diff mbox series

[v4,16/16] accel/tcg: Add aarch64 store_atom_insert_al16

Message ID 20230526002334.1760495-17-richard.henderson@linaro.org
State Superseded
Headers show
Series tcg: Improvements to atomic128 | expand

Commit Message

Richard Henderson May 26, 2023, 12:23 a.m. UTC 
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 host/include/aarch64/host/store-insert-al16.h | 47 +++++++++++++++++++
 1 file changed, 47 insertions(+)
 create mode 100644 host/include/aarch64/host/store-insert-al16.h

Comments

Peter Maydell May 30, 2023, 2:04 p.m. UTC | #1 
On Fri, 26 May 2023 at 01:24, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM
diff mbox series

Patch

diff --git a/host/include/aarch64/host/store-insert-al16.h b/host/include/aarch64/host/store-insert-al16.h
new file mode 100644
index 0000000000..1943155bc6
--- /dev/null
+++ b/host/include/aarch64/host/store-insert-al16.h
@@ -0,0 +1,47 @@ 
+/*
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ * Atomic store insert into 128-bit, AArch64 version.
+ *
+ * Copyright (C) 2023 Linaro, Ltd.
+ */
+
+#ifndef AARCH64_STORE_INSERT_AL16_H
+#define AARCH64_STORE_INSERT_AL16_H
+
+/**
+ * store_atom_insert_al16:
+ * @p: host address
+ * @val: shifted value to store
+ * @msk: mask for value to store
+ *
+ * Atomically store @val to @p masked by @msk.
+ */
+static inline void ATTRIBUTE_ATOMIC128_OPT
+store_atom_insert_al16(Int128 *ps, Int128 val, Int128 msk)
+{
+    /*
+     * GCC only implements __sync* primitives for int128 on aarch64.
+     * We can do better without the barriers, and integrating the
+     * arithmetic into the load-exclusive/store-conditional pair.
+     */
+    uint64_t tl, th, vl, vh, ml, mh;
+    uint32_t fail;
+
+    qemu_build_assert(!HOST_BIG_ENDIAN);
+    vl = int128_getlo(val);
+    vh = int128_gethi(val);
+    ml = int128_getlo(msk);
+    mh = int128_gethi(msk);
+
+    asm("0: ldxp %[l], %[h], %[mem]\n\t"
+        "bic %[l], %[l], %[ml]\n\t"
+        "bic %[h], %[h], %[mh]\n\t"
+        "orr %[l], %[l], %[vl]\n\t"
+        "orr %[h], %[h], %[vh]\n\t"
+        "stxp %w[f], %[l], %[h], %[mem]\n\t"
+        "cbnz %w[f], 0b\n"
+        : [mem] "+Q"(*ps), [f] "=&r"(fail), [l] "=&r"(tl), [h] "=&r"(th)
+        : [vl] "r"(vl), [vh] "r"(vh), [ml] "r"(ml), [mh] "r"(mh));
+}
+
+#endif /* AARCH64_STORE_INSERT_AL16_H */