From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from server.nextmovesoftware.com (server.nextmovesoftware.com [162.254.253.69]) by sourceware.org (Postfix) with ESMTPS id C08653858406 for ; Mon, 22 Nov 2021 08:39:44 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org C08653858406 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=nextmovesoftware.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=nextmovesoftware.com DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=nextmovesoftware.com; s=default; h=Content-Type:MIME-Version:Message-ID: Date:Subject:To:From:Sender:Reply-To:Cc:Content-Transfer-Encoding:Content-ID: Content-Description:Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc :Resent-Message-ID:In-Reply-To:References:List-Id:List-Help:List-Unsubscribe: List-Subscribe:List-Post:List-Owner:List-Archive; bh=kTBNgODbppTidg4JFfzXJsJ0QHOlL6kidXYg6+xWWZ8=; b=E+MvhyeSvJTzOT4Wp3t78sXULE 76IQvaR5lhmCaotL8if1+p0fPToCYRbk4bJFcQOCHRmpZEtXQKSYtrZk4Hmbc5qGZ5QNDyQuSXDwE +fXk+Qv2ckrBHrYV1yLTDUd+k+4SgYiqg9q2h+mt8wFXziHMpT6gmBFs55qsRmDsKoECaMy5HcRHJ DPmfu6oJse73jpg3hJtLDjxeUi+fMh3ED1FM8byiAssTgcWdyns0VBt0DbAynof1QgBzhH2Rw4hZm sj+70vC95yl1sD4RA3FI3QJkYeFgOBAJZAVaPtyLsJw5Yw0BBEQbvUiEzXWo6iymgBJpFGEzSrb8q v+immjwA==; Received: from host81-157-90-156.range81-157.btcentralplus.com ([81.157.90.156]:59730 helo=Dell) by server.nextmovesoftware.com with esmtpsa (TLS1.2) tls TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 (Exim 4.94.2) (envelope-from ) id 1mp4rY-0000Tj-1N for gcc-patches@gcc.gnu.org; Mon, 22 Nov 2021 03:39:44 -0500 From: "Roger Sayle" To: "'GCC Patches'" Subject: [PATCH] tree-optimization/103345: Improved load merging Date: Mon, 22 Nov 2021 08:39:42 -0000 Message-ID: <001601d7df7c$7f739920$7e5acb60$@nextmovesoftware.com> MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_NextPart_000_0017_01D7DF7C.7F739920" X-Mailer: Microsoft Outlook 16.0 Thread-Index: Adffe+tI/5IpmQXbSgKhv4MZkfj+2w== Content-Language: en-gb X-AntiAbuse: This header was added to track abuse, please include it with any abuse report X-AntiAbuse: Primary Hostname - server.nextmovesoftware.com X-AntiAbuse: Original Domain - gcc.gnu.org X-AntiAbuse: Originator/Caller UID/GID - [47 12] / [47 12] X-AntiAbuse: Sender Address Domain - nextmovesoftware.com X-Get-Message-Sender-Via: server.nextmovesoftware.com: authenticated_id: roger@nextmovesoftware.com X-Authenticated-Sender: server.nextmovesoftware.com: roger@nextmovesoftware.com X-Source: X-Source-Args: X-Source-Dir: X-Spam-Status: No, score=-12.6 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, SPF_HELO_NONE, SPF_PASS, TXREP autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: gcc-patches@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 22 Nov 2021 08:39:46 -0000 This is a multipart message in MIME format. ------=_NextPart_000_0017_01D7DF7C.7F739920 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit This patch implements PR tree-optimization/103345 to merge adjacent loads when combined with addition or bitwise xor. The current code in gimple-ssa-store-merging.c's find_bswap_or_nop alreay handles ior, so that all that's required is to treat PLUS_EXPR and BIT_XOR_EXPR in the same way at BIT_IOR_EXPR. Many thanks to Andrew Pinski for pointing out that this also resolves PR target/98953. This patch has been tested on x86_64-pc-linux-gnu with make bootstrap and make -k check with no new failures. The new testcases should also pass (but haven't been tested) on other endian targets. Ok for mainline? 2021-11-22 Roger Sayle gcc/ChangeLog PR tree-optimization/98953 PR tree-optimization/103345 * gimple-ssa-store-merging.c (find_bswap_or_nop_1): Handle BIT_XOR_EXPR and PLUS_EXPR the same as BIT_IOR_EXPR. (pass_optimize_bswap::execute): Likewise. gcc/testsuite/ChangeLog PR tree-optimization/98953 PR tree-optimization/103345 * gcc.dg/tree-ssa/pr98953.c: New test case. * gcc.dg/tree-ssa/pr103345.c: New test case. Thanks in advance, Roger -- ------=_NextPart_000_0017_01D7DF7C.7F739920 Content-Type: text/plain; name="patchx.txt" Content-Transfer-Encoding: quoted-printable Content-Disposition: attachment; filename="patchx.txt" diff --git a/gcc/gimple-ssa-store-merging.c = b/gcc/gimple-ssa-store-merging.c=0A= index 4efa200..1740c9e 100644=0A= --- a/gcc/gimple-ssa-store-merging.c=0A= +++ b/gcc/gimple-ssa-store-merging.c=0A= @@ -742,10 +742,7 @@ find_bswap_or_nop_1 (gimple *stmt, struct = symbolic_number *n, int limit)=0A= struct symbolic_number n1, n2;=0A= gimple *source_stmt, *source_stmt2;=0A= =0A= - if (code !=3D BIT_IOR_EXPR)=0A= - return NULL;=0A= -=0A= - if (TREE_CODE (rhs2) !=3D SSA_NAME)=0A= + if (!rhs2 || TREE_CODE (rhs2) !=3D SSA_NAME)=0A= return NULL;=0A= =0A= rhs2_stmt =3D SSA_NAME_DEF_STMT (rhs2);=0A= @@ -753,6 +750,8 @@ find_bswap_or_nop_1 (gimple *stmt, struct = symbolic_number *n, int limit)=0A= switch (code)=0A= {=0A= case BIT_IOR_EXPR:=0A= + case BIT_XOR_EXPR:=0A= + case PLUS_EXPR:=0A= source_stmt1 =3D find_bswap_or_nop_1 (rhs1_stmt, &n1, limit - 1);=0A= =0A= if (!source_stmt1)=0A= @@ -1495,6 +1494,8 @@ pass_optimize_bswap::execute (function *fun)=0A= continue;=0A= /* Fall through. */=0A= case BIT_IOR_EXPR:=0A= + case BIT_XOR_EXPR:=0A= + case PLUS_EXPR:=0A= break;=0A= case CONSTRUCTOR:=0A= {=0A= diff --git a/gcc/testsuite/gcc.dg/tree-ssa/pr103345.c = b/gcc/testsuite/gcc.dg/tree-ssa/pr103345.c=0A= new file mode 100644=0A= index 0000000..94388b5=0A= --- /dev/null=0A= +++ b/gcc/testsuite/gcc.dg/tree-ssa/pr103345.c=0A= @@ -0,0 +1,53 @@=0A= +/* { dg-do compile } */=0A= +/* { dg-options "-O2 -fdump-tree-bswap-details" } */=0A= +=0A= +typedef unsigned int uint32_t;=0A= +typedef unsigned char uint8_t;=0A= +=0A= +uint32_t load_le_32_or(const uint8_t *ptr)=0A= +{=0A= +#if __BYTE_ORDER__ =3D=3D __ORDER_LITTLE_ENDIAN__=0A= + return ((uint32_t)ptr[0]) |=0A= + ((uint32_t)ptr[1] << 8) |=0A= + ((uint32_t)ptr[2] << 16) |=0A= + ((uint32_t)ptr[3] << 24);=0A= +#else=0A= + return ((uint32_t)ptr[3]) |=0A= + ((uint32_t)ptr[2] << 8) |=0A= + ((uint32_t)ptr[1] << 16) |=0A= + ((uint32_t)ptr[0] << 24);=0A= +#endif=0A= +}=0A= +=0A= +uint32_t load_le_32_add(const uint8_t *ptr)=0A= +{=0A= +#if __BYTE_ORDER__ =3D=3D __ORDER_LITTLE_ENDIAN__=0A= + return ((uint32_t)ptr[0]) +=0A= + ((uint32_t)ptr[1] << 8) +=0A= + ((uint32_t)ptr[2] << 16) +=0A= + ((uint32_t)ptr[3] << 24);=0A= +#else=0A= + return ((uint32_t)ptr[3]) +=0A= + ((uint32_t)ptr[2] << 8) +=0A= + ((uint32_t)ptr[1] << 16) +=0A= + ((uint32_t)ptr[0] << 24);=0A= +#endif=0A= +}=0A= +=0A= +uint32_t load_le_32_xor(const uint8_t *ptr)=0A= +{=0A= +#if __BYTE_ORDER__ =3D=3D __ORDER_LITTLE_ENDIAN__=0A= + return ((uint32_t)ptr[0]) ^=0A= + ((uint32_t)ptr[1] << 8) ^=0A= + ((uint32_t)ptr[2] << 16) ^=0A= + ((uint32_t)ptr[3] << 24);=0A= +#else=0A= + return ((uint32_t)ptr[0]) ^=0A= + ((uint32_t)ptr[1] << 8) ^=0A= + ((uint32_t)ptr[2] << 16) ^=0A= + ((uint32_t)ptr[3] << 24);=0A= +#endif=0A= +}=0A= +=0A= +/* { dg-final { scan-tree-dump-times "32 bit load in target endianness = found" 3 "bswap" } } */=0A= +=0A= diff --git a/gcc/testsuite/gcc.dg/tree-ssa/pr98953.c = b/gcc/testsuite/gcc.dg/tree-ssa/pr98953.c=0A= new file mode 100644=0A= index 0000000..7687dc2=0A= --- /dev/null=0A= +++ b/gcc/testsuite/gcc.dg/tree-ssa/pr98953.c=0A= @@ -0,0 +1,14 @@=0A= +/* { dg-do compile } */=0A= +/* { dg-options "-O2 -fdump-tree-bswap-details" } */=0A= +=0A= +int foo(unsigned char *ptr)=0A= +{=0A= +#if __BYTE_ORDER__ =3D=3D __ORDER_LITTLE_ENDIAN__=0A= + return ptr[0] + (ptr[1] << 8);=0A= +#else=0A= + return ptr[1] + (ptr[0] << 8);=0A= +#endif=0A= +}=0A= +=0A= +/* { dg-final { scan-tree-dump "16 bit load in target endianness found" = "bswap" } } */=0A= +=0A= ------=_NextPart_000_0017_01D7DF7C.7F739920--