vdev_ashift should only be set once

== Motivation and Context

The new vdev ashift optimization prevents the removal of devices when
a zfs configuration is comprised of disks which have different logical
and physical block sizes. This is caused because we set 'spa_min_ashift'
in vdev_open and then later call 'vdev_ashift_optimize'. This would
result in an inconsistency between spa's ashift calculations and that
of the top-level vdev.

In addition, the optimization logical ignores the overridden ashift
value that would be provided by '-o ashift=<val>'.

== Description

This change reworks the vdev ashift optimization so that it's only
set the first time the device is configured. It still allows the
physical and logical ahsift values to be set every time the device
is opened but those values are only consulted on first open.

Reviewed-by: Matthew Ahrens <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Reviewed-by: Cedric Berger <[email protected]>
Signed-off-by: George Wilson <[email protected]>
External-Issue: DLPX-71831
Closes #10932 

3 files changed, 31 insertions, 6 deletions

diff --git a/module/os/freebsd/zfs/sysctl_os.c b/module/os/freebsd/zfs/sysctl_os.c
index 200bbf43d..b3cb7e7e4 100644
--- a/module/os/freebsd/zfs/sysctl_os.c
+++ b/module/os/freebsd/zfs/sysctl_os.c
@@ -121,6 +121,7 @@ SYSCTL_NODE(_vfs_zfs, OID_AUTO, zio, CTLFLAG_RW, 0, "ZFS ZIO");
 SYSCTL_NODE(_vfs_zfs_livelist, OID_AUTO, condense, CTLFLAG_RW, 0,
     "ZFS livelist condense");
 SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, cache, CTLFLAG_RW, 0, "ZFS VDEV Cache");
+SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, file, CTLFLAG_RW, 0, "ZFS VDEV file");
 SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, mirror, CTLFLAG_RD, 0,
     "ZFS VDEV mirror");
 
diff --git a/module/os/freebsd/zfs/vdev_file.c b/module/os/freebsd/zfs/vdev_file.c
index 4d27751c8..cf762c5fd 100644
--- a/module/os/freebsd/zfs/vdev_file.c
+++ b/module/os/freebsd/zfs/vdev_file.c
@@ -20,7 +20,7 @@
  */
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2011, 2016 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2020 by Delphix. All rights reserved.
  */
 
 #include <sys/zfs_context.h>
@@ -40,6 +40,9 @@
 
 static taskq_t *vdev_file_taskq;
 
+unsigned long vdev_file_logical_ashift = SPA_MINBLOCKSHIFT;
+unsigned long vdev_file_physical_ashift = SPA_MINBLOCKSHIFT;
+
 void
 vdev_file_init(void)
 {
@@ -167,8 +170,8 @@ skip_open:
 	}
 
 	*max_psize = *psize = zfa.zfa_size;
-	*logical_ashift = SPA_MINBLOCKSHIFT;
-	*physical_ashift = SPA_MINBLOCKSHIFT;
+	*logical_ashift = vdev_file_logical_ashift;
+	*physical_ashift = vdev_file_physical_ashift;
 
 	return (0);
 }
@@ -326,3 +329,8 @@ vdev_ops_t vdev_disk_ops = {
 };
 
 #endif
+
+ZFS_MODULE_PARAM(zfs_vdev_file, vdev_file_, logical_ashift, ULONG, ZMOD_RW,
+	"Logical ashift for file-based devices");
+ZFS_MODULE_PARAM(zfs_vdev_file, vdev_file_, physical_ashift, ULONG, ZMOD_RW,
+	"Physical ashift for file-based devices");
diff --git a/module/os/linux/zfs/vdev_file.c b/module/os/linux/zfs/vdev_file.c
index a4e71ca40..423ce8581 100644
--- a/module/os/linux/zfs/vdev_file.c
+++ b/module/os/linux/zfs/vdev_file.c
@@ -20,7 +20,7 @@
  */
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2011, 2016 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2020 by Delphix. All rights reserved.
  */
 
 #include <sys/zfs_context.h>
@@ -45,6 +45,17 @@
 
 static taskq_t *vdev_file_taskq;
 
+/*
+ * By default, the logical/physical ashift for file vdevs is set to
+ * SPA_MINBLOCKSHIFT (9). This allows all file vdevs to use 512B (1 << 9)
+ * blocksizes. Users may opt to change one or both of these for testing
+ * or performance reasons. Care should be taken as these values will
+ * impact the vdev_ashift setting which can only be set at vdev creation
+ * time.
+ */
+unsigned long vdev_file_logical_ashift = SPA_MINBLOCKSHIFT;
+unsigned long vdev_file_physical_ashift = SPA_MINBLOCKSHIFT;
+
 static void
 vdev_file_hold(vdev_t *vd)
 {
@@ -159,8 +170,8 @@ skip_open:
 	}
 
 	*max_psize = *psize = zfa.zfa_size;
-	*logical_ashift = SPA_MINBLOCKSHIFT;
-	*physical_ashift = SPA_MINBLOCKSHIFT;
+	*logical_ashift = vdev_file_logical_ashift;
+	*physical_ashift = vdev_file_physical_ashift;
 
 	return (0);
 }
@@ -346,3 +357,8 @@ vdev_ops_t vdev_disk_ops = {
 };
 
 #endif
+
+ZFS_MODULE_PARAM(zfs_vdev_file, vdev_file_, logical_ashift, ULONG, ZMOD_RW,
+	"Logical ashift for file-based devices");
+ZFS_MODULE_PARAM(zfs_vdev_file, vdev_file_, physical_ashift, ULONG, ZMOD_RW,
+	"Physical ashift for file-based devices");