Merge http://nfsclient.bkbits.net/linux-2.5

into ppc970.osdl.org:/home/torvalds/v2.6/linux

fs/nfs/direct.c

@@ -32,6 +32,7 @@
  * 18 Dec 2001	Initial implementation for 2.4  --cel
  * 08 Jul 2002	Version for 2.4.19, with bug fixes --trondmy
  * 08 Jun 2003	Port to 2.5 APIs  --cel
+ * 31 Mar 2004	Handle direct I/O without VFS support  --cel
  *
  */
 
@@ -252,9 +253,7 @@ nfs_direct_write_seg(struct inode *inode, struct file *file,
 {
 	const unsigned int wsize = NFS_SERVER(inode)->wsize;
 	size_t request;
-	int need_commit;
-	int tot_bytes;
-	int curpage;
+	int curpage, need_commit, result, tot_bytes;
 	struct nfs_writeverf first_verf;
 	struct nfs_write_data	wdata = {
 		.inode		= inode,
@@ -281,8 +280,6 @@ nfs_direct_write_seg(struct inode *inode, struct file *file,
 	wdata.args.pgbase = user_addr & ~PAGE_MASK;
 	wdata.args.offset = file_offset;
         do {
-		int result;
-
 		wdata.args.count = request;
                 if (wdata.args.count > wsize)
                         wdata.args.count = wsize;
@@ -299,7 +296,7 @@ nfs_direct_write_seg(struct inode *inode, struct file *file,
 		if (result <= 0) {
 			if (tot_bytes > 0)
 				break;
-			return result;
+			goto out;
 		}
 
 		if (tot_bytes == 0)
@@ -324,8 +321,6 @@ nfs_direct_write_seg(struct inode *inode, struct file *file,
 	 * Commit data written so far, even in the event of an error
 	 */
 	if (need_commit) {
-		int result;
-
 		wdata.args.count = tot_bytes;
 		wdata.args.offset = file_offset;
 
@@ -338,9 +333,12 @@ nfs_direct_write_seg(struct inode *inode, struct file *file,
 						VERF_SIZE) != 0)
 			goto sync_retry;
 	}
+	result = tot_bytes;
+
+out:
 	nfs_end_data_update_defer(inode);
 
-	return tot_bytes;
+	return result;
 
 sync_retry:
 	wdata.args.stable = NFS_FILE_SYNC;
@@ -409,12 +407,6 @@ nfs_direct_write(struct inode *inode, struct file *file,
  * file_offset: offset in file to begin the operation
  * nr_segs: size of iovec array
  *
- * Usually a file system implements direct I/O by calling out to
- * blockdev_direct_IO.  The NFS client doesn't have a backing block
- * device, so we do everything by hand instead.
- *
- * The inode's i_sem is no longer held by the VFS layer before it calls
- * this function to do a write.
  */
 ssize_t
 nfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
@@ -429,11 +421,7 @@ nfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	 * No support for async yet
 	 */
 	if (!is_sync_kiocb(iocb))
-		goto out;
-
-	result = nfs_revalidate_inode(NFS_SERVER(inode), inode);
-	if (result < 0)
-		goto out;
+		return result;
 
 	switch (rw) {
 	case READ:
@@ -453,8 +441,160 @@ nfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	default:
 		break;
 	}
+	return result;
+}
+
+/**
+ * nfs_file_direct_read - file direct read operation for NFS files
+ * @iocb: target I/O control block
+ * @buf: user's buffer into which to read data
+ * count: number of bytes to read
+ * pos: byte offset in file where reading starts
+ *
+ * We use this function for direct reads instead of calling
+ * generic_file_aio_read() in order to avoid gfar's check to see if
+ * the request starts before the end of the file.  For that check
+ * to work, we must generate a GETATTR before each direct read, and
+ * even then there is a window between the GETATTR and the subsequent
+ * READ where the file size could change.  So our preference is simply
+ * to do all reads the application wants, and the server will take
+ * care of managing the end of file boundary.
+ * 
+ * This function also eliminates unnecessarily updating the file's
+ * atime locally, as the NFS server sets the file's atime, and this
+ * client must read the updated atime from the server back into its
+ * cache.
+ */
+ssize_t
+nfs_file_direct_read(struct kiocb *iocb, char *buf, size_t count, loff_t pos)
+{
+	ssize_t retval = -EINVAL;
+	loff_t *ppos = &iocb->ki_pos;
+	struct file *file = iocb->ki_filp;
+	struct dentry *dentry = file->f_dentry;
+	struct address_space *mapping = file->f_mapping;
+	struct inode *inode = mapping->host;
+	struct iovec iov = {
+		.iov_base = buf,
+		.iov_len = count,
+	};
+
+	dprintk("nfs: direct read(%s/%s, %lu@%lu)\n",
+		dentry->d_parent->d_name.name, dentry->d_name.name,
+		(unsigned long) count, (unsigned long) pos);
+
+	if (!is_sync_kiocb(iocb))
+		goto out;
+	if (count < 0)
+		goto out;
+	retval = -EFAULT;
+	if (!access_ok(VERIFY_WRITE, iov.iov_base, iov.iov_len))
+		goto out;
+	retval = 0;
+	if (!count)
+		goto out;
+
+	if (mapping->nrpages) {
+		retval = filemap_fdatawrite(mapping);
+		if (retval == 0)
+			retval = filemap_fdatawait(mapping);
+		if (retval)
+			goto out;
+	}
+
+	retval = nfs_direct_read(inode, file, &iov, pos, 1);
+	if (retval > 0)
+		*ppos = pos + retval;
 
 out:
-	dprintk("NFS: direct_IO result=%zd\n", result);
-	return result;
+	return retval;
+}
+
+/**
+ * nfs_file_direct_write - file direct write operation for NFS files
+ * @iocb: target I/O control block
+ * @buf: user's buffer from which to write data
+ * count: number of bytes to write
+ * pos: byte offset in file where writing starts
+ *
+ * We use this function for direct writes instead of calling
+ * generic_file_aio_write() in order to avoid taking the inode
+ * semaphore and updating the i_size.  The NFS server will set
+ * the new i_size and this client must read the updated size
+ * back into its cache.  We let the server do generic write
+ * parameter checking and report problems.
+ *
+ * We also avoid an unnecessary invocation of generic_osync_inode(),
+ * as it is fairly meaningless to sync the metadata of an NFS file.
+ *
+ * We eliminate local atime updates, see direct read above.
+ *
+ * We avoid unnecessary page cache invalidations for normal cached
+ * readers of this file.
+ *
+ * Note that O_APPEND is not supported for NFS direct writes, as there
+ * is no atomic O_APPEND write facility in the NFS protocol.
+ */
+ssize_t
+nfs_file_direct_write(struct kiocb *iocb, const char *buf, size_t count, loff_t pos)
+{
+	ssize_t retval = -EINVAL;
+	loff_t *ppos = &iocb->ki_pos;
+	unsigned long limit = current->rlim[RLIMIT_FSIZE].rlim_cur;
+	struct file *file = iocb->ki_filp;
+	struct dentry *dentry = file->f_dentry;
+	struct address_space *mapping = file->f_mapping;
+	struct inode *inode = mapping->host;
+	struct iovec iov = {
+		.iov_base = (void __user *)buf,
+		.iov_len = count,
+	};
+
+	dfprintk(VFS, "nfs: direct write(%s/%s(%ld), %lu@%lu)\n",
+		dentry->d_parent->d_name.name, dentry->d_name.name,
+		inode->i_ino, (unsigned long) count, (unsigned long) pos);
+
+	if (!is_sync_kiocb(iocb))
+		goto out;
+	if (count < 0)
+		goto out;
+        if (pos < 0)
+		goto out;
+	retval = -EFAULT;
+	if (!access_ok(VERIFY_READ, iov.iov_base, iov.iov_len))
+		goto out;
+        if (file->f_error) {
+                retval = file->f_error;
+                file->f_error = 0;
+                goto out;
+        }
+	retval = -EFBIG;
+	if (limit != RLIM_INFINITY) {
+		if (pos >= limit) {
+			send_sig(SIGXFSZ, current, 0);
+			goto out;
+		}
+		if (count > limit - (unsigned long) pos)
+			count = limit - (unsigned long) pos;
+	}
+	retval = 0;
+	if (!count)
+		goto out;
+
+	if (mapping->nrpages) {
+		retval = filemap_fdatawrite(mapping);
+		if (retval == 0)
+			retval = filemap_fdatawait(mapping);
+		if (retval)
+			goto out;
+	}
+
+	retval = nfs_direct_write(inode, file, &iov, pos, 1);
+	if (mapping->nrpages)
+		invalidate_inode_pages2(mapping);
+	if (retval > 0)
+		*ppos = pos + retval;
+
+out:
+	return retval;
 }

fs/nfs/file.c

@@ -154,6 +154,11 @@ nfs_file_read(struct kiocb *iocb, char * buf, size_t count, loff_t pos)
 	struct inode * inode = dentry->d_inode;
 	ssize_t result;
 
+#ifdef CONFIG_NFS_DIRECTIO
+	if (iocb->ki_filp->f_flags & O_DIRECT)
+		return nfs_file_direct_read(iocb, buf, count, pos);
+#endif
+
 	dfprintk(VFS, "nfs: read(%s/%s, %lu@%lu)\n",
 		dentry->d_parent->d_name.name, dentry->d_name.name,
 		(unsigned long) count, (unsigned long) pos);
@@ -268,6 +273,11 @@ nfs_file_write(struct kiocb *iocb, const char *buf, size_t count, loff_t pos)
 	struct inode * inode = dentry->d_inode;
 	ssize_t result;
 
+#ifdef CONFIG_NFS_DIRECTIO
+	if (iocb->ki_filp->f_flags & O_DIRECT)
+		return nfs_file_direct_write(iocb, buf, count, pos);
+#endif
+
 	dfprintk(VFS, "nfs: write(%s/%s(%ld), %lu@%lu)\n",
 		dentry->d_parent->d_name.name, dentry->d_name.name,
 		inode->i_ino, (unsigned long) count, (unsigned long) pos);

fs/nfs/inode.c

@@ -237,7 +237,7 @@ nfs_get_root(struct super_block *sb, struct nfs_fh *rootfh, struct nfs_fsinfo *f
 
 	error = server->rpc_ops->getroot(server, rootfh, fsinfo);
 	if (error < 0) {
-		printk(KERN_NOTICE "nfs_get_root: getattr error = %d\n", -error);
+		dprintk("nfs_get_root: getattr error = %d\n", -error);
 		return ERR_PTR(error);
 	}
 
@@ -262,6 +262,7 @@ nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor)
 	struct nfs_pathconf pathinfo = {
 			.fattr = &fattr,
 	};
+	int no_root_error = 0;
 
 	/* We probably want something more informative here */
 	snprintf(sb->s_id, sizeof(sb->s_id), "%x:%x", MAJOR(sb->s_dev), MINOR(sb->s_dev));
@@ -272,12 +273,15 @@ nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor)
 
 	root_inode = nfs_get_root(sb, &server->fh, &fsinfo);
 	/* Did getting the root inode fail? */
-	if (IS_ERR(root_inode))
+	if (IS_ERR(root_inode)) {
+		no_root_error = PTR_ERR(root_inode);
 		goto out_no_root;
+	}
 	sb->s_root = d_alloc_root(root_inode);
-	if (!sb->s_root)
+	if (!sb->s_root) {
+		no_root_error = -ENOMEM;
 		goto out_no_root;
-
+	}
 	sb->s_root->d_op = server->rpc_ops->dentry_ops;
 
 	/* Get some general file system info */
@@ -337,10 +341,10 @@ nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor)
 	return 0;
 	/* Yargs. It didn't work out. */
 out_no_root:
-	printk("nfs_read_super: get root inode failed\n");
+	dprintk("nfs_sb_init: get root inode failed: errno %d\n", -no_root_error);
 	if (!IS_ERR(root_inode))
 		iput(root_inode);
-	return -EINVAL;
+	return no_root_error;
 }
 
 /*

fs/nfs/nfs4state.c

@@ -731,6 +731,8 @@ nfs4_reclaim_open_state(struct nfs4_state_owner *sp)
 	int status = 0;
 
 	list_for_each_entry(state, &sp->so_states, open_states) {
+		if (state->state == 0)
+			continue;
 		status = nfs4_open_reclaim(sp, state);
 		if (status >= 0)
 			continue;

fs/nfs/read.c

@@ -43,14 +43,12 @@ static mempool_t *nfs_rdata_mempool;
 
 #define MIN_POOL_READ	(32)
 
-static __inline__ struct nfs_read_data *nfs_readdata_alloc(void)
+static struct nfs_read_data *nfs_readdata_alloc(void)
 {
 	struct nfs_read_data   *p;
 	p = (struct nfs_read_data *)mempool_alloc(nfs_rdata_mempool, SLAB_NOFS);
-	if (p) {
+	if (p)
 		memset(p, 0, sizeof(*p));
-		INIT_LIST_HEAD(&p->pages);
-	}
 	return p;
 }
 
@@ -99,10 +97,17 @@ nfs_readpage_sync(struct file *file, struct inode *inode, struct page *page)
 	unsigned int	rsize = NFS_SERVER(inode)->rsize;
 	unsigned int	count = PAGE_CACHE_SIZE;
 	int		result;
-	struct nfs_read_data	rdata = {
+	struct nfs_read_data *rdata;
+
+	rdata = nfs_readdata_alloc();
+	if (!rdata)
+		return -ENOMEM;
+
+	*rdata = (struct nfs_read_data) {
 		.flags		= (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0),
 		.cred		= NULL,
 		.inode		= inode,
+		.pages		= LIST_HEAD_INIT(rdata->pages),
 		.args		= {
 			.fh		= NFS_FH(inode),
 			.lockowner	= current->files,
@@ -111,7 +116,7 @@ nfs_readpage_sync(struct file *file, struct inode *inode, struct page *page)
 			.count		= rsize,
 		},
 		.res		= {
-			.fattr		= &rdata.fattr,
+			.fattr		= &rdata->fattr,
 		}
 	};
 
@@ -123,19 +128,19 @@ nfs_readpage_sync(struct file *file, struct inode *inode, struct page *page)
 	 */
 	do {
 		if (count < rsize)
-			rdata.args.count = count;
-		rdata.res.count = rdata.args.count;
-		rdata.args.offset = page_offset(page) + rdata.args.pgbase;
+			rdata->args.count = count;
+		rdata->res.count = rdata->args.count;
+		rdata->args.offset = page_offset(page) + rdata->args.pgbase;
 
 		dprintk("NFS: nfs_proc_read(%s, (%s/%Ld), %Lu, %u)\n",
 			NFS_SERVER(inode)->hostname,
 			inode->i_sb->s_id,
 			(long long)NFS_FILEID(inode),
-			(unsigned long long)rdata.args.pgbase,
-			rdata.args.count);
+			(unsigned long long)rdata->args.pgbase,
+			rdata->args.count);
 
 		lock_kernel();
-		result = NFS_PROTO(inode)->read(&rdata, file);
+		result = NFS_PROTO(inode)->read(rdata, file);
 		unlock_kernel();
 
 		/*
@@ -148,17 +153,17 @@ nfs_readpage_sync(struct file *file, struct inode *inode, struct page *page)
 			goto io_error;
 		}
 		count -= result;
-		rdata.args.pgbase += result;
+		rdata->args.pgbase += result;
 		/* Note: result == 0 should only happen if we're caching
 		 * a write that extends the file and punches a hole.
 		 */
-		if (rdata.res.eof != 0 || result == 0)
+		if (rdata->res.eof != 0 || result == 0)
 			break;
 	} while (count);
 	NFS_FLAGS(inode) |= NFS_INO_INVALID_ATIME;
 
 	if (count)
-		memclear_highpage_flush(page, rdata.args.pgbase, count);
+		memclear_highpage_flush(page, rdata->args.pgbase, count);
 	SetPageUptodate(page);
 	if (PageError(page))
 		ClearPageError(page);
@@ -166,6 +171,7 @@ nfs_readpage_sync(struct file *file, struct inode *inode, struct page *page)
 
 io_error:
 	unlock_page(page);
+	nfs_readdata_free(rdata);
 	return result;
 }
 
@@ -305,6 +311,7 @@ static int nfs_pagein_multi(struct list_head *head, struct inode *inode)
 		data = nfs_readdata_alloc();
 		if (!data)
 			goto out_bad;
+		INIT_LIST_HEAD(&data->pages);
 		list_add(&data->pages, &list);
 		requests++;
 		if (nbytes <= rsize)
@@ -361,6 +368,7 @@ static int nfs_pagein_one(struct list_head *head, struct inode *inode)
 	if (!data)
 		goto out_bad;
 
+	INIT_LIST_HEAD(&data->pages);
 	pages = data->pagevec;
 	count = 0;
 	while (!list_empty(head)) {

fs/nfs/write.c

@@ -179,7 +179,13 @@ static int nfs_writepage_sync(struct file *file, struct inode *inode,
 {
 	unsigned int	wsize = NFS_SERVER(inode)->wsize;
 	int		result, written = 0;
-	struct nfs_write_data	wdata = {
+	struct nfs_write_data *wdata;
+
+	wdata = kmalloc(sizeof(*wdata), GFP_NOFS);
+	if (!wdata)
+		return -ENOMEM;
+
+	*wdata = (struct nfs_write_data) {
 		.flags		= how,
 		.cred		= NULL,
 		.inode		= inode,
@@ -192,8 +198,8 @@ static int nfs_writepage_sync(struct file *file, struct inode *inode,
 			.count		= wsize,
 		},
 		.res		= {
-			.fattr		= &wdata.fattr,
-			.verf		= &wdata.verf,
+			.fattr		= &wdata->fattr,
+			.verf		= &wdata->verf,
 		},
 	};
 
@@ -205,22 +211,22 @@ static int nfs_writepage_sync(struct file *file, struct inode *inode,
 	nfs_begin_data_update(inode);
 	do {
 		if (count < wsize)
-			wdata.args.count = count;
-		wdata.args.offset = page_offset(page) + wdata.args.pgbase;
+			wdata->args.count = count;
+		wdata->args.offset = page_offset(page) + wdata->args.pgbase;
 
-		result = NFS_PROTO(inode)->write(&wdata, file);
+		result = NFS_PROTO(inode)->write(wdata, file);
 
 		if (result < 0) {
 			/* Must mark the page invalid after I/O error */
 			ClearPageUptodate(page);
 			goto io_error;
 		}
-		if (result < wdata.args.count)
+		if (result < wdata->args.count)
 			printk(KERN_WARNING "NFS: short write, count=%u, result=%d\n",
-					wdata.args.count, result);
+					wdata->args.count, result);
 
-		wdata.args.offset += result;
-	        wdata.args.pgbase += result;
+		wdata->args.offset += result;
+	        wdata->args.pgbase += result;
 		written += result;
 		count -= result;
 	} while (count);
@@ -234,9 +240,10 @@ static int nfs_writepage_sync(struct file *file, struct inode *inode,
 
 io_error:
 	nfs_end_data_update_defer(inode);
-	if (wdata.cred)
-		put_rpccred(wdata.cred);
+	if (wdata->cred)
+		put_rpccred(wdata->cred);
 
+	kfree(wdata);
 	return written ? written : result;
 }
 

include/linux/nfs_fs.h

@@ -306,6 +306,10 @@ nfs_file_cred(struct file *file)
  */
 extern ssize_t nfs_direct_IO(int, struct kiocb *, const struct iovec *, loff_t,
 			unsigned long);
+extern ssize_t nfs_file_direct_read(struct kiocb *iocb, char *buf,
+			size_t count, loff_t pos);
+extern ssize_t nfs_file_direct_write(struct kiocb *iocb, const char *buf,
+			size_t count, loff_t pos);
 
 /*
  * linux/fs/nfs/dir.c

include/linux/sunrpc/gss_api.h

@@ -50,46 +50,36 @@ u32 gss_verify_mic(
 u32 gss_delete_sec_context(
 		struct gss_ctx		**ctx_id);
 
-/* We maintain a list of the pseudoflavors (equivalently, mechanism-qop-service
- * triples) that we currently support: */
-
-struct sup_sec_triple {
-	struct list_head	triples;
-	u32			pseudoflavor;
-	struct gss_api_mech	*mech;
-	u32			qop;
-	u32			service;
+struct gss_api_mech * gss_mech_get_by_name(char *name);
+struct gss_api_mech * gss_mech_get_by_pseudoflavor(u32 pseudoflavor);
+u32 gss_pseudoflavor_to_service(struct gss_api_mech *, u32 pseudoflavor);
+char *gss_service_to_auth_domain_name(struct gss_api_mech *, u32 service);
+
+struct pf_desc {
+	u32	pseudoflavor;
+	u32	qop;
+	u32	service;
+	char	*name;
+	char	*auth_domain_name;
 };
 
-int gss_register_triple(u32 pseudoflavor, struct gss_api_mech *mech, u32 qop,
-			u32 service);
-int gss_unregister_triple(u32 pseudoflavor);
-int gss_pseudoflavor_supported(u32 pseudoflavor);
-u32 gss_cmp_triples(u32 oid_len, char *oid_data, u32 qop, u32 service);
-u32 gss_get_pseudoflavor(struct gss_ctx *ctx_id, u32 qop, u32 service);
-u32 gss_pseudoflavor_to_service(u32 pseudoflavor);
-/* Both return NULL on failure: */
-struct gss_api_mech * gss_pseudoflavor_to_mech(u32 pseudoflavor);
-int gss_pseudoflavor_to_mechOID(u32 pseudoflavor, struct xdr_netobj *mech);
-
 /* Different mechanisms (e.g., krb5 or spkm3) may implement gss-api, and
- * mechanisms may be dynamically registered or unregistered by modules.
- * Our only built-in mechanism is a trivial debugging mechanism that provides
- * no actual security; the following function registers that mechanism: */
-
-void gss_mech_register_debug(void);
+ * mechanisms may be dynamically registered or unregistered by modules. */
 
 /* Each mechanism is described by the following struct: */
 struct gss_api_mech {
-	struct xdr_netobj	gm_oid;
 	struct list_head	gm_list;
-	atomic_t		gm_count;
+	struct module		*gm_owner;
+	struct xdr_netobj	gm_oid;
+	char			*gm_name;
 	struct gss_api_ops	*gm_ops;
+	/* pseudoflavors supported by this mechanism: */
+	int			gm_pf_num;
+	struct pf_desc		gm_pfs[];
 };
 
 /* and must provide the following operations: */
 struct gss_api_ops {
-	char *name;
 	u32 (*gss_import_sec_context)(
 			struct xdr_netobj	*input_token,
 			struct gss_ctx		*ctx_id);
@@ -107,29 +97,25 @@ struct gss_api_ops {
 			void			*internal_ctx_id);
 };
 
-/* Returns nonzero on failure. */
-int gss_mech_register(struct xdr_netobj *, struct gss_api_ops *);
+int gss_mech_register(struct gss_api_mech *);
+void gss_mech_unregister(struct gss_api_mech *);
 
-/* Returns nonzero iff someone still has a reference to this mech. */
-int gss_mech_unregister(struct gss_api_mech *);
-
-/* Returns nonzer iff someone still has a reference to some mech. */
-int gss_mech_unregister_all(void);
-
-/* returns a mechanism descriptor given an OID, an increments the mechanism's
+/* returns a mechanism descriptor given an OID, and increments the mechanism's
  * reference count. */
 struct gss_api_mech * gss_mech_get_by_OID(struct xdr_netobj *);
 
-/* Similar, but get by name like "krb5", "spkm", etc., instead of OID. */
+/* Returns a reference to a mechanism, given a name like "krb5" etc. */
 struct gss_api_mech *gss_mech_get_by_name(char *);
 
+/* Similar, but get by pseudoflavor. */
+struct gss_api_mech *gss_mech_get_by_pseudoflavor(u32);
+
 /* Just increments the mechanism's reference count and returns its input: */
 struct gss_api_mech * gss_mech_get(struct gss_api_mech *);
 
-/* Returns nonzero iff you've released the last reference to this mech.
- * Note that for every succesful gss_get_mech call there must be exactly
- * one corresponding call to gss_mech_put.*/
-int gss_mech_put(struct gss_api_mech *);
+/* For every succesful gss_mech_get or gss_mech_get_by_* call there must be a
+ * corresponding call to gss_mech_put. */
+void gss_mech_put(struct gss_api_mech *);
 
 #endif /* __KERNEL__ */
 #endif /* _LINUX_SUNRPC_GSS_API_H */

include/linux/sunrpc/gss_asn1.h

@@ -81,5 +81,4 @@ int g_token_size(
 void g_make_token_header(
      struct xdr_netobj *mech,
      int body_size,
-     unsigned char **buf,
-     int tok_type);
+     unsigned char **buf);

include/linux/sunrpc/gss_krb5.h

@@ -115,7 +115,7 @@ enum seal_alg {
 #define ENCTYPE_UNKNOWN         0x01ff
 
 s32
-krb5_make_checksum(s32 cksumtype, char *header, struct xdr_buf *body,
+make_checksum(s32 cksumtype, char *header, int hdrlen, struct xdr_buf *body,
 		   struct xdr_netobj *cksum);
 
 u32

include/linux/sunrpc/xprt.h

@@ -69,8 +69,7 @@ extern unsigned int xprt_tcp_slot_table_entries;
  * This describes a timeout strategy
  */
 struct rpc_timeout {
-	unsigned long		to_current,		/* current timeout */
-				to_initval,		/* initial timeout */
+	unsigned long		to_initval,		/* initial timeout */
 				to_maxval,		/* max timeout */
 				to_increment;		/* if !exponential */
 	unsigned int		to_retries;		/* max # of retries */
@@ -85,7 +84,6 @@ struct rpc_rqst {
 	 * This is the user-visible part
 	 */
 	struct rpc_xprt *	rq_xprt;		/* RPC client */
-	struct rpc_timeout	rq_timeout;		/* timeout parms */
 	struct xdr_buf		rq_snd_buf;		/* send buffer */
 	struct xdr_buf		rq_rcv_buf;		/* recv buffer */
 
@@ -103,6 +101,9 @@ struct rpc_rqst {
 	struct xdr_buf		rq_private_buf;		/* The receive buffer
 							 * used in the softirq.
 							 */
+	unsigned long		rq_majortimeo;	/* major timeout alarm */
+	unsigned long		rq_timeout;	/* Current timeout value */
+	unsigned int		rq_retries;	/* # of retries */
 	/*
 	 * For authentication (e.g. auth_des)
 	 */
@@ -115,7 +116,6 @@ struct rpc_rqst {
 	u32			rq_bytes_sent;	/* Bytes we have sent */
 
 	unsigned long		rq_xtime;	/* when transmitted */
-	int			rq_ntimeo;
 	int			rq_ntrans;
 };
 #define rq_svec			rq_snd_buf.head
@@ -210,7 +210,7 @@ void			xprt_reserve(struct rpc_task *);
 int			xprt_prepare_transmit(struct rpc_task *);
 void			xprt_transmit(struct rpc_task *);
 void			xprt_receive(struct rpc_task *);
-int			xprt_adjust_timeout(struct rpc_timeout *);
+int			xprt_adjust_timeout(struct rpc_rqst *req);
 void			xprt_release(struct rpc_task *);
 void			xprt_connect(struct rpc_task *);
 int			xprt_clear_backlog(struct rpc_xprt *);

net/sunrpc/auth_gss/Makefile

@@ -4,11 +4,11 @@
 
 obj-$(CONFIG_SUNRPC_GSS) += auth_rpcgss.o
 
-auth_rpcgss-objs := auth_gss.o gss_pseudoflavors.o gss_generic_token.o \
-	sunrpcgss_syms.o gss_mech_switch.o svcauth_gss.o
+auth_rpcgss-objs := auth_gss.o gss_generic_token.o \
+	gss_mech_switch.o svcauth_gss.o gss_krb5_crypto.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_crypto.o gss_krb5_seqnum.o
+	gss_krb5_seqnum.o
 

net/sunrpc/auth_gss/gss_generic_token.c

@@ -32,6 +32,7 @@
  */
 
 #include <linux/types.h>
+#include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/string.h>
 #include <linux/sunrpc/sched.h>
@@ -151,22 +152,23 @@ g_token_size(struct xdr_netobj *mech, unsigned int body_size)
 	return(1 + der_length_size(body_size) + body_size);
 }
 
+EXPORT_SYMBOL(g_token_size);
+
 /* fills in a buffer with the token header.  The buffer is assumed to
    be the right size.  buf is advanced past the token header */
 
 void
-g_make_token_header(struct xdr_netobj *mech, int body_size, unsigned char **buf,
-		int tok_type)
+g_make_token_header(struct xdr_netobj *mech, int body_size, unsigned char **buf)
 {
 	*(*buf)++ = 0x60;
 	der_write_length(buf, 4 + mech->len + body_size);
 	*(*buf)++ = 0x06;
 	*(*buf)++ = (unsigned char) mech->len;
 	TWRITE_STR(*buf, mech->data, ((int) mech->len));
-	*(*buf)++ = (unsigned char) ((tok_type>>8)&0xff);
-	*(*buf)++ = (unsigned char) (tok_type&0xff);
 }
 
+EXPORT_SYMBOL(g_make_token_header);
+
 /*
  * Given a buffer containing a token, reads and verifies the token,
  * leaving buf advanced past the token header, and setting body_size
@@ -221,9 +223,6 @@ g_verify_token_header(struct xdr_netobj *mech, int *body_size,
 	if (ret)
 		return(ret);
 
-	if ((*buf++ != ((tok_type>>8)&0xff)) || (*buf++ != (tok_type&0xff))) 
-		return(G_WRONG_TOKID);
-
 	if (!ret) {
 		*buf_in = buf;
 		*body_size = toksize;
@@ -232,6 +231,8 @@ g_verify_token_header(struct xdr_netobj *mech, int *body_size,
 	return(ret);
 }
 
+EXPORT_SYMBOL(g_verify_token_header);
+
 /* Given a buffer containing a token, returns a copy of the mech oid in
  * the parameter mech. */
 u32

net/sunrpc/auth_gss/gss_krb5_crypto.c

@@ -59,14 +59,14 @@ krb5_encrypt(
         struct scatterlist sg[1];
 	u8 local_iv[16] = {0};
 
-	dprintk("RPC: krb5_encrypt: input data:\n");
+	dprintk("RPC:      krb5_encrypt: input data:\n");
 	print_hexl((u32 *)in, length, 0);
 
 	if (length % crypto_tfm_alg_blocksize(tfm) != 0)
 		goto out;
 
 	if (crypto_tfm_alg_ivsize(tfm) > 16) {
-		dprintk("RPC: gss_k5encrypt: tfm iv size to large %d\n",
+		dprintk("RPC:      gss_k5encrypt: tfm iv size to large %d\n",
 		         crypto_tfm_alg_ivsize(tfm));
 		goto out;
 	}
@@ -81,13 +81,15 @@ krb5_encrypt(
 
 	ret = crypto_cipher_encrypt_iv(tfm, sg, sg, length, local_iv);
 
-	dprintk("RPC: krb5_encrypt: output data:\n");
+	dprintk("RPC:      krb5_encrypt: output data:\n");
 	print_hexl((u32 *)out, length, 0);
 out:
-	dprintk("krb5_encrypt returns %d\n",ret);
+	dprintk("RPC:      krb5_encrypt returns %d\n",ret);
 	return(ret);
 }
 
+EXPORT_SYMBOL(krb5_encrypt);
+
 u32
 krb5_decrypt(
      struct crypto_tfm *tfm,
@@ -100,14 +102,14 @@ krb5_decrypt(
 	struct scatterlist sg[1];
 	u8 local_iv[16] = {0};
 
-	dprintk("RPC: krb5_decrypt: input data:\n");
+	dprintk("RPC:      krb5_decrypt: input data:\n");
 	print_hexl((u32 *)in, length, 0);
 
 	if (length % crypto_tfm_alg_blocksize(tfm) != 0)
 		goto out;
 
 	if (crypto_tfm_alg_ivsize(tfm) > 16) {
-		dprintk("RPC: gss_k5decrypt: tfm iv size to large %d\n",
+		dprintk("RPC:      gss_k5decrypt: tfm iv size to large %d\n",
 			crypto_tfm_alg_ivsize(tfm));
 		goto out;
 	}
@@ -121,13 +123,15 @@ krb5_decrypt(
 
 	ret = crypto_cipher_decrypt_iv(tfm, sg, sg, length, local_iv);
 
-	dprintk("RPC: krb5_decrypt: output_data:\n");
+	dprintk("RPC:      krb5_decrypt: output_data:\n");
 	print_hexl((u32 *)out, length, 0);
 out:
-	dprintk("gss_k5decrypt returns %d\n",ret);
+	dprintk("RPC:      gss_k5decrypt returns %d\n",ret);
 	return(ret);
 }
 
+EXPORT_SYMBOL(krb5_decrypt);
+
 void
 buf_to_sg(struct scatterlist *sg, char *ptr, int len) {
 	sg->page = virt_to_page(ptr);
@@ -135,10 +139,9 @@ buf_to_sg(struct scatterlist *sg, char *ptr, int len) {
 	sg->length = len;
 }
 
-/* checksum the plaintext data and the first 8 bytes of the krb5 token header,
- * as specified by the rfc: */
+/* checksum the plaintext data and hdrlen bytes of the token header */
 s32
-krb5_make_checksum(s32 cksumtype, char *header, struct xdr_buf *body,
+make_checksum(s32 cksumtype, char *header, int hdrlen, struct xdr_buf *body,
 		   struct xdr_netobj *cksum)
 {
 	char                            *cksumname;
@@ -153,7 +156,7 @@ krb5_make_checksum(s32 cksumtype, char *header, struct xdr_buf *body,
 			cksumname = "md5";
 			break;
 		default:
-			dprintk("RPC: krb5_make_checksum:"
+			dprintk("RPC:      krb5_make_checksum:"
 				" unsupported checksum %d", cksumtype);
 			goto out;
 	}
@@ -164,7 +167,7 @@ krb5_make_checksum(s32 cksumtype, char *header, struct xdr_buf *body,
 		goto out;
 
 	crypto_digest_init(tfm);
-	buf_to_sg(sg, header, 8);
+	buf_to_sg(sg, header, hdrlen);
 	crypto_digest_update(tfm, sg, 1);
 	if (body->head[0].iov_len) {
 		buf_to_sg(sg, body->head[0].iov_base, body->head[0].iov_len);
@@ -202,3 +205,5 @@ krb5_make_checksum(s32 cksumtype, char *header, struct xdr_buf *body,
 		crypto_free_tfm(tfm);
 	return code;
 }
+
+EXPORT_SYMBOL(make_checksum);

net/sunrpc/auth_gss/gss_krb5_mech.c

@@ -40,7 +40,6 @@
 #include <linux/slab.h>
 #include <linux/sunrpc/auth.h>
 #include <linux/in.h>
-#include <linux/sunrpc/svcauth_gss.h>
 #include <linux/sunrpc/gss_krb5.h>
 #include <linux/sunrpc/xdr.h>
 #include <linux/crypto.h>
@@ -100,7 +99,7 @@ get_key(char **p, char *end, struct crypto_tfm **res)
 			alg_mode = CRYPTO_TFM_MODE_CBC;
 			break;
 		default:
-			dprintk("RPC: get_key: unsupported algorithm %d\n", alg);
+			dprintk("RPC:      get_key: unsupported algorithm %d\n", alg);
 			goto out_err_free_key;
 	}
 	if (!(*res = crypto_alloc_tfm(alg_name, alg_mode)))
@@ -155,7 +154,7 @@ gss_import_sec_context_kerberos(struct xdr_netobj *inbuf,
 		goto out_err_free_key2;
 
 	ctx_id->internal_ctx_id = ctx;
-	dprintk("Succesfully imported new context.\n");
+	dprintk("RPC:      Succesfully imported new context.\n");
 	return 0;
 
 out_err_free_key2:
@@ -197,7 +196,7 @@ gss_verify_mic_kerberos(struct gss_ctx		*ctx,
 	if (!maj_stat && qop_state)
 	    *qstate = qop_state;
 
-	dprintk("RPC: gss_verify_mic_kerberos returning %d\n", maj_stat);
+	dprintk("RPC:      gss_verify_mic_kerberos returning %d\n", maj_stat);
 	return maj_stat;
 }
 
@@ -211,41 +210,42 @@ gss_get_mic_kerberos(struct gss_ctx	*ctx,
 
 	err = krb5_make_token(kctx, qop, message, mic_token, KG_TOK_MIC_MSG);
 
-	dprintk("RPC: gss_get_mic_kerberos returning %d\n",err);
+	dprintk("RPC:      gss_get_mic_kerberos returning %d\n",err);
 
 	return err;
 }
 
 static struct gss_api_ops gss_kerberos_ops = {
-	.name			= "krb5",
 	.gss_import_sec_context	= gss_import_sec_context_kerberos,
 	.gss_get_mic		= gss_get_mic_kerberos,
 	.gss_verify_mic		= gss_verify_mic_kerberos,
 	.gss_delete_sec_context	= gss_delete_sec_context_kerberos,
 };
 
-/* XXX error checking? reference counting? */
+static struct gss_api_mech gss_kerberos_mech = {
+	.gm_name	= "krb5",
+	.gm_owner	= THIS_MODULE,
+	.gm_ops		= &gss_kerberos_ops,
+	.gm_pf_num	= 2,
+	.gm_pfs		= {
+		{RPC_AUTH_GSS_KRB5, 0, RPC_GSS_SVC_NONE, "krb5"},
+		{RPC_AUTH_GSS_KRB5I, 0, RPC_GSS_SVC_INTEGRITY, "krb5i"},
+	},
+};
+
 static int __init init_kerberos_module(void)
 {
-	struct gss_api_mech *gm;
+	int status;
 
-	if (gss_mech_register(&gss_mech_krb5_oid, &gss_kerberos_ops))
+	status = gss_mech_register(&gss_kerberos_mech);
+	if (status)
 		printk("Failed to register kerberos gss mechanism!\n");
-	gm = gss_mech_get_by_OID(&gss_mech_krb5_oid);
-	gss_register_triple(RPC_AUTH_GSS_KRB5 , gm, 0, RPC_GSS_SVC_NONE);
-	gss_register_triple(RPC_AUTH_GSS_KRB5I, gm, 0, RPC_GSS_SVC_INTEGRITY);
-	if (svcauth_gss_register_pseudoflavor(RPC_AUTH_GSS_KRB5, "krb5"))
-		printk("Failed to register %s with server!\n", "krb5");
-	if (svcauth_gss_register_pseudoflavor(RPC_AUTH_GSS_KRB5I, "krb5i"))
-		printk("Failed to register %s with server!\n", "krb5i");
-	gss_mech_put(gm);
-	return 0;
+	return status;
 }
 
 static void __exit cleanup_kerberos_module(void)
 {
-	gss_unregister_triple(RPC_AUTH_GSS_KRB5I);
-	gss_unregister_triple(RPC_AUTH_GSS_KRB5);
+	gss_mech_unregister(&gss_kerberos_mech);
 }
 
 MODULE_LICENSE("GPL");

net/sunrpc/auth_gss/gss_krb5_seal.c

@@ -91,7 +91,7 @@ krb5_make_token(struct krb5_ctx *ctx, int qop_req,
 
 	dprintk("RPC:     gss_krb5_seal\n");
 
-	now = jiffies;
+	now = get_seconds();
 
 	if (qop_req != 0)
 		goto out_err;
@@ -101,12 +101,12 @@ krb5_make_token(struct krb5_ctx *ctx, int qop_req,
 			checksum_type = CKSUMTYPE_RSA_MD5;
 			break;
 		default:
-			dprintk("RPC: gss_krb5_seal: ctx->signalg %d not"
+			dprintk("RPC:      gss_krb5_seal: ctx->signalg %d not"
 				" supported\n", ctx->signalg);
 			goto out_err;
 	}
 	if (ctx->sealalg != SEAL_ALG_NONE && ctx->sealalg != SEAL_ALG_DES) {
-		dprintk("RPC: gss_krb5_seal: ctx->sealalg %d not supported\n",
+		dprintk("RPC:      gss_krb5_seal: ctx->sealalg %d not supported\n",
 			ctx->sealalg);
 		goto out_err;
 	}
@@ -122,7 +122,10 @@ krb5_make_token(struct krb5_ctx *ctx, int qop_req,
 	token->len = g_token_size(&ctx->mech_used, 22 + tmsglen);
 
 	ptr = token->data;
-	g_make_token_header(&ctx->mech_used, 22 + tmsglen, &ptr, toktype);
+	g_make_token_header(&ctx->mech_used, 22 + tmsglen, &ptr);
+
+	*ptr++ = (unsigned char) ((toktype>>8)&0xff);
+	*ptr++ = (unsigned char) (toktype&0xff);
 
 	/* ptr now at byte 2 of header described in rfc 1964, section 1.2.1: */
 	krb5_hdr = ptr - 2;
@@ -137,7 +140,7 @@ krb5_make_token(struct krb5_ctx *ctx, int qop_req,
 		/* XXX removing support for now */
 		goto out_err;
 	} else { /* Sign only.  */
-		if (krb5_make_checksum(checksum_type, krb5_hdr, text,
+		if (make_checksum(checksum_type, krb5_hdr, 8, text,
 				       &md5cksum))
 			goto out_err;
 	}
@@ -151,7 +154,7 @@ krb5_make_token(struct krb5_ctx *ctx, int qop_req,
 		       md5cksum.data + md5cksum.len - KRB5_CKSUM_LENGTH,
 		       KRB5_CKSUM_LENGTH);
 
-		dprintk("make_seal_token: cksum data: \n");
+		dprintk("RPC:      make_seal_token: cksum data: \n");
 		print_hexl((u32 *) (krb5_hdr + 16), KRB5_CKSUM_LENGTH, 0);
 		break;
 	default:

net/sunrpc/auth_gss/gss_krb5_seqnum.c

@@ -70,7 +70,7 @@ krb5_get_seq_num(struct crypto_tfm *key,
 	s32 code;
 	unsigned char plain[8];
 
-	dprintk("krb5_get_seq_num: \n");
+	dprintk("RPC:      krb5_get_seq_num:\n");
 
 	if ((code = krb5_decrypt(key, cksum, buf, plain, 8)))
 		return code;

net/sunrpc/auth_gss/gss_krb5_unseal.c

@@ -99,6 +99,10 @@ krb5_read_token(struct krb5_ctx *ctx,
 	if (g_verify_token_header(&ctx->mech_used, &bodysize, &ptr, toktype,
 					read_token->len))
 		goto out;
+
+	if ((*ptr++ != ((toktype>>8)&0xff)) || (*ptr++ != (toktype&0xff)))
+		goto out;
+
 	/* XXX sanity-check bodysize?? */
 
 	if (toktype == KG_TOK_WRAP_MSG) {
@@ -149,7 +153,7 @@ krb5_read_token(struct krb5_ctx *ctx,
 
 	switch (signalg) {
 	case SGN_ALG_DES_MAC_MD5:
-		ret = krb5_make_checksum(checksum_type, ptr - 2,
+		ret = make_checksum(checksum_type, ptr - 2, 8,
 					 message_buffer, &md5cksum);
 		if (ret)
 			goto out;
@@ -174,7 +178,7 @@ krb5_read_token(struct krb5_ctx *ctx,
 	if (qop_state)
 		*qop_state = GSS_C_QOP_DEFAULT;
 
-	now = jiffies;
+	now = get_seconds();
 
 	ret = GSS_S_CONTEXT_EXPIRED;
 	if (now > ctx->endtime)

net/sunrpc/auth_gss/gss_mech_switch.c

@@ -36,9 +36,11 @@
 #include <linux/types.h>
 #include <linux/slab.h>
 #include <linux/socket.h>
+#include <linux/module.h>
 #include <linux/sunrpc/msg_prot.h>
 #include <linux/sunrpc/gss_asn1.h>
 #include <linux/sunrpc/auth_gss.h>
+#include <linux/sunrpc/svcauth_gss.h>
 #include <linux/sunrpc/gss_err.h>
 #include <linux/sunrpc/sched.h>
 #include <linux/sunrpc/gss_api.h>
@@ -51,146 +53,186 @@
 static LIST_HEAD(registered_mechs);
 static spinlock_t registered_mechs_lock = SPIN_LOCK_UNLOCKED;
 
-/* Reference counting: The reference count includes the reference in the
- * global registered_mechs list.  That reference will never diseappear
- * (so the reference count will never go below 1) until after the mech
- * is removed from the list.  Nothing can be removed from the list without
- * first getting the registered_mechs_lock, so a gss_api_mech won't diseappear
- * from underneath us while we hold the registered_mech_lock.  */
-
-int
-gss_mech_register(struct xdr_netobj * mech_type, struct gss_api_ops * ops)
+static void
+gss_mech_free(struct gss_api_mech *gm)
 {
-	struct gss_api_mech *gm;
+	struct pf_desc *pf;
+	int i;
 
-	if (!(gm = kmalloc(sizeof(*gm), GFP_KERNEL))) {
-		printk("Failed to allocate memory in gss_mech_register");
-		return -1;
+	for (i = 0; i < gm->gm_pf_num; i++) {
+		pf = &gm->gm_pfs[i];
+		if (pf->auth_domain_name)
+			kfree(pf->auth_domain_name);
+		pf->auth_domain_name = NULL;
 	}
-	gm->gm_oid.len = mech_type->len;
-	if (!(gm->gm_oid.data = kmalloc(mech_type->len, GFP_KERNEL))) {
-		kfree(gm);
-		printk("Failed to allocate memory in gss_mech_register");
-		return -1;
-	}
-	memcpy(gm->gm_oid.data, mech_type->data, mech_type->len);
-	/* We're counting the reference in the registered_mechs list: */
-	atomic_set(&gm->gm_count, 1);
-	gm->gm_ops = ops;
-	
-	spin_lock(&registered_mechs_lock);
-	list_add(&gm->gm_list, &registered_mechs);
-	spin_unlock(&registered_mechs_lock);
-	dprintk("RPC: gss_mech_register: registered mechanism with oid:\n");
-	print_hexl((u32 *)mech_type->data, mech_type->len, 0);
-	return 0;
 }
 
-/* The following must be called with spinlock held: */
-int
-do_gss_mech_unregister(struct gss_api_mech *gm)
+static inline char *
+make_auth_domain_name(char *name)
 {
+	static char	*prefix = "gss/";
+	char		*new;
 
-	list_del(&gm->gm_list);
+	new = kmalloc(strlen(name) + strlen(prefix) + 1, GFP_KERNEL);
+	if (new) {
+		strcpy(new, prefix);
+		strcat(new, name);
+	}
+	return new;
+}
 
-	dprintk("RPC: unregistered mechanism with oid:\n");
-	print_hexl((u32 *)gm->gm_oid.data, gm->gm_oid.len, 0);
-	if (!gss_mech_put(gm)) {
-		dprintk("RPC: We just unregistered a gss_mechanism which"
-				" someone is still using.\n");
-		return -1;
-	} else {
-		return 0;
+static int
+gss_mech_svc_setup(struct gss_api_mech *gm)
+{
+	struct pf_desc *pf;
+	int i, status;
+
+	for (i = 0; i < gm->gm_pf_num; i++) {
+		pf = &gm->gm_pfs[i];
+		pf->auth_domain_name = make_auth_domain_name(pf->name);
+		status = -ENOMEM;
+		if (pf->auth_domain_name == NULL)
+			goto out;
+		status = svcauth_gss_register_pseudoflavor(pf->pseudoflavor,
+							pf->auth_domain_name);
+		if (status)
+			goto out;
 	}
+	return 0;
+out:
+	gss_mech_free(gm);
+	return status;
 }
 
 int
-gss_mech_unregister(struct gss_api_mech *gm)
+gss_mech_register(struct gss_api_mech *gm)
 {
 	int status;
 
+	status = gss_mech_svc_setup(gm);
+	if (status)
+		return status;
 	spin_lock(&registered_mechs_lock);
-	status = do_gss_mech_unregister(gm);
+	list_add(&gm->gm_list, &registered_mechs);
 	spin_unlock(&registered_mechs_lock);
-	return status;
+	dprintk("RPC:      registered gss mechanism %s\n", gm->gm_name);
+	return 0;
 }
 
-int
-gss_mech_unregister_all(void)
-{
-	struct list_head	*pos;
-	struct gss_api_mech	*gm;
-	int			status = 0;
+EXPORT_SYMBOL(gss_mech_register);
 
+void
+gss_mech_unregister(struct gss_api_mech *gm)
+{
 	spin_lock(&registered_mechs_lock);
-	while (!list_empty(&registered_mechs)) {
-		pos = registered_mechs.next;
-		gm = list_entry(pos, struct gss_api_mech, gm_list);
-		if (do_gss_mech_unregister(gm))
-			status = -1;
-	}
+	list_del(&gm->gm_list);
 	spin_unlock(&registered_mechs_lock);
-	return status;
+	dprintk("RPC:      unregistered gss mechanism %s\n", gm->gm_name);
+	gss_mech_free(gm);
 }
 
+EXPORT_SYMBOL(gss_mech_unregister);
+
 struct gss_api_mech *
 gss_mech_get(struct gss_api_mech *gm)
 {
-	atomic_inc(&gm->gm_count);
+	__module_get(gm->gm_owner);
 	return gm;
 }
 
+EXPORT_SYMBOL(gss_mech_get);
+
 struct gss_api_mech *
-gss_mech_get_by_OID(struct xdr_netobj *mech_type)
+gss_mech_get_by_name(char *name)
 {
-	struct gss_api_mech 	*pos, *gm = NULL;
+	struct gss_api_mech	*pos, *gm = NULL;
 
-	dprintk("RPC: gss_mech_get_by_OID searching for mechanism with OID:\n");
-	print_hexl((u32 *)mech_type->data, mech_type->len, 0);
 	spin_lock(&registered_mechs_lock);
 	list_for_each_entry(pos, &registered_mechs, gm_list) {
-		if ((pos->gm_oid.len == mech_type->len)
-			&& !memcmp(pos->gm_oid.data, mech_type->data,
-							mech_type->len)) {
-			gm = gss_mech_get(pos);
+		if (0 == strcmp(name, pos->gm_name)) {
+			if (!try_module_get(pos->gm_owner))
+				continue;
+			gm = pos;
 			break;
 		}
 	}
 	spin_unlock(&registered_mechs_lock);
-	dprintk("RPC: gss_mech_get_by_OID %s it\n", gm ? "found" : "didn't find");
 	return gm;
+
+}
+
+EXPORT_SYMBOL(gss_mech_get_by_name);
+
+static inline int
+mech_supports_pseudoflavor(struct gss_api_mech *gm, u32 pseudoflavor)
+{
+	int i;
+
+	for (i = 0; i < gm->gm_pf_num; i++) {
+		if (gm->gm_pfs[i].pseudoflavor == pseudoflavor)
+			return 1;
+	}
+	return 0;
 }
 
 struct gss_api_mech *
-gss_mech_get_by_name(char *name)
+gss_mech_get_by_pseudoflavor(u32 pseudoflavor)
 {
-	struct gss_api_mech	*pos, *gm = NULL;
+	struct gss_api_mech *pos, *gm = NULL;
 
 	spin_lock(&registered_mechs_lock);
 	list_for_each_entry(pos, &registered_mechs, gm_list) {
-		if (0 == strcmp(name, pos->gm_ops->name)) {
-			gm = gss_mech_get(pos);
-			break;
+		if (!try_module_get(pos->gm_owner))
+			continue;
+		if (!mech_supports_pseudoflavor(pos, pseudoflavor)) {
+			module_put(pos->gm_owner);
+			continue;
 		}
+		gm = pos;
+		break;
 	}
 	spin_unlock(&registered_mechs_lock);
 	return gm;
+}
+
+EXPORT_SYMBOL(gss_mech_get_by_pseudoflavor);
 
+u32
+gss_pseudoflavor_to_service(struct gss_api_mech *gm, u32 pseudoflavor)
+{
+	int i;
+
+	for (i = 0; i < gm->gm_pf_num; i++) {
+		if (gm->gm_pfs[i].pseudoflavor == pseudoflavor)
+			return gm->gm_pfs[i].service;
+	}
+	return 0;
 }
 
-int
-gss_mech_put(struct gss_api_mech * gm)
+EXPORT_SYMBOL(gss_pseudoflavor_to_service);
+
+char *
+gss_service_to_auth_domain_name(struct gss_api_mech *gm, u32 service)
 {
-	if (atomic_dec_and_test(&gm->gm_count)) {
-		if (gm->gm_oid.len >0)
-			kfree(gm->gm_oid.data);
-		kfree(gm);
-		return 1;
-	} else {
-		return 0;
+	int i;
+
+	for (i = 0; i < gm->gm_pf_num; i++) {
+		if (gm->gm_pfs[i].service == service)
+			return gm->gm_pfs[i].auth_domain_name;
 	}
+	return NULL;
 }
 
+EXPORT_SYMBOL(gss_service_to_auth_domain_name);
+
+void
+gss_mech_put(struct gss_api_mech * gm)
+{
+	module_put(gm->gm_owner);
+}
+
+EXPORT_SYMBOL(gss_mech_put);
+
 /* The mech could probably be determined from the token instead, but it's just
  * as easy for now to pass it in. */
 u32
@@ -244,7 +286,8 @@ gss_verify_mic(struct gss_ctx		*context_handle,
 u32
 gss_delete_sec_context(struct gss_ctx	**context_handle)
 {
-	dprintk("gss_delete_sec_context deleting %p\n",*context_handle);
+	dprintk("RPC:      gss_delete_sec_context deleting %p\n",
+			*context_handle);
 
 	if (!*context_handle)
 		return(GSS_S_NO_CONTEXT);

net/sunrpc/auth_gss/gss_pseudoflavors.c

@@ -82,12 +82,13 @@ gss_register_triple(u32 pseudoflavor, struct gss_api_mech *mech,
 
 	spin_lock(&registered_triples_lock);
 	if (do_lookup_triple_by_pseudoflavor(pseudoflavor)) {
-		printk("Registered pseudoflavor %d again\n", pseudoflavor);
+		printk(KERN_WARNING "RPC: Registered pseudoflavor %d again\n",
+				pseudoflavor);
 		goto err_unlock;
 	}
 	list_add(&triple->triples, &registered_triples);
 	spin_unlock(&registered_triples_lock);
-	dprintk("RPC: registered pseudoflavor %d\n", pseudoflavor);
+	dprintk("RPC:      registered pseudoflavor %d\n", pseudoflavor);
 
 	return 0;
 
@@ -145,7 +146,7 @@ gss_cmp_triples(u32 oid_len, char *oid_data, u32 qop, u32 service)
 	oid.len = oid_len;
 	oid.data = oid_data;
 
-	dprintk("RPC: gss_cmp_triples \n");
+	dprintk("RPC:      gss_cmp_triples\n");
 	print_sec_triple(&oid,qop,service);
 
 	spin_lock(&registered_triples_lock);
@@ -158,7 +159,7 @@ gss_cmp_triples(u32 oid_len, char *oid_data, u32 qop, u32 service)
 		}
 	}
 	spin_unlock(&registered_triples_lock);
-	dprintk("RPC: gss_cmp_triples return %d\n", pseudoflavor);
+	dprintk("RPC:      gss_cmp_triples return %d\n", pseudoflavor);
 	return pseudoflavor;
 }
 
@@ -193,8 +194,8 @@ gss_pseudoflavor_to_service(u32 pseudoflavor)
 	triple = do_lookup_triple_by_pseudoflavor(pseudoflavor);
 	spin_unlock(&registered_triples_lock);
 	if (!triple) {
-		dprintk("RPC: gss_pseudoflavor_to_service called with"
-			" unsupported pseudoflavor %d\n", pseudoflavor);
+		dprintk("RPC:      gss_pseudoflavor_to_service called with unsupported pseudoflavor %d\n",
+				pseudoflavor);
 		return 0;
 	}
 	return triple->service;
@@ -211,8 +212,8 @@ gss_pseudoflavor_to_mech(u32 pseudoflavor) {
 	if (triple)
 		mech = gss_mech_get(triple->mech);
 	else
-		dprintk("RPC: gss_pseudoflavor_to_mech called with"
-			" unsupported pseudoflavor %d\n", pseudoflavor);
+		dprintk("RPC:      gss_pseudoflavor_to_mech called with unsupported pseudoflavor %d\n",
+				pseudoflavor);
 	return mech;
 }
 
@@ -223,8 +224,8 @@ gss_pseudoflavor_to_mechOID(u32 pseudoflavor, struct xdr_netobj * oid)
 
 	mech = gss_pseudoflavor_to_mech(pseudoflavor);
 	if (!mech)  {
-		dprintk("RPC: gss_pseudoflavor_to_mechOID called with"
-			" unsupported pseudoflavor %d\n", pseudoflavor);
+		dprintk("RPC:      gss_pseudoflavor_to_mechOID called with unsupported pseudoflavor %d\n",
+				pseudoflavor);
 		        return -1;
 	}
 	oid->len = mech->gm_oid.len;

net/sunrpc/auth_gss/sunrpcgss_syms.c

@@ -10,26 +10,28 @@
 #include <linux/sunrpc/auth_gss.h>
 #include <linux/sunrpc/svcauth_gss.h>
 #include <linux/sunrpc/gss_asn1.h>
+#include <linux/sunrpc/gss_krb5.h>
 
-/* sec_triples: */
-EXPORT_SYMBOL(gss_register_triple);
-EXPORT_SYMBOL(gss_unregister_triple);
-EXPORT_SYMBOL(gss_cmp_triples);
-EXPORT_SYMBOL(gss_pseudoflavor_to_mechOID);
-EXPORT_SYMBOL(gss_pseudoflavor_supported);
-EXPORT_SYMBOL(gss_pseudoflavor_to_service);
+/* svcauth_gss.c: */
 EXPORT_SYMBOL(svcauth_gss_register_pseudoflavor);
 
 /* registering gss mechanisms to the mech switching code: */
 EXPORT_SYMBOL(gss_mech_register);
+EXPORT_SYMBOL(gss_mech_unregister);
 EXPORT_SYMBOL(gss_mech_get);
-EXPORT_SYMBOL(gss_mech_get_by_OID);
+EXPORT_SYMBOL(gss_mech_get_by_pseudoflavor);
+EXPORT_SYMBOL(gss_mech_get_by_name);
 EXPORT_SYMBOL(gss_mech_put);
+EXPORT_SYMBOL(gss_pseudoflavor_to_service);
+EXPORT_SYMBOL(gss_service_to_auth_domain_name);
 
 /* generic functionality in gss code: */
 EXPORT_SYMBOL(g_make_token_header);
 EXPORT_SYMBOL(g_verify_token_header);
 EXPORT_SYMBOL(g_token_size);
+EXPORT_SYMBOL(make_checksum);
+EXPORT_SYMBOL(krb5_encrypt);
+EXPORT_SYMBOL(krb5_decrypt);
 
 /* debug */
 EXPORT_SYMBOL(print_hexl);

net/sunrpc/auth_gss/svcauth_gss.c

@@ -570,14 +570,14 @@ gss_verify_header(struct svc_rqst *rqstp, struct rsc *rsci,
 	}
 
 	if (gc->gc_seq > MAXSEQ) {
-		dprintk("svcauth_gss: discarding request with large"
-			" sequence number %d\n", gc->gc_seq);
+		dprintk("RPC:      svcauth_gss: discarding request with large sequence number %d\n",
+				gc->gc_seq);
 		*authp = rpcsec_gsserr_ctxproblem;
 		return SVC_DENIED;
 	}
 	if (!gss_check_seq_num(rsci, gc->gc_seq)) {
-		dprintk("svcauth_gss: discarding request with old"
-				" sequence number %d\n", gc->gc_seq);
+		dprintk("RPC:      svcauth_gss: discarding request with old sequence number %d\n",
+				gc->gc_seq);
 		return SVC_DROP;
 	}
 	return SVC_OK;
@@ -617,19 +617,15 @@ struct gss_domain {
 	u32			pseudoflavor;
 };
 
-/* XXX this should be done in gss_pseudoflavors, and shouldn't be hardcoded: */
 static struct auth_domain *
 find_gss_auth_domain(struct gss_ctx *ctx, u32 svc)
 {
-	switch(gss_get_pseudoflavor(ctx, 0, svc)) {
-		case RPC_AUTH_GSS_KRB5:
-			return auth_domain_find("gss/krb5");
-		case RPC_AUTH_GSS_KRB5I:
-			return auth_domain_find("gss/krb5i");
-		case RPC_AUTH_GSS_KRB5P:
-			return auth_domain_find("gss/krb5p");
-	}
-	return NULL;
+	char *name;
+
+	name = gss_service_to_auth_domain_name(ctx->mech_type, svc);
+	if (!name)
+		return NULL;
+	return auth_domain_find(name);
 }
 
 int
@@ -637,19 +633,17 @@ svcauth_gss_register_pseudoflavor(u32 pseudoflavor, char * name)
 {
 	struct gss_domain	*new;
 	struct auth_domain	*test;
-	static char		*prefix = "gss/";
-	int			stat = -1;
+	int			stat = -ENOMEM;
 
 	new = kmalloc(sizeof(*new), GFP_KERNEL);
 	if (!new)
 		goto out;
 	cache_init(&new->h.h);
 	atomic_inc(&new->h.h.refcnt);
-	new->h.name = kmalloc(strlen(name) + strlen(prefix) + 1, GFP_KERNEL);
+	new->h.name = kmalloc(strlen(name) + 1, GFP_KERNEL);
 	if (!new->h.name)
 		goto out_free_dom;
-	strcpy(new->h.name, prefix);
-	strcat(new->h.name, name);
+	strcpy(new->h.name, name);
 	new->h.flavour = RPC_AUTH_GSS;
 	new->pseudoflavor = pseudoflavor;
 	new->h.h.expiry_time = NEVER;
@@ -670,6 +664,8 @@ svcauth_gss_register_pseudoflavor(u32 pseudoflavor, char * name)
 	return stat;
 }
 
+EXPORT_SYMBOL(svcauth_gss_register_pseudoflavor);
+
 static inline int
 read_u32_from_xdr_buf(struct xdr_buf *buf, int base, u32 *obj)
 {
@@ -755,7 +751,7 @@ svcauth_gss_accept(struct svc_rqst *rqstp, u32 *authp)
 	u32		*reject_stat = resv->iov_base + resv->iov_len;
 	int		ret;
 
-	dprintk("RPC: svcauth_gss: argv->iov_len = %zd\n",argv->iov_len);
+	dprintk("RPC:      svcauth_gss: argv->iov_len = %zd\n",argv->iov_len);
 
 	*authp = rpc_autherr_badcred;
 	if (!svcdata)

net/sunrpc/clnt.c

@@ -788,13 +788,11 @@ static void
 call_timeout(struct rpc_task *task)
 {
 	struct rpc_clnt	*clnt = task->tk_client;
-	struct rpc_timeout *to = &task->tk_rqstp->rq_timeout;
 
-	if (xprt_adjust_timeout(to)) {
+	if (xprt_adjust_timeout(task->tk_rqstp) == 0) {
 		dprintk("RPC: %4d call_timeout (minor)\n", task->tk_pid);
 		goto retry;
 	}
-	to->to_retries = clnt->cl_timeout.to_retries;
 
 	dprintk("RPC: %4d call_timeout (major)\n", task->tk_pid);
 	if (RPC_IS_SOFT(task)) {

net/sunrpc/timer.c

@@ -39,6 +39,7 @@ rpc_init_rtt(struct rpc_rtt *rt, unsigned long timeo)
 	for (i = 0; i < 5; i++) {
 		rt->srtt[i] = init;
 		rt->sdrtt[i] = RPC_RTO_INIT;
+		rt->ntimeouts[i] = 0;
 	}
 }
 

net/sunrpc/xprt.c

@@ -351,36 +351,58 @@ xprt_adjust_cwnd(struct rpc_xprt *xprt, int result)
 	xprt->cwnd = cwnd;
 }
 
+/*
+ * Reset the major timeout value
+ */
+static void xprt_reset_majortimeo(struct rpc_rqst *req)
+{
+	struct rpc_timeout *to = &req->rq_xprt->timeout;
+
+	req->rq_majortimeo = req->rq_timeout;
+	if (to->to_exponential)
+		req->rq_majortimeo <<= to->to_retries;
+	else
+		req->rq_majortimeo += to->to_increment * to->to_retries;
+	if (req->rq_majortimeo > to->to_maxval || req->rq_majortimeo == 0)
+		req->rq_majortimeo = to->to_maxval;
+	req->rq_majortimeo += jiffies;
+}
+
 /*
  * Adjust timeout values etc for next retransmit
  */
-int
-xprt_adjust_timeout(struct rpc_timeout *to)
+int xprt_adjust_timeout(struct rpc_rqst *req)
 {
-	if (to->to_retries > 0) {
+	struct rpc_xprt *xprt = req->rq_xprt;
+	struct rpc_timeout *to = &xprt->timeout;
+	int status = 0;
+
+	if (time_before(jiffies, req->rq_majortimeo)) {
 		if (to->to_exponential)
-			to->to_current <<= 1;
+			req->rq_timeout <<= 1;
 		else
-			to->to_current += to->to_increment;
-		if (to->to_maxval && to->to_current >= to->to_maxval)
-			to->to_current = to->to_maxval;
+			req->rq_timeout += to->to_increment;
+		if (to->to_maxval && req->rq_timeout >= to->to_maxval)
+			req->rq_timeout = to->to_maxval;
+		req->rq_retries++;
+		pprintk("RPC: %lu retrans\n", jiffies);
 	} else {
-		if (to->to_exponential)
-			to->to_initval <<= 1;
-		else
-			to->to_initval += to->to_increment;
-		if (to->to_maxval && to->to_initval >= to->to_maxval)
-			to->to_initval = to->to_maxval;
-		to->to_current = to->to_initval;
+		req->rq_timeout = to->to_initval;
+		req->rq_retries = 0;
+		xprt_reset_majortimeo(req);
+		/* Reset the RTT counters == "slow start" */
+		spin_lock_bh(&xprt->sock_lock);
+		rpc_init_rtt(req->rq_task->tk_client->cl_rtt, to->to_initval);
+		spin_unlock_bh(&xprt->sock_lock);
+		pprintk("RPC: %lu timeout\n", jiffies);
+		status = -ETIMEDOUT;
 	}
 
-	if (!to->to_current) {
-		printk(KERN_WARNING "xprt_adjust_timeout: to_current = 0!\n");
-		to->to_current = 5 * HZ;
+	if (req->rq_timeout == 0) {
+		printk(KERN_WARNING "xprt_adjust_timeout: rq_timeout = 0!\n");
+		req->rq_timeout = 5 * HZ;
 	}
-	pprintk("RPC: %lu %s\n", jiffies,
-			to->to_retries? "retrans" : "timeout");
-	return to->to_retries-- > 0;
+	return status;
 }
 
 /*
@@ -537,8 +559,17 @@ void xprt_connect(struct rpc_task *task)
 
 	task->tk_timeout = RPC_CONNECT_TIMEOUT;
 	rpc_sleep_on(&xprt->pending, task, xprt_connect_status, NULL);
-	if (!test_and_set_bit(XPRT_CONNECTING, &xprt->sockstate))
-		schedule_work(&xprt->sock_connect);
+	if (!test_and_set_bit(XPRT_CONNECTING, &xprt->sockstate)) {
+		/* Note: if we are here due to a dropped connection
+		 * 	 we delay reconnecting by RPC_REESTABLISH_TIMEOUT/HZ
+		 * 	 seconds
+		 */
+		if (xprt->sock != NULL)
+			schedule_delayed_work(&xprt->sock_connect,
+					RPC_REESTABLISH_TIMEOUT);
+		else
+			schedule_work(&xprt->sock_connect);
+	}
 	return;
  out_write:
 	xprt_release_write(xprt, task);
@@ -566,7 +597,6 @@ xprt_connect_status(struct rpc_task *task)
 	case -ECONNREFUSED:
 	case -ECONNRESET:
 	case -ENOTCONN:
-		rpc_delay(task, RPC_REESTABLISH_TIMEOUT);
 		return;
 	case -ETIMEDOUT:
 		dprintk("RPC: %4d xprt_connect_status: timed out\n",
@@ -1166,6 +1196,7 @@ xprt_transmit(struct rpc_task *task)
 			/* Add request to the receive list */
 			list_add_tail(&req->rq_list, &xprt->recv);
 			spin_unlock_bh(&xprt->sock_lock);
+			xprt_reset_majortimeo(req);
 		}
 	} else if (!req->rq_bytes_sent)
 		return;
@@ -1221,7 +1252,7 @@ xprt_transmit(struct rpc_task *task)
 			if (!xprt_connected(xprt))
 				task->tk_status = -ENOTCONN;
 			else if (test_bit(SOCK_NOSPACE, &xprt->sock->flags)) {
-				task->tk_timeout = req->rq_timeout.to_current;
+				task->tk_timeout = req->rq_timeout;
 				rpc_sleep_on(&xprt->pending, task, NULL, NULL);
 			}
 			spin_unlock_bh(&xprt->sock_lock);
@@ -1248,13 +1279,11 @@ xprt_transmit(struct rpc_task *task)
 	if (!xprt->nocong) {
 		int timer = task->tk_msg.rpc_proc->p_timer;
 		task->tk_timeout = rpc_calc_rto(clnt->cl_rtt, timer);
-		task->tk_timeout <<= rpc_ntimeo(clnt->cl_rtt, timer);
-		task->tk_timeout <<= clnt->cl_timeout.to_retries
-			- req->rq_timeout.to_retries;
-		if (task->tk_timeout > req->rq_timeout.to_maxval)
-			task->tk_timeout = req->rq_timeout.to_maxval;
+		task->tk_timeout <<= rpc_ntimeo(clnt->cl_rtt, timer) + req->rq_retries;
+		if (task->tk_timeout > xprt->timeout.to_maxval || task->tk_timeout == 0)
+			task->tk_timeout = xprt->timeout.to_maxval;
 	} else
-		task->tk_timeout = req->rq_timeout.to_current;
+		task->tk_timeout = req->rq_timeout;
 	/* Don't race with disconnect */
 	if (!xprt_connected(xprt))
 		task->tk_status = -ENOTCONN;
@@ -1324,7 +1353,7 @@ xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt)
 {
 	struct rpc_rqst	*req = task->tk_rqstp;
 
-	req->rq_timeout = xprt->timeout;
+	req->rq_timeout = xprt->timeout.to_initval;
 	req->rq_task	= task;
 	req->rq_xprt    = xprt;
 	req->rq_xid     = xprt_alloc_xid(xprt);
@@ -1381,7 +1410,6 @@ xprt_default_timeout(struct rpc_timeout *to, int proto)
 void
 xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long incr)
 {
-	to->to_current   = 
 	to->to_initval   = 
 	to->to_increment = incr;
 	to->to_maxval    = incr * retr;
@@ -1446,7 +1474,6 @@ xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to)
 	/* Set timeout parameters */
 	if (to) {
 		xprt->timeout = *to;
-		xprt->timeout.to_current = to->to_initval;
 	} else
 		xprt_default_timeout(&xprt->timeout, xprt->prot);