Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending

Pull SCSI target updates from Nicholas Bellinger:
 "It's been usually busy for summer, with most of the efforts centered
  around TCMU developments and various target-core + fabric driver bug
  fixing activities. Not particularly large in terms of LoC, but lots of
  smaller patches from many different folks.

  The highlights include:

   - ibmvscsis logical partition manager support (Michael Cyr + Bryant
     Ly)

   - Convert target/iblock WRITE_SAME to blkdev_issue_zeroout (hch +
     nab)

   - Add support for TMR percpu LUN reference counting (nab)

   - Fix a potential deadlock between EXTENDED_COPY and iscsi shutdown
     (Bart)

   - Fix COMPARE_AND_WRITE caw_sem leak during se_cmd quiesce (Jiang Yi)

   - Fix TMCU module removal (Xiubo Li)

   - Fix iser-target OOPs during login failure (Andrea Righi + Sagi)

   - Breakup target-core free_device backend driver callback (mnc)

   - Perform TCMU add/delete/reconfig synchronously (mnc)

   - Fix TCMU multiple UIO open/close sequences (mnc)

   - Fix TCMU CHECK_CONDITION sense handling (mnc)

   - Fix target-core SAM_STAT_BUSY + TASK_SET_FULL handling (mnc + nab)

   - Introduce TYPE_ZBC support in PSCSI (Damien Le Moal)

   - Fix possible TCMU memory leak + OOPs when recalculating cmd base
     size (Xiubo Li + Bryant Ly + Damien Le Moal + mnc)

   - Add login_keys_workaround attribute for non RFC initiators (Robert
     LeBlanc + Arun Easi + nab)"

* 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending: (68 commits)
  iscsi-target: Add login_keys_workaround attribute for non RFC initiators
  Revert "qla2xxx: Fix incorrect tcm_qla2xxx_free_cmd use during TMR ABORT"
  tcmu: clean up the code and with one small fix
  tcmu: Fix possbile memory leak / OOPs when recalculating cmd base size
  target: export lio pgr/alua support as device attr
  target: Fix return sense reason in target_scsi3_emulate_pr_out
  target: Fix cmd size for PR-OUT in passthrough_parse_cdb
  tcmu: Fix dev_config_store
  target: pscsi: Introduce TYPE_ZBC support
  target: Use macro for WRITE_VERIFY_32 operation codes
  target: fix SAM_STAT_BUSY/TASK_SET_FULL handling
  target: remove transport_complete
  pscsi: finish cmd processing from pscsi_req_done
  tcmu: fix sense handling during completion
  target: add helper to copy sense to se_cmd buffer
  target: do not require a transport_complete for SCF_TRANSPORT_TASK_SENSE
  target: make device_mutex and device_list static
  tcmu: Fix flushing cmd entry dcache page
  tcmu: fix multiple uio open/close sequences
  tcmu: drop configured check in destroy
  ...

1 files changed, 118 insertions, 27 deletions

diff --git a/drivers/target/target_core_device.c b/drivers/target/target_core_device.c
index 8add07f387f9..e8dd6da164b2 100644
--- a/drivers/target/target_core_device.c
+++ b/drivers/target/target_core_device.c
@@ -49,8 +49,9 @@
 #include "target_core_pr.h"
 #include "target_core_ua.h"
 
-DEFINE_MUTEX(g_device_mutex);
-LIST_HEAD(g_device_list);
+static DEFINE_MUTEX(device_mutex);
+static LIST_HEAD(device_list);
+static DEFINE_IDR(devices_idr);
 
 static struct se_hba *lun0_hba;
 /* not static, needed by tpg.c */
@@ -168,11 +169,20 @@ int transport_lookup_tmr_lun(struct se_cmd *se_cmd, u64 unpacked_lun)
 	rcu_read_lock();
 	deve = target_nacl_find_deve(nacl, unpacked_lun);
 	if (deve) {
-		se_cmd->se_lun = rcu_dereference(deve->se_lun);
 		se_lun = rcu_dereference(deve->se_lun);
+
+		if (!percpu_ref_tryget_live(&se_lun->lun_ref)) {
+			se_lun = NULL;
+			goto out_unlock;
+		}
+
+		se_cmd->se_lun = rcu_dereference(deve->se_lun);
 		se_cmd->pr_res_key = deve->pr_res_key;
 		se_cmd->orig_fe_lun = unpacked_lun;
+		se_cmd->se_cmd_flags |= SCF_SE_LUN_CMD;
+		se_cmd->lun_ref_active = true;
 	}
+out_unlock:
 	rcu_read_unlock();
 
 	if (!se_lun) {
@@ -182,9 +192,6 @@ int transport_lookup_tmr_lun(struct se_cmd *se_cmd, u64 unpacked_lun)
 			unpacked_lun);
 		return -ENODEV;
 	}
-	/*
-	 * XXX: Add percpu se_lun->lun_ref reference count for TMR
-	 */
 	se_cmd->se_dev = rcu_dereference_raw(se_lun->lun_se_dev);
 	se_tmr->tmr_dev = rcu_dereference_raw(se_lun->lun_se_dev);
 
@@ -756,19 +763,16 @@ struct se_device *target_alloc_device(struct se_hba *hba, const char *name)
 	if (!dev)
 		return NULL;
 
-	dev->dev_link_magic = SE_DEV_LINK_MAGIC;
 	dev->se_hba = hba;
 	dev->transport = hba->backend->ops;
 	dev->prot_length = sizeof(struct t10_pi_tuple);
 	dev->hba_index = hba->hba_index;
 
-	INIT_LIST_HEAD(&dev->dev_list);
 	INIT_LIST_HEAD(&dev->dev_sep_list);
 	INIT_LIST_HEAD(&dev->dev_tmr_list);
 	INIT_LIST_HEAD(&dev->delayed_cmd_list);
 	INIT_LIST_HEAD(&dev->state_list);
 	INIT_LIST_HEAD(&dev->qf_cmd_list);
-	INIT_LIST_HEAD(&dev->g_dev_node);
 	spin_lock_init(&dev->execute_task_lock);
 	spin_lock_init(&dev->delayed_cmd_lock);
 	spin_lock_init(&dev->dev_reservation_lock);
@@ -851,7 +855,7 @@ bool target_configure_unmap_from_queue(struct se_dev_attrib *attrib,
 	attrib->unmap_granularity = q->limits.discard_granularity / block_size;
 	attrib->unmap_granularity_alignment = q->limits.discard_alignment /
 								block_size;
-	attrib->unmap_zeroes_data = 0;
+	attrib->unmap_zeroes_data = (q->limits.max_write_zeroes_sectors);
 	return true;
 }
 EXPORT_SYMBOL(target_configure_unmap_from_queue);
@@ -875,10 +879,79 @@ sector_t target_to_linux_sector(struct se_device *dev, sector_t lb)
 }
 EXPORT_SYMBOL(target_to_linux_sector);
 
+/**
+ * target_find_device - find a se_device by its dev_index
+ * @id: dev_index
+ * @do_depend: true if caller needs target_depend_item to be done
+ *
+ * If do_depend is true, the caller must do a target_undepend_item
+ * when finished using the device.
+ *
+ * If do_depend is false, the caller must be called in a configfs
+ * callback or during removal.
+ */
+struct se_device *target_find_device(int id, bool do_depend)
+{
+	struct se_device *dev;
+
+	mutex_lock(&device_mutex);
+	dev = idr_find(&devices_idr, id);
+	if (dev && do_depend && target_depend_item(&dev->dev_group.cg_item))
+		dev = NULL;
+	mutex_unlock(&device_mutex);
+	return dev;
+}
+EXPORT_SYMBOL(target_find_device);
+
+struct devices_idr_iter {
+	int (*fn)(struct se_device *dev, void *data);
+	void *data;
+};
+
+static int target_devices_idr_iter(int id, void *p, void *data)
+{
+	struct devices_idr_iter *iter = data;
+	struct se_device *dev = p;
+
+	/*
+	 * We add the device early to the idr, so it can be used
+	 * by backend modules during configuration. We do not want
+	 * to allow other callers to access partially setup devices,
+	 * so we skip them here.
+	 */
+	if (!(dev->dev_flags & DF_CONFIGURED))
+		return 0;
+
+	return iter->fn(dev, iter->data);
+}
+
+/**
+ * target_for_each_device - iterate over configured devices
+ * @fn: iterator function
+ * @data: pointer to data that will be passed to fn
+ *
+ * fn must return 0 to continue looping over devices. non-zero will break
+ * from the loop and return that value to the caller.
+ */
+int target_for_each_device(int (*fn)(struct se_device *dev, void *data),
+			   void *data)
+{
+	struct devices_idr_iter iter;
+	int ret;
+
+	iter.fn = fn;
+	iter.data = data;
+
+	mutex_lock(&device_mutex);
+	ret = idr_for_each(&devices_idr, target_devices_idr_iter, &iter);
+	mutex_unlock(&device_mutex);
+	return ret;
+}
+
 int target_configure_device(struct se_device *dev)
 {
 	struct se_hba *hba = dev->se_hba;
-	int ret;
+	int ret, id;
 
 	if (dev->dev_flags & DF_CONFIGURED) {
 		pr_err("se_dev->se_dev_ptr already set for storage"
@@ -886,9 +959,26 @@ int target_configure_device(struct se_device *dev)
 		return -EEXIST;
 	}
 
+	/*
+	 * Add early so modules like tcmu can use during its
+	 * configuration.
+	 */
+	mutex_lock(&device_mutex);
+	/*
+	 * Use cyclic to try and avoid collisions with devices
+	 * that were recently removed.
+	 */
+	id = idr_alloc_cyclic(&devices_idr, dev, 0, INT_MAX, GFP_KERNEL);
+	mutex_unlock(&device_mutex);
+	if (id < 0) {
+		ret = -ENOMEM;
+		goto out;
+	}
+	dev->dev_index = id;
+
 	ret = dev->transport->configure_device(dev);
 	if (ret)
-		goto out;
+		goto out_free_index;
 	/*
 	 * XXX: there is not much point to have two different values here..
 	 */
@@ -903,12 +993,11 @@ int target_configure_device(struct se_device *dev)
 					 dev->dev_attrib.hw_block_size);
 	dev->dev_attrib.optimal_sectors = dev->dev_attrib.hw_max_sectors;
 
-	dev->dev_index = scsi_get_new_index(SCSI_DEVICE_INDEX);
 	dev->creation_time = get_jiffies_64();
 
 	ret = core_setup_alua(dev);
 	if (ret)
-		goto out;
+		goto out_free_index;
 
 	/*
 	 * Startup the struct se_device processing thread
@@ -946,16 +1035,16 @@ int target_configure_device(struct se_device *dev)
 	hba->dev_count++;
 	spin_unlock(&hba->device_lock);
 
-	mutex_lock(&g_device_mutex);
-	list_add_tail(&dev->g_dev_node, &g_device_list);
-	mutex_unlock(&g_device_mutex);
-
 	dev->dev_flags |= DF_CONFIGURED;
 
 	return 0;
 
 out_free_alua:
 	core_alua_free_lu_gp_mem(dev);
+out_free_index:
+	mutex_lock(&device_mutex);
+	idr_remove(&devices_idr, dev->dev_index);
+	mutex_unlock(&device_mutex);
 out:
 	se_release_vpd_for_dev(dev);
 	return ret;
@@ -970,9 +1059,11 @@ void target_free_device(struct se_device *dev)
 	if (dev->dev_flags & DF_CONFIGURED) {
 		destroy_workqueue(dev->tmr_wq);
 
-		mutex_lock(&g_device_mutex);
-		list_del(&dev->g_dev_node);
-		mutex_unlock(&g_device_mutex);
+		dev->transport->destroy_device(dev);
+
+		mutex_lock(&device_mutex);
+		idr_remove(&devices_idr, dev->dev_index);
+		mutex_unlock(&device_mutex);
 
 		spin_lock(&hba->device_lock);
 		hba->dev_count--;
@@ -1087,19 +1178,19 @@ passthrough_parse_cdb(struct se_cmd *cmd,
 	      TRANSPORT_FLAG_PASSTHROUGH_PGR)) {
 		if (cdb[0] == PERSISTENT_RESERVE_IN) {
 			cmd->execute_cmd = target_scsi3_emulate_pr_in;
-			size = (cdb[7] << 8) + cdb[8];
+			size = get_unaligned_be16(&cdb[7]);
 			return target_cmd_size_check(cmd, size);
 		}
 		if (cdb[0] == PERSISTENT_RESERVE_OUT) {
 			cmd->execute_cmd = target_scsi3_emulate_pr_out;
-			size = (cdb[7] << 8) + cdb[8];
+			size = get_unaligned_be32(&cdb[5]);
 			return target_cmd_size_check(cmd, size);
 		}
 
 		if (cdb[0] == RELEASE || cdb[0] == RELEASE_10) {
 			cmd->execute_cmd = target_scsi2_reservation_release;
 			if (cdb[0] == RELEASE_10)
-				size = (cdb[7] << 8) | cdb[8];
+				size = get_unaligned_be16(&cdb[7]);
 			else
 				size = cmd->data_length;
 			return target_cmd_size_check(cmd, size);
@@ -1107,7 +1198,7 @@ passthrough_parse_cdb(struct se_cmd *cmd,
 		if (cdb[0] == RESERVE || cdb[0] == RESERVE_10) {
 			cmd->execute_cmd = target_scsi2_reservation_reserve;
 			if (cdb[0] == RESERVE_10)
-				size = (cdb[7] << 8) | cdb[8];
+				size = get_unaligned_be16(&cdb[7]);
 			else
 				size = cmd->data_length;
 			return target_cmd_size_check(cmd, size);
@@ -1126,7 +1217,7 @@ passthrough_parse_cdb(struct se_cmd *cmd,
 	case WRITE_16:
 	case WRITE_VERIFY:
 	case WRITE_VERIFY_12:
-	case 0x8e: /* WRITE_VERIFY_16 */
+	case WRITE_VERIFY_16:
 	case COMPARE_AND_WRITE:
 	case XDWRITEREAD_10:
 		cmd->se_cmd_flags |= SCF_SCSI_DATA_CDB;
@@ -1135,7 +1226,7 @@ passthrough_parse_cdb(struct se_cmd *cmd,
 		switch (get_unaligned_be16(&cdb[8])) {
 		case READ_32:
 		case WRITE_32:
-		case 0x0c: /* WRITE_VERIFY_32 */
+		case WRITE_VERIFY_32:
 		case XDWRITEREAD_32:
 			cmd->se_cmd_flags |= SCF_SCSI_DATA_CDB;
 			break;