diff --git a/Makefile b/Makefile
index 8ebbb60f2078a..2213fe336705f 100644
--- a/Makefile
+++ b/Makefile
@@ -1,6 +1,6 @@
 VERSION = 4
 PATCHLEVEL = 9
-SUBLEVEL = 251
+SUBLEVEL = 252
 EXTRAVERSION =
 NAME = Roaring Lionus
 
diff --git a/arch/arm/mach-omap2/omap_device.c b/arch/arm/mach-omap2/omap_device.c
index f989145480c8f..bf236b7af8c1a 100644
--- a/arch/arm/mach-omap2/omap_device.c
+++ b/arch/arm/mach-omap2/omap_device.c
@@ -224,10 +224,12 @@ static int _omap_device_notifier_call(struct notifier_block *nb,
 		break;
 	case BUS_NOTIFY_BIND_DRIVER:
 		od = to_omap_device(pdev);
-		if (od && (od->_state == OMAP_DEVICE_STATE_ENABLED) &&
-		    pm_runtime_status_suspended(dev)) {
+		if (od) {
 			od->_driver_status = BUS_NOTIFY_BIND_DRIVER;
-			pm_runtime_set_active(dev);
+			if (od->_state == OMAP_DEVICE_STATE_ENABLED &&
+			    pm_runtime_status_suspended(dev)) {
+				pm_runtime_set_active(dev);
+			}
 		}
 		break;
 	case BUS_NOTIFY_ADD_DEVICE:
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 0ad81fa13688f..10d80456f38f1 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -450,6 +450,10 @@ static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
 {
 	u64 pmcr, val;
 
+	/* No PMU available, PMCR_EL0 may UNDEF... */
+	if (!kvm_arm_support_pmu_v3())
+		return;
+
 	pmcr = read_sysreg(pmcr_el0);
 	/*
 	 * Writable bits of PMCR_EL0 (ARMV8_PMU_PMCR_MASK) are reset to UNKNOWN
diff --git a/arch/powerpc/include/asm/book3s/32/pgtable.h b/arch/powerpc/include/asm/book3s/32/pgtable.h
index 6b8b2d57fdc8c..e588028922a83 100644
--- a/arch/powerpc/include/asm/book3s/32/pgtable.h
+++ b/arch/powerpc/include/asm/book3s/32/pgtable.h
@@ -411,9 +411,9 @@ static inline void __set_pte_at(struct mm_struct *mm, unsigned long addr,
 	if (pte_val(*ptep) & _PAGE_HASHPTE)
 		flush_hash_entry(mm, ptep, addr);
 	__asm__ __volatile__("\
-		stw%U0%X0 %2,%0\n\
+		stw%X0 %2,%0\n\
 		eieio\n\
-		stw%U0%X0 %L2,%1"
+		stw%X1 %L2,%1"
 	: "=m" (*ptep), "=m" (*((unsigned char *)ptep+4))
 	: "r" (pte) : "memory");
 
diff --git a/arch/powerpc/include/asm/nohash/pgtable.h b/arch/powerpc/include/asm/nohash/pgtable.h
index 1263c22d60d85..330fe178c0c5e 100644
--- a/arch/powerpc/include/asm/nohash/pgtable.h
+++ b/arch/powerpc/include/asm/nohash/pgtable.h
@@ -155,9 +155,9 @@ static inline void __set_pte_at(struct mm_struct *mm, unsigned long addr,
 		flush_hash_entry(mm, ptep, addr);
 #endif
 	__asm__ __volatile__("\
-		stw%U0%X0 %2,%0\n\
+		stw%X0 %2,%0\n\
 		eieio\n\
-		stw%U0%X0 %L2,%1"
+		stw%X1 %L2,%1"
 	: "=m" (*ptep), "=m" (*((unsigned char *)ptep+4))
 	: "r" (pte) : "memory");
 
diff --git a/block/genhd.c b/block/genhd.c
index fcd6d4fae657c..9c1adfd768d2c 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -159,14 +159,17 @@ struct hd_struct *disk_part_iter_next(struct disk_part_iter *piter)
 		part = rcu_dereference(ptbl->part[piter->idx]);
 		if (!part)
 			continue;
+		get_device(part_to_dev(part));
+		piter->part = part;
 		if (!part_nr_sects_read(part) &&
 		    !(piter->flags & DISK_PITER_INCL_EMPTY) &&
 		    !(piter->flags & DISK_PITER_INCL_EMPTY_PART0 &&
-		      piter->idx == 0))
+		      piter->idx == 0)) {
+			put_device(part_to_dev(part));
+			piter->part = NULL;
 			continue;
+		}
 
-		get_device(part_to_dev(part));
-		piter->part = part;
 		piter->idx += inc;
 		break;
 	}
diff --git a/drivers/block/Kconfig b/drivers/block/Kconfig
index 39dd30b6ef86e..894102fd5a069 100644
--- a/drivers/block/Kconfig
+++ b/drivers/block/Kconfig
@@ -530,6 +530,7 @@ config BLK_DEV_RBD
 config BLK_DEV_RSXX
 	tristate "IBM Flash Adapter 900GB Full Height PCIe Device Driver"
 	depends on PCI
+	select CRC32
 	help
 	  Device driver for IBM's high speed PCIe SSD
 	  storage device: Flash Adapter 900GB Full Height.
diff --git a/drivers/cpufreq/powernow-k8.c b/drivers/cpufreq/powernow-k8.c
index 0b5bf135b0907..59f16807921ad 100644
--- a/drivers/cpufreq/powernow-k8.c
+++ b/drivers/cpufreq/powernow-k8.c
@@ -887,9 +887,9 @@ static int get_transition_latency(struct powernow_k8_data *data)
 
 /* Take a frequency, and issue the fid/vid transition command */
 static int transition_frequency_fidvid(struct powernow_k8_data *data,
-		unsigned int index)
+		unsigned int index,
+		struct cpufreq_policy *policy)
 {
-	struct cpufreq_policy *policy;
 	u32 fid = 0;
 	u32 vid = 0;
 	int res;
@@ -921,9 +921,6 @@ static int transition_frequency_fidvid(struct powernow_k8_data *data,
 	freqs.old = find_khz_freq_from_fid(data->currfid);
 	freqs.new = find_khz_freq_from_fid(fid);
 
-	policy = cpufreq_cpu_get(smp_processor_id());
-	cpufreq_cpu_put(policy);
-
 	cpufreq_freq_transition_begin(policy, &freqs);
 	res = transition_fid_vid(data, fid, vid);
 	cpufreq_freq_transition_end(policy, &freqs, res);
@@ -978,7 +975,7 @@ static long powernowk8_target_fn(void *arg)
 
 	powernow_k8_acpi_pst_values(data, newstate);
 
-	ret = transition_frequency_fidvid(data, newstate);
+	ret = transition_frequency_fidvid(data, newstate, pol);
 
 	if (ret) {
 		pr_err("transition frequency failed\n");
diff --git a/drivers/dma/xilinx/xilinx_dma.c b/drivers/dma/xilinx/xilinx_dma.c
index ef99ef0bb1ca2..f00652585ee31 100644
--- a/drivers/dma/xilinx/xilinx_dma.c
+++ b/drivers/dma/xilinx/xilinx_dma.c
@@ -2357,7 +2357,7 @@ static int xilinx_dma_chan_probe(struct xilinx_dma_device *xdev,
 		has_dre = false;
 
 	if (!has_dre)
-		xdev->common.copy_align = fls(width - 1);
+		xdev->common.copy_align = (enum dmaengine_alignment)fls(width - 1);
 
 	if (of_device_is_compatible(node, "xlnx,axi-vdma-mm2s-channel") ||
 	    of_device_is_compatible(node, "xlnx,axi-dma-mm2s-channel") ||
@@ -2630,7 +2630,11 @@ static int xilinx_dma_probe(struct platform_device *pdev)
 	}
 
 	/* Register the DMA engine with the core */
-	dma_async_device_register(&xdev->common);
+	err = dma_async_device_register(&xdev->common);
+	if (err) {
+		dev_err(xdev->dev, "failed to register the dma device\n");
+		goto error;
+	}
 
 	err = of_dma_controller_register(node, of_dma_xilinx_xlate,
 					 xdev);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 4548d89abcdc3..ff8168c60b35a 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -882,7 +882,7 @@ eb_vma_misplaced(struct i915_vma *vma)
 		return !only_mappable_for_reloc(entry->flags);
 
 	if ((entry->flags & EXEC_OBJECT_SUPPORTS_48B_ADDRESS) == 0 &&
-	    (vma->node.start + vma->node.size - 1) >> 32)
+	    (vma->node.start + vma->node.size + 4095) >> 32)
 		return true;
 
 	return false;
diff --git a/drivers/iommu/intel_irq_remapping.c b/drivers/iommu/intel_irq_remapping.c
index ce125ec23d2a5..88ba1a65c2830 100644
--- a/drivers/iommu/intel_irq_remapping.c
+++ b/drivers/iommu/intel_irq_remapping.c
@@ -1350,6 +1350,8 @@ static int intel_irq_remapping_alloc(struct irq_domain *domain,
 		irq_data = irq_domain_get_irq_data(domain, virq + i);
 		irq_cfg = irqd_cfg(irq_data);
 		if (!irq_data || !irq_cfg) {
+			if (!i)
+				kfree(data);
 			ret = -EINVAL;
 			goto out_free_data;
 		}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
index 8cd7227fbdfce..3dd0bc8804c1a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
@@ -930,6 +930,7 @@ err_destroy_groups:
 	ft->g[ft->num_groups] = NULL;
 	mlx5e_destroy_groups(ft);
 	kvfree(in);
+	kfree(ft->g);
 
 	return err;
 }
diff --git a/drivers/net/wan/Kconfig b/drivers/net/wan/Kconfig
index 4e9fe75d70675..069f933b0add2 100644
--- a/drivers/net/wan/Kconfig
+++ b/drivers/net/wan/Kconfig
@@ -295,6 +295,7 @@ config SLIC_DS26522
 	tristate "Slic Maxim ds26522 card support"
 	depends on SPI
 	depends on FSL_SOC || ARCH_MXC || ARCH_LAYERSCAPE || COMPILE_TEST
+	select BITREVERSE
 	help
 	  This module initializes and configures the slic maxim card
 	  in T1 or E1 mode.
diff --git a/drivers/net/wireless/ath/wil6210/Kconfig b/drivers/net/wireless/ath/wil6210/Kconfig
index 6dfedc8bd6a3d..7df13a684d2df 100644
--- a/drivers/net/wireless/ath/wil6210/Kconfig
+++ b/drivers/net/wireless/ath/wil6210/Kconfig
@@ -1,6 +1,7 @@
 config WIL6210
 	tristate "Wilocity 60g WiFi card wil6210 support"
 	select WANT_DEV_COREDUMP
+	select CRC32
 	depends on CFG80211
 	depends on PCI
 	default n
diff --git a/drivers/spi/spi-pxa2xx.c b/drivers/spi/spi-pxa2xx.c
index da3834fe5e570..9bb6a574ab2fe 100644
--- a/drivers/spi/spi-pxa2xx.c
+++ b/drivers/spi/spi-pxa2xx.c
@@ -1606,7 +1606,7 @@ static int pxa2xx_spi_probe(struct platform_device *pdev)
 		return -ENODEV;
 	}
 
-	master = spi_alloc_master(dev, sizeof(struct driver_data));
+	master = devm_spi_alloc_master(dev, sizeof(*drv_data));
 	if (!master) {
 		dev_err(&pdev->dev, "cannot alloc spi_master\n");
 		pxa_ssp_free(ssp);
@@ -1788,7 +1788,6 @@ out_error_clock_enabled:
 	free_irq(ssp->irq, drv_data);
 
 out_error_master_alloc:
-	spi_master_put(master);
 	pxa_ssp_free(ssp);
 	return status;
 }
diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c
index e738b4621cbba..ecd707f74ddcb 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -1736,6 +1736,10 @@ void transport_generic_request_failure(struct se_cmd *cmd,
 	case TCM_LOGICAL_BLOCK_APP_TAG_CHECK_FAILED:
 	case TCM_LOGICAL_BLOCK_REF_TAG_CHECK_FAILED:
 	case TCM_COPY_TARGET_DEVICE_NOT_REACHABLE:
+	case TCM_TOO_MANY_TARGET_DESCS:
+	case TCM_UNSUPPORTED_TARGET_DESC_TYPE_CODE:
+	case TCM_TOO_MANY_SEGMENT_DESCS:
+	case TCM_UNSUPPORTED_SEGMENT_DESC_TYPE_CODE:
 		break;
 	case TCM_OUT_OF_RESOURCES:
 		sense_reason = TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE;
@@ -2886,6 +2890,26 @@ static const struct sense_info sense_info_table[] = {
 		.key = ILLEGAL_REQUEST,
 		.asc = 0x26, /* INVALID FIELD IN PARAMETER LIST */
 	},
+	[TCM_TOO_MANY_TARGET_DESCS] = {
+		.key = ILLEGAL_REQUEST,
+		.asc = 0x26,
+		.ascq = 0x06, /* TOO MANY TARGET DESCRIPTORS */
+	},
+	[TCM_UNSUPPORTED_TARGET_DESC_TYPE_CODE] = {
+		.key = ILLEGAL_REQUEST,
+		.asc = 0x26,
+		.ascq = 0x07, /* UNSUPPORTED TARGET DESCRIPTOR TYPE CODE */
+	},
+	[TCM_TOO_MANY_SEGMENT_DESCS] = {
+		.key = ILLEGAL_REQUEST,
+		.asc = 0x26,
+		.ascq = 0x08, /* TOO MANY SEGMENT DESCRIPTORS */
+	},
+	[TCM_UNSUPPORTED_SEGMENT_DESC_TYPE_CODE] = {
+		.key = ILLEGAL_REQUEST,
+		.asc = 0x26,
+		.ascq = 0x09, /* UNSUPPORTED SEGMENT DESCRIPTOR TYPE CODE */
+	},
 	[TCM_PARAMETER_LIST_LENGTH_ERROR] = {
 		.key = ILLEGAL_REQUEST,
 		.asc = 0x1a, /* PARAMETER LIST LENGTH ERROR */
diff --git a/drivers/target/target_core_xcopy.c b/drivers/target/target_core_xcopy.c
index 18848ba8d2ba0..84e3bf1132fd5 100644
--- a/drivers/target/target_core_xcopy.c
+++ b/drivers/target/target_core_xcopy.c
@@ -52,64 +52,87 @@ static int target_xcopy_gen_naa_ieee(struct se_device *dev, unsigned char *buf)
 	return 0;
 }
 
-static int target_xcopy_locate_se_dev_e4(struct se_cmd *se_cmd, struct xcopy_op *xop,
-					bool src)
+/**
+ * target_xcopy_locate_se_dev_e4_iter - compare XCOPY NAA device identifiers
+ *
+ * @se_dev: device being considered for match
+ * @dev_wwn: XCOPY requested NAA dev_wwn
+ * @return: 1 on match, 0 on no-match
+ */
+static int target_xcopy_locate_se_dev_e4_iter(struct se_device *se_dev,
+					      const unsigned char *dev_wwn)
 {
-	struct se_device *se_dev;
-	unsigned char tmp_dev_wwn[XCOPY_NAA_IEEE_REGEX_LEN], *dev_wwn;
+	unsigned char tmp_dev_wwn[XCOPY_NAA_IEEE_REGEX_LEN];
 	int rc;
 
-	if (src)
-		dev_wwn = &xop->dst_tid_wwn[0];
-	else
-		dev_wwn = &xop->src_tid_wwn[0];
-
-	mutex_lock(&g_device_mutex);
-	list_for_each_entry(se_dev, &g_device_list, g_dev_node) {
-
-		if (!se_dev->dev_attrib.emulate_3pc)
-			continue;
+	if (!se_dev->dev_attrib.emulate_3pc) {
+		pr_debug("XCOPY: emulate_3pc disabled on se_dev %p\n", se_dev);
+		return 0;
+	}
 
-		memset(&tmp_dev_wwn[0], 0, XCOPY_NAA_IEEE_REGEX_LEN);
-		target_xcopy_gen_naa_ieee(se_dev, &tmp_dev_wwn[0]);
+	memset(&tmp_dev_wwn[0], 0, XCOPY_NAA_IEEE_REGEX_LEN);
+	target_xcopy_gen_naa_ieee(se_dev, &tmp_dev_wwn[0]);
 
-		rc = memcmp(&tmp_dev_wwn[0], dev_wwn, XCOPY_NAA_IEEE_REGEX_LEN);
-		if (rc != 0)
-			continue;
+	rc = memcmp(&tmp_dev_wwn[0], dev_wwn, XCOPY_NAA_IEEE_REGEX_LEN);
+	if (rc != 0) {
+		pr_debug("XCOPY: skip non-matching: %*ph\n",
+			 XCOPY_NAA_IEEE_REGEX_LEN, tmp_dev_wwn);
+		return 0;
+	}
+	pr_debug("XCOPY 0xe4: located se_dev: %p\n", se_dev);
 
-		if (src) {
-			xop->dst_dev = se_dev;
-			pr_debug("XCOPY 0xe4: Setting xop->dst_dev: %p from located"
-				" se_dev\n", xop->dst_dev);
-		} else {
-			xop->src_dev = se_dev;
-			pr_debug("XCOPY 0xe4: Setting xop->src_dev: %p from located"
-				" se_dev\n", xop->src_dev);
-		}
+	return 1;
+}
 
-		rc = target_depend_item(&se_dev->dev_group.cg_item);
-		if (rc != 0) {
-			pr_err("configfs_depend_item attempt failed:"
-				" %d for se_dev: %p\n", rc, se_dev);
-			mutex_unlock(&g_device_mutex);
-			return rc;
+static int target_xcopy_locate_se_dev_e4(struct se_session *sess,
+					const unsigned char *dev_wwn,
+					struct se_device **_found_dev,
+					struct percpu_ref **_found_lun_ref)
+{
+	struct se_dev_entry *deve;
+	struct se_node_acl *nacl;
+	struct se_lun *this_lun = NULL;
+	struct se_device *found_dev = NULL;
+
+	/* cmd with NULL sess indicates no associated $FABRIC_MOD */
+	if (!sess)
+		goto err_out;
+
+	pr_debug("XCOPY 0xe4: searching for: %*ph\n",
+		 XCOPY_NAA_IEEE_REGEX_LEN, dev_wwn);
+
+	nacl = sess->se_node_acl;
+	rcu_read_lock();
+	hlist_for_each_entry_rcu(deve, &nacl->lun_entry_hlist, link) {
+		struct se_device *this_dev;
+		int rc;
+
+		this_lun = rcu_dereference(deve->se_lun);
+		this_dev = rcu_dereference_raw(this_lun->lun_se_dev);
+
+		rc = target_xcopy_locate_se_dev_e4_iter(this_dev, dev_wwn);
+		if (rc) {
+			if (percpu_ref_tryget_live(&this_lun->lun_ref))
+				found_dev = this_dev;
+			break;
 		}
-
-		pr_debug("Called configfs_depend_item for se_dev: %p"
-			" se_dev->se_dev_group: %p\n", se_dev,
-			&se_dev->dev_group);
-
-		mutex_unlock(&g_device_mutex);
-		return 0;
 	}
-	mutex_unlock(&g_device_mutex);
-
+	rcu_read_unlock();
+	if (found_dev == NULL)
+		goto err_out;
+
+	pr_debug("lun_ref held for se_dev: %p se_dev->se_dev_group: %p\n",
+		 found_dev, &found_dev->dev_group);
+	*_found_dev = found_dev;
+	*_found_lun_ref = &this_lun->lun_ref;
+	return 0;
+err_out:
 	pr_debug_ratelimited("Unable to locate 0xe4 descriptor for EXTENDED_COPY\n");
 	return -EINVAL;
 }
 
 static int target_xcopy_parse_tiddesc_e4(struct se_cmd *se_cmd, struct xcopy_op *xop,
-				unsigned char *p, bool src)
+				unsigned char *p, unsigned short cscd_index)
 {
 	unsigned char *desc = p;
 	unsigned short ript;
@@ -154,7 +177,13 @@ static int target_xcopy_parse_tiddesc_e4(struct se_cmd *se_cmd, struct xcopy_op
 		return -EINVAL;
 	}
 
-	if (src) {
+	if (cscd_index != xop->stdi && cscd_index != xop->dtdi) {
+		pr_debug("XCOPY 0xe4: ignoring CSCD entry %d - neither src nor "
+			 "dest\n", cscd_index);
+		return 0;
+	}
+
+	if (cscd_index == xop->stdi) {
 		memcpy(&xop->src_tid_wwn[0], &desc[8], XCOPY_NAA_IEEE_REGEX_LEN);
 		/*
 		 * Determine if the source designator matches the local device
@@ -166,10 +195,15 @@ static int target_xcopy_parse_tiddesc_e4(struct se_cmd *se_cmd, struct xcopy_op
 			pr_debug("XCOPY 0xe4: Set xop->src_dev %p from source"
 					" received xop\n", xop->src_dev);
 		}
-	} else {
+	}
+
+	if (cscd_index == xop->dtdi) {
 		memcpy(&xop->dst_tid_wwn[0], &desc[8], XCOPY_NAA_IEEE_REGEX_LEN);
 		/*
-		 * Determine if the destination designator matches the local device
+		 * Determine if the destination designator matches the local
+		 * device. If @cscd_index corresponds to both source (stdi) and
+		 * destination (dtdi), or dtdi comes after stdi, then
+		 * XCOL_DEST_RECV_OP wins.
 		 */
 		if (!memcmp(&xop->local_dev_wwn[0], &xop->dst_tid_wwn[0],
 				XCOPY_NAA_IEEE_REGEX_LEN)) {
@@ -189,9 +223,9 @@ static int target_xcopy_parse_target_descriptors(struct se_cmd *se_cmd,
 {
 	struct se_device *local_dev = se_cmd->se_dev;
 	unsigned char *desc = p;
-	int offset = tdll % XCOPY_TARGET_DESC_LEN, rc, ret = 0;
+	int offset = tdll % XCOPY_TARGET_DESC_LEN, rc;
+	unsigned short cscd_index = 0;
 	unsigned short start = 0;
-	bool src = true;
 
 	*sense_ret = TCM_INVALID_PARAMETER_LIST;
 
@@ -214,25 +248,19 @@ static int target_xcopy_parse_target_descriptors(struct se_cmd *se_cmd,
 
 	while (start < tdll) {
 		/*
-		 * Check target descriptor identification with 0xE4 type with
-		 * use VPD 0x83 WWPN matching ..
+		 * Check target descriptor identification with 0xE4 type, and
+		 * compare the current index with the CSCD descriptor IDs in
+		 * the segment descriptor. Use VPD 0x83 WWPN matching ..
 		 */
 		switch (desc[0]) {
 		case 0xe4:
 			rc = target_xcopy_parse_tiddesc_e4(se_cmd, xop,
-							&desc[0], src);
+							&desc[0], cscd_index);
 			if (rc != 0)
 				goto out;
-			/*
-			 * Assume target descriptors are in source -> destination order..
-			 */
-			if (src)
-				src = false;
-			else
-				src = true;
 			start += XCOPY_TARGET_DESC_LEN;
 			desc += XCOPY_TARGET_DESC_LEN;
-			ret++;
+			cscd_index++;
 			break;
 		default:
 			pr_err("XCOPY unsupported descriptor type code:"
@@ -241,10 +269,25 @@ static int target_xcopy_parse_target_descriptors(struct se_cmd *se_cmd,
 		}
 	}
 
-	if (xop->op_origin == XCOL_SOURCE_RECV_OP)
-		rc = target_xcopy_locate_se_dev_e4(se_cmd, xop, true);
-	else
-		rc = target_xcopy_locate_se_dev_e4(se_cmd, xop, false);
+	switch (xop->op_origin) {
+	case XCOL_SOURCE_RECV_OP:
+		rc = target_xcopy_locate_se_dev_e4(se_cmd->se_sess,
+						xop->dst_tid_wwn,
+						&xop->dst_dev,
+						&xop->remote_lun_ref);
+		break;
+	case XCOL_DEST_RECV_OP:
+		rc = target_xcopy_locate_se_dev_e4(se_cmd->se_sess,
+						xop->src_tid_wwn,
+						&xop->src_dev,
+						&xop->remote_lun_ref);
+		break;
+	default:
+		pr_err("XCOPY CSCD descriptor IDs not found in CSCD list - "
+			"stdi: %hu dtdi: %hu\n", xop->stdi, xop->dtdi);
+		rc = -EINVAL;
+		break;
+	}
 	/*
 	 * If a matching IEEE NAA 0x83 descriptor for the requested device
 	 * is not located on this node, return COPY_ABORTED with ASQ/ASQC
@@ -261,7 +304,7 @@ static int target_xcopy_parse_target_descriptors(struct se_cmd *se_cmd,
 	pr_debug("XCOPY TGT desc: Dest dev: %p NAA IEEE WWN: 0x%16phN\n",
 		 xop->dst_dev, &xop->dst_tid_wwn[0]);
 
-	return ret;
+	return cscd_index;
 
 out:
 	return -EINVAL;
@@ -305,17 +348,26 @@ static int target_xcopy_parse_segdesc_02(struct se_cmd *se_cmd, struct xcopy_op
 
 static int target_xcopy_parse_segment_descriptors(struct se_cmd *se_cmd,
 				struct xcopy_op *xop, unsigned char *p,
-				unsigned int sdll)
+				unsigned int sdll, sense_reason_t *sense_ret)
 {
 	unsigned char *desc = p;
 	unsigned int start = 0;
 	int offset = sdll % XCOPY_SEGMENT_DESC_LEN, rc, ret = 0;
 
+	*sense_ret = TCM_INVALID_PARAMETER_LIST;
+
 	if (offset != 0) {
 		pr_err("XCOPY segment descriptor list length is not"
 			" multiple of %d\n", XCOPY_SEGMENT_DESC_LEN);
 		return -EINVAL;
 	}
+	if (sdll > RCR_OP_MAX_SG_DESC_COUNT * XCOPY_SEGMENT_DESC_LEN) {
+		pr_err("XCOPY supports %u segment descriptor(s), sdll: %u too"
+			" large..\n", RCR_OP_MAX_SG_DESC_COUNT, sdll);
+		/* spc4r37 6.4.3.5 SEGMENT DESCRIPTOR LIST LENGTH field */
+		*sense_ret = TCM_TOO_MANY_SEGMENT_DESCS;
+		return -EINVAL;
+	}
 
 	while (start < sdll) {
 		/*
@@ -372,18 +424,12 @@ static int xcopy_pt_get_cmd_state(struct se_cmd *se_cmd)
 
 static void xcopy_pt_undepend_remotedev(struct xcopy_op *xop)
 {
-	struct se_device *remote_dev;
-
 	if (xop->op_origin == XCOL_SOURCE_RECV_OP)
-		remote_dev = xop->dst_dev;
+		pr_debug("putting dst lun_ref for %p\n", xop->dst_dev);
 	else
-		remote_dev = xop->src_dev;
+		pr_debug("putting src lun_ref for %p\n", xop->src_dev);
 
-	pr_debug("Calling configfs_undepend_item for"
-		  " remote_dev: %p remote_dev->dev_group: %p\n",
-		  remote_dev, &remote_dev->dev_group.cg_item);
-
-	target_undepend_item(&remote_dev->dev_group.cg_item);
+	percpu_ref_put(xop->remote_lun_ref);
 }
 
 static void xcopy_pt_release_cmd(struct se_cmd *se_cmd)
@@ -893,6 +939,20 @@ sense_reason_t target_do_xcopy(struct se_cmd *se_cmd)
 		" tdll: %hu sdll: %u inline_dl: %u\n", list_id, list_id_usage,
 		tdll, sdll, inline_dl);
 
+	/*
+	 * skip over the target descriptors until segment descriptors
+	 * have been passed - CSCD ids are needed to determine src and dest.
+	 */
+	seg_desc = &p[16] + tdll;
+
+	rc = target_xcopy_parse_segment_descriptors(se_cmd, xop, seg_desc,
+						    sdll, &ret);
+	if (rc <= 0)
+		goto out;
+
+	pr_debug("XCOPY: Processed %d segment descriptors, length: %u\n", rc,
+				rc * XCOPY_SEGMENT_DESC_LEN);
+
 	rc = target_xcopy_parse_target_descriptors(se_cmd, xop, &p[16], tdll, &ret);
 	if (rc <= 0)
 		goto out;
@@ -910,18 +970,8 @@ sense_reason_t target_do_xcopy(struct se_cmd *se_cmd)
 
 	pr_debug("XCOPY: Processed %d target descriptors, length: %u\n", rc,
 				rc * XCOPY_TARGET_DESC_LEN);
-	seg_desc = &p[16];
-	seg_desc += (rc * XCOPY_TARGET_DESC_LEN);
-
-	rc = target_xcopy_parse_segment_descriptors(se_cmd, xop, seg_desc, sdll);
-	if (rc <= 0) {
-		xcopy_pt_undepend_remotedev(xop);
-		goto out;
-	}
 	transport_kunmap_data_sg(se_cmd);
 
-	pr_debug("XCOPY: Processed %d segment descriptors, length: %u\n", rc,
-				rc * XCOPY_SEGMENT_DESC_LEN);
 	INIT_WORK(&xop->xop_work, target_xcopy_do_work);
 	queue_work(xcopy_wq, &xop->xop_work);
 	return TCM_NO_SENSE;
diff --git a/drivers/target/target_core_xcopy.h b/drivers/target/target_core_xcopy.h
index 700a981c7b415..7db8d0c9223f8 100644
--- a/drivers/target/target_core_xcopy.h
+++ b/drivers/target/target_core_xcopy.h
@@ -19,6 +19,7 @@ struct xcopy_op {
 	struct se_device *dst_dev;
 	unsigned char dst_tid_wwn[XCOPY_NAA_IEEE_REGEX_LEN];
 	unsigned char local_dev_wwn[XCOPY_NAA_IEEE_REGEX_LEN];
+	struct percpu_ref *remote_lun_ref;
 
 	sector_t src_lba;
 	sector_t dst_lba;
diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
index 9213a9e046ae0..99caaae01caba 100644
--- a/fs/ubifs/io.c
+++ b/fs/ubifs/io.c
@@ -331,7 +331,7 @@ void ubifs_pad(const struct ubifs_info *c, void *buf, int pad)
 {
 	uint32_t crc;
 
-	ubifs_assert(pad >= 0 && !(pad & 7));
+	ubifs_assert(pad >= 0);
 
 	if (pad >= UBIFS_PAD_NODE_SZ) {
 		struct ubifs_ch *ch = buf;
@@ -721,6 +721,10 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 		 * write-buffer.
 		 */
 		memcpy(wbuf->buf + wbuf->used, buf, len);
+		if (aligned_len > len) {
+			ubifs_assert(aligned_len - len < 8);
+			ubifs_pad(c, wbuf->buf + wbuf->used + len, aligned_len - len);
+		}
 
 		if (aligned_len == wbuf->avail) {
 			dbg_io("flush jhead %s wbuf to LEB %d:%d",
@@ -813,13 +817,18 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 	}
 
 	spin_lock(&wbuf->lock);
-	if (aligned_len)
+	if (aligned_len) {
 		/*
 		 * And now we have what's left and what does not take whole
 		 * max. write unit, so write it to the write-buffer and we are
 		 * done.
 		 */
 		memcpy(wbuf->buf, buf + written, len);
+		if (aligned_len > len) {
+			ubifs_assert(aligned_len - len < 8);
+			ubifs_pad(c, wbuf->buf + len, aligned_len - len);
+		}
+	}
 
 	if (c->leb_size - wbuf->offs >= c->max_write_size)
 		wbuf->size = c->max_write_size;
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index 4fdb1d9848444..36198563fb8bc 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -460,7 +460,10 @@
  */
 #define TEXT_TEXT							\
 		ALIGN_FUNCTION();					\
-		*(.text.hot TEXT_MAIN .text.fixup .text.unlikely)	\
+		*(.text.hot .text.hot.*)				\
+		*(TEXT_MAIN .text.fixup)				\
+		*(.text.unlikely .text.unlikely.*)			\
+		*(.text.unknown .text.unknown.*)			\
 		*(.text..ftrace)					\
 		*(TEXT_CFI_MAIN)					\
 		*(.ref.text)						\
diff --git a/include/target/target_core_base.h b/include/target/target_core_base.h
index 30f99ce4c6cea..8a70d38f13329 100644
--- a/include/target/target_core_base.h
+++ b/include/target/target_core_base.h
@@ -178,6 +178,10 @@ enum tcm_sense_reason_table {
 	TCM_LOGICAL_BLOCK_APP_TAG_CHECK_FAILED	= R(0x16),
 	TCM_LOGICAL_BLOCK_REF_TAG_CHECK_FAILED	= R(0x17),
 	TCM_COPY_TARGET_DEVICE_NOT_REACHABLE	= R(0x18),
+	TCM_TOO_MANY_TARGET_DESCS		= R(0x19),
+	TCM_UNSUPPORTED_TARGET_DESC_TYPE_CODE	= R(0x1a),
+	TCM_TOO_MANY_SEGMENT_DESCS		= R(0x1b),
+	TCM_UNSUPPORTED_SEGMENT_DESC_TYPE_CODE	= R(0x1c),
 #undef R
 };
 
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index a4c4234976862..026f4525063c1 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1592,6 +1592,12 @@ int pskb_trim_rcsum_slow(struct sk_buff *skb, unsigned int len)
 		skb->csum = csum_block_sub(skb->csum,
 					   skb_checksum(skb, len, delta, 0),
 					   len);
+	} else if (skb->ip_summed == CHECKSUM_PARTIAL) {
+		int hdlen = (len > skb_headlen(skb)) ? skb_headlen(skb) : len;
+		int offset = skb_checksum_start_offset(skb) + skb->csum_offset;
+
+		if (offset + sizeof(__sum16) > hdlen)
+			return -EINVAL;
 	}
 	return __pskb_trim(skb, len);
 }
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index c37e9598262e5..3164bae4024a4 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -300,7 +300,7 @@ static int ip_finish_output(struct net *net, struct sock *sk, struct sk_buff *sk
 	if (skb_is_gso(skb))
 		return ip_finish_output_gso(net, sk, skb, mtu);
 
-	if (skb->len > mtu || (IPCB(skb)->flags & IPSKB_FRAG_PMTU))
+	if (skb->len > mtu || IPCB(skb)->frag_max_size)
 		return ip_fragment(net, sk, skb, mtu, ip_finish_output2);
 
 	return ip_finish_output2(net, sk, skb);
diff --git a/net/ipv4/ip_tunnel.c b/net/ipv4/ip_tunnel.c
index 5f2e3334cccec..26e1dbc958189 100644
--- a/net/ipv4/ip_tunnel.c
+++ b/net/ipv4/ip_tunnel.c
@@ -743,7 +743,11 @@ void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
 		goto tx_error;
 	}
 
-	if (tnl_update_pmtu(dev, skb, rt, tnl_params->frag_off, inner_iph)) {
+	df = tnl_params->frag_off;
+	if (skb->protocol == htons(ETH_P_IP) && !tunnel->ignore_df)
+		df |= (inner_iph->frag_off & htons(IP_DF));
+
+	if (tnl_update_pmtu(dev, skb, rt, df, inner_iph)) {
 		ip_rt_put(rt);
 		goto tx_error;
 	}
@@ -771,10 +775,6 @@ void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
 			ttl = ip4_dst_hoplimit(&rt->dst);
 	}
 
-	df = tnl_params->frag_off;
-	if (skb->protocol == htons(ETH_P_IP) && !tunnel->ignore_df)
-		df |= (inner_iph->frag_off&htons(IP_DF));
-
 	max_headroom = LL_RESERVED_SPACE(rt->dst.dev) + sizeof(struct iphdr)
 			+ rt->dst.header_len + ip_encap_hlen(&tunnel->encap);
 	if (max_headroom > dev->needed_headroom)