[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <84fc99dbf937ae21e9b99e083d56c4bf@smtp.hushmail.com>
Date: Sun, 17 Jun 2012 17:11:08 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Re: [patch] optional new raw sha1 implemetation
It's gone again anyway, Tavis reverted to defining _GNU_SOURCE and using
memrchr. Will this work on *BSD? I have no idea. For a Sparc it's no
issue since the format won't be included anyway.
magnum
On 2012-06-17 17:00, jfoug wrote:
> Why are we doing a memset here? A simply *ptr=0 is all that is needed. We
> simply need to return a ASCIIZ string, changing the 0x80 to 0x00.
>
> + strrchr((char*)key, 0x80) = 0;
>
> We are assured that there will be a 0x80 within first 16 bytes, and we have
> an array that now is 20 bytes long.
>
> I also think the key[4] = 0 could be eliminated by doing: static uint32_t
> key[5] = {0}; or by using a pointer and calloc(17).
>
> Jim.
>
>> @@ -270,7 +270,7 @@ static void sha1_fmt_set_key(char *key, int index)
>>
>> static char * sha1_fmt_get_key(int index)
>> {
>> - static uint32_t key[4];
>> + static uint32_t key[5];
>>
>> // This function is not hot, we can do this slowly. First, restore
>> // endianness.
>> @@ -279,6 +279,9 @@ static char * sha1_fmt_get_key(int index)
>> key[2] = __builtin_bswap32(M[index][2]);
>> key[3] = __builtin_bswap32(M[index][3]);
>>
>> + // We need a null byte for the strrchr
>> + key[4] = 0;
>> +
>> // Skip backwards until we hit the trailing bit, then remove it.
>> memset(strrchr((char*)key, 0x80), 0x00, 1);
>>
>>
>> magnum
>
>
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
