From nobody Wed Feb 11 00:55:39 2026 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass(p=none dis=none) header.from=linaro.org ARC-Seal: i=1; a=rsa-sha256; t=1633638782; cv=none; d=zohomail.com; s=zohoarc; b=mQXUF40IlFvvkebGUPLq4C+58CzT2imtPheOxdmcaImC+R9t1WOQX2iskuiI+NHenpUgcKS+CZWSRk4a830eYEXb/FDLGoLmLksT/zWMAKcfW0lIfPeU6X/5PRetCT5WWikH6UpccAXR2PedG6v/zQ3qXBGAwGFlAY7c/c/7aVc= ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com; s=zohoarc; t=1633638782; h=Content-Transfer-Encoding:Date:From:In-Reply-To:List-Subscribe:List-Post:List-Id:List-Archive:List-Help:List-Unsubscribe:MIME-Version:Message-ID:References:Sender:Subject:To; bh=NxFgWT4ru6x0IdGSyZDen0BhvRtc1g1hOUCqOyHF/68=; b=JRa/ct3pAPFCw1V7NLjfEVoDOUEylwbs8mIx69HwPrZp6OcgVRm3eVlOsAnqnY/WAIxnG3birDzYnQixnyDkHXJFhOujxFj/VsEV4ATZHUsOfIDwlKJoi5t1t9Q0ebH48aBULROaYIxMOUhZB5duKt6PJOwAgWHlGBv0LFyl1kQ= ARC-Authentication-Results: i=1; mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass header.from= (p=none dis=none) Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 1633638781846822.6057767084923; Thu, 7 Oct 2021 13:33:01 -0700 (PDT) Received: from localhost ([::1]:33406 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1mYa4a-000675-KS for importer@patchew.org; Thu, 07 Oct 2021 16:33:00 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:60192) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1mYZUJ-0000RR-Oe for qemu-devel@nongnu.org; Thu, 07 Oct 2021 15:55:31 -0400 Received: from mail-pl1-x634.google.com ([2607:f8b0:4864:20::634]:36591) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1mYZUC-0007Oe-7J for qemu-devel@nongnu.org; Thu, 07 Oct 2021 15:55:29 -0400 Received: by mail-pl1-x634.google.com with SMTP id y5so4645610pll.3 for ; Thu, 07 Oct 2021 12:55:23 -0700 (PDT) Received: from localhost.localdomain ([71.212.134.125]) by smtp.gmail.com with ESMTPSA id y4sm216555pfn.193.2021.10.07.12.55.21 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 07 Oct 2021 12:55:21 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; h=from:to:subject:date:message-id:in-reply-to:references:mime-version :content-transfer-encoding; bh=NxFgWT4ru6x0IdGSyZDen0BhvRtc1g1hOUCqOyHF/68=; b=zZkrurEKQhCUEVfSlVlbcPaVJ4WnHZHOYdFePYdQ58zuxAB5PtHkkBuVTZoql0an4q HOXRv12PCS08/cosuvcrGfhsVGAD7dhq+Hv5DCOKPfJrEF7kVX+YzPaiFc94EczX22Eh Eqsuzuboqf71glJA0BZB9jGgXIUp1v7RUyWC3Ij+zrRjxDL4feLb0NxC1z0yKtysXL0L diINX9Fwo4KyATed+RRNxqt5MMz7eHXMEMmGjh61QZje5douZN/6PGS/bitZwyZMS6Z1 YDgt5VIUMiM40oQ42YX2tkRmQalB1oITfLlaA+7j0JpER3t5m0jQJVBlqHc47dKH8ZtN 4xSA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:from:to:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=NxFgWT4ru6x0IdGSyZDen0BhvRtc1g1hOUCqOyHF/68=; b=aIFJ+u/ubUUkGrtvRaHIafv0FSQJPlT9xYYOI44ChPRZXwzyzaERj7VIiUOXlv8G+R I+8gfyxCH92RLbsCzZwAMUa9jwGgGDqaSE5W/3HDhQb6hJ/gzorzCNhadm/I8wLTxqOm cfLmiqfimJ0+GusTmPvlIybkCFx3UFcwlNaaPIPVdZeHypYBPE+RY4f0YFIEDIY1/yGC 5w5uzwe8HcNFMbZu13NtBADkhGViqvFmSAbJU7M5VXFc+M6HAp2eMg2kbROY0rquh+95 6ASxZ2UYBCRZv+JhBHdJVqnIr9VvPXEFfDh+xDh8ZtpUxt6kSZy3ml9cHYX1+7esxQUp kceQ== X-Gm-Message-State: AOAM530DJaoV2Kk8EnebG0gssCCftsLtbX7FkF+ixGCbGmqY3nZ0qkpg mfPDsHGY6uy2mmRpZ3Zo621odUDio2YIoA== X-Google-Smtp-Source: ABdhPJyzUDrkYJug42yr4r/Du6DonyRV9eZ9fGWHdsG/Zr6ZKb3JN+3yt1eE2ZvMMAxtfLO+DI+wsA== X-Received: by 2002:a17:90b:1b03:: with SMTP id nu3mr7125266pjb.76.1633636522128; Thu, 07 Oct 2021 12:55:22 -0700 (PDT) From: Richard Henderson To: qemu-devel@nongnu.org Subject: [PATCH v2 33/48] tcg/optimize: Add type to OptContext Date: Thu, 7 Oct 2021 12:54:41 -0700 Message-Id: <20211007195456.1168070-34-richard.henderson@linaro.org> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20211007195456.1168070-1-richard.henderson@linaro.org> References: <20211007195456.1168070-1-richard.henderson@linaro.org> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: pass client-ip=2607:f8b0:4864:20::634; envelope-from=richard.henderson@linaro.org; helo=mail-pl1-x634.google.com X-Spam_score_int: -20 X-Spam_score: -2.1 X-Spam_bar: -- X-Spam_report: (-2.1 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_NONE=-0.0001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" X-ZohoMail-DKIM: pass (identity @linaro.org) X-ZM-MESSAGEID: 1633638783124100003 Content-Type: text/plain; charset="utf-8" Compute the type of the operation early. There are at least 4 places that used a def->flags ladder to determine the type of the operation being optimized. There were two places that assumed !TCG_OPF_64BIT means TCG_TYPE_I32, and so could potentially compute incorrect results for vector operations. Signed-off-by: Richard Henderson --- tcg/optimize.c | 130 ++++++++++++++++++++++++++++--------------------- 1 file changed, 75 insertions(+), 55 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index 0734bcb42e..324866cbf1 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -51,6 +51,7 @@ typedef struct OptContext { =20 /* In flight values from optimization. */ uint64_t z_mask; + TCGType type; } OptContext; =20 static inline TempOptInfo *ts_info(TCGTemp *ts) @@ -187,7 +188,6 @@ static bool tcg_opt_gen_mov(OptContext *ctx, TCGOp *op,= TCGArg dst, TCGArg src) { TCGTemp *dst_ts =3D arg_temp(dst); TCGTemp *src_ts =3D arg_temp(src); - const TCGOpDef *def; TempOptInfo *di; TempOptInfo *si; uint64_t z_mask; @@ -201,16 +201,24 @@ static bool tcg_opt_gen_mov(OptContext *ctx, TCGOp *o= p, TCGArg dst, TCGArg src) reset_ts(dst_ts); di =3D ts_info(dst_ts); si =3D ts_info(src_ts); - def =3D &tcg_op_defs[op->opc]; - if (def->flags & TCG_OPF_VECTOR) { - new_op =3D INDEX_op_mov_vec; - } else if (def->flags & TCG_OPF_64BIT) { - new_op =3D INDEX_op_mov_i64; - } else { + + switch (ctx->type) { + case TCG_TYPE_I32: new_op =3D INDEX_op_mov_i32; + break; + case TCG_TYPE_I64: + new_op =3D INDEX_op_mov_i64; + break; + case TCG_TYPE_V64: + case TCG_TYPE_V128: + case TCG_TYPE_V256: + /* TCGOP_VECL and TCGOP_VECE remain unchanged. */ + new_op =3D INDEX_op_mov_vec; + break; + default: + g_assert_not_reached(); } op->opc =3D new_op; - /* TCGOP_VECL and TCGOP_VECE remain unchanged. */ op->args[0] =3D dst; op->args[1] =3D src; =20 @@ -237,20 +245,9 @@ static bool tcg_opt_gen_mov(OptContext *ctx, TCGOp *op= , TCGArg dst, TCGArg src) static bool tcg_opt_gen_movi(OptContext *ctx, TCGOp *op, TCGArg dst, uint64_t val) { - const TCGOpDef *def =3D &tcg_op_defs[op->opc]; - TCGType type; - TCGTemp *tv; - - if (def->flags & TCG_OPF_VECTOR) { - type =3D TCGOP_VECL(op) + TCG_TYPE_V64; - } else if (def->flags & TCG_OPF_64BIT) { - type =3D TCG_TYPE_I64; - } else { - type =3D TCG_TYPE_I32; - } - /* Convert movi to mov with constant temp. */ - tv =3D tcg_constant_internal(type, val); + TCGTemp *tv =3D tcg_constant_internal(ctx->type, val); + init_ts_info(ctx, tv); return tcg_opt_gen_mov(ctx, op, dst, temp_arg(tv)); } @@ -420,11 +417,11 @@ static uint64_t do_constant_folding_2(TCGOpcode op, u= int64_t x, uint64_t y) } } =20 -static uint64_t do_constant_folding(TCGOpcode op, uint64_t x, uint64_t y) +static uint64_t do_constant_folding(TCGOpcode op, TCGType type, + uint64_t x, uint64_t y) { - const TCGOpDef *def =3D &tcg_op_defs[op]; uint64_t res =3D do_constant_folding_2(op, x, y); - if (!(def->flags & TCG_OPF_64BIT)) { + if (type =3D=3D TCG_TYPE_I32) { res =3D (int32_t)res; } return res; @@ -510,19 +507,21 @@ static bool do_constant_folding_cond_eq(TCGCond c) * Return -1 if the condition can't be simplified, * and the result of the condition (0 or 1) if it can. */ -static int do_constant_folding_cond(TCGOpcode op, TCGArg x, +static int do_constant_folding_cond(TCGType type, TCGArg x, TCGArg y, TCGCond c) { uint64_t xv =3D arg_info(x)->val; uint64_t yv =3D arg_info(y)->val; =20 if (arg_is_const(x) && arg_is_const(y)) { - const TCGOpDef *def =3D &tcg_op_defs[op]; - tcg_debug_assert(!(def->flags & TCG_OPF_VECTOR)); - if (def->flags & TCG_OPF_64BIT) { - return do_constant_folding_cond_64(xv, yv, c); - } else { + switch (type) { + case TCG_TYPE_I32: return do_constant_folding_cond_32(xv, yv, c); + case TCG_TYPE_I64: + return do_constant_folding_cond_64(xv, yv, c); + default: + /* Only scalar comparisons are optimizable */ + return -1; } } else if (args_are_copies(x, y)) { return do_constant_folding_cond_eq(c); @@ -677,7 +676,7 @@ static bool fold_const1(OptContext *ctx, TCGOp *op) uint64_t t; =20 t =3D arg_info(op->args[1])->val; - t =3D do_constant_folding(op->opc, t, 0); + t =3D do_constant_folding(op->opc, ctx->type, t, 0); return tcg_opt_gen_movi(ctx, op, op->args[0], t); } return false; @@ -689,7 +688,7 @@ static bool fold_const2(OptContext *ctx, TCGOp *op) uint64_t t1 =3D arg_info(op->args[1])->val; uint64_t t2 =3D arg_info(op->args[2])->val; =20 - t1 =3D do_constant_folding(op->opc, t1, t2); + t1 =3D do_constant_folding(op->opc, ctx->type, t1, t2); return tcg_opt_gen_movi(ctx, op, op->args[0], t1); } return false; @@ -791,7 +790,7 @@ static bool fold_andc(OptContext *ctx, TCGOp *op) static bool fold_brcond(OptContext *ctx, TCGOp *op) { TCGCond cond =3D op->args[2]; - int i =3D do_constant_folding_cond(op->opc, op->args[0], op->args[1], = cond); + int i =3D do_constant_folding_cond(ctx->type, op->args[0], op->args[1]= , cond); =20 if (i =3D=3D 0) { tcg_op_remove(ctx->tcg, op); @@ -838,7 +837,7 @@ static bool fold_brcond2(OptContext *ctx, TCGOp *op) * Simplify EQ/NE comparisons where one of the pairs * can be simplified. */ - i =3D do_constant_folding_cond(INDEX_op_brcond_i32, op->args[0], + i =3D do_constant_folding_cond(TCG_TYPE_I32, op->args[0], op->args[2], cond); switch (i ^ inv) { case 0: @@ -847,7 +846,7 @@ static bool fold_brcond2(OptContext *ctx, TCGOp *op) goto do_brcond_high; } =20 - i =3D do_constant_folding_cond(INDEX_op_brcond_i32, op->args[1], + i =3D do_constant_folding_cond(TCG_TYPE_I32, op->args[1], op->args[3], cond); switch (i ^ inv) { case 0: @@ -889,7 +888,7 @@ static bool fold_bswap(OptContext *ctx, TCGOp *op) if (arg_is_const(op->args[1])) { uint64_t t =3D arg_info(op->args[1])->val; =20 - t =3D do_constant_folding(op->opc, t, op->args[2]); + t =3D do_constant_folding(op->opc, ctx->type, t, op->args[2]); return tcg_opt_gen_movi(ctx, op, op->args[0], t); } return false; @@ -933,7 +932,7 @@ static bool fold_count_zeros(OptContext *ctx, TCGOp *op) uint64_t t =3D arg_info(op->args[1])->val; =20 if (t !=3D 0) { - t =3D do_constant_folding(op->opc, t, 0); + t =3D do_constant_folding(op->opc, ctx->type, t, 0); return tcg_opt_gen_movi(ctx, op, op->args[0], t); } return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[2]); @@ -1060,9 +1059,8 @@ static bool fold_mov(OptContext *ctx, TCGOp *op) =20 static bool fold_movcond(OptContext *ctx, TCGOp *op) { - TCGOpcode opc =3D op->opc; TCGCond cond =3D op->args[5]; - int i =3D do_constant_folding_cond(opc, op->args[1], op->args[2], cond= ); + int i =3D do_constant_folding_cond(ctx->type, op->args[1], op->args[2]= , cond); =20 if (i >=3D 0) { return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[4 - i]); @@ -1071,9 +1069,18 @@ static bool fold_movcond(OptContext *ctx, TCGOp *op) if (arg_is_const(op->args[3]) && arg_is_const(op->args[4])) { uint64_t tv =3D arg_info(op->args[3])->val; uint64_t fv =3D arg_info(op->args[4])->val; + TCGOpcode opc; =20 - opc =3D (opc =3D=3D INDEX_op_movcond_i32 - ? INDEX_op_setcond_i32 : INDEX_op_setcond_i64); + switch (ctx->type) { + case TCG_TYPE_I32: + opc =3D INDEX_op_setcond_i32; + break; + case TCG_TYPE_I64: + opc =3D INDEX_op_setcond_i64; + break; + default: + g_assert_not_reached(); + } =20 if (tv =3D=3D 1 && fv =3D=3D 0) { op->opc =3D opc; @@ -1164,7 +1171,7 @@ static bool fold_qemu_st(OptContext *ctx, TCGOp *op) static bool fold_setcond(OptContext *ctx, TCGOp *op) { TCGCond cond =3D op->args[3]; - int i =3D do_constant_folding_cond(op->opc, op->args[1], op->args[2], = cond); + int i =3D do_constant_folding_cond(ctx->type, op->args[1], op->args[2]= , cond); =20 if (i >=3D 0) { return tcg_opt_gen_movi(ctx, op, op->args[0], i); @@ -1203,7 +1210,7 @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op) * Simplify EQ/NE comparisons where one of the pairs * can be simplified. */ - i =3D do_constant_folding_cond(INDEX_op_setcond_i32, op->args[1], + i =3D do_constant_folding_cond(TCG_TYPE_I32, op->args[1], op->args[3], cond); switch (i ^ inv) { case 0: @@ -1212,7 +1219,7 @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op) goto do_setcond_high; } =20 - i =3D do_constant_folding_cond(INDEX_op_setcond_i32, op->args[2], + i =3D do_constant_folding_cond(TCG_TYPE_I32, op->args[2], op->args[4], cond); switch (i ^ inv) { case 0: @@ -1314,6 +1321,15 @@ void tcg_optimize(TCGContext *s) init_arguments(&ctx, op, def->nb_oargs + def->nb_iargs); copy_propagate(&ctx, op, def->nb_oargs, def->nb_iargs); =20 + /* Pre-compute the type of the operation. */ + if (def->flags & TCG_OPF_VECTOR) { + ctx.type =3D TCG_TYPE_V64 + TCGOP_VECL(op); + } else if (def->flags & TCG_OPF_64BIT) { + ctx.type =3D TCG_TYPE_I64; + } else { + ctx.type =3D TCG_TYPE_I32; + } + /* For commutative operations make constant second argument */ switch (opc) { CASE_OP_32_64_VEC(add): @@ -1394,18 +1410,18 @@ void tcg_optimize(TCGContext *s) /* Proceed with possible constant folding. */ break; } - if (opc =3D=3D INDEX_op_sub_i32) { + switch (ctx.type) { + case TCG_TYPE_I32: neg_op =3D INDEX_op_neg_i32; have_neg =3D TCG_TARGET_HAS_neg_i32; - } else if (opc =3D=3D INDEX_op_sub_i64) { + break; + case TCG_TYPE_I64: neg_op =3D INDEX_op_neg_i64; have_neg =3D TCG_TARGET_HAS_neg_i64; - } else if (TCG_TARGET_HAS_neg_vec) { - TCGType type =3D TCGOP_VECL(op) + TCG_TYPE_V64; - unsigned vece =3D TCGOP_VECE(op); + break; + default: neg_op =3D INDEX_op_neg_vec; - have_neg =3D tcg_can_emit_vec_op(neg_op, type, vece) >= 0; - } else { + have_neg =3D tcg_can_emit_vec_op(neg_op, ctx.type, TCG= OP_VECE(op)) > 0; break; } if (!have_neg) { @@ -1459,15 +1475,19 @@ void tcg_optimize(TCGContext *s) TCGOpcode not_op; bool have_not; =20 - if (def->flags & TCG_OPF_VECTOR) { + switch (ctx.type) { + default: not_op =3D INDEX_op_not_vec; have_not =3D TCG_TARGET_HAS_not_vec; - } else if (def->flags & TCG_OPF_64BIT) { + break; + case TCG_TYPE_I64: not_op =3D INDEX_op_not_i64; have_not =3D TCG_TARGET_HAS_not_i64; - } else { + break; + case TCG_TYPE_I32: not_op =3D INDEX_op_not_i32; have_not =3D TCG_TARGET_HAS_not_i32; + break; } if (!have_not) { break; @@ -1738,7 +1758,7 @@ void tcg_optimize(TCGContext *s) below, we can ignore high bits, but for further optimizations we need to record that the high bits contain garbage. */ partmask =3D z_mask; - if (!(def->flags & TCG_OPF_64BIT)) { + if (ctx.type =3D=3D TCG_TYPE_I32) { z_mask |=3D ~(tcg_target_ulong)0xffffffffu; partmask &=3D 0xffffffffu; affected &=3D 0xffffffffu; --=20 2.25.1