Thanks to Zdenek Sojka for reporting PR target/106640 where an RTL checking build reveals a thinko in my recent patch to support TImode shifts/rotates in STV. My "senior moment" was to inappropriately use XINT where I should be using INTVAL of XEXP. I can't imagine what I was thinking. Corrected by the patch below, tested on x86_64-pc-linux-gnu with make bootstrap, both with and without --enable-checking=rtl, and regression tested, both with and without --target_board=unix{-m32}, with no new failures. Committed to mainline as obvious.
2022-08-17 Roger Sayle <ro...@nextmovesoftware.com> gcc/ChangeLog PR target/106640 * config/i386/i386-features.cc (timde_scalar_chain::compute_convert_gain): Replace incorrect use of XINT with INTVAL (XEXP (src, 1)). Roger --
diff --git a/gcc/config/i386/i386-features.cc b/gcc/config/i386/i386-features.cc index 821d8c7..d6bb66c 100644 --- a/gcc/config/i386/i386-features.cc +++ b/gcc/config/i386/i386-features.cc @@ -1248,7 +1248,7 @@ timode_scalar_chain::compute_convert_gain () case ASHIFT: case LSHIFTRT: /* See ix86_expand_v1ti_shift. */ - op1val = XINT (src, 1); + op1val = INTVAL (XEXP (src, 1)); if (optimize_insn_for_size_p ()) { if (op1val == 64 || op1val == 65) @@ -1282,7 +1282,7 @@ timode_scalar_chain::compute_convert_gain () case ASHIFTRT: /* See ix86_expand_v1ti_ashiftrt. */ - op1val = XINT (src, 1); + op1val = INTVAL (XEXP (src, 1)); if (optimize_insn_for_size_p ()) { if (op1val == 64 || op1val == 127) @@ -1355,7 +1355,7 @@ timode_scalar_chain::compute_convert_gain () case ROTATE: case ROTATERT: /* See ix86_expand_v1ti_rotate. */ - op1val = XINT (src, 1); + op1val = INTVAL (XEXP (src, 1)); if (optimize_insn_for_size_p ()) { scost = COSTS_N_BYTES (13);