On 10 Sep 2022, at 17:14, Al Viro wrote:
similar to one that used to be in copy_page_to_iter(). Could you try
the following:
Yes, this fixes up generic/551. No crash or corruption. I'll send it
through a full run of xfstests as well.
Ben
nfsd_splice_actor(): handle compound pages
pipe_buffer might refer to a compound page (and contain more than a
PAGE_SIZE
worth of data). Theoretically it had been possible since way back,
but
nfsd_splice_actor() hadn't run into that until copy_page_to_iter()
change.
Fortunately, the only thing that changes for compound pages is that we
need to stuff each relevant subpage in and convert the offset into
offset
in the first subpage.
Hopefully-fixes: f0f6b614f83d "copy_page_to_iter(): don't split
high-order page in case of ITER_PIPE"
Signed-off-by: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
---
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 9f486b788ed0..b16aed158ba6 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -846,10 +846,14 @@ nfsd_splice_actor(struct pipe_inode_info *pipe,
struct pipe_buffer *buf,
struct splice_desc *sd)
{
struct svc_rqst *rqstp = sd->u.data;
-
- svc_rqst_replace_page(rqstp, buf->page);
- if (rqstp->rq_res.page_len == 0)
- rqstp->rq_res.page_base = buf->offset;
+ struct page *page = buf->page; // may be a compound one
+ unsigned offset = buf->offset;
+
+ page += offset / PAGE_SIZE;
+ for (int i = sd->len; i > 0; i -= PAGE_SIZE)
+ svc_rqst_replace_page(rqstp, page++);
+ if (rqstp->rq_res.page_len == 0) // first call
+ rqstp->rq_res.page_base = offset % PAGE_SIZE;
rqstp->rq_res.page_len += sd->len;
return sd->len;
}