diff --git a/include/common/buf.h b/include/common/buf.h
index 564e0eb..fccad45 100644
--- a/include/common/buf.h
+++ b/include/common/buf.h
@@ -329,6 +329,47 @@
 		b->p = b->data;
 }
 
+/* b_slow_realign() : this function realigns a possibly wrapping buffer so that
+ * the part remaining to be parsed is contiguous and starts at the beginning of
+ * the buffer and the already parsed output part ends at the end of the buffer.
+ * This provides the best conditions since it allows the largest inputs to be
+ * processed at once and ensures that once the output data leaves, the whole
+ * buffer is available at once. The number of output bytes supposedly present
+ * at the beginning of the buffer and which need to be moved to the end must be
+ * passed in <output>. A temporary swap area at least as large as b->size must
+ * be provided in <swap>. It's up to the caller to ensure <output> is no larger
+ * than the difference between the whole buffer's length and its input.
+ */
+static inline void b_slow_realign(struct buffer *b, char *swap, size_t output)
+{
+	size_t block1 = output;
+	size_t block2 = 0;
+
+	/* process output data in two steps to cover wrapping */
+	if (block1 > b_size(b) - b_head_ofs(b)) {
+		block2 = b_size(b) - b_head_ofs(b);
+		block1 -= block2;
+	}
+	memcpy(swap + b_size(b) - output, b_head(b), block1);
+	memcpy(swap + b_size(b) - block2, b_orig(b), block2);
+
+	/* process input data in two steps to cover wrapping */
+	block1 = b_data(b) - output;
+	block2 = 0;
+
+	if (block1 > b_tail_ofs(b)) {
+		block2 = b_tail_ofs(b);
+		block1 = block1 - block2;
+	}
+	memcpy(swap, b_peek(b, output), block1);
+	memcpy(swap + block1, b_orig(b), block2);
+
+	/* reinject changes into the buffer */
+	memcpy(b_orig(b), swap, b_data(b) - output);
+	memcpy(b_wrap(b) - output, swap + b_size(b) - output, output);
+
+	b->p = b->data;
+}
 
 #endif /* _COMMON_BUF_H */
 
diff --git a/include/common/buffer.h b/include/common/buffer.h
index 0218afa..3b34267 100644
--- a/include/common/buffer.h
+++ b/include/common/buffer.h
@@ -52,7 +52,7 @@
 int buffer_replace2(struct buffer *b, char *pos, char *end, const char *str, int len);
 int buffer_insert_line2(struct buffer *b, char *pos, const char *str, int len);
 void buffer_dump(FILE *o, struct buffer *b, int from, int to);
-void buffer_slow_realign(struct buffer *buf);
+void buffer_slow_realign(struct buffer *buf, size_t output);
 
 /*****************************************************************/
 /* These functions are used to compute various buffer area sizes */
diff --git a/include/proto/channel.h b/include/proto/channel.h
index 3f30cc0..c7135de 100644
--- a/include/proto/channel.h
+++ b/include/proto/channel.h
@@ -701,6 +701,17 @@
 	chn->buf->i = 0;
 }
 
+/* This function realigns a possibly wrapping channel buffer so that the input
+ * part is contiguous and starts at the beginning of the buffer and the output
+ * part ends at the end of the buffer. This provides the best conditions since
+ * it allows the largest inputs to be processed at once and ensures that once
+ * the output data leaves, the whole buffer is available at once.
+ */
+static inline void channel_slow_realign(struct channel *chn)
+{
+	return buffer_slow_realign(chn->buf, co_data(chn));
+}
+
 /*
  * Advance the channel buffer's read pointer by <len> bytes. This is useful
  * when data have been read directly from the buffer. It is illegal to call
diff --git a/src/buffer.c b/src/buffer.c
index b6ece4b..becaa10 100644
--- a/src/buffer.c
+++ b/src/buffer.c
@@ -172,37 +172,15 @@
  * contiguous and starts at the beginning of the buffer and the output part
  * ends at the end of the buffer. This provides the best conditions since it
  * allows the largest inputs to be processed at once and ensures that once the
- * output data leaves, the whole buffer is available at once.
+ * output data leaves, the whole buffer is available at once. The number of
+ * output bytes supposedly present at the beginning of the buffer and which
+ * need to be moved to the end must be passed in <output>. It's up to the
+ * caller to ensure <output> is no larger than the difference between the
+ * while buffer's length and its input.
  */
-void buffer_slow_realign(struct buffer *buf)
+void buffer_slow_realign(struct buffer *buf, size_t output)
 {
-	int block1 = buf->o;
-	int block2 = 0;
-
-	/* process output data in two steps to cover wrapping */
-	if (block1 > buf->p - buf->data) {
-		block2 = buf->p - buf->data;
-		block1 -= block2;
-	}
-	memcpy(swap_buffer + buf->size - buf->o, bo_ptr(buf), block1);
-	memcpy(swap_buffer + buf->size - block2, buf->data, block2);
-
-	/* process input data in two steps to cover wrapping */
-	block1 = buf->i;
-	block2 = 0;
-
-	if (block1 > buf->data + buf->size - buf->p) {
-		block1 = buf->data + buf->size - buf->p;
-		block2 = buf->i - block1;
-	}
-	memcpy(swap_buffer, bi_ptr(buf), block1);
-	memcpy(swap_buffer + block1, buf->data, block2);
-
-	/* reinject changes into the buffer */
-	memcpy(buf->data, swap_buffer, buf->i);
-	memcpy(buf->data + buf->size - buf->o, swap_buffer + buf->size - buf->o, buf->o);
-
-	buf->p = buf->data;
+	return b_slow_realign(buf, swap_buffer, output);
 }
 
 /*
diff --git a/src/hlua.c b/src/hlua.c
index 4715639..cdb5ff6 100644
--- a/src/hlua.c
+++ b/src/hlua.c
@@ -3041,7 +3041,7 @@
 	 * detects a non contiguous buffer and realign it.
 	 */
 	if (bi_space_for_replace(chn->buf) < max)
-		buffer_slow_realign(chn->buf);
+		channel_slow_realign(chn);
 
 	/* Copy input data in the buffer. */
 	max = buffer_replace2(chn->buf, chn->buf->p, chn->buf->p, str + l, max);
diff --git a/src/mux_h2.c b/src/mux_h2.c
index 1eba104..3210302 100644
--- a/src/mux_h2.c
+++ b/src/mux_h2.c
@@ -2688,7 +2688,7 @@
 		/* it doesn't fit and the buffer is fragmented,
 		 * so let's defragment it and try again.
 		 */
-		buffer_slow_realign(buf);
+		buffer_slow_realign(buf, 0);
 	}
 
 	/* first check if we have some room after p+i */
@@ -2995,7 +2995,7 @@
 		if (outbuf.size >= 9 || !buffer_space_wraps(h2c->mbuf))
 			break;
 	realign_again:
-		buffer_slow_realign(h2c->mbuf);
+		buffer_slow_realign(h2c->mbuf, h2c->mbuf->o);
 	}
 
 	if (outbuf.size < 9) {
@@ -3153,7 +3153,7 @@
 		if (outbuf.size >= 9 || !buffer_space_wraps(h2c->mbuf))
 			break;
 	realign_again:
-		buffer_slow_realign(h2c->mbuf);
+		buffer_slow_realign(h2c->mbuf, h2c->mbuf->o);
 	}
 
 	if (outbuf.size < 9) {
diff --git a/src/proto_http.c b/src/proto_http.c
index c447929..a4db33a 100644
--- a/src/proto_http.c
+++ b/src/proto_http.c
@@ -1652,7 +1652,7 @@
 			}
 			if (unlikely(bi_end(req->buf) < b_ptr(req->buf, msg->next) ||
 			             bi_end(req->buf) > req->buf->data + req->buf->size - global.tune.maxrewrite))
-				buffer_slow_realign(req->buf);
+				channel_slow_realign(req);
 		}
 
 		if (likely(msg->next < req->buf->i)) /* some unparsed data are available */
@@ -5139,7 +5139,7 @@
 
 		if (unlikely(bi_end(rep->buf) < b_ptr(rep->buf, msg->next) ||
 		             bi_end(rep->buf) > rep->buf->data + rep->buf->size - global.tune.maxrewrite))
-			buffer_slow_realign(rep->buf);
+			channel_slow_realign(rep);
 
 		if (likely(msg->next < rep->buf->i))
 			http_msg_analyzer(msg, &txn->hdr_idx);
@@ -9517,7 +9517,7 @@
 		 */
 		if (s->req.buf->p > s->req.buf->data &&
 		    s->req.buf->i + s->req.buf->p > s->req.buf->data + s->req.buf->size - global.tune.maxrewrite)
-			buffer_slow_realign(s->req.buf);
+			channel_slow_realign(&s->req);
 
 		if (unlikely(txn->req.msg_state < HTTP_MSG_BODY)) {
 			if (msg->msg_state == HTTP_MSG_ERROR)