From patchwork Wed Mar 1 09:19:48 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jakub Jelinek X-Patchwork-Id: 62808 Return-Path: Delivered-To: ouuuleilei@gmail.com Received: by 2002:a5d:5915:0:0:0:0:0 with SMTP id v21csp3526085wrd; Wed, 1 Mar 2023 01:20:40 -0800 (PST) X-Google-Smtp-Source: AK7set9XK8bxw40LxZWlLdjI1/R3IA97CNANTycw/BZFGEt0WY3haZyaZXuZB7eKfT8LCFCjKAf+ X-Received: by 2002:a17:907:d046:b0:8f3:8bfd:a8e with SMTP id vb6-20020a170907d04600b008f38bfd0a8emr6178591ejc.26.1677662440241; Wed, 01 Mar 2023 01:20:40 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1677662440; cv=none; d=google.com; s=arc-20160816; b=mLDU4TlAVUeG61D+QxstpUA96jlhbtn/kVkruD9SLD0Qz13/9KmkbD1IYrXy3ObJz1 cLsyI6Ykft7zlN74He/Xqdc4dDLxnRq1epZ4fFrB5RFQ8c3ivlVCh0BSuSsja94PybMj vRPNowWa7jEMz9QOhTAYuK91cKO7D3+VAqw2V4i5+VSnPKlOWl9gueKHZh0DIEri0bPI fwu06vWaSHeOjEEHjxylJeJvpuvBlWIJnFtyGqHJWvg5UNXZ+ncmseUrMqrMIGIu7OSq rN+cEnNyS0JFqCg0CQfDfhDpe62gmpWUB6J8bZ59sONvdm3snuoU5cpKeGEN01k1I9+K yVow== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=sender:errors-to:reply-to:from:list-subscribe:list-help:list-post :list-archive:list-unsubscribe:list-id:precedence :content-disposition:in-reply-to:mime-version:references:message-id :subject:to:date:dmarc-filter:delivered-to:dkim-signature :dkim-filter; bh=tLRrt41XpexO1GsDa44aLW/Q/1aJUik47ziqb6B/HXY=; b=HO+dzPKdCm/fC+ieQ9UN1KqWFuZ9JN+4ECerZA9PGhcTPYwtzyg5KH/1BXuP8BIWpi PkFURLd115netduL3HJ0CPZ5SdwDq9XxiX1QPYFhE7BuaN6lBmmEOe/wEJeStOngayds scJyuXAmEQvyerFDrOcfJSefQKoVhxC9rycxB5Zlq+rZmRPMCGujNhtnUDK1JvIph+2B rF6YCRlwJdh7EickWuo/+vzLhf+TlIhz9H7eFq3XfHaJmcK8CNtn9UFc8Q71vh+cKHY+ QFneoIZyTHN37FAcZguc+9FJ8wMK54lC74XWIdG4Po2L5iTr4DZ38KOjZlsg7Fi9qvrr c8og== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@gcc.gnu.org header.s=default header.b=DdDXm4Wp; spf=pass (google.com: domain of gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org designates 8.43.85.97 as permitted sender) smtp.mailfrom="gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org"; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=gnu.org Received: from sourceware.org (server2.sourceware.org. [8.43.85.97]) by mx.google.com with ESMTPS id k19-20020a17090646d300b008e176493a7fsi13431744ejs.358.2023.03.01.01.20.40 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 01 Mar 2023 01:20:40 -0800 (PST) Received-SPF: pass (google.com: domain of gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org designates 8.43.85.97 as permitted sender) client-ip=8.43.85.97; Authentication-Results: mx.google.com; dkim=pass header.i=@gcc.gnu.org header.s=default header.b=DdDXm4Wp; spf=pass (google.com: domain of gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org designates 8.43.85.97 as permitted sender) smtp.mailfrom="gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org"; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=gnu.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 51B093858C2C for ; Wed, 1 Mar 2023 09:20:39 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 51B093858C2C DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1677662439; bh=tLRrt41XpexO1GsDa44aLW/Q/1aJUik47ziqb6B/HXY=; h=Date:To:Subject:References:In-Reply-To:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To: From; b=DdDXm4WphdXr8+qcIfRzouIORGlNT7GuOu5xyQ2yZR0XLv6c07/qrasLwlEEvc8lL 7f468s1Wy2YKf09Bw8e6FFEpHzQFkPxFVNwi8Jt6GYdmsxJ2QIEDoKaplp0kkEAcdM a1n0vwX5/6lMc3l7W4x93tahb88v/uVX/T5ZITAU= X-Original-To: gcc-patches@gcc.gnu.org Delivered-To: gcc-patches@gcc.gnu.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) by sourceware.org (Postfix) with ESMTPS id C895A3858D33 for ; Wed, 1 Mar 2023 09:19:53 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org C895A3858D33 Received: from mimecast-mx02.redhat.com (mx3-rdu2.redhat.com [66.187.233.73]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-394-U1CMRsM_NlW-RenLQiqRjA-1; Wed, 01 Mar 2023 04:19:52 -0500 X-MC-Unique: U1CMRsM_NlW-RenLQiqRjA-1 Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.rdu2.redhat.com [10.11.54.6]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 10F0B280A322 for ; Wed, 1 Mar 2023 09:19:52 +0000 (UTC) Received: from tucnak.zalov.cz (unknown [10.39.192.16]) by smtp.corp.redhat.com (Postfix) with ESMTPS id C84322166B26; Wed, 1 Mar 2023 09:19:51 +0000 (UTC) Received: from tucnak.zalov.cz (localhost [127.0.0.1]) by tucnak.zalov.cz (8.17.1/8.17.1) with ESMTPS id 3219Jniw1620836 (version=TLSv1.3 cipher=TLS_AES_256_GCM_SHA384 bits=256 verify=NOT); Wed, 1 Mar 2023 10:19:49 +0100 Received: (from jakub@localhost) by tucnak.zalov.cz (8.17.1/8.17.1/Submit) id 3219JnGb1620835; Wed, 1 Mar 2023 10:19:49 +0100 Date: Wed, 1 Mar 2023 10:19:48 +0100 To: Jason Merrill , Jonathan Wakely , gcc-patches@gcc.gnu.org Subject: [PATCH] c++, v2: Emit fundamental tinfos for all _Float*/decltype(0.0bf16) types unconditionally [PR108883] Message-ID: References: <7163c5db-3025-01cb-c0ad-5526e46eff8c@redhat.com> MIME-Version: 1.0 In-Reply-To: X-Scanned-By: MIMEDefang 3.1 on 10.11.54.6 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Disposition: inline X-Spam-Status: No, score=-15.0 required=5.0 tests=BAYES_00, DKIM_INVALID, DKIM_SIGNED, KAM_DMARC_NONE, KAM_DMARC_STATUS, RCVD_IN_DNSWL_NONE, RCVD_IN_MSPIKE_H2, SPF_HELO_NONE, SPF_NONE, TXREP autolearn=no 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: 2.1.29 Precedence: list List-Id: Gcc-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Jakub Jelinek via Gcc-patches From: Jakub Jelinek Reply-To: Jakub Jelinek Errors-To: gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org Sender: "Gcc-patches" X-getmail-retrieved-from-mailbox: =?utf-8?q?INBOX?= X-GMAIL-THRID: =?utf-8?q?1759156571063541348?= X-GMAIL-MSGID: =?utf-8?q?1759156571063541348?= On Tue, Feb 28, 2023 at 11:04:42AM +0100, Jakub Jelinek via Gcc-patches wrote: > On Tue, Feb 28, 2023 at 12:51:06AM +0100, Jakub Jelinek via Gcc-patches wrote: > > And then there is a question whether we want to emit rtti for > > _Float{16,32,64,128}, _Float{32,64,128}x and decltype(0.0bf16) regardless > > of whether the target supports them at all or not. > > If the answer to that is yes, we want them all, then the following patch > arranges that (had to force immediate emit_tinfo_decl to get rid of the > fallback_* stuff), these days it doesn't mean it will be emitted right away > anyway, just it will be registered with varpool and have initializers > created. > > If the answer is yes, except _Float128x, we could in the patch just move > that &float128x_type_node to the other initializer. > > If the answer is no, then I think we need a target hook which would tell us > which fundamentals should be forced this way, could have e.g. a form > of returning a vector of global tree addresses to be treated that way in > addition to the standard ones (dfloat*), or a hook that would take > a callback pointer and be called with &emit_support_tinfo_1 and the backend > would call the callback on any types it wants to handle that way. Here is an actually bootstrapped/regtested version of that patch - had to add bfloat16_type_node mangling to cp/mangle.cc, otherwise it ICEd e.g. on i686-linux when not compiled with -msse2, but also on any non-x86 target. Though, if the answer to the earlier question is no, I think the version with a callback passed to target hook would be better because e.g. target might need to specify some particular TYPE_MODE or similar things on the fallback type for a particular case. 2023-03-01 Jakub Jelinek PR target/108883 * cp-tree.h (enum cp_tree_index): Remove CPTI_FALLBACK_DFLOAT*_TYPE enumerators. (fallback_dfloat32_type, fallback_dfloat64_type, fallback_dfloat128_type): Remove. * rtti.cc (emit_support_tinfo_1): If not emitted already, call emit_tinfo_decl and remove from unemitted_tinfo_decls right away. (emit_support_tinfos): Move &dfloat*_type_node, &bfloat16-type_node and &float[0-9]*_type_node from fundamentals array into new fundamentals_with_fallback array. Call emit_support_tinfo_1 on elements of that array too, with the difference that if the type is NULL, use a fallback REAL_TYPE for it temporarily. Drop the !targetm.decimal_float_supported_p () handling. * mangle.cc (write_builtin_type): Remove references to fallback_dfloat*_type. Handle bfloat16_type_node mangling. Jakub --- gcc/cp/cp-tree.h.jj 2023-02-18 12:38:30.910023526 +0100 +++ gcc/cp/cp-tree.h 2023-02-28 10:39:22.377379948 +0100 @@ -235,10 +235,6 @@ enum cp_tree_index CPTI_PSEUDO_CONTRACT_VIOLATION, - CPTI_FALLBACK_DFLOAT32_TYPE, - CPTI_FALLBACK_DFLOAT64_TYPE, - CPTI_FALLBACK_DFLOAT128_TYPE, - CPTI_MAX }; @@ -397,13 +393,6 @@ extern GTY(()) tree cp_global_trees[CPTI access nodes in tree.h. */ #define access_default_node null_node - -/* Variant of dfloat{32,64,128}_type_node only used for fundamental - rtti purposes if DFP is disabled. */ -#define fallback_dfloat32_type cp_global_trees[CPTI_FALLBACK_DFLOAT32_TYPE] -#define fallback_dfloat64_type cp_global_trees[CPTI_FALLBACK_DFLOAT64_TYPE] -#define fallback_dfloat128_type cp_global_trees[CPTI_FALLBACK_DFLOAT128_TYPE] - #include "name-lookup.h" --- gcc/cp/rtti.cc.jj 2023-02-22 15:58:50.137998090 +0100 +++ gcc/cp/rtti.cc 2023-02-28 10:31:53.693893021 +0100 @@ -1577,6 +1577,15 @@ emit_support_tinfo_1 (tree bltn) gcc_assert (TREE_PUBLIC (tinfo) && !DECL_COMDAT (tinfo)); DECL_INTERFACE_KNOWN (tinfo) = 1; } + + /* Emit it right away if not emitted already. */ + if (DECL_INITIAL (tinfo) == NULL_TREE) + { + gcc_assert (unemitted_tinfo_decls->last () == tinfo); + bool ok = emit_tinfo_decl (tinfo); + gcc_assert (ok); + unemitted_tinfo_decls->pop (); + } } } @@ -1602,10 +1611,17 @@ emit_support_tinfos (void) &long_integer_type_node, &long_unsigned_type_node, &long_long_integer_type_node, &long_long_unsigned_type_node, &float_type_node, &double_type_node, &long_double_type_node, + &nullptr_type_node, + 0 + }; + /* Similar, but for floating point types only which should get type info + regardless whether they are non-NULL or NULL. */ + static tree *const fundamentals_with_fallback[] = + { &dfloat32_type_node, &dfloat64_type_node, &dfloat128_type_node, &bfloat16_type_node, &float16_type_node, &float32_type_node, &float64_type_node, &float128_type_node, &float32x_type_node, - &float64x_type_node, &float128x_type_node, &nullptr_type_node, + &float64x_type_node, &float128x_type_node, 0 }; int ix; @@ -1627,8 +1643,20 @@ emit_support_tinfos (void) location_t saved_loc = input_location; input_location = BUILTINS_LOCATION; doing_runtime = 1; + tree fallback = NULL_TREE; for (ix = 0; fundamentals[ix]; ix++) emit_support_tinfo_1 (*fundamentals[ix]); + for (ix = 0; fundamentals_with_fallback[ix]; ix++) + if (*fundamentals_with_fallback[ix]) + emit_support_tinfo_1 (*fundamentals_with_fallback[ix]); + else + { + if (fallback == NULL_TREE) + fallback = make_node (REAL_TYPE); + *fundamentals_with_fallback[ix] = fallback; + emit_support_tinfo_1 (fallback); + *fundamentals_with_fallback[ix] = NULL_TREE; + } for (ix = 0; ix < NUM_INT_N_ENTS; ix ++) if (int_n_enabled_p[ix]) { @@ -1637,20 +1665,6 @@ emit_support_tinfos (void) } for (tree t = registered_builtin_types; t; t = TREE_CHAIN (t)) emit_support_tinfo_1 (TREE_VALUE (t)); - /* For compatibility, emit DFP typeinfos even when DFP isn't enabled, - because we've emitted that in the past. */ - if (!targetm.decimal_float_supported_p ()) - { - gcc_assert (dfloat32_type_node == NULL_TREE - && dfloat64_type_node == NULL_TREE - && dfloat128_type_node == NULL_TREE); - fallback_dfloat32_type = make_node (REAL_TYPE); - fallback_dfloat64_type = make_node (REAL_TYPE); - fallback_dfloat128_type = make_node (REAL_TYPE); - emit_support_tinfo_1 (fallback_dfloat32_type); - emit_support_tinfo_1 (fallback_dfloat64_type); - emit_support_tinfo_1 (fallback_dfloat128_type); - } input_location = saved_loc; } --- gcc/cp/mangle.cc.jj 2023-02-09 09:31:48.900375193 +0100 +++ gcc/cp/mangle.cc 2023-02-28 23:55:06.912039890 +0100 @@ -2732,11 +2732,11 @@ write_builtin_type (tree type) write_char ('d'); else if (type == long_double_type_node) write_char ('e'); - else if (type == dfloat32_type_node || type == fallback_dfloat32_type) + else if (type == dfloat32_type_node) write_string ("Df"); - else if (type == dfloat64_type_node || type == fallback_dfloat64_type) + else if (type == dfloat64_type_node) write_string ("Dd"); - else if (type == dfloat128_type_node || type == fallback_dfloat128_type) + else if (type == dfloat128_type_node) write_string ("De"); else if (type == float16_type_node) write_string ("DF16_"); @@ -2752,6 +2752,8 @@ write_builtin_type (tree type) write_string ("DF64x"); else if (type == float128x_type_node) write_string ("DF128x"); + else if (type == bfloat16_type_node) + write_string ("DF16b"); else gcc_unreachable (); break;