[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 25 Mar 2013 02:07:59 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Bug in dynamic (was: new warnings with gcc 4.8.0)
On 24 Mar, 2013, at 9:45 , Dhiru Kholia <dhiru.kholia@...il.com> wrote:
> I tried to build unstable-jumbo with gcc 4.8.0 and got some new
> warnings.
> dynamic_fmt.c: In function ‘init’:
> dynamic_fmt.c:842:27: warning: iteration 8u invokes undefined behavior [-Waggressive-loop-optimizations]
> total_len2[i*MMX_COEF] = 32;
This one is an actual bug that must be fixed:
429 static unsigned int total_len2[BLOCK_LOOPS];
...
833 if (curdat.input2_set_len32) {
834 for (i = 0; i < MAX_KEYS_PER_CRYPT_X86; ++i)
835 total_len2_X86[i] = 32;
836 #ifdef MMX_COEF
837 for (i = 0; i < BLOCK_LOOPS; ++i) {
838 input_buf2[i].c[GETPOS(32,0)] = 0x80;
839 input_buf2[i].c[GETPOS(57,0)] = 0x1;
840 input_buf2[i].c[GETPOS(32,1)] = 0x80;
841 input_buf2[i].c[GETPOS(57,1)] = 0x1;
842 total_len2[i*MMX_COEF] = 32;
843 total_len2[i*MMX_COEF+1] = 32;
844 #if (MMX_COEF==4)
845 input_buf2[i].c[GETPOS(32,2)] = 0x80;
846 input_buf2[i].c[GETPOS(57,2)] = 0x1;
847 input_buf2[i].c[GETPOS(32,3)] = 0x80;
848 input_buf2[i].c[GETPOS(57,3)] = 0x1;
849 total_len2[i*MMX_COEF+2] = 32;
850 total_len2[i*MMX_COEF+3] = 32;
851 #endif
852 }
853 #endif
854 }
total_len2 is defined as [BLOCK_LOOPS] but in the loop it's used as if it were MMX_COEF*BLOCK_LOOPS. I'm not sure in what end it's wrong - Jim?
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.
