This patch resolves PR middle-end/124637, a wrong code regression when
passing a struct as a register on big-endian targets. On big-endian
targets, store_constructor fills fields from the most significant bits,
so for structs narrower than word size, any padding is incorrectly
placed in the least significant bytes. This issue is fixed (on
affected targets) by using a (unsigned) right shift on the value
determined by store_constructor to correctly align the structure in
the least significant bytes, and place the padding in the high bits.
Many thanks to Manjunath Matti for testing this patch on real hardware,
and Drea Pinski for reviewing/approving it. The new test case may be
a little fragile, but currently "works for me". Please feel free to
tweak it for powerpc variants/environments I've not consider/encountered.
2026-05-08 Roger Sayle <[email protected]>
gcc/ChangeLog
PR middle-end/124637
* calls.cc (load_register_parameters): If using store_constructor
to place a constant structure in a register, use a right shift to
align the structure/padding if required on big-endian targets.
gcc/testsuite/ChangeLog
PR middle-end/124637
* gcc.target/powerpc/pr124637.c: New test case.
Many thanks,
Roger
--
diff --git a/gcc/calls.cc b/gcc/calls.cc
index 4cdc2361a4b..d491a414611 100644
--- a/gcc/calls.cc
+++ b/gcc/calls.cc
@@ -2252,10 +2252,20 @@ load_register_parameters (struct arg_data *args, int
num_actuals,
&& !TREE_SIDE_EFFECTS (tree_value)
&& immediate_const_ctor_p (DECL_INITIAL (tree_value)))
{
+ HOST_WIDE_INT size = int_expr_size (DECL_INITIAL (tree_value));
rtx target = gen_reg_rtx (word_mode);
store_constructor (DECL_INITIAL (tree_value), target, 0,
- int_expr_size (DECL_INITIAL (tree_value)),
- false);
+ size, false);
+ /* On big-endian targets, store_constructor places fields
+ from the MSB, which places any padding bits in the least
+ significant bytes. If required, use a logical right shift
+ to place things where expected in a register parameter. */
+ if (BYTES_BIG_ENDIAN
+ && size < UNITS_PER_WORD
+ && args[i].locate.where_pad == PAD_DOWNWARD)
+ target = expand_shift (RSHIFT_EXPR, word_mode, target,
+ (UNITS_PER_WORD - size) * BITS_PER_UNIT,
+ NULL_RTX, 1);
reg = gen_rtx_REG (word_mode, REGNO (reg));
emit_move_insn (reg, target);
}
diff --git a/gcc/testsuite/gcc.target/powerpc/pr124637.c
b/gcc/testsuite/gcc.target/powerpc/pr124637.c
new file mode 100644
index 00000000000..f38c7fa7126
--- /dev/null
+++ b/gcc/testsuite/gcc.target/powerpc/pr124637.c
@@ -0,0 +1,26 @@
+/* { dg-do compile { target { powerpc64-*-* } } } */
+/* { dg-options "-O2 -m64 -mbig-endian" } */
+
+typedef struct {
+ unsigned short a;
+ unsigned short b;
+ unsigned short c;
+} S_t;
+
+static const S_t m1 = { 0x20, 1, 1 };
+
+void ext(S_t r);
+void foo() { ext(m1); }
+
+/* Apologies in advance that these may be a little fragile. */
+
+/* { dg-final { scan-assembler "lis 3,0x20" } } */
+/* { dg-final { scan-assembler "ori 3,3,0x1" } } */
+/* { dg-final { scan-assembler "sldi 3,3,16" } } */
+/* { dg-final { scan-assembler "ori 3,3,0x1" } } */
+
+/* { dg-final { scan-assembler-not "lis 9,0x20" } } */
+/* { dg-final { scan-assembler-not "ori 9,9,0x1" } } */
+/* { dg-final { scan-assembler-not "lis 3,0x1" } } */
+/* { dg-final { scan-assembler-not "rldimi 3,9,32,0" } } */
+
