On 25/8/25 01:32, Richard Henderson wrote:
On 8/25/25 08:27, Edgar E. Iglesias wrote:
From: "Edgar E. Iglesias" <edgar.igles...@amd.com>

Rename and reorder function args to better match with spec
and pseudo code.

No functional change.

Signed-off-by: Edgar E. Iglesias <edgar.igles...@amd.com>
---
  target/microblaze/op_helper.c | 18 +++++++++---------
  target/microblaze/translate.c | 12 ++----------
  2 files changed, 11 insertions(+), 19 deletions(-)

diff --git a/target/microblaze/op_helper.c b/target/microblaze/ op_helper.c
index 470526ee92..092977b3e1 100644
--- a/target/microblaze/op_helper.c
+++ b/target/microblaze/op_helper.c
@@ -69,9 +69,9 @@ void helper_raise_exception(CPUMBState *env, uint32_t index)
      cpu_loop_exit(cs);
  }
-static bool check_divz(CPUMBState *env, uint32_t b, uintptr_t ra)
+static bool check_divz(CPUMBState *env, uint32_t divisor, uintptr_t pc)

The name GETPC notwithstanding, I don't think ra -> pc is a good rename.

PC often gets confused about whether that's a host or guest value.
RA (return address) is less so; it gets used often in core tcg.

If you really don't like RA, then perhaps "retaddr" (also with quite a bit of usage) or "unwind_pc" (no existing usage, but very descriptive).

+1 for 'unwind_pc'.

While thinking more about it, I noticed we don't document the
prototypes in "exec/cpu-common.h". Doing so we could rename
s/pc/unwind_pc/ and document @unwind_pc. Also these helpers
would be better declared in a TCG specific header. Anyway,
note for Richard, orthogonal to this patch ;)


Reply via email to