From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 23015 invoked by alias); 19 Jun 2013 14:10:51 -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 23000 invoked by uid 89); 19 Jun 2013 14:10:49 -0000 X-Spam-SWARE-Status: No, score=-8.1 required=5.0 tests=AWL,BAYES_00,KHOP_THREADED,RCVD_IN_HOSTKARMA_W,RCVD_IN_HOSTKARMA_WL,RP_MATCHES_RCVD,SPF_HELO_PASS,SPF_PASS,TW_RG autolearn=ham version=3.3.1 Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.84/v0.84-167-ge50287c) with ESMTP; Wed, 19 Jun 2013 14:10:49 +0000 Received: from int-mx02.intmail.prod.int.phx2.redhat.com (int-mx02.intmail.prod.int.phx2.redhat.com [10.5.11.12]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id r5JEAkR8025864 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 19 Jun 2013 10:10:46 -0400 Received: from [127.0.0.1] (ovpn01.gateway.prod.ext.ams2.redhat.com [10.39.146.11]) by int-mx02.intmail.prod.int.phx2.redhat.com (8.13.8/8.13.8) with ESMTP id r5JEAilQ016926; Wed, 19 Jun 2013 10:10:45 -0400 Message-ID: <51C1BBE4.5080107@redhat.com> Date: Wed, 19 Jun 2013 14:53:00 -0000 From: Pedro Alves User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130514 Thunderbird/17.0.6 MIME-Version: 1.0 To: lgustavo@codesourcery.com CC: "'gdb-patches@sourceware.org'" Subject: Re: [PATCH, gdbsim] Avoid silly crash when no binary is loaded References: <51C0C7E3.1030603@codesourcery.com> <51C193AE.7010608@redhat.com> <51C19FF0.8000005@codesourcery.com> In-Reply-To: <51C19FF0.8000005@codesourcery.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-SW-Source: 2013-06/txt/msg00492.txt.bz2 On 06/19/2013 01:11 PM, Luis Machado wrote: > Hi, > > On 06/19/2013 08:19 AM, Pedro Alves wrote: >> On 06/18/2013 09:49 PM, Luis Machado wrote: >>> Hi, >>> >>> This patch prevents the long-standing crash scenario where we start >>> gdbsim and "run" without any binaries. Warnings are issued, but those >>> don't prevent the simulator from proceeding with garbage data. >> >> Which sim and backtrace? I suspect this to be sim/arch dependent. > > This is arm. Other simulators (mips and powerpc) have different > behaviors. No crashes, but they go all over the place in terms of messages. > > I'm questioning the use case of attempting to let the simulator go > without loading any image to it. If it is useful, then we should state > that and make it stop crashing. I don't really know. All I see is that from the code at it was supported at least at some point. > > There is already a barrier, see > remote-sim.c:gdbsim_xfer_inferior_memory. The same message will be > displayed there with an error. > > if (!sim_data->program_loaded) > error (_("No program loaded.")); > > So, in a way, we're already preventing this scenario later on. If we > want to keep the old behavior, for whatever old reason that may be, i'm > ok with it. > > #0 0x00000000006a0580 in ARMul_SetPC (state=0x0, value=0) at > ../../../gdb-head/sim/arm/armsupp.c:83 Curious. 'state' is initialized by the ARM sim's 'init' function in the same file, and init is called only by sim_write, sim_read, sim_store_register and sim_fetch_register. 'init' ends up getting called by "load", through sim_load -> sim_load_file -> sim_write. > #1 0x0000000000690cef in sim_create_inferior (sd=0x1, abfd=0x0, > argv=0x0, env=0xc21d90) at ../../../gdb-head/sim/arm/wrapper.c:249 > #2 0x0000000000456a93 in gdbsim_create_inferior (target=0xb58100 > , exec_file=0x0, args=0xc39df0 "", env=0xc21d90, from_tty=1) > at ../../gdb-head/gdb/remote-sim.c:646 >>> >>> Replacing those warnings with error calls seems to be the most >>> appropriate here. >> >> Well, the code seems to have been written like that for a reason. >> >> Real boards can be powered on with no real program in memory >> too... >> > > Of course. The question is if there is any useful use case of letting > the simulator run without any images loaded. I'll leave that up to Mike. > >>> if (exec_file == 0 || exec_bfd == 0) >>> - warning (_("No executable file specified.")); >>> + error (_("No executable file specified.")); >>> if (!sim_data->program_loaded) >>> - warning (_("No program loaded.")); >>> + error (_("No program loaded.")); >>> >> >> There's code just below that does: >> >>> if (remote_debug) >>> printf_filtered ("gdbsim_create_inferior: exec_file \"%s\", args \"%s\"\n", >> ... >>> if (exec_file != NULL) >>> { >>> len = strlen (exec_file) + 1 + strlen (args) + 1 + /*slop */ 10; >>> arg_buf = (char *) alloca (len); >>> arg_buf[0] = '\0'; >>> strcat (arg_buf, exec_file); >>> strcat (arg_buf, " "); >>> strcat (arg_buf, args); >>> argv = gdb_buildargv (arg_buf); >>> make_cleanup_freeargv (argv); >>> } >>> else >>> argv = NULL; >> >> So if we error out, then these NULL checks are now dead. >> > > Right. This may turn to be dead code and may need removal. I have no doubt it ends up as dead code. ;-) The patch just looks obviously incomplete as is, and that prompted my reply. > Is there a good reason why bfin would allow things to proceed without > any image? It doesn't even run past that point really. > > All i see, for whatever operation, is "No memory". Leaving to Mike. I just picked bfin because it's a maintained sim. > ppc gives me "No program loaded", "The program is not being run" and > "The program has no registers now" > > mips says "sim_monitor: unhandled reason = 0, pc = 0xbfc00000", then > falls into the old "Cannot execute this command while the selected > thread is running" or "sim-events.c:231: assertion failed - > events->resume_wallclock == 0". > > If running, and by that i mean issuing run/start/continue/step commands, > the simulators with no image is a valid use case, then sounds like > steering the arm simulator to just do more or less what the other > simulators do is the right thing. > > If the use case is not useful at all, i think we should just wipe it out > rather than preserve some old unclear feature. Thank you -- all this analysis is much clearer and a stronger rationale than the original "silly", or just calling out that things seem appropriate with no backing. ;-) > Then again, these simulators are old and not used that often. -- Pedro Alves