Merge tag 'nfsd-4.8' of git://linux-nfs.org/~bfields/linux

Pull nfsd updates from Bruce Fields:
 "Highlights:

   - Trond made a change to the server's tcp logic that allows a fast
     client to better take advantage of high bandwidth networks, but may
     increase the risk that a single client could starve other clients;
     a new sunrpc.svc_rpc_per_connection_limit parameter should help
     mitigate this in the (hopefully unlikely) event this becomes a
     problem in practice.

   - Tom Haynes added a minimal flex-layout pnfs server, which is of no
     use in production for now--don't build it unless you're doing
     client testing or further server development"

* tag 'nfsd-4.8' of git://linux-nfs.org/~bfields/linux: (32 commits)
  nfsd: remove some dead code in nfsd_create_locked()
  nfsd: drop unnecessary MAY_EXEC check from create
  nfsd: clean up bad-type check in nfsd_create_locked
  nfsd: remove unnecessary positive-dentry check
  nfsd: reorganize nfsd_create
  nfsd: check d_can_lookup in fh_verify of directories
  nfsd: remove redundant zero-length check from create
  nfsd: Make creates return EEXIST instead of EACCES
  SUNRPC: Detect immediate closure of accepted sockets
  SUNRPC: accept() may return sockets that are still in SYN_RECV
  nfsd: allow nfsd to advertise multiple layout types
  nfsd: Close race between nfsd4_release_lockowner and nfsd4_lock
  nfsd/blocklayout: Make sure calculate signature/designator length aligned
  xfs: abstract block export operations from nfsd layouts
  SUNRPC: Remove unused callback xpo_adjust_wspace()
  SUNRPC: Change TCP socket space reservation
  SUNRPC: Add a server side per-connection limit
  SUNRPC: Micro optimisation for svc_data_ready
  SUNRPC: Call the default socket callbacks instead of open coding
  SUNRPC: lock the socket while detaching it
  ...

fs/nfsd/state.h

@@ -345,6 +345,7 @@ struct nfs4_client {
 	u32			cl_exchange_flags;
 	/* number of rpc's in progress over an associated session: */
 	atomic_t		cl_refcount;
+	struct nfs4_op_map      cl_spo_must_allow;
 
 	/* for nfs41 callbacks */
 	/* We currently support a single back channel with a single slot */

fs/nfsd/vfs.c

@@ -1135,96 +1135,37 @@ nfsd_check_ignore_resizing(struct iattr *iap)
 		iap->ia_valid &= ~ATTR_SIZE;
 }
 
-/*
- * Create a file (regular, directory, device, fifo); UNIX sockets 
- * not yet implemented.
- * If the response fh has been verified, the parent directory should
- * already be locked. Note that the parent directory is left locked.
- *
- * N.B. Every call to nfsd_create needs an fh_put for _both_ fhp and resfhp
- */
+/* The parent directory should already be locked: */
 __be32
-nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
+nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp,
 		char *fname, int flen, struct iattr *iap,
 		int type, dev_t rdev, struct svc_fh *resfhp)
 {
-	struct dentry	*dentry, *dchild = NULL;
+	struct dentry	*dentry, *dchild;
 	struct inode	*dirp;
 	__be32		err;
 	__be32		err2;
 	int		host_err;
 
-	err = nfserr_perm;
-	if (!flen)
-		goto out;
-	err = nfserr_exist;
-	if (isdotent(fname, flen))
-		goto out;
-
-	err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_CREATE);
-	if (err)
-		goto out;
-
 	dentry = fhp->fh_dentry;
 	dirp = d_inode(dentry);
 
-	err = nfserr_notdir;
-	if (!dirp->i_op->lookup)
-		goto out;
-	/*
-	 * Check whether the response file handle has been verified yet.
-	 * If it has, the parent directory should already be locked.
-	 */
-	if (!resfhp->fh_dentry) {
-		host_err = fh_want_write(fhp);
-		if (host_err)
-			goto out_nfserr;
-
-		/* called from nfsd_proc_mkdir, or possibly nfsd3_proc_create */
-		fh_lock_nested(fhp, I_MUTEX_PARENT);
-		dchild = lookup_one_len(fname, dentry, flen);
-		host_err = PTR_ERR(dchild);
-		if (IS_ERR(dchild))
-			goto out_nfserr;
-		err = fh_compose(resfhp, fhp->fh_export, dchild, fhp);
-		if (err)
-			goto out;
-	} else {
-		/* called from nfsd_proc_create */
-		dchild = dget(resfhp->fh_dentry);
-		if (!fhp->fh_locked) {
-			/* not actually possible */
-			printk(KERN_ERR
-				"nfsd_create: parent %pd2 not locked!\n",
+	dchild = dget(resfhp->fh_dentry);
+	if (!fhp->fh_locked) {
+		WARN_ONCE(1, "nfsd_create: parent %pd2 not locked!\n",
 				dentry);
-			err = nfserr_io;
-			goto out;
-		}
-	}
-	/*
-	 * Make sure the child dentry is still negative ...
-	 */
-	err = nfserr_exist;
-	if (d_really_is_positive(dchild)) {
-		dprintk("nfsd_create: dentry %pd/%pd not negative!\n",
-			dentry, dchild);
-		goto out; 
+		err = nfserr_io;
+		goto out;
 	}
 
+	err = nfsd_permission(rqstp, fhp->fh_export, dentry, NFSD_MAY_CREATE);
+	if (err)
+		goto out;
+
 	if (!(iap->ia_valid & ATTR_MODE))
 		iap->ia_mode = 0;
 	iap->ia_mode = (iap->ia_mode & S_IALLUGO) | type;
 
-	err = nfserr_inval;
-	if (!S_ISREG(type) && !S_ISDIR(type) && !special_file(type)) {
-		printk(KERN_WARNING "nfsd: bad file type %o in nfsd_create\n",
-		       type);
-		goto out;
-	}
-
-	/*
-	 * Get the dir op function pointer.
-	 */
 	err = 0;
 	host_err = 0;
 	switch (type) {
@@ -1242,6 +1183,10 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	case S_IFSOCK:
 		host_err = vfs_mknod(dirp, dchild, iap->ia_mode, rdev);
 		break;
+	default:
+		printk(KERN_WARNING "nfsd: bad file type %o in nfsd_create\n",
+		       type);
+		host_err = -EINVAL;
 	}
 	if (host_err < 0)
 		goto out_nfserr;
@@ -1251,7 +1196,7 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	/*
 	 * nfsd_create_setattr already committed the child.  Transactional
 	 * filesystems had a chance to commit changes for both parent and
-	 * child * simultaneously making the following commit_metadata a
+	 * child simultaneously making the following commit_metadata a
 	 * noop.
 	 */
 	err2 = nfserrno(commit_metadata(fhp));
@@ -1263,8 +1208,7 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	if (!err)
 		err = fh_update(resfhp);
 out:
-	if (dchild && !IS_ERR(dchild))
-		dput(dchild);
+	dput(dchild);
 	return err;
 
 out_nfserr:
@@ -1272,6 +1216,50 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	goto out;
 }
 
+/*
+ * Create a filesystem object (regular, directory, special).
+ * Note that the parent directory is left locked.
+ *
+ * N.B. Every call to nfsd_create needs an fh_put for _both_ fhp and resfhp
+ */
+__be32
+nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
+		char *fname, int flen, struct iattr *iap,
+		int type, dev_t rdev, struct svc_fh *resfhp)
+{
+	struct dentry	*dentry, *dchild = NULL;
+	struct inode	*dirp;
+	__be32		err;
+	int		host_err;
+
+	if (isdotent(fname, flen))
+		return nfserr_exist;
+
+	err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_NOP);
+	if (err)
+		return err;
+
+	dentry = fhp->fh_dentry;
+	dirp = d_inode(dentry);
+
+	host_err = fh_want_write(fhp);
+	if (host_err)
+		return nfserrno(host_err);
+
+	fh_lock_nested(fhp, I_MUTEX_PARENT);
+	dchild = lookup_one_len(fname, dentry, flen);
+	host_err = PTR_ERR(dchild);
+	if (IS_ERR(dchild))
+		return nfserrno(host_err);
+	err = fh_compose(resfhp, fhp->fh_export, dchild, fhp);
+	if (err) {
+		dput(dchild);
+		return err;
+	}
+	return nfsd_create_locked(rqstp, fhp, fname, flen, iap, type,
+					rdev, resfhp);
+}
+
 #ifdef CONFIG_NFSD_V3
 
 /*
@@ -1304,12 +1292,6 @@ do_nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	dentry = fhp->fh_dentry;
 	dirp = d_inode(dentry);
 
-	/* Get all the sanity checks out of the way before
-	 * we lock the parent. */
-	err = nfserr_notdir;
-	if (!dirp->i_op->lookup)
-		goto out;
-
 	host_err = fh_want_write(fhp);
 	if (host_err)
 		goto out_nfserr;

fs/nfsd/vfs.h

@@ -59,6 +59,9 @@ __be32		nfsd4_vfs_fallocate(struct svc_rqst *, struct svc_fh *,
 __be32		nfsd4_clone_file_range(struct file *, u64, struct file *,
 			u64, u64);
 #endif /* CONFIG_NFSD_V4 */
+__be32		nfsd_create_locked(struct svc_rqst *, struct svc_fh *,
+				char *name, int len, struct iattr *attrs,
+				int type, dev_t rdev, struct svc_fh *res);
 __be32		nfsd_create(struct svc_rqst *, struct svc_fh *,
 				char *name, int len, struct iattr *attrs,
 				int type, dev_t rdev, struct svc_fh *res);

fs/nfsd/xdr4.h

@@ -59,6 +59,7 @@ struct nfsd4_compound_state {
 	struct nfsd4_session	*session;
 	struct nfsd4_slot	*slot;
 	int			data_offset;
+	bool                    spo_must_allowed;
 	size_t			iovlen;
 	u32			minorversion;
 	__be32			status;
@@ -403,6 +404,8 @@ struct nfsd4_exchange_id {
 	clientid_t	clientid;
 	u32		seqid;
 	int		spa_how;
+	u32             spo_must_enforce[3];
+	u32             spo_must_allow[3];
 };
 
 struct nfsd4_sequence {
@@ -654,6 +657,8 @@ set_change_info(struct nfsd4_change_info *cinfo, struct svc_fh *fhp)
 
 }
 
+
+bool nfsd4_mach_creds_match(struct nfs4_client *cl, struct svc_rqst *rqstp);
 int nfs4svc_encode_voidres(struct svc_rqst *, __be32 *, void *);
 int nfs4svc_decode_compoundargs(struct svc_rqst *, __be32 *,
 		struct nfsd4_compoundargs *);

fs/xfs/Makefile

@@ -121,5 +121,4 @@ xfs-$(CONFIG_XFS_RT)		+= xfs_rtalloc.o
 xfs-$(CONFIG_XFS_POSIX_ACL)	+= xfs_acl.o
 xfs-$(CONFIG_SYSCTL)		+= xfs_sysctl.o
 xfs-$(CONFIG_COMPAT)		+= xfs_ioctl32.o
-xfs-$(CONFIG_NFSD_BLOCKLAYOUT)	+= xfs_pnfs.o
-xfs-$(CONFIG_NFSD_SCSILAYOUT)	+= xfs_pnfs.o
+xfs-$(CONFIG_EXPORTFS_BLOCK_OPS)	+= xfs_pnfs.o

fs/xfs/xfs_export.c

@@ -246,7 +246,7 @@ const struct export_operations xfs_export_operations = {
 	.fh_to_parent		= xfs_fs_fh_to_parent,
 	.get_parent		= xfs_fs_get_parent,
 	.commit_metadata	= xfs_fs_nfs_commit_metadata,
-#ifdef CONFIG_NFSD_BLOCKLAYOUT
+#ifdef CONFIG_EXPORTFS_BLOCK_OPS
 	.get_uuid		= xfs_fs_get_uuid,
 	.map_blocks		= xfs_fs_map_blocks,
 	.commit_blocks		= xfs_fs_commit_blocks,

fs/xfs/xfs_pnfs.h

 #ifndef _XFS_PNFS_H
 #define _XFS_PNFS_H 1
 
-#if defined(CONFIG_NFSD_BLOCKLAYOUT) || defined(CONFIG_NFSD_SCSILAYOUT)
+#ifdef CONFIG_EXPORTFS_BLOCK_OPS
 int xfs_fs_get_uuid(struct super_block *sb, u8 *buf, u32 *len, u64 *offset);
 int xfs_fs_map_blocks(struct inode *inode, loff_t offset, u64 length,
 		struct iomap *iomap, bool write, u32 *device_generation);
@@ -15,5 +15,5 @@ xfs_break_layouts(struct inode *inode, uint *iolock, bool with_imutex)
 {
 	return 0;
 }
-#endif /* CONFIG_NFSD_PNFS */
+#endif /* CONFIG_EXPORTFS_BLOCK_OPS */
 #endif /* _XFS_PNFS_H */

include/linux/nfs4.h

@@ -643,4 +643,15 @@ enum pnfs_update_layout_reason {
 	PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET,
 };
 
+#define NFS4_OP_MAP_NUM_LONGS					\
+	DIV_ROUND_UP(LAST_NFS4_OP, 8 * sizeof(unsigned long))
+#define NFS4_OP_MAP_NUM_WORDS \
+	(NFS4_OP_MAP_NUM_LONGS * sizeof(unsigned long) / sizeof(u32))
+struct nfs4_op_map {
+	union {
+		unsigned long longs[NFS4_OP_MAP_NUM_LONGS];
+		u32 words[NFS4_OP_MAP_NUM_WORDS];
+	} u;
+};
+
 #endif

include/linux/nfs_xdr.h

@@ -1185,17 +1185,6 @@ struct pnfs_ds_commit_info {
 	struct pnfs_commit_bucket *buckets;
 };
 
-#define NFS4_OP_MAP_NUM_LONGS \
-	DIV_ROUND_UP(LAST_NFS4_OP, 8 * sizeof(unsigned long))
-#define NFS4_OP_MAP_NUM_WORDS \
-	(NFS4_OP_MAP_NUM_LONGS * sizeof(unsigned long) / sizeof(u32))
-struct nfs4_op_map {
-	union {
-		unsigned long longs[NFS4_OP_MAP_NUM_LONGS];
-		u32 words[NFS4_OP_MAP_NUM_WORDS];
-	} u;
-};
-
 struct nfs41_state_protection {
 	u32 how;
 	struct nfs4_op_map enforce;

include/linux/sunrpc/cache.h

@@ -78,8 +78,6 @@ struct cache_detail {
 	struct hlist_head *	hash_table;
 	rwlock_t		hash_lock;
 
-	atomic_t		inuse; /* active user-space update or lookup */
-
 	char			*name;
 	void			(*cache_put)(struct kref *);
 

include/linux/sunrpc/svc.h

@@ -268,6 +268,7 @@ struct svc_rqst {
 						 * cache pages */
 #define	RQ_VICTIM	(5)			/* about to be shut down */
 #define	RQ_BUSY		(6)			/* request is busy */
+#define	RQ_DATA		(7)			/* request has data */
 	unsigned long		rq_flags;	/* flags field */
 
 	void *			rq_argp;	/* decoded arguments */

include/linux/sunrpc/svc_xprt.h

@@ -25,7 +25,6 @@ struct svc_xprt_ops {
 	void		(*xpo_detach)(struct svc_xprt *);
 	void		(*xpo_free)(struct svc_xprt *);
 	int		(*xpo_secure_port)(struct svc_rqst *);
-	void		(*xpo_adjust_wspace)(struct svc_xprt *);
 };
 
 struct svc_xprt_class {
@@ -69,6 +68,7 @@ struct svc_xprt {
 
 	struct svc_serv		*xpt_server;	/* service for transport */
 	atomic_t    	    	xpt_reserved;	/* space on outq that is rsvd */
+	atomic_t		xpt_nr_rqsts;	/* Number of requests */
 	struct mutex		xpt_mutex;	/* to serialize sending data */
 	spinlock_t		xpt_lock;	/* protects sk_deferred
 						 * and xpt_auth_cache */

include/trace/events/sunrpc.h

@@ -473,6 +473,39 @@ TRACE_EVENT(svc_recv,
 			show_rqstp_flags(__entry->flags))
 );
 
+DECLARE_EVENT_CLASS(svc_rqst_event,
+
+	TP_PROTO(struct svc_rqst *rqst),
+
+	TP_ARGS(rqst),
+
+	TP_STRUCT__entry(
+		__field(__be32, xid)
+		__field(unsigned long, flags)
+		__dynamic_array(unsigned char, addr, rqst->rq_addrlen)
+	),
+
+	TP_fast_assign(
+		__entry->xid = rqst->rq_xid;
+		__entry->flags = rqst->rq_flags;
+		memcpy(__get_dynamic_array(addr),
+			&rqst->rq_addr, rqst->rq_addrlen);
+	),
+
+	TP_printk("addr=%pIScp rq_xid=0x%x flags=%s",
+		(struct sockaddr *)__get_dynamic_array(addr),
+		be32_to_cpu(__entry->xid),
+		show_rqstp_flags(__entry->flags))
+);
+
+DEFINE_EVENT(svc_rqst_event, svc_defer,
+	TP_PROTO(struct svc_rqst *rqst),
+	TP_ARGS(rqst));
+
+DEFINE_EVENT(svc_rqst_event, svc_drop,
+	TP_PROTO(struct svc_rqst *rqst),
+	TP_ARGS(rqst));
+
 DECLARE_EVENT_CLASS(svc_rqst_status,
 
 	TP_PROTO(struct svc_rqst *rqst, int status),
@@ -529,45 +562,67 @@ TRACE_EVENT(svc_xprt_do_enqueue,
 
 	TP_STRUCT__entry(
 		__field(struct svc_xprt *, xprt)
-		__field_struct(struct sockaddr_storage, ss)
 		__field(int, pid)
 		__field(unsigned long, flags)
+		__dynamic_array(unsigned char, addr, xprt != NULL ?
+			xprt->xpt_remotelen : 0)
 	),
 
 	TP_fast_assign(
 		__entry->xprt = xprt;
-		xprt ? memcpy(&__entry->ss, &xprt->xpt_remote, sizeof(__entry->ss)) : memset(&__entry->ss, 0, sizeof(__entry->ss));
 		__entry->pid = rqst? rqst->rq_task->pid : 0;
-		__entry->flags = xprt ? xprt->xpt_flags : 0;
+		if (xprt) {
+			memcpy(__get_dynamic_array(addr),
+				&xprt->xpt_remote,
+				xprt->xpt_remotelen);
+			__entry->flags = xprt->xpt_flags;
+		} else
+			__entry->flags = 0;
 	),
 
 	TP_printk("xprt=0x%p addr=%pIScp pid=%d flags=%s", __entry->xprt,
-		(struct sockaddr *)&__entry->ss,
+		__get_dynamic_array_len(addr) != 0 ?
+			(struct sockaddr *)__get_dynamic_array(addr) : NULL,
 		__entry->pid, show_svc_xprt_flags(__entry->flags))
 );
 
-TRACE_EVENT(svc_xprt_dequeue,
+DECLARE_EVENT_CLASS(svc_xprt_event,
 	TP_PROTO(struct svc_xprt *xprt),
 
 	TP_ARGS(xprt),
 
 	TP_STRUCT__entry(
 		__field(struct svc_xprt *, xprt)
-		__field_struct(struct sockaddr_storage, ss)
 		__field(unsigned long, flags)
+		__dynamic_array(unsigned char, addr, xprt != NULL ?
+			xprt->xpt_remotelen : 0)
 	),
 
 	TP_fast_assign(
-		__entry->xprt = xprt,
-		xprt ? memcpy(&__entry->ss, &xprt->xpt_remote, sizeof(__entry->ss)) : memset(&__entry->ss, 0, sizeof(__entry->ss));
-		__entry->flags = xprt ? xprt->xpt_flags : 0;
+		__entry->xprt = xprt;
+		if (xprt) {
+			memcpy(__get_dynamic_array(addr),
+					&xprt->xpt_remote,
+					xprt->xpt_remotelen);
+			__entry->flags = xprt->xpt_flags;
+		} else
+			__entry->flags = 0;
 	),
 
 	TP_printk("xprt=0x%p addr=%pIScp flags=%s", __entry->xprt,
-		(struct sockaddr *)&__entry->ss,
+		__get_dynamic_array_len(addr) != 0 ?
+			(struct sockaddr *)__get_dynamic_array(addr) : NULL,
 		show_svc_xprt_flags(__entry->flags))
 );
 
+DEFINE_EVENT(svc_xprt_event, svc_xprt_dequeue,
+	TP_PROTO(struct svc_xprt *xprt),
+	TP_ARGS(xprt));
+
+DEFINE_EVENT(svc_xprt_event, svc_xprt_no_write_space,
+	TP_PROTO(struct svc_xprt *xprt),
+	TP_ARGS(xprt));
+
 TRACE_EVENT(svc_wake_up,
 	TP_PROTO(int pid),
 
@@ -592,21 +647,56 @@ TRACE_EVENT(svc_handle_xprt,
 	TP_STRUCT__entry(
 		__field(struct svc_xprt *, xprt)
 		__field(int, len)
-		__field_struct(struct sockaddr_storage, ss)
 		__field(unsigned long, flags)
+		__dynamic_array(unsigned char, addr, xprt != NULL ?
+			xprt->xpt_remotelen : 0)
 	),
 
 	TP_fast_assign(
 		__entry->xprt = xprt;
-		xprt ? memcpy(&__entry->ss, &xprt->xpt_remote, sizeof(__entry->ss)) : memset(&__entry->ss, 0, sizeof(__entry->ss));
 		__entry->len = len;
-		__entry->flags = xprt ? xprt->xpt_flags : 0;
+		if (xprt) {
+			memcpy(__get_dynamic_array(addr),
+					&xprt->xpt_remote,
+					xprt->xpt_remotelen);
+			__entry->flags = xprt->xpt_flags;
+		} else
+			__entry->flags = 0;
 	),
 
 	TP_printk("xprt=0x%p addr=%pIScp len=%d flags=%s", __entry->xprt,
-		(struct sockaddr *)&__entry->ss,
+		__get_dynamic_array_len(addr) != 0 ?
+			(struct sockaddr *)__get_dynamic_array(addr) : NULL,
 		__entry->len, show_svc_xprt_flags(__entry->flags))
 );
+
+
+DECLARE_EVENT_CLASS(svc_deferred_event,
+	TP_PROTO(struct svc_deferred_req *dr),
+
+	TP_ARGS(dr),
+
+	TP_STRUCT__entry(
+		__field(__be32, xid)
+		__dynamic_array(unsigned char, addr, dr->addrlen)
+	),
+
+	TP_fast_assign(
+		__entry->xid = *(__be32 *)(dr->args + (dr->xprt_hlen>>2));
+		memcpy(__get_dynamic_array(addr), &dr->addr, dr->addrlen);
+	),
+
+	TP_printk("addr=%pIScp xid=0x%x",
+		(struct sockaddr *)__get_dynamic_array(addr),
+		be32_to_cpu(__entry->xid))
+);
+
+DEFINE_EVENT(svc_deferred_event, svc_drop_deferred,
+	TP_PROTO(struct svc_deferred_req *dr),
+	TP_ARGS(dr));
+DEFINE_EVENT(svc_deferred_event, svc_revisit_deferred,
+	TP_PROTO(struct svc_deferred_req *dr),
+	TP_ARGS(dr));
 #endif /* _TRACE_SUNRPC_H */
 
 #include <trace/define_trace.h>

net/sunrpc/auth_gss/svcauth_gss.c

@@ -1230,8 +1230,9 @@ static int svcauth_gss_proxy_init(struct svc_rqst *rqstp,
 	if (status)
 		goto out;
 
-	dprintk("RPC:       svcauth_gss: gss major status = %d\n",
-			ud.major_status);
+	dprintk("RPC:       svcauth_gss: gss major status = %d "
+			"minor status = %d\n",
+			ud.major_status, ud.minor_status);
 
 	switch (ud.major_status) {
 	case GSS_S_CONTINUE_NEEDED:

net/sunrpc/cache.c

@@ -362,7 +362,7 @@ void sunrpc_destroy_cache_detail(struct cache_detail *cd)
 	cache_purge(cd);
 	spin_lock(&cache_list_lock);
 	write_lock(&cd->hash_lock);
-	if (cd->entries || atomic_read(&cd->inuse)) {
+	if (cd->entries) {
 		write_unlock(&cd->hash_lock);
 		spin_unlock(&cache_list_lock);
 		goto out;

net/sunrpc/svc_xprt.c

@@ -21,6 +21,10 @@
 
 #define RPCDBG_FACILITY	RPCDBG_SVCXPRT
 
+static unsigned int svc_rpc_per_connection_limit __read_mostly;
+module_param(svc_rpc_per_connection_limit, uint, 0644);
+
+
 static struct svc_deferred_req *svc_deferred_dequeue(struct svc_xprt *xprt);
 static int svc_deferred_recv(struct svc_rqst *rqstp);
 static struct cache_deferred_req *svc_defer(struct cache_req *req);
@@ -329,12 +333,45 @@ char *svc_print_addr(struct svc_rqst *rqstp, char *buf, size_t len)
 }
 EXPORT_SYMBOL_GPL(svc_print_addr);
 
+static bool svc_xprt_slots_in_range(struct svc_xprt *xprt)
+{
+	unsigned int limit = svc_rpc_per_connection_limit;
+	int nrqsts = atomic_read(&xprt->xpt_nr_rqsts);
+
+	return limit == 0 || (nrqsts >= 0 && nrqsts < limit);
+}
+
+static bool svc_xprt_reserve_slot(struct svc_rqst *rqstp, struct svc_xprt *xprt)
+{
+	if (!test_bit(RQ_DATA, &rqstp->rq_flags)) {
+		if (!svc_xprt_slots_in_range(xprt))
+			return false;
+		atomic_inc(&xprt->xpt_nr_rqsts);
+		set_bit(RQ_DATA, &rqstp->rq_flags);
+	}
+	return true;
+}
+
+static void svc_xprt_release_slot(struct svc_rqst *rqstp)
+{
+	struct svc_xprt	*xprt = rqstp->rq_xprt;
+	if (test_and_clear_bit(RQ_DATA, &rqstp->rq_flags)) {
+		atomic_dec(&xprt->xpt_nr_rqsts);
+		svc_xprt_enqueue(xprt);
+	}
+}
+
 static bool svc_xprt_has_something_to_do(struct svc_xprt *xprt)
 {
 	if (xprt->xpt_flags & ((1<<XPT_CONN)|(1<<XPT_CLOSE)))
 		return true;
-	if (xprt->xpt_flags & ((1<<XPT_DATA)|(1<<XPT_DEFERRED)))
-		return xprt->xpt_ops->xpo_has_wspace(xprt);
+	if (xprt->xpt_flags & ((1<<XPT_DATA)|(1<<XPT_DEFERRED))) {
+		if (xprt->xpt_ops->xpo_has_wspace(xprt) &&
+		    svc_xprt_slots_in_range(xprt))
+			return true;
+		trace_svc_xprt_no_write_space(xprt);
+		return false;
+	}
 	return false;
 }
 
@@ -480,8 +517,6 @@ void svc_reserve(struct svc_rqst *rqstp, int space)
 		atomic_sub((rqstp->rq_reserved - space), &xprt->xpt_reserved);
 		rqstp->rq_reserved = space;
 
-		if (xprt->xpt_ops->xpo_adjust_wspace)
-			xprt->xpt_ops->xpo_adjust_wspace(xprt);
 		svc_xprt_enqueue(xprt);
 	}
 }
@@ -512,8 +547,8 @@ static void svc_xprt_release(struct svc_rqst *rqstp)
 
 	rqstp->rq_res.head[0].iov_len = 0;
 	svc_reserve(rqstp, 0);
+	svc_xprt_release_slot(rqstp);
 	rqstp->rq_xprt = NULL;
-
 	svc_xprt_put(xprt);
 }
 
@@ -781,7 +816,7 @@ static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt)
 			svc_add_new_temp_xprt(serv, newxpt);
 		else
 			module_put(xprt->xpt_class->xcl_owner);
-	} else {
+	} else if (svc_xprt_reserve_slot(rqstp, xprt)) {
 		/* XPT_DATA|XPT_DEFERRED case: */
 		dprintk("svc: server %p, pool %u, transport %p, inuse=%d\n",
 			rqstp, rqstp->rq_pool->sp_id, xprt,
@@ -871,6 +906,7 @@ EXPORT_SYMBOL_GPL(svc_recv);
  */
 void svc_drop(struct svc_rqst *rqstp)
 {
+	trace_svc_drop(rqstp);
 	dprintk("svc: xprt %p dropped request\n", rqstp->rq_xprt);
 	svc_xprt_release(rqstp);
 }
@@ -1148,6 +1184,7 @@ static void svc_revisit(struct cache_deferred_req *dreq, int too_many)
 		spin_unlock(&xprt->xpt_lock);
 		dprintk("revisit canceled\n");
 		svc_xprt_put(xprt);
+		trace_svc_drop_deferred(dr);
 		kfree(dr);
 		return;
 	}
@@ -1205,6 +1242,7 @@ static struct cache_deferred_req *svc_defer(struct cache_req *req)
 	set_bit(RQ_DROPME, &rqstp->rq_flags);
 
 	dr->handle.revisit = svc_revisit;
+	trace_svc_defer(rqstp);
 	return &dr->handle;
 }
 
@@ -1245,6 +1283,7 @@ static struct svc_deferred_req *svc_deferred_dequeue(struct svc_xprt *xprt)
 				struct svc_deferred_req,
 				handle.recent);
 		list_del_init(&dr->handle.recent);
+		trace_svc_revisit_deferred(dr);
 	} else
 		clear_bit(XPT_DEFERRED, &xprt->xpt_flags);
 	spin_unlock(&xprt->xpt_lock);

net/sunrpc/svcsock.c

@@ -60,7 +60,6 @@
 
 static struct svc_sock *svc_setup_socket(struct svc_serv *, struct socket *,
 					 int flags);
-static void		svc_udp_data_ready(struct sock *);
 static int		svc_udp_recvfrom(struct svc_rqst *);
 static int		svc_udp_sendto(struct svc_rqst *);
 static void		svc_sock_detach(struct svc_xprt *);
@@ -398,48 +397,21 @@ static int svc_sock_secure_port(struct svc_rqst *rqstp)
 	return svc_port_is_privileged(svc_addr(rqstp));
 }
 
-static bool sunrpc_waitqueue_active(wait_queue_head_t *wq)
-{
-	if (!wq)
-		return false;
-	/*
-	 * There should normally be a memory * barrier here--see
-	 * wq_has_sleeper().
-	 *
-	 * It appears that isn't currently necessary, though, basically
-	 * because callers all appear to have sufficient memory barriers
-	 * between the time the relevant change is made and the
-	 * time they call these callbacks.
-	 *
-	 * The nfsd code itself doesn't actually explicitly wait on
-	 * these waitqueues, but it may wait on them for example in
-	 * sendpage() or sendmsg() calls.  (And those may be the only
-	 * places, since it it uses nonblocking reads.)
-	 *
-	 * Maybe we should add the memory barriers anyway, but these are
-	 * hot paths so we'd need to be convinced there's no sigificant
-	 * penalty.
-	 */
-	return waitqueue_active(wq);
-}
-
 /*
  * INET callback when data has been received on the socket.
  */
-static void svc_udp_data_ready(struct sock *sk)
+static void svc_data_ready(struct sock *sk)
 {
 	struct svc_sock	*svsk = (struct svc_sock *)sk->sk_user_data;
-	wait_queue_head_t *wq = sk_sleep(sk);
 
 	if (svsk) {
 		dprintk("svc: socket %p(inet %p), busy=%d\n",
 			svsk, sk,
 			test_bit(XPT_BUSY, &svsk->sk_xprt.xpt_flags));
-		set_bit(XPT_DATA, &svsk->sk_xprt.xpt_flags);
-		svc_xprt_enqueue(&svsk->sk_xprt);
+		svsk->sk_odata(sk);
+		if (!test_and_set_bit(XPT_DATA, &svsk->sk_xprt.xpt_flags))
+			svc_xprt_enqueue(&svsk->sk_xprt);
 	}
-	if (sunrpc_waitqueue_active(wq))
-		wake_up_interruptible(wq);
 }
 
 /*
@@ -448,56 +420,22 @@ static void svc_udp_data_ready(struct sock *sk)
 static void svc_write_space(struct sock *sk)
 {
 	struct svc_sock	*svsk = (struct svc_sock *)(sk->sk_user_data);
-	wait_queue_head_t *wq = sk_sleep(sk);
 
 	if (svsk) {
 		dprintk("svc: socket %p(inet %p), write_space busy=%d\n",
 			svsk, sk, test_bit(XPT_BUSY, &svsk->sk_xprt.xpt_flags));
+		svsk->sk_owspace(sk);
 		svc_xprt_enqueue(&svsk->sk_xprt);
 	}
-
-	if (sunrpc_waitqueue_active(wq)) {
-		dprintk("RPC svc_write_space: someone sleeping on %p\n",
-		       svsk);
-		wake_up_interruptible(wq);
-	}
 }
 
 static int svc_tcp_has_wspace(struct svc_xprt *xprt)
 {
-	struct svc_sock *svsk =	container_of(xprt, struct svc_sock, sk_xprt);
-	struct svc_serv *serv = svsk->sk_xprt.xpt_server;
-	int required;
+	struct svc_sock *svsk = container_of(xprt, struct svc_sock, sk_xprt);
 
 	if (test_bit(XPT_LISTENER, &xprt->xpt_flags))
 		return 1;
-	required = atomic_read(&xprt->xpt_reserved) + serv->sv_max_mesg;
-	if (sk_stream_wspace(svsk->sk_sk) >= required ||
-	    (sk_stream_min_wspace(svsk->sk_sk) == 0 &&
-	     atomic_read(&xprt->xpt_reserved) == 0))
-		return 1;
-	set_bit(SOCK_NOSPACE, &svsk->sk_sock->flags);
-	return 0;
-}
-
-static void svc_tcp_write_space(struct sock *sk)
-{
-	struct svc_sock *svsk = (struct svc_sock *)(sk->sk_user_data);
-	struct socket *sock = sk->sk_socket;
-
-	if (!sk_stream_is_writeable(sk) || !sock)
-		return;
-	if (!svsk || svc_tcp_has_wspace(&svsk->sk_xprt))
-		clear_bit(SOCK_NOSPACE, &sock->flags);
-	svc_write_space(sk);
-}
-
-static void svc_tcp_adjust_wspace(struct svc_xprt *xprt)
-{
-	struct svc_sock *svsk = container_of(xprt, struct svc_sock, sk_xprt);
-
-	if (svc_tcp_has_wspace(xprt))
-		clear_bit(SOCK_NOSPACE, &svsk->sk_sock->flags);
+	return !test_bit(SOCK_NOSPACE, &svsk->sk_sock->flags);
 }
 
 /*
@@ -746,7 +684,7 @@ static void svc_udp_init(struct svc_sock *svsk, struct svc_serv *serv)
 	svc_xprt_init(sock_net(svsk->sk_sock->sk), &svc_udp_class,
 		      &svsk->sk_xprt, serv);
 	clear_bit(XPT_CACHE_AUTH, &svsk->sk_xprt.xpt_flags);
-	svsk->sk_sk->sk_data_ready = svc_udp_data_ready;
+	svsk->sk_sk->sk_data_ready = svc_data_ready;
 	svsk->sk_sk->sk_write_space = svc_write_space;
 
 	/* initialise setting must have enough space to
@@ -786,11 +724,12 @@ static void svc_udp_init(struct svc_sock *svsk, struct svc_serv *serv)
 static void svc_tcp_listen_data_ready(struct sock *sk)
 {
 	struct svc_sock	*svsk = (struct svc_sock *)sk->sk_user_data;
-	wait_queue_head_t *wq;
 
 	dprintk("svc: socket %p TCP (listen) state change %d\n",
 		sk, sk->sk_state);
 
+	if (svsk)
+		svsk->sk_odata(sk);
 	/*
 	 * This callback may called twice when a new connection
 	 * is established as a child socket inherits everything
@@ -808,10 +747,6 @@ static void svc_tcp_listen_data_ready(struct sock *sk)
 		} else
 			printk("svc: socket %p: no user data\n", sk);
 	}
-
-	wq = sk_sleep(sk);
-	if (sunrpc_waitqueue_active(wq))
-		wake_up_interruptible_all(wq);
 }
 
 /*
@@ -820,7 +755,6 @@ static void svc_tcp_listen_data_ready(struct sock *sk)
 static void svc_tcp_state_change(struct sock *sk)
 {
 	struct svc_sock	*svsk = (struct svc_sock *)sk->sk_user_data;
-	wait_queue_head_t *wq = sk_sleep(sk);
 
 	dprintk("svc: socket %p TCP (connected) state change %d (svsk %p)\n",
 		sk, sk->sk_state, sk->sk_user_data);
@@ -828,26 +762,12 @@ static void svc_tcp_state_change(struct sock *sk)
 	if (!svsk)
 		printk("svc: socket %p: no user data\n", sk);
 	else {
-		set_bit(XPT_CLOSE, &svsk->sk_xprt.xpt_flags);
-		svc_xprt_enqueue(&svsk->sk_xprt);
-	}
-	if (sunrpc_waitqueue_active(wq))
-		wake_up_interruptible_all(wq);
-}
-
-static void svc_tcp_data_ready(struct sock *sk)
-{
-	struct svc_sock *svsk = (struct svc_sock *)sk->sk_user_data;
-	wait_queue_head_t *wq = sk_sleep(sk);
-
-	dprintk("svc: socket %p TCP data ready (svsk %p)\n",
-		sk, sk->sk_user_data);
-	if (svsk) {
-		set_bit(XPT_DATA, &svsk->sk_xprt.xpt_flags);
-		svc_xprt_enqueue(&svsk->sk_xprt);
+		svsk->sk_ostate(sk);
+		if (sk->sk_state != TCP_ESTABLISHED) {
+			set_bit(XPT_CLOSE, &svsk->sk_xprt.xpt_flags);
+			svc_xprt_enqueue(&svsk->sk_xprt);
+		}
 	}
-	if (sunrpc_waitqueue_active(wq))
-		wake_up_interruptible(wq);
 }
 
 /*
@@ -901,6 +821,11 @@ static struct svc_xprt *svc_tcp_accept(struct svc_xprt *xprt)
 	dprintk("%s: connect from %s\n", serv->sv_name,
 		__svc_print_addr(sin, buf, sizeof(buf)));
 
+	/* Reset the inherited callbacks before calling svc_setup_socket */
+	newsock->sk->sk_state_change = svsk->sk_ostate;
+	newsock->sk->sk_data_ready = svsk->sk_odata;
+	newsock->sk->sk_write_space = svsk->sk_owspace;
+
 	/* make sure that a write doesn't block forever when
 	 * low on memory
 	 */
@@ -1317,7 +1242,6 @@ static struct svc_xprt_ops svc_tcp_ops = {
 	.xpo_has_wspace = svc_tcp_has_wspace,
 	.xpo_accept = svc_tcp_accept,
 	.xpo_secure_port = svc_sock_secure_port,
-	.xpo_adjust_wspace = svc_tcp_adjust_wspace,
 };
 
 static struct svc_xprt_class svc_tcp_class = {
@@ -1357,8 +1281,8 @@ static void svc_tcp_init(struct svc_sock *svsk, struct svc_serv *serv)
 	} else {
 		dprintk("setting up TCP socket for reading\n");
 		sk->sk_state_change = svc_tcp_state_change;
-		sk->sk_data_ready = svc_tcp_data_ready;
-		sk->sk_write_space = svc_tcp_write_space;
+		sk->sk_data_ready = svc_data_ready;
+		sk->sk_write_space = svc_write_space;
 
 		svsk->sk_reclen = 0;
 		svsk->sk_tcplen = 0;
@@ -1368,8 +1292,13 @@ static void svc_tcp_init(struct svc_sock *svsk, struct svc_serv *serv)
 		tcp_sk(sk)->nonagle |= TCP_NAGLE_OFF;
 
 		set_bit(XPT_DATA, &svsk->sk_xprt.xpt_flags);
-		if (sk->sk_state != TCP_ESTABLISHED)
+		switch (sk->sk_state) {
+		case TCP_SYN_RECV:
+		case TCP_ESTABLISHED:
+			break;
+		default:
 			set_bit(XPT_CLOSE, &svsk->sk_xprt.xpt_flags);
+		}
 	}
 }
 
@@ -1428,17 +1357,14 @@ static struct svc_sock *svc_setup_socket(struct svc_serv *serv,
 	/* Initialize the socket */
 	if (sock->type == SOCK_DGRAM)
 		svc_udp_init(svsk, serv);
-	else {
-		/* initialise setting must have enough space to
-		 * receive and respond to one request.
-		 */
-		svc_sock_setbufsize(svsk->sk_sock, 4 * serv->sv_max_mesg,
-					4 * serv->sv_max_mesg);
+	else
 		svc_tcp_init(svsk, serv);
-	}
 
-	dprintk("svc: svc_setup_socket created %p (inet %p)\n",
-				svsk, svsk->sk_sk);
+	dprintk("svc: svc_setup_socket created %p (inet %p), "
+			"listen %d close %d\n",
+			svsk, svsk->sk_sk,
+			test_bit(XPT_LISTENER, &svsk->sk_xprt.xpt_flags),
+			test_bit(XPT_CLOSE, &svsk->sk_xprt.xpt_flags));
 
 	return svsk;
 }
@@ -1606,18 +1532,16 @@ static void svc_sock_detach(struct svc_xprt *xprt)
 {
 	struct svc_sock *svsk = container_of(xprt, struct svc_sock, sk_xprt);
 	struct sock *sk = svsk->sk_sk;
-	wait_queue_head_t *wq;
 
 	dprintk("svc: svc_sock_detach(%p)\n", svsk);
 
 	/* put back the old socket callbacks */
+	lock_sock(sk);
 	sk->sk_state_change = svsk->sk_ostate;
 	sk->sk_data_ready = svsk->sk_odata;
 	sk->sk_write_space = svsk->sk_owspace;
-
-	wq = sk_sleep(sk);
-	if (sunrpc_waitqueue_active(wq))
-		wake_up_interruptible(wq);
+	sk->sk_user_data = NULL;
+	release_sock(sk);
 }
 
 /*