[PATCH] files: break up files struct

In order for the RCU to work, the file table array, sets and their sizes must
be updated atomically.  Instead of ensuring this through too many memory
barriers, we put the arrays and their sizes in a separate structure.  This
patch takes the first step of putting the file table elements in a separate
structure fdtable that is embedded withing files_struct.  It also changes all
the users to refer to the file table using files_fdtable() macro.  Subsequent
applciation of RCU becomes easier after this.

Signed-off-by: Dipankar Sarma <dipankar@in.ibm.com>
Signed-Off-By: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

2 files changed, 63 insertions, 40 deletions

diff --git a/kernel/exit.c b/kernel/exit.c
index 5b0fb9f..83beb1e 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -368,17 +368,19 @@ EXPORT_SYMBOL(daemonize);
 static inline void close_files(struct files_struct * files)
 {
 	int i, j;
+	struct fdtable *fdt;
 
 	j = 0;
+	fdt = files_fdtable(files);
 	for (;;) {
 		unsigned long set;
 		i = j * __NFDBITS;
-		if (i >= files->max_fdset || i >= files->max_fds)
+		if (i >= fdt->max_fdset || i >= fdt->max_fds)
 			break;
-		set = files->open_fds->fds_bits[j++];
+		set = fdt->open_fds->fds_bits[j++];
 		while (set) {
 			if (set & 1) {
-				struct file * file = xchg(&files->fd[i], NULL);
+				struct file * file = xchg(&fdt->fd[i], NULL);
 				if (file)
 					filp_close(file, files);
 			}
@@ -403,16 +405,19 @@ struct files_struct *get_files_struct(struct task_struct *task)
 
 void fastcall put_files_struct(struct files_struct *files)
 {
+	struct fdtable *fdt;
+
 	if (atomic_dec_and_test(&files->count)) {
 		close_files(files);
 		/*
 		 * Free the fd and fdset arrays if we expanded them.
 		 */
-		if (files->fd != &files->fd_array[0])
-			free_fd_array(files->fd, files->max_fds);
-		if (files->max_fdset > __FD_SETSIZE) {
-			free_fdset(files->open_fds, files->max_fdset);
-			free_fdset(files->close_on_exec, files->max_fdset);
+		fdt = files_fdtable(files);
+		if (fdt->fd != &files->fd_array[0])
+			free_fd_array(fdt->fd, fdt->max_fds);
+		if (fdt->max_fdset > __FD_SETSIZE) {
+			free_fdset(fdt->open_fds, fdt->max_fdset);
+			free_fdset(fdt->close_on_exec, fdt->max_fdset);
 		}
 		kmem_cache_free(files_cachep, files);
 	}
diff --git a/kernel/fork.c b/kernel/fork.c
index b258020..ecc694d 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -568,21 +568,47 @@ static inline int copy_fs(unsigned long clone_flags, struct task_struct * tsk)
 static int count_open_files(struct files_struct *files, int size)
 {
 	int i;
+	struct fdtable *fdt;
 
 	/* Find the last open fd */
+	fdt = files_fdtable(files);
 	for (i = size/(8*sizeof(long)); i > 0; ) {
-		if (files->open_fds->fds_bits[--i])
+		if (fdt->open_fds->fds_bits[--i])
 			break;
 	}
 	i = (i+1) * 8 * sizeof(long);
 	return i;
 }
 
+static struct files_struct *alloc_files(void)
+{
+	struct files_struct *newf;
+	struct fdtable *fdt;
+
+	newf = kmem_cache_alloc(files_cachep, SLAB_KERNEL);
+	if (!newf)
+		goto out;
+
+	atomic_set(&newf->count, 1);
+
+	spin_lock_init(&newf->file_lock);
+	fdt = files_fdtable(newf);
+	fdt->next_fd = 0;
+	fdt->max_fds = NR_OPEN_DEFAULT;
+	fdt->max_fdset = __FD_SETSIZE;
+	fdt->close_on_exec = &newf->close_on_exec_init;
+	fdt->open_fds = &newf->open_fds_init;
+	fdt->fd = &newf->fd_array[0];
+out:
+	return newf;
+}
+
 static int copy_files(unsigned long clone_flags, struct task_struct * tsk)
 {
 	struct files_struct *oldf, *newf;
 	struct file **old_fds, **new_fds;
 	int open_files, size, i, error = 0, expand;
+	struct fdtable *old_fdt, *new_fdt;
 
 	/*
 	 * A background process may not have any files ...
@@ -603,35 +629,27 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk)
 	 */
 	tsk->files = NULL;
 	error = -ENOMEM;
-	newf = kmem_cache_alloc(files_cachep, SLAB_KERNEL);
-	if (!newf) 
+	newf = alloc_files();
+	if (!newf)
 		goto out;
 
-	atomic_set(&newf->count, 1);
-
-	spin_lock_init(&newf->file_lock);
-	newf->next_fd	    = 0;
-	newf->max_fds	    = NR_OPEN_DEFAULT;
-	newf->max_fdset	    = __FD_SETSIZE;
-	newf->close_on_exec = &newf->close_on_exec_init;
-	newf->open_fds	    = &newf->open_fds_init;
-	newf->fd	    = &newf->fd_array[0];
-
 	spin_lock(&oldf->file_lock);
-
-	open_files = count_open_files(oldf, oldf->max_fdset);
+	old_fdt = files_fdtable(oldf);
+	new_fdt = files_fdtable(newf);
+	size = old_fdt->max_fdset;
+	open_files = count_open_files(oldf, old_fdt->max_fdset);
 	expand = 0;
 
 	/*
 	 * Check whether we need to allocate a larger fd array or fd set.
 	 * Note: we're not a clone task, so the open count won't  change.
 	 */
-	if (open_files > newf->max_fdset) {
-		newf->max_fdset = 0;
+	if (open_files > new_fdt->max_fdset) {
+		new_fdt->max_fdset = 0;
 		expand = 1;
 	}
-	if (open_files > newf->max_fds) {
-		newf->max_fds = 0;
+	if (open_files > new_fdt->max_fds) {
+		new_fdt->max_fds = 0;
 		expand = 1;
 	}
 
@@ -646,11 +664,11 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk)
 		spin_lock(&oldf->file_lock);
 	}
 
-	old_fds = oldf->fd;
-	new_fds = newf->fd;
+	old_fds = old_fdt->fd;
+	new_fds = new_fdt->fd;
 
-	memcpy(newf->open_fds->fds_bits, oldf->open_fds->fds_bits, open_files/8);
-	memcpy(newf->close_on_exec->fds_bits, oldf->close_on_exec->fds_bits, open_files/8);
+	memcpy(new_fdt->open_fds->fds_bits, old_fdt->open_fds->fds_bits, open_files/8);
+	memcpy(new_fdt->close_on_exec->fds_bits, old_fdt->close_on_exec->fds_bits, open_files/8);
 
 	for (i = open_files; i != 0; i--) {
 		struct file *f = *old_fds++;
@@ -663,24 +681,24 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk)
 			 * is partway through open().  So make sure that this
 			 * fd is available to the new process.
 			 */
-			FD_CLR(open_files - i, newf->open_fds);
+			FD_CLR(open_files - i, new_fdt->open_fds);
 		}
 		*new_fds++ = f;
 	}
 	spin_unlock(&oldf->file_lock);
 
 	/* compute the remainder to be cleared */
-	size = (newf->max_fds - open_files) * sizeof(struct file *);
+	size = (new_fdt->max_fds - open_files) * sizeof(struct file *);
 
 	/* This is long word aligned thus could use a optimized version */ 
 	memset(new_fds, 0, size); 
 
-	if (newf->max_fdset > open_files) {
-		int left = (newf->max_fdset-open_files)/8;
+	if (new_fdt->max_fdset > open_files) {
+		int left = (new_fdt->max_fdset-open_files)/8;
 		int start = open_files / (8 * sizeof(unsigned long));
 
-		memset(&newf->open_fds->fds_bits[start], 0, left);
-		memset(&newf->close_on_exec->fds_bits[start], 0, left);
+		memset(&new_fdt->open_fds->fds_bits[start], 0, left);
+		memset(&new_fdt->close_on_exec->fds_bits[start], 0, left);
 	}
 
 	tsk->files = newf;
@@ -689,9 +707,9 @@ out:
 	return error;
 
 out_release:
-	free_fdset (newf->close_on_exec, newf->max_fdset);
-	free_fdset (newf->open_fds, newf->max_fdset);
-	free_fd_array(newf->fd, newf->max_fds);
+	free_fdset (new_fdt->close_on_exec, new_fdt->max_fdset);
+	free_fdset (new_fdt->open_fds, new_fdt->max_fdset);
+	free_fd_array(new_fdt->fd, new_fdt->max_fds);
 	kmem_cache_free(files_cachep, newf);
 	goto out;
 }