:
goto auth_err;
@@ -1170,8 +1170,6 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp)
goto out;
}
garbage_args:
- /* Restore write pointer to its original value: */
- xdr_ressize_check(rqstp, reject_stat);
ret = SVC_GARBAGE;
goto out;
auth_err:
diff --git a/net/sunrpc/auth_unix.c b/net/sunrpc/auth_unix.c
index 44920b90bdc4..46b2647c5bd2 100644
--- a/net/sunrpc/auth_unix.c
+++ b/net/sunrpc/auth_unix.c
@@ -66,7 +66,7 @@ unx_create_cred(struct rpc_auth *auth, struct auth_cred *acred, int flags)
dprintk("RPC: allocating UNIX cred for uid %d gid %d\n",
acred->uid, acred->gid);
- if (!(cred = kmalloc(sizeof(*cred), GFP_KERNEL)))
+ if (!(cred = kmalloc(sizeof(*cred), GFP_NOFS)))
return ERR_PTR(-ENOMEM);
rpcauth_init_cred(&cred->uc_base, acred, auth, &unix_credops);
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 8945307556ec..76739e928d0d 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -25,6 +25,7 @@
#include <linux/module.h>
#include <linux/types.h>
+#include <linux/kallsyms.h>
#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/smp_lock.h>
@@ -58,7 +59,6 @@ static void call_start(struct rpc_task *task);
static void call_reserve(struct rpc_task *task);
static void call_reserveresult(struct rpc_task *task);
static void call_allocate(struct rpc_task *task);
-static void call_encode(struct rpc_task *task);
static void call_decode(struct rpc_task *task);
static void call_bind(struct rpc_task *task);
static void call_bind_status(struct rpc_task *task);
@@ -70,9 +70,9 @@ static void call_refreshresult(struct rpc_task *task);
static void call_timeout(struct rpc_task *task);
static void call_connect(struct rpc_task *task);
static void call_connect_status(struct rpc_task *task);
-static __be32 * call_header(struct rpc_task *task);
-static __be32 * call_verify(struct rpc_task *task);
+static __be32 *rpc_encode_header(struct rpc_task *task);
+static __be32 *rpc_verify_header(struct rpc_task *task);
static int rpc_ping(struct rpc_clnt *clnt, int flags);
static void rpc_register_client(struct rpc_clnt *clnt)
@@ -324,6 +324,8 @@ struct rpc_clnt *rpc_create(struct rpc_create_args *args)
clnt->cl_autobind = 1;
if (args->flags & RPC_CLNT_CREATE_DISCRTRY)
clnt->cl_discrtry = 1;
+ if (!(args->flags & RPC_CLNT_CREATE_QUIET))
+ clnt->cl_chatty = 1;
return clnt;
}
@@ -690,6 +692,21 @@ rpc_restart_call(struct rpc_task *task)
}
EXPORT_SYMBOL_GPL(rpc_restart_call);
+#ifdef RPC_DEBUG
+static const char *rpc_proc_name(const struct rpc_task *task)
+{
+ const struct rpc_procinfo *proc = task->tk_msg.rpc_proc;
+
+ if (proc) {
+ if (proc->p_name)
+ return proc->p_name;
+ else
+ return "NULL";
+ } else
+ return "no proc";
+}
+#endif
+
/*
* 0. Initial state
*
@@ -701,9 +718,9 @@ call_start(struct rpc_task *task)
{
struct rpc_clnt *clnt = task->tk_client;
- dprintk("RPC: %5u call_start %s%d proc %d (%s)\n", task->tk_pid,
+ dprintk("RPC: %5u call_start %s%d proc %s (%s)\n", task->tk_pid,
clnt->cl_protname, clnt->cl_vers,
- task->tk_msg.rpc_proc->p_proc,
+ rpc_proc_name(task),
(RPC_IS_ASYNC(task) ? "async" : "sync"));
/* Increment call count */
@@ -861,7 +878,7 @@ rpc_xdr_buf_init(struct xdr_buf *buf, void *start, size_t len)
* 3. Encode arguments of an RPC call
*/
static void
-call_encode(struct rpc_task *task)
+rpc_xdr_encode(struct rpc_task *task)
{
struct rpc_rqst *req = task->tk_rqstp;
kxdrproc_t encode;
@@ -876,23 +893,19 @@ call_encode(struct rpc_task *task)
(char *)req->rq_buffer + req->rq_callsize,
req->rq_rcvsize);
- /* Encode header and provided arguments */
- encode = task->tk_msg.rpc_proc->p_encode;
- if (!(p = call_header(task))) {
- printk(KERN_INFO "RPC: call_header failed, exit EIO\n");
+ p = rpc_encode_header(task);
+ if (p == NULL) {
+ printk(KERN_INFO "RPC: couldn't encode RPC header, exit EIO\n");
rpc_exit(task, -EIO);
return;
}
+
+ encode = task->tk_msg.rpc_proc->p_encode;
if (encode == NULL)
return;
task->tk_status = rpcauth_wrap_req(task, encode, req, p,
task->tk_msg.rpc_argp);
- if (task->tk_status == -ENOMEM) {
- /* XXX: Is this sane? */
- rpc_delay(task, 3*HZ);
- task->tk_status = -EAGAIN;
- }
}
/*
@@ -929,11 +942,9 @@ call_bind_status(struct rpc_task *task)
}
switch (task->tk_status) {
- case -EAGAIN:
- dprintk("RPC: %5u rpcbind waiting for another request "
- "to finish\n", task->tk_pid);
- /* avoid busy-waiting here -- could be a network outage. */
- rpc_delay(task, 5*HZ);
+ case -ENOMEM:
+ dprintk("RPC: %5u rpcbind out of memory\n", task->tk_pid);
+ rpc_delay(task, HZ >> 2);
goto retry_timeout;
case -EACCES:
dprintk("RPC: %5u remote rpcbind: RPC program/version "
@@ -1046,10 +1057,16 @@ call_transmit(struct rpc_task *task)
/* Encode here so that rpcsec_gss can use correct sequence number. */
if (rpc_task_need_encode(task)) {
BUG_ON(task->tk_rqstp->rq_bytes_sent != 0);
- call_encode(task);
+ rpc_xdr_encode(task);
/* Did the encode result in an error condition? */
- if (task->tk_status != 0)
+ if (task->tk_status != 0) {
+ /* Was the error nonfatal? */
+ if (task->tk_status == -EAGAIN)
+ rpc_delay(task, HZ >> 4);
+ else
+ rpc_exit(task, task->tk_status);
return;
+ }
}
xprt_transmit(task);
if (task->tk_status < 0)
@@ -1132,7 +1149,8 @@ call_status(struct rpc_task *task)
rpc_exit(task, status);
break;
default:
- printk("%s: RPC call returned error %d\n",
+ if (clnt->cl_chatty)
+ printk("%s: RPC call returned error %d\n",
clnt->cl_protname, -status);
rpc_exit(task, status);
}
@@ -1157,7 +1175,8 @@ call_timeout(struct rpc_task *task)
task->tk_timeouts++;
if (RPC_IS_SOFT(task)) {
- printk(KERN_NOTICE "%s: server %s not responding, timed out\n",
+ if (clnt->cl_chatty)
+ printk(KERN_NOTICE "%s: server %s not responding, timed out\n",
clnt->cl_protname, clnt->cl_server);
rpc_exit(task, -EIO);
return;
@@ -1165,7 +1184,8 @@ call_timeout(struct rpc_task *task)
if (!(task->tk_flags & RPC_CALL_MAJORSEEN)) {
task->tk_flags |= RPC_CALL_MAJORSEEN;
- printk(KERN_NOTICE "%s: server %s not responding, still trying\n",
+ if (clnt->cl_chatty)
+ printk(KERN_NOTICE "%s: server %s not responding, still trying\n",
clnt->cl_protname, clnt->cl_server);
}
rpc_force_rebind(clnt);
@@ -1196,8 +1216,9 @@ call_decode(struct rpc_task *task)
task->tk_pid, task->tk_status);
if (task->tk_flags & RPC_CALL_MAJORSEEN) {
- printk(KERN_NOTICE "%s: server %s OK\n",
- clnt->cl_protname, clnt->cl_server);
+ if (clnt->cl_chatty)
+ printk(KERN_NOTICE "%s: server %s OK\n",
+ clnt->cl_protname, clnt->cl_server);
task->tk_flags &= ~RPC_CALL_MAJORSEEN;
}
@@ -1224,8 +1245,7 @@ call_decode(struct rpc_task *task)
goto out_retry;
}
- /* Verify the RPC header */
- p = call_verify(task);
+ p = rpc_verify_header(task);
if (IS_ERR(p)) {
if (p == ERR_PTR(-EAGAIN))
goto out_retry;
@@ -1243,7 +1263,7 @@ call_decode(struct rpc_task *task)
return;
out_retry:
task->tk_status = 0;
- /* Note: call_verify() may have freed the RPC slot */
+ /* Note: rpc_verify_header() may have freed the RPC slot */
if (task->tk_rqstp == req) {
req->rq_received = req->rq_rcv_buf.len = 0;
if (task->tk_client->cl_discrtry)
@@ -1290,11 +1310,8 @@ call_refreshresult(struct rpc_task *task)
return;
}
-/*
- * Call header serialization
- */
static __be32 *
-call_header(struct rpc_task *task)
+rpc_encode_header(struct rpc_task *task)
{
struct rpc_clnt *clnt = task->tk_client;
struct rpc_rqst *req = task->tk_rqstp;
@@ -1314,11 +1331,8 @@ call_header(struct rpc_task *task)
return p;
}
-/*
- * Reply header verification
- */
static __be32 *
-call_verify(struct rpc_task *task)
+rpc_verify_header(struct rpc_task *task)
{
struct kvec *iov = &task->tk_rqstp->rq_rcv_buf.head[0];
int len = task->tk_rqstp->rq_rcv_buf.len >> 2;
@@ -1392,7 +1406,7 @@ call_verify(struct rpc_task *task)
task->tk_action = call_bind;
goto out_retry;
case RPC_AUTH_TOOWEAK:
- printk(KERN_NOTICE "call_verify: server %s requires stronger "
+ printk(KERN_NOTICE "RPC: server %s requires stronger "
"authentication.\n", task->tk_client->cl_server);
break;
default:
@@ -1431,10 +1445,10 @@ call_verify(struct rpc_task *task)
error = -EPROTONOSUPPORT;
goto out_err;
case RPC_PROC_UNAVAIL:
- dprintk("RPC: %5u %s: proc %p unsupported by program %u, "
+ dprintk("RPC: %5u %s: proc %s unsupported by program %u, "
"version %u on server %s\n",
task->tk_pid, __func__,
- task->tk_msg.rpc_proc,
+ rpc_proc_name(task),
task->tk_client->cl_prog,
task->tk_client->cl_vers,
task->tk_client->cl_server);
@@ -1517,44 +1531,53 @@ struct rpc_task *rpc_call_null(struct rpc_clnt *clnt, struct rpc_cred *cred, int
EXPORT_SYMBOL_GPL(rpc_call_null);
#ifdef RPC_DEBUG
+static void rpc_show_header(void)
+{
+ printk(KERN_INFO "-pid- flgs status -client- --rqstp- "
+ "-timeout ---ops--\n");
+}
+
+static void rpc_show_task(const struct rpc_clnt *clnt,
+ const struct rpc_task *task)
+{
+ const char *rpc_waitq = "none";
+ char *p, action[KSYM_SYMBOL_LEN];
+
+ if (RPC_IS_QUEUED(task))
+ rpc_waitq = rpc_qname(task->tk_waitqueue);
+
+ /* map tk_action pointer to a function name; then trim off
+ * the "+0x0 [sunrpc]" */
+ sprint_symbol(action, (unsigned long)task->tk_action);
+ p = strchr(action, '+');
+ if (p)
+ *p = '\0';
+
+ printk(KERN_INFO "%5u %04x %6d %8p %8p %8ld %8p %sv%u %s a:%s q:%s\n",
+ task->tk_pid, task->tk_flags, task->tk_status,
+ clnt, task->tk_rqstp, task->tk_timeout, task->tk_ops,
+ clnt->cl_protname, clnt->cl_vers, rpc_proc_name(task),
+ action, rpc_waitq);
+}
+
void rpc_show_tasks(void)
{
struct rpc_clnt *clnt;
- struct rpc_task *t;
+ struct rpc_task *task;
+ int header = 0;
spin_lock(&rpc_client_lock);
- if (list_empty(&all_clients))
- goto out;
- printk("-pid- proc flgs status -client- -prog- --rqstp- -timeout "
- "-rpcwait -action- ---ops--\n");
list_for_each_entry(clnt, &all_clients, cl_clients) {
- if (list_empty(&clnt->cl_tasks))
- continue;
spin_lock(&clnt->cl_lock);
- list_for_each_entry(t, &clnt->cl_tasks, tk_task) {
- const char *rpc_waitq = "none";
- int proc;
-
- if (t->tk_msg.rpc_proc)
- proc = t->tk_msg.rpc_proc->p_proc;
- else
- proc = -1;
-
- if (RPC_IS_QUEUED(t))
- rpc_waitq = rpc_qname(t->tk_waitqueue);
-
- printk("%5u %04d %04x %6d %8p %6d %8p %8ld %8s %8p %8p\n",
- t->tk_pid, proc,
- t->tk_flags, t->tk_status,
- t->tk_client,
- (t->tk_client ? t->tk_client->cl_prog : 0),
- t->tk_rqstp, t->tk_timeout,
- rpc_waitq,
- t->tk_action, t->tk_ops);
+ list_for_each_entry(task, &clnt->cl_tasks, tk_task) {
+ if (!header) {
+ rpc_show_header();
+ header++;
+ }
+ rpc_show_task(clnt, task);
}
spin_unlock(&clnt->cl_lock);
}
-out:
spin_unlock(&rpc_client_lock);
}
#endif
diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c
index 5a9b0e7828cd..23a2b8f6dc49 100644
--- a/net/sunrpc/rpc_pipe.c
+++ b/net/sunrpc/rpc_pipe.c
@@ -897,7 +897,7 @@ static struct file_system_type rpc_pipe_fs_type = {
};
static void
-init_once(struct kmem_cache * cachep, void *foo)
+init_once(void *foo)
{
struct rpc_inode *rpci = (struct rpc_inode *) foo;
diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c
index 0517967a68bf..24db2b4d12d3 100644
--- a/net/sunrpc/rpcb_clnt.c
+++ b/net/sunrpc/rpcb_clnt.c
@@ -32,6 +32,10 @@
#define RPCBIND_PROGRAM (100000u)
#define RPCBIND_PORT (111u)
+#define RPCBVERS_2 (2u)
+#define RPCBVERS_3 (3u)
+#define RPCBVERS_4 (4u)
+
enum {
RPCBPROC_NULL,
RPCBPROC_SET,
@@ -64,6 +68,7 @@ enum {
#define RPCB_MAXOWNERLEN sizeof(RPCB_OWNER_STRING)
static void rpcb_getport_done(struct rpc_task *, void *);
+static void rpcb_map_release(void *data);
static struct rpc_program rpcb_program;
struct rpcbind_args {
@@ -76,41 +81,73 @@ struct rpcbind_args {
const char * r_netid;
const char * r_addr;
const char * r_owner;
+
+ int r_status;
};
static struct rpc_procinfo rpcb_procedures2[];
static struct rpc_procinfo rpcb_procedures3[];
+static struct rpc_procinfo rpcb_procedures4[];
struct rpcb_info {
- int rpc_vers;
+ u32 rpc_vers;
struct rpc_procinfo * rpc_proc;
};
static struct rpcb_info rpcb_next_version[];
static struct rpcb_info rpcb_next_version6[];
+static const struct rpc_call_ops rpcb_getport_ops = {
+ .rpc_call_done = rpcb_getport_done,
+ .rpc_release = rpcb_map_release,
+};
+
+static void rpcb_wake_rpcbind_waiters(struct rpc_xprt *xprt, int status)
+{
+ xprt_clear_binding(xprt);
+ rpc_wake_up_status(&xprt->binding, status);
+}
+
static void rpcb_map_release(void *data)
{
struct rpcbind_args *map = data;
+ rpcb_wake_rpcbind_waiters(map->r_xprt, map->r_status);
xprt_put(map->r_xprt);
kfree(map);
}
-static const struct rpc_call_ops rpcb_getport_ops = {
- .rpc_call_done = rpcb_getport_done,
- .rpc_release = rpcb_map_release,
+static const struct sockaddr_in rpcb_inaddr_loopback = {
+ .sin_family = AF_INET,
+ .sin_addr.s_addr = htonl(INADDR_LOOPBACK),
+ .sin_port = htons(RPCBIND_PORT),
};
-static void rpcb_wake_rpcbind_waiters(struct rpc_xprt *xprt, int status)
+static const struct sockaddr_in6 rpcb_in6addr_loopback = {
+ .sin6_family = AF_INET6,
+ .sin6_addr = IN6ADDR_LOOPBACK_INIT,
+ .sin6_port = htons(RPCBIND_PORT),
+};
+
+static struct rpc_clnt *rpcb_create_local(struct sockaddr *addr,
+ size_t addrlen, u32 version)
{
- xprt_clear_binding(xprt);
- rpc_wake_up_status(&xprt->binding, status);
+ struct rpc_create_args args = {
+ .protocol = XPRT_TRANSPORT_UDP,
+ .address = addr,
+ .addrsize = addrlen,
+ .servername = "localhost",
+ .program = &rpcb_program,
+ .version = version,
+ .authflavor = RPC_AUTH_UNIX,
+ .flags = RPC_CLNT_CREATE_NOPING,
+ };
+
+ return rpc_create(&args);
}
static struct rpc_clnt *rpcb_create(char *hostname, struct sockaddr *srvaddr,
- size_t salen, int proto, u32 version,
- int privileged)
+ size_t salen, int proto, u32 version)
{
struct rpc_create_args args = {
.protocol = proto,
@@ -120,7 +157,8 @@ static struct rpc_clnt *rpcb_create(char *hostname, struct sockaddr *srvaddr,
.program = &rpcb_program,
.version = version,
.authflavor = RPC_AUTH_UNIX,
- .flags = RPC_CLNT_CREATE_NOPING,
+ .flags = (RPC_CLNT_CREATE_NOPING |
+ RPC_CLNT_CREATE_NONPRIVPORT),
};
switch (srvaddr->sa_family) {
@@ -134,29 +172,72 @@ static struct rpc_clnt *rpcb_create(char *hostname, struct sockaddr *srvaddr,
return NULL;
}
- if (!privileged)
- args.flags |= RPC_CLNT_CREATE_NONPRIVPORT;
return rpc_create(&args);
}
+static int rpcb_register_call(struct sockaddr *addr, size_t addrlen,
+ u32 version, struct rpc_message *msg,
+ int *result)
+{
+ struct rpc_clnt *rpcb_clnt;
+ int error = 0;
+
+ *result = 0;
+
+ rpcb_clnt = rpcb_create_local(addr, addrlen, version);
+ if (!IS_ERR(rpcb_clnt)) {
+ error = rpc_call_sync(rpcb_clnt, msg, 0);
+ rpc_shutdown_client(rpcb_clnt);
+ } else
+ error = PTR_ERR(rpcb_clnt);
+
+ if (error < 0)
+ printk(KERN_WARNING "RPC: failed to contact local rpcbind "
+ "server (errno %d).\n", -error);
+ dprintk("RPC: registration status %d/%d\n", error, *result);
+
+ return error;
+}
+
/**
* rpcb_register - set or unset a port registration with the local rpcbind svc
* @prog: RPC program number to bind
* @vers: RPC version number to bind
- * @prot: transport protocol to use to make this request
+ * @prot: transport protocol to register
* @port: port value to register
- * @okay: result code
+ * @okay: OUT: result code
+ *
+ * RPC services invoke this function to advertise their contact
+ * information via the system's rpcbind daemon. RPC services
+ * invoke this function once for each [program, version, transport]
+ * tuple they wish to advertise.
+ *
+ * Callers may also unregister RPC services that are no longer
+ * available by setting the passed-in port to zero. This removes
+ * all registered transports for [program, version] from the local
+ * rpcbind database.
+ *
+ * Returns zero if the registration request was dispatched
+ * successfully and a reply was received. The rpcbind daemon's
+ * boolean result code is stored in *okay.
+ *
+ * Returns an errno value and sets *result to zero if there was
+ * some problem that prevented the rpcbind request from being
+ * dispatched, or if the rpcbind daemon did not respond within
+ * the timeout.
*
- * port == 0 means unregister, port != 0 means register.
+ * This function uses rpcbind protocol version 2 to contact the
+ * local rpcbind daemon.
*
- * This routine supports only rpcbind version 2.
+ * Registration works over both AF_INET and AF_INET6, and services
+ * registered via this function are advertised as available for any
+ * address. If the local rpcbind daemon is listening on AF_INET6,
+ * services registered via this function will be advertised on
+ * IN6ADDR_ANY (ie available for all AF_INET and AF_INET6
+ * addresses).
*/
int rpcb_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay)
{
- struct sockaddr_in sin = {
- .sin_family = AF_INET,
- .sin_addr.s_addr = htonl(INADDR_LOOPBACK),
- };
struct rpcbind_args map = {
.r_prog = prog,
.r_vers = vers,
@@ -164,32 +245,159 @@ int rpcb_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay)
.r_port = port,
};
struct rpc_message msg = {
- .rpc_proc = &rpcb_procedures2[port ?
- RPCBPROC_SET : RPCBPROC_UNSET],
.rpc_argp = &map,
.rpc_resp = okay,
};
- struct rpc_clnt *rpcb_clnt;
- int error = 0;
dprintk("RPC: %sregistering (%u, %u, %d, %u) with local "
"rpcbind\n", (port ? "" : "un"),
prog, vers, prot, port);
- rpcb_clnt = rpcb_create("localhost", (struct sockaddr *) &sin,
- sizeof(sin), XPRT_TRANSPORT_UDP, 2, 1);
- if (IS_ERR(rpcb_clnt))
- return PTR_ERR(rpcb_clnt);
+ msg.rpc_proc = &rpcb_procedures2[RPCBPROC_UNSET];
+ if (port)
+ msg.rpc_proc = &rpcb_procedures2[RPCBPROC_SET];
- error = rpc_call_sync(rpcb_clnt, &msg, 0);
+ return rpcb_register_call((struct sockaddr *)&rpcb_inaddr_loopback,
+ sizeof(rpcb_inaddr_loopback),
+ RPCBVERS_2, &msg, okay);
+}
- rpc_shutdown_client(rpcb_clnt);
- if (error < 0)
- printk(KERN_WARNING "RPC: failed to contact local rpcbind "
- "server (errno %d).\n", -error);
- dprintk("RPC: registration status %d/%d\n", error, *okay);
+/*
+ * Fill in AF_INET family-specific arguments to register
+ */
+static int rpcb_register_netid4(struct sockaddr_in *address_to_register,
+ struct rpc_message *msg)
+{
+ struct rpcbind_args *map = msg->rpc_argp;
+ unsigned short port = ntohs(address_to_register->sin_port);
+ char buf[32];
+
+ /* Construct AF_INET universal address */
+ snprintf(buf, sizeof(buf),
+ NIPQUAD_FMT".%u.%u",
+ NIPQUAD(address_to_register->sin_addr.s_addr),
+ port >> 8, port & 0xff);
+ map->r_addr = buf;
+
+ dprintk("RPC: %sregistering [%u, %u, %s, '%s'] with "
+ "local rpcbind\n", (port ? "" : "un"),
+ map->r_prog, map->r_vers,
+ map->r_addr, map->r_netid);
+
+ msg->rpc_proc = &rpcb_procedures4[RPCBPROC_UNSET];
+ if (port)
+ msg->rpc_proc = &rpcb_procedures4[RPCBPROC_SET];
+
+ return rpcb_register_call((struct sockaddr *)&rpcb_inaddr_loopback,
+ sizeof(rpcb_inaddr_loopback),
+ RPCBVERS_4, msg, msg->rpc_resp);
+}
- return error;
+/*
+ * Fill in AF_INET6 family-specific arguments to register
+ */
+static int rpcb_register_netid6(struct sockaddr_in6 *address_to_register,
+ struct rpc_message *msg)
+{
+ struct rpcbind_args *map = msg->rpc_argp;
+ unsigned short port = ntohs(address_to_register->sin6_port);
+ char buf[64];
+
+ /* Construct AF_INET6 universal address */
+ snprintf(buf, sizeof(buf),
+ NIP6_FMT".%u.%u",
+ NIP6(address_to_register->sin6_addr),
+ port >> 8, port & 0xff);
+ map->r_addr = buf;
+
+ dprintk("RPC: %sregistering [%u, %u, %s, '%s'] with "
+ "local rpcbind\n", (port ? "" : "un"),
+ map->r_prog, map->r_vers,
+ map->r_addr, map->r_netid);
+
+ msg->rpc_proc = &rpcb_procedures4[RPCBPROC_UNSET];
+ if (port)
+ msg->rpc_proc = &rpcb_procedures4[RPCBPROC_SET];
+
+ return rpcb_register_call((struct sockaddr *)&rpcb_in6addr_loopback,
+ sizeof(rpcb_in6addr_loopback),
+ RPCBVERS_4, msg, msg->rpc_resp);
+}
+
+/**
+ * rpcb_v4_register - set or unset a port registration with the local rpcbind
+ * @program: RPC program number of service to (un)register
+ * @version: RPC version number of service to (un)register
+ * @address: address family, IP address, and port to (un)register
+ * @netid: netid of transport protocol to (un)register
+ * @result: result code from rpcbind RPC call
+ *
+ * RPC services invoke this function to advertise their contact
+ * information via the system's rpcbind daemon. RPC services
+ * invoke this function once for each [program, version, address,
+ * netid] tuple they wish to advertise.
+ *
+ * Callers may also unregister RPC services that are no longer
+ * available by setting the port number in the passed-in address
+ * to zero. Callers pass a netid of "" to unregister all
+ * transport netids associated with [program, version, address].
+ *
+ * Returns zero if the registration request was dispatched
+ * successfully and a reply was received. The rpcbind daemon's
+ * result code is stored in *result.
+ *
+ * Returns an errno value and sets *result to zero if there was
+ * some problem that prevented the rpcbind request from being
+ * dispatched, or if the rpcbind daemon did not respond within
+ * the timeout.
+ *
+ * This function uses rpcbind protocol version 4 to contact the
+ * local rpcbind daemon. The local rpcbind daemon must support
+ * version 4 of the rpcbind protocol in order for these functions
+ * to register a service successfully.
+ *
+ * Supported netids include "udp" and "tcp" for UDP and TCP over
+ * IPv4, and "udp6" and "tcp6" for UDP and TCP over IPv6,
+ * respectively.
+ *
+ * The contents of @address determine the address family and the
+ * port to be registered. The usual practice is to pass INADDR_ANY
+ * as the raw address, but specifying a non-zero address is also
+ * supported by this API if the caller wishes to advertise an RPC
+ * service on a specific network interface.
+ *
+ * Note that passing in INADDR_ANY does not create the same service
+ * registration as IN6ADDR_ANY. The former advertises an RPC
+ * service on any IPv4 address, but not on IPv6. The latter
+ * advertises the service on all IPv4 and IPv6 addresses.
+ */
+int rpcb_v4_register(const u32 program, const u32 version,
+ const struct sockaddr *address, const char *netid,
+ int *result)
+{
+ struct rpcbind_args map = {
+ .r_prog = program,
+ .r_vers = version,
+ .r_netid = netid,
+ .r_owner = RPCB_OWNER_STRING,
+ };
+ struct rpc_message msg = {
+ .rpc_argp = &map,
+ .rpc_resp = result,
+ };
+
+ *result = 0;
+
+ switch (address->sa_family) {
+ case AF_INET:
+ return rpcb_register_netid4((struct sockaddr_in *)address,
+ &msg);
+ case AF_INET6:
+ return rpcb_register_netid6((struct sockaddr_in6 *)address,
+ &msg);
+ }
+
+ return -EAFNOSUPPORT;
}
/**
@@ -227,7 +435,7 @@ int rpcb_getport_sync(struct sockaddr_in *sin, u32 prog, u32 vers, int prot)
__func__, NIPQUAD(sin->sin_addr.s_addr), prog, vers, prot);
rpcb_clnt = rpcb_create(NULL, (struct sockaddr *)sin,
- sizeof(*sin), prot, 2, 0);
+ sizeof(*sin), prot, RPCBVERS_2);
if (IS_ERR(rpcb_clnt))
return PTR_ERR(rpcb_clnt);
@@ -243,10 +451,10 @@ int rpcb_getport_sync(struct sockaddr_in *sin, u32 prog, u32 vers, int prot)
}
EXPORT_SYMBOL_GPL(rpcb_getport_sync);
-static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbind_args *map, int version)
+static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbind_args *map, struct rpc_procinfo *proc)
{
struct rpc_message msg = {
- .rpc_proc = rpcb_next_version[version].rpc_proc,
+ .rpc_proc = proc,
.rpc_argp = map,
.rpc_resp = &map->r_port,
};
@@ -271,6 +479,7 @@ static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbi
void rpcb_getport_async(struct rpc_task *task)
{
struct rpc_clnt *clnt = task->tk_client;
+ struct rpc_procinfo *proc;
u32 bind_version;
struct rpc_xprt *xprt = task->tk_xprt;
struct rpc_clnt *rpcb_clnt;
@@ -280,7 +489,6 @@ void rpcb_getport_async(struct rpc_task *task)
struct sockaddr *sap = (struct sockaddr *)&addr;
size_t salen;
int status;
- struct rpcb_info *info;
dprintk("RPC: %5u %s(%s, %u, %u, %d)\n",
task->tk_pid, __func__,
@@ -289,17 +497,16 @@ void rpcb_getport_async(struct rpc_task *task)
/* Autobind on cloned rpc clients is discouraged */
BUG_ON(clnt->cl_parent != clnt);
+ /* Put self on the wait queue to ensure we get notified if
+ * some other task is already attempting to bind the port */
+ rpc_sleep_on(&xprt->binding, task, NULL);
+
if (xprt_test_and_set_binding(xprt)) {
- status = -EAGAIN; /* tell caller to check again */
dprintk("RPC: %5u %s: waiting for another binder\n",
task->tk_pid, __func__);
- goto bailout_nowake;
+ return;
}
- /* Put self on queue before sending rpcbind request, in case
- * rpcb_getport_done completes before we return from rpc_run_task */
- rpc_sleep_on(&xprt->binding, task, NULL);
-
/* Someone else may have bound if we slept */
if (xprt_bound(xprt)) {
status = 0;
@@ -313,10 +520,12 @@ void rpcb_getport_async(struct rpc_task *task)
/* Don't ever use rpcbind v2 for AF_INET6 requests */
switch (sap->sa_family) {
case AF_INET:
- info = rpcb_next_version;
+ proc = rpcb_next_version[xprt->bind_index].rpc_proc;
+ bind_version = rpcb_next_version[xprt->bind_index].rpc_vers;
break;
case AF_INET6:
- info = rpcb_next_version6;
+ proc = rpcb_next_version6[xprt->bind_index].rpc_proc;
+ bind_version = rpcb_next_version6[xprt->bind_index].rpc_vers;
break;
default:
status = -EAFNOSUPPORT;
@@ -324,20 +533,19 @@ void rpcb_getport_async(struct rpc_task *task)
task->tk_pid, __func__);
goto bailout_nofree;
}
- if (info[xprt->bind_index].rpc_proc == NULL) {
+ if (proc == NULL) {
xprt->bind_index = 0;
status = -EPFNOSUPPORT;
dprintk("RPC: %5u %s: no more getport versions available\n",
task->tk_pid, __func__);
goto bailout_nofree;
}
- bind_version = info[xprt->bind_index].rpc_vers;
dprintk("RPC: %5u %s: trying rpcbind version %u\n",
task->tk_pid, __func__, bind_version);
rpcb_clnt = rpcb_create(clnt->cl_server, sap, salen, xprt->prot,
- bind_version, 0);
+ bind_version);
if (IS_ERR(rpcb_clnt)) {
status = PTR_ERR(rpcb_clnt);
dprintk("RPC: %5u %s: rpcb_create failed, error %ld\n",
@@ -360,26 +568,23 @@ void rpcb_getport_async(struct rpc_task *task)
map->r_netid = rpc_peeraddr2str(clnt, RPC_DISPLAY_NETID);
map->r_addr = rpc_peeraddr2str(rpcb_clnt, RPC_DISPLAY_UNIVERSAL_ADDR);
map->r_owner = RPCB_OWNER_STRING; /* ignored for GETADDR */
+ map->r_status = -EIO;
- child = rpcb_call_async(rpcb_clnt, map, xprt->bind_index);
+ child = rpcb_call_async(rpcb_clnt, map, proc);
rpc_release_client(rpcb_clnt);
if (IS_ERR(child)) {
- status = -EIO;
+ /* rpcb_map_release() has freed the arguments */
dprintk("RPC: %5u %s: rpc_run_task failed\n",
task->tk_pid, __func__);
- goto bailout;
+ return;
}
rpc_put_task(child);
task->tk_xprt->stat.bind_count++;
return;
-bailout:
- kfree(map);
- xprt_put(xprt);
bailout_nofree:
rpcb_wake_rpcbind_waiters(xprt, status);
-bailout_nowake:
task->tk_status = status;
}
EXPORT_SYMBOL_GPL(rpcb_getport_async);
@@ -418,9 +623,13 @@ static void rpcb_getport_done(struct rpc_task *child, void *data)
dprintk("RPC: %5u rpcb_getport_done(status %d, port %u)\n",
child->tk_pid, status, map->r_port);
- rpcb_wake_rpcbind_waiters(xprt, status);
+ map->r_status = status;
}
+/*
+ * XDR functions for rpcbind
+ */
+
static int rpcb_encode_mapping(struct rpc_rqst *req, __be32 *p,
struct rpcbind_args *rpcb)
{
@@ -439,7 +648,7 @@ static int rpcb_decode_getport(struct rpc_rqst *req, __be32 *p,
unsigned short *portp)
{
*portp = (unsigned short) ntohl(*p++);
- dprintk("RPC: rpcb_decode_getport result %u\n",
+ dprintk("RPC: rpcb_decode_getport result %u\n",
*portp);
return 0;
}
@@ -448,8 +657,8 @@ static int rpcb_decode_set(struct rpc_rqst *req, __be32 *p,
unsigned int *boolp)
{
*boolp = (unsigned int) ntohl(*p++);
- dprintk("RPC: rpcb_decode_set result %u\n",
- *boolp);
+ dprintk("RPC: rpcb_decode_set: call %s\n",
+ (*boolp ? "succeeded" : "failed"));
return 0;
}
@@ -572,52 +781,60 @@ out_err:
static struct rpc_procinfo rpcb_procedures2[] = {
PROC(SET, mapping, set),
PROC(UNSET, mapping, set),
- PROC(GETADDR, mapping, getport),
+ PROC(GETPORT, mapping, getport),
};
static struct rpc_procinfo rpcb_procedures3[] = {
- PROC(SET, mapping, set),
- PROC(UNSET, mapping, set),
+ PROC(SET, getaddr, set),
+ PROC(UNSET, getaddr, set),
PROC(GETADDR, getaddr, getaddr),
};
static struct rpc_procinfo rpcb_procedures4[] = {
- PROC(SET, mapping, set),
- PROC(UNSET, mapping, set),
+ PROC(SET, getaddr, set),
+ PROC(UNSET, getaddr, set),
+ PROC(GETADDR, getaddr, getaddr),
PROC(GETVERSADDR, getaddr, getaddr),
};
static struct rpcb_info rpcb_next_version[] = {
-#ifdef CONFIG_SUNRPC_BIND34
- { 4, &rpcb_procedures4[RPCBPROC_GETVERSADDR] },
- { 3, &rpcb_procedures3[RPCBPROC_GETADDR] },
-#endif
- { 2, &rpcb_procedures2[RPCBPROC_GETPORT] },
- { 0, NULL },
+ {
+ .rpc_vers = RPCBVERS_2,
+ .rpc_proc = &rpcb_procedures2[RPCBPROC_GETPORT],
+ },
+ {
+ .rpc_proc = NULL,
+ },
};
static struct rpcb_info rpcb_next_version6[] = {
-#ifdef CONFIG_SUNRPC_BIND34
- { 4, &rpcb_procedures4[RPCBPROC_GETVERSADDR] },
- { 3, &rpcb_procedures3[RPCBPROC_GETADDR] },
-#endif
- { 0, NULL },
+ {
+ .rpc_vers = RPCBVERS_4,
+ .rpc_proc = &rpcb_procedures4[RPCBPROC_GETADDR],
+ },
+ {
+ .rpc_vers = RPCBVERS_3,
+ .rpc_proc = &rpcb_procedures3[RPCBPROC_GETADDR],
+ },
+ {
+ .rpc_proc = NULL,
+ },
};
static struct rpc_version rpcb_version2 = {
- .number = 2,
+ .number = RPCBVERS_2,
.nrprocs = RPCB_HIGHPROC_2,
.procs = rpcb_procedures2
};
static struct rpc_version rpcb_version3 = {
- .number = 3,
+ .number = RPCBVERS_3,
.nrprocs = RPCB_HIGHPROC_3,
.procs = rpcb_procedures3
};
static struct rpc_version rpcb_version4 = {
- .number = 4,
+ .number = RPCBVERS_4,
.nrprocs = RPCB_HIGHPROC_4,
.procs = rpcb_procedures4
};
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 6eab9bf94baf..385f427bedad 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -576,9 +576,7 @@ EXPORT_SYMBOL_GPL(rpc_delay);
*/
static void rpc_prepare_task(struct rpc_task *task)
{
- lock_kernel();
task->tk_ops->rpc_call_prepare(task, task->tk_calldata);
- unlock_kernel();
}
/*
@@ -588,9 +586,7 @@ void rpc_exit_task(struct rpc_task *task)
{
task->tk_action = NULL;
if (task->tk_ops->rpc_call_done != NULL) {
- lock_kernel();
task->tk_ops->rpc_call_done(task, task->tk_calldata);
- unlock_kernel();
if (task->tk_action != NULL) {
WARN_ON(RPC_ASSASSINATED(task));
/* Always release the RPC slot and buffer memory */
@@ -602,11 +598,8 @@ EXPORT_SYMBOL_GPL(rpc_exit_task);
void rpc_release_calldata(const struct rpc_call_ops *ops, void *calldata)
{
- if (ops->rpc_release != NULL) {
- lock_kernel();
+ if (ops->rpc_release != NULL)
ops->rpc_release(calldata);
- unlock_kernel();
- }
}
/*
@@ -626,19 +619,15 @@ static void __rpc_execute(struct rpc_task *task)
/*
* Execute any pending callback.
*/
- if (RPC_DO_CALLBACK(task)) {
- /* Define a callback save pointer */
+ if (task->tk_callback) {
void (*save_callback)(struct rpc_task *);
/*
- * If a callback exists, save it, reset it,
- * call it.
- * The save is needed to stop from resetting
- * another callback set within the callback handler
- * - Dave
+ * We set tk_callback to NULL before calling it,
+ * in case it sets the tk_callback field itself:
*/
- save_callback=task->tk_callback;
- task->tk_callback=NULL;
+ save_callback = task->tk_callback;
+ task->tk_callback = NULL;
save_callback(task);
}
diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
index 01c7e311b904..5a32cb7c4bb4 100644
--- a/net/sunrpc/svc.c
+++ b/net/sunrpc/svc.c
@@ -18,6 +18,7 @@
#include <linux/mm.h>
#include <linux/interrupt.h>
#include <linux/module.h>
+#include <linux/kthread.h>
#include <linux/sunrpc/types.h>
#include <linux/sunrpc/xdr.h>
@@ -291,15 +292,14 @@ svc_pool_map_put(void)
/*
- * Set the current thread's cpus_allowed mask so that it
+ * Set the given thread's cpus_allowed mask so that it
* will only run on cpus in the given pool.
- *
- * Returns 1 and fills in oldmask iff a cpumask was applied.
*/
-static inline int
-svc_pool_map_set_cpumask(unsigned int pidx, cpumask_t *oldmask)
+static inline void
+svc_pool_map_set_cpumask(struct task_struct *task, unsigned int pidx)
{
struct svc_pool_map *m = &svc_pool_map;
+ unsigned int node = m->pool_to[pidx];
/*
* The caller checks for sv_nrpools > 1, which
@@ -307,26 +307,17 @@ svc_pool_map_set_cpumask(unsigned int pidx, cpumask_t *oldmask)
*/
BUG_ON(m->count == 0);
- switch (m->mode)
- {
- default:
- return 0;
+ switch (m->mode) {
case SVC_POOL_PERCPU:
{
- unsigned int cpu = m->pool_to[pidx];
-
- *oldmask = current->cpus_allowed;
- set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
- return 1;
+ set_cpus_allowed_ptr(task, &cpumask_of_cpu(node));
+ break;
}
case SVC_POOL_PERNODE:
{
- unsigned int node = m->pool_to[pidx];
node_to_cpumask_ptr(nodecpumask, node);
-
- *oldmask = current->cpus_allowed;
- set_cpus_allowed_ptr(current, nodecpumask);
- return 1;
+ set_cpus_allowed_ptr(task, nodecpumask);
+ break;
}
}
}
@@ -443,7 +434,7 @@ EXPORT_SYMBOL(svc_create);
struct svc_serv *
svc_create_pooled(struct svc_program *prog, unsigned int bufsize,
void (*shutdown)(struct svc_serv *serv),
- svc_thread_fn func, int sig, struct module *mod)
+ svc_thread_fn func, struct module *mod)
{
struct svc_serv *serv;
unsigned int npools = svc_pool_map_get();
@@ -452,7 +443,6 @@ svc_create_pooled(struct svc_program *prog, unsigned int bufsize,
if (serv != NULL) {
serv->sv_function = func;
- serv->sv_kill_signal = sig;
serv->sv_module = mod;
}
@@ -461,7 +451,8 @@ svc_create_pooled(struct svc_program *prog, unsigned int bufsize,
EXPORT_SYMBOL(svc_create_pooled);
/*
- * Destroy an RPC service. Should be called with the BKL held
+ * Destroy an RPC service. Should be called with appropriate locking to
+ * protect the sv_nrthreads, sv_permsocks and sv_tempsocks.
*/
void
svc_destroy(struct svc_serv *serv)
@@ -578,46 +569,6 @@ out_enomem:
EXPORT_SYMBOL(svc_prepare_thread);
/*
- * Create a thread in the given pool. Caller must hold BKL.
- * On a NUMA or SMP machine, with a multi-pool serv, the thread
- * will be restricted to run on the cpus belonging to the pool.
- */
-static int
-__svc_create_thread(svc_thread_fn func, struct svc_serv *serv,
- struct svc_pool *pool)
-{
- struct svc_rqst *rqstp;
- int error = -ENOMEM;
- int have_oldmask = 0;
- cpumask_t uninitialized_var(oldmask);
-
- rqstp = svc_prepare_thread(serv, pool);
- if (IS_ERR(rqstp)) {
- error = PTR_ERR(rqstp);
- goto out;
- }
-
- if (serv->sv_nrpools > 1)
- have_oldmask = svc_pool_map_set_cpumask(pool->sp_id, &oldmask);
-
- error = kernel_thread((int (*)(void *)) func, rqstp, 0);
-
- if (have_oldmask)
- set_cpus_allowed(current, oldmask);
-
- if (error < 0)
- goto out_thread;
- svc_sock_update_bufs(serv);
- error = 0;
-out:
- return error;
-
-out_thread:
- svc_exit_thread(rqstp);
- goto out;
-}
-
-/*
* Choose a pool in which to create a new thread, for svc_set_num_threads
*/
static inline struct svc_pool *
@@ -674,7 +625,7 @@ found_pool:
* of threads the given number. If `pool' is non-NULL, applies
* only to threads in that pool, otherwise round-robins between
* all pools. Must be called with a svc_get() reference and
- * the BKL held.
+ * the BKL or another lock to protect access to svc_serv fields.
*
* Destroying threads relies on the service threads filling in
* rqstp->rq_task, which only the nfs ones do. Assumes the serv
@@ -686,7 +637,9 @@ found_pool:
int
svc_set_num_threads(struct svc_serv *serv, struct svc_pool *pool, int nrservs)
{
- struct task_struct *victim;
+ struct svc_rqst *rqstp;
+ struct task_struct *task;
+ struct svc_pool *chosen_pool;
int error = 0;
unsigned int state = serv->sv_nrthreads-1;
@@ -702,18 +655,34 @@ svc_set_num_threads(struct svc_serv *serv, struct svc_pool *pool, int nrservs)
/* create new threads */
while (nrservs > 0) {
nrservs--;
+ chosen_pool = choose_pool(serv, pool, &state);
+
+ rqstp = svc_prepare_thread(serv, chosen_pool);
+ if (IS_ERR(rqstp)) {
+ error = PTR_ERR(rqstp);
+ break;
+ }
+
__module_get(serv->sv_module);
- error = __svc_create_thread(serv->sv_function, serv,
- choose_pool(serv, pool, &state));
- if (error < 0) {
+ task = kthread_create(serv->sv_function, rqstp, serv->sv_name);
+ if (IS_ERR(task)) {
+ error = PTR_ERR(task);
module_put(serv->sv_module);
+ svc_exit_thread(rqstp);
break;
}
+
+ rqstp->rq_task = task;
+ if (serv->sv_nrpools > 1)
+ svc_pool_map_set_cpumask(task, chosen_pool->sp_id);
+
+ svc_sock_update_bufs(serv);
+ wake_up_process(task);
}
/* destroy old threads */
while (nrservs < 0 &&
- (victim = choose_victim(serv, pool, &state)) != NULL) {
- send_sig(serv->sv_kill_signal, victim, 1);
+ (task = choose_victim(serv, pool, &state)) != NULL) {
+ send_sig(SIGINT, task, 1);
nrservs++;
}
@@ -722,7 +691,8 @@ svc_set_num_threads(struct svc_serv *serv, struct svc_pool *pool, int nrservs)
EXPORT_SYMBOL(svc_set_num_threads);
/*
- * Called from a server thread as it's exiting. Caller must hold BKL.
+ * Called from a server thread as it's exiting. Caller must hold the BKL or
+ * the "service mutex", whichever is appropriate for the service.
*/
void
svc_exit_thread(struct svc_rqst *rqstp)
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index e1770f7ba0b3..99a52aabe332 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -690,7 +690,7 @@ static void xprt_connect_status(struct rpc_task *task)
{
struct rpc_xprt *xprt = task->tk_xprt;
- if (task->tk_status >= 0) {
+ if (task->tk_status == 0) {
xprt->stat.connect_count++;
xprt->stat.connect_time += (long)jiffies - xprt->stat.connect_start;
dprintk("RPC: %5u xprt_connect_status: connection established\n",
@@ -699,12 +699,6 @@ static void xprt_connect_status(struct rpc_task *task)
}
switch (task->tk_status) {
- case -ECONNREFUSED:
- case -ECONNRESET:
- dprintk("RPC: %5u xprt_connect_status: server %s refused "
- "connection\n", task->tk_pid,
- task->tk_client->cl_server);
- break;
case -ENOTCONN:
dprintk("RPC: %5u xprt_connect_status: connection broken\n",
task->tk_pid);
@@ -878,6 +872,7 @@ void xprt_transmit(struct rpc_task *task)
return;
req->rq_connect_cookie = xprt->connect_cookie;
+ req->rq_xtime = jiffies;
status = xprt->ops->send_request(task);
if (status == 0) {
dprintk("RPC: %5u xmit complete\n", task->tk_pid);
diff --git a/net/sunrpc/xprtrdma/svc_rdma.c b/net/sunrpc/xprtrdma/svc_rdma.c
index 88c0ca20bb1e..87101177825b 100644
--- a/net/sunrpc/xprtrdma/svc_rdma.c
+++ b/net/sunrpc/xprtrdma/svc_rdma.c
@@ -69,6 +69,10 @@ atomic_t rdma_stat_rq_prod;
atomic_t rdma_stat_sq_poll;
atomic_t rdma_stat_sq_prod;
+/* Temporary NFS request map and context caches */
+struct kmem_cache *svc_rdma_map_cachep;
+struct kmem_cache *svc_rdma_ctxt_cachep;
+
/*
* This function implements reading and resetting an atomic_t stat
* variable through read/write to a proc file. Any write to the file
@@ -236,11 +240,14 @@ static ctl_table svcrdma_root_table[] = {
void svc_rdma_cleanup(void)
{
dprintk("SVCRDMA Module Removed, deregister RPC RDMA transport\n");
+ flush_scheduled_work();
if (svcrdma_table_header) {
unregister_sysctl_table(svcrdma_table_header);
svcrdma_table_header = NULL;
}
svc_unreg_xprt_class(&svc_rdma_class);
+ kmem_cache_destroy(svc_rdma_map_cachep);
+ kmem_cache_destroy(svc_rdma_ctxt_cachep);
}
int svc_rdma_init(void)
@@ -255,9 +262,37 @@ int svc_rdma_init(void)
svcrdma_table_header =
register_sysctl_table(svcrdma_root_table);
+ /* Create the temporary map cache */
+ svc_rdma_map_cachep = kmem_cache_create("svc_rdma_map_cache",
+ sizeof(struct svc_rdma_req_map),
+ 0,
+ SLAB_HWCACHE_ALIGN,
+ NULL);
+ if (!svc_rdma_map_cachep) {
+ printk(KERN_INFO "Could not allocate map cache.\n");
+ goto err0;
+ }
+
+ /* Create the temporary context cache */
+ svc_rdma_ctxt_cachep =
+ kmem_cache_create("svc_rdma_ctxt_cache",
+ sizeof(struct svc_rdma_op_ctxt),
+ 0,
+ SLAB_HWCACHE_ALIGN,
+ NULL);
+ if (!svc_rdma_ctxt_cachep) {
+ printk(KERN_INFO "Could not allocate WR ctxt cache.\n");
+ goto err1;
+ }
+
/* Register RDMA with the SVC transport switch */
svc_reg_xprt_class(&svc_rdma_class);
return 0;
+ err1:
+ kmem_cache_destroy(svc_rdma_map_cachep);
+ err0:
+ unregister_sysctl_table(svcrdma_table_header);
+ return -ENOMEM;
}
MODULE_AUTHOR("Tom Tucker <tom@opengridcomputing.com>");
MODULE_DESCRIPTION("SVC RDMA Transport");
diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
index 06ab4841537b..b4b17f44cb29 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
@@ -112,11 +112,6 @@ static void rdma_build_arg_xdr(struct svc_rqst *rqstp,
rqstp->rq_arg.tail[0].iov_len = 0;
}
-struct chunk_sge {
- int start; /* sge no for this chunk */
- int count; /* sge count for this chunk */
-};
-
/* Encode a read-chunk-list as an array of IB SGE
*
* Assumptions:
@@ -134,8 +129,8 @@ static int rdma_rcl_to_sge(struct svcxprt_rdma *xprt,
struct svc_rqst *rqstp,
struct svc_rdma_op_ctxt *head,
struct rpcrdma_msg *rmsgp,
- struct ib_sge *sge,
- struct chunk_sge *ch_sge_ary,
+ struct svc_rdma_req_map *rpl_map,
+ struct svc_rdma_req_map *chl_map,
int ch_count,
int byte_count)
{
@@ -156,22 +151,18 @@ static int rdma_rcl_to_sge(struct svcxprt_rdma *xprt,
head->arg.head[0] = rqstp->rq_arg.head[0];
head->arg.tail[0] = rqstp->rq_arg.tail[0];
head->arg.pages = &head->pages[head->count];
- head->sge[0].length = head->count; /* save count of hdr pages */
+ head->hdr_count = head->count; /* save count of hdr pages */
head->arg.page_base = 0;
head->arg.page_len = ch_bytes;
head->arg.len = rqstp->rq_arg.len + ch_bytes;
head->arg.buflen = rqstp->rq_arg.buflen + ch_bytes;
head->count++;
- ch_sge_ary[0].start = 0;
+ chl_map->ch[0].start = 0;
while (byte_count) {
+ rpl_map->sge[sge_no].iov_base =
+ page_address(rqstp->rq_arg.pages[page_no]) + page_off;
sge_bytes = min_t(int, PAGE_SIZE-page_off, ch_bytes);
- sge[sge_no].addr =
- ib_dma_map_page(xprt->sc_cm_id->device,
- rqstp->rq_arg.pages[page_no],
- page_off, sge_bytes,
- DMA_FROM_DEVICE);
- sge[sge_no].length = sge_bytes;
- sge[sge_no].lkey = xprt->sc_phys_mr->lkey;
+ rpl_map->sge[sge_no].iov_len = sge_bytes;
/*
* Don't bump head->count here because the same page
* may be used by multiple SGE.
@@ -187,11 +178,11 @@ static int rdma_rcl_to_sge(struct svcxprt_rdma *xprt,
* SGE, move to the next SGE
*/
if (ch_bytes == 0) {
- ch_sge_ary[ch_no].count =
- sge_no - ch_sge_ary[ch_no].start;
+ chl_map->ch[ch_no].count =
+ sge_no - chl_map->ch[ch_no].start;
ch_no++;
ch++;
- ch_sge_ary[ch_no].start = sge_no;
+ chl_map->ch[ch_no].start = sge_no;
ch_bytes = ch->rc_target.rs_length;
/* If bytes remaining account for next chunk */
if (byte_count) {
@@ -220,18 +211,25 @@ static int rdma_rcl_to_sge(struct svcxprt_rdma *xprt,
return sge_no;
}
-static void rdma_set_ctxt_sge(struct svc_rdma_op_ctxt *ctxt,
- struct ib_sge *sge,
+static void rdma_set_ctxt_sge(struct svcxprt_rdma *xprt,
+ struct svc_rdma_op_ctxt *ctxt,
+ struct kvec *vec,
u64 *sgl_offset,
int count)
{
int i;
ctxt->count = count;
+ ctxt->direction = DMA_FROM_DEVICE;
for (i = 0; i < count; i++) {
- ctxt->sge[i].addr = sge[i].addr;
- ctxt->sge[i].length = sge[i].length;
- *sgl_offset = *sgl_offset + sge[i].length;
+ atomic_inc(&xprt->sc_dma_used);
+ ctxt->sge[i].addr =
+ ib_dma_map_single(xprt->sc_cm_id->device,
+ vec[i].iov_base, vec[i].iov_len,
+ DMA_FROM_DEVICE);
+ ctxt->sge[i].length = vec[i].iov_len;
+ ctxt->sge[i].lkey = xprt->sc_phys_mr->lkey;
+ *sgl_offset = *sgl_offset + vec[i].iov_len;
}
}
@@ -282,34 +280,29 @@ static int rdma_read_xdr(struct svcxprt_rdma *xprt,
struct ib_send_wr read_wr;
int err = 0;
int ch_no;
- struct ib_sge *sge;
int ch_count;
int byte_count;
int sge_count;
u64 sgl_offset;
struct rpcrdma_read_chunk *ch;
struct svc_rdma_op_ctxt *ctxt = NULL;
- struct svc_rdma_op_ctxt *tmp_sge_ctxt;
- struct svc_rdma_op_ctxt *tmp_ch_ctxt;
- struct chunk_sge *ch_sge_ary;
+ struct svc_rdma_req_map *rpl_map;
+ struct svc_rdma_req_map *chl_map;
/* If no read list is present, return 0 */
ch = svc_rdma_get_read_chunk(rmsgp);
if (!ch)
return 0;
- /* Allocate temporary contexts to keep SGE */
- BUG_ON(sizeof(struct ib_sge) < sizeof(struct chunk_sge));
- tmp_sge_ctxt = svc_rdma_get_context(xprt);
- sge = tmp_sge_ctxt->sge;
- tmp_ch_ctxt = svc_rdma_get_context(xprt);
- ch_sge_ary = (struct chunk_sge *)tmp_ch_ctxt->sge;
+ /* Allocate temporary reply and chunk maps */
+ rpl_map = svc_rdma_get_req_map();
+ chl_map = svc_rdma_get_req_map();
svc_rdma_rcl_chunk_counts(ch, &ch_count, &byte_count);
if (ch_count > RPCSVC_MAXPAGES)
return -EINVAL;
sge_count = rdma_rcl_to_sge(xprt, rqstp, hdr_ctxt, rmsgp,
- sge, ch_sge_ary,
+ rpl_map, chl_map,
ch_count, byte_count);
sgl_offset = 0;
ch_no = 0;
@@ -331,14 +324,15 @@ next_sge:
read_wr.wr.rdma.remote_addr =
get_unaligned(&(ch->rc_target.rs_offset)) +
sgl_offset;
- read_wr.sg_list = &sge[ch_sge_ary[ch_no].start];
+ read_wr.sg_list = ctxt->sge;
read_wr.num_sge =
- rdma_read_max_sge(xprt, ch_sge_ary[ch_no].count);
- rdma_set_ctxt_sge(ctxt, &sge[ch_sge_ary[ch_no].start],
+ rdma_read_max_sge(xprt, chl_map->ch[ch_no].count);
+ rdma_set_ctxt_sge(xprt, ctxt,
+ &rpl_map->sge[chl_map->ch[ch_no].start],
&sgl_offset,
read_wr.num_sge);
if (((ch+1)->rc_discrim == 0) &&
- (read_wr.num_sge == ch_sge_ary[ch_no].count)) {
+ (read_wr.num_sge == chl_map->ch[ch_no].count)) {
/*
* Mark the last RDMA_READ with a bit to
* indicate all RPC data has been fetched from
@@ -358,9 +352,9 @@ next_sge:
}
atomic_inc(&rdma_stat_read);
- if (read_wr.num_sge < ch_sge_ary[ch_no].count) {
- ch_sge_ary[ch_no].count -= read_wr.num_sge;
- ch_sge_ary[ch_no].start += read_wr.num_sge;
+ if (read_wr.num_sge < chl_map->ch[ch_no].count) {
+ chl_map->ch[ch_no].count -= read_wr.num_sge;
+ chl_map->ch[ch_no].start += read_wr.num_sge;
goto next_sge;
}
sgl_offset = 0;
@@ -368,8 +362,8 @@ next_sge:
}
out:
- svc_rdma_put_context(tmp_sge_ctxt, 0);
- svc_rdma_put_context(tmp_ch_ctxt, 0);
+ svc_rdma_put_req_map(rpl_map);
+ svc_rdma_put_req_map(chl_map);
/* Detach arg pages. svc_recv will replenish them */
for (ch_no = 0; &rqstp->rq_pages[ch_no] < rqstp->rq_respages; ch_no++)
@@ -399,7 +393,7 @@ static int rdma_read_complete(struct svc_rqst *rqstp,
rqstp->rq_pages[page_no] = head->pages[page_no];
}
/* Point rq_arg.pages past header */
- rqstp->rq_arg.pages = &rqstp->rq_pages[head->sge[0].length];
+ rqstp->rq_arg.pages = &rqstp->rq_pages[head->hdr_count];
rqstp->rq_arg.page_len = head->arg.page_len;
rqstp->rq_arg.page_base = head->arg.page_base;
diff --git a/net/sunrpc/xprtrdma/svc_rdma_sendto.c b/net/sunrpc/xprtrdma/svc_rdma_sendto.c
index fb82b1b683f8..84d328329d98 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_sendto.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_sendto.c
@@ -63,52 +63,44 @@
* SGE[2..sge_count-2] data from xdr->pages[]
* SGE[sge_count-1] data from xdr->tail.
*
+ * The max SGE we need is the length of the XDR / pagesize + one for
+ * head + one for tail + one for RPCRDMA header. Since RPCSVC_MAXPAGES
+ * reserves a page for both the request and the reply header, and this
+ * array is only concerned with the reply we are assured that we have
+ * on extra page for the RPCRMDA header.
*/
-static struct ib_sge *xdr_to_sge(struct svcxprt_rdma *xprt,
- struct xdr_buf *xdr,
- struct ib_sge *sge,
- int *sge_count)
+static void xdr_to_sge(struct svcxprt_rdma *xprt,
+ struct xdr_buf *xdr,
+ struct svc_rdma_req_map *vec)
{
- /* Max we need is the length of the XDR / pagesize + one for
- * head + one for tail + one for RPCRDMA header
- */
int sge_max = (xdr->len+PAGE_SIZE-1) / PAGE_SIZE + 3;
int sge_no;
- u32 byte_count = xdr->len;
u32 sge_bytes;
u32 page_bytes;
- int page_off;
+ u32 page_off;
int page_no;
+ BUG_ON(xdr->len !=
+ (xdr->head[0].iov_len + xdr->page_len + xdr->tail[0].iov_len));
+
/* Skip the first sge, this is for the RPCRDMA header */
sge_no = 1;
/* Head SGE */
- sge[sge_no].addr = ib_dma_map_single(xprt->sc_cm_id->device,
- xdr->head[0].iov_base,
- xdr->head[0].iov_len,
- DMA_TO_DEVICE);
- sge_bytes = min_t(u32, byte_count, xdr->head[0].iov_len);
- byte_count -= sge_bytes;
- sge[sge_no].length = sge_bytes;
- sge[sge_no].lkey = xprt->sc_phys_mr->lkey;
+ vec->sge[sge_no].iov_base = xdr->head[0].iov_base;
+ vec->sge[sge_no].iov_len = xdr->head[0].iov_len;
sge_no++;
/* pages SGE */
page_no = 0;
page_bytes = xdr->page_len;
page_off = xdr->page_base;
- while (byte_count && page_bytes) {
- sge_bytes = min_t(u32, byte_count, (PAGE_SIZE-page_off));
- sge[sge_no].addr =
- ib_dma_map_page(xprt->sc_cm_id->device,
- xdr->pages[page_no], page_off,
- sge_bytes, DMA_TO_DEVICE);
- sge_bytes = min(sge_bytes, page_bytes);
- byte_count -= sge_bytes;
+ while (page_bytes) {
+ vec->sge[sge_no].iov_base =
+ page_address(xdr->pages[page_no]) + page_off;
+ sge_bytes = min_t(u32, page_bytes, (PAGE_SIZE - page_off));
page_bytes -= sge_bytes;
- sge[sge_no].length = sge_bytes;
- sge[sge_no].lkey = xprt->sc_phys_mr->lkey;
+ vec->sge[sge_no].iov_len = sge_bytes;
sge_no++;
page_no++;
@@ -116,36 +108,24 @@ static struct ib_sge *xdr_to_sge(struct svcxprt_rdma *xprt,
}
/* Tail SGE */
- if (byte_count && xdr->tail[0].iov_len) {
- sge[sge_no].addr =
- ib_dma_map_single(xprt->sc_cm_id->device,
- xdr->tail[0].iov_base,
- xdr->tail[0].iov_len,
- DMA_TO_DEVICE);
- sge_bytes = min_t(u32, byte_count, xdr->tail[0].iov_len);
- byte_count -= sge_bytes;
- sge[sge_no].length = sge_bytes;
- sge[sge_no].lkey = xprt->sc_phys_mr->lkey;
+ if (xdr->tail[0].iov_len) {
+ vec->sge[sge_no].iov_base = xdr->tail[0].iov_base;
+ vec->sge[sge_no].iov_len = xdr->tail[0].iov_len;
sge_no++;
}
BUG_ON(sge_no > sge_max);
- BUG_ON(byte_count != 0);
-
- *sge_count = sge_no;
- return sge;
+ vec->count = sge_no;
}
-
/* Assumptions:
* - The specified write_len can be represented in sc_max_sge * PAGE_SIZE
*/
static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp,
u32 rmr, u64 to,
u32 xdr_off, int write_len,
- struct ib_sge *xdr_sge, int sge_count)
+ struct svc_rdma_req_map *vec)
{
- struct svc_rdma_op_ctxt *tmp_sge_ctxt;
struct ib_send_wr write_wr;
struct ib_sge *sge;
int xdr_sge_no;
@@ -154,25 +134,23 @@ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp,
int sge_off;
int bc;
struct svc_rdma_op_ctxt *ctxt;
- int ret = 0;
- BUG_ON(sge_count > RPCSVC_MAXPAGES);
+ BUG_ON(vec->count > RPCSVC_MAXPAGES);
dprintk("svcrdma: RDMA_WRITE rmr=%x, to=%llx, xdr_off=%d, "
- "write_len=%d, xdr_sge=%p, sge_count=%d\n",
+ "write_len=%d, vec->sge=%p, vec->count=%lu\n",
rmr, (unsigned long long)to, xdr_off,
- write_len, xdr_sge, sge_count);
+ write_len, vec->sge, vec->count);
ctxt = svc_rdma_get_context(xprt);
- ctxt->count = 0;
- tmp_sge_ctxt = svc_rdma_get_context(xprt);
- sge = tmp_sge_ctxt->sge;
+ ctxt->direction = DMA_TO_DEVICE;
+ sge = ctxt->sge;
/* Find the SGE associated with xdr_off */
- for (bc = xdr_off, xdr_sge_no = 1; bc && xdr_sge_no < sge_count;
+ for (bc = xdr_off, xdr_sge_no = 1; bc && xdr_sge_no < vec->count;
xdr_sge_no++) {
- if (xdr_sge[xdr_sge_no].length > bc)
+ if (vec->sge[xdr_sge_no].iov_len > bc)
break;
- bc -= xdr_sge[xdr_sge_no].length;
+ bc -= vec->sge[xdr_sge_no].iov_len;
}
sge_off = bc;
@@ -180,21 +158,29 @@ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp,
sge_no = 0;
/* Copy the remaining SGE */
- while (bc != 0 && xdr_sge_no < sge_count) {
- sge[sge_no].addr = xdr_sge[xdr_sge_no].addr + sge_off;
- sge[sge_no].lkey = xdr_sge[xdr_sge_no].lkey;
+ while (bc != 0 && xdr_sge_no < vec->count) {
+ sge[sge_no].lkey = xprt->sc_phys_mr->lkey;
sge_bytes = min((size_t)bc,
- (size_t)(xdr_sge[xdr_sge_no].length-sge_off));
+ (size_t)(vec->sge[xdr_sge_no].iov_len-sge_off));
sge[sge_no].length = sge_bytes;
-
+ atomic_inc(&xprt->sc_dma_used);
+ sge[sge_no].addr =
+ ib_dma_map_single(xprt->sc_cm_id->device,
+ (void *)
+ vec->sge[xdr_sge_no].iov_base + sge_off,
+ sge_bytes, DMA_TO_DEVICE);
+ if (dma_mapping_error(xprt->sc_cm_id->device->dma_device,
+ sge[sge_no].addr))
+ goto err;
sge_off = 0;
sge_no++;
+ ctxt->count++;
xdr_sge_no++;
bc -= sge_bytes;
}
BUG_ON(bc != 0);
- BUG_ON(xdr_sge_no > sge_count);
+ BUG_ON(xdr_sge_no > vec->count);
/* Prepare WRITE WR */
memset(&write_wr, 0, sizeof write_wr);
@@ -209,21 +195,20 @@ static int send_write(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp,
/* Post It */
atomic_inc(&rdma_stat_write);
- if (svc_rdma_send(xprt, &write_wr)) {
- svc_rdma_put_context(ctxt, 1);
- /* Fatal error, close transport */
- ret = -EIO;
- }
- svc_rdma_put_context(tmp_sge_ctxt, 0);
- return ret;
+ if (svc_rdma_send(xprt, &write_wr))
+ goto err;
+ return 0;
+ err:
+ svc_rdma_put_context(ctxt, 0);
+ /* Fatal error, close transport */
+ return -EIO;
}
static int send_write_chunks(struct svcxprt_rdma *xprt,
struct rpcrdma_msg *rdma_argp,
struct rpcrdma_msg *rdma_resp,
struct svc_rqst *rqstp,
- struct ib_sge *sge,
- int sge_count)
+ struct svc_rdma_req_map *vec)
{
u32 xfer_len = rqstp->rq_res.page_len + rqstp->rq_res.tail[0].iov_len;
int write_len;
@@ -269,8 +254,7 @@ static int send_write_chunks(struct svcxprt_rdma *xprt,
rs_offset + chunk_off,
xdr_off,
this_write,
- sge,
- sge_count);
+ vec);
if (ret) {
dprintk("svcrdma: RDMA_WRITE failed, ret=%d\n",
ret);
@@ -292,8 +276,7 @@ static int send_reply_chunks(struct svcxprt_rdma *xprt,
struct rpcrdma_msg *rdma_argp,
struct rpcrdma_msg *rdma_resp,
struct svc_rqst *rqstp,
- struct ib_sge *sge,
- int sge_count)
+ struct svc_rdma_req_map *vec)
{
u32 xfer_len = rqstp->rq_res.len;
int write_len;
@@ -341,8 +324,7 @@ static int send_reply_chunks(struct svcxprt_rdma *xprt,
rs_offset + chunk_off,
xdr_off,
this_write,
- sge,
- sge_count);
+ vec);
if (ret) {
dprintk("svcrdma: RDMA_WRITE failed, ret=%d\n",
ret);
@@ -380,7 +362,7 @@ static int send_reply(struct svcxprt_rdma *rdma,
struct page *page,
struct rpcrdma_msg *rdma_resp,
struct svc_rdma_op_ctxt *ctxt,
- int sge_count,
+ struct svc_rdma_req_map *vec,
int byte_count)
{
struct ib_send_wr send_wr;
@@ -405,6 +387,7 @@ static int send_reply(struct svcxprt_rdma *rdma,
ctxt->count = 1;
/* Prepare the SGE for the RPCRDMA Header */
+ atomic_inc(&rdma->sc_dma_used);
ctxt->sge[0].addr =
ib_dma_map_page(rdma->sc_cm_id->device,
page, 0, PAGE_SIZE, DMA_TO_DEVICE);
@@ -413,10 +396,16 @@ static int send_reply(struct svcxprt_rdma *rdma,
ctxt->sge[0].lkey = rdma->sc_phys_mr->lkey;
/* Determine how many of our SGE are to be transmitted */
- for (sge_no = 1; byte_count && sge_no < sge_count; sge_no++) {
- sge_bytes = min((size_t)ctxt->sge[sge_no].length,
- (size_t)byte_count);
+ for (sge_no = 1; byte_count && sge_no < vec->count; sge_no++) {
+ sge_bytes = min_t(size_t, vec->sge[sge_no].iov_len, byte_count);
byte_count -= sge_bytes;
+ atomic_inc(&rdma->sc_dma_used);
+ ctxt->sge[sge_no].addr =
+ ib_dma_map_single(rdma->sc_cm_id->device,
+ vec->sge[sge_no].iov_base,
+ sge_bytes, DMA_TO_DEVICE);
+ ctxt->sge[sge_no].length = sge_bytes;
+ ctxt->sge[sge_no].lkey = rdma->sc_phys_mr->lkey;
}
BUG_ON(byte_count != 0);
@@ -428,8 +417,10 @@ static int send_reply(struct svcxprt_rdma *rdma,
ctxt->pages[page_no+1] = rqstp->rq_respages[page_no];
ctxt->count++;
rqstp->rq_respages[page_no] = NULL;
+ /* If there are more pages than SGE, terminate SGE list */
+ if (page_no+1 >= sge_no)
+ ctxt->sge[page_no+1].length = 0;
}
-
BUG_ON(sge_no > rdma->sc_max_sge);
memset(&send_wr, 0, sizeof send_wr);
ctxt->wr_op = IB_WR_SEND;
@@ -473,20 +464,20 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
enum rpcrdma_proc reply_type;
int ret;
int inline_bytes;
- struct ib_sge *sge;
- int sge_count = 0;
struct page *res_page;
struct svc_rdma_op_ctxt *ctxt;
+ struct svc_rdma_req_map *vec;
dprintk("svcrdma: sending response for rqstp=%p\n", rqstp);
/* Get the RDMA request header. */
rdma_argp = xdr_start(&rqstp->rq_arg);
- /* Build an SGE for the XDR */
+ /* Build an req vec for the XDR */
ctxt = svc_rdma_get_context(rdma);
ctxt->direction = DMA_TO_DEVICE;
- sge = xdr_to_sge(rdma, &rqstp->rq_res, ctxt->sge, &sge_count);
+ vec = svc_rdma_get_req_map();
+ xdr_to_sge(rdma, &rqstp->rq_res, vec);
inline_bytes = rqstp->rq_res.len;
@@ -503,7 +494,7 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
/* Send any write-chunk data and build resp write-list */
ret = send_write_chunks(rdma, rdma_argp, rdma_resp,
- rqstp, sge, sge_count);
+ rqstp, vec);
if (ret < 0) {
printk(KERN_ERR "svcrdma: failed to send write chunks, rc=%d\n",
ret);
@@ -513,7 +504,7 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
/* Send any reply-list data and update resp reply-list */
ret = send_reply_chunks(rdma, rdma_argp, rdma_resp,
- rqstp, sge, sge_count);
+ rqstp, vec);
if (ret < 0) {
printk(KERN_ERR "svcrdma: failed to send reply chunks, rc=%d\n",
ret);
@@ -521,11 +512,13 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
}
inline_bytes -= ret;
- ret = send_reply(rdma, rqstp, res_page, rdma_resp, ctxt, sge_count,
+ ret = send_reply(rdma, rqstp, res_page, rdma_resp, ctxt, vec,
inline_bytes);
+ svc_rdma_put_req_map(vec);
dprintk("svcrdma: send_reply returns %d\n", ret);
return ret;
error:
+ svc_rdma_put_req_map(vec);
svc_rdma_put_context(ctxt, 0);
put_page(res_page);
return ret;
diff --git a/net/sunrpc/xprtrdma/svc_rdma_transport.c b/net/sunrpc/xprtrdma/svc_rdma_transport.c
index e132509d1db0..19ddc382b777 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_transport.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_transport.c
@@ -84,70 +84,37 @@ struct svc_xprt_class svc_rdma_class = {
.xcl_max_payload = RPCSVC_MAXPAYLOAD_TCP,
};
-static int rdma_bump_context_cache(struct svcxprt_rdma *xprt)
+/* WR context cache. Created in svc_rdma.c */
+extern struct kmem_cache *svc_rdma_ctxt_cachep;
+
+struct svc_rdma_op_ctxt *svc_rdma_get_context(struct svcxprt_rdma *xprt)
{
- int target;
- int at_least_one = 0;
struct svc_rdma_op_ctxt *ctxt;
- target = min(xprt->sc_ctxt_cnt + xprt->sc_ctxt_bump,
- xprt->sc_ctxt_max);
-
- spin_lock_bh(&xprt->sc_ctxt_lock);
- while (xprt->sc_ctxt_cnt < target) {
- xprt->sc_ctxt_cnt++;
- spin_unlock_bh(&xprt->sc_ctxt_lock);
-
- ctxt = kmalloc(sizeof(*ctxt), GFP_KERNEL);
-
- spin_lock_bh(&xprt->sc_ctxt_lock);
- if (ctxt) {
- at_least_one = 1;
- INIT_LIST_HEAD(&ctxt->free_list);
- list_add(&ctxt->free_list, &xprt->sc_ctxt_free);
- } else {
- /* kmalloc failed...give up for now */
- xprt->sc_ctxt_cnt--;
+ while (1) {
+ ctxt = kmem_cache_alloc(svc_rdma_ctxt_cachep, GFP_KERNEL);
+ if (ctxt)
break;
- }
+ schedule_timeout_uninterruptible(msecs_to_jiffies(500));
}
- spin_unlock_bh(&xprt->sc_ctxt_lock);
- dprintk("svcrdma: sc_ctxt_max=%d, sc_ctxt_cnt=%d\n",
- xprt->sc_ctxt_max, xprt->sc_ctxt_cnt);
- return at_least_one;
+ ctxt->xprt = xprt;
+ INIT_LIST_HEAD(&ctxt->dto_q);
+ ctxt->count = 0;
+ atomic_inc(&xprt->sc_ctxt_used);
+ return ctxt;
}
-struct svc_rdma_op_ctxt *svc_rdma_get_context(struct svcxprt_rdma *xprt)
+static void svc_rdma_unmap_dma(struct svc_rdma_op_ctxt *ctxt)
{
- struct svc_rdma_op_ctxt *ctxt;
-
- while (1) {
- spin_lock_bh(&xprt->sc_ctxt_lock);
- if (unlikely(list_empty(&xprt->sc_ctxt_free))) {
- /* Try to bump my cache. */
- spin_unlock_bh(&xprt->sc_ctxt_lock);
-
- if (rdma_bump_context_cache(xprt))
- continue;
-
- printk(KERN_INFO "svcrdma: sleeping waiting for "
- "context memory on xprt=%p\n",
- xprt);
- schedule_timeout_uninterruptible(msecs_to_jiffies(500));
- continue;
- }
- ctxt = list_entry(xprt->sc_ctxt_free.next,
- struct svc_rdma_op_ctxt,
- free_list);
- list_del_init(&ctxt->free_list);
- spin_unlock_bh(&xprt->sc_ctxt_lock);
- ctxt->xprt = xprt;
- INIT_LIST_HEAD(&ctxt->dto_q);
- ctxt->count = 0;
- atomic_inc(&xprt->sc_ctxt_used);
- break;
+ struct svcxprt_rdma *xprt = ctxt->xprt;
+ int i;
+ for (i = 0; i < ctxt->count && ctxt->sge[i].length; i++) {
+ atomic_dec(&xprt->sc_dma_used);
+ ib_dma_unmap_single(xprt->sc_cm_id->device,
+ ctxt->sge[i].addr,
+ ctxt->sge[i].length,
+ ctxt->direction);
}
- return ctxt;
}
void svc_rdma_put_context(struct svc_rdma_op_ctxt *ctxt, int free_pages)
@@ -161,18 +128,36 @@ void svc_rdma_put_context(struct svc_rdma_op_ctxt *ctxt, int free_pages)
for (i = 0; i < ctxt->count; i++)
put_page(ctxt->pages[i]);
- for (i = 0; i < ctxt->count; i++)
- ib_dma_unmap_single(xprt->sc_cm_id->device,
- ctxt->sge[i].addr,
- ctxt->sge[i].length,
- ctxt->direction);
-
- spin_lock_bh(&xprt->sc_ctxt_lock);
- list_add(&ctxt->free_list, &xprt->sc_ctxt_free);
- spin_unlock_bh(&xprt->sc_ctxt_lock);
+ kmem_cache_free(svc_rdma_ctxt_cachep, ctxt);
atomic_dec(&xprt->sc_ctxt_used);
}
+/* Temporary NFS request map cache. Created in svc_rdma.c */
+extern struct kmem_cache *svc_rdma_map_cachep;
+
+/*
+ * Temporary NFS req mappings are shared across all transport
+ * instances. These are short lived and should be bounded by the number
+ * of concurrent server threads * depth of the SQ.
+ */
+struct svc_rdma_req_map *svc_rdma_get_req_map(void)
+{
+ struct svc_rdma_req_map *map;
+ while (1) {
+ map = kmem_cache_alloc(svc_rdma_map_cachep, GFP_KERNEL);
+ if (map)
+ break;
+ schedule_timeout_uninterruptible(msecs_to_jiffies(500));
+ }
+ map->count = 0;
+ return map;
+}
+
+void svc_rdma_put_req_map(struct svc_rdma_req_map *map)
+{
+ kmem_cache_free(svc_rdma_map_cachep, map);
+}
+
/* ib_cq event handler */
static void cq_event_handler(struct ib_event *event, void *context)
{
@@ -302,6 +287,7 @@ static void rq_cq_reap(struct svcxprt_rdma *xprt)
ctxt = (struct svc_rdma_op_ctxt *)(unsigned long)wc.wr_id;
ctxt->wc_status = wc.status;
ctxt->byte_len = wc.byte_len;
+ svc_rdma_unmap_dma(ctxt);
if (wc.status != IB_WC_SUCCESS) {
/* Close the transport */
dprintk("svcrdma: transport closing putting ctxt %p\n", ctxt);
@@ -351,6 +337,7 @@ static void sq_cq_reap(struct svcxprt_rdma *xprt)
ctxt = (struct svc_rdma_op_ctxt *)(unsigned long)wc.wr_id;
xprt = ctxt->xprt;
+ svc_rdma_unmap_dma(ctxt);
if (wc.status != IB_WC_SUCCESS)
/* Close the transport */
set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags);
@@ -361,10 +348,13 @@ static void sq_cq_reap(struct svcxprt_rdma *xprt)
switch (ctxt->wr_op) {
case IB_WR_SEND:
- case IB_WR_RDMA_WRITE:
svc_rdma_put_context(ctxt, 1);
break;
+ case IB_WR_RDMA_WRITE:
+ svc_rdma_put_context(ctxt, 0);
+ break;
+
case IB_WR_RDMA_READ:
if (test_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags)) {
struct svc_rdma_op_ctxt *read_hdr = ctxt->read_hdr;
@@ -423,40 +413,6 @@ static void sq_comp_handler(struct ib_cq *cq, void *cq_context)
tasklet_schedule(&dto_tasklet);
}
-static void create_context_cache(struct svcxprt_rdma *xprt,
- int ctxt_count, int ctxt_bump, int ctxt_max)
-{
- struct svc_rdma_op_ctxt *ctxt;
- int i;
-
- xprt->sc_ctxt_max = ctxt_max;
- xprt->sc_ctxt_bump = ctxt_bump;
- xprt->sc_ctxt_cnt = 0;
- atomic_set(&xprt->sc_ctxt_used, 0);
-
- INIT_LIST_HEAD(&xprt->sc_ctxt_free);
- for (i = 0; i < ctxt_count; i++) {
- ctxt = kmalloc(sizeof(*ctxt), GFP_KERNEL);
- if (ctxt) {
- INIT_LIST_HEAD(&ctxt->free_list);
- list_add(&ctxt->free_list, &xprt->sc_ctxt_free);
- xprt->sc_ctxt_cnt++;
- }
- }
-}
-
-static void destroy_context_cache(struct svcxprt_rdma *xprt)
-{
- while (!list_empty(&xprt->sc_ctxt_free)) {
- struct svc_rdma_op_ctxt *ctxt;
- ctxt = list_entry(xprt->sc_ctxt_free.next,
- struct svc_rdma_op_ctxt,
- free_list);
- list_del_init(&ctxt->free_list);
- kfree(ctxt);
- }
-}
-
static struct svcxprt_rdma *rdma_create_xprt(struct svc_serv *serv,
int listener)
{
@@ -473,7 +429,6 @@ static struct svcxprt_rdma *rdma_create_xprt(struct svc_serv *serv,
spin_lock_init(&cma_xprt->sc_lock);
spin_lock_init(&cma_xprt->sc_read_complete_lock);
- spin_lock_init(&cma_xprt->sc_ctxt_lock);
spin_lock_init(&cma_xprt->sc_rq_dto_lock);
cma_xprt->sc_ord = svcrdma_ord;
@@ -482,21 +437,9 @@ static struct svcxprt_rdma *rdma_create_xprt(struct svc_serv *serv,
cma_xprt->sc_max_requests = svcrdma_max_requests;
cma_xprt->sc_sq_depth = svcrdma_max_requests * RPCRDMA_SQ_DEPTH_MULT;
atomic_set(&cma_xprt->sc_sq_count, 0);
+ atomic_set(&cma_xprt->sc_ctxt_used, 0);
- if (!listener) {
- int reqs = cma_xprt->sc_max_requests;
- create_context_cache(cma_xprt,
- reqs << 1, /* starting size */
- reqs, /* bump amount */
- reqs +
- cma_xprt->sc_sq_depth +
- RPCRDMA_MAX_THREADS + 1); /* max */
- if (list_empty(&cma_xprt->sc_ctxt_free)) {
- kfree(cma_xprt);
- return NULL;
- }
- clear_bit(XPT_LISTENER, &cma_xprt->sc_xprt.xpt_flags);
- } else
+ if (listener)
set_bit(XPT_LISTENER, &cma_xprt->sc_xprt.xpt_flags);
return cma_xprt;
@@ -532,6 +475,7 @@ int svc_rdma_post_recv(struct svcxprt_rdma *xprt)
BUG_ON(sge_no >= xprt->sc_max_sge);
page = svc_rdma_get_page();
ctxt->pages[sge_no] = page;
+ atomic_inc(&xprt->sc_dma_used);
pa = ib_dma_map_page(xprt->sc_cm_id->device,
page, 0, PAGE_SIZE,
DMA_FROM_DEVICE);
@@ -566,7 +510,7 @@ int svc_rdma_post_recv(struct svcxprt_rdma *xprt)
* will call the recvfrom method on the listen xprt which will accept the new
* connection.
*/
-static void handle_connect_req(struct rdma_cm_id *new_cma_id)
+static void handle_connect_req(struct rdma_cm_id *new_cma_id, size_t client_ird)
{
struct svcxprt_rdma *listen_xprt = new_cma_id->context;
struct svcxprt_rdma *newxprt;
@@ -583,6 +527,9 @@ static void handle_connect_req(struct rdma_cm_id *new_cma_id)
dprintk("svcrdma: Creating newxprt=%p, cm_id=%p, listenxprt=%p\n",
newxprt, newxprt->sc_cm_id, listen_xprt);
+ /* Save client advertised inbound read limit for use later in accept. */
+ newxprt->sc_ord = client_ird;
+
/* Set the local and remote addresses in the transport */
sa = (struct sockaddr *)&newxprt->sc_cm_id->route.addr.dst_addr;
svc_xprt_set_remote(&newxprt->sc_xprt, sa, svc_addr_len(sa));
@@ -619,7 +566,8 @@ static int rdma_listen_handler(struct rdma_cm_id *cma_id,
case RDMA_CM_EVENT_CONNECT_REQUEST:
dprintk("svcrdma: Connect request on cma_id=%p, xprt = %p, "
"event=%d\n", cma_id, cma_id->context, event->event);
- handle_connect_req(cma_id);
+ handle_connect_req(cma_id,
+ event->param.conn.responder_resources);
break;
case RDMA_CM_EVENT_ESTABLISHED:
@@ -793,8 +741,12 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt)
(size_t)svcrdma_max_requests);
newxprt->sc_sq_depth = RPCRDMA_SQ_DEPTH_MULT * newxprt->sc_max_requests;
- newxprt->sc_ord = min((size_t)devattr.max_qp_rd_atom,
- (size_t)svcrdma_ord);
+ /*
+ * Limit ORD based on client limit, local device limit, and
+ * configured svcrdma limit.
+ */
+ newxprt->sc_ord = min_t(size_t, devattr.max_qp_rd_atom, newxprt->sc_ord);
+ newxprt->sc_ord = min_t(size_t, svcrdma_ord, newxprt->sc_ord);
newxprt->sc_pd = ib_alloc_pd(newxprt->sc_cm_id->device);
if (IS_ERR(newxprt->sc_pd)) {
@@ -987,7 +939,6 @@ static void __svc_rdma_free(struct work_struct *work)
* cm_id because the device ptr is needed to unmap the dma in
* svc_rdma_put_context.
*/
- spin_lock_bh(&rdma->sc_read_complete_lock);
while (!list_empty(&rdma->sc_read_complete_q)) {
struct svc_rdma_op_ctxt *ctxt;
ctxt = list_entry(rdma->sc_read_complete_q.next,
@@ -996,10 +947,8 @@ static void __svc_rdma_free(struct work_struct *work)
list_del_init(&ctxt->dto_q);
svc_rdma_put_context(ctxt, 1);
}
- spin_unlock_bh(&rdma->sc_read_complete_lock);
/* Destroy queued, but not processed recv completions */
- spin_lock_bh(&rdma->sc_rq_dto_lock);
while (!list_empty(&rdma->sc_rq_dto_q)) {
struct svc_rdma_op_ctxt *ctxt;
ctxt = list_entry(rdma->sc_rq_dto_q.next,
@@ -1008,10 +957,10 @@ static void __svc_rdma_free(struct work_struct *work)
list_del_init(&ctxt->dto_q);
svc_rdma_put_context(ctxt, 1);
}
- spin_unlock_bh(&rdma->sc_rq_dto_lock);
/* Warn if we leaked a resource or under-referenced */
WARN_ON(atomic_read(&rdma->sc_ctxt_used) != 0);
+ WARN_ON(atomic_read(&rdma->sc_dma_used) != 0);
/* Destroy the QP if present (not a listener) */
if (rdma->sc_qp && !IS_ERR(rdma->sc_qp))
@@ -1032,7 +981,6 @@ static void __svc_rdma_free(struct work_struct *work)
/* Destroy the CM ID */
rdma_destroy_id(rdma->sc_cm_id);
- destroy_context_cache(rdma);
kfree(rdma);
}
@@ -1132,6 +1080,7 @@ void svc_rdma_send_error(struct svcxprt_rdma *xprt, struct rpcrdma_msg *rmsgp,
length = svc_rdma_xdr_encode_error(xprt, rmsgp, err, va);
/* Prepare SGE for local address */
+ atomic_inc(&xprt->sc_dma_used);
sge.addr = ib_dma_map_page(xprt->sc_cm_id->device,
p, 0, PAGE_SIZE, DMA_FROM_DEVICE);
sge.lkey = xprt->sc_phys_mr->lkey;
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index ddbe981ab516..4486c59c3aca 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -579,7 +579,6 @@ static int xs_udp_send_request(struct rpc_task *task)
req->rq_svec->iov_base,
req->rq_svec->iov_len);
- req->rq_xtime = jiffies;
status = xs_sendpages(transport->sock,
xs_addr(xprt),
xprt->addrlen, xdr,
@@ -671,7 +670,6 @@ static int xs_tcp_send_request(struct rpc_task *task)
* to cope with writespace callbacks arriving _after_ we have
* called sendmsg(). */
while (1) {
- req->rq_xtime = jiffies;
status = xs_sendpages(transport->sock,
NULL, 0, xdr, req->rq_bytes_sent);
generated by cgit v1.2.3 (git 2.39.0) at 2024-12-21 03:53:04 +0300