Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 29 Sep 2015 16:18:57 +0530
From: Sayantan Datta <std2048@...il.com>
To: john-dev <john-dev@...ts.openwall.com>
Subject: Re: Reading loaded password hashes inside crypt-all.

On Tue, Sep 29, 2015 at 3:05 PM, Sayantan Datta <std2048@...il.com> wrote:

> Hi Solar,
>
> Are password hashes removed from database as they are cracked ? Apparently
> there are discrepancies between salt->count and number of remaining hashes
> in salt->list. While salt-count is updated, it seems like salt->list is not
> updated to reflect the cracked hashes.
>
> pw = salt -> list;
> i = 0;
> do {
>         bin = (int *)pw -> binary;
>         if (bin != NULL) {
>           store_binary();
>           i++ ;
>         }
>     } while ((pw = pw -> next)) ;
>
> Running this loop in the middle of cracking results in 'i' being greater
> than 'salt->count' !! Some recent change might have triggered this issue.
> Earlier, such behavior were not noticed.
>
> This broke an optimization where hashes are re-loaded onto GPU after some
> of them are cracked, resulting in better speed!!
>
> Regadrs,
> Sayantan
>

Sorry, must have sent twice due to network issues!!

This one line patch fixes the issues but I'm not sure to commit it.

Regards,
Sayantan

Content of type "text/html" skipped

View attachment "patch.diff" of type "text/plain" (433 bytes)

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.