Add a comment explaining why we call dmu_write() unconditionally, even if

uiomove() fails, especially that it is different from what OpenSolaris
does (I'm not entirely sure they are right).

Suggested by:	darrenr

1 files changed, 9 insertions, 0 deletions

diff --git a/sys/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c b/sys/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
index 5888e06..288bf9a 100644
--- a/sys/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
+++ b/sys/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
@@ -327,6 +327,15 @@ again:
 				va = (caddr_t)sf_buf_kva(sf);
 				woff = uio->uio_loffset - off;
 				error = uiomove(va + off, bytes, UIO_WRITE, uio);
+				/*
+				 * The uiomove() above could have been partially
+				 * successful, that's why we call dmu_write()
+				 * below unconditionally. The page was marked
+				 * non-dirty above and we would lose the changes
+				 * without doing so. If the uiomove() failed
+				 * entirely, well, we just write what we got
+				 * before one more time.
+				 */
 				dmu_write(os, zp->z_id, woff,
 				    MIN(PAGESIZE, fsize - woff), va, tx);
 				sf_buf_free(sf);