From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:40574)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <Yongbok.Kim@imgtec.com>) id 1YzNNs-0000T8-KX
	for qemu-devel@nongnu.org; Mon, 01 Jun 2015 06:55:57 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <Yongbok.Kim@imgtec.com>) id 1YzNNo-0007VS-UO
	for qemu-devel@nongnu.org; Mon, 01 Jun 2015 06:55:56 -0400
Received: from mailapp01.imgtec.com ([195.59.15.196]:5656)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <Yongbok.Kim@imgtec.com>) id 1YzNNo-0007Uj-Op
	for qemu-devel@nongnu.org; Mon, 01 Jun 2015 06:55:52 -0400
Message-ID: <556C17EC.4040907@imgtec.com>
Date: Mon, 1 Jun 2015 09:29:32 +0100
From: Yongbok Kim <yongbok.kim@imgtec.com>
MIME-Version: 1.0
References: <1432733342-64176-1-git-send-email-yongbok.kim@imgtec.com>
	<1432733342-64176-2-git-send-email-yongbok.kim@imgtec.com>
	<55685C94.6020009@imgtec.com>
In-Reply-To: <55685C94.6020009@imgtec.com>
Content-Type: text/plain; charset="windows-1252"
Content-Transfer-Encoding: quoted-printable
Subject: Re: [Qemu-devel] [PATCH v6 1/3] target-mips: Misaligned memory
	accesses for R6
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Leon Alrae <leon.alrae@imgtec.com>, qemu-devel@nongnu.org
Cc: peter.maydell@linaro.org, afaerber@suse.de, rth@twiddle.net

On 29/05/2015 13:33, Leon Alrae wrote:
> On 27/05/2015 14:29, Yongbok Kim wrote:
>> @@ -2143,7 +2146,8 @@ static void gen_ld(DisasContext *ctx, uint32_t o=
pc,
>>          t1 =3D tcg_const_tl(pc_relative_pc(ctx));
>>          gen_op_addr_add(ctx, t0, t0, t1);
>>          tcg_temp_free(t1);
>> -        tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, MO_TEQ);
>> +        tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, MO_TEQ |
>> +                           ctx->default_tcg_memop_mask);
>>          gen_store_gpr(t0, rt);
>>          opn =3D "ldpc";
>>          break;
>> @@ -2152,22 +2156,26 @@ static void gen_ld(DisasContext *ctx, uint32_t=
 opc,
>>          t1 =3D tcg_const_tl(pc_relative_pc(ctx));
>>          gen_op_addr_add(ctx, t0, t0, t1);
>>          tcg_temp_free(t1);
>> -        tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, MO_TESL);
>> +        tcg_gen_qemu_ld_tl(t0, t0, ctx->mem_idx, MO_TESL |
>> +                           ctx->default_tcg_memop_mask);
>>          gen_store_gpr(t0, rt);
>>          opn =3D "lwpc";
>>          break;
>=20
> As I can see in other places you skipped load/store instructions not
> present in R6 spec (like pre-R6 DSP or microMIPS loads/stores), which
> probably is fine. However, IIUC these two instructions LWPC and LDPC ar=
e
> from mips16 ASE, so probably you want to skip them as well? (note that
> for R6 we=92ve got R6_OPC_LWPC and R6_OPC_LDPC and they are naturally
> aligned).

Nice finding!

>=20
> Apart from that,
>=20
> Reviewed-by: Leon Alrae <leon.alrae@imgtec.com>
>=20
> BTW these OPC_LWPC and OPC_LDPC are very confusing... I presume these
> fake instructions were created in order to reuse gen_ldst() function fo=
r
> mips16 M16_OPC_LWPC and I64_LDPC instructions. It would be nice to clea=
n
> this up at some point (the same as OPC_JALRC).
>=20
> Leon
>=20

Yes, indeed. I have removed those fictional branch instructions similar t=
o
these before, still we have more to remove.

Regards,
Yongbok