On Fri, Aug 13, 2021 at 05:31:55PM +0800, Xie Yongji wrote:
> The invalidate_inode_pages2() might be called with FUSE_NOWRITE
> set in fuse_finish_open(), which can lead to deadlock in
> fuse_launder_page().
>
> To fix it, this tries to delay calling invalidate_inode_pages2()
> until FUSE_NOWRITE is removed.
Thanks for the report and the patch. I think it doesn't make sense to delay the
invalidate_inode_pages2() call since the inode has been truncated in this case,
there's no data worth writing out.
This patch replaces the invalidate_inode_pages2() with a truncate_pagecache()
call. This makes sense regardless of FOPEN_KEEP_CACHE or fc->writeback cache,
so do it unconditionally.
Can you please check out the following patch?
Thanks,
Miklos
---
fs/fuse/file.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -198,12 +198,11 @@ void fuse_finish_open(struct inode *inod
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 *inod
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);
}
