[PATCH 1/2] Fix resource leak found by Coverity

[PATCH 1/2] Fix resource leak found by Coverity

[Gary Benson]

This commit fixes a resource leak found by Coverity, where cli_interp's
constructor creates a new cli_ui_out, but there was no corresponding
destructor to free it.

gdb/ChangeLog:

	* cli/cli-interp.c (cli_interp::~cli_interp): New function.
---
 gdb/ChangeLog        | 4 ++++
 gdb/cli/cli-interp.c | 6 ++++++
 2 files changed, 10 insertions(+)

diff --git a/gdb/cli/cli-interp.c b/gdb/cli/cli-interp.c
index 2aa41d6..75d8093 100644
--- a/gdb/cli/cli-interp.c
+++ b/gdb/cli/cli-interp.c
@@ -44,6 +44,7 @@ class cli_interp final : public cli_interp_base
 {
  public:
   explicit cli_interp (const char *name);
+  ~cli_interp ();
 
   void init (bool top_level) override;
   void resume () override;
@@ -62,6 +63,11 @@ cli_interp::cli_interp (const char *name)
   this->cli_uiout = cli_out_new (gdb_stdout);
 }
 
+cli_interp::~cli_interp ()
+{
+  delete cli_uiout;
+}
+
 /* Suppress notification struct.  */
 struct cli_suppress_notification cli_suppress_notification =
   {
-- 
1.8.3.1

[PATCH 0/2] Fix resource leaks found by Coverity

[Gary Benson]

Hi all,

These two patches fix resource leaks found by Coverity,
where resources allocated by constructors were not freed
by the relevant destructors.

Built and regtested on RHEL 7.5 x86_64.

Ok to commit?

Thanks,
Gary

[PATCH 2/2] Fix resource leak found by Coverity

[Gary Benson]

This commit fixes a resource leak found by Coverity, where interp's
constructor allocated memory for m_name that interp's destructor did
not free.

gdb/ChangeLog:

	* interps.h (interp::m_name): Make private and mutable.
	* interps.c (interp::~interp): Free m_name.
---
 gdb/ChangeLog | 5 +++++
 gdb/interps.c | 4 +++-
 gdb/interps.h | 4 +++-
 3 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/gdb/interps.c b/gdb/interps.c
index 6fe4c74..883e042 100644
--- a/gdb/interps.c
+++ b/gdb/interps.c
@@ -84,7 +84,9 @@ interp::interp (const char *name)
 }
 
 interp::~interp ()
-{}
+{
+  xfree (m_name);
+}
 
 /* An interpreter factory.  Maps an interpreter name to the factory
    function that instantiates an interpreter by that name.  */
diff --git a/gdb/interps.h b/gdb/interps.h
index 74c9a80..dbf91f1 100644
--- a/gdb/interps.h
+++ b/gdb/interps.h
@@ -80,10 +80,12 @@ public:
   }
 
   /* This is the name in "-i=" and "set interpreter".  */
-  const char *m_name;
+private:
+  char *m_name;
 
   /* Interpreters are stored in a linked list, this is the next
      one...  */
+public:
   struct interp *next;
 
   /* Has the init method been run?  */
-- 
1.8.3.1

Re: [PATCH 2/2] Fix resource leak found by Coverity

[Tom Tromey]

>>>>> "Gary" == Gary Benson <gbenson@redhat.com> writes:

Gary> This commit fixes a resource leak found by Coverity, where interp's
Gary> constructor allocated memory for m_name that interp's destructor did
Gary> not free.

Elsewhere Pedro asked for a different patch subject, so I think that
would apply here as well.

Gary> 	* interps.h (interp::m_name): Make private and mutable.
Gary> 	* interps.c (interp::~interp): Free m_name.

This is ok.

Tom

Re: [PATCH 1/2] Fix resource leak found by Coverity

[Tom Tromey]

>>>>> "Gary" == Gary Benson <gbenson@redhat.com> writes:

Gary> This commit fixes a resource leak found by Coverity, where cli_interp's
Gary> constructor creates a new cli_ui_out, but there was no corresponding
Gary> destructor to free it.

Same deal with the patch title.

Gary> 	* cli/cli-interp.c (cli_interp::~cli_interp): New function.

I wonder if it's possible for the ui-out to be used after the
interpreter is destroyed.  A simple check with valgrind would show that,
I think.

This is ok if that succeeds.

Tom