6.15-stable review patch. If anyone has any objections, please let me know.
------------------
From: David Howells <dhowells@xxxxxxxxxx>
[ Upstream commit 2e0658940d90a3dc130bb3b7f75bae9f4100e01f ]
Fix the updating of i_size, particularly in regard to the completion of DIO
writes and especially async DIO writes by using a lock.
The bug is triggered occasionally by the generic/207 xfstest as it chucks a
bunch of AIO DIO writes at the filesystem and then checks that fstat()
returns a reasonable st_size as each completes.
The problem is that netfs is trying to do "if new_size > inode->i_size,
update inode->i_size" sort of thing but without a lock around it.
This can be seen with cifs, but shouldn't be seen with kafs because kafs
serialises modification ops on the client whereas cifs sends the requests
to the server as they're generated and lets the server order them.
Fixes: 153a9961b551 ("netfs: Implement unbuffered/DIO write support")
Signed-off-by: David Howells <dhowells@xxxxxxxxxx>
Link: https://lore.kernel.org/20250701163852.2171681-11-dhowells@xxxxxxxxxx
Reviewed-by: Paulo Alcantara (Red Hat) <pc@xxxxxxxxxxxxx>
cc: Steve French <sfrench@xxxxxxxxx>
cc: Paulo Alcantara <pc@xxxxxxxxxxxxx>
cc: linux-cifs@xxxxxxxxxxxxxxx
cc: netfs@xxxxxxxxxxxxxxx
cc: linux-fsdevel@xxxxxxxxxxxxxxx
Signed-off-by: Christian Brauner <brauner@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
fs/netfs/buffered_write.c | 2 ++
fs/netfs/direct_write.c | 8 ++++++--
2 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/fs/netfs/buffered_write.c b/fs/netfs/buffered_write.c
index dbb544e183d13..9f22ff890a8cd 100644
--- a/fs/netfs/buffered_write.c
+++ b/fs/netfs/buffered_write.c
@@ -64,6 +64,7 @@ static void netfs_update_i_size(struct netfs_inode *ctx, struct inode *inode,
return;
}
+ spin_lock(&inode->i_lock);
i_size_write(inode, pos);
#if IS_ENABLED(CONFIG_FSCACHE)
fscache_update_cookie(ctx->cache, NULL, &pos);
@@ -77,6 +78,7 @@ static void netfs_update_i_size(struct netfs_inode *ctx, struct inode *inode,
DIV_ROUND_UP(pos, SECTOR_SIZE),
inode->i_blocks + add);
}
+ spin_unlock(&inode->i_lock);
}
/**
diff --git a/fs/netfs/direct_write.c b/fs/netfs/direct_write.c
index fa9a5bf3c6d51..3efa5894b2c07 100644
--- a/fs/netfs/direct_write.c
+++ b/fs/netfs/direct_write.c
@@ -14,13 +14,17 @@ static void netfs_cleanup_dio_write(struct netfs_io_request *wreq)
struct inode *inode = wreq->inode;
unsigned long long end = wreq->start + wreq->transferred;
- if (!wreq->error &&
- i_size_read(inode) < end) {
+ if (wreq->error || end <= i_size_read(inode))
+ return;
+
+ spin_lock(&inode->i_lock);
+ if (end > i_size_read(inode)) {
if (wreq->netfs_ops->update_i_size)
wreq->netfs_ops->update_i_size(inode, end);
else
i_size_write(inode, end);
}
+ spin_unlock(&inode->i_lock);
}
/*
--
2.39.5
