From: Max Filippov <jcmvb...@gmail.com>
s32c1i must load and store value with target endianness, not host.
This results in an infinite loop in atomic cmpxchg sequences when target
endianness doesn't match host endianness.
Fixes: 9fb40342d4b3 ("target/xtensa: support MTTCG")
Cc: qemu-sta...@nongnu.org
Signed-off-by: Max Filippov <jcmvb...@gmail.com>
Reviewed-by: Richard Henderson <richard.hender...@linaro.org>
(cherry picked from commit 7a54cfbcee8dd7aa87ce655a321b622107556326)
Signed-off-by: Michael Roth <mdr...@linux.vnet.ibm.com>
---
target/xtensa/translate.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/target/xtensa/translate.c b/target/xtensa/translate.c
index 25399058a0..c626583cd9 100644
--- a/target/xtensa/translate.c
+++ b/target/xtensa/translate.c
@@ -2288,7 +2288,7 @@ static void translate_s32c1i(DisasContext *dc, const
uint32_t arg[],
gen_load_store_alignment(dc, 2, addr, true);
gen_check_atomctl(dc, addr);
tcg_gen_atomic_cmpxchg_i32(cpu_R[arg[0]], addr, cpu_SR[SCOMPARE1],
- tmp, dc->cring, MO_32);
+ tmp, dc->cring, MO_TEUL);
tcg_temp_free(addr);
tcg_temp_free(tmp);
}
--
2.17.1
