From: Chandan Rajendra <chandan@linux.vnet.ibm.com>
To: clm@fb.com, jbacik@fb.com, dsterba@suse.com
Cc: Chandan Rajendra <chandan@linux.vnet.ibm.com>,
linux-btrfs@vger.kernel.org
Subject: [PATCH V21 14/19] Btrfs: subpage-blocksize: Fix file defragmentation code
Date: Sun, 2 Oct 2016 18:54:23 +0530 [thread overview]
Message-ID: <1475414668-25954-15-git-send-email-chandan@linux.vnet.ibm.com> (raw)
In-Reply-To: <1475414668-25954-1-git-send-email-chandan@linux.vnet.ibm.com>
This commit gets file defragmentation code to work in subpage-blocksize
scenario. It does this by keeping track of page offsets that mark block
boundaries and passing them as arguments to the functions that implement
the defragmentation logic.
Signed-off-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
---
fs/btrfs/ioctl.c | 198 ++++++++++++++++++++++++++++++++++++++-----------------
1 file changed, 136 insertions(+), 62 deletions(-)
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index a222bad..4077fc1 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -902,12 +902,13 @@ out_unlock:
static int check_defrag_in_cache(struct inode *inode, u64 offset, u32 thresh)
{
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
struct extent_map *em = NULL;
struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
u64 end;
read_lock(&em_tree->lock);
- em = lookup_extent_mapping(em_tree, offset, PAGE_SIZE);
+ em = lookup_extent_mapping(em_tree, offset, root->sectorsize);
read_unlock(&em_tree->lock);
if (em) {
@@ -997,7 +998,7 @@ static struct extent_map *defrag_lookup_extent(struct inode *inode, u64 start)
struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
struct extent_map *em;
- u64 len = PAGE_SIZE;
+ u64 len = BTRFS_I(inode)->root->sectorsize;
/*
* hopefully we have this extent in the tree already, try without
@@ -1116,37 +1117,47 @@ out:
* before calling this.
*/
static int cluster_pages_for_defrag(struct inode *inode,
- struct page **pages,
- unsigned long start_index,
- unsigned long num_pages)
+ struct page **pages,
+ unsigned long start_index,
+ size_t pg_offset,
+ unsigned long num_blks)
{
- unsigned long file_end;
u64 isize = i_size_read(inode);
+ u64 start_blk;
+ u64 end_blk;
u64 page_start;
u64 page_end;
u64 page_cnt;
+ u64 blk_cnt;
int ret;
int i;
int i_done;
struct btrfs_ordered_extent *ordered;
struct extent_state *cached_state = NULL;
struct extent_io_tree *tree;
+ struct btrfs_root *root;
gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
- file_end = (isize - 1) >> PAGE_SHIFT;
- if (!isize || start_index > file_end)
+ root = BTRFS_I(inode)->root;
+ start_blk = (start_index << PAGE_SHIFT) + pg_offset;
+ start_blk >>= inode->i_blkbits;
+ end_blk = (isize - 1) >> inode->i_blkbits;
+ if (!isize || start_blk > end_blk)
return 0;
- page_cnt = min_t(u64, (u64)num_pages, (u64)file_end - start_index + 1);
+ blk_cnt = min_t(u64, (u64)num_blks, (u64)end_blk - start_blk + 1);
ret = btrfs_delalloc_reserve_space(inode,
- start_index << PAGE_SHIFT,
- page_cnt << PAGE_SHIFT);
+ start_blk << inode->i_blkbits,
+ blk_cnt << inode->i_blkbits);
if (ret)
return ret;
i_done = 0;
tree = &BTRFS_I(inode)->io_tree;
+ page_cnt = DIV_ROUND_UP(pg_offset + (blk_cnt << inode->i_blkbits),
+ PAGE_SIZE);
+
/* step one, lock all the pages */
for (i = 0; i < page_cnt; i++) {
struct page *page;
@@ -1157,12 +1168,22 @@ again:
break;
page_start = page_offset(page);
- page_end = page_start + PAGE_SIZE - 1;
+
+ if (i == 0)
+ page_start += pg_offset;
+
+ if (i == page_cnt - 1) {
+ page_end = (start_index << PAGE_SHIFT) + pg_offset;
+ page_end += (blk_cnt << inode->i_blkbits) - 1;
+ } else {
+ page_end = page_offset(page) + PAGE_SIZE - 1;
+ }
+
while (1) {
lock_extent_bits(tree, page_start, page_end,
&cached_state);
- ordered = btrfs_lookup_ordered_extent(inode,
- page_start);
+ ordered = btrfs_lookup_ordered_range(inode, page_start,
+ page_end - page_start + 1);
unlock_extent_cached(tree, page_start, page_end,
&cached_state, GFP_NOFS);
if (!ordered)
@@ -1201,7 +1222,7 @@ again:
}
pages[i] = page;
- i_done++;
+ i_done += (page_end - page_start + 1) >> inode->i_blkbits;
}
if (!i_done || ret)
goto out;
@@ -1213,55 +1234,77 @@ again:
* so now we have a nice long stream of locked
* and up to date pages, lets wait on them
*/
- for (i = 0; i < i_done; i++)
+ page_cnt = DIV_ROUND_UP(pg_offset + (i_done << inode->i_blkbits),
+ PAGE_SIZE);
+ for (i = 0; i < page_cnt; i++)
wait_on_page_writeback(pages[i]);
- page_start = page_offset(pages[0]);
- page_end = page_offset(pages[i_done - 1]) + PAGE_SIZE;
+ page_start = page_offset(pages[0]) + pg_offset;
+ page_end = page_start + (i_done << inode->i_blkbits) - 1;
lock_extent_bits(&BTRFS_I(inode)->io_tree,
- page_start, page_end - 1, &cached_state);
+ page_start, page_end, &cached_state);
clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start,
- page_end - 1, EXTENT_DIRTY | EXTENT_DELALLOC |
+ page_end, EXTENT_DIRTY | EXTENT_DELALLOC |
EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG, 0, 0,
&cached_state, GFP_NOFS);
- if (i_done != page_cnt) {
+ if (i_done != blk_cnt) {
spin_lock(&BTRFS_I(inode)->lock);
BTRFS_I(inode)->outstanding_extents++;
spin_unlock(&BTRFS_I(inode)->lock);
btrfs_delalloc_release_space(inode,
- start_index << PAGE_SHIFT,
- (page_cnt - i_done) << PAGE_SHIFT);
+ start_blk << inode->i_blkbits,
+ (blk_cnt - i_done) << inode->i_blkbits);
}
- set_extent_defrag(&BTRFS_I(inode)->io_tree, page_start, page_end - 1,
- &cached_state);
+ set_extent_defrag(&BTRFS_I(inode)->io_tree, page_start, page_end,
+ &cached_state);
unlock_extent_cached(&BTRFS_I(inode)->io_tree,
- page_start, page_end - 1, &cached_state,
+ page_start, page_end, &cached_state,
GFP_NOFS);
- for (i = 0; i < i_done; i++) {
+ for (i = 0; i < page_cnt; i++) {
clear_page_dirty_for_io(pages[i]);
ClearPageChecked(pages[i]);
set_page_extent_mapped(pages[i]);
+
+ page_start = page_offset(pages[i]);
+ if (i == 0)
+ page_start += pg_offset;
+
+ if (i == page_cnt - 1) {
+ page_end = page_offset(pages[0]) + pg_offset;
+ page_end += (i_done << inode->i_blkbits) - 1;
+ } else {
+ page_end = page_offset(pages[i]) + PAGE_SIZE - 1;
+ }
+
+ if (root->sectorsize < PAGE_SIZE)
+ set_page_blks_state(pages[i],
+ 1 << BLK_STATE_UPTODATE | 1 << BLK_STATE_DIRTY,
+ page_start, page_end);
set_page_dirty(pages[i]);
unlock_page(pages[i]);
put_page(pages[i]);
}
return i_done;
out:
- for (i = 0; i < i_done; i++) {
- unlock_page(pages[i]);
- put_page(pages[i]);
+ if (i_done) {
+ page_cnt = DIV_ROUND_UP(pg_offset + (i_done << inode->i_blkbits),
+ PAGE_SIZE);
+ for (i = 0; i < page_cnt; i++) {
+ unlock_page(pages[i]);
+ put_page(pages[i]);
+ }
}
+
btrfs_delalloc_release_space(inode,
- start_index << PAGE_SHIFT,
- page_cnt << PAGE_SHIFT);
+ start_blk << inode->i_blkbits,
+ blk_cnt << inode->i_blkbits);
return ret;
-
}
int btrfs_defrag_file(struct inode *inode, struct file *file,
@@ -1270,19 +1313,24 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct file_ra_state *ra = NULL;
+ unsigned long first_off, last_off;
+ unsigned long first_block, last_block;
unsigned long last_index;
u64 isize = i_size_read(inode);
u64 last_len = 0;
u64 skip = 0;
u64 defrag_end = 0;
u64 newer_off = range->start;
+ u64 start;
+ u64 page_cnt;
unsigned long i;
unsigned long ra_index = 0;
+ size_t pg_offset;
int ret;
int defrag_count = 0;
int compress_type = BTRFS_COMPRESS_ZLIB;
u32 extent_thresh = range->extent_thresh;
- unsigned long max_cluster = SZ_256K >> PAGE_SHIFT;
+ unsigned long max_cluster = SZ_256K >> inode->i_blkbits;
unsigned long cluster = max_cluster;
u64 new_align = ~((u64)SZ_128K - 1);
struct page **pages = NULL;
@@ -1316,8 +1364,14 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
ra = &file->f_ra;
}
- pages = kmalloc_array(max_cluster, sizeof(struct page *),
- GFP_NOFS);
+ /*
+ * In subpage-blocksize scenario the first of "max_cluster" blocks
+ * may start on a non-zero page offset. In such scenarios we need one
+ * page more than what would be needed in the case where the first block
+ * maps to first block of a page.
+ */
+ page_cnt = (max_cluster >> (PAGE_SHIFT - inode->i_blkbits)) + 1;
+ pages = kmalloc_array(page_cnt, sizeof(struct page *), GFP_NOFS);
if (!pages) {
ret = -ENOMEM;
goto out_ra;
@@ -1325,12 +1379,15 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
/* find the last page to defrag */
if (range->start + range->len > range->start) {
- last_index = min_t(u64, isize - 1,
- range->start + range->len - 1) >> PAGE_SHIFT;
+ last_off = min_t(u64, isize - 1, range->start + range->len - 1);
} else {
- last_index = (isize - 1) >> PAGE_SHIFT;
+ last_off = isize - 1;
}
+ last_off = round_up(last_off, root->sectorsize) - 1;
+ last_block = last_off >> inode->i_blkbits;
+ last_index = last_off >> PAGE_SHIFT;
+
if (newer_than) {
ret = find_new_extents(root, inode, newer_than,
&newer_off, SZ_64K);
@@ -1340,14 +1397,20 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
* we always align our defrag to help keep
* the extents in the file evenly spaced
*/
- i = (newer_off & new_align) >> PAGE_SHIFT;
+ first_off = newer_off & new_align;
} else
goto out_ra;
} else {
- i = range->start >> PAGE_SHIFT;
+ first_off = range->start;
}
+
+ first_off = round_down(first_off, root->sectorsize);
+ first_block = first_off >> inode->i_blkbits;
+ i = first_off >> PAGE_SHIFT;
+ pg_offset = first_off & (PAGE_SIZE - 1);
+
if (!max_to_defrag)
- max_to_defrag = last_index - i + 1;
+ max_to_defrag = last_block - first_block + 1;
/*
* make writeback starts from i, so the defrag range can be
@@ -1371,39 +1434,50 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
break;
}
- if (!should_defrag_range(inode, (u64)i << PAGE_SHIFT,
- extent_thresh, &last_len, &skip,
- &defrag_end, range->flags &
- BTRFS_DEFRAG_RANGE_COMPRESS)) {
+ start = pg_offset + ((u64)i << PAGE_SHIFT);
+ if (!should_defrag_range(inode, start,
+ extent_thresh, &last_len, &skip,
+ &defrag_end, range->flags &
+ BTRFS_DEFRAG_RANGE_COMPRESS)) {
unsigned long next;
/*
* the should_defrag function tells us how much to skip
* bump our counter by the suggested amount
*/
- next = DIV_ROUND_UP(skip, PAGE_SIZE);
- i = max(i + 1, next);
+ next = max(skip, start + root->sectorsize);
+ next >>= inode->i_blkbits;
+
+ first_off = next << inode->i_blkbits;
+ i = first_off >> PAGE_SHIFT;
+ pg_offset = first_off & (PAGE_SIZE - 1);
continue;
}
if (!newer_than) {
- cluster = (PAGE_ALIGN(defrag_end) >>
- PAGE_SHIFT) - i;
+ cluster = (defrag_end >> inode->i_blkbits)
+ - (start >> inode->i_blkbits);
+
cluster = min(cluster, max_cluster);
} else {
cluster = max_cluster;
}
- if (i + cluster > ra_index) {
+ page_cnt = pg_offset + (cluster << inode->i_blkbits) - 1;
+ page_cnt = DIV_ROUND_UP(page_cnt, PAGE_SIZE);
+ if (i + page_cnt > ra_index) {
ra_index = max(i, ra_index);
btrfs_force_ra(inode->i_mapping, ra, file, ra_index,
- cluster);
- ra_index += cluster;
+ page_cnt);
+ ra_index += DIV_ROUND_UP(pg_offset +
+ (cluster << inode->i_blkbits),
+ PAGE_SIZE);
}
inode_lock(inode);
if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)
BTRFS_I(inode)->force_compress = compress_type;
- ret = cluster_pages_for_defrag(inode, pages, i, cluster);
+ ret = cluster_pages_for_defrag(inode, pages, i, pg_offset,
+ cluster);
if (ret < 0) {
inode_unlock(inode);
goto out_ra;
@@ -1418,29 +1492,29 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
if (newer_off == (u64)-1)
break;
- if (ret > 0)
- i += ret;
-
newer_off = max(newer_off + 1,
- (u64)i << PAGE_SHIFT);
+ start + (ret << inode->i_blkbits));
ret = find_new_extents(root, inode, newer_than,
&newer_off, SZ_64K);
if (!ret) {
range->start = newer_off;
- i = (newer_off & new_align) >> PAGE_SHIFT;
+ first_off = newer_off & new_align;
} else {
break;
}
} else {
if (ret > 0) {
- i += ret;
- last_len += ret << PAGE_SHIFT;
+ first_off = start + (ret << inode->i_blkbits);
+ last_len += ret << inode->i_blkbits;
} else {
- i++;
+ first_off = start + root->sectorsize;
last_len = 0;
}
}
+ first_off = round_down(first_off, root->sectorsize);
+ i = first_off >> PAGE_SHIFT;
+ pg_offset = first_off & (PAGE_SIZE - 1);
}
if ((range->flags & BTRFS_DEFRAG_RANGE_START_IO)) {
--
2.5.5
next prev parent reply other threads:[~2016-10-02 13:26 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-02 13:24 [PATCH V21 00/19] Allow I/O on blocks whose size is less than page size Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 01/19] Btrfs: subpage-blocksize: extent_clear_unlock_delalloc: Prevent page from being unlocked more than once Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 02/19] Btrfs: subpage-blocksize: Make sure delalloc range intersects with the locked page's range Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 03/19] Btrfs: subpage-blocksize: Use PG_Uptodate flag to track block uptodate status Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 04/19] Btrfs: Remove extent_io_tree's track_uptodate member Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 05/19] Btrfs: subpage-blocksize: Fix whole page read Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 06/19] Btrfs: subpage-blocksize: Fix whole page write Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 07/19] Btrfs: subpage-blocksize: Use kmalloc()-ed memory to hold metadata blocks Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 08/19] Btrfs: subpage-blocksize: Execute sanity tests on all possible block sizes Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 09/19] Btrfs: subpage-blocksize: Compute free space tree BITMAP_RANGE based on sectorsize Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 10/19] Btrfs: subpage-blocksize: Allow mounting filesystems where sectorsize < PAGE_SIZE Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 11/19] Btrfs: subpage-blocksize: Deal with partial ordered extent allocations Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 12/19] Btrfs: subpage-blocksize: Explicitly track I/O status of blocks of an ordered extent Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 13/19] Btrfs: subpage-blocksize: btrfs_punch_hole: Fix uptodate blocks check Chandan Rajendra
2016-10-02 13:24 ` Chandan Rajendra [this message]
2016-10-02 13:24 ` [PATCH V21 15/19] Btrfs: subpage-blocksize: Enable dedupe ioctl Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 16/19] Btrfs: subpage-blocksize: btrfs_clone: Flush dirty blocks of a page that do not map the clone range Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 17/19] Btrfs: subpage-blocksize: Make file extent relocate code subpage blocksize aware Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 18/19] Btrfs: subpage-blocksize: __btrfs_lookup_bio_sums: Set offset when moving to a new bio_vec Chandan Rajendra
2016-10-02 13:24 ` [PATCH V21 19/19] Btrfs: subpage-blocksize: Disable compression Chandan Rajendra
2017-06-19 10:19 ` [PATCH V21 00/19] Allow I/O on blocks whose size is less than page size Chandan Rajendra
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=1475414668-25954-15-git-send-email-chandan@linux.vnet.ibm.com \
--to=chandan@linux.vnet.ibm.com \
--cc=clm@fb.com \
--cc=dsterba@suse.com \
--cc=jbacik@fb.com \
--cc=linux-btrfs@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.