-diff -u -r linux-stage.orig/fs/ext4/ext4.h linux-stage/fs/ext4/ext4.h
---- linux-stage.orig/fs/ext4/ext4.h 2012-12-31 11:12:36.000000000 -0500
-+++ linux-stage/fs/ext4/ext4.h 2012-12-31 11:12:48.000000000 -0500
-@@ -1170,11 +1170,14 @@
-
+Index: linux-stage/fs/ext4/ext4.h
+===================================================================
+--- linux-stage.orig/fs/ext4/ext4.h
++++ linux-stage/fs/ext4/ext4.h
+@@ -1173,11 +1173,14 @@ struct ext4_sb_info {
+
/* tunables */
unsigned long s_stripe;
- unsigned int s_mb_stream_request;
unsigned int s_mb_group_prealloc;
unsigned int s_max_writeback_mb_bump;
/* where last allocation was done - for stream allocation */
-diff -u -r linux-stage.orig/fs/ext4/inode.c linux-stage/fs/ext4/inode.c
---- linux-stage.orig/fs/ext4/inode.c 2012-12-31 11:12:36.000000000 -0500
-+++ linux-stage/fs/ext4/inode.c 2012-12-31 11:12:48.000000000 -0500
-@@ -2937,6 +2937,11 @@
+Index: linux-stage/fs/ext4/inode.c
+===================================================================
+--- linux-stage.orig/fs/ext4/inode.c
++++ linux-stage/fs/ext4/inode.c
+@@ -2948,6 +2948,11 @@ static int ext4_da_writepages(struct add
if (unlikely(sbi->s_mount_flags & EXT4_MF_FS_ABORTED))
return -EROFS;
-
+
+ if (wbc->nr_to_write < sbi->s_mb_small_req) {
+ nr_to_writebump = sbi->s_mb_small_req - wbc->nr_to_write;
+ wbc->nr_to_write = sbi->s_mb_small_req;
+
if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
range_whole = 1;
-
-diff -u -r linux-stage.orig/fs/ext4/mballoc.c linux-stage/fs/ext4/mballoc.c
---- linux-stage.orig/fs/ext4/mballoc.c 2012-12-31 11:12:36.000000000 -0500
-+++ linux-stage/fs/ext4/mballoc.c 2012-12-31 11:20:51.000000000 -0500
-@@ -1799,6 +1799,25 @@
+
+Index: linux-stage/fs/ext4/mballoc.c
+===================================================================
+--- linux-stage.orig/fs/ext4/mballoc.c
++++ linux-stage/fs/ext4/mballoc.c
+@@ -1802,6 +1802,26 @@ void ext4_mb_simple_scan_group(struct ex
}
}
-
-+static void ext4_mb_prealloc_table_add(struct ext4_sb_info *sbi, int value)
+
++static int ext4_mb_prealloc_table_add(struct ext4_sb_info *sbi, int value)
+{
+ int i;
+
+ if (value > (sbi->s_blocks_per_group - 1 - 1 - sbi->s_itb_per_group))
-+ return;
++ return -1;
+
+ for (i = 0; i < sbi->s_mb_prealloc_table_size; i++) {
+ if (sbi->s_mb_prealloc_table[i] == 0) {
+ sbi->s_mb_prealloc_table[i] = value;
-+ return;
++ return 0;
+ }
+
+ /* they should add values in order */
+ if (value <= sbi->s_mb_prealloc_table[i])
-+ return;
++ return -1;
+ }
++ return -1;
+}
+
/*
* The routine scans the group and measures all found extents.
* In order to optimize scanning, caller must pass number of
-@@ -2172,6 +2191,80 @@
+@@ -2179,6 +2199,82 @@ static const struct seq_operations ext4_
.show = ext4_mb_seq_groups_show,
};
-
+
+#define EXT4_MB_PREALLOC_TABLE "prealloc_table"
+
+static int ext4_mb_prealloc_table_proc_read(char *page, char **start, off_t off,
+ while (cur < end && i < num) {
+ while ((cur < end) && (*cur == ' ')) cur++;
+ value = simple_strtol(cur, &cur, 0);
-+ ext4_mb_prealloc_table_add(sbi, value);
-+ i++;
++ if (ext4_mb_prealloc_table_add(sbi, value) == 0)
++ i++;
+ }
++ if (i != num)
++ sbi->s_mb_prealloc_table_size = i;
+
+ return cnt;
+}
static int ext4_mb_seq_groups_open(struct inode *inode, struct file *file)
{
struct super_block *sb = PDE(inode)->data;
-@@ -2469,9 +2562,52 @@
+@@ -2425,7 +2521,7 @@ static int ext4_groupinfo_create_slab(si
+ int ext4_mb_init(struct super_block *sb, int needs_recovery)
+ {
+ struct ext4_sb_info *sbi = EXT4_SB(sb);
+- unsigned i, j;
++ unsigned i, j, k, l;
+ unsigned offset;
+ unsigned max;
+ int ret;
+@@ -2476,9 +2572,51 @@ int ext4_mb_init(struct super_block *sb,
sbi->s_mb_max_to_scan = MB_DEFAULT_MAX_TO_SCAN;
sbi->s_mb_min_to_scan = MB_DEFAULT_MIN_TO_SCAN;
sbi->s_mb_stats = MB_DEFAULT_STATS;
+ }
+ memset(sbi->s_mb_prealloc_table, 0, i);
+
-+ ext4_mb_prealloc_table_add(sbi, 4);
-+ ext4_mb_prealloc_table_add(sbi, 8);
-+ ext4_mb_prealloc_table_add(sbi, 16);
-+ ext4_mb_prealloc_table_add(sbi, 32);
-+ ext4_mb_prealloc_table_add(sbi, 64);
-+ ext4_mb_prealloc_table_add(sbi, 128);
-+ ext4_mb_prealloc_table_add(sbi, 256);
-+ ext4_mb_prealloc_table_add(sbi, 512);
-+ ext4_mb_prealloc_table_add(sbi, 1024);
-+ ext4_mb_prealloc_table_add(sbi, 2048);
++ for (k = 0, l = 4; k <= 9; ++k, l *= 2) {
++ if (ext4_mb_prealloc_table_add(sbi, l) < 0) {
++ sbi->s_mb_prealloc_table_size = k;
++ break;
++ }
++ }
+
+ sbi->s_mb_small_req = 256;
+ sbi->s_mb_large_req = 1024;
+ }
+ memset(sbi->s_mb_prealloc_table, 0, i);
+
-+ ext4_mb_prealloc_table_add(sbi, sbi->s_stripe);
-+ ext4_mb_prealloc_table_add(sbi, sbi->s_stripe * 2);
-+ ext4_mb_prealloc_table_add(sbi, sbi->s_stripe * 4);
++ for (k = 0, l = sbi->s_stripe; k <= 2; ++k, l *= 2) {
++ if (ext4_mb_prealloc_table_add(sbi, l) < 0) {
++ sbi->s_mb_prealloc_table_size = k;
++ break;
++ }
++ }
+
+ sbi->s_mb_small_req = sbi->s_stripe;
+ sbi->s_mb_large_req = sbi->s_stripe * 8;
+ sbi->s_mb_group_prealloc = sbi->s_stripe * 4;
+ }
-
+
sbi->s_locality_groups = alloc_percpu(struct ext4_locality_group);
if (sbi->s_locality_groups == NULL) {
-@@ -2487,12 +2623,22 @@
+@@ -2494,12 +2632,22 @@ int ext4_mb_init(struct super_block *sb,
spin_lock_init(&lg->lg_prealloc_lock);
}
-
+
- if (sbi->s_proc)
+ if (sbi->s_proc) {
+ struct proc_dir_entry *p;
+ p->write_proc = ext4_mb_prealloc_table_proc_write;
+ }
+ }
-
+
out:
if (ret) {
+ kfree(sbi->s_mb_prealloc_table);
kfree(sbi->s_mb_offsets);
kfree(sbi->s_mb_maxs);
}
-@@ -2528,8 +2674,10 @@
+@@ -2533,8 +2681,10 @@ int ext4_mb_release(struct super_block *
struct ext4_sb_info *sbi = EXT4_SB(sb);
struct kmem_cache *cachep = get_groupinfo_cache(sb->s_blocksize_bits);
-
+
- if (sbi->s_proc)
+ if (sbi->s_proc) {
remove_proc_entry("mb_groups", sbi->s_proc);
+ remove_proc_entry(EXT4_MB_PREALLOC_TABLE, sbi->s_proc);
+ }
-
+
if (sbi->s_group_info) {
for (i = 0; i < ngroups; i++) {
-@@ -2859,11 +3007,12 @@
+@@ -2870,11 +3020,12 @@ static noinline_for_stack void
ext4_mb_normalize_request(struct ext4_allocation_context *ac,
struct ext4_allocation_request *ar)
{
struct ext4_inode_info *ei = EXT4_I(ac->ac_inode);
+ struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb);
struct ext4_prealloc_space *pa;
-
+
/* do normalize only data requests, metadata requests
-@@ -2894,49 +3043,34 @@
+@@ -2905,49 +3056,34 @@ ext4_mb_normalize_request(struct ext4_al
if (size < i_size_read(ac->ac_inode))
size = i_size_read(ac->ac_inode);
orig_size = size;
+ size = (size + ac->ac_sb->s_blocksize - 1) >> bsbits;
-
+
- /* max size of free chunks */
- max = 2 << bsbits;
+ start = wind = 0;
-
+
-#define NRL_CHECK_SIZE(req, size, max, chunk_size) \
- (req <= (size) || max <= (chunk_size))
+ /* let's choose preallocation window depending on file size */
+ }
+ }
+ size = wind;
-
+
- /* first, try to predict filesize */
- /* XXX: should this table be tunable? */
- start_off = 0;
- size = size >> bsbits;
- start = start_off >> bsbits;
+ orig_size = size;
-
+
/* don't cover already allocated blocks in selected range */
if (ar->pleft && start <= ar->lleft) {
-@@ -3008,7 +3143,6 @@
+@@ -3020,7 +3156,6 @@ ext4_mb_normalize_request(struct ext4_al
}
BUG_ON(start + size <= ac->ac_o_ex.fe_logical &&
start > ac->ac_o_ex.fe_logical);
- BUG_ON(size <= 0 || size > EXT4_BLOCKS_PER_GROUP(ac->ac_sb));
-
+
/* now prepare goal request */
-
-@@ -3940,11 +4074,19 @@
-
+
+@@ -3956,11 +4091,19 @@ static void ext4_mb_group_or_file(struct
+
/* don't use group allocation for large files */
size = max(size, isize);
- if (size > sbi->s_mb_stream_request) {
ac->ac_flags |= EXT4_MB_STREAM_ALLOC;
return;
}
-
+
+ /*
+ * request is so large that we don't care about
+ * streaming - it overweights any possible seek
BUG_ON(ac->ac_lg != NULL);
/*
* locality group prealloc space are per cpu. The reason for having
-diff -u -r linux-stage.orig/fs/ext4/super.c linux-stage/fs/ext4/super.c
---- linux-stage.orig/fs/ext4/super.c 2012-12-31 11:12:36.000000000 -0500
-+++ linux-stage/fs/ext4/super.c 2012-12-31 11:12:48.000000000 -0500
-@@ -2531,7 +2531,8 @@
+Index: linux-stage/fs/ext4/super.c
+===================================================================
+--- linux-stage.orig/fs/ext4/super.c
++++ linux-stage/fs/ext4/super.c
+@@ -2595,7 +2595,8 @@ EXT4_RW_ATTR_SBI_UI(mb_stats, s_mb_stats
EXT4_RW_ATTR_SBI_UI(mb_max_to_scan, s_mb_max_to_scan);
EXT4_RW_ATTR_SBI_UI(mb_min_to_scan, s_mb_min_to_scan);
EXT4_RW_ATTR_SBI_UI(mb_order2_req, s_mb_order2_reqs);
+EXT4_RW_ATTR_SBI_UI(mb_large_req, s_mb_large_req);
EXT4_RW_ATTR_SBI_UI(mb_group_prealloc, s_mb_group_prealloc);
EXT4_RW_ATTR_SBI_UI(max_writeback_mb_bump, s_max_writeback_mb_bump);
-
-@@ -2548,7 +2549,8 @@
+
+@@ -2611,7 +2612,8 @@ static struct attribute *ext4_attrs[] =
ATTR_LIST(mb_max_to_scan),
ATTR_LIST(mb_min_to_scan),
ATTR_LIST(mb_order2_req),
git://git.whamcloud.com / fs / lustre-release.git / commitdiff