{"id":2225513,"url":"http://patchwork.ozlabs.org/api/patches/2225513/?format=json","web_url":"http://patchwork.ozlabs.org/project/gcc/patch/aecVxk86j5b4Wh-L@tucnak/","project":{"id":17,"url":"http://patchwork.ozlabs.org/api/projects/17/?format=json","name":"GNU Compiler Collection","link_name":"gcc","list_id":"gcc-patches.gcc.gnu.org","list_email":"gcc-patches@gcc.gnu.org","web_url":null,"scm_url":null,"webscm_url":null,"list_archive_url":"","list_archive_url_format":"","commit_url_format":""},"msgid":"<aecVxk86j5b4Wh-L@tucnak>","list_archive_url":null,"date":"2026-04-21T06:14:30","name":"bitintlower: Padding bit fixes, part 4 [PR123635]","commit_ref":null,"pull_url":null,"state":"new","archived":false,"hash":"91e65baf2527268532df908ca8e9047abb49d51b","submitter":{"id":671,"url":"http://patchwork.ozlabs.org/api/people/671/?format=json","name":"Jakub Jelinek","email":"jakub@redhat.com"},"delegate":null,"mbox":"http://patchwork.ozlabs.org/project/gcc/patch/aecVxk86j5b4Wh-L@tucnak/mbox/","series":[{"id":500738,"url":"http://patchwork.ozlabs.org/api/series/500738/?format=json","web_url":"http://patchwork.ozlabs.org/project/gcc/list/?series=500738","date":"2026-04-21T06:14:30","name":"bitintlower: Padding bit fixes, part 4 [PR123635]","version":1,"mbox":"http://patchwork.ozlabs.org/series/500738/mbox/"}],"comments":"http://patchwork.ozlabs.org/api/patches/2225513/comments/","check":"pending","checks":"http://patchwork.ozlabs.org/api/patches/2225513/checks/","tags":{},"related":[],"headers":{"Return-Path":"<gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org>","X-Original-To":["incoming@patchwork.ozlabs.org","gcc-patches@gcc.gnu.org"],"Delivered-To":["patchwork-incoming@legolas.ozlabs.org","gcc-patches@gcc.gnu.org"],"Authentication-Results":["legolas.ozlabs.org;\n\tdkim=pass (1024-bit key;\n unprotected) header.d=redhat.com header.i=@redhat.com header.a=rsa-sha256\n header.s=mimecast20190719 header.b=VIDO+MuT;\n\tdkim-atps=neutral","legolas.ozlabs.org;\n spf=pass (sender SPF authorized) smtp.mailfrom=gcc.gnu.org\n (client-ip=2620:52:6:3111::32; helo=vm01.sourceware.org;\n envelope-from=gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org;\n receiver=patchwork.ozlabs.org)","sourceware.org;\n\tdkim=pass (1024-bit key,\n unprotected) header.d=redhat.com header.i=@redhat.com header.a=rsa-sha256\n header.s=mimecast20190719 header.b=VIDO+MuT","sourceware.org; dmarc=pass (p=quarantine dis=none)\n header.from=redhat.com","sourceware.org; spf=pass smtp.mailfrom=redhat.com","server2.sourceware.org;\n arc=none smtp.remote-ip=170.10.129.124"],"Received":["from vm01.sourceware.org (vm01.sourceware.org\n [IPv6:2620:52:6:3111::32])\n\t(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n\t key-exchange x25519 server-signature ECDSA (secp384r1) server-digest SHA384)\n\t(No client certificate requested)\n\tby legolas.ozlabs.org (Postfix) with ESMTPS id 4g0Bsq0Dsrz1yCv\n\tfor <incoming@patchwork.ozlabs.org>; Tue, 21 Apr 2026 16:15:09 +1000 (AEST)","from vm01.sourceware.org (localhost [127.0.0.1])\n\tby sourceware.org (Postfix) with ESMTP id A810A4BA23F1\n\tfor <incoming@patchwork.ozlabs.org>; Tue, 21 Apr 2026 06:15:07 +0000 (GMT)","from us-smtp-delivery-124.mimecast.com\n (us-smtp-delivery-124.mimecast.com [170.10.129.124])\n by sourceware.org (Postfix) with ESMTP id 4CBC74BA23CB\n for <gcc-patches@gcc.gnu.org>; Tue, 21 Apr 2026 06:14:39 +0000 (GMT)","from mx-prod-mc-05.mail-002.prod.us-west-2.aws.redhat.com\n (ec2-54-186-198-63.us-west-2.compute.amazonaws.com [54.186.198.63]) by\n relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3,\n cipher=TLS_AES_256_GCM_SHA384) id us-mta-687-dMmurl0yM0yZV7-bzKs04g-1; Tue,\n 21 Apr 2026 02:14:35 -0400","from mx-prod-int-08.mail-002.prod.us-west-2.aws.redhat.com\n (mx-prod-int-08.mail-002.prod.us-west-2.aws.redhat.com [10.30.177.111])\n (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest\n SHA256)\n (No client certificate requested)\n by mx-prod-mc-05.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTPS\n id 2663D1956088; Tue, 21 Apr 2026 06:14:34 +0000 (UTC)","from tucnak.zalov.cz (unknown [10.44.48.12])\n by mx-prod-int-08.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with\n ESMTPS\n id 7A0F51800446; Tue, 21 Apr 2026 06:14:33 +0000 (UTC)","from tucnak.zalov.cz (localhost [127.0.0.1])\n by tucnak.zalov.cz (8.18.1/8.18.1) with ESMTPS id 63L6EU4P886190\n (version=TLSv1.3 cipher=TLS_AES_256_GCM_SHA384 bits=256 verify=NOT);\n Tue, 21 Apr 2026 08:14:30 +0200","(from jakub@localhost)\n by tucnak.zalov.cz (8.18.1/8.18.1/Submit) id 63L6EUWl886189;\n Tue, 21 Apr 2026 08:14:30 +0200"],"DKIM-Filter":["OpenDKIM Filter v2.11.0 sourceware.org A810A4BA23F1","OpenDKIM Filter v2.11.0 sourceware.org 4CBC74BA23CB"],"DMARC-Filter":"OpenDMARC Filter v1.4.2 sourceware.org 4CBC74BA23CB","ARC-Filter":"OpenARC Filter v1.0.0 sourceware.org 4CBC74BA23CB","ARC-Seal":"i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1776752079; cv=none;\n b=Z1JLocb9R9V/Ix8OypdKDXjjVcz1EsBBKQzWGgd5yZAd78K6Isos0NhmL1mYmpcu+pLU6+Ux1hJlXePrZOPq9DjE0CVbWzdmcmUsgJ7oJC7CE1nF5ODDEAfAB5V/5AvogmFLD+Z905FvRllkJq/22v2bjJdyOr2eD47yj++k/7Q=","ARC-Message-Signature":"i=1; a=rsa-sha256; d=sourceware.org; s=key;\n t=1776752079; c=relaxed/simple;\n bh=wUwm5tAm+Q2RPGUJ9OnJI1X+0+d+ShFmBJkyLCdW+Ok=;\n h=DKIM-Signature:Date:From:To:Subject:Message-ID:MIME-Version;\n b=aDPGNgxbKU12qKhXPvMlpMTqShL8OEwodcpZ4+4PYCzKBh85tuDF3secCcGV9YLvl2IdAKnK5FUEt5eKM9b2eXOkofSnnFnrXoSI1KUFzi9yHDtznb4GYDt/KKfwQ+l2eYRpJhQKr/sWcj0gfp1DZVYtI+3UtLyeAaEt8GcTRok=","ARC-Authentication-Results":"i=1; server2.sourceware.org","DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com;\n s=mimecast20190719; t=1776752079;\n h=from:from:reply-to:reply-to:subject:subject:date:date:\n message-id:message-id:to:to:cc:cc:mime-version:mime-version:\n content-type:content-type; bh=fm+KtijdAPpmGaf1wA8b83AmeRwTtajYexvewacIhe4=;\n b=VIDO+MuT/OGGXjDlPcu8tb0PGNRdVPSWBp/ahQ21TUeRJzH8fgGSGRQqs6CJXZ7xU7rCbJ\n pFvGr5OIkmojqTRoiB8lNAx28+bN3CLlu+AvqbmHw1qCj4ZxbBuLXg2r7fx0cMKhnvQL3e\n 0wWKsHH87ellvm4QfeFRsldo2GwWffM=","X-MC-Unique":"dMmurl0yM0yZV7-bzKs04g-1","X-Mimecast-MFC-AGG-ID":"dMmurl0yM0yZV7-bzKs04g_1776752074","Date":"Tue, 21 Apr 2026 08:14:30 +0200","From":"Jakub Jelinek <jakub@redhat.com>","To":"Richard Biener <rguenther@suse.de>,\n Andrew Pinski <andrew.pinski@oss.qualcomm.com>","Cc":"gcc-patches@gcc.gnu.org","Subject":"[PATCH] bitintlower: Padding bit fixes, part 4 [PR123635]","Message-ID":"<aecVxk86j5b4Wh-L@tucnak>","MIME-Version":"1.0","X-Scanned-By":"MIMEDefang 3.4.1 on 10.30.177.111","X-Mimecast-Spam-Score":"0","X-Mimecast-MFC-PROC-ID":"uS99IcCynX9KJudNlCmONn2gTAzdBKMEk2eG892xkno_1776752074","X-Mimecast-Originator":"redhat.com","Content-Type":"text/plain; charset=us-ascii","Content-Disposition":"inline","X-BeenThere":"gcc-patches@gcc.gnu.org","X-Mailman-Version":"2.1.30","Precedence":"list","List-Id":"Gcc-patches mailing list <gcc-patches.gcc.gnu.org>","List-Unsubscribe":"<https://gcc.gnu.org/mailman/options/gcc-patches>,\n <mailto:gcc-patches-request@gcc.gnu.org?subject=unsubscribe>","List-Archive":"<https://gcc.gnu.org/pipermail/gcc-patches/>","List-Post":"<mailto:gcc-patches@gcc.gnu.org>","List-Help":"<mailto:gcc-patches-request@gcc.gnu.org?subject=help>","List-Subscribe":"<https://gcc.gnu.org/mailman/listinfo/gcc-patches>,\n <mailto:gcc-patches-request@gcc.gnu.org?subject=subscribe>","Reply-To":"Jakub Jelinek <jakub@redhat.com>","Errors-To":"gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org"},"content":"Hi!\n\nAs the following testcase shows, not clearing the padding bits after\nsigned MULT_EXPR (or signed division) is reasonable when overflow actually is\nundefined behavior because then anything can happen.  But when it is not\nundefined behavior due to -fwrapv, we need to clear the padding bits\non targets which chose that behavior.  It isn't only signed MULT_EXPR,\nbut also division because smallest negative / -1 overflows and in that\ncase the padding bits aren't correct for bitint_extended targets either.\n\nBootstrapped/regtested on x86_64-linux and i686-linux and tested on\nmake -j8 -k check-gcc GCC_TEST_RUN_EXPENSIVE=1 RUNTESTFLAGS=\"GCC_TEST_RUN_EXPENSIVE=1 dg.exp='*bitint* pr112673.c builtin-stdc-bit-*.c pr112566-2.c pr112511.c pr116588.c pr116003.c pr113693.c pr113602.c flex-array-counted-by-7.c' dg-torture.exp='*bitint* pr116480-2.c pr114312.c pr114121.c' dfp.exp=*bitint* vect.exp='vect-early-break_99-pr113287.c' tree-ssa.exp=pr113735.c\"\non riscv64-linux, ok for trunk?\n\n2026-04-21  Jakub Jelinek  <jakub@redhat.com>\n\n\tPR middle-end/123635\n\t* gimple-lower-bitint.cc (bitint_large_huge::lower_muldiv_stmt):\n\tExtend the padding bits not just for unsigned MULT_EXPR but for any\n\tTYPE_OVERFLOW_WRAPS MULT_EXPR and signed TYPE_OVERFLOW_WRAPS division.\n\n\t* gcc.dg/torture/bitint-90.c: New test.\n\n\n\tJakub","diff":"--- gcc/gimple-lower-bitint.cc.jj\t2026-04-20 09:11:19.132955449 +0200\n+++ gcc/gimple-lower-bitint.cc\t2026-04-20 18:29:23.419243836 +0200\n@@ -4085,7 +4085,8 @@ bitint_large_huge::lower_muldiv_stmt (tr\n   gcc_assert (TREE_CODE (type) == BITINT_TYPE\n \t      && bitint_precision_kind (type) >= bitint_prec_large);\n   int prec = TYPE_PRECISION (type), prec1, prec2;\n-  bool zero_ms_limb = false;\n+  bool ext_ms_limb = false;\n+  bool do_ext = false;\n   rhs1 = handle_operand_addr (rhs1, stmt, NULL, &prec1);\n   rhs2 = handle_operand_addr (rhs2, stmt, NULL, &prec2);\n   if (obj == NULL_TREE)\n@@ -4101,15 +4102,26 @@ bitint_large_huge::lower_muldiv_stmt (tr\n       lhs = force_gimple_operand_gsi (&m_gsi, lhs, true,\n \t\t\t\t      NULL_TREE, true, GSI_SAME_STMT);\n     }\n+  if (bitint_extended && TYPE_OVERFLOW_WRAPS (type))\n+    {\n+      if (rhs_code == MULT_EXPR)\n+\tdo_ext = true;\n+      /* For signed division with -fwrapv, minimum negative / -1 needs\n+\t is minimum negative and the padding bits above it should be all\n+\t set.  */\n+      else if (!TYPE_UNSIGNED (type)\n+\t       && (rhs_code == TRUNC_DIV_EXPR || rhs_code == EXACT_DIV_EXPR))\n+\tdo_ext = true;\n+    }\n   if (bitint_extended == bitint_ext_full\n       && abi_limb_prec > limb_prec\n       && (CEIL (prec, abi_limb_prec) * abi_limb_prec\n \t  > CEIL (prec, limb_prec) * limb_prec))\n     {\n       /* unsigned multiplication needs to wrap around, so we can't\n-\t increase prec.  */\n-      if (rhs_code == MULT_EXPR && TYPE_UNSIGNED (type))\n-\tzero_ms_limb = true;\n+\t increase prec.  Similarly for -fwrapv.  */\n+      if (do_ext)\n+\text_ms_limb = true;\n       else\n \tprec = CEIL (prec, abi_limb_prec) * abi_limb_prec;\n     }\n@@ -4166,10 +4178,8 @@ bitint_large_huge::lower_muldiv_stmt (tr\n \t  add_eh_edge (e2->src, e1);\n \t}\n     }\n-  if (bitint_extended\n-      && rhs_code == MULT_EXPR\n-      && TYPE_UNSIGNED (type)\n-      && (prec % limb_prec) != 0)\n+  if (do_ext\n+      && ((prec % limb_prec) != 0 || (ext_ms_limb && !TYPE_UNSIGNED (type))))\n     {\n       /* Unsigned multiplication wraps, but libgcc function will return the\n \t bits beyond prec within the top limb as another limb of the full\n@@ -4180,13 +4190,33 @@ bitint_large_huge::lower_muldiv_stmt (tr\n       tree v = make_ssa_name (m_limb_type);\n       g = gimple_build_assign (v, l);\n       insert_before (g);\n-      v = add_cast (ctype, v);\n-      l = limb_access (type, obj, idx, true);\n-      v = add_cast (m_limb_type, v);\n-      g = gimple_build_assign (l, v);\n-      insert_before (g);\n+      tree v2 = v;\n+      if ((prec % limb_prec) != 0)\n+\t{\n+\t  v = add_cast (ctype, v);\n+\t  l = limb_access (type, obj, idx, true);\n+\t  v = add_cast (m_limb_type, v);\n+\t  v2 = v;\n+\t  g = gimple_build_assign (l, v);\n+\t  insert_before (g);\n+\t}\n+      if (ext_ms_limb && !TYPE_UNSIGNED (type))\n+\t{\n+\t  v2 = add_cast (signed_type_for (m_limb_type), v2);\n+\t  tree lpm1 = build_int_cst (unsigned_type_node, limb_prec - 1);\n+\t  v = make_ssa_name (TREE_TYPE (v2));\n+\t  g = gimple_build_assign (v, RSHIFT_EXPR, v2, lpm1);\n+\t  insert_before (g);\n+\t  unsigned int i\n+\t    = CEIL (prec, abi_limb_prec) * abi_limb_prec / limb_prec;\n+\t  v = add_cast (m_limb_type, v);\n+\t  g = gimple_build_assign (limb_access (type, obj, size_int (i - 1),\n+\t\t\t\t\t\ttrue), v);\n+\t  insert_before (g);\n+\t  ext_ms_limb = false;\n+\t}\n     }\n-  if (zero_ms_limb)\n+  if (ext_ms_limb)\n     {\n       unsigned int i = CEIL (prec, abi_limb_prec) * abi_limb_prec / limb_prec;\n       g = gimple_build_assign (limb_access (type, obj, size_int (i - 1), true),\n--- gcc/testsuite/gcc.dg/torture/bitint-90.c.jj\t2026-04-20 14:55:37.538638849 +0200\n+++ gcc/testsuite/gcc.dg/torture/bitint-90.c\t2026-04-20 15:05:46.279391694 +0200\n@@ -0,0 +1,57 @@\n+/* PR middle-end/123635 */\n+/* { dg-do run { target bitint } } */\n+/* { dg-options \"-std=c23 -pedantic-errors -fwrapv\" } */\n+/* { dg-skip-if \"\" { ! run_expensive_tests }  { \"*\" } { \"-O0\" \"-O2\" } } */\n+/* { dg-skip-if \"\" { ! run_expensive_tests } { \"-flto\" } { \"\" } } */\n+\n+#if __BITINT_MAXWIDTH__ >= 513\n+_BitInt(513) a, b, c, d, e, f;\n+#endif\n+\n+#include \"../bitintext.h\"\n+\n+#if __BITINT_MAXWIDTH__ >= 513\n+[[gnu::noipa]] void\n+f1 (_BitInt(513) q, _BitInt(513) r, _BitInt(513) s, _BitInt(513) t)\n+{\n+  a = q * r;\n+  BEXTC (a);\n+  b = s * t;\n+  BEXTC (b);\n+  c = q / r;\n+  BEXTC (c);\n+  d = s / t;\n+  BEXTC (d);\n+  e = q % r;\n+  BEXTC (e);\n+  f = s % t;\n+  BEXTC (f);\n+}\n+#endif\n+\n+int\n+main ()\n+{\n+#if __BITINT_MAXWIDTH__ >= 513\n+  __builtin_memset (&a, 0x55, sizeof (a));\n+  __builtin_memset (&b, 0xaa, sizeof (b));\n+  __builtin_memset (&c, 0x55, sizeof (c));\n+  __builtin_memset (&d, 0xaa, sizeof (d));\n+  __builtin_memset (&e, 0x55, sizeof (e));\n+  __builtin_memset (&f, 0xaa, sizeof (f));\n+  f1 (-2569442167708526883032321549154973033917914038435073849581606377835896531300464410672941576709766056731856560073179286923714310748484353548564505224275735wb,\n+      2168258207301112123000422415459500116718298091852207369853333917415269374034197937707930701823104675221367858731870896332002496458560727651061447629500290wb,\n+      -13407807929942597099574024998205846127479365820592393377723561443721764030073546976801874298166903427690031858186486050853753882811946569946433649006084095wb - 1,\n+      -1wb);\n+  __builtin_memset (&a, 0x55, sizeof (a));\n+  __builtin_memset (&b, 0xaa, sizeof (b));\n+  __builtin_memset (&c, 0x55, sizeof (c));\n+  __builtin_memset (&d, 0xaa, sizeof (d));\n+  __builtin_memset (&e, 0x55, sizeof (e));\n+  __builtin_memset (&f, 0xaa, sizeof (f));\n+  f1 (11182751058163266900062523691585082292935101550776465927580156147436400116827277563246395892657460272253702154027346376026845302222986890807879025959971622wb,\n+      8164167338045748506210025483965624248110931799582481638963258497469488589147418794642910866480254020122096903001279978332046522756597851796198650349835611wb,\n+      -10069248634646183494877194696541551363329286638973022393951487924262590826071711445633243661699089159257055025645295791432791410100972241125609033584314602wb,\n+      -8066635266964063729231989957397398910332013431298049172762795336443651271867788598667551930170370590334905534762032999936946715583168838873523754981312676wb);\n+#endif\n+}\n","prefixes":[]}