From patchwork Thu Jul 25 03:18:48 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jiufu Guo X-Patchwork-Id: 1964586 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@legolas.ozlabs.org Authentication-Results: legolas.ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=ibm.com header.i=@ibm.com header.a=rsa-sha256 header.s=pp1 header.b=kiAYZv6O; dkim-atps=neutral Authentication-Results: legolas.ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=gcc.gnu.org (client-ip=2620:52:3:1:0:246e:9693:128c; helo=server2.sourceware.org; envelope-from=gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org; receiver=patchwork.ozlabs.org) Received: from server2.sourceware.org (server2.sourceware.org [IPv6:2620:52:3:1:0:246e:9693:128c]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (secp384r1) server-digest SHA384) (No client certificate requested) by legolas.ozlabs.org (Postfix) with ESMTPS id 4WTx29543yz1ybY for ; Thu, 25 Jul 2024 13:19:28 +1000 (AEST) Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 668E63858430 for ; Thu, 25 Jul 2024 03:19:26 +0000 (GMT) X-Original-To: gcc-patches@gcc.gnu.org Delivered-To: gcc-patches@gcc.gnu.org Received: from mx0b-001b2d01.pphosted.com (mx0b-001b2d01.pphosted.com [148.163.158.5]) by sourceware.org (Postfix) with ESMTPS id C4A713858D29; Thu, 25 Jul 2024 03:19:01 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org C4A713858D29 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=linux.ibm.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=linux.ibm.com ARC-Filter: OpenARC Filter v1.0.0 sourceware.org C4A713858D29 Authentication-Results: server2.sourceware.org; arc=none smtp.remote-ip=148.163.158.5 ARC-Seal: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1721877546; cv=none; b=qnZXqf2gJIQelXoMXxeMxYNoR2p7ZiHNwRTSYOFo4aPjTxPyY0SQo+hmC2zmO26rpIZAZqdSv4LT0clTuOtPvE0E5a5EWJkAs/FftgyOBTS+7vLP018dK6brQ7For1/sx0gV9uOd3Gqy86wJWJNBvjUqMi8Y4SENGg22ADi2tt4= ARC-Message-Signature: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1721877546; c=relaxed/simple; bh=909hPkretuHvgVJuVt4xbSHZyIq6KzGwXIsl+PMpGw8=; h=DKIM-Signature:From:To:Subject:Date:Message-ID:MIME-Version; b=CsbuiCqHtaTDxGZreCplOdn37gGYvTCvMxhCyoQ3e4pdrGl/ZJxMg0CXwhvxJOhE4Iw4XELj85OnVzmMpeFFLSt6Geg1c1oLFkQidd2GhpAoFJr/sill/J3TG2Wl/H9VpZfeIAcx7IHL+oYBxZaHB7kUi4nE9oTWHkoOGyka1xQ= ARC-Authentication-Results: i=1; server2.sourceware.org Received: from pps.filterd (m0353722.ppops.net [127.0.0.1]) by mx0a-001b2d01.pphosted.com (8.18.1.2/8.18.1.2) with ESMTP id 46P20C2D017764; Thu, 25 Jul 2024 03:18:59 GMT DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=ibm.com; h=from :to:cc:subject:date:message-id:content-transfer-encoding :mime-version; s=pp1; bh=HTObxkqaraX6Vx7RSTxnUChPvb1DIjXyeUvrZqv DcJk=; b=kiAYZv6OXWL3EmBmhpK+umiKPqYCNj18VsuDTiBPZ4fEWQ7tkpt62SG FvDDDDhof6HULU3BFbZhInsRVJsKqR+MvgWnfu9YgWSbp3fIWaWHZNbs3D3JSbVz 5azFcs6P+oa4aKf6W6dp4Xsu/j4hLUUvEL1lloVgAT2sRSGcOvE+d5yTE/HguKn0 Jpal3G8eA+j+lxJmwX8FKrGmuSTlVkw/vw8lHt0Xqme9CQzdbt5/PV/fkIFwqlme LVOIrIf0L25zm1BR0trXkg5qrmsZcnAR9lfj71ZwrqI2FayAlgk0fW8iTXTcvZ3Q YlqCKYPoouHNsEw4/hwzspicZuVw+Qw== Received: from pps.reinject (localhost [127.0.0.1]) by mx0a-001b2d01.pphosted.com (PPS) with ESMTPS id 40kc6qrad1-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Thu, 25 Jul 2024 03:18:58 +0000 (GMT) Received: from m0353722.ppops.net (m0353722.ppops.net [127.0.0.1]) by pps.reinject (8.18.0.8/8.18.0.8) with ESMTP id 46P3Iwtr023265; Thu, 25 Jul 2024 03:18:58 GMT Received: from ppma22.wdc07v.mail.ibm.com (5c.69.3da9.ip4.static.sl-reverse.com [169.61.105.92]) by mx0a-001b2d01.pphosted.com (PPS) with ESMTPS id 40kc6qracy-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Thu, 25 Jul 2024 03:18:58 +0000 (GMT) Received: from pps.filterd (ppma22.wdc07v.mail.ibm.com [127.0.0.1]) by ppma22.wdc07v.mail.ibm.com (8.17.1.19/8.17.1.19) with ESMTP id 46P38QpB007929; Thu, 25 Jul 2024 03:18:57 GMT Received: from smtprelay02.fra02v.mail.ibm.com ([9.218.2.226]) by ppma22.wdc07v.mail.ibm.com (PPS) with ESMTPS id 40gxna3hc5-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Thu, 25 Jul 2024 03:18:57 +0000 Received: from smtpav05.fra02v.mail.ibm.com (smtpav05.fra02v.mail.ibm.com [10.20.54.104]) by smtprelay02.fra02v.mail.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id 46P3IpTP55705868 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Thu, 25 Jul 2024 03:18:53 GMT Received: from smtpav05.fra02v.mail.ibm.com (unknown [127.0.0.1]) by IMSVA (Postfix) with ESMTP id 4B87B20040; Thu, 25 Jul 2024 03:18:51 +0000 (GMT) Received: from smtpav05.fra02v.mail.ibm.com (unknown [127.0.0.1]) by IMSVA (Postfix) with ESMTP id 7758E20043; Thu, 25 Jul 2024 03:18:49 +0000 (GMT) Received: from ltcden2-lp1.aus.stglabs.ibm.com (unknown [9.3.90.43]) by smtpav05.fra02v.mail.ibm.com (Postfix) with ESMTP; Thu, 25 Jul 2024 03:18:49 +0000 (GMT) From: Jiufu Guo To: gcc-patches@gcc.gnu.org Cc: rguenther@suse.de, jeffreyalaw@gmail.com, richard.sandiford@arm.com, pinskia@gcc.gnu.org, segher@kernel.crashing.org, linkw@gcc.gnu.org, dje.gcc@gmail.com, bergner@linux.ibm.com, guojiufu@linux.ibm.com Subject: [PATCH V3] fsra: gimple final sra pass for paramters and returns Date: Thu, 25 Jul 2024 11:18:48 +0800 Message-ID: <20240725031848.1792449-1-guojiufu@linux.ibm.com> X-Mailer: git-send-email 2.43.5 X-TM-AS-GCONF: 00 X-Proofpoint-GUID: rN_qFdHPQ6qetjS_-dl1rhdbLTT87438 X-Proofpoint-ORIG-GUID: 16ONVwXPcLb-iVPa4KmZscZKZTxFBGUQ X-Proofpoint-UnRewURL: 0 URL was un-rewritten MIME-Version: 1.0 X-Proofpoint-Virus-Version: vendor=baseguard engine=ICAP:2.0.293,Aquarius:18.0.1039,Hydra:6.0.680,FMLib:17.12.28.16 definitions=2024-07-25_03,2024-07-24_01,2024-05-17_01 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 lowpriorityscore=0 bulkscore=0 impostorscore=0 adultscore=0 suspectscore=0 spamscore=0 malwarescore=0 priorityscore=1501 phishscore=0 clxscore=1011 mlxscore=0 mlxlogscore=999 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.19.0-2407110000 definitions=main-2407250015 X-Spam-Status: No, score=-10.7 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_EF, GIT_PATCH_0, KAM_SHORT, RCVD_IN_MSPIKE_H3, RCVD_IN_MSPIKE_WL, SPF_HELO_NONE, SPF_PASS, TXREP autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: gcc-patches@gcc.gnu.org X-Mailman-Version: 2.1.30 Precedence: list List-Id: Gcc-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org Hi, There are a few PRs (meta-bug PR101926) about accessing aggregate parameters/returns which are passed through registers. A major reason of those issues is when access the aggregate, the temporary stack slots are used without leverage the information about the incoming/outgoing registers. We could use the current SRA pass in a special mode right before GIMPLE->RTL expansion for the parameters/returns, and scalarize the access according to the incoming/outgoing registers. Some discussion in: https://gcc.gnu.org/pipermail/gcc-patches/2023-November/637935.html This patch adds a FINAL mode for tree-sra; and introduces IFN ARG_PARTS /SET_RET_PARTS for scalar(s) access on parameters/returns. And expand the IFNs according to the incoming/outgoing registers. Compare with previous version: https://gcc.gnu.org/pipermail/gcc-patches/2024-July/656359.html This version: * Support to cover mutli-registers in one access * Support more bitfield access * Support aggregate assignment to 'return' * Support aggregate assignment from 'parameter' or 'sub-parameter' * Remove SET_RET_LAST_PARTS, only keep SET_RET_PARTS * Accept caller's parameters on callee's arguments. * Refactor the code, e.g. style, comments. Again there would be a few thing could be enhanced. * More optimization for access parameter which in memory * Enhancemore parameter accesses, e.g. grp_write. * More cases/targets checking Bootstrapped/regtested on ppc64{,le}, x86_64. With known behavior changes in pr88873.c, pr101908-3.c and bfxil_1.c(aarch64). Is this ok for trunk? BR, Jeff (Jiufu Guo) PR target/108073 PR target/65421 PR target/69143 gcc/ChangeLog: * calls.cc (precompute_register_parameters): Prepare callee argument from caller parameter. * cfgexpand.cc (expand_value_return): Update 'rtx eq' checking. (expand_return): Checking sclarized returns. * internal-fn.cc (query_position_in_series): New function. (assign_from_regs): New function. (reference_alias_ptr_type): Extern declare. (expand_ARG_PARTS): New IFN expand. (assign_to_regs): New function. (expand_SET_RET_PARTS): New IFN expand. * internal-fn.def (ARG_PARTS): New IFN (SET_RET_PARTS): New IFN * passes.def (pass_sra_final): Add new pass. * tree-pass.h (make_pass_sra_final): New function. * tree-sra.cc (enum sra_mode): New enum item SRA_MODE_FINAL_INTRA. (build_access_from_expr_1): Check TARGET_MEM_REF. (build_accesses_from_assign): Accept SRA_MODE_FINAL_INTRA. (scan_function): Check SRA_MODE_FINAL_INTRA. (find_var_candidates): Add condidates for fsra. (fsra_analyze): New function. (analyze_access_subtree): Check fsra_analyze. (generate_subtree_copies): Generate IFNs for fsra. (final_intra_sra): New function. (class pass_sra_final): New pass class. (make_pass_sra_final): New function. gcc/testsuite/ChangeLog: * g++.target/powerpc/pr102024.C: Update instructions checking. * gcc.target/powerpc/pr108073-1.c: New test. * gcc.target/powerpc/pr108073.c: New test. * gcc.target/powerpc/pr65421.c: New test. * gcc.target/powerpc/pr69143.c: New test. --- gcc/calls.cc | 9 + gcc/cfgexpand.cc | 6 +- gcc/internal-fn.cc | 413 ++++++++++++++++++ gcc/internal-fn.def | 6 + gcc/passes.def | 1 + gcc/tree-pass.h | 1 + gcc/tree-sra.cc | 170 ++++++- gcc/testsuite/g++.target/powerpc/pr102024.C | 3 +- gcc/testsuite/gcc.target/powerpc/pr108073-1.c | 76 ++++ gcc/testsuite/gcc.target/powerpc/pr108073.c | 74 ++++ gcc/testsuite/gcc.target/powerpc/pr65421.c | 26 ++ gcc/testsuite/gcc.target/powerpc/pr69143.c | 22 + 12 files changed, 795 insertions(+), 12 deletions(-) create mode 100644 gcc/testsuite/gcc.target/powerpc/pr108073-1.c create mode 100644 gcc/testsuite/gcc.target/powerpc/pr108073.c create mode 100644 gcc/testsuite/gcc.target/powerpc/pr65421.c create mode 100644 gcc/testsuite/gcc.target/powerpc/pr69143.c diff --git a/gcc/calls.cc b/gcc/calls.cc index 21d78f9779f..9035b12f429 100644 --- a/gcc/calls.cc +++ b/gcc/calls.cc @@ -996,6 +996,15 @@ precompute_register_parameters (int num_actuals, struct arg_data *args, pop_temp_slots (); } + if (GET_CODE (args[i].value) == PARALLEL) + { + tree type = TREE_TYPE (args[i].tree_value); + int size = int_size_in_bytes (type); + rtx mem = assign_stack_temp (BLKmode, size); + emit_group_store (mem, args[i].value, type, size); + args[i].value = mem; + } + /* If we are to promote the function arg to a wider mode, do it now. */ diff --git a/gcc/cfgexpand.cc b/gcc/cfgexpand.cc index dad3ae1b7c6..2940431cb16 100644 --- a/gcc/cfgexpand.cc +++ b/gcc/cfgexpand.cc @@ -3789,7 +3789,7 @@ expand_value_return (rtx val) tree decl = DECL_RESULT (current_function_decl); rtx return_reg = DECL_RTL (decl); - if (return_reg != val) + if (!rtx_equal_p (return_reg, val)) { tree funtype = TREE_TYPE (current_function_decl); tree type = TREE_TYPE (decl); @@ -3862,6 +3862,10 @@ expand_return (tree retval) been stored into it, so we don't have to do anything special. */ if (TREE_CODE (retval_rhs) == RESULT_DECL) expand_value_return (result_rtl); + /* return is scalarized by fsra. */ + else if (DECL_RTL_SET_P (retval_rhs) + && rtx_equal_p (result_rtl, DECL_RTL (retval_rhs))) + expand_null_return_1 (); /* If the result is an aggregate that is being returned in one (or more) registers, load the registers here. */ diff --git a/gcc/internal-fn.cc b/gcc/internal-fn.cc index 95946bfd683..78d9d7d557a 100644 --- a/gcc/internal-fn.cc +++ b/gcc/internal-fn.cc @@ -3483,6 +3483,419 @@ expand_ACCESS_WITH_SIZE (internal_fn, gcall *stmt) expand_assignment (lhs, ref_to_obj, false); } +/* From the registers series REGS, compute which registers are touched + at the position {BITPOS, BITSIZE}. The results are stored into + START_INDEX, END_INDEX, LEFT_BITS and RIGHT_BITS. + Return the first register at the expected position. */ + +static rtx +query_position_in_series (rtx regs, HOST_WIDE_INT bitpos, HOST_WIDE_INT bitsize, + int &start_index, int &end_index, + HOST_WIDE_INT &left_bits, HOST_WIDE_INT &right_bits) +{ + if (GET_CODE (regs) == PARALLEL) + { + int cur_index = XEXP (XVECEXP (regs, 0, 0), 0) ? 0 : 1; + for (; cur_index < XVECLEN (regs, 0); cur_index++) + { + rtx slot = XVECEXP (regs, 0, cur_index); + HOST_WIDE_INT off = UINTVAL (XEXP (slot, 1)) * BITS_PER_UNIT; + machine_mode mode = GET_MODE (XEXP (slot, 0)); + HOST_WIDE_INT size = GET_MODE_BITSIZE (mode).to_constant (); + if (off <= bitpos && off + size > bitpos) + { + start_index = cur_index; + left_bits = bitpos - off; + } + if (off + size >= bitpos + bitsize) + { + end_index = cur_index; + right_bits = off + size - (bitpos + bitsize); + break; + } + } + /* Does not hit the incoming reg, e.g. accessing the padding. */ + if (start_index < 0 || end_index < start_index) + return NULL_RTX; + return XEXP (XVECEXP (regs, 0, start_index), 0); + } + + if (REG_P (regs) && GET_MODE (regs) == BLKmode) + { + HOST_WIDE_INT end_bits = bitpos + bitsize - 1; + start_index = bitpos / BITS_PER_WORD; + left_bits = bitpos % BITS_PER_WORD; + end_index = end_bits / BITS_PER_WORD; + right_bits = BITS_PER_WORD - 1 - (end_bits % BITS_PER_WORD); + return gen_rtx_REG (word_mode, REGNO (regs) + start_index); + } + + /* Complex concat. */ + if (GET_CODE (regs) == CONCAT && COMPLEX_MODE_P (GET_MODE (regs))) + { + machine_mode mode = GET_MODE (XEXP (regs, 0)); + HOST_WIDE_INT mode_size = GET_MODE_BITSIZE (mode).to_constant (); + if ((bitsize == mode_size && (bitpos == 0 || bitpos == mode_size)) + || (bitsize == mode_size * 2 && bitpos == 0)) + { + start_index = end_index = bitpos == 0 ? 0 : 1; + left_bits = right_bits = 0; + if (bitsize == mode_size) + return XEXP (regs, start_index); + + /* bitsize == mode_size * 2 */ + rtx reg = gen_reg_rtx (GET_MODE (regs)); + emit_move_insn (reg, regs); + return reg; + } + + /* Take middle bits of complex. */ + return NULL_RTX; + } + + if (REG_P (regs)) + { + start_index = end_index = 0; + left_bits = bitpos; + machine_mode mode = GET_MODE (regs); + right_bits = GET_MODE_BITSIZE (mode).to_constant () - bitpos - bitsize; + return regs; + } + + return NULL_RTX; +} + +/* For an access on ARG at {BITPOS, BITSIZE}, compute a RTX + expression for the access and assign it to LHS. + ARG is an aggregate parameter of a function, and it should + be passed through registers. + Return true if able to figure out an effective way to assign + the access to LHS, otherwise return false. */ + +static bool +assign_from_regs (tree arg, HOST_WIDE_INT bitpos, HOST_WIDE_INT bitsize, + bool reversep, tree lhs) +{ + rtx regs = DECL_INCOMING_RTL (arg); + if (MEM_P (regs)) + return false; + int start_index = -1; + int end_index = -1; + HOST_WIDE_INT left_bits = 0; + HOST_WIDE_INT right_bits = 0; + + /* Prefer pseudo on DECL_RTL instead of single hard register. */ + if (REG_P (regs)) + { + rtx arg_rtx = arg->decl_with_rtl.rtl; + if (GET_MODE (regs) != BLKmode) + regs = arg_rtx; + else if (int_size_in_bytes (TREE_TYPE (arg)) < UNITS_PER_WORD) + { + regs = gen_reg_rtx (word_mode); + emit_move_insn (regs, adjust_address (arg_rtx, word_mode, 0)); + } + } + rtx reg1st = query_position_in_series (regs, bitpos, bitsize, start_index, + end_index, left_bits, right_bits); + if (!reg1st) + return false; + machine_mode mode = word_mode; + + tree type = TREE_TYPE (lhs); + machine_mode expr_mode = TYPE_MODE (type); + + /* Just need one reg for the access. */ + if (end_index == start_index) + { + rtx reg = reg1st; + /* Perfer pesudo */ + if (HARD_REGISTER_P (reg)) + { + rtx tmp = gen_reg_rtx (GET_MODE (reg)); + emit_move_insn (tmp, reg); + reg = tmp; + } + + if (left_bits == 0 && right_bits == 0) + { + if (reversep) + reg = flip_storage_order (GET_MODE (reg), reg); + if (GET_MODE (reg) != expr_mode) + reg = gen_lowpart (expr_mode, reg); + + rtx dest = expand_expr (lhs, NULL_RTX, VOIDmode, EXPAND_WRITE); + emit_move_insn (dest, reg); + return true; + } + + /* Need to extract bitfield part reg for the access. + left_bits != 0 or right_bits != 0 */ + scalar_int_mode imode; + if (!int_mode_for_mode (expr_mode).exists (&imode)) + return false; + + bool sgn = TYPE_UNSIGNED (type); + rtx res = extract_bit_field (reg, bitsize, left_bits, sgn, NULL_RTX, + imode, imode, reversep, NULL); + + if (GET_MODE (res) != imode) + res = gen_lowpart (imode, res); + + if (expr_mode != imode) + { + /* expr_mode != imode, e.g. SF != SI. */ + rtx tmp = gen_reg_rtx (imode); + emit_move_insn (tmp, res); + res = gen_lowpart (expr_mode, tmp); + } + + rtx dest = expand_expr (lhs, NULL_RTX, VOIDmode, EXPAND_WRITE); + emit_move_insn (dest, res); + return true; + } + + /* Access mult-registers. */ + if (left_bits != 0 || right_bits != 0 || bitsize % BITS_PER_WORD) + return false; + + int len = end_index - start_index + 1; + if (!VECTOR_MODE_P (expr_mode)) + { + rtvec vec = rtvec_alloc (len); + int off0 = 0; + if (GET_CODE (regs) == PARALLEL) + off0 = INTVAL (XEXP (XVECEXP (regs, 0, start_index), 1)); + for (int i = 0; i < len; i++) + { + int index = start_index + i; + rtx src; + rtx off; + if (REG_P (regs)) + { + src = gen_rtx_REG (mode, REGNO (regs) + index); + off = GEN_INT (UNITS_PER_WORD * i); + } + else + { + rtx e = XVECEXP (regs, 0, index); + src = XEXP (e, 0); + off = GEN_INT (INTVAL (XEXP (e, 1)) - off0); + } + rtx tmp = gen_reg_rtx (GET_MODE (src)); + emit_move_insn (tmp, src); + RTVEC_ELT (vec, i) = alloc_EXPR_LIST (REG_DEP_TRUE, tmp, off); + } + rtx res = gen_rtx_PARALLEL (expr_mode, vec); + if (expr_mode == BLKmode) + lhs->decl_with_rtl.rtl = res; // SET_DECL_RTL + else + { + rtx dest = expand_expr (lhs, NULL_RTX, VOIDmode, EXPAND_WRITE); + emit_group_store (dest, res, type, bitsize / BITS_PER_UNIT); + } + return true; + } + + /* Check vector mode and element mode. */ + if (GET_CODE (regs) == PARALLEL) + mode = GET_MODE (XEXP (XVECEXP (regs, 0, start_index), 0)); + scalar_mode smode = as_a (mode); + machine_mode vmode; + if (!related_vector_mode (expr_mode, smode, len).exists (&vmode)) + return false; + insn_code icode = convert_optab_handler (vec_init_optab, vmode, smode); + if (icode == CODE_FOR_nothing) + return false; + + rtvec vec = rtvec_alloc (len); + for (int i = 0; i < len; i++) + { + int index = start_index + i; + rtx src = REG_P (regs) ? gen_rtx_REG (mode, REGNO (regs) + index) + : XEXP (XVECEXP (regs, 0, index), 0); + RTVEC_ELT (vec, i) = src; + } + + rtx res = gen_reg_rtx (vmode); + emit_insn (GEN_FCN (icode) (res, gen_rtx_PARALLEL (vmode, vec))); + if (expr_mode != vmode) + res = gen_lowpart (expr_mode, res); + rtx dest = expand_expr (lhs, NULL_RTX, VOIDmode, EXPAND_WRITE); + emit_move_insn (dest, res); + return true; +} + +/* Extern function for building MEM_REF rtx. */ +tree +reference_alias_ptr_type (tree t); + +/* Expand the IFN_ARG_PARTS function: + LHS = .ARG_PARTS(INCOMING_ARG, BIT_OFFSET, BIT_SIZE, REVERSEP). */ + +static void +expand_ARG_PARTS (internal_fn, gcall *stmt) +{ + tree lhs = gimple_call_lhs (stmt); + tree arg = gimple_call_arg (stmt, 0); + HOST_WIDE_INT offset = tree_to_shwi (gimple_call_arg (stmt, 1)); + HOST_WIDE_INT size = tree_to_shwi (gimple_call_arg (stmt, 2)); + int reversep = tree_to_shwi (gimple_call_arg (stmt, 3)); + if (assign_from_regs (arg, offset, size, reversep, lhs)) + return; + + /* Access bitfiled. */ + tree type = TREE_TYPE (lhs); + if ((INTEGRAL_TYPE_P (type) && !type_has_mode_precision_p (type)) + || offset % BITS_PER_UNIT != 0 || size % BITS_PER_UNIT != 0) + { + machine_mode mode = TYPE_MODE (type); + rtx src + = expand_expr_real (arg, NULL, VOIDmode, EXPAND_NORMAL, NULL, true); + src = extract_bit_field (src, size, offset, TYPE_UNSIGNED (type), NULL, + mode, mode, reversep, NULL); + rtx dest = expand_expr (lhs, NULL, VOIDmode, EXPAND_WRITE); + if (GET_CODE (dest) == SUBREG && SUBREG_PROMOTED_VAR_P (dest)) + convert_move (SUBREG_REG (dest), src, SUBREG_PROMOTED_SIGN (dest)); + else + emit_move_insn (dest, src); + return; + } + + /* Fall to original expand. */ + gcc_assert (offset % BITS_PER_UNIT == 0 && size % BITS_PER_UNIT == 0); + tree base = build_fold_addr_expr (arg); + tree atype = reference_alias_ptr_type (arg); + tree off = build_int_cst (atype, offset / BITS_PER_UNIT); + location_t loc = EXPR_LOCATION (arg); + tree rhs = fold_build2_loc (loc, MEM_REF, type, base, off); + REF_REVERSE_STORAGE_ORDER (rhs) = reversep; + expand_assignment (lhs, rhs, false); +} + +/* REGS constains the function return registers, compute which register(s) + are access via. {BITPOS, BITSIZE}. REV indicates the aggrgates's storeage + order. + If the accessed register(s) are whole register(s) or lowpart of a + register, then assign rhs to it(them) and return true. + Otherwise return NULL_RTX. */ + +static bool +assign_to_regs (rtx regs, HOST_WIDE_INT bitpos, HOST_WIDE_INT bitsize, bool rev, + tree rhs) +{ + auto store_into_reg + = [rev] (rtx dest, HOST_WIDE_INT size, HOST_WIDE_INT pos, rtx src) -> bool { + machine_mode mode = GET_MODE (dest); + if (known_eq (size, GET_MODE_BITSIZE (mode))) + emit_move_insn (dest, gen_lowpart (mode, src)); + else + store_bit_field (dest, size, pos, 0, 0, mode, src, rev, false); + return true; + }; + + rtx src = expand_expr (rhs, NULL_RTX, VOIDmode, EXPAND_NORMAL); + gcc_assert (src); + if (REG_P (regs)) + return store_into_reg (regs, bitsize, bitpos, src); + + gcc_assert (GET_CODE (regs) == PARALLEL); + + int start_index = -1; + int end_index = -1; + HOST_WIDE_INT left_bits = 0; + HOST_WIDE_INT right_bits = 0; + rtx reg1st = query_position_in_series (regs, bitpos, bitsize, start_index, + end_index, left_bits, right_bits); + + gcc_assert (reg1st); + + if (end_index == start_index) + return store_into_reg (reg1st, bitsize, left_bits, src); + + /* Multi-registers */ + machine_mode src_mode = GET_MODE (src); + if (src_mode != BLKmode && !REG_P (src)) + { + /* To extract bitfield, convert src. */ + poly_int64 size = GET_MODE_SIZE (src_mode); + rtx mem = assign_stack_temp (src_mode, size); + emit_move_insn (mem, src); + src = mem; + } + + int off0 = INTVAL (XEXP (XVECEXP (regs, 0, start_index), 1)) * BITS_PER_UNIT; + if (left_bits != 0) + { + rtx reg = reg1st; + machine_mode mode = GET_MODE (reg); + int size = GET_MODE_BITSIZE (mode).to_constant (); + size -= left_bits; + rtx elt + = extract_bit_field (src, size, 0, 1, NULL, mode, mode, false, NULL); + store_into_reg (reg, size, left_bits, elt); + + off0 += left_bits; + start_index++; + } + + if (right_bits != 0) + { + rtx e = XVECEXP (regs, 0, end_index); + rtx reg = XEXP (e, 0); + machine_mode mode = GET_MODE (reg); + int size = GET_MODE_BITSIZE (mode).to_constant (); + size -= right_bits; + int off = INTVAL (XEXP (e, 1)) * BITS_PER_UNIT - off0; + rtx elt + = extract_bit_field (src, size, off, 1, NULL, mode, mode, false, NULL); + store_into_reg (reg, size, 0, elt); + + end_index--; + } + + for (int i = start_index; i <= end_index; i++) + { + rtx e = XVECEXP (regs, 0, i); + rtx reg = XEXP (e, 0); + int off = INTVAL (XEXP (e, 1)) * BITS_PER_UNIT - off0; + machine_mode mode = GET_MODE (reg); + int size = GET_MODE_BITSIZE (mode).to_constant (); + rtx elt + = extract_bit_field (src, size, off, 1, NULL, mode, mode, false, NULL); + emit_move_insn (reg, elt); + } + return true; +} + +/* Expand the IFN_SET_RET_PARTS function: + .SET_REG_PARTS(RET_BASE, BIT_OFFSET, BIT_SIZE, SRC). + e.g. .SET_RET_PARTS (D.2774, 0, 8, SRC); return D.2774; */ + +static void +expand_SET_RET_PARTS (internal_fn, gcall *stmt) +{ + HOST_WIDE_INT offset = tree_to_shwi (gimple_call_arg (stmt, 1)); + HOST_WIDE_INT size = tree_to_shwi (gimple_call_arg (stmt, 2)); + tree rhs = gimple_call_arg (stmt, 3); + + tree decl = DECL_RESULT (current_function_decl); + rtx dest_regs = decl->decl_with_rtl.rtl; // DECL_RTL (base); + if (GET_CODE (dest_regs) != PARALLEL + && (REG_P (dest_regs) ? REGNO (dest_regs) >= FIRST_PSEUDO_REGISTER + : DECL_REGISTER (decl)) + && GET_CODE (crtl->return_rtx) == PARALLEL) + { + dest_regs = crtl->return_rtx; + decl->decl_with_rtl.rtl = dest_regs; + } + tree base = gimple_call_arg (stmt, 0); + base->decl_with_rtl.rtl = dest_regs; // SET_DECL_RTL + + bool rev = TYPE_REVERSE_STORAGE_ORDER (TREE_TYPE (decl)); + assign_to_regs (dest_regs, offset, size, rev, rhs); +} + /* The size of an OpenACC compute dimension. */ static void diff --git a/gcc/internal-fn.def b/gcc/internal-fn.def index 0b45f322f0d..555ac9576d8 100644 --- a/gcc/internal-fn.def +++ b/gcc/internal-fn.def @@ -520,6 +520,12 @@ DEF_INTERNAL_FN (DEFERRED_INIT, ECF_CONST | ECF_LEAF | ECF_NOTHROW, NULL) 2nd argument. */ DEF_INTERNAL_FN (ACCESS_WITH_SIZE, ECF_PURE | ECF_LEAF | ECF_NOTHROW, NULL) +/* A function to extract elemet(s) from an aggregate argument in fsra. */ +DEF_INTERNAL_FN (ARG_PARTS, ECF_CONST | ECF_LEAF | ECF_NOTHROW, NULL) + +/* Functions to set/construct elemet(s) for an 'return' aggregate. */ +DEF_INTERNAL_FN (SET_RET_PARTS, ECF_LEAF | ECF_NOTHROW, NULL) + /* DIM_SIZE and DIM_POS return the size of a particular compute dimension and the executing thread's position within that dimension. DIM_POS is pure (and not const) so that it isn't diff --git a/gcc/passes.def b/gcc/passes.def index 94386ba5457..d2a6b421418 100644 --- a/gcc/passes.def +++ b/gcc/passes.def @@ -450,6 +450,7 @@ along with GCC; see the file COPYING3. If not see NEXT_PASS (pass_harden_conditional_branches); NEXT_PASS (pass_harden_compares); NEXT_PASS (pass_warn_access, /*early=*/false); + NEXT_PASS (pass_sra_final); NEXT_PASS (pass_cleanup_cfg_post_optimizing); NEXT_PASS (pass_warn_function_noreturn); diff --git a/gcc/tree-pass.h b/gcc/tree-pass.h index 84bc91b51e9..58ee0333594 100644 --- a/gcc/tree-pass.h +++ b/gcc/tree-pass.h @@ -366,6 +366,7 @@ extern gimple_opt_pass *make_pass_early_tree_profile (gcc::context *ctxt); extern gimple_opt_pass *make_pass_cleanup_eh (gcc::context *ctxt); extern gimple_opt_pass *make_pass_sra (gcc::context *ctxt); extern gimple_opt_pass *make_pass_sra_early (gcc::context *ctxt); +extern gimple_opt_pass *make_pass_sra_final (gcc::context *ctxt); extern gimple_opt_pass *make_pass_tail_recursion (gcc::context *ctxt); extern gimple_opt_pass *make_pass_tail_calls (gcc::context *ctxt); extern gimple_opt_pass *make_pass_fix_loops (gcc::context *ctxt); diff --git a/gcc/tree-sra.cc b/gcc/tree-sra.cc index 8040b0c5645..132304428b9 100644 --- a/gcc/tree-sra.cc +++ b/gcc/tree-sra.cc @@ -21,14 +21,16 @@ along with GCC; see the file COPYING3. If not see . */ /* This file implements Scalar Reduction of Aggregates (SRA). SRA is run - twice, once in the early stages of compilation (early SRA) and once in the - late stages (late SRA). The aim of both is to turn references to scalar - parts of aggregates into uses of independent scalar variables. + three times, once in the early stages of compilation (early SRA) and once + in the late stages (late SRA). The aim of them is to turn references to + scalar parts of aggregates into uses of independent scalar variables. - The two passes are nearly identical, the only difference is that early SRA + The three passes are nearly identical, the difference are that early SRA does not scalarize unions which are used as the result in a GIMPLE_RETURN statement because together with inlining this can lead to weird type - conversions. + conversions. The third pass is more care about parameters and returns, + it would be helpful for the parameters and returns which are passed through + registers. Both passes operate in four stages: @@ -104,6 +106,7 @@ along with GCC; see the file COPYING3. If not see /* Enumeration of all aggregate reductions we can do. */ enum sra_mode { SRA_MODE_EARLY_IPA, /* early call regularization */ SRA_MODE_EARLY_INTRA, /* early intraprocedural SRA */ + SRA_MODE_FINAL_INTRA, /* final gimple intraprocedural SRA */ SRA_MODE_INTRA }; /* late intraprocedural SRA */ /* Global variable describing which aggregate reduction we are performing at @@ -1269,6 +1272,12 @@ build_access_from_expr_1 (tree expr, gimple *stmt, bool write) return NULL; } + if (TREE_CODE (expr) == TARGET_MEM_REF) + { + disqualify_base_of_expr (expr, "variable offset on base."); + return NULL; + } + struct access *ret = NULL; bool partial_ref; @@ -1549,7 +1558,8 @@ build_accesses_from_assign (gimple *stmt) } if (lacc && racc - && (sra_mode == SRA_MODE_EARLY_INTRA || sra_mode == SRA_MODE_INTRA) + && (sra_mode == SRA_MODE_EARLY_INTRA || sra_mode == SRA_MODE_INTRA + || sra_mode == SRA_MODE_FINAL_INTRA) && !lacc->grp_unscalarizable_region && !racc->grp_unscalarizable_region && AGGREGATE_TYPE_P (TREE_TYPE (lhs)) @@ -1616,7 +1626,8 @@ scan_function (void) tree t; unsigned i; - if (gimple_code (stmt) != GIMPLE_CALL) + if (gimple_code (stmt) != GIMPLE_CALL + || sra_mode == SRA_MODE_FINAL_INTRA) walk_stmt_load_store_addr_ops (stmt, NULL, NULL, NULL, scan_visit_addr); @@ -2261,6 +2272,26 @@ find_var_candidates (void) parm = DECL_CHAIN (parm)) ret |= maybe_add_sra_candidate (parm); + /* fsra only care about parameters and returns */ + if (sra_mode == SRA_MODE_FINAL_INTRA) + { + tree res = DECL_RESULT (current_function_decl); + if (!res || !AGGREGATE_TYPE_P (TREE_TYPE (res)) + || aggregate_value_p (res, current_function_decl)) + return ret; + + edge_iterator ei; + edge e; + FOR_EACH_EDGE (e, ei, EXIT_BLOCK_PTR_FOR_FN (cfun)->preds) + if (greturn *r = safe_dyn_cast (*gsi_last_bb (e->src))) + { + tree val = gimple_return_retval (r); + if (val && VAR_P (val)) + ret |= maybe_add_sra_candidate (val); + } + return ret; + } + FOR_EACH_LOCAL_DECL (cfun, i, var) { if (!VAR_P (var)) @@ -2771,6 +2802,33 @@ expr_with_var_bounded_array_refs_p (tree expr) return false; } +/* Analyze if the access ROOT should be replaced. */ +static bool +fsra_analyze (struct access *root) +{ + gcc_assert (sra_mode == SRA_MODE_FINAL_INTRA); + + bool scalar = is_gimple_reg_type (root->type); + if (scalar) + { + /* Parameter is scalarizable even no writing to it. */ + if (TREE_CODE (root->base) == PARM_DECL) + return root->grp_scalar_read || root->grp_assignment_read; + + /* Returns is scalarizable even no reading on it. */ + else + return root->grp_scalar_write || root->grp_assignment_write; + } + + /* fsra could handle assignment on aggregate, e.g for *p = arg; */ + if (TREE_CODE (root->base) == PARM_DECL) + return !root->grp_scalar_read && root->grp_assignment_read; + + /* For "D.3988 = *pa_2(D); return * D.3988;". */ + return !root->grp_scalar_write && root->grp_assignment_write + && root->base == root->expr; +} + /* Analyze the subtree of accesses rooted in ROOT, scheduling replacements when both seeming beneficial and when ALLOW_REPLACEMENTS allows it. If TOTALLY is set, we are totally scalarizing the aggregate. Also set all sorts of @@ -2903,6 +2961,23 @@ analyze_access_subtree (struct access *root, struct access *parent, sth_created = true; hole = false; } + + /* Support aggregate assignment, e.g. "D.3988 = *pa_2(D); return * D.3988;" + or "*p = arg;" */ + else if (allow_replacements && sra_mode == SRA_MODE_FINAL_INTRA + && !root->first_child && fsra_analyze (root)) + { + root->grp_to_be_replaced = 1; + if (scalar || TREE_CODE (root->base) == PARM_DECL) + root->replacement_decl = create_access_replacement (root); + + /* For function assign to whole 'return', directly use DECL_RESULT. */ + else + root->replacement_decl = DECL_RESULT (current_function_decl); + sth_created = true; + hole = false; + } + else { if (allow_replacements @@ -2932,6 +3007,16 @@ analyze_access_subtree (struct access *root, struct access *parent, root->grp_covered = 1; else if (root->grp_write || comes_initialized_p (root->base)) root->grp_unscalarized_data = 1; /* not covered and written to */ + + if (sra_mode == SRA_MODE_FINAL_INTRA) + { + /* Does not support writen to PARAM and partial-unscalarized RET yet. */ + if (root->grp_unscalarized_data && (VAR_P (root->base))) + return false; + if (root->grp_write && TREE_CODE (root->base) == PARM_DECL) + return false; + } + return sth_created; } @@ -3887,7 +3972,7 @@ generate_subtree_copies (struct access *access, tree agg, || access->offset + access->size > start_offset)) { tree expr, repl = get_access_replacement (access); - gassign *stmt; + gimple *stmt; expr = build_ref_for_model (loc, agg, access->offset - top_offset, access, gsi, insert_after); @@ -3899,7 +3984,19 @@ generate_subtree_copies (struct access *access, tree agg, !insert_after, insert_after ? GSI_NEW_STMT : GSI_SAME_STMT); - stmt = gimple_build_assign (repl, expr); + if (sra_mode == SRA_MODE_FINAL_INTRA + && TREE_CODE (access->base) == PARM_DECL + && (access->grp_scalar_read || access->grp_assignment_read)) + { + stmt = gimple_build_call_internal ( + IFN_ARG_PARTS, 4, access->base, + wide_int_to_tree (sizetype, access->offset), + wide_int_to_tree (sizetype, access->size), + wide_int_to_tree (sizetype, access->reverse)); + gimple_call_set_lhs (stmt, repl); + } + else + stmt = gimple_build_assign (repl, expr); } else { @@ -3909,7 +4006,14 @@ generate_subtree_copies (struct access *access, tree agg, !insert_after, insert_after ? GSI_NEW_STMT : GSI_SAME_STMT); - stmt = gimple_build_assign (expr, repl); + if (sra_mode == SRA_MODE_FINAL_INTRA && VAR_P (access->base) + && (access->grp_scalar_write || access->grp_assignment_write)) + stmt = gimple_build_call_internal ( + IFN_SET_RET_PARTS, 4, access->base, + wide_int_to_tree (sizetype, access->offset), + wide_int_to_tree (sizetype, access->size), repl); + else + stmt = gimple_build_assign (expr, repl); } gimple_set_location (stmt, loc); @@ -5134,6 +5238,14 @@ late_intra_sra (void) return perform_intra_sra (); } +/* Perform "final sra" SRA just before expander. */ +static unsigned int +final_intra_sra (void) +{ + sra_mode = SRA_MODE_FINAL_INTRA; + return perform_intra_sra (); +} + static bool gate_intra_sra (void) @@ -5217,6 +5329,44 @@ make_pass_sra (gcc::context *ctxt) return new pass_sra (ctxt); } +namespace +{ +const pass_data pass_data_sra_final = { + GIMPLE_PASS, /* type */ + "fsra", /* name */ + OPTGROUP_NONE, /* optinfo_flags */ + TV_TREE_SRA, /* tv_id */ + (PROP_cfg | PROP_ssa), /* properties_required */ + 0, /* properties_provided */ + 0, /* properties_destroyed */ + 0, /* todo_flags_start */ + TODO_update_ssa, /* todo_flags_finish */ +}; + +class pass_sra_final : public gimple_opt_pass +{ +public: + pass_sra_final (gcc::context *ctxt) + : gimple_opt_pass (pass_data_sra_final, ctxt) + { + } + + /* opt_pass methods: */ + bool gate (function *) final override { return gate_intra_sra (); } + unsigned int execute (function *) final override + { + return final_intra_sra (); + } + +}; // class pass_sra_final + +} // namespace + +gimple_opt_pass * +make_pass_sra_final (gcc::context *ctxt) +{ + return new pass_sra_final (ctxt); +} /* If type T cannot be totally scalarized, return false. Otherwise return true and push to the vector within PC offsets and lengths of all padding in the diff --git a/gcc/testsuite/g++.target/powerpc/pr102024.C b/gcc/testsuite/g++.target/powerpc/pr102024.C index 769585052b5..4d9bbb0f050 100644 --- a/gcc/testsuite/g++.target/powerpc/pr102024.C +++ b/gcc/testsuite/g++.target/powerpc/pr102024.C @@ -5,7 +5,8 @@ // Test that a zero-width bit field in an otherwise homogeneous aggregate // generates a psabi warning and passes arguments in GPRs. -// { dg-final { scan-assembler-times {\mstd\M} 4 } } +// { dg-final { scan-assembler-times {\mstd\M} 4 {target { ! has_arch_pwr8 } } } } +// { dg-final { scan-assembler-times {\mmtvsrd\M} 4 {target { has_arch_pwr8 } } } } struct a_thing { diff --git a/gcc/testsuite/gcc.target/powerpc/pr108073-1.c b/gcc/testsuite/gcc.target/powerpc/pr108073-1.c new file mode 100644 index 00000000000..4892716e85f --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr108073-1.c @@ -0,0 +1,76 @@ +/* { dg-do run } */ +/* { dg-require-effective-target hard_float } */ +/* { dg-options "-O2 -save-temps" } */ + +typedef struct DF +{ + double a[4]; + short s1; + short s2; + short s3; + short s4; +} DF; +typedef struct SF +{ + float a[4]; + int i1; + int i2; +} SF; + +/* { dg-final { scan-assembler-times {\mmtvsrd|mtvsrws\M} 3 {target { lp64 && has_arch_pwr8 } } } } */ +/* { dg-final { scan-assembler-not {\mlwz\M} {target { lp64 && has_arch_pwr8 } } } } */ +/* { dg-final { scan-assembler-not {\mlhz\M} {target { lp64 && has_arch_pwr8 } } } } */ + +#define NOIPA __attribute__ ((noipa)) + +short NOIPA +foo_hi (DF a, int flag) +{ + if (flag == 2) + return a.s2 + a.s3; + return 0; +} +int NOIPA +foo_si (SF a, int flag) +{ + if (flag == 2) + return a.i2 + a.i1; + return 0; +} +double NOIPA +foo_df (DF arg, int flag) +{ + if (flag == 2) + return arg.a[3]; + else + return 0.0; +} +float NOIPA +foo_sf (SF arg, int flag) +{ + if (flag == 2) + return arg.a[2]; + return 0; +} +float NOIPA +foo_sf1 (SF arg, int flag) +{ + if (flag == 2) + return arg.a[1]; + return 0; +} + +DF gdf = {{1.0, 2.0, 3.0, 4.0}, 1, 2, 3, 4}; +SF gsf = {{1.0f, 2.0f, 3.0f, 4.0f}, 1, 2}; + +int +main () +{ + if (!(foo_hi (gdf, 2) == 5 && foo_si (gsf, 2) == 3 && foo_df (gdf, 2) == 4.0 + && foo_sf (gsf, 2) == 3.0 && foo_sf1 (gsf, 2) == 2.0)) + __builtin_abort (); + if (!(foo_hi (gdf, 1) == 0 && foo_si (gsf, 1) == 0 && foo_df (gdf, 1) == 0 + && foo_sf (gsf, 1) == 0 && foo_sf1 (gsf, 1) == 0)) + __builtin_abort (); + return 0; +} diff --git a/gcc/testsuite/gcc.target/powerpc/pr108073.c b/gcc/testsuite/gcc.target/powerpc/pr108073.c new file mode 100644 index 00000000000..4e7feaa6810 --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr108073.c @@ -0,0 +1,74 @@ +/* { dg-do run } */ +/* { dg-require-effective-target hard_float } */ +/* { dg-options "-O2 -save-temps" } */ + +/* { dg-final { scan-assembler-times {\mmtvsrd|mtvsrws\M} 5 {target { lp64 && { has_arch_pwr8 && be } } } } } */ +/* { dg-final { scan-assembler-times {\mxscvspdpn\M} 4 {target { lp64 && { has_arch_pwr8 && be } } } } } */ +/* { dg-final { scan-assembler-times {\mmtvsrd|mtvsrws\M} 3 {target { lp64 && { has_arch_pwr8 && le } } } } } */ +/* { dg-final { scan-assembler-times {\mxscvspdpn\M} 2 {target { lp64 && { has_arch_pwr8 && le } } } } } */ +/* { dg-final { scan-assembler-times {\mfadds\M} 2 {target { lp64 && has_arch_pwr8 } } } } */ + +#define NOIPA __attribute__ ((noipa)) +typedef struct X +{ + float x; + float y; +} X; + +float NOIPA +fooX (X y) +{ + y.x += 1; + return y.x + y.y; +} + +typedef struct Y +{ + double a[4]; + long l; +} Y; + +double NOIPA +fooY (Y arg) +{ + return arg.a[3]; +} + +typedef struct Z +{ + float a[4]; + short l; +} Z; + +float NOIPA +fooZ (Z arg) +{ + return arg.a[3]; +} + +float NOIPA +fooZ2 (Z arg) +{ + return arg.a[2]; +} + +X x = {1.0f, 2.0f}; +Y y = {1.0, 2.0, 3.0, 4.0, 1}; +Z z = {1.0f, 2.0f, 3.0f, 4.0f, 1}; +int +main () +{ + if (fooX (x) != 4.0f) + __builtin_abort (); + + if (fooY (y) != 4.0) + __builtin_abort (); + + if (fooZ (z) != 4.0f) + __builtin_abort (); + + if (fooZ2 (z) != 3.0f) + __builtin_abort (); + + return 0; +} diff --git a/gcc/testsuite/gcc.target/powerpc/pr65421.c b/gcc/testsuite/gcc.target/powerpc/pr65421.c new file mode 100644 index 00000000000..25293610680 --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr65421.c @@ -0,0 +1,26 @@ +/* { dg-require-effective-target hard_float } */ +/* { dg-options "-O2" } */ + +/* generate lfd for fun1, fun4 and fun5 */ +/* { dg-final { scan-assembler-times {\mlfd\M} 12 {target { lp64 && powerpc_elfv2 } } } } */ + +/* generate stfd for fun2 */ +/* { dg-final { scan-assembler-times {\mstfd\M} 4 {target { lp64 && powerpc_elfv2 } } } } */ + +/* no other load/store */ +/* { dg-final { scan-assembler-not {\m[p?l|st]d\M} {target { lp64 && powerpc_elfv2 } } } } */ +/* { dg-final { scan-assembler-not {\m[l|st]xvd2x\M} {target { lp64 && powerpc_elfv2 } } } } */ +/* { dg-final { scan-assembler-not {\m[p?l|st]xv\M} {target { lp64 && powerpc_elfv2 } } } } */ + +typedef struct { double a[4]; } A; +typedef struct { A a; double b; } B; +typedef struct { double c; A a;} C; + +A fun1 (const A *pa) { return *pa; } +void fun2 (const A a, A *pa) { *pa = a; } +A fun3 (const A a) { return a; } +A fun4 (const B *b) { return b->a; } +A fun5 (const C *b) { return b->a; } +A fun6 (const B b) { return b.a; } +B fun7 (const A a) { B b; b.a = a; b.b = 0; return b; } +C fun8 (const A a) { C c; c.a = a; c.c = 0; return c; } diff --git a/gcc/testsuite/gcc.target/powerpc/pr69143.c b/gcc/testsuite/gcc.target/powerpc/pr69143.c new file mode 100644 index 00000000000..62db4b0bd82 --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr69143.c @@ -0,0 +1,22 @@ +/* { dg-require-effective-target hard_float } */ +/* { dg-options "-O2" } */ + +/* { dg-final { scan-assembler-times {\mfmr\M} 3 {target { { lp64 && powerpc_elfv2 } && has_arch_pwr8 } } } } */ +/* { dg-final { scan-assembler-not {\mxscvspdpn\M} {target { { lp64 && powerpc_elfv2 } && has_arch_pwr8 } } } } */ + +struct foo1 +{ + float x; + float y; +}; + +struct foo1 +blah1 (struct foo1 y) +{ + struct foo1 x; + + x.x = y.y; + x.y = y.x; + + return x; +}