Merge tag 'notifications-pipe-prep-20191115' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs

Pull pipe rework from David Howells:
 "This is my set of preparatory patches for building a general
  notification queue on top of pipes. It makes a number of significant
  changes:

   - It removes the nr_exclusive argument from __wake_up_sync_key() as
     this is always 1. This prepares for the next step:

   - Adds wake_up_interruptible_sync_poll_locked() so that poll can be
     woken up from a function that's holding the poll waitqueue
     spinlock.

   - Change the pipe buffer ring to be managed in terms of unbounded
     head and tail indices rather than bounded index and length. This
     means that reading the pipe only needs to modify one index, not
     two.

   - A selection of helper functions are provided to query the state of
     the pipe buffer, plus a couple to apply updates to the pipe
     indices.

   - The pipe ring is allowed to have kernel-reserved slots. This allows
     many notification messages to be spliced in by the kernel without
     allowing userspace to pin too many pages if it writes to the same
     pipe.

   - Advance the head and tail indices inside the pipe waitqueue lock
     and use wake_up_interruptible_sync_poll_locked() to poke poll
     without having to take the lock twice.

   - Rearrange pipe_write() to preallocate the buffer it is going to
     write into and then drop the spinlock. This allows kernel
     notifications to then be added the ring whilst it is filling the
     buffer it allocated. The read side is stalled because the pipe
     mutex is still held.

   - Don't wake up readers on a pipe if there was already data in it
     when we added more.

   - Don't wake up writers on a pipe if the ring wasn't full before we
     removed a buffer"

* tag 'notifications-pipe-prep-20191115' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
  pipe: Remove sync on wake_ups
  pipe: Increase the writer-wakeup threshold to reduce context-switch count
  pipe: Check for ring full inside of the spinlock in pipe_write()
  pipe: Remove redundant wakeup from pipe_write()
  pipe: Rearrange sequence in pipe_write() to preallocate slot
  pipe: Conditionalise wakeup in pipe_read()
  pipe: Advance tail pointer inside of wait spinlock in pipe_read()
  pipe: Allow pipes to have kernel-reserved slots
  pipe: Use head and tail pointers for the ring, not cursor and length
  Add wake_up_interruptible_sync_poll_locked()
  Remove the nr_exclusive argument from __wake_up_sync_key()
  pipe: Reduce #inclusion of pipe_fs_i.h

1 files changed, 139 insertions, 93 deletions

diff --git a/fs/pipe.c b/fs/pipe.c
index a9149199e0e7..648ce440ca85 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -43,10 +43,12 @@ unsigned long pipe_user_pages_hard;
 unsigned long pipe_user_pages_soft = PIPE_DEF_BUFFERS * INR_OPEN_CUR;
 
 /*
- * We use a start+len construction, which provides full use of the 
- * allocated memory.
- * -- Florian Coosmann (FGC)
- * 
+ * We use head and tail indices that aren't masked off, except at the point of
+ * dereference, but rather they're allowed to wrap naturally.  This means there
+ * isn't a dead spot in the buffer, but the ring has to be a power of two and
+ * <= 2^31.
+ * -- David Howells 2019-09-23.
+ *
  * Reads with count = 0 should always return 0.
  * -- Julian Bradfield 1999-06-07.
  *
@@ -285,10 +287,12 @@ pipe_read(struct kiocb *iocb, struct iov_iter *to)
 	ret = 0;
 	__pipe_lock(pipe);
 	for (;;) {
-		int bufs = pipe->nrbufs;
-		if (bufs) {
-			int curbuf = pipe->curbuf;
-			struct pipe_buffer *buf = pipe->bufs + curbuf;
+		unsigned int head = pipe->head;
+		unsigned int tail = pipe->tail;
+		unsigned int mask = pipe->ring_size - 1;
+
+		if (!pipe_empty(head, tail)) {
+			struct pipe_buffer *buf = &pipe->bufs[tail & mask];
 			size_t chars = buf->len;
 			size_t written;
 			int error;
@@ -320,18 +324,27 @@ pipe_read(struct kiocb *iocb, struct iov_iter *to)
 			}
 
 			if (!buf->len) {
+				bool wake;
 				pipe_buf_release(pipe, buf);
-				curbuf = (curbuf + 1) & (pipe->buffers - 1);
-				pipe->curbuf = curbuf;
-				pipe->nrbufs = --bufs;
+				spin_lock_irq(&pipe->wait.lock);
+				tail++;
+				pipe->tail = tail;
 				do_wakeup = 1;
+				wake = head - (tail - 1) == pipe->max_usage / 2;
+				if (wake)
+					wake_up_locked_poll(
+						&pipe->wait, EPOLLOUT | EPOLLWRNORM);
+				spin_unlock_irq(&pipe->wait.lock);
+				if (wake)
+					kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
 			}
 			total_len -= chars;
 			if (!total_len)
 				break;	/* common path: read succeeded */
+			if (!pipe_empty(head, tail))	/* More to do? */
+				continue;
 		}
-		if (bufs)	/* More to do? */
-			continue;
+
 		if (!pipe->writers)
 			break;
 		if (!pipe->waiting_writers) {
@@ -352,17 +365,13 @@ pipe_read(struct kiocb *iocb, struct iov_iter *to)
 				ret = -ERESTARTSYS;
 			break;
 		}
-		if (do_wakeup) {
-			wake_up_interruptible_sync_poll(&pipe->wait, EPOLLOUT | EPOLLWRNORM);
- 			kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
-		}
 		pipe_wait(pipe);
 	}
 	__pipe_unlock(pipe);
 
 	/* Signal writers asynchronously that there is more room. */
 	if (do_wakeup) {
-		wake_up_interruptible_sync_poll(&pipe->wait, EPOLLOUT | EPOLLWRNORM);
+		wake_up_interruptible_poll(&pipe->wait, EPOLLOUT | EPOLLWRNORM);
 		kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
 	}
 	if (ret > 0)
@@ -380,6 +389,7 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 {
 	struct file *filp = iocb->ki_filp;
 	struct pipe_inode_info *pipe = filp->private_data;
+	unsigned int head, max_usage, mask;
 	ssize_t ret = 0;
 	int do_wakeup = 0;
 	size_t total_len = iov_iter_count(from);
@@ -397,12 +407,14 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 		goto out;
 	}
 
+	head = pipe->head;
+	max_usage = pipe->max_usage;
+	mask = pipe->ring_size - 1;
+
 	/* We try to merge small writes */
 	chars = total_len & (PAGE_SIZE-1); /* size of the last buffer */
-	if (pipe->nrbufs && chars != 0) {
-		int lastbuf = (pipe->curbuf + pipe->nrbufs - 1) &
-							(pipe->buffers - 1);
-		struct pipe_buffer *buf = pipe->bufs + lastbuf;
+	if (!pipe_empty(head, pipe->tail) && chars != 0) {
+		struct pipe_buffer *buf = &pipe->bufs[(head - 1) & mask];
 		int offset = buf->offset + buf->len;
 
 		if (pipe_buf_can_merge(buf) && offset + chars <= PAGE_SIZE) {
@@ -423,18 +435,16 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 	}
 
 	for (;;) {
-		int bufs;
-
 		if (!pipe->readers) {
 			send_sig(SIGPIPE, current, 0);
 			if (!ret)
 				ret = -EPIPE;
 			break;
 		}
-		bufs = pipe->nrbufs;
-		if (bufs < pipe->buffers) {
-			int newbuf = (pipe->curbuf + bufs) & (pipe->buffers-1);
-			struct pipe_buffer *buf = pipe->bufs + newbuf;
+
+		head = pipe->head;
+		if (!pipe_full(head, pipe->tail, max_usage)) {
+			struct pipe_buffer *buf = &pipe->bufs[head & mask];
 			struct page *page = pipe->tmp_page;
 			int copied;
 
@@ -446,38 +456,64 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 				}
 				pipe->tmp_page = page;
 			}
+
+			/* Allocate a slot in the ring in advance and attach an
+			 * empty buffer.  If we fault or otherwise fail to use
+			 * it, either the reader will consume it or it'll still
+			 * be there for the next write.
+			 */
+			spin_lock_irq(&pipe->wait.lock);
+
+			head = pipe->head;
+			if (pipe_full(head, pipe->tail, max_usage)) {
+				spin_unlock_irq(&pipe->wait.lock);
+				continue;
+			}
+
+			pipe->head = head + 1;
+
 			/* Always wake up, even if the copy fails. Otherwise
 			 * we lock up (O_NONBLOCK-)readers that sleep due to
 			 * syscall merging.
 			 * FIXME! Is this really true?
 			 */
-			do_wakeup = 1;
-			copied = copy_page_from_iter(page, 0, PAGE_SIZE, from);
-			if (unlikely(copied < PAGE_SIZE && iov_iter_count(from))) {
-				if (!ret)
-					ret = -EFAULT;
-				break;
-			}
-			ret += copied;
+			wake_up_locked_poll(
+				&pipe->wait, EPOLLIN | EPOLLRDNORM);
+
+			spin_unlock_irq(&pipe->wait.lock);
+			kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
 
 			/* Insert it into the buffer array */
+			buf = &pipe->bufs[head & mask];
 			buf->page = page;
 			buf->ops = &anon_pipe_buf_ops;
 			buf->offset = 0;
-			buf->len = copied;
+			buf->len = 0;
 			buf->flags = 0;
 			if (is_packetized(filp)) {
 				buf->ops = &packet_pipe_buf_ops;
 				buf->flags = PIPE_BUF_FLAG_PACKET;
 			}
-			pipe->nrbufs = ++bufs;
 			pipe->tmp_page = NULL;
 
+			copied = copy_page_from_iter(page, 0, PAGE_SIZE, from);
+			if (unlikely(copied < PAGE_SIZE && iov_iter_count(from))) {
+				if (!ret)
+					ret = -EFAULT;
+				break;
+			}
+			ret += copied;
+			buf->offset = 0;
+			buf->len = copied;
+
 			if (!iov_iter_count(from))
 				break;
 		}
-		if (bufs < pipe->buffers)
+
+		if (!pipe_full(head, pipe->tail, max_usage))
 			continue;
+
+		/* Wait for buffer space to become available. */
 		if (filp->f_flags & O_NONBLOCK) {
 			if (!ret)
 				ret = -EAGAIN;
@@ -488,11 +524,6 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 				ret = -ERESTARTSYS;
 			break;
 		}
-		if (do_wakeup) {
-			wake_up_interruptible_sync_poll(&pipe->wait, EPOLLIN | EPOLLRDNORM);
-			kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
-			do_wakeup = 0;
-		}
 		pipe->waiting_writers++;
 		pipe_wait(pipe);
 		pipe->waiting_writers--;
@@ -500,7 +531,7 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 out:
 	__pipe_unlock(pipe);
 	if (do_wakeup) {
-		wake_up_interruptible_sync_poll(&pipe->wait, EPOLLIN | EPOLLRDNORM);
+		wake_up_interruptible_poll(&pipe->wait, EPOLLIN | EPOLLRDNORM);
 		kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
 	}
 	if (ret > 0 && sb_start_write_trylock(file_inode(filp)->i_sb)) {
@@ -515,17 +546,19 @@ out:
 static long pipe_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 {
 	struct pipe_inode_info *pipe = filp->private_data;
-	int count, buf, nrbufs;
+	int count, head, tail, mask;
 
 	switch (cmd) {
 		case FIONREAD:
 			__pipe_lock(pipe);
 			count = 0;
-			buf = pipe->curbuf;
-			nrbufs = pipe->nrbufs;
-			while (--nrbufs >= 0) {
-				count += pipe->bufs[buf].len;
-				buf = (buf+1) & (pipe->buffers - 1);
+			head = pipe->head;
+			tail = pipe->tail;
+			mask = pipe->ring_size - 1;
+
+			while (tail != head) {
+				count += pipe->bufs[tail & mask].len;
+				tail++;
 			}
 			__pipe_unlock(pipe);
 
@@ -541,21 +574,25 @@ pipe_poll(struct file *filp, poll_table *wait)
 {
 	__poll_t mask;
 	struct pipe_inode_info *pipe = filp->private_data;
-	int nrbufs;
+	unsigned int head = READ_ONCE(pipe->head);
+	unsigned int tail = READ_ONCE(pipe->tail);
 
 	poll_wait(filp, &pipe->wait, wait);
 
+	BUG_ON(pipe_occupancy(head, tail) > pipe->ring_size);
+
 	/* Reading only -- no need for acquiring the semaphore.  */
-	nrbufs = pipe->nrbufs;
 	mask = 0;
 	if (filp->f_mode & FMODE_READ) {
-		mask = (nrbufs > 0) ? EPOLLIN | EPOLLRDNORM : 0;
+		if (!pipe_empty(head, tail))
+			mask |= EPOLLIN | EPOLLRDNORM;
 		if (!pipe->writers && filp->f_version != pipe->w_counter)
 			mask |= EPOLLHUP;
 	}
 
 	if (filp->f_mode & FMODE_WRITE) {
-		mask |= (nrbufs < pipe->buffers) ? EPOLLOUT | EPOLLWRNORM : 0;
+		if (!pipe_full(head, tail, pipe->max_usage))
+			mask |= EPOLLOUT | EPOLLWRNORM;
 		/*
 		 * Most Unices do not set EPOLLERR for FIFOs but on Linux they
 		 * behave exactly like pipes for poll().
@@ -679,7 +716,8 @@ struct pipe_inode_info *alloc_pipe_info(void)
 	if (pipe->bufs) {
 		init_waitqueue_head(&pipe->wait);
 		pipe->r_counter = pipe->w_counter = 1;
-		pipe->buffers = pipe_bufs;
+		pipe->max_usage = pipe_bufs;
+		pipe->ring_size = pipe_bufs;
 		pipe->user = user;
 		mutex_init(&pipe->mutex);
 		return pipe;
@@ -697,9 +735,9 @@ void free_pipe_info(struct pipe_inode_info *pipe)
 {
 	int i;
 
-	(void) account_pipe_buffers(pipe->user, pipe->buffers, 0);
+	(void) account_pipe_buffers(pipe->user, pipe->ring_size, 0);
 	free_uid(pipe->user);
-	for (i = 0; i < pipe->buffers; i++) {
+	for (i = 0; i < pipe->ring_size; i++) {
 		struct pipe_buffer *buf = pipe->bufs + i;
 		if (buf->ops)
 			pipe_buf_release(pipe, buf);
@@ -882,7 +920,7 @@ SYSCALL_DEFINE1(pipe, int __user *, fildes)
 
 static int wait_for_partner(struct pipe_inode_info *pipe, unsigned int *cnt)
 {
-	int cur = *cnt;	
+	int cur = *cnt;
 
 	while (cur == *cnt) {
 		pipe_wait(pipe);
@@ -957,7 +995,7 @@ static int fifo_open(struct inode *inode, struct file *filp)
 			}
 		}
 		break;
-	
+
 	case FMODE_WRITE:
 	/*
 	 *  O_WRONLY
@@ -977,7 +1015,7 @@ static int fifo_open(struct inode *inode, struct file *filp)
 				goto err_wr;
 		}
 		break;
-	
+
 	case FMODE_READ | FMODE_WRITE:
 	/*
 	 *  O_RDWR
@@ -1056,14 +1094,14 @@ unsigned int round_pipe_size(unsigned long size)
 static long pipe_set_size(struct pipe_inode_info *pipe, unsigned long arg)
 {
 	struct pipe_buffer *bufs;
-	unsigned int size, nr_pages;
+	unsigned int size, nr_slots, head, tail, mask, n;
 	unsigned long user_bufs;
 	long ret = 0;
 
 	size = round_pipe_size(arg);
-	nr_pages = size >> PAGE_SHIFT;
+	nr_slots = size >> PAGE_SHIFT;
 
-	if (!nr_pages)
+	if (!nr_slots)
 		return -EINVAL;
 
 	/*
@@ -1073,13 +1111,13 @@ static long pipe_set_size(struct pipe_inode_info *pipe, unsigned long arg)
 	 * Decreasing the pipe capacity is always permitted, even
 	 * if the user is currently over a limit.
 	 */
-	if (nr_pages > pipe->buffers &&
+	if (nr_slots > pipe->ring_size &&
 			size > pipe_max_size && !capable(CAP_SYS_RESOURCE))
 		return -EPERM;
 
-	user_bufs = account_pipe_buffers(pipe->user, pipe->buffers, nr_pages);
+	user_bufs = account_pipe_buffers(pipe->user, pipe->ring_size, nr_slots);
 
-	if (nr_pages > pipe->buffers &&
+	if (nr_slots > pipe->ring_size &&
 			(too_many_pipe_buffers_hard(user_bufs) ||
 			 too_many_pipe_buffers_soft(user_bufs)) &&
 			is_unprivileged_user()) {
@@ -1088,17 +1126,21 @@ static long pipe_set_size(struct pipe_inode_info *pipe, unsigned long arg)
 	}
 
 	/*
-	 * We can shrink the pipe, if arg >= pipe->nrbufs. Since we don't
-	 * expect a lot of shrink+grow operations, just free and allocate
-	 * again like we would do for growing. If the pipe currently
+	 * We can shrink the pipe, if arg is greater than the ring occupancy.
+	 * Since we don't expect a lot of shrink+grow operations, just free and
+	 * allocate again like we would do for growing.  If the pipe currently
 	 * contains more buffers than arg, then return busy.
 	 */
-	if (nr_pages < pipe->nrbufs) {
+	mask = pipe->ring_size - 1;
+	head = pipe->head;
+	tail = pipe->tail;
+	n = pipe_occupancy(pipe->head, pipe->tail);
+	if (nr_slots < n) {
 		ret = -EBUSY;
 		goto out_revert_acct;
 	}
 
-	bufs = kcalloc(nr_pages, sizeof(*bufs),
+	bufs = kcalloc(nr_slots, sizeof(*bufs),
 		       GFP_KERNEL_ACCOUNT | __GFP_NOWARN);
 	if (unlikely(!bufs)) {
 		ret = -ENOMEM;
@@ -1107,33 +1149,37 @@ static long pipe_set_size(struct pipe_inode_info *pipe, unsigned long arg)
 
 	/*
 	 * The pipe array wraps around, so just start the new one at zero
-	 * and adjust the indexes.
+	 * and adjust the indices.
 	 */
-	if (pipe->nrbufs) {
-		unsigned int tail;
-		unsigned int head;
-
-		tail = pipe->curbuf + pipe->nrbufs;
-		if (tail < pipe->buffers)
-			tail = 0;
-		else
-			tail &= (pipe->buffers - 1);
-
-		head = pipe->nrbufs - tail;
-		if (head)
-			memcpy(bufs, pipe->bufs + pipe->curbuf, head * sizeof(struct pipe_buffer));
-		if (tail)
-			memcpy(bufs + head, pipe->bufs, tail * sizeof(struct pipe_buffer));
+	if (n > 0) {
+		unsigned int h = head & mask;
+		unsigned int t = tail & mask;
+		if (h > t) {
+			memcpy(bufs, pipe->bufs + t,
+			       n * sizeof(struct pipe_buffer));
+		} else {
+			unsigned int tsize = pipe->ring_size - t;
+			if (h > 0)
+				memcpy(bufs + tsize, pipe->bufs,
+				       h * sizeof(struct pipe_buffer));
+			memcpy(bufs, pipe->bufs + t,
+			       tsize * sizeof(struct pipe_buffer));
+		}
 	}
 
-	pipe->curbuf = 0;
+	head = n;
+	tail = 0;
+
 	kfree(pipe->bufs);
 	pipe->bufs = bufs;
-	pipe->buffers = nr_pages;
-	return nr_pages * PAGE_SIZE;
+	pipe->ring_size = nr_slots;
+	pipe->max_usage = nr_slots;
+	pipe->tail = tail;
+	pipe->head = head;
+	return pipe->max_usage * PAGE_SIZE;
 
 out_revert_acct:
-	(void) account_pipe_buffers(pipe->user, nr_pages, pipe->buffers);
+	(void) account_pipe_buffers(pipe->user, nr_slots, pipe->ring_size);
 	return ret;
 }
 
@@ -1163,7 +1209,7 @@ long pipe_fcntl(struct file *file, unsigned int cmd, unsigned long arg)
 		ret = pipe_set_size(pipe, arg);
 		break;
 	case F_GETPIPE_SZ:
-		ret = pipe->buffers * PAGE_SIZE;
+		ret = pipe->max_usage * PAGE_SIZE;
 		break;
 	default:
 		ret = -EINVAL;