[PATCH] reiserfs: on-demand bitmap loading
This is the patch the three previous ones have been leading up to.
It changes the behavior of ReiserFS from loading and caching all the bitmaps
as special, to treating the bitmaps like any other bit of metadata and just
letting the system-wide caches figure out what to hang on to.
Buffer heads are allocated on the fly, so there is no need to retain pointers
to all of them. The caching of the metadata occurs when the data is read and
updated, and is considered invalid and uncached until then.
I needed to remove the vs-4040 check for performing a duplicate operation on a
particular bit. The reason is that while the other sites for working with
bitmaps are allowed to schedule, is_reusable() is called from do_balance(),
which will panic if a schedule occurs in certain places.
The benefit of on-demand bitmaps clearly outweighs a sanity check that depends
on a compile-time option that is discouraged.
[akpm@osdl.org: warning fix]
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Cc: <reiserfs-dev@namesys.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/fs/reiserfs/bitmap.c b/fs/reiserfs/bitmap.c
index abdd6d9..cca1dbf 100644
--- a/fs/reiserfs/bitmap.c
+++ b/fs/reiserfs/bitmap.c
@@ -60,7 +60,6 @@
int is_reusable(struct super_block *s, b_blocknr_t block, int bit_value)
{
int bmap, offset;
- struct buffer_head *bh;
if (block == 0 || block >= SB_BLOCK_COUNT(s)) {
reiserfs_warning(s,
@@ -98,22 +97,6 @@
return 0;
}
- bh = SB_AP_BITMAP(s)[bmap].bh;
- get_bh(bh);
-
- if ((bit_value == 0 && reiserfs_test_le_bit(offset, bh->b_data)) ||
- (bit_value == 1 && reiserfs_test_le_bit(offset, bh->b_data) == 0)) {
- reiserfs_warning(s,
- "vs-4040: is_reusable: corresponding bit of block %lu does not "
- "match required value (bmap==%d, offset==%d) test_bit==%d",
- block, bmap, offset,
- reiserfs_test_le_bit(offset, bh->b_data));
-
- brelse(bh);
- return 0;
- }
- brelse(bh);
-
if (bit_value == 0 && block == SB_ROOT_BLOCK(s)) {
reiserfs_warning(s,
"vs-4050: is_reusable: this is root block (%u), "
@@ -173,13 +156,10 @@
bmap_n);
return 0;
}
- bh = bi->bh;
- get_bh(bh);
- if (buffer_locked(bh)) {
- PROC_INFO_INC(s, scan_bitmap.wait);
- __wait_on_buffer(bh);
- }
+ bh = reiserfs_read_bitmap_block(s, bmap_n);
+ if (bh == NULL)
+ return 0;
while (1) {
cont:
@@ -285,9 +265,20 @@
*/
static inline int block_group_used(struct super_block *s, u32 id)
{
- int bm;
- bm = bmap_hash_id(s, id);
- if (SB_AP_BITMAP(s)[bm].free_count > ((s->s_blocksize << 3) * 60 / 100)) {
+ int bm = bmap_hash_id(s, id);
+ struct reiserfs_bitmap_info *info = &SB_AP_BITMAP(s)[bm];
+
+ /* If we don't have cached information on this bitmap block, we're
+ * going to have to load it later anyway. Loading it here allows us
+ * to make a better decision. This favors long-term performace gain
+ * with a better on-disk layout vs. a short term gain of skipping the
+ * read and potentially having a bad placement. */
+ if (info->first_zero_hint == 0) {
+ struct buffer_head *bh = reiserfs_read_bitmap_block(s, bm);
+ brelse(bh);
+ }
+
+ if (info->free_count > ((s->s_blocksize << 3) * 60 / 100)) {
return 0;
}
return 1;
@@ -413,8 +404,9 @@
return;
}
- bmbh = apbi[nr].bh;
- get_bh(bmbh);
+ bmbh = reiserfs_read_bitmap_block(s, nr);
+ if (!bmbh)
+ return;
reiserfs_prepare_for_journal(s, bmbh, 1);
@@ -1320,6 +1312,7 @@
unsigned int bitmap)
{
b_blocknr_t block = (sb->s_blocksize << 3) * bitmap;
+ struct reiserfs_bitmap_info *info = SB_AP_BITMAP(sb) + bitmap;
struct buffer_head *bh;
/* Way old format filesystems had the bitmaps packed up front.
@@ -1330,9 +1323,21 @@
else if (bitmap == 0)
block = (REISERFS_DISK_OFFSET_IN_BYTES >> sb->s_blocksize_bits) + 1;
- bh = sb_getblk(sb, block);
- if (!buffer_uptodate(bh))
- ll_rw_block(READ, 1, &bh);
+ bh = sb_bread(sb, block);
+ if (bh == NULL)
+ reiserfs_warning(sb, "sh-2029: %s: bitmap block (#%lu) "
+ "reading failed", __FUNCTION__, bh->b_blocknr);
+ else {
+ if (buffer_locked(bh)) {
+ PROC_INFO_INC(sb, scan_bitmap.wait);
+ __wait_on_buffer(bh);
+ }
+ BUG_ON(!buffer_uptodate(bh));
+ BUG_ON(atomic_read(&bh->b_count) == 0);
+
+ if (info->first_zero_hint == 0)
+ reiserfs_cache_bitmap_metadata(sb, bh, info);
+ }
return bh;
}
@@ -1340,7 +1345,6 @@
int reiserfs_init_bitmap_cache(struct super_block *sb)
{
struct reiserfs_bitmap_info *bitmap;
- int i;
bitmap = vmalloc(sizeof (*bitmap) * SB_BMAP_NR(sb));
if (bitmap == NULL)
@@ -1348,28 +1352,15 @@
memset(bitmap, 0, sizeof (*bitmap) * SB_BMAP_NR(sb));
- for (i = 0; i < SB_BMAP_NR(sb); i++)
- bitmap[i].bh = reiserfs_read_bitmap_block(sb, i);
-
- /* make sure we have them all */
- for (i = 0; i < SB_BMAP_NR(sb); i++) {
- wait_on_buffer(bitmap[i].bh);
- if (!buffer_uptodate(bitmap[i].bh)) {
- reiserfs_warning(sb, "sh-2029: %s: "
- "bitmap block (#%lu) reading failed",
- __FUNCTION__, bitmap[i].bh->b_blocknr);
- for (i = 0; i < SB_BMAP_NR(sb); i++)
- brelse(bitmap[i].bh);
- vfree(bitmap);
- return -EIO;
- }
- }
-
- /* Cache the info on the bitmaps before we get rolling */
- for (i = 0; i < SB_BMAP_NR(sb); i++)
- reiserfs_cache_bitmap_metadata(sb, bitmap[i].bh, &bitmap[i]);
-
SB_AP_BITMAP(sb) = bitmap;
return 0;
}
+
+void reiserfs_free_bitmap_cache(struct super_block *sb)
+{
+ if (SB_AP_BITMAP(sb)) {
+ vfree(SB_AP_BITMAP(sb));
+ SB_AP_BITMAP(sb) = NULL;
+ }
+}
diff --git a/fs/reiserfs/resize.c b/fs/reiserfs/resize.c
index 90d39fd..3156847 100644
--- a/fs/reiserfs/resize.c
+++ b/fs/reiserfs/resize.c
@@ -128,8 +128,9 @@
* transaction begins, and the new bitmaps don't matter if the
* transaction fails. */
for (i = bmap_nr; i < bmap_nr_new; i++) {
- bh = sb_getblk(s, i * s->s_blocksize * 8);
- get_bh(bh);
+ /* don't use read_bitmap_block since it will cache
+ * the uninitialized bitmap */
+ bh = sb_bread(s, i * s->s_blocksize * 8);
memset(bh->b_data, 0, sb_blocksize(sb));
reiserfs_test_and_set_le_bit(0, bh->b_data);
reiserfs_cache_bitmap_metadata(s, bh, bitmap + i);
@@ -140,7 +141,6 @@
// update bitmap_info stuff
bitmap[i].first_zero_hint = 1;
bitmap[i].free_count = sb_blocksize(sb) * 8 - 1;
- bitmap[i].bh = bh;
brelse(bh);
}
/* free old bitmap blocks array */
@@ -157,8 +157,13 @@
/* Extend old last bitmap block - new blocks have been made available */
info = SB_AP_BITMAP(s) + bmap_nr - 1;
- bh = info->bh;
- get_bh(bh);
+ bh = reiserfs_read_bitmap_block(s, bmap_nr - 1);
+ if (!bh) {
+ int jerr = journal_end(&th, s, 10);
+ if (jerr)
+ return jerr;
+ return -EIO;
+ }
reiserfs_prepare_for_journal(s, bh, 1);
for (i = block_r; i < s->s_blocksize * 8; i++)
@@ -172,8 +177,13 @@
/* Correct new last bitmap block - It may not be full */
info = SB_AP_BITMAP(s) + bmap_nr_new - 1;
- bh = info->bh;
- get_bh(bh);
+ bh = reiserfs_read_bitmap_block(s, bmap_nr_new - 1);
+ if (!bh) {
+ int jerr = journal_end(&th, s, 10);
+ if (jerr)
+ return jerr;
+ return -EIO;
+ }
reiserfs_prepare_for_journal(s, bh, 1);
for (i = block_r_new; i < s->s_blocksize * 8; i++)
diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c
index c78e99e..c89aa23 100644
--- a/fs/reiserfs/super.c
+++ b/fs/reiserfs/super.c
@@ -432,7 +432,6 @@
static void reiserfs_put_super(struct super_block *s)
{
- int i;
struct reiserfs_transaction_handle th;
th.t_trans_id = 0;
@@ -462,10 +461,7 @@
*/
journal_release(&th, s);
- for (i = 0; i < SB_BMAP_NR(s); i++)
- brelse(SB_AP_BITMAP(s)[i].bh);
-
- vfree(SB_AP_BITMAP(s));
+ reiserfs_free_bitmap_cache(s);
brelse(SB_BUFFER_WITH_SB(s));
@@ -1344,7 +1340,6 @@
/* after journal replay, reread all bitmap and super blocks */
static int reread_meta_blocks(struct super_block *s)
{
- int i;
ll_rw_block(READ, 1, &(SB_BUFFER_WITH_SB(s)));
wait_on_buffer(SB_BUFFER_WITH_SB(s));
if (!buffer_uptodate(SB_BUFFER_WITH_SB(s))) {
@@ -1353,20 +1348,7 @@
return 1;
}
- for (i = 0; i < SB_BMAP_NR(s); i++) {
- ll_rw_block(READ, 1, &(SB_AP_BITMAP(s)[i].bh));
- wait_on_buffer(SB_AP_BITMAP(s)[i].bh);
- if (!buffer_uptodate(SB_AP_BITMAP(s)[i].bh)) {
- reiserfs_warning(s,
- "reread_meta_blocks, error reading bitmap block number %d at %llu",
- i,
- (unsigned long long)SB_AP_BITMAP(s)[i].
- bh->b_blocknr);
- return 1;
- }
- }
return 0;
-
}
/////////////////////////////////////////////////////
@@ -1547,7 +1529,6 @@
static int reiserfs_fill_super(struct super_block *s, void *data, int silent)
{
struct inode *root_inode;
- int j;
struct reiserfs_transaction_handle th;
int old_format = 0;
unsigned long blocks;
@@ -1793,19 +1774,17 @@
if (jinit_done) { /* kill the commit thread, free journal ram */
journal_release_error(NULL, s);
}
- if (SB_DISK_SUPER_BLOCK(s)) {
- for (j = 0; j < SB_BMAP_NR(s); j++) {
- if (SB_AP_BITMAP(s))
- brelse(SB_AP_BITMAP(s)[j].bh);
- }
- vfree(SB_AP_BITMAP(s));
- }
+
+ reiserfs_free_bitmap_cache(s);
if (SB_BUFFER_WITH_SB(s))
brelse(SB_BUFFER_WITH_SB(s));
#ifdef CONFIG_QUOTA
- for (j = 0; j < MAXQUOTAS; j++) {
- kfree(sbi->s_qf_names[j]);
- sbi->s_qf_names[j] = NULL;
+ {
+ int j;
+ for (j = 0; j < MAXQUOTAS; j++) {
+ kfree(sbi->s_qf_names[j]);
+ sbi->s_qf_names[j] = NULL;
+ }
}
#endif
kfree(sbi);
diff --git a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h
index 4f21ad3..73e0bec 100644
--- a/include/linux/reiserfs_fs_sb.h
+++ b/include/linux/reiserfs_fs_sb.h
@@ -267,7 +267,6 @@
// FIXME: Won't work with block sizes > 8K
__u16 first_zero_hint;
__u16 free_count;
- struct buffer_head *bh; /* the actual bitmap */
};
struct proc_dir_entry;