Skip to contentSkip to conversation

Fix ext2 non-working

PATCH New unsik Kim 2 participants 2 messages Jan 30, 2009 – Feb 21, 2009

Subsystems (2)
lore
Share & tools
lore

unsik Kim

author
#0 lore raw

PATCH 1/1 Fix ext2 non-working

Hello, all?

Currently ext2ls, ext2load not working (I think sometimes it will work.) The reason is incorrect operations of ext2fs_read_inode and ext2fs_blockgroup.

I fix miscalculated block number(blkno) and offset (blkoff), and some hard coded values.

unsik Kim

---

Makefile +1 -0
index d26a004..9e9ba36 100644
@@ -403,6 +403,7 @@ TAG_SUBDIRS += fs/fat
 TAG_SUBDIRS += fs/fdos
 TAG_SUBDIRS += fs/jffs2
 TAG_SUBDIRS += fs/yaffs2
+TAG_SUBDIRS += fs/ext2
 TAG_SUBDIRS += net
 TAG_SUBDIRS += disk
 TAG_SUBDIRS += common
fs/ext2/ext2fs.c +22 -16
index 436f4a4..4ac2bdd 100644
@@ -110,6 +110,7 @@ struct ext2_block_group {
 	uint32_t inode_table_id;
 	uint16_t free_blocks;
 	uint16_t free_inodes;
+	uint16_t used_dir_cnt;
 	uint16_t pad;
 	uint32_t reserved[3];
 };
@@ -182,14 +183,22 @@ int indir2_blkno = -1;
 
 static int ext2fs_blockgroup
 	(struct ext2_data *data, int group, struct ext2_block_group *blkgrp) {
+	unsigned int blkno;
+	unsigned int blkoff;
+	unsigned int desc_per_blk;
+
+	desc_per_blk = EXT2_BLOCK_SIZE(data) / sizeof(struct ext2_block_group);
+
+	blkno = __le32_to_cpu(data->sblock.first_data_block) + 1 +
+		group / desc_per_blk;
+	blkoff = (group % desc_per_blk) * sizeof(struct ext2_block_group);
 #ifdef DEBUG
-	printf ("ext2fs read blockgroup\n");
+	printf ("ext2fs read %d group descriptor (blkno %d blkoff %d)\n",
+		 group, blkno, blkoff);
 #endif
 	return (ext2fs_devread
-		(((__le32_to_cpu (data->sblock.first_data_block) +
-		   1) << LOG2_EXT2_BLOCK_SIZE (data)),
-		 group * sizeof (struct ext2_block_group),
-		 sizeof (struct ext2_block_group), (char *) blkgrp));
+		(blkno << LOG2_EXT2_BLOCK_SIZE(data),
+		 blkoff, sizeof(struct ext2_block_group), (char *)blkgrp));
 }
 
 
@@ -203,11 +212,11 @@ static int ext2fs_read_inode
 	unsigned int blkno;
 	unsigned int blkoff;
 
-	/* It is easier to calculate if the first inode is 0.  */
-	ino--;
 #ifdef DEBUG
 	printf ("ext2fs read inode %d\n", ino);
 #endif
+	/* It is easier to calculate if the first inode is 0.  */
+	ino--;
 	status = ext2fs_blockgroup (data,
 				    ino /
 				    __le32_to_cpu (sblock->inodes_per_group),
@@ -215,19 +224,16 @@ static int ext2fs_read_inode
 	if (status == 0) {
 		return (0);
 	}
-	inodes_per_block = EXT2_BLOCK_SIZE (data) / 128;
-	blkno = (ino % __le32_to_cpu (sblock->inodes_per_group)) /
-		inodes_per_block;
-	blkoff = (ino % __le32_to_cpu (sblock->inodes_per_group)) %
-		inodes_per_block;
+	inodes_per_block = EXT2_BLOCK_SIZE(data) / __le32_to_cpu(sblock->inode_size);
+	blkno = blkgrp.inode_table_id +
+		(ino % __le32_to_cpu(sblock->inodes_per_group)) / inodes_per_block;
+	blkoff = (ino % inodes_per_block) * __le32_to_cpu(sblock->inode_size);
 #ifdef DEBUG
 	printf ("ext2fs read inode blkno %d blkoff %d\n", blkno, blkoff);
 #endif
 	/* Read the inode.  */
-	status = ext2fs_devread (((__le32_to_cpu (blkgrp.inode_table_id) +
-				   blkno) << LOG2_EXT2_BLOCK_SIZE (data)),
-				 sizeof (struct ext2_inode) * blkoff,
-				 sizeof (struct ext2_inode), (char *) inode);
+	status = ext2fs_devread (blkno << LOG2_EXT2_BLOCK_SIZE (data), blkoff,
+				sizeof (struct ext2_inode), (char *) inode);
 	if (status == 0) {
 		return (0);
 	}
#1 lore raw

PATCH 1/1 Fix ext2 non-working

Dear unsik Kim,

Quoted 7 lines · In message <[email protected]> you…

Hello, all?

Currently ext2ls, ext2load not working (I think sometimes it will work.) The reason is incorrect operations of ext2fs_read_inode and ext2fs_blockgroup.

I fix miscalculated block number(blkno) and offset (blkoff), and some hard coded values.

Hm.... I tried the current code without your patch but I could not find any error case.

But with your patch applied, I got this:

=> ext2ls ide 0:4 bin
** ext2fs_devread() read outside partition sector 536870912
Failed to mount ext2 filesystem...
** Bad ext2 partition or disk - ide 0:4 **

I tested this on a PowerPC system (big endian). Any chance there might be some endianess issues with your patch?

Best regards,

Wolfgang Denk

Signature · 4 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
I read part of it all the way through.