From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 76608 invoked by alias); 23 Aug 2018 16:01:36 -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 76274 invoked by uid 89); 23 Aug 2018 16:01:35 -0000 Authentication-Results: sourceware.org; auth=none X-Spam-SWARE-Status: No, score=-24.5 required=5.0 tests=AWL,BAYES_00,GIT_PATCH_0,GIT_PATCH_1,GIT_PATCH_2,GIT_PATCH_3,KAM_NUMSUBJECT,SPF_HELO_PASS,SPF_PASS autolearn=ham version=3.3.2 spammy=attaching, UD:all_image.notifier, sk:all_ima, solib X-HELO: smtp.polymtl.ca Received: from smtp.polymtl.ca (HELO smtp.polymtl.ca) (132.207.4.11) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Thu, 23 Aug 2018 16:01:34 +0000 Received: from simark.ca (simark.ca [158.69.221.121]) (authenticated bits=0) by smtp.polymtl.ca (8.14.7/8.14.7) with ESMTP id w7NG1RK3029534 (version=TLSv1/SSLv3 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT) for ; Thu, 23 Aug 2018 12:01:32 -0400 Received: by simark.ca (Postfix, from userid 112) id 6FFD61EB3F; Thu, 23 Aug 2018 12:01:27 -0400 (EDT) Received: from simark.ca (localhost [127.0.0.1]) by simark.ca (Postfix) with ESMTP id 016F01E012; Thu, 23 Aug 2018 12:01:24 -0400 (EDT) MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII; format=flowed Content-Transfer-Encoding: 7bit Date: Thu, 23 Aug 2018 16:01:00 -0000 From: Simon Marchi To: Tom Tromey Cc: gdb-patches@sourceware.org, roirand@adacore.com Subject: Re: [RFC] Make "run" work on macOS 10.13 In-Reply-To: <20180629205532.25377-1-tom@tromey.com> References: <20180629205532.25377-1-tom@tromey.com> Message-ID: X-Sender: simon.marchi@polymtl.ca User-Agent: Roundcube Webmail/1.3.6 X-IsSubscribed: yes X-SW-Source: 2018-08/txt/msg00567.txt.bz2 On 2018-06-29 16:55, Tom Tromey wrote: > I would like some feedback on this patch. > > On macOS 10.13.5, "run" does not work in gdb. There are two cases: > > 1. If I forget to "set startup-with-shell off", then gdb will fail due > to the system integrity protection feature. I believe this happens > because gdb is not allowed to debug the shell. > > You can find many sites advocating "set startup-with-shell off", > but it seems to me that it is friendlier for gdb to simply do it by > default. > > One option here might be to do this conditionally based on the > version of the OS. > > 2. I found that gdb was setting the solib breakpoint incorrectly, > causing a failure. Adding the load address to the notifier address > makes this work for me. I suspect this would regress earlier > versions of macOS, but I have no way to test that; one idea might > be to only do this when gdb_dyld_all_image_infos::version == 15. > > gdb/ChangeLog > 2018-06-29 Tom Tromey > > * solib-darwin.c (darwin_solib_create_inferior_hook): Create solib > breakpoint later. Add load_addr to the notifier address. > * darwin-nat.c (darwin_nat_target::create_inferior): Bind > startup_with_shell to 0. > --- > gdb/ChangeLog | 7 +++++++ > gdb/darwin-nat.c | 5 +++++ > gdb/solib-darwin.c | 9 +++++---- > 3 files changed, 17 insertions(+), 4 deletions(-) > > diff --git a/gdb/ChangeLog b/gdb/ChangeLog > index 4c04d0ba728..c6462259fe0 100644 > --- a/gdb/ChangeLog > +++ b/gdb/ChangeLog > @@ -1,3 +1,10 @@ > +2018-06-29 Tom Tromey > + > + * solib-darwin.c (darwin_solib_create_inferior_hook): Create solib > + breakpoint later. Add load_addr to the notifier address. > + * darwin-nat.c (darwin_nat_target::create_inferior): Bind > + startup_with_shell to 0. > + > 2018-06-28 Tom Tromey > > * NEWS: Mention --enable-codesign. > diff --git a/gdb/darwin-nat.c b/gdb/darwin-nat.c > index 7dccce73926..542c8389ef0 100644 > --- a/gdb/darwin-nat.c > +++ b/gdb/darwin-nat.c > @@ -1809,6 +1809,11 @@ darwin_nat_target::create_inferior (const char > *exec_file, > const std::string &allargs, > char **env, int from_tty) > { > + /* Starting with Sierra, SIP prevents gdb from attaching to the > + shell, so users have to disable startup-with-shell. */ > + scoped_restore save_startup > + = make_scoped_restore (&startup_with_shell, 0); > + > /* Do the hard work. */ > fork_inferior (exec_file, allargs, env, darwin_ptrace_me, I think this part is good. I would suggest printing a message/warnings to indicate that we are disabling startup-with-shell (only if startup_with_shell is 1 in the first place). > darwin_ptrace_him, darwin_pre_ptrace, NULL, > diff --git a/gdb/solib-darwin.c b/gdb/solib-darwin.c > index ed8e0c13365..a4e15dc6b5b 100644 > --- a/gdb/solib-darwin.c > +++ b/gdb/solib-darwin.c > @@ -528,10 +528,6 @@ darwin_solib_create_inferior_hook (int from_tty) > return; > } > > - /* Add the breakpoint which is hit by dyld when the list of solib is > - modified. */ > - create_solib_event_breakpoint (target_gdbarch (), > info->all_image.notifier); > - > if (info->all_image.count != 0) > { > /* Possible relocate the main executable (PIE). */ > @@ -547,6 +543,11 @@ darwin_solib_create_inferior_hook (int from_tty) > load_addr = darwin_read_exec_load_addr_at_init (info); > } > > + /* Add the breakpoint which is hit by dyld when the list of solib is > + modified. */ > + create_solib_event_breakpoint (target_gdbarch (), > + info->all_image.notifier + load_addr); > + > if (load_addr != 0 && symfile_objfile != NULL) > { > CORE_ADDR vmaddr; About the dynamic loader relocation, I am trying to compare your approach with Xavier's approach here: https://sourceware.org/ml/gdb-patches/2018-08/msg00519.html If I print the resulting notifier address, I get two different values: With Tom's patch: 0x10000f782 With Xavier's patch: 0x100012782 The unrelocated value of the symbol is 0xf782. That breakpoint is used for "set stop-on-solib-events", it seems, so I tried to enable that with both of your patches. I got a stop with Xavier's patch and none with Tom's, which leads me to think that Xavier's patch gets it right. I think you may be using the executable base address, while we actually want to use dyld's base address? This is not very clear to me yet. Simon