From: Andrew Burgess <andrew.burgess@embecosm.com>
To: gdb-patches@sourceware.org
Cc: Andrew Burgess <andrew.burgess@embecosm.com>
Subject: [PATCH] gdb: Don't leak memory with TYPE_ALLOC / TYPE_ZALLOC
Date: Tue, 11 Sep 2018 14:12:00 -0000 [thread overview]
Message-ID: <20180911141147.14107-1-andrew.burgess@embecosm.com> (raw)
I ran into the following issue after runnig valgrind on GDB.
Calling TYPE_ALLOC or TYPE_ZALLOC currently allocates memory from
either the objfile obstack or by using malloc. The problem with this
is that types are allocated either on the objfile obstack, or on the
gdbarch obstack.
As a result, if we discard a type associated with an objfile then
auxiliary data allocated with TYPE_(Z)ALLOC will be correctly
discarded. But, if we were ever to discard a gdbarch then any
auxiliary type data would be leaked.
This commit ensures that auxiliary type data is allocated from the
same obstack as the type itself, which should reduce leaked memory.
The one problem case that I found with this change was in eval.c,
where in one place we allocate a local type structure, and then used
TYPE_ZALLOC to allocate some space for the type. This local type is
neither object file owned, nor gdbarch owned, and so the updated
TYPE_ALLOC code is unable to find an objstack to allocate space on.
My proposed solution for this issue is that the space should be
allocated with a direct call to xzalloc. We could extend TYPE_ALLOC
to check for type->gdbarch being null, and then fall back to a direct
call to xzalloc, however, I think making the rare case of a local type
requiring special handling is not a bad thing, this serves to
highlight that clearing up the memory will require special handling
too.
Tested on X86-64 GNU/Linux with no regressions.
gdb/ChangeLog:
* gdbtypes.h (TYPE_ALLOC): Allocate space on either the objfile
obstack, or the gdbarch obstack.
(TYPE_ZALLOC): Rewrite using TYPE_ALLOC.
* eval.c (fake_method::fake_method): Call xzalloc directly for a
type that is neither object file owned, nor gdbarch owned.
---
gdb/ChangeLog | 8 ++++++++
gdb/eval.c | 2 +-
gdb/gdbtypes.h | 36 ++++++++++++++++++++----------------
3 files changed, 29 insertions(+), 17 deletions(-)
diff --git a/gdb/eval.c b/gdb/eval.c
index 2e08e9355f5..baa98e17545 100644
--- a/gdb/eval.c
+++ b/gdb/eval.c
@@ -685,7 +685,7 @@ fake_method::fake_method (type_instance_flags flags,
TYPE_NFIELDS (type) = num_types;
TYPE_FIELDS (type) = (struct field *)
- TYPE_ZALLOC (type, sizeof (struct field) * num_types);
+ xzalloc (sizeof (struct field) * num_types);
while (num_types-- > 0)
TYPE_FIELD_TYPE (type, num_types) = param_types[num_types];
diff --git a/gdb/gdbtypes.h b/gdb/gdbtypes.h
index eb7c365b71d..3778b46839c 100644
--- a/gdb/gdbtypes.h
+++ b/gdb/gdbtypes.h
@@ -49,6 +49,7 @@
#include "common/enum-flags.h"
#include "common/underlying.h"
#include "common/print-utils.h"
+#include "gdbarch.h"
/* Forward declarations for prototypes. */
struct field;
@@ -1717,26 +1718,29 @@ extern const struct floatformat *floatformats_vax_d[BFD_ENDIAN_UNKNOWN];
extern const struct floatformat *floatformats_ibm_long_double[BFD_ENDIAN_UNKNOWN];
-/* * Allocate space for storing data associated with a particular
+/* Allocate space for storing data associated with a particular
type. We ensure that the space is allocated using the same
mechanism that was used to allocate the space for the type
structure itself. I.e. if the type is on an objfile's
objfile_obstack, then the space for data associated with that type
- will also be allocated on the objfile_obstack. If the type is not
- associated with any particular objfile (such as builtin types),
- then the data space will be allocated with xmalloc, the same as for
- the type structure. */
-
-#define TYPE_ALLOC(t,size) \
- (TYPE_OBJFILE_OWNED (t) \
- ? obstack_alloc (&TYPE_OBJFILE (t) -> objfile_obstack, size) \
- : xmalloc (size))
-
-#define TYPE_ZALLOC(t,size) \
- (TYPE_OBJFILE_OWNED (t) \
- ? memset (obstack_alloc (&TYPE_OBJFILE (t)->objfile_obstack, size), \
- 0, size) \
- : xzalloc (size))
+ will also be allocated on the objfile_obstack. If the type is
+ associated with a gdbarch, then the space for data associated with that
+ type will also be allocated on the gdbarch_obstack.
+
+ If a type is not associated with neither an objfile or a gdbarch then
+ you should not use this macro to allocate space for data, instead you
+ should call xmalloc directly, and ensure the memory is correctly freed
+ when it is no longer needed. */
+
+#define TYPE_ALLOC(t,size) \
+ (obstack_alloc ((TYPE_OBJFILE_OWNED (t) \
+ ? &TYPE_OBJFILE (t)->objfile_obstack \
+ : gdbarch_obstack (TYPE_OWNER (t).gdbarch)), \
+ size))
+
+/* See comment on TYPE_ALLOC. */
+
+#define TYPE_ZALLOC(t,size) (memset (TYPE_ALLOC (t, size), 0, size))
/* Use alloc_type to allocate a type owned by an objfile. Use
alloc_type_arch to allocate a type owned by an architecture. Use
--
2.14.4
next reply other threads:[~2018-09-11 14:12 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-11 14:12 Andrew Burgess [this message]
2018-09-11 20:38 ` Tom Tromey
2018-09-13 0:31 ` Andrew Burgess
2018-09-13 3:31 ` Tom Tromey
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=20180911141147.14107-1-andrew.burgess@embecosm.com \
--to=andrew.burgess@embecosm.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