public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Al Viro <viro@ZenIV.linux.org.uk>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: "Alan J. Wylie" <alan@wylie.me.uk>,
	Thorsten Leemhuis <regressions@leemhuis.info>,
	linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: 4.9.0 regression in pipe-backed iov_iter with systemd-nspawn
Date: Fri, 13 Jan 2017 22:50:58 +0000	[thread overview]
Message-ID: <20170113225058.GW1555@ZenIV.linux.org.uk> (raw)
In-Reply-To: <20170113221308.GV1555@ZenIV.linux.org.uk>

On Fri, Jan 13, 2017 at 10:13:08PM +0000, Al Viro wrote:
> On Fri, Jan 13, 2017 at 09:59:19PM +0000, Al Viro wrote:
> > +		if (off) {
> > +			pipe->bufs[idx].len = off - pipe->bufs[idx].offset;
> > +			/* free all after idx; n can't be 0 */
> 
> Yes, it can - full pipe, truncation to the part of the first buffer ;-/
> OTOH, handling the wraparound can be done uniformly for both "empty all"
> and "leave something" cases...  Fixed variant follows:

Or, even better, we can get rid of all wraparound-related crap if we
calculate the final value of pipe->nrbufs and watch for _that_ as
loop condition:

diff --git a/lib/iov_iter.c b/lib/iov_iter.c
index 25f572303801..e68604ae3ced 100644
--- a/lib/iov_iter.c
+++ b/lib/iov_iter.c
@@ -730,43 +730,50 @@ size_t iov_iter_copy_from_user_atomic(struct page *page,
 }
 EXPORT_SYMBOL(iov_iter_copy_from_user_atomic);
 
+static inline void pipe_truncate(struct iov_iter *i)
+{
+	struct pipe_inode_info *pipe = i->pipe;
+	if (pipe->nrbufs) {
+		size_t off = i->iov_offset;
+		int idx = i->idx;
+		int nrbufs = (idx - pipe->curbuf) & (pipe->buffers - 1);
+		if (off) {
+			pipe->bufs[idx].len = off - pipe->bufs[idx].offset;
+			idx = next_idx(idx, pipe);
+			nrbufs++;
+		}
+		while (pipe->nrbufs > nrbufs) {
+			pipe_buf_release(pipe, &pipe->bufs[idx]);
+			idx = next_idx(idx, pipe);
+			pipe->nrbufs--;
+		}
+	}
+}
+
 static void pipe_advance(struct iov_iter *i, size_t size)
 {
 	struct pipe_inode_info *pipe = i->pipe;
-	struct pipe_buffer *buf;
-	int idx = i->idx;
-	size_t off = i->iov_offset, orig_sz;
-	
 	if (unlikely(i->count < size))
 		size = i->count;
-	orig_sz = size;
-
 	if (size) {
+		struct pipe_buffer *buf;
+		size_t off = i->iov_offset, left = size;
+		int idx = i->idx;
 		if (off) /* make it relative to the beginning of buffer */
-			size += off - pipe->bufs[idx].offset;
+			left += off - pipe->bufs[idx].offset;
 		while (1) {
 			buf = &pipe->bufs[idx];
-			if (size <= buf->len)
+			if (left <= buf->len)
 				break;
-			size -= buf->len;
+			left -= buf->len;
 			idx = next_idx(idx, pipe);
 		}
-		buf->len = size;
 		i->idx = idx;
-		off = i->iov_offset = buf->offset + size;
-	}
-	if (off)
-		idx = next_idx(idx, pipe);
-	if (pipe->nrbufs) {
-		int unused = (pipe->curbuf + pipe->nrbufs) & (pipe->buffers - 1);
-		/* [curbuf,unused) is in use.  Free [idx,unused) */
-		while (idx != unused) {
-			pipe_buf_release(pipe, &pipe->bufs[idx]);
-			idx = next_idx(idx, pipe);
-			pipe->nrbufs--;
-		}
+		i->iov_offset = buf->offset + left;
 	}
-	i->count -= orig_sz;
+	i->count -= size;
+	/* ... and discard everything past that point */
+	pipe_truncate(i);
 }
 
 void iov_iter_advance(struct iov_iter *i, size_t size)
@@ -826,6 +833,7 @@ void iov_iter_pipe(struct iov_iter *i, int direction,
 			size_t count)
 {
 	BUG_ON(direction != ITER_PIPE);
+	WARN_ON(pipe->nrbufs == pipe->buffers);
 	i->type = direction;
 	i->pipe = pipe;
 	i->idx = (pipe->curbuf + pipe->nrbufs) & (pipe->buffers - 1);

  reply	other threads:[~2017-01-13 22:51 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-01-12 20:26 4.9.0 regression in pipe-backed iov_iter with systemd-nspawn Alan J. Wylie
2017-01-12 20:31 ` Al Viro
2017-01-12 20:38   ` Alan J. Wylie
2017-01-12 22:26 ` Linus Torvalds
2017-01-12 22:37   ` Al Viro
2017-01-12 22:46     ` Al Viro
2017-01-12 23:02       ` Linus Torvalds
2017-01-12 23:14         ` Al Viro
2017-01-12 23:14         ` Linus Torvalds
2017-01-12 23:27           ` Al Viro
2017-01-12 22:46   ` Alan J. Wylie
2017-01-12 22:58     ` Al Viro
2017-01-12 23:28     ` Linus Torvalds
2017-01-13  4:00       ` Al Viro
2017-01-13  7:38         ` Alan J. Wylie
2017-01-13  7:23       ` Alan J. Wylie
2017-01-13  9:33         ` Al Viro
2017-01-13  9:54           ` Alan J. Wylie
2017-01-13 10:20             ` Al Viro
2017-01-13 10:32               ` Alan J. Wylie
2017-01-13 11:25                 ` Al Viro
2017-01-13 11:18               ` Al Viro
2017-01-13 19:33                 ` Linus Torvalds
2017-01-13 20:08                   ` Al Viro
2017-01-13 20:11                     ` Al Viro
2017-01-13 20:32                       ` Linus Torvalds
2017-01-13 20:47                         ` Al Viro
2017-01-13 21:55                           ` Al Viro
2017-01-13 21:59                             ` Al Viro
2017-01-13 22:13                               ` Al Viro
2017-01-13 22:50                                 ` Al Viro [this message]
2017-01-14  0:59                                   ` Linus Torvalds
2017-01-14  1:24                                     ` Al Viro
2017-01-14  1:43                                       ` Al Viro
2017-01-14  1:46                                       ` Linus Torvalds
2017-01-14  1:57                                         ` Al Viro
2017-01-15  0:53                                           ` Al Viro
2017-01-14 13:16                                   ` Alan J. Wylie
2017-01-14 16:29                                     ` Alan J. Wylie
2017-01-14 17:57                                       ` Linus Torvalds
2017-01-13 20:08                   ` Linus Torvalds
2017-01-13 20:16                     ` Al Viro

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20170113225058.GW1555@ZenIV.linux.org.uk \
    --to=viro@zeniv.linux.org.uk \
    --cc=alan@wylie.me.uk \
    --cc=linux-kernel@vger.kernel.org \
    --cc=regressions@leemhuis.info \
    --cc=torvalds@linux-foundation.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox