From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wm1-x329.google.com (mail-wm1-x329.google.com [IPv6:2a00:1450:4864:20::329]) by sourceware.org (Postfix) with ESMTPS id DBC963858C56 for ; Thu, 21 Jul 2022 16:54:45 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org DBC963858C56 Received: by mail-wm1-x329.google.com with SMTP id w8-20020a05600c014800b003a32e89bc4eso1143744wmm.5 for ; Thu, 21 Jul 2022 09:54:45 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:from:date:to:cc:message-id:subject:mime-version :list-id; bh=tVP3gl35RXTy9GTzM9ef8GEmcodBVVfl5PGklxxQOc0=; b=jqZWUFUxi59VS1XMnuz8br3U0TRb4a56p/mlHHmDJIc0BLg1qeXzUgq5tNDBfBX/6x DQ56t3anscfD/46iYyj28+G0ljSZXNpyAPGp4FDHnOUAb+Me0GUbp/YXEqH0j8QLIVPQ Zomb9v9OuhY/lxp6QT9tQvacxq4AeJBhQFpn2pwqBHDBimpuwsWelXZbP3YYqQKsgxm/ mNItXDRXXqKwdR6aIG19XGe9q1co7edcRhKn841PONIoDcZPhQgHdNEfxN73w1zljxYK tINmfVrkjYPP0mBpyw7o+lCFOMn2qrloY9ID/opf87/MqmEyADz0URx4VGMzXF1Bu2/e xzHA== X-Gm-Message-State: AJIora9LW9Ls0O20NQ1Iu25e9r8PGfG4ogIMgXTlU4hhReR5aXrlYDWC Dr97CcYegHmZ4/FFskOovyr6YA== X-Google-Smtp-Source: AGRyM1vpaq29JmnomZMs47o9L/NSsFm8mBSXzKDA4uCUlCWgl+PNGbI2E4ZMrXbXVqWIHcZuwxWdIg== X-Received: by 2002:a7b:cb0c:0:b0:3a3:2a38:5066 with SMTP id u12-20020a7bcb0c000000b003a32a385066mr7877591wmj.74.1658422483877; Thu, 21 Jul 2022 09:54:43 -0700 (PDT) Received: from jenkins.jenkins (ci.linaro.org. [88.99.136.175]) by smtp.gmail.com with ESMTPSA id r67-20020a1c2b46000000b003a308e9a192sm6268006wmr.30.2022.07.21.09.54.42 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 21 Jul 2022 09:54:43 -0700 (PDT) From: ci_notify@linaro.org X-Google-Original-From: linaro-infrastructure-errors@lists.linaro.org Date: Thu, 21 Jul 2022 16:54:41 +0000 (UTC) To: Richard Biener Cc: gcc-regression@gcc.gnu.org Message-ID: <279373770.2069.1658422483393@jenkins.jenkins> Subject: [TCWG CI] Regression caused by gcc: Teach VN about masked/len stores MIME-Version: 1.0 X-Jenkins-Job: TCWG Bisect tcwg_gcc_bootstrap/master-arm-bootstrap_O3 X-Jenkins-Result: SUCCESS X-Spam-Status: No, score=-13.5 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, KAM_LOTSOFHASH, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit X-Content-Filtered-By: Mailman/MimeDel 2.1.29 X-BeenThere: gcc-regression@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-regression mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 21 Jul 2022 16:54:49 -0000 [TCWG CI] Regression caused by gcc: Teach VN about masked/len stores: commit bd9837bc3ca1344c32aef7ba9f8fa1785063132e Author: Richard Biener Teach VN about masked/len stores Results regressed to # reset_artifacts: -10 # true: 0 # build_abe binutils: 1 # First few build errors in logs: # 00:06:14 make[3]: [Makefile:1787: armv8l-unknown-linux-gnueabihf/bits/largefile-config.h] Error 1 (ignored) # 00:08:57 /home/tcwg-buildslave/workspace/tcwg_gnu_2/abe/snapshots/gcc.git~master/libiberty/../include/xregex.h:8:19: internal compiler error: Segmentation fault # 00:08:58 make[3]: *** [Makefile:1218: regex.o] Error 1 # 00:09:30 make[2]: *** [Makefile:12585: all-stage2-libiberty] Error 2 # 00:09:30 make[1]: *** [Makefile:25739: stage2-bubble] Error 2 # 00:09:30 make: *** [Makefile:1072: all] Error 2 from # reset_artifacts: -10 # true: 0 # build_abe binutils: 1 # build_abe bootstrap_O3: 2 THIS IS THE END OF INTERESTING STUFF. BELOW ARE LINKS TO BUILDS, REPRODUCTION INSTRUCTIONS, AND THE RAW COMMIT. This commit has regressed these CI configurations: - tcwg_gcc_bootstrap/master-arm-bootstrap_O3 First_bad build: https://ci.linaro.org/job/tcwg_gcc_bootstrap-bisect-master-arm-bootstrap_O3/11/artifact/artifacts/build-bd9837bc3ca1344c32aef7ba9f8fa1785063132e/ Last_good build: https://ci.linaro.org/job/tcwg_gcc_bootstrap-bisect-master-arm-bootstrap_O3/11/artifact/artifacts/build-f4ed610d02aaf8cfcdcb5cf03e0cde65f1f5f890/ Baseline build: https://ci.linaro.org/job/tcwg_gcc_bootstrap-bisect-master-arm-bootstrap_O3/11/artifact/artifacts/build-baseline/ Even more details: https://ci.linaro.org/job/tcwg_gcc_bootstrap-bisect-master-arm-bootstrap_O3/11/artifact/artifacts/ Reproduce builds: mkdir investigate-gcc-bd9837bc3ca1344c32aef7ba9f8fa1785063132e cd investigate-gcc-bd9837bc3ca1344c32aef7ba9f8fa1785063132e # Fetch scripts git clone https://git.linaro.org/toolchain/jenkins-scripts # Fetch manifests and test.sh script mkdir -p artifacts/manifests curl -o artifacts/manifests/build-baseline.sh https://ci.linaro.org/job/tcwg_gcc_bootstrap-bisect-master-arm-bootstrap_O3/11/artifact/artifacts/manifests/build-baseline.sh --fail curl -o artifacts/manifests/build-parameters.sh https://ci.linaro.org/job/tcwg_gcc_bootstrap-bisect-master-arm-bootstrap_O3/11/artifact/artifacts/manifests/build-parameters.sh --fail curl -o artifacts/test.sh https://ci.linaro.org/job/tcwg_gcc_bootstrap-bisect-master-arm-bootstrap_O3/11/artifact/artifacts/test.sh --fail chmod +x artifacts/test.sh # Reproduce the baseline build (build all pre-requisites) ./jenkins-scripts/tcwg_gnu-build.sh @@ artifacts/manifests/build-baseline.sh # Save baseline build state (which is then restored in artifacts/test.sh) mkdir -p ./bisect rsync -a --del --delete-excluded --exclude /bisect/ --exclude /artifacts/ --exclude /gcc/ ./ ./bisect/baseline/ cd gcc # Reproduce first_bad build git checkout --detach bd9837bc3ca1344c32aef7ba9f8fa1785063132e ../artifacts/test.sh # Reproduce last_good build git checkout --detach f4ed610d02aaf8cfcdcb5cf03e0cde65f1f5f890 ../artifacts/test.sh cd .. Full commit (up to 1000 lines): commit bd9837bc3ca1344c32aef7ba9f8fa1785063132e Author: Richard Biener Date: Wed Jul 20 12:28:26 2022 +0200 Teach VN about masked/len stores The following teaches VN to handle reads from .MASK_STORE and .LEN_STORE. For this push_partial_def is extended first for convenience so we don't have to handle the full def case in the caller (possibly other paths can be simplified then). Also the partial definition stored value can have an offset applied so we don't have to build a fake RHS when we register the pieces of an existing store. PR tree-optimization/106365 * tree-ssa-sccvn.cc (pd_data::rhs_off): New field determining the offset to start encoding of RHS from. (vn_walk_cb_data::vn_walk_cb_data): Initialize it. (vn_walk_cb_data::push_partial_def): Allow the first partial definition to be fully providing the def. Offset RHS before encoding if requested. (vn_reference_lookup_3): Initialize def_rhs everywhere. Add support for .MASK_STORE and .LEN_STORE (partial) definitions. * gcc.target/i386/vec-maskstore-vn.c: New testcase. --- gcc/testsuite/gcc.target/i386/vec-maskstore-vn.c | 30 +++ gcc/tree-ssa-sccvn.cc | 255 ++++++++++++++++++----- 2 files changed, 228 insertions(+), 57 deletions(-) diff --git a/gcc/testsuite/gcc.target/i386/vec-maskstore-vn.c b/gcc/testsuite/gcc.target/i386/vec-maskstore-vn.c new file mode 100644 index 00000000000..98213905ece --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/vec-maskstore-vn.c @@ -0,0 +1,30 @@ +/* { dg-do compile } */ +/* { dg-options "-O3 -mavx2 -fdump-tree-fre5" } */ + +void __attribute__((noinline,noclone)) +foo (int *out, int *res) +{ + int mask[] = { 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1 }; + int i; + for (i = 0; i < 16; ++i) + { + if (mask[i]) + out[i] = i; + } + int o0 = out[0]; + int o7 = out[7]; + int o14 = out[14]; + int o15 = out[15]; + res[0] = o0; + res[2] = o7; + res[4] = o14; + res[6] = o15; +} + +/* Vectorization produces .MASK_STORE, unrolling will unroll the two + vector iterations. FRE5 after that should be able to CSE + out[7] and out[15], but leave out[0] and out[14] alone. */ +/* { dg-final { scan-tree-dump " = o0_\[0-9\]+;" "fre5" } } */ +/* { dg-final { scan-tree-dump " = 7;" "fre5" } } */ +/* { dg-final { scan-tree-dump " = o14_\[0-9\]+;" "fre5" } } */ +/* { dg-final { scan-tree-dump " = 15;" "fre5" } } */ diff --git a/gcc/tree-ssa-sccvn.cc b/gcc/tree-ssa-sccvn.cc index f41d5031365..7d947b55a27 100644 --- a/gcc/tree-ssa-sccvn.cc +++ b/gcc/tree-ssa-sccvn.cc @@ -1790,6 +1790,7 @@ struct pd_range struct pd_data { tree rhs; + HOST_WIDE_INT rhs_off; HOST_WIDE_INT offset; HOST_WIDE_INT size; }; @@ -1816,6 +1817,7 @@ struct vn_walk_cb_data unsigned int pos = 0, prec = w.get_precision (); pd_data pd; pd.rhs = build_constructor (NULL_TREE, NULL); + pd.rhs_off = 0; /* When bitwise and with a constant is done on a memory load, we don't really need all the bits to be defined or defined to constants, we don't really care what is in the position @@ -1976,6 +1978,7 @@ vn_walk_cb_data::push_partial_def (pd_data pd, bool pd_constant_p = (TREE_CODE (pd.rhs) == CONSTRUCTOR || CONSTANT_CLASS_P (pd.rhs)); + pd_range *r; if (partial_defs.is_empty ()) { /* If we get a clobber upfront, fail. */ @@ -1989,65 +1992,70 @@ vn_walk_cb_data::push_partial_def (pd_data pd, first_set = set; first_base_set = base_set; last_vuse_ptr = NULL; - /* Continue looking for partial defs. */ - return NULL; - } - - if (!known_ranges) - { - /* ??? Optimize the case where the 2nd partial def completes things. */ - gcc_obstack_init (&ranges_obstack); - known_ranges = splay_tree_new_with_allocator (pd_range_compare, 0, 0, - pd_tree_alloc, - pd_tree_dealloc, this); - splay_tree_insert (known_ranges, - (splay_tree_key)&first_range.offset, - (splay_tree_value)&first_range); - } - - pd_range newr = { pd.offset, pd.size }; - splay_tree_node n; - pd_range *r; - /* Lookup the predecessor of offset + 1 and see if we need to merge. */ - HOST_WIDE_INT loffset = newr.offset + 1; - if ((n = splay_tree_predecessor (known_ranges, (splay_tree_key)&loffset)) - && ((r = (pd_range *)n->value), true) - && ranges_known_overlap_p (r->offset, r->size + 1, - newr.offset, newr.size)) - { - /* Ignore partial defs already covered. Here we also drop shadowed - clobbers arriving here at the floor. */ - if (known_subrange_p (newr.offset, newr.size, r->offset, r->size)) - return NULL; - r->size = MAX (r->offset + r->size, newr.offset + newr.size) - r->offset; + r = &first_range; + /* Go check if the first partial definition was a full one in case + the caller didn't optimize for this. */ } else { - /* newr.offset wasn't covered yet, insert the range. */ - r = XOBNEW (&ranges_obstack, pd_range); - *r = newr; - splay_tree_insert (known_ranges, (splay_tree_key)&r->offset, - (splay_tree_value)r); - } - /* Merge r which now contains newr and is a member of the splay tree with - adjacent overlapping ranges. */ - pd_range *rafter; - while ((n = splay_tree_successor (known_ranges, (splay_tree_key)&r->offset)) - && ((rafter = (pd_range *)n->value), true) - && ranges_known_overlap_p (r->offset, r->size + 1, - rafter->offset, rafter->size)) - { - r->size = MAX (r->offset + r->size, - rafter->offset + rafter->size) - r->offset; - splay_tree_remove (known_ranges, (splay_tree_key)&rafter->offset); - } - /* If we get a clobber, fail. */ - if (TREE_CLOBBER_P (pd.rhs)) - return (void *)-1; - /* Non-constants are OK as long as they are shadowed by a constant. */ - if (!pd_constant_p) - return (void *)-1; - partial_defs.safe_push (pd); + if (!known_ranges) + { + /* ??? Optimize the case where the 2nd partial def completes + things. */ + gcc_obstack_init (&ranges_obstack); + known_ranges = splay_tree_new_with_allocator (pd_range_compare, 0, 0, + pd_tree_alloc, + pd_tree_dealloc, this); + splay_tree_insert (known_ranges, + (splay_tree_key)&first_range.offset, + (splay_tree_value)&first_range); + } + + pd_range newr = { pd.offset, pd.size }; + splay_tree_node n; + /* Lookup the predecessor of offset + 1 and see if we need to merge. */ + HOST_WIDE_INT loffset = newr.offset + 1; + if ((n = splay_tree_predecessor (known_ranges, (splay_tree_key)&loffset)) + && ((r = (pd_range *)n->value), true) + && ranges_known_overlap_p (r->offset, r->size + 1, + newr.offset, newr.size)) + { + /* Ignore partial defs already covered. Here we also drop shadowed + clobbers arriving here at the floor. */ + if (known_subrange_p (newr.offset, newr.size, r->offset, r->size)) + return NULL; + r->size + = MAX (r->offset + r->size, newr.offset + newr.size) - r->offset; + } + else + { + /* newr.offset wasn't covered yet, insert the range. */ + r = XOBNEW (&ranges_obstack, pd_range); + *r = newr; + splay_tree_insert (known_ranges, (splay_tree_key)&r->offset, + (splay_tree_value)r); + } + /* Merge r which now contains newr and is a member of the splay tree with + adjacent overlapping ranges. */ + pd_range *rafter; + while ((n = splay_tree_successor (known_ranges, + (splay_tree_key)&r->offset)) + && ((rafter = (pd_range *)n->value), true) + && ranges_known_overlap_p (r->offset, r->size + 1, + rafter->offset, rafter->size)) + { + r->size = MAX (r->offset + r->size, + rafter->offset + rafter->size) - r->offset; + splay_tree_remove (known_ranges, (splay_tree_key)&rafter->offset); + } + /* If we get a clobber, fail. */ + if (TREE_CLOBBER_P (pd.rhs)) + return (void *)-1; + /* Non-constants are OK as long as they are shadowed by a constant. */ + if (!pd_constant_p) + return (void *)-1; + partial_defs.safe_push (pd); + } /* Now we have merged newr into the range tree. When we have covered [offseti, sizei] then the tree will contain exactly one node which has @@ -2081,7 +2089,8 @@ vn_walk_cb_data::push_partial_def (pd_data pd, else { len = native_encode_expr (pd.rhs, this_buffer, bufsize, - MAX (0, -pd.offset) / BITS_PER_UNIT); + (MAX (0, -pd.offset) + + pd.rhs_off) / BITS_PER_UNIT); if (len <= 0 || len < (ROUND_UP (pd.size, BITS_PER_UNIT) / BITS_PER_UNIT - MAX (0, -pd.offset) / BITS_PER_UNIT)) @@ -2906,6 +2915,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, { pd_data pd; pd.rhs = build_constructor (NULL_TREE, NULL); + pd.rhs_off = 0; pd.offset = offset2i; pd.size = leni << LOG2_BITS_PER_UNIT; return data->push_partial_def (pd, 0, 0, offseti, maxsizei); @@ -2955,6 +2965,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, by a later def. */ pd_data pd; pd.rhs = gimple_assign_rhs1 (def_stmt); + pd.rhs_off = 0; pd.offset = offset2i; pd.size = size2i; return data->push_partial_def (pd, ao_ref_alias_set (&lhs_ref), @@ -3107,6 +3118,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, if (TREE_CODE (rhs) == SSA_NAME) rhs = SSA_VAL (rhs); pd.rhs = rhs; + pd.rhs_off = 0; pd.offset = offset2i; pd.size = size2i; return data->push_partial_def (pd, ao_ref_alias_set (&lhs_ref), @@ -3186,6 +3198,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, { pd_data pd; pd.rhs = SSA_VAL (def_rhs); + pd.rhs_off = 0; pd.offset = offset2i; pd.size = size2i; return data->push_partial_def (pd, ao_ref_alias_set (&lhs_ref), @@ -3195,6 +3208,133 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, } } + /* 4b) Assignment done via one of the vectorizer internal store + functions where we may be able to access pieces from or we can + combine to a larger entity. */ + else if (known_eq (ref->size, maxsize) + && is_gimple_reg_type (vr->type) + && !reverse_storage_order_for_component_p (vr->operands) + && !contains_storage_order_barrier_p (vr->operands) + && is_gimple_call (def_stmt) + && gimple_call_internal_p (def_stmt) + && internal_store_fn_p (gimple_call_internal_fn (def_stmt))) + { + gcall *call = as_a (def_stmt); + internal_fn fn = gimple_call_internal_fn (call); + tree def_rhs = gimple_call_arg (call, + internal_fn_stored_value_index (fn)); + def_rhs = vn_valueize (def_rhs); + if (TREE_CODE (def_rhs) != VECTOR_CST) + return (void *)-1; + + tree mask = NULL_TREE, len = NULL_TREE, bias = NULL_TREE; + switch (fn) + { + case IFN_MASK_STORE: + mask = gimple_call_arg (call, internal_fn_mask_index (fn)); + mask = vn_valueize (mask); + if (TREE_CODE (mask) != VECTOR_CST) + return (void *)-1; + break; + case IFN_LEN_STORE: + len = gimple_call_arg (call, 2); + bias = gimple_call_arg (call, 4); + if (!tree_fits_uhwi_p (len) || !tree_fits_shwi_p (bias)) + return (void *)-1; + break; + default: + return (void *)-1; + } + ao_ref_init_from_ptr_and_size (&lhs_ref, + vn_valueize (gimple_call_arg (call, 0)), + TYPE_SIZE_UNIT (TREE_TYPE (def_rhs))); + tree base2; + poly_int64 offset2, size2, maxsize2; + HOST_WIDE_INT offset2i, size2i, offseti; + base2 = ao_ref_base (&lhs_ref); + offset2 = lhs_ref.offset; + size2 = lhs_ref.size; + maxsize2 = lhs_ref.max_size; + if (known_size_p (maxsize2) + && known_eq (maxsize2, size2) + && adjust_offsets_for_equal_base_address (base, &offset, + base2, &offset2) + && maxsize.is_constant (&maxsizei) + && offset.is_constant (&offseti) + && offset2.is_constant (&offset2i) + && size2.is_constant (&size2i)) + { + if (!ranges_maybe_overlap_p (offset, maxsize, offset2, size2)) + /* Poor-mans disambiguation. */ + return NULL; + else if (ranges_known_overlap_p (offset, maxsize, offset2, size2)) + { + pd_data pd; + pd.rhs = def_rhs; + tree aa = gimple_call_arg (call, 1); + alias_set_type set = get_deref_alias_set (TREE_TYPE (aa)); + tree vectype = TREE_TYPE (def_rhs); + unsigned HOST_WIDE_INT elsz + = tree_to_uhwi (TYPE_SIZE (TREE_TYPE (vectype))); + if (mask) + { + HOST_WIDE_INT start = 0, len = 0; + unsigned mask_idx = 0; + do + { + if (integer_zerop (VECTOR_CST_ELT (mask, mask_idx))) + { + if (len != 0) + { + pd.rhs_off = start; + pd.offset = offset2i + start; + pd.size = len; + if (ranges_known_overlap_p + (offset, maxsize, pd.offset, pd.size)) + { + void *res = data->push_partial_def + (pd, set, set, offseti, maxsizei); + if (res != NULL) + return res; + } + } + start = (mask_idx + 1) * elsz; + len = 0; + } + else + len += elsz; + mask_idx++; + } + while (known_lt (mask_idx, TYPE_VECTOR_SUBPARTS (vectype))); + if (len != 0) + { + pd.rhs_off = start; + pd.offset = offset2i + start; + pd.size = len; + if (ranges_known_overlap_p (offset, maxsize, + pd.offset, pd.size)) + return data->push_partial_def (pd, set, set, + offseti, maxsizei); + } + } + else if (fn == IFN_LEN_STORE) + { + pd.rhs_off = 0; + pd.offset = offset2i; + pd.size = (tree_to_uhwi (len) + + -tree_to_shwi (bias)) * BITS_PER_UNIT; + if (ranges_known_overlap_p (offset, maxsize, + pd.offset, pd.size)) + return data->push_partial_def (pd, set, set, + offseti, maxsizei); + } + else + gcc_unreachable (); + return NULL; + } + } + } + /* 5) For aggregate copies translate the reference through them if the copy kills ref. */ else if (data->vn_walk_kind == VN_WALKREWRITE @@ -3327,6 +3467,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, { pd_data pd; pd.rhs = val; + pd.rhs_off = 0; pd.offset = 0; pd.size = maxsizei; return data->push_partial_def (pd, ao_ref_alias_set (&lhs_ref), >>From ci_notify@linaro.org Thu Jul 21 20:13:05 2022 Return-Path: X-Original-To: gcc-regression@gcc.gnu.org Delivered-To: gcc-regression@gcc.gnu.org Received: from mail-wm1-x330.google.com (mail-wm1-x330.google.com [IPv6:2a00:1450:4864:20::330]) by sourceware.org (Postfix) with ESMTPS id E8DCD3838F33 for ; Thu, 21 Jul 2022 20:13:01 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org E8DCD3838F33 Received: by mail-wm1-x330.google.com with SMTP id j29-20020a05600c1c1d00b003a2fdafdefbso1405360wms.2 for ; Thu, 21 Jul 2022 13:13:01 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:from:date:to:cc:message-id:subject:mime-version :list-id; bh=0dGlzuU7g06SjR54MAC9azlcQdLs6GWjR4AIqnvaIbw=; b=HjQLpvUsZcDJeYAoDLQVnol7r0+ZBkwoI76x+5POW6+GM3b6h0xPEcUYroC19IlQPi svQ7G1g94nJI4mtO4ZeeylvEdlX9xtmyZS3MAFFRDdr6qg3UN8dhgh4mY8mxOZD9seJ5 /W4g9PWMsO+fuPBCroGeBMOzywjLxhXoK1gyZFT2L4P0ZtYOTGPYbOS0cSiY8wn7mr0C P4GU+x+iubFSlfH0uD/SbQtGNu12JnVyKY7gEswOfCxGV6hb4IWBDO6m88ZXtQPAwJey o7TVgB9m2zVDYG6JfRUKbJsdGlYGM6RSq+ExCISaafGcDlC0+qtgoyI+mykBPraizv+4 GcQA== X-Gm-Message-State: AJIora/+7WrS4TG/AKklOLkVeFIdiUhGEjpmUxAo9rm3umrI67Sa5m6O gJ/ZWX/ELuTdTaMlsAJ/r6keeA== X-Google-Smtp-Source: AGRyM1s2SBBko519Zfln9WPhaDxhT36i042sc+MlemAtcPSTuTwET6jYIR1kByCNdj12Ni5ufNrB9w== X-Received: by 2002:a7b:c450:0:b0:3a3:2121:c690 with SMTP id l16-20020a7bc450000000b003a32121c690mr9520353wmi.99.1658434380135; Thu, 21 Jul 2022 13:13:00 -0700 (PDT) Received: from jenkins.jenkins (ci.linaro.org. [88.99.136.175]) by smtp.gmail.com with ESMTPSA id e10-20020adfe7ca000000b0021dbaa4f38dsm3519093wrn.18.2022.07.21.13.12.59 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 21 Jul 2022 13:12:59 -0700 (PDT) From: ci_notify@linaro.org X-Google-Original-From: linaro-infrastructure-errors@lists.linaro.org Date: Thu, 21 Jul 2022 20:12:57 +0000 (UTC) To: Richard Biener Cc: gcc-regression@gcc.gnu.org Message-ID: <2000386721.2106.1658434379326@jenkins.jenkins> Subject: [TCWG CI] Regression caused by gcc: Teach VN about masked/len stores MIME-Version: 1.0 X-Jenkins-Job: TCWG Bisect tcwg_gnu_native_build/master-arm X-Jenkins-Result: SUCCESS X-Spam-Status: No, score=-13.5 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, KAM_LOTSOFHASH, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit X-Content-Filtered-By: Mailman/MimeDel 2.1.29 X-BeenThere: gcc-regression@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-regression mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 21 Jul 2022 20:13:05 -0000 [TCWG CI] Regression caused by gcc: Teach VN about masked/len stores: commit bd9837bc3ca1344c32aef7ba9f8fa1785063132e Author: Richard Biener Teach VN about masked/len stores Results regressed to # reset_artifacts: -10 # true: 0 # build_abe binutils: 1 # build_abe gcc: 2 # build_abe linux: 4 # First few build errors in logs: # 00:02:21 crypt_util.c:607:1: internal compiler error: Segmentation fault # 00:02:21 crypt_util.c:607:1: internal compiler error: Segmentation fault # 00:02:21 make[2]: *** [../o-iterator.mk:9: /home/tcwg-buildslave/workspace/tcwg_gnu_13/abe/builds/armv8l-unknown-linux-gnueabihf/armv8l-unknown-linux-gnueabihf/glibc-glibc.git~master/crypt/crypt_util.o] Error 1 # 00:02:21 make[2]: *** [../o-iterator.mk:9: /home/tcwg-buildslave/workspace/tcwg_gnu_13/abe/builds/armv8l-unknown-linux-gnueabihf/armv8l-unknown-linux-gnueabihf/glibc-glibc.git~master/crypt/crypt_util.os] Error 1 # 00:02:21 make[1]: *** [Makefile:484: crypt/others] Error 2 # 00:02:21 make: *** [Makefile:9: all] Error 2 from # reset_artifacts: -10 # true: 0 # build_abe binutils: 1 # build_abe gcc: 2 # build_abe linux: 4 # build_abe glibc: 5 # build_abe gdb: 6 THIS IS THE END OF INTERESTING STUFF. BELOW ARE LINKS TO BUILDS, REPRODUCTION INSTRUCTIONS, AND THE RAW COMMIT. This commit has regressed these CI configurations: - tcwg_gnu_native_build/master-arm First_bad build: https://ci.linaro.org/job/tcwg_gnu_native_build-bisect-master-arm/11/artifact/artifacts/build-bd9837bc3ca1344c32aef7ba9f8fa1785063132e/ Last_good build: https://ci.linaro.org/job/tcwg_gnu_native_build-bisect-master-arm/11/artifact/artifacts/build-f4ed610d02aaf8cfcdcb5cf03e0cde65f1f5f890/ Baseline build: https://ci.linaro.org/job/tcwg_gnu_native_build-bisect-master-arm/11/artifact/artifacts/build-baseline/ Even more details: https://ci.linaro.org/job/tcwg_gnu_native_build-bisect-master-arm/11/artifact/artifacts/ Reproduce builds: mkdir investigate-gcc-bd9837bc3ca1344c32aef7ba9f8fa1785063132e cd investigate-gcc-bd9837bc3ca1344c32aef7ba9f8fa1785063132e # Fetch scripts git clone https://git.linaro.org/toolchain/jenkins-scripts # Fetch manifests and test.sh script mkdir -p artifacts/manifests curl -o artifacts/manifests/build-baseline.sh https://ci.linaro.org/job/tcwg_gnu_native_build-bisect-master-arm/11/artifact/artifacts/manifests/build-baseline.sh --fail curl -o artifacts/manifests/build-parameters.sh https://ci.linaro.org/job/tcwg_gnu_native_build-bisect-master-arm/11/artifact/artifacts/manifests/build-parameters.sh --fail curl -o artifacts/test.sh https://ci.linaro.org/job/tcwg_gnu_native_build-bisect-master-arm/11/artifact/artifacts/test.sh --fail chmod +x artifacts/test.sh # Reproduce the baseline build (build all pre-requisites) ./jenkins-scripts/tcwg_gnu-build.sh @@ artifacts/manifests/build-baseline.sh # Save baseline build state (which is then restored in artifacts/test.sh) mkdir -p ./bisect rsync -a --del --delete-excluded --exclude /bisect/ --exclude /artifacts/ --exclude /gcc/ ./ ./bisect/baseline/ cd gcc # Reproduce first_bad build git checkout --detach bd9837bc3ca1344c32aef7ba9f8fa1785063132e ../artifacts/test.sh # Reproduce last_good build git checkout --detach f4ed610d02aaf8cfcdcb5cf03e0cde65f1f5f890 ../artifacts/test.sh cd .. Full commit (up to 1000 lines): commit bd9837bc3ca1344c32aef7ba9f8fa1785063132e Author: Richard Biener Date: Wed Jul 20 12:28:26 2022 +0200 Teach VN about masked/len stores The following teaches VN to handle reads from .MASK_STORE and .LEN_STORE. For this push_partial_def is extended first for convenience so we don't have to handle the full def case in the caller (possibly other paths can be simplified then). Also the partial definition stored value can have an offset applied so we don't have to build a fake RHS when we register the pieces of an existing store. PR tree-optimization/106365 * tree-ssa-sccvn.cc (pd_data::rhs_off): New field determining the offset to start encoding of RHS from. (vn_walk_cb_data::vn_walk_cb_data): Initialize it. (vn_walk_cb_data::push_partial_def): Allow the first partial definition to be fully providing the def. Offset RHS before encoding if requested. (vn_reference_lookup_3): Initialize def_rhs everywhere. Add support for .MASK_STORE and .LEN_STORE (partial) definitions. * gcc.target/i386/vec-maskstore-vn.c: New testcase. --- gcc/testsuite/gcc.target/i386/vec-maskstore-vn.c | 30 +++ gcc/tree-ssa-sccvn.cc | 255 ++++++++++++++++++----- 2 files changed, 228 insertions(+), 57 deletions(-) diff --git a/gcc/testsuite/gcc.target/i386/vec-maskstore-vn.c b/gcc/testsuite/gcc.target/i386/vec-maskstore-vn.c new file mode 100644 index 00000000000..98213905ece --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/vec-maskstore-vn.c @@ -0,0 +1,30 @@ +/* { dg-do compile } */ +/* { dg-options "-O3 -mavx2 -fdump-tree-fre5" } */ + +void __attribute__((noinline,noclone)) +foo (int *out, int *res) +{ + int mask[] = { 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1 }; + int i; + for (i = 0; i < 16; ++i) + { + if (mask[i]) + out[i] = i; + } + int o0 = out[0]; + int o7 = out[7]; + int o14 = out[14]; + int o15 = out[15]; + res[0] = o0; + res[2] = o7; + res[4] = o14; + res[6] = o15; +} + +/* Vectorization produces .MASK_STORE, unrolling will unroll the two + vector iterations. FRE5 after that should be able to CSE + out[7] and out[15], but leave out[0] and out[14] alone. */ +/* { dg-final { scan-tree-dump " = o0_\[0-9\]+;" "fre5" } } */ +/* { dg-final { scan-tree-dump " = 7;" "fre5" } } */ +/* { dg-final { scan-tree-dump " = o14_\[0-9\]+;" "fre5" } } */ +/* { dg-final { scan-tree-dump " = 15;" "fre5" } } */ diff --git a/gcc/tree-ssa-sccvn.cc b/gcc/tree-ssa-sccvn.cc index f41d5031365..7d947b55a27 100644 --- a/gcc/tree-ssa-sccvn.cc +++ b/gcc/tree-ssa-sccvn.cc @@ -1790,6 +1790,7 @@ struct pd_range struct pd_data { tree rhs; + HOST_WIDE_INT rhs_off; HOST_WIDE_INT offset; HOST_WIDE_INT size; }; @@ -1816,6 +1817,7 @@ struct vn_walk_cb_data unsigned int pos = 0, prec = w.get_precision (); pd_data pd; pd.rhs = build_constructor (NULL_TREE, NULL); + pd.rhs_off = 0; /* When bitwise and with a constant is done on a memory load, we don't really need all the bits to be defined or defined to constants, we don't really care what is in the position @@ -1976,6 +1978,7 @@ vn_walk_cb_data::push_partial_def (pd_data pd, bool pd_constant_p = (TREE_CODE (pd.rhs) == CONSTRUCTOR || CONSTANT_CLASS_P (pd.rhs)); + pd_range *r; if (partial_defs.is_empty ()) { /* If we get a clobber upfront, fail. */ @@ -1989,65 +1992,70 @@ vn_walk_cb_data::push_partial_def (pd_data pd, first_set = set; first_base_set = base_set; last_vuse_ptr = NULL; - /* Continue looking for partial defs. */ - return NULL; - } - - if (!known_ranges) - { - /* ??? Optimize the case where the 2nd partial def completes things. */ - gcc_obstack_init (&ranges_obstack); - known_ranges = splay_tree_new_with_allocator (pd_range_compare, 0, 0, - pd_tree_alloc, - pd_tree_dealloc, this); - splay_tree_insert (known_ranges, - (splay_tree_key)&first_range.offset, - (splay_tree_value)&first_range); - } - - pd_range newr = { pd.offset, pd.size }; - splay_tree_node n; - pd_range *r; - /* Lookup the predecessor of offset + 1 and see if we need to merge. */ - HOST_WIDE_INT loffset = newr.offset + 1; - if ((n = splay_tree_predecessor (known_ranges, (splay_tree_key)&loffset)) - && ((r = (pd_range *)n->value), true) - && ranges_known_overlap_p (r->offset, r->size + 1, - newr.offset, newr.size)) - { - /* Ignore partial defs already covered. Here we also drop shadowed - clobbers arriving here at the floor. */ - if (known_subrange_p (newr.offset, newr.size, r->offset, r->size)) - return NULL; - r->size = MAX (r->offset + r->size, newr.offset + newr.size) - r->offset; + r = &first_range; + /* Go check if the first partial definition was a full one in case + the caller didn't optimize for this. */ } else { - /* newr.offset wasn't covered yet, insert the range. */ - r = XOBNEW (&ranges_obstack, pd_range); - *r = newr; - splay_tree_insert (known_ranges, (splay_tree_key)&r->offset, - (splay_tree_value)r); - } - /* Merge r which now contains newr and is a member of the splay tree with - adjacent overlapping ranges. */ - pd_range *rafter; - while ((n = splay_tree_successor (known_ranges, (splay_tree_key)&r->offset)) - && ((rafter = (pd_range *)n->value), true) - && ranges_known_overlap_p (r->offset, r->size + 1, - rafter->offset, rafter->size)) - { - r->size = MAX (r->offset + r->size, - rafter->offset + rafter->size) - r->offset; - splay_tree_remove (known_ranges, (splay_tree_key)&rafter->offset); - } - /* If we get a clobber, fail. */ - if (TREE_CLOBBER_P (pd.rhs)) - return (void *)-1; - /* Non-constants are OK as long as they are shadowed by a constant. */ - if (!pd_constant_p) - return (void *)-1; - partial_defs.safe_push (pd); + if (!known_ranges) + { + /* ??? Optimize the case where the 2nd partial def completes + things. */ + gcc_obstack_init (&ranges_obstack); + known_ranges = splay_tree_new_with_allocator (pd_range_compare, 0, 0, + pd_tree_alloc, + pd_tree_dealloc, this); + splay_tree_insert (known_ranges, + (splay_tree_key)&first_range.offset, + (splay_tree_value)&first_range); + } + + pd_range newr = { pd.offset, pd.size }; + splay_tree_node n; + /* Lookup the predecessor of offset + 1 and see if we need to merge. */ + HOST_WIDE_INT loffset = newr.offset + 1; + if ((n = splay_tree_predecessor (known_ranges, (splay_tree_key)&loffset)) + && ((r = (pd_range *)n->value), true) + && ranges_known_overlap_p (r->offset, r->size + 1, + newr.offset, newr.size)) + { + /* Ignore partial defs already covered. Here we also drop shadowed + clobbers arriving here at the floor. */ + if (known_subrange_p (newr.offset, newr.size, r->offset, r->size)) + return NULL; + r->size + = MAX (r->offset + r->size, newr.offset + newr.size) - r->offset; + } + else + { + /* newr.offset wasn't covered yet, insert the range. */ + r = XOBNEW (&ranges_obstack, pd_range); + *r = newr; + splay_tree_insert (known_ranges, (splay_tree_key)&r->offset, + (splay_tree_value)r); + } + /* Merge r which now contains newr and is a member of the splay tree with + adjacent overlapping ranges. */ + pd_range *rafter; + while ((n = splay_tree_successor (known_ranges, + (splay_tree_key)&r->offset)) + && ((rafter = (pd_range *)n->value), true) + && ranges_known_overlap_p (r->offset, r->size + 1, + rafter->offset, rafter->size)) + { + r->size = MAX (r->offset + r->size, + rafter->offset + rafter->size) - r->offset; + splay_tree_remove (known_ranges, (splay_tree_key)&rafter->offset); + } + /* If we get a clobber, fail. */ + if (TREE_CLOBBER_P (pd.rhs)) + return (void *)-1; + /* Non-constants are OK as long as they are shadowed by a constant. */ + if (!pd_constant_p) + return (void *)-1; + partial_defs.safe_push (pd); + } /* Now we have merged newr into the range tree. When we have covered [offseti, sizei] then the tree will contain exactly one node which has @@ -2081,7 +2089,8 @@ vn_walk_cb_data::push_partial_def (pd_data pd, else { len = native_encode_expr (pd.rhs, this_buffer, bufsize, - MAX (0, -pd.offset) / BITS_PER_UNIT); + (MAX (0, -pd.offset) + + pd.rhs_off) / BITS_PER_UNIT); if (len <= 0 || len < (ROUND_UP (pd.size, BITS_PER_UNIT) / BITS_PER_UNIT - MAX (0, -pd.offset) / BITS_PER_UNIT)) @@ -2906,6 +2915,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, { pd_data pd; pd.rhs = build_constructor (NULL_TREE, NULL); + pd.rhs_off = 0; pd.offset = offset2i; pd.size = leni << LOG2_BITS_PER_UNIT; return data->push_partial_def (pd, 0, 0, offseti, maxsizei); @@ -2955,6 +2965,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, by a later def. */ pd_data pd; pd.rhs = gimple_assign_rhs1 (def_stmt); + pd.rhs_off = 0; pd.offset = offset2i; pd.size = size2i; return data->push_partial_def (pd, ao_ref_alias_set (&lhs_ref), @@ -3107,6 +3118,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, if (TREE_CODE (rhs) == SSA_NAME) rhs = SSA_VAL (rhs); pd.rhs = rhs; + pd.rhs_off = 0; pd.offset = offset2i; pd.size = size2i; return data->push_partial_def (pd, ao_ref_alias_set (&lhs_ref), @@ -3186,6 +3198,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, { pd_data pd; pd.rhs = SSA_VAL (def_rhs); + pd.rhs_off = 0; pd.offset = offset2i; pd.size = size2i; return data->push_partial_def (pd, ao_ref_alias_set (&lhs_ref), @@ -3195,6 +3208,133 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, } } + /* 4b) Assignment done via one of the vectorizer internal store + functions where we may be able to access pieces from or we can + combine to a larger entity. */ + else if (known_eq (ref->size, maxsize) + && is_gimple_reg_type (vr->type) + && !reverse_storage_order_for_component_p (vr->operands) + && !contains_storage_order_barrier_p (vr->operands) + && is_gimple_call (def_stmt) + && gimple_call_internal_p (def_stmt) + && internal_store_fn_p (gimple_call_internal_fn (def_stmt))) + { + gcall *call = as_a (def_stmt); + internal_fn fn = gimple_call_internal_fn (call); + tree def_rhs = gimple_call_arg (call, + internal_fn_stored_value_index (fn)); + def_rhs = vn_valueize (def_rhs); + if (TREE_CODE (def_rhs) != VECTOR_CST) + return (void *)-1; + + tree mask = NULL_TREE, len = NULL_TREE, bias = NULL_TREE; + switch (fn) + { + case IFN_MASK_STORE: + mask = gimple_call_arg (call, internal_fn_mask_index (fn)); + mask = vn_valueize (mask); + if (TREE_CODE (mask) != VECTOR_CST) + return (void *)-1; + break; + case IFN_LEN_STORE: + len = gimple_call_arg (call, 2); + bias = gimple_call_arg (call, 4); + if (!tree_fits_uhwi_p (len) || !tree_fits_shwi_p (bias)) + return (void *)-1; + break; + default: + return (void *)-1; + } + ao_ref_init_from_ptr_and_size (&lhs_ref, + vn_valueize (gimple_call_arg (call, 0)), + TYPE_SIZE_UNIT (TREE_TYPE (def_rhs))); + tree base2; + poly_int64 offset2, size2, maxsize2; + HOST_WIDE_INT offset2i, size2i, offseti; + base2 = ao_ref_base (&lhs_ref); + offset2 = lhs_ref.offset; + size2 = lhs_ref.size; + maxsize2 = lhs_ref.max_size; + if (known_size_p (maxsize2) + && known_eq (maxsize2, size2) + && adjust_offsets_for_equal_base_address (base, &offset, + base2, &offset2) + && maxsize.is_constant (&maxsizei) + && offset.is_constant (&offseti) + && offset2.is_constant (&offset2i) + && size2.is_constant (&size2i)) + { + if (!ranges_maybe_overlap_p (offset, maxsize, offset2, size2)) + /* Poor-mans disambiguation. */ + return NULL; + else if (ranges_known_overlap_p (offset, maxsize, offset2, size2)) + { + pd_data pd; + pd.rhs = def_rhs; + tree aa = gimple_call_arg (call, 1); + alias_set_type set = get_deref_alias_set (TREE_TYPE (aa)); + tree vectype = TREE_TYPE (def_rhs); + unsigned HOST_WIDE_INT elsz + = tree_to_uhwi (TYPE_SIZE (TREE_TYPE (vectype))); + if (mask) + { + HOST_WIDE_INT start = 0, len = 0; + unsigned mask_idx = 0; + do + { + if (integer_zerop (VECTOR_CST_ELT (mask, mask_idx))) + { + if (len != 0) + { + pd.rhs_off = start; + pd.offset = offset2i + start; + pd.size = len; + if (ranges_known_overlap_p + (offset, maxsize, pd.offset, pd.size)) + { + void *res = data->push_partial_def + (pd, set, set, offseti, maxsizei); + if (res != NULL) + return res; + } + } + start = (mask_idx + 1) * elsz; + len = 0; + } + else + len += elsz; + mask_idx++; + } + while (known_lt (mask_idx, TYPE_VECTOR_SUBPARTS (vectype))); + if (len != 0) + { + pd.rhs_off = start; + pd.offset = offset2i + start; + pd.size = len; + if (ranges_known_overlap_p (offset, maxsize, + pd.offset, pd.size)) + return data->push_partial_def (pd, set, set, + offseti, maxsizei); + } + } + else if (fn == IFN_LEN_STORE) + { + pd.rhs_off = 0; + pd.offset = offset2i; + pd.size = (tree_to_uhwi (len) + + -tree_to_shwi (bias)) * BITS_PER_UNIT; + if (ranges_known_overlap_p (offset, maxsize, + pd.offset, pd.size)) + return data->push_partial_def (pd, set, set, + offseti, maxsizei); + } + else + gcc_unreachable (); + return NULL; + } + } + } + /* 5) For aggregate copies translate the reference through them if the copy kills ref. */ else if (data->vn_walk_kind == VN_WALKREWRITE @@ -3327,6 +3467,7 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_, { pd_data pd; pd.rhs = val; + pd.rhs_off = 0; pd.offset = 0; pd.size = maxsizei; return data->push_partial_def (pd, ao_ref_alias_set (&lhs_ref), >>From hjl@sc.intel.com Thu Jul 21 23:51:43 2022 Return-Path: X-Original-To: gcc-regression@gcc.gnu.org Delivered-To: gcc-regression@gcc.gnu.org Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by sourceware.org (Postfix) with ESMTPS id E526D3835839 for ; Thu, 21 Jul 2022 23:51:42 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org E526D3835839 X-IronPort-AV: E=McAfee;i="6400,9594,10415"; a="287202289" X-IronPort-AV: E=Sophos;i="5.93,184,1654585200"; d="scan'208";a="287202289" Received: from fmsmga007.fm.intel.com ([10.253.24.52]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 21 Jul 2022 16:51:25 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.93,184,1654585200"; d="scan'208";a="602490863" Received: from scymds01.sc.intel.com ([10.148.94.138]) by fmsmga007.fm.intel.com with ESMTP; 21 Jul 2022 16:51:25 -0700 Received: from gnu-clx-1.sc.intel.com (gnu-clx-1.sc.intel.com [172.25.70.216]) by scymds01.sc.intel.com with ESMTP id 26LNpPsA007953; Thu, 21 Jul 2022 16:51:25 -0700 Received: by gnu-clx-1.sc.intel.com (Postfix, from userid 1000) id 4F8A93E001E; Thu, 21 Jul 2022 16:51:25 -0700 (PDT) Date: Thu, 21 Jul 2022 16:51:25 -0700 To: skpgkp2@gmail.com, hjl.tools@gmail.com, gcc-regression@gcc.gnu.org Subject: Regressions on native/master at commit r13-1786 vs commit r13-1780 on Linux/x86_64 User-Agent: Heirloom mailx 12.5 7/5/10 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: <20220721235125.4F8A93E001E@gnu-clx-1.sc.intel.com> From: "H. J. Lu" X-Spam-Status: No, score=-3459.2 required=5.0 tests=BAYES_00, KAM_DMARC_STATUS, KAM_LAZY_DOMAIN_SECURITY, KAM_NUMSUBJECT, SPF_HELO_NONE, SPF_NONE, TXREP autolearn=no autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: gcc-regression@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-regression mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 21 Jul 2022 23:51:44 -0000 New failures: FAIL: gcc.dg/analyzer/stdarg-3.c (test for excess errors) FAIL: gcc.dg/analyzer/stdarg-3.c (test for excess errors) FAIL: gcc.dg/analyzer/stdarg-3.c (test for excess errors) New passes: