All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Matias Bjørling" <mb@lightnvm.io>
To: hch@infradead.org, axboe@fb.com, linux-fsdevel@vger.kernel.org,
	linux-kernel@vger.kernel.org, linux-nvme@lists.infradead.org
Cc: Stephen.Bates@pmcs.com, keith.busch@intel.com,
	"Matias Bjørling" <mb@lightnvm.io>,
	"Javier González" <jg@lightnvm.io>
Subject: [PATCH v5 5/5] nvme: LightNVM support
Date: Wed, 22 Jul 2015 19:51:04 +0200	[thread overview]
Message-ID: <1437587464-7964-6-git-send-email-mb@lightnvm.io> (raw)
In-Reply-To: <1437587464-7964-1-git-send-email-mb@lightnvm.io>

The first generation of Open-Channel SSDs will be based on NVMe. The
integration requires that a NVMe device exposes itself as a LightNVM
device. The way this is done currently is by hooking into the
Controller Capabilities (CAP register) and a bit in NSFEAT for each
namespace.

After detection, vendor specific codes are used to identify the device
and enumerate supported features.

Signed-off-by: Javier González <jg@lightnvm.io>
Signed-off-by: Matias Bjørling <mb@lightnvm.io>
---
 drivers/block/Makefile        |   2 +-
 drivers/block/nvme-core.c     |  22 +-
 drivers/block/nvme-lightnvm.c | 504 ++++++++++++++++++++++++++++++++++++++++++
 include/linux/nvme.h          |   6 +
 include/uapi/linux/nvme.h     |   3 +
 5 files changed, 533 insertions(+), 4 deletions(-)
 create mode 100644 drivers/block/nvme-lightnvm.c

diff --git a/drivers/block/Makefile b/drivers/block/Makefile
index 02b688d..a01d7d8 100644
--- a/drivers/block/Makefile
+++ b/drivers/block/Makefile
@@ -44,6 +44,6 @@ obj-$(CONFIG_BLK_DEV_RSXX) += rsxx/
 obj-$(CONFIG_BLK_DEV_NULL_BLK)	+= null_blk.o
 obj-$(CONFIG_ZRAM) += zram/
 
-nvme-y		:= nvme-core.o nvme-scsi.o
+nvme-y		:= nvme-core.o nvme-scsi.o nvme-lightnvm.o
 skd-y		:= skd_main.o
 swim_mod-y	:= swim.o swim_asm.o
diff --git a/drivers/block/nvme-core.c b/drivers/block/nvme-core.c
index 666e994..ab799b0 100644
--- a/drivers/block/nvme-core.c
+++ b/drivers/block/nvme-core.c
@@ -40,6 +40,7 @@
 #include <linux/slab.h>
 #include <linux/t10-pi.h>
 #include <linux/types.h>
+#include <linux/lightnvm.h>
 #include <scsi/sg.h>
 #include <asm-generic/io-64-nonatomic-lo-hi.h>
 
@@ -1751,7 +1752,8 @@ static int nvme_configure_admin_queue(struct nvme_dev *dev)
 
 	dev->page_size = 1 << page_shift;
 
-	dev->ctrl_config = NVME_CC_CSS_NVM;
+	dev->ctrl_config = NVME_CAP_LIGHTNVM(cap) ?
+					NVME_CC_CSS_LIGHTNVM : NVME_CC_CSS_NVM;
 	dev->ctrl_config |= (page_shift - 12) << NVME_CC_MPS_SHIFT;
 	dev->ctrl_config |= NVME_CC_ARB_RR | NVME_CC_SHN_NONE;
 	dev->ctrl_config |= NVME_CC_IOSQES | NVME_CC_IOCQES;
@@ -1997,6 +1999,16 @@ static int nvme_revalidate_disk(struct gendisk *disk)
 		return -ENODEV;
 	}
 
+	if ((dev->ctrl_config & NVME_CC_CSS_LIGHTNVM) &&
+		id->nsfeat & NVME_NS_FEAT_NVM && ns->type != NVME_NS_NVM) {
+		if (nvme_nvm_register(ns->queue, disk)) {
+			dev_warn(dev->dev,
+				    "%s: LightNVM init failure\n", __func__);
+			return -ENODEV;
+		}
+		ns->type = NVME_NS_NVM;
+	}
+
 	old_ms = ns->ms;
 	lbaf = id->flbas & NVME_NS_FLBAS_LBA_MASK;
 	ns->lba_shift = id->lbaf[lbaf].ds;
@@ -2028,7 +2040,7 @@ static int nvme_revalidate_disk(struct gendisk *disk)
 								!ns->ext)
 		nvme_init_integrity(ns);
 
-	if (ns->ms && !blk_get_integrity(disk))
+	if ((ns->ms && !blk_get_integrity(disk)) || ns->type == NVME_NS_NVM)
 		set_capacity(disk, 0);
 	else
 		set_capacity(disk, le64_to_cpup(&id->nsze) << (ns->lba_shift - 9));
@@ -2146,7 +2158,8 @@ static void nvme_alloc_ns(struct nvme_dev *dev, unsigned nsid)
 	if (nvme_revalidate_disk(ns->disk))
 		goto out_free_disk;
 
-	add_disk(ns->disk);
+	if (ns->type != NVME_NS_NVM)
+		add_disk(ns->disk);
 	if (ns->ms) {
 		struct block_device *bd = bdget_disk(ns->disk, 0);
 		if (!bd)
@@ -2345,6 +2358,9 @@ static void nvme_free_namespace(struct nvme_ns *ns)
 {
 	list_del(&ns->list);
 
+	if (ns->type == NVME_NS_NVM)
+		nvm_unregister(ns->disk);
+
 	spin_lock(&dev_list_lock);
 	ns->disk->private_data = NULL;
 	spin_unlock(&dev_list_lock);
diff --git a/drivers/block/nvme-lightnvm.c b/drivers/block/nvme-lightnvm.c
new file mode 100644
index 0000000..397a66e
--- /dev/null
+++ b/drivers/block/nvme-lightnvm.c
@@ -0,0 +1,504 @@
+/*
+ * nvme-lightnvm.c - LightNVM NVMe device
+ *
+ * Copyright (C) 2014-2015 IT University of Copenhagen
+ * Initial release: Matias Bjorling <mb@lightnvm.io>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version
+ * 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; see the file COPYING.  If not, write to
+ * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139,
+ * USA.
+ *
+ */
+
+#include <linux/nvme.h>
+#include <linux/bitops.h>
+#include <linux/lightnvm.h>
+
+#ifdef CONFIG_NVM
+
+enum nvme_nvm_opcode {
+	nvme_nvm_cmd_hb_write	= 0x81,
+	nvme_nvm_cmd_hb_read	= 0x02,
+	nvme_nvm_cmd_phys_write	= 0x91,
+	nvme_nvm_cmd_phys_read	= 0x92,
+	nvme_nvm_cmd_erase	= 0x90,
+};
+
+enum nvme_nvm_admin_opcode {
+	nvme_nvm_admin_identify		= 0xe2,
+	nvme_nvm_admin_get_features	= 0xe6,
+	nvme_nvm_admin_set_resp		= 0xe5,
+	nvme_nvm_admin_get_l2p_tbl	= 0xea,
+	nvme_nvm_admin_get_bb_tbl	= 0xf2,
+	nvme_nvm_admin_set_bb_tbl	= 0xf1,
+};
+
+struct nvme_nvm_hb_rw {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd2;
+	__le64			metadata;
+	__le64			prp1;
+	__le64			prp2;
+	__le64			slba;
+	__le16			length;
+	__le16			control;
+	__le32			dsmgmt;
+	__le64			phys_addr;
+};
+
+struct nvme_nvm_identify {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd[2];
+	__le64			prp1;
+	__le64			prp2;
+	__le32			chnl_off;
+	__u32			rsvd11[5];
+};
+
+struct nvme_nvm_l2ptbl {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__le32			cdw2[4];
+	__le64			prp1;
+	__le64			prp2;
+	__le64			slba;
+	__le32			nlb;
+	__le16			cdw14[6];
+};
+
+struct nvme_nvm_bbtbl {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd[2];
+	__le64			prp1;
+	__le64			prp2;
+	__le32			prp1_len;
+	__le32			prp2_len;
+	__le32			lbb;
+	__u32			rsvd11[3];
+};
+
+struct nvme_nvm_set_resp {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd[2];
+	__le64			prp1;
+	__le64			prp2;
+	__le64			resp;
+	__u32			rsvd11[4];
+};
+
+struct nvme_nvm_erase_blk {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd[2];
+	__le64			prp1;
+	__le64			prp2;
+	__le64			blk_addr;
+	__u32			rsvd11[4];
+};
+
+struct nvme_nvm_command {
+	union {
+		struct nvme_common_command common;
+		struct nvme_nvm_identify nvm_identify;
+		struct nvme_nvm_hb_rw nvm_hb_rw;
+		struct nvme_nvm_l2ptbl nvm_l2p;
+		struct nvme_nvm_bbtbl nvm_get_bb;
+		struct nvme_nvm_bbtbl nvm_set_bb;
+		struct nvme_nvm_set_resp nvm_resp;
+		struct nvme_nvm_erase_blk nvm_erase;
+	};
+};
+
+/*
+ * Check we didin't inadvertently grow the command struct
+ */
+static inline void _nvme_nvm_check_size(void)
+{
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_identify) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_hb_rw) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_l2ptbl) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_bbtbl) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_set_resp) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_erase_blk) != 64);
+}
+
+struct nvme_nvm_id_chnl {
+	__le64			laddr_begin;
+	__le64			laddr_end;
+	__le32			oob_size;
+	__le32			queue_size;
+	__le32			gran_read;
+	__le32			gran_write;
+	__le32			gran_erase;
+	__le32			t_r;
+	__le32			t_sqr;
+	__le32			t_w;
+	__le32			t_sqw;
+	__le32			t_e;
+	__le16			chnl_parallelism;
+	__u8			io_sched;
+	__u8			reserved[133];
+} __packed;
+
+struct nvme_nvm_id {
+	__u8			ver_id;
+	__u8			nvm_type;
+	__le16			nchannels;
+	__u8			reserved[252];
+	struct nvme_nvm_id_chnl	chnls[];
+} __packed;
+
+#define NVME_NVM_CHNLS_PR_REQ ((4096U - sizeof(struct nvme_nvm_id)) \
+					/ sizeof(struct nvme_nvm_id_chnl))
+
+
+static int init_chnls(struct request_queue *q, struct nvm_id *nvm_id,
+						struct nvme_nvm_id *nvme_nvm_id)
+{
+	struct nvme_nvm_id_chnl *src = nvme_nvm_id->chnls;
+	struct nvm_id_chnl *dst = nvm_id->chnls;
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_command c = {
+		.nvm_identify.opcode = nvme_nvm_admin_identify,
+		.nvm_identify.nsid = cpu_to_le32(ns->ns_id),
+	};
+	unsigned int len = nvm_id->nchannels;
+	int i, end, ret, off = 0;
+
+	while (len) {
+		end = min_t(u32, NVME_NVM_CHNLS_PR_REQ, len);
+
+		for (i = 0; i < end; i++, dst++, src++) {
+			dst->laddr_begin = le64_to_cpu(src->laddr_begin);
+			dst->laddr_end = le64_to_cpu(src->laddr_end);
+			dst->oob_size = le32_to_cpu(src->oob_size);
+			dst->queue_size = le32_to_cpu(src->queue_size);
+			dst->gran_read = le32_to_cpu(src->gran_read);
+			dst->gran_write = le32_to_cpu(src->gran_write);
+			dst->gran_erase = le32_to_cpu(src->gran_erase);
+			dst->t_r = le32_to_cpu(src->t_r);
+			dst->t_sqr = le32_to_cpu(src->t_sqr);
+			dst->t_w = le32_to_cpu(src->t_w);
+			dst->t_sqw = le32_to_cpu(src->t_sqw);
+			dst->t_e = le32_to_cpu(src->t_e);
+			dst->io_sched = src->io_sched;
+		}
+
+		len -= end;
+		if (!len)
+			break;
+
+		off += end;
+
+		c.nvm_identify.chnl_off = off;
+
+		ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c,
+							nvme_nvm_id, 4096);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+static int nvme_nvm_identify(struct request_queue *q, struct nvm_id *nvm_id)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_id *nvme_nvm_id;
+	struct nvme_nvm_command c = {
+		.nvm_identify.opcode = nvme_nvm_admin_identify,
+		.nvm_identify.nsid = cpu_to_le32(ns->ns_id),
+		.nvm_identify.chnl_off = 0,
+	};
+	int ret;
+
+	nvme_nvm_id = kmalloc(4096, GFP_KERNEL);
+	if (!nvme_nvm_id)
+		return -ENOMEM;
+
+	ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, nvme_nvm_id,
+									4096);
+	if (ret) {
+		ret = -EIO;
+		goto out;
+	}
+
+	nvm_id->ver_id = nvme_nvm_id->ver_id;
+	nvm_id->nvm_type = nvme_nvm_id->nvm_type;
+	nvm_id->nchannels = le16_to_cpu(nvme_nvm_id->nchannels);
+
+	if (!nvm_id->chnls)
+		nvm_id->chnls = kmalloc(sizeof(struct nvm_id_chnl)
+					* nvm_id->nchannels, GFP_KERNEL);
+	if (!nvm_id->chnls) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = init_chnls(q, nvm_id, nvme_nvm_id);
+out:
+	kfree(nvme_nvm_id);
+	return ret;
+}
+
+static int nvme_nvm_get_features(struct request_queue *q,
+						struct nvm_get_features *gf)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_command c = {
+		.common.opcode = nvme_nvm_admin_get_features,
+		.common.nsid = ns->ns_id,
+	};
+	int sz = sizeof(struct nvm_get_features);
+	int ret;
+	u64 *resp;
+
+	resp = kmalloc(sz, GFP_KERNEL);
+	if (!resp)
+		return -ENOMEM;
+
+	ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, resp, sz);
+	if (ret)
+		goto done;
+
+	gf->rsp = le64_to_cpu(resp[0]);
+	gf->ext = le64_to_cpu(resp[1]);
+
+done:
+	kfree(resp);
+	return ret;
+}
+
+static int nvme_nvm_set_resp(struct request_queue *q, u64 resp)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_command c = {
+		.nvm_resp.opcode = nvme_nvm_admin_set_resp,
+		.nvm_resp.nsid = cpu_to_le32(ns->ns_id),
+		.nvm_resp.resp = cpu_to_le64(resp),
+	};
+
+	return nvme_submit_sync_cmd(q, (struct nvme_command *)&c, NULL, 0);
+}
+
+static int nvme_nvm_get_l2p_tbl(struct request_queue *q, u64 slba, u64 nlb,
+				nvm_l2p_update_fn *update_l2p, void *priv)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_dev *dev = ns->dev;
+	struct nvme_nvm_command c = {
+		.nvm_l2p.opcode = nvme_nvm_admin_get_l2p_tbl,
+		.nvm_l2p.nsid = cpu_to_le32(ns->ns_id),
+	};
+	u32 len = queue_max_hw_sectors(q) << 9;
+	u64 nlb_pr_rq = len / sizeof(u64);
+	u64 cmd_slba = slba;
+	void *entries;
+	int ret = 0;
+
+	entries = kmalloc(len, GFP_KERNEL);
+	if (!entries)
+		return -ENOMEM;
+
+	while (nlb) {
+		u64 cmd_nlb = min_t(u64, nlb_pr_rq, nlb);
+
+		c.nvm_l2p.slba = cmd_slba;
+		c.nvm_l2p.nlb = cmd_nlb;
+
+		ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c,
+								entries, len);
+		if (ret) {
+			dev_err(dev->dev, "L2P table transfer failed (%d)\n",
+									ret);
+			ret = -EIO;
+			goto out;
+		}
+
+		if (update_l2p(cmd_slba, cmd_nlb, entries, priv)) {
+			ret = -EINTR;
+			goto out;
+		}
+
+		cmd_slba += cmd_nlb;
+		nlb -= cmd_nlb;
+	}
+
+out:
+	kfree(entries);
+	return ret;
+}
+
+static int nvme_nvm_set_bb_tbl(struct request_queue *q, int lunid,
+	unsigned int nr_blocks, nvm_bb_update_fn *update_bbtbl, void *priv)
+{
+	return 0;
+}
+
+static int nvme_nvm_get_bb_tbl(struct request_queue *q, int lunid,
+	unsigned int nr_blocks, nvm_bb_update_fn *update_bbtbl, void *priv)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_dev *dev = ns->dev;
+	struct nvme_nvm_command c = {
+		.nvm_get_bb.opcode = nvme_nvm_admin_get_bb_tbl,
+		.nvm_get_bb.nsid = cpu_to_le32(ns->ns_id),
+		.nvm_get_bb.lbb = cpu_to_le32(lunid),
+	};
+	void *bb_bitmap;
+	u16 bb_bitmap_size;
+	int ret = 0;
+
+	bb_bitmap_size = ((nr_blocks >> 15) + 1) * PAGE_SIZE;
+	bb_bitmap = kmalloc(bb_bitmap_size, GFP_KERNEL);
+	if (!bb_bitmap)
+		return -ENOMEM;
+
+	bitmap_zero(bb_bitmap, nr_blocks);
+
+	ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, bb_bitmap,
+								bb_bitmap_size);
+	if (ret) {
+		dev_err(dev->dev, "get bad block table failed (%d)\n", ret);
+		ret = -EIO;
+		goto out;
+	}
+
+	ret = update_bbtbl(lunid, bb_bitmap, nr_blocks, priv);
+	if (ret) {
+		ret = -EINTR;
+		goto out;
+	}
+
+out:
+	kfree(bb_bitmap);
+	return ret;
+}
+
+static inline void nvme_nvm_rqtocmd(struct request *rq, struct nvm_rq *rqd,
+				struct nvme_ns *ns, struct nvme_nvm_command *c)
+{
+	c->nvm_hb_rw.opcode = (rq_data_dir(rq) ?
+				nvme_nvm_cmd_hb_write : nvme_nvm_cmd_hb_read);
+	c->nvm_hb_rw.nsid = cpu_to_le32(ns->ns_id);
+	c->nvm_hb_rw.slba = cpu_to_le64(nvme_block_nr(ns,
+						rqd->bio->bi_iter.bi_sector));
+	c->nvm_hb_rw.length = cpu_to_le16(
+		(blk_rq_bytes(rq) >> ns->lba_shift) - 1);
+	c->nvm_hb_rw.phys_addr =
+		cpu_to_le64(nvme_block_nr(ns, rqd->phys_sector));
+}
+
+static void nvme_nvm_end_io(struct request *rq, int error)
+{
+	struct nvm_rq *rqd = rq->end_io_data;
+	struct nvm_tgt_instance *ins = rqd->ins;
+
+	ins->tt->end_io(rq->end_io_data, error);
+
+	kfree(rq->cmd);
+	blk_mq_free_request(rq);
+}
+
+static int nvme_nvm_submit_io(struct request_queue *q, struct nvm_rq *rqd)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct request *rq;
+	struct bio *bio = rqd->bio;
+	struct nvme_nvm_command *cmd;
+
+	rq = blk_mq_alloc_request(q, bio_rw(bio), GFP_KERNEL, 0);
+	if (IS_ERR(rq))
+		return -ENOMEM;
+
+	cmd = kzalloc(sizeof(struct nvme_nvm_command), GFP_KERNEL);
+	if (!cmd) {
+		blk_mq_free_request(rq);
+		return -ENOMEM;
+	}
+
+	rq->cmd_type = REQ_TYPE_DRV_PRIV;
+	rq->ioprio = bio_prio(bio);
+
+	if (bio_has_data(bio))
+		rq->nr_phys_segments = bio_phys_segments(q, bio);
+
+	rq->__data_len = bio->bi_iter.bi_size;
+	rq->bio = rq->biotail = bio;
+
+	nvme_nvm_rqtocmd(rq, rqd, ns, cmd);
+
+	rq->cmd = (unsigned char *)cmd;
+	rq->cmd_len = sizeof(struct nvme_nvm_command);
+	rq->special = (void *)0;
+
+	rq->end_io_data = rqd;
+
+	blk_execute_rq_nowait(q, NULL, rq, 0, nvme_nvm_end_io);
+
+	return 0;
+}
+
+static int nvme_nvm_erase_block(struct request_queue *q, sector_t block_id)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_command c = {
+		.nvm_erase.opcode = nvme_nvm_cmd_erase,
+		.nvm_erase.nsid = cpu_to_le32(ns->ns_id),
+		.nvm_erase.blk_addr = cpu_to_le64(block_id),
+	};
+
+	return nvme_submit_sync_cmd(q, (struct nvme_command *)&c, NULL, 0);
+}
+
+static struct nvm_dev_ops nvme_nvm_dev_ops = {
+	.identify		= nvme_nvm_identify,
+
+	.get_features		= nvme_nvm_get_features,
+	.set_responsibility	= nvme_nvm_set_resp,
+
+	.get_l2p_tbl		= nvme_nvm_get_l2p_tbl,
+
+	.set_bb_tbl		= nvme_nvm_set_bb_tbl,
+	.get_bb_tbl		= nvme_nvm_get_bb_tbl,
+
+	.submit_io		= nvme_nvm_submit_io,
+	.erase_block		= nvme_nvm_erase_block,
+};
+
+int nvme_nvm_register(struct request_queue *q, struct gendisk *disk)
+{
+	return nvm_register(q, disk, &nvme_nvm_dev_ops);
+}
+#else
+int nvme_nvm_register(struct request_queue *q, struct gendisk *disk)
+{
+	return 0;
+}
+
+#endif /* CONFIG_NVM */
diff --git a/include/linux/nvme.h b/include/linux/nvme.h
index fa3fe16..fa58242 100644
--- a/include/linux/nvme.h
+++ b/include/linux/nvme.h
@@ -19,6 +19,7 @@
 #include <linux/pci.h>
 #include <linux/kref.h>
 #include <linux/blk-mq.h>
+#include <linux/lightnvm.h>
 
 struct nvme_bar {
 	__u64			cap;	/* Controller Capabilities */
@@ -41,6 +42,7 @@ struct nvme_bar {
 #define NVME_CAP_STRIDE(cap)	(((cap) >> 32) & 0xf)
 #define NVME_CAP_MPSMIN(cap)	(((cap) >> 48) & 0xf)
 #define NVME_CAP_MPSMAX(cap)	(((cap) >> 52) & 0xf)
+#define NVME_CAP_LIGHTNVM(cap)	(((cap) >> 38) & 0x1)
 
 #define NVME_CMB_BIR(cmbloc)	((cmbloc) & 0x7)
 #define NVME_CMB_OFST(cmbloc)	(((cmbloc) >> 12) & 0xfffff)
@@ -56,6 +58,7 @@ struct nvme_bar {
 enum {
 	NVME_CC_ENABLE		= 1 << 0,
 	NVME_CC_CSS_NVM		= 0 << 4,
+	NVME_CC_CSS_LIGHTNVM	= 1 << 4,
 	NVME_CC_MPS_SHIFT	= 7,
 	NVME_CC_ARB_RR		= 0 << 11,
 	NVME_CC_ARB_WRRU	= 1 << 11,
@@ -138,6 +141,7 @@ struct nvme_ns {
 	u16 ms;
 	bool ext;
 	u8 pi_type;
+	int type;
 	u64 mode_select_num_blocks;
 	u32 mode_select_block_len;
 };
@@ -184,4 +188,6 @@ int nvme_sg_io(struct nvme_ns *ns, struct sg_io_hdr __user *u_hdr);
 int nvme_sg_io32(struct nvme_ns *ns, unsigned long arg);
 int nvme_sg_get_version_num(int __user *ip);
 
+int nvme_nvm_register(struct request_queue *q, struct gendisk *disk);
+
 #endif /* _LINUX_NVME_H */
diff --git a/include/uapi/linux/nvme.h b/include/uapi/linux/nvme.h
index 732b32e..0374f11 100644
--- a/include/uapi/linux/nvme.h
+++ b/include/uapi/linux/nvme.h
@@ -130,6 +130,7 @@ struct nvme_id_ns {
 
 enum {
 	NVME_NS_FEAT_THIN	= 1 << 0,
+	NVME_NS_FEAT_NVM	= 1 << 3,
 	NVME_NS_FLBAS_LBA_MASK	= 0xf,
 	NVME_NS_FLBAS_META_EXT	= 0x10,
 	NVME_LBAF_RP_BEST	= 0,
@@ -146,6 +147,8 @@ enum {
 	NVME_NS_DPS_PI_TYPE1	= 1,
 	NVME_NS_DPS_PI_TYPE2	= 2,
 	NVME_NS_DPS_PI_TYPE3	= 3,
+
+	NVME_NS_NVM		= 1,
 };
 
 struct nvme_smart_log {
-- 
2.1.4


WARNING: multiple messages have this Message-ID (diff)
From: mb@lightnvm.io (Matias Bjørling)
Subject: [PATCH v5 5/5] nvme: LightNVM support
Date: Wed, 22 Jul 2015 19:51:04 +0200	[thread overview]
Message-ID: <1437587464-7964-6-git-send-email-mb@lightnvm.io> (raw)
In-Reply-To: <1437587464-7964-1-git-send-email-mb@lightnvm.io>

The first generation of Open-Channel SSDs will be based on NVMe. The
integration requires that a NVMe device exposes itself as a LightNVM
device. The way this is done currently is by hooking into the
Controller Capabilities (CAP register) and a bit in NSFEAT for each
namespace.

After detection, vendor specific codes are used to identify the device
and enumerate supported features.

Signed-off-by: Javier Gonz?lez <jg at lightnvm.io>
Signed-off-by: Matias Bj?rling <mb at lightnvm.io>
---
 drivers/block/Makefile        |   2 +-
 drivers/block/nvme-core.c     |  22 +-
 drivers/block/nvme-lightnvm.c | 504 ++++++++++++++++++++++++++++++++++++++++++
 include/linux/nvme.h          |   6 +
 include/uapi/linux/nvme.h     |   3 +
 5 files changed, 533 insertions(+), 4 deletions(-)
 create mode 100644 drivers/block/nvme-lightnvm.c

diff --git a/drivers/block/Makefile b/drivers/block/Makefile
index 02b688d..a01d7d8 100644
--- a/drivers/block/Makefile
+++ b/drivers/block/Makefile
@@ -44,6 +44,6 @@ obj-$(CONFIG_BLK_DEV_RSXX) += rsxx/
 obj-$(CONFIG_BLK_DEV_NULL_BLK)	+= null_blk.o
 obj-$(CONFIG_ZRAM) += zram/
 
-nvme-y		:= nvme-core.o nvme-scsi.o
+nvme-y		:= nvme-core.o nvme-scsi.o nvme-lightnvm.o
 skd-y		:= skd_main.o
 swim_mod-y	:= swim.o swim_asm.o
diff --git a/drivers/block/nvme-core.c b/drivers/block/nvme-core.c
index 666e994..ab799b0 100644
--- a/drivers/block/nvme-core.c
+++ b/drivers/block/nvme-core.c
@@ -40,6 +40,7 @@
 #include <linux/slab.h>
 #include <linux/t10-pi.h>
 #include <linux/types.h>
+#include <linux/lightnvm.h>
 #include <scsi/sg.h>
 #include <asm-generic/io-64-nonatomic-lo-hi.h>
 
@@ -1751,7 +1752,8 @@ static int nvme_configure_admin_queue(struct nvme_dev *dev)
 
 	dev->page_size = 1 << page_shift;
 
-	dev->ctrl_config = NVME_CC_CSS_NVM;
+	dev->ctrl_config = NVME_CAP_LIGHTNVM(cap) ?
+					NVME_CC_CSS_LIGHTNVM : NVME_CC_CSS_NVM;
 	dev->ctrl_config |= (page_shift - 12) << NVME_CC_MPS_SHIFT;
 	dev->ctrl_config |= NVME_CC_ARB_RR | NVME_CC_SHN_NONE;
 	dev->ctrl_config |= NVME_CC_IOSQES | NVME_CC_IOCQES;
@@ -1997,6 +1999,16 @@ static int nvme_revalidate_disk(struct gendisk *disk)
 		return -ENODEV;
 	}
 
+	if ((dev->ctrl_config & NVME_CC_CSS_LIGHTNVM) &&
+		id->nsfeat & NVME_NS_FEAT_NVM && ns->type != NVME_NS_NVM) {
+		if (nvme_nvm_register(ns->queue, disk)) {
+			dev_warn(dev->dev,
+				    "%s: LightNVM init failure\n", __func__);
+			return -ENODEV;
+		}
+		ns->type = NVME_NS_NVM;
+	}
+
 	old_ms = ns->ms;
 	lbaf = id->flbas & NVME_NS_FLBAS_LBA_MASK;
 	ns->lba_shift = id->lbaf[lbaf].ds;
@@ -2028,7 +2040,7 @@ static int nvme_revalidate_disk(struct gendisk *disk)
 								!ns->ext)
 		nvme_init_integrity(ns);
 
-	if (ns->ms && !blk_get_integrity(disk))
+	if ((ns->ms && !blk_get_integrity(disk)) || ns->type == NVME_NS_NVM)
 		set_capacity(disk, 0);
 	else
 		set_capacity(disk, le64_to_cpup(&id->nsze) << (ns->lba_shift - 9));
@@ -2146,7 +2158,8 @@ static void nvme_alloc_ns(struct nvme_dev *dev, unsigned nsid)
 	if (nvme_revalidate_disk(ns->disk))
 		goto out_free_disk;
 
-	add_disk(ns->disk);
+	if (ns->type != NVME_NS_NVM)
+		add_disk(ns->disk);
 	if (ns->ms) {
 		struct block_device *bd = bdget_disk(ns->disk, 0);
 		if (!bd)
@@ -2345,6 +2358,9 @@ static void nvme_free_namespace(struct nvme_ns *ns)
 {
 	list_del(&ns->list);
 
+	if (ns->type == NVME_NS_NVM)
+		nvm_unregister(ns->disk);
+
 	spin_lock(&dev_list_lock);
 	ns->disk->private_data = NULL;
 	spin_unlock(&dev_list_lock);
diff --git a/drivers/block/nvme-lightnvm.c b/drivers/block/nvme-lightnvm.c
new file mode 100644
index 0000000..397a66e
--- /dev/null
+++ b/drivers/block/nvme-lightnvm.c
@@ -0,0 +1,504 @@
+/*
+ * nvme-lightnvm.c - LightNVM NVMe device
+ *
+ * Copyright (C) 2014-2015 IT University of Copenhagen
+ * Initial release: Matias Bjorling <mb at lightnvm.io>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version
+ * 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; see the file COPYING.  If not, write to
+ * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139,
+ * USA.
+ *
+ */
+
+#include <linux/nvme.h>
+#include <linux/bitops.h>
+#include <linux/lightnvm.h>
+
+#ifdef CONFIG_NVM
+
+enum nvme_nvm_opcode {
+	nvme_nvm_cmd_hb_write	= 0x81,
+	nvme_nvm_cmd_hb_read	= 0x02,
+	nvme_nvm_cmd_phys_write	= 0x91,
+	nvme_nvm_cmd_phys_read	= 0x92,
+	nvme_nvm_cmd_erase	= 0x90,
+};
+
+enum nvme_nvm_admin_opcode {
+	nvme_nvm_admin_identify		= 0xe2,
+	nvme_nvm_admin_get_features	= 0xe6,
+	nvme_nvm_admin_set_resp		= 0xe5,
+	nvme_nvm_admin_get_l2p_tbl	= 0xea,
+	nvme_nvm_admin_get_bb_tbl	= 0xf2,
+	nvme_nvm_admin_set_bb_tbl	= 0xf1,
+};
+
+struct nvme_nvm_hb_rw {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd2;
+	__le64			metadata;
+	__le64			prp1;
+	__le64			prp2;
+	__le64			slba;
+	__le16			length;
+	__le16			control;
+	__le32			dsmgmt;
+	__le64			phys_addr;
+};
+
+struct nvme_nvm_identify {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd[2];
+	__le64			prp1;
+	__le64			prp2;
+	__le32			chnl_off;
+	__u32			rsvd11[5];
+};
+
+struct nvme_nvm_l2ptbl {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__le32			cdw2[4];
+	__le64			prp1;
+	__le64			prp2;
+	__le64			slba;
+	__le32			nlb;
+	__le16			cdw14[6];
+};
+
+struct nvme_nvm_bbtbl {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd[2];
+	__le64			prp1;
+	__le64			prp2;
+	__le32			prp1_len;
+	__le32			prp2_len;
+	__le32			lbb;
+	__u32			rsvd11[3];
+};
+
+struct nvme_nvm_set_resp {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd[2];
+	__le64			prp1;
+	__le64			prp2;
+	__le64			resp;
+	__u32			rsvd11[4];
+};
+
+struct nvme_nvm_erase_blk {
+	__u8			opcode;
+	__u8			flags;
+	__u16			command_id;
+	__le32			nsid;
+	__u64			rsvd[2];
+	__le64			prp1;
+	__le64			prp2;
+	__le64			blk_addr;
+	__u32			rsvd11[4];
+};
+
+struct nvme_nvm_command {
+	union {
+		struct nvme_common_command common;
+		struct nvme_nvm_identify nvm_identify;
+		struct nvme_nvm_hb_rw nvm_hb_rw;
+		struct nvme_nvm_l2ptbl nvm_l2p;
+		struct nvme_nvm_bbtbl nvm_get_bb;
+		struct nvme_nvm_bbtbl nvm_set_bb;
+		struct nvme_nvm_set_resp nvm_resp;
+		struct nvme_nvm_erase_blk nvm_erase;
+	};
+};
+
+/*
+ * Check we didin't inadvertently grow the command struct
+ */
+static inline void _nvme_nvm_check_size(void)
+{
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_identify) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_hb_rw) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_l2ptbl) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_bbtbl) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_set_resp) != 64);
+	BUILD_BUG_ON(sizeof(struct nvme_nvm_erase_blk) != 64);
+}
+
+struct nvme_nvm_id_chnl {
+	__le64			laddr_begin;
+	__le64			laddr_end;
+	__le32			oob_size;
+	__le32			queue_size;
+	__le32			gran_read;
+	__le32			gran_write;
+	__le32			gran_erase;
+	__le32			t_r;
+	__le32			t_sqr;
+	__le32			t_w;
+	__le32			t_sqw;
+	__le32			t_e;
+	__le16			chnl_parallelism;
+	__u8			io_sched;
+	__u8			reserved[133];
+} __packed;
+
+struct nvme_nvm_id {
+	__u8			ver_id;
+	__u8			nvm_type;
+	__le16			nchannels;
+	__u8			reserved[252];
+	struct nvme_nvm_id_chnl	chnls[];
+} __packed;
+
+#define NVME_NVM_CHNLS_PR_REQ ((4096U - sizeof(struct nvme_nvm_id)) \
+					/ sizeof(struct nvme_nvm_id_chnl))
+
+
+static int init_chnls(struct request_queue *q, struct nvm_id *nvm_id,
+						struct nvme_nvm_id *nvme_nvm_id)
+{
+	struct nvme_nvm_id_chnl *src = nvme_nvm_id->chnls;
+	struct nvm_id_chnl *dst = nvm_id->chnls;
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_command c = {
+		.nvm_identify.opcode = nvme_nvm_admin_identify,
+		.nvm_identify.nsid = cpu_to_le32(ns->ns_id),
+	};
+	unsigned int len = nvm_id->nchannels;
+	int i, end, ret, off = 0;
+
+	while (len) {
+		end = min_t(u32, NVME_NVM_CHNLS_PR_REQ, len);
+
+		for (i = 0; i < end; i++, dst++, src++) {
+			dst->laddr_begin = le64_to_cpu(src->laddr_begin);
+			dst->laddr_end = le64_to_cpu(src->laddr_end);
+			dst->oob_size = le32_to_cpu(src->oob_size);
+			dst->queue_size = le32_to_cpu(src->queue_size);
+			dst->gran_read = le32_to_cpu(src->gran_read);
+			dst->gran_write = le32_to_cpu(src->gran_write);
+			dst->gran_erase = le32_to_cpu(src->gran_erase);
+			dst->t_r = le32_to_cpu(src->t_r);
+			dst->t_sqr = le32_to_cpu(src->t_sqr);
+			dst->t_w = le32_to_cpu(src->t_w);
+			dst->t_sqw = le32_to_cpu(src->t_sqw);
+			dst->t_e = le32_to_cpu(src->t_e);
+			dst->io_sched = src->io_sched;
+		}
+
+		len -= end;
+		if (!len)
+			break;
+
+		off += end;
+
+		c.nvm_identify.chnl_off = off;
+
+		ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c,
+							nvme_nvm_id, 4096);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+static int nvme_nvm_identify(struct request_queue *q, struct nvm_id *nvm_id)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_id *nvme_nvm_id;
+	struct nvme_nvm_command c = {
+		.nvm_identify.opcode = nvme_nvm_admin_identify,
+		.nvm_identify.nsid = cpu_to_le32(ns->ns_id),
+		.nvm_identify.chnl_off = 0,
+	};
+	int ret;
+
+	nvme_nvm_id = kmalloc(4096, GFP_KERNEL);
+	if (!nvme_nvm_id)
+		return -ENOMEM;
+
+	ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, nvme_nvm_id,
+									4096);
+	if (ret) {
+		ret = -EIO;
+		goto out;
+	}
+
+	nvm_id->ver_id = nvme_nvm_id->ver_id;
+	nvm_id->nvm_type = nvme_nvm_id->nvm_type;
+	nvm_id->nchannels = le16_to_cpu(nvme_nvm_id->nchannels);
+
+	if (!nvm_id->chnls)
+		nvm_id->chnls = kmalloc(sizeof(struct nvm_id_chnl)
+					* nvm_id->nchannels, GFP_KERNEL);
+	if (!nvm_id->chnls) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = init_chnls(q, nvm_id, nvme_nvm_id);
+out:
+	kfree(nvme_nvm_id);
+	return ret;
+}
+
+static int nvme_nvm_get_features(struct request_queue *q,
+						struct nvm_get_features *gf)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_command c = {
+		.common.opcode = nvme_nvm_admin_get_features,
+		.common.nsid = ns->ns_id,
+	};
+	int sz = sizeof(struct nvm_get_features);
+	int ret;
+	u64 *resp;
+
+	resp = kmalloc(sz, GFP_KERNEL);
+	if (!resp)
+		return -ENOMEM;
+
+	ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, resp, sz);
+	if (ret)
+		goto done;
+
+	gf->rsp = le64_to_cpu(resp[0]);
+	gf->ext = le64_to_cpu(resp[1]);
+
+done:
+	kfree(resp);
+	return ret;
+}
+
+static int nvme_nvm_set_resp(struct request_queue *q, u64 resp)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_command c = {
+		.nvm_resp.opcode = nvme_nvm_admin_set_resp,
+		.nvm_resp.nsid = cpu_to_le32(ns->ns_id),
+		.nvm_resp.resp = cpu_to_le64(resp),
+	};
+
+	return nvme_submit_sync_cmd(q, (struct nvme_command *)&c, NULL, 0);
+}
+
+static int nvme_nvm_get_l2p_tbl(struct request_queue *q, u64 slba, u64 nlb,
+				nvm_l2p_update_fn *update_l2p, void *priv)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_dev *dev = ns->dev;
+	struct nvme_nvm_command c = {
+		.nvm_l2p.opcode = nvme_nvm_admin_get_l2p_tbl,
+		.nvm_l2p.nsid = cpu_to_le32(ns->ns_id),
+	};
+	u32 len = queue_max_hw_sectors(q) << 9;
+	u64 nlb_pr_rq = len / sizeof(u64);
+	u64 cmd_slba = slba;
+	void *entries;
+	int ret = 0;
+
+	entries = kmalloc(len, GFP_KERNEL);
+	if (!entries)
+		return -ENOMEM;
+
+	while (nlb) {
+		u64 cmd_nlb = min_t(u64, nlb_pr_rq, nlb);
+
+		c.nvm_l2p.slba = cmd_slba;
+		c.nvm_l2p.nlb = cmd_nlb;
+
+		ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c,
+								entries, len);
+		if (ret) {
+			dev_err(dev->dev, "L2P table transfer failed (%d)\n",
+									ret);
+			ret = -EIO;
+			goto out;
+		}
+
+		if (update_l2p(cmd_slba, cmd_nlb, entries, priv)) {
+			ret = -EINTR;
+			goto out;
+		}
+
+		cmd_slba += cmd_nlb;
+		nlb -= cmd_nlb;
+	}
+
+out:
+	kfree(entries);
+	return ret;
+}
+
+static int nvme_nvm_set_bb_tbl(struct request_queue *q, int lunid,
+	unsigned int nr_blocks, nvm_bb_update_fn *update_bbtbl, void *priv)
+{
+	return 0;
+}
+
+static int nvme_nvm_get_bb_tbl(struct request_queue *q, int lunid,
+	unsigned int nr_blocks, nvm_bb_update_fn *update_bbtbl, void *priv)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_dev *dev = ns->dev;
+	struct nvme_nvm_command c = {
+		.nvm_get_bb.opcode = nvme_nvm_admin_get_bb_tbl,
+		.nvm_get_bb.nsid = cpu_to_le32(ns->ns_id),
+		.nvm_get_bb.lbb = cpu_to_le32(lunid),
+	};
+	void *bb_bitmap;
+	u16 bb_bitmap_size;
+	int ret = 0;
+
+	bb_bitmap_size = ((nr_blocks >> 15) + 1) * PAGE_SIZE;
+	bb_bitmap = kmalloc(bb_bitmap_size, GFP_KERNEL);
+	if (!bb_bitmap)
+		return -ENOMEM;
+
+	bitmap_zero(bb_bitmap, nr_blocks);
+
+	ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, bb_bitmap,
+								bb_bitmap_size);
+	if (ret) {
+		dev_err(dev->dev, "get bad block table failed (%d)\n", ret);
+		ret = -EIO;
+		goto out;
+	}
+
+	ret = update_bbtbl(lunid, bb_bitmap, nr_blocks, priv);
+	if (ret) {
+		ret = -EINTR;
+		goto out;
+	}
+
+out:
+	kfree(bb_bitmap);
+	return ret;
+}
+
+static inline void nvme_nvm_rqtocmd(struct request *rq, struct nvm_rq *rqd,
+				struct nvme_ns *ns, struct nvme_nvm_command *c)
+{
+	c->nvm_hb_rw.opcode = (rq_data_dir(rq) ?
+				nvme_nvm_cmd_hb_write : nvme_nvm_cmd_hb_read);
+	c->nvm_hb_rw.nsid = cpu_to_le32(ns->ns_id);
+	c->nvm_hb_rw.slba = cpu_to_le64(nvme_block_nr(ns,
+						rqd->bio->bi_iter.bi_sector));
+	c->nvm_hb_rw.length = cpu_to_le16(
+		(blk_rq_bytes(rq) >> ns->lba_shift) - 1);
+	c->nvm_hb_rw.phys_addr =
+		cpu_to_le64(nvme_block_nr(ns, rqd->phys_sector));
+}
+
+static void nvme_nvm_end_io(struct request *rq, int error)
+{
+	struct nvm_rq *rqd = rq->end_io_data;
+	struct nvm_tgt_instance *ins = rqd->ins;
+
+	ins->tt->end_io(rq->end_io_data, error);
+
+	kfree(rq->cmd);
+	blk_mq_free_request(rq);
+}
+
+static int nvme_nvm_submit_io(struct request_queue *q, struct nvm_rq *rqd)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct request *rq;
+	struct bio *bio = rqd->bio;
+	struct nvme_nvm_command *cmd;
+
+	rq = blk_mq_alloc_request(q, bio_rw(bio), GFP_KERNEL, 0);
+	if (IS_ERR(rq))
+		return -ENOMEM;
+
+	cmd = kzalloc(sizeof(struct nvme_nvm_command), GFP_KERNEL);
+	if (!cmd) {
+		blk_mq_free_request(rq);
+		return -ENOMEM;
+	}
+
+	rq->cmd_type = REQ_TYPE_DRV_PRIV;
+	rq->ioprio = bio_prio(bio);
+
+	if (bio_has_data(bio))
+		rq->nr_phys_segments = bio_phys_segments(q, bio);
+
+	rq->__data_len = bio->bi_iter.bi_size;
+	rq->bio = rq->biotail = bio;
+
+	nvme_nvm_rqtocmd(rq, rqd, ns, cmd);
+
+	rq->cmd = (unsigned char *)cmd;
+	rq->cmd_len = sizeof(struct nvme_nvm_command);
+	rq->special = (void *)0;
+
+	rq->end_io_data = rqd;
+
+	blk_execute_rq_nowait(q, NULL, rq, 0, nvme_nvm_end_io);
+
+	return 0;
+}
+
+static int nvme_nvm_erase_block(struct request_queue *q, sector_t block_id)
+{
+	struct nvme_ns *ns = q->queuedata;
+	struct nvme_nvm_command c = {
+		.nvm_erase.opcode = nvme_nvm_cmd_erase,
+		.nvm_erase.nsid = cpu_to_le32(ns->ns_id),
+		.nvm_erase.blk_addr = cpu_to_le64(block_id),
+	};
+
+	return nvme_submit_sync_cmd(q, (struct nvme_command *)&c, NULL, 0);
+}
+
+static struct nvm_dev_ops nvme_nvm_dev_ops = {
+	.identify		= nvme_nvm_identify,
+
+	.get_features		= nvme_nvm_get_features,
+	.set_responsibility	= nvme_nvm_set_resp,
+
+	.get_l2p_tbl		= nvme_nvm_get_l2p_tbl,
+
+	.set_bb_tbl		= nvme_nvm_set_bb_tbl,
+	.get_bb_tbl		= nvme_nvm_get_bb_tbl,
+
+	.submit_io		= nvme_nvm_submit_io,
+	.erase_block		= nvme_nvm_erase_block,
+};
+
+int nvme_nvm_register(struct request_queue *q, struct gendisk *disk)
+{
+	return nvm_register(q, disk, &nvme_nvm_dev_ops);
+}
+#else
+int nvme_nvm_register(struct request_queue *q, struct gendisk *disk)
+{
+	return 0;
+}
+
+#endif /* CONFIG_NVM */
diff --git a/include/linux/nvme.h b/include/linux/nvme.h
index fa3fe16..fa58242 100644
--- a/include/linux/nvme.h
+++ b/include/linux/nvme.h
@@ -19,6 +19,7 @@
 #include <linux/pci.h>
 #include <linux/kref.h>
 #include <linux/blk-mq.h>
+#include <linux/lightnvm.h>
 
 struct nvme_bar {
 	__u64			cap;	/* Controller Capabilities */
@@ -41,6 +42,7 @@ struct nvme_bar {
 #define NVME_CAP_STRIDE(cap)	(((cap) >> 32) & 0xf)
 #define NVME_CAP_MPSMIN(cap)	(((cap) >> 48) & 0xf)
 #define NVME_CAP_MPSMAX(cap)	(((cap) >> 52) & 0xf)
+#define NVME_CAP_LIGHTNVM(cap)	(((cap) >> 38) & 0x1)
 
 #define NVME_CMB_BIR(cmbloc)	((cmbloc) & 0x7)
 #define NVME_CMB_OFST(cmbloc)	(((cmbloc) >> 12) & 0xfffff)
@@ -56,6 +58,7 @@ struct nvme_bar {
 enum {
 	NVME_CC_ENABLE		= 1 << 0,
 	NVME_CC_CSS_NVM		= 0 << 4,
+	NVME_CC_CSS_LIGHTNVM	= 1 << 4,
 	NVME_CC_MPS_SHIFT	= 7,
 	NVME_CC_ARB_RR		= 0 << 11,
 	NVME_CC_ARB_WRRU	= 1 << 11,
@@ -138,6 +141,7 @@ struct nvme_ns {
 	u16 ms;
 	bool ext;
 	u8 pi_type;
+	int type;
 	u64 mode_select_num_blocks;
 	u32 mode_select_block_len;
 };
@@ -184,4 +188,6 @@ int nvme_sg_io(struct nvme_ns *ns, struct sg_io_hdr __user *u_hdr);
 int nvme_sg_io32(struct nvme_ns *ns, unsigned long arg);
 int nvme_sg_get_version_num(int __user *ip);
 
+int nvme_nvm_register(struct request_queue *q, struct gendisk *disk);
+
 #endif /* _LINUX_NVME_H */
diff --git a/include/uapi/linux/nvme.h b/include/uapi/linux/nvme.h
index 732b32e..0374f11 100644
--- a/include/uapi/linux/nvme.h
+++ b/include/uapi/linux/nvme.h
@@ -130,6 +130,7 @@ struct nvme_id_ns {
 
 enum {
 	NVME_NS_FEAT_THIN	= 1 << 0,
+	NVME_NS_FEAT_NVM	= 1 << 3,
 	NVME_NS_FLBAS_LBA_MASK	= 0xf,
 	NVME_NS_FLBAS_META_EXT	= 0x10,
 	NVME_LBAF_RP_BEST	= 0,
@@ -146,6 +147,8 @@ enum {
 	NVME_NS_DPS_PI_TYPE1	= 1,
 	NVME_NS_DPS_PI_TYPE2	= 2,
 	NVME_NS_DPS_PI_TYPE3	= 3,
+
+	NVME_NS_NVM		= 1,
 };
 
 struct nvme_smart_log {
-- 
2.1.4

  parent reply	other threads:[~2015-07-22 17:52 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-07-22 17:50 [PATCH v5 0/5] Support for Open-Channel SSDs Matias Bjørling
2015-07-22 17:50 ` Matias Bjørling
2015-07-22 17:51 ` [PATCH v5 1/5] lightnvm: " Matias Bjørling
2015-07-22 17:51   ` Matias Bjørling
2015-07-22 17:51   ` Matias Bjørling
2015-07-23  9:53   ` Christoph Hellwig
2015-07-23  9:53     ` Christoph Hellwig
2015-07-23 10:38     ` Matias Bjørling
2015-07-23 10:38       ` Matias Bjørling
2015-07-22 17:51 ` [PATCH v5 2/5] rrpc: Hybrid Open-Channel SSD RRPC target Matias Bjørling
2015-07-22 17:51   ` Matias Bjørling
2015-07-22 17:51 ` [PATCH v5 3/5] bm_hb: Hybrid Open-Channel SSD block manager Matias Bjørling
2015-07-22 17:51   ` Matias Bjørling
2015-07-22 17:51 ` [PATCH v5 4/5] null_blk: LightNVM support Matias Bjørling
2015-07-22 17:51   ` Matias Bjørling
2015-07-23  9:53   ` Christoph Hellwig
2015-07-23  9:53     ` Christoph Hellwig
2015-07-23 10:48     ` Matias Bjørling
2015-07-23 10:48       ` Matias Bjørling
2015-07-25  6:25       ` Christoph Hellwig
2015-07-25  6:25         ` Christoph Hellwig
2015-07-22 17:51 ` Matias Bjørling [this message]
2015-07-22 17:51   ` [PATCH v5 5/5] nvme: " Matias Bjørling

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1437587464-7964-6-git-send-email-mb@lightnvm.io \
    --to=mb@lightnvm.io \
    --cc=Stephen.Bates@pmcs.com \
    --cc=axboe@fb.com \
    --cc=hch@infradead.org \
    --cc=jg@lightnvm.io \
    --cc=keith.busch@intel.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-nvme@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.