https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87812
--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> ---
Without AVX, v32qi should be passed the same way before AVX was
added to compiler. For example,
[hjl@gnu-cfl-1 tmp]$ cat z.c
typedef int __attribute__((mode(SI))) si;
typedef si __attribute__((vector_size (128))) v;
v
foo (int x)
{
v y = {'0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f',
'0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f'};
return y;
}
[hjl@gnu-cfl-1 tmp]$ gcc -S -O2 z.c
[hjl@gnu-cfl-1 tmp]$ cat z.s
.file "z.c"
.text
.p2align 4,,15
.globl foo
.type foo, @function
foo:
.LFB0:
.cfi_startproc
movq %rdi, %rax
movl $48, (%rdi)
movl $49, 4(%rdi)
movl $50, 8(%rdi)
movl $51, 12(%rdi)
movl $52, 16(%rdi)
movl $53, 20(%rdi)
movl $54, 24(%rdi)
movl $55, 28(%rdi)
movl $56, 32(%rdi)
movl $57, 36(%rdi)
movl $97, 40(%rdi)
movl $98, 44(%rdi)
movl $99, 48(%rdi)
movl $100, 52(%rdi)
movl $101, 56(%rdi)
movl $102, 60(%rdi)
movl $48, 64(%rdi)
movl $49, 68(%rdi)
movl $50, 72(%rdi)
movl $51, 76(%rdi)
movl $52, 80(%rdi)
movl $53, 84(%rdi)
movl $54, 88(%rdi)
movl $55, 92(%rdi)
movl $56, 96(%rdi)
movl $57, 100(%rdi)
movl $97, 104(%rdi)
movl $98, 108(%rdi)
movl $99, 112(%rdi)
movl $100, 116(%rdi)
movl $101, 120(%rdi)
movl $102, 124(%rdi)
ret
.cfi_endproc
.LFE0:
.size foo, .-foo
.ident "GCC: (GNU) 8.2.1 20181011 (Red Hat 8.2.1-4)"
.section .note.GNU-stack,"",@progbits
[hjl@gnu-cfl-1 tmp]$
If there were AVX1024, GCC would pass y in vector register. When AVX1024
were disabled, GCC would pass y in memory and issue an ABI change warning.
So LLVM is wrong.
