diff --git a/core/src/apps/bitcoin/common.py b/core/src/apps/bitcoin/common.py
index e1c88b536..9fceb2dff 100644
--- a/core/src/apps/bitcoin/common.py
+++ b/core/src/apps/bitcoin/common.py
@@ -5,7 +5,8 @@ from trezor.crypto import bech32, bip32, der
 from trezor.crypto.curve import secp256k1
 from trezor.utils import ensure
 
-from apps.common.coininfo import CoinInfo
+if False:
+    from apps.common.coininfo import CoinInfo
 
 # supported witness version for bech32 addresses
 _BECH32_WITVER = const(0x00)
diff --git a/core/src/trezor/crypto/der.py b/core/src/trezor/crypto/der.py
index ac99aa3e7..06f6d2c06 100644
--- a/core/src/trezor/crypto/der.py
+++ b/core/src/trezor/crypto/der.py
@@ -1,3 +1,9 @@
+from apps.common.readers import BytearrayReader
+
+if False:
+    from typing import List
+
+
 def encode_length(l: int) -> bytes:
     if l < 0x80:
         return bytes([l])
@@ -9,15 +15,80 @@ def encode_length(l: int) -> bytes:
         raise ValueError
 
 
+def decode_length(r: BytearrayReader) -> int:
+    init = r.get()
+    if init < 0x80:
+        # short form encodes length in initial octet
+        return init
+
+    if init == 0x80 or init == 0xFF or r.peek() == 0x00:
+        raise ValueError  # indefinite length, RFU or not shortest possible
+
+    # long form
+    n = 0
+    for _ in range(init & 0x7F):
+        n = n * 0x100 + r.get()
+
+    if n < 128:
+        raise ValueError  # encoding is not the shortest possible
+
+    return n
+
+
 def encode_int(i: bytes) -> bytes:
     i = i.lstrip(b"\x00")
+    if not i:
+        i = b"\00"
+
     if i[0] >= 0x80:
         i = b"\x00" + i
     return b"\x02" + encode_length(len(i)) + i
 
 
+def decode_int(r: BytearrayReader) -> bytes:
+    if r.get() != 0x02:
+        raise ValueError
+
+    n = decode_length(r)
+    if n == 0:
+        raise ValueError
+
+    if r.peek() & 0x80:
+        raise ValueError  # negative integer
+
+    if r.peek() == 0x00 and n > 1:
+        r.get()
+        n -= 1
+        if r.peek() & 0x80 == 0x00:
+            raise ValueError  # excessive zero-padding
+
+        if r.peek() == 0x00:
+            raise ValueError  # excessive zero-padding
+
+    return r.read(n)
+
+
 def encode_seq(seq: tuple) -> bytes:
     res = b""
     for i in seq:
         res += encode_int(i)
     return b"\x30" + encode_length(len(res)) + res
+
+
+def decode_seq(data: bytes) -> List[bytes]:
+    r = BytearrayReader(data)
+
+    if r.get() != 0x30:
+        raise ValueError
+    n = decode_length(r)
+
+    seq = []
+    end = r.offset + n
+    while r.offset < end:
+        i = decode_int(r)
+        seq.append(i)
+
+    if r.offset != end or r.remaining_count():
+        raise ValueError
+
+    return seq
diff --git a/core/tests/test_trezor.crypto.der.py b/core/tests/test_trezor.crypto.der.py
index 6adf42d06..d935b85b9 100644
--- a/core/tests/test_trezor.crypto.der.py
+++ b/core/tests/test_trezor.crypto.der.py
@@ -46,11 +46,26 @@ class TestCryptoDer(unittest.TestCase):
     def test_der_encode_seq(self):
 
         for s, d in self.vectors_seq:
-            s = (unhexlify(i) for i in s)
+            s = tuple(unhexlify(i) for i in s)
             d = unhexlify(d)
             d2 = der.encode_seq(s)
-            self.assertEqual(d, d2)
-
+            self.assertEqual(d2, d)
+            s = [i.lstrip(b"\x00") for i in s]
+            s2 = der.decode_seq(d)
+            self.assertEqual(s2, s)
+
+    def test_der_encode_decode_long_seq(self):
+        for length in (1, 127, 128, 129, 255, 256, 257):
+            raw_int = bytes((i & 0xfe) + 1 for i in range(length))
+            for leading_zeros in range(3):
+                encoded = der.encode_seq((b"\x00" * leading_zeros + raw_int,))
+                decoded = der.decode_seq(encoded)
+                self.assertEqual(decoded, [raw_int])
+
+        for zeroes in range(3):
+            encoded = der.encode_seq((b"\x00" * zeroes,))
+            decoded = der.decode_seq(encoded)
+            self.assertEqual(decoded, [b"\x00"])
 
 if __name__ == '__main__':
     unittest.main()