fusefs: fix mmap'd writes in direct_io mode

If a FUSE server returns FOPEN_DIRECT_IO in response to FUSE_OPEN, that
instructs the kernel to bypass the page cache for that file. This feature
is also known by libfuse's name: "direct_io".

However, when accessing a file via mmap, there is no possible way to bypass
the cache completely. This change fixes a deadlock that would happen when
an mmap'd write tried to invalidate a portion of the cache, wrongly assuming
that a write couldn't possibly come from cache if direct_io were set.

Arguably, we could instead disable mmap for files with FOPEN_DIRECT_IO set.
But allowing it is less likely to cause user complaints, and is more in
keeping with the spirit of open(2), where O_DIRECT instructs the kernel to
"reduce", not "eliminate" cache effects.

PR:		247276
Reported by:	trapexit@spawn.link
Reviewed by:	cem
MFC after:	3 days
Differential Revision:	https://reviews.freebsd.org/D26485

1 files changed, 8 insertions, 8 deletions

diff --git a/sys/fs/fuse/fuse_io.c b/sys/fs/fuse/fuse_io.c
index ac0235e8c718..4e178bb5340a 100644
--- a/sys/fs/fuse/fuse_io.c
+++ b/sys/fs/fuse/fuse_io.c
@@ -291,6 +291,7 @@ fuse_io_dispatch(struct vnode *vp, struct uio *uio, int ioflag,
 		fuse_vnode_update(vp, FN_MTIMECHANGE | FN_CTIMECHANGE);
 		if (directio) {
 			off_t start, end, filesize;
+			bool pages = (ioflag & IO_VMIO) != 0;
 
 			SDT_PROBE2(fusefs, , io, trace, 1,
 				"direct write of vnode");
@@ -301,15 +302,14 @@ fuse_io_dispatch(struct vnode *vp, struct uio *uio, int ioflag,
 
 			start = uio->uio_offset;
 			end = start + uio->uio_resid;
-			KASSERT((ioflag & (IO_VMIO | IO_DIRECT)) !=
-				(IO_VMIO | IO_DIRECT),
-			    ("IO_DIRECT used for a cache flush?"));
-			/* Invalidate the write cache when writing directly */
-			err = fuse_inval_buf_range(vp, filesize, start, end);
-			if (err)
-				return (err);
+			if (!pages) {
+				err = fuse_inval_buf_range(vp, filesize, start,
+				    end);
+				if (err)
+					return (err);
+			}
 			err = fuse_write_directbackend(vp, uio, cred, fufh,
-				filesize, ioflag, false);
+				filesize, ioflag, pages);
 		} else {
 			SDT_PROBE2(fusefs, , io, trace, 1,
 				"buffered write of vnode");