[PATCH 07/11] netfs: Fix folio->private handling in netfs_perform_write()

[David Howells]

Under some circumstances, netfs_perform_write() doesn't correctly
manipulate folio->private between NULL, NETFS_FOLIO_COPY_TO_CACHE, pointing
to a group and pointing to a netfs_folio struct, leading to potential
multiple attachments of private data with associated folio ref leaks and
also leaks of netfs_folio structs or netfs_group refs.

Fix this by consolidating the place at which a folio is marked uptodate in
one place and having that look at what's attached to folio->private and
decide how to clean it up and then set the new group.  Also, the content
shouldn't be flushed if group is NULL, even if a group is specified in the
netfs_group parameter, as that would be the case for a new folio.  A
filesystem should always specify netfs_group or never specify netfs_group.

The Sashiko auto-review tool noted that it was theoretically possible that
the fpos >= ctx->zero_point section might leak if it modified a streaming
write folio.  This is unlikely, but with a network filesystem, third party
changes can happen.  It also pointed out that __netfs_set_group() would
leak if called multiple times on the same folio from the "whole folio
modify section".

Fixes: 8f52de0077ba ("netfs: Reduce number of conditional branches in netfs_perform_write()")
Closes: https://sashiko.dev/#/patchset/20260414082004.3756080-1-dhowells%40redhat.com
Signed-off-by: David Howells <dhowells@xxxxxxxxxx>
cc: Paulo Alcantara <pc@xxxxxxxxxxxxx>
cc: Matthew Wilcox <willy@xxxxxxxxxxxxx>
cc: netfs@xxxxxxxxxxxxxxx
cc: linux-fsdevel@xxxxxxxxxxxxxxx
---
 fs/netfs/buffered_write.c    | 77 +++++++++++++++++++++---------------
 include/trace/events/netfs.h |  2 +
 2 files changed, 48 insertions(+), 31 deletions(-)

diff --git a/fs/netfs/buffered_write.c b/fs/netfs/buffered_write.c
index de82eb6ed473..0ca8e922790d 100644
--- a/fs/netfs/buffered_write.c
+++ b/fs/netfs/buffered_write.c
@@ -13,12 +13,6 @@
 #include <linux/pagevec.h>
 #include "internal.h"
 
-static void __netfs_set_group(struct folio *folio, struct netfs_group *netfs_group)
-{
-	if (netfs_group)
-		folio_attach_private(folio, netfs_get_group(netfs_group));
-}
-
 static void netfs_set_group(struct folio *folio, struct netfs_group *netfs_group)
 {
 	void *priv = folio_get_private(folio);
@@ -158,6 +152,7 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 		size_t offset;	/* Offset into pagecache folio */
 		size_t part;	/* Bytes to write to folio */
 		size_t copied;	/* Bytes copied from user */
+		void *priv;
 
 		offset = pos & (max_chunk - 1);
 		part = min(max_chunk - offset, iov_iter_count(iter));
@@ -213,8 +208,9 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 		finfo = netfs_folio_info(folio);
 		group = netfs_folio_group(folio);
 
-		if (unlikely(group != netfs_group) &&
-		    group != NETFS_FOLIO_COPY_TO_CACHE)
+		if (unlikely(group) &&
+		    group != NETFS_FOLIO_COPY_TO_CACHE &&
+		    group != netfs_group)
 			goto flush_content;
 
 		if (folio_test_uptodate(folio)) {
@@ -238,24 +234,22 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 			if (unlikely(copied == 0))
 				goto copy_failed;
 			folio_zero_segment(folio, offset + copied, flen);
-			__netfs_set_group(folio, netfs_group);
-			folio_mark_uptodate(folio);
-			trace = netfs_modify_and_clear;
-			goto copied;
+			if (finfo)
+				trace = netfs_modify_and_clear_rm_finfo;
+			else
+				trace = netfs_modify_and_clear;
+			goto mark_uptodate;
 		}
 
 		/* See if we can write a whole folio in one go. */
 		if (!maybe_trouble && offset == 0 && part >= flen) {
 			copied = copy_folio_from_iter_atomic(folio, offset, part, iter);
 			if (likely(copied == part)) {
-				if (finfo) {
+				if (finfo)
 					trace = netfs_whole_folio_modify_filled;
-					goto folio_now_filled;
-				}
-				__netfs_set_group(folio, netfs_group);
-				folio_mark_uptodate(folio);
-				trace = netfs_whole_folio_modify;
-				goto copied;
+				else
+					trace = netfs_whole_folio_modify;
+				goto mark_uptodate;
 			}
 			if (copied == 0)
 				goto copy_failed;
@@ -271,8 +265,10 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 			 * accept the partial write.
 			 */
 			finfo->dirty_len += finfo->dirty_offset;
-			if (finfo->dirty_len == flen)
-				goto folio_now_filled;
+			if (finfo->dirty_len == flen) {
+				trace = netfs_whole_folio_modify_filled_efault;
+				goto mark_uptodate;
+			}
 			if (copied > finfo->dirty_len)
 				finfo->dirty_len = copied;
 			finfo->dirty_offset = 0;
@@ -306,6 +302,7 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 			goto copied;
 		}
 
+		/* Do a streaming write on a folio that has nothing in it yet. */
 		if (!finfo) {
 			ret = -EIO;
 			if (WARN_ON(folio_get_private(folio)))
@@ -314,10 +311,8 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 			if (unlikely(copied == 0))
 				goto copy_failed;
 			if (offset == 0 && copied == flen) {
-				__netfs_set_group(folio, netfs_group);
-				folio_mark_uptodate(folio);
 				trace = netfs_streaming_filled_page;
-				goto copied;
+				goto mark_uptodate;
 			}
 
 			finfo = kzalloc_obj(*finfo);
@@ -346,7 +341,7 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 			finfo->dirty_len += copied;
 			if (finfo->dirty_offset == 0 && finfo->dirty_len == flen) {
 				trace = netfs_streaming_cont_filled_page;
-				goto folio_now_filled;
+				goto mark_uptodate;
 			}
 			trace = netfs_streaming_write_cont;
 			goto copied;
@@ -362,13 +357,33 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
 			goto out;
 		continue;
 
-	folio_now_filled:
-		if (finfo->netfs_group)
-			folio_change_private(folio, finfo->netfs_group);
-		else
-			folio_detach_private(folio);
+		/* Mark a folio as being up to data when we've filled it
+		 * completely.  If the folio has a group attached, then it must
+		 * be the same group, otherwise we should have flushed it out
+		 * above.  We have to get rid of the netfs_folio struct if
+		 * there was one.
+		 */
+	mark_uptodate:
+		priv = folio_get_private(folio);
+		if (likely(priv == netfs_group)) {
+			/* Already set correctly; no change required. */
+		} else if (priv == NETFS_FOLIO_COPY_TO_CACHE) {
+			if (!netfs_group)
+				folio_detach_private(folio);
+			else
+				folio_change_private(folio, netfs_get_group(netfs_group));
+		} else if (!priv) {
+			folio_attach_private(folio, netfs_get_group(netfs_group));
+		} else {
+			WARN_ON_ONCE(!finfo);
+			if (netfs_group)
+				/* finfo->netfs_group has a ref */
+				folio_change_private(folio, netfs_group);
+			else
+				folio_detach_private(folio);
+			kfree(finfo);
+		}
 		folio_mark_uptodate(folio);
-		kfree(finfo);
 	copied:
 		trace_netfs_folio(folio, trace);
 		flush_dcache_folio(folio);
diff --git a/include/trace/events/netfs.h b/include/trace/events/netfs.h
index 67f6d56c94ce..1f5e3a5af08a 100644
--- a/include/trace/events/netfs.h
+++ b/include/trace/events/netfs.h
@@ -179,7 +179,9 @@
 	EM(netfs_whole_folio_modify,		"mod-whole-f")	\
 	EM(netfs_whole_folio_modify_efault,	"mod-whole-f!")	\
 	EM(netfs_whole_folio_modify_filled,	"mod-whole-f+")	\
+	EM(netfs_whole_folio_modify_filled_efault, "mod-whole-f+!")	\
 	EM(netfs_modify_and_clear,		"mod-n-clear")	\
+	EM(netfs_modify_and_clear_rm_finfo,	"mod-n-clear+")	\
 	EM(netfs_streaming_write,		"mod-streamw")	\
 	EM(netfs_streaming_write_cont,		"mod-streamw+")	\
 	EM(netfs_flush_content,			"flush")	\