From nobody Thu Feb 20 23:49:57 2025 X-Original-To: freebsd-hackers@mlmmj.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mlmmj.nyi.freebsd.org (Postfix) with ESMTP id 4YzVPK0kRBz5pbbk for ; Thu, 20 Feb 2025 23:50:13 +0000 (UTC) (envelope-from kib@freebsd.org) Received: from kib.kiev.ua (kib.kiev.ua [IPv6:2001:470:d5e7:1::1]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 4YzVPH3KSLz4332 for ; Thu, 20 Feb 2025 23:50:11 +0000 (UTC) (envelope-from kib@freebsd.org) Authentication-Results: mx1.freebsd.org; dkim=none; dmarc=fail reason="No valid SPF, No valid DKIM" header.from=freebsd.org (policy=none); spf=softfail (mx1.freebsd.org: 2001:470:d5e7:1::1 is neither permitted nor denied by domain of kib@freebsd.org) smtp.mailfrom=kib@freebsd.org Received: from tom.home (kib@localhost [127.0.0.1] (may be forged)) by kib.kiev.ua (8.18.1/8.18.1) with ESMTP id 51KNnvAp082342; Fri, 21 Feb 2025 01:50:00 +0200 (EET) (envelope-from kib@freebsd.org) DKIM-Filter: OpenDKIM Filter v2.10.3 kib.kiev.ua 51KNnvAp082342 Received: (from kostik@localhost) by tom.home (8.18.1/8.18.1/Submit) id 51KNnvFO082341; Fri, 21 Feb 2025 01:49:57 +0200 (EET) (envelope-from kib@freebsd.org) X-Authentication-Warning: tom.home: kostik set sender to kib@freebsd.org using -f Date: Fri, 21 Feb 2025 01:49:57 +0200 From: Konstantin Belousov To: =?utf-8?B?Vmluw61jaXVz?= dos Santos Oliveira Cc: freebsd-hackers@freebsd.org Subject: Re: Capsicum and weak libc symbols Message-ID: References: List-Id: Technical discussions relating to FreeBSD List-Archive: https://lists.freebsd.org/archives/freebsd-hackers List-Help: List-Post: List-Subscribe: List-Unsubscribe: Sender: owner-freebsd-hackers@FreeBSD.org MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: X-Spam-Status: No, score=-2.8 required=5.0 tests=ALL_TRUSTED,BAYES_00, URIBL_SBL_A autolearn=no autolearn_force=no version=4.0.1 X-Spam-Checker-Version: SpamAssassin 4.0.1 (2024-03-26) on tom.home X-Spamd-Result: default: False [-2.70 / 15.00]; NEURAL_HAM_MEDIUM(-1.00)[-1.000]; NEURAL_HAM_LONG(-1.00)[-1.000]; NEURAL_HAM_SHORT(-0.70)[-0.702]; DMARC_POLICY_SOFTFAIL(0.10)[freebsd.org : No valid SPF, No valid DKIM,none]; MIME_GOOD(-0.10)[text/plain]; TAGGED_RCPT(0.00)[]; RCPT_COUNT_TWO(0.00)[2]; ARC_NA(0.00)[]; FREEFALL_USER(0.00)[kib]; ASN(0.00)[asn:6939, ipnet:2001:470::/32, country:US]; TO_DN_SOME(0.00)[]; MIME_TRACE(0.00)[0:+]; HAS_XAW(0.00)[]; R_DKIM_NA(0.00)[]; MLMMJ_DEST(0.00)[freebsd-hackers@freebsd.org]; R_SPF_SOFTFAIL(0.00)[~all]; FROM_EQ_ENVFROM(0.00)[]; FREEMAIL_TO(0.00)[gmail.com]; TO_MATCH_ENVRCPT_SOME(0.00)[]; RCVD_TLS_LAST(0.00)[]; FROM_HAS_DN(0.00)[]; RCVD_COUNT_TWO(0.00)[2]; MISSING_XM_UA(0.00)[] X-Rspamd-Queue-Id: 4YzVPH3KSLz4332 X-Spamd-Bar: -- On Wed, Feb 19, 2025 at 06:57:57AM -0300, Vinícius dos Santos Oliveira wrote: > Em sex., 7 de fev. de 2025 às 00:08, Konstantin Belousov > escreveu: > > On Thu, Feb 06, 2025 at 10:40:52PM -0300, Vinícius dos Santos Oliveira wrote: > > > [...] > > > > The purpose of the weak attribute is to allow the weak symbol to be undefined. > > It is not about the order of resolution (by default our rtld indeed prefers > > non-weak over weak, but it is a bug, and there is knob to turn this behavior > > off). > > > > If you need to interpose symbol, just link the interposing object before the > > object that supplies the symbol to be preempted by your definition. > > There are two use cases for interposing symbols: > > * with rtld > * without rtld (fully static binaries) > > I don't want to change what rtld sees. I don't want to change > /lib/libc.so.7. It works fines as is right now. > > I want to change what /usr/bin/ld.lld sees. I only want the symbols to > be weak in /usr/lib/libc.a. > > For instance, the following code fails to compile when you try to > build a static binary (it works fine on the default mode though): > > > #include > #include > #include > #include > > DIR *opendir(const char *name) > { > if (name[0] == '\0') { > errno = ENOENT; > return NULL; > } > > int fd = open(name, O_RDONLY | O_NONBLOCK | O_DIRECTORY | O_CLOEXEC); > if (fd == -1) { > return NULL; > } > > DIR *ret = fdopendir(fd); > if (!ret) { > int saved_errno = errno; > close(fd); > errno = saved_errno; > } > return ret; > } > > int main(int argc, char *argv[]) > { > (void)argc; > (void)argv; > } > > > The build output for for clang -static will be: > > > ld: error: duplicate symbol: opendir > >>> defined at t2.c > >>> /tmp/t2-3cf81c.o:(opendir) > >>> defined at opendir.c:55 (/usr/src/lib/libc/gen/opendir.c:55) > >>> opendir.o:(.text+0x0) in archive /usr/lib/libc.a > clang: error: linker command failed with exit code 1 (use -v to see invocation) > > > Meanwhile symbols such as unlink are weak (in /usr/lib/libc.a) and I > can interpose them easily in static builds as well. The proper fix there is not to make symbols weak, but to move the opendir() and fdopendir() into dedicated source files, so that they are located in different .o files in libc.a. https://reviews.freebsd.org/D49089 > > However with further investigation I've realised that weak symbols in > /usr/lib/libc.a are also weak in /lib/libc.so.7. Does FreeBSD use the > same weak attributes whether it's building libc.so.7 or libc.a? > > > -- > Vinícius dos Santos Oliveira > https://vinipsmaker.github.io/