->s_group_desc have to be zero filled because if sb_read() failed
we jump to following error path.
failed_mount2:
        for (i = 0; i < db_count; i++)
                brelse(sbi->s_group_desc[i]);<< Bad things may happen here

Signed-off-by: Dmitry Monakhov <[EMAIL PROTECTED]>
---
 fs/ext3/super.c |    2 +-
 fs/ext4/super.c |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index f8ac18f..208738e 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -1718,7 +1718,7 @@ static int ext3_fill_super (struct super_block *sb, void 
*data, int silent)
                                       / EXT3_BLOCKS_PER_GROUP(sb)) + 1;
        db_count = (sbi->s_groups_count + EXT3_DESC_PER_BLOCK(sb) - 1) /
                   EXT3_DESC_PER_BLOCK(sb);
-       sbi->s_group_desc = kmalloc(db_count * sizeof (struct buffer_head *),
+       sbi->s_group_desc = kzalloc(db_count * sizeof (struct buffer_head *),
                                    GFP_KERNEL);
        if (sbi->s_group_desc == NULL) {
                printk (KERN_ERR "EXT3-fs: not enough memory\n");
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 8f1d2f6..fefffc0 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -1830,7 +1830,7 @@ static int ext4_fill_super (struct super_block *sb, void 
*data, int silent)
        sbi->s_groups_count = blocks_count;
        db_count = (sbi->s_groups_count + EXT4_DESC_PER_BLOCK(sb) - 1) /
                   EXT4_DESC_PER_BLOCK(sb);
-       sbi->s_group_desc = kmalloc(db_count * sizeof (struct buffer_head *),
+       sbi->s_group_desc = kzalloc(db_count * sizeof (struct buffer_head *),
                                    GFP_KERNEL);
        if (sbi->s_group_desc == NULL) {
                printk (KERN_ERR "EXT4-fs: not enough memory\n");
-- 
1.5.2.2


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Reply via email to