[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PATCH v5 45/54] tcg/ppc: Use atom_and_align_for_opc
From: |
Richard Henderson |
Subject: |
[PATCH v5 45/54] tcg/ppc: Use atom_and_align_for_opc |
Date: |
Mon, 15 May 2023 07:33:04 -0700 |
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/ppc/tcg-target.c.inc | 19 ++++++++++++++++++-
1 file changed, 18 insertions(+), 1 deletion(-)
diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc
index b62a163014..b5c49895f3 100644
--- a/tcg/ppc/tcg-target.c.inc
+++ b/tcg/ppc/tcg-target.c.inc
@@ -2015,6 +2015,7 @@ static bool tcg_out_qemu_st_slow_path(TCGContext *s,
TCGLabelQemuLdst *lb)
typedef struct {
TCGReg base;
TCGReg index;
+ TCGAtomAlign aa;
} HostAddress;
bool tcg_target_has_memory_bswap(MemOp memop)
@@ -2034,7 +2035,23 @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext
*s, HostAddress *h,
{
TCGLabelQemuLdst *ldst = NULL;
MemOp opc = get_memop(oi);
- unsigned a_bits = get_alignment_bits(opc);
+ MemOp a_bits;
+
+ /*
+ * Book II, Section 1.4, Single-Copy Atomicity, specifies:
+ *
+ * Before 3.0, "An access that is not atomic is performed as a set of
+ * smaller disjoint atomic accesses. In general, the number and alignment
+ * of these accesses are implementation-dependent." Thus MO_ATOM_IFALIGN.
+ *
+ * As of 3.0, "the non-atomic access is performed as described in
+ * the corresponding list", which matches MO_ATOM_SUBALIGN.
+ */
+ h->aa = atom_and_align_for_opc(s, opc,
+ have_isa_3_00 ? MO_ATOM_SUBALIGN
+ : MO_ATOM_IFALIGN,
+ false);
+ a_bits = h->aa.align;
#ifdef CONFIG_SOFTMMU
int mem_index = get_mmuidx(oi);
--
2.34.1
- Re: [PATCH v5 31/54] tcg/loongarch64: Check the host supports unaligned accesses, (continued)
- [PATCH v5 35/54] tcg: Add INDEX_op_qemu_{ld,st}_i128, Richard Henderson, 2023/05/15
- [PATCH v5 34/54] tcg: Introduce tcg_target_has_memory_bswap, Richard Henderson, 2023/05/15
- [PATCH v5 30/54] accel/tcg: Remove helper_unaligned_{ld,st}, Richard Henderson, 2023/05/15
- [PATCH v5 42/54] tcg/arm: Use atom_and_align_for_opc, Richard Henderson, 2023/05/15
- [PATCH v5 40/54] tcg/i386: Use atom_and_align_for_opc, Richard Henderson, 2023/05/15
- [PATCH v5 43/54] tcg/loongarch64: Use atom_and_align_for_opc, Richard Henderson, 2023/05/15
- [PATCH v5 41/54] tcg/aarch64: Use atom_and_align_for_opc, Richard Henderson, 2023/05/15
- [PATCH v5 44/54] tcg/mips: Use atom_and_align_for_opc, Richard Henderson, 2023/05/15
- [PATCH v5 45/54] tcg/ppc: Use atom_and_align_for_opc,
Richard Henderson <=
- [PATCH v5 50/54] tcg/i386: Support 128-bit load/store with have_atomic16, Richard Henderson, 2023/05/15
- [PATCH v5 51/54] tcg/aarch64: Rename temporaries, Richard Henderson, 2023/05/15
- [PATCH v5 47/54] tcg/s390x: Use atom_and_align_for_opc, Richard Henderson, 2023/05/15
- [PATCH v5 46/54] tcg/riscv: Use atom_and_align_for_opc, Richard Henderson, 2023/05/15
- [PATCH v5 48/54] tcg/sparc64: Use atom_and_align_for_opc, Richard Henderson, 2023/05/15
- [PATCH v5 49/54] tcg/i386: Honor 64-bit atomicity in 32-bit mode, Richard Henderson, 2023/05/15
- [PATCH v5 54/54] tcg/s390x: Support 128-bit load/store, Richard Henderson, 2023/05/15