cxlflash: Virtual LUN support

Add support for physical LUN segmentation (virtual LUNs) to device
driver supporting the IBM CXL Flash adapter. This patch allows user
space applications to virtually segment a physical LUN into N virtual
LUNs, taking advantage of the translation features provided by this
adapter.

Signed-off-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Manoj N. Kumar <manoj@linux.vnet.ibm.com>
Reviewed-by: Michael Neuling <mikey@neuling.org>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

9 files changed, 1454 insertions, 16 deletions

diff --git a/drivers/scsi/cxlflash/Makefile b/drivers/scsi/cxlflash/Makefile
index c14d24c720d6..9e39866d473b 100644
--- a/drivers/scsi/cxlflash/Makefile
+++ b/drivers/scsi/cxlflash/Makefile
@@ -1,2 +1,2 @@
 obj-$(CONFIG_CXLFLASH) += cxlflash.o
-cxlflash-y += main.o superpipe.o lunmgt.o
+cxlflash-y += main.o superpipe.o lunmgt.o vlun.o
diff --git a/drivers/scsi/cxlflash/common.h b/drivers/scsi/cxlflash/common.h
index d3e54e61c7a5..1c56037146e1 100644
--- a/drivers/scsi/cxlflash/common.h
+++ b/drivers/scsi/cxlflash/common.h
@@ -116,6 +116,9 @@ struct cxlflash_cfg {
 
 	atomic_t num_user_contexts;
 
+	/* Parameters that are LUN table related */
+	int last_lun_index[CXLFLASH_NUM_FC_PORTS];
+	int promote_lun_index;
 	struct list_head lluns; /* list of llun_info structs */
 
 	wait_queue_head_t tmf_waitq;
@@ -200,5 +203,6 @@ int cxlflash_ioctl(struct scsi_device *, int, void __user *);
 void cxlflash_stop_term_user_contexts(struct cxlflash_cfg *);
 int cxlflash_mark_contexts_error(struct cxlflash_cfg *);
 void cxlflash_term_local_luns(struct cxlflash_cfg *);
+void cxlflash_restore_luntable(struct cxlflash_cfg *);
 
 #endif /* ifndef _CXLFLASH_COMMON_H */
diff --git a/drivers/scsi/cxlflash/lunmgt.c b/drivers/scsi/cxlflash/lunmgt.c
index 66d5bef11ee6..d98ad0ff64c1 100644
--- a/drivers/scsi/cxlflash/lunmgt.c
+++ b/drivers/scsi/cxlflash/lunmgt.c
@@ -20,6 +20,7 @@
 
 #include "sislite.h"
 #include "common.h"
+#include "vlun.h"
 #include "superpipe.h"
 
 /**
@@ -42,6 +43,7 @@ static struct llun_info *create_local(struct scsi_device *sdev, u8 *wwid)
 	lli->sdev = sdev;
 	lli->newly_created = true;
 	lli->host_no = sdev->host->host_no;
+	lli->in_table = false;
 
 	memcpy(lli->wwid, wwid, DK_CXLFLASH_MANAGE_LUN_WWID_LEN);
 out:
@@ -208,6 +210,7 @@ void cxlflash_term_global_luns(void)
 	mutex_lock(&global.mutex);
 	list_for_each_entry_safe(gli, temp, &global.gluns, list) {
 		list_del(&gli->list);
+		cxlflash_ba_terminate(&gli->blka.ba_lun);
 		kfree(gli);
 	}
 	mutex_unlock(&global.mutex);
diff --git a/drivers/scsi/cxlflash/main.c b/drivers/scsi/cxlflash/main.c
index 02d464f41b7f..458ed838f83a 100644
--- a/drivers/scsi/cxlflash/main.c
+++ b/drivers/scsi/cxlflash/main.c
@@ -1989,6 +1989,8 @@ static int init_afu(struct cxlflash_cfg *cfg)
 	afu_err_intr_init(cfg->afu);
 	atomic64_set(&afu->room, readq_be(&afu->host_map->cmd_room));
 
+	/* Restore the LUN mappings */
+	cxlflash_restore_luntable(cfg);
 err1:
 	pr_debug("%s: returning rc=%d\n", __func__, rc);
 	return rc;
@@ -2286,6 +2288,17 @@ static int cxlflash_probe(struct pci_dev *pdev,
 
 	cfg->init_state = INIT_STATE_NONE;
 	cfg->dev = pdev;
+
+	/*
+	 * The promoted LUNs move to the top of the LUN table. The rest stay
+	 * on the bottom half. The bottom half grows from the end
+	 * (index = 255), whereas the top half grows from the beginning
+	 * (index = 0).
+	 */
+	cfg->promote_lun_index  = 0;
+	cfg->last_lun_index[0] = CXLFLASH_NUM_VLUNS/2 - 1;
+	cfg->last_lun_index[1] = CXLFLASH_NUM_VLUNS/2 - 1;
+
 	cfg->dev_id = (struct pci_device_id *)dev_id;
 	cfg->mcctx = NULL;
 
diff --git a/drivers/scsi/cxlflash/sislite.h b/drivers/scsi/cxlflash/sislite.h
index 66b889151a4c..63bf394fe78c 100644
--- a/drivers/scsi/cxlflash/sislite.h
+++ b/drivers/scsi/cxlflash/sislite.h
@@ -397,16 +397,17 @@ struct cxlflash_afu_map {
 	};
 };
 
-/* LBA translation control blocks */
-
+/*
+ * LXT - LBA Translation Table
+ * LXT control blocks
+ */
 struct sisl_lxt_entry {
 	u64 rlba_base;	/* bits 0:47 is base
-				 * b48:55 is lun index
-				 * b58:59 is write & read perms
-				 * (if no perm, afu_rc=0x15)
-				 * b60:63 is port_sel mask
-				 */
-
+			 * b48:55 is lun index
+			 * b58:59 is write & read perms
+			 * (if no perm, afu_rc=0x15)
+			 * b60:63 is port_sel mask
+			 */
 };
 
 /*
@@ -465,4 +466,7 @@ struct sisl_rht_entry_f1 {
 #define TMF_LUN_RESET  0x1U
 #define TMF_CLEAR_ACA  0x2U
 
+
+#define SISLITE_MAX_WS_BLOCKS 512
+
 #endif /* _SISLITE_H */
diff --git a/drivers/scsi/cxlflash/superpipe.c b/drivers/scsi/cxlflash/superpipe.c
index 3c8bce8bbb0b..f1b62cea75b1 100644
--- a/drivers/scsi/cxlflash/superpipe.c
+++ b/drivers/scsi/cxlflash/superpipe.c
@@ -26,11 +26,25 @@
 
 #include "sislite.h"
 #include "common.h"
+#include "vlun.h"
 #include "superpipe.h"
 
 struct cxlflash_global global;
 
 /**
+ * marshal_rele_to_resize() - translate release to resize structure
+ * @rele:	Source structure from which to translate/copy.
+ * @resize:	Destination structure for the translate/copy.
+ */
+static void marshal_rele_to_resize(struct dk_cxlflash_release *release,
+				   struct dk_cxlflash_resize *resize)
+{
+	resize->hdr = release->hdr;
+	resize->context_id = release->context_id;
+	resize->rsrc_handle = release->rsrc_handle;
+}
+
+/**
  * marshal_det_to_rele() - translate detach to release structure
  * @detach:	Destination structure for the translate/copy.
  * @rele:	Source structure from which to translate/copy.
@@ -449,6 +463,7 @@ void rhte_checkin(struct ctx_info *ctxi,
 	rhte->fp = 0;
 	ctxi->rht_out--;
 	ctxi->rht_lun[rsrc_handle] = NULL;
+	ctxi->rht_needs_ws[rsrc_handle] = false;
 }
 
 /**
@@ -526,13 +541,21 @@ out:
 /**
  * cxlflash_lun_detach() - detaches a user from a LUN and resets the LUN's mode
  * @gli:	LUN to detach.
+ *
+ * When resetting the mode, terminate block allocation resources as they
+ * are no longer required (service is safe to call even when block allocation
+ * resources were not present - such as when transitioning from physical mode).
+ * These resources will be reallocated when needed (subsequent transition to
+ * virtual mode).
  */
 void cxlflash_lun_detach(struct glun_info *gli)
 {
 	mutex_lock(&gli->mutex);
 	WARN_ON(gli->mode == MODE_NONE);
-	if (--gli->users == 0)
+	if (--gli->users == 0) {
 		gli->mode = MODE_NONE;
+		cxlflash_ba_terminate(&gli->blka.ba_lun);
+	}
 	pr_debug("%s: gli->users=%u\n", __func__, gli->users);
 	WARN_ON(gli->users < 0);
 	mutex_unlock(&gli->mutex);
@@ -544,10 +567,12 @@ void cxlflash_lun_detach(struct glun_info *gli)
  * @ctxi:	Context owning resources.
  * @release:	Release ioctl data structure.
  *
- * Note that the AFU sync should _not_ be performed when the context is sitting
- * on the error recovery list. A context on the error recovery list is not known
- * to the AFU due to reset. When the context is recovered, it will be reattached
- * and made known again to the AFU.
+ * For LUNs in virtual mode, the virtual LUN associated with the specified
+ * resource handle is resized to 0 prior to releasing the RHTE. Note that the
+ * AFU sync should _not_ be performed when the context is sitting on the error
+ * recovery list. A context on the error recovery list is not known to the AFU
+ * due to reset. When the context is recovered, it will be reattached and made
+ * known again to the AFU.
  *
  * Return: 0 on success, -errno on failure
  */
@@ -562,6 +587,7 @@ int _cxlflash_disk_release(struct scsi_device *sdev,
 	struct afu *afu = cfg->afu;
 	bool put_ctx = false;
 
+	struct dk_cxlflash_resize size;
 	res_hndl_t rhndl = release->rsrc_handle;
 
 	int rc = 0;
@@ -594,7 +620,24 @@ int _cxlflash_disk_release(struct scsi_device *sdev,
 		goto out;
 	}
 
+	/*
+	 * Resize to 0 for virtual LUNS by setting the size
+	 * to 0. This will clear LXT_START and LXT_CNT fields
+	 * in the RHT entry and properly sync with the AFU.
+	 *
+	 * Afterwards we clear the remaining fields.
+	 */
 	switch (gli->mode) {
+	case MODE_VIRTUAL:
+		marshal_rele_to_resize(release, &size);
+		size.req_size = 0;
+		rc = _cxlflash_vlun_resize(sdev, ctxi, &size);
+		if (rc) {
+			dev_dbg(dev, "%s: resize failed rc %d\n", __func__, rc);
+			goto out;
+		}
+
+		break;
 	case MODE_PHYSICAL:
 		/*
 		 * Clear the Format 1 RHT entry for direct access
@@ -666,6 +709,7 @@ static void destroy_context(struct cxlflash_cfg *cfg,
 
 	/* Free memory associated with context */
 	free_page((ulong)ctxi->rht_start);
+	kfree(ctxi->rht_needs_ws);
 	kfree(ctxi->rht_lun);
 	kfree(ctxi);
 	atomic_dec_if_positive(&cfg->num_user_contexts);
@@ -693,11 +737,13 @@ static struct ctx_info *create_context(struct cxlflash_cfg *cfg,
 	struct afu *afu = cfg->afu;
 	struct ctx_info *ctxi = NULL;
 	struct llun_info **lli = NULL;
+	bool *ws = NULL;
 	struct sisl_rht_entry *rhte;
 
 	ctxi = kzalloc(sizeof(*ctxi), GFP_KERNEL);
 	lli = kzalloc((MAX_RHT_PER_CONTEXT * sizeof(*lli)), GFP_KERNEL);
-	if (unlikely(!ctxi || !lli)) {
+	ws = kzalloc((MAX_RHT_PER_CONTEXT * sizeof(*ws)), GFP_KERNEL);
+	if (unlikely(!ctxi || !lli || !ws)) {
 		dev_err(dev, "%s: Unable to allocate context!\n", __func__);
 		goto err;
 	}
@@ -709,6 +755,7 @@ static struct ctx_info *create_context(struct cxlflash_cfg *cfg,
 	}
 
 	ctxi->rht_lun = lli;
+	ctxi->rht_needs_ws = ws;
 	ctxi->rht_start = rhte;
 	ctxi->rht_perms = perms;
 
@@ -728,6 +775,7 @@ out:
 	return ctxi;
 
 err:
+	kfree(ws);
 	kfree(lli);
 	kfree(ctxi);
 	ctxi = NULL;
@@ -1729,6 +1777,12 @@ static int cxlflash_disk_verify(struct scsi_device *sdev,
 	case MODE_PHYSICAL:
 		last_lba = gli->max_lba;
 		break;
+	case MODE_VIRTUAL:
+		/* Cast lxt_cnt to u64 for multiply to be treated as 64bit op */
+		last_lba = ((u64)rhte->lxt_cnt * MC_CHUNK_SIZE * gli->blk_len);
+		last_lba /= CXLFLASH_BLOCK_SIZE;
+		last_lba--;
+		break;
 	default:
 		WARN(1, "Unsupported LUN mode!");
 	}
@@ -1756,12 +1810,18 @@ static char *decode_ioctl(int cmd)
 		return __stringify_1(DK_CXLFLASH_ATTACH);
 	case DK_CXLFLASH_USER_DIRECT:
 		return __stringify_1(DK_CXLFLASH_USER_DIRECT);
+	case DK_CXLFLASH_USER_VIRTUAL:
+		return __stringify_1(DK_CXLFLASH_USER_VIRTUAL);
+	case DK_CXLFLASH_VLUN_RESIZE:
+		return __stringify_1(DK_CXLFLASH_VLUN_RESIZE);
 	case DK_CXLFLASH_RELEASE:
 		return __stringify_1(DK_CXLFLASH_RELEASE);
 	case DK_CXLFLASH_DETACH:
 		return __stringify_1(DK_CXLFLASH_DETACH);
 	case DK_CXLFLASH_VERIFY:
 		return __stringify_1(DK_CXLFLASH_VERIFY);
+	case DK_CXLFLASH_VLUN_CLONE:
+		return __stringify_1(DK_CXLFLASH_VLUN_CLONE);
 	case DK_CXLFLASH_RECOVER_AFU:
 		return __stringify_1(DK_CXLFLASH_RECOVER_AFU);
 	case DK_CXLFLASH_MANAGE_LUN:
@@ -1876,6 +1936,7 @@ static int ioctl_common(struct scsi_device *sdev, int cmd)
 	rc = check_state(cfg);
 	if (unlikely(rc) && (cfg->state == STATE_FAILTERM)) {
 		switch (cmd) {
+		case DK_CXLFLASH_VLUN_RESIZE:
 		case DK_CXLFLASH_RELEASE:
 		case DK_CXLFLASH_DETACH:
 			dev_dbg(dev, "%s: Command override! (%d)\n",
@@ -1923,12 +1984,18 @@ int cxlflash_ioctl(struct scsi_device *sdev, int cmd, void __user *arg)
 	{sizeof(struct dk_cxlflash_verify), (sioctl)cxlflash_disk_verify},
 	{sizeof(struct dk_cxlflash_recover_afu), (sioctl)cxlflash_afu_recover},
 	{sizeof(struct dk_cxlflash_manage_lun), (sioctl)cxlflash_manage_lun},
+	{sizeof(struct dk_cxlflash_uvirtual), cxlflash_disk_virtual_open},
+	{sizeof(struct dk_cxlflash_resize), (sioctl)cxlflash_vlun_resize},
+	{sizeof(struct dk_cxlflash_clone), (sioctl)cxlflash_disk_clone},
 	};
 
 	/* Restrict command set to physical support only for internal LUN */
 	if (afu->internal_lun)
 		switch (cmd) {
 		case DK_CXLFLASH_RELEASE:
+		case DK_CXLFLASH_USER_VIRTUAL:
+		case DK_CXLFLASH_VLUN_RESIZE:
+		case DK_CXLFLASH_VLUN_CLONE:
 			dev_dbg(dev, "%s: %s not supported for lun_mode=%d\n",
 				__func__, decode_ioctl(cmd), afu->internal_lun);
 			rc = -EINVAL;
@@ -1942,6 +2009,9 @@ int cxlflash_ioctl(struct scsi_device *sdev, int cmd, void __user *arg)
 	case DK_CXLFLASH_DETACH:
 	case DK_CXLFLASH_VERIFY:
 	case DK_CXLFLASH_RECOVER_AFU:
+	case DK_CXLFLASH_USER_VIRTUAL:
+	case DK_CXLFLASH_VLUN_RESIZE:
+	case DK_CXLFLASH_VLUN_CLONE:
 		dev_dbg(dev, "%s: %s (%08X) on dev(%d/%d/%d/%llu)\n",
 			__func__, decode_ioctl(cmd), cmd, shost->host_no,
 			sdev->channel, sdev->id, sdev->lun);
diff --git a/drivers/scsi/cxlflash/superpipe.h b/drivers/scsi/cxlflash/superpipe.h
index ae39b9627118..d7dc88bc64a4 100644
--- a/drivers/scsi/cxlflash/superpipe.h
+++ b/drivers/scsi/cxlflash/superpipe.h
@@ -31,9 +31,11 @@ extern struct cxlflash_global global;
 #define MC_DISCOVERY_TIMEOUT 5  /* 5 secs */
 
 #define CHAN2PORT(_x)	((_x) + 1)
+#define PORT2CHAN(_x)	((_x) - 1)
 
 enum lun_mode {
 	MODE_NONE = 0,
+	MODE_VIRTUAL,
 	MODE_PHYSICAL
 };
 
@@ -41,13 +43,14 @@ enum lun_mode {
 struct glun_info {
 	u64 max_lba;		/* from read cap(16) */
 	u32 blk_len;		/* from read cap(16) */
-	enum lun_mode mode;	/* NONE, PHYSICAL */
+	enum lun_mode mode;	/* NONE, VIRTUAL, PHYSICAL */
 	int users;		/* Number of users w/ references to LUN */
 
 	u8 wwid[16];
 
 	struct mutex mutex;
 
+	struct blka blka;
 	struct list_head list;
 };
 
@@ -58,6 +61,7 @@ struct llun_info {
 	u32 host_no;		/* host_no from Scsi_host */
 	u32 port_sel;		/* What port to use for this LUN */
 	bool newly_created;	/* Whether the LUN was just discovered */
+	bool in_table;		/* Whether a LUN table entry was created */
 
 	u8 wwid[16];		/* Keep a duplicate copy here? */
 
@@ -90,6 +94,7 @@ struct ctx_info {
 	u32 rht_out;		/* Number of checked out RHT entries */
 	u32 rht_perms;		/* User-defined permissions for RHT entries */
 	struct llun_info **rht_lun;       /* Mapping of RHT entries to LUNs */
+	bool *rht_needs_ws;	/* User-desired write-same function per RHTE */
 
 	struct cxl_ioctl_start_work work;
 	u64 ctxid;
@@ -111,10 +116,18 @@ struct cxlflash_global {
 	struct page *err_page; /* One page of all 0xF for error notification */
 };
 
+int cxlflash_vlun_resize(struct scsi_device *, struct dk_cxlflash_resize *);
+int _cxlflash_vlun_resize(struct scsi_device *, struct ctx_info *,
+			  struct dk_cxlflash_resize *);
+
 int cxlflash_disk_release(struct scsi_device *, struct dk_cxlflash_release *);
 int _cxlflash_disk_release(struct scsi_device *, struct ctx_info *,
 			   struct dk_cxlflash_release *);
 
+int cxlflash_disk_clone(struct scsi_device *, struct dk_cxlflash_clone *);
+
+int cxlflash_disk_virtual_open(struct scsi_device *, void *);
+
 int cxlflash_lun_attach(struct glun_info *, enum lun_mode, bool);
 void cxlflash_lun_detach(struct glun_info *);
 
@@ -127,6 +140,8 @@ struct sisl_rht_entry *get_rhte(struct ctx_info *, res_hndl_t,
 struct sisl_rht_entry *rhte_checkout(struct ctx_info *, struct llun_info *);
 void rhte_checkin(struct ctx_info *, struct sisl_rht_entry *);
 
+void cxlflash_ba_terminate(struct ba_lun *);
+
 int cxlflash_manage_lun(struct scsi_device *, struct dk_cxlflash_manage_lun *);
 
 #endif /* ifndef _CXLFLASH_SUPERPIPE_H */
diff --git a/drivers/scsi/cxlflash/vlun.c b/drivers/scsi/cxlflash/vlun.c
new file mode 100644
index 000000000000..6155cb1d4ed3
--- /dev/null
+++ b/drivers/scsi/cxlflash/vlun.c
@@ -0,0 +1,1243 @@
+/*
+ * CXL Flash Device Driver
+ *
+ * Written by: Manoj N. Kumar <manoj@linux.vnet.ibm.com>, IBM Corporation
+ *             Matthew R. Ochs <mrochs@linux.vnet.ibm.com>, IBM Corporation
+ *
+ * Copyright (C) 2015 IBM Corporation
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+
+#include <linux/syscalls.h>
+#include <misc/cxl.h>
+#include <asm/unaligned.h>
+#include <asm/bitsperlong.h>
+
+#include <scsi/scsi_cmnd.h>
+#include <scsi/scsi_host.h>
+#include <uapi/scsi/cxlflash_ioctl.h>
+
+#include "sislite.h"
+#include "common.h"
+#include "vlun.h"
+#include "superpipe.h"
+
+/**
+ * marshal_virt_to_resize() - translate uvirtual to resize structure
+ * @virt:	Source structure from which to translate/copy.
+ * @resize:	Destination structure for the translate/copy.
+ */
+static void marshal_virt_to_resize(struct dk_cxlflash_uvirtual *virt,
+				   struct dk_cxlflash_resize *resize)
+{
+	resize->hdr = virt->hdr;
+	resize->context_id = virt->context_id;
+	resize->rsrc_handle = virt->rsrc_handle;
+	resize->req_size = virt->lun_size;
+	resize->last_lba = virt->last_lba;
+}
+
+/**
+ * marshal_clone_to_rele() - translate clone to release structure
+ * @clone:	Source structure from which to translate/copy.
+ * @rele:	Destination structure for the translate/copy.
+ */
+static void marshal_clone_to_rele(struct dk_cxlflash_clone *clone,
+				  struct dk_cxlflash_release *release)
+{
+	release->hdr = clone->hdr;
+	release->context_id = clone->context_id_dst;
+}
+
+/**
+ * ba_init() - initializes a block allocator
+ * @ba_lun:	Block allocator to initialize.
+ *
+ * Return: 0 on success, -errno on failure
+ */
+static int ba_init(struct ba_lun *ba_lun)
+{
+	struct ba_lun_info *bali = NULL;
+	int lun_size_au = 0, i = 0;
+	int last_word_underflow = 0;
+	u64 *lam;
+
+	pr_debug("%s: Initializing LUN: lun_id = %llX, "
+		 "ba_lun->lsize = %lX, ba_lun->au_size = %lX\n",
+		__func__, ba_lun->lun_id, ba_lun->lsize, ba_lun->au_size);
+
+	/* Calculate bit map size */
+	lun_size_au = ba_lun->lsize / ba_lun->au_size;
+	if (lun_size_au == 0) {
+		pr_debug("%s: Requested LUN size of 0!\n", __func__);
+		return -EINVAL;
+	}
+
+	/* Allocate lun information container */
+	bali = kzalloc(sizeof(struct ba_lun_info), GFP_KERNEL);
+	if (unlikely(!bali)) {
+		pr_err("%s: Failed to allocate lun_info for lun_id %llX\n",
+		       __func__, ba_lun->lun_id);
+		return -ENOMEM;
+	}
+
+	bali->total_aus = lun_size_au;
+	bali->lun_bmap_size = lun_size_au / BITS_PER_LONG;
+
+	if (lun_size_au % BITS_PER_LONG)
+		bali->lun_bmap_size++;
+
+	/* Allocate bitmap space */
+	bali->lun_alloc_map = kzalloc((bali->lun_bmap_size * sizeof(u64)),
+				      GFP_KERNEL);
+	if (unlikely(!bali->lun_alloc_map)) {
+		pr_err("%s: Failed to allocate lun allocation map: "
+		       "lun_id = %llX\n", __func__, ba_lun->lun_id);
+		kfree(bali);
+		return -ENOMEM;
+	}
+
+	/* Initialize the bit map size and set all bits to '1' */
+	bali->free_aun_cnt = lun_size_au;
+
+	for (i = 0; i < bali->lun_bmap_size; i++)
+		bali->lun_alloc_map[i] = 0xFFFFFFFFFFFFFFFFULL;
+
+	/* If the last word not fully utilized, mark extra bits as allocated */
+	last_word_underflow = (bali->lun_bmap_size * BITS_PER_LONG);
+	last_word_underflow -= bali->free_aun_cnt;
+	if (last_word_underflow > 0) {
+		lam = &bali->lun_alloc_map[bali->lun_bmap_size - 1];
+		for (i = (HIBIT - last_word_underflow + 1);
+		     i < BITS_PER_LONG;
+		     i++)
+			clear_bit(i, (ulong *)lam);
+	}
+
+	/* Initialize high elevator index, low/curr already at 0 from kzalloc */
+	bali->free_high_idx = bali->lun_bmap_size;
+
+	/* Allocate clone map */
+	bali->aun_clone_map = kzalloc((bali->total_aus * sizeof(u8)),
+				      GFP_KERNEL);
+	if (unlikely(!bali->aun_clone_map)) {
+		pr_err("%s: Failed to allocate clone map: lun_id = %llX\n",
+		       __func__, ba_lun->lun_id);
+		kfree(bali->lun_alloc_map);
+		kfree(bali);
+		return -ENOMEM;
+	}
+
+	/* Pass the allocated lun info as a handle to the user */
+	ba_lun->ba_lun_handle = bali;
+
+	pr_debug("%s: Successfully initialized the LUN: "
+		 "lun_id = %llX, bitmap size = %X, free_aun_cnt = %llX\n",
+		__func__, ba_lun->lun_id, bali->lun_bmap_size,
+		bali->free_aun_cnt);
+	return 0;
+}
+
+/**
+ * find_free_range() - locates a free bit within the block allocator
+ * @low:	First word in block allocator to start search.
+ * @high:	Last word in block allocator to search.
+ * @bali:	LUN information structure owning the block allocator to search.
+ * @bit_word:	Passes back the word in the block allocator owning the free bit.
+ *
+ * Return: The bit position within the passed back word, -1 on failure
+ */
+static int find_free_range(u32 low,
+			   u32 high,
+			   struct ba_lun_info *bali, int *bit_word)
+{
+	int i;
+	u64 bit_pos = -1;
+	ulong *lam, num_bits;
+
+	for (i = low; i < high; i++)
+		if (bali->lun_alloc_map[i] != 0) {
+			lam = (ulong *)&bali->lun_alloc_map[i];
+			num_bits = (sizeof(*lam) * BITS_PER_BYTE);
+			bit_pos = find_first_bit(lam, num_bits);
+
+			pr_devel("%s: Found free bit %llX in lun "
+				 "map entry %llX at bitmap index = %X\n",
+				 __func__, bit_pos, bali->lun_alloc_map[i],
+				 i);
+
+			*bit_word = i;
+			bali->free_aun_cnt--;
+			clear_bit(bit_pos, lam);
+			break;
+		}
+
+	return bit_pos;
+}
+
+/**
+ * ba_alloc() - allocates a block from the block allocator
+ * @ba_lun:	Block allocator from which to allocate a block.
+ *
+ * Return: The allocated block, -1 on failure
+ */
+static u64 ba_alloc(struct ba_lun *ba_lun)
+{
+	u64 bit_pos = -1;
+	int bit_word = 0;
+	struct ba_lun_info *bali = NULL;
+
+	bali = ba_lun->ba_lun_handle;
+
+	pr_debug("%s: Received block allocation request: "
+		 "lun_id = %llX, free_aun_cnt = %llX\n",
+		 __func__, ba_lun->lun_id, bali->free_aun_cnt);
+
+	if (bali->free_aun_cnt == 0) {
+		pr_debug("%s: No space left on LUN: lun_id = %llX\n",
+			 __func__, ba_lun->lun_id);
+		return -1ULL;
+	}
+
+	/* Search to find a free entry, curr->high then low->curr */
+	bit_pos = find_free_range(bali->free_curr_idx,
+				  bali->free_high_idx, bali, &bit_word);
+	if (bit_pos == -1) {
+		bit_pos = find_free_range(bali->free_low_idx,
+					  bali->free_curr_idx,
+					  bali, &bit_word);
+		if (bit_pos == -1) {
+			pr_debug("%s: Could not find an allocation unit on LUN:"
+				 " lun_id = %llX\n", __func__, ba_lun->lun_id);
+			return -1ULL;
+		}
+	}
+
+	/* Update the free_curr_idx */
+	if (bit_pos == HIBIT)
+		bali->free_curr_idx = bit_word + 1;
+	else
+		bali->free_curr_idx = bit_word;
+
+	pr_debug("%s: Allocating AU number %llX, on lun_id %llX, "
+		 "free_aun_cnt = %llX\n", __func__,
+		 ((bit_word * BITS_PER_LONG) + bit_pos), ba_lun->lun_id,
+		 bali->free_aun_cnt);
+
+	return (u64) ((bit_word * BITS_PER_LONG) + bit_pos);
+}
+
+/**
+ * validate_alloc() - validates the specified block has been allocated
+ * @ba_lun_info:	LUN info owning the block allocator.
+ * @aun:		Block to validate.
+ *
+ * Return: 0 on success, -1 on failure
+ */
+static int validate_alloc(struct ba_lun_info *bali, u64 aun)
+{
+	int idx = 0, bit_pos = 0;
+
+	idx = aun / BITS_PER_LONG;
+	bit_pos = aun % BITS_PER_LONG;
+
+	if (test_bit(bit_pos, (ulong *)&bali->lun_alloc_map[idx]))
+		return -1;
+
+	return 0;
+}
+
+/**
+ * ba_free() - frees a block from the block allocator
+ * @ba_lun:	Block allocator from which to allocate a block.
+ * @to_free:	Block to free.
+ *
+ * Return: 0 on success, -1 on failure
+ */
+static int ba_free(struct ba_lun *ba_lun, u64 to_free)
+{
+	int idx = 0, bit_pos = 0;
+	struct ba_lun_info *bali = NULL;
+
+	bali = ba_lun->ba_lun_handle;
+
+	if (validate_alloc(bali, to_free)) {
+		pr_debug("%s: The AUN %llX is not allocated on lun_id %llX\n",
+			 __func__, to_free, ba_lun->lun_id);
+		return -1;
+	}
+
+	pr_debug("%s: Received a request to free AU %llX on lun_id %llX, "
+		 "free_aun_cnt = %llX\n", __func__, to_free, ba_lun->lun_id,
+		 bali->free_aun_cnt);
+
+	if (bali->aun_clone_map[to_free] > 0) {
+		pr_debug("%s: AUN %llX on lun_id %llX has been cloned. Clone "
+			 "count = %X\n", __func__, to_free, ba_lun->lun_id,
+			 bali->aun_clone_map[to_free]);
+		bali->aun_clone_map[to_free]--;
+		return 0;
+	}
+
+	idx = to_free / BITS_PER_LONG;
+	bit_pos = to_free % BITS_PER_LONG;
+
+	set_bit(bit_pos, (ulong *)&bali->lun_alloc_map[idx]);
+	bali->free_aun_cnt++;
+
+	if (idx < bali->free_low_idx)
+		bali->free_low_idx = idx;
+	else if (idx > bali->free_high_idx)
+		bali->free_high_idx = idx;
+
+	pr_debug("%s: Successfully freed AU at bit_pos %X, bit map index %X on "
+		 "lun_id %llX, free_aun_cnt = %llX\n", __func__, bit_pos, idx,
+		 ba_lun->lun_id, bali->free_aun_cnt);
+
+	return 0;
+}
+
+/**
+ * ba_clone() - Clone a chunk of the block allocation table
+ * @ba_lun:	Block allocator from which to allocate a block.
+ * @to_free:	Block to free.
+ *
+ * Return: 0 on success, -1 on failure
+ */
+static int ba_clone(struct ba_lun *ba_lun, u64 to_clone)
+{
+	struct ba_lun_info *bali = ba_lun->ba_lun_handle;
+
+	if (validate_alloc(bali, to_clone)) {
+		pr_debug("%s: AUN %llX is not allocated on lun_id %llX\n",
+			 __func__, to_clone, ba_lun->lun_id);
+		return -1;
+	}
+
+	pr_debug("%s: Received a request to clone AUN %llX on lun_id %llX\n",
+		 __func__, to_clone, ba_lun->lun_id);
+
+	if (bali->aun_clone_map[to_clone] == MAX_AUN_CLONE_CNT) {
+		pr_debug("%s: AUN %llX on lun_id %llX hit max clones already\n",
+			 __func__, to_clone, ba_lun->lun_id);
+		return -1;
+	}
+
+	bali->aun_clone_map[to_clone]++;
+
+	return 0;
+}
+
+/**
+ * ba_space() - returns the amount of free space left in the block allocator
+ * @ba_lun:	Block allocator.
+ *
+ * Return: Amount of free space in block allocator
+ */
+static u64 ba_space(struct ba_lun *ba_lun)
+{
+	struct ba_lun_info *bali = ba_lun->ba_lun_handle;
+
+	return bali->free_aun_cnt;
+}
+
+/**
+ * cxlflash_ba_terminate() - frees resources associated with the block allocator
+ * @ba_lun:	Block allocator.
+ *
+ * Safe to call in a partially allocated state.
+ */
+void cxlflash_ba_terminate(struct ba_lun *ba_lun)
+{
+	struct ba_lun_info *bali = ba_lun->ba_lun_handle;
+
+	if (bali) {
+		kfree(bali->aun_clone_map);
+		kfree(bali->lun_alloc_map);
+		kfree(bali);
+		ba_lun->ba_lun_handle = NULL;
+	}
+}
+
+/**
+ * init_vlun() - initializes a LUN for virtual use
+ * @lun_info:	LUN information structure that owns the block allocator.
+ *
+ * Return: 0 on success, -errno on failure
+ */
+static int init_vlun(struct llun_info *lli)
+{
+	int rc = 0;
+	struct glun_info *gli = lli->parent;
+	struct blka *blka = &gli->blka;
+
+	memset(blka, 0, sizeof(*blka));
+	mutex_init(&blka->mutex);
+
+	/* LUN IDs are unique per port, save the index instead */
+	blka->ba_lun.lun_id = lli->lun_index;
+	blka->ba_lun.lsize = gli->max_lba + 1;
+	blka->ba_lun.lba_size = gli->blk_len;
+
+	blka->ba_lun.au_size = MC_CHUNK_SIZE;
+	blka->nchunk = blka->ba_lun.lsize / MC_CHUNK_SIZE;
+
+	rc = ba_init(&blka->ba_lun);
+	if (unlikely(rc))
+		pr_debug("%s: cannot init block_alloc, rc=%d\n", __func__, rc);
+
+	pr_debug("%s: returning rc=%d lli=%p\n", __func__, rc, lli);
+	return rc;
+}
+
+/**
+ * write_same16() - sends a SCSI WRITE_SAME16 (0) command to specified LUN
+ * @sdev:	SCSI device associated with LUN.
+ * @lba:	Logical block address to start write same.
+ * @nblks:	Number of logical blocks to write same.
+ *
+ * Return: 0 on success, -errno on failure
+ */
+static int write_same16(struct scsi_device *sdev,
+			u64 lba,
+			u32 nblks)
+{
+	u8 *cmd_buf = NULL;
+	u8 *scsi_cmd = NULL;
+	u8 *sense_buf = NULL;
+	int rc = 0;
+	int result = 0;
+	int ws_limit = SISLITE_MAX_WS_BLOCKS;
+	u64 offset = lba;
+	int left = nblks;
+	u32 tout = sdev->request_queue->rq_timeout;
+	struct cxlflash_cfg *cfg = (struct cxlflash_cfg *)sdev->host->hostdata;
+	struct device *dev = &cfg->dev->dev;
+
+	cmd_buf = kzalloc(CMD_BUFSIZE, GFP_KERNEL);
+	scsi_cmd = kzalloc(MAX_COMMAND_SIZE, GFP_KERNEL);
+	sense_buf = kzalloc(SCSI_SENSE_BUFFERSIZE, GFP_KERNEL);
+	if (unlikely(!cmd_buf || !scsi_cmd || !sense_buf)) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	while (left > 0) {
+
+		scsi_cmd[0] = WRITE_SAME_16;
+		put_unaligned_be64(offset, &scsi_cmd[2]);
+		put_unaligned_be32(ws_limit < left ? ws_limit : left,
+				   &scsi_cmd[10]);
+
+		result = scsi_execute(sdev, scsi_cmd, DMA_TO_DEVICE, cmd_buf,
+				      CMD_BUFSIZE, sense_buf, tout, 5, 0, NULL);
+		if (result) {
+			dev_err_ratelimited(dev, "%s: command failed for "
+					    "offset %lld result=0x%x\n",
+					    __func__, offset, result);
+			rc = -EIO;
+			goto out;
+		}
+		left -= ws_limit;
+		offset += ws_limit;
+	}
+
+out:
+	kfree(cmd_buf);
+	kfree(scsi_cmd);
+	kfree(sense_buf);
+	pr_debug("%s: returning rc=%d\n", __func__, rc);
+	return rc;
+}
+
+/**
+ * grow_lxt() - expands the translation table associated with the specified RHTE
+ * @afu:	AFU associated with the host.
+ * @sdev:	SCSI device associated with LUN.
+ * @ctxid:	Context ID of context owning the RHTE.
+ * @rhndl:	Resource handle associated with the RHTE.
+ * @rhte:	Resource handle entry (RHTE).
+ * @new_size:	Number of translation entries associated with RHTE.
+ *
+ * By design, this routine employs a 'best attempt' allocation and will
+ * truncate the requested size down if there is not sufficient space in
+ * the block allocator to satisfy the request but there does exist some
+ * amount of space. The user is made aware of this by returning the size
+ * allocated.
+ *
+ * Return: 0 on success, -errno on failure
+ */
+static int grow_lxt(struct afu *afu,
+		    struct scsi_device *sdev,
+		    ctx_hndl_t ctxid,
+		    res_hndl_t rhndl,
+		    struct sisl_rht_entry *rhte,
+		    u64 *new_size)
+{
+	struct sisl_lxt_entry *lxt = NULL, *lxt_old = NULL;
+	struct llun_info *lli = sdev->hostdata;
+	struct glun_info *gli = lli->parent;
+	struct blka *blka = &gli->blka;
+	u32 av_size;
+	u32 ngrps, ngrps_old;
+	u64 aun;		/* chunk# allocated by block allocator */
+	u64 delta = *new_size - rhte->lxt_cnt;
+	u64 my_new_size;
+	int i, rc = 0;
+
+	/*
+	 * Check what is available in the block allocator before re-allocating
+	 * LXT array. This is done up front under the mutex which must not be
+	 * released until after allocation is complete.
+	 */
+	mutex_lock(&blka->mutex);
+	av_size = ba_space(&blka->ba_lun);
+	if (unlikely(av_size <= 0)) {
+		pr_debug("%s: ba_space error: av_size %d\n", __func__, av_size);
+		mutex_unlock(&blka->mutex);
+		rc = -ENOSPC;
+		goto out;
+	}
+
+	if (av_size < delta)
+		delta = av_size;
+
+	lxt_old = rhte->lxt_start;
+	ngrps_old = LXT_NUM_GROUPS(rhte->lxt_cnt);
+	ngrps = LXT_NUM_GROUPS(rhte->lxt_cnt + delta);
+
+	if (ngrps != ngrps_old) {
+		/* reallocate to fit new size */
+		lxt = kzalloc((sizeof(*lxt) * LXT_GROUP_SIZE * ngrps),
+			      GFP_KERNEL);
+		if (unlikely(!lxt)) {
+			mutex_unlock(&blka->mutex);
+			rc = -ENOMEM;
+			goto out;
+		}
+
+		/* copy over all old entries */
+		memcpy(lxt, lxt_old, (sizeof(*lxt) * rhte->lxt_cnt));
+	} else
+		lxt = lxt_old;
+
+	/* nothing can fail from now on */
+	my_new_size = rhte->lxt_cnt + delta;
+
+	/* add new entries to the end */
+	for (i = rhte->lxt_cnt; i < my_new_size; i++) {
+		/*
+		 * Due to the earlier check of available space, ba_alloc
+		 * cannot fail here. If it did due to internal error,
+		 * leave a rlba_base of -1u which will likely be a
+		 * invalid LUN (too large).
+		 */
+		aun = ba_alloc(&blka->ba_lun);
+		if ((aun == -1ULL) || (aun >= blka->nchunk))
+			pr_debug("%s: ba_alloc error: allocated chunk# %llX, "
+				 "max %llX\n", __func__, aun, blka->nchunk - 1);
+
+		/* select both ports, use r/w perms from RHT */
+		lxt[i].rlba_base = ((aun << MC_CHUNK_SHIFT) |
+				    (lli->lun_index << LXT_LUNIDX_SHIFT) |
+				    (RHT_PERM_RW << LXT_PERM_SHIFT |
+				     lli->port_sel));
+	}
+
+	mutex_unlock(&blka->mutex);
+
+	/*
+	 * The following sequence is prescribed in the SISlite spec
+	 * for syncing up with the AFU when adding LXT entries.
+	 */
+	dma_wmb(); /* Make LXT updates are visible */
+
+	rhte->lxt_start = lxt;
+	dma_wmb(); /* Make RHT entry's LXT table update visible */
+
+	rhte->lxt_cnt = my_new_size;
+	dma_wmb(); /* Make RHT entry's LXT table size update visible */
+
+	cxlflash_afu_sync(afu, ctxid, rhndl, AFU_LW_SYNC);
+
+	/* free old lxt if reallocated */
+	if (lxt != lxt_old)
+		kfree(lxt_old);
+	*new_size = my_new_size;
+out:
+	pr_debug("%s: returning rc=%d\n", __func__, rc);
+	return rc;
+}
+
+/**
+ * shrink_lxt() - reduces translation table associated with the specified RHTE
+ * @afu:	AFU associated with the host.
+ * @sdev:	SCSI device associated with LUN.
+ * @rhndl:	Resource handle associated with the RHTE.
+ * @rht