Hi @namjaejeon,

Thank you for your quick reply. bd->sector_size is 4096. However, some parts of the code in mkfs.c only write 512 bytes because they rely on the size of some structures (e.g. struct pbr or struct exbs) whose size is 512 bytes.

I have written a patch and it fixes the issue. It requires more testing though.

Thanks again,

from exfatprogs.

Okay, Can you check if bd->sector_size is not 4096 ?

int exfat_get_blk_dev_info(struct exfat_user_input *ui,
		struct exfat_blk_dev *bd)
...
	if (ioctl(fd, BLKSSZGET, &bd->sector_size) < 0)
		bd->sector_size = DEFAULT_SECTOR_SIZE;

from exfatprogs.

Sound good!

from exfatprogs.

Fixed

This issue and other issues observed with 4Kn HDDs and reported by @sandeen on the Linux fsdevel mailing list are fixed by the following commits:

• dump: use BytesPerSectorShift of boot sector
• fsck: use BytesPerSectorShift of boot sector to verify boot region
• exfatprogs: libexfat: fix wrong bd->num_sectors calculation
• exfatprogs: tune: use sector size extracted from the boot sector
• mkfs: fix boot sector checksum when the sector size is 4 KB

These changes are included in exfatprogs 1.1.2.

from exfatprogs.