public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
* [PATCH] Use scoped_value_mark in two more places
@ 2023-05-25 12:57 Tom Tromey
  2023-06-09 13:42 ` Tom Tromey
  0 siblings, 1 reply; 2+ messages in thread
From: Tom Tromey @ 2023-05-25 12:57 UTC (permalink / raw)
  To: gdb-patches; +Cc: Tom Tromey

I found a couple of spots that could use scoped_value_mark.  One of
them is a spot that didn't consider the possibility that value_mark
can return NULL.  I tend to doubt this can be seen in this context,
but nevertheless this is safer.

Regression tested on x86-64 Fedora 36.
---
 gdb/f-lang.c | 15 +++++++--------
 gdb/value.c  |  8 +++-----
 2 files changed, 10 insertions(+), 13 deletions(-)

diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 365e0c0489b..9f579c5c990 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -261,8 +261,8 @@ class fortran_array_repacker_base_impl
   {
     if (inner_p)
       {
-	gdb_assert (m_mark == nullptr);
-	m_mark = value_mark ();
+	gdb_assert (!m_mark.has_value ());
+	m_mark.emplace ();
       }
   }
 
@@ -272,9 +272,8 @@ class fortran_array_repacker_base_impl
   {
     if (inner_p)
       {
-	gdb_assert (m_mark != nullptr);
-	value_free_to_mark (m_mark);
-	m_mark = nullptr;
+	gdb_assert (m_mark.has_value ());
+	m_mark.reset ();
       }
   }
 
@@ -295,9 +294,9 @@ class fortran_array_repacker_base_impl
      written.  */
   LONGEST m_dest_offset;
 
-  /* Set with a call to VALUE_MARK, and then reset after calling
-     VALUE_FREE_TO_MARK.  */
-  struct value *m_mark = nullptr;
+  /* Set and reset to handle removing intermediate values from the
+     value chain.  */
+  gdb::optional<scoped_value_mark> m_mark;
 };
 
 /* A class used by FORTRAN_VALUE_SUBARRAY when repacking Fortran array
diff --git a/gdb/value.c b/gdb/value.c
index 980722a6dd7..fb9466493b5 100644
--- a/gdb/value.c
+++ b/gdb/value.c
@@ -3878,7 +3878,9 @@ value::fetch_lazy_register ()
   frame_info_ptr next_frame;
   int regnum;
   struct type *type = check_typedef (this->type ());
-  struct value *new_val = this, *mark = value_mark ();
+  struct value *new_val = this;
+
+  scoped_value_mark mark;
 
   /* Offsets are not supported here; lazy register values must
      refer to the entire register.  */
@@ -3982,10 +3984,6 @@ value::fetch_lazy_register ()
 
       frame_debug_printf ("%s", debug_file.c_str ());
     }
-
-  /* Dispose of the intermediate values.  This prevents
-     watchpoints from trying to watch the saved frame pointer.  */
-  value_free_to_mark (mark);
 }
 
 /* See value.h.  */
-- 
2.40.0


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] Use scoped_value_mark in two more places
  2023-05-25 12:57 [PATCH] Use scoped_value_mark in two more places Tom Tromey
@ 2023-06-09 13:42 ` Tom Tromey
  0 siblings, 0 replies; 2+ messages in thread
From: Tom Tromey @ 2023-06-09 13:42 UTC (permalink / raw)
  To: Tom Tromey via Gdb-patches; +Cc: Tom Tromey

>>>>> "Tom" == Tom Tromey via Gdb-patches <gdb-patches@sourceware.org> writes:

Tom> I found a couple of spots that could use scoped_value_mark.  One of
Tom> them is a spot that didn't consider the possibility that value_mark
Tom> can return NULL.  I tend to doubt this can be seen in this context,
Tom> but nevertheless this is safer.

Tom> Regression tested on x86-64 Fedora 36.

I'm checking this in.

Tom

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2023-06-09 13:42 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-05-25 12:57 [PATCH] Use scoped_value_mark in two more places Tom Tromey
2023-06-09 13:42 ` Tom Tromey

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).