[Bug target/67995] New: __attribute__ ((target("arch=XXX"))) enables unsupported ISA

[Bug target/67995] New: __attribute__ ((target("arch=XXX"))) enables unsupported ISA

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67995

            Bug ID: 67995
           Summary: __attribute__ ((target("arch=XXX"))) enables
                    unsupported ISA
           Product: gcc
           Version: 5.3.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: hjl.tools at gmail dot com
  Target Milestone: ---

[hjl@gnu-6 pr67985]$ cat r.i
unsigned int
__attribute__ ((target("arch=core2")))
__x86_rdrand(void)
{
  unsigned int retries = 100;
  unsigned int val;

  while (__builtin_ia32_rdrand32_step(&val) == 0)
    if (--retries == 0)
      return 0;

  return val;
}
[hjl@gnu-6 pr67985]$ /export/build/gnu/gcc-test/build-x86_64-linux/gcc/xgcc
-B/export/build/gnu/gcc-test/build-x86_64-linux/gcc/ -O2 -march=haswell -S -o
r.s r.i
[hjl@gnu-6 pr67985]$ cat r.s
        .file   "r.i"
        .text
        .p2align 4,,15
        .globl  __x86_rdrand
        .type   __x86_rdrand, @function
__x86_rdrand:
.LFB0:
        .cfi_startproc
        movl    $100, %eax
        movl    $1, %ecx
        jmp     .L2
        .p2align 4
.L4:
        subl    $1, %eax
        je      .L8
.L2:
        rdrand  %edx
        movl    %edx, -4(%rsp)
        cmovc   %ecx, %edx
        testl   %edx, %edx
        je      .L4
        movl    -4(%rsp), %eax
.L8:
        ret
        .cfi_endproc
.LFE0:
        .size   __x86_rdrand, .-__x86_rdrand
        .ident  "GCC: (GNU) 6.0.0 20151016 (experimental)"
        .section        .note.GNU-stack,"",@progbits
[hjl@gnu-6 pr67985]$ 

There should be a warning since rdrand isn't supported on Intel Core 2.

[Bug target/67995] __attribute__ ((target("arch=XXX"))) enables unsupported ISA

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67995

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> ---
Created attachment 36532
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36532&action=edit
A patch

I am testing this patch.

[Bug target/67995] __attribute__ ((target("arch=XXX"))) enables unsupported ISA

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67995

--- Comment #2 from H.J. Lu <hjl.tools at gmail dot com> ---
*** Bug 67994 has been marked as a duplicate of this bug. ***

[Bug target/67995] __attribute__ ((target("arch=XXX"))) enables unsupported ISA

[hjl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67995

--- Comment #3 from hjl at gcc dot gnu.org <hjl at gcc dot gnu.org> ---
Author: hjl
Date: Mon Oct 19 11:18:14 2015
New Revision: 228967

URL: https://gcc.gnu.org/viewcvs?rev=228967&root=gcc&view=rev
Log:
Don't leak ISA to __attribute__ ((target("arch=XXX")))

When processing __attribute__ ((target("arch=XXX"))), we should clear
the ISA bits in x_ix86_isa_flags first to avoid leaking ISA from
command line.

gcc/

        PR target/67995
        * config/i386/i386.c (ix86_valid_target_attribute_tree): If
        arch= is set,  clear all bits in x_ix86_isa_flags, except for
        ISA_64BIT, ABI_64, ABI_X32, and CODE16.

gcc/testsuite/

        PR target/67995
        * gcc.target/i386/pr67995-1.c: New test.
        * gcc.target/i386/pr67995-2.c: Likewise.
        * gcc.target/i386/pr67995-3.c: Likewise.

Added:
    trunk/gcc/testsuite/gcc.target/i386/pr67995-1.c
    trunk/gcc/testsuite/gcc.target/i386/pr67995-2.c
    trunk/gcc/testsuite/gcc.target/i386/pr67995-3.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/i386/i386.c
    trunk/gcc/testsuite/ChangeLog

[Bug target/67995] __attribute__ ((target("arch=XXX"))) enables unsupported ISA

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67995

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-10-19
      Known to work|                            |6.0
     Ever confirmed|0                           |1
      Known to fail|                            |4.9.4, 5.3.0

--- Comment #4 from H.J. Lu <hjl.tools at gmail dot com> ---
Fixed on trunk so far.