From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: 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 14DFD3849AE9 for ; Wed, 24 Apr 2024 09:16:24 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 14DFD3849AE9 Authentication-Results: sourceware.org; dmarc=pass (p=none dis=none) header.from=redhat.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=redhat.com ARC-Filter: OpenARC Filter v1.0.0 sourceware.org 14DFD3849AE9 Authentication-Results: server2.sourceware.org; arc=none smtp.remote-ip=170.10.133.124 ARC-Seal: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1713950185; cv=none; b=B+/j7iOb/cbZy3Iem4Y806hCEd810E/k2Kb3rPsFaARF6M3tvw7KGFBhn8cNsgHuM2ZJWc6wcA7TdOi0xaR9wFjHvGuo8SwDbiM/WNfSPFFtaZJrA80SWMWTmRiXIGeDMZMTLoqJe92StgfGEVKfyVYi9jxULUWxctTZCu+Dt2s= ARC-Message-Signature: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1713950185; c=relaxed/simple; bh=ZYDgO5dRSEwsCH4dUqv2pT8XjG+lbwxR698b+hZEY5k=; h=DKIM-Signature:MIME-Version:From:Date:Message-ID:Subject:To; b=vi31rBFeKUBVQ6hFKPDK0T+KHwD0RoyHhgeQW+nqecNj6TWHj8HjFgjGItlb/g4EjQSeRt7MJ3KvpUBSoFiGr5fmjGtoMXA3ScD046NgyBjeStbFddI3FS+cyaNYWOr5xJEng2xF+2QdRQapQ22N3/2SPGAsZgmMOkOwslAm6ck= ARC-Authentication-Results: i=1; server2.sourceware.org DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1713950183; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: in-reply-to:in-reply-to:references:references; bh=LBbl+rwiXgcPmkp1gb2V1lFCHuELzXDVeuQ7kOKwAo4=; b=XVHMkhbTEL+MPJLPL6YJ4nEwpccjKuS9/jQ6nNmFzzxI5qHgclyMGtUxFrDawS9+fdsYVw Sg6A/oP2/fY4asp3jGDTNwO2GVlgHyd/2ArWlhJg1G1TdHz68JubNlD+5r+r+LeIZ0UmTP yQfpH4Zab+usRqtBNtbvAXWbeo+zENM= Received: from mail-yb1-f199.google.com (mail-yb1-f199.google.com [209.85.219.199]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-596-AOQxFfRuNFqPB7eRVS1tYQ-1; Wed, 24 Apr 2024 05:16:22 -0400 X-MC-Unique: AOQxFfRuNFqPB7eRVS1tYQ-1 Received: by mail-yb1-f199.google.com with SMTP id 3f1490d57ef6-de46f4eacd3so12475664276.3 for ; Wed, 24 Apr 2024 02:16:22 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1713950181; x=1714554981; h=cc:to:subject:message-id:date:from:in-reply-to:references :mime-version:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=LBbl+rwiXgcPmkp1gb2V1lFCHuELzXDVeuQ7kOKwAo4=; b=YUd3DT5RypgWh/Xn4zHPxob/ZJ1lS6hl0Yw5O4LqW2zX8YDog30aGxFDPKa14TkX+n xptWIcdRTBFklfZ+wkVsRg1DwCT4dByScrW5pk0EHq7SSs38KPCqjlVkBwQGs6d5cx4F tIyoQZDzTy5MJ2w4cBwMo9GZagB5I7H2dwmvLHEmyegLHHiI3CFV6jLa5ZwVVRer8EWt eGgi8jepkm/4BMVVIm91o8rzZAcMA29d7T+zYr+xGyqzYs2zwF9nLMU9oslpcbhuQrwa yjPX35yJihhEyXNG57IVYk2S9N7RA+WhXsrSujncCgF8AYQlymfZMCoNC9UqTNB+43IY LfPQ== X-Forwarded-Encrypted: i=1; AJvYcCUvUtV51KXmg+i6NaY/8eBFw1QdK1d0x7u1N8WTlPIFwA7o8H/Env+F5ss5zb2tiAhcf0CxjOs4s9V3yjVP0x8GHsW3zIXlow== X-Gm-Message-State: AOJu0YyYfRopiunUojQcoibN4ZP3Ps5lM5latgKxJZyq4YSmU7GF1WcN FZhsHxUgEjGR+GR3p9OunptYAG79KkUGH5rsgN+vU22m34QBEV/chZ65ba8+MYcMvJrx46UxD4o /cyrU1NgDRp+yL2TeNox24Zmd8AHgupkLyAq5GGiw5v82z5iGVeiSmx4aNA5Ztb35yBAxIucQwj r+82B5vLIivQMEyAaUtOa1oipCeFPrXg== X-Received: by 2002:a25:ce12:0:b0:dca:e4fd:b6d5 with SMTP id x18-20020a25ce12000000b00dcae4fdb6d5mr2145901ybe.27.1713950181600; Wed, 24 Apr 2024 02:16:21 -0700 (PDT) X-Google-Smtp-Source: AGHT+IE/Y2Al+leQFWGxyaETjX/qbP7KAGzUTxtSyVZItb8KpTENEdDvRrLa7wJ7REyK/Dj/+Q8+z6plCkY0WI2o0ks= X-Received: by 2002:a25:ce12:0:b0:dca:e4fd:b6d5 with SMTP id x18-20020a25ce12000000b00dcae4fdb6d5mr2145888ybe.27.1713950181294; Wed, 24 Apr 2024 02:16:21 -0700 (PDT) MIME-Version: 1.0 References: <5fe9db40-4bff-4be4-a434-6509133f26e3@redhat.com> In-Reply-To: From: Jonathan Wakely Date: Wed, 24 Apr 2024 10:16:05 +0100 Message-ID: Subject: Re: [PATCH] c++, v2: Retry the aliasing of base/complete cdtor optimization at import_export_decl time [PR113208] To: Jakub Jelinek Cc: Jason Merrill , gcc-patches@gcc.gnu.org, Jan Hubicka , Richard Biener , Patrick Palka X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Type: text/plain; charset="UTF-8" X-Spam-Status: No, score=-6.5 required=5.0 tests=BAYES_00,DKIMWL_WL_HIGH,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,RCVD_IN_DNSWL_NONE,RCVD_IN_MSPIKE_H4,RCVD_IN_MSPIKE_WL,SPF_HELO_NONE,SPF_NONE,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 List-Id: On Tue, 23 Apr 2024 at 17:05, Jakub Jelinek wrote: > > On Mon, Apr 22, 2024 at 11:14:35PM -0400, Jason Merrill wrote: > > > > The following testcase regressed with Marek's r14-5979 change, > > > > when pr113208_0.C is compiled where the ctor is marked constexpr, > > > > we no longer perform this optimization, where > > > > _ZN6vectorI12QualityValueEC2ERKS1_ was emitted in the > > > > _ZN6vectorI12QualityValueEC5ERKS1_ comdat group and > > > > _ZN6vectorI12QualityValueEC1ERKS1_ was made an alias to it, > > > > instead we emit _ZN6vectorI12QualityValueEC2ERKS1_ in > > > > _ZN6vectorI12QualityValueEC2ERKS1_ comdat group and the same > > > > content _ZN6vectorI12QualityValueEC1ERKS1_ as separate symbol in > > > > _ZN6vectorI12QualityValueEC1ERKS1_ comdat group. > > > > This seems like an ABI bug that could use a non-LTO testcase. > > Well, except for the issues it causes to LTO I think it is compatible, > worst case we get the body of the ctor duplicated in the executable > and the linker picks some of the weak symbols as the symbol definitions. > Anyway, I've added a non-LTO testcase for that in the patch below. > > > Hmm, cloning the bodies and then discarding them later seems like more extra > > work than creating the cgraph nodes. > > So, I've tried to handle that in tentative_decl_linkage, like that function > already handles functions declared inline except for implicit template > instantiations. If we expect that import_export_decl will do comdat_linkage > for the ctor later on do it right away. > > That fixes the testcases too, but seems to regress > +FAIL: libstdc++-abi/abi_check > on both x86_64-linux and i686-linux, in each case 8 symbols disappeared from > libstdc++.so.6: > _ZNSt12__shared_ptrINSt10filesystem7__cxx1128recursive_directory_iterator10_Dir_stackELN9__gnu_cxx12_Lock_policyE2EEC1Ev > _ZNSt12__shared_ptrINSt10filesystem4_DirELN9__gnu_cxx12_Lock_policyE2EEC1Ev > _ZNSt12__shared_ptrINSt10filesystem28recursive_directory_iterator10_Dir_stackELN9__gnu_cxx12_Lock_policyE2EEC1Ev > _ZNSt12__shared_ptrINSt10filesystem4_DirELN9__gnu_cxx12_Lock_policyE2EEC2Ev > _ZNSt12__shared_ptrINSt10filesystem7__cxx114_DirELN9__gnu_cxx12_Lock_policyE2EEC1Ev > _ZNSt12__shared_ptrINSt10filesystem7__cxx1128recursive_directory_iterator10_Dir_stackELN9__gnu_cxx12_Lock_policyE2EEC2Ev > _ZNSt12__shared_ptrINSt10filesystem28recursive_directory_iterator10_Dir_stackELN9__gnu_cxx12_Lock_policyE2EEC2Ev > _ZNSt12__shared_ptrINSt10filesystem7__cxx114_DirELN9__gnu_cxx12_Lock_policyE2EEC2Ev > > Will need to study why that happened, it might be that it was ok because > I think the filesystem stuff is unlike the rest compiled with no exported > templates, but would need at least some hacks in libstdc++ to preserve > previously exported symbols. There are explicit instantiation definitions that should instantiate those types: src/c++17/fs_dir.cc:template class std::__shared_ptr; src/c++17/fs_dir.cc:template class std::__shared_ptr; src/c++17/fs_path.cc:template class std::__shared_ptr; So the missing symbols should be present in cow-fs_dir.o and cow-fs_path.o > Still, feels like a risky change this late if it wouldn't break ABI of other > libraries. > > 2024-04-23 Jakub Jelinek > > PR lto/113208 > * decl2.cc (tentative_decl_linkage): Use comdat_linkage also > for implicit instantiations of maybe in charge ctors/dtors > if -fimplicit-templates or -fimplicit-inline-templates and > -fweak and target supports aliases. > > * g++.dg/abi/comdat2.C: New test. > * g++.dg/lto/pr113208_0.C: New test. > * g++.dg/lto/pr113208_1.C: New file. > * g++.dg/lto/pr113208.h: New file. > > --- gcc/cp/decl2.cc.jj 2024-04-22 15:16:55.328548807 +0200 > +++ gcc/cp/decl2.cc 2024-04-23 09:52:18.993250442 +0200 > @@ -3314,7 +3314,16 @@ tentative_decl_linkage (tree decl) > to mark the functions at this point. */ > if (DECL_DECLARED_INLINE_P (decl) > && (!DECL_IMPLICIT_INSTANTIATION (decl) > - || DECL_DEFAULTED_FN (decl))) > + || DECL_DEFAULTED_FN (decl) > + /* For implicit instantiations of cdtors, > + if import_export_decl would use comdat linkage, > + make sure to use it right away, so that maybe_clone_body > + can use aliases. See PR113208. */ > + || (DECL_MAYBE_IN_CHARGE_CDTOR_P (decl) > + && (flag_implicit_templates > + || flag_implicit_inline_templates) > + && flag_weak > + && TARGET_SUPPORTS_ALIASES))) > { > /* This function must have external linkage, as > otherwise DECL_INTERFACE_KNOWN would have been > --- gcc/testsuite/g++.dg/abi/comdat2.C.jj 2024-04-23 10:04:28.485964610 +0200 > +++ gcc/testsuite/g++.dg/abi/comdat2.C 2024-04-23 10:05:24.757171194 +0200 > @@ -0,0 +1,26 @@ > +// PR lto/113208 > +// { dg-do compile { target { c++11 && { *-*-*gnu* } } } } > +// { dg-additional-options "-O2 -fkeep-inline-functions" } > +// { dg-final { scan-assembler "_ZN1BI1CEC5ERKS1_,comdat" } } > +// { dg-final { scan-assembler-not "_ZN1BI1CEC1ERKS1_,comdat" } } > +// { dg-final { scan-assembler-not "_ZN1BI1CEC2ERKS1_,comdat" } } > + > +template > +struct A { > + int foo () const; > + A (int, int); > +}; > +template > +struct B : A { > + constexpr B (const B &x) : A (1, x.foo ()) {} > + B () : A (1, 2) {} > +}; > +struct C; > +struct D : B {}; > +void bar (D); > + > +void > +baz (D x) > +{ > + bar (x); > +} > --- gcc/testsuite/g++.dg/lto/pr113208.h.jj 2024-04-23 09:44:07.523179110 +0200 > +++ gcc/testsuite/g++.dg/lto/pr113208.h 2024-04-23 09:44:07.523179110 +0200 > @@ -0,0 +1,10 @@ > +template struct _Vector_base { > + int g() const; > + _Vector_base(int, int); > +}; > +template > +struct vector : _Vector_base<_Tp> { > + CONSTEXPR vector(const vector &__x) > + : _Vector_base<_Tp>(1, __x.g()) {} > + vector() : _Vector_base<_Tp>(1, 2) {} > +}; > --- gcc/testsuite/g++.dg/lto/pr113208_1.C.jj 2024-04-23 09:44:07.523179110 +0200 > +++ gcc/testsuite/g++.dg/lto/pr113208_1.C 2024-04-23 09:44:07.523179110 +0200 > @@ -0,0 +1,6 @@ > +#define CONSTEXPR > +#include "pr113208.h" > + > +struct QualityValue; > +vector values1; > +vector values{values1}; > --- gcc/testsuite/g++.dg/lto/pr113208_0.C.jj 2024-04-23 09:44:07.523179110 +0200 > +++ gcc/testsuite/g++.dg/lto/pr113208_0.C 2024-04-23 09:44:07.523179110 +0200 > @@ -0,0 +1,13 @@ > +// { dg-lto-do link } > +// { dg-lto-options { {-O1 -std=c++20 -flto}} } > +// { dg-extra-ld-options "-r -nostdlib -flinker-output=nolto-rel" } > +// { dg-require-linker-plugin "" } > + > +#define CONSTEXPR constexpr > +#include "pr113208.h" > + > +struct QualityValue; > +struct k : vector {}; > + > +void m(k); > +void n(k i) { m(i); } > > > Jakub >