[RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[David Howells]

We don't need to use O_LARGEFILE anymore inside the kernel and can for the
most part just let it sit in the UAPI headers and ignore it everywhere else.
Indeed on 64-bit arches, we enforce the use of O_LARGEFILE.

Quite a few places turn it on, but only the following places actually take
any notice of it directly:

 (1) fs/9p/vfs_inode_dotl.c.  Converts O_LARGEFILE to P9_DOTL_LARGEFILE but
     otherwise seems to ignore it.  Possibly it gets sent to a server.

 (2) fs/notify/fanotify/fanotify_user.c.  Pass in event_f_flags.

 (3) fs/hfsplus/inode.c.  Length check in hfsplus_file_open().

 (4) fs/open.c:  Length check in ftruncate().

 (5) fs/open.c:  Length check in generic_file_open().

 (6) fs/xfs/xfs_file.c:  Length check in xfs_file_open().

 (7) mm/filemap.c:  Length check in generic_write_checks().

All but the first two are just making length checks that are waived
unconditionally on a 64-bit system.  Just skip the length checks, assuming
that O_LARGEFILE is actually set.

There's also things like fcntl(F_GETFL) and fuse that pass it to userspace.
NFS may be in the same boat, but if it is, I can't see where the
flag->protocol translation takes place for it.  For F_GETFL, fuse and
fanotify, just set it unconditionally in the places that then pass it to
userspace.

We don't actually then set it in file->f_flags - except when userspace
passes it in - since nothing then examines the bit in f_flags.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/compat.c                        |    9 +++------
 fs/fcntl.c                         |    2 +-
 fs/fhandle.c                       |    8 +-------
 fs/fuse/file.c                     |    7 ++++---
 fs/hfsplus/inode.c                 |    2 --
 fs/notify/fanotify/fanotify_user.c |    3 +--
 fs/open.c                          |   28 ++++------------------------
 fs/xfs/xfs_file.c                  |    2 --
 fs/xfs/xfs_ioctl.c                 |    4 ----
 mm/filemap.c                       |   10 ----------
 10 files changed, 14 insertions(+), 61 deletions(-)

diff --git a/fs/compat.c b/fs/compat.c
index 6fd272d455e4..51579f09a135 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -1084,8 +1084,7 @@ COMPAT_SYSCALL_DEFINE3(getdents64, unsigned int, fd,
 #endif /* __ARCH_WANT_COMPAT_SYS_GETDENTS64 */
 
 /*
- * Exactly like fs/open.c:sys_open(), except that it doesn't set the
- * O_LARGEFILE flag.
+ * Exactly like fs/open.c:sys_open().
  */
 COMPAT_SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, umode_t, mode)
 {
@@ -1093,8 +1092,7 @@ COMPAT_SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, umode_t,
 }
 
 /*
- * Exactly like fs/open.c:sys_openat(), except that it doesn't set the
- * O_LARGEFILE flag.
+ * Exactly like fs/open.c:sys_openat().
  */
 COMPAT_SYSCALL_DEFINE4(openat, int, dfd, const char __user *, filename, int, flags, umode_t, mode)
 {
@@ -1470,8 +1468,7 @@ COMPAT_SYSCALL_DEFINE5(ppoll, struct pollfd __user *, ufds,
 
 #ifdef CONFIG_FHANDLE
 /*
- * Exactly like fs/open.c:sys_open_by_handle_at(), except that it
- * doesn't set the O_LARGEFILE flag.
+ * Exactly like fs/open.c:sys_open_by_handle_at().
  */
 COMPAT_SYSCALL_DEFINE3(open_by_handle_at, int, mountdirfd,
 			     struct file_handle __user *, handle, int, flags)
diff --git a/fs/fcntl.c b/fs/fcntl.c
index ee85cd4e136a..a7a8e07e0d59 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -260,7 +260,7 @@ static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
 		set_close_on_exec(fd, arg & FD_CLOEXEC);
 		break;
 	case F_GETFL:
-		err = filp->f_flags;
+		err = filp->f_flags | O_LARGEFILE;
 		break;
 	case F_SETFL:
 		err = setfl(fd, filp, arg);
diff --git a/fs/fhandle.c b/fs/fhandle.c
index d59712dfa3e7..93948172d186 100644
--- a/fs/fhandle.c
+++ b/fs/fhandle.c
@@ -256,11 +256,5 @@ SYSCALL_DEFINE3(open_by_handle_at, int, mountdirfd,
 		struct file_handle __user *, handle,
 		int, flags)
 {
-	long ret;
-
-	if (force_o_largefile())
-		flags |= O_LARGEFILE;
-
-	ret = do_handle_open(mountdirfd, handle, flags);
-	return ret;
+	return do_handle_open(mountdirfd, handle, flags);
 }
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index f523f2f04c19..680bcbb9613b 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -28,6 +28,7 @@ static int fuse_send_open(struct fuse_conn *fc, u64 nodeid, struct file *file,
 
 	memset(&inarg, 0, sizeof(inarg));
 	inarg.flags = file->f_flags & ~(O_CREAT | O_EXCL | O_NOCTTY);
+	inarg.flags |= O_LARGEFILE;
 	if (!fc->atomic_o_trunc)
 		inarg.flags &= ~O_TRUNC;
 	args.in.h.opcode = opcode;
@@ -235,7 +236,7 @@ static void fuse_prepare_release(struct fuse_file *ff, int flags, int opcode)
 	wake_up_interruptible_all(&ff->poll_wait);
 
 	inarg->fh = ff->fh;
-	inarg->flags = flags;
+	inarg->flags = flags | O_LARGEFILE;
 	req->in.h.opcode = opcode;
 	req->in.h.nodeid = ff->nodeid;
 	req->in.numargs = 1;
@@ -505,7 +506,7 @@ void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos,
 	inarg->fh = ff->fh;
 	inarg->offset = pos;
 	inarg->size = count;
-	inarg->flags = file->f_flags;
+	inarg->flags = file->f_flags | O_LARGEFILE;
 	req->in.h.opcode = opcode;
 	req->in.h.nodeid = ff->nodeid;
 	req->in.numargs = 1;
@@ -947,7 +948,7 @@ static size_t fuse_send_write(struct fuse_req *req, struct fuse_io_priv *io,
 	struct fuse_write_in *inarg = &req->misc.write.in;
 
 	fuse_write_fill(req, ff, pos, count);
-	inarg->flags = file->f_flags;
+	inarg->flags = file->f_flags | O_LARGEFILE;
 	if (owner != NULL) {
 		inarg->write_flags |= FUSE_WRITE_LOCKOWNER;
 		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c
index 6dd107d7421e..0d73a3732532 100644
--- a/fs/hfsplus/inode.c
+++ b/fs/hfsplus/inode.c
@@ -216,8 +216,6 @@ static int hfsplus_file_open(struct inode *inode, struct file *file)
 {
 	if (HFSPLUS_IS_RSRC(inode))
 		inode = HFSPLUS_I(inode)->rsrc_inode;
-	if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
-		return -EOVERFLOW;
 	atomic_inc(&HFSPLUS_I(inode)->opencnt);
 	return 0;
 }
diff --git a/fs/notify/fanotify/fanotify_user.c b/fs/notify/fanotify/fanotify_user.c
index 8e8e6bcd1d43..eaf202ef3d99 100644
--- a/fs/notify/fanotify/fanotify_user.c
+++ b/fs/notify/fanotify/fanotify_user.c
@@ -748,8 +748,7 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 	}
 	group->overflow_event = &oevent->fse;
 
-	if (force_o_largefile())
-		event_f_flags |= O_LARGEFILE;
+	event_f_flags |= O_LARGEFILE;
 	group->fanotify_data.f_flags = event_f_flags;
 #ifdef CONFIG_FANOTIFY_ACCESS_PERMISSIONS
 	spin_lock_init(&group->fanotify_data.access_lock);
diff --git a/fs/open.c b/fs/open.c
index b6f1e96a7c0b..d1fbb8ee69f6 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -151,7 +151,7 @@ COMPAT_SYSCALL_DEFINE2(truncate, const char __user *, path, compat_off_t, length
 }
 #endif
 
-static long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
+static long do_sys_ftruncate(unsigned int fd, loff_t length)
 {
 	struct inode *inode;
 	struct dentry *dentry;
@@ -166,21 +166,12 @@ static long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
 	if (!f.file)
 		goto out;
 
-	/* explicitly opened as large or we are on 64-bit box */
-	if (f.file->f_flags & O_LARGEFILE)
-		small = 0;
-
 	dentry = f.file->f_path.dentry;
 	inode = dentry->d_inode;
 	error = -EINVAL;
 	if (!S_ISREG(inode->i_mode) || !(f.file->f_mode & FMODE_WRITE))
 		goto out_putf;
 
-	error = -EINVAL;
-	/* Cannot ftruncate over 2^31 bytes without large file support */
-	if (small && length > MAX_NON_LFS)
-		goto out_putf;
-
 	error = -EPERM;
 	if (IS_APPEND(inode))
 		goto out_putf;
@@ -200,13 +191,13 @@ out:
 
 SYSCALL_DEFINE2(ftruncate, unsigned int, fd, unsigned long, length)
 {
-	return do_sys_ftruncate(fd, length, 1);
+	return do_sys_ftruncate(fd, length);
 }
 
 #ifdef CONFIG_COMPAT
 COMPAT_SYSCALL_DEFINE2(ftruncate, unsigned int, fd, compat_ulong_t, length)
 {
-	return do_sys_ftruncate(fd, length, 1);
+	return do_sys_ftruncate(fd, length);
 }
 #endif
 
@@ -219,7 +210,7 @@ SYSCALL_DEFINE2(truncate64, const char __user *, path, loff_t, length)
 
 SYSCALL_DEFINE2(ftruncate64, unsigned int, fd, loff_t, length)
 {
-	return do_sys_ftruncate(fd, length, 0);
+	return do_sys_ftruncate(fd, length);
 }
 #endif /* BITS_PER_LONG == 32 */
 
@@ -1037,18 +1028,12 @@ long do_sys_open(int dfd, const char __user *filename, int flags, umode_t mode)
 
 SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, umode_t, mode)
 {
-	if (force_o_largefile())
-		flags |= O_LARGEFILE;
-
 	return do_sys_open(AT_FDCWD, filename, flags, mode);
 }
 
 SYSCALL_DEFINE4(openat, int, dfd, const char __user *, filename, int, flags,
 		umode_t, mode)
 {
-	if (force_o_largefile())
-		flags |= O_LARGEFILE;
-
 	return do_sys_open(dfd, filename, flags, mode);
 }
 
@@ -1126,14 +1111,9 @@ SYSCALL_DEFINE0(vhangup)
 
 /*
  * Called when an inode is about to be open.
- * We use this to disallow opening large files on 32bit systems if
- * the caller didn't specify O_LARGEFILE.  On 64bit systems we force
- * on this flag in sys_open.
  */
 int generic_file_open(struct inode * inode, struct file * filp)
 {
-	if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
-		return -EOVERFLOW;
 	return 0;
 }
 
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index e78feb400e22..7c8d9b4e44eb 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -1031,8 +1031,6 @@ xfs_file_open(
 	struct inode	*inode,
 	struct file	*file)
 {
-	if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
-		return -EFBIG;
 	if (XFS_FORCED_SHUTDOWN(XFS_M(inode->i_sb)))
 		return -EIO;
 	return 0;
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index ea7d85af5310..0c5b4ae746c0 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -218,10 +218,6 @@ xfs_open_by_handle(
 		goto out_dput;
 	}
 
-#if BITS_PER_LONG != 32
-	hreq->oflags |= O_LARGEFILE;
-#endif
-
 	permflag = hreq->oflags;
 	fmode = OPEN_FMODE(permflag);
 	if ((!(permflag & O_APPEND) || (permflag & O_TRUNC)) &&
diff --git a/mm/filemap.c b/mm/filemap.c
index 72940fb38666..1d1e75aa49cf 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2311,16 +2311,6 @@ inline ssize_t generic_write_checks(struct kiocb *iocb, struct iov_iter *from)
 	}
 
 	/*
-	 * LFS rule
-	 */
-	if (unlikely(pos + iov_iter_count(from) > MAX_NON_LFS &&
-				!(file->f_flags & O_LARGEFILE))) {
-		if (pos >= MAX_NON_LFS)
-			return -EFBIG;
-		iov_iter_truncate(from, MAX_NON_LFS - (unsigned long)pos);
-	}
-
-	/*
 	 * Are we about to exceed the fs block limit ?
 	 *
 	 * If we have written data it becomes a short write.  If we have

[RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[David Howells]

We don't need to use O_LARGEFILE anymore inside the kernel and can for the
most part just let it sit in the UAPI headers and ignore it everywhere else.
Indeed on 64-bit arches, we enforce the use of O_LARGEFILE.

Quite a few places turn it on, but only the following places actually take
any notice of it directly:

 (1) fs/9p/vfs_inode_dotl.c.  Converts O_LARGEFILE to P9_DOTL_LARGEFILE but
     otherwise seems to ignore it.  Possibly it gets sent to a server.

 (2) fs/notify/fanotify/fanotify_user.c.  Pass in event_f_flags.

 (3) fs/hfsplus/inode.c.  Length check in hfsplus_file_open().

 (4) fs/open.c:  Length check in ftruncate().

 (5) fs/open.c:  Length check in generic_file_open().

 (6) fs/xfs/xfs_file.c:  Length check in xfs_file_open().

 (7) mm/filemap.c:  Length check in generic_write_checks().

All but the first two are just making length checks that are waived
unconditionally on a 64-bit system.  Just skip the length checks, assuming
that O_LARGEFILE is actually set.

There's also things like fcntl(F_GETFL) and fuse that pass it to userspace.
NFS may be in the same boat, but if it is, I can't see where the
flag->protocol translation takes place for it.  For F_GETFL, fuse and
fanotify, just set it unconditionally in the places that then pass it to
userspace.

We don't actually then set it in file->f_flags - except when userspace
passes it in - since nothing then examines the bit in f_flags.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/compat.c                        |    9 +++------
 fs/fcntl.c                         |    2 +-
 fs/fhandle.c                       |    8 +-------
 fs/fuse/file.c                     |    7 ++++---
 fs/hfsplus/inode.c                 |    2 --
 fs/notify/fanotify/fanotify_user.c |    3 +--
 fs/open.c                          |   28 ++++------------------------
 fs/xfs/xfs_file.c                  |    2 --
 fs/xfs/xfs_ioctl.c                 |    4 ----
 mm/filemap.c                       |   10 ----------
 10 files changed, 14 insertions(+), 61 deletions(-)

diff --git a/fs/compat.c b/fs/compat.c
index 6fd272d455e4..51579f09a135 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -1084,8 +1084,7 @@ COMPAT_SYSCALL_DEFINE3(getdents64, unsigned int, fd,
 #endif /* __ARCH_WANT_COMPAT_SYS_GETDENTS64 */
 
 /*
- * Exactly like fs/open.c:sys_open(), except that it doesn't set the
- * O_LARGEFILE flag.
+ * Exactly like fs/open.c:sys_open().
  */
 COMPAT_SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, umode_t, mode)
 {
@@ -1093,8 +1092,7 @@ COMPAT_SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, umode_t,
 }
 
 /*
- * Exactly like fs/open.c:sys_openat(), except that it doesn't set the
- * O_LARGEFILE flag.
+ * Exactly like fs/open.c:sys_openat().
  */
 COMPAT_SYSCALL_DEFINE4(openat, int, dfd, const char __user *, filename, int, flags, umode_t, mode)
 {
@@ -1470,8 +1468,7 @@ COMPAT_SYSCALL_DEFINE5(ppoll, struct pollfd __user *, ufds,
 
 #ifdef CONFIG_FHANDLE
 /*
- * Exactly like fs/open.c:sys_open_by_handle_at(), except that it
- * doesn't set the O_LARGEFILE flag.
+ * Exactly like fs/open.c:sys_open_by_handle_at().
  */
 COMPAT_SYSCALL_DEFINE3(open_by_handle_at, int, mountdirfd,
 			     struct file_handle __user *, handle, int, flags)
diff --git a/fs/fcntl.c b/fs/fcntl.c
index ee85cd4e136a..a7a8e07e0d59 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -260,7 +260,7 @@ static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
 		set_close_on_exec(fd, arg & FD_CLOEXEC);
 		break;
 	case F_GETFL:
-		err = filp->f_flags;
+		err = filp->f_flags | O_LARGEFILE;
 		break;
 	case F_SETFL:
 		err = setfl(fd, filp, arg);
diff --git a/fs/fhandle.c b/fs/fhandle.c
index d59712dfa3e7..93948172d186 100644
--- a/fs/fhandle.c
+++ b/fs/fhandle.c
@@ -256,11 +256,5 @@ SYSCALL_DEFINE3(open_by_handle_at, int, mountdirfd,
 		struct file_handle __user *, handle,
 		int, flags)
 {
-	long ret;
-
-	if (force_o_largefile())
-		flags |= O_LARGEFILE;
-
-	ret = do_handle_open(mountdirfd, handle, flags);
-	return ret;
+	return do_handle_open(mountdirfd, handle, flags);
 }
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index f523f2f04c19..680bcbb9613b 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -28,6 +28,7 @@ static int fuse_send_open(struct fuse_conn *fc, u64 nodeid, struct file *file,
 
 	memset(&inarg, 0, sizeof(inarg));
 	inarg.flags = file->f_flags & ~(O_CREAT | O_EXCL | O_NOCTTY);
+	inarg.flags |= O_LARGEFILE;
 	if (!fc->atomic_o_trunc)
 		inarg.flags &= ~O_TRUNC;
 	args.in.h.opcode = opcode;
@@ -235,7 +236,7 @@ static void fuse_prepare_release(struct fuse_file *ff, int flags, int opcode)
 	wake_up_interruptible_all(&ff->poll_wait);
 
 	inarg->fh = ff->fh;
-	inarg->flags = flags;
+	inarg->flags = flags | O_LARGEFILE;
 	req->in.h.opcode = opcode;
 	req->in.h.nodeid = ff->nodeid;
 	req->in.numargs = 1;
@@ -505,7 +506,7 @@ void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos,
 	inarg->fh = ff->fh;
 	inarg->offset = pos;
 	inarg->size = count;
-	inarg->flags = file->f_flags;
+	inarg->flags = file->f_flags | O_LARGEFILE;
 	req->in.h.opcode = opcode;
 	req->in.h.nodeid = ff->nodeid;
 	req->in.numargs = 1;
@@ -947,7 +948,7 @@ static size_t fuse_send_write(struct fuse_req *req, struct fuse_io_priv *io,
 	struct fuse_write_in *inarg = &req->misc.write.in;
 
 	fuse_write_fill(req, ff, pos, count);
-	inarg->flags = file->f_flags;
+	inarg->flags = file->f_flags | O_LARGEFILE;
 	if (owner != NULL) {
 		inarg->write_flags |= FUSE_WRITE_LOCKOWNER;
 		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c
index 6dd107d7421e..0d73a3732532 100644
--- a/fs/hfsplus/inode.c
+++ b/fs/hfsplus/inode.c
@@ -216,8 +216,6 @@ static int hfsplus_file_open(struct inode *inode, struct file *file)
 {
 	if (HFSPLUS_IS_RSRC(inode))
 		inode = HFSPLUS_I(inode)->rsrc_inode;
-	if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
-		return -EOVERFLOW;
 	atomic_inc(&HFSPLUS_I(inode)->opencnt);
 	return 0;
 }
diff --git a/fs/notify/fanotify/fanotify_user.c b/fs/notify/fanotify/fanotify_user.c
index 8e8e6bcd1d43..eaf202ef3d99 100644
--- a/fs/notify/fanotify/fanotify_user.c
+++ b/fs/notify/fanotify/fanotify_user.c
@@ -748,8 +748,7 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 	}
 	group->overflow_event = &oevent->fse;
 
-	if (force_o_largefile())
-		event_f_flags |= O_LARGEFILE;
+	event_f_flags |= O_LARGEFILE;
 	group->fanotify_data.f_flags = event_f_flags;
 #ifdef CONFIG_FANOTIFY_ACCESS_PERMISSIONS
 	spin_lock_init(&group->fanotify_data.access_lock);
diff --git a/fs/open.c b/fs/open.c
index b6f1e96a7c0b..d1fbb8ee69f6 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -151,7 +151,7 @@ COMPAT_SYSCALL_DEFINE2(truncate, const char __user *, path, compat_off_t, length
 }
 #endif
 
-static long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
+static long do_sys_ftruncate(unsigned int fd, loff_t length)
 {
 	struct inode *inode;
 	struct dentry *dentry;
@@ -166,21 +166,12 @@ static long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
 	if (!f.file)
 		goto out;
 
-	/* explicitly opened as large or we are on 64-bit box */
-	if (f.file->f_flags & O_LARGEFILE)
-		small = 0;
-
 	dentry = f.file->f_path.dentry;
 	inode = dentry->d_inode;
 	error = -EINVAL;
 	if (!S_ISREG(inode->i_mode) || !(f.file->f_mode & FMODE_WRITE))
 		goto out_putf;
 
-	error = -EINVAL;
-	/* Cannot ftruncate over 2^31 bytes without large file support */
-	if (small && length > MAX_NON_LFS)
-		goto out_putf;
-
 	error = -EPERM;
 	if (IS_APPEND(inode))
 		goto out_putf;
@@ -200,13 +191,13 @@ out:
 
 SYSCALL_DEFINE2(ftruncate, unsigned int, fd, unsigned long, length)
 {
-	return do_sys_ftruncate(fd, length, 1);
+	return do_sys_ftruncate(fd, length);
 }
 
 #ifdef CONFIG_COMPAT
 COMPAT_SYSCALL_DEFINE2(ftruncate, unsigned int, fd, compat_ulong_t, length)
 {
-	return do_sys_ftruncate(fd, length, 1);
+	return do_sys_ftruncate(fd, length);
 }
 #endif
 
@@ -219,7 +210,7 @@ SYSCALL_DEFINE2(truncate64, const char __user *, path, loff_t, length)
 
 SYSCALL_DEFINE2(ftruncate64, unsigned int, fd, loff_t, length)
 {
-	return do_sys_ftruncate(fd, length, 0);
+	return do_sys_ftruncate(fd, length);
 }
 #endif /* BITS_PER_LONG == 32 */
 
@@ -1037,18 +1028,12 @@ long do_sys_open(int dfd, const char __user *filename, int flags, umode_t mode)
 
 SYSCALL_DEFINE3(open, const char __user *, filename, int, flags, umode_t, mode)
 {
-	if (force_o_largefile())
-		flags |= O_LARGEFILE;
-
 	return do_sys_open(AT_FDCWD, filename, flags, mode);
 }
 
 SYSCALL_DEFINE4(openat, int, dfd, const char __user *, filename, int, flags,
 		umode_t, mode)
 {
-	if (force_o_largefile())
-		flags |= O_LARGEFILE;
-
 	return do_sys_open(dfd, filename, flags, mode);
 }
 
@@ -1126,14 +1111,9 @@ SYSCALL_DEFINE0(vhangup)
 
 /*
  * Called when an inode is about to be open.
- * We use this to disallow opening large files on 32bit systems if
- * the caller didn't specify O_LARGEFILE.  On 64bit systems we force
- * on this flag in sys_open.
  */
 int generic_file_open(struct inode * inode, struct file * filp)
 {
-	if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
-		return -EOVERFLOW;
 	return 0;
 }
 
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index e78feb400e22..7c8d9b4e44eb 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -1031,8 +1031,6 @@ xfs_file_open(
 	struct inode	*inode,
 	struct file	*file)
 {
-	if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
-		return -EFBIG;
 	if (XFS_FORCED_SHUTDOWN(XFS_M(inode->i_sb)))
 		return -EIO;
 	return 0;
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index ea7d85af5310..0c5b4ae746c0 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -218,10 +218,6 @@ xfs_open_by_handle(
 		goto out_dput;
 	}
 
-#if BITS_PER_LONG != 32
-	hreq->oflags |= O_LARGEFILE;
-#endif
-
 	permflag = hreq->oflags;
 	fmode = OPEN_FMODE(permflag);
 	if ((!(permflag & O_APPEND) || (permflag & O_TRUNC)) &&
diff --git a/mm/filemap.c b/mm/filemap.c
index 72940fb38666..1d1e75aa49cf 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -2311,16 +2311,6 @@ inline ssize_t generic_write_checks(struct kiocb *iocb, struct iov_iter *from)
 	}
 
 	/*
-	 * LFS rule
-	 */
-	if (unlikely(pos + iov_iter_count(from) > MAX_NON_LFS &&
-				!(file->f_flags & O_LARGEFILE))) {
-		if (pos >= MAX_NON_LFS)
-			return -EFBIG;
-		iov_iter_truncate(from, MAX_NON_LFS - (unsigned long)pos);
-	}
-
-	/*
 	 * Are we about to exceed the fs block limit ?
 	 *
 	 * If we have written data it becomes a short write.  If we have

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

[RFC PATCH 2/2] VFS: Don't pass O_LARGEFILE when opening a file internally

[David Howells]

Don't pass O_LARGEFILE when opening a file internally within the kernel
because due to the preceeding commit, the flag is now assumed fixed on for
all arches, not just 64-bit ones.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 drivers/media/pci/cx25821/cx25821-audio-upstream.c |    4 ++--
 drivers/mtd/nand/nandsim.c                         |    2 +-
 drivers/staging/lustre/lustre/libcfs/tracefile.c   |    4 ++--
 drivers/target/target_core_file.c                  |    4 ++--
 drivers/usb/gadget/function/storage_common.c       |    4 ++--
 fs/block_dev.c                                     |    2 --
 fs/cachefiles/rdwr.c                               |    2 +-
 fs/coredump.c                                      |    3 +--
 fs/ecryptfs/kthread.c                              |    4 ++--
 fs/exec.c                                          |    4 ++--
 fs/nfsd/vfs.c                                      |    6 +++---
 kernel/acct.c                                      |    2 +-
 mm/shmem.c                                         |    2 +-
 mm/swapfile.c                                      |    4 ++--
 tools/lguest/lguest.c                              |    2 +-
 15 files changed, 23 insertions(+), 26 deletions(-)

diff --git a/drivers/media/pci/cx25821/cx25821-audio-upstream.c b/drivers/media/pci/cx25821/cx25821-audio-upstream.c
index 68dbc2dbc982..aa6011bb88cf 100644
--- a/drivers/media/pci/cx25821/cx25821-audio-upstream.c
+++ b/drivers/media/pci/cx25821/cx25821-audio-upstream.c
@@ -271,7 +271,7 @@ static int cx25821_get_audio_data(struct cx25821_dev *dev,
 	if (dev->_audiofile_status == END_OF_FILE)
 		return 0;
 
-	file = filp_open(dev->_audiofilename, O_RDONLY | O_LARGEFILE, 0);
+	file = filp_open(dev->_audiofilename, O_RDONLY, 0);
 	if (IS_ERR(file)) {
 		pr_err("%s(): ERROR opening file(%s) with errno = %ld!\n",
 		       __func__, dev->_audiofilename, -PTR_ERR(file));
@@ -326,7 +326,7 @@ static int cx25821_openfile_audio(struct cx25821_dev *dev,
 	loff_t offset;
 	int i, j;
 
-	file = filp_open(dev->_audiofilename, O_RDONLY | O_LARGEFILE, 0);
+	file = filp_open(dev->_audiofilename, O_RDONLY, 0);
 	if (IS_ERR(file)) {
 		pr_err("%s(): ERROR opening file(%s) with errno = %ld!\n",
 			__func__, dev->_audiofilename, PTR_ERR(file));
diff --git a/drivers/mtd/nand/nandsim.c b/drivers/mtd/nand/nandsim.c
index 95d0cc49cfc2..f932022e2ff0 100644
--- a/drivers/mtd/nand/nandsim.c
+++ b/drivers/mtd/nand/nandsim.c
@@ -575,7 +575,7 @@ static int alloc_device(struct nandsim *ns)
 	int i, err;
 
 	if (cache_file) {
-		cfile = filp_open(cache_file, O_CREAT | O_RDWR | O_LARGEFILE, 0600);
+		cfile = filp_open(cache_file, O_CREAT | O_RDWR, 0600);
 		if (IS_ERR(cfile))
 			return PTR_ERR(cfile);
 		if (!(cfile->f_mode & FMODE_CAN_READ)) {
diff --git a/drivers/staging/lustre/lustre/libcfs/tracefile.c b/drivers/staging/lustre/lustre/libcfs/tracefile.c
index effa2af58c13..a862b2f21111 100644
--- a/drivers/staging/lustre/lustre/libcfs/tracefile.c
+++ b/drivers/staging/lustre/lustre/libcfs/tracefile.c
@@ -683,7 +683,7 @@ int cfs_tracefile_dump_all_pages(char *filename)
 
 	cfs_tracefile_write_lock();
 
-	filp = filp_open(filename, O_CREAT|O_EXCL|O_WRONLY|O_LARGEFILE, 0600);
+	filp = filp_open(filename, O_CREAT|O_EXCL|O_WRONLY, 0600);
 	if (IS_ERR(filp)) {
 		rc = PTR_ERR(filp);
 		filp = NULL;
@@ -985,7 +985,7 @@ static int tracefiled(void *arg)
 		cfs_tracefile_read_lock();
 		if (cfs_tracefile[0] != 0) {
 			filp = filp_open(cfs_tracefile,
-					 O_CREAT | O_RDWR | O_LARGEFILE,
+					 O_CREAT | O_RDWR,
 					 0600);
 			if (IS_ERR(filp)) {
 				rc = PTR_ERR(filp);
diff --git a/drivers/target/target_core_file.c b/drivers/target/target_core_file.c
index e3195700211a..e7474fc79ea8 100644
--- a/drivers/target/target_core_file.c
+++ b/drivers/target/target_core_file.c
@@ -114,7 +114,7 @@ static int fd_configure_device(struct se_device *dev)
 	 * Use O_DSYNC by default instead of O_SYNC to forgo syncing
 	 * of pure timestamp updates.
 	 */
-	flags = O_RDWR | O_CREAT | O_LARGEFILE | O_DSYNC;
+	flags = O_RDWR | O_CREAT | O_DSYNC;
 
 	/*
 	 * Optionally allow fd_buffered_io=1 to be enabled for people
@@ -732,7 +732,7 @@ static int fd_init_prot(struct se_device *dev)
 	struct fd_dev *fd_dev = FD_DEV(dev);
 	struct file *prot_file, *file = fd_dev->fd_file;
 	struct inode *inode;
-	int ret, flags = O_RDWR | O_CREAT | O_LARGEFILE | O_DSYNC;
+	int ret, flags = O_RDWR | O_CREAT | O_DSYNC;
 	char buf[FD_MAX_DEV_PROT_NAME];
 
 	if (!file) {
diff --git a/drivers/usb/gadget/function/storage_common.c b/drivers/usb/gadget/function/storage_common.c
index d62683017cf3..73d33930beb8 100644
--- a/drivers/usb/gadget/function/storage_common.c
+++ b/drivers/usb/gadget/function/storage_common.c
@@ -196,12 +196,12 @@ int fsg_lun_open(struct fsg_lun *curlun, const char *filename)
 	/* R/W if we can, R/O if we must */
 	ro = curlun->initially_ro;
 	if (!ro) {
-		filp = filp_open(filename, O_RDWR | O_LARGEFILE, 0);
+		filp = filp_open(filename, O_RDWR, 0);
 		if (PTR_ERR(filp) == -EROFS || PTR_ERR(filp) == -EACCES)
 			ro = 1;
 	}
 	if (ro)
-		filp = filp_open(filename, O_RDONLY | O_LARGEFILE, 0);
+		filp = filp_open(filename, O_RDONLY, 0);
 	if (IS_ERR(filp)) {
 		LINFO(curlun, "unable to open backing file: %s\n", filename);
 		return PTR_ERR(filp);
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 22ea424ee741..170abedc983a 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -1462,8 +1462,6 @@ static int blkdev_open(struct inode * inode, struct file * filp)
 	 * binary needs it. We might want to drop this workaround
 	 * during an unstable branch.
 	 */
-	filp->f_flags |= O_LARGEFILE;
-
 	if (filp->f_flags & O_NDELAY)
 		filp->f_mode |= FMODE_NDELAY;
 	if (filp->f_flags & O_EXCL)
diff --git a/fs/cachefiles/rdwr.c b/fs/cachefiles/rdwr.c
index 3cbb0e834694..9b0c7348cf22 100644
--- a/fs/cachefiles/rdwr.c
+++ b/fs/cachefiles/rdwr.c
@@ -909,7 +909,7 @@ int cachefiles_write_page(struct fscache_storage *op, struct page *page)
 	 * own time */
 	path.mnt = cache->mnt;
 	path.dentry = object->backer;
-	file = dentry_open(&path, O_RDWR | O_LARGEFILE, cache->cache_cred);
+	file = dentry_open(&path, O_RDWR, cache->cache_cred);
 	if (IS_ERR(file)) {
 		ret = PTR_ERR(file);
 	} else {
diff --git a/fs/coredump.c b/fs/coredump.c
index a8f75640ac86..8e4a0feb145d 100644
--- a/fs/coredump.c
+++ b/fs/coredump.c
@@ -667,8 +667,7 @@ void do_coredump(const siginfo_t *siginfo)
 		 * writes its coredump successfully, not which one.
 		 */
 		cprm.file = filp_open(cn.corename,
-				 O_CREAT | 2 | O_NOFOLLOW |
-				 O_LARGEFILE | O_EXCL,
+				 O_CREAT | 2 | O_NOFOLLOW | O_EXCL,
 				 0600);
 		if (IS_ERR(cprm.file))
 			goto fail_unlock;
diff --git a/fs/ecryptfs/kthread.c b/fs/ecryptfs/kthread.c
index 866bb18efefe..3fef39d3345f 100644
--- a/fs/ecryptfs/kthread.c
+++ b/fs/ecryptfs/kthread.c
@@ -74,7 +74,7 @@ static int ecryptfs_threadfn(void *ignored)
 					       kthread_ctl_list);
 			list_del(&req->kthread_ctl_list);
 			*req->lower_file = dentry_open(&req->path,
-				(O_RDWR | O_LARGEFILE), current_cred());
+						       O_RDWR, current_cred());
 			complete(&req->done);
 		}
 		mutex_unlock(&ecryptfs_kthread_ctl.mux);
@@ -133,7 +133,7 @@ int ecryptfs_privileged_open(struct file **lower_file,
 			     const struct cred *cred)
 {
 	struct ecryptfs_open_req req;
-	int flags = O_LARGEFILE;
+	int flags = 0;
 	int rc = 0;
 
 	init_completion(&req.done);
diff --git a/fs/exec.c b/fs/exec.c
index b06623a9347f..35668aaa3400 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -118,7 +118,7 @@ SYSCALL_DEFINE1(uselib, const char __user *, library)
 	struct filename *tmp = getname(library);
 	int error = PTR_ERR(tmp);
 	static const struct open_flags uselib_flags = {
-		.open_flag = O_LARGEFILE | O_RDONLY | __FMODE_EXEC,
+		.open_flag = O_RDONLY | __FMODE_EXEC,
 		.acc_mode = MAY_READ | MAY_EXEC | MAY_OPEN,
 		.intent = LOOKUP_OPEN,
 		.lookup_flags = LOOKUP_FOLLOW,
@@ -762,7 +762,7 @@ static struct file *do_open_execat(int fd, struct filename *name, int flags)
 	struct file *file;
 	int err;
 	struct open_flags open_exec_flags = {
-		.open_flag = O_LARGEFILE | O_RDONLY | __FMODE_EXEC,
+		.open_flag = O_RDONLY | __FMODE_EXEC,
 		.acc_mode = MAY_EXEC | MAY_OPEN,
 		.intent = LOOKUP_OPEN,
 		.lookup_flags = LOOKUP_FOLLOW,
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 45c04979e7b3..41adbadf9a01 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -640,7 +640,7 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type,
 	struct path	path;
 	struct inode	*inode;
 	struct file	*file;
-	int		flags = O_RDONLY|O_LARGEFILE;
+	int		flags = O_RDONLY;
 	__be32		err;
 	int		host_err = 0;
 
@@ -690,9 +690,9 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type,
 
 	if (may_flags & NFSD_MAY_WRITE) {
 		if (may_flags & NFSD_MAY_READ)
-			flags = O_RDWR|O_LARGEFILE;
+			flags = O_RDWR;
 		else
-			flags = O_WRONLY|O_LARGEFILE;
+			flags = O_WRONLY;
 	}
 
 	file = dentry_open(&path, flags, current_cred());
diff --git a/kernel/acct.c b/kernel/acct.c
index 74963d192c5d..c292fdf35757 100644
--- a/kernel/acct.c
+++ b/kernel/acct.c
@@ -201,7 +201,7 @@ static int acct_on(struct filename *pathname)
 		return -ENOMEM;
 
 	/* Difference from BSD - they don't do O_APPEND */
-	file = file_open_name(pathname, O_WRONLY|O_APPEND|O_LARGEFILE, 0);
+	file = file_open_name(pathname, O_WRONLY|O_APPEND, 0);
 	if (IS_ERR(file)) {
 		kfree(acct);
 		return PTR_ERR(file);
diff --git a/mm/shmem.c b/mm/shmem.c
index 48ce82926d93..4073e785cda4 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -2968,7 +2968,7 @@ SYSCALL_DEFINE2(memfd_create,
 	}
 	info = SHMEM_I(file_inode(file));
 	file->f_mode |= FMODE_LSEEK | FMODE_PREAD | FMODE_PWRITE;
-	file->f_flags |= O_RDWR | O_LARGEFILE;
+	file->f_flags |= O_RDWR;
 	if (flags & MFD_ALLOW_SEALING)
 		info->seals &= ~F_SEAL_SEAL;
 
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 58877312cf6b..53e854ab0960 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -1863,7 +1863,7 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 	if (IS_ERR(pathname))
 		return PTR_ERR(pathname);
 
-	victim = file_open_name(pathname, O_RDWR|O_LARGEFILE, 0);
+	victim = file_open_name(pathname, O_RDWR, 0);
 	err = PTR_ERR(victim);
 	if (IS_ERR(victim))
 		goto out;
@@ -2419,7 +2419,7 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 		name = NULL;
 		goto bad_swap;
 	}
-	swap_file = file_open_name(name, O_RDWR|O_LARGEFILE, 0);
+	swap_file = file_open_name(name, O_RDWR, 0);
 	if (IS_ERR(swap_file)) {
 		error = PTR_ERR(swap_file);
 		swap_file = NULL;
diff --git a/tools/lguest/lguest.c b/tools/lguest/lguest.c
index 80159e6811c2..a82e6bc0978e 100644
--- a/tools/lguest/lguest.c
+++ b/tools/lguest/lguest.c
@@ -3046,7 +3046,7 @@ static void setup_block_file(const char *filename)
 	vblk = dev->priv = malloc(sizeof(*vblk));
 
 	/* First we open the file and store the length. */
-	vblk->fd = open_or_die(filename, O_RDWR|O_LARGEFILE);
+	vblk->fd = open_or_die(filename, O_RDWR);
 	vblk->len = lseek64(vblk->fd, 0, SEEK_END);
 
 	/* Tell Guest how many sectors this device has. */

[RFC PATCH 2/2] VFS: Don't pass O_LARGEFILE when opening a file internally

[David Howells]

Don't pass O_LARGEFILE when opening a file internally within the kernel
because due to the preceeding commit, the flag is now assumed fixed on for
all arches, not just 64-bit ones.

Signed-off-by: David Howells <dhowells@redhat.com>
---

 drivers/media/pci/cx25821/cx25821-audio-upstream.c |    4 ++--
 drivers/mtd/nand/nandsim.c                         |    2 +-
 drivers/staging/lustre/lustre/libcfs/tracefile.c   |    4 ++--
 drivers/target/target_core_file.c                  |    4 ++--
 drivers/usb/gadget/function/storage_common.c       |    4 ++--
 fs/block_dev.c                                     |    2 --
 fs/cachefiles/rdwr.c                               |    2 +-
 fs/coredump.c                                      |    3 +--
 fs/ecryptfs/kthread.c                              |    4 ++--
 fs/exec.c                                          |    4 ++--
 fs/nfsd/vfs.c                                      |    6 +++---
 kernel/acct.c                                      |    2 +-
 mm/shmem.c                                         |    2 +-
 mm/swapfile.c                                      |    4 ++--
 tools/lguest/lguest.c                              |    2 +-
 15 files changed, 23 insertions(+), 26 deletions(-)

diff --git a/drivers/media/pci/cx25821/cx25821-audio-upstream.c b/drivers/media/pci/cx25821/cx25821-audio-upstream.c
index 68dbc2dbc982..aa6011bb88cf 100644
--- a/drivers/media/pci/cx25821/cx25821-audio-upstream.c
+++ b/drivers/media/pci/cx25821/cx25821-audio-upstream.c
@@ -271,7 +271,7 @@ static int cx25821_get_audio_data(struct cx25821_dev *dev,
 	if (dev->_audiofile_status == END_OF_FILE)
 		return 0;
 
-	file = filp_open(dev->_audiofilename, O_RDONLY | O_LARGEFILE, 0);
+	file = filp_open(dev->_audiofilename, O_RDONLY, 0);
 	if (IS_ERR(file)) {
 		pr_err("%s(): ERROR opening file(%s) with errno = %ld!\n",
 		       __func__, dev->_audiofilename, -PTR_ERR(file));
@@ -326,7 +326,7 @@ static int cx25821_openfile_audio(struct cx25821_dev *dev,
 	loff_t offset;
 	int i, j;
 
-	file = filp_open(dev->_audiofilename, O_RDONLY | O_LARGEFILE, 0);
+	file = filp_open(dev->_audiofilename, O_RDONLY, 0);
 	if (IS_ERR(file)) {
 		pr_err("%s(): ERROR opening file(%s) with errno = %ld!\n",
 			__func__, dev->_audiofilename, PTR_ERR(file));
diff --git a/drivers/mtd/nand/nandsim.c b/drivers/mtd/nand/nandsim.c
index 95d0cc49cfc2..f932022e2ff0 100644
--- a/drivers/mtd/nand/nandsim.c
+++ b/drivers/mtd/nand/nandsim.c
@@ -575,7 +575,7 @@ static int alloc_device(struct nandsim *ns)
 	int i, err;
 
 	if (cache_file) {
-		cfile = filp_open(cache_file, O_CREAT | O_RDWR | O_LARGEFILE, 0600);
+		cfile = filp_open(cache_file, O_CREAT | O_RDWR, 0600);
 		if (IS_ERR(cfile))
 			return PTR_ERR(cfile);
 		if (!(cfile->f_mode & FMODE_CAN_READ)) {
diff --git a/drivers/staging/lustre/lustre/libcfs/tracefile.c b/drivers/staging/lustre/lustre/libcfs/tracefile.c
index effa2af58c13..a862b2f21111 100644
--- a/drivers/staging/lustre/lustre/libcfs/tracefile.c
+++ b/drivers/staging/lustre/lustre/libcfs/tracefile.c
@@ -683,7 +683,7 @@ int cfs_tracefile_dump_all_pages(char *filename)
 
 	cfs_tracefile_write_lock();
 
-	filp = filp_open(filename, O_CREAT|O_EXCL|O_WRONLY|O_LARGEFILE, 0600);
+	filp = filp_open(filename, O_CREAT|O_EXCL|O_WRONLY, 0600);
 	if (IS_ERR(filp)) {
 		rc = PTR_ERR(filp);
 		filp = NULL;
@@ -985,7 +985,7 @@ static int tracefiled(void *arg)
 		cfs_tracefile_read_lock();
 		if (cfs_tracefile[0] != 0) {
 			filp = filp_open(cfs_tracefile,
-					 O_CREAT | O_RDWR | O_LARGEFILE,
+					 O_CREAT | O_RDWR,
 					 0600);
 			if (IS_ERR(filp)) {
 				rc = PTR_ERR(filp);
diff --git a/drivers/target/target_core_file.c b/drivers/target/target_core_file.c
index e3195700211a..e7474fc79ea8 100644
--- a/drivers/target/target_core_file.c
+++ b/drivers/target/target_core_file.c
@@ -114,7 +114,7 @@ static int fd_configure_device(struct se_device *dev)
 	 * Use O_DSYNC by default instead of O_SYNC to forgo syncing
 	 * of pure timestamp updates.
 	 */
-	flags = O_RDWR | O_CREAT | O_LARGEFILE | O_DSYNC;
+	flags = O_RDWR | O_CREAT | O_DSYNC;
 
 	/*
 	 * Optionally allow fd_buffered_io=1 to be enabled for people
@@ -732,7 +732,7 @@ static int fd_init_prot(struct se_device *dev)
 	struct fd_dev *fd_dev = FD_DEV(dev);
 	struct file *prot_file, *file = fd_dev->fd_file;
 	struct inode *inode;
-	int ret, flags = O_RDWR | O_CREAT | O_LARGEFILE | O_DSYNC;
+	int ret, flags = O_RDWR | O_CREAT | O_DSYNC;
 	char buf[FD_MAX_DEV_PROT_NAME];
 
 	if (!file) {
diff --git a/drivers/usb/gadget/function/storage_common.c b/drivers/usb/gadget/function/storage_common.c
index d62683017cf3..73d33930beb8 100644
--- a/drivers/usb/gadget/function/storage_common.c
+++ b/drivers/usb/gadget/function/storage_common.c
@@ -196,12 +196,12 @@ int fsg_lun_open(struct fsg_lun *curlun, const char *filename)
 	/* R/W if we can, R/O if we must */
 	ro = curlun->initially_ro;
 	if (!ro) {
-		filp = filp_open(filename, O_RDWR | O_LARGEFILE, 0);
+		filp = filp_open(filename, O_RDWR, 0);
 		if (PTR_ERR(filp) == -EROFS || PTR_ERR(filp) == -EACCES)
 			ro = 1;
 	}
 	if (ro)
-		filp = filp_open(filename, O_RDONLY | O_LARGEFILE, 0);
+		filp = filp_open(filename, O_RDONLY, 0);
 	if (IS_ERR(filp)) {
 		LINFO(curlun, "unable to open backing file: %s\n", filename);
 		return PTR_ERR(filp);
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 22ea424ee741..170abedc983a 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -1462,8 +1462,6 @@ static int blkdev_open(struct inode * inode, struct file * filp)
 	 * binary needs it. We might want to drop this workaround
 	 * during an unstable branch.
 	 */
-	filp->f_flags |= O_LARGEFILE;
-
 	if (filp->f_flags & O_NDELAY)
 		filp->f_mode |= FMODE_NDELAY;
 	if (filp->f_flags & O_EXCL)
diff --git a/fs/cachefiles/rdwr.c b/fs/cachefiles/rdwr.c
index 3cbb0e834694..9b0c7348cf22 100644
--- a/fs/cachefiles/rdwr.c
+++ b/fs/cachefiles/rdwr.c
@@ -909,7 +909,7 @@ int cachefiles_write_page(struct fscache_storage *op, struct page *page)
 	 * own time */
 	path.mnt = cache->mnt;
 	path.dentry = object->backer;
-	file = dentry_open(&path, O_RDWR | O_LARGEFILE, cache->cache_cred);
+	file = dentry_open(&path, O_RDWR, cache->cache_cred);
 	if (IS_ERR(file)) {
 		ret = PTR_ERR(file);
 	} else {
diff --git a/fs/coredump.c b/fs/coredump.c
index a8f75640ac86..8e4a0feb145d 100644
--- a/fs/coredump.c
+++ b/fs/coredump.c
@@ -667,8 +667,7 @@ void do_coredump(const siginfo_t *siginfo)
 		 * writes its coredump successfully, not which one.
 		 */
 		cprm.file = filp_open(cn.corename,
-				 O_CREAT | 2 | O_NOFOLLOW |
-				 O_LARGEFILE | O_EXCL,
+				 O_CREAT | 2 | O_NOFOLLOW | O_EXCL,
 				 0600);
 		if (IS_ERR(cprm.file))
 			goto fail_unlock;
diff --git a/fs/ecryptfs/kthread.c b/fs/ecryptfs/kthread.c
index 866bb18efefe..3fef39d3345f 100644
--- a/fs/ecryptfs/kthread.c
+++ b/fs/ecryptfs/kthread.c
@@ -74,7 +74,7 @@ static int ecryptfs_threadfn(void *ignored)
 					       kthread_ctl_list);
 			list_del(&req->kthread_ctl_list);
 			*req->lower_file = dentry_open(&req->path,
-				(O_RDWR | O_LARGEFILE), current_cred());
+						       O_RDWR, current_cred());
 			complete(&req->done);
 		}
 		mutex_unlock(&ecryptfs_kthread_ctl.mux);
@@ -133,7 +133,7 @@ int ecryptfs_privileged_open(struct file **lower_file,
 			     const struct cred *cred)
 {
 	struct ecryptfs_open_req req;
-	int flags = O_LARGEFILE;
+	int flags = 0;
 	int rc = 0;
 
 	init_completion(&req.done);
diff --git a/fs/exec.c b/fs/exec.c
index b06623a9347f..35668aaa3400 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -118,7 +118,7 @@ SYSCALL_DEFINE1(uselib, const char __user *, library)
 	struct filename *tmp = getname(library);
 	int error = PTR_ERR(tmp);
 	static const struct open_flags uselib_flags = {
-		.open_flag = O_LARGEFILE | O_RDONLY | __FMODE_EXEC,
+		.open_flag = O_RDONLY | __FMODE_EXEC,
 		.acc_mode = MAY_READ | MAY_EXEC | MAY_OPEN,
 		.intent = LOOKUP_OPEN,
 		.lookup_flags = LOOKUP_FOLLOW,
@@ -762,7 +762,7 @@ static struct file *do_open_execat(int fd, struct filename *name, int flags)
 	struct file *file;
 	int err;
 	struct open_flags open_exec_flags = {
-		.open_flag = O_LARGEFILE | O_RDONLY | __FMODE_EXEC,
+		.open_flag = O_RDONLY | __FMODE_EXEC,
 		.acc_mode = MAY_EXEC | MAY_OPEN,
 		.intent = LOOKUP_OPEN,
 		.lookup_flags = LOOKUP_FOLLOW,
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 45c04979e7b3..41adbadf9a01 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -640,7 +640,7 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type,
 	struct path	path;
 	struct inode	*inode;
 	struct file	*file;
-	int		flags = O_RDONLY|O_LARGEFILE;
+	int		flags = O_RDONLY;
 	__be32		err;
 	int		host_err = 0;
 
@@ -690,9 +690,9 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type,
 
 	if (may_flags & NFSD_MAY_WRITE) {
 		if (may_flags & NFSD_MAY_READ)
-			flags = O_RDWR|O_LARGEFILE;
+			flags = O_RDWR;
 		else
-			flags = O_WRONLY|O_LARGEFILE;
+			flags = O_WRONLY;
 	}
 
 	file = dentry_open(&path, flags, current_cred());
diff --git a/kernel/acct.c b/kernel/acct.c
index 74963d192c5d..c292fdf35757 100644
--- a/kernel/acct.c
+++ b/kernel/acct.c
@@ -201,7 +201,7 @@ static int acct_on(struct filename *pathname)
 		return -ENOMEM;
 
 	/* Difference from BSD - they don't do O_APPEND */
-	file = file_open_name(pathname, O_WRONLY|O_APPEND|O_LARGEFILE, 0);
+	file = file_open_name(pathname, O_WRONLY|O_APPEND, 0);
 	if (IS_ERR(file)) {
 		kfree(acct);
 		return PTR_ERR(file);
diff --git a/mm/shmem.c b/mm/shmem.c
index 48ce82926d93..4073e785cda4 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -2968,7 +2968,7 @@ SYSCALL_DEFINE2(memfd_create,
 	}
 	info = SHMEM_I(file_inode(file));
 	file->f_mode |= FMODE_LSEEK | FMODE_PREAD | FMODE_PWRITE;
-	file->f_flags |= O_RDWR | O_LARGEFILE;
+	file->f_flags |= O_RDWR;
 	if (flags & MFD_ALLOW_SEALING)
 		info->seals &= ~F_SEAL_SEAL;
 
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 58877312cf6b..53e854ab0960 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -1863,7 +1863,7 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 	if (IS_ERR(pathname))
 		return PTR_ERR(pathname);
 
-	victim = file_open_name(pathname, O_RDWR|O_LARGEFILE, 0);
+	victim = file_open_name(pathname, O_RDWR, 0);
 	err = PTR_ERR(victim);
 	if (IS_ERR(victim))
 		goto out;
@@ -2419,7 +2419,7 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 		name = NULL;
 		goto bad_swap;
 	}
-	swap_file = file_open_name(name, O_RDWR|O_LARGEFILE, 0);
+	swap_file = file_open_name(name, O_RDWR, 0);
 	if (IS_ERR(swap_file)) {
 		error = PTR_ERR(swap_file);
 		swap_file = NULL;
diff --git a/tools/lguest/lguest.c b/tools/lguest/lguest.c
index 80159e6811c2..a82e6bc0978e 100644
--- a/tools/lguest/lguest.c
+++ b/tools/lguest/lguest.c
@@ -3046,7 +3046,7 @@ static void setup_block_file(const char *filename)
 	vblk = dev->priv = malloc(sizeof(*vblk));
 
 	/* First we open the file and store the length. */
-	vblk->fd = open_or_die(filename, O_RDWR|O_LARGEFILE);
+	vblk->fd = open_or_die(filename, O_RDWR);
 	vblk->len = lseek64(vblk->fd, 0, SEEK_END);
 
 	/* Tell Guest how many sectors this device has. */

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[Theodore Ts'o]

On Tue, Sep 22, 2015 at 04:24:50PM +0100, David Howells wrote:
> 
>  (4) fs/open.c:  Length check in ftruncate().
> 
>  (5) fs/open.c:  Length check in generic_file_open().
> 
> All but the first two are just making length checks that are waived
> unconditionally on a 64-bit system.  Just skip the length checks, assuming
> that O_LARGEFILE is actually set.

So what this means is that on 32-bit systems, if we have a userspace
program which isn't using the Largefile-enabled, and it opens a file
which is larger than can be addressed with a 32-bit off_t, it can get
surprised and possibly cause data loss.

Is this something we are willing to live with?  After all, there was a
originally a really good reason for the O_LARGEFILE flag in the first
place, and it was primarily about making sure that a non-LARGEFILE
capable program would hard fail on the open, instead of after it had
trashed the user's data.  Granted that 32-bit systems are rarer these
days, and hopefully this isn't a situation that would come up that
often in embedded systems, but if breaking this functionality is
something that we are deliberately going to be doing, we should
discuss it explicitly, and document the decision in the commit
message.

Was there a reason that motivated this change, other than just an
clean up?

						- Ted

Re: [RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[Theodore Ts'o]

On Tue, Sep 22, 2015 at 04:24:50PM +0100, David Howells wrote:
> 
>  (4) fs/open.c:  Length check in ftruncate().
> 
>  (5) fs/open.c:  Length check in generic_file_open().
> 
> All but the first two are just making length checks that are waived
> unconditionally on a 64-bit system.  Just skip the length checks, assuming
> that O_LARGEFILE is actually set.

So what this means is that on 32-bit systems, if we have a userspace
program which isn't using the Largefile-enabled, and it opens a file
which is larger than can be addressed with a 32-bit off_t, it can get
surprised and possibly cause data loss.

Is this something we are willing to live with?  After all, there was a
originally a really good reason for the O_LARGEFILE flag in the first
place, and it was primarily about making sure that a non-LARGEFILE
capable program would hard fail on the open, instead of after it had
trashed the user's data.  Granted that 32-bit systems are rarer these
days, and hopefully this isn't a situation that would come up that
often in embedded systems, but if breaking this functionality is
something that we are deliberately going to be doing, we should
discuss it explicitly, and document the decision in the commit
message.

Was there a reason that motivated this change, other than just an
clean up?

						- Ted

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[David Howells]

Theodore Ts'o <tytso@mit.edu> wrote:

> So what this means is that on 32-bit systems, if we have a userspace
> program which isn't using the Largefile-enabled, and it opens a file
> which is larger than can be addressed with a 32-bit off_t, it can get
> surprised and possibly cause data loss.

Good point.  I was initially thinking that 32-bit userspace on a 64-bit system
would have O_LARGEFILE automatically enabled - but I guess it'll trap through
the compat entry points which avoid that.

That said, fanotify and xfs_open_by_handle() will both automatically set
O_LARGEFILE irrespectively of the 32-bitness of the original caller.

Further, path-based truncate() makes no checks based on file-largeness, unlike
ftruncate().

> Is this something we are willing to live with?  After all, there was a
> originally a really good reason for the O_LARGEFILE flag in the first
> place, and it was primarily about making sure that a non-LARGEFILE
> capable program would hard fail on the open, instead of after it had
> trashed the user's data.

Okay, that seems reasonable - but it still leaves truncate() dangling.  I'm
not sure there's a good answer to that, though.

> Was there a reason that motivated this change, other than just an
> clean up?

Overlayfs and one or two other places need to potentially apply O_LARGEFILE to
the things that they do on behalf of userspace - but other than suppressing
some size checks, it seems to be ignored by the filesystems and the VM.

I vaguely seem to remember that at one point there were still filesystems that
couldn't handle large files and would reject such opens - but they appear to
all have been fixed.

David

Re: [RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[David Howells]

Theodore Ts'o <tytso@mit.edu> wrote:

> So what this means is that on 32-bit systems, if we have a userspace
> program which isn't using the Largefile-enabled, and it opens a file
> which is larger than can be addressed with a 32-bit off_t, it can get
> surprised and possibly cause data loss.

Good point.  I was initially thinking that 32-bit userspace on a 64-bit system
would have O_LARGEFILE automatically enabled - but I guess it'll trap through
the compat entry points which avoid that.

That said, fanotify and xfs_open_by_handle() will both automatically set
O_LARGEFILE irrespectively of the 32-bitness of the original caller.

Further, path-based truncate() makes no checks based on file-largeness, unlike
ftruncate().

> Is this something we are willing to live with?  After all, there was a
> originally a really good reason for the O_LARGEFILE flag in the first
> place, and it was primarily about making sure that a non-LARGEFILE
> capable program would hard fail on the open, instead of after it had
> trashed the user's data.

Okay, that seems reasonable - but it still leaves truncate() dangling.  I'm
not sure there's a good answer to that, though.

> Was there a reason that motivated this change, other than just an
> clean up?

Overlayfs and one or two other places need to potentially apply O_LARGEFILE to
the things that they do on behalf of userspace - but other than suppressing
some size checks, it seems to be ignored by the filesystems and the VM.

I vaguely seem to remember that at one point there were still filesystems that
couldn't handle large files and would reject such opens - but they appear to
all have been fixed.

David

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[Theodore Ts'o]

On Tue, Sep 22, 2015 at 05:12:42PM +0100, David Howells wrote:
> 
> Further, path-based truncate() makes no checks based on file-largeness, unlike
> ftruncate().

Right, but truncate in general is used to make files *smaller* so I'm
having trouble thinking of a scenario where a largefile-oblivious
program could get in trouble by truncating a file > 2TB to some
hard-coded length (normally zero).

> Overlayfs and one or two other places need to potentially apply O_LARGEFILE to
> the things that they do on behalf of userspace - but other than suppressing
> some size checks, it seems to be ignored by the filesystems and the VM.

The size checks really were the primary points of O_LARGEFILE.  As I
recall the primary system calls where this really matters is open(2)
and stat(2) (since if st_size is too small to represent the size of
the file, then the user space program could get really confused).

Essentially O_LARGEFILE is an assertion that userspace can handle
handle 64-bit files, and won't get confused by system call interfaces
where off_t is 32-bit wide, because it will use the 64-bit variants.
So it's not at all surprising that the file systems and the VM in
general doesn't need to worry about the flag.

    	       	   	      	      - Ted

Re: [RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[Theodore Ts'o]

On Tue, Sep 22, 2015 at 05:12:42PM +0100, David Howells wrote:
> 
> Further, path-based truncate() makes no checks based on file-largeness, unlike
> ftruncate().

Right, but truncate in general is used to make files *smaller* so I'm
having trouble thinking of a scenario where a largefile-oblivious
program could get in trouble by truncating a file > 2TB to some
hard-coded length (normally zero).

> Overlayfs and one or two other places need to potentially apply O_LARGEFILE to
> the things that they do on behalf of userspace - but other than suppressing
> some size checks, it seems to be ignored by the filesystems and the VM.

The size checks really were the primary points of O_LARGEFILE.  As I
recall the primary system calls where this really matters is open(2)
and stat(2) (since if st_size is too small to represent the size of
the file, then the user space program could get really confused).

Essentially O_LARGEFILE is an assertion that userspace can handle
handle 64-bit files, and won't get confused by system call interfaces
where off_t is 32-bit wide, because it will use the 64-bit variants.
So it's not at all surprising that the file systems and the VM in
general doesn't need to worry about the flag.

    	       	   	      	      - Ted

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[Dave Chinner]

On Tue, Sep 22, 2015 at 05:12:42PM +0100, David Howells wrote:
> Theodore Ts'o <tytso@mit.edu> wrote:
> 
> > So what this means is that on 32-bit systems, if we have a userspace
> > program which isn't using the Largefile-enabled, and it opens a file
> > which is larger than can be addressed with a 32-bit off_t, it can get
> > surprised and possibly cause data loss.
> 
> Good point.  I was initially thinking that 32-bit userspace on a 64-bit system
> would have O_LARGEFILE automatically enabled - but I guess it'll trap through
> the compat entry points which avoid that.
> 
> That said, fanotify and xfs_open_by_handle() will both automatically set
> O_LARGEFILE irrespectively of the 32-bitness of the original caller.

Any binaries that use xfs_open_by_handle() and then don't support
greater than 32bit file offsets are simply broken. No ifs or buts -
if you are using low level XFS specific file access ioctls, you need
to build binaries that support 64 bit offsets.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: [RFC PATCH 1/2] VFS: Kill use of O_LARGEFILE inside the kernel

[Dave Chinner]

On Tue, Sep 22, 2015 at 05:12:42PM +0100, David Howells wrote:
> Theodore Ts'o <tytso@mit.edu> wrote:
> 
> > So what this means is that on 32-bit systems, if we have a userspace
> > program which isn't using the Largefile-enabled, and it opens a file
> > which is larger than can be addressed with a 32-bit off_t, it can get
> > surprised and possibly cause data loss.
> 
> Good point.  I was initially thinking that 32-bit userspace on a 64-bit system
> would have O_LARGEFILE automatically enabled - but I guess it'll trap through
> the compat entry points which avoid that.
> 
> That said, fanotify and xfs_open_by_handle() will both automatically set
> O_LARGEFILE irrespectively of the 32-bitness of the original caller.

Any binaries that use xfs_open_by_handle() and then don't support
greater than 32bit file offsets are simply broken. No ifs or buts -
if you are using low level XFS specific file access ioctls, you need
to build binaries that support 64 bit offsets.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs