[Webkit-unassigned] [Bug 208186] [OpenSSL] Implement WebCrypto APIs for AES-CTR and AES-KW

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=208186

--- Comment #3 from Yoshiaki Jitsukawa <yoshiaki.jitsukawa at sony.com> ---
Comment on attachment 391641
  --> https://bugs.webkit.org/attachment.cgi?id=391641
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=391641&action=review

> Source/WebCore/crypto/openssl/CryptoAlgorithmAES_CTROpenSSL.cpp:66
> +    // Disable padding

Do we need to explicitly disable padding in CTR mode?

> Source/WebCore/crypto/openssl/CryptoAlgorithmAES_CTROpenSSL.cpp:70
> +    size_t blocks = inputText.size() / blockSize + 1;

Is there any particular reason why you don't use roundUpToMultipleOf() like other AES variants?
e.g. const size_t blocks= roundUpToMultipleOf(blockSize, plainSize) / blockSize

> Source/WebCore/crypto/openssl/CryptoAlgorithmAES_CTROpenSSL.cpp:73
> +    if (counterLength < sizeof(size_t) * 8 && blocks > ((size_t) 1 << counterLength))

Could you remove the space between "(size_t)" and "1"?

> Source/WebCore/crypto/openssl/CryptoAlgorithmAES_CTROpenSSL.cpp:86
> +    Vector<uint8_t> outputText(headSize);

outputText should be reserved and used both in first and second parts to avoid reallocation.

> Source/WebCore/crypto/openssl/CryptoAlgorithmAES_CTROpenSSL.cpp:102
> +    {

Could you early skip the second part by doing condition check like:
 if (capacity < block)

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20200302/be78d461/attachment-0001.htm>