From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from simark.ca by simark.ca with LMTP id HpkqGir1w2XxrhYAWB0awg (envelope-from ) for ; Wed, 07 Feb 2024 16:24:58 -0500 Authentication-Results: simark.ca; dkim=pass (1024-bit key; unprotected) header.d=redhat.com header.i=@redhat.com header.a=rsa-sha256 header.s=mimecast20190719 header.b=Ia9m7tf6; dkim-atps=neutral Received: by simark.ca (Postfix, from userid 112) id 586871E0C3; Wed, 7 Feb 2024 16:24:58 -0500 (EST) Received: from server2.sourceware.org (server2.sourceware.org [8.43.85.97]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (prime256v1) server-digest SHA256) (No client certificate requested) by simark.ca (Postfix) with ESMTPS id 38BB31E092 for ; Wed, 7 Feb 2024 16:24:56 -0500 (EST) Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 811E3385829B for ; Wed, 7 Feb 2024 21:24:55 +0000 (GMT) Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by sourceware.org (Postfix) with ESMTPS id 485A13858C5F for ; Wed, 7 Feb 2024 21:24:26 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 485A13858C5F 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 485A13858C5F Authentication-Results: server2.sourceware.org; arc=none smtp.remote-ip=170.10.129.124 ARC-Seal: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1707341069; cv=none; b=vb5S3lwzstrysXZJbeM4V6CxyZIU3x9zYC4xewNmzbn1JHkBsokbEiiBwD+s7STSIKu88/0UcRiuSaBTOQ+1zGvy5TO6mj3aaYAKOMXG6tqhv2Rceok2AM519CbnW1Le5d/lxseC/zlooqEbUauvW5thg63kRNdvEcAY1sHCux8= ARC-Message-Signature: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1707341069; c=relaxed/simple; bh=nIvpKLK4z4GtAcAZKU3b8bO15XRVD4EfMU4lzO9l7pU=; h=DKIM-Signature:MIME-Version:From:Date:Message-ID:Subject:To; b=jtbcH67tquP36XG2SnzmAKOBt3nZQ900+erQl0KWbgg4m16vEWAEjnUW/4FRbzI9Oj1KFbqzKOraWoTYEjEbG4gZAULrzOgg3xYj49Z8XdLTgLun7ctHYoA/1HRT8M/S1E0oX6bfOxJEZQq3Sg5aY/YQy9hk5BMD9TMhXzI13WU= ARC-Authentication-Results: i=1; server2.sourceware.org DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1707341065; 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: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=vY4ZV7hM2p6ff3LOslPRl0vX6P0VxvJPQkuv6c9QNyg=; b=Ia9m7tf6Y9r9nJPkJE13QxmXKYSMyBOZtFIppZGBCb6hCkd4zw3AE8vGMVZvNGkOzR6t63 IhrPyQcjpmoWl1sBrYcqvsRIkzmKO+LTZ2jw1cKDKdirG+ZhU0ZflnPcsF3WFVXh9p2FlC /un7zmTqG7eiObfUVglGTEIrqj4Mxzk= Received: from mail-pg1-f197.google.com (mail-pg1-f197.google.com [209.85.215.197]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-359-A1o2DqBeO1WPb7PzGq7Eyg-1; Wed, 07 Feb 2024 16:24:24 -0500 X-MC-Unique: A1o2DqBeO1WPb7PzGq7Eyg-1 Received: by mail-pg1-f197.google.com with SMTP id 41be03b00d2f7-5cfc2041cdfso841470a12.2 for ; Wed, 07 Feb 2024 13:24:24 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1707341063; x=1707945863; h=content-transfer-encoding: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=vY4ZV7hM2p6ff3LOslPRl0vX6P0VxvJPQkuv6c9QNyg=; b=YPO+LPCkftStKXAwb1xgw8RXomm3ht+YVw8KFdiQQjXuveprA3HONUUatyziwvvMpD /MqHqojYssuxhhuSRpHVCy+QPUtzaYPi6g6EdlJpldhpQfSZQmcuopzHRmZSN/sqp+zm dVF51uFpVgk1dCwnlT4CJbNkKlg6Pndoabcuhx/lIfOmgMk5VYwi8KYKoyC56LtBSoGl uDwi0ZduecP9IJaWIFSEy2ENGIDAmNHzGmVpHbpjqC+fnbUXXgmKACPE/7qQoUsCHfMA DIMNKaL1L1Dsq/GEjbsv9ckk2/qrkgd/synVMBRELECO2px5/WFXyvoJCorzZLAa3B8L gOrg== X-Gm-Message-State: AOJu0YyOCGjZSN/2N837hX52A/OLeyOUck0FiSIoHPNOvz46hPfgHtis DEkURnCE79kWVdQg3UjSQR47FiCsPPshbiBLKNFuE3vQfsY8kYKQTahQBUFnNc80vBAQCCH1SfC jwIMjJrYxmajaqlfIBvoRr3I4rY5ef1oNFGUGFUVPqFXhDhNaA8dfHIBEsMl0qVYgpwA3qWHg7a z0BKIZtDa2P24MsZXDYVjepwVWjgfQj/fIZa+Y+ya4 X-Received: by 2002:a05:6a20:d49a:b0:19c:888b:b74e with SMTP id im26-20020a056a20d49a00b0019c888bb74emr8541621pzb.61.1707341063120; Wed, 07 Feb 2024 13:24:23 -0800 (PST) X-Google-Smtp-Source: AGHT+IEBK8X3hTC4q423ceO65KJZ97N03UDGnJ7MQqs6zTvxJ40/hxSWOrYBmLMPv5YZDpsFH/Ho6f9iYH4gylJW+ME= X-Received: by 2002:a05:6a20:d49a:b0:19c:888b:b74e with SMTP id im26-20020a056a20d49a00b0019c888bb74emr8541595pzb.61.1707341062598; Wed, 07 Feb 2024 13:24:22 -0800 (PST) MIME-Version: 1.0 References: <87o7di6e8h.fsf@redhat.com> <20240119050923.2502614-1-amerey@redhat.com> In-Reply-To: <20240119050923.2502614-1-amerey@redhat.com> From: Aaron Merey Date: Wed, 7 Feb 2024 16:24:11 -0500 Message-ID: Subject: [PING][PATCH 2/4 v3] gdb/progspace: Add reverse safe iterator To: Andrew Burgess Cc: gdb-patches@sourceware.org, Thiago Jung Bauermann X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Spam-Status: No, score=-14.9 required=5.0 tests=BAYES_00, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, RCVD_IN_DNSWL_NONE, RCVD_IN_MSPIKE_H2, RCVD_IN_SORBS_WEB, SPF_HELO_NONE, SPF_NONE, TXREP, T_SCC_BODY_TEXT_LINE 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: gdb-patches@sourceware.org X-Mailman-Version: 2.1.30 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: gdb-patches-bounces+public-inbox=simark.ca@sourceware.org Ping Thanks, Aaron On Fri, Jan 19, 2024 at 12:09=E2=80=AFAM Aaron Merey wr= ote: > > Hi Andrew, > > Thanks for the review. > > On Thu, Jan 18, 2024 at 7:03=E2=80=AFAM Andrew Burgess wrote: > > > > Aaron Merey writes: > > > +using qf_list =3D std::forward_list; > > > +using qf_range =3D iterator_range; > > > +using qf_safe_range =3D basic_safe_range; > > > + > > > /* Sections in an objfile. The section offsets are stored in the > > > OBJFILE. */ > > > > > > @@ -764,6 +770,10 @@ struct objfile > > > reader. */ > > > std::forward_list qf; > > > > I think this should be updated to use qf_list. > > Done. > > > > --- a/gdb/symfile-debug.c > > > +++ b/gdb/symfile-debug.c > > > @@ -85,7 +85,7 @@ objfile::has_partial_symbols () > > > them, then that is an indication that they are in fact availabl= e. Without > > > this function the symbols may have been already read in but the= y also may > > > not be present in this objfile. */ > > > - for (const auto &iter : qf) > > > + for (const auto &iter : qf_safe ()) > > > > I read the commit message a couple of times, but I'm still not 100% sur= e > > I understand what's happening here. Is the change from qf -> qf_safe() > > because calling (in this case has_symbols()) might download the full > > debug info, and in doing do, might delete a previously downloaded > > .debug_info section? > > quick_functions_up made from a separately downloaded .gdb_index is at > first associated with the parent objfile (i.e. the objfile of the shared > library containing a .gnu_debuglink that refers to the separate debug inf= o > which actually contains this .gdb_index section). > > When the separate debuginfo is downloaded (possibly during iteration > over the parent's qf_list), we need to remove the gdb_index qf from > the parent's qf_list. This prevents gdb from later attempting to read > the parent objfile for debug info that is actually in the corresponding > separate debug objfile. > > > If this is the case then is it true that we will almost never want to > > iterate over `qf` directly as most quick_symbol_functions methods might > > result in this download/delete behaviour? > > > > If I'm right, then I think it might be useful to place a warning in the > > comment before `qf` explaining this. > > Done. > > > > +/* A reverse basic_safe_iterator. See basic_safe_iterator for inten= ded use. */ > > > + > > > +template > > > +class basic_safe_reverse_iterator > > > +{ > > > +public: > > > + typedef basic_safe_reverse_iterator self_type; > > > + typedef typename Iterator::value_type value_type; > > > + typedef typename Iterator::reference reference; > > > + typedef typename Iterator::pointer pointer; > > > + typedef typename Iterator::iterator_category iterator_category; > > > + typedef typename Iterator::difference_type difference_type; > > > + > > > + /* Construct the iterator using ARG, and construct the end iterato= r > > > + using ARG2. */ > > > > Maybe this is a standard thing, and I'm just showing my lack of > > experience, but, given the existence of methods like rbegin() and > > rend(), I think the comments here should be doing more to explain what'= s > > expected of this input. > > > > Plus this comment seems confusing, though I can sort of see what it's > > trying to say. > > > > We're creating a *_iterator object, which itself contains other > > iterators. We need to be super clear which particular "iterator" we're > > talking about in the comments. > > > > Also, I think we need to be explicit that for this reverse iterator, th= e > > first argument to the constructor is going to be the LAST element that > > is iterated over. One might be tempted to think that a safe reverse it= erator > > should be initialised with rbegin() and rend(), which isn't going to be > > correct (unless I'm not understand, please correct me if so). > > That's right, the arguments to basic_safe_reverse_iterator should be > forward iterators. This will not work with reverse iterators. > > I added more information to this comment to hopefully clear up any > confusion. > > > > + > > > +/* A range adapter that wraps a forward range, and then returns > > > + safe reverse iterators wrapping the original range's iterators. = */ > > > > I think the fact that this takes a FORWARD range is really important, > > and we should do more to advertise this. E.g. I think there should be = a > > comment on the constructor saying that RANGE is a forward range over > > which we will iterate backwards. > > > > Done. The updated patch is included below. > > Aaron > > Commit message: > > This patch changes progspace objfile_list insertion so that separate > debug objfiles are placed into the list after the parent objfile, > instead of before. Additionally qf_require_partial_symbols now returns > a safe_range. > > These changes are intended to prepare gdb for on-demand debuginfo > downloading and the downloading of .gdb_index sections. > > With on-demand downloading enabled, gdb might need to delete a > .gdb_index quick_symbol_functions from a parent objfile while looping > the objfile's list of quick_symbol_functions because the separate > debug objfile has just been downloaded. The use of a safe_range > prevents this removal from causing iterator invalidation. > > gdb might also download a debuginfo file during symtab expansion. > In this case an objfile will be added to the current progspace's > objfiles_list during iteration over the list (for example, in > iterate_over_symtabs). We want these loops to also iterate over > newly downloaded objfiles. So objfiles need to be inserted into > objfiles_list after their parent since it is during the search of > the parent objfile for some symbol or filename that the separate > debug objfile might be downloaded. > > To facilitate the safe deletion of objfiles, this patch also adds > basic_safe_reverse_range and basic_safe_reverse_iterator. This allows > objfiles to be removed from the objfiles_list in a loop without iterator > invalidation. > > If a forward safe iterator were to be used, the deletion of an > objfile could invalidate the safe iterator's reference to the next > objfile in the objfiles_list. This can happen when the deletion > of an objfile causes the deletion of a separate debug objfile that > happens to the be next element in the objfiles_list. > > The standard reverse iterator is not suitable for safe objfile deletion. > In order to safely delete the first objfile in the objfiles_list, the > standard reverse iterator's underlying begin iterator would have to be > decremented, resulting in undefined behavior. > > A small change was also made to a testcase in py-objfile.exp to > account for the new placement of separate debug objfiles in > objfiles_list. > --- > gdb/jit.c | 7 +- > gdb/objfiles.c | 14 +-- > gdb/objfiles.h | 17 +++- > gdb/progspace.c | 19 +++- > gdb/progspace.h | 31 ++++--- > gdb/symfile-debug.c | 34 +++---- > gdb/testsuite/gdb.python/py-objfile.exp | 2 +- > gdbsupport/safe-iterator.h | 115 ++++++++++++++++++++++++ > 8 files changed, 195 insertions(+), 44 deletions(-) > > diff --git a/gdb/jit.c b/gdb/jit.c > index 85a10be3055..10ee5f9b749 100644 > --- a/gdb/jit.c > +++ b/gdb/jit.c > @@ -1240,11 +1240,10 @@ jit_breakpoint_re_set (void) > static void > jit_inferior_exit_hook (struct inferior *inf) > { > - for (objfile *objf : current_program_space->objfiles_safe ()) > + current_program_space->unlink_objfiles_if ([&] (const objfile *objf) > { > - if (objf->jited_data !=3D nullptr && objf->jited_data->addr !=3D 0= ) > - objf->unlink (); > - } > + return (objf->jited_data !=3D nullptr) && (objf->jited_data->addr = !=3D 0); > + }); > } > > void > diff --git a/gdb/objfiles.c b/gdb/objfiles.c > index 8f085b1bb7c..d53dfaca90b 100644 > --- a/gdb/objfiles.c > +++ b/gdb/objfiles.c > @@ -470,6 +470,12 @@ objfile::unlink () > current_program_space->remove_objfile (this); > } > > +qf_safe_range > +objfile::qf_safe () > +{ > + return qf_safe_range (qf_range (qf.begin (), qf.end ())); > +} > + > /* Free all separate debug objfile of OBJFILE, but don't free OBJFILE > itself. */ > > @@ -793,14 +799,12 @@ have_full_symbols (void) > void > objfile_purge_solibs (void) > { > - for (objfile *objf : current_program_space->objfiles_safe ()) > + current_program_space->unlink_objfiles_if ([&] (const objfile *objf) > { > /* We assume that the solib package has been purged already, or wi= ll > be soon. */ > - > - if (!(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED= )) > - objf->unlink (); > - } > + return !(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHA= RED); > + }); > } > > > diff --git a/gdb/objfiles.h b/gdb/objfiles.h > index 621342d22b7..b917886988d 100644 > --- a/gdb/objfiles.h > +++ b/gdb/objfiles.h > @@ -370,6 +370,12 @@ class separate_debug_iterator > > typedef iterator_range separate_debug_range; > > +/* See objfile::qf_safe. */ > + > +using qf_list =3D std::forward_list; > +using qf_range =3D iterator_range; > +using qf_safe_range =3D basic_safe_range; > + > /* Sections in an objfile. The section offsets are stored in the > OBJFILE. */ > > @@ -761,8 +767,15 @@ struct objfile > const struct sym_fns *sf =3D nullptr; > > /* The "quick" (aka partial) symbol functions for this symbol > - reader. */ > - std::forward_list qf; > + reader. Many quick_symbol_functions methods may result > + in the deletion of a quick_symbol_functions from this > + qf_list. It is recommended that qf_safe be used to iterate > + over the qf_list. */ > + qf_list qf; > + > + /* Returns an iterable object that allows for safe deletion during > + iteration. See gdbsupport/safe-iterator.h. */ > + qf_safe_range qf_safe (); > > /* Per objfile data-pointers required by other GDB modules. */ > > diff --git a/gdb/progspace.c b/gdb/progspace.c > index 0fc1fdd57ab..c4cc2e2e8af 100644 > --- a/gdb/progspace.c > +++ b/gdb/progspace.c > @@ -141,19 +141,19 @@ program_space::free_all_objfiles () > > void > program_space::add_objfile (std::unique_ptr &&objfile, > - struct objfile *before) > + struct objfile *after) > { > - if (before =3D=3D nullptr) > + if (after =3D=3D nullptr) > objfiles_list.push_back (std::move (objfile)); > else > { > auto iter =3D std::find_if (objfiles_list.begin (), objfiles_list.= end (), > [=3D] (const std::unique_ptr<::objfile> &= objf) > { > - return objf.get () =3D=3D before; > + return objf.get () =3D=3D after; > }); > gdb_assert (iter !=3D objfiles_list.end ()); > - objfiles_list.insert (iter, std::move (objfile)); > + objfiles_list.insert (++iter, std::move (objfile)); > } > } > > @@ -182,6 +182,17 @@ program_space::remove_objfile (struct objfile *objfi= le) > > /* See progspace.h. */ > > +void > +program_space::unlink_objfiles_if > + (gdb::function_view predicate) > +{ > + for (auto &it : objfiles_safe ()) > + if (predicate (it.get ())) > + it->unlink (); > +} > + > +/* See progspace.h. */ > + > struct objfile * > program_space::objfile_for_address (CORE_ADDR address) > { > diff --git a/gdb/progspace.h b/gdb/progspace.h > index 163cbf6a0f8..f23a57dc842 100644 > --- a/gdb/progspace.h > +++ b/gdb/progspace.h > @@ -250,28 +250,32 @@ struct program_space > unwrapping_objfile_iterator (objfiles_list.end ())); > } > > - using objfiles_safe_range =3D basic_safe_range; > + using objfiles_it_range =3D iterator_range; > + using objfiles_safe_reverse_range > + =3D basic_safe_reverse_range; > > /* An iterable object that can be used to iterate over all objfiles. > The basic use is in a foreach, like: > > for (objfile *objf : pspace->objfiles_safe ()) { ... } > > - This variant uses a basic_safe_iterator so that objfiles can be > - deleted during iteration. */ > - objfiles_safe_range objfiles_safe () > + This variant uses a basic_safe_reverse_iterator so that objfiles > + can be deleted during iteration. > + > + The use of a reverse iterator helps ensure that separate debug > + objfiles are deleted before their parent objfile. This prevents > + iterator invalidation due to the deletion of a parent objfile. */ > + objfiles_safe_reverse_range objfiles_safe () > { > - return objfiles_safe_range > - (objfiles_range > - (unwrapping_objfile_iterator (objfiles_list.begin ()), > - unwrapping_objfile_iterator (objfiles_list.end ()))); > + return objfiles_safe_reverse_range > + (objfiles_it_range (objfiles_list.begin (), objfiles_list.end ()))= ; > } > > - /* Add OBJFILE to the list of objfiles, putting it just before > - BEFORE. If BEFORE is nullptr, it will go at the end of the > + /* Add OBJFILE to the list of objfiles, putting it just after > + AFTER. If AFTER is nullptr, it will go at the end of the > list. */ > void add_objfile (std::unique_ptr &&objfile, > - struct objfile *before); > + struct objfile *after); > > /* Remove OBJFILE from the list of objfiles. */ > void remove_objfile (struct objfile *objfile); > @@ -286,6 +290,11 @@ struct program_space > /* Free all the objfiles associated with this program space. */ > void free_all_objfiles (); > > + /* Unlink all objfiles associated with this program space for which > + PREDICATE evaluates to true. */ > + void unlink_objfiles_if > + (gdb::function_view predicate); > + > /* Return the objfile containing ADDRESS, or nullptr if the address > is outside all objfiles in this progspace. */ > struct objfile *objfile_for_address (CORE_ADDR address); > diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c > index 32c3ed9d349..c684f9cdd43 100644 > --- a/gdb/symfile-debug.c > +++ b/gdb/symfile-debug.c > @@ -85,7 +85,7 @@ objfile::has_partial_symbols () > them, then that is an indication that they are in fact available. = Without > this function the symbols may have been already read in but they al= so may > not be present in this objfile. */ > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > { > retval =3D iter->has_symbols (this); > if (retval) > @@ -108,7 +108,7 @@ objfile::has_unexpanded_symtabs () > objfile_debug_name (this)); > > bool result =3D false; > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > { > if (iter->has_unexpanded_symtabs (this)) > { > @@ -133,7 +133,7 @@ objfile::find_last_source_symtab () > gdb_printf (gdb_stdlog, "qf->find_last_source_symtab (%s)\n", > objfile_debug_name (this)); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > { > retval =3D iter->find_last_source_symtab (this); > if (retval !=3D nullptr) > @@ -166,7 +166,7 @@ objfile::forget_cached_source_info () > } > } > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > iter->forget_cached_source_info (this); > } > > @@ -213,7 +213,7 @@ objfile::map_symtabs_matching_filename > return result; > }; > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > { > if (!iter->expand_symtabs_matching (this, > match_one_filename, > @@ -278,7 +278,7 @@ objfile::lookup_symbol (block_enum kind, const char *= name, domain_enum domain) > return true; > }; > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > { > if (!iter->expand_symtabs_matching (this, > nullptr, > @@ -309,7 +309,7 @@ objfile::print_stats (bool print_bcache) > gdb_printf (gdb_stdlog, "qf->print_stats (%s, %d)\n", > objfile_debug_name (this), print_bcache); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > iter->print_stats (this, print_bcache); > } > > @@ -320,7 +320,7 @@ objfile::dump () > gdb_printf (gdb_stdlog, "qf->dump (%s)\n", > objfile_debug_name (this)); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > iter->dump (this); > } > > @@ -335,7 +335,7 @@ objfile::expand_symtabs_for_function (const char *fun= c_name) > lookup_name_info base_lookup (func_name, symbol_name_match_type::FULL)= ; > lookup_name_info lookup_name =3D base_lookup.make_ignore_params (); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > iter->expand_symtabs_matching (this, > nullptr, > &lookup_name, > @@ -354,7 +354,7 @@ objfile::expand_all_symtabs () > gdb_printf (gdb_stdlog, "qf->expand_all_symtabs (%s)\n", > objfile_debug_name (this)); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > iter->expand_all_symtabs (this); > } > > @@ -372,7 +372,7 @@ objfile::expand_symtabs_with_fullname (const char *fu= llname) > return filename_cmp (basenames ? basename : fullname, filename) =3D= =3D 0; > }; > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > iter->expand_symtabs_matching (this, > file_matcher, > nullptr, > @@ -406,7 +406,7 @@ objfile::expand_symtabs_matching > host_address_to_string (&expansion_notify), > search_domain_name (kind)); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > if (!iter->expand_symtabs_matching (this, file_matcher, lookup_name, > symbol_matcher, expansion_notify, > search_flags, domain, kind)) > @@ -431,7 +431,7 @@ objfile::find_pc_sect_compunit_symtab (struct bound_m= inimal_symbol msymbol, > host_address_to_string (section), > warn_if_readin); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > { > retval =3D iter->find_pc_sect_compunit_symtab (this, msymbol, pc, = section, > warn_if_readin); > @@ -459,7 +459,7 @@ objfile::map_symbol_filenames (gdb::function_view fun, > objfile_debug_name (this), > need_fullname); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > iter->map_symbol_filenames (this, fun, need_fullname); > } > > @@ -471,7 +471,7 @@ objfile::compute_main_name () > "qf->compute_main_name (%s)\n", > objfile_debug_name (this)); > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > iter->compute_main_name (this); > } > > @@ -485,7 +485,7 @@ objfile::find_compunit_symtab_by_address (CORE_ADDR a= ddress) > hex_string (address)); > > struct compunit_symtab *result =3D NULL; > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > { > result =3D iter->find_compunit_symtab_by_address (this, address); > if (result !=3D nullptr) > @@ -510,7 +510,7 @@ objfile::lookup_global_symbol_language (const char *n= ame, > enum language result =3D language_unknown; > *symbol_found_p =3D false; > > - for (const auto &iter : qf) > + for (const auto &iter : qf_safe ()) > { > result =3D iter->lookup_global_symbol_language (this, name, domain= , > symbol_found_p); > diff --git a/gdb/testsuite/gdb.python/py-objfile.exp b/gdb/testsuite/gdb.= python/py-objfile.exp > index 61b9942de79..0bf49976b73 100644 > --- a/gdb/testsuite/gdb.python/py-objfile.exp > +++ b/gdb/testsuite/gdb.python/py-objfile.exp > @@ -135,7 +135,7 @@ gdb_test "p main" "=3D {} $hex
" \ > gdb_py_test_silent_cmd "python objfile.add_separate_debug_file(\"${binfi= le}\")" \ > "Add separate debug file file" 1 > > -gdb_py_test_silent_cmd "python sep_objfile =3D gdb.objfiles()\[0\]" \ > +gdb_py_test_silent_cmd "python sep_objfile =3D gdb.objfiles()\[1\]" \ > "Get separate debug info objfile" 1 > > gdb_test "python print (sep_objfile.owner.filename)" "${testfile}2" \ > diff --git a/gdbsupport/safe-iterator.h b/gdbsupport/safe-iterator.h > index b4891328f1a..d1b9ce21ffc 100644 > --- a/gdbsupport/safe-iterator.h > +++ b/gdbsupport/safe-iterator.h > @@ -136,4 +136,119 @@ class basic_safe_range > Range m_range; > }; > > +/* A reverse basic_safe_iterator. See basic_safe_iterator for intended = use. */ > + > +template > +class basic_safe_reverse_iterator > +{ > +public: > + typedef basic_safe_reverse_iterator self_type; > + typedef typename Iterator::value_type value_type; > + typedef typename Iterator::reference reference; > + typedef typename Iterator::pointer pointer; > + typedef typename Iterator::iterator_category iterator_category; > + typedef typename Iterator::difference_type difference_type; > + > + /* Construct the iterator using ARG, and construct the end iterator > + using ARG2. ARG and ARG2 should be forward iterators, typically > + from begin and end methods, respectively. > + > + For example if ARG1 is created with container.begin and ARG2 is > + is created with container.end, then the basic_safe_reverse_iterator > + will traverse from the last element in the container to the first > + element in the container. */ > + template > + explicit basic_safe_reverse_iterator (Arg &&arg, Arg &&arg2) > + : m_begin (std::forward (arg)), > + m_end (std::forward (arg2)), > + m_it (m_end), > + m_next (m_end) > + { > + /* M_IT and M_NEXT are initialized as one-past-end. Set M_IT to poi= nt > + to the last element and set M_NEXT to point to the second last el= ement, > + if such elements exist. */ > + if (m_it !=3D m_begin) > + { > + --m_it; > + > + if (m_it !=3D m_begin) > + { > + --m_next; > + --m_next; > + } > + } > + } > + > + typename std::invoke_result:= :type > + operator* () const > + { return *m_it; } > + > + self_type &operator++ () > + { > + m_it =3D m_next; > + > + if (m_it !=3D m_end) > + { > + /* Use M_BEGIN only if we sure that it is valid. */ > + if (m_it =3D=3D m_begin) > + m_next =3D m_end; > + else > + --m_next; > + } > + > + return *this; > + } > + > + bool operator=3D=3D (const self_type &other) const > + { return m_it =3D=3D other.m_it; } > + > + bool operator!=3D (const self_type &other) const > + { return m_it !=3D other.m_it; } > + > +private: > + /* The first element. */ > + Iterator m_begin {}; > + > + /* A one-past-end iterator. */ > + Iterator m_end {}; > + > + /* The current element. */ > + Iterator m_it {}; > + > + /* The next element. Always one element ahead of M_IT. */ > + Iterator m_next {}; > +}; > + > +/* A range adapter that wraps a forward range, and then returns > + safe reverse iterators wrapping the original range's iterators. */ > + > +template > +class basic_safe_reverse_range > +{ > +public: > + > + typedef basic_safe_reverse_iterator iterator= ; > + > + /* RANGE must be a forward range. basic_safe_reverse_iterators > + will be used to traverse the forward range from the last element > + to the first. */ > + explicit basic_safe_reverse_range (Range range) > + : m_range (range) > + { > + } > + > + iterator begin () > + { > + return iterator (m_range.begin (), m_range.end ()); > + } > + > + iterator end () > + { > + return iterator (m_range.end (), m_range.end ()); > + } > + > +private: > + > + Range m_range; > +}; > #endif /* COMMON_SAFE_ITERATOR_H */ > -- > 2.43.0 >