Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 21 Sep 2015 18:22:35 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Regression problem from john-huge-prefetch.diff

On 21/09/15 12:41, Solar Designer wrote:
> On Sun, Sep 20, 2015 at 11:02:41PM +0200, magnum wrote:
>> I get a segfault (after a while, and after cracking a good number) using
>> single mode with LM. Bisected it to a2294f9f which is
>> john-huge-prefetch.diff. So I tried just defining CRK_PREFETCH to 0 in
>> current code - and problem goes away.
>
> The attached patch should fix this.  I've explained the issue in a
> comment.  An alternative approach would be to check for stale entries,
> which could be simpler to do and would avoid the need to re-prefetch,
> but it'd add a check to a semi-hot path, whereas this patch only adds
> code to a cold path (successful guess).

Yep, that fixed it. Committed now, and workaround removed.

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.