lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:	Tue,  2 Feb 2016 02:30:51 +0000
From:	Bean Huo <jackyard88@...il.com>
To:	richard@....at, dedekind1@...il.com, adrian.hunter@...el.com,
	computersforpeace@...il.com, boris.brezillon@...e-electrons.com
Cc:	beanhuo@...ron.com, linux-mtd@...ts.infradead.org,
	linux-kernel@...r.kernel.org, zszubbocsev@...ron.com,
	peterpandong@...ron.com
Subject: [PATCH v2 16/17] driver:mtd:ubi:add new bakvol module in ubi layer

From: Bean Huo <beanhuo@...ron.com>

Bakvol(bakckup volume) module is for MLC NAND paired page power loss issue.
This patch file is to add new file in mtd/ubi folder.

Signed-off-by: BeanHuo <beanhuo@...ron.com>
---
 drivers/mtd/ubi/bakvol.c | 1296 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 1296 insertions(+)
 create mode 100644 drivers/mtd/ubi/bakvol.c

diff --git a/drivers/mtd/ubi/bakvol.c b/drivers/mtd/ubi/bakvol.c
new file mode 100644
index 0000000..a6b1a31
--- /dev/null
+++ b/drivers/mtd/ubi/bakvol.c
@@ -0,0 +1,1296 @@
+/* This version ported to the Linux-UBI system by Micron
+ *
+ * Based on: Micorn APPARATUSES AND METHODS FOR MEMORY MANAGEMENT
+ *
+ */
+
+/*===========================================
+  A UBI solution for MLC NAND powerloss
+
+  This driver implements backup lower page to an UBI internal bakvol volume.
+
+  The contents of this file are subject to the Mozilla Public
+  License Version 1.1 (the "License"); You may obtain a copy of
+  the License at http://www.mozilla.org/MPL/.But you can use this
+  file, no matter in compliance with the License or not.
+
+  The initial developer of the original code is Beanhuo
+  <beanhuo@...orn.com>. Portions created by Micorn.
+
+  Alternatively, the contents of this file may be used under the
+  terms of the GNU General Public License version 2 (the "GPL"), in
+  which case the provisions of the GPL are applicable instead of the
+  above.  If you wish to allow the use of your version of this file
+  only under the terms of the GPL and not to allow others to use
+  your version of this file under the MPL, indicate your decision
+  by deleting the provisions above and replace them with the notice
+  and other provisions required by the GPL.  If you do not delete
+  the provisions above, a recipient may use your version of this
+  file under either the MPL or the GPL.
+===========================================*/
+#include <asm/div64.h>
+
+#include <linux/crc32.h>
+#include <linux/err.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/mtd/mtd.h>
+
+#include "ubi.h"
+
+#ifdef CONFIG_MTD_UBI_MLC_NAND_BAKVOL
+
+/* Global point for dual plane programming operation */
+struct mtd_oob_ops *oob_ops_bak, *oob_ops_src;
+struct mtd_oob_ops *oob_ops; /* Global read/write ops point */
+
+int Recovery_done; /* 1: bakvol recovery already done. 0: need to do recovery */
+
+/* Temporary variables used during scanning PEB */
+struct ubi_ec_hdr *ech;
+struct ubi_vid_hdr *vidh;
+
+/* Which generation of Micron NAND device will be used */
+#define MICRON_NAND_VERSION_L7X_CONFIG 0
+#define MICRON_NAND_VERSION_L8X_CONFIG 1
+
+/**
+ * is_lowerpage - check page belongs to lower or uppoer page.
+ *		For the detail algorithm, please refer to Micron
+ *		product datasheet "Shared Pages".
+ * @page_num: physical page number within PEB
+ *
+ *  if paeg_num is lower page, return 1
+ *  if page_num is upper page, return 0
+ *  if page_num is SLC page, return -1
+ */
+static int is_lowerpage(int page_num)
+{
+	int ret;
+#if MICRON_NAND_VERSION_L8X_CONFIG
+	/* used for Micron L8x series parallel nand */
+	switch (page_num) {
+	case 2:
+	case 3:
+	case 248:
+	case 249:
+	case 252:
+	case 253:
+	case 254:
+	case 255:
+		ret = -1;/* This page belongs to SLC page. */
+		break;
+	default:
+		if (page_num % 4 < 2) {
+			/* If remainder is 2,3 ,this page belongs to lower page,
+			 * else if 0,1, it belongs to upper page */
+			ret = 1; /* Lower page */
+		} else {
+			ret = 0;
+		}
+		break;
+	}
+#elif MICRON_NAND_VERSION_L7X_CONFIG
+	/* Used for Micron L7X series parallel NAND */
+	switch (page_num) {
+	case 0:
+	case 1:
+		ret = 1;
+		break;
+	case 4:
+	case 5:
+	case 254:
+	case 255: /* 4,5,254, 255 is upper page */
+		ret = 0;
+		break;
+	default:
+		if (page_num % 4 > 1)
+			/* 2 ,3 is lower page, 0,1 is upper page */
+			ret = 1; /* Lower page */
+		else
+			ret = 0;
+		break;
+	}
+#endif
+	return ret;
+}
+
+/**
+ * get_next_lowerpage - return next lower page number.
+ * @page: physical page number within PEB
+ *
+ * This function returns corresponding next lower page nubmer,
+ * if exist lower page; return -1, if this page is already last
+ * page or no lower  page after this page.
+ * Note: For the detail algorithm, please refer to Micron product
+ * datasheet "Shared Pages".
+ */
+static int get_next_lowerpage(int page)
+{
+	int ret = -1;
+#if MICRON_NAND_VERSION_L8X_CONFIG
+	if (page == 254)
+		ret = -1;
+	if (page >= 255)
+		return -1;
+	/* Used for L8x  Micron series NAND */
+	if (page % 4 < 2) {
+		if (page % 2 == 0)
+			ret = page + 1;
+		else {
+			if ((page == 1) || (page == 3))
+				ret = page + 1;
+			else
+				ret = page + 3;
+		}
+
+	} else {
+		if (page % 2 == 0)
+			ret = page + 2;
+		else
+			ret = page + 1;
+	}
+
+	/* Skip SLC page */
+	switch (ret) {
+	case 2:
+	case 3:
+		ret = 4;
+		break;
+	case 248:
+	case 249:
+	case 252:
+	case 253:
+	case 254:
+	case 255:
+		ret = -1;
+		break;
+	default:
+		break;
+	}
+#elif MICRON_NAND_VERSION_L7X_CONFIG
+	/* Used for Micron L7X series NAND */
+	switch (page) {
+	case 0:
+	case 1:
+	case 2:
+		ret = page + 1;
+		break;
+	case 4:
+	case 5:
+		ret = page + 2;
+		break;
+	case 254:
+	case 255:   /* 254, 255 is upper page */
+		ret = -1;
+		break;
+	default:
+		if (page % 4 > 1) {
+			if (page % 2 == 0)
+				ret = page + 1;
+			else
+				ret = page + 3;
+		} else {
+			if (page % 2 == 0)
+				ret = page + 2;
+			else
+				ret = page + 1;
+
+		}
+		break;
+	}
+#endif
+	return ret;
+}
+
+/**
+ * get_oppo_plane_seq - return opposite plane number according to PEB number
+ *
+ *
+ */
+static int get_oppo_plane_seq(struct ubi_device *ubi, loff_t addr)
+{
+	int peb;
+
+	peb = (int)(addr >> ubi->mtd->erasesize_shift);
+
+	if (peb % 2)
+		return 0; /* plane 0 */
+	else
+		return 1; /* plane 1 */
+}
+
+/**
+ * check_original_data - check original data whether being corrupted.
+ * @ubi: UBI device description object
+ * @bak_addr: address of backup page
+ *		if source block not erased, user oob data
+ *		should equal to bak_addr.
+ * @src_addr: source page address
+ * @datbuf: point of backup data
+ */
+static int check_original_data(struct ubi_device *ubi, loff_t bak_addr,
+					loff_t src_addr, uint8_t *datbuf)
+{
+	int i;
+	loff_t read_addr;
+	struct bakvol_oob_info *oob_info;
+	void *buf = datbuf;
+	int ret = -1;
+
+	struct mtd_oob_ops *ops = kmalloc(sizeof(struct mtd_oob_ops),
+						GFP_KERNEL);
+
+	if (!ops) {
+		ubi_err(ubi, "[%d]Error,kmalloc error!\n", __LINE__);
+		goto out;
+	}
+
+	ops->datbuf = kmalloc(ubi->min_io_size, GFP_KERNEL);
+	ops->oobbuf = kmalloc(ubi->mtd->oobsize, GFP_KERNEL);
+	if (!ops->datbuf || !ops->oobbuf) {
+		ubi_err(ubi, "[%d]Error,kmalloc error!\n", __LINE__);
+		goto free;
+	}
+
+	ops->mode = MTD_OPS_AUTO_OOB;
+	ops->ooblen = UBI_BAKVOL_OOB_SIZE;
+	ops->len = ubi->min_io_size;
+	ops->ooboffs = 0;
+	ops->retlen = 0;
+	ops->oobretlen = 0;
+
+	dbg_gen("%s:Source page addr = 0x%llx\n", __func__, src_addr);
+
+	ret = ubi->mtd->_read_oob(ubi->mtd, src_addr, ops);
+
+	if (ret < 0) {
+		/*
+		 * Read error, means this page being corrupted,
+		 * need to recover.
+		 */
+		ret = 1;
+		goto free;
+	}
+
+	oob_info = (struct bakvol_oob_info *)ops->oobbuf;
+	read_addr = be64_to_cpu(oob_info->addr);
+	if (read_addr != bak_addr) {
+		/*
+		 * source page oobbuf != bak_addr, means this PEB already
+		 * being erased or re-programmed. what's more, exsit bitflips
+		 * in user oob area, anyway, current backup page data
+		 * already is not corresponding source page data.
+		*/
+		dbg_gen("[%d] backup page address not match.\n", __LINE__);
+		ret = 0;
+		goto free;
+	}
+
+	for (i = 0; i < ubi->min_io_size; i += 4) {
+		if (*(u32 *)ops->datbuf != *(u32 *)buf) {
+			ret = 1;
+			goto free;
+		}
+	}
+	ret = 0;
+	dbg_gen("Original data not being corrupted.\n");
+
+free:
+	kfree(ops->oobbuf);
+	kfree(ops->datbuf);
+	kfree(ops);
+out:
+	return ret;
+
+}
+/**
+ * find_last_programmed_page - find last page being programmed.
+ * @ubi: UBI device description object
+ * @pnum: PEB number
+ * @ret_page: return last programmed page number
+ */
+static int find_last_programmed_page(struct ubi_device *ubi,
+					int pnum, int *ret_page)
+{
+	struct mtd_info *mtd = ubi->mtd;
+	int page, start_page, err;
+
+	oob_ops->mode = MTD_OPS_AUTO_OOB;
+	oob_ops->ooblen = UBI_BAKVOL_OOB_SIZE;
+	oob_ops->len = ubi->min_io_size;
+	oob_ops->ooboffs = 0;
+	oob_ops->retlen = 0;
+	oob_ops->oobretlen = 0;
+
+	page = (mtd->erasesize - 1) >> mtd->writesize_shift;
+	start_page = ubi->leb_start >> mtd->writesize_shift;
+	for (; page >= start_page; page--) {
+		err = ubi->mtd->_read_oob(ubi->mtd,
+				(pnum << mtd->erasesize_shift) |
+				(page << mtd->writesize_shift), oob_ops);
+
+		if (err < 0 && err != -EUCLEAN) {
+			*ret_page = page;
+			return -1;
+		}
+
+		if (!ubi_check_pattern(oob_ops->oobbuf, 0xFF,
+						oob_ops->ooblen)) {
+			*ret_page = page;
+			return 1;
+		}
+	}
+	return 0;
+}
+
+/**
+ * get_peb_from_bakvol - get a PEB that already being opened by bakvol.
+ * @ubi: UBI device description object
+ * @plane: indicates this PEB should belong to which plane
+ * @page_num: indicates which page should not be programmed
+ */
+static struct ubi_bkblk_info *get_peb_from_bakvol(struct ubi_device *ubi,
+						char plane, int page_num)
+{
+	struct ubi_bkblk_tbl *backup_info = ubi->bkblk_tbl;
+	struct list_head *head = &backup_info->head;
+	struct ubi_bkblk_info *bbi, *tempbbi = NULL;
+	int page;
+
+	if (list_empty(head)) {
+		dbg_gen("%s:bakvol no already opened PEB.\n", __func__);
+		return NULL;
+	}
+
+	list_for_each_entry(bbi, head, node) {
+		if (bbi->plane == plane) {
+			page = get_next_lowerpage(bbi->pgnum);
+			if (page == -1)
+				continue;
+			if (page == page_num) {
+				/*
+				 * Perfecct match
+				 */
+				tempbbi = bbi;
+				break;
+			} else if (page < page_num) {
+				/*
+				 * Make sure this page not be programmed.
+				 */
+				if (tempbbi == NULL)
+					tempbbi = bbi;
+				else {
+					if (tempbbi->pgnum > bbi->pgnum)
+						/*
+						 * As far as possible choose
+						 * the PEB with smallest page
+						 * number.
+						 */
+						tempbbi = bbi;
+				}
+			}
+		}
+	}
+
+	if (tempbbi)
+		dbg_gen("Get bakvol PEB %d:%d for original lower page %d.\n",
+				tempbbi->peb, tempbbi->pgnum, page_num);
+	else
+		dbg_gen("Cannot get free bakvol PEB for plane %d:page %d.\n",
+					plane, page_num);
+
+	return tempbbi;
+}
+
+static void prepare_bakvol_oob_info(loff_t *addr,
+			struct bakvol_oob_info *bakvol_oob)
+{
+	uint32_t crc;
+
+	ubi_assert(addr != NULL &&  bakvol_oob != NULL);
+
+	bakvol_oob->addr = cpu_to_be64(*(loff_t *)addr);
+	crc = crc32(UBI_CRC32_INIT, bakvol_oob, UBI_BAKVOL_OOB_SIZE_CRC);
+	bakvol_oob->crc = cpu_to_be32(crc);
+}
+
+/**
+ * validate_bakvol_oob_info- validate bakvol oob user info.
+ * @ubi: UBI device description object
+ * @oob_info: bakvol oob info
+ *
+ * This function returns zero if bakvol oob user info is valid,
+ * and return 1 if not.
+ */
+static int validate_bakvol_oob_info(struct ubi_device *ubi,
+			const struct bakvol_oob_info *oob_info)
+{
+	uint32_t crc, read_crc;
+	int ret = 0;
+
+	crc = crc32(UBI_CRC32_INIT, oob_info, UBI_BAKVOL_OOB_SIZE_CRC);
+	read_crc = be32_to_cpu(oob_info->crc);
+	if (read_crc != crc) {
+		ubi_err(ubi, "OOB info CRC error,calculate 0x%x, read 0x%0x",
+							crc, read_crc);
+		ret = 1;
+	}
+
+	return ret;
+}
+
+static struct
+ubi_bkblk_info *find_min_free_space_peb(struct ubi_bkblk_tbl *backup_info)
+{
+	struct list_head *head = &backup_info->head;
+	struct ubi_bkblk_info *bbi, *retbbi;
+
+	retbbi = list_first_entry(head, struct ubi_bkblk_info, node);
+
+	list_for_each_entry(bbi, head, node) {
+		if (bbi->pgnum > retbbi->pgnum)
+			retbbi = bbi;
+	}
+
+	return retbbi;
+}
+
+static struct
+ubi_bkblk_info *allo_new_block_for_bakvol(struct ubi_device *ubi, u8 plane)
+{
+
+	struct ubi_vid_hdr *vid_hdr;
+	struct ubi_volume *vol;
+	struct ubi_bkblk_tbl *backup_info = ubi->bkblk_tbl;
+	struct ubi_bkblk_info *bbin, *newbbin;
+	int bc_plane0, bc_plane1;
+	int pnum;
+	int err;
+	int tries = 1;
+
+	vol = ubi->volumes[vol_id2idx(ubi, UBI_BACKUP_VOLUME_ID)];
+
+	vid_hdr = ubi_zalloc_vid_hdr(ubi, GFP_NOFS);
+	if (!vid_hdr)
+		return NULL;
+
+	newbbin = kmalloc(sizeof(*newbbin), GFP_ATOMIC);
+	if (!newbbin) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	vid_hdr->vol_type = UBI_VID_DYNAMIC;
+	vid_hdr->sqnum = cpu_to_be64(ubi_next_sqnum(ubi));
+	vid_hdr->vol_id = cpu_to_be32(UBI_BACKUP_VOLUME_ID);
+	vid_hdr->compat = UBI_BACKUP_VOLUME_COMPAT;
+	vid_hdr->data_pad = cpu_to_be32(0);
+	vid_hdr->data_size = vid_hdr->used_ebs = vid_hdr->data_pad;
+
+retry:
+	pnum = ubi_wl_get_plane_peb(ubi, plane);
+
+	if (pnum < 0) {
+		err = -ENOMEM;
+		goto free_mem;
+	}
+
+	if (pnum%2 != plane) {
+		err = pnum;
+		goto free_mem;
+	}
+
+	bc_plane0 = backup_info->bcount_of_plane[0];
+	bc_plane1 = backup_info->bcount_of_plane[1];
+
+	if (bc_plane0 + bc_plane1 >= vol->reserved_pebs) {
+		/**
+		 * we make sure bakvol opened PEBs cannot overflow
+		 * bakvol reserved PEB.
+		 */
+		dbg_gen("bakvol opened PEB over reserved PEB.\n");
+		dbg_gen("Will put PEB with minimum free space.\n");
+		bbin = find_min_free_space_peb(backup_info);
+
+		if (vol->eba_tbl[bbin->leb] >= 0) {
+			dbg_gen("bakvol put PEB %d.\n", bbin->peb);
+			err = ubi_wl_put_peb(ubi, vol->vol_id, bbin->leb,
+							bbin->peb, 0);
+			if (err)
+				goto free_mem;
+		}
+
+		vol->eba_tbl[bbin->leb] = UBI_LEB_UNMAPPED;
+		backup_info->bcount_of_plane[bbin->plane]--;
+		newbbin->leb = bbin->leb;
+
+		list_del(&bbin->node);
+
+	} else {
+		int i = 0;
+
+		for (i = 0; i < vol->reserved_pebs; i++)
+			if (vol->eba_tbl[i] == UBI_LEB_UNMAPPED) {
+				newbbin->leb = i;
+				break;
+			}
+	}
+
+	vid_hdr->lnum = cpu_to_be32(newbbin->leb);
+	if ((newbbin->leb > vol->reserved_pebs) || (newbbin->leb < 0)) {
+		ubi_err(ubi, "BUG: logic block number [%d] error.\n", newbbin->leb);
+		panic("BUG!");
+	}
+
+	err = ubi_io_write_vid_hdr(ubi, pnum, vid_hdr);
+	if (err) {
+		ubi_err(ubi, "Failed to write VID header to PEB %d.\n", pnum);
+		goto write_error;
+	}
+
+	vol->eba_tbl[newbbin->leb] = pnum;
+
+	newbbin->peb = pnum;
+	newbbin->plane = plane;
+	newbbin->pgnum = get_next_lowerpage(1); /* Skip page0 and page1 */
+
+	/* Update bakvol variable */
+	backup_info->bcount_of_plane[plane]++;
+	list_add(&newbbin->node, &backup_info->head);
+
+out:
+	ubi_free_vid_hdr(ubi, vid_hdr);
+	return newbbin;
+
+write_error:
+	/*
+	 * Bad luck? This physical eraseblock is bad too?  Let's try to
+	 * get another one.
+	 */
+	ubi_err(ubi, "Failed to write to PEB %d", pnum);
+	ubi_wl_put_peb(ubi, vol->vol_id, newbbin->leb, newbbin->peb, 1);
+	if (++tries > UBI_IO_RETRIES)
+		goto free_mem;
+
+	ubi_err(ubi, "Try again");
+	goto retry;
+
+free_mem:
+	ubi_free_vid_hdr(ubi, vid_hdr);
+	kfree(newbbin);
+	ubi_err(ubi, "%s:%d %d times failed to alloc a new block!",
+			__func__, __LINE__, tries);
+	return NULL;
+}
+
+/**
+ * is_backup_need -check if this page date need to be backup.
+ * @ubi: UBI device description object
+ * @addr: page address
+ *
+ * return 0, this page data shouldn't  backup
+ * return 1, this page data should  backup
+ */
+int is_backup_need(struct ubi_device *ubi, loff_t addr)
+{
+	int page_num, ret;
+
+	page_num = (addr & ubi->mtd->erasesize_mask) >>
+						ubi->mtd->writesize_shift;
+	if (page_num == 0 || page_num == 1)
+		/* Currently,we don't backup EC head and VID head */
+		return 0;
+
+	ret = is_lowerpage(page_num);
+	if (ret != 1) {
+		dbg_gen("Page %d is %s page.\n", page_num,
+				(ret == -1 ? "SLC" : "upper"));
+		ret = 0;
+	} else
+		dbg_gen("Page %d is lower page.\n", page_num);
+
+	return ret;
+}
+
+/**
+ * ubi_check_bakvol_module - check if bakvol has been buildup
+ * @ubi: UBI device description object
+ *
+ * returns 1 in case of  backup volume have been built,
+ * 0 in case of don't find
+ */
+int ubi_check_bakvol_module(struct ubi_device *ubi)
+{
+	return (ubi->bkblk_tbl->bakvol_flag & UBI_BAKVOL_ENABLE);
+}
+
+/**
+ * ubi_duplicate_data_to_bakvol - dual plane program data into bakvol area
+ *			and main area respectively.
+ * @ubi: UBI device description object
+ * @addr: address will be programmed
+ * @len: date length
+ * @retlen: already programmed data length
+ * @buf: data buffer, points to data that should be programmed
+ */
+int ubi_duplicate_data_to_bakvol(struct ubi_device *ubi, loff_t addr,
+				size_t len, size_t *retlen, const void *buf)
+{
+	struct ubi_bkblk_tbl *backup_info = ubi->bkblk_tbl;
+	struct mtd_info *mtd = ubi->mtd;
+	int err = 0, nobak = 0;
+	int pnum;
+	u8 oppe_plane;
+	loff_t lpage_addr; /* Lower page byte address */
+	struct ubi_bkblk_info *pbk;
+	int page_num;
+	struct bakvol_oob_info *oob_bak = NULL, *oob_src = NULL;
+
+	if (len > ubi->min_io_size) {
+		ubi_err(ubi, "%d: Write data len overflow [%d]\n",
+							__LINE__, len);
+		return -EROFS;
+	}
+
+	if (!buf) {
+		ubi_err(ubi, "%d: Write buf is NULL!\n", __LINE__);
+		return -EROFS;
+
+	}
+	oppe_plane = get_oppo_plane_seq(ubi, addr);
+
+	page_num = (int)(addr & ubi->mtd->erasesize_mask) >>
+						ubi->mtd->writesize_shift;
+	pnum = (int)(addr >> ubi->mtd->erasesize_shift);
+
+	oob_bak = kzalloc(UBI_BAKVOL_OOB_SIZE, GFP_KERNEL);
+	if (!oob_bak) {
+		err = -ENOMEM;
+		ubi_err(ubi, "%s:%d: kzalloc error.\n",
+						__func__, __LINE__);
+		goto free;
+	}
+
+	oob_src = kzalloc(UBI_BAKVOL_OOB_SIZE, GFP_KERNEL);
+	if (!oob_src) {
+		err = -ENOMEM;
+		ubi_err(ubi, "%s:%d: kzalloc error.\n",
+						__func__, __LINE__);
+		goto free;
+	}
+
+	if (backup_info->bcount_of_plane[oppe_plane] == 0)
+		pbk = NULL;
+	else
+		pbk = get_peb_from_bakvol(ubi, oppe_plane, page_num);
+
+	if (pbk == NULL) {
+		dbg_gen("Allocate new PEB for Bakvol.\n");
+		pbk = allo_new_block_for_bakvol(ubi, oppe_plane);
+		if (!pbk) {
+			ubi_err(ubi, "Allocate new PEB failed.\n");
+			nobak = 1;
+			goto Only_source;
+		}
+	}
+
+	/*
+	 * Here original page address stores in user oob area of bakvol page,
+	 * and corresponding its backup page address stores in user oob area of
+	 * main data area page.
+	*/
+	lpage_addr = (pbk->peb << mtd->erasesize_shift) |
+				(page_num << mtd->writesize_shift);
+
+	/*
+	 * organize bakvol page ops
+	*/
+	oob_ops_bak->datbuf = (uint8_t *)buf;
+	oob_ops_bak->mode = MTD_OPS_AUTO_OOB;
+	oob_ops_bak->ooblen = UBI_BAKVOL_OOB_SIZE;
+	prepare_bakvol_oob_info(&addr, oob_bak);
+	oob_ops_bak->oobbuf = (uint8_t *)oob_bak; /* Original page addr */
+	oob_ops_bak->len = len;
+	oob_ops_bak->ooboffs = 0;
+	oob_ops_bak->retlen = 0;
+	oob_ops_bak->oobretlen = 0;
+
+Only_source:
+	/*
+	 * Organize main data area page ops
+	*/
+	oob_ops_src->datbuf = (uint8_t *)buf;
+	oob_ops_src->mode = MTD_OPS_AUTO_OOB;
+	oob_ops_src->ooblen = UBI_BAKVOL_OOB_SIZE;
+	prepare_bakvol_oob_info(&lpage_addr, oob_src);
+	oob_ops_src->oobbuf = (uint8_t *)oob_src; /* backup page addr */
+	oob_ops_src->len = len;
+	oob_ops_src->ooboffs = 0;
+	oob_ops_src->retlen = 0;
+	oob_ops_src->oobretlen = 0;
+
+	if (!nobak)
+		err = mtd_write_dual_plane_oob(ubi->mtd, lpage_addr,
+						oob_ops_bak, addr, oob_ops_src);
+
+	if ((err == -EOPNOTSUPP) || (nobak == 1)) {
+		/*
+		 * If dual plane program function not enable, or get
+		 * bakvol peb failed, we only program original data to
+		 * main data area.
+		*/
+		ubi_err(ubi, "Only program source data.\n");
+		err = mtd_write_oob(ubi->mtd, addr, oob_ops_src);
+		goto out;
+	}
+
+	pbk->pgnum = page_num; /* updata bakvol PEB programmed page info */
+
+out:
+	if (err)
+		*retlen = 0;
+	else
+		*retlen = len;
+free:
+	kfree(oob_bak);
+	kfree(oob_src);
+	return err;
+}
+
+
+int ubi_bakvol_module_init(struct ubi_device *ubi)
+{
+	int step;
+
+	if ((ubi->mtd->type != MTD_MLCNANDFLASH) ||
+		(ubi->mtd->oobavail < UBI_BAKVOL_OOB_SIZE)) {
+		ubi_err(ubi, "NAND cann't meet Bakvol module requirement.\n");
+		step = 0;
+		goto out_init;
+	}
+
+	ech = kzalloc(ubi->ec_hdr_alsize, GFP_KERNEL);
+	if (!ech) {
+		step = 1;
+		goto out_init;
+	}
+
+	vidh = ubi_zalloc_vid_hdr(ubi, GFP_KERNEL);
+	if (!vidh) {
+		step = 2;
+		goto out_ech;
+	}
+
+	ubi->bkblk_tbl = kzalloc(sizeof(struct ubi_bkblk_tbl), GFP_KERNEL);
+	if (!ubi->bkblk_tbl) {
+		step = 3;
+		goto out_vidh;
+	}
+	ubi->bkblk_tbl->bcount_of_plane[0] = 0;
+	ubi->bkblk_tbl->bcount_of_plane[1] = 0;
+	INIT_LIST_HEAD(&ubi->bkblk_tbl->head);
+
+	oob_ops_bak = kzalloc(sizeof(struct mtd_oob_ops), GFP_KERNEL);
+	if (!oob_ops_bak) {
+		step = 4;
+		goto out_backup_info;
+	}
+
+	oob_ops_bak->oobbuf = kzalloc(ubi->mtd->oobsize, GFP_KERNEL);
+	if (!oob_ops_bak->oobbuf) {
+		step = 5;
+		goto out_back_oob_ops;
+	}
+
+	oob_ops_src = kzalloc(sizeof(struct mtd_oob_ops), GFP_KERNEL);
+	if (!oob_ops_src) {
+		step = 6;
+		goto out_obs_src;
+	}
+
+	oob_ops_src->oobbuf = kzalloc(ubi->mtd->oobsize, GFP_KERNEL);
+	if (!oob_ops_src->oobbuf) {
+		step = 7;
+		goto out_oob_ops;
+	}
+
+	oob_ops = kmalloc(sizeof(struct mtd_oob_ops), GFP_KERNEL);
+	if (!oob_ops) {
+		step = 8;
+		goto out0;
+	}
+
+	oob_ops->datbuf = kmalloc(ubi->min_io_size, GFP_KERNEL);
+	if (!oob_ops->datbuf) {
+		step = 9;
+		goto out1;
+	}
+
+	oob_ops->oobbuf = kmalloc(ubi->mtd->oobsize, GFP_KERNEL);
+	if (!oob_ops->oobbuf) {
+		step = 10;
+		goto out2;
+	}
+
+	ubi_free_vid_hdr(ubi, vidh);
+	kfree(ech);
+
+	ubi->bkblk_tbl->bakvol_flag = UBI_BAKVOL_INIT_START;
+	return 0;
+
+out2:
+	kfree(oob_ops->datbuf);
+out1:
+	kfree(oob_ops);
+out0:
+	kfree(oob_ops_src->oobbuf);
+out_oob_ops:
+	kfree(oob_ops_src);
+out_obs_src:
+	kfree(oob_ops_bak->oobbuf);
+out_back_oob_ops:
+	kfree(oob_ops_bak);
+out_backup_info:
+	kfree(ubi->bkblk_tbl);
+out_vidh:
+	ubi_free_vid_hdr(ubi, vidh);
+out_ech:
+	kfree(ech);
+out_init:
+	ubi_err(ubi, "bakvol module init error,init step [%d].\n", step);
+	ubi->bkblk_tbl->bakvol_flag = UBI_BAKVOL_REJECT;
+	return -1;
+}
+
+/**
+ * ubi_bakvol_peb_scan - check VID header,see if this PEB
+ *				belongs to bakvol
+ *
+ * This function returns 1 if this PEB not belongs to bakvol
+ * return -1 in case of allocate resource error
+ * return 0  if scan complete, and it belongs to bakvol
+ */
+int ubi_bakvol_peb_scan(struct ubi_device *ubi, struct ubi_vid_hdr *vidh,
+								int pnum)
+{
+	int page;
+	int vol_id;
+	struct ubi_bkblk_info *bbi;
+	int lnum;
+	int ret;
+
+	page = 0;
+
+	vol_id = be32_to_cpu(vidh->vol_id);
+	if (vol_id == UBI_BACKUP_VOLUME_ID) {
+		/* Found bakvol PEB */
+		lnum = be32_to_cpu(vidh->lnum);
+		dbg_gen("%s:Found backup PEB,pnum is [%d],lnum is[%d].\n",
+							__func__, pnum, lnum);
+
+		/* Unsupported internal volume */
+		if (vidh->compat != UBI_COMPAT_REJECT) {
+			ubi_err(ubi, "Backup volume compat != UBI_COMPAT_REJECT\n");
+			return -1;
+		}
+	} else {
+		return 1;
+	}
+
+	ret = find_last_programmed_page(ubi, pnum, &page);
+	dbg_gen("bakvol PEB %d last programmed page [%d]\n", pnum, page);
+
+	if (ret == 1) {
+		bbi = kmalloc(sizeof(*bbi), GFP_ATOMIC);
+		if (!bbi)
+			goto out;
+
+		bbi->peb = pnum;
+		bbi->leb = lnum;
+		bbi->pgnum = page;
+		bbi->plane = pnum % 2;
+		bbi->plane ? (ubi->bkblk_tbl->bcount_of_plane[1]++) :
+					(ubi->bkblk_tbl->bcount_of_plane[0]++);
+		list_add(&bbi->node, &ubi->bkblk_tbl->head);
+	} else if (ret == 0) {
+		/* This backup PEB has not been programmed. */
+		bbi = kmalloc(sizeof(*bbi), GFP_ATOMIC);
+		if (!bbi)
+			goto out;
+		bbi->peb = pnum;
+		bbi->leb = lnum;
+		bbi->pgnum = get_next_lowerpage(1);
+		bbi->plane = pnum % 2;
+		bbi->plane ? (ubi->bkblk_tbl->bcount_of_plane[1]++) :
+					(ubi->bkblk_tbl->bcount_of_plane[0]++);
+		list_add(&bbi->node, &ubi->bkblk_tbl->head);
+	} else {
+		/* Backup block is corrupted,maybe power cut happened during
+		 * programming lower page. so this backup block with corrupted
+		 * page can be removed from bakvol volume.
+		 */
+		dbg_gen("PEB %d will be removed from backup volume later.\n",
+									pnum);
+
+		bbi = kmalloc(sizeof(*bbi), GFP_ATOMIC);
+		if (!bbi)
+			goto out;
+		bbi->peb = pnum;
+		bbi->leb = lnum;
+		bbi->pgnum = page;
+		bbi->plane = pnum % 2;
+		bbi->plane ? (ubi->bkblk_tbl->bcount_of_plane[1]++) :
+				(ubi->bkblk_tbl->bcount_of_plane[0]++);
+		list_add(&bbi->node, &ubi->bkblk_tbl->head);
+	}
+
+	return 0;
+out:
+	return -1;
+
+}
+
+/**
+ * ubi_bakvol_module_init_tail - init backup volume.
+ * @ubi: UBI device description object
+ * @si: scanning information
+ *
+ * This function init the backup volume
+ * Returns zero in case of success and a negative
+ * error code in case of failure
+ */
+int ubi_bakvol_module_init_tail(struct ubi_device *ubi,
+					struct ubi_attach_info *si)
+{
+	int reserved_pebs = 0;
+	struct ubi_volume *vol;
+	struct ubi_bkblk_info *bbi;
+
+	if (ubi->bkblk_tbl->bakvol_flag & UBI_BAKVOL_REJECT)
+		return 0;
+
+	if (!(ubi->bkblk_tbl->bakvol_flag & UBI_BAKVOL_INIT_DONE)) {
+
+		/* And add the layout volume */
+		vol = kzalloc(sizeof(struct ubi_volume), GFP_KERNEL);
+		if (!vol)
+			return -ENOMEM;
+
+		vol->reserved_pebs = UBI_BACKUP_VOLUME_EBS;
+		vol->alignment = 1;
+		vol->vol_type = UBI_DYNAMIC_VOLUME;
+		vol->name_len = sizeof(UBI_BACKUP_VOLUME_NAME) - 1;
+		vol->data_pad = 0;
+		memcpy(vol->name, UBI_BACKUP_VOLUME_NAME, vol->name_len + 1);
+
+		vol->usable_leb_size = ubi->leb_size;
+		vol->used_ebs = vol->reserved_pebs;
+		vol->last_eb_bytes = vol->reserved_pebs;
+		vol->used_bytes =
+		    (long long)vol->used_ebs * (ubi->leb_size - vol->data_pad);
+		vol->vol_id = UBI_BACKUP_VOLUME_ID;
+		vol->ref_count = UBI_BACKUP_VOLUME_EBS;
+		ubi->volumes[vol_id2idx(ubi, vol->vol_id)] = vol;
+		reserved_pebs += vol->reserved_pebs;
+		ubi->vol_count += 1;
+		vol->ubi = ubi;
+		if (reserved_pebs > ubi->avail_pebs) {
+			ubi_err(ubi, "No enough PEBs, required %d, available %d",
+				reserved_pebs, ubi->avail_pebs);
+			return -1;
+		}
+		ubi->rsvd_pebs += reserved_pebs;
+		ubi->avail_pebs -= reserved_pebs;
+
+		ubi->bkblk_tbl->bakvol_flag = UBI_BAKVOL_INIT_DONE;
+
+		ubi_msg(ubi, "bakvol module opened PEB list:");
+		list_for_each_entry(bbi, &(ubi->bkblk_tbl->head), node) {
+			ubi_msg(ubi, "peb %d,pgnum %d,plane %d,;leb %d.",
+				bbi->peb, bbi->pgnum, bbi->plane, bbi->leb);
+		}
+	}
+	return 0;
+}
+
+int ubi_corrupted_data_recovery(struct ubi_volume_desc *desc)
+{
+	u32 base_offset, buff_offset, pnum, offset, page;
+	u32 correct, uncorrect_l, uncorrect_h;
+	int ret, err;
+	struct ubi_device *ubi = desc->vol->ubi;
+	struct mtd_info *mtd = ubi->mtd;
+	struct ubi_bkblk_info *bbi;
+	unsigned char *buf = ubi->peb_buf;
+	struct ubi_volume_desc volume_desc;
+	int i, move;
+	loff_t bakpage_addr;
+	struct bakvol_oob_info *oob_info;
+	off_t src_addr;
+
+	move = 0;
+	correct = 0;
+	uncorrect_h = 0;
+	uncorrect_l = 0;
+
+	if ((ubi->bkblk_tbl->bakvol_flag & UBI_BAKVOL_RECOVERY) ||
+		!(ubi->bkblk_tbl->bakvol_flag & UBI_BAKVOL_INIT_DONE))
+		return 0;
+
+	vidh = ubi_zalloc_vid_hdr(ubi, GFP_KERNEL);
+	if (!vidh)
+		return -1;
+
+	oob_ops->mode = MTD_OPS_AUTO_OOB;
+	oob_ops->ooblen = UBI_BAKVOL_OOB_SIZE;
+	oob_ops->len = ubi->min_io_size;
+
+	list_for_each_entry(bbi, &(ubi->bkblk_tbl->head), node) {
+		dbg_gen("Processing bakvol PEB %d,pgnum %d,plane %d.\n",
+			bbi->peb, bbi->pgnum, bbi->plane);
+		i = get_next_lowerpage(1);
+
+		for ( ; (i <= bbi->pgnum) && (i != -1);
+				i = get_next_lowerpage(i)) {
+			/*
+			 * Read backup data from bakvol PEB with OOB.
+			*/
+			oob_ops->ooboffs = 0;
+			oob_ops->retlen = 0;
+			oob_ops->oobretlen = 0;
+
+			err = ubi->mtd->_read_oob(ubi->mtd,
+					(bbi->peb << mtd->erasesize_shift) |
+					(i << mtd->writesize_shift), oob_ops);
+
+			if (err < 0 && err != -EUCLEAN) {
+				dbg_gen("Read bakvol PEB %d:%d error %d.\n",
+						bbi->peb, i, err);
+				move = 1;
+				continue;
+			}
+
+			if (ubi_check_pattern(oob_ops->oobbuf, 0xFF,
+						oob_ops->ooblen)) {
+				dbg_gen("Bakvol PEB %d ever skip lower page %d.\n",
+						bbi->peb, i);
+				continue;
+			}
+
+			oob_info = (struct bakvol_oob_info *)oob_ops->oobbuf;
+
+			if (validate_bakvol_oob_info(ubi, oob_info)) {
+				dbg_gen("Bakvol PEB %d page %d user oob area exist bitflips.\n",
+						bbi->peb, i);
+				continue;
+			}
+
+		src_addr = be64_to_cpu(oob_info->addr);
+
+		/*
+		 * Calculate backup page address, used for check if source block
+		 * already being erased or re-mapped.
+		*/
+		bakpage_addr = (bbi->peb << mtd->erasesize_shift) |
+					(i << mtd->writesize_shift);
+
+		ret = check_original_data(ubi, bakpage_addr,
+						src_addr, oob_ops->datbuf);
+		  if (ret == 1) {
+			/* The original data isn't correct anymore,
+			 * already being damaged, and oobbuf store original page
+			 * address.
+			 */
+			pnum = src_addr >> mtd->erasesize_shift;
+			offset = src_addr & mtd->erasesize_mask;
+			buff_offset = 0;
+
+			/**
+			 * Recover operation
+			 */
+
+			for (base_offset = ubi->leb_start;
+				base_offset < ubi->peb_size;
+				base_offset += mtd->writesize) {
+
+				ret = ubi_io_read(ubi, buf + buff_offset, pnum,
+						base_offset, mtd->writesize);
+
+			if (ret == -EBADMSG || ret == -EIO) {
+				if (base_offset == offset) {
+				    dbg_gen("Bakvol recover offset 0x%x.....\n",
+						base_offset);
+				    memcpy(buf + buff_offset, oob_ops->datbuf,
+							mtd->writesize);
+				    correct++;
+				} else {
+				    page = base_offset >> mtd->writesize_shift;
+					if (is_lowerpage(page)) {
+						/* More low-page also
+						 * be corrupted */
+						uncorrect_l++;
+					ubi_err(ubi, "PEB %d unrecovery lower page %d\n",
+								pnum, page);
+					uncorrect_l++;
+					} else {
+					/**
+					 * Unbackup upper page also
+					 * be corruptted */
+					ubi_err(ubi, "PEB %d unrecovery upper page %d.\n",
+								pnum, page);
+					uncorrect_h++;
+					}
+				}
+			}
+
+			ret = ubi_check_pattern(buf + buff_offset, 0xFF,
+							mtd->writesize);
+			if (ret)
+				/* This page is empty, so the last of pages
+				 * also are empty, no need to read.*/
+				break;
+
+			buff_offset += mtd->writesize;
+			}
+
+		if ((correct == 0) || (uncorrect_h > 1) || (uncorrect_l > 0)) {
+			/* Could not recover. We only accept two corrupted
+			 * pages, they are paired-page. If there are two
+			 * corrupted lower pages within one PEB, also give
+			 * up recovery.
+			 */
+			dbg_gen("Can not be recoverred.\n");
+			uncorrect_l = 0;
+			uncorrect_h = 0;
+			continue;
+			}
+
+		ret = ubi_io_read_vid_hdr(ubi, pnum, vidh, 0);
+
+		volume_desc.vol = ubi->volumes[be32_to_cpu(vidh->vol_id)];
+		volume_desc.mode = UBI_READWRITE;
+		ret = ubi_leb_change(&volume_desc, be32_to_cpu(vidh->lnum),
+							buf, buff_offset);
+		correct = 0;
+		uncorrect_h = 0;
+
+		if (ret) {
+			ubi_err(ubi, "Changing %d bytes in LEB %d failed",
+						buff_offset, vidh->lnum);
+			ubi_err(ubi, "Ret error:%d.", ret);
+			dump_stack();
+		} else
+			dbg_gen(".....Done\n");
+
+	  }
+
+	}
+
+		if (move == 1) {
+		    bbi->pgnum = (ubi->mtd->erasesize - 1) >>
+					ubi->mtd->writesize_shift;
+		    move = 0;
+		}
+	}
+	ubi->bkblk_tbl->bakvol_flag |= UBI_BAKVOL_RECOVERY;
+	ubi_free_vid_hdr(ubi, vidh);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(ubi_corrupted_data_recovery);
+
+void clear_bakvol(struct ubi_device *ubi)
+{
+	kfree(oob_ops->datbuf);
+	kfree(oob_ops);
+	kfree(oob_ops_src->oobbuf);
+	kfree(oob_ops_src);
+	kfree(oob_ops_bak->oobbuf);
+	kfree(oob_ops_bak);
+	kfree(ubi->bkblk_tbl);
+	ubi->bkblk_tbl->bakvol_flag = UBI_BAKVOL_INIT_START;
+}
+
+void init_bakvol(struct ubi_volume_desc *desc, uint8_t choice)
+{
+	struct ubi_volume *vol = desc->vol;
+	struct ubi_device *ubi = vol->ubi;
+
+	if (choice) {
+		if (ubi->bkblk_tbl->bakvol_flag & UBI_BAKVOL_INIT_DONE ||
+			ubi->bkblk_tbl->bakvol_flag & UBI_BAKVOL_DISABLE) {
+			dbg_gen("[%s][%d] Enable bakvol module successfully!\n",
+						__func__, __LINE__);
+			ubi->bkblk_tbl->bakvol_flag &= ~UBI_BAKVOL_DISABLE;
+			ubi->bkblk_tbl->bakvol_flag |= UBI_BAKVOL_ENABLE;
+
+		} else
+			dbg_gen("[%s][%d] Enable bakvol module failed!\n",
+					__func__, __LINE__);
+
+	} else {
+		ubi->bkblk_tbl->bakvol_flag &= ~UBI_BAKVOL_ENABLE;
+		ubi->bkblk_tbl->bakvol_flag |= UBI_BAKVOL_DISABLE;
+		dbg_gen("[%s][%d] Disable bakvol module!\n",
+						__func__, __LINE__);
+	}
+}
+EXPORT_SYMBOL_GPL(init_bakvol);
+#else
+int is_backup_need(struct ubi_device *ubi, loff_t addr)
+{
+	return 0;
+}
+
+int ubi_check_bakvol_module(struct ubi_device *ubi)
+{
+
+	return 0;
+}
+
+int ubi_duplicate_data_to_bakvol(struct ubi_device *ubi, loff_t addr,
+				size_t len, size_t *retlen, const void *buf)
+{
+	return 0;
+}
+int ubi_bakvol_module_init(struct ubi_device *ubi)
+{
+	return 0;
+
+}
+
+int ubi_bakvol_peb_scan(struct ubi_device *ubi, struct ubi_vid_hdr *vidh,
+								int pnum)
+{
+	return 0;
+}
+
+int ubi_bakvol_module_init_tail(struct ubi_device *ubi,
+					struct ubi_attach_info *si)
+{
+	return 0;
+}
+
+int ubi_corrupted_data_recovery(struct ubi_volume_desc *desc)
+{
+	return 0;
+
+}
+EXPORT_SYMBOL_GPL(ubi_corrupted_data_recovery);
+
+
+void clear_bakvol(struct ubi_device *ubi)
+{
+
+}
+
+void init_bakvol(struct ubi_volume_desc *desc, uint8_t choice)
+{
+
+}
+EXPORT_SYMBOL_GPL(init_bakvol);
+#endif
+
+MODULE_LICENSE("Dual MPL/GPL");
+MODULE_AUTHOR("Bean Huo <beanhuo@...ron.com>");
+MODULE_DESCRIPTION("Support code for MLC NAND pair page powerloss protection");
-- 
1.9.1

Powered by blists - more mailing lists