Re: splice vs O_APPEND

[Miklos Szeredi]

On Thu, 9 Oct 2008, Linus Torvalds wrote:
> And it turns out that handling O_APPEND is actually pretty easy, so 
> instead of doing -EINVAL, we can just implement it. Something like this 
> (untested, of course).
> 
> Does this look better?

Yeah, only the append is now racy because the O_APPEND check is
outside i_mutex.  So maybe just stick with -EINVAL in
do_splice_from()?  That also covers do_splice_direct(), which is used
in NFS and sendfile() and a couple of other places.

We know that nobody is currently relying on O_APPEND semantics with
splice, so this should be OK.

Untested patch...

Miklos

Index: linux-2.6/fs/splice.c
===================================================================
--- linux-2.6.orig/fs/splice.c	2008-08-29 14:39:20.000000000 +0200
+++ linux-2.6/fs/splice.c	2008-10-09 18:19:25.000000000 +0200
@@ -892,6 +892,9 @@ static long do_splice_from(struct pipe_i
 {
 	int ret;
 
+	if (out->f_flags & O_APPEND)
+		return -EINVAL;
+
 	if (unlikely(!out->f_op || !out->f_op->splice_write))
 		return -EINVAL;
 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/