Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1555146
| From | Jan Kara <jack@suse.cz> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | Re: [PATCH 07/12 linux-next] udf: store allocation offset in udf_prealloc_extents() |
| Date | 2017-01-10 12:00 +0100 |
| Message-ID | <sY2jL-he-11@gated-at.bofh.it> (permalink) |
| References | <sWJMd-7eR-3@gated-at.bofh.it> <sWJMe-7eR-49@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
On Fri 06-01-17 21:53:55, Fabian Frederick wrote:
> recalculate offset if needed.
>
> Signed-off-by: Fabian Frederick <fabf@skynet.be>
IMHO not really worth it since numalloc gets updated and forgetting to
update the byte variant is just too easy. Skipped.
Honza
> ---
> fs/udf/inode.c | 21 +++++++++------------
> 1 file changed, 9 insertions(+), 12 deletions(-)
>
> diff --git a/fs/udf/inode.c b/fs/udf/inode.c
> index f563e97..66514ad 100644
> --- a/fs/udf/inode.c
> +++ b/fs/udf/inode.c
> @@ -998,10 +998,10 @@ static void udf_prealloc_extents(struct inode *inode, int c, int lastblock,
> length : UDF_DEFAULT_PREALLOC_BLOCKS) -
> currlength);
> if (numalloc) {
> + loff_t pos = numalloc << inode->i_sb->s_blocksize_bits;
> +
> if (start == (c + 1))
> - laarr[start].extLength +=
> - (numalloc <<
> - inode->i_sb->s_blocksize_bits);
> + laarr[start].extLength += pos;
> else {
> memmove(&laarr[c + 2], &laarr[c + 1],
> sizeof(struct long_ad) * (*endnum - (c + 1)));
> @@ -1011,9 +1011,7 @@ static void udf_prealloc_extents(struct inode *inode, int c, int lastblock,
> laarr[c].extLocation.
> partitionReferenceNum;
> laarr[c + 1].extLength =
> - EXT_NOT_RECORDED_ALLOCATED |
> - (numalloc <<
> - inode->i_sb->s_blocksize_bits);
> + EXT_NOT_RECORDED_ALLOCATED | pos;
> start = c + 1;
> }
>
> @@ -1024,12 +1022,12 @@ static void udf_prealloc_extents(struct inode *inode, int c, int lastblock,
> inode->i_sb->s_blocksize_bits;
>
> if (elen > numalloc) {
> - laarr[i].extLength -=
> - (numalloc <<
> - inode->i_sb->s_blocksize_bits);
> - numalloc = 0;
> + laarr[i].extLength -= pos;
> + numalloc = pos = 0;
> } else {
> numalloc -= elen;
> + pos = numalloc <<
> + inode->i_sb->s_blocksize_bits;
> if (*endnum > (i + 1))
> memmove(&laarr[i],
> &laarr[i + 1],
> @@ -1039,8 +1037,7 @@ static void udf_prealloc_extents(struct inode *inode, int c, int lastblock,
> (*endnum)--;
> }
> }
> - UDF_I(inode)->i_lenExtents +=
> - numalloc << inode->i_sb->s_blocksize_bits;
> + UDF_I(inode)->i_lenExtents += pos;
> }
> }
> }
> --
> 2.7.4
>
>
--
Jan Kara <jack@suse.com>
SUSE Labs, CR
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[PATCH 00/12 linux-next] udf: clean-up Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
[PATCH 09/12 linux-next] udf: merge module informations in super.c Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
Re: [PATCH 09/12 linux-next] udf: merge module informations in super.c Jan Kara <jack@suse.cz> - 2017-01-10 12:00 +0100
[PATCH 04/12 linux-next] udf: remove unneeded line break Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
Re: [PATCH 04/12 linux-next] udf: remove unneeded line break Jan Kara <jack@suse.cz> - 2017-01-10 11:40 +0100
[PATCH 05/12 linux-next] udf: remove empty condition Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
Re: [PATCH 05/12 linux-next] udf: remove empty condition Jan Kara <jack@suse.cz> - 2017-01-10 11:40 +0100
[PATCH 08/12 linux-next] udf: remove next_epos from udf_update_extent_cache() Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
Re: [PATCH 08/12 linux-next] udf: remove next_epos from udf_update_extent_cache() Jan Kara <jack@suse.cz> - 2017-01-10 12:00 +0100
[PATCH 01/12 linux-next] udf: use __packed instead of __attribute__ ((packed)) Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
Re: [PATCH 01/12 linux-next] udf: use __packed instead of __attribute__ ((packed)) Jan Kara <jack@suse.cz> - 2017-01-10 11:40 +0100
[PATCH 02/12 linux-next] udf: use pointer for kernel_long_ad argument Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
Re: [PATCH 02/12 linux-next] udf: use pointer for kernel_long_ad argument Jan Kara <jack@suse.cz> - 2017-01-10 11:40 +0100
[PATCH 10/12 linux-next] udf: atomically read inode size Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
[PATCH 11/12 linux-next] udf: replace 0xFFFFFFFF by ~0 Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
Re: [PATCH 10/12 linux-next] udf: atomically read inode size Jan Kara <jack@suse.cz> - 2017-01-10 12:00 +0100
[PATCH 07/12 linux-next] udf: store allocation offset in udf_prealloc_extents() Fabian Frederick <fabf@skynet.be> - 2017-01-06 22:00 +0100
Re: [PATCH 07/12 linux-next] udf: store allocation offset in udf_prealloc_extents() Jan Kara <jack@suse.cz> - 2017-01-10 12:00 +0100
Re: [PATCH 00/12 linux-next] udf: clean-up Jan Kara <jack@suse.cz> - 2017-01-10 12:10 +0100
csiph-web