Similar to ocfs2, btrfs also supports that extents can be shared by
different inodes, and there are some userspace tools requesting
for this kind of ''space shared infomation''.[1]
ocfs2 uses flag FIEMAP_EXTENT_SHARED, so does btrfs.
[1]:
http://thr3ads.net/ocfs2-devel/2010/09/489052-PATCH-3-3-shared-du-using-fiemap-to-figure-up-the-shared-extents-per-file-and-the-footprint-in
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
---
RFC->PATCH: use type ''unsigned long'' instead of
''int'', as David suggested.
fs/btrfs/extent_io.c | 34 +++++++++++++++++++++++++++++++++-
1 files changed, 33 insertions(+), 1 deletions(-)
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 09582b8..efd91d7 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -20,6 +20,7 @@
#include "check-integrity.h"
#include "locking.h"
#include "rcu-string.h"
+#include "backref.h"
static struct kmem_cache *extent_state_cache;
static struct kmem_cache *extent_buffer_cache;
@@ -4056,6 +4057,19 @@ static struct extent_map *get_extent_skip_holes(struct
inode *inode,
return NULL;
}
+static noinline int count_ext_ref(u64 inum, u64 offset, u64 root_id, void *ctx)
+{
+ int cnt = *((int *)ctx);
+
+ cnt++;
+ *((int *)ctx) = cnt;
+
+ /* Now we''re sure that the extent is shared. */
+ if (cnt > 1)
+ return 1;
+ return 0;
+}
+
int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
__u64 start, __u64 len, get_extent_t *get_extent)
{
@@ -4122,7 +4136,7 @@ int extent_fiemap(struct inode *inode, struct
fiemap_extent_info *fieinfo,
last = found_key.offset;
last_for_get_extent = last + 1;
}
- btrfs_free_path(path);
+ btrfs_release_path(path);
/*
* we might have some extents allocated but more delalloc past those
@@ -4192,7 +4206,24 @@ int extent_fiemap(struct inode *inode, struct
fiemap_extent_info *fieinfo,
flags |= (FIEMAP_EXTENT_DELALLOC |
FIEMAP_EXTENT_UNKNOWN);
} else {
+ unsigned long ref_cnt = 0;
+
disko = em->block_start + offset_in_extent;
+
+ /*
+ * As btrfs supports shared space, this information
+ * can be exported to userspace tools via
+ * flag FIEMAP_EXTENT_SHARED.
+ */
+ ret = iterate_inodes_from_logical(
+ em->block_start,
+ BTRFS_I(inode)->root->fs_info,
+ path, count_ext_ref, &ref_cnt);
+ if (ret < 0)
+ goto out_free;
+
+ if (ref_cnt > 1)
+ flags |= FIEMAP_EXTENT_SHARED;
}
if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
flags |= FIEMAP_EXTENT_ENCODED;
@@ -4226,6 +4257,7 @@ out_free:
out:
unlock_extent_cached(&BTRFS_I(inode)->io_tree, start, start + len - 1,
&cached_state, GFP_NOFS);
+ btrfs_free_path(path);
return ret;
}
--
1.7.7
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Josef Bacik
2013-Sep-18 20:53 UTC
Re: [PATCH] Btrfs: export btrfs space shared info to userspace
On Mon, Sep 16, 2013 at 03:40:12PM +0800, Liu Bo wrote:> Similar to ocfs2, btrfs also supports that extents can be shared by > different inodes, and there are some userspace tools requesting > for this kind of ''space shared infomation''.[1] > > ocfs2 uses flag FIEMAP_EXTENT_SHARED, so does btrfs. > > [1]: http://thr3ads.net/ocfs2-devel/2010/09/489052-PATCH-3-3-shared-du-using-fiemap-to-figure-up-the-shared-extents-per-file-and-the-footprint-in > > Signed-off-by: Liu Bo <bo.li.liu@oracle.com> > --- > RFC->PATCH: use type ''unsigned long'' instead of ''int'', as David suggested.This broke any xfstest that used fiemap, so generic/255, generic/316 and shared/218. Thanks, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Liu Bo
2013-Sep-20 00:22 UTC
Re: [PATCH] Btrfs: export btrfs space shared info to userspace
On Wed, Sep 18, 2013 at 04:53:04PM -0400, Josef Bacik wrote:> On Mon, Sep 16, 2013 at 03:40:12PM +0800, Liu Bo wrote: > > Similar to ocfs2, btrfs also supports that extents can be shared by > > different inodes, and there are some userspace tools requesting > > for this kind of ''space shared infomation''.[1] > > > > ocfs2 uses flag FIEMAP_EXTENT_SHARED, so does btrfs. > > > > [1]: http://thr3ads.net/ocfs2-devel/2010/09/489052-PATCH-3-3-shared-du-using-fiemap-to-figure-up-the-shared-extents-per-file-and-the-footprint-in > > > > Signed-off-by: Liu Bo <bo.li.liu@oracle.com> > > --- > > RFC->PATCH: use type ''unsigned long'' instead of ''int'', as David suggested. > > This broke any xfstest that used fiemap, so generic/255, generic/316 and > shared/218. Thanks,I''ll look into it ASAP, thanks, Josef. -liubo -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html