Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Thu, 30 Aug 2012 20:20:34 +0200
From: magnum <>
Subject: Re: sha256crypt bug?

On 30 Aug, 2012, at 15:18 , Claudio André <> wrote:

> Em 29-08-2012 18:55, magnum escreveu:
>> The sha256crypt-opencl format emits the following warning:
>> OpenCL platform 0: Apple, 2 device(s).
>> Using device 1: GeForce GT 650M
>> Building the kernel, this could take a while
>> Compilation log: In file included from <program source>:15:
>> <program source>:223:5: warning: array index of '15' indexes past the end of an array (that contains 1 element)
>>     ctx->buffer->mem_32[15] = SWAP32(ctx->total * 8);
>>     ^                   ~~
> mem_32 is a field inside a union. The real array is "buffer", and it can handle the ctx->buffer->mem_32[15]. It is:
>    buffer_32                   buffer[16];     //512bits
> I haven't seen this warning on bull GTX card (i wasn't there for a while, and it was refusing connection when i tried it some time ago).
> I expect it to succeed on TS. Have you tried?.

Not yet, but I expect it to work just fine.

> I can change it to something like this: ctx->buffer[15].mem_32[0].

I tried this now and it works fine (and warnings are gone). Apple's OpenCL compiler is the noisiest I've seen. Sometimes that is a good thing, sometimes not :)


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.