Saner handling of "lseek should fail with ESPIPE" - gets rid of

magical no_llseek thing and makes checks consistent.  In particular,
 ad-hoc "can we do splice via internal pipe" checks got saner (and
 somewhat more permissive, which is what Jason had been after, AFAICT)
 
 Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
 -----BEGIN PGP SIGNATURE-----
 
 iHUEABYIAB0WIQQqUNBr3gm4hGXdBJlZ7Krx/gZQ6wUCYug2xgAKCRBZ7Krx/gZQ
 6wxWAQDqeg+xMq2FGPXmgjCa+Cp3PXH96Lp6f3hHzakIDx+t8gEAxvuiXAD22Mct
 6S1SKuGj0iDIuM4L7hUiWTiY/bDXSAc=
 =3EC/
 -----END PGP SIGNATURE-----

Merge tag 'pull-work.lseek' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull vfs lseek updates from Al Viro:
 "Jason's lseek series.

  Saner handling of 'lseek should fail with ESPIPE' - this gets rid of
  the magical no_llseek thing and makes checks consistent.

  In particular, the ad-hoc "can we do splice via internal pipe" checks
  got saner (and somewhat more permissive, which is what Jason had been
  after, AFAICT)"

* tag 'pull-work.lseek' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  fs: remove no_llseek
  fs: check FMODE_LSEEK to control internal pipe splicing
  vfio: do not set FMODE_LSEEK flag
  dma-buf: remove useless FMODE_LSEEK flag
  fs: do not compare against ->llseek
  fs: clear or set FMODE_LSEEK based on llseek function

Documentation/filesystems/porting.rst

@@ -914,3 +914,11 @@ Calling conventions for file_open_root() changed; now it takes struct path *
 instead of passing mount and dentry separately.  For callers that used to
 pass <mnt, mnt->mnt_root> pair (i.e. the root of given mount), a new helper
 is provided - file_open_root_mnt().  In-tree users adjusted.
+
+---
+
+**mandatory**
+
+no_llseek is gone; don't set .llseek to that - just leave it NULL instead.
+Checks for "does that file have llseek(2), or should it fail with ESPIPE"
+should be done by looking at FMODE_LSEEK in file->f_mode.

drivers/dma-buf/dma-buf.c

@@ -549,7 +549,6 @@ struct dma_buf *dma_buf_export(const struct dma_buf_export_info *exp_info)
 		goto err_dmabuf;
 	}
 
-	file->f_mode |= FMODE_LSEEK;
 	dmabuf->file = file;
 
 	mutex_init(&dmabuf->lock);

drivers/gpu/drm/drm_file.c

@@ -552,8 +552,7 @@ EXPORT_SYMBOL(drm_release_noglobal);
  * Since events are used by the KMS API for vblank and page flip completion this
  * means all modern display drivers must use it.
  *
- * @offset is ignored, DRM events are read like a pipe. Therefore drivers also
+ * @offset is ignored, DRM events are read like a pipe. Polling support is
- * must set the &file_operation.llseek to no_llseek(). Polling support is
  * provided by drm_poll().
  *
  * This function will only ever read a full event. Therefore userspace must

drivers/vfio/vfio.c

@@ -1132,7 +1132,7 @@ static struct file *vfio_device_open(struct vfio_device *device)
 	 * Appears to be missing by lack of need rather than
 	 * explicitly prevented.  Now there's need.
 	 */
-	filep->f_mode |= (FMODE_LSEEK | FMODE_PREAD | FMODE_PWRITE);
+	filep->f_mode |= (FMODE_PREAD | FMODE_PWRITE);
 
 	if (device->group->type == VFIO_NO_IOMMU)
 		dev_warn(device->dev, "vfio-noiommu device opened by user "

fs/coredump.c

@@ -816,9 +816,9 @@ static int __dump_skip(struct coredump_params *cprm, size_t nr)
 {
 	static char zeroes[PAGE_SIZE];
 	struct file *file = cprm->file;
-	if (file->f_op->llseek && file->f_op->llseek != no_llseek) {
+	if (file->f_mode & FMODE_LSEEK) {
 		if (dump_interrupted() ||
-		    file->f_op->llseek(file, nr, SEEK_CUR) < 0)
+		    vfs_llseek(file, nr, SEEK_CUR) < 0)
 			return 0;
 		cprm->pos += nr;
 		return 1;

fs/file_table.c

@@ -235,6 +235,8 @@ static struct file *alloc_file(const struct path *path, int flags,
 	file->f_mapping = path->dentry->d_inode->i_mapping;
 	file->f_wb_err = filemap_sample_wb_err(file->f_mapping);
 	file->f_sb_err = file_sample_sb_err(file);
+	if (fop->llseek)
+		file->f_mode |= FMODE_LSEEK;
 	if ((file->f_mode & FMODE_READ) &&
 	     likely(fop->read || fop->read_iter))
 		file->f_mode |= FMODE_CAN_READ;

fs/open.c

@@ -888,6 +888,8 @@ static int do_dentry_open(struct file *f,
 	if ((f->f_mode & FMODE_WRITE) &&
 	     likely(f->f_op->write || f->f_op->write_iter))
 		f->f_mode |= FMODE_CAN_WRITE;
+	if ((f->f_mode & FMODE_LSEEK) && !f->f_op->llseek)
+		f->f_mode &= ~FMODE_LSEEK;
 	if (f->f_mapping->a_ops && f->f_mapping->a_ops->direct_IO)
 		f->f_mode |= FMODE_CAN_ODIRECT;
 

fs/overlayfs/copy_up.c

@@ -226,8 +226,7 @@ static int ovl_copy_up_data(struct ovl_fs *ofs, struct path *old,
 	/* Couldn't clone, so now we try to copy the data */
 
 	/* Check if lower fs supports seek operation */
-	if (old_file->f_mode & FMODE_LSEEK &&
+	if (old_file->f_mode & FMODE_LSEEK)
-	    old_file->f_op->llseek)
 		skip_hole = true;
 
 	while (len) {

fs/read_write.c

@@ -227,12 +227,6 @@ loff_t noop_llseek(struct file *file, loff_t offset, int whence)
 }
 EXPORT_SYMBOL(noop_llseek);
 
-loff_t no_llseek(struct file *file, loff_t offset, int whence)
-{
-	return -ESPIPE;
-}
-EXPORT_SYMBOL(no_llseek);
-
 loff_t default_llseek(struct file *file, loff_t offset, int whence)
 {
 	struct inode *inode = file_inode(file);
@@ -290,14 +284,9 @@ EXPORT_SYMBOL(default_llseek);
 
 loff_t vfs_llseek(struct file *file, loff_t offset, int whence)
 {
-	loff_t (*fn)(struct file *, loff_t, int);
+	if (!(file->f_mode & FMODE_LSEEK))
-
+		return -ESPIPE;
-	fn = no_llseek;
+	return file->f_op->llseek(file, offset, whence);
-	if (file->f_mode & FMODE_LSEEK) {
-		if (file->f_op->llseek)
-			fn = file->f_op->llseek;
-	}
-	return fn(file, offset, whence);
 }
 EXPORT_SYMBOL(vfs_llseek);
 

fs/splice.c

@@ -814,17 +814,15 @@ ssize_t splice_direct_to_actor(struct file *in, struct splice_desc *sd,
 {
 	struct pipe_inode_info *pipe;
 	long ret, bytes;
-	umode_t i_mode;
 	size_t len;
 	int i, flags, more;
 
 	/*
-	 * We require the input being a regular file, as we don't want to
+	 * We require the input to be seekable, as we don't want to randomly
-	 * randomly drop data for eg socket -> socket splicing. Use the
+	 * drop data for eg socket -> socket splicing. Use the piped splicing
-	 * piped splicing for that!
+	 * for that!
 	 */
-	i_mode = file_inode(in)->i_mode;
+	if (unlikely(!(in->f_mode & FMODE_LSEEK)))
-	if (unlikely(!S_ISREG(i_mode) && !S_ISBLK(i_mode)))
 		return -EINVAL;
 
 	/*

include/linux/fs.h

@@ -3152,7 +3152,7 @@ extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
 extern void
 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
 extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
-extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
+#define no_llseek NULL
 extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
 extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,

kernel/bpf/bpf_iter.c

@@ -81,10 +81,9 @@ static bool bpf_iter_support_resched(struct seq_file *seq)
 #define MAX_ITER_OBJECTS	1000000
 
 /* bpf_seq_read, a customized and simpler version for bpf iterator.
- * no_llseek is assumed for this file.
  * The following are differences from seq_read():
  *  . fixed buffer size (PAGE_SIZE)
- *  . assuming no_llseek
+ *  . assuming NULL ->llseek()
  *  . stop() may call bpf program, handling potential overflow there
  */
 static ssize_t bpf_seq_read(struct file *file, char __user *buf, size_t size,