From: Yao Qi <qiyaoltc@gmail.com>
To: Alan Hayward <Alan.Hayward@arm.com>
Cc: Pedro Alves <palves@redhat.com>,
Joel Brobecker <brobecker@adacore.com>,
"gdb-patches\@sourceware.org" <gdb-patches@sourceware.org>,
Subject: Re: [PATCH] Removal of uses of MAX_REGISTER_SIZE
Date: Wed, 08 Feb 2017 12:24:00 -0000 [thread overview]
Message-ID: <86inokq39w.fsf@gmail.com> (raw)
In-Reply-To: <5F3D30AE-9A53-493A-B6DC-DF594C2FAB18@arm.com> (Alan Hayward's message of "Tue, 7 Feb 2017 16:33:19 +0000")
Alan Hayward <Alan.Hayward@arm.com> writes:
> diff --git a/gdb/stack.c b/gdb/stack.c
> index e00e2972cf20bc63917af19f86bf57f1c6b0b5b0..7ba7d68bde8d83ea1e700faa466c6951979e0f76 100644
> --- a/gdb/stack.c
> +++ b/gdb/stack.c
> @@ -1650,33 +1650,35 @@ frame_info (char *addr_exp, int from_tty)
> int count;
> int i;
> int need_nl = 1;
> + int sp_regnum = gdbarch_sp_regnum (gdbarch);
>
> /* The sp is special; what's displayed isn't the save address, but
> the value of the previous frame's sp. This is a legacy thing,
> at one stage the frame cached the previous frame's SP instead
> of its address, hence it was easiest to just display the cached
> value. */
> - if (gdbarch_sp_regnum (gdbarch) >= 0)
> + if (sp_regnum >= 0)
> {
> /* Find out the location of the saved stack pointer with out
> actually evaluating it. */
> - frame_register_unwind (fi, gdbarch_sp_regnum (gdbarch),
> - &optimized, &unavailable, &lval, &addr,
> - &realnum, NULL);
> + frame_register_unwind (fi, sp_regnum, &optimized, &unavailable, &lval,
> + &addr, &realnum, NULL);
> if (!optimized && !unavailable && lval == not_lval)
> {
> enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
> - int sp_size = register_size (gdbarch, gdbarch_sp_regnum (gdbarch));
> - gdb_byte value[MAX_REGISTER_SIZE];
> + int sp_size = register_size (gdbarch, sp_regnum);
> CORE_ADDR sp;
> + struct value *value = frame_unwind_register_value (fi, sp_regnum);
>
> - frame_register_unwind (fi, gdbarch_sp_regnum (gdbarch),
> - &optimized, &unavailable, &lval, &addr,
> - &realnum, value);
> + gdb_assert (value != NULL);
Why don't you hoist frame_unwind_register_value above?, so the
frame_register_unwind call is no longer needed,
struct value *value = frame_unwind_register_value (fi, sp_regnum);
gdb_assert (value != NULL);
if (!value_optimized_out (value) && value_entirely_available (value))
{
if (VALUE_LVAL (value) == not_lval)
{
sp = extract_unsigned_integer (value_contents_all (value),
sp_size, byte_order);
}
else if (VALUE_LVAL (value) == lval_memory)
{
// use value_address (value);
}
else if (VALUE_LVAL (value) == lval_register)
{
// use VALUE_REGNUM (value);
}
}
/* else keep quiet. */
release_value (value);
value_free (value);
> /* NOTE: cagney/2003-05-22: This is assuming that the
> stack pointer was packed as an unsigned integer. That
> may or may not be valid. */
> - sp = extract_unsigned_integer (value, sp_size, byte_order);
> + sp = extract_unsigned_integer (value_contents_all (value), sp_size,
> + byte_order);
> + release_value (value);
> + value_free (value);
> +
> printf_filtered (" Previous frame's sp is ");
> fputs_filtered (paddress (gdbarch, sp), gdb_stdout);
> printf_filtered ("\n");
> @@ -1702,7 +1704,7 @@ frame_info (char *addr_exp, int from_tty)
> numregs = gdbarch_num_regs (gdbarch)
> + gdbarch_num_pseudo_regs (gdbarch);
> for (i = 0; i < numregs; i++)
> - if (i != gdbarch_sp_regnum (gdbarch)
> + if (i != sp_regnum
> && gdbarch_register_reggroup_p (gdbarch, i, all_reggroup))
> {
> /* Find out the location of the saved register without
--
Yao (齐尧)
next prev parent reply other threads:[~2017-02-08 12:24 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-01-24 10:31 Alan Hayward
2017-01-27 11:49 ` Pedro Alves
2017-01-27 12:11 ` Pedro Alves
2017-01-27 16:46 ` Alan Hayward
2017-02-01 15:49 ` Pedro Alves
2017-02-01 12:45 ` Yao Qi
2017-02-01 15:48 ` Pedro Alves
2017-02-02 9:40 ` Joel Brobecker
2017-02-03 9:59 ` Alan Hayward
2017-02-03 10:28 ` Yao Qi
2017-02-03 11:00 ` Pedro Alves
2017-02-03 11:25 ` Alan Hayward
2017-02-03 16:50 ` Yao Qi
2017-02-06 9:33 ` Alan Hayward
[not found] ` <20170206152635.GE11916@E107787-LIN>
2017-02-07 16:33 ` Alan Hayward
2017-02-08 10:47 ` Yao Qi
2017-02-08 14:17 ` Alan Hayward
2017-02-08 12:06 ` Yao Qi
2017-02-08 12:24 ` Yao Qi [this message]
2017-02-08 14:44 ` Alan Hayward
2017-02-18 23:19 ` Yao Qi
2017-02-20 11:19 ` Alan Hayward
2017-02-08 15:32 ` Yao Qi
2017-02-08 17:10 ` Yao Qi
2017-02-09 13:26 ` Alan Hayward
2017-02-14 11:24 ` Alan Hayward
2017-02-08 17:36 ` Yao Qi
2017-02-13 11:59 ` Alan Hayward
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=86inokq39w.fsf@gmail.com \
--to=qiyaoltc@gmail.com \
--cc=Alan.Hayward@arm.com \
--cc=brobecker@adacore.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).