On Jun 20, 2019, at 1:42 PM, Nick Radcliffe <nradclif@cray.com> wrote:

Hui,

The library was built on one of our internal build servers, and run on a different machine of course. But I don't think it's a configure issue. I did a quick test and it's definitely possible for the address of a pointer to not be 8-byte aligned:

[nradclif@heron ~]$ cat test_align.c
#include <stdlib.h>
#include <stdio.h>

int main()
{
    int *c;

    struct
    {
        int a;
        int *b;
    } __attribute__((packed)) s;

    c = (int *)&s.b;

    fprintf(stdout, "c = %p\n", c);

    return 0;
}
[nradclif@heron ~]$ ./test_align.x
c = 0x7ffe3d4e2c44

I'll have to talk to the user to get more details about their code.

Nick Radcliffe  Software Engineer | Cray Inc. 

2131 Lindau Ln #1000 | Bloomington, MN 55425 

+1-651-605-8864  nradclif@cray.com  www.cray.com

---

From: Zhou, Hui <zhouh@anl.gov>
Sent: Thursday, June 20, 2019 1:13:43 PM
To: discuss@mpich.org
Cc: Nick Radcliffe
Subject: Re: [mpich-discuss] Incorrect error checking?

 

That’s interesting. It means the `configure` was finding pointers has to 8-bytes aligned but the user code is having pointers that are 4-byte aligned. I suspect the user did not build the library on the same machine it runs? I am also curious that which architecture actually are not allowing the 4-byte aligned pointers. Nick, is it possible to trace where the questioned mpi library was configured?

— 

Hui Zhou

T: 630-252-3430

On Jun 20, 2019, at 1:04 PM, Nick Radcliffe via discuss <discuss@mpich.org> wrote:

Ken,

The reason I brought this up is that a user's (valid) code was failing due to this parameter check. It's possible to use an address that's not 8 byte aligned to store a pointer to an int, i.e.,

    int *x;

    int *y;

    x = malloc(16);
    y = &x[1];

I would generally expect y to be 4 byte aligned in the above case. So sometimes this parameter check will cause a correct program to fail, right?

Nick Radcliffe  Software Engineer | Cray Inc. 

2131 Lindau Ln #1000 | Bloomington, MN 55425 

+1-651-605-8864  nradclif@cray.com  www.cray.com

---

From: Raffenetti, Kenneth J. via discuss <discuss@mpich.org>
Sent: Wednesday, June 19, 2019 2:44:40 PM
To: discuss@mpich.org
Cc: Raffenetti, Kenneth J.
Subject: Re: [mpich-discuss] Incorrect error checking?

 

Upon further thought, it's just trying to catch a common error, and uses 
alignment to detect it since the type information is lost thru the 
interface. In testing on x86_64, it doesn't appear that it will catch 
the intended error.

As Thomas pointed out in another reply, there are also some issues with 
the implementation, though it is correct that we require intptr_t to 
build MPICH. At the very least, we could patch to use sizeof(void *) vs. 
sizeof(intptr_t). Or we can still remove the check since the use-case is 
so narrow. The preprocessor guard also depends on a AC_TRY_RUN that I'd 
prefer to delete, to be honest.

Ken

On 6/19/19 9:27 AM, Raffenetti, Kenneth J. via discuss wrote:
> Nick,
> 
> I agree this check looks like nonsense. IMO, we can just delete it and
> the others like it. I will create a PR.
> 
> Ken
> 
> On 6/18/19 5:25 PM, Nick Radcliffe via discuss wrote:
>> I recently noticed that a parameter check in MPII_Win_get_attr looks a
>> bit off:
>>
>>
>>               /* A common user error is to pass the address of a 4-byte
>>                * int when the address of a pointer (or an address-sized int)
>>                * should have been used.  We can test for this specific
>>                * case.  Note that this code assumes sizeof(intptr_t) is
>>                * a power of 2. */
>>               if ((intptr_t) attribute_val & (sizeof(intptr_t) - 1)) {
>>                   MPIR_ERR_SETANDSTMT(mpi_errno, MPI_ERR_ARG, goto
>> fn_fail, "**attrnotptr");
>>               }
>>
>>
>> The comment indicates that the check is testing "attribute_val" to see
>> if a pointer to int was passed in, rather than a double pointer or
>> pointer to an integer that can store an address. But the check seems to
>> only be testing for 8 byte alignment (and failing if not aligned). Am I
>> missing something here?
>>
>>
>>
>> Nick Radcliffe  Software Engineer| Cray Inc.
>>
>> 2131 Lindau Ln #1000 | Bloomington, MN55425
>>
>> +1-651-605-8864  nradclif@cray.com <mailto:email@cray.com>  www.cray.com
>> <http://www.cray.com>
>>
>>
>> _______________________________________________
>> discuss mailing list     discuss@mpich.org
>> To manage subscription options or unsubscribe:
>> https://lists.mpich.org/mailman/listinfo/discuss
>>
> _______________________________________________
> discuss mailing list     discuss@mpich.org
> To manage subscription options or unsubscribe:
> https://lists.mpich.org/mailman/listinfo/discuss
> 
_______________________________________________
discuss mailing list     discuss@mpich.org
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss

_______________________________________________
discuss mailing list     discuss@mpich.org
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss