mke2fs: batch zeroing inode table

For flex_bg enabled fs, we could merge the
inode table blocks into a contiguous range,
this improves mke2fs time on large devices
when lazy_itable_init is disabled.

On a 977TB device, unpatched mke2fs was running
for 449m10s before getting terminated manually.
strace shows huge number of fallocate, given the
offset from fallocate it has done 41% of the inode
tables, the estimated time needed would be 1082m.

	unpatched 	patched
real	449m10.954s	4m20.531s
user	0m18.217s	0m16.147s
sys	0m20.311s	0m8.944s

Signed-off-by: Li Dongyang <dongyangli@ddn.com>
Link: https://lore.kernel.org/r/20230904045806.827621-1-dongyangli@ddn.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>

1 files changed, 31 insertions, 7 deletions

diff --git a/misc/mke2fs.c b/misc/mke2fs.c
index 4a9c1b092..aebf050f6 100644
--- a/misc/mke2fs.c
+++ b/misc/mke2fs.c
@@ -415,9 +415,9 @@ static errcode_t packed_allocate_tables(ext2_filsys fs)
 static void write_inode_tables(ext2_filsys fs, int lazy_flag, int itable_zeroed)
 {
 	errcode_t	retval;
-	blk64_t		blk;
+	blk64_t		start = 0;
 	dgrp_t		i;
-	int		num;
+	int		len = 0;
 	struct ext2fs_numeric_progress_struct progress;
 
 	ext2fs_numeric_progress_init(fs, &progress,
@@ -425,10 +425,10 @@ static void write_inode_tables(ext2_filsys fs, int lazy_flag, int itable_zeroed)
 				     fs->group_desc_count);
 
 	for (i = 0; i < fs->group_desc_count; i++) {
-		ext2fs_numeric_progress_update(fs, &progress, i);
+		blk64_t blk = ext2fs_inode_table_loc(fs, i);
+		int num = fs->inode_blocks_per_group;
 
-		blk = ext2fs_inode_table_loc(fs, i);
-		num = fs->inode_blocks_per_group;
+		ext2fs_numeric_progress_update(fs, &progress, i);
 
 		if (lazy_flag)
 			num = ext2fs_div_ceil((fs->super->s_inodes_per_group -
@@ -441,14 +441,26 @@ static void write_inode_tables(ext2_filsys fs, int lazy_flag, int itable_zeroed)
 			ext2fs_group_desc_csum_set(fs, i);
 		}
 		if (!itable_zeroed) {
-			retval = ext2fs_zero_blocks2(fs, blk, num, &blk, &num);
+			if (len == 0) {
+				start = blk;
+				len = num;
+				continue;
+			}
+			/* 'len' must not overflow 2^31 blocks for ext2fs_zero_blocks2() */
+			if (start + len == blk && len + num >= len) {
+				len += num;
+				continue;
+			}
+			retval = ext2fs_zero_blocks2(fs, start, len, &start, &len);
 			if (retval) {
 				fprintf(stderr, _("\nCould not write %d "
 					  "blocks in inode table starting at %llu: %s\n"),
-					num, (unsigned long long) blk,
+					len, (unsigned long long) start,
 					error_message(retval));
 				exit(1);
 			}
+			start = blk;
+			len = num;
 		}
 		if (sync_kludge) {
 			if (sync_kludge == 1)
@@ -457,6 +469,18 @@ static void write_inode_tables(ext2_filsys fs, int lazy_flag, int itable_zeroed)
 				io_channel_flush(fs->io);
 		}
 	}
+	if (len) {
+		retval = ext2fs_zero_blocks2(fs, start, len, &start, &len);
+		if (retval) {
+			fprintf(stderr, _("\nCould not write %d "
+				  "blocks in inode table starting at %llu: %s\n"),
+				len, (unsigned long long) start,
+				error_message(retval));
+			exit(1);
+		}
+		if (sync_kludge)
+			io_channel_flush(fs->io);
+	}
 	ext2fs_numeric_progress_close(fs, &progress,
 				      _("done                            \n"));