From: Antoine Tremblay <antoine.tremblay@ericsson.com>
To: Pedro Alves <palves@redhat.com>, <gdb-patches@sourceware.org>
Subject: Re: [PATCH 06/10] Replace breakpoint_reinsert_addr by get_next_pcs operation in GDBServer.
Date: Thu, 05 Nov 2015 14:35:00 -0000 [thread overview]
Message-ID: <563B6911.8090903@ericsson.com> (raw)
In-Reply-To: <563A52F9.9040108@redhat.com>
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 <http://www.gnu.org/licenses/>. */
>> +
>> +#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
next prev parent reply other threads:[~2015-11-05 14:35 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-29 17:14 [PATCH 0/10] Support software single step and conditional breakpoints on ARM " Antoine Tremblay
2015-10-29 17:15 ` [PATCH 10/10] Enable conditional breakpoints for targets that support software single step " Antoine Tremblay
2015-11-04 18:58 ` Pedro Alves
2015-10-29 17:15 ` [PATCH 07/10] Share some ARM target dependant code from GDB with GDBServer Antoine Tremblay
2015-11-04 19:56 ` Pedro Alves
2015-11-05 14:44 ` Antoine Tremblay
2015-10-29 17:15 ` [PATCH 04/10] Remove support for thread events without TRACE_EVENT_CLONE in GDBServer Antoine Tremblay
2015-11-03 17:05 ` Yao Qi
2015-11-03 17:24 ` Antoine Tremblay
2015-11-03 17:27 ` [PATCH v2] " Antoine Tremblay
2015-11-04 18:19 ` [PATCH 04/10] " Pedro Alves
2015-11-05 11:43 ` Antoine Tremblay
2015-10-29 17:16 ` [PATCH 09/10] Enable software single stepping for while-stepping actions " Antoine Tremblay
2015-11-04 18:55 ` Pedro Alves
2015-10-29 17:28 ` [PATCH 03/10] Refactor queries for hardware and software single stepping support " Antoine Tremblay
2015-11-04 18:47 ` Pedro Alves
2015-10-29 17:31 ` [PATCH 05/10] Remove too simple breakpoint_reinsert_addr implementations Antoine Tremblay
2015-11-04 18:21 ` Pedro Alves
2015-10-29 17:36 ` [PATCH 01/10] Fix breakpoint size when stepping over a permanent breakpoint in GDBServer Antoine Tremblay
2015-11-03 16:22 ` Yao Qi
2015-11-03 17:05 ` Antoine Tremblay
2015-10-29 17:43 ` [PATCH 08/10] Support software single step on ARM " Antoine Tremblay
2015-11-04 18:46 ` Pedro Alves
2015-10-29 17:45 ` [PATCH 06/10] Replace breakpoint_reinsert_addr by get_next_pcs operation " Antoine Tremblay
2015-11-04 18:48 ` Pedro Alves
2015-11-05 14:35 ` Antoine Tremblay [this message]
2015-10-29 17:48 ` [PATCH 02/10] Fix instruction skipping when using software single step " Antoine Tremblay
2015-11-04 18:08 ` Pedro Alves
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=563B6911.8090903@ericsson.com \
--to=antoine.tremblay@ericsson.com \
--cc=gdb-patches@sourceware.org \
--cc=palves@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox