From: Yao Qi <yao@codesourcery.com>
To: Jan Kratochvil <jan.kratochvil@redhat.com>
Cc: Tom Tromey <tromey@redhat.com>, Doug Evans <dje@google.com>,
<gdb-patches@sourceware.org>
Subject: Re: [RFC] Don't allow setting register in non-innermost frame
Date: Mon, 10 Sep 2012 02:02:00 -0000 [thread overview]
Message-ID: <504D49DA.6070006@codesourcery.com> (raw)
In-Reply-To: <20120907164544.GA18234@host2.jankratochvil.net>
On 09/08/2012 12:45 AM, Jan Kratochvil wrote:
> (gdb) set variable i=20
> warning: Assigning to register in non-innermost frame.
>
> Why?
>
> I would find correct to print such warning for:
> (gdb) set $rbx=20
> There it should happen only if user has explicitly specified the register
> itself.
I checked the difference of value of register and value of variable
(saved in register), TYPE_OBJFILE_OWNED (value_type (toval)) is the
only difference between them. IIUC, type of variable is owned by an
objfile, while type of register is not (alloc_type vs. alloc_type_arch),
so looks we can use it to differentiate values in register and values
in variables.
This new patch includes TYPE_OBJFILE_OWNED into the condition, and
adjust the NEWS entry discussed with Eli before.
--
Yao
gdb/testsuite:
2012-09-10 Yao Qi <yao@codesourcery.com>
* gdb.base/set-reg.exp: New.
* gdb.base/set-reg.c: New.
gdb:
2012-09-10 Yao Qi <yao@codesourcery.com>
Doug Evans <dje@google.com>
* valops.c (value_assign): Emit warning when assigning to
registers in non-innermost frames.
* NEWS: Mention it.
---
gdb/NEWS | 3 +
gdb/testsuite/gdb.base/set-reg.c | 40 ++++++++++
gdb/testsuite/gdb.base/set-reg.exp | 139 ++++++++++++++++++++++++++++++++++++
gdb/valops.c | 16 ++++
4 files changed, 198 insertions(+), 0 deletions(-)
create mode 100644 gdb/testsuite/gdb.base/set-reg.c
create mode 100644 gdb/testsuite/gdb.base/set-reg.exp
diff --git a/gdb/NEWS b/gdb/NEWS
index dba6937..f7476e4 100644
--- a/gdb/NEWS
+++ b/gdb/NEWS
@@ -17,6 +17,9 @@
* The 'cd' command now defaults to using '~' (the home directory) if not
given an argument.
+* GDB will display a warning when assigning to registers by CLI command
+ "set" or MI command "-var-assign" in non-innermost frames.
+
* New configure options
--enable-libmcheck/--disable-libmcheck
diff --git a/gdb/testsuite/gdb.base/set-reg.c b/gdb/testsuite/gdb.base/set-reg.c
new file mode 100644
index 0000000..03938eb
--- /dev/null
+++ b/gdb/testsuite/gdb.base/set-reg.c
@@ -0,0 +1,40 @@
+/* This test is part of GDB, the GNU debugger.
+
+ Copyright 2012 Free Software Foundation, Inc.
+
+ 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/>. */
+
+static int
+inner (int x)
+{
+ return x + 2;
+}
+
+static int
+middle (int x)
+{
+ return 2 * inner (x);
+}
+
+static int
+top (int x)
+{
+ return 2 * middle (x);
+}
+
+int
+main (int argc, char **argv)
+{
+ return top (argc);
+}
diff --git a/gdb/testsuite/gdb.base/set-reg.exp b/gdb/testsuite/gdb.base/set-reg.exp
new file mode 100644
index 0000000..5ea65da
--- /dev/null
+++ b/gdb/testsuite/gdb.base/set-reg.exp
@@ -0,0 +1,139 @@
+# Copyright 2012 Free Software Foundation, Inc.
+#
+# 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/>.
+
+standard_testfile
+set executable $testfile
+
+# A value in register in previous frame. E.g,. call-clobbered register.
+set reg_prev_frame_in_reg "unknown"
+# A (register) value in memory in previous frame. E.g., call-saved register.
+set reg_prev_frame_in_memory "unknown"
+
+if [is_amd64_regs_target] {
+ set reg_prev_frame_in_reg "rax"
+ set reg_prev_frame_in_memory "rbp"
+} elseif [is_x86_like_target] {
+ set reg_prev_frame_in_reg "eax"
+ set reg_prev_frame_in_memory "ebp"
+} else {
+ # set reg_prev_frame_in_reg for other arch.
+}
+
+# Name of register is not set, skip the rest of test.
+if { [string equal $reg_prev_frame_in_reg "unknown"] } {
+ return
+}
+
+if { [gdb_compile $srcdir/$subdir/$srcfile $binfile executable {debug}] != "" } {
+ untested "Failed to compile"
+ return -1
+}
+
+proc set_register { register check_value } {with_test_prefix "${register} ${check_value}" {
+ global decimal
+ global gdb_prompt
+
+ set reg_val ""
+ gdb_test_multiple "p/d \$${register}" "print register ${register}" {
+ -re ".*${decimal} = ($decimal).*$gdb_prompt $" {
+ set reg_val $expect_out(1,string)
+ }
+ }
+ if { $reg_val == "" } {
+ fail "get the value of register ${register}"
+ return
+ }
+
+ # Assign the same value to register.
+ gdb_test_no_output "set \$${register}=${reg_val}" \
+ "set register ${register} on inner frame (unchanged)"
+
+ # Assign the different value to register.
+ set reg_val [expr $reg_val + 4]
+ gdb_test_no_output "set \$${register}=${reg_val}" \
+ "set register ${register} on inner frame (changed)"
+ gdb_test "p/d \$${register}" ".*${decimal} = $reg_val.*" \
+ "register ${register} is changed"
+
+ # Restore register.
+ set reg_val [expr $reg_val - 4]
+ gdb_test_no_output "set \$${register}=${reg_val}" \
+ "set register ${register} on inner frame (restored)"
+ gdb_test "p/d \$${register}" ".*${decimal} = $reg_val.*" \
+ "register ${register} is restored"
+
+ gdb_test "up"
+
+ gdb_test_multiple "p/d \$${register}" "print register ${register}" {
+ -re ".*${decimal} = ($decimal).*$gdb_prompt $" {
+ set reg_val $expect_out(1,string)
+ }
+ }
+ if { $reg_val == "" } {
+ fail "get the value of register ${register}"
+ return
+ }
+ # Assign the same value to register.
+ gdb_test "set \$${register}=${reg_val}" \
+ "Assigning to register in non-innermost frame.*" \
+ "set register ${register} on non-inner frame"
+
+ if { $check_value } {
+ # Assign the different value to register.
+ set reg_val [expr $reg_val + 4]
+ gdb_test "set \$${register}=${reg_val}" \
+ "Assigning to register in non-innermost frame.*" \
+ "set register ${register} on non-inner frame (changed)"
+ gdb_test "p/d \$${register}" ".*${decimal} = $reg_val.*" \
+ "register ${register} is changed"
+ # Restore register.
+ set reg_val [expr $reg_val - 4]
+ gdb_test "set \$${register}=${reg_val}" \
+ "Assigning to register in non-innermost frame.*" \
+ "set register ${register} on non-inner frame (restored)"
+ gdb_test "p/d \$${register}" ".*${decimal} = $reg_val.*" \
+ "register ${register} is restored"
+ }
+ gdb_test "down"
+}}
+
+# Test setting register on innermost frame and non-innermost frame.
+
+proc test_set_reg_on_frame {} {with_test_prefix "on frame" {
+ global executable
+ global reg_prev_frame_in_reg
+ global reg_prev_frame_in_memory
+ global decimal
+ global hex
+ global expect_out
+ global gdb_prompt
+
+ clean_restart $executable
+
+ if ![runto_main] {
+ return
+ }
+
+ gdb_test "break inner" "Breakpoint.*at.* file .*, line.*"
+ gdb_test "continue" "Continuing.\r\n\r\nBreakpoint.*inner.*at.*" \
+ "continue to inner"
+
+ set_register ${reg_prev_frame_in_reg} 1
+ # Change the register saved on frame may clobber the frame, and get
+ # the contents of registers in mess. Don't check register value.
+ set_register ${reg_prev_frame_in_memory} 0
+}}
+
+test_set_reg_on_frame
diff --git a/gdb/valops.c b/gdb/valops.c
index 17696ee..7ac23e3 100644
--- a/gdb/valops.c
+++ b/gdb/valops.c
@@ -1241,6 +1241,22 @@ value_assign (struct value *toval, struct value *fromval)
and then restore the new frame afterwards. */
old_frame = get_frame_id (deprecated_safe_get_selected_frame ());
+ /* TOVAL is a register, although VALUE_LVAL(TOVAL) may not be
+ lval_register. A call-saved register saved in memory will have
+ 'VALUE_REGNUM >= 0' but 'VALUE_LVAL == lval_memory'. We also have to
+ avoid emitting warning when assign value to some local variables which
+ are stored in registers, TYPE_OBJFILE_OWNED helps to differentiate
+ we are assigning to a register explicitly or to a variable saved in
+ register. */
+ if (VALUE_REGNUM (toval) >= 0 && !TYPE_OBJFILE_OWNED (type))
+ {
+ /* Figure out which frame this is in currently. */
+ struct frame_info *frame = frame_find_by_id (VALUE_FRAME_ID (toval));
+
+ if (get_next_frame (frame) != NULL)
+ warning (_("Assigning to register in non-innermost frame."));
+ }
+
switch (VALUE_LVAL (toval))
{
case lval_internalvar:
--
1.7.7.6
next prev parent reply other threads:[~2012-09-10 2:02 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-08-17 2:21 Yao Qi
2012-08-20 20:19 ` Doug Evans
2012-08-21 3:27 ` Yao Qi
2012-08-23 16:25 ` Tom Tromey
2012-08-29 9:51 ` Yao Qi
2012-09-04 22:37 ` dje
2012-09-07 10:01 ` Yao Qi
2012-09-07 10:11 ` Eli Zaretskii
2012-09-07 10:21 ` Yao Qi
2012-09-07 11:27 ` Eli Zaretskii
2012-09-07 13:14 ` Yao Qi
2012-09-07 14:32 ` Eli Zaretskii
2012-09-07 16:46 ` Jan Kratochvil
2012-09-09 2:31 ` Yao Qi
2012-09-10 2:02 ` Yao Qi [this message]
2012-09-10 7:47 ` Jan Kratochvil
2012-09-10 19:43 ` Jan Kratochvil
2012-09-11 17:12 ` Tom Tromey
2012-09-11 17:19 ` Jan Kratochvil
2012-09-11 17:23 ` Tom Tromey
2012-09-12 0:51 ` Yao Qi
2012-09-12 13:19 ` Jan Kratochvil
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=504D49DA.6070006@codesourcery.com \
--to=yao@codesourcery.com \
--cc=dje@google.com \
--cc=gdb-patches@sourceware.org \
--cc=jan.kratochvil@redhat.com \
--cc=tromey@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