From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from EUR04-HE1-obe.outbound.protection.outlook.com (mail-oln040092073087.outbound.protection.outlook.com [40.92.73.87]) by sourceware.org (Postfix) with ESMTPS id 8E03E3857C44 for ; Fri, 21 May 2021 05:42:31 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.3.2 sourceware.org 8E03E3857C44 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=hotmail.de Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=bernd.edlinger@hotmail.de ARC-Seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=ELg0QTkmC2ynaUqMQ91zC4eOVg7+REQYSEb8KCZNB9l944nl5bZhvfyHP4oPrY0ty7VxRaTxNDqi2FxRcvFbq1Z5DUVrVzIEcTtI9hWRvFcabk7XVR9oNguqwJ8WNPG81PiIrKbPpPeOTCvvYa410NLIR/mookSR1EuIfeNriVjRSIRcF8SzJB+xbnz5MbyxMHsKC8OY1697/CBSIHgPEDc/xbxTwhBVJM0go2B/G6IWE4eOtmIiqKsFoHeLPnmFuXMZ+3jrvWujtLiuimhZSUJH1fxtGtK7O/ouYHR6J3BqHYwXLa4Ma6Q67p+75NIJlX2y4y79LuyugTzjYGO2sw== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=Y+sX4h/KVUJBMCi0+E2QIn4G/PUUP+qxwYpTZOsET0s=; b=mCHFQT6SGIBe+YPwx0RYW3IiI3JgWJgZgLtQrzim8aWzF+l2oK1oZpEZbgWmxkVuMLb/LOuU9yHaUv37zUK8M3lQbNDJuNHAsHyB5ag0ZDa8kAMZbrTDiKEn7X+dnKRspLliwuwDo8Z0nSKV8uBif1I6s+h6wF9rOkSNh5TPaqjF9gFXeOyVIBzMoY/j/NcH4/OnZhcugWBfn9RvGYih46douAz7Rr4ajZJWZXQqt/SNuXc7IuGWhIuhLj/yeNvbNWYsRQiupaV2U9mBskCE8A3ag8LR9kjcFvg7+OCiNcrFzp0ZjnbEHyST/SWaVR4iHK3IU6ON8T0IHh4SLpvebQ== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=none; dmarc=none; dkim=none; arc=none Received: from DB3EUR04FT031.eop-eur04.prod.protection.outlook.com (2a01:111:e400:7e0c::40) by DB3EUR04HT225.eop-eur04.prod.protection.outlook.com (2a01:111:e400:7e0c::209) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.4129.25; Fri, 21 May 2021 05:42:29 +0000 Received: from AM8PR10MB4708.EURPRD10.PROD.OUTLOOK.COM (2a01:111:e400:7e0c::43) by DB3EUR04FT031.mail.protection.outlook.com (2a01:111:e400:7e0c::239) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.4129.25 via Frontend Transport; Fri, 21 May 2021 05:42:29 +0000 X-IncomingTopHeaderMarker: OriginalChecksum:6EA7E3E908F2B0C047D2F6A3CC933EAEADD51E169F1F8A948583621A33667F5F; UpperCasedChecksum:583FC10A296EBF24A9F23AD82F0EB5D96908FFFD27C8795E8D93DFE3BDD2F56E; SizeAsReceived:9047; Count:48 Received: from AM8PR10MB4708.EURPRD10.PROD.OUTLOOK.COM ([fe80::e41b:107f:af82:150a]) by AM8PR10MB4708.EURPRD10.PROD.OUTLOOK.COM ([fe80::e41b:107f:af82:150a%7]) with mapi id 15.20.4129.035; Fri, 21 May 2021 05:42:29 +0000 Subject: Re: [PATCH] Add 3 target hooks for memset To: "H.J. Lu" , Richard Biener Cc: GCC Patches , Richard Sandiford , Uros Bizjak References: <20210518191646.3071005-1-hjl.tools@gmail.com> <20210518191646.3071005-2-hjl.tools@gmail.com> From: Bernd Edlinger Message-ID: Date: Fri, 21 May 2021 07:42:27 +0200 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.6.1 In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit X-TMN: [6facd7zc4cZSpVFK1AfITozkNUiu/S+W] X-ClientProxiedBy: FR2P281CA0032.DEUP281.PROD.OUTLOOK.COM (2603:10a6:d10:14::19) To AM8PR10MB4708.EURPRD10.PROD.OUTLOOK.COM (2603:10a6:20b:364::23) X-Microsoft-Original-Message-ID: <18c41513-eaa9-cf31-37f7-6dccdac5e755@hotmail.de> MIME-Version: 1.0 X-MS-Exchange-MessageSentRepresentingType: 1 Received: from [192.168.1.101] (84.57.61.94) by FR2P281CA0032.DEUP281.PROD.OUTLOOK.COM (2603:10a6:d10:14::19) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.4173.11 via Frontend Transport; Fri, 21 May 2021 05:42:28 +0000 X-MS-PublicTrafficType: Email X-IncomingHeaderCount: 48 X-EOPAttributedMessage: 0 X-MS-Office365-Filtering-Correlation-Id: 6c1e78bf-3d78-4d50-dff3-08d91c1b38b8 X-MS-TrafficTypeDiagnostic: DB3EUR04HT225: X-Microsoft-Antispam: BCL:0; X-Microsoft-Antispam-Message-Info: XWU8zAhAw/cQMQUUPvmlVQDjbj17m5YqLQ6N7dzu/DzZWBqdomsL79bZb1pvtX9JCJRhJX9GvT8EB22Ah2LbrTv+0WxYkM1wbG9U1IngVck39AtALD1KHg4knbb0djLTrQ3tvq0n5i94LQkEW5dapT9mBPLtc3UenbF2b7FrGInqvyQ5lhI0VGdWzUkBajo/AjrlEzbRapW3TiRgROtufRwv/VqaMRWt8fP9jQDJSBPehoFQKppUgtsfezPS9FcFcohudSE/7i3lZ40Bja5cORBpSOJ1cQcdG6ms+C5rGLnN+xnVsmeJf3LIoRmVecl5cub2H8t5DHQ08EK/shoDBhj7rPKNWVrXuHSeJtOMvecca+b5sl7e1VcH8yc2tHFs5Fd0ecX4Zs1HPUo3SnJbOA== X-MS-Exchange-AntiSpam-MessageData: Nm9UwYkxr83p6wALqAM53ItX9XZpO7aJn0HMQ4I55i2B5Vag8IkeRSfauympr0Nt76Ic1cWpVKDgfCft9kvjGOVZlaI7px3MA6bhQlDDoZ5DXMOMp/hjbD8x9cGGdpTNYTnewGuchrv4odVyWjTDHA== X-OriginatorOrg: outlook.com X-MS-Exchange-CrossTenant-Network-Message-Id: 6c1e78bf-3d78-4d50-dff3-08d91c1b38b8 X-MS-Exchange-CrossTenant-OriginalArrivalTime: 21 May 2021 05:42:29.4325 (UTC) X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted X-MS-Exchange-CrossTenant-Id: 84df9e7f-e9f6-40af-b435-aaaaaaaaaaaa X-MS-Exchange-CrossTenant-AuthSource: DB3EUR04FT031.eop-eur04.prod.protection.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Anonymous X-MS-Exchange-CrossTenant-FromEntityHeader: Internet X-MS-Exchange-CrossTenant-RMS-PersistedConsumerOrg: 00000000-0000-0000-0000-000000000000 X-MS-Exchange-Transport-CrossTenantHeadersStamped: DB3EUR04HT225 X-Spam-Status: No, score=-8.9 required=5.0 tests=BAYES_00, FORGED_MUA_MOZILLA, FREEMAIL_FROM, GIT_PATCH_0, KAM_DMARC_STATUS, MSGID_FROM_MTA_HEADER, NICE_REPLY_A, RCVD_IN_DNSWL_NONE, RCVD_IN_MSPIKE_H2, SPF_HELO_PASS, SPF_PASS, TXREP autolearn=ham autolearn_force=no version=3.4.2 X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) 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: Fri, 21 May 2021 05:42:34 -0000 On 5/20/21 10:49 PM, H.J. Lu wrote: > On Wed, May 19, 2021 at 5:55 AM H.J. Lu wrote: >> >> On Wed, May 19, 2021 at 2:25 AM Richard Biener >> wrote: >>> >>> On Tue, May 18, 2021 at 9:16 PM H.J. Lu wrote: >>>> >>>> Add TARGET_READ_MEMSET_VALUE and TARGET_GEN_MEMSET_VALUE to support >>>> target instructions to duplicate QImode value to TImode/OImode/XImode >>>> value for memmset. >>>> >>>> PR middle-end/90773 >>>> * builtins.c (builtin_memset_read_str): Call >>>> targetm.read_memset_value. >>>> (builtin_memset_gen_str): Call targetm.gen_memset_value. >>>> * target.def (read_memset_value): New hook. >>>> (gen_memset_value): Likewise. >>>> * targhooks.c: Inclue "builtins.h". >>>> (default_read_memset_value): New function. >>>> (default_gen_memset_value): Likewise. >>>> * targhooks.h (default_read_memset_value): New prototype. >>>> (default_gen_memset_value): Likewise. >>>> * doc/tm.texi.in: Add TARGET_READ_MEMSET_VALUE and >>>> TARGET_GEN_MEMSET_VALUE hooks. >>>> * doc/tm.texi: Regenerated. >>>> --- >>>> gcc/builtins.c | 47 ++++---------------------------------- >>>> gcc/doc/tm.texi | 16 +++++++++++++ >>>> gcc/doc/tm.texi.in | 4 ++++ >>>> gcc/target.def | 20 +++++++++++++++++ >>>> gcc/targhooks.c | 56 ++++++++++++++++++++++++++++++++++++++++++++++ >>>> gcc/targhooks.h | 4 ++++ >>>> 6 files changed, 104 insertions(+), 43 deletions(-) >>>> >>>> diff --git a/gcc/builtins.c b/gcc/builtins.c >>>> index e1b284846b1..f78a36478ef 100644 >>>> --- a/gcc/builtins.c >>>> +++ b/gcc/builtins.c >>>> @@ -6584,24 +6584,11 @@ expand_builtin_strncpy (tree exp, rtx target) >>>> previous iteration. */ >>>> >>>> rtx >>>> -builtin_memset_read_str (void *data, void *prevp, >>>> +builtin_memset_read_str (void *data, void *prev, >>>> HOST_WIDE_INT offset ATTRIBUTE_UNUSED, >>>> scalar_int_mode mode) >>>> { >>>> - by_pieces_prev *prev = (by_pieces_prev *) prevp; >>>> - if (prev != nullptr && prev->data != nullptr) >>>> - { >>>> - /* Use the previous data in the same mode. */ >>>> - if (prev->mode == mode) >>>> - return prev->data; >>>> - } >>>> - >>>> - const char *c = (const char *) data; >>>> - char *p = XALLOCAVEC (char, GET_MODE_SIZE (mode)); >>>> - >>>> - memset (p, *c, GET_MODE_SIZE (mode)); >>>> - >>>> - return c_readstr (p, mode); >>>> + return targetm.read_memset_value ((const char *) data, prev, mode); >>>> } >>>> >>>> /* Callback routine for store_by_pieces. Return the RTL of a register >>>> @@ -6611,37 +6598,11 @@ builtin_memset_read_str (void *data, void *prevp, >>>> nullptr, it has the RTL info from the previous iteration. */ >>>> >>>> static rtx >>>> -builtin_memset_gen_str (void *data, void *prevp, >>>> +builtin_memset_gen_str (void *data, void *prev, >>>> HOST_WIDE_INT offset ATTRIBUTE_UNUSED, >>>> scalar_int_mode mode) >>>> { >>>> - rtx target, coeff; >>>> - size_t size; >>>> - char *p; >>>> - >>>> - by_pieces_prev *prev = (by_pieces_prev *) prevp; >>>> - if (prev != nullptr && prev->data != nullptr) >>>> - { >>>> - /* Use the previous data in the same mode. */ >>>> - if (prev->mode == mode) >>>> - return prev->data; >>>> - >>>> - target = simplify_gen_subreg (mode, prev->data, prev->mode, 0); >>>> - if (target != nullptr) >>>> - return target; >>>> - } >>>> - >>>> - size = GET_MODE_SIZE (mode); >>>> - if (size == 1) >>>> - return (rtx) data; >>>> - >>>> - p = XALLOCAVEC (char, size); >>>> - memset (p, 1, size); >>>> - coeff = c_readstr (p, mode); >>>> - >>>> - target = convert_to_mode (mode, (rtx) data, 1); >>>> - target = expand_mult (mode, target, coeff, NULL_RTX, 1); >>>> - return force_reg (mode, target); >>>> + return targetm.gen_memset_value ((rtx) data, prev, mode); >>>> } >>>> >>>> /* Expand expression EXP, which is a call to the memset builtin. Return >>>> diff --git a/gcc/doc/tm.texi b/gcc/doc/tm.texi >>>> index 85ea9395560..51385044e76 100644 >>>> --- a/gcc/doc/tm.texi >>>> +++ b/gcc/doc/tm.texi >>>> @@ -11868,6 +11868,22 @@ This function prepares to emit a conditional comparison within a sequence >>>> @var{bit_code} is @code{AND} or @code{IOR}, which is the op on the compares. >>>> @end deftypefn >>>> >>>> +@deftypefn {Target Hook} rtx TARGET_READ_MEMSET_VALUE (const char *@var{c}, void *@var{prev}, scalar_int_mode @var{mode}) >>>> +This function returns the RTL of a constant integer corresponding to >>>> +target reading @code{GET_MODE_SIZE (@var{mode})} bytes from the stringn >>>> +constant @var{str}. If @var{prev} is not @samp{nullptr}, it contains >>>> +the RTL information from the previous interation. >>>> +@end deftypefn >>>> + >>>> +@deftypefn {Target Hook} rtx TARGET_GEN_MEMSET_VALUE (rtx @var{data}, void *@var{prev}, scalar_int_mode @var{mode}) >>>> +This function returns the RTL of a register containing >>>> +@code{GET_MODE_SIZE (@var{mode})} consecutive copies of the unsigned >>>> +char value given in the RTL register @var{data}. For example, if >>>> +@var{mode} is 4 bytes wide, return the RTL for 0x01010101*@var{data}. >>>> +If @var{PREV} is not @samp{nullptr}, it is the RTL information from >>>> +the previous iteration. >>>> +@end deftypefn >>>> + >>>> @deftypefn {Target Hook} unsigned TARGET_LOOP_UNROLL_ADJUST (unsigned @var{nunroll}, class loop *@var{loop}) >>>> This target hook returns a new value for the number of times @var{loop} >>>> should be unrolled. The parameter @var{nunroll} is the number of times >>>> diff --git a/gcc/doc/tm.texi.in b/gcc/doc/tm.texi.in >>>> index d8e3de14af1..8d4c3949fbf 100644 >>>> --- a/gcc/doc/tm.texi.in >>>> +++ b/gcc/doc/tm.texi.in >>>> @@ -7956,6 +7956,10 @@ lists. >>>> >>>> @hook TARGET_GEN_CCMP_NEXT >>>> >>>> +@hook TARGET_READ_MEMSET_VALUE >>>> + >>>> +@hook TARGET_GEN_MEMSET_VALUE >>>> + >>>> @hook TARGET_LOOP_UNROLL_ADJUST >>>> >>>> @defmac POWI_MAX_MULTS >>>> diff --git a/gcc/target.def b/gcc/target.def >>>> index bbaf6b4f3a0..c9aca40fa88 100644 >>>> --- a/gcc/target.def >>>> +++ b/gcc/target.def >>>> @@ -2694,6 +2694,26 @@ DEFHOOK >>>> rtx, (rtx_insn **prep_seq, rtx_insn **gen_seq, rtx prev, int cmp_code, tree op0, tree op1, int bit_code), >>>> NULL) >>>> >>>> +DEFHOOK >>>> +(read_memset_value, >>>> + "This function returns the RTL of a constant integer corresponding to\n\ >>>> +target reading @code{GET_MODE_SIZE (@var{mode})} bytes from the stringn\n\ >>>> +constant @var{str}. If @var{prev} is not @samp{nullptr}, it contains\n\ >>> >>> where is 'str' defined? I can't really tell what's the difference >> >> Fixed with >> >> diff --git a/gcc/target.def b/gcc/target.def >> index c9aca40fa88..4c3a5fcc634 100644 >> --- a/gcc/target.def >> +++ b/gcc/target.def >> @@ -2699,8 +2699,8 @@ DEFHOOK >> "This function returns the RTL of a constant integer corresponding to\n\ >> target reading @code{GET_MODE_SIZE (@var{mode})} bytes from the string\n\ >> constant @var{str}. If @var{prev} is not @samp{nullptr}, it contains\n\ >> -the RTL information from the previous interation.", >> - rtx, (const char *c, void *prev, scalar_int_mode mode), >> +the RTL information from the previous iteration.", >> + rtx, (const char *str, void *prev, scalar_int_mode mode), >> default_read_memset_value) >> >> DEFHOOK >> >>> from read_memset_value >>> and gen_memset_value. >> >> The difference is that input of read_memset_value is a string constant >> like "123" and input of gen_memset_value is an RTL register. >> >>> Somehow I feel that an optab for the "splat" operation similar >>> to vec_duplicate might be a better way to expose this - of course >>> that doesn't handle the "prev" thing. >> >> The x86 backend has ix86_expand_vector_init_duplicate () to >> broadcast QImode to TImode/OImode/XImode: >> >> /* A subroutine of ix86_expand_vector_init. Store into TARGET a vector >> with all elements equal to VAR. Return true if successful. */ >> >> bool >> ix86_expand_vector_init_duplicate (bool mmx_ok, machine_mode mode, >> rtx target, rtx val) >> >>> So how's this the right point of abstraction to the target? >> >> I can add 2 target hooks, one for scratch register and one for >> broadcasting QImode to TImode/OImode/XImode. Then I can >> move x86 codes to the middle-end. >> > > Here is the patch to add 3 target hooks: > > -- Target Hook: rtx TARGET_READ_MEMSET_VALUE (const char *C, > scalar_int_mode MODE) > This function returns the RTL of a constant integer corresponding > to target reading 'GET_MODE_SIZE (MODE)' bytes from the string > constant C. > > -- Target Hook: rtx TARGET_GEN_MEMSET_VALUE (rtx DATA, scalar_int_mode > MODE) > This function returns the RTL of a register containing > 'GET_MODE_SIZE (MODE)' consecutive copies of the unsigned char > value given in the RTL register DATA. For example, if MODE is 4 > bytes wide, return the RTL for 0x01010101*DATA. > > -- Target Hook: rtx TARGET_GEN_MEMSET_VALUE_FROM_PREV (void *PREV, > scalar_int_mode MODE) > This function returns the RTL of a register in MODE generated from > PREV in the previous iteration. > > with > > /* Return the RTL of a register in MODE generated from PREV in the > previous iteration. */ > > static rtx > gen_memset_value_from_prev (void *prevp, scalar_int_mode mode) > { > by_pieces_prev *prev = (by_pieces_prev *) prevp; > rtx value; > if (prev != nullptr && prev->data != nullptr) > { > /* Use the previous data in the same mode. */ > if (prev->mode == mode) > return prev->data; > > value = targetm.gen_memset_value_from_prev (prevp, mode); > } > else > value = nullptr; > return value; > } > > /* Callback routine for store_by_pieces. Read GET_MODE_BITSIZE (MODE) > bytes from constant string DATA + OFFSET and return it as target > constant. If PREV isn't nullptr, it has the RTL info from the > previous iteration. */ > > rtx > builtin_memset_read_str (void *data, void *prev, > HOST_WIDE_INT offset ATTRIBUTE_UNUSED, > scalar_int_mode mode) > { > const char *str = (const char *) data; > > /* Don't use the previous value if size is 1. */ > if (GET_MODE_SIZE (mode) == 1) > return default_read_memset_value (str, mode); > > rtx value = gen_memset_value_from_prev (prev, mode); > if (value) > return value; > > return targetm.read_memset_value (str, mode); > } > > /* Callback routine for store_by_pieces. Return the RTL of a register > containing GET_MODE_SIZE (MODE) consecutive copies of the unsigned > char value given in the RTL register data. For example, if mode is > 4 bytes wide, return the RTL for 0x01010101*data. If PREV isn't > nullptr, it has the RTL info from the previous iteration. */ > > static rtx > builtin_memset_gen_str (void *datap, void *prev, > HOST_WIDE_INT offset ATTRIBUTE_UNUSED, > scalar_int_mode mode) > { > rtx data = (rtx) datap; > > /* Don't use the previous value if size is 1. */ > if (GET_MODE_SIZE (mode) == 1) > return data; > > rtx value = gen_memset_value_from_prev (prev, mode); > if (value) > return value; > > return targetm.gen_memset_value (data, mode); > }> +/* Default implementation of TARGET_GEN_MEMSET_VALUE. */ > + > +rtx > +default_gen_memset_value (rtx data, scalar_int_mode mode) > +{ > + rtx target, coeff; > + size_t size; > + char *p; > + > + size = GET_MODE_SIZE (mode); > + if (size == 1) > + return data; > + > + p = XALLOCAVEC (char, size); > + memset (p, 1, size); > + coeff = c_readstr (p, mode); > + > + target = convert_to_mode (mode, data, 1); > + target = expand_mult (mode, target, coeff, NULL_RTX, 1); Note this formula does not work for data = -1 for instance, since 0x01010101U * -1 = 0xFEFEFEFFU but memset(str, n, -1) set it to 0xFFFFFFFFU, right? So are we sure that the value of "data" is always in the range [0..255] ? Bernd.