Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Fri, 18 Sep 2015 18:51:55 +0200
From: Frank Dittrich <>
Subject: Re: multi-threaded hash table initialization

On 09/18/2015 06:45 PM, magnum wrote:
> On 2015-09-18 18:23, Solar Designer wrote:
>> Well, let's try reverting "multi-threaded hash table initialization",
>> which we should anyway.  Then if the problem persists, investigate
>> further.
> It's now reverted. I also saw very weird problems with OMP and forking
> (similar test to the raw-md5 but using LM). The children got stuck in
> limbo.

May be it's good to revert it, because of questionable performance
improvements and possible performance regressions.

Regarding the failing non-OMP tests, I must admit that I feel somewhat
stupid now.
It may have been a combination of PEBKAC and a bug:

After disabling SessionFileProtect = Always in my config, I failed to
reproduce the problem.
(May be JimF also has SessionFileProtect set in his config.)

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.