From: Yao Qi <qiyaoltc@gmail.com>
To: Antoine Tremblay <antoine.tremblay@ericsson.com>
Cc: Pedro Alves <palves@redhat.com>,
"gdb-patches\@sourceware.org" <gdb-patches@sourceware.org>
Subject: Re: [PATCH 1/2] This patch fixes GDBServer's run control for single stepping
Date: Mon, 03 Apr 2017 15:18:00 -0000 [thread overview]
Message-ID: <8637dpldta.fsf@gmail.com> (raw)
In-Reply-To: <wwokshlplje0.fsf@ericsson.com> (Antoine Tremblay's message of "Mon, 3 Apr 2017 09:18:15 -0400")
Antoine Tremblay <antoine.tremblay@ericsson.com> writes:
> + if ((inst1 & 0xff00) == 0xbf00 && (inst1 & 0x000f) != 0)
> + {
> + /* An IT instruction. Because this instruction does not
> + modify the flags, we can accurately predict the next
> + executed instruction. */
> + itstate = inst1 & 0x00ff;
> + pc += thumb_insn_size (inst1);
> +
> + while (itstate != 0 && ! condition_true (itstate >> 4, status))
> + {
> + inst1 = read_mem_uint (pc, 2,byte_order_for_code);
> + pc += thumb_insn_size (inst1);
> + itstate = thumb_advance_itstate (itstate);
> + }
> + next_pcs.push_back (std::pair<CORE_ADDR, arm_breakpoint_kinds>
> + (MAKE_THUMB_ADDR (pc), ARM_BP_KIND_THUMB2));
It is incorrect to choose ARM_BP_KIND_THUMB2 if the instruction is
16-bit. IMO, this function should only tell whether PC is in IT block
nor not. It shouldn't involve any breakpoint kinds selection.
> + return next_pcs;
> + }
> + else if (itstate != 0)
> + {
> + /* We are in a conditional block. Check the condition. */
> + if (! condition_true (itstate >> 4, status))
> + {
> + /* Advance to the next executed instruction. */
> + pc += thumb_insn_size (inst1);
> + itstate = thumb_advance_itstate (itstate);
> +
> + while (itstate != 0 && ! condition_true (itstate >> 4, status))
> + {
> + inst1 = read_mem_uint (pc, 2, byte_order_for_code);
> +
> + pc += thumb_insn_size (inst1);
> + itstate = thumb_advance_itstate (itstate);
> + }
> +
If all the following instructions' condition is false, breakpoint should
be set on the first instruction out side of IT block. We can still use
16-bit thumb breakpoint.
> + next_pcs.push_back (std::pair<CORE_ADDR, arm_breakpoint_kinds>
> + (MAKE_THUMB_ADDR (pc),
> ARM_BP_KIND_THUMB2));
The same issue.
> + return next_pcs;
> + }
> + else if ((itstate & 0x0f) == 0x08)
> + {
> + /* This is the last instruction of the conditional
> + block, and it is executed. We can handle it normally
> + because the following instruction is not conditional,
> + and we must handle it normally because it is
> + permitted to branch. Fall through. */
How do we fall through now?
> + }
> + else
> + {
> + int cond_negated;
> +
> + /* There are conditional instructions after this one.
> + If this instruction modifies the flags, then we can
> + not predict what the next executed instruction will
> + be. Fortunately, this instruction is archi2tecturally
> + forbidden to branch; we know it will fall through.
> + Start by skipping past it. */
> + pc += thumb_insn_size (inst1);
> + itstate = thumb_advance_itstate (itstate);
> +
> + /* Set a breakpoint on the following instruction. */
> + gdb_assert ((itstate & 0x0f) != 0);
> + next_pcs.push_back (std::pair<CORE_ADDR, arm_breakpoint_kinds>
> + (MAKE_THUMB_ADDR (pc), ARM_BP_KIND_THUMB2));
> +
> + cond_negated = (itstate >> 4) & 1;
> +
> + /* Skip all following instructions with the same
> + condition. If there is a later instruction in the IT
> + block with the opposite condition, set the other
> + breakpoint there. If not, then set a breakpoint on
> + the instruction after the IT block. */
> + do
> + {
> + inst1 = read_mem_uint (pc, 2, byte_order_for_code);
> + pc += thumb_insn_size (inst1);
> + itstate = thumb_advance_itstate (itstate);
> + }
> + while (itstate != 0 && ((itstate >> 4) & 1) == cond_negated);
> +
> + if (itstate != 0 && ((itstate >> 4) & 1) == cond_negated)
> + {
> + next_pcs.push_back (std::pair<CORE_ADDR, arm_breakpoint_kinds>
> + (MAKE_THUMB_ADDR (pc), ARM_BP_KIND_THUMB2));
> + }
> + else
> + {
> + next_pcs.push_back (std::pair<CORE_ADDR, arm_breakpoint_kinds>
> + (MAKE_THUMB_ADDR (pc), ARM_BP_KIND_THUMB));
> + }
Why do you choose breakpoint in this way?
> diff --git a/gdb/gdbserver/mem-break.c b/gdb/gdbserver/mem-break.c
> index 6e6926a..f3845cf 100644
> --- a/gdb/gdbserver/mem-break.c
> +++ b/gdb/gdbserver/mem-break.c
> @@ -855,7 +855,21 @@ set_breakpoint_type_at (enum bkpt_type type, CORE_ADDR where,
> {
> int err_ignored;
> CORE_ADDR placed_address = where;
> - int breakpoint_kind = target_breakpoint_kind_from_pc (&placed_address);
> + int breakpoint_kind;
> +
> + /* Get the kind of breakpoint to PLACED_ADDRESS except single-step
> + breakpoint. Get the kind of single-step breakpoint according to
> + the current register state. */
> + if (type == single_step_breakpoint)
> + {
> + breakpoint_kind
> + = target_breakpoint_kind_from_current_state (&placed_address);
I read my patch again, but it looks wrong. If we single-step an
instruction with a state change, like bx or blx, current get_next_pcs
correctly marked the address bit. However, with the change like this,
we'll get the wrong breakpoint kind.
--
Yao (齐尧)
next prev parent reply other threads:[~2017-04-03 15:18 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-11-29 12:07 Antoine Tremblay
2016-11-29 12:07 ` [PATCH 2/2] Avoid step-over infinite loop in GDBServer Antoine Tremblay
2017-01-16 17:27 ` Antoine Tremblay
2017-01-18 16:31 ` Antoine Tremblay
2017-02-03 16:21 ` Pedro Alves
2017-02-17 3:39 ` Antoine Tremblay
2017-02-22 10:15 ` Yao Qi
2017-03-27 13:28 ` Antoine Tremblay
2016-11-29 12:12 ` [PATCH 1/2] This patch fixes GDBServer's run control for single stepping Antoine Tremblay
2017-01-16 17:28 ` Antoine Tremblay
2017-01-27 15:01 ` Yao Qi
2017-01-27 16:07 ` Antoine Tremblay
2017-01-27 17:01 ` Yao Qi
2017-01-27 18:24 ` Antoine Tremblay
2017-01-29 21:41 ` Yao Qi
2017-01-30 13:29 ` Antoine Tremblay
2017-02-03 16:13 ` Pedro Alves
2017-02-17 1:42 ` Antoine Tremblay
2017-02-17 2:05 ` Pedro Alves
2017-02-17 3:06 ` Antoine Tremblay
2017-02-17 22:19 ` Yao Qi
2017-02-18 0:19 ` Antoine Tremblay
2017-02-18 22:49 ` Yao Qi
2017-02-19 19:40 ` Antoine Tremblay
2017-02-19 20:31 ` Antoine Tremblay
2017-03-29 12:41 ` Antoine Tremblay
2017-03-29 14:11 ` Antoine Tremblay
2017-03-29 17:54 ` Antoine Tremblay
2017-03-30 16:06 ` Yao Qi
2017-03-30 18:31 ` Antoine Tremblay
2017-03-31 16:31 ` Yao Qi
2017-03-31 18:22 ` Antoine Tremblay
2017-04-03 12:41 ` Yao Qi
2017-04-03 13:18 ` Antoine Tremblay
2017-04-03 15:18 ` Yao Qi [this message]
2017-04-03 16:57 ` Antoine Tremblay
2017-02-16 22:32 ` Yao Qi
2017-02-17 2:17 ` Antoine Tremblay
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=8637dpldta.fsf@gmail.com \
--to=qiyaoltc@gmail.com \
--cc=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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).