Skip to contentSkip to conversation

JFFS2: use MTD_OOB_AUTO

RFC PATCH In review Artem Bityutskiy 2 participants 3 messages Feb 1, 2007 – Feb 2, 2007

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 JFFS2: use MTD_OOB_AUTO

From: Artem Bityutskiy <[email protected]>
Date: Wed, 31 Jan 2007 11:38:53 +0200
Subject: [PATCH] [JFFS2] use MTD_OOB_AUTO

Nowadays MTD supports an MTD_OOB_AUTO option which allows users to access free bytes in NAND's OOB as a contiguous buffer, although it may be highly discontinuous.

This patch teaches JFFS2 to use this nice feature instead of the old MTD_OOB_PLACE option. This for example caused problems with OneNAND. Now JFFS2 does not care how are the free bytes situated.

This may change position of the clean marker on some flashes, but this is not a problem. JFFS2 will just re-erase the empty eraseblocks and write the new (correct) clean marker.

---
 fs/jffs2/jffs2_fs_sb.h |   12 +--
 fs/jffs2/scan.c        |   10 ++-
 fs/jffs2/wbuf.c        |  206 +++++++++++++++++-------------------------------
 3 files changed, 84 insertions(+), 144 deletions(-)
fs/jffs2/jffs2_fs_sb.h +3 -9
Index: dedekind-mtd-2.6.git/fs/jffs2/jffs2_fs_sb.h
===================================================================
@@ -98,20 +98,14 @@ struct jffs2_sb_info {
 	uint32_t wbuf_pagesize; /* 0 for NOR and other flashes with no wbuf */
 
 #ifdef CONFIG_JFFS2_FS_WRITEBUFFER
-	/* Write-behind buffer for NAND flash */
-	unsigned char *wbuf;
-	unsigned char *oobbuf;
+	unsigned char *wbuf; /* Write-behind buffer for NAND flash */
 	uint32_t wbuf_ofs;
 	uint32_t wbuf_len;
 	struct jffs2_inodirty *wbuf_inodes;
-
 	struct rw_semaphore wbuf_sem;	/* Protects the write buffer */
 
-	/* Information about out-of-band area usage... */
-	struct nand_ecclayout *ecclayout;
-	uint32_t badblock_pos;
-	uint32_t fsdata_pos;
-	uint32_t fsdata_len;
+	unsigned char *oobbuf;
+	int oobavail; /* How many bytes are available for JFFS2 in OOB */
 #endif
 
 	struct jffs2_summary *summary;		/* Summary information */
fs/jffs2/scan.c +7 -3
Index: dedekind-mtd-2.6.git/fs/jffs2/scan.c
===================================================================
@@ -450,16 +450,20 @@ static int jffs2_scan_eraseblock (struct
 
 #ifdef CONFIG_JFFS2_FS_WRITEBUFFER
 	if (jffs2_cleanmarker_oob(c)) {
-		int ret = jffs2_check_nand_cleanmarker(c, jeb);
+		int ret;
+
+		if (c->mtd->block_isbad(c->mtd, jeb->offset))
+			return BLK_STATE_BADBLOCK;
+
+		ret = jffs2_check_nand_cleanmarker(c, jeb);
 		D2(printk(KERN_NOTICE "jffs_check_nand_cleanmarker returned %d\n",ret));
+
 		/* Even if it's not found, we still scan to see
 		   if the block is empty. We use this information
 		   to decide whether to erase it or not. */
 		switch (ret) {
 		case 0:		cleanmarkerfound = 1; break;
 		case 1: 	break;
-		case 2: 	return BLK_STATE_BADBLOCK;
-		case 3:		return BLK_STATE_ALLDIRTY; /* Block has failed to erase min. once */
 		default: 	return ret;
 		}
 	}
fs/jffs2/wbuf.c +74 -132

291 lines not on this page · raw message

#1 lore raw
Quoted 8 lines
fs/jffs2/wbuf.c +3 -1
-#define NR_OOB_SCAN_PAGES	4
+#define NR_OOB_SCAN_PAGES 4
+
+/* For historical reasons we use only 12 bytes for OOB clean 

marker */

fs/jffs2/wbuf.c exit: · line 970

Quoted 7 lines
fs/jffs2/wbuf.c +7 -0
+#define OOB_CM_SIZE 12
+
+static struct jffs2_unknown_node oob_cleanmarker = {
+	.magic = cpu_to_je16(JFFS2_MAGIC_BITMASK),
+	.nodetype = cpu_to_je16(JFFS2_NODETYPE_CLEANMARKER),
+	.totlen = cpu_to_je32(8)
+};

Looks good, how about make it const?

-

Quoted 6 lines
fs/jffs2/wbuf.c +1 -5
-	n.magic = cpu_to_je16 (JFFS2_MAGIC_BITMASK);
-	n.nodetype = cpu_to_je16 (JFFS2_NODETYPE_CLEANMARKER);
-	n.totlen = cpu_to_je32 (8);
-	p = (unsigned char *) &n;
-	b = c->oobbuf + c->fsdata_pos;
+	p = (unsigned char *)&oob_cleanmarker;

fs/jffs2/wbuf.c int jffs2_check_oob_empty(struct jffs2_s · line 1051

Quoted 10 lines
fs/jffs2/wbuf.c +5 -3
+	b = c->oobbuf;
 
-	for (i = c->fsdata_len; i; i--) {
-		if (*b++ != *p++)
+	for (i = 0; i < cmlen; i++)
+		if (*b++ != *p++) {
 			ret = 1;
-	}
+			break;
+		}

Why we compare it with character pointer? just use memcmp

if (memcmp(p, b, cmlen))

	ret = 1;

Are there any reasons?

Thank you,
Kyungmin Park

#2 lore raw
Quoted 1 line

Looks good, how about make it const?

Ok.

Quoted 1 line

Why we compare it with character pointer? just use memcmp

Ok, thanks

I'll send updated patch. We've tested it with our device - works like a charm. May you please try it on NAND and confirm it is OK?

Signature · 2 lines
Best regards,
Artem Bityutskiy (Битюцкий Артём)