From b6e1373bd34aebbb512a03ea9a4e3c7acd955382 Mon Sep 17 00:00:00 2001 From: Roger Sayle Date: Wed, 8 Jun 2022 20:43:03 +0100 Subject: [PATCH] PR middle-end/105874: Use EXPAND_MEMORY to fix ada bootstrap. Many thanks to Tamar Christina for filing PR middle-end/105874 indicating that SPECcpu 2017's Leela is failing on x86_64 due to a miscompilation of FastBoard::is_eye. This function is much smaller and easier to work with than my previous hunt for the cause of the Ada bootstrap failures due to miscompilation somewhere in GCC (or one of the 131 places that the problematic form of optimization triggers during an ada bootstrap). It turns out the source of the miscompilation introduced by my recent patch is the distinction (during RTL expansion) of l-values and r-values. According to the documentation above expand_modifier, EXPAND_MEMORY should be used for lvalues (when a memory is required), and EXPAND_NORMAL for rvalues when a constant is permissible. In what I'd like to consider a latent bug, the recursive call to expand_expr_real on line 11188 of expr.cc, in the case handling ARRAY_REF, COMPONENT_REF, BIT_FIELD_REF and ARRARY_RANGE_REF was passing EXPAND_NORMAL when it really required (the semantics of) EXPAND_MEMORY. All the time that VAR_DECLs were being returned as memory this was fine, but as soon as we're able to optimize sort arrays into immediate constants, bad things happen. In the test case from Leela, we notice that the array s_eyemask always has DImode constant value { 4, 64 }, which is useful as an rvalue, but not when we need to index it as an lvalue, as in s_eyemask[color]. This also explains why everything being accepted by immediate_const_ctor_p (during an ada bootstrap) looks reasonable, what's incorrect is that we don't know how these structs/arrays are to be used. The fix is to ensure that we call expand_expr with EXPAND_MEMORY when processing the VAR_DECL's returned by get_inner_reference. 2022-06-08 Roger Sayle gcc/ChangeLog PR middle-end/105874 * expr.cc (expand_expr_real_1) : New local variable tem_modifier for calculating the expand_modifier enum to use for expanding tem. If tem is a VAR_DECL, use EXPAND_MEMORY. gcc/testsuite/ChangeLog PR middle-end/105874 * g++.dg/opt/pr105874.C: New test case. --- gcc/expr.cc | 11 ++++++++--- gcc/testsuite/g++.dg/opt/pr105874.C | 30 ++++++++++++++++++++++++++++++ 2 files changed, 38 insertions(+), 3 deletions(-) create mode 100644 gcc/testsuite/g++.dg/opt/pr105874.C diff --git a/gcc/expr.cc b/gcc/expr.cc index 85cb414..c37a999 100644 --- a/gcc/expr.cc +++ b/gcc/expr.cc @@ -11181,6 +11181,13 @@ expand_expr_real_1 (tree exp, rtx target, machine_mode tmode, infinitely recurse. */ gcc_assert (tem != exp); + /* If tem is a VAR_DECL, we need a memory reference. */ + enum expand_modifier tem_modifier = modifier; + if (tem_modifier == EXPAND_SUM) + tem_modifier = EXPAND_NORMAL; + if (TREE_CODE (tem) == VAR_DECL) + tem_modifier = EXPAND_MEMORY; + /* If TEM's type is a union of variable size, pass TARGET to the inner computation, since it will need a temporary and TARGET is known to have to do. This occurs in unchecked conversion in Ada. */ @@ -11192,9 +11199,7 @@ expand_expr_real_1 (tree exp, rtx target, machine_mode tmode, != INTEGER_CST) && modifier != EXPAND_STACK_PARM ? target : NULL_RTX), - VOIDmode, - modifier == EXPAND_SUM ? EXPAND_NORMAL : modifier, - NULL, true); + VOIDmode, tem_modifier, NULL, true); /* If the field has a mode, we want to access it in the field's mode, not the computed mode. diff --git a/gcc/testsuite/g++.dg/opt/pr105874.C b/gcc/testsuite/g++.dg/opt/pr105874.C new file mode 100644 index 0000000..58699a6 --- /dev/null +++ b/gcc/testsuite/g++.dg/opt/pr105874.C @@ -0,0 +1,30 @@ +/* { dg-do compile } */ +/* { dg-options "-O2 -std=c++11" } */ +#include + +static constexpr int NBR_SHIFT = 4; + +static constexpr int MAXBOARDSIZE = 25; + +static constexpr int MAXSQ = ((MAXBOARDSIZE + 2) * (MAXBOARDSIZE + 2)); + +enum square_t : char { + BLACK = 0, WHITE = 1, EMPTY = 2, INVAL = 3 + }; + +const std::array s_eyemask = { + 4 * (1 << (NBR_SHIFT * BLACK)), + 4 * (1 << (NBR_SHIFT * WHITE)) +}; + +/* counts of neighboring stones */ +std::array m_neighbours; + +int is_eye(const int color, const int i) { + /* check for 4 neighbors of the same color */ + int ownsurrounded = (m_neighbours[i] & s_eyemask[color]); + + return ownsurrounded; +} + +/* { dg-final { scan-assembler "s_eyemask" } } */ -- 2.7.4