[mpich-discuss] Incorrect error checking?

[Nick Radcliffe]

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, MN 55425

+1-651-605-8864  nradclif at cray.com<mailto:email at cray.com>  www.cray.com<http://www.cray.com>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20190618/a8c589dc/attachment.html>