From: Xi Ruoyao <xry111@xry111.site>
Date: Sat, 1 Mar 2025 03:46:49 +0000 (+0800)
Subject: LoongArch: Don't expand atomic_fetch_sub_{hi, qi} to LL-SC loop if -mlam-bh
X-Git-Tag: basepoints/gcc-17~5548
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=6d67bbce340351cfe7f30f1b17cba7635f18c8d5;p=thirdparty%2Fgcc.git

LoongArch: Don't expand atomic_fetch_sub_{hi, qi} to LL-SC loop if -mlam-bh

With -mlam-bh, we should negate the addend first, and use an amadd
instruction.  Disabling the expander makes the compiler do it correctly.

gcc/ChangeLog:

	* config/loongarch/sync.md (atomic_fetch_sub<SHORT:mode>):
	Disable if ISA_HAS_LAM_BH.
---

diff --git a/gcc/config/loongarch/sync.md b/gcc/config/loongarch/sync.md
index b6acfff3a61..60038258f29 100644
--- a/gcc/config/loongarch/sync.md
+++ b/gcc/config/loongarch/sync.md
@@ -661,7 +661,7 @@
 			(match_operand:SHORT 2 "reg_or_0_operand" "rJ"))
 	   (match_operand:SI 3 "const_int_operand")] ;; model
 	 UNSPEC_SYNC_OLD_OP))]
-  ""
+  "!ISA_HAS_LAM_BH"
 {
   union loongarch_gen_fn_ptrs generator;
   generator.fn_7 = gen_atomic_cas_value_sub_7_si;