FS DAX requires file systems to call into the DAX layout prior to
unlinking inodes to ensure there is no ongoing DMA or other remote
access to the direct mapped page. The fuse file system implements
fuse_dax_break_layouts() to do this which includes a comment
indicating that passing dmap_end == 0 leads to unmapping of the whole
file.

However this is not true - passing dmap_end == 0 will not unmap
anything before dmap_start, and further more
dax_layout_busy_page_range() will not scan any of the range to see if
there maybe ongoing DMA access to the range. Fix this by checking for
dmap_end == 0 in fuse_dax_break_layouts() and pass the entire file
range to dax_layout_busy_page_range().

Signed-off-by: Alistair Popple <apop...@nvidia.com>
Fixes: 6ae330cad6ef ("virtiofs: serialize truncate/punch_hole and dax fault 
path")
Cc: Vivek Goyal <vgo...@redhat.com>

---

I am not at all familiar with the fuse file system driver so I have no
idea if the comment is relevant or not and whether the documented
behaviour for dmap_end == 0 is ever relied upon. However this seemed
like the safest fix unless someone more familiar with fuse can confirm
that dmap_end == 0 is never used.
---
 fs/fuse/dax.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/fuse/dax.c b/fs/fuse/dax.c
index 12ef91d..501a097 100644
--- a/fs/fuse/dax.c
+++ b/fs/fuse/dax.c
@@ -693,6 +693,10 @@ int fuse_dax_break_layouts(struct inode *inode, u64 
dmap_start,
                ret = __fuse_dax_break_layouts(inode, &retry, dmap_start,
                                               dmap_end);
        } while (ret == 0 && retry);
+       if (!dmap_end) {
+               dmap_start = 0;
+               dmap_end = LLONG_MAX;
+       }
 
        return ret;
 }
-- 
git-series 0.9.1

Reply via email to