|
|
Message-ID: <f74099bf3a46432a6b5f8ef3535b60fc@smtp.hushmail.com>
Date: Thu, 05 Jul 2012 15:22:14 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: sha512crypt-opencl seems flakey
Yes, that fixed it! I can't see why we did not hit this until now. Here
is the patch:
--- a/src/common-opencl.c
+++ b/src/common-opencl.c
@@ -40,7 +40,7 @@ static void read_kernel_source(char *kernel_filename)
fseek(fp, 0, SEEK_END);
source_size = ftell(fp);
fseek(fp, 0, SEEK_SET);
- kernel_source = mem_calloc_tiny(source_size, MEM_ALIGN_NONE);
+ kernel_source = mem_calloc_tiny(source_size + 1, MEM_ALIGN_NONE);
read_size = fread(kernel_source, sizeof(char), source_size, fp);
if (read_size != source_size)
fprintf(stderr,
Until now I have never had a problem with this, but today I got random
[runtime-]build errors in the OpenCL formats. Different formats
different times.
magnum
On 2012-07-05 14:40, magnum wrote:
> Maybe it was fixed with 73dffd4e, committed now. Weird. I will test now.
>
> magnum
>
>
> On 2012-07-05 03:16, Claudio André wrote:
>> Will check.
>>
>> Thanks.
>>
>> Em 04-07-2012 19:04, magnum escreveu:
>>> Claudio,
>>>
>>> It seems sha512crypt-opencl sometimes pass Test Suite, and sometimes not?
>>>
>>> form=sha512crypt-opencl guesses: 1389 time: 0:00:01:02
>>> [PASSED]
>>> .pot CHK:sha512crypt-opencl guesses: 0 unk unk : Expected
>>> count(s) (1389) [!!!FAILED!!!]
>>>
>>> Here it passed first pass, but failed .pot check
>>>
>>> Next run:
>>>
>>> form=sha512crypt-opencl guesses: 0 unk unk : Expected
>>> count(s) (1389) [!!!FAILED!!!]
>>> sh: 1: cannot open ./tst.pot: No such file
>>> .pot CHK:sha512crypt-opencl guesses: 0 unk unk [PASSED]
>>>
>>> Here it failed in the first pass.
>>>
>>> This was on bull's Nvidia, 1.7.9-jumbo-6-fixes branch
>>>
>>> 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.