zfs+sendfile: populate all requested pages, not just those already cached

kern_sendfile() uses vm_rdwr() to read-ahead blocks of data to populate
page cache.  When sendfile stumbles upon a page that is not populated
yet, it sends out all the mbufs that it collected so far.  This
resulted in very poor performance with ZFS when file data is not in the
page cache, because ZFS vop_read for UIO_NOCOPY case populated only
those pages that are already in cache, but not valid.  Which means that
most of the time it populated only the first requested page in the
described above scenario.

Reported by:	Alexander Zagrebin <alexz@visp.ru>
Tested by:	Alexander Zagrebin <alexz@visp.ru>,
		Artemiev Igor <ai@kliksys.ru>
MFC after:	12 days

1 files changed, 14 insertions, 4 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 ba0f0eb48717..85647a89ef15 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
@@ -67,6 +67,7 @@
 #include <sys/sf_buf.h>
 #include <sys/sched.h>
 #include <sys/acl.h>
+#include <vm/vm_pageout.h>
 
 /*
  * Programming rules.
@@ -481,7 +482,7 @@ again:
 				uiomove_fromphys(&m, off, bytes, uio);
 			VM_OBJECT_LOCK(obj);
 			vm_page_wakeup(m);
-		} else if (m != NULL && uio->uio_segflg == UIO_NOCOPY) {
+		} else if (uio->uio_segflg == UIO_NOCOPY) {
 			/*
 			 * The code below is here to make sendfile(2) work
 			 * correctly with ZFS. As pointed out by ups@
@@ -491,7 +492,7 @@ again:
 			 */
 			KASSERT(off == 0,
 			    ("unexpected offset in mappedread for sendfile"));
-			if ((m->oflags & VPO_BUSY) != 0) {
+			if (m != NULL && (m->oflags & VPO_BUSY) != 0) {
 				/*
 				 * Reference the page before unlocking and
 				 * sleeping so that the page daemon is less
@@ -501,8 +502,17 @@ again:
 				vm_page_flag_set(m, PG_REFERENCED);
 				vm_page_sleep(m, "zfsmrb");
 				goto again;
+			} else if (m == NULL) {
+				m = vm_page_alloc(obj, OFF_TO_IDX(start),
+				    VM_ALLOC_NOBUSY | VM_ALLOC_NORMAL);
+				if (m == NULL) {
+					VM_OBJECT_UNLOCK(obj);
+					VM_WAIT;
+					VM_OBJECT_LOCK(obj);
+					goto again;
+				}
 			}
-			vm_page_busy(m);
+			vm_page_io_start(m);
 			VM_OBJECT_UNLOCK(obj);
 			if (dirbytes > 0) {
 				error = dmu_read_uio(os, zp->z_id, uio,
@@ -520,7 +530,7 @@ again:
 			VM_OBJECT_LOCK(obj);
 			if (error == 0)
 				m->valid = VM_PAGE_BITS_ALL;
-			vm_page_wakeup(m);
+			vm_page_io_finish(m);
 			if (error == 0) {
 				uio->uio_resid -= bytes;
 				uio->uio_offset += bytes;