diff --git a/src/disc/mod.rs b/src/disc/mod.rs
index bca3d8f..da04c79 100644
--- a/src/disc/mod.rs
+++ b/src/disc/mod.rs
@@ -2435,7 +2435,7 @@ impl Disc {
     pub fn aacs_disc_hash(&self) -> String {
         self.aacs
             .as_ref()
-            .map(|a| a.disc_hash.trim_start_matches("0x").to_string())
+            .map(|a| crate::hex::strip_hex_prefix(&a.disc_hash).to_string())
             .unwrap_or_default()
     }
 
diff --git a/src/hex.rs b/src/hex.rs
index a94e49b..9b2859a 100644
--- a/src/hex.rs
+++ b/src/hex.rs
@@ -16,7 +16,7 @@
 /// (case-insensitive), then requires an even run of ASCII hex digits. Any
 /// non-hex byte, or an odd length, yields `None`.
 pub fn parse_hex_bytes(s: &str) -> Option<Vec<u8>> {
-    let body = strip_prefix(s.trim());
+    let body = strip_hex_prefix(s.trim());
     let bytes = body.as_bytes();
     // Empty → empty Vec (a legitimately-empty variable-length field); odd length
     // is malformed. (`parse_hex_fixed` enforces a concrete length separately.)
@@ -34,7 +34,7 @@ pub fn parse_hex_bytes(s: &str) -> Option<Vec<u8>> {
 /// prefix; requires EXACTLY `2*N` ASCII hex digits after it. `None` on any
 /// non-hex byte or a length mismatch.
 pub fn parse_hex_fixed<const N: usize>(s: &str) -> Option<[u8; N]> {
-    let body = strip_prefix(s.trim());
+    let body = strip_hex_prefix(s.trim());
     let bytes = body.as_bytes();
     if bytes.len() != 2 * N {
         return None;
@@ -46,8 +46,33 @@ pub fn parse_hex_fixed<const N: usize>(s: &str) -> Option<[u8; N]> {
     Some(out)
 }
 
-/// Strip a single leading `0x` / `0X` if present (case-insensitive).
-fn strip_prefix(s: &str) -> &str {
+/// Parse a hex string into a `u16`. Accepts an optional `0x`/`0X` prefix
+/// (case-insensitive) via the same [`strip_hex_prefix`] the byte parsers use.
+/// `None` on any non-hex content or overflow.
+///
+/// Exists so callers never hand-roll `from_str_radix(s.trim_start_matches("0x"), 16)`
+/// — a **case-sensitive** strip that silently dropped an uppercase-`0X` value.
+/// (That reintroduced-in-keydb bug is exactly what this module was built to kill;
+/// the integer fields now share the one prefix rule.)
+pub fn parse_hex_u16(s: &str) -> Option<u16> {
+    u16::from_str_radix(strip_hex_prefix(s.trim()), 16).ok()
+}
+
+/// Parse a hex string into a `u32`. See [`parse_hex_u16`].
+pub fn parse_hex_u32(s: &str) -> Option<u32> {
+    u32::from_str_radix(strip_hex_prefix(s.trim()), 16).ok()
+}
+
+/// Parse a hex string into a `u8`. See [`parse_hex_u16`].
+pub fn parse_hex_u8(s: &str) -> Option<u8> {
+    u8::from_str_radix(strip_hex_prefix(s.trim()), 16).ok()
+}
+
+/// Strip a single leading `0x` / `0X` if present (case-insensitive). Public so
+/// callers that only need the prefix rule (e.g. normalizing a disc hash) reuse
+/// the one definition instead of hand-rolling a case-sensitive
+/// `trim_start_matches("0x")`.
+pub fn strip_hex_prefix(s: &str) -> &str {
     s.strip_prefix("0x")
         .or_else(|| s.strip_prefix("0X"))
         .unwrap_or(s)
@@ -95,6 +120,23 @@ mod tests {
         assert_eq!(parse_hex_fixed::<16>(&s), None);
     }
 
+    #[test]
+    fn hex_ints_accept_both_prefix_cases_and_bare() {
+        // The regression the keydb device-key bug hit: uppercase `0X` must parse
+        // identically to `0x` and to a bare value.
+        assert_eq!(parse_hex_u16("0x0001"), Some(1));
+        assert_eq!(parse_hex_u16("0X0001"), Some(1));
+        assert_eq!(parse_hex_u16("0001"), Some(1));
+        assert_eq!(parse_hex_u16(" 0XABCD "), Some(0xABCD));
+        assert_eq!(parse_hex_u32("0X00000002"), Some(2));
+        assert_eq!(parse_hex_u32("deadbeef"), Some(0xDEAD_BEEF));
+        assert_eq!(parse_hex_u8("0X03"), Some(3));
+        assert_eq!(parse_hex_u8("ff"), Some(0xFF));
+        // Overflow / non-hex → None.
+        assert_eq!(parse_hex_u8("0x1FF"), None);
+        assert_eq!(parse_hex_u16("0xzz"), None);
+    }
+
     #[test]
     fn bytes_variable_length_and_odd_rejected() {
         assert_eq!(parse_hex_bytes("0xAABBCC"), Some(vec![0xAA, 0xBB, 0xCC]));