[Patch 6/7] 2.4.20-pre4/ext3: jbd commit interval tuning

[Patch 6/7] 2.4.20-pre4/ext3: jbd commit interval tuning

[Stephen Tweedie]

Allow an arbitrary commit interval to be set when mounting or remounting
a filesystem.

Note that if this is greater than the system bdflush interval, then the
regular sync()s will beat the commit timer and you won't get longer
commit timeouts.

--- linux-2.4-ext3merge/fs/ext3/super.c.=K0005=.orig	Thu Sep 26 12:25:37 2002
+++ linux-2.4-ext3merge/fs/ext3/super.c	Thu Sep 26 12:25:37 2002
@@ -646,6 +646,11 @@
 				*mount_options &= ~EXT3_MOUNT_DATA_FLAGS;
 				*mount_options |= data_opt;
 			}
+		} else if (!strcmp (this_char, "commit")) {
+			unsigned long v;
+			if (want_numeric(value, "commit", &v))
+				return 0;
+			sbi->s_commit_interval = (HZ * v);
 		} else {
 			printk (KERN_ERR 
 				"EXT3-fs: Unrecognized mount option %s\n",
@@ -1228,6 +1233,22 @@
 	return NULL;
 }
 
+/*
+ * Setup any per-fs journal parameters now.  We'll do this both on
+ * initial mount, once the journal has been initialised but before we've
+ * done any recovery; and again on any subsequent remount. 
+ */
+static void ext3_init_journal_params(struct ext3_sb_info *sbi, 
+				     journal_t *journal)
+{
+	if (sbi->s_commit_interval)
+		journal->j_commit_interval = sbi->s_commit_interval;
+	/* We could also set up an ext3-specific default for the commit
+	 * interval here, but for now we'll just fall back to the jbd
+	 * default. */
+}
+
+
 static journal_t *ext3_get_journal(struct super_block *sb, int journal_inum)
 {
 	struct inode *journal_inode;
@@ -1262,7 +1283,7 @@
 		printk(KERN_ERR "EXT3-fs: Could not load journal inode\n");
 		iput(journal_inode);
 	}
-	
+	ext3_init_journal_params(EXT3_SB(sb), journal);
 	return journal;
 }
 
@@ -1340,6 +1361,7 @@
 		goto out_journal;
 	}
 	EXT3_SB(sb)->journal_bdev = bdev;
+	ext3_init_journal_params(EXT3_SB(sb), journal);
 	return journal;
 out_journal:
 	journal_destroy(journal);
@@ -1637,6 +1659,8 @@
 
 	es = sbi->s_es;
 
+	ext3_init_journal_params(sbi, sbi->s_journal);
+	
 	if ((*flags & MS_RDONLY) != (sb->s_flags & MS_RDONLY)) {
 		if (sbi->s_mount_opt & EXT3_MOUNT_ABORT)
 			return -EROFS;

Re: [Patch 6/7] 2.4.20-pre4/ext3: jbd commit interval tuning

[Marcelo Tosatti]

On Thu, 26 Sep 2002, Stephen Tweedie wrote:

> Allow an arbitrary commit interval to be set when mounting or remounting
> a filesystem.
>
> Note that if this is greater than the system bdflush interval, then the
> regular sync()s will beat the commit timer and you won't get longer
> commit timeouts.

Applied. For 2.5 we probably want all fs'es to have the same string to set
the commit interval.