initrd/etc/functions:875

• disk_info_sysfs() builds disk_info lines using "\n" and then echoes with echo -n, which produces literal backslash-n sequences unless the consumer interprets escapes. To make output consistent across callers, prefer emitting real newlines via printf (e.g., printf 'Disk /dev/%s: %s GB\n' ...) or use printf '%b' when printing an escape-containing buffer.

				local size_gb=$(((size_bytes * sector_size + 500000000) / 1000000000))
				disk_info="${disk_info}Disk /dev/${devname}: ${size_gb} GB\n"
				DEBUG "disk_info_sysfs: /dev/${devname} calculated as ${size_gb} GB (${size_bytes} sectors * ${sector_size} bytes/sector = $((size_bytes * sector_size)) bytes)"
			fi
		fi
	done
	echo -n "$disk_info"

initrd/etc/functions:832

• list_block_devices() always exits 0 (even when it outputs nothing). Callers like config-gui.sh currently treat enumeration failure as an error (if ! list_block_devices ...; then ...), but that branch will never run now—an empty device list will instead reach file_selector, which exits with a different error path. Consider having list_block_devices return non-zero when no devices are found (or update callers to check for an empty output file explicitly).

# List all block devices using sysfs
# Outputs one device path per line (e.g., /dev/sda, /dev/vda, /dev/nvme0n1)
list_block_devices() {
	TRACE_FUNC
	for dev in /sys/block/sd* /sys/block/nvme* /sys/block/vd* /sys/block/hd*; do
		if [ -e "$dev" ]; then
			echo "/dev/$(basename "$dev")"
		fi
	done | sort
}

initrd/etc/functions:1347

• The comment says “otherwise check for grub type”, but the fdisk-based fallback was removed and the function now only checks PARTTYPENAME=BIOS boot and then returns 1. Please update the comment (or reintroduce an actual fallback) so it matches the current behavior.

	# Check partition type using sysfs if available, otherwise check for grub type
	# For GPT disks, check /sys/class/block/<dev>/<partition>/uevent for PARTTYPENAME
	local part_sys="/sys/class/block/$DEVICE/$(basename "$PART_DEV")"
	if [ -e "$part_sys/uevent" ]; then
		if grep -q "PARTTYPENAME=BIOS boot" "$part_sys/uevent"; then
			TRACE "$PART_DEV is a GPT BIOS grub partition"
			return 0
		fi
	fi

	# Fallback: try to detect using other sysfs attributes if available
	# For MBR disks, we would need to check partition type differently
	DEBUG "$PART_DEV - unable to confirm it's a bios-grub partition via sysfs"

initrd/etc/functions:1393

• The comment says p<num> which looks like an HTML-escaped fragment and is confusing in a shell script comment. Replace it with a plain-text example like p<num> or p1.

	# match nvme style (p<num>) or normal (digit suffix)

initrd/etc/functions:958

• Avoid logging any information about secrets, including password length. Even in debug logs this can leak sensitive metadata and is unnecessary for normal troubleshooting. Drop this log line or replace it with a generic message that doesn’t include derived properties of the password.

	read -r -s -p $'\nTPM Owner Password: ' tpm_owner_password
	echo
	# Cache the password externally to be reused by who needs it
	DEBUG "password length ${#tpm_owner_password}"

initrd/etc/functions:1484

• mounted_boot_dev is assigned without local, which can leak/overwrite a global variable in this large shared functions file and make later logic harder to reason about. Declare it as local mounted_boot_dev (and consider local CONFIG_BOOT_DEV behavior) to keep detect_boot_device side effects intentional.

	mounted_boot_dev="$(awk '$2=="/boot" {print $1; exit}' /proc/mounts)"
	if [ -n "$mounted_boot_dev" ] && ls -d /boot/grub* >/dev/null 2>&1; then
		CONFIG_BOOT_DEV="$mounted_boot_dev"
		return 0