From: Pedro Alves <palves@redhat.com>
To: Antoine Tremblay <antoine.tremblay@ericsson.com>,
gdb-patches@sourceware.org
Subject: Re: [PATCH 06/10] Replace breakpoint_reinsert_addr by get_next_pcs operation in GDBServer.
Date: Wed, 04 Nov 2015 18:48:00 -0000 [thread overview]
Message-ID: <563A52F9.9040108@redhat.com> (raw)
In-Reply-To: <1446138583-13268-7-git-send-email-antoine.tremblay@ericsson.com>
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.
>
> 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.)
>
> 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.
> +{
> + /* 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.
> +};
> +
> +#endif /* GET_NEXT_PCS_H */
> diff --git a/gdb/gdbserver/linux-aarch64-low.c b/gdb/gdbserver/linux-aarch64-low.c
> index a0bc1c2..3e21042 100644
> --- a/gdb/gdbserver/linux-aarch64-low.c
> +++ b/gdb/gdbserver/linux-aarch64-low.c
> @@ -2959,7 +2959,7 @@ struct linux_target_ops the_low_target =
> aarch64_set_pc,
> NULL, /* breakpoint_kind_from_pc */
> aarch64_sw_breakpoint_from_kind,
> - NULL, /* breakpoint_reinsert_addr */
> + NULL, /* get_next_pcs */
> 0, /* decr_pc_after_break */
> aarch64_breakpoint_at,
> aarch64_supports_z_point_type,
> diff --git a/gdb/gdbserver/linux-arm-low.c b/gdb/gdbserver/linux-arm-low.c
> index 8e32fae..0b87367 100644
> --- a/gdb/gdbserver/linux-arm-low.c
> +++ b/gdb/gdbserver/linux-arm-low.c
> @@ -1027,7 +1027,7 @@ struct linux_target_ops the_low_target = {
> arm_set_pc,
> arm_breakpoint_kind_from_pc,
> arm_sw_breakpoint_from_kind,
> - NULL, /* breakpoint_reinsert_addr */
> + NULL, /* get_next_pcs */
> 0,
> arm_breakpoint_at,
> arm_supports_z_point_type,
> diff --git a/gdb/gdbserver/linux-bfin-low.c b/gdb/gdbserver/linux-bfin-low.c
> index 912d253..1de2f78 100644
> --- a/gdb/gdbserver/linux-bfin-low.c
> +++ b/gdb/gdbserver/linux-bfin-low.c
> @@ -141,7 +141,7 @@ struct linux_target_ops the_low_target = {
> bfin_set_pc,
> NULL, /* breakpoint_kind_from_pc */
> bfin_sw_breakpoint_from_kind,
> - NULL, /* breakpoint_reinsert_addr */
> + NULL, /* get_next_pcs */
> 2,
> bfin_breakpoint_at,
> NULL, /* supports_z_point_type */
> diff --git a/gdb/gdbserver/linux-cris-low.c b/gdb/gdbserver/linux-cris-low.c
> index 9f4519c..6902a45 100644
> --- a/gdb/gdbserver/linux-cris-low.c
> +++ b/gdb/gdbserver/linux-cris-low.c
> @@ -139,7 +139,7 @@ struct linux_target_ops the_low_target = {
> cris_set_pc,
> NULL, /* breakpoint_kind_from_pc */
> cris_sw_breakpoint_from_kind,
> - NULL, /* breakpoint_reinsert_addr */
> + NULL, /* get_next_pcs */
> 0,
> cris_breakpoint_at,
> };
> diff --git a/gdb/gdbserver/linux-crisv32-low.c b/gdb/gdbserver/linux-crisv32-low.c
> index 2404d0e..28c1981 100644
> --- a/gdb/gdbserver/linux-crisv32-low.c
> +++ b/gdb/gdbserver/linux-crisv32-low.c
> @@ -422,7 +422,7 @@ struct linux_target_ops the_low_target = {
> cris_set_pc,
> NULL, /* breakpoint_kind_from_pc */
> cris_sw_breakpoint_from_kind,
> - NULL, /* breakpoint_reinsert_addr */
> + NULL, /* get_next_pcs */
> 0,
> cris_breakpoint_at,
> cris_supports_z_point_type,
> diff --git a/gdb/gdbserver/linux-low.c b/gdb/gdbserver/linux-low.c
> index a3b0f3a..1f423ba 100644
> --- a/gdb/gdbserver/linux-low.c
> +++ b/gdb/gdbserver/linux-low.c
> @@ -54,6 +54,7 @@
> #include <elf.h>
> #endif
> #include "nat/linux-namespaces.h"
> +#include "common/get-next-pcs.h"
>
> #ifndef SPUFS_MAGIC
> #define SPUFS_MAGIC 0x23c9b64e
> @@ -281,12 +282,12 @@ can_hardware_single_step (void)
> }
>
> /* True if the low target can software single-step. Such targets
> - implement the BREAKPOINT_REINSERT_ADDR callback. */
> + implement the GET_NEXT_PCS callback. */
>
> static int
> can_software_single_step (void)
> {
> - return (the_low_target.breakpoint_reinsert_addr != NULL);
> + return (the_low_target.get_next_pcs != NULL);
> }
>
> /* True if the low target supports memory breakpoints. If so, we'll
> @@ -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.
> + 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?
> + (*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.
> +
> + VEC_free (CORE_ADDR, next_pcs.result);
> +}
> +
Thanks,
Pedro Alves
next prev parent reply other threads:[~2015-11-04 18:48 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 04/10] Remove support for thread events without TRACE_EVENT_CLONE " 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 17:15 ` [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 10/10] Enable conditional breakpoints for targets that support software single step in GDBServer Antoine Tremblay
2015-11-04 18:58 ` Pedro Alves
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 [this message]
2015-11-05 14:35 ` Antoine Tremblay
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=563A52F9.9040108@redhat.com \
--to=palves@redhat.com \
--cc=antoine.tremblay@ericsson.com \
--cc=gdb-patches@sourceware.org \
/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