My sincere apologies to everyone (again). As diagnosed by Jakub Jelinek, my recent patch to fold the signedness of LSHIFT_EXPR needs to be careful not to attempt transforming a left shift in an integer type into an invalid left shift of a pointer type.
This patch has been committed as obvious, after testing that a "make bootstrap" on x86_64-pc-linux-gnu is unaffected [make -k check is underway]. Sorry (yet) again. 2021-08-23 Roger Sayle <ro...@nextmovesoftware.com> Jakub Jelinek <ja...@redhat.com> gcc/ChangeLog PR middle-end/102029 * match.pd (shift transformations): Add an additional check for !POINTER_TYPE_P in the recently added left shift transformation. gcc/testsuite/ChangeLog PR middle-end/102029 * gcc.dg/fold-convlshift-3.c: New test case. Roger --
diff --git a/gcc/match.pd b/gcc/match.pd index 978a1b0..e5bbb12 100644 --- a/gcc/match.pd +++ b/gcc/match.pd @@ -3389,7 +3389,9 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT) the form that minimizes the number of conversions. */ (simplify (convert (lshift:s@0 (convert:s@1 @2) INTEGER_CST@3)) - (if (tree_nop_conversion_p (type, TREE_TYPE (@0)) + (if (INTEGRAL_TYPE_P (type) + && !POINTER_TYPE_P (type) + && tree_nop_conversion_p (type, TREE_TYPE (@0)) && INTEGRAL_TYPE_P (TREE_TYPE (@2)) && TYPE_PRECISION (TREE_TYPE (@2)) <= TYPE_PRECISION (type)) (lshift (convert @2) @3)))
/* PR middle-end/102029 */ /* { dg-do compile } */ /* { dg-options "-O2" } */ int * foo (const __PTRDIFF_TYPE__ l) { return (int *) (l << 2); }