From: "Martin Liška" <mliska@suse.cz>
To: Jakub Jelinek <jakub@redhat.com>
Cc: Jonathan Wakely <jwakely.gcc@gmail.com>,
Marc Glisse <marc.glisse@inria.fr>,
Richard Biener <richard.guenther@gmail.com>,
Jason Merrill <jason@redhat.com>, Jan Hubicka <hubicka@ucw.cz>,
GCC Patches <gcc-patches@gcc.gnu.org>,
Nathan Sidwell <nathan@acm.org>
Subject: Re: [PATCH] Allow new/delete operator deletion only for replaceable.
Date: Tue, 14 Apr 2020 09:09:43 +0200 [thread overview]
Message-ID: <069c1c26-d74a-0eea-554b-1aeb6881dca7@suse.cz> (raw)
In-Reply-To: <20200410091754.GQ2212@tucnak>
[-- Attachment #1: Type: text/plain, Size: 4513 bytes --]
On 4/10/20 11:17 AM, Jakub Jelinek wrote:
> On Fri, Apr 10, 2020 at 10:29:29AM +0200, Martin Liška wrote:
>> +/* Valid pairs of new and delete operators for DCE. */
>> +static hash_set<nofree_string_hash> *valid_pairs = NULL;
>> +
>> +/* Return that NEW_CALL and DELETE_CALL are a valid pair of new
>> + and delete operators. */
>> +
>> +static bool
>> +valid_new_delete_pair_p (gimple *new_call, gimple *delete_call)
>> +{
>> + const char *new_name
>> + = IDENTIFIER_POINTER (DECL_ASSEMBLER_NAME (gimple_call_fndecl (new_call)));
>> + const char *delete_name
>> + = IDENTIFIER_POINTER (DECL_ASSEMBLER_NAME (gimple_call_fndecl (delete_call)));
>> +
>> + if (new_name[0] == '_' and new_name[1] == '_')
>> + ++new_name;
>> + if (delete_name[0] == '_' and delete_name[1] == '_')
>> + ++delete_name;
>> +
>> + char *needle = concat (new_name, ":", delete_name, NULL);
>> +
>> + if (valid_pairs == NULL)
>> + {
>> + valid_pairs = new hash_set<nofree_string_hash> ();
>> + /* Invalid pairs:
>> + non-[] and []
>> + aligned and non-aligned
>> + */
>> +
>> + const char *pairs[] = {
>> + /* non-[] operators. */
>> + "_Znwm:_ZdlPv" ,
>> + "_Znwm:_ZdlPvm" ,
>
> Formatting, I think the /* and "_Z should be two columns from const char,
> and no space before ,
> Not sure I like using the : and using a hash table looks like an overkill to
> me, don't we have only 8 options, and most of them with different identifier
> lengths? The concat itself will take a while...
> And, what's worse, the m in there is really different on different targets.
> It can be j, m or y depending on what fundamental type size_t is.
>
> So why not (completely untested):
Hi.
I've updated a bit the Jakub's patch.
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
Ready to be installed?
Thanks,
Martin
>
> static bool
> valid_new_delete_pair_p (gimple *new_call, gimple *delete_call)
> {
> tree new_asm = DECL_ASSEMBLER_NAME (gimple_call_fndecl (new_call));
> tree delete_asm = DECL_ASSEMBLER_NAME (gimple_call_fndecl (delete_call));
> const char *new_name = IDENTIFIER_POINTER (new_asm);
> const char *delete_name = IDENTIFIER_POINTER (delete_asm);
> unsigned int new_len = IDENTIFIER_LENGTH (mew_asm);
> unsigned int delete_len = IDENTIFIER_LENGTH (delete_asm);
> if (new_name < 5 || delete_len < 6)
> return false;
> if (new_name[0] == '_')
> ++new_name, --new_len;
> if (new_name[0] == '_')
> ++new_name, --new_len;
> if (delete_name[0] == '_')
> ++delete_name, --delete_len;
> if (delete_name[0] == '_')
> ++delete_name, --delete_len;
> if (new_len < 4 || delete_len < 5)
> return false;
> /* *_len is now just the length after initial underscores. */
> if (new_name[0] != 'Z' || new_name[1] != 'n')
> return false;
> if (delete_name[0] != 'Z' || delete_name[1] != 'd')
> return false;
> /* _Znw must match _Zdl, _Zna must match _Zda. */
> if ((new_name[2] != 'w' || delete_name[2] != 'l')
> && (new_name[2] != 'a' || delete_name[2] != 'a'))
> return false;
> if (new_name[3] != 'j' && new_name[3] != 'm' && new_name[3] != 'y')
> return false;
> if (delete_name[3] != 'P' || delete_name[4] != 'v')
> return false;
> if (new_len == 4
> || (new_len == 18 && !memcmp (new_name + 4, "RKSt9nothrow_t", 14)))
> {
> /* _ZnXY or _ZnXYRKSt9nothrow_t matches
> _ZdXPv, _ZdXPvY and _ZdXPvRKSt9nothrow_t. */
> if (delete_len == 5)
> return true;
> if (delete_len == 6 && delete_name[5] == new_name[3])
> return true;
> if (delete_len == 19 && !memcmp (delete_name + 5, "RKSt9nothrow_t", 14))
> return true;
> }
> else if ((new_len == 19 && !memcmp (new_name + 4, "St11align_val_t", 15))
> || (new_len == 33
> && !memcmp (new_name + 4, "St11align_val_tRKSt9nothrow_t", 29)))
> {
> /* _ZnXYSt11align_val_t or _ZnXYSt11align_val_tRKSt9nothrow_t matches
> _ZdXPvSt11align_val_t or _ZdXPvYSt11align_val_t or or
> _ZdXPvSt11align_val_tRKSt9nothrow_t. */
> if (delete_len == 20 && !memcmp (delete_name + 5, "St11align_val_t", 15))
> return true;
> if (delete_len == 21
> && delete_name[5] == new_name[3]
> && !memcmp (delete_name + 6, "St11align_val_t", 15))
> return true;
> if (delete_len == 34
> && !memcmp (delete_name + 5, "St11align_val_tRKSt9nothrow_t", 29))
> return true;
> }
> return false;
> }
>
> Jakub
>
[-- Attachment #2: 0001-List-valid-pairs-for-new-and-delete-operators.patch --]
[-- Type: text/x-patch, Size: 7492 bytes --]
From 60626d4b040bc502c563b7c014b62751a517705a Mon Sep 17 00:00:00 2001
From: Martin Liska <mliska@suse.cz>
Date: Thu, 9 Apr 2020 15:50:58 +0200
Subject: [PATCH] List valid pairs for new and delete operators.
gcc/ChangeLog:
2020-04-09 Martin Liska <mliska@suse.cz>
Jakub Jelinek <jakub@redhat.com>
PR c++/94314
* cgraphclones.c (set_new_clone_decl_and_node_flags): Drop
DECL_IS_REPLACEABLE_OPERATOR during cloning.
* tree-ssa-dce.c (valid_new_delete_pair_p): New function.
(propagate_necessity): Check operator names.
(perform_tree_ssa_dce): Delete valid_pairs.
gcc/testsuite/ChangeLog:
2020-04-09 Martin Liska <mliska@suse.cz>
Jakub Jelinek <jakub@redhat.com>
PR c++/94314
* g++.dg/pr94314-4.C: New test.
---
gcc/cgraphclones.c | 2 +
gcc/testsuite/g++.dg/pr94314-4.C | 33 ++++++++++
gcc/tree-ssa-dce.c | 103 ++++++++++++++++++++++++++++---
3 files changed, 128 insertions(+), 10 deletions(-)
create mode 100644 gcc/testsuite/g++.dg/pr94314-4.C
diff --git a/gcc/cgraphclones.c b/gcc/cgraphclones.c
index c73b8f810f0..8f541a28b6e 100644
--- a/gcc/cgraphclones.c
+++ b/gcc/cgraphclones.c
@@ -165,6 +165,7 @@ set_new_clone_decl_and_node_flags (cgraph_node *new_node)
DECL_STATIC_DESTRUCTOR (new_node->decl) = 0;
DECL_SET_IS_OPERATOR_NEW (new_node->decl, 0);
DECL_SET_IS_OPERATOR_DELETE (new_node->decl, 0);
+ DECL_IS_REPLACEABLE_OPERATOR (new_node->decl) = 0;
new_node->externally_visible = 0;
new_node->local = 1;
@@ -1030,6 +1031,7 @@ cgraph_node::create_version_clone_with_body
DECL_STATIC_DESTRUCTOR (new_decl) = 0;
DECL_SET_IS_OPERATOR_NEW (new_decl, 0);
DECL_SET_IS_OPERATOR_DELETE (new_decl, 0);
+ DECL_IS_REPLACEABLE_OPERATOR (new_decl) = 0;
/* Create the new version's call-graph node.
and update the edges of the new node. */
diff --git a/gcc/testsuite/g++.dg/pr94314-4.C b/gcc/testsuite/g++.dg/pr94314-4.C
new file mode 100644
index 00000000000..afa2a443dc4
--- /dev/null
+++ b/gcc/testsuite/g++.dg/pr94314-4.C
@@ -0,0 +1,33 @@
+/* PR c++/94314. */
+/* { dg-do run } */
+/* { dg-options "-O2 -fdump-tree-cddce-details -std=c++14" } */
+/* { dg-additional-options "-fdelete-null-pointer-checks" } */
+
+#include <stdio.h>
+
+int count = 0;
+
+__attribute__((malloc, noinline)) void* operator new[](__SIZE_TYPE__ sz) {
+ ++count;
+ return ::operator new(sz);
+}
+
+void operator delete[](void* ptr) noexcept {
+ --count;
+ ::operator delete(ptr);
+}
+
+void operator delete[](void* ptr, __SIZE_TYPE__ sz) noexcept {
+ --count;
+ ::operator delete(ptr, sz);
+}
+
+int main() {
+ delete[] new int[1];
+ if (count != 0)
+ __builtin_abort ();
+
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-not "Deleting : operator delete" "cddce1"} } */
diff --git a/gcc/tree-ssa-dce.c b/gcc/tree-ssa-dce.c
index fd5f24c746c..69566cd50e2 100644
--- a/gcc/tree-ssa-dce.c
+++ b/gcc/tree-ssa-dce.c
@@ -646,6 +646,80 @@ degenerate_phi_p (gimple *phi)
return true;
}
+/* Valid pairs of new and delete operators for DCE. */
+static hash_set<nofree_string_hash> *valid_pairs = NULL;
+
+/* Return that NEW_CALL and DELETE_CALL are a valid pair of new
+ and delete operators. */
+
+static bool
+valid_new_delete_pair_p (gimple *new_call, gimple *delete_call)
+{
+ tree new_asm = DECL_ASSEMBLER_NAME (gimple_call_fndecl (new_call));
+ tree delete_asm = DECL_ASSEMBLER_NAME (gimple_call_fndecl (delete_call));
+ const char *new_name = IDENTIFIER_POINTER (new_asm);
+ const char *delete_name = IDENTIFIER_POINTER (delete_asm);
+ unsigned int new_len = IDENTIFIER_LENGTH (new_asm);
+ unsigned int delete_len = IDENTIFIER_LENGTH (delete_asm);
+
+ if (new_len < 5 || delete_len < 6)
+ return false;
+ if (new_name[0] == '_')
+ ++new_name, --new_len;
+ if (new_name[0] == '_')
+ ++new_name, --new_len;
+ if (delete_name[0] == '_')
+ ++delete_name, --delete_len;
+ if (delete_name[0] == '_')
+ ++delete_name, --delete_len;
+ if (new_len < 4 || delete_len < 5)
+ return false;
+ /* *_len is now just the length after initial underscores. */
+ if (new_name[0] != 'Z' || new_name[1] != 'n')
+ return false;
+ if (delete_name[0] != 'Z' || delete_name[1] != 'd')
+ return false;
+ /* _Znw must match _Zdl, _Zna must match _Zda. */
+ if ((new_name[2] != 'w' || delete_name[2] != 'l')
+ && (new_name[2] != 'a' || delete_name[2] != 'a'))
+ return false;
+ /* 'j', 'm' and 'y' correspond to size_t. */
+ if (new_name[3] != 'j' && new_name[3] != 'm' && new_name[3] != 'y')
+ return false;
+ if (delete_name[3] != 'P' || delete_name[4] != 'v')
+ return false;
+ if (new_len == 4
+ || (new_len == 18 && !memcmp (new_name + 4, "RKSt9nothrow_t", 14)))
+ {
+ /* _ZnXY or _ZnXYRKSt9nothrow_t matches
+ _ZdXPv, _ZdXPvY and _ZdXPvRKSt9nothrow_t. */
+ if (delete_len == 5)
+ return true;
+ if (delete_len == 6 && delete_name[5] == new_name[3])
+ return true;
+ if (delete_len == 19 && !memcmp (delete_name + 5, "RKSt9nothrow_t", 14))
+ return true;
+ }
+ else if ((new_len == 19 && !memcmp (new_name + 4, "St11align_val_t", 15))
+ || (new_len == 33
+ && !memcmp (new_name + 4, "St11align_val_tRKSt9nothrow_t", 29)))
+ {
+ /* _ZnXYSt11align_val_t or _ZnXYSt11align_val_tRKSt9nothrow_t matches
+ _ZdXPvSt11align_val_t or _ZdXPvYSt11align_val_t or or
+ _ZdXPvSt11align_val_tRKSt9nothrow_t. */
+ if (delete_len == 20 && !memcmp (delete_name + 5, "St11align_val_t", 15))
+ return true;
+ if (delete_len == 21
+ && delete_name[5] == new_name[3]
+ && !memcmp (delete_name + 6, "St11align_val_t", 15))
+ return true;
+ if (delete_len == 34
+ && !memcmp (delete_name + 5, "St11align_val_tRKSt9nothrow_t", 29))
+ return true;
+ }
+ return false;
+}
+
/* Propagate necessity using the operands of necessary statements.
Process the uses on each statement in the worklist, and add all
feeding statements which contribute to the calculation of this
@@ -824,16 +898,23 @@ propagate_necessity (bool aggressive)
|| DECL_FUNCTION_CODE (def_callee) == BUILT_IN_CALLOC))
|| DECL_IS_REPLACEABLE_OPERATOR_NEW_P (def_callee)))
{
- /* Delete operators can have alignment and (or) size as next
- arguments. When being a SSA_NAME, they must be marked
- as necessary. */
- if (is_delete_operator && gimple_call_num_args (stmt) >= 2)
- for (unsigned i = 1; i < gimple_call_num_args (stmt); i++)
- {
- tree arg = gimple_call_arg (stmt, i);
- if (TREE_CODE (arg) == SSA_NAME)
- mark_operand_necessary (arg);
- }
+ if (is_delete_operator)
+ {
+ if (!valid_new_delete_pair_p (def_stmt, stmt))
+ mark_operand_necessary (gimple_call_arg (stmt, 0));
+
+ /* Delete operators can have alignment and (or) size
+ as next arguments. When being a SSA_NAME, they
+ must be marked as necessary. */
+ if (gimple_call_num_args (stmt) >= 2)
+ for (unsigned i = 1; i < gimple_call_num_args (stmt);
+ i++)
+ {
+ tree arg = gimple_call_arg (stmt, i);
+ if (TREE_CODE (arg) == SSA_NAME)
+ mark_operand_necessary (arg);
+ }
+ }
continue;
}
@@ -1662,6 +1743,8 @@ perform_tree_ssa_dce (bool aggressive)
visited = BITMAP_ALLOC (NULL);
propagate_necessity (aggressive);
BITMAP_FREE (visited);
+ delete valid_pairs;
+ valid_pairs = NULL;
something_changed |= eliminate_unnecessary_stmts ();
something_changed |= cfg_altered;
--
2.26.0
next prev parent reply other threads:[~2020-04-14 7:09 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-30 8:40 [PATCH] Check DECL_CONTEXT of new/delete operators Martin Liška
2020-03-30 8:53 ` Richard Biener
2020-03-31 12:29 ` Jan Hubicka
2020-03-31 12:38 ` Martin Liška
2020-04-03 15:26 ` Jan Hubicka
2020-04-03 15:42 ` Jan Hubicka
2020-04-04 11:53 ` Jan Hubicka
2020-04-06 9:27 ` Richard Biener
2020-04-06 15:10 ` Jason Merrill
2020-04-06 8:34 ` Martin Liška
2020-04-06 12:45 ` Nathan Sidwell
2020-04-07 8:26 ` Jonathan Wakely
2020-04-07 9:29 ` Richard Biener
2020-04-07 9:49 ` Jan Hubicka
2020-04-07 10:22 ` Richard Biener
2020-04-07 10:42 ` Martin Liška
2020-04-07 11:41 ` Jonathan Wakely
2020-04-07 10:46 ` Martin Liška
2020-04-07 11:29 ` Jonathan Wakely
2020-04-07 11:40 ` Richard Biener
2020-04-07 11:46 ` Jonathan Wakely
2020-04-07 11:57 ` Richard Biener
2020-04-07 15:00 ` [PATCH] Allow new/delete operator deletion only for replaceable Martin Liška
2020-04-08 8:47 ` Richard Biener
2020-04-08 13:20 ` Jason Merrill
2020-04-08 13:32 ` Jakub Jelinek
2020-04-08 13:34 ` Jason Merrill
2020-04-08 15:16 ` Martin Liška
2020-04-08 15:46 ` Jan Hubicka
2020-04-08 16:06 ` Jakub Jelinek
2020-04-09 5:05 ` Martin Liška
2020-04-09 6:45 ` Richard Biener
2020-04-09 6:59 ` Martin Liška
2020-04-09 7:21 ` Richard Biener
2020-04-09 7:55 ` Jakub Jelinek
2020-04-09 8:04 ` Marc Glisse
2020-04-09 8:13 ` Jonathan Wakely
2020-04-10 8:08 ` Martin Liška
2020-04-10 8:18 ` Jonathan Wakely
2020-04-10 8:29 ` Martin Liška
2020-04-10 9:17 ` Jakub Jelinek
2020-04-14 7:09 ` Martin Liška [this message]
2020-04-14 7:11 ` Martin Liška
2020-04-14 8:37 ` Jakub Jelinek
2020-04-14 10:54 ` Martin Liška
2020-04-17 7:05 ` Jakub Jelinek
2020-04-17 8:12 ` Jonathan Wakely
2020-04-10 8:37 ` Marc Glisse
2020-04-10 9:11 ` Iain Sandoe
2020-04-09 16:55 ` Jason Merrill
2020-04-07 15:16 ` [PATCH] Check DECL_CONTEXT of new/delete operators Jonathan Wakely
2020-04-08 7:34 ` Richard Biener
2020-04-08 8:11 ` Jonathan Wakely
2020-04-07 14:11 ` Nathan Sidwell
2020-03-30 9:29 ` Marc Glisse
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=069c1c26-d74a-0eea-554b-1aeb6881dca7@suse.cz \
--to=mliska@suse.cz \
--cc=gcc-patches@gcc.gnu.org \
--cc=hubicka@ucw.cz \
--cc=jakub@redhat.com \
--cc=jason@redhat.com \
--cc=jwakely.gcc@gmail.com \
--cc=marc.glisse@inria.fr \
--cc=nathan@acm.org \
--cc=richard.guenther@gmail.com \
/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).