From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 61815 invoked by alias); 4 May 2017 10:19:36 -0000 Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org Received: (qmail 61795 invoked by uid 89); 4 May 2017 10:19:35 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-26.9 required=5.0 tests=BAYES_00,GIT_PATCH_0,GIT_PATCH_1,GIT_PATCH_2,GIT_PATCH_3,RP_MATCHES_RCVD,SPF_HELO_PASS autolearn=ham version=3.3.2 spammy=releasing, 4856, WHICH X-HELO: mx1.redhat.com Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Thu, 04 May 2017 10:19:34 +0000 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 4B35C4E4C6; Thu, 4 May 2017 10:19:35 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com 4B35C4E4C6 Authentication-Results: ext-mx09.extmail.prod.ext.phx2.redhat.com; dmarc=none (p=none dis=none) header.from=redhat.com Authentication-Results: ext-mx09.extmail.prod.ext.phx2.redhat.com; spf=pass smtp.mailfrom=palves@redhat.com DKIM-Filter: OpenDKIM Filter v2.11.0 mx1.redhat.com 4B35C4E4C6 Received: from [127.0.0.1] (ovpn04.gateway.prod.ext.ams2.redhat.com [10.39.146.4]) by smtp.corp.redhat.com (Postfix) with ESMTP id 7BB634FA45; Thu, 4 May 2017 10:19:34 +0000 (UTC) Subject: Re: [PATCH obv] Remove some superfluous code in corelow.c To: Andreas Arnez , gdb-patches@sourceware.org References: From: Pedro Alves Message-ID: Date: Thu, 04 May 2017 10:19:00 -0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.4.0 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit X-SW-Source: 2017-05/txt/msg00105.txt.bz2 On 05/04/2017 10:09 AM, Andreas Arnez wrote: > Pushed this as obvious. > > > In corelow.c I stumbled upon an extra semicolon and an xfree of a NULL > pointer. Remove them. > @@ -129,7 +129,7 @@ sniff_core_bfd (bfd *abfd) > { > struct core_fns *cf; > struct core_fns *yummy = NULL; > - int matches = 0;; > + int matches = 0; > > /* Don't sniff if we have support for register sets in > CORE_GDBARCH. */ > @@ -511,7 +511,7 @@ get_core_register_section (struct regcache *regcache, > const char *human_name, > int required) > { > - static char *section_name = NULL; > + static char *section_name; But this variable is static.. > struct bfd_section *section; > bfd_size_type size; > char *contents; > @@ -519,8 +519,6 @@ get_core_register_section (struct regcache *regcache, > && regset->flags & REGSET_VARIABLE_SIZE); > ptid_t ptid = regcache_get_ptid (regcache); > > - xfree (section_name); > - ..so this was releasing the memory allocated in a previous invocation of get_core_register_section, AFAICS. So, this wasn't superfluous, and we now leak. I'm not sure why the variable is static but from [1] I suspect that the reason was just "laziness" to avoid having to think about freeing it on every function return path: https://sourceware.org/ml/gdb-patches/2005-03/msg00237.html The easiest to fix that nowadays is to use a std::string. I don't see a need to xstrdup the section name in the single-threaded case though, and also there's more than one place that computes a multi-threaded section name in the same way, so how about the patch below? >From 7c91ec5530e157c33d6e5edc3dc040d9fba1818b Mon Sep 17 00:00:00 2001 From: Pedro Alves Date: Thu, 4 May 2017 10:47:04 +0100 Subject: [PATCH] fix leak, introduce thread_section_name --- gdb/corelow.c | 76 +++++++++++++++++++++++++++++++++++++++++------------------ 1 file changed, 53 insertions(+), 23 deletions(-) diff --git a/gdb/corelow.c b/gdb/corelow.c index 2266f24..0aff02d 100644 --- a/gdb/corelow.c +++ b/gdb/corelow.c @@ -485,6 +485,51 @@ core_detach (struct target_ops *ops, const char *args, int from_tty) printf_filtered (_("No core file now.\n")); } +/* Build either a single-thread or multi-threaded section name for + PTID. + + If ptid's lwp member is zero, we want to do the single-threaded + thing: look for a section named NAME (as passed to the + constructor). If ptid's lwp member is non-zero, we'll want do the + multi-threaded thing: look for a section named "NAME/LWP", where + LWP is the shortest ASCII decimal representation of ptid's lwp + member. */ + +class thread_section_name +{ +public: + /* NAME is the single-threaded section name. If PTID represents an + LWP, then the build section name is "NAME/LWP", otherwise it's + just "NAME" unmodified. */ + thread_section_name (const char *name, ptid_t ptid) + { + if (ptid.lwp_p ()) + { + m_storage = string_printf ("%s/%ld", name, ptid.lwp ()); + m_section_name = m_storage.c_str (); + } + else + m_section_name = name; + } + + /* Return the computed section name. The result is valid as long as + this thread_section_name object is live. */ + const char *c_str () const + { return m_section_name; } + + /* Disable copy. */ + thread_section_name (const thread_section_name &) = delete; + void operator= (const thread_section_name &) = delete; + +private: + /* Either a pointer into M_STORAGE, or a pointer to the name passed + as parameter to the constructor. */ + const char *m_section_name; + /* If we need to build a new section name, this is where we store + it. */ + std::string m_storage; +}; + /* Try to retrieve registers from a section in core_bfd, and supply them to core_vec->core_read_registers, as the register set numbered WHICH. @@ -511,21 +556,15 @@ get_core_register_section (struct regcache *regcache, const char *human_name, int required) { - static char *section_name; struct bfd_section *section; bfd_size_type size; char *contents; bool variable_size_section = (regset != NULL && regset->flags & REGSET_VARIABLE_SIZE); - ptid_t ptid = regcache_get_ptid (regcache); - if (ptid_get_lwp (ptid)) - section_name = xstrprintf ("%s/%ld", name, - ptid_get_lwp (ptid)); - else - section_name = xstrdup (name); + thread_section_name section_name (name, regcache->ptid ()); - section = bfd_get_section_by_name (core_bfd, section_name); + section = bfd_get_section_by_name (core_bfd, section_name.c_str ()); if (! section) { if (required) @@ -537,13 +576,14 @@ get_core_register_section (struct regcache *regcache, size = bfd_section_size (core_bfd, section); if (size < min_size) { - warning (_("Section `%s' in core file too small."), section_name); + warning (_("Section `%s' in core file too small."), + section_name.c_str ()); return; } if (size != min_size && !variable_size_section) { warning (_("Unexpected size of section `%s' in core file."), - section_name); + section_name.c_str ()); } contents = (char *) alloca (size); @@ -551,7 +591,7 @@ get_core_register_section (struct regcache *regcache, (file_ptr) 0, size)) { warning (_("Couldn't read %s registers from `%s' section in core file."), - human_name, name); + human_name, section_name.c_str ()); return; } @@ -681,18 +721,8 @@ add_to_spuid_list (bfd *abfd, asection *asect, void *list_p) static LONGEST get_core_siginfo (bfd *abfd, gdb_byte *readbuf, ULONGEST offset, ULONGEST len) { - asection *section; - char *section_name; - const char *name = ".note.linuxcore.siginfo"; - - if (ptid_get_lwp (inferior_ptid)) - section_name = xstrprintf ("%s/%ld", name, - ptid_get_lwp (inferior_ptid)); - else - section_name = xstrdup (name); - - section = bfd_get_section_by_name (abfd, section_name); - xfree (section_name); + thread_section_name section_name (".note.linuxcore.siginfo", inferior_ptid); + asection *section = bfd_get_section_by_name (abfd, section_name.c_str ()); if (section == NULL) return -1; -- 2.5.5