LU-14305 ldiskfs: add parameters for mb_c123_threshold

Add mount options for /sys/fs/ldiskfs/*/mb_c[123]_threshold values
so that they can be set persistently via mount options.

The /sys/fs/ldiskfs/*/mb_c[123]_threshold values are always shown
rounded down to the next lower percentage value due to integer
division, since internal values are stored as blocks for efficiency.

Round up the values shown to the next percent to match what was
used to originally set these parameters.

Fixes: 95f8ae567749 ("LU-12103 ldiskfs: don't search large block range if disk full")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ie36a6667f8bca7481aa8179ab5b97c85d449d619
Reviewed-on: https://review.whamcloud.com/41193
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

diff --git a/ldiskfs/kernel_patches/patches/rhel7.6/ext4-simple-blockalloc.patch b/ldiskfs/kernel_patches/patches/rhel7.6/ext4-simple-blockalloc.patch
index f7a11a6..a516663 100644 (file)
--- a/ldiskfs/kernel_patches/patches/rhel7.6/ext4-simple-blockalloc.patch
+++ b/ldiskfs/kernel_patches/patches/rhel7.6/ext4-simple-blockalloc.patch
@@ -158,21 +158,28 @@ Index: linux-stage/fs/ext4/mballoc.c
  #define EXT4_MB_PREALLOC_TABLE          "prealloc_table"
  
  static int ext4_mb_check_and_update_prealloc(struct ext4_sb_info *sbi,
-@@ -2730,6 +2851,7 @@ static int ext4_groupinfo_create_slab(si
+@@ -2730,6 +2851,8 @@ static int ext4_groupinfo_create_slab(si
        return 0;
  }
  
-+#define THRESHOLD_BLOCKS(ts) (ext4_blocks_count(sbi->s_es) / 100 * ts)
++#define THRESHOLD_BLOCKS(sbi, percent)                                        \
++      (ext4_blocks_count((sbi)->s_es) / 100 * (percent))
  int ext4_mb_init(struct super_block *sb)
  {
        struct ext4_sb_info *sbi = EXT4_SB(sb);
-@@ -2781,6 +2903,9 @@ int ext4_mb_init(struct super_block *sb)
+@@ -2781,6 +2903,15 @@ int ext4_mb_init(struct super_block *sb)
        sbi->s_mb_min_to_scan = MB_DEFAULT_MIN_TO_SCAN;
        sbi->s_mb_stats = MB_DEFAULT_STATS;
        sbi->s_mb_order2_reqs = MB_DEFAULT_ORDER2_REQS;
-+      sbi->s_mb_c1_blocks = THRESHOLD_BLOCKS(MB_DEFAULT_C1_THRESHOLD);
-+      sbi->s_mb_c2_blocks = THRESHOLD_BLOCKS(MB_DEFAULT_C2_THRESHOLD);
-+      sbi->s_mb_c3_blocks = THRESHOLD_BLOCKS(MB_DEFAULT_C3_THRESHOLD);
++      if (!sbi->s_mb_c1_blocks)
++              sbi->s_mb_c1_blocks =
++                      THRESHOLD_BLOCKS(sbi, MB_DEFAULT_C1_THRESHOLD);
++      if (!sbi->s_mb_c2_blocks)
++              sbi->s_mb_c2_blocks =
++                      THRESHOLD_BLOCKS(sbi, MB_DEFAULT_C2_THRESHOLD);
++      if (!sbi->s_mb_c3_blocks)
++              sbi->s_mb_c3_blocks =
++                      THRESHOLD_BLOCKS(sbi, MB_DEFAULT_C3_THRESHOLD);
        /*
         * The default group preallocation is 512, which for 4k block
         * sizes translates to 2 megabytes.  However for bigalloc file
@@ -234,19 +241,69 @@ Index: linux-stage/fs/ext4/ext4.h
        spinlock_t s_bal_lock;
        unsigned long s_mb_buddies_generated;
        unsigned long long s_mb_generation_time;
+@@ -2115,6 +2121,8 @@ struct ext4_sb_info {
+ extern void ext4_end_bitmap_read(struct buffer_head *bh, int uptodate);
+ 
+ /* mballoc.c */
++extern int save_threshold_percent(struct ext4_sb_info *sbi, const char *buf,
++                                ext4_fsblk_t *blocks);
+ extern long ext4_mb_stats;
+ extern long ext4_mb_max_to_scan;
+ extern int ext4_mb_init(struct super_block *);
 Index: linux-stage/fs/ext4/super.c
 ===================================================================
 --- linux-stage.orig/fs/ext4/super.c
 +++ linux-stage/fs/ext4/super.c
-@@ -2734,6 +2734,73 @@ static ssize_t sbi_deprecated_show(struc
+@@ -1208,6 +1208,7 @@ enum {
+       Opt_nomblk_io_submit, Opt_block_validity, Opt_noblock_validity,
+       Opt_inode_readahead_blks, Opt_journal_ioprio,
+       Opt_dioread_nolock, Opt_dioread_lock,
++      Opt_mb_c1_threshold, Opt_mb_c2_threshold, Opt_mb_c3_threshold,
+       Opt_no_mbcache,
+       Opt_discard, Opt_nodiscard, Opt_init_itable, Opt_noinit_itable,
+       Opt_max_dir_size_kb, Opt_nojournal_checksum,
+@@ -1287,6 +1288,9 @@ static const match_table_t tokens = {
+       {Opt_nodiscard, "nodiscard"},
+       {Opt_init_itable, "init_itable=%u"},
+       {Opt_no_mbcache, "no_mbcache"},
++      {Opt_mb_c1_threshold, "mb_c1_threshold=%s"},
++      {Opt_mb_c2_threshold, "mb_c2_threshold=%s"},
++      {Opt_mb_c3_threshold, "mb_c3_threshold=%s"},
+       {Opt_init_itable, "init_itable"},
+       {Opt_noinit_itable, "noinit_itable"},
+       {Opt_max_dir_size_kb, "max_dir_size_kb=%u"},
+@@ -1449,6 +1453,9 @@ static const struct mount_opts {
+       {Opt_auto_da_alloc, EXT4_MOUNT_NO_AUTO_DA_ALLOC, MOPT_CLEAR},
+       {Opt_noinit_itable, EXT4_MOUNT_INIT_INODE_TABLE, MOPT_CLEAR},
+       {Opt_no_mbcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET},
++      {Opt_mb_c1_threshold, 0, MOPT_STRING},
++      {Opt_mb_c2_threshold, 0, MOPT_STRING},
++      {Opt_mb_c3_threshold, 0, MOPT_STRING},
+       {Opt_commit, 0, MOPT_GTE0},
+       {Opt_max_batch_time, 0, MOPT_GTE0},
+       {Opt_min_batch_time, 0, MOPT_GTE0},
+@@ -1571,6 +1578,12 @@ static const struct mount_opts {
+               sbi->s_max_dir_size_kb = arg;
+               /* reset s_warning_dir_size and make it re-calculated */
+               sbi->s_warning_dir_size = 0;
++      } else if (token == Opt_mb_c1_threshold) {
++              save_threshold_percent(sbi, args[0].from, &sbi->s_mb_c1_blocks);
++      } else if (token == Opt_mb_c2_threshold) {
++              save_threshold_percent(sbi, args[0].from, &sbi->s_mb_c2_blocks);
++      } else if (token == Opt_mb_c3_threshold) {
++              save_threshold_percent(sbi, args[0].from, &sbi->s_mb_c3_blocks);
+       } else if (token == Opt_stripe) {
+               sbi->s_stripe = arg;
+       } else if (token == Opt_resuid) {
+@@ -2734,6 +2747,74 @@ static ssize_t sbi_deprecated_show(struc
        return snprintf(buf, PAGE_SIZE, "%d\n", a->u.deprecated_val);
  }
  
-+static int save_threshold(struct ext4_sb_info *sbi, const char *buf,
-+                       ext4_fsblk_t *blocks) {
++int save_threshold_percent(struct ext4_sb_info *sbi, const char *buf,
++                         ext4_fsblk_t *blocks) {
 +      unsigned long long val;
 +
-+      if (!parse_strtoull(buf, 100, &val)) {
++      if (!parse_strtoull(buf, 100, &val) && val <= 100) {
 +              *blocks = val * ext4_blocks_count(sbi->s_es) / 100;
 +              return 0;
 +      }
@@ -254,14 +311,15 @@ Index: linux-stage/fs/ext4/super.c
 +      return -EINVAL;
 +}
 +
-+#define THRESHOLD_PERCENT(ts) (ts * 100 / ext4_blocks_count(sbi->s_es))
++#define THRESHOLD_PERCENT(sbi, blocks)                                        \
++      (((blocks) - 1) * 100 / ext4_blocks_count((sbi)->s_es) + 1)
 +static ssize_t mb_c1_threshold_store(struct ext4_attr *a,
 +                                  struct ext4_sb_info *sbi,
 +                                  const char *buf, size_t count)
 +{
 +      int ret;
 +
-+      ret = save_threshold(sbi, buf, &sbi->s_mb_c1_blocks);
++      ret = save_threshold_percent(sbi, buf, &sbi->s_mb_c1_blocks);
 +
 +      return ret ? ret : count;
 +}
@@ -270,7 +328,7 @@ Index: linux-stage/fs/ext4/super.c
 +                                 struct ext4_sb_info *sbi, char *buf)
 +{
 +      return snprintf(buf, PAGE_SIZE, "%llu\n",
-+                      THRESHOLD_PERCENT(sbi->s_mb_c1_blocks));
++                      THRESHOLD_PERCENT(sbi, sbi->s_mb_c1_blocks));
 +}
 +
 +static ssize_t mb_c2_threshold_store(struct ext4_attr *a,
@@ -279,7 +337,7 @@ Index: linux-stage/fs/ext4/super.c
 +{
 +      int ret;
 +
-+      ret = save_threshold(sbi, buf, &sbi->s_mb_c2_blocks);
++      ret = save_threshold_percent(sbi, buf, &sbi->s_mb_c2_blocks);
 +      return ret ? ret : count;
 +}
 +
@@ -287,7 +345,7 @@ Index: linux-stage/fs/ext4/super.c
 +                                 struct ext4_sb_info *sbi, char *buf)
 +{
 +              return snprintf(buf, PAGE_SIZE, "%llu\n",
-+                              THRESHOLD_PERCENT(sbi->s_mb_c2_blocks));
++                              THRESHOLD_PERCENT(sbi, sbi->s_mb_c2_blocks));
 +}
 +
 +static ssize_t mb_c3_threshold_store(struct ext4_attr *a,
@@ -296,7 +354,7 @@ Index: linux-stage/fs/ext4/super.c
 +{
 +      int ret;
 +
-+      ret = save_threshold(sbi, buf, &sbi->s_mb_c3_blocks);
++      ret = save_threshold_percent(sbi, buf, &sbi->s_mb_c3_blocks);
 +
 +      return ret ? ret : count;
 +}
@@ -305,7 +363,7 @@ Index: linux-stage/fs/ext4/super.c
 +                                 struct ext4_sb_info *sbi, char *buf)
 +{
 +              return snprintf(buf, PAGE_SIZE, "%llu\n",
-+                              THRESHOLD_PERCENT(sbi->s_mb_c3_blocks));
++                              THRESHOLD_PERCENT(sbi, sbi->s_mb_c3_blocks));
 +}
 +
 +

diff --git a/ldiskfs/kernel_patches/patches/rhel8/ext4-simple-blockalloc.patch b/ldiskfs/kernel_patches/patches/rhel8/ext4-simple-blockalloc.patch
index f5ecabd..1d9b22a 100644 (file)
--- a/ldiskfs/kernel_patches/patches/rhel8/ext4-simple-blockalloc.patch
+++ b/ldiskfs/kernel_patches/patches/rhel8/ext4-simple-blockalloc.patch
@@ -22,11 +22,13 @@ index 3b9ec24..64dc5fd 100644
        spinlock_t s_bal_lock;
        unsigned long s_mb_buddies_generated;
        unsigned long long s_mb_generation_time;
-@@ -2563,6 +2569,7 @@ extern void ext4_end_bitmap_read(struct buffer_head *bh, int uptodate);
+@@ -2563,6 +2569,9 @@ extern void ext4_end_bitmap_read(struct buffer_head *bh, int uptodate);
  /* mballoc.c */
  extern const struct file_operations ext4_seq_prealloc_table_fops;
  extern const struct seq_operations ext4_mb_seq_groups_ops;
 +extern const struct file_operations ext4_mb_seq_alloc_fops;
++extern int save_threshold_percent(struct ext4_sb_info *sbi, const char *buf,
++                                ext4_fsblk_t *blocks);
  extern const struct file_operations ext4_seq_mb_last_group_fops;
  extern int ext4_mb_seq_last_start_seq_show(struct seq_file *m, void *v);
  extern long ext4_mb_stats;
@@ -189,21 +191,28 @@ index 15c962f..7870406 100644
  int ext4_mb_seq_last_start_seq_show(struct seq_file *m, void *v)
  {
        struct ext4_sb_info *sbi = EXT4_SB(m->private);
-@@ -2734,6 +2854,7 @@ static int ext4_groupinfo_create_slab(size_t size)
+@@ -2734,6 +2854,8 @@ static int ext4_groupinfo_create_slab(size_t size)
        return 0;
  }
  
-+#define THRESHOLD_BLOCKS(ts) (ext4_blocks_count(sbi->s_es) / 100 * ts)
++#define THRESHOLD_BLOCKS(sbi, percent)                                        \
++      (ext4_blocks_count((sbi)->s_es) / 100 * (percent))
  int ext4_mb_init(struct super_block *sb)
  {
        struct ext4_sb_info *sbi = EXT4_SB(sb);
-@@ -2787,6 +2908,9 @@ int ext4_mb_init(struct super_block *sb)
+@@ -2787,6 +2908,15 @@ int ext4_mb_init(struct super_block *sb)
        sbi->s_mb_min_to_scan = MB_DEFAULT_MIN_TO_SCAN;
        sbi->s_mb_stats = MB_DEFAULT_STATS;
        sbi->s_mb_order2_reqs = MB_DEFAULT_ORDER2_REQS;
-+      sbi->s_mb_c1_blocks = THRESHOLD_BLOCKS(MB_DEFAULT_C1_THRESHOLD);
-+      sbi->s_mb_c2_blocks = THRESHOLD_BLOCKS(MB_DEFAULT_C2_THRESHOLD);
-+      sbi->s_mb_c3_blocks = THRESHOLD_BLOCKS(MB_DEFAULT_C3_THRESHOLD);
++      if (!sbi->s_mb_c1_blocks)
++              sbi->s_mb_c1_blocks =
++                      THRESHOLD_BLOCKS(sbi, MB_DEFAULT_C1_THRESHOLD);
++      if (!sbi->s_mb_c2_blocks)
++              sbi->s_mb_c2_blocks =
++                      THRESHOLD_BLOCKS(sbi, MB_DEFAULT_C2_THRESHOLD);
++      if (!sbi->s_mb_c3_blocks)
++              sbi->s_mb_c3_blocks =
++                      THRESHOLD_BLOCKS(sbi, MB_DEFAULT_C3_THRESHOLD);
        /*
         * The default group preallocation is 512, which for 4k block
         * sizes translates to 2 megabytes.  However for bigalloc file
@@ -238,6 +247,51 @@ index e00c3b7..d02daaf 100644
  
  /*
   * default group prealloc size 512 blocks
+diff --git a/fs/ext4/super.c b/fs/ext4/super.c
+===================================================================
+--- linux-stage.orig/fs/ext4/super.c
++++ linux-stage/fs/ext4/super.c
+@@ -1450,6 +1450,7 @@ enum {
+       Opt_nomblk_io_submit, Opt_block_validity, Opt_noblock_validity,
+       Opt_inode_readahead_blks, Opt_journal_ioprio,
+       Opt_dioread_nolock, Opt_dioread_lock,
++      Opt_mb_c1_threshold, Opt_mb_c2_threshold, Opt_mb_c3_threshold,
+       Opt_discard, Opt_nodiscard, Opt_init_itable, Opt_noinit_itable,
+       Opt_max_dir_size_kb, Opt_nojournal_checksum, Opt_nombcache,
+ };
+@@ -1604,6 +1605,9 @@ static const match_table_t tokens = {
+       {Opt_init_itable, "init_itable"},
+       {Opt_noinit_itable, "noinit_itable"},
+       {Opt_max_dir_size_kb, "max_dir_size_kb=%u"},
++      {Opt_mb_c1_threshold, "mb_c1_threshold=%s"},
++      {Opt_mb_c2_threshold, "mb_c2_threshold=%s"},
++      {Opt_mb_c3_threshold, "mb_c3_threshold=%s"},
+       {Opt_test_dummy_encryption, "test_dummy_encryption"},
+       {Opt_nombcache, "nombcache"},
+       {Opt_nombcache, "no_mbcache"},  /* for backward compatibility */
+@@ -1748,6 +1752,9 @@ static const struct mount_opts {
+       {Opt_jqfmt_vfsv0, QFMT_VFS_V0, MOPT_QFMT},
+       {Opt_jqfmt_vfsv1, QFMT_VFS_V1, MOPT_QFMT},
+       {Opt_max_dir_size_kb, 0, MOPT_GTE0},
++      {Opt_mb_c1_threshold, 0, MOPT_STRING},
++      {Opt_mb_c2_threshold, 0, MOPT_STRING},
++      {Opt_mb_c3_threshold, 0, MOPT_STRING},
+       {Opt_test_dummy_encryption, 0, MOPT_GTE0},
+       {Opt_nombcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET},
+       {Opt_err, 0, 0}
+@@ -1874,6 +1881,12 @@ static const struct mount_opts {
+               sbi->s_max_dir_size_kb = arg;
+               /* reset s_warning_dir_size and make it re-calculated */
+               sbi->s_warning_dir_size = 0;
++      } else if (token == Opt_mb_c1_threshold) {
++              save_threshold_percent(sbi, args[0].from, &sbi->s_mb_c1_blocks);
++      } else if (token == Opt_mb_c2_threshold) {
++              save_threshold_percent(sbi, args[0].from, &sbi->s_mb_c2_blocks);
++      } else if (token == Opt_mb_c3_threshold) {
++              save_threshold_percent(sbi, args[0].from, &sbi->s_mb_c3_blocks);
+       } else if (token == Opt_stripe) {
+               sbi->s_stripe = arg;
+       } else if (token == Opt_resuid) {
 diff --git a/fs/ext4/sysfs.c b/fs/ext4/sysfs.c
 index 417b33a..f49821e 100644
 --- a/fs/ext4/sysfs.c
@@ -256,10 +310,8 @@ index 417b33a..f49821e 100644
                        task_pid_vnr(sbi->s_journal->j_task));
  }
  
-+#define THRESHOLD_PERCENT(ts) (ts * 100 / ext4_blocks_count(sbi->s_es))
-+
-+static int save_threshold_percent(struct ext4_sb_info *sbi, const char *buf,
-+                                ext4_fsblk_t *blocks)
++int save_threshold_percent(struct ext4_sb_info *sbi, const char *buf,
++                         ext4_fsblk_t *blocks)
 +{
 +      unsigned long long val;
 +
@@ -273,6 +325,8 @@ index 417b33a..f49821e 100644
 +      return 0;
 +}
 +
++#define THRESHOLD_PERCENT(sbi, blocks)                                        \
++      (((blocks) - 1) * 100 / ext4_blocks_count((sbi)->s_es) + 1)
 +static ssize_t mb_threshold_store(struct ext4_sb_info *sbi,
 +                                const char *buf, size_t count,
 +                                ext4_fsblk_t *blocks)
@@ -311,13 +365,13 @@ index 417b33a..f49821e 100644
                       percpu_counter_sum(&sbi->s_dirtyclusters_counter)));
 +      case attr_mb_c1_threshold:
 +              return scnprintf(buf, PAGE_SIZE, "%llu\n",
-+                               THRESHOLD_PERCENT(sbi->s_mb_c1_blocks));
++                               THRESHOLD_PERCENT(sbi, sbi->s_mb_c1_blocks));
 +      case attr_mb_c2_threshold:
 +              return scnprintf(buf, PAGE_SIZE, "%llu\n",
-+                               THRESHOLD_PERCENT(sbi->s_mb_c2_blocks));
++                               THRESHOLD_PERCENT(sbi, sbi->s_mb_c2_blocks));
 +      case attr_mb_c3_threshold:
 +              return scnprintf(buf, PAGE_SIZE, "%llu\n",
-+                               THRESHOLD_PERCENT(sbi->s_mb_c3_blocks));
++                               THRESHOLD_PERCENT(sbi, sbi->s_mb_c3_blocks));
        case attr_session_write_kbytes:
                return session_write_kbytes_show(sbi, buf);
        case attr_lifetime_write_kbytes: