Import vdev ashift optimization from FreeBSD

Many modern devices use physical allocation units that are much
larger than the minimum logical allocation size accessible by
external commands. Two prevalent examples of this are 512e disk
drives (512b logical sector, 4K physical sector) and flash devices
(512b logical sector, 4K or larger allocation block size, and 128k
or larger erase block size). Operations that modify less than the
physical sector size result in a costly read-modify-write or garbage
collection sequence on these devices.

Simply exporting the true physical sector of the device to ZFS would
yield optimal performance, but has two serious drawbacks:

 1. Existing pools created with devices that have different logical
    and physical block sizes, but were configured to use the logical
    block size (e.g. because the OS version used for pool construction
    reported the logical block size instead of the physical block
    size) will suddenly find that the vdev allocation size has
    increased. This can be easily tolerated for active members of
    the array, but ZFS would prevent replacement of a vdev with
    another identical device because it now appears that the smaller
    allocation size required by the pool is not supported by the new
    device.

 2. The device's physical block size may be too large to be supported
    by ZFS. The optimal allocation size for the vdev may be quite
    large. For example, a RAID controller may export a vdev that
    requires read-modify-write cycles unless accessed using 64k
    aligned/sized requests. ZFS currently has an 8k minimum block
    size limit.

Reporting both the logical and physical allocation sizes for vdevs
solves these problems. A device may be used so long as the logical
block size is compatible with the configuration. By comparing the
logical and physical block sizes, new configurations can be optimized
and administrators can be notified of any existing pools that are
sub-optimal.

Reviewed-by: Ryan Moeller <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Co-authored-by: Matthew Macy <[email protected]>
Signed-off-by: Matt Macy <[email protected]>
Closes #10619 

3 files changed, 35 insertions, 35 deletions

diff --git a/module/os/freebsd/zfs/sysctl_os.c b/module/os/freebsd/zfs/sysctl_os.c
index d76b49de9..200bbf43d 100644
--- a/module/os/freebsd/zfs/sysctl_os.c
+++ b/module/os/freebsd/zfs/sysctl_os.c
@@ -52,6 +52,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/spa.h>
 #include <sys/spa_impl.h>
 #include <sys/vdev.h>
+#include <sys/vdev_impl.h>
 #include <sys/dmu.h>
 #include <sys/dsl_dir.h>
 #include <sys/dsl_dataset.h>
@@ -518,56 +519,53 @@ SYSCTL_INT(_vfs_zfs, OID_AUTO, space_map_ibs, CTLFLAG_RWTUN,
 
 
 /* vdev.c */
-#ifdef notyet
-extern uint64_t zfs_max_auto_ashift;
-extern uint64_t zfs_min_auto_ashift;
-
-static int
-sysctl_vfs_zfs_max_auto_ashift(SYSCTL_HANDLER_ARGS)
+int
+param_set_min_auto_ashift(SYSCTL_HANDLER_ARGS)
 {
 	uint64_t val;
 	int err;
 
-	val = zfs_max_auto_ashift;
+	val = zfs_vdev_min_auto_ashift;
 	err = sysctl_handle_64(oidp, &val, 0, req);
 	if (err != 0 || req->newptr == NULL)
-		return (err);
+		return (SET_ERROR(err));
 
-	if (val > ASHIFT_MAX || val < zfs_min_auto_ashift)
-		return (EINVAL);
+	if (val < ASHIFT_MIN || val > zfs_vdev_max_auto_ashift)
+		return (SET_ERROR(EINVAL));
 
-	zfs_max_auto_ashift = val;
+	zfs_vdev_min_auto_ashift = val;
 
 	return (0);
 }
-SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift,
-    CTLTYPE_U64 | CTLFLAG_MPSAFE | CTLFLAG_RW, 0, sizeof (uint64_t),
-    sysctl_vfs_zfs_max_auto_ashift, "QU",
-    "Max ashift used when optimising for logical -> physical sectors size on "
-    "new top-level vdevs.");
-static int
-sysctl_vfs_zfs_min_auto_ashift(SYSCTL_HANDLER_ARGS)
+
+int
+param_set_max_auto_ashift(SYSCTL_HANDLER_ARGS)
 {
 	uint64_t val;
 	int err;
 
-	val = zfs_min_auto_ashift;
+	val = zfs_vdev_max_auto_ashift;
 	err = sysctl_handle_64(oidp, &val, 0, req);
 	if (err != 0 || req->newptr == NULL)
-		return (err);
+		return (SET_ERROR(err));
 
-	if (val < ASHIFT_MIN || val > zfs_max_auto_ashift)
-		return (EINVAL);
+	if (val > ASHIFT_MAX || val < zfs_vdev_min_auto_ashift)
+		return (SET_ERROR(EINVAL));
 
-	zfs_min_auto_ashift = val;
+	zfs_vdev_max_auto_ashift = val;
 
 	return (0);
 }
-SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift,
-    CTLTYPE_U64 | CTLFLAG_MPSAFE | CTLFLAG_RW, 0, sizeof (uint64_t),
-    sysctl_vfs_zfs_min_auto_ashift, "QU",
-    "Min ashift used when creating new top-level vdevs.");
-#endif
+
+SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift, CTLTYPE_U64 | CTLFLAG_RWTUN,
+    &zfs_vdev_min_auto_ashift, sizeof (zfs_vdev_min_auto_ashift),
+    param_set_min_auto_ashift, "QU",
+    "Min ashift used when creating new top-level vdev. (LEGACY)");
+SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift, CTLTYPE_U64 | CTLFLAG_RWTUN,
+    &zfs_vdev_max_auto_ashift, sizeof (zfs_vdev_max_auto_ashift),
+    param_set_max_auto_ashift, "QU",
+    "Max ashift used when optimizing for logical -> physical sector size on "
+    "new top-level vdevs. (LEGACY)");
 
 /*
  * Since the DTL space map of a vdev is not expected to have a lot of
diff --git a/module/os/freebsd/zfs/vdev_file.c b/module/os/freebsd/zfs/vdev_file.c
index cca6bffd9..4d27751c8 100644
--- a/module/os/freebsd/zfs/vdev_file.c
+++ b/module/os/freebsd/zfs/vdev_file.c
@@ -83,7 +83,7 @@ vdev_file_open_mode(spa_mode_t spa_mode)
 
 static int
 vdev_file_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize,
-    uint64_t *ashift)
+    uint64_t *logical_ashift, uint64_t *physical_ashift)
 {
 	vdev_file_t *vf;
 	zfs_file_t *fp;
@@ -167,7 +167,8 @@ skip_open:
 	}
 
 	*max_psize = *psize = zfa.zfa_size;
-	*ashift = SPA_MINBLOCKSHIFT;
+	*logical_ashift = SPA_MINBLOCKSHIFT;
+	*physical_ashift = SPA_MINBLOCKSHIFT;
 
 	return (0);
 }
diff --git a/module/os/freebsd/zfs/vdev_geom.c b/module/os/freebsd/zfs/vdev_geom.c
index 8462755f8..bf06f6919 100644
--- a/module/os/freebsd/zfs/vdev_geom.c
+++ b/module/os/freebsd/zfs/vdev_geom.c
@@ -802,7 +802,7 @@ vdev_geom_open_by_path(vdev_t *vd, int check_guid)
 
 static int
 vdev_geom_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize,
-    uint64_t *logical_ashift)
+    uint64_t *logical_ashift, uint64_t *physical_ashift)
 {
 	struct g_provider *pp;
 	struct g_consumer *cp;
@@ -949,11 +949,12 @@ skip_open:
 	 * transfer size.
 	 */
 	*logical_ashift = highbit(MAX(pp->sectorsize, SPA_MINBLOCKSIZE)) - 1;
-#ifdef notyet
-	if (pp->stripesize > (1 << *logical_ashift) && ISP2(pp->stripesize) &&
-	    pp->stripesize <= (1 << ASHIFT_MAX) && pp->stripeoffset == 0)
+	*physical_ashift = 0;
+	if (pp->stripesize && pp->stripesize > (1 << *logical_ashift) &&
+	    ISP2(pp->stripesize) && pp->stripesize <= (1 << ASHIFT_MAX) &&
+	    pp->stripeoffset == 0)
 		*physical_ashift = highbit(pp->stripesize) - 1;
-#endif
+
 	/*
 	 * Clear the nowritecache settings, so that on a vdev_reopen()
 	 * we will try again.