From: Aurelien Jarno <aurelien@aurel32.net>
To: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH][ppc] convert SPE logical instructions to TCG
Date: Wed, 15 Oct 2008 19:00:55 +0200 [thread overview]
Message-ID: <48F621C7.1020906@aurel32.net> (raw)
In-Reply-To: <20081015020915.GM18389@codesourcery.com>
Nathan Froyd a écrit :
> On Tue, Oct 14, 2008 at 09:16:17PM +0200, Aurelien Jarno wrote:
>> Yes, you should use local variable, cpu_T array will eventually
>> disappear. However, I am more concerned by the fact that it may not be a
>> good idea to expand the macro. IMHO you should put the common code in
>> the macro and pass the name of the function that does the logical
>> operation as an argument.
Applied with the change suggested by Paul Brook. Thanks.
> Done thusly.
>
> -Nathan
>
> diff -r 5ed4e3fd0fe7 -r cac4af009435 target-ppc/op.c
> --- a/target-ppc/op.c Tue Oct 14 15:23:35 2008 -0400
> +++ b/target-ppc/op.c Tue Oct 14 22:06:28 2008 -0400
> @@ -2513,54 +2513,6 @@
> RETURN();
> }
>
> -void OPPROTO op_evand (void)
> -{
> - T0_64 &= T1_64;
> - RETURN();
> -}
> -
> -void OPPROTO op_evandc (void)
> -{
> - T0_64 &= ~T1_64;
> - RETURN();
> -}
> -
> -void OPPROTO op_evor (void)
> -{
> - T0_64 |= T1_64;
> - RETURN();
> -}
> -
> -void OPPROTO op_evxor (void)
> -{
> - T0_64 ^= T1_64;
> - RETURN();
> -}
> -
> -void OPPROTO op_eveqv (void)
> -{
> - T0_64 = ~(T0_64 ^ T1_64);
> - RETURN();
> -}
> -
> -void OPPROTO op_evnor (void)
> -{
> - T0_64 = ~(T0_64 | T1_64);
> - RETURN();
> -}
> -
> -void OPPROTO op_evorc (void)
> -{
> - T0_64 |= ~T1_64;
> - RETURN();
> -}
> -
> -void OPPROTO op_evnand (void)
> -{
> - T0_64 = ~(T0_64 & T1_64);
> - RETURN();
> -}
> -
> void OPPROTO op_evsrws (void)
> {
> do_evsrws();
> diff -r 5ed4e3fd0fe7 -r cac4af009435 target-ppc/translate.c
> --- a/target-ppc/translate.c Tue Oct 14 15:23:35 2008 -0400
> +++ b/target-ppc/translate.c Tue Oct 14 22:06:28 2008 -0400
> @@ -5427,6 +5427,23 @@
> gen_store_gpr64(rD(ctx->opcode), cpu_T64[0]); \
> }
>
> +#define GEN_SPEOP_TCG_ARITH2(name) \
> +static always_inline void gen_##name (DisasContext *ctx) \
> +{ \
> + if (unlikely(!ctx->spe_enabled)) { \
> + GEN_EXCP_NO_AP(ctx); \
> + return; \
> + } \
> + TCGv t0 = tcg_temp_local_new(TCG_TYPE_I64); \
> + TCGv t1 = tcg_temp_local_new(TCG_TYPE_I64); \
> + gen_load_gpr64(t0, rA(ctx->opcode)); \
> + gen_load_gpr64(t1, rB(ctx->opcode)); \
> + gen_op_##name(t0, t1); \
> + gen_store_gpr64(rD(ctx->opcode), t0); \
> + tcg_temp_free(t0); \
> + tcg_temp_free(t1); \
> +}
> +
> #define GEN_SPEOP_ARITH1(name) \
> static always_inline void gen_##name (DisasContext *ctx) \
> { \
> @@ -5453,14 +5470,59 @@
> }
>
> /* Logical */
> -GEN_SPEOP_ARITH2(evand);
> -GEN_SPEOP_ARITH2(evandc);
> -GEN_SPEOP_ARITH2(evxor);
> -GEN_SPEOP_ARITH2(evor);
> -GEN_SPEOP_ARITH2(evnor);
> -GEN_SPEOP_ARITH2(eveqv);
> -GEN_SPEOP_ARITH2(evorc);
> -GEN_SPEOP_ARITH2(evnand);
> +static always_inline void gen_op_evand (TCGv t0, TCGv t1)
> +{
> + tcg_gen_and_i64(t0, t0, t1);
> +}
> +
> +static always_inline void gen_op_evandc (TCGv t0, TCGv t1)
> +{
> + tcg_gen_not_i64(t1, t1);
> + tcg_gen_and_i64(t0, t0, t1);
> +}
> +
> +static always_inline void gen_op_evxor (TCGv t0, TCGv t1)
> +{
> + tcg_gen_xor_i64(t0, t0, t1);
> +}
> +
> +static always_inline void gen_op_evor (TCGv t0, TCGv t1)
> +{
> + tcg_gen_or_i64(t0, t0, t1);
> +}
> +
> +static always_inline void gen_op_evnor (TCGv t0, TCGv t1)
> +{
> + tcg_gen_or_i64(t0, t0, t1);
> + tcg_gen_not_i64(t0, t0);
> +}
> +
> +static always_inline void gen_op_eveqv (TCGv t0, TCGv t1)
> +{
> + tcg_gen_xor_i64(t0, t0, t1);
> + tcg_gen_not_i64(t0, t0);
> +}
> +
> +static always_inline void gen_op_evorc (TCGv t0, TCGv t1)
> +{
> + tcg_gen_not_i64(t1, t1);
> + tcg_gen_or_i64(t0, t0, t1);
> +}
> +
> +static always_inline void gen_op_evnand (TCGv t0, TCGv t1)
> +{
> + tcg_gen_and_i64(t0, t0, t1);
> + tcg_gen_not_i64(t0, t0);
> +}
> +
> +GEN_SPEOP_TCG_ARITH2(evand);
> +GEN_SPEOP_TCG_ARITH2(evandc);
> +GEN_SPEOP_TCG_ARITH2(evxor);
> +GEN_SPEOP_TCG_ARITH2(evor);
> +GEN_SPEOP_TCG_ARITH2(evnor);
> +GEN_SPEOP_TCG_ARITH2(eveqv);
> +GEN_SPEOP_TCG_ARITH2(evorc);
> +GEN_SPEOP_TCG_ARITH2(evnand);
> GEN_SPEOP_ARITH2(evsrwu);
> GEN_SPEOP_ARITH2(evsrws);
> GEN_SPEOP_ARITH2(evslw);
>
>
>
--
.''`. Aurelien Jarno | GPG: 1024D/F1BCDB73
: :' : Debian developer | Electrical Engineer
`. `' aurel32@debian.org | aurelien@aurel32.net
`- people.debian.org/~aurel32 | www.aurel32.net
prev parent reply other threads:[~2008-10-15 17:01 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-10-14 17:19 [Qemu-devel] [PATCH][ppc] convert SPE logical instructions to TCG Nathan Froyd
[not found] ` <CC7516B4-D44E-472F-BADE-D30481322C38@hotmail.com>
2008-10-14 17:33 ` C.W. Betts
2008-10-14 19:16 ` Aurelien Jarno
2008-10-15 2:09 ` Nathan Froyd
2008-10-15 9:32 ` Paul Brook
2008-10-15 17:00 ` Aurelien Jarno [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=48F621C7.1020906@aurel32.net \
--to=aurelien@aurel32.net \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).