From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from simark.ca by simark.ca with LMTP id 0DFLJandv1/4MAAAWB0awg (envelope-from ) for ; Thu, 26 Nov 2020 11:54:01 -0500 Received: by simark.ca (Postfix, from userid 112) id 8D0051F0AB; Thu, 26 Nov 2020 11:54:01 -0500 (EST) X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on simark.ca X-Spam-Level: X-Spam-Status: No, score=-1.0 required=5.0 tests=MAILING_LIST_MULTI autolearn=ham autolearn_force=no version=3.4.2 Received: from 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 RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by simark.ca (Postfix) with ESMTPS id 7E9171E552 for ; Thu, 26 Nov 2020 11:54:00 -0500 (EST) Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id F14E3396EC7E; Thu, 26 Nov 2020 16:53:59 +0000 (GMT) Received: from simark.ca (simark.ca [158.69.221.121]) by sourceware.org (Postfix) with ESMTPS id 65FC6388A422 for ; Thu, 26 Nov 2020 16:53:57 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.3.2 sourceware.org 65FC6388A422 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=simark.ca Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=simark@simark.ca Received: from [10.0.0.11] (173-246-6-90.qc.cable.ebox.net [173.246.6.90]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by simark.ca (Postfix) with ESMTPSA id 075841E552; Thu, 26 Nov 2020 11:53:56 -0500 (EST) Subject: Re: [PATCH v2] Search for DWZ files in debug-file-directories as well To: Sergio Durigan Junior , gdb-patches@sourceware.org References: <20201114234842.2334396-1-sergiodj@sergiodj.net> <20201119022708.3627287-1-sergiodj@sergiodj.net> From: Simon Marchi Message-ID: Date: Thu, 26 Nov 2020 11:53:56 -0500 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.12.0 MIME-Version: 1.0 In-Reply-To: <20201119022708.3627287-1-sergiodj@sergiodj.net> Content-Type: text/plain; charset=utf-8 Content-Language: fr Content-Transfer-Encoding: 7bit X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Mark Wielaard Errors-To: gdb-patches-bounces@sourceware.org Sender: "Gdb-patches" On 2020-11-18 9:27 p.m., Sergio Durigan Junior via Gdb-patches wrote: > Changes from v1: > > - Addressed Simon's comments (new comment explaining how we try to > match for the current debug-file-directory; properly use > .erase/.insert methods -- keeping in mind that they modify the > string in-place). > > > When Debian (and Ubuntu) builds its binaries, it (still) doesn't use > dwz's "--relative" option. This causes their debuginfo files to > carry a .gnu_debugaltlink section containing a full pathname to the > DWZ alt debug file, like this: > > $ readelf -wk /usr/bin/cat > Contents of the .gnu_debugaltlink section: > > Separate debug info file: /usr/lib/debug/.dwz/x86_64-linux-gnu/coreutils.debug > Build-ID (0x14 bytes): > ee 76 5d 71 97 37 ce 46 99 44 32 bb e8 a9 1a ef 99 96 88 db > > Contents of the .gnu_debuglink section: > > Separate debug info file: 06d3bee37b8c7e67b31cb2689cb351102ae73b.debug > CRC value: 0x53267655 > > This usually works OK, because most of the debuginfo files installed > via apt will be present in /usr/lib/debug anyway. However, imagine > the following scenario: > > - You are using /usr/bin/cat, it crashes on you and generates a > corefile. > > - You don't want/need to "apt install" the debuginfo file for > coreutils from the repositories. Instead, you already have the > debuginfo files in a separate directory (e.g., $HOME/dbgsym). > > - You start GDB and "set debug-file-directory $HOME/dbgsym". Should the above read $HOME/dbgsym/usr/lib/debug to be consistent with the example below? > diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c > index 3c59826291..c462b9bb2c 100644 > --- a/gdb/dwarf2/read.c > +++ b/gdb/dwarf2/read.c > @@ -2190,7 +2190,7 @@ locate_dwz_sections (bfd *abfd, asection *sectp, dwz_file *dwz_file) > struct dwz_file * > dwarf2_get_dwz_file (dwarf2_per_bfd *per_bfd) > { > - const char *filename; > + std::string filename; > bfd_size_type buildid_len_arg; > size_t buildid_len; > bfd_byte *buildid; > @@ -2216,19 +2216,17 @@ dwarf2_get_dwz_file (dwarf2_per_bfd *per_bfd) > > filename = data.get (); Declare the filename variable here. > > - std::string abs_storage; > - if (!IS_ABSOLUTE_PATH (filename)) > + if (!IS_ABSOLUTE_PATH (filename.c_str ())) > { > gdb::unique_xmalloc_ptr abs > = gdb_realpath (bfd_get_filename (per_bfd->obfd)); > > - abs_storage = ldirname (abs.get ()) + SLASH_STRING + filename; > - filename = abs_storage.c_str (); > + filename = ldirname (abs.get ()) + SLASH_STRING + filename; > } > > /* First try the file name given in the section. If that doesn't > work, try to use the build-id instead. */ > - gdb_bfd_ref_ptr dwz_bfd (gdb_bfd_open (filename, gnutarget)); > + gdb_bfd_ref_ptr dwz_bfd (gdb_bfd_open (filename.c_str (), gnutarget)); > if (dwz_bfd != NULL) > { > if (!build_id_verify (dwz_bfd.get (), buildid_len, buildid)) > @@ -2238,6 +2236,69 @@ dwarf2_get_dwz_file (dwarf2_per_bfd *per_bfd) > if (dwz_bfd == NULL) > dwz_bfd = build_id_to_debug_bfd (buildid_len, buildid); > > + if (dwz_bfd == nullptr) > + { > + /* If the user has provided us with different > + debug-file-directories, we can try them in order. */ > + size_t dwz_pos = filename.find ("/.dwz/"); > + > + if (dwz_pos != std::string::npos) > + { > + std::vector> debugdir_vec > + = dirnames_to_char_ptr_vec (debug_file_directory); > + > + for (const gdb::unique_xmalloc_ptr &debugdir : debugdir_vec) > + { > + /* The idea is to iterate over the > + debug-file-directories provided by the user and > + replace the hard-coded path in the "filename" by each > + debug-file-directory. > + > + For example, suppose that filename is: > + > + /usr/lib/debug/.dwz/foo.dwz > + > + And suppose that we have "$HOME/bar" as the > + debug-file-directory. We would then adjust filename > + to look like: > + > + $HOME/bar/.dwz/foo.dwz > + > + which would hopefully allow us to find the alt debug > + file. */ > + std::string ddir = debugdir.get (); > + > + /* Check whether the beginning of FILENAME is DDIR. If > + it is, then we are dealing with a file which we > + already attempted to open before, so we just skip it > + and continue processing the reamining > + debug-file-directories. */ > + if (filename.size () > ddir.size () > + && filename.compare (0, ddir.size (), ddir) == 0) > + continue; Corner case, but what if file name is /usr/lib/abcde/.dwz/foo.dwz and ddir is /usr/lib/abc? We will wrongfully skip that debug dir I guess? Filesystem paths are hard :(. Is there some "is parent of" function we can use? Worst case, we skip that check and do an unnecessary attempt. > + > + /* Replace FILENAME's default debug-file-directory with > + DDIR. */ > + std::string new_filename = filename; > + new_filename.erase (0, dwz_pos); > + new_filename.insert (0, ddir); I think it would be more readable and efficient (less bytes copied around) to do just build the new_filename string with something like: std::string new_filename = debugdir + &filename[dwz_pos]; This is untested, but I think you'll get the point. &filename[dwz_pos] could also be put into a variable with a meaningful name, for clarity (though I can't think of a good name right now). And we probably don't need the allocated `ddir` std::string, you construct it but never really use it (you could just refer to debugdir). > + > + dwz_bfd = gdb_bfd_open (new_filename.c_str (), gnutarget); > + > + if (dwz_bfd != nullptr) > + { > + if (!build_id_verify (dwz_bfd.get (), buildid_len, buildid)) > + { > + dwz_bfd.reset (nullptr); Just ".reset ()". > + continue; > + } > + /* Found it. */ > + break; > + } Try to minimize the indentation levels where possible, by using early returns or continues. For example, here: if (dwz_bfd == nullptr) continue; if (!build_id_verify (...)) { dwz_bfd.reset (); continue; } /* Found it. */ break; Simon