path: root/drivers/lightnvm/pblk.h

/*
 * Copyright (C) 2015 IT University of Copenhagen (rrpc.h)
 * Copyright (C) 2016 CNEX Labs
 * Initial release: Matias Bjorling <matias@cnexlabs.com>
 * Write buffering: Javier Gonzalez <javier@cnexlabs.com>
 *
 * 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.
 *
 * Implementation of a Physical Block-device target for Open-channel SSDs.
 *
 */

#ifndef PBLK_H_
#define PBLK_H_

#include <linux/blkdev.h>
#include <linux/blk-mq.h>
#include <linux/bio.h>
#include <linux/module.h>
#include <linux/kthread.h>
#include <linux/vmalloc.h>
#include <linux/crc32.h>
#include <linux/uuid.h>

#include <linux/lightnvm.h>

/* Run only GC if less than 1/X blocks are free */
#define GC_LIMIT_INVERSE 5
#define GC_TIME_MSECS 1000

#define PBLK_SECTOR (512)
#define PBLK_EXPOSED_PAGE_SIZE (4096)
#define PBLK_MAX_REQ_ADDRS (64)
#define PBLK_MAX_REQ_ADDRS_PW (6)

#define PBLK_META_POOL_SIZE (128)
#define PBLK_READ_REQ_POOL_SIZE (1024)

#define PBLK_NR_CLOSE_JOBS (4)

#define PBLK_CACHE_NAME_LEN (DISK_NAME_LEN + 16)

#define PBLK_COMMAND_TIMEOUT_MS 30000

/* Max 512 LUNs per device */
#define PBLK_MAX_LUNS_BITMAP (4)

#define NR_PHY_IN_LOG (PBLK_EXPOSED_PAGE_SIZE / PBLK_SECTOR)

#define pblk_for_each_lun(pblk, rlun, i) \
		for ((i) = 0, rlun = &(pblk)->luns[0]; \
			(i) < (pblk)->nr_luns; (i)++, rlun = &(pblk)->luns[(i)])

#define ERASE 2 /* READ = 0, WRITE = 1 */

#define PBLK_GEN_WS_POOL_SIZE (2)

enum {
	/* IO Types */
	PBLK_IOTYPE_USER	= 1 << 0,
	PBLK_IOTYPE_GC		= 1 << 1,

	/* Write buffer flags */
	PBLK_FLUSH_ENTRY	= 1 << 2,
	PBLK_WRITTEN_DATA	= 1 << 3,
	PBLK_SUBMITTED_ENTRY	= 1 << 4,
	PBLK_WRITABLE_ENTRY	= 1 << 5,
};

enum {
	PBLK_BLK_ST_OPEN =	0x1,
	PBLK_BLK_ST_CLOSED =	0x2,
};

struct pblk_sec_meta {
	u64 reserved;
	__le64 lba;
};

/* The number of GC lists and the rate-limiter states go together. This way the
 * rate-limiter can dictate how much GC is needed based on resource utilization.
 */
#define PBLK_GC_NR_LISTS 3

enum {
	PBLK_RL_HIGH = 1,
	PBLK_RL_MID = 2,
	PBLK_RL_LOW = 3,
};

#define pblk_dma_meta_size (sizeof(struct pblk_sec_meta) * PBLK_MAX_REQ_ADDRS)

/* write buffer completion context */
struct pblk_c_ctx {
	struct list_head list;		/* Head for out-of-order completion */

	unsigned long *lun_bitmap;	/* Luns used on current request */
	unsigned int sentry;
	unsigned int nr_valid;
	unsigned int nr_padded;
};

/* generic context */
struct pblk_g_ctx {
	void *private;
};

/* Pad context */
struct pblk_pad_rq {
	struct pblk *pblk;
	struct completion wait;
	struct kref ref;
};

/* Recovery context */
struct pblk_rec_ctx {
	struct pblk *pblk;
	struct nvm_rq *rqd;
	struct list_head failed;
	struct work_struct ws_rec;
};

/* Write context */
struct pblk_w_ctx {
	struct bio_list bios;		/* Original bios - used for completion
					 * in REQ_FUA, REQ_FLUSH case
					 */
	u64 lba;			/* Logic addr. associated with entry */
	struct ppa_addr ppa;		/* Physic addr. associated with entry */
	int flags;			/* Write context flags */
};

struct pblk_rb_entry {
	struct ppa_addr cacheline;	/* Cacheline for this entry */
	void *data;			/* Pointer to data on this entry */
	struct pblk_w_ctx w_ctx;	/* Context for this entry */
	struct list_head index;		/* List head to enable indexes */
};

#define EMPTY_ENTRY (~0U)

struct pblk_rb_pages {
	struct page *pages;
	int order;
	struct list_head list;
};

struct pblk_rb {
	struct pblk_rb_entry *entries;	/* Ring buffer entries */
	unsigned int mem;		/* Write offset - points to next
					 * writable entry in memory
					 */
	unsigned int subm;		/* Read offset - points to last entry
					 * that has been submitted to the media
					 * to be persisted
					 */
	unsigned int sync;		/* Synced - backpointer that signals
					 * the last submitted entry that has
					 * been successfully persisted to media
					 */
	unsigned int sync_point;	/* Sync point - last entry that must be
					 * flushed to the media. Used with
					 * REQ_FLUSH and REQ_FUA
					 */
	unsigned int l2p_update;	/* l2p update point - next entry for
					 * which l2p mapping will be updated to
					 * conta