Yarn: Implement fiber context switching for Windows, macOS, Linux and Android.

Windows uses OSFiber_windows.hpp.
macOS and Linux uses OSFiber_asm_* or OSFiber_ucontext.hpp
Android has no support for makecontext / swapcontext, and so relies on the OSFiber_asm_* implementations.

Unlike makecontext, swapcontext or the Windows fiber APIs, these assembly implementations *do not* save or restore signal masks, floating-point control or status registers, FS and GS segment registers, thread-local storage state nor any SIMD registers.

This should not be a problem as the yarn scheduler requires fibers to be executed on a single thread.

Bug: b/139010488
Change-Id: I5f890ffe4f069361a7eeae64567a77be3a4ba008
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/34771
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

src/Yarn/OSFiber_asm_aarch64.c

diff --git a/src/Yarn/OSFiber_asm_aarch64.c b/src/Yarn/OSFiber_asm_aarch64.c
new file mode 100644
index 0000000..7b09c58
--- /dev/null
+++ b/src/Yarn/OSFiber_asm_aarch64.c
@@ -0,0 +1,33 @@
+// Copyright 2019 The SwiftShader Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//    http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#if defined(__aarch64__)
+
+#include "OSFiber_asm_aarch64.h"
+
+void yarn_fiber_trampoline(void(*target)(void*), void* arg)
+{
+    target(arg);
+}
+
+void yarn_fiber_set_target(struct yarn_fiber_context* ctx, void* stack, uint32_t stack_size, void(*target)(void*), void* arg)
+{
+    uintptr_t* stack_top = (uintptr_t*)((uint8_t*)(stack) + stack_size);
+    ctx->LR = (uintptr_t)&yarn_fiber_trampoline;
+    ctx->r0 = (uintptr_t)target;
+    ctx->r1 = (uintptr_t)arg;
+    ctx->SP = ((uintptr_t)stack_top) & ~(uintptr_t)15;
+}
+
+#endif // defined(__aarch64__)