Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
[Crypto in CTF] Bleichenbacher RSA Signature Fo...
Search
oalieno
October 31, 2020
Technology
0
570
[Crypto in CTF] Bleichenbacher RSA Signature Forgery
https://github.com/oalieno/Crypto-Course/tree/master/RSA
oalieno
October 31, 2020
Tweet
Share
More Decks by oalieno
See All by oalieno
[Crypto in CTF] Classical Cipher
oalieno
0
420
[Crypto in CTF] Block Cipher Mode
oalieno
0
950
[Crypto in CTF] HASH
oalieno
0
260
[Crypto in CTF] LFSR
oalieno
0
470
[Crypto in CTF] RSA
oalieno
0
660
[Crypto in CTF] Blockchain Security
oalieno
0
400
滲透測試基本技巧與經驗分享
oalieno
2
1.1k
Other Decks in Technology
See All in Technology
✨敗北解法コレクション✨〜Expertだった頃に足りなかった知識と技術〜
nanachi
1
780
メルカリIBIS:AIが拓く次世代インシデント対応
0gm
2
460
プロダクトエンジニアリングで開発の楽しさを拡張する話
barometrica
0
210
LLM時代の検索とコンテキストエンジニアリング
shibuiwilliam
2
750
【OptimizationNight】数理最適化のラストワンマイルとしてのUIUX
brainpadpr
2
560
Oracle Exadata Database Service on Cloud@Customer X11M (ExaDB-C@C) サービス概要
oracle4engineer
PRO
2
6.4k
ABEMAにおける 生成AI活用の現在地 / The Current Status of Generative AI at ABEMA
dekatotoro
0
390
20250807_Kiroと私の反省会
riz3f7
0
270
モノレポにおけるエラー管理 ~Runbook自動生成とチームメンションの最適化
biwashi
0
380
AIは変更差分からユニットテスト_結合テスト_システムテストでテストすべきことが出せるのか?
mineo_matsuya
5
2.6k
React Server ComponentsでAPI不要の開発体験
polidog
PRO
0
350
ロールが細分化された組織でSREと協働するインフラエンジニアは何をするか? / SRE Lounge #18
kossykinto
0
240
Featured
See All Featured
How to train your dragon (web standard)
notwaldorf
96
6.2k
Balancing Empowerment & Direction
lara
2
570
The Cost Of JavaScript in 2023
addyosmani
53
8.8k
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
rmw
34
3.1k
Intergalactic Javascript Robots from Outer Space
tanoku
272
27k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.8k
YesSQL, Process and Tooling at Scale
rocio
173
14k
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
marktimemedia
31
2.5k
Testing 201, or: Great Expectations
jmmastey
45
7.6k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
46
7.6k
Evolution of real-time – Irina Nazarova, EuRuKo, 2024
irinanazarova
8
890
The Myth of the Modular Monolith - Day 2 Keynote - Rails World 2024
eileencodes
26
3k
Transcript
Bleichenbacher RSA Signature Forgery ( 2006 ) oalieno
PKCS
PKCS • PKCS ( Public Key Cryptography Standards ) 是公鑰密碼標準
• 制定了了⼀一系列列從 PKCS#1 到 PKCS#15 的標準 • 其中 PKCS#1 是 RSA Cryptography Standard
ASN.1 • ASN.1 是⾼高階的抽象標準 • 具體的實作編碼規則有 : BER, CER, DER,
PER, XER
PKCS#1 1.5 Signature https://tools.ietf.org/html/rfc2313 Step 1 : Message Digest M
H(M) HASH Sign
• ASN.1 是編碼數據的格式,這裡紀錄了了使⽤用的 hash 演算法 H(M) ASN.1 01 FF …
00 FF D = 00 padding Step 2 : Data Encoding Sign PKCS#1 1.5 Signature https://tools.ietf.org/html/rfc2313
Step 3 : RSA encryption D d % n =
S Sign PKCS#1 1.5 Signature https://tools.ietf.org/html/rfc2313
Step 1 : RSA decryption Verify S e % n
= D PKCS#1 1.5 Signature https://tools.ietf.org/html/rfc2313
Step 2 : Data Decoding Verify • 需要 parse 這個格式取出
H(M) • 這個標準沒有說要怎麼 parse • 如果 e 太⼩小且沒有正確的 parse,就有機會偽造簽章 H(M) ASN.1 01 FF … 00 FF D = 00 PKCS#1 1.5 Signature https://tools.ietf.org/html/rfc2313
Step 3 : Message digesting and comparison M' H(M)' H(M)
Verify compare PKCS#1 1.5 Signature https://tools.ietf.org/html/rfc2313
Bleichenbacher RSA Signature Forgery ( 2006 )
Bleichenbacher RSA Signature Forgery ( 2006 ) https://mailarchive.ietf.org/arch/msg/openpgp/5rnE9ZRN1AokBVj3VqblGlP63QE • ⼜又稱作
BB06 • 針對 PKCS#1 1.5 ( RFC 2313 ) • RSA 簽章偽造 06
Bleichenbacher RSA Signature Forgery ( 2006 ) https://mailarchive.ietf.org/arch/msg/openpgp/5rnE9ZRN1AokBVj3VqblGlP63QE • 實作缺陷
: 可以有多餘的字元在後⾯面 • parse 的時候直接取出後⾯面固定長度的 H(M) • 沒有檢查後⾯面還有沒有東⻄西 H(M) ASN.1 01 FF … 00 FF 00 Garbage
• 在 e = 3 的情況下可以 forge signature • 嘗試構造
ED 讓 ED 的三次⽅方不超過 n 且滿⾜足以下格式 S 3 % n = H(M) ASN.1 01 FF … 00 FF 00 Garbage Bleichenbacher RSA Signature Forgery ( 2006 ) https://mailarchive.ietf.org/arch/msg/openpgp/5rnE9ZRN1AokBVj3VqblGlP63QE
H(M) ASN.1 01 FF … 00 FF Garbage 00 D
( length d ) G ( length g ) 2t−15 G + total length t (x + y)3 x3 3x2y + 2g ⋅ D + −2d+g 3xy2 y3 + + = Bleichenbacher RSA Signature Forgery ( 2006 ) https://mailarchive.ietf.org/arch/msg/openpgp/5rnE9ZRN1AokBVj3VqblGlP63QE
x = 2t − 15 3 y = (D −
2d) ⋅ 2g 3 ⋅ 22(t − 15) 3 Bleichenbacher RSA Signature Forgery ( 2006 ) https://mailarchive.ietf.org/arch/msg/openpgp/5rnE9ZRN1AokBVj3VqblGlP63QE
x = 21019 y = (D − 2288) ⋅ 234
3 • 假設 • Key 長度為 3072 bit • Garbage 長度為 2072 bit • 使⽤用 SHA-1 的話,D 的長度是 288 bit • 最後 ED = x + y 就是我們構造出的合法簽章 Bleichenbacher RSA Signature Forgery ( 2006 ) https://mailarchive.ietf.org/arch/msg/openpgp/5rnE9ZRN1AokBVj3VqblGlP63QE
RSA Signature Forgery in python-rsa ( 2016 ) CVE-2016-1494
RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/ •
實作缺陷 : padding bytes 可以是任意字元 直接取第⼆二個 0x00 沒有檢查中間的 padding bytes
• 在 e = 3 的情況下可以 forge signature • 嘗試構造
ED 讓 ED 的三次⽅方不超過 n 且滿⾜足以下格式 • ED3 的後綴是 ASN.1 + H(M) • ED3 的前綴是 \x00\x01 H(M) ASN.1 01 ?? … 00 ?? 00 S 3 % n = RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
0 S S3 ⽬目標 0 0 1 0 0 0
1 1 1 0 1 match RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
0 S S3 ⽬目標 0 0 1 0 0 0
1 1 1 0 1 match RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
0 S S3 ⽬目標 0 0 1 0 0 0
1 1 1 0 1 mismatch RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
0 S S3 ⽬目標 1 0 1 1 1 0
1 1 1 0 1 match RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
0 S S3 ⽬目標 1 0 1 1 1 0
1 1 1 0 1 match RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
0 S S3 ⽬目標 1 0 1 1 1 0
1 1 1 0 1 01013 = 1111101 RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
01 … … 00 3 = 92 3f … 68
04 bc 28 76 e4 50 … = 3 • 要讓 ED3 的前綴是 \x00\x01 只要把 \x00\x01... 開三次⽅方 • 最後再把開完三次⽅方的值的後綴換成前⾯面算出來來的後綴 • 就可以成功⾃自⼰己構造合法簽章了了 RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
H(M) ASN.1 01 ?? … 00 ?? 00 92 3f
… bc 28 3 = RSA Signature Forgery in python-rsa ( 2016 ) https://blog.filippo.io/bleichenbacher-06-signature-forgery-in-python-rsa/
A Decade After Bleichenbacher '06, RSA Signature Forgery Still Works
( 2019 )
A Decade After Bleichenbacher '06, RSA Signature Forgery Still Works
( 2019 ) https://i.blackhat.com/USA-19/Wednesday/us-19-Chau-A-Decade-After-Bleichenbacher-06-RSA-Signature-Forgery-Still-Works.pdf • 整個格式固定是 n 這麼長 • ⽤用 Symbolic Execution 去找到可以任意亂塞的部分有多長
A Decade After Bleichenbacher '06, RSA Signature Forgery Still Works
( 2019 ) https://i.blackhat.com/USA-19/Wednesday/us-19-Chau-A-Decade-After-Bleichenbacher-06-RSA-Signature-Forgery-Still-Works.pdf • 實作缺陷 : padding bytes 可以是任意字元 H(M) ASN.1 01 ?? … 00 ?? 00 CVE-2018-15836 Openswan 2.6.50
CVE-2018-16152 strongSwan 5.6.3 A Decade After Bleichenbacher '06, RSA Signature
Forgery Still Works ( 2019 ) https://i.blackhat.com/USA-19/Wednesday/us-19-Chau-A-Decade-After-Bleichenbacher-06-RSA-Signature-Forgery-Still-Works.pdf • 實作缺陷 : • Algorithm Parameter 可以是任意字元 • Algorithm OID 後⾯面可以有多餘的字元 H(M) 01 FF … 00 FF 00 ASN.1 00 03 20 03 0c Algorithm Parameter 04 10 Algorithm OID
CVE-2018-16150 axTLS 2.1.3 A Decade After Bleichenbacher '06, RSA Signature
Forgery Still Works ( 2019 ) https://i.blackhat.com/USA-19/Wednesday/us-19-Chau-A-Decade-After-Bleichenbacher-06-RSA-Signature-Forgery-Still-Works.pdf • 實作缺陷 : • 可以有多餘的字元在後⾯面 • Algorithm Identifier 可以是任意字元 H(M) 01 FF … 00 FF 00 ASN.1 00 03 20 03 0c Algorithm Identifier 04 10 Garbage
Defense against RSA Signature Forgery
How to defense? • ⽤用其他的簽章演算法,比如說 ECDSA • ⽤用更更⼤大的 e,比如 65537
• parsing based → comparison based H(M) ASN.1 01 FF … 00 FF 00 H(M) ASN.1 01 FF … 00 FF 00 compare