From: Yao Qi <qiyaoltc@gmail.com>
To: Pedro Alves <palves@redhat.com>
Cc: gdb-patches@sourceware.org
Subject: Re: [PATCH v2 28/31] Use ui_file_as_string throughout more
Date: Thu, 23 Feb 2017 10:23:00 -0000 [thread overview]
Message-ID: <86r32pb40h.fsf@gmail.com> (raw)
In-Reply-To: <1476968896-13600-1-git-send-email-palves@redhat.com> (Pedro Alves's message of "Thu, 20 Oct 2016 14:08:16 +0100")
Pedro Alves <palves@redhat.com> writes:
Hi Pedro,
> std::string as member of a struct, I needed to adjust
> allocation/destruction of said struct to use new/delete instead of
> xmalloc/xfree.
>
> diff --git a/gdb/varobj-iter.h b/gdb/varobj-iter.h
> index bdbf661..1e7889b 100644
> --- a/gdb/varobj-iter.h
> +++ b/gdb/varobj-iter.h
> @@ -19,7 +19,7 @@
> typedef struct varobj_item
> {
> /* Name of this item. */
> - char *name;
> + std::string name;
>
> /* Value of this item. */
> struct value *value;
> @@ -67,6 +67,6 @@ struct varobj_iter_ops
> if ((ITER) != NULL) \
> { \
> (ITER)->ops->dtor (ITER); \
> - xfree (ITER); \
> + delete (ITER); \
We don't need to use "delete" for varobj_iter, because you didn't add
std::string member in varobj_iter. This change causes an ASAN error,
-var-create container @ c^M
=================================================================^M
^[[1m^[[31m==12975==ERROR: AddressSanitizer: alloc-dealloc-mismatch (malloc vs operator delete) on 0x603000085e40^M
^[[1m^[[0m #0 0x2b8ce74d15d7 in operator delete(void*) (/usr/lib/x86_64-linux-gnu/libasan.so.1+0x555d7)^M
#1 0xc13d3c in update_dynamic_varobj_children /home/yao/SourceCode/gnu/gdb/git/gdb/varobj.c:786^M
#2 0xc1425f in varobj_get_num_children(varobj*) /home/yao/SourceCode/gnu/gdb/git/gdb/varobj.c:846^M
#3 0x59ffeb in print_varobj /home/yao/SourceCode/gnu/gdb/git/gdb/mi/mi-cmd-var.c:61^M
#4 0x5a0689 in mi_cmd_var_create(char*, char**, int) /home/yao/SourceCode/gnu/gdb/git/gdb/mi/mi-cmd-var.c:145^M
#5 0x5b5ac9 in mi_cmd_execute /home/yao/SourceCode/gnu/gdb/git/gdb/mi/mi-main.c:2295^M
#6 0x5b3de0 in captured_mi_execute_command /home/yao/SourceCode/gnu/gdb/git/gdb/mi/mi-main.c:1997^M
.....
^M
^[[1m^[[32m0x603000085e40 is located 0 bytes inside of 32-byte region [0x603000085e40,0x603000085e60)^M
^[[1m^[[0m^[[1m^[[35mallocated by thread T0 here:^[[1m^[[0m^M
#0 0x2b8ce74d0862 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.1+0x54862)^M
#1 0x80df7f in xmalloc /home/yao/SourceCode/gnu/gdb/git/gdb/common/common-utils.c:43^M
#2 0x6364ec in py_varobj_iter_new /home/yao/SourceCode/gnu/gdb/git/gdb/python/py-varobj.c:154^M
#3 0x636710 in py_varobj_get_iterator(varobj*, _object*) /home/yao/SourceCode/gnu/gdb/git/gdb/python/py-varobj.c:188^M
#4 0xc13580 in varobj_get_iterator /home/yao/SourceCode/gnu/gdb/git/gdb/varobj.c:712^M
#5 0xc138fa in update_dynamic_varobj_children /home/yao/SourceCode/gnu/gdb/git/gdb/varobj.c:750^M
#6 0xc1425f in varobj_get_num_children(varobj*) /home/yao/SourceCode/gnu/gdb/git/gdb/varobj.c:846^M
reverting the change above fixes the error. We can also C++-fy
varobj_iter, but I'd like to do in another patch. How is the patch below?
> } \
> } while (0)
> diff --git a/gdb/varobj.h b/gdb/varobj.h
> index 6b9a71f..7f4aad2 100644
> --- a/gdb/varobj.h
> +++ b/gdb/varobj.h
> @@ -86,22 +86,22 @@ struct varobj_dynamic;
>
> /* Every variable in the system has a structure of this type defined
> for it. This structure holds all information necessary to manipulate
> - a particular object variable. Members which must be freed are noted. */
> + a particular object variable. */
> struct varobj
> {
> - /* Alloc'd name of the variable for this object. If this variable is a
> + /* Name of the variable for this object. If this variable is a
> child, then this name will be the child's source name.
> (bar, not foo.bar). */
> /* NOTE: This is the "expression". */
> - char *name;
> + std::string name;
>
> - /* Alloc'd expression for this child. Can be used to create a
> - root variable corresponding to this child. */
> - char *path_expr;
> + /* Expression for this child. Can be used to create a root variable
> + corresponding to this child. */
> + std::string path_expr;
>
> - /* The alloc'd name for this variable's object. This is here for
> + /* The name for this variable's object. This is here for
> convenience when constructing this object's children. */
> - char *obj_name;
> + std::string obj_name;
>
> /* Index of this variable in its parent or -1. */
> int index;
> @@ -137,7 +137,7 @@ struct varobj
> int updated;
>
> /* Last print value. */
> - char *print_value;
> + std::string print_value;
>
> /* Is this variable frozen. Frozen variables are never implicitly
> updated by -var-update *
> @@ -170,18 +170,15 @@ struct lang_varobj_ops
> /* The number of children of PARENT. */
> int (*number_of_children) (const struct varobj *parent);
>
> - /* The name (expression) of a root varobj. The returned value must be freed
> - by the caller. */
> - char *(*name_of_variable) (const struct varobj *parent);
> + /* The name (expression) of a root varobj. */
> + std::string (*name_of_variable) (const struct varobj *parent);
>
> - /* The name of the INDEX'th child of PARENT. The returned value must be
> - freed by the caller. */
> - char *(*name_of_child) (const struct varobj *parent, int index);
> + /* The name of the INDEX'th child of PARENT. */
> + std::string (*name_of_child) (const struct varobj *parent, int index);
>
> /* Returns the rooted expression of CHILD, which is a variable
> - obtain that has some parent. The returned value must be freed by the
> - caller. */
> - char *(*path_expr_of_child) (const struct varobj *child);
> + obtain that has some parent. */
> + std::string (*path_expr_of_child) (const struct varobj *child);
>
> /* The ``struct value *'' of the INDEX'th child of PARENT. */
> struct value *(*value_of_child) (const struct varobj *parent, int index);
> @@ -189,10 +186,9 @@ struct lang_varobj_ops
> /* The type of the INDEX'th child of PARENT. */
> struct type *(*type_of_child) (const struct varobj *parent, int index);
>
> - /* The current value of VAR. The returned value must be freed by the
> - caller. */
> - char *(*value_of_variable) (const struct varobj *var,
> - enum varobj_display_formats format);
> + /* The current value of VAR. */
> + std::string (*value_of_variable) (const struct varobj *var,
> + enum varobj_display_formats format);
>
> /* Return non-zero if changes in value of VAR must be detected and
> reported by -var-update. Return zero if -var-update should never
--
Yao (齐尧)
From 71dc64a061cfb1ec57b753e495c02665827c7a01 Mon Sep 17 00:00:00 2001
From: Yao Qi <yao.qi@linaro.org>
Date: Thu, 23 Feb 2017 09:29:48 +0000
Subject: [PATCH] Use xfree rather than delete for varobj_iter
ASAN reports an error,
-var-create container @ c^M
=================================================================^M
^[[1m^[[31m==21639==ERROR: AddressSanitizer: alloc-dealloc-mismatch (malloc vs operator delete) on 0x6030000805c0^M
^[[1m^[[0m #0 0x7f2449b01b2a in operator delete(void*) (/usr/lib/x86_64-linux-gnu/libasan.so.2+0x99b2a)^M
#1 0xbb601d in update_dynamic_varobj_children ../../binutils-gdb/gdb/varobj.c:794^M
#2 0xbb6556 in varobj_get_num_children(varobj*) ../../binutils-gdb/gdb/varobj.c:854^M
#3 0x580cb4 in print_varobj ../../binutils-gdb/gdb/mi/mi-cmd-var.c:61^M
#4 0x58138b in mi_cmd_var_create(char*, char**, int) ../../binutils-gdb/gdb/mi/mi-cmd-var.c:145^M
#5 0x5967ce in mi_cmd_execute ../../binutils-gdb/gdb/mi/mi-main.c:2301^M
#6 0x594b05 in captured_mi_execute_command ../../binutils-gdb/gdb/mi/mi-main.c:2001
....
^M
^[[1m^[[32m0x6030000805c0 is located 0 bytes inside of 32-byte region [0x6030000805c0,0x6030000805e0)^M
^[[1m^[[0m^[[1m^[[35mallocated by thread T0 here:^[[1m^[[0m^M
#0 0x7f2449b00602 in malloc (/usr/lib/x86_64-linux-gnu/libasan.so.2+0x98602)^M
#1 0x7d1596 in xmalloc ../../binutils-gdb/gdb/common/common-utils.c:43^M
#2 0x604176 in py_varobj_iter_new ../../binutils-gdb/gdb/python/py-varobj.c:159^M
#3 0x6042da in py_varobj_get_iterator(varobj*, _object*) ../../binutils-gdb/gdb/python/py-varobj.c:198^M
#4 0xbb5806 in varobj_get_iterator ../../binutils-gdb/gdb/varobj.c:720^M
#5 0xbb5b9b in update_dynamic_varobj_children ../../binutils-gdb/gdb/varobj.c:758^M
gdb:
2017-02-23 Yao Qi <yao.qi@linaro.org>
* varobj-iter.h (varobj_iter_delete): Call xfree instead of
delete.
diff --git a/gdb/varobj-iter.h b/gdb/varobj-iter.h
index 34182e0..7eed4f1 100644
--- a/gdb/varobj-iter.h
+++ b/gdb/varobj-iter.h
@@ -67,6 +67,6 @@ struct varobj_iter_ops
if ((ITER) != NULL) \
{ \
(ITER)->ops->dtor (ITER); \
- delete (ITER); \
+ xfree (ITER); \
} \
} while (0)
next prev parent reply other threads:[~2017-02-23 10:23 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-19 1:13 [PATCH v2 00/31] More cleanup elimination & unlimited args to user-defined funcs Pedro Alves
2016-10-19 1:12 ` [PATCH v2 22/31] Use ui_file_as_string in gdb/c-exp.y Pedro Alves
2016-10-19 1:12 ` [PATCH v2 21/31] Use ui_file_as_string in gdb/compile/ Pedro Alves
2016-10-19 23:08 ` Simon Marchi
2016-10-19 23:48 ` Pedro Alves
2016-10-20 3:17 ` Simon Marchi
2016-10-20 12:13 ` Pedro Alves
2016-10-19 1:12 ` [PATCH v2 19/31] Use ui_file_as_string in gdb/remote.c Pedro Alves
2016-10-19 1:12 ` [PATCH v2 07/31] Clean up tracepoint.h/c:collection_list Pedro Alves
2016-10-19 1:12 ` [PATCH v2 24/31] Use ui_file_as_string in gdb/ada-lang.c Pedro Alves
2016-10-19 1:12 ` [PATCH v2 06/31] Introduce ui_file_as_string Pedro Alves
2016-10-19 1:12 ` [PATCH v2 01/31] Introduce string_printf Pedro Alves
2016-10-19 13:43 ` Trevor Saunders
2016-10-19 14:41 ` Pedro Alves
2016-10-19 17:18 ` Simon Marchi
2016-10-19 21:02 ` Pedro Alves
2016-11-08 15:35 ` Pedro Alves
2016-10-19 1:12 ` [PATCH v2 20/31] Use ui_file_as_string in gdb/cli/cli-setshow.c Pedro Alves
2016-10-19 1:12 ` [PATCH v2 12/31] Use ui_file_as_string in gdb/utils.c Pedro Alves
2016-10-19 1:13 ` [PATCH v2 23/31] Use ui_file_as_string in gdbarch.sh/gdbarch.c Pedro Alves
2016-10-19 1:13 ` [PATCH v2 26/31] Use ui_file_as_string in gdb/rust-lang.c Pedro Alves
2016-10-19 1:13 ` [PATCH v2 03/31] breakpoint.c:commands_command_1 constification and cleanup Pedro Alves
2016-10-19 1:13 ` [PATCH v2 14/31] Use ui_file_as_string in gdb/guile/ Pedro Alves
2016-10-19 1:13 ` [PATCH v2 29/31] 'struct agent_expr *' -> unique_ptr<agent_expr> Pedro Alves
2016-10-19 23:19 ` Simon Marchi
2016-10-19 23:58 ` Pedro Alves
2016-10-19 1:13 ` [PATCH v2 27/31] Use ui_file_as_string in gdb/language.c Pedro Alves
2016-10-19 1:13 ` [PATCH v2 15/31] Use ui_file_as_string in execute_command_to_string Pedro Alves
2016-10-19 1:13 ` [PATCH v2 08/31] Use ui_file_as_string in dwarf2_compute_name Pedro Alves
2016-10-19 1:13 ` [PATCH v2 02/31] cli/cli-script.c: Remove some dead NULL checks Pedro Alves
2016-10-19 17:24 ` Simon Marchi
2016-10-19 21:18 ` Pedro Alves
2016-10-19 1:13 ` [PATCH v2 31/31] Support an "unlimited" number of user-defined arguments Pedro Alves
2016-10-19 6:29 ` Eli Zaretskii
2016-10-19 11:33 ` Philipp Rudo
2016-10-19 12:47 ` Pedro Alves
2016-10-19 17:40 ` Philipp Rudo
2016-10-19 17:45 ` Pedro Alves
2016-11-08 15:41 ` Pedro Alves
2016-10-19 1:17 ` [PATCH v2 10/31] Use ui_file_as_string in gdb/ada-valprint.c Pedro Alves
2016-10-19 1:17 ` [PATCH v2 11/31] Use ui_file_as_string in gdb/ui-out.c Pedro Alves
2016-10-19 1:17 ` [PATCH v2 25/31] Use ui_file_as_string in gdb/infrun.c Pedro Alves
2016-10-19 1:17 ` [PATCH v2 09/31] Use ui_file_as_string in gdb/xtensa-tdep.c Pedro Alves
2016-10-19 1:18 ` [PATCH v2 05/31] 'struct expression *' -> gdb::unique_xmalloc_ptr<expression> Pedro Alves
2016-10-19 18:45 ` Simon Marchi
2016-10-19 21:50 ` Pedro Alves
2016-10-19 22:25 ` Simon Marchi
2016-10-19 22:36 ` Pedro Alves
2016-10-19 1:18 ` [PATCH v2 30/31] Eliminate agent_expr_p; VEC -> std::vector in struct bp_target_info Pedro Alves
2016-10-19 1:19 ` [PATCH v2 04/31] cli-script.c: Simplify using std::string, eliminate cleanups Pedro Alves
2016-10-19 18:25 ` Simon Marchi
2016-10-19 21:45 ` Pedro Alves
2016-10-19 1:21 ` [PATCH v2 13/31] Use ui_file_as_string in gdb/arm-tdep.c Pedro Alves
2016-10-19 22:54 ` Simon Marchi
2016-10-19 1:21 ` [PATCH v2 18/31] Use ui_file_as_string in gdb/python/ Pedro Alves
2016-10-19 1:21 ` [PATCH v2 16/31] Use ui_file_as_string in gdb/top.c Pedro Alves
2016-10-19 1:21 ` [PATCH v2 17/31] Use ui_file_as_string in gdb/printcmd.c Pedro Alves
2016-10-20 13:08 ` [PATCH v2 28/31] Use ui_file_as_string throughout more Pedro Alves
2017-02-23 10:23 ` Yao Qi [this message]
2017-02-23 10:53 ` Pedro Alves
2017-02-23 10:35 ` Yao Qi
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=86r32pb40h.fsf@gmail.com \
--to=qiyaoltc@gmail.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).