- Always try to write one whole page at a time.

- vm_page_undirty() is enough (instead of vm_page_set_validclean()), but it has
  to be called before we write the data in case someone makes page dirty after
  our write, but before our vm_page_undirty() call.
- Always dmu_write, not matter if uiomove() succeeded, because it could
  partially be ok and we would lose some changes.

All good ideas from:	ups

2 files changed, 18 insertions, 14 deletions

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
index d434cfc..77e6545 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
@@ -298,29 +298,31 @@ mappedwrite(vnode_t *vp, int nbytes, uio_t *uio, dmu_tx_t *tx)
 	VM_OBJECT_LOCK(obj);
 	for (start &= PAGEMASK; len > 0; start += PAGESIZE) {
 		uint64_t bytes = MIN(PAGESIZE - off, len);
-		uint64_t woff = uio->uio_loffset;
 
 again:
 		if ((m = vm_page_lookup(obj, OFF_TO_IDX(start))) != NULL &&
 		    vm_page_is_valid(m, (vm_offset_t)off, bytes)) {
+			uint64_t woff, dmubytes;
 			caddr_t va;
 
 			if (vm_page_sleep_if_busy(m, FALSE, "zfsmwb"))
 				goto again;
+			woff = uio->uio_loffset;
+			dmubytes = MIN(PAGESIZE,
+			    obj->un_pager.vnp.vnp_size - (woff - off));
 			vm_page_busy(m);
+			vm_page_lock_queues();
+			vm_page_undirty(m);
+			vm_page_unlock_queues();
 			VM_OBJECT_UNLOCK(obj);
 			sched_pin();
 			sf = sf_buf_alloc(m, SFB_CPUPRIVATE);
 			va = (caddr_t)sf_buf_kva(sf);
-			error = uiomove(va+off, bytes, UIO_WRITE, uio);
-			if (error == 0)
-				dmu_write(os, zp->z_id, woff, bytes, va+off, tx);
+			error = uiomove(va + off, bytes, UIO_WRITE, uio);
+			dmu_write(os, zp->z_id, woff - off, dmubytes, va, tx);
 			sf_buf_free(sf);
 			sched_unpin();
 			VM_OBJECT_LOCK(obj);
-			vm_page_lock_queues();
-			vm_page_set_validclean(m, off, bytes);
-			vm_page_unlock_queues();
 			vm_page_wakeup(m);
 		} else {
 			VM_OBJECT_UNLOCK(obj);
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 d434cfc..77e6545 100644
--- a/sys/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
+++ b/sys/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
@@ -298,29 +298,31 @@ mappedwrite(vnode_t *vp, int nbytes, uio_t *uio, dmu_tx_t *tx)
 	VM_OBJECT_LOCK(obj);
 	for (start &= PAGEMASK; len > 0; start += PAGESIZE) {
 		uint64_t bytes = MIN(PAGESIZE - off, len);
-		uint64_t woff = uio->uio_loffset;
 
 again:
 		if ((m = vm_page_lookup(obj, OFF_TO_IDX(start))) != NULL &&
 		    vm_page_is_valid(m, (vm_offset_t)off, bytes)) {
+			uint64_t woff, dmubytes;
 			caddr_t va;
 
 			if (vm_page_sleep_if_busy(m, FALSE, "zfsmwb"))
 				goto again;
+			woff = uio->uio_loffset;
+			dmubytes = MIN(PAGESIZE,
+			    obj->un_pager.vnp.vnp_size - (woff - off));
 			vm_page_busy(m);
+			vm_page_lock_queues();
+			vm_page_undirty(m);
+			vm_page_unlock_queues();
 			VM_OBJECT_UNLOCK(obj);
 			sched_pin();
 			sf = sf_buf_alloc(m, SFB_CPUPRIVATE);
 			va = (caddr_t)sf_buf_kva(sf);
-			error = uiomove(va+off, bytes, UIO_WRITE, uio);
-			if (error == 0)
-				dmu_write(os, zp->z_id, woff, bytes, va+off, tx);
+			error = uiomove(va + off, bytes, UIO_WRITE, uio);
+			dmu_write(os, zp->z_id, woff - off, dmubytes, va, tx);
 			sf_buf_free(sf);
 			sched_unpin();
 			VM_OBJECT_LOCK(obj);
-			vm_page_lock_queues();
-			vm_page_set_validclean(m, off, bytes);
-			vm_page_unlock_queues();
 			vm_page_wakeup(m);
 		} else {
 			VM_OBJECT_UNLOCK(obj);