[PATCH v4 3/6] fat: zero out seek range on _fat_get_block

[PATCH v4 3/6] fat: zero out seek range on _fat_get_block

[Namjae Jeon]

From: Namjae Jeon <namjae.jeon@samsung.com>

For normal buffered write operations, normally if we try to write to an
offset > than file size, it does a cont_expand_zero till that offset.
Now, in case of fallocated regions, since the blocks are already
allocated.  So, make it zero out that buffers for those blocks till the
seek'ed offset.

Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
---
 fs/fat/inode.c | 24 +++++++++++++++++++++++-
 1 file changed, 23 insertions(+), 1 deletion(-)

diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index ca222ef..a594dae 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -54,6 +54,25 @@ static int fat_add_cluster(struct inode *inode)
 	return err;
 }
 
+static void check_fallocated_region(struct inode *inode, sector_t iblock,
+		unsigned long *max_blocks, struct buffer_head *bh_result)
+{
+	struct super_block *sb = inode->i_sb;
+	sector_t last_block, disk_block;
+	const unsigned long blocksize = sb->s_blocksize;
+	const unsigned char blocksize_bits = sb->s_blocksize_bits;
+
+	last_block = (MSDOS_I(inode)->mmu_private + (blocksize - 1))
+		>> blocksize_bits;
+	disk_block = (MSDOS_I(inode)->i_disksize + (blocksize - 1))
+		>> blocksize_bits;
+	if (iblock >= last_block && iblock <= disk_block) {
+		MSDOS_I(inode)->mmu_private += *max_blocks << blocksize_bits;
+		set_buffer_new(bh_result);
+	}
+
+}
+
 static inline int __fat_get_block(struct inode *inode, sector_t iblock,
 				  unsigned long *max_blocks,
 				  struct buffer_head *bh_result, int create)
@@ -68,8 +87,11 @@ static inline int __fat_get_block(struct inode *inode, sector_t iblock,
 	if (err)
 		return err;
 	if (phys) {
-		map_bh(bh_result, sb, phys);
 		*max_blocks = min(mapped_blocks, *max_blocks);
+		if (create)
+			check_fallocated_region(inode, iblock, max_blocks,
+				bh_result);
+		map_bh(bh_result, sb, phys);
 		return 0;
 	}
 	if (!create)
-- 
1.7.11-rc0

Re: [PATCH v4 3/6] fat: zero out seek range on _fat_get_block

[OGAWA Hirofumi]

Namjae Jeon <linkinjeon@gmail.com> writes:

> +static void check_fallocated_region(struct inode *inode, sector_t iblock,
> +		unsigned long *max_blocks, struct buffer_head *bh_result)
> +{
> +	struct super_block *sb = inode->i_sb;
> +	sector_t last_block, disk_block;
> +	const unsigned long blocksize = sb->s_blocksize;
> +	const unsigned char blocksize_bits = sb->s_blocksize_bits;
> +
> +	last_block = (MSDOS_I(inode)->mmu_private + (blocksize - 1))
> +		>> blocksize_bits;
> +	disk_block = (MSDOS_I(inode)->i_disksize + (blocksize - 1))
> +		>> blocksize_bits;
> +	if (iblock >= last_block && iblock <= disk_block) {

Maybe off-by-one error. If iblock == disk_block, phys should not be set
by fat_bmap()?
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

Re: [PATCH v4 3/6] fat: zero out seek range on _fat_get_block

[Namjae Jeon]

2014-03-18 23:59 GMT+09:00, OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>:
> Namjae Jeon <linkinjeon@gmail.com> writes:
>
>> +static void check_fallocated_region(struct inode *inode, sector_t
>> iblock,
>> +		unsigned long *max_blocks, struct buffer_head *bh_result)
>> +{
>> +	struct super_block *sb = inode->i_sb;
>> +	sector_t last_block, disk_block;
>> +	const unsigned long blocksize = sb->s_blocksize;
>> +	const unsigned char blocksize_bits = sb->s_blocksize_bits;
>> +
>> +	last_block = (MSDOS_I(inode)->mmu_private + (blocksize - 1))
>> +		>> blocksize_bits;
>> +	disk_block = (MSDOS_I(inode)->i_disksize + (blocksize - 1))
>> +		>> blocksize_bits;
>> +	if (iblock >= last_block && iblock <= disk_block) {
>
Hi OGAWA.
> Maybe off-by-one error. If iblock == disk_block, phys should not be set
> by fat_bmap()?
Yes, right, iblock == disk_block case is not needed in this condition
because fat_bmap don't set to phys when iblock==diskblock.(phys will
be 0)
So I will change it like this.

if (iblock >= last_block && iblock < disk_block) {

And I will fix your review comments in other patches.

Thanks for review!
> --
> OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
>