From: Laurent Desnogues <laurent.desnogues@gmail.com>
To: juha.riihimaki@nokia.com
Cc: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH v2 02/10] target-arm: add support for neon vld1.64/vst1.64 instructions
Date: Sun, 25 Oct 2009 12:11:10 +0100 [thread overview]
Message-ID: <761ea48b0910250411k41466bd9g75945660ebf5cea3@mail.gmail.com> (raw)
In-Reply-To: <1256386749-85299-3-git-send-email-juha.riihimaki@nokia.com>
On Sat, Oct 24, 2009 at 1:19 PM, <juha.riihimaki@nokia.com> wrote:
> From: Juha Riihimäki <juha.riihimaki@nokia.com>
>
> Add support for NEON vld1.64 and vst1.64 instructions. This patch is
> revised to follow more closely the specification and raises
> undefined exception if 64bit element size is used for vld2/vst2 or
> vld4/vst4 instructions.
>
> Signed-off-by: Juha Riihimäki <juha.riihimaki@nokia.com>
Acked-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Laurent
> ---
> target-arm/translate.c | 133 +++++++++++++++++++++++++++++-------------------
> 1 files changed, 81 insertions(+), 52 deletions(-)
>
> diff --git a/target-arm/translate.c b/target-arm/translate.c
> index 8a85db6..09c996d 100644
> --- a/target-arm/translate.c
> +++ b/target-arm/translate.c
> @@ -795,6 +795,12 @@ static inline TCGv gen_ld32(TCGv addr, int index)
> tcg_gen_qemu_ld32u(tmp, addr, index);
> return tmp;
> }
> +static inline TCGv_i64 gen_ld64(TCGv addr, int index)
> +{
> + TCGv_i64 tmp = tcg_temp_new_i64();
> + tcg_gen_qemu_ld64(tmp, addr, index);
> + return tmp;
> +}
> static inline void gen_st8(TCGv val, TCGv addr, int index)
> {
> tcg_gen_qemu_st8(val, addr, index);
> @@ -810,6 +816,11 @@ static inline void gen_st32(TCGv val, TCGv addr, int index)
> tcg_gen_qemu_st32(val, addr, index);
> dead_tmp(val);
> }
> +static inline void gen_st64(TCGv_i64 val, TCGv addr, int index)
> +{
> + tcg_gen_qemu_st64(val, addr, index);
> + tcg_temp_free_i64(val);
> +}
>
> static inline void gen_set_pc_im(uint32_t val)
> {
> @@ -3680,6 +3691,7 @@ static int disas_neon_ls_insn(CPUState * env, DisasContext *s, uint32_t insn)
> int op;
> int nregs;
> int interleave;
> + int spacing;
> int stride;
> int size;
> int reg;
> @@ -3690,6 +3702,7 @@ static int disas_neon_ls_insn(CPUState * env, DisasContext *s, uint32_t insn)
> TCGv addr;
> TCGv tmp;
> TCGv tmp2;
> + TCGv_i64 tmp64;
>
> if (!vfp_enabled(env))
> return 1;
> @@ -3702,10 +3715,13 @@ static int disas_neon_ls_insn(CPUState * env, DisasContext *s, uint32_t insn)
> /* Load store all elements. */
> op = (insn >> 8) & 0xf;
> size = (insn >> 6) & 3;
> - if (op > 10 || size == 3)
> + if (op > 10)
> return 1;
> nregs = neon_ls_element_type[op].nregs;
> interleave = neon_ls_element_type[op].interleave;
> + spacing = neon_ls_element_type[op].spacing;
> + if (size == 3 && (interleave | spacing) != 1)
> + return 1;
> load_reg_var(s, addr, rn);
> stride = (1 << size) * interleave;
> for (reg = 0; reg < nregs; reg++) {
> @@ -3716,65 +3732,78 @@ static int disas_neon_ls_insn(CPUState * env, DisasContext *s, uint32_t insn)
> load_reg_var(s, addr, rn);
> tcg_gen_addi_i32(addr, addr, 1 << size);
> }
> - for (pass = 0; pass < 2; pass++) {
> - if (size == 2) {
> - if (load) {
> - tmp = gen_ld32(addr, IS_USER(s));
> - neon_store_reg(rd, pass, tmp);
> - } else {
> - tmp = neon_load_reg(rd, pass);
> - gen_st32(tmp, addr, IS_USER(s));
> - }
> - tcg_gen_addi_i32(addr, addr, stride);
> - } else if (size == 1) {
> - if (load) {
> - tmp = gen_ld16u(addr, IS_USER(s));
> - tcg_gen_addi_i32(addr, addr, stride);
> - tmp2 = gen_ld16u(addr, IS_USER(s));
> - tcg_gen_addi_i32(addr, addr, stride);
> - gen_bfi(tmp, tmp, tmp2, 16, 0xffff);
> - dead_tmp(tmp2);
> - neon_store_reg(rd, pass, tmp);
> - } else {
> - tmp = neon_load_reg(rd, pass);
> - tmp2 = new_tmp();
> - tcg_gen_shri_i32(tmp2, tmp, 16);
> - gen_st16(tmp, addr, IS_USER(s));
> - tcg_gen_addi_i32(addr, addr, stride);
> - gen_st16(tmp2, addr, IS_USER(s));
> + if (size == 3) {
> + if (load) {
> + tmp64 = gen_ld64(addr, IS_USER(s));
> + neon_store_reg64(tmp64, rd);
> + tcg_temp_free_i64(tmp64);
> + } else {
> + tmp64 = tcg_temp_new_i64();
> + neon_load_reg64(tmp64, rd);
> + gen_st64(tmp64, addr, IS_USER(s));
> + }
> + tcg_gen_addi_i32(addr, addr, stride);
> + } else {
> + for (pass = 0; pass < 2; pass++) {
> + if (size == 2) {
> + if (load) {
> + tmp = gen_ld32(addr, IS_USER(s));
> + neon_store_reg(rd, pass, tmp);
> + } else {
> + tmp = neon_load_reg(rd, pass);
> + gen_st32(tmp, addr, IS_USER(s));
> + }
> tcg_gen_addi_i32(addr, addr, stride);
> - }
> - } else /* size == 0 */ {
> - if (load) {
> - TCGV_UNUSED(tmp2);
> - for (n = 0; n < 4; n++) {
> - tmp = gen_ld8u(addr, IS_USER(s));
> + } else if (size == 1) {
> + if (load) {
> + tmp = gen_ld16u(addr, IS_USER(s));
> + tcg_gen_addi_i32(addr, addr, stride);
> + tmp2 = gen_ld16u(addr, IS_USER(s));
> + tcg_gen_addi_i32(addr, addr, stride);
> + gen_bfi(tmp, tmp, tmp2, 16, 0xffff);
> + dead_tmp(tmp2);
> + neon_store_reg(rd, pass, tmp);
> + } else {
> + tmp = neon_load_reg(rd, pass);
> + tmp2 = new_tmp();
> + tcg_gen_shri_i32(tmp2, tmp, 16);
> + gen_st16(tmp, addr, IS_USER(s));
> + tcg_gen_addi_i32(addr, addr, stride);
> + gen_st16(tmp2, addr, IS_USER(s));
> tcg_gen_addi_i32(addr, addr, stride);
> - if (n == 0) {
> - tmp2 = tmp;
> - } else {
> - gen_bfi(tmp2, tmp2, tmp, n * 8, 0xff);
> - dead_tmp(tmp);
> - }
> }
> - neon_store_reg(rd, pass, tmp2);
> - } else {
> - tmp2 = neon_load_reg(rd, pass);
> - for (n = 0; n < 4; n++) {
> - tmp = new_tmp();
> - if (n == 0) {
> - tcg_gen_mov_i32(tmp, tmp2);
> - } else {
> - tcg_gen_shri_i32(tmp, tmp2, n * 8);
> + } else /* size == 0 */ {
> + if (load) {
> + TCGV_UNUSED(tmp2);
> + for (n = 0; n < 4; n++) {
> + tmp = gen_ld8u(addr, IS_USER(s));
> + tcg_gen_addi_i32(addr, addr, stride);
> + if (n == 0) {
> + tmp2 = tmp;
> + } else {
> + gen_bfi(tmp2, tmp2, tmp, n * 8, 0xff);
> + dead_tmp(tmp);
> + }
> }
> - gen_st8(tmp, addr, IS_USER(s));
> - tcg_gen_addi_i32(addr, addr, stride);
> + neon_store_reg(rd, pass, tmp2);
> + } else {
> + tmp2 = neon_load_reg(rd, pass);
> + for (n = 0; n < 4; n++) {
> + tmp = new_tmp();
> + if (n == 0) {
> + tcg_gen_mov_i32(tmp, tmp2);
> + } else {
> + tcg_gen_shri_i32(tmp, tmp2, n * 8);
> + }
> + gen_st8(tmp, addr, IS_USER(s));
> + tcg_gen_addi_i32(addr, addr, stride);
> + }
> + dead_tmp(tmp2);
> }
> - dead_tmp(tmp2);
> }
> }
> }
> - rd += neon_ls_element_type[op].spacing;
> + rd += spacing;
> }
> stride = nregs * 8;
> } else {
> --
> 1.6.5
>
>
>
>
next prev parent reply other threads:[~2009-10-25 11:11 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-10-24 12:18 [Qemu-devel] [PATCH v2 00/10] target-arm: miscellaneous fixes juha.riihimaki
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 01/10] target-arm: fix neon vshrn/vrshrn ops juha.riihimaki
2009-10-25 10:58 ` Laurent Desnogues
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 02/10] target-arm: add support for neon vld1.64/vst1.64 instructions juha.riihimaki
2009-10-25 11:11 ` Laurent Desnogues [this message]
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 03/10] target-arm: allow modifying vfp fpexc en bit only juha.riihimaki
2009-10-25 12:23 ` Laurent Desnogues
2009-10-26 7:32 ` Juha.Riihimaki
2009-10-26 9:08 ` Laurent Desnogues
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 04/10] target-arm: optimize vfp load/store multiple ops juha.riihimaki
2009-10-24 17:36 ` Laurent Desnogues
2009-10-27 8:42 ` Aurelien Jarno
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 05/10] target-arm: optimize arm " juha.riihimaki
2009-10-27 8:39 ` Aurelien Jarno
2009-10-27 8:48 ` Juha.Riihimaki
2009-10-27 9:00 ` Aurelien Jarno
2009-10-27 9:05 ` Juha.Riihimaki
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 06/10] target-arm: fix neon vsri, vshl and vsli ops juha.riihimaki
2009-10-24 17:44 ` Laurent Desnogues
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 07/10] target-arm: optimize thumb2 load/store multiple ops juha.riihimaki
2009-10-24 17:32 ` Laurent Desnogues
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 08/10] target-arm: optimize thumb push/pop ops juha.riihimaki
2009-10-24 17:34 ` Laurent Desnogues
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 09/10] target-arm: optimize neon vld/vst ops juha.riihimaki
2009-10-25 14:01 ` Laurent Desnogues
2009-10-26 7:46 ` Juha.Riihimaki
2009-10-26 9:11 ` Laurent Desnogues
2009-10-26 21:05 ` Aurelien Jarno
2009-10-29 13:45 ` Juha.Riihimaki
2009-10-29 13:52 ` Laurent Desnogues
2009-10-24 12:19 ` [Qemu-devel] [PATCH v2 10/10] target-arm: fix neon shift helper functions juha.riihimaki
2009-10-25 12:16 ` Laurent Desnogues
2009-10-25 19:17 ` [Qemu-devel] [PATCH v2 00/10] target-arm: miscellaneous fixes Aurelien Jarno
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=761ea48b0910250411k41466bd9g75945660ebf5cea3@mail.gmail.com \
--to=laurent.desnogues@gmail.com \
--cc=juha.riihimaki@nokia.com \
--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).