[Bug tdep/25099] New: Amd64 inferior calls with __m256 and __m512 unsupported

vries at gcc dot gnu.org sourceware-bugzilla@sourceware.org
Sun Oct 13 11:36:00 GMT 2019


https://sourceware.org/bugzilla/show_bug.cgi?id=25099

            Bug ID: 25099
           Summary: Amd64 inferior calls with __m256 and __m512
                    unsupported
           Product: gdb
           Version: HEAD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: tdep
          Assignee: unassigned at sourceware dot org
          Reporter: vries at gcc dot gnu.org
  Target Milestone: ---

In amd64-tdep.c we have:
...
static void
amd64_classify_aggregate (struct type *type, enum amd64_reg_class theclass[2])
{
  /* 1. If the size of an object is larger than two eightbytes, or it has       
        unaligned fields, it has class memory.  */
  if (TYPE_LENGTH (type) > 16 || amd64_has_unaligned_fields (type))
    {
      theclass[0] = theclass[1] = AMD64_MEMORY;
      return;
    }
...

where the rule 1 quoted in the comment is from the 'The classification of
aggregate (structures and arrays) and union types' part of section 3.2.3
Parameter Passing of the SYSV x86_64 psABI doc.

The 'two eightbytes' part in the docs by now has been updated twice, once to
'four' for __m256, and once to 'eight' for __m512.

-- 
You are receiving this mail because:
You are on the CC list for the bug.


More information about the Gdb-prs mailing list