fuse: break infinite loop in fuse_fill_write_pages()

I got a report about unkillable task eating CPU. Further
investigation shows, that the problem is in the fuse_fill_write_pages()
function. If iov's first segment has zero length, we get an infinite
loop, because we never reach iov_iter_advance() call.

Fix this by calling iov_iter_advance() before repeating an attempt to
copy data from userspace.

A similar problem is described in 124d3b7 ("fix writev regression:
pan hanging unkillable and un-straceable"). If zero-length segmend
is followed by segment with invalid address,
iov_iter_fault_in_readable() checks only first segment (zero-length),
iov_iter_copy_from_user_atomic() skips it, fails at second and
returns zero -> goto again without skipping zero-length segment.

Patch calls iov_iter_advance() before goto again: we'll skip zero-length
segment at second iteraction and iov_iter_fault_in_readable() will detect
invalid address.

Special thanks to Konstantin Khlebnikov, who helped a lot with the commit
description.

Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Maxim Patlasov <mpatlasov@parallels.com>
Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Signed-off-by: Roman Gushchin <klamm@yandex-team.ru>
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Fixes: ea9b990 ("fuse: implement perform_write")
Cc: <stable@vger.kernel.org>

fs/fuse/file.c

@@ -1049,6 +1049,7 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req,
 		tmp = iov_iter_copy_from_user_atomic(page, ii, offset, bytes);
 		flush_dcache_page(page);
 
+		iov_iter_advance(ii, tmp);
 		if (!tmp) {
 			unlock_page(page);
 			page_cache_release(page);
@@ -1061,7 +1062,6 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req,
 		req->page_descs[req->num_pages].length = tmp;
 		req->num_pages++;
 
-		iov_iter_advance(ii, tmp);
 		count += tmp;
 		pos += tmp;
 		offset += tmp;