diff --git a/iscsi-scst/kernel/conn.c b/iscsi-scst/kernel/conn.c
index e563838fc..5f3bc93a0 100644
--- a/iscsi-scst/kernel/conn.c
+++ b/iscsi-scst/kernel/conn.c
@@ -850,7 +850,7 @@ static int iscsi_conn_alloc(struct iscsi_session *session,
 	struct iscsi_conn *conn;
 	int res = 0;
 
-	conn = kzalloc(sizeof(*conn), GFP_KERNEL);
+	conn = kzalloc(L1_CACHE_ALIGN(sizeof(*conn)), GFP_KERNEL);
 	if (!conn) {
 		res = -ENOMEM;
 		goto out_err;
diff --git a/iscsi-scst/kernel/iscsi.c b/iscsi-scst/kernel/iscsi.c
index b2f605b1c..a03e3734c 100644
--- a/iscsi-scst/kernel/iscsi.c
+++ b/iscsi-scst/kernel/iscsi.c
@@ -3918,7 +3918,7 @@ int iscsi_threads_pool_get(const cpumask_t *cpu_mask,
 
 	TRACE_DBG("%s", "Creating new iSCSI thread pool");
 
-	p = kzalloc(sizeof(*p), GFP_KERNEL);
+	p = kzalloc(L1_CACHE_ALIGN(sizeof(*p)), GFP_KERNEL);
 	if (p == NULL) {
 		PRINT_ERROR("Unable to allocate iSCSI thread pool (size %zd)",
 			sizeof(*p));
diff --git a/iscsi-scst/kernel/iscsi.h b/iscsi-scst/kernel/iscsi.h
index 72b7976cc..e14f1fd5f 100644
--- a/iscsi-scst/kernel/iscsi.h
+++ b/iscsi-scst/kernel/iscsi.h
@@ -77,7 +77,8 @@ struct iscsi_thread_pool {
 	struct list_head rd_list;
 	wait_queue_head_t rd_waitQ;
 
-	spinlock_t wr_lock;
+	/* It's used by another thread, hence aligned */
+	spinlock_t wr_lock ____cacheline_aligned_in_smp;
 	struct list_head wr_list;
 	wait_queue_head_t wr_waitQ;
 
diff --git a/iscsi-scst/kernel/session.c b/iscsi-scst/kernel/session.c
index fa279ed9e..7a6ba0175 100644
--- a/iscsi-scst/kernel/session.c
+++ b/iscsi-scst/kernel/session.c
@@ -43,7 +43,7 @@ static int iscsi_session_alloc(struct iscsi_target *target,
 	struct iscsi_session *session;
 	char *name = NULL;
 
-	session = kzalloc(sizeof(*session), GFP_KERNEL);
+	session = kzalloc(L1_CACHE_ALIGN(sizeof(*session)), GFP_KERNEL);
 	if (!session)
 		return -ENOMEM;
 
diff --git a/qla2x00t/qla2x00-target/qla2x00t.c b/qla2x00t/qla2x00-target/qla2x00t.c
index 72743b3bc..299af2e46 100644
--- a/qla2x00t/qla2x00-target/qla2x00t.c
+++ b/qla2x00t/qla2x00-target/qla2x00t.c
@@ -1155,7 +1155,7 @@ static struct q2t_sess *q2t_create_sess(scsi_qla_host_t *ha, fc_port_t *fcport,
 
 	/* We are under tgt_mutex, so a new sess can't be added behind us */
 
-	sess = kzalloc(sizeof(*sess), GFP_KERNEL);
+	sess = kzalloc(L1_CACHE_ALIGN(sizeof(*sess)), GFP_KERNEL);
 	if (sess == NULL) {
 		PRINT_ERROR("qla2x00t(%ld): session allocation failed, "
 			"all commands from port %02x:%02x:%02x:%02x:"
@@ -5736,7 +5736,7 @@ static int q2t_add_target(scsi_qla_host_t *ha)
 
 	sBUG_ON((ha->q2t_tgt != NULL) || (ha->tgt != NULL));
 
-	tgt = kzalloc(sizeof(*tgt), GFP_KERNEL);
+	tgt = kzalloc(L1_CACHE_ALIGN(sizeof(*tgt)), GFP_KERNEL);
 	if (tgt == NULL) {
 		PRINT_ERROR("qla2x00t: %s", "Allocation of tgt failed");
 		res = -ENOMEM;
diff --git a/scst/src/dev_handlers/scst_user.c b/scst/src/dev_handlers/scst_user.c
index 4e9bcfd3f..b8b330954 100644
--- a/scst/src/dev_handlers/scst_user.c
+++ b/scst/src/dev_handlers/scst_user.c
@@ -1196,9 +1196,8 @@ static int dev_user_map_buf(struct scst_user_cmd *ucmd, unsigned long ubuff,
 
 	ucmd->num_data_pages = num_pg;
 
-	ucmd->data_pages =
-		kmalloc(sizeof(*ucmd->data_pages) * ucmd->num_data_pages,
-			  GFP_KERNEL);
+	ucmd->data_pages = kmalloc(L1_CACHE_ALIGN(sizeof(*ucmd->data_pages) * ucmd->num_data_pages),
+				  GFP_KERNEL);
 	if (ucmd->data_pages == NULL) {
 		TRACE(TRACE_OUT_OF_MEM, "Unable to allocate data_pages array "
 			"(num_data_pages=%d)", ucmd->num_data_pages);
@@ -2931,7 +2930,7 @@ static int dev_user_register_dev(struct file *file,
 		goto out;
 	}
 
-	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
+	dev = kzalloc(L1_CACHE_ALIGN(sizeof(*dev)), GFP_KERNEL);
 	if (dev == NULL) {
 		res = -ENOMEM;
 		goto out_put;
diff --git a/scst/src/dev_handlers/scst_vdisk.c b/scst/src/dev_handlers/scst_vdisk.c
index 066ac7d09..03ee51fa9 100644
--- a/scst/src/dev_handlers/scst_vdisk.c
+++ b/scst/src/dev_handlers/scst_vdisk.c
@@ -1629,7 +1629,7 @@ static struct scatterlist *alloc_sg(size_t size, unsigned off, gfp_t gfp_mask,
 
 	sg_cnt = PAGE_ALIGN(size + off) >> PAGE_SHIFT;
 	sg = sg_cnt <= small_sg_size ? small_sg :
-		kmalloc(sg_cnt * sizeof(*sg), gfp_mask);
+		kmalloc(L1_CACHE_ALIGN(sg_cnt * sizeof(*sg)), gfp_mask);
 	if (!sg)
 		goto out;
 
@@ -3387,7 +3387,7 @@ static struct iovec *vdisk_alloc_iv(struct scst_cmd *cmd,
 		p->iv_count = 0;
 		/* It can't be called in atomic context */
 		p->iv = (iv_count <= ARRAY_SIZE(p->small_iv)) ? p->small_iv :
-			kmalloc(sizeof(*p->iv) * iv_count, GFP_KERNEL);
+			kmalloc(L1_CACHE_ALIGN(sizeof(*p->iv) * iv_count), GFP_KERNEL);
 		if (p->iv == NULL) {
 			PRINT_ERROR("Unable to allocate iv (%d)", iv_count);
 			goto out;
@@ -4307,7 +4307,7 @@ static int vdev_create(struct scst_dev_type *devt,
 	if (vdev_find(name))
 		goto out;
 
-	virt_dev = kzalloc(sizeof(*virt_dev), GFP_KERNEL);
+	virt_dev = kzalloc(L1_CACHE_ALIGN(sizeof(*virt_dev)), GFP_KERNEL);
 	if (virt_dev == NULL) {
 		PRINT_ERROR("Allocation of virtual device %s failed",
 			devt->name);
diff --git a/scst/src/scst_lib.c b/scst/src/scst_lib.c
index 662ec60bc..4efbf751d 100644
--- a/scst/src/scst_lib.c
+++ b/scst/src/scst_lib.c
@@ -3373,7 +3373,7 @@ int scst_alloc_device(gfp_t gfp_mask, struct scst_device **out_dev)
 
 	TRACE_ENTRY();
 
-	dev = kzalloc(sizeof(*dev), gfp_mask);
+	dev = kzalloc(L1_CACHE_ALIGN(sizeof(*dev)), gfp_mask);
 	if (dev == NULL) {
 		PRINT_ERROR("%s", "Allocation of scst_device failed");
 		res = -ENOMEM;
@@ -5249,9 +5249,10 @@ void scst_cmd_set_ext_cdb(struct scst_cmd *cmd,
 		goto out;
 	}
 
-	cmd->cdb = kmalloc(len, gfp_mask);
+	cmd->cdb = kmalloc(L1_CACHE_ALIGN(len), gfp_mask);
 	if (unlikely(cmd->cdb == NULL)) {
-		PRINT_ERROR("Unable to alloc extended CDB (size %d)", len);
+		PRINT_ERROR("Unable to alloc extended CDB (size %d)",
+			L1_CACHE_ALIGN(len));
 		goto out_err;
 	}
 
@@ -5325,10 +5326,10 @@ int scst_pre_init_cmd(struct scst_cmd *cmd, const uint8_t *cdb,
 			res = -EINVAL;
 			goto out;
 		}
-		cmd->cdb = kmalloc(cdb_len, gfp_mask);
+		cmd->cdb = kmalloc(L1_CACHE_ALIGN(cdb_len), gfp_mask);
 		if (unlikely(cmd->cdb == NULL)) {
 			PRINT_ERROR("Unable to alloc extended CDB (size %d)",
-				cdb_len);
+				L1_CACHE_ALIGN(cdb_len));
 			res = -ENOMEM;
 			goto out;
 		}
diff --git a/scst/src/scst_mem.c b/scst/src/scst_mem.c
index a6fa99faf..a2841553d 100644
--- a/scst/src/scst_mem.c
+++ b/scst/src/scst_mem.c
@@ -660,7 +660,7 @@ static int sgv_alloc_arrays(struct sgv_pool_obj *obj,
 
 	sz = pages_to_alloc * sizeof(obj->sg_entries[0]);
 
-	obj->sg_entries = kmalloc(sz, gfp_mask);
+	obj->sg_entries = kmalloc(L1_CACHE_ALIGN(sz), gfp_mask);
 	if (unlikely(obj->sg_entries == NULL)) {
 		TRACE(TRACE_OUT_OF_MEM, "Allocation of sgv_pool_obj "
 			"SG vector failed (size %d)", sz);
@@ -680,7 +680,7 @@ static int sgv_alloc_arrays(struct sgv_pool_obj *obj,
 			 */
 		} else {
 			tsz = pages_to_alloc * sizeof(obj->trans_tbl[0]);
-			obj->trans_tbl = kzalloc(tsz, gfp_mask);
+			obj->trans_tbl = kzalloc(L1_CACHE_ALIGN(tsz), gfp_mask);
 			if (unlikely(obj->trans_tbl == NULL)) {
 				TRACE(TRACE_OUT_OF_MEM, "Allocation of "
 					"trans_tbl failed (size %d)", tsz);
@@ -1021,7 +1021,7 @@ struct scatterlist *sgv_pool_alloc(struct sgv_pool *pool, unsigned int size,
 
 		sz = sizeof(*obj) + pages * sizeof(obj->sg_entries[0]);
 
-		obj = kmalloc(sz, gfp_mask);
+		obj = kmalloc(L1_CACHE_ALIGN(sz), gfp_mask);
 		if (unlikely(obj == NULL)) {
 			TRACE(TRACE_OUT_OF_MEM, "Allocation of "
 				"sgv_pool_obj failed (size %d)", size);
@@ -1374,7 +1374,7 @@ static void sgv_pool_init_cache(struct sgv_pool *pool, int cache_num)
 		"%s-%uK", pool->name, (pages << PAGE_SHIFT) >> 10);
 	pool->caches[cache_num] = kmem_cache_create(
 		pool->cache_names[cache_num], size,
-		__alignof__(struct sgv_pool_obj), SCST_SLAB_FLAGS, NULL
+		0, SCST_SLAB_FLAGS|SLAB_HWCACHE_ALIGN, NULL
 #if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 23))
 		, NULL);
 #else
@@ -1651,7 +1651,7 @@ struct sgv_pool *sgv_pool_create(const char *name,
 		}
 	}
 
-	pool = kzalloc(sizeof(*pool), GFP_KERNEL);
+	pool = kzalloc(L1_CACHE_ALIGN(sizeof(*pool)), GFP_KERNEL);
 	if (pool == NULL) {
 		PRINT_ERROR("Allocation of sgv_pool failed (size %zd)",
 			sizeof(*pool));
diff --git a/scst_local/scst_local.c b/scst_local/scst_local.c
index 1569f7df1..326214c12 100644
--- a/scst_local/scst_local.c
+++ b/scst_local/scst_local.c
@@ -1623,7 +1623,7 @@ static int __scst_local_add_adapter(struct scst_local_tgt *tgt,
 
 	TRACE_ENTRY();
 
-	sess = kzalloc(sizeof(*sess), GFP_KERNEL);
+	sess = kzalloc(L1_CACHE_ALIGN(sizeof(*sess)), GFP_KERNEL);
 	if (NULL == sess) {
 		PRINT_ERROR("Unable to alloc scst_lcl_host (size %zu)",
 			sizeof(*sess));