From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 43512 invoked by alias); 5 Nov 2015 14:35:14 -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 43499 invoked by uid 89); 5 Nov 2015 14:35:14 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-1.7 required=5.0 tests=AWL,BAYES_00,SPF_PASS autolearn=ham version=3.3.2 X-HELO: usevmg20.ericsson.net Received: from usevmg20.ericsson.net (HELO usevmg20.ericsson.net) (198.24.6.45) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with (AES256-SHA encrypted) ESMTPS; Thu, 05 Nov 2015 14:35:07 +0000 Received: from EUSAAHC002.ericsson.se (Unknown_Domain [147.117.188.78]) by usevmg20.ericsson.net (Symantec Mail Security) with SMTP id F8.CF.32596.AB70B365; Thu, 5 Nov 2015 08:39:38 +0100 (CET) Received: from [142.133.110.95] (147.117.188.8) by smtp-am.internal.ericsson.com (147.117.188.80) with Microsoft SMTP Server id 14.3.248.2; Thu, 5 Nov 2015 09:34:57 -0500 Subject: Re: [PATCH 06/10] Replace breakpoint_reinsert_addr by get_next_pcs operation in GDBServer. To: Pedro Alves , References: <1446138583-13268-1-git-send-email-antoine.tremblay@ericsson.com> <1446138583-13268-7-git-send-email-antoine.tremblay@ericsson.com> <563A52F9.9040108@redhat.com> From: Antoine Tremblay Message-ID: <563B6911.8090903@ericsson.com> Date: Thu, 05 Nov 2015 14:35:00 -0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.3.0 MIME-Version: 1.0 In-Reply-To: <563A52F9.9040108@redhat.com> Content-Type: text/plain; charset="windows-1252"; format=flowed Content-Transfer-Encoding: 7bit X-IsSubscribed: yes X-SW-Source: 2015-11/txt/msg00184.txt.bz2 On 11/04/2015 01:48 PM, Pedro Alves wrote: > On 10/29/2015 05:09 PM, Antoine Tremblay wrote: >> This patch in preparation for software single step support on ARM. It refactors >> breakpoint_reinsert_addr into get_next_pcs so that multiple location can be >> returned. >> >> When software single stepping there can be multiple possible next addresses >> because we're stepping over a branch instruction for example. > > ITYM mean a _conditional_ branch instruction. > Yes, fixed. >> >> The operation get_next_pcs handles that by returning a vector of all the >> possible next addresses. >> >> Software breakpoints are installed at each location returned. >> >> No regressions, tested on ubuntu 14.04 ARMv7 and x86. >> With gdbserver-{native,extended} / { -marm -mthumb } >> >> gdb/ChangeLog: >> >> * common/get-next-pcs.h: New file. >> * common/gdb_vecs.h: Add CORE_ADDR vect. >> * symtab.h: Move CORE_ADDR vect to gdb_vecs.h > > Missing period in last line. (Pedantically, AFAICS, in this patch, > nothing in gdb includes common/get-next-pcs.h, and nothing > in gdbserver includes symtab.h. So seems like putting the VEC > in common/gdb_vecs.h here is premature.) > Yes I could do another patch with just that change since I really wanted it to be out of the big 08/10 Support software single step on ARM in GDBServer patch, but that would kinda remove some context and it would still be premature in the same way. Same applies to 07/10 Share some ARM target dependant code from GDB with GDBServer... I'm keeping it there in the context that this is a preparatory patch unless there's an objection. >> >> gdb/gdbserver/ChangeLog: >> * linux-aarch64-low.c (struct linux_target_ops): Rename >> breakpoint_reinsert_addr to get_next_pcs. >> * linux-arm-low.c (struct linux_target_ops): Likewise. >> * linux-bfin-low.c (struct linux_target_ops): Likewise. >> * linux-cris-low.c (struct linux_target_ops): Likewise. >> * linux-crisv32-low.c (struct linux_target_ops): Likewise. >> * linux-low.c (can_software_single_step): Likewise. >> (install_software_single_step_breakpoints): New function. >> (start_step_over): Use isntall_software_single_step_breakpoints. >> * linux-low.h: Rename breakpoint_reinsert_addr to >> get_next_pcs. >> * linux-mips-low.c (struct linux_target_ops): Likewise. >> * linux-nios2-low.c (struct linux_target_ops): Likewise. >> * linux-sparc-low.c (struct linux_target_ops): Likewise. > > > This LGTM. Some nits below. > >> +++ b/gdb/common/get-next-pcs.h >> @@ -0,0 +1,34 @@ >> +/* Common code for software single stepping support. >> + >> + Copyright (C) 2015 Free Software Foundation, Inc. >> + >> + This file is part of GDB. >> + >> + This program is free software; you can redistribute it and/or modify >> + it under the terms of the GNU General Public License as published by >> + the Free Software Foundation; either version 3 of the License, or >> + (at your option) any later version. >> + >> + This program is distributed in the hope that it will be useful, >> + but WITHOUT ANY WARRANTY; without even the implied warranty of >> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the >> + GNU General Public License for more details. >> + >> + You should have received a copy of the GNU General Public License >> + along with this program. If not, see . */ >> + >> +#ifndef GET_NEXT_PCS_H >> +#define GET_NEXT_PCS_H 1 >> + >> +#include "gdb_vecs.h" >> + >> +struct get_next_pcs > > There should be an intro comment mentioning the purpose of > this structure. > Done as : /* This structure contains the shared context of a get_next_pcs call, as used in linux target operation "get_next_pcs" in GDBServer and the software_single_step architecture dependent function in GDB. */ >> +{ >> + /* Resulting vector of possible next addresses. */ >> + VEC (CORE_ADDR) *result; >> + /* Base PC from which to get the next pcs. */ >> + CORE_ADDR pc; >> + struct regcache* regcache; > > This field should be documented too. > Done. >> @@ -3876,6 +3877,29 @@ enqueue_pending_signal (struct lwp_info *lwp, int signal, siginfo_t *info) >> lwp->pending_signals = p_sig; >> } >> >> +/* Install breakpoints for software single stepping. */ >> + >> +static void >> +install_software_single_step_breakpoints (struct lwp_info *lwp) >> +{ >> + struct get_next_pcs next_pcs; >> + int i; >> + CORE_ADDR pc; >> + >> + struct regcache *regcache = get_thread_regcache (current_thread, 1); > > Formatting. Variable should be declared in the declaration block > along with the others, and then there should be a line break. > Done thanks, >> + next_pcs.pc = get_pc (lwp); >> + next_pcs.regcache = regcache; >> + next_pcs.result = VEC_alloc (CORE_ADDR, 1); > > Is there a reason to allocate upfront instead of setting it to NULL? > No that was a mistake, fixed. >> + (*the_low_target.get_next_pcs) (&next_pcs); >> + >> + for (i = 0; VEC_iterate (CORE_ADDR, next_pcs.result, i, pc); ++i) >> + { >> + set_reinsert_breakpoint (pc); >> + } > > At some point we should do a pass and rename "reinsert breakpoint" > to "software single-step breakpoint" throughout. It's no longer > just about reinserting a breakpoint that we step over. > Ok, noted. Thanks, Antoine Tremblay