From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mo4-p00-ob.smtp.rzone.de (mo4-p00-ob.smtp.rzone.de [81.169.146.216]) by sourceware.org (Postfix) with ESMTPS id A95273858C3A; Sun, 23 Jan 2022 13:18:22 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org A95273858C3A Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=nieper-wisskirchen.de Authentication-Results: sourceware.org; spf=none smtp.mailfrom=nieper-wisskirchen.de DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; t=1642943901; s=strato-dkim-0002; d=nieper-wisskirchen.de; h=Cc:To:Subject:Message-ID:Date:From:In-Reply-To:References:Cc:Date: From:Subject:Sender; bh=zEDp7I485nXLMFZdl+gmDco7sJlORnbjJOkWkMjw8wQ=; b=qx5B1rTpZ/nnkmnt9oZq7F2yy3L0l3zr/JAomaeqvOcHWfByinmcR4kzwBEgfiPNlc EiYLbMR1cmq4L1/OA+JOV3/twnqLDTB4bjFDMqe9QwyQaeLIb0UJ+1Sa/5E69dijedQu YA7kZG5uW5Ag6KUTUsOrApx8yYNLRRY0ba4PNggrowMhV69ZLULf7StSXqm5uOnlxWfL raDwC8+OrOxorhnipLNygGG3rRW/oRT9y//Hng1fLX/CHAIFw+eTJN3OjibVNreHvXSt Jj7+4O0s3ZZGGtc8/ESOxx7y01/VVpS5z1ZCalxvv9F+TlFxkEhR6ndAdvEGSA2DFzxT lfcw== Authentication-Results: strato.com; dkim=none X-RZG-AUTH: ":IW0WdmCmcvpIrP2+VJuPtIhjJvc4Ig+QdhX22iZVwSDOx4Kp3cYsBVGy6CZgmO/guIaKV8N5653pcA==" X-RZG-CLASS-ID: mo00 Received: from mail-yb1-f182.google.com by smtp.strato.de (RZmta 47.38.0 AUTH) with ESMTPSA id 262004y0NDILGni (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256 bits)) (Client did not present a certificate); Sun, 23 Jan 2022 14:18:21 +0100 (CET) Received: by mail-yb1-f182.google.com with SMTP id m1so42602304ybo.5; Sun, 23 Jan 2022 05:18:20 -0800 (PST) X-Gm-Message-State: AOAM531y5cDOaIr9spsZp/iTIUDna2VH1dGXkKYZ5Tc4O5W+2wNTntMj ArE1z+4LYFPRV2IrI3krrVB6Jymb6PAV36FChS4= X-Google-Smtp-Source: ABdhPJz85oDtAU+m0rSSOHhf/FoB7NZT4ycTUTH0Cze7Cuzrr9HZYcCys4hdtKhsNldz8mu5RDA3BP9cRF9qf96GU/A= X-Received: by 2002:a5b:dc7:: with SMTP id t7mr16955086ybr.17.1642943900085; Sun, 23 Jan 2022 05:18:20 -0800 (PST) MIME-Version: 1.0 References: <20211219213010.17113-1-marc@nieper-wisskirchen.de> <2a54fac9b37d87afb009b8eb339d5ad6927454dd.camel@redhat.com> In-Reply-To: From: =?UTF-8?Q?Marc_Nieper=2DWi=C3=9Fkirchen?= Date: Sun, 23 Jan 2022 14:18:09 +0100 X-Gmail-Original-Message-ID: Message-ID: Subject: Re: [PATCH] gcc: pass-manager: Fix memory leak. [PR jit/63854] To: Jeff Law Cc: David Malcolm , gcc-patches@gcc.gnu.org, =?UTF-8?Q?Marc_Nieper=2DWi=C3=9Fkirchen_via_Jit?= Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Spam-Status: No, score=-3.8 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, RCVD_IN_DNSWL_LOW, RCVD_IN_MSPIKE_H3, RCVD_IN_MSPIKE_WL, SPF_HELO_PASS, SPF_NONE, TXREP autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: jit@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Jit mailing list List-Unsubscribe: , List-Archive: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 23 Jan 2022 13:18:25 -0000 Am Sa., 15. Jan. 2022 um 14:56 Uhr schrieb Marc Nieper-Wi=C3=9Fkirchen : > > Jeff, David, do you need any more input from my side? > > -- Marc > > Am Sa., 8. Jan. 2022 um 17:32 Uhr schrieb Jeff Law : > > > > > > > > On 1/6/2022 6:53 AM, David Malcolm via Gcc-patches wrote: > > > On Sun, 2021-12-19 at 22:30 +0100, Marc Nieper-Wi=C3=9Fkirchen wrote: > > >> This patch fixes a memory leak in the pass manager. In the existing > > >> code, > > >> the m_name_to_pass_map is allocated in > > >> pass_manager::register_pass_name, but > > >> never deallocated. This is fixed by adding a deletion in > > >> pass_manager::~pass_manager. Moreover the string keys in > > >> m_name_to_pass_map are > > >> all dynamically allocated. To free them, this patch adds a new hash > > >> trait for > > >> string hashes that are to be freed when the corresponding hash entry > > >> is removed. > > >> > > >> This fix is particularly relevant for using GCC as a library through > > >> libgccjit. > > >> The memory leak also occurs when libgccjit is instructed to use an > > >> external > > >> driver. > > >> > > >> Before the patch, compiling the hello world example of libgccjit wit= h > > >> the external driver under Valgrind shows a loss of 12,611 (48 direct= ) > > >> bytes. After the patch, no memory leaks are reported anymore. > > >> (Memory leaks occurring when using the internal driver are mostly in > > >> the driver code in gcc/gcc.c and have to be fixed separately.) > > >> > > >> The patch has been tested by fully bootstrapping the compiler with > > >> the > > >> frontends C, C++, Fortran, LTO, ObjC, JIT and running the test suite > > >> under a x86_64-pc-linux-gnu host. > > > Thanks for the patch. > > > > > > It looks correct to me, given that pass_manager::register_pass_name > > > does an xstrdup and puts the result in the map. > > > > > > That said: > > > - I'm not officially a reviewer for this part of gcc (though I probab= ly > > > touched this code last) > > > - is it cleaner to instead change m_name_to_pass_map's key type from > > > const char * to char *, to convey that the map "owns" the name? That > > > way we probably wouldn't need struct typed_const_free_remove, and (I > > > hope) works better with the type system. > > > > > > Dave > > > > > >> gcc/ChangeLog: > > >> > > >> PR jit/63854 > > >> * hash-traits.h (struct typed_const_free_remove): New. > > >> (struct free_string_hash): New. > > >> * pass_manager.h: Use free_string_hash. > > >> * passes.c (pass_manager::register_pass_name): Use > > >> free_string_hash. > > >> (pass_manager::~pass_manager): Delete allocated > > >> m_name_to_pass_map. > > My concern (and what I hadn't had time to dig into) was we initially > > used nofree_string_hash -- I wanted to make sure there wasn't any path > > where the name came from the stack (can't be free'd), was saved > > elsewhere (danging pointer) and the like. ie, why were we using > > nofree_string_hash to begin with? I've never really mucked around with > > these bits, so the analysis side kept falling off the daily todo list. The only occurrences of m_name_to_pass_map are in pass-manager.h (where it is defined as a private field of the class pass_manager) and in passes.cc. There is just one instance where a name is added to the map in passes.cc, namely through the put method. There, the name has been xstrdup'ed. The name (as a const char *) escapes the pass map in pass_manager::create_pass_tab through the call to m_name_pass_map->traverse. This inserts the name into the pass_tab, which is a static vec of const char *s. The pass_tab does not escape the translation unit of passes.c. It is used in dump_one_pass where the name is used as an argument to fprintf. The important point is that it is not freed and not further copied. > > > > If/once you're comfortable with the patch David, then go ahead and appl= y > > it on Marc's behalf. > > > > jeff > >