Re: [PATCH] RISC-V: Fix testcase failed when default -mcmodel=medany

[Lehua Ding <lehua.ding@rivai.ai>]

Hi Robin,

> In general I'm fine with this small change of course, I just wonder if
> the testcase is not brittle anyway. From what I can tell the respective
> change is independent of the actual number of registers so maybe it's enough to
> not compare the fully body but just make sure the addis are not present?
> That way, the test could also work for -march=rv64 (which saves one
> register less anyway regardless of mcmodel - but the change still helps)
> or maybe even with instruction scheduling.  Would you mind checking this still?

I think you are rigth, I would like to remove the `-mcmodel=medany` option and
relax assert from `__riscv_save/restore_4` to `__riscv_save/restore_(3|4)` to let
this testcase not brittle on any -mcmodel.  Then I'm also going to add another
testcase (I dont known how to run -march=rv32imafc and -march=rv64imafc on
the same testcase) that uses -march=rv64imafc.

Removing scheduling option will result in a change in the order of the assert
assembly, and I don't feel like removing it because the order may be different for
different microarchitectures.

Best,
Lehua

V2 patch:

gcc/testsuite/ChangeLog:

	* gcc.target/riscv/stack_save_restore.c: Moved to...
	* gcc.target/riscv/stack_save_restore_2.c: ...here.
	* gcc.target/riscv/stack_save_restore_1.c: New test.

---
 .../gcc.target/riscv/stack_save_restore_1.c   | 40 +++++++++++++++++++
 ..._save_restore.c => stack_save_restore_2.c} |  6 +--
 2 files changed, 43 insertions(+), 3 deletions(-)
 create mode 100644 gcc/testsuite/gcc.target/riscv/stack_save_restore_1.c
 rename gcc/testsuite/gcc.target/riscv/{stack_save_restore.c => stack_save_restore_2.c} (90%)

diff --git a/gcc/testsuite/gcc.target/riscv/stack_save_restore_1.c b/gcc/testsuite/gcc.target/riscv/stack_save_restore_1.c
new file mode 100644
index 00000000000..255ce5f40c9
--- /dev/null
+++ b/gcc/testsuite/gcc.target/riscv/stack_save_restore_1.c
@@ -0,0 +1,40 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64imafc -mabi=lp64f -msave-restore -O2 -fno-schedule-insns -fno-schedule-insns2 -fno-unroll-loops -fno-peel-loops -fno-lto" } */
+/* { dg-final { check-function-bodies "**" "" } } */
+
+char my_getchar();
+float getf();
+
+/*
+** bar:
+**	call	t0,__riscv_save_(3|4)
+**	addi	sp,sp,-2032
+**	...
+**	li	t0,-12288
+**	add	sp,sp,t0
+**	...
+**	li	t0,12288
+**	add	sp,sp,t0
+**	...
+**	addi	sp,sp,2032
+**	tail	__riscv_restore_(3|4)
+*/
+int bar()
+{
+  float volatile farray[3568];
+
+  float sum = 0;
+  float f1 = getf();
+  float f2 = getf();
+  float f3 = getf();
+  float f4 = getf();
+
+  for (int i = 0; i < 3568; i++)
+  {
+    farray[i] = my_getchar() * 1.2;
+    sum += farray[i];
+  }
+
+  return sum + f1 + f2 + f3 + f4;
+}
+
diff --git a/gcc/testsuite/gcc.target/riscv/stack_save_restore.c b/gcc/testsuite/gcc.target/riscv/stack_save_restore_2.c
similarity index 90%
rename from gcc/testsuite/gcc.target/riscv/stack_save_restore.c
rename to gcc/testsuite/gcc.target/riscv/stack_save_restore_2.c
index 522e706cfbf..4ce5e0118a4 100644
--- a/gcc/testsuite/gcc.target/riscv/stack_save_restore.c
+++ b/gcc/testsuite/gcc.target/riscv/stack_save_restore_2.c
@@ -6,8 +6,8 @@ char my_getchar();
 float getf();
 
 /*
-**bar:
-**	call	t0,__riscv_save_4
+** bar:
+**	call	t0,__riscv_save_(3|4)
 **	addi	sp,sp,-2032
 **	...
 **	li	t0,-12288
@@ -17,7 +17,7 @@ float getf();
 **	add	sp,sp,t0
 **	...
 **	addi	sp,sp,2032
-**	tail	__riscv_restore_4
+**	tail	__riscv_restore_(3|4)
 */
 int bar()
 {
-- 
2.36.3