https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113100
Kewen Lin <linkw at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |linkw at gcc dot gnu.org
Status|UNCONFIRMED |NEW
Ever confirmed|0 |1
Last reconfirmed| |2023-12-21
--- Comment #2 from Kewen Lin <linkw at gcc dot gnu.org> ---
Confirmed, but it needs an explicit cpu type like -mcpu=power9 for
reproduction, otherwise it could pass on power10 as it can work with pcrel (so
no toc base r2 needed). The change can extend the end of scrubbing, it cleans
the saved toc base unexpectedly.
I noticed that there is one macro SPARC_STACK_BOUNDARY_HACK, which aims to
indicate this SPARC64 specific behavior. Could we leverage this macro (guarded
the biasing with it)? like:
diff --git a/gcc/builtins.cc b/gcc/builtins.cc
index 125ea158ebf..9bad1e962b4 100644
--- a/gcc/builtins.cc
+++ b/gcc/builtins.cc
@@ -5450,6 +5450,7 @@ expand_builtin_stack_address ()
rtx ret = convert_to_mode (ptr_mode, copy_to_reg (stack_pointer_rtx),
STACK_UNSIGNED);
+#ifdef SPARC_STACK_BOUNDARY_HACK
/* Unbias the stack pointer, bringing it to the boundary between the
stack area claimed by the active function calling this builtin,
and stack ranges that could get clobbered if it called another
@@ -5476,7 +5477,9 @@ expand_builtin_stack_address ()
(caller) function's active area as well, whereas those pushed or
allocated temporarily for a call are regarded as part of the
callee's stack range, rather than the caller's. */
- ret = plus_constant (ptr_mode, ret, STACK_POINTER_OFFSET);
+ if (SPARC_STACK_BOUNDARY_HACK)
+ ret = plus_constant (ptr_mode, ret, STACK_POINTER_OFFSET);
+#endif
return force_reg (ptr_mode, ret);
}
