add prefetch property

ZFS prefetch is currently governed by the zfs_prefetch_disable
tunable. However, this is a module-wide settings - if a specific
dataset benefits from prefetch, while others have issue with it,
an optimal solution does not exists.

This commit introduce the "prefetch" tri-state property, which enable
granular control (at dataset/volume level) for prefetching.

This patch does not remove the zfs_prefetch_disable, which reimains
a system-wide switch for enable/disable prefetch. However, to avoid
duplication, it would be preferable to deprecate and then remove
the module tunable.

Please note that during pool import, os_prefetch defaults to 0.
For this reason ZFS_PREFETCH_ALL was enumerated as 0, otherwise
prefetch would be disabled for pool import (causing slower import).

Signed-off-by: Gionatan Danti <[email protected]>

include/sys/dmu_objset.h

@@ -197,6 +197,9 @@ struct objset {
  dmu_objset_upgrade_cb_t os_upgrade_cb;
  boolean_t os_upgrade_exit;
  int os_upgrade_status;
+
+ /* prefetch */
+ zfs_prefetch_type_t os_prefetch;
 };
 
 #define DMU_META_OBJSET 0

include/sys/fs/zfs.h

@@ -191,6 +191,7 @@ typedef enum {
  ZFS_PROP_REDACTED,
  ZFS_PROP_REDACT_SNAPS,
  ZFS_PROP_SNAPSHOTS_CHANGED,
+ ZFS_PROP_PREFETCH,
  ZFS_NUM_PROPS
 } zfs_prop_t;
 
@@ -543,6 +544,12 @@ typedef enum zfs_key_location {
  ZFS_KEYLOCATION_LOCATIONS
 } zfs_keylocation_t;
 
+typedef enum {
+ ZFS_PREFETCH_ALL = 0,
+ ZFS_PREFETCH_METADATA = 1,
+ ZFS_PREFETCH_NONE = 2
+} zfs_prefetch_type_t;
+
 #define DEFAULT_PBKDF2_ITERATIONS 350000
 #define MIN_PBKDF2_ITERATIONS 100000
 

module/zcommon/zfs_prop.c

@@ -388,6 +388,13 @@ zfs_prop_init(void)
  { NULL }
  };
 
+ static const zprop_index_t prefetch_table[] = {
+ { "all", ZFS_PREFETCH_ALL },
+ { "metadata", ZFS_PREFETCH_METADATA },
+ { "none", ZFS_PREFETCH_NONE },
+ { NULL }
+ };
+
  struct zfs_mod_supported_features *sfeatures =
  zfs_mod_list_supported(ZFS_SYSFS_DATASET_PROPERTIES);
 
@@ -468,6 +475,10 @@ zfs_prop_init(void)
  ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME,
  "default | full | geom | dev | none", "VOLMODE", volmode_table,
  sfeatures);
+ zprop_register_index(ZFS_PROP_PREFETCH, "prefetch",
+ ZFS_PREFETCH_ALL, PROP_INHERIT, 
+ ZFS_TYPE_FILESYSTEM | ZFS_TYPE_SNAPSHOT | ZFS_TYPE_VOLUME,
+ "all | metadata | none", "PREFETCH", prefetch_table, sfeatures);
 
  /* inherit index (boolean) properties */
  zprop_register_index(ZFS_PROP_ATIME, "atime", 1, PROP_INHERIT,

module/zfs/dmu_objset.c

@@ -263,6 +263,19 @@ secondary_cache_changed_cb(void *arg, uint64_t newval)
  os->os_secondary_cache = newval;
 }
 
+static void
+prefetch_changed_cb(void *arg, uint64_t newval)
+{
+ objset_t *os = arg;
+
+ /*
+ * Inheritance should have been done by now.
+ */
+ ASSERT(newval == ZFS_PREFETCH_ALL || newval == ZFS_PREFETCH_NONE ||
+ newval == ZFS_PREFETCH_METADATA);
+ os->os_prefetch = newval;
+}
+
 static void
 sync_changed_cb(void *arg, uint64_t newval)
 {
@@ -562,6 +575,11 @@ dmu_objset_open_impl(spa_t *spa, dsl_dataset_t *ds, blkptr_t *bp,
  zfs_prop_to_name(ZFS_PROP_SECONDARYCACHE),
  secondary_cache_changed_cb, os);
  }
+ if (err == 0) {
+ err = dsl_prop_register(ds,
+ zfs_prop_to_name(ZFS_PROP_PREFETCH),
+ prefetch_changed_cb, os);
+ }
  if (!ds->ds_is_snapshot) {
  if (err == 0) {
  err = dsl_prop_register(ds,

module/zfs/dmu_zfetch.c

@@ -329,9 +329,14 @@ dmu_zfetch_prepare(zfetch_t *zf, uint64_t blkid, uint64_t nblks,
 {
  zstream_t *zs;
  spa_t *spa = zf->zf_dnode->dn_objset->os_spa;
+ zfs_prefetch_type_t os_prefetch = zf->zf_dnode->dn_objset->os_prefetch;
 
- if (zfs_prefetch_disable)
+ if (zfs_prefetch_disable || os_prefetch == ZFS_PREFETCH_NONE)
  return (NULL);
+
+ if (os_prefetch == ZFS_PREFETCH_METADATA)
+ fetch_data = B_FALSE;
+
  /*
  * If we haven't yet loaded the indirect vdevs' mappings, we
  * can only read from blocks that we carefully ensure are on