On 10/03/2013 01:57 PM, Miklos Szeredi wrote:
On Wed, Oct 02, 2013 at 09:38:32PM +0400, Maxim Patlasov wrote:
If writeback happens while fuse is in FUSE_NOWRITE condition, the request
will be queued but not processed immediately (see fuse_flush_writepages()).
Until FUSE_NOWRITE becomes relaxed, more writebacks can happen. They will
be queued as "secondary" requests to that first ("primary") request.
Existing implementation crops only primary request. This is not correct
because a subsequent extending write(2) may increase i_size and then secondary
requests won't be cropped properly. The result would be stale data written to
the server to a file offset where zeros must be.
Similar problem may happen if secondary requests are attached to an in-flight
request that was already cropped.
The patch solves the issue by cropping all secondary requests in
fuse_writepage_end(). Thanks to Miklos for idea.
How about this, even simpler, one?
Very cute, but unfortunately it has a flaw. See please inline comment below.
Thanks,
Miklos
Index: linux/fs/fuse/file.c
===================================================================
--- linux.orig/fs/fuse/file.c 2013-10-03 11:27:00.597084704 +0200
+++ linux/fs/fuse/file.c 2013-10-03 11:53:30.477208467 +0200
@@ -1436,12 +1436,12 @@ static void fuse_writepage_finish(struct
}
/* Called under fc->lock, may release and reacquire it */
-static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
+static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req,
+ loff_t size)
__releases(fc->lock)
__acquires(fc->lock)
{
struct fuse_inode *fi = get_fuse_inode(req->inode);
- loff_t size = i_size_read(req->inode);
struct fuse_write_in *inarg = &req->misc.write.in;
__u64 data_size = req->num_pages * PAGE_CACHE_SIZE;
@@ -1476,7 +1476,7 @@ __acquires(fc->lock)
*
* Called with fc->lock
*/
-void fuse_flush_writepages(struct inode *inode)
+void __fuse_flush_writepages(struct inode *inode, loff_t crop)
__releases(fc->lock)
__acquires(fc->lock)
{
@@ -1487,9 +1487,15 @@ __acquires(fc->lock)
while (fi->writectr >= 0 && !list_empty(&fi->queued_writes)) {
req = list_entry(fi->queued_writes.next, struct fuse_req, list);
list_del_init(&req->list);
- fuse_send_writepage(fc, req);
+ fuse_send_writepage(fc, req, crop);
}
}
+void fuse_flush_writepages(struct inode *inode)
+__releases(fc->lock)
+__acquires(fc->lock)
+{
+ __fuse_flush_writepages(inode, i_size_read(inode));
+}
static void fuse_writepage_end(struct fuse_conn *fc, struct fuse_req *req)
{
@@ -1499,12 +1505,13 @@ static void fuse_writepage_end(struct fu
mapping_set_error(inode->i_mapping, req->out.h.error);
spin_lock(&fc->lock);
while (req->misc.write.next) {
+ struct fuse_write_in *inarg = &req->misc.write.in;
struct fuse_req *next = req->misc.write.next;
req->misc.write.next = next->misc.write.next;
next->misc.write.next = NULL;
list_add(&next->writepages_entry, &fi->writepages);
list_add_tail(&next->list, &fi->queued_writes);
- fuse_flush_writepages(inode);
+ __fuse_flush_writepages(inode, inarg->offset + inarg->size);
__fuse_flush_writepages() will ignore its 'crop' arg if fi->writectr is
below zero. This can easily happen if a request is finalized after
fuse_set_nowrite(). So in a scenario like this:
1. There is an in-flight primary request with a chain of secondary ones.
2. User calls ftruncate(2) to extend file; fuse_set_nowrite() makes
fi->writectr negative and starts waiting for completion of that
in-flight request
3. Userspace fuse daemon ACKs the request and fuse_writepage_end() is
called; it calls __fuse_flush_writepages(), but the latter does nothing
because fi->writectr < 0
4. fuse_do_setattr() proceeds extending i_size and calling
__fuse_release_nowrite(). But now new (increased) i_size will be used as
'crop' arg of __fuse_flush_writepages()
stale data can leak to the server.
Thanks,
Maxim
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
