Skip to contentSkip to conversation

FAILED: patch "[PATCH] fuse: truncate pagecache on atomic_o_trunc" failed to apply to 4.19-stable tree

PATCH [email protected] 2 participants 3 messages Sep 13, 2021 – Nov 15, 2021

Subsystems (2)
Stable branch FUSE filesystem core
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 4.19-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 76224355db7570cbe6b6f75c8929a1558828dd55 Mon Sep 17 00:00:00 2001
From: Miklos Szeredi <[email protected]>
Date: Tue, 17 Aug 2021 21:05:16 +0200
Subject: [PATCH] fuse: truncate pagecache on atomic_o_trunc

fuse_finish_open() will be called with FUSE_NOWRITE in case of atomic O_TRUNC. This can deadlock with fuse_wait_on_page_writeback() in fuse_launder_page() triggered by invalidate_inode_pages2().

Fix by replacing invalidate_inode_pages2() in fuse_finish_open() with a truncate_pagecache() call. This makes sense regardless of FOPEN_KEEP_CACHE or fc->writeback cache, so do it unconditionally.

Reported-by Xie Yongji <[email protected]>
Reported-and-tested-by [email protected]
Fixes e4648309b85a ("fuse: truncate pending writes on O_TRUNC")
Cc <[email protected]>
Signed-off-by Miklos Szeredi <[email protected]>
fs/fuse/file.c +5 -2
index 97f860cfc195..5e5efb66c7d7 100644
@@ -198,12 +198,11 @@ void fuse_finish_open(struct inode *inode, struct file *file)
 	struct fuse_file *ff = file->private_data;
 	struct fuse_conn *fc = get_fuse_conn(inode);
 
-	if (!(ff->open_flags & FOPEN_KEEP_CACHE))
-		invalidate_inode_pages2(inode->i_mapping);
 	if (ff->open_flags & FOPEN_STREAM)
 		stream_open(inode, file);
 	else if (ff->open_flags & FOPEN_NONSEEKABLE)
 		nonseekable_open(inode, file);
+
 	if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) {
 		struct fuse_inode *fi = get_fuse_inode(inode);
 

@@ -211,10 +210,14 @@ void fuse_finish_open(struct inode *inode, struct file *file)
 		fi->attr_version = atomic64_inc_return(&fc->attr_version);
 		i_size_write(inode, 0);
 		spin_unlock(&fi->lock);
+		truncate_pagecache(inode, 0);
 		fuse_invalidate_attr(inode);
 		if (fc->writeback_cache)
 			file_update_time(file);
+	} else if (!(ff->open_flags & FOPEN_KEEP_CACHE)) {
+		invalidate_inode_pages2(inode->i_mapping);
 	}
+
 	if ((file->f_mode & FMODE_WRITE) && fc->writeback_cache)
 		fuse_link_write_file(file);
 }
#1 lore raw
Quoted 6 lines · On Mon, Sep 13, 2021 at 1:32 PM <[email protected]> wrote:

The patch below does not apply to the 4.19-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Hi Greg,

Attaching the backport of commit 76224355db75 ("fuse: truncate pagecache on atomic_o_trunc") to v4.19.217.

Thanks,
Miklos

From 56d08204f7da642c3c914fb5b5208bffe9debe3a Mon Sep 17 00:00:00 2001
From: Miklos Szeredi <[email protected]>
Date: Tue, 17 Aug 2021 21:05:16 +0200
Subject: [PATCH] fuse: truncate pagecache on atomic_o_trunc

fuse_finish_open() will be called with FUSE_NOWRITE in case of atomic O_TRUNC. This can deadlock with fuse_wait_on_page_writeback() in fuse_launder_page() triggered by invalidate_inode_pages2().

Fix by replacing invalidate_inode_pages2() in fuse_finish_open() with a truncate_pagecache() call. This makes sense regardless of FOPEN_KEEP_CACHE or fc->writeback cache, so do it unconditionally.

Reported-by Xie Yongji <[email protected]>
Reported-and-tested-by [email protected]
Fixes e4648309b85a ("fuse: truncate pending writes on O_TRUNC")
Cc <[email protected]>
Signed-off-by Miklos Szeredi <[email protected]>
(cherry picked from commit 76224355db7570cbe6b6f75c8929a1558828dd55)
---
 fs/fuse/file.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
fs/fuse/file.c +5 -2
index 6a3d89672ff7..13371a40f7a1 100644
@@ -178,12 +178,11 @@ void fuse_finish_open(struct inode *inode, struct file *file)
 
 	if (ff->open_flags & FOPEN_DIRECT_IO)
 		file->f_op = &fuse_direct_io_file_operations;
-	if (!(ff->open_flags & FOPEN_KEEP_CACHE))
-		invalidate_inode_pages2(inode->i_mapping);
 	if (ff->open_flags & FOPEN_STREAM)
 		stream_open(inode, file);
 	else if (ff->open_flags & FOPEN_NONSEEKABLE)
 		nonseekable_open(inode, file);
+
 	if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) {
 		struct fuse_inode *fi = get_fuse_inode(inode);
 

@@ -191,10 +190,14 @@ void fuse_finish_open(struct inode *inode, struct file *file)
 		fi->attr_version = ++fc->attr_version;
 		i_size_write(inode, 0);
 		spin_unlock(&fc->lock);
+		truncate_pagecache(inode, 0);
 		fuse_invalidate_attr(inode);
 		if (fc->writeback_cache)
 			file_update_time(file);
+	} else if (!(ff->open_flags & FOPEN_KEEP_CACHE)) {
+		invalidate_inode_pages2(inode->i_mapping);
 	}
+
 	if ((file->f_mode & FMODE_WRITE) && fc->writeback_cache)
 		fuse_link_write_file(file);
 }
Signature · 1 line
2.31.1
author maintainer
#2 lore raw
Quoted 8 lines · Miklos Szeredi
Quoted 6 lines · On Mon, Sep 13, 2021 at 1:32 PM <[email protected]> wrote:

The patch below does not apply to the 4.19-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Quoted 4 lines

Hi Greg,

Attaching the backport of commit 76224355db75 ("fuse: truncate pagecache on atomic_o_trunc") to v4.19.217.

Now queued up, thanks.

greg k-h