# This introduces the filehandle abstraction, which provides a centralized
# facility for managing file descriptors.  This is needed because descriptors
# are a scarce resource, especially when host-based filesystems are used within
# UML.  A find, kernel build, or updatedb will easily consume the available
# descriptors.  Unrelated, and non-piggy parts of UML, such as process 
# creation, will then start failing.
# This alleviates the problem by allowing file descriptors to be reclaimed
# when the limit is reached.  Host resources which will persist when 
# descriptors to them are closed, such as non-deleted files, will have their
# file handles marked reclaimable, and will provide the information needed
# to reopen the file if it becomes active again.
# Also, some names in hostfs are changed temporarily to avoid name clashes.
Index: um/arch/um/include/filehandle.h
===================================================================
--- um.orig/arch/um/include/filehandle.h	2003-09-15 09:40:47.000000000 -0400
+++ um/arch/um/include/filehandle.h	2004-08-05 20:39:23.000000000 -0400
@@ -0,0 +1,51 @@
+/* 
+ * Copyright (C) 2004 Jeff Dike (jdike@addtoit.com)
+ * Licensed under the GPL
+ */
+
+#ifndef __FILEHANDLE_H__
+#define __FILEHANDLE_H__
+
+#include "linux/list.h"
+#include "linux/fs.h"
+#include "os.h"
+
+struct file_handle {
+	struct list_head list;
+	int fd;
+	char *(*get_name)(struct inode *);
+	struct inode *inode;
+	struct openflags flags;
+};
+
+extern struct file_handle bad_filehandle;
+
+extern int open_file(char *name, struct openflags flags, int mode);
+extern void *open_dir(char *file);
+extern int open_filehandle(char *name, struct openflags flags, int mode, 
+			   struct file_handle *fh);
+extern int read_file(struct file_handle *fh, unsigned long long offset, 
+		     char *buf, int len);
+extern int write_file(struct file_handle *fh, unsigned long long offset, 
+		      const char *buf, int len);
+extern int truncate_file(struct file_handle *fh, unsigned long long size);
+extern int close_file(struct file_handle *fh);
+extern void not_reclaimable(struct file_handle *fh);
+extern void is_reclaimable(struct file_handle *fh, 
+			   char *(name_proc)(struct inode *),
+			   struct inode *inode);
+extern int filehandle_fd(struct file_handle *fh);
+extern int make_pipe(struct file_handle *fhs);
+
+#endif
+
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only.  This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-file-style: "linux"
+ * End:
+ */
Index: um/arch/um/include/os.h
===================================================================
--- um.orig/arch/um/include/os.h	2004-08-05 20:38:09.000000000 -0400
+++ um/arch/um/include/os.h	2004-08-05 20:39:23.000000000 -0400
@@ -133,7 +133,10 @@
 extern int os_mode_fd(int fd, int mode);
 
 extern int os_seek_file(int fd, __u64 offset);
+extern void *os_open_dir(char *dir, int *err_out);
 extern int os_open_file(char *file, struct openflags flags, int mode);
+extern int os_truncate_file(const char *file, unsigned long long len);
+extern int os_truncate_fd(int fd, unsigned long long len);
 extern int os_read_file(int fd, void *buf, int len);
 extern int os_write_file(int fd, const void *buf, int count);
 extern int os_file_size(char *file, long long *size_out);
Index: um/arch/um/kernel/Makefile
===================================================================
--- um.orig/arch/um/kernel/Makefile	2004-08-05 20:38:09.000000000 -0400
+++ um/arch/um/kernel/Makefile	2004-08-05 20:39:23.000000000 -0400
@@ -10,13 +10,13 @@
 
 subdir- = ../util
 
-obj-y = checksum.o config.o exec_kern.o exitcode.o frame_kern.o frame.o \
-	helper.o init_task.o irq.o irq_user.o ksyms.o mem.o mem_user.o \
-	physmem.o process.o process_kern.o ptrace.o reboot.o resource.o \
-	sigio_user.o sigio_kern.o signal_kern.o signal_user.o smp.o \
-	syscall_kern.o syscall_user.o sysrq.o sys_call_table.o tempfile.o \
-	time.o time_kern.o tlb.o trap_kern.o trap_user.o uaccess_user.o \
-	um_arch.o umid.o user_util.o
+obj-y = checksum.o config.o exec_kern.o exitcode.o filehandle.o frame_kern.o \
+	frame.o helper.o init_task.o irq.o irq_user.o ksyms.o mem.o \
+	mem_user.o physmem.o process.o process_kern.o ptrace.o reboot.o \
+	resource.o sigio_user.o sigio_kern.o signal_kern.o signal_user.o \
+	smp.o syscall_kern.o syscall_user.o sysrq.o sys_call_table.o \
+	tempfile.o time.o time_kern.o tlb.o trap_kern.o trap_user.o \
+	uaccess_user.o um_arch.o umid.o user_util.o
 
 obj-$(CONFIG_BLK_DEV_INITRD) += initrd_kern.o initrd_user.o
 obj-$(CONFIG_GPROF)	+= gprof_syms.o
Index: um/arch/um/kernel/filehandle.c
===================================================================
--- um.orig/arch/um/kernel/filehandle.c	2003-09-15 09:40:47.000000000 -0400
+++ um/arch/um/kernel/filehandle.c	2004-08-05 20:39:23.000000000 -0400
@@ -0,0 +1,250 @@
+/* 
+ * Copyright (C) 2004 Jeff Dike (jdike@karaya.com)
+ * Licensed under the GPL
+ */
+
+#include "linux/slab.h"
+#include "linux/list.h"
+#include "linux/spinlock.h"
+#include "linux/fs.h"
+#include "linux/errno.h"
+#include "filehandle.h"
+#include "os.h"
+#include "kern_util.h"
+
+static spinlock_t open_files_lock = SPIN_LOCK_UNLOCKED;
+static struct list_head open_files = LIST_HEAD_INIT(open_files);
+
+#define NUM_RECLAIM 128
+
+static void reclaim_fds(void)
+{
+	struct file_handle *victim;
+	int closed = NUM_RECLAIM;
+
+	spin_lock(&open_files_lock);
+	while(!list_empty(&open_files) && closed--){
+		victim = list_entry(open_files.prev, struct file_handle, list);
+		os_close_file(victim->fd);
+		victim->fd = -1;
+		list_del_init(&victim->list);
+	}
+	spin_unlock(&open_files_lock);
+}
+
+int open_file(char *name, struct openflags flags, int mode)
+{
+	int fd;
+
+	fd = os_open_file(name, flags, mode);
+	if(fd != -EMFILE)
+		return(fd);
+
+	reclaim_fds();
+	fd = os_open_file(name, flags, mode);
+
+	return(fd);
+}
+
+void *open_dir(char *file)
+{
+	void *dir;
+	int err;
+
+	dir = os_open_dir(file, &err);
+	if(dir != NULL)
+		return(dir);
+	if(err != -EMFILE)
+		return(ERR_PTR(err));
+
+	reclaim_fds();
+
+	dir = os_open_dir(file, &err);
+	if(dir == NULL)
+		dir = ERR_PTR(err);
+
+	return(dir);
+}
+
+void not_reclaimable(struct file_handle *fh)
+{
+	char *name;
+
+	if(fh->get_name == NULL)
+		return;
+
+	if(list_empty(&fh->list)){
+		name = (*fh->get_name)(fh->inode);
+		if(name != NULL){
+			fh->fd = open_file(name, fh->flags, 0);
+			kfree(name);
+		}
+		else printk("File descriptor %d has no name\n", fh->fd);
+	}
+	else {
+		spin_lock(&open_files_lock);
+		list_del_init(&fh->list);
+		spin_unlock(&open_files_lock);
+	}
+}
+
+void is_reclaimable(struct file_handle *fh, char *(name_proc)(struct inode *),
+		    struct inode *inode)
+{
+	fh->get_name = name_proc;
+	fh->inode = inode;
+
+	spin_lock(&open_files_lock);
+	list_add(&fh->list, &open_files);
+	spin_unlock(&open_files_lock);
+}
+
+static int active_handle(struct file_handle *fh)
+{
+	int fd;
+	char *name;
+
+	if(!list_empty(&fh->list))
+		list_move(&fh->list, &open_files);
+
+	if(fh->fd != -1)
+		return(0);
+
+	if(fh->inode == NULL)
+		return(-ENOENT);
+
+	name = (*fh->get_name)(fh->inode);
+	if(name == NULL)
+		return(-ENOMEM);
+
+	fd = open_file(name, fh->flags, 0);
+	kfree(name);
+	if(fd < 0)
+		return(fd);
+
+	fh->fd = fd;
+	is_reclaimable(fh, fh->get_name, fh->inode);
+
+	return(0);
+}
+
+int filehandle_fd(struct file_handle *fh)
+{
+	int err;
+
+	err = active_handle(fh);
+	if(err)
+		return(err);
+
+	return(fh->fd);
+}
+
+static void init_fh(struct file_handle *fh, int fd, struct openflags flags)
+{
+	flags.c = 0;
+	*fh = ((struct file_handle) { .list	= LIST_HEAD_INIT(fh->list),
+				      .fd	= fd,
+				      .get_name	= NULL,
+				      .inode	= NULL,
+				      .flags	= flags });
+}
+
+int open_filehandle(char *name, struct openflags flags, int mode, 
+		    struct file_handle *fh)
+{
+	int fd;
+
+	fd = open_file(name, flags, mode);
+	if(fd < 0)
+		return(fd);
+
+	init_fh(fh, fd, flags);
+	return(0);
+}
+
+int close_file(struct file_handle *fh)
+{
+	spin_lock(&open_files_lock);
+	list_del(&fh->list);
+	spin_unlock(&open_files_lock);
+
+	os_close_file(fh->fd);
+
+	fh->fd = -1;
+	return(0);
+}
+
+int read_file(struct file_handle *fh, unsigned long long offset, char *buf,
+	      int len)
+{
+	int err;
+
+	err = active_handle(fh);
+	if(err)
+		return(err);
+
+	err = os_seek_file(fh->fd, offset);
+	if(err)
+		return(err);
+
+	return(os_read_file(fh->fd, buf, len));
+}
+
+int write_file(struct file_handle *fh, unsigned long long offset, 
+	       const char *buf, int len)
+{
+	int err;
+
+	err = active_handle(fh);
+	if(err)
+		return(err);
+
+	if(offset != -1)
+		err = os_seek_file(fh->fd, offset);
+	if(err)
+		return(err);
+
+	return(os_write_file(fh->fd, buf, len));
+}
+
+int truncate_file(struct file_handle *fh, unsigned long long size)
+{
+	int err;
+
+	err = active_handle(fh);
+	if(err)
+		return(err);
+
+	return(os_truncate_fd(fh->fd, size));
+}
+
+int make_pipe(struct file_handle *fhs)
+{
+	int fds[2], err;
+
+	err = os_pipe(fds, 1, 1);
+	if(err && (err != -EMFILE))
+		return(err);
+
+	if(err){
+		reclaim_fds();
+		err = os_pipe(fds, 1, 1);
+	}
+	if(err)
+		return(err);
+
+	init_fh(&fhs[0], fds[0], OPENFLAGS());
+	init_fh(&fhs[1], fds[1], OPENFLAGS());
+	return(0);
+}
+
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only.  This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-file-style: "linux"
+ * End:
+ */
Index: um/arch/um/os-Linux/file.c
===================================================================
--- um.orig/arch/um/os-Linux/file.c	2004-08-05 20:38:09.000000000 -0400
+++ um/arch/um/os-Linux/file.c	2004-08-05 20:39:23.000000000 -0400
@@ -278,6 +278,25 @@
 	return(fd);
 }
 
+void *os_open_dir(char *path, int *err_out)
+{
+	void *dir;
+
+	dir = opendir(path);
+	*err_out = -errno;
+	return(dir);
+}
+
+int os_truncate_fd(int fd, unsigned long long len)
+{
+	int err;
+
+	err = ftruncate(fd, len);
+	if(err)
+		return(-errno);
+	return(0);
+}
+
 int os_connect_socket(char *name)
 {
 	struct sockaddr_un sock;
Index: um/fs/hostfs/hostfs.h
===================================================================
--- um.orig/fs/hostfs/hostfs.h	2004-08-05 20:38:09.000000000 -0400
+++ um/fs/hostfs/hostfs.h	2004-08-05 20:39:23.000000000 -0400
@@ -37,15 +37,15 @@
 		     struct timespec *mtime_out, struct timespec *ctime_out, 
 		     int *blksize_out, unsigned long long *blocks_out);
 extern int access_file(char *path, int r, int w, int x);
-extern int open_file(char *path, int r, int w, int append);
+extern int hostfs_open_file(char *path, int r, int w, int append);
 extern int file_type(const char *path, int *rdev);
-extern void *open_dir(char *path, int *err_out);
+extern void *hostfs_open_dir(char *path, int *err_out);
 extern char *read_dir(void *stream, unsigned long long *pos, 
 		      unsigned long long *ino_out, int *len_out);
-extern void close_file(void *stream);
+extern void hostfs_close_file(void *stream);
 extern void close_dir(void *stream);
-extern int read_file(int fd, unsigned long long *offset, char *buf, int len);
-extern int write_file(int fd, unsigned long long *offset, const char *buf,
+extern int hostfs_read_file(int fd, unsigned long long *offset, char *buf, int len);
+extern int hostfs_write_file(int fd, unsigned long long *offset, const char *buf,
 		      int len);
 extern int lseek_file(int fd, long long offset, int whence);
 extern int file_create(char *name, int ur, int uw, int ux, int gr, 
Index: um/fs/hostfs/hostfs_kern.c
===================================================================
--- um.orig/fs/hostfs/hostfs_kern.c	2004-08-05 20:38:09.000000000 -0400
+++ um/fs/hostfs/hostfs_kern.c	2004-08-05 20:39:23.000000000 -0400
@@ -290,7 +290,7 @@
 		kfree(HOSTFS_I(inode)->host_filename);
 
 	if(HOSTFS_I(inode)->fd != -1) 
-		close_file(&HOSTFS_I(inode)->fd);
+		hostfs_close_file(&HOSTFS_I(inode)->fd);
 
 	kfree(HOSTFS_I(inode));
 }
@@ -316,7 +316,7 @@
 
 	name = dentry_name(file->f_dentry, 0);
 	if(name == NULL) return(-ENOMEM);
-	dir = open_dir(name, &error);
+	dir = hostfs_open_dir(name, &error);
 	kfree(name);
 	if(dir == NULL) return(-error);
 	next = file->f_pos;
@@ -343,7 +343,7 @@
 	 * so this resets things and reopens the file with the new access.
 	 */
 	if(HOSTFS_I(ino)->fd != -1){
-		close_file(&HOSTFS_I(ino)->fd);
+		hostfs_close_file(&HOSTFS_I(ino)->fd);
 		HOSTFS_I(ino)->fd = -1;
 	}
 
@@ -359,7 +359,7 @@
 	if(name == NULL) 
 		return(-ENOMEM);
 
-	fd = open_file(name, r, w, append);
+	fd = hostfs_open_file(name, r, w, append);
 	kfree(name);
 	if(fd < 0) return(fd);
 	FILE_HOSTFS_I(file)->fd = fd;
@@ -403,7 +403,7 @@
 	buffer = kmap(page);
 	base = ((unsigned long long) page->index) << PAGE_CACHE_SHIFT;
 
-	err = write_file(HOSTFS_I(inode)->fd, &base, buffer, count);
+	err = hostfs_write_file(HOSTFS_I(inode)->fd, &base, buffer, count);
 	if(err != count){
 		ClearPageUptodate(page);
 		goto out;
@@ -431,7 +431,7 @@
 
 	start = (long long) page->index << PAGE_CACHE_SHIFT;
 	buffer = kmap(page);
-	err = read_file(FILE_HOSTFS_I(file)->fd, &start, buffer,
+	err = hostfs_read_file(FILE_HOSTFS_I(file)->fd, &start, buffer,
 			PAGE_CACHE_SIZE);
 	if(err < 0) goto out;
 
@@ -458,13 +458,13 @@
 	buffer = kmap(page);
 	if(from != 0){
 		tmp = start;
-		err = read_file(FILE_HOSTFS_I(file)->fd, &tmp, buffer,
+		err = hostfs_read_file(FILE_HOSTFS_I(file)->fd, &tmp, buffer,
 				from);
 		if(err < 0) goto out;
 	}
 	if(to != PAGE_CACHE_SIZE){
 		start += to;
-		err = read_file(FILE_HOSTFS_I(file)->fd, &start, buffer + to,
+		err = hostfs_read_file(FILE_HOSTFS_I(file)->fd, &start, buffer + to,
 				PAGE_CACHE_SIZE - to);
 		if(err < 0) goto out;		
 	}
@@ -485,7 +485,7 @@
 
 	start = (long long) (page->index << PAGE_CACHE_SHIFT) + from;
 	buffer = kmap(page);
-	err = write_file(FILE_HOSTFS_I(file)->fd, &start, buffer + from, 
+	err = hostfs_write_file(FILE_HOSTFS_I(file)->fd, &start, buffer + from, 
 			 to - from);
 	if(err > 0) err = 0;
 	if(!err && (start > inode->i_size))
Index: um/fs/hostfs/hostfs_user.c
===================================================================
--- um.orig/fs/hostfs/hostfs_user.c	2004-08-05 20:38:09.000000000 -0400
+++ um/fs/hostfs/hostfs_user.c	2004-08-05 20:39:23.000000000 -0400
@@ -83,7 +83,7 @@
 	else return(0);
 }
 
-int open_file(char *path, int r, int w, int append)
+int hostfs_open_file(char *path, int r, int w, int append)
 {
 	int mode = 0, fd;
 
@@ -102,7 +102,7 @@
 	else return(fd);
 }
 
-void *open_dir(char *path, int *err_out)
+void *hostfs_open_dir(char *path, int *err_out)
 {
 	DIR *dir;
 
@@ -127,7 +127,7 @@
 	return(ent->d_name);
 }
 
-int read_file(int fd, unsigned long long *offset, char *buf, int len)
+int hostfs_read_file(int fd, unsigned long long *offset, char *buf, int len)
 {
 	int n;
 
@@ -137,7 +137,7 @@
 	return(n);
 }
 
-int write_file(int fd, unsigned long long *offset, const char *buf, int len)
+int hostfs_write_file(int fd, unsigned long long *offset, const char *buf, int len)
 {
 	int n;
 
@@ -156,7 +156,7 @@
 	return(0);
 }
 
-void close_file(void *stream)
+void hostfs_close_file(void *stream)
 {
 	close(*((int *) stream));
 }