public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
* [PATCH] gdbsupport: fix memory leak in create_file_handler when re-using file handler
@ 2021-12-02 16:51 Simon Marchi
  2021-12-02 17:31 ` Tom Tromey
  0 siblings, 1 reply; 5+ messages in thread
From: Simon Marchi @ 2021-12-02 16:51 UTC (permalink / raw)
  To: gdb-patches; +Cc: Simon Marchi

From: Simon Marchi <simon.marchi@efficios.com>

ASan made me notice a memory leak, where the memory tied to the file
handle name string wasn't freed.  When register a file handler with an
fd that is already registered, we re-use the file_handler object, so we
ended up creating a new std::string object and overwriting the
file_handler::name pointer, without free-ing the old std::string.

Change it so that if re-using a file_handler object, we assign the
name to the existing std::string object.

Change-Id: Ie304cc78ab5ae5dfad9a1366e9890c09de651f43
---
 gdbsupport/event-loop.cc | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/gdbsupport/event-loop.cc b/gdbsupport/event-loop.cc
index 98d1ada52cd8..48e5e0031919 100644
--- a/gdbsupport/event-loop.cc
+++ b/gdbsupport/event-loop.cc
@@ -319,6 +319,7 @@ create_file_handler (int fd, int mask, handler_func * proc,
       file_ptr->fd = fd;
       file_ptr->ready_mask = 0;
       file_ptr->next_file = gdb_notifier.first_file_handler;
+      file_ptr->name = new std::string (std::move (name));
       gdb_notifier.first_file_handler = file_ptr;
 
       if (use_poll)
@@ -362,11 +363,16 @@ create_file_handler (int fd, int mask, handler_func * proc,
 	    gdb_notifier.num_fds = fd + 1;
 	}
     }
+  else
+    {
+      /* Re-using a file_handler object, re-use previous name object.  */
+      *file_ptr->name = std::move (name);
+    }
+
 
   file_ptr->proc = proc;
   file_ptr->client_data = client_data;
   file_ptr->mask = mask;
-  file_ptr->name = new std::string (std::move (name));
   file_ptr->is_ui = is_ui;
 }
 
-- 
2.33.1


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] gdbsupport: fix memory leak in create_file_handler when re-using file handler
  2021-12-02 16:51 [PATCH] gdbsupport: fix memory leak in create_file_handler when re-using file handler Simon Marchi
@ 2021-12-02 17:31 ` Tom Tromey
  2021-12-02 19:10   ` Simon Marchi
  0 siblings, 1 reply; 5+ messages in thread
From: Tom Tromey @ 2021-12-02 17:31 UTC (permalink / raw)
  To: Simon Marchi via Gdb-patches; +Cc: Simon Marchi, Simon Marchi

>>>>> "Simon" == Simon Marchi via Gdb-patches <gdb-patches@sourceware.org> writes:

Simon> Change it so that if re-using a file_handler object, we assign the
Simon> name to the existing std::string object.

This seems fine, but it seems to me that this would also be fixed if
file handlers were allocated with new and could hold smart pointers.

Tom

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] gdbsupport: fix memory leak in create_file_handler when re-using file handler
  2021-12-02 17:31 ` Tom Tromey
@ 2021-12-02 19:10   ` Simon Marchi
  2021-12-04 18:18     ` Tom Tromey
  0 siblings, 1 reply; 5+ messages in thread
From: Simon Marchi @ 2021-12-02 19:10 UTC (permalink / raw)
  To: Tom Tromey, Simon Marchi via Gdb-patches

On 2021-12-02 12:31, Tom Tromey wrote:
>>>>>> "Simon" == Simon Marchi via Gdb-patches <gdb-patches@sourceware.org> writes:
> 
> Simon> Change it so that if re-using a file_handler object, we assign the
> Simon> name to the existing std::string object.
> 
> This seems fine, but it seems to me that this would also be fixed if
> file handlers were allocated with new and could hold smart pointers.
> 
> Tom

I thought it would be more difficult, but not really.  Here's the
updated patch.

From 3c588d67961bca5ccf8b8824917e381c79dca505 Mon Sep 17 00:00:00 2001
From: Simon Marchi <simon.marchi@polymtl.ca>
Date: Thu, 2 Dec 2021 14:04:18 -0500
Subject: [PATCH] gdbsupport: fix memory leak in create_file_handler when
 re-using file handler

ASan made me notice a memory leak, where the memory tied to the file
handle name string wasn't freed.  When register a file handler with an
fd that is already registered, we re-use the file_handler object, so we
ended up creating a new std::string object and overwriting the
file_handler::name pointer, without free-ing the old std::string.

Fix this by allocating file_handler with new, deleting it with
delete, and making file_handler::name not a pointer.

Change-Id: Ie304cc78ab5ae5dfad9a1366e9890c09de651f43
---
 gdbsupport/event-loop.cc | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/gdbsupport/event-loop.cc b/gdbsupport/event-loop.cc
index 98d1ada52cd8..51cfaec4c7ea 100644
--- a/gdbsupport/event-loop.cc
+++ b/gdbsupport/event-loop.cc
@@ -65,8 +65,8 @@ struct file_handler
   /* Argument to pass to proc.  */
   gdb_client_data client_data;
 
-  /* User-friendly name of this handler.  Heap-allocated, owned by this.*/
-  std::string *name;
+  /* User-friendly name of this handler.  */
+  std::string name;
 
   /* If set, this file descriptor is used for a user interface.  */
   bool is_ui;
@@ -315,7 +315,7 @@ create_file_handler (int fd, int mask, handler_func * proc,
      change the data associated with it.  */
   if (file_ptr == NULL)
     {
-      file_ptr = XNEW (file_handler);
+      file_ptr = new file_handler;
       file_ptr->fd = fd;
       file_ptr->ready_mask = 0;
       file_ptr->next_file = gdb_notifier.first_file_handler;
@@ -366,7 +366,7 @@ create_file_handler (int fd, int mask, handler_func * proc,
   file_ptr->proc = proc;
   file_ptr->client_data = client_data;
   file_ptr->mask = mask;
-  file_ptr->name = new std::string (std::move (name));
+  file_ptr->name = std::move (name);
   file_ptr->is_ui = is_ui;
 }
 
@@ -500,8 +500,7 @@ delete_file_handler (int fd)
       prev_ptr->next_file = file_ptr->next_file;
     }
 
-  delete file_ptr->name;
-  xfree (file_ptr);
+  delete file_ptr;
 }
 
 /* Handle the given event by calling the procedure associated to the
@@ -571,7 +570,7 @@ handle_file_event (file_handler *file_ptr, int ready_mask)
 	    {
 	      event_loop_ui_debug_printf (file_ptr->is_ui,
 					  "invoking fd file handler `%s`",
-					  file_ptr->name->c_str ());
+					  file_ptr->name.c_str ());
 	      file_ptr->proc (file_ptr->error, file_ptr->client_data);
 	    }
 	}
-- 
2.33.1


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] gdbsupport: fix memory leak in create_file_handler when re-using file handler
  2021-12-02 19:10   ` Simon Marchi
@ 2021-12-04 18:18     ` Tom Tromey
  2021-12-05  2:47       ` Simon Marchi
  0 siblings, 1 reply; 5+ messages in thread
From: Tom Tromey @ 2021-12-04 18:18 UTC (permalink / raw)
  To: Simon Marchi; +Cc: Tom Tromey, Simon Marchi via Gdb-patches, Simon Marchi

>>>>> "Simon" == Simon Marchi <simon.marchi@efficios.com> writes:

Simon> I thought it would be more difficult, but not really.  Here's the
Simon> updated patch.

Looks great, thank you.

Tom

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] gdbsupport: fix memory leak in create_file_handler when re-using file handler
  2021-12-04 18:18     ` Tom Tromey
@ 2021-12-05  2:47       ` Simon Marchi
  0 siblings, 0 replies; 5+ messages in thread
From: Simon Marchi @ 2021-12-05  2:47 UTC (permalink / raw)
  To: Tom Tromey; +Cc: Simon Marchi via Gdb-patches, Simon Marchi



On 2021-12-04 13:18, Tom Tromey wrote:
>>>>>> "Simon" == Simon Marchi <simon.marchi@efficios.com> writes:
> 
> Simon> I thought it would be more difficult, but not really.  Here's the
> Simon> updated patch.
> 
> Looks great, thank you.
> 
> Tom
> 

Thanks, pushed.

Simon

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2021-12-05  2:48 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-12-02 16:51 [PATCH] gdbsupport: fix memory leak in create_file_handler when re-using file handler Simon Marchi
2021-12-02 17:31 ` Tom Tromey
2021-12-02 19:10   ` Simon Marchi
2021-12-04 18:18     ` Tom Tromey
2021-12-05  2:47       ` Simon Marchi

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).