From: Theodore Ts'o <tytso@mit.edu>
To: linux-ext4@vger.kernel.org
Cc: Theodore Ts'o <tytso@mit.edu>
Subject: [PATCH, RFC 09/12] ext4: teach ext4_ext_map_blocks() about the bigalloc feature
Date: Sat, 19 Mar 2011 17:28:34 -0400 [thread overview]
Message-ID: <1300570117-24048-10-git-send-email-tytso@mit.edu> (raw)
In-Reply-To: <1300570117-24048-1-git-send-email-tytso@mit.edu>
If we need to allocate a new block in ext4_ext_map_blocks(), the
function needs to see if the cluster has already been allocated.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
---
fs/ext4/extents.c | 132 +++++++++++++++++++++++++++++++++++++++++++++++-----
1 files changed, 119 insertions(+), 13 deletions(-)
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 9ea1bc6..3221b79 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -1273,7 +1273,8 @@ static int ext4_ext_search_left(struct inode *inode,
*/
static int ext4_ext_search_right(struct inode *inode,
struct ext4_ext_path *path,
- ext4_lblk_t *logical, ext4_fsblk_t *phys)
+ ext4_lblk_t *logical, ext4_fsblk_t *phys,
+ struct ext4_extent **ret_ex)
{
struct buffer_head *bh = NULL;
struct ext4_extent_header *eh;
@@ -1315,9 +1316,7 @@ static int ext4_ext_search_right(struct inode *inode,
return -EIO;
}
}
- *logical = le32_to_cpu(ex->ee_block);
- *phys = ext4_ext_pblock(ex);
- return 0;
+ goto found_extent;
}
if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
@@ -1330,9 +1329,7 @@ static int ext4_ext_search_right(struct inode *inode,
if (ex != EXT_LAST_EXTENT(path[depth].p_hdr)) {
/* next allocated block in this leaf */
ex++;
- *logical = le32_to_cpu(ex->ee_block);
- *phys = ext4_ext_pblock(ex);
- return 0;
+ goto found_extent;
}
/* go up and search for index to the right */
@@ -1375,9 +1372,12 @@ got_index:
return -EIO;
}
ex = EXT_FIRST_EXTENT(eh);
+found_extent:
*logical = le32_to_cpu(ex->ee_block);
*phys = ext4_ext_pblock(ex);
- put_bh(bh);
+ *ret_ex = ex;
+ if (bh)
+ put_bh(bh);
return 0;
}
@@ -1606,7 +1606,8 @@ static int ext4_ext_try_to_merge(struct inode *inode,
* such that there will be no overlap, and then returns 1.
* If there is no overlap found, it returns 0.
*/
-static unsigned int ext4_ext_check_overlap(struct inode *inode,
+static unsigned int ext4_ext_check_overlap(struct ext4_sb_info *sbi,
+ struct inode *inode,
struct ext4_extent *newext,
struct ext4_ext_path *path)
{
@@ -1620,6 +1621,7 @@ static unsigned int ext4_ext_check_overlap(struct inode *inode,
if (!path[depth].p_ext)
goto out;
b2 = le32_to_cpu(path[depth].p_ext->ee_block);
+ b2 &= ~(sbi->s_cluster_ratio - 1);
/*
* get the next allocated block if the extent in the path
@@ -3274,6 +3276,82 @@ out2:
}
/*
+ * get_implied_cluster_alloc - check to see if the requested
+ * allocation (in the map structure) overlaps with a cluster already
+ * allocated in an extent.
+ * @map The requested lblk->pblk mapping
+ * @c_len The number of blocks in a cluster
+ * @ex The extent structure which might contain an implied
+ * cluster allocation
+ *
+ * This function is called by ext4_ext_map_blocks() after we failed to
+ * find blocks that were already in the inode's extent tree. Hence,
+ * we know that the beginning of the requested region cannot overlap
+ * the extent from the inode's extent tree. There are two cases we
+ * want to catch. The first is this case:
+ *
+ * |--- cluster # N--|
+ * |--- extent ---| |---- requested region ---|
+ * |==========|
+ *
+ * The second case that we need to test for is this one:
+ *
+ * |--------- cluster # N ----------------|
+ * |--- requested region --| |------- extent ----|
+ * |=======================|
+ *
+ * In each case, we need to return the extent labelled as "|====|"
+ * from cluster #N.
+ */
+static int get_implied_cluster_alloc(struct ext4_sb_info *sbi,
+ struct ext4_map_blocks *map,
+ struct ext4_extent *ex)
+{
+ ext4_lblk_t c_offset = map->m_lblk & (sbi->s_cluster_ratio-1);
+ ext4_lblk_t ex_cluster_start, ex_cluster_end;
+ ext4_lblk_t rr_cluster_start, rr_cluster_end;
+ ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
+ ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
+ unsigned short ee_len = ext4_ext_get_actual_len(ex);
+
+ if (!c_offset)
+ return 0;
+
+ /* The extent passed in that we are trying to match */
+ ex_cluster_start = EXT4_B2C(sbi, ee_block);
+ ex_cluster_end = EXT4_B2C(sbi, ee_block + ee_len - 1);
+
+ /* The requested region passed into ext4_map_blocks() */
+ rr_cluster_start = EXT4_B2C(sbi, map->m_lblk);
+ rr_cluster_end = EXT4_B2C(sbi, map->m_lblk + map->m_len - 1);
+
+ if ((rr_cluster_start == ex_cluster_end) ||
+ (rr_cluster_start == ex_cluster_start)) {
+ if (rr_cluster_start == ex_cluster_end)
+ ee_start += ee_len - 1;
+ map->m_pblk = (ee_start & ~(sbi->s_cluster_ratio - 1)) +
+ c_offset;
+ map->m_len = min(map->m_len,
+ (unsigned) sbi->s_cluster_ratio - c_offset);
+ /*
+ * Check for and handle this case:
+ *
+ * |--------- cluster # N-------------|
+ * |------- extent ----|
+ * |--- requested region ---|
+ * |===========|
+ */
+
+ if ((map->m_lblk < ee_block) &&
+ (map->m_lblk + map->m_len > ee_block))
+ map->m_len = ee_block - map->m_lblk;
+ return 1;
+ }
+ return 0;
+}
+
+
+/*
* Block allocation/map/preallocation routine for extents based files
*
*
@@ -3296,12 +3374,14 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
{
struct ext4_ext_path *path = NULL;
struct ext4_extent_header *eh;
- struct ext4_extent newex, *ex;
+ struct ext4_extent newex, *ex, *ex2;
+ struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
ext4_fsblk_t newblock;
int err = 0, depth, ret;
unsigned int allocated = 0;
struct ext4_allocation_request ar;
ext4_io_end_t *io = EXT4_I(inode)->cur_aio_dio;
+ ext4_lblk_t cluster_offset;
ext_debug("blocks %u/%u requested for inode %lu\n",
map->m_lblk, map->m_len, inode->i_ino);
@@ -3398,9 +3478,21 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
ext4_ext_put_gap_in_cache(inode, path, map->m_lblk);
goto out2;
}
+
/*
* Okay, we need to do block allocation.
*/
+ cluster_offset = map->m_lblk & (sbi->s_cluster_ratio-1);
+ /*
+ * If we are doing bigalloc, check to see if the extent returned
+ * by ext4_find_extent() implies a cluster we can use.
+ */
+ if (cluster_offset && ex &&
+ get_implied_cluster_alloc(sbi, map, ex)) {
+ ar.len = allocated = map->m_len;
+ newblock = map->m_pblk;
+ goto got_allocated_blocks;
+ }
/* find neighbour allocated blocks */
ar.lleft = map->m_lblk;
@@ -3408,10 +3500,20 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
if (err)
goto out2;
ar.lright = map->m_lblk;
- err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright);
+ ex2 = 0;
+ err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright, &ex2);
if (err)
goto out2;
+ /* Check if the extent after searching to the right implies a
+ * cluster we can use. */
+ if (cluster_offset && ex2 &&
+ get_implied_cluster_alloc(sbi, map, ex2)) {
+ ar.len = allocated = map->m_len;
+ newblock = map->m_pblk;
+ goto got_allocated_blocks;
+ }
+
/*
* See if request is beyond maximum number of blocks we can have in
* a single extent. For an initialized extent this limit is
@@ -3428,7 +3530,7 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
/* Check if we can really insert (m_lblk)::(m_lblk + m_len) extent */
newex.ee_block = cpu_to_le32(map->m_lblk);
newex.ee_len = cpu_to_le16(map->m_len);
- err = ext4_ext_check_overlap(inode, &newex, path);
+ err = ext4_ext_check_overlap(sbi, inode, &newex, path);
if (err)
allocated = ext4_ext_get_actual_len(&newex);
else
@@ -3438,7 +3540,7 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
ar.inode = inode;
ar.goal = ext4_ext_find_goal(inode, path, map->m_lblk);
ar.logical = map->m_lblk;
- ar.len = allocated;
+ ar.len = EXT4_NUM_B2C(sbi, allocated);
if (S_ISREG(inode->i_mode))
ar.flags = EXT4_MB_HINT_DATA;
else
@@ -3449,7 +3551,11 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
goto out2;
ext_debug("allocate new block: goal %llu, found %llu/%u\n",
ar.goal, newblock, allocated);
+ ar.len <<= sbi->s_cluster_bits;
+ if (ar.len > allocated)
+ ar.len = allocated;
+got_allocated_blocks:
/* try to insert new extent into found leaf and return */
ext4_ext_store_pblock(&newex, newblock);
newex.ee_len = cpu_to_le16(ar.len);
--
1.7.3.1
next prev parent reply other threads:[~2011-03-19 21:37 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-03-19 21:28 [PATCH, RFC 00/12] bigalloc patchset Theodore Ts'o
2011-03-19 21:28 ` [PATCH, RFC 01/12] ext4: read-only support for bigalloc file systems Theodore Ts'o
2011-03-21 19:35 ` Lukas Czerner
2011-03-22 17:02 ` Ted Ts'o
2011-03-23 10:28 ` Lukas Czerner
2011-03-19 21:28 ` [PATCH, RFC 02/12] ext4: enforce bigalloc restrictions (e.g., no online resizing, etc.) Theodore Ts'o
2011-03-19 21:28 ` [PATCH, RFC 03/12] ext4: Convert instances of EXT4_BLOCKS_PER_GROUP to EXT4_CLUSTERS_PER_GROUP Theodore Ts'o
2011-03-20 10:26 ` Amir Goldstein
2011-03-21 13:12 ` Ted Ts'o
2011-03-19 21:28 ` [PATCH, RFC 04/12] ext4: Remove block bitmap initialization in ext4_new_inode() Theodore Ts'o
2011-03-19 21:28 ` [PATCH, RFC 05/12] ext4: factor out block group accounting into functions Theodore Ts'o
2011-03-19 21:28 ` [PATCH, RFC 06/12] ext4: split out ext4_free_blocks_after_init() Theodore Ts'o
2011-03-19 21:28 ` [PATCH, RFC 07/12] ext4: bigalloc changes to block bitmap initialization functions Theodore Ts'o
2011-03-19 21:28 ` [PATCH, RFC 08/12] ext4: Convert block group-relative offsets to use clusters Theodore Ts'o
2011-03-19 21:28 ` Theodore Ts'o [this message]
2011-03-19 21:28 ` [PATCH, RFC 10/12] ext4: teach ext4_statfs() to deal with clusters if bigalloc is enabled Theodore Ts'o
2011-03-21 20:17 ` Lukas Czerner
2011-03-22 22:09 ` Ted Ts'o
2011-03-19 21:28 ` [PATCH, RFC 11/12] ext4: tune mballoc's default group prealloc size for bigalloc file systems Theodore Ts'o
2011-03-19 21:28 ` [PATCH, RFC 12/12] ext4: enable mounting bigalloc as read/write Theodore Ts'o
2011-03-20 10:33 ` [PATCH, RFC 00/12] bigalloc patchset Amir Goldstein
2011-03-21 8:55 ` Andreas Dilger
2011-03-21 11:31 ` Rogier Wolff
2011-03-21 13:24 ` Ted Ts'o
2011-03-21 23:42 ` Andreas Dilger
2011-04-05 17:23 ` Coly Li
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1300570117-24048-10-git-send-email-tytso@mit.edu \
--to=tytso@mit.edu \
--cc=linux-ext4@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).