On Fri, 10 Jan 2025, Richard Biener wrote:
> The following were found compiling SPEC CPU 2017 with valgrind.
>
> Bootstrap and regtest pending on x86_64-unknown-linux-gnu.
>
> * tree-vect-slp.cc (vect_analyze_slp): Release saved_stmts
> vector.
> (vect_build_slp_tree_2): Release new_oprnds_info when not
> used.
I sneaked in one more fix, pushed as follows after bootstrapping
and regtesting on x86_64-unknown-linux-gnu.
>From 0b0a460f80b16c0d7dba402a0f5dfa2cff7a8a56 Mon Sep 17 00:00:00 2001
From: Richard Biener <rguent...@suse.de>
Date: Fri, 10 Jan 2025 16:25:35 +0100
Subject: [PATCH] More memory leak fixes
To: gcc-patches@gcc.gnu.org
The following were found compiling SPEC CPU 2017 with valgrind.
* tree-vect-slp.cc (vect_analyze_slp): Release saved_stmts
vector.
(vect_build_slp_tree_2): Release new_oprnds_info when not
used.
(vect_analyze_slp): Release root_stmts when gcond SLP
build fails.
---
gcc/tree-vect-slp.cc | 13 ++++++++-----
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/gcc/tree-vect-slp.cc b/gcc/tree-vect-slp.cc
index 337506419d9..02e7f5c4d58 100644
--- a/gcc/tree-vect-slp.cc
+++ b/gcc/tree-vect-slp.cc
@@ -2678,6 +2678,8 @@ out:
nops = 1;
has_two_operators_perm = true;
}
+ else
+ vect_free_oprnd_info (new_oprnds_info);
}
auto_vec<slp_tree, 4> children;
@@ -4951,8 +4953,8 @@ vect_analyze_slp (vec_info *vinfo, unsigned max_tree_size,
max_tree_size, &limit,
bst_map, NULL, force_single_lane);
}
- saved_stmts.release ();
}
+ saved_stmts.release ();
}
/* Make sure to vectorize only-live stmts, usually inductions. */
@@ -5013,10 +5015,11 @@ vect_analyze_slp (vec_info *vinfo, unsigned
max_tree_size,
stmts.create (1);
stmts.quick_push (vect_stmt_to_vectorize (varg));
- vect_build_slp_instance (vinfo, slp_inst_kind_gcond,
- stmts, roots, remain,
- max_tree_size, &limit,
- bst_map, NULL, force_single_lane);
+ if (! vect_build_slp_instance (vinfo, slp_inst_kind_gcond,
+ stmts, roots, remain,
+ max_tree_size, &limit,
+ bst_map, NULL, force_single_lane))
+ roots.release ();
}
/* Find and create slp instances for inductions that have been forced
--
2.43.0
