Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1672035 > unrolled thread

Re: [PATCH v5 27/28] ext4: xattr inode deduplication

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH v5 27/28] ext4: xattr inode deduplication Andreas Dilger <adilger@dilger.ca> - 2017-06-21 23:20 +0200
    Re: [PATCH v5 27/28] ext4: xattr inode deduplication Tahsin Erdogan <tahsin@google.com> - 2017-06-21 23:40 +0200
      Re: [PATCH v5 27/28] ext4: xattr inode deduplication Andreas Dilger <adilger@dilger.ca> - 2017-06-21 23:50 +0200

#1672035 — Re: [PATCH v5 27/28] ext4: xattr inode deduplication

[Multipart message — attachments visible in raw view] — view raw

On Jun 20, 2017, at 3:07 AM, Tahsin Erdogan <tahsin@google.com> wrote:
> 
> Ext4 now supports xattr values that are up to 64k in size (vfs limit).
> Large xattr values are stored in external inodes each one holding a
> single value. Once written the data blocks of these inodes are immutable.
> 
> The real world use cases are expected to have a lot of value duplication
> such as inherited acls etc. To reduce data duplication on disk, this patch
> implements a deduplicator that allows sharing of xattr inodes.
> 
> The deduplication is based on an in-memory hash lookup that is a best
> effort sharing scheme. When a xattr inode is read from disk (i.e.
> getxattr() call), its crc32c hash is added to a hash table. Before
> creating a new xattr inode for a value being set, the hash table is
> checked to see if an existing inode holds an identical value. If such an
> inode is found, the ref count on that inode is incremented. On value
> removal the ref count is decremented and if it reaches zero the inode is
> deleted.
> 
> The quota charging for such inodes is manually managed. Every reference
> holder is charged the full size as if there was no sharing happening.
> This is consistent with how xattr blocks are also charged.
> 
> Signed-off-by: Tahsin Erdogan <tahsin@google.com>
> ---
> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
> index d79d8d7bee88..59e9488c4876 100644
> --- a/fs/ext4/ext4.h
> +++ b/fs/ext4/ext4.h
> @@ -1517,6 +1517,7 @@ struct ext4_sb_info {
> 	long s_es_nr_inode;
> 	struct ext4_es_stats s_es_stats;
> 	struct mb_cache *s_mb_cache;
> +	struct mb_cache *s_ea_inode_cache;

These names should be consistent, like "s_ea_block_cache".

> diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c
> index 0484df8dadd1..d7e60358ec91 100644
> --- a/fs/ext4/xattr.c
> +++ b/fs/ext4/xattr.c
> @@ -108,6 +108,9 @@ const struct xattr_handler *ext4_xattr_handlers[] = {
> #define EXT4_GET_MB_CACHE(inode)	(((struct ext4_sb_info *) \
> 				inode->i_sb->s_fs_info)->s_mb_cache)
> 
> +#define EA_INODE_CACHE(inode)	(((struct ext4_sb_info *) \
> +				inode->i_sb->s_fs_info)->s_ea_inode_cache)

These names should be consistent, like EXT4_GET_EA_CACHE() or maybe
EXT4_GET_EA_BLOCK_CACHE() and EXT4_GET_EA_INODE_CACHE().

The rest of the changes look reasonable.

Cheers, Andreas

[toc] | [next] | [standalone]

#1672082

> Tashin, we are already using the "no_mbcache" option name, so would prefer
> to keep that working.  It would be OK to accept both option names to mean
> the same thing, and only document the "nombcache" option.

Updated patch to accept both nombcache and no_mbcache.

>>       struct mb_cache *s_mb_cache;
>> +     struct mb_cache *s_ea_inode_cache;
>
> These names should be consistent, like "s_ea_block_cache".

Yes, I will rename this to s_ea_block_cache.

>> #define EXT4_GET_MB_CACHE(inode)      (((struct ext4_sb_info *) \
>>                               inode->i_sb->s_fs_info)->s_mb_cache)
>>
>> +#define EA_INODE_CACHE(inode)        (((struct ext4_sb_info *) \
>> +                             inode->i_sb->s_fs_info)->s_ea_inode_cache)
>
> These names should be consistent, like EXT4_GET_EA_CACHE() or maybe
> EXT4_GET_EA_BLOCK_CACHE() and EXT4_GET_EA_INODE_CACHE().

How about EA_BLOCK_CACHE() and EA_INODE_CACHE() to keep them short?

[toc] | [prev] | [next] | [standalone]

#1672085

[Multipart message — attachments visible in raw view] — view raw

> On Jun 21, 2017, at 3:34 PM, Tahsin Erdogan <tahsin@google.com> wrote:
> 
>> Tashin, we are already using the "no_mbcache" option name, so would prefer
>> to keep that working.  It would be OK to accept both option names to mean
>> the same thing, and only document the "nombcache" option.
> 
> Updated patch to accept both nombcache and no_mbcache.
> 
>>>      struct mb_cache *s_mb_cache;
>>> +     struct mb_cache *s_ea_inode_cache;
>> 
>> These names should be consistent, like "s_ea_block_cache".
> 
> Yes, I will rename this to s_ea_block_cache.
> 
>>> #define EXT4_GET_MB_CACHE(inode)      (((struct ext4_sb_info *) \
>>>                              inode->i_sb->s_fs_info)->s_mb_cache)
>>> 
>>> +#define EA_INODE_CACHE(inode)        (((struct ext4_sb_info *) \
>>> +                             inode->i_sb->s_fs_info)->s_ea_inode_cache)
>> 
>> These names should be consistent, like EXT4_GET_EA_CACHE() or maybe
>> EXT4_GET_EA_BLOCK_CACHE() and EXT4_GET_EA_INODE_CACHE().
> 
> How about EA_BLOCK_CACHE() and EA_INODE_CACHE() to keep them short?

Sure, that is fine since these macros are local to xattr.c.

Cheers, Andreas

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web