[RFC PATCH v1 3/4] aarch64: Support for Static Trampolines

[madvenka@linux.microsoft.com]

From: "Madhavan T. Venkataraman" <madvenka@linux.microsoft.com>

	- Define the arch-specific initialization function ffi_tramp_arch ()
	  that returns trampoline size information to common code.

	- Define the trampoline code and data mapping sizes.

	- Introduce a tiny amount of code at the beginning of each ABI
	  handler to retrieve the information saved by the trampoline on
	  stack.

	- Define the trampoline code table statically.

	- Call ffi_closure_tramp_init () to initialize static trampoline
	  parameters from ffi_prep_closure_loc ().

Signed-off-by: Madhavan T. Venkataraman <madvenka@linux.microsoft.com>
---
 src/aarch64/ffi.c      | 16 +++++++++++++++
 src/aarch64/internal.h | 10 +++++++++
 src/aarch64/sysv.S     | 46 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 72 insertions(+)

diff --git a/src/aarch64/ffi.c b/src/aarch64/ffi.c
index ef09f4d..64d36b6 100644
--- a/src/aarch64/ffi.c
+++ b/src/aarch64/ffi.c
@@ -817,6 +817,9 @@ ffi_prep_closure_loc (ffi_closure *closure,
   };
   char *tramp = closure->tramp;
   
+  if (ffi_closure_tramp_set_parms (closure, start))
+    goto out;
+
   memcpy (tramp, trampoline, sizeof(trampoline));
   
   *(UINT64 *)(tramp + 16) = (uintptr_t)start;
@@ -832,6 +835,7 @@ ffi_prep_closure_loc (ffi_closure *closure,
   unsigned char *tramp_code = ffi_data_to_code_pointer (tramp);
   #endif
   ffi_clear_cache (tramp_code, tramp_code + FFI_TRAMPOLINE_SIZE);
+out:
 #endif
 
   closure->cif = cif;
@@ -1022,4 +1026,16 @@ ffi_closure_SYSV_inner (ffi_cif *cif,
   return flags;
 }
 
+#if defined(FFI_EXEC_STATIC_TRAMP)
+void *
+ffi_tramp_arch (size_t *tramp_size, size_t *map_size)
+{
+  extern void *trampoline_code_table;
+
+  *tramp_size = AARCH64_TRAMP_SIZE;
+  *map_size = AARCH64_TRAMP_MAP_SIZE;
+  return &trampoline_code_table;
+}
+#endif
+
 #endif /* (__aarch64__) || defined(__arm64__)|| defined (_M_ARM64)*/
diff --git a/src/aarch64/internal.h b/src/aarch64/internal.h
index 3d4d035..de55755 100644
--- a/src/aarch64/internal.h
+++ b/src/aarch64/internal.h
@@ -66,3 +66,13 @@ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.  */
 #define N_X_ARG_REG		8
 #define N_V_ARG_REG		8
 #define CALL_CONTEXT_SIZE	(N_V_ARG_REG * 16 + N_X_ARG_REG * 8)
+
+#if defined(FFI_EXEC_STATIC_TRAMP)
+/*
+ * For the trampoline code table mapping, a mapping size of 16K is chosen to
+ * cover the base page sizes of 4K and 16K.
+ */
+#define AARCH64_TRAMP_MAP_SHIFT	14
+#define AARCH64_TRAMP_MAP_SIZE	(1 << AARCH64_TRAMP_MAP_SHIFT)
+#define AARCH64_TRAMP_SIZE	32
+#endif
diff --git a/src/aarch64/sysv.S b/src/aarch64/sysv.S
index b720a92..fece3f9 100644
--- a/src/aarch64/sysv.S
+++ b/src/aarch64/sysv.S
@@ -232,6 +232,10 @@ CNAME(ffi_call_SYSV):
 	.align 4
 CNAME(ffi_closure_SYSV_V):
 	cfi_startproc
+#if defined(FFI_EXEC_STATIC_TRAMP)
+	ldr	x17, [sp, #8]
+	add	sp, sp, #16
+#endif
 	stp     x29, x30, [sp, #-ffi_closure_SYSV_FS]!
 	cfi_adjust_cfa_offset (ffi_closure_SYSV_FS)
 	cfi_rel_offset (x29, 0)
@@ -255,6 +259,10 @@ CNAME(ffi_closure_SYSV_V):
 	.align	4
 	cfi_startproc
 CNAME(ffi_closure_SYSV):
+#if defined(FFI_EXEC_STATIC_TRAMP)
+	ldr	x17, [sp, #8]
+	add	sp, sp, #16
+#endif
 	stp     x29, x30, [sp, #-ffi_closure_SYSV_FS]!
 	cfi_adjust_cfa_offset (ffi_closure_SYSV_FS)
 	cfi_rel_offset (x29, 0)
@@ -367,6 +375,44 @@ CNAME(ffi_closure_SYSV):
 	.size	CNAME(ffi_closure_SYSV), . - CNAME(ffi_closure_SYSV)
 #endif
 
+#if defined(FFI_EXEC_STATIC_TRAMP)
+/*
+ * The trampoline uses register x17. It saves the original value of x17 on
+ * the stack.
+ *
+ * The trampoline has two parameters - target code to jump to and data for
+ * the target code. The trampoline extracts the parameters from its parameter
+ * block (see tramp_table_map()). The trampoline saves the data address on
+ * the stack. Finally, it jumps to the target code.
+ *
+ * The target code can choose to:
+ *
+ * - restore the value of x17
+ * - load the data address in a register
+ * - restore the stack pointer to what it was when the trampoline was invoked.
+ */
+	.align	AARCH64_TRAMP_MAP_SHIFT
+CNAME(trampoline_code_table):
+	.rept	AARCH64_TRAMP_MAP_SIZE / AARCH64_TRAMP_SIZE
+	sub	sp, sp, #16		/* Make space on the stack */
+	str	x17, [sp]		/* Save x17 on stack */
+	adr	x17, #16376		/* Get data address */
+	ldr	x17, [x17]		/* Copy data into x17 */
+	str	x17, [sp, #8]		/* Save data on stack */
+	adr	x17, #16372		/* Get code address */
+	ldr	x17, [x17]		/* Load code address into x17 */
+	br	x17			/* Jump to code */
+	.endr
+
+	.globl CNAME(trampoline_code_table)
+	FFI_HIDDEN(CNAME(trampoline_code_table))
+#ifdef __ELF__
+	.type	CNAME(trampoline_code_table), #function
+	.size	CNAME(trampoline_code_table), . - CNAME(trampoline_code_table)
+#endif
+	.align	AARCH64_TRAMP_MAP_SHIFT
+#endif /* FFI_EXEC_STATIC_TRAMP */
+
 #if FFI_EXEC_TRAMPOLINE_TABLE
 
 #ifdef __MACH__
-- 
2.25.1