Reduce size of zfs_sb_t: allocate z_hold_mtx separately

zfs_sb_t has grown to the point where using kmem_zalloc() for allocations
is triggering the 32k warning threshold.

We can't safely convert this entire allocation to use vmem_alloc() instead
of kmem_alloc() because the backing_dev_info structure is embedded here.
It depends on the bit_waitqueue() function which won't behave properly
when given a virtual address.

Instead, use vmem_alloc() to allocate the z_hold_mtx array separately.

Signed-off-by: Brian Behlendorf <[email protected]>
Signed-off-by: Chris Dunlop <[email protected]>
Closes #3178

3 files changed, 10 insertions, 1 deletions

diff --git a/include/sys/zfs_vfsops.h b/include/sys/zfs_vfsops.h
index 4b88260de..c4db2a911 100644
--- a/include/sys/zfs_vfsops.h
+++ b/include/sys/zfs_vfsops.h
@@ -92,7 +92,7 @@ typedef struct zfs_sb {
 	uint64_t	z_replay_eof;	/* New end of file - replay only */
 	sa_attr_type_t	*z_attr_table;	/* SA attr mapping->id */
 #define	ZFS_OBJ_MTX_SZ	256
-	kmutex_t	z_hold_mtx[ZFS_OBJ_MTX_SZ];	/* znode hold locks */
+	kmutex_t	*z_hold_mtx;	/* znode hold locks */
 } zfs_sb_t;
 
 #define	ZFS_SUPER_MAGIC	0x2fc12fc1
diff --git a/module/zfs/zfs_vfsops.c b/module/zfs/zfs_vfsops.c
index e98f4bf6a..41a1c4d88 100644
--- a/module/zfs/zfs_vfsops.c
+++ b/module/zfs/zfs_vfsops.c
@@ -776,6 +776,9 @@ zfs_sb_create(const char *osname, zfs_sb_t **zsbp)
 	rrw_init(&zsb->z_teardown_lock, B_FALSE);
 	rw_init(&zsb->z_teardown_inactive_lock, NULL, RW_DEFAULT, NULL);
 	rw_init(&zsb->z_fuid_lock, NULL, RW_DEFAULT, NULL);
+
+	zsb->z_hold_mtx = vmem_zalloc(sizeof (kmutex_t) * ZFS_OBJ_MTX_SZ,
+	    KM_SLEEP);
 	for (i = 0; i != ZFS_OBJ_MTX_SZ; i++)
 		mutex_init(&zsb->z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL);
 
@@ -789,6 +792,8 @@ zfs_sb_create(const char *osname, zfs_sb_t **zsbp)
 out:
 	dmu_objset_disown(os, zsb);
 	*zsbp = NULL;
+
+	vmem_free(zsb->z_hold_mtx, sizeof (kmutex_t) * ZFS_OBJ_MTX_SZ);
 	kmem_free(zsb, sizeof (zfs_sb_t));
 	return (error);
 }
@@ -892,6 +897,7 @@ zfs_sb_free(zfs_sb_t *zsb)
 	rw_destroy(&zsb->z_fuid_lock);
 	for (i = 0; i != ZFS_OBJ_MTX_SZ; i++)
 		mutex_destroy(&zsb->z_hold_mtx[i]);
+	vmem_free(zsb->z_hold_mtx, sizeof (kmutex_t) * ZFS_OBJ_MTX_SZ);
 	mutex_destroy(&zsb->z_ctldir_lock);
 	avl_destroy(&zsb->z_ctldir_snaps);
 	kmem_free(zsb, sizeof (zfs_sb_t));
diff --git a/module/zfs/zfs_znode.c b/module/zfs/zfs_znode.c
index 3a7c30db2..a96ac8338 100644
--- a/module/zfs/zfs_znode.c
+++ b/module/zfs/zfs_znode.c
@@ -1731,6 +1731,8 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
 	list_create(&zsb->z_all_znodes, sizeof (znode_t),
 	    offsetof(znode_t, z_link_node));
 
+	zsb->z_hold_mtx = vmem_zalloc(sizeof (kmutex_t) * ZFS_OBJ_MTX_SZ,
+	    KM_SLEEP);
 	for (i = 0; i != ZFS_OBJ_MTX_SZ; i++)
 		mutex_init(&zsb->z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL);
 
@@ -1755,6 +1757,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
 	for (i = 0; i != ZFS_OBJ_MTX_SZ; i++)
 		mutex_destroy(&zsb->z_hold_mtx[i]);
 
+	vmem_free(zsb->z_hold_mtx, sizeof (kmutex_t) * ZFS_OBJ_MTX_SZ);
 	kmem_free(sb, sizeof (struct super_block));
 	kmem_free(zsb, sizeof (zfs_sb_t));
 }