From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id BF9003858C50; Thu,  9 Feb 2023 12:22:56 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org BF9003858C50
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1675945376;
	bh=9vuilwlXYyzuhurgb5CuBKDAyuT0u2EkfGHgT90vwX8=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=e2G3HQCwcGkqquR6x0ek7AbUYgxf0PAEXkTlxTZ3l2Ufjw/RX2wDs1dw56VQ1uGZb
	 DyIjPcTE9CPqS93HVsijXoluMkwuELIsOFyGg9IioGJ9nq6Vo0pYkWz64qPadqmXu+
	 i4Qab2CbQ8uhTF4LdAndi9fNmuQABO0DcTWv4xLY=
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/100758] __builtin_cpu_supports does not (always) detect
 "sse2"
Date: Thu, 09 Feb 2023 12:22:56 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 11.1.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: jakub at gcc dot gnu.org
X-Bugzilla-Status: RESOLVED
X-Bugzilla-Resolution: WONTFIX
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: marxin at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-100758-4-DmLS5znXfV@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-100758-4@http.gcc.gnu.org/bugzilla/>
References: <bug-100758-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D100758

--- Comment #16 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to Erich Eckner from comment #4)
> Created attachment 50871 [details]
> cpuid probing
>=20
> Does the attached program yield, what you need? (Sry, I'm quite unfamiliar
> with asm in gcc)
>=20
> It gives:
>=20
> 00000001 746e6543 736c7561 48727561=20
> 000006d0 00000800 00004181 a7c9bbff=20
>=20
> and
>=20
> 0000000a 746e6543 736c7561 48727561=20
> 000006fa 00010800 008863a9 afc9fbff=20
>=20
> on the two machines, respectively.

Erich, if you still have access to those CPUs, could you retry with:
#include <stdio.h>

static void cpuid( unsigned int ax, unsigned int cx, unsigned int *p )
{
  unsigned int flags;

  __asm __volatile
    ("movl %%ebx, %%esi\n\t"
     "cpuid\n\t"
      "xchgl %%ebx, %%esi"
      : "=3Da" (p[0]), "=3DS" (p[1]),
      "=3Dc" (p[2]), "=3Dd" (p[3])
      : "0" (ax), "2" (cx));
}

int main() {
  unsigned int regs[4];
  for (int j=3D0; j<=3D18; j++) {
    int k =3D j > 10 ? -__INT_MAX__ - 1 + (j - 11) : j;
    cpuid( k, regs, 0 );
    printf("%d %d ", k, 0);
    for (int i=3D0; i<4; i++) {
      printf("%08x ", regs[i]);
    }
    printf("\n");
    cpuid( k, regs, 1 );
    printf("%d %d ", k, 1);
    for (int i=3D0; i<4; i++) {
      printf("%08x ", regs[i]);
    }
    printf("\n");
  }
  return 0;
}

Thanks.

I've posted a different patch to gcc-patches:
https://gcc.gnu.org/pipermail/gcc-patches/2023-February/611632.html=