public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: "Marc Poulhiès" <poulhies@adacore.com>
To: gcc-patches@gcc.gnu.org
Cc: Gary Dismukes <dismukes@adacore.com>
Subject: [COMMITTED] ada: Excess finalization on return of call to dispatching constructor
Date: Mon,  6 May 2024 11:18:43 +0200	[thread overview]
Message-ID: <20240506091843.1586654-1-poulhies@adacore.com> (raw)

From: Gary Dismukes <dismukes@adacore.com>

The compiler expands a too-early finalization call for the result
object of an extended return statement that returns a call to a
dispatching constructor function for a limited interface type,
resulting in premature (and extra) finalization of the result.
The temporary object that the compiler creates to hold the result
of the build-in-place call loses the fact that it comes from a
return, and the wrong BIP allocation form may be passed in the
call to the dispatching constructor, and the later code for dealing
with finalization in Exp_Ch7.Build_Finalizer incorrectly finalizes
the result object.

gcc/ada/

	* exp_ch6.adb
	(Make_Build_In_Place_Iface_Call_In_Object_Declaration): Set the
	Is_Return_Object flag on the entity of the temp object created to
	hold the BIP call result, from the flag on the passed-in object
	declaration's entity. Update copyright notice to 2024.

Tested on x86_64-pc-linux-gnu, committed on master.

---
 gcc/ada/exp_ch6.adb | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/gcc/ada/exp_ch6.adb b/gcc/ada/exp_ch6.adb
index fcfd1d7f0bf..a89c9af0bb2 100644
--- a/gcc/ada/exp_ch6.adb
+++ b/gcc/ada/exp_ch6.adb
@@ -9393,6 +9393,16 @@ package body Exp_Ch6 is
       Insert_Action (Obj_Decl, Tmp_Decl);
       Expander_Mode_Restore;
 
+      --  Inherit Is_Return_Object from the parent object to the temp object,
+      --  so that Make_In_Build_Place_Call_In_Object_Declaration will handle
+      --  the temp properly in cases where there's a BIP_Alloc_Form formal of
+      --  an enclosing function that should be passed along (and which also
+      --  ensures that if the BIP call is used as a function result and it
+      --  requires finalization, then it will not be finalized prematurely
+      --  or redundantly).
+
+      Set_Is_Return_Object (Tmp_Id, Is_Return_Object (Obj_Id));
+
       Make_Build_In_Place_Call_In_Object_Declaration
         (Obj_Decl      => Tmp_Decl,
          Function_Call => Expression (Tmp_Decl));
-- 
2.43.2


                 reply	other threads:[~2024-05-06  9:18 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20240506091843.1586654-1-poulhies@adacore.com \
    --to=poulhies@adacore.com \
    --cc=dismukes@adacore.com \
    --cc=gcc-patches@gcc.gnu.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).