Mailing-List: contact dev-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@httpd.apache.org
Subject: Re: 2.4.17 alignment issue sparc/ia64
To: dev@httpd.apache.org
References: 
 <CALK=YjMWSTLMzr=5nKws8M8=CkrhqQJ8S1xCgyQycms-V1pNDQ@mail.gmail.com>
 <CAKQ1sVP6HPyQj4Tq-47DQ27tbCTRk+005UoAnxmgL7xdWW6+_A@mail.gmail.com>
 <CAKQ1sVO5=kaZitCpQ0Vuqx8E+nj69cv5+pEuzDo1_jhMZRB4Cg@mail.gmail.com>
From: Rainer Jung <rainer.jung@kippdata.de>
Message-ID: <5620EF0E.7000109@kippdata.de>
Date: Fri, 16 Oct 2015 14:35:26 +0200
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:38.0) Gecko/20100101
 Thunderbird/38.3.0
MIME-Version: 1.0
In-Reply-To: 
 <CAKQ1sVO5=kaZitCpQ0Vuqx8E+nj69cv5+pEuzDo1_jhMZRB4Cg@mail.gmail.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit

Am 16.10.2015 um 13:54 schrieb Yann Ylavic:
> On Fri, Oct 16, 2015 at 10:02 AM, Yann Ylavic <ylavic.dev@gmail.com> wrote:
>>
>> We should do something like the following patch:
>>
>> Index: server/scoreboard.c
>> ===================================================================
>> --- server/scoreboard.c    (revision 1708095)
>> +++ server/scoreboard.c    (working copy)
>> @@ -129,14 +129,19 @@ static apr_status_t ap_cleanup_shared_mem(void *d)
>>       return APR_SUCCESS;
>>   }
>>
>> +#define SIZE_OF_scoreboard    APR_ALIGN_DEFAULT(sizeof(scoreboard))
>> +#define SIZE_OF_global_score  APR_ALIGN_DEFAULT(sizeof(global_score))
>> +#define SIZE_OF_process_score APR_ALIGN_DEFAULT(sizeof(process_score))
>> +#define SIZE_OF_worker_score  APR_ALIGN_DEFAULT(sizeof(worker_score))
>
> Maybe the following would be more correct:
>
> +#define SCOREBOARD_ALIGN(size) APR_ALIGN((size),sizeof(void *))
> +#define SIZE_OF_scoreboard     SCOREBOARD_ALIGN(sizeof(scoreboard))
> +#define SIZE_OF_global_score   SCOREBOARD_ALIGN(sizeof(global_score))
> +#define SIZE_OF_process_score  SCOREBOARD_ALIGN(sizeof(process_score))
> +#define SIZE_OF_worker_score   SCOREBOARD_ALIGN(sizeof(worker_score))
>
> since APR_ALIGN_DEFAULT seems to align to 8 bytes whatever the
> platform requires...
>
> Maybe some APR_ALIGN_NATIVE macro is missing in APR, like:
> #define APR_ALIGN_NATIVE(size) APR_ALIGN((size), APR_SIZEOF_VOIDP)
> but that's another story :)

I didn't yet have the time to reproduce and test your patch, but the 
APR_ALIGN((size),sizeof(void *)) align approach would not work. The 
problem here is that even or especially when building for 32 Bits and 
then using a 64 Bit data type, the alignment fails. So the pointer size 
would be 4 here, but the requirement alignment 8 bytes.

Fun with hardware.

Rainer