diff --git a/README.adoc b/README.adoc
index 9f8f139..41a5759 100644
--- a/README.adoc
+++ b/README.adoc
@@ -11647,7 +11647,7 @@ It is then a big copy paste for most other data instructions.
 Then, modify that program to make the assertion fail:
 
 ....
-ASSERT_EQ(%rax, $4)
+LKMC_ASSERT_EQ(%rax, $4)
 ....
 
 because 1 + 2 tends to equal 3 instead of 4.
@@ -11676,10 +11676,16 @@ and notice how the error message gives both:
 Other infrastructure sanity checks that you might want to look into include:
 
 * link:userland/arch/empty.S[]
-* `FAIL` tests
+* `LKMC_FAIL` tests
 ** link:userland/arch/fail.S[]
-* `ASSERT_MEMCMP` tests
+* `LKMC_ASSERT_EQ` tests
+** link:userland/arch/x86_64/lkmc_assert_eq_fail.S[]
+** link:userland/arch/arm/lkmc_assert_eq_fail.S[]
+** link:userland/arch/aarch64/lkmc_assert_eq_fail.S[]
+* `LKMC_ASSERT_MEMCMP` tests
 ** link:userland/arch/x86_64/lkmc_assert_memcmp_fail.S[]
+** link:userland/arch/arm/lkmc_assert_memcmp_fail.S[]
+** link:userland/arch/aarch64/lkmc_assert_memcmp_fail.S[]
 
 === Assembly registers
 
@@ -11799,13 +11805,13 @@ This allows using the C standard library for IO, which is very convenient and po
 
 It also exposes other non-IO functionality that is very convenient such as `memcmp`.
 
-The C standard library infrastructure is implemented in the following files:
+The C standard library infrastructure is implemented in the common userland / baremetal source files:
 
-* link:userland/arch/main.c[]
-* link:userland/arch/common.h[]
-* link:userland/arch/x86_64/common_arch.h[]
-* link:userland/arch/arm/common_arch.h[]
-* link:userland/arch/aarch64/common_arch.h[]
+* link:lkmc.c[]
+* link:lkmc.h[]
+* link:lkmc/aarch64.h[]
+* link:lkmc/arm.h[]
+* link:lkmc/x86_64.h[]
 
 ==== Freestanding programs
 
@@ -11914,23 +11920,51 @@ Questions about the C inline assembly examples:
 * x86_64: https://stackoverflow.com/questions/9506353/how-to-invoke-a-system-call-via-sysenter-in-inline-assembly/54956854#54956854
 * ARM: https://stackoverflow.com/questions/21729497/doing-a-syscall-without-libc-using-arm-inline-assembly
 
-=== Calling conventions
+=== Linux calling conventions
+
+Summary:
+
+[options="header"]
+|===
+|arch |arguments |return value |callee saved registers
+
+|x86_64
+|rdi, rsi, rdx, rcx, r8, r9, xmm0–7
+|rax, rdx
+|rbx, rbp, r12–r15
+
+|arm
+|r0-r3
+|r0-r3
+|r4-r11
+
+|aarch64
+|x0-x7
+|x0-x7
+|x19-x29
+
+|===
 
 ==== x86_64 calling convention
 
 Examples:
 
-* link:userland/arch/x86_64/common_arch.h[] `ENTRY` and `EXIT`
+* link:lkmc/x86_64.h[] `ENTRY` and `EXIT`
+
+Bibliography:
+
+* https://en.wikipedia.org/wiki/X86_calling_conventions#System_V_AMD64_ABI
+* https://stackoverflow.com/questions/18024672/what-registers-are-preserved-through-a-linux-x86-64-function-call/55207335#55207335
 
 ==== ARM calling convention
 
 Call C standard library functions from assembly and vice versa.
 
 * arm
-** link:userland/arch/arm/common_arch.h[] `ENTRY` and `EXIT`
+** link:lkmc/arm.h[] `ENTRY` and `EXIT`
 ** link:userland/arch/arm/linux/c_from_asm.S[]
 * aarch64
-** link:userland/arch/aarch64/common_arch.h[] `ENTRY` and `EXIT`
+** link:lkmc/aarch64.h[] `ENTRY` and `EXIT`
 ** link:userland/arch/aarch64/c/linux/asm_from_c.c[]
 
 ARM Architecture Procedure Call Standard (AAPCS) is the name that ARM Holdings gives to the calling convention.
@@ -12685,7 +12719,7 @@ Guaranteed undefined! Therefore raise illegal instruction signal. Used by GCC `_
 * link:userland/arch/arm/udf.S[]
 * link:userland/arch/aarch64/udf.S[]
 
-Why GNU GAS 2.29 does not have a mnemonic for it in A64 because it is very recent: shows in [[armarm8-db]] but not `ca`.
+Why GNU GAS 2.29 does not have a mnemonic for it in A64 because it is very recent: shows in <<armarm8-db>> but not `ca`.
 
 === ARM SIMD
 
diff --git a/build-userland b/build-userland
index 22c4c52..4d9a390 100755
--- a/build-userland
+++ b/build-userland
@@ -60,25 +60,6 @@ Default: build all examples that have their package dependencies met, e.g.:
             extra_deps=[self.env['common_h']],
             link=False,
         )
-        extra_obj_userland_asm = os.path.join(
-            build_dir,
-            'arch',
-            'main' + self.env['obj_ext']
-        )
-        extra_obj_userland_asm_relpath = os.path.join(
-            'arch',
-            'main' + self.env['c_ext']
-        )
-        self._build_one(
-            in_path=os.path.join(
-                self.env['userland_source_dir'],
-                extra_obj_userland_asm_relpath
-            ),
-            out_path=extra_obj_userland_asm,
-            cc_flags=cc_flags,
-            extra_deps=[self.env['common_h']],
-            link=False,
-        )
         with thread_pool.ThreadPool(
             self._build_one,
             nthreads=self.env['nproc'],
@@ -94,7 +75,6 @@ Default: build all examples that have their package dependencies met, e.g.:
                         my_thread_pool.submit({
                             'cc_flags': cc_flags,
                             'extra_objs_lkmc_common': [extra_obj_lkmc_common],
-                            'extra_objs_userland_asm': [extra_obj_userland_asm],
                             'in_path': in_path,
                             'out_path': self.resolve_userland_executable(in_path),
                         })
diff --git a/common.py b/common.py
index 25bfaa7..e2c00f6 100644
--- a/common.py
+++ b/common.py
@@ -1497,8 +1497,6 @@ https://github.com/cirosantilli/linux-kernel-module-cheat#gem5-debug-build
             if link:
                 if my_path_properties['extra_objs_lkmc_common']:
                     extra_objs.extend(extra_objs_lkmc_common)
-                if my_path_properties['extra_objs_userland_asm']:
-                    extra_objs.extend(extra_objs_userland_asm)
                 if my_path_properties['extra_objs_baremetal_bootloader']:
                     extra_objs.extend(extra_objs_baremetal_bootloader)
             if self.need_rebuild([in_path] + extra_objs + extra_deps, out_path):
@@ -1518,12 +1516,7 @@ https://github.com/cirosantilli/linux-kernel-module-cheat#gem5-debug-build
                 if dirpath_relative_root_components_len > 0:
                     if dirpath_relative_root_components[0] == 'userland':
                         if dirpath_relative_root_components_len > 1:
-                            if dirpath_relative_root_components[1] == 'arch':
-                                cc_flags.extend([
-                                    '-I', os.path.join(self.env['userland_source_arch_arch_dir']), LF,
-                                    '-I', os.path.join(self.env['userland_source_arch_dir']), LF,
-                                ])
-                            elif dirpath_relative_root_components[1] == 'libs':
+                            if dirpath_relative_root_components[1] == 'libs':
                                 if dirpath_relative_root_components_len > 1:
                                     if self.env['gcc_which'] == 'host':
                                         eigen_root = '/'
diff --git a/lkmc.c b/lkmc.c
index d00ea3e..a82414f 100644
--- a/lkmc.c
+++ b/lkmc.c
@@ -2,10 +2,60 @@
 
 #include <math.h>
 #include <stdio.h>
+#include <stdio.h>
 #include <stdlib.h>
 
 #include <lkmc.h>
 
+#define LKMC_ASSERT_EQ_DEFINE(bits) \
+    LKMC_ASSERT_EQ_DECLARE(bits) \
+    { \
+        if (val1 != val2) { \
+            printf("%s failed\n", __func__); \
+            printf("val1 0x%" PRIX ## bits "\n", val1); \
+            printf("val2 0x%" PRIX ## bits "\n", val2); \
+            lkmc_assert_fail(line); \
+        } \
+    }
+LKMC_ASSERT_EQ_DEFINE(32)
+LKMC_ASSERT_EQ_DEFINE(64)
+#undef ASSERT_EQ_DEFINE
+
+void lkmc_assert_fail(uint32_t line) {
+    printf("error: assertion failed at line: %" PRIu32 "\n", line);
+    exit(1);
+}
+
+void lkmc_assert_memcmp(
+    const void *s1,
+    const void *s2,
+    size_t n,
+    uint32_t line
+) {
+    size_t i;
+    uint8_t *s1b, *s2b;
+    uint8_t b1, b2;
+
+    s1b = (uint8_t *)s1;
+    s2b = (uint8_t *)s2;
+    for (i = 0; i < n; ++i) {
+        b1 = s1b[i];
+        b2 = s2b[i];
+        if (b1 != b2) {
+            printf(
+                "%s failed: "
+                "byte1, byte2, index: "
+                "0x%02" PRIX8 " 0x%02" PRIX8 " 0x%zX\n",
+                __func__,
+                b1,
+                b2,
+                i
+            );
+            lkmc_assert_fail(line);
+        }
+    }
+}
+
 void lkmc_baremetal_on_exit_callback(int status, void *arg) {
     (void)arg;
     if (status != 0) {
diff --git a/lkmc.h b/lkmc.h
index a642b6d..7fb5be6 100644
--- a/lkmc.h
+++ b/lkmc.h
@@ -3,17 +3,33 @@
 #ifndef LKMC_H
 #define LKMC_H
 
-/* Common C definitions. */
-#define LKMC_UNUSED(x) (void)x
-
 #if !defined(__ASSEMBLER__)
 #include <inttypes.h>
 #include <stdbool.h>
+#include <stdint.h>
+#include <stdio.h>
 #include <stdlib.h>
 
+#define LKMC_ASSERT_EQ_DECLARE(bits) \
+    void lkmc_assert_eq_ ## bits( \
+        uint ## bits ## _t val1, \
+        uint ## bits ## _t val2, \
+        uint32_t line \
+    )
+LKMC_ASSERT_EQ_DECLARE(32);
+LKMC_ASSERT_EQ_DECLARE(64);
+void lkmc_assert_fail(uint32_t line);
+void lkmc_assert_memcmp(const void *s1, const void *s2, size_t n, uint32_t line);
 void lkmc_baremetal_on_exit_callback(int status, void *arg);
 #endif
 
+/* Assert that the given branch instruction is taken. */
+#define LKMC_ASSERT(branch_if_pass) \
+    branch_if_pass 1f; \
+    LKMC_ASSERT_FAIL; \
+1: \
+;
+
 /* https://stackoverflow.com/questions/1489932/how-to-concatenate-twice-with-the-c-preprocessor-and-expand-a-macro-as-in-arg */
 #define LKMC_CONCAT_EVAL(a,b) a ## b
 #define LKMC_CONCAT(a,b) LKMC_CONCAT_EVAL(a, b)
@@ -22,6 +38,9 @@ void lkmc_baremetal_on_exit_callback(int status, void *arg);
     .global name; \
     name:
 
+/* Common C definitions. */
+#define LKMC_UNUSED(x) (void)x
+
 /* Weak means that if any other file defines it as a non-weak global,
  * that one will take precedence:
  * https://stackoverflow.com/questions/274753/how-to-make-weak-linking-work-with-gcc/54601464#54601464
@@ -30,238 +49,14 @@ void lkmc_baremetal_on_exit_callback(int status, void *arg);
     .weak name; \
     name:
 
-/* aarch64 specifics. */
-#if defined(__aarch64__)
-
-/* LKMC_VECTOR_TABLE
- *
- * Adapted from: https://github.com/takeharukato/sample-tsk-sw/blob/ce7973aa5d46c9eedb58309de43df3b09d4f8d8d/hal/aarch64/vector.S
- */
-
-#define LKMC_VECTOR_SYNC_SP0     (0x1)
-#define LKMC_VECTOR_IRQ_SP0      (0x2)
-#define LKMC_VECTOR_FIQ_SP0      (0x3)
-#define LKMC_VECTOR_SERR_SP0     (0x4)
-
-#define LKMC_VECTOR_SYNC_SPX     (0x11)
-#define LKMC_VECTOR_IRQ_SPX      (0x12)
-#define LKMC_VECTOR_FIQ_SPX      (0x13)
-#define LKMC_VECTOR_SERR_SPX     (0x14)
-
-#define LKMC_VECTOR_SYNC_AARCH64 (0x21)
-#define LKMC_VECTOR_IRQ_AARCH64  (0x22)
-#define LKMC_VECTOR_FIQ_AARCH64  (0x23)
-#define LKMC_VECTOR_SERR_AARCH64 (0x24)
-
-#define LKMC_VECTOR_SYNC_AARCH32 (0x31)
-#define LKMC_VECTOR_IRQ_AARCH32  (0x32)
-#define LKMC_VECTOR_FIQ_AARCH32  (0x33)
-#define LKMC_VECTOR_SERR_AARCH32 (0x34)
-
-#define LKMC_VECTOR_EXC_FRAME_SIZE      (288) /* sizeof(lkmc_vector_exception_frame)             */
-#define LKMC_VECTOR_EXC_EXC_TYPE_OFFSET (0)   /* offsetof(lkmc_vector_exception_frame, exc_type) */
-#define LKMC_VECTOR_EXC_EXC_ESR_OFFSE   (8)   /* offsetof(lkmc_vector_exception_frame, exc_esr)  */
-#define LKMC_VECTOR_EXC_EXC_SP_OFFSET   (16)  /* offsetof(lkmc_vector_exception_frame, exc_sp)   */
-#define LKMC_VECTOR_EXC_EXC_ELR_OFFSET  (24)  /* offsetof(lkmc_vector_exception_frame, exc_elr)  */
-#define LKMC_VECTOR_EXC_EXC_SPSR_OFFSET (32)  /* offsetof(lkmc_vector_exception_frame, exc_spsr) */
-
-#define LKMC_VECTOR_FUNC_ALIGN .align 2
-#define LKMC_VECTOR_SYMBOL_PREFIX lkmc_vector_
-
-#define LKMC_VECTOR_BUILD_TRAPFRAME(exc_type) \
-    stp x29, x30, [sp, -16]!; \
-    stp x27, x28, [sp, -16]!; \
-    stp x25, x26, [sp, -16]!; \
-    stp x23, x24, [sp, -16]!; \
-    stp x21, x22, [sp, -16]!; \
-    stp x19, x20, [sp, -16]!; \
-    stp x17, x18, [sp, -16]!; \
-    stp x15, x16, [sp, -16]!; \
-    stp x13, x14, [sp, -16]!; \
-    stp x11, x12, [sp, -16]!; \
-    stp x9, x10, [sp, -16]!; \
-    stp x7, x8, [sp, -16]!; \
-    stp x5, x6, [sp, -16]!; \
-    stp x3, x4, [sp, -16]!; \
-    stp x1, x2, [sp, -16]!; \
-    mrs x21, spsr_el1; \
-    stp x21, x0, [sp, -16]!; \
-    mrs x21, elr_el1; \
-    stp xzr, x21, [sp, -16]!; \
-    mov x21, exc_type; \
-    mrs x22, esr_el1; \
-    stp x21, x22, [sp, -16]!
-
-#define LKMC_VECTOR_STORE_TRAPED_SP \
-    mrs x21, sp_el0; \
-    str x21, [sp, LKMC_VECTOR_EXC_EXC_SP_OFFSET]
-
-#define LKMC_VECTOR_CALL_TRAP_HANDLER \
-    mov x0, sp; \
-    bl lkmc_vector_trap_handler
-
-#define LKMC_VECTOR_STORE_NESTED_SP \
-    mov x21, sp; \
-    add x21, x21, LKMC_VECTOR_EXC_FRAME_SIZE; \
-    str x21, [sp, LKMC_VECTOR_EXC_EXC_SP_OFFSET]
-
-#define LKMC_VECTOR_RESTORE_TRAPED_SP \
-    ldr x21, [sp, LKMC_VECTOR_EXC_EXC_SP_OFFSET]; \
-    msr sp_el0, x21
-
-#define LKMC_VECTOR_RESTORE_TRAPFRAME \
-    add sp, sp, 16; \
-    ldp x21, x22, [sp], 16; \
-    msr elr_el1, x22; \
-    ldp x21, x0, [sp], 16; \
-    msr spsr_el1, x21; \
-    ldp x1, x2, [sp], 16; \
-    ldp x3, x4, [sp], 16; \
-    ldp x5, x6, [sp], 16; \
-    ldp x7, x8, [sp], 16; \
-    ldp x9, x10, [sp], 16; \
-    ldp x11, x12, [sp], 16; \
-    ldp x13, x14, [sp], 16; \
-    ldp x15, x16, [sp], 16; \
-    ldp x17, x18, [sp], 16; \
-    ldp x19, x20, [sp], 16; \
-    ldp x21, x22, [sp], 16; \
-    ldp x23, x24, [sp], 16; \
-    ldp x25, x26, [sp], 16; \
-    ldp x27, x28, [sp], 16; \
-    ldp x29, x30, [sp], 16; \
-    eret
-
-#define LKMC_VECTOR_ENTRY(func_name) \
-    .align 7; \
-    b LKMC_VECTOR_SYMBOL_PREFIX ## func_name
-
-#define LKMC_VECTOR_FUNC(func_name, func_id) \
-    LKMC_VECTOR_FUNC_ALIGN; \
-LKMC_VECTOR_SYMBOL_PREFIX ## func_name:; \
-    LKMC_VECTOR_BUILD_TRAPFRAME(func_id); \
-    LKMC_VECTOR_STORE_TRAPED_SP; \
-    LKMC_VECTOR_CALL_TRAP_HANDLER; \
-    LKMC_VECTOR_RESTORE_TRAPED_SP; \
-    LKMC_VECTOR_RESTORE_TRAPFRAME
-
-#define LKMC_VECTOR_FUNC_NESTED(func_name, func_id) \
-    LKMC_VECTOR_FUNC_ALIGN; \
-LKMC_VECTOR_SYMBOL_PREFIX ## func_name:; \
-    LKMC_VECTOR_BUILD_TRAPFRAME(func_id); \
-    LKMC_VECTOR_STORE_NESTED_SP; \
-    LKMC_VECTOR_CALL_TRAP_HANDLER; \
-    LKMC_VECTOR_RESTORE_TRAPFRAME
-
-/* Define the actual vector table. */
-#define LKMC_VECTOR_TABLE \
-    .align 11; \
-    LKMC_GLOBAL(lkmc_vector_table); \
-    ; \
-    LKMC_VECTOR_ENTRY(curr_el_sp0_sync); \
-    LKMC_VECTOR_ENTRY(curr_el_sp0_irq); \
-    LKMC_VECTOR_ENTRY(curr_el_sp0_fiq); \
-    LKMC_VECTOR_ENTRY(curr_el_sp0_serror); \
-    ; \
-    LKMC_VECTOR_ENTRY(curr_el_spx_sync); \
-    LKMC_VECTOR_ENTRY(curr_el_spx_irq); \
-    LKMC_VECTOR_ENTRY(curr_el_spx_fiq); \
-    LKMC_VECTOR_ENTRY(curr_el_spx_serror); \
-    ; \
-    LKMC_VECTOR_ENTRY(lower_el_aarch64_sync); \
-    LKMC_VECTOR_ENTRY(lower_el_aarch64_irq); \
-    LKMC_VECTOR_ENTRY(lower_el_aarch64_fiq); \
-    LKMC_VECTOR_ENTRY(lower_el_aarch64_serror); \
-    ; \
-    LKMC_VECTOR_ENTRY(lower_el_aarch32_sync); \
-    LKMC_VECTOR_ENTRY(lower_el_aarch32_irq); \
-    LKMC_VECTOR_ENTRY(lower_el_aarch32_fiq); \
-    LKMC_VECTOR_ENTRY(lower_el_aarch32_serror); \
-    ; \
-    LKMC_VECTOR_FUNC(curr_el_sp0_sync, LKMC_VECTOR_SYNC_SP0); \
-    LKMC_VECTOR_FUNC(curr_el_sp0_irq, LKMC_VECTOR_IRQ_SP0); \
-    LKMC_VECTOR_FUNC(curr_el_sp0_fiq, LKMC_VECTOR_FIQ_SP0); \
-    LKMC_VECTOR_FUNC(curr_el_sp0_serror, LKMC_VECTOR_SERR_SP0); \
-    ; \
-    LKMC_VECTOR_FUNC_NESTED(curr_el_spx_sync, LKMC_VECTOR_SYNC_SPX); \
-    LKMC_VECTOR_FUNC_NESTED(curr_el_spx_irq, LKMC_VECTOR_IRQ_SPX); \
-    LKMC_VECTOR_FUNC_NESTED(curr_el_spx_fiq, LKMC_VECTOR_FIQ_SPX); \
-    LKMC_VECTOR_FUNC_NESTED(curr_el_spx_serror, LKMC_VECTOR_SERR_SPX); \
-    ; \
-    LKMC_VECTOR_FUNC(lower_el_aarch64_sync, LKMC_VECTOR_SYNC_AARCH64); \
-    LKMC_VECTOR_FUNC(lower_el_aarch64_irq, LKMC_VECTOR_IRQ_AARCH64); \
-    LKMC_VECTOR_FUNC(lower_el_aarch64_fiq, LKMC_VECTOR_FIQ_AARCH64); \
-    LKMC_VECTOR_FUNC(lower_el_aarch64_serror, LKMC_VECTOR_SERR_AARCH64); \
-    ; \
-    LKMC_VECTOR_FUNC(lower_el_aarch32_sync, LKMC_VECTOR_SYNC_AARCH32); \
-    LKMC_VECTOR_FUNC(lower_el_aarch32_irq, LKMC_VECTOR_IRQ_AARCH32); \
-    LKMC_VECTOR_FUNC(lower_el_aarch32_fiq, LKMC_VECTOR_FIQ_AARCH32); \
-    LKMC_VECTOR_FUNC(lower_el_aarch32_serror, LKMC_VECTOR_SERR_AARCH32)
-
-/* aarch64 C definitions. */
-#if !defined(__ASSEMBLER__)
-#include <stdint.h>
-
-typedef struct {
-    uint64_t exc_type;
-    uint64_t exc_esr;
-    uint64_t exc_sp;
-    uint64_t exc_elr;
-    uint64_t exc_spsr;
-    uint64_t x0;
-    uint64_t x1;
-    uint64_t x2;
-    uint64_t x3;
-    uint64_t x4;
-    uint64_t x5;
-    uint64_t x6;
-    uint64_t x7;
-    uint64_t x8;
-    uint64_t x9;
-    uint64_t x10;
-    uint64_t x11;
-    uint64_t x12;
-    uint64_t x13;
-    uint64_t x14;
-    uint64_t x15;
-    uint64_t x16;
-    uint64_t x17;
-    uint64_t x18;
-    uint64_t x19;
-    uint64_t x20;
-    uint64_t x21;
-    uint64_t x22;
-    uint64_t x23;
-    uint64_t x24;
-    uint64_t x25;
-    uint64_t x26;
-    uint64_t x27;
-    uint64_t x28;
-    uint64_t x29;
-    uint64_t x30;
-} LkmcVectorExceptionFrame;
-
-void lkmc_vector_trap_handler(LkmcVectorExceptionFrame *exception);
-
-#define LKMC_SVC(immediate) __asm__ __volatile__("svc " #immediate : : : )
-
-#define LKMC_SYSREG_SYMBOL_PREFIX lkmc_sysreg_
-#define LKMC_SYSREG_READ_WRITE(type, name) \
-    type LKMC_CONCAT(LKMC_CONCAT(LKMC_SYSREG_SYMBOL_PREFIX, name), _read)(void); \
-    void LKMC_CONCAT(LKMC_CONCAT(LKMC_SYSREG_SYMBOL_PREFIX, name), _write)(type name);
-#define LKMC_SYSREG_OPS \
-    LKMC_SYSREG_READ_WRITE(uint32_t, cntv_ctl_el0) \
-    LKMC_SYSREG_READ_WRITE(uint32_t, daif) \
-    LKMC_SYSREG_READ_WRITE(uint32_t, spsel) \
-    LKMC_SYSREG_READ_WRITE(uint64_t, cntfrq_el0) \
-    LKMC_SYSREG_READ_WRITE(uint64_t, cntv_cval_el0) \
-    LKMC_SYSREG_READ_WRITE(uint64_t, cntv_tval_el0) \
-    LKMC_SYSREG_READ_WRITE(uint64_t, cntvct_el0) \
-    LKMC_SYSREG_READ_WRITE(uint64_t, sp_el1) \
-    LKMC_SYSREG_READ_WRITE(uint64_t, vbar_el1)
-LKMC_SYSREG_OPS
-#undef LKMC_SYSREG_READ_WRITE
+#if defined(__x86_64__)
+#include <lkmc/x86_64.h>
+#elif defined(__arm__)
+#include <lkmc/arm.h>
+#elif defined(__aarch64__)
+#include <lkmc/aarch64.h>
+#else
+#error
+#endif
 
 #endif
-#endif
-#endif
diff --git a/lkmc/aarch64.h b/lkmc/aarch64.h
new file mode 100644
index 0000000..475643f
--- /dev/null
+++ b/lkmc/aarch64.h
@@ -0,0 +1,299 @@
+#ifndef LKMC_AARCH64_H
+#define LKMC_AARCH64_H
+
+#define LKMC_ASSERT_EQ(reg, const) \
+    mov x0, reg; \
+    ldr x1, const; \
+    ldr w2, =__LINE__; \
+    bl lkmc_assert_eq_64; \
+;
+
+#define LKMC_ASSERT_EQ_REG(reg1, reg2) \
+    str reg2, [sp, -16]!; \
+    mov x0, reg1; \
+    ldr x1, [sp], 16; \
+    ldr w2, =__LINE__; \
+    bl lkmc_assert_eq_64; \
+;
+
+#define LKMC_ASSERT_EQ_REG_32(reg1, reg2) \
+    str reg2, [sp, -4]!; \
+    mov w0, reg1; \
+    ldr w1, [sp], 4; \
+    ldr w2, =__LINE__; \
+    bl lkmc_assert_eq_32; \
+;
+
+#define LKMC_ASSERT_FAIL \
+    ldr w0, =__LINE__; \
+    bl lkmc_assert_fail; \
+;
+
+#define LKMC_ASSERT_MEMCMP(label1, label2, size) \
+    adr x0, label1; \
+    adr x1, label2; \
+    ldr x2, size; \
+    ldr x3, =__LINE__; \
+    bl lkmc_assert_memcmp; \
+;
+
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-calling-convention */
+#define LKMC_EPILOGUE \
+    ldp x19, x20, [sp, 0x50]; \
+    ldp x21, x22, [sp, 0x40]; \
+    ldp x23, x24, [sp, 0x30]; \
+    ldp x25, x26, [sp, 0x20]; \
+    ldp x27, x28, [sp, 0x10]; \
+    ldp x29, x30, [sp]; \
+    add sp, sp, 0x60; \
+    mov x0, 0; \
+    ret; \
+;
+
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-calling-convention */
+#define LKMC_PROLOGUE \
+.text; \
+.global main; \
+main: \
+    sub sp, sp, 0x60; \
+    stp x29, x30, [sp]; \
+    stp x27, x28, [sp, 0x10]; \
+    stp x25, x26, [sp, 0x20]; \
+    stp x23, x24, [sp, 0x30]; \
+    stp x21, x22, [sp, 0x40]; \
+    stp x19, x20, [sp, 0x50]; \
+main_after_prologue: \
+;
+
+/* LKMC_VECTOR_TABLE
+ *
+ * Adapted from: https://github.com/takeharukato/sample-tsk-sw/blob/ce7973aa5d46c9eedb58309de43df3b09d4f8d8d/hal/aarch64/vector.S
+ */
+
+#define LKMC_VECTOR_SYNC_SP0     (0x1)
+#define LKMC_VECTOR_IRQ_SP0      (0x2)
+#define LKMC_VECTOR_FIQ_SP0      (0x3)
+#define LKMC_VECTOR_SERR_SP0     (0x4)
+
+#define LKMC_VECTOR_SYNC_SPX     (0x11)
+#define LKMC_VECTOR_IRQ_SPX      (0x12)
+#define LKMC_VECTOR_FIQ_SPX      (0x13)
+#define LKMC_VECTOR_SERR_SPX     (0x14)
+
+#define LKMC_VECTOR_SYNC_AARCH64 (0x21)
+#define LKMC_VECTOR_IRQ_AARCH64  (0x22)
+#define LKMC_VECTOR_FIQ_AARCH64  (0x23)
+#define LKMC_VECTOR_SERR_AARCH64 (0x24)
+
+#define LKMC_VECTOR_SYNC_AARCH32 (0x31)
+#define LKMC_VECTOR_IRQ_AARCH32  (0x32)
+#define LKMC_VECTOR_FIQ_AARCH32  (0x33)
+#define LKMC_VECTOR_SERR_AARCH32 (0x34)
+
+#define LKMC_VECTOR_EXC_FRAME_SIZE      (288) /* sizeof(lkmc_vector_exception_frame)             */
+#define LKMC_VECTOR_EXC_EXC_TYPE_OFFSET (0)   /* offsetof(lkmc_vector_exception_frame, exc_type) */
+#define LKMC_VECTOR_EXC_EXC_ESR_OFFSE   (8)   /* offsetof(lkmc_vector_exception_frame, exc_esr)  */
+#define LKMC_VECTOR_EXC_EXC_SP_OFFSET   (16)  /* offsetof(lkmc_vector_exception_frame, exc_sp)   */
+#define LKMC_VECTOR_EXC_EXC_ELR_OFFSET  (24)  /* offsetof(lkmc_vector_exception_frame, exc_elr)  */
+#define LKMC_VECTOR_EXC_EXC_SPSR_OFFSET (32)  /* offsetof(lkmc_vector_exception_frame, exc_spsr) */
+
+#define LKMC_VECTOR_FUNC_ALIGN .align 2
+#define LKMC_VECTOR_SYMBOL_PREFIX lkmc_vector_
+
+#define LKMC_VECTOR_BUILD_TRAPFRAME(exc_type) \
+    stp x29, x30, [sp, -16]!; \
+    stp x27, x28, [sp, -16]!; \
+    stp x25, x26, [sp, -16]!; \
+    stp x23, x24, [sp, -16]!; \
+    stp x21, x22, [sp, -16]!; \
+    stp x19, x20, [sp, -16]!; \
+    stp x17, x18, [sp, -16]!; \
+    stp x15, x16, [sp, -16]!; \
+    stp x13, x14, [sp, -16]!; \
+    stp x11, x12, [sp, -16]!; \
+    stp x9, x10, [sp, -16]!; \
+    stp x7, x8, [sp, -16]!; \
+    stp x5, x6, [sp, -16]!; \
+    stp x3, x4, [sp, -16]!; \
+    stp x1, x2, [sp, -16]!; \
+    mrs x21, spsr_el1; \
+    stp x21, x0, [sp, -16]!; \
+    mrs x21, elr_el1; \
+    stp xzr, x21, [sp, -16]!; \
+    mov x21, exc_type; \
+    mrs x22, esr_el1; \
+    stp x21, x22, [sp, -16]!
+
+#define LKMC_VECTOR_STORE_TRAPED_SP \
+    mrs x21, sp_el0; \
+    str x21, [sp, LKMC_VECTOR_EXC_EXC_SP_OFFSET]
+
+#define LKMC_VECTOR_CALL_TRAP_HANDLER \
+    mov x0, sp; \
+    bl lkmc_vector_trap_handler
+
+#define LKMC_VECTOR_STORE_NESTED_SP \
+    mov x21, sp; \
+    add x21, x21, LKMC_VECTOR_EXC_FRAME_SIZE; \
+    str x21, [sp, LKMC_VECTOR_EXC_EXC_SP_OFFSET]
+
+#define LKMC_VECTOR_RESTORE_TRAPED_SP \
+    ldr x21, [sp, LKMC_VECTOR_EXC_EXC_SP_OFFSET]; \
+    msr sp_el0, x21
+
+#define LKMC_VECTOR_RESTORE_TRAPFRAME \
+    add sp, sp, 16; \
+    ldp x21, x22, [sp], 16; \
+    msr elr_el1, x22; \
+    ldp x21, x0, [sp], 16; \
+    msr spsr_el1, x21; \
+    ldp x1, x2, [sp], 16; \
+    ldp x3, x4, [sp], 16; \
+    ldp x5, x6, [sp], 16; \
+    ldp x7, x8, [sp], 16; \
+    ldp x9, x10, [sp], 16; \
+    ldp x11, x12, [sp], 16; \
+    ldp x13, x14, [sp], 16; \
+    ldp x15, x16, [sp], 16; \
+    ldp x17, x18, [sp], 16; \
+    ldp x19, x20, [sp], 16; \
+    ldp x21, x22, [sp], 16; \
+    ldp x23, x24, [sp], 16; \
+    ldp x25, x26, [sp], 16; \
+    ldp x27, x28, [sp], 16; \
+    ldp x29, x30, [sp], 16; \
+    eret
+
+#define LKMC_VECTOR_ENTRY(func_name) \
+    .align 7; \
+    b LKMC_VECTOR_SYMBOL_PREFIX ## func_name
+
+#define LKMC_VECTOR_FUNC(func_name, func_id) \
+    LKMC_VECTOR_FUNC_ALIGN; \
+LKMC_VECTOR_SYMBOL_PREFIX ## func_name:; \
+    LKMC_VECTOR_BUILD_TRAPFRAME(func_id); \
+    LKMC_VECTOR_STORE_TRAPED_SP; \
+    LKMC_VECTOR_CALL_TRAP_HANDLER; \
+    LKMC_VECTOR_RESTORE_TRAPED_SP; \
+    LKMC_VECTOR_RESTORE_TRAPFRAME
+
+#define LKMC_VECTOR_FUNC_NESTED(func_name, func_id) \
+    LKMC_VECTOR_FUNC_ALIGN; \
+LKMC_VECTOR_SYMBOL_PREFIX ## func_name:; \
+    LKMC_VECTOR_BUILD_TRAPFRAME(func_id); \
+    LKMC_VECTOR_STORE_NESTED_SP; \
+    LKMC_VECTOR_CALL_TRAP_HANDLER; \
+    LKMC_VECTOR_RESTORE_TRAPFRAME
+
+/* Define the actual vector table. */
+#define LKMC_VECTOR_TABLE \
+    .align 11; \
+    LKMC_GLOBAL(lkmc_vector_table); \
+    ; \
+    LKMC_VECTOR_ENTRY(curr_el_sp0_sync); \
+    LKMC_VECTOR_ENTRY(curr_el_sp0_irq); \
+    LKMC_VECTOR_ENTRY(curr_el_sp0_fiq); \
+    LKMC_VECTOR_ENTRY(curr_el_sp0_serror); \
+    ; \
+    LKMC_VECTOR_ENTRY(curr_el_spx_sync); \
+    LKMC_VECTOR_ENTRY(curr_el_spx_irq); \
+    LKMC_VECTOR_ENTRY(curr_el_spx_fiq); \
+    LKMC_VECTOR_ENTRY(curr_el_spx_serror); \
+    ; \
+    LKMC_VECTOR_ENTRY(lower_el_aarch64_sync); \
+    LKMC_VECTOR_ENTRY(lower_el_aarch64_irq); \
+    LKMC_VECTOR_ENTRY(lower_el_aarch64_fiq); \
+    LKMC_VECTOR_ENTRY(lower_el_aarch64_serror); \
+    ; \
+    LKMC_VECTOR_ENTRY(lower_el_aarch32_sync); \
+    LKMC_VECTOR_ENTRY(lower_el_aarch32_irq); \
+    LKMC_VECTOR_ENTRY(lower_el_aarch32_fiq); \
+    LKMC_VECTOR_ENTRY(lower_el_aarch32_serror); \
+    ; \
+    LKMC_VECTOR_FUNC(curr_el_sp0_sync, LKMC_VECTOR_SYNC_SP0); \
+    LKMC_VECTOR_FUNC(curr_el_sp0_irq, LKMC_VECTOR_IRQ_SP0); \
+    LKMC_VECTOR_FUNC(curr_el_sp0_fiq, LKMC_VECTOR_FIQ_SP0); \
+    LKMC_VECTOR_FUNC(curr_el_sp0_serror, LKMC_VECTOR_SERR_SP0); \
+    ; \
+    LKMC_VECTOR_FUNC_NESTED(curr_el_spx_sync, LKMC_VECTOR_SYNC_SPX); \
+    LKMC_VECTOR_FUNC_NESTED(curr_el_spx_irq, LKMC_VECTOR_IRQ_SPX); \
+    LKMC_VECTOR_FUNC_NESTED(curr_el_spx_fiq, LKMC_VECTOR_FIQ_SPX); \
+    LKMC_VECTOR_FUNC_NESTED(curr_el_spx_serror, LKMC_VECTOR_SERR_SPX); \
+    ; \
+    LKMC_VECTOR_FUNC(lower_el_aarch64_sync, LKMC_VECTOR_SYNC_AARCH64); \
+    LKMC_VECTOR_FUNC(lower_el_aarch64_irq, LKMC_VECTOR_IRQ_AARCH64); \
+    LKMC_VECTOR_FUNC(lower_el_aarch64_fiq, LKMC_VECTOR_FIQ_AARCH64); \
+    LKMC_VECTOR_FUNC(lower_el_aarch64_serror, LKMC_VECTOR_SERR_AARCH64); \
+    ; \
+    LKMC_VECTOR_FUNC(lower_el_aarch32_sync, LKMC_VECTOR_SYNC_AARCH32); \
+    LKMC_VECTOR_FUNC(lower_el_aarch32_irq, LKMC_VECTOR_IRQ_AARCH32); \
+    LKMC_VECTOR_FUNC(lower_el_aarch32_fiq, LKMC_VECTOR_FIQ_AARCH32); \
+    LKMC_VECTOR_FUNC(lower_el_aarch32_serror, LKMC_VECTOR_SERR_AARCH32)
+
+/* aarch64 C definitions. */
+#if !defined(__ASSEMBLER__)
+#include <stdint.h>
+
+typedef struct {
+    uint64_t exc_type;
+    uint64_t exc_esr;
+    uint64_t exc_sp;
+    uint64_t exc_elr;
+    uint64_t exc_spsr;
+    uint64_t x0;
+    uint64_t x1;
+    uint64_t x2;
+    uint64_t x3;
+    uint64_t x4;
+    uint64_t x5;
+    uint64_t x6;
+    uint64_t x7;
+    uint64_t x8;
+    uint64_t x9;
+    uint64_t x10;
+    uint64_t x11;
+    uint64_t x12;
+    uint64_t x13;
+    uint64_t x14;
+    uint64_t x15;
+    uint64_t x16;
+    uint64_t x17;
+    uint64_t x18;
+    uint64_t x19;
+    uint64_t x20;
+    uint64_t x21;
+    uint64_t x22;
+    uint64_t x23;
+    uint64_t x24;
+    uint64_t x25;
+    uint64_t x26;
+    uint64_t x27;
+    uint64_t x28;
+    uint64_t x29;
+    uint64_t x30;
+} LkmcVectorExceptionFrame;
+
+void lkmc_vector_trap_handler(LkmcVectorExceptionFrame *exception);
+
+#define LKMC_SVC(immediate) __asm__ __volatile__("svc " #immediate : : : )
+
+#define LKMC_SYSREG_SYMBOL_PREFIX lkmc_sysreg_
+#define LKMC_SYSREG_READ_WRITE(type, name) \
+    type LKMC_CONCAT(LKMC_CONCAT(LKMC_SYSREG_SYMBOL_PREFIX, name), _read)(void); \
+    void LKMC_CONCAT(LKMC_CONCAT(LKMC_SYSREG_SYMBOL_PREFIX, name), _write)(type name);
+#define LKMC_SYSREG_OPS \
+    LKMC_SYSREG_READ_WRITE(uint32_t, cntv_ctl_el0) \
+    LKMC_SYSREG_READ_WRITE(uint32_t, daif) \
+    LKMC_SYSREG_READ_WRITE(uint32_t, spsel) \
+    LKMC_SYSREG_READ_WRITE(uint64_t, cntfrq_el0) \
+    LKMC_SYSREG_READ_WRITE(uint64_t, cntv_cval_el0) \
+    LKMC_SYSREG_READ_WRITE(uint64_t, cntv_tval_el0) \
+    LKMC_SYSREG_READ_WRITE(uint64_t, cntvct_el0) \
+    LKMC_SYSREG_READ_WRITE(uint64_t, sp_el1) \
+    LKMC_SYSREG_READ_WRITE(uint64_t, vbar_el1)
+LKMC_SYSREG_OPS
+#undef LKMC_SYSREG_READ_WRITE
+
+#endif
+
+#endif
diff --git a/lkmc/arm.h b/lkmc/arm.h
new file mode 100644
index 0000000..80ffc7c
--- /dev/null
+++ b/lkmc/arm.h
@@ -0,0 +1,53 @@
+#ifndef LKMC_ARM_H
+#define LKMC_ARM_H
+
+#if defined(__ASSEMBLER__)
+.syntax unified
+#endif
+
+#define LKMC_ASSERT_EQ(reg, const) \
+    mov r0, reg; \
+    ldr r1, const; \
+    ldr r2, =__LINE__; \
+    bl lkmc_assert_eq_32; \
+;
+
+#define LKMC_ASSERT_EQ_REG(reg1, reg2) \
+    str reg2, [sp, -4]!; \
+    mov r0, reg1; \
+    ldr r1, [sp], 4; \
+    ldr r2, =__LINE__; \
+    bl lkmc_assert_eq_32; \
+;
+
+#define LKMC_ASSERT_FAIL \
+    ldr r0, =__LINE__; \
+    bl lkmc_assert_fail; \
+;
+
+#define LKMC_ASSERT_MEMCMP(label1, label2, size) \
+    ldr r0, =label1; \
+    ldr r1, =label2; \
+    ldr r2, size; \
+    ldr r3, =__LINE__; \
+    bl lkmc_assert_memcmp; \
+;
+
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-calling-convention */
+#define LKMC_EPILOGUE \
+    add sp, 16; \
+    ldmia sp!, {r4-r12, lr}; \
+    mov r0, 0; \
+    bx lr; \
+;
+
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-calling-convention */
+#define LKMC_PROLOGUE \
+.text; \
+.global main; \
+main: \
+    stmdb sp!, {r0-r12, lr}; \
+main_after_prologue: \
+;
+
+#endif
diff --git a/lkmc/x86_64.h b/lkmc/x86_64.h
new file mode 100644
index 0000000..5ad67fa
--- /dev/null
+++ b/lkmc/x86_64.h
@@ -0,0 +1,62 @@
+#ifndef LKMC_X86_64_H
+#define LKMC_X86_64_H
+
+/* This and other macros may make C function calls, and therefore can destroy
+ * non-callee saved registers. */
+#define LKMC_ASSERT_EQ(general1, general2) \
+    mov general2, %rdi; \
+    push %rdi; \
+    mov general1, %rdi; \
+    pop %rsi; \
+    mov $__LINE__, %edx; \
+    call lkmc_assert_eq_64; \
+;
+
+#define LKMC_ASSERT_FAIL \
+    mov $__LINE__, %edi; \
+    call lkmc_assert_fail; \
+;
+
+/* Assert that two memory arrays are the same. */
+#define LKMC_ASSERT_MEMCMP(label1, label2, const_size) \
+    lea label1(%rip), %rdi; \
+    lea label2(%rip), %rsi; \
+    mov const_size, %rdx; \
+    mov $__LINE__, %ecx; \
+    call lkmc_assert_memcmp; \
+;
+
+/* Function epilogue.
+ *
+ * https://github.com/cirosantilli/linux-kernel-module-cheat#x86_64-calling-convention
+ */
+#define LKMC_EPILOGUE \
+    pop %rbx; \
+    pop %r12; \
+    pop %r13; \
+    pop %r14; \
+    pop %r15; \
+    pop %rbp; \
+    mov $0, %rax; \
+    ret; \
+;
+
+/* Function prologue.
+ *
+ * https://github.com/cirosantilli/linux-kernel-module-cheat#x86_64-calling-convention
+ */
+#define LKMC_PROLOGUE \
+.text; \
+.global main; \
+main: \
+    push %rbp; \
+    mov %rsp, %rbp; \
+    push %r15; \
+    push %r14; \
+    push %r13; \
+    push %r12; \
+    push %rbx; \
+main_after_prologue: \
+;
+
+#endif
diff --git a/path_properties.py b/path_properties.py
index a67b25d..aa665de 100644
--- a/path_properties.py
+++ b/path_properties.py
@@ -36,7 +36,6 @@ class PathProperties:
         'extra_objs_baremetal_bootloader': False,
         # We should get rid of this if we ever properly implement dependency graphs.
         'extra_objs_lkmc_common': False,
-        'extra_objs_userland_asm': False,
         'interactive': False,
         # The script takes a perceptible amount of time to run. Possibly an infinite loop.
         'more_than_1s': False,
@@ -209,7 +208,7 @@ freestanding_properties = {
         '-nostdlib', LF,
         '-static', LF,
     ],
-    'extra_objs_userland_asm': False,
+    'extra_objs_lkmc_common': False,
 }
 # See: https://github.com/cirosantilli/linux-kernel-module-cheat#path-properties
 path_properties_tuples = (
@@ -284,7 +283,7 @@ path_properties_tuples = (
             {
                 'arch': (
                     {
-                        'extra_objs_userland_asm': True,
+                        'extra_objs_lkmc_common': True,
                     },
                     {
                         'arm': (
@@ -311,13 +310,14 @@ path_properties_tuples = (
                             {
                                 'c': (
                                     {
-                                        'extra_objs_userland_asm': False,
                                     },
                                     {
                                         'freestanding': freestanding_properties,
                                     },
                                 ),
                                 'freestanding': freestanding_properties,
+                                'lkmc_assert_eq_fail.S': {'exit_status': 1},
+                                'lkmc_assert_memcmp_fail.S': {'exit_status': 1},
                                 'udf.S': {
                                     'exit_status': 132,
                                     'receives_signal': True
@@ -329,13 +329,14 @@ path_properties_tuples = (
                             {
                                 'c': (
                                     {
-                                        'extra_objs_userland_asm': False,
                                     },
                                     {
                                         'freestanding': freestanding_properties,
                                     },
                                 ),
                                 'freestanding': freestanding_properties,
+                                'lkmc_assert_eq_fail.S': {'exit_status': 1},
+                                'lkmc_assert_memcmp_fail.S': {'exit_status': 1},
                                 'udf.S': {
                                     'exit_status': 132,
                                     'receives_signal': True
@@ -343,16 +344,11 @@ path_properties_tuples = (
                             }
                         ),
                         'fail.S': {'exit_status': 1},
-                        'main.c': {
-                            'extra_objs_userland_asm': False,
-                            'no_executable': True
-                        },
                         'x86_64': (
                             {'allowed_archs': {'x86_64'}},
                             {
                                 'c': (
                                     {
-                                        'extra_objs_userland_asm': False,
                                     },
                                     {
                                         'freestanding': freestanding_properties,
@@ -383,11 +379,6 @@ path_properties_tuples = (
                 ),
                 'gcc': gnu_extension_properties,
                 'kernel_modules': {**gnu_extension_properties, **{'requires_kernel_modules': True}},
-                'lkmc': (
-                    {'extra_objs_lkmc_common': True},
-                    {
-                    }
-                ),
                 'libs': (
                     {'requires_dynamic_library': True},
                     {
diff --git a/userland/arch/aarch64/add.S b/userland/arch/aarch64/add.S
index a2754c9..76dac10 100644
--- a/userland/arch/aarch64/add.S
+++ b/userland/arch/aarch64/add.S
@@ -1,9 +1,9 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     mov x0, 1
     add x1, x0, 2
-    LKMC_ASSERT_EQ(x1, 3)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x1, =3)
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/add_vector.S b/userland/arch/aarch64/add_vector.S
index aa61c5d..4606025 100644
--- a/userland/arch/aarch64/add_vector.S
+++ b/userland/arch/aarch64/add_vector.S
@@ -3,9 +3,9 @@
  * Add a bunch of integers in one go.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 .data
     input0:    .long 0xF1F1F1F1, 0xF2F2F2F2, 0xF3F3F3F3, 0xF4F4F4F4
     input1:    .long 0x12121212, 0x13131313, 0x14141414, 0x15151515
@@ -22,11 +22,11 @@ LKMC_ENTRY
     add v2. ## size, v0. ## size, v1. ## size; \
     adr x0, output; \
     st1 {v2. ## size}, [x0]; \
-    LKMC_ASSERT_MEMCMP(output, expect_ ## size, 0x10)
+    LKMC_ASSERT_MEMCMP(output, expect_ ## size, =0x10)
 
     /* 4x 32-bit */
     TEST(4s)
     /* 2x 64-bit */
     TEST(2d)
 #undef TEST
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/adr.S b/userland/arch/aarch64/adr.S
index 62159cd..4b36342 100644
--- a/userland/arch/aarch64/adr.S
+++ b/userland/arch/aarch64/adr.S
@@ -1,11 +1,11 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-adr-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
 .data
 data_label:
     .word 0x1234678
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* This is not possible in v7 because the label is in another section.
      * objdump says that this generates a R_AARCH64_ADR_PRE relocation.
      * which looks specific to ADR, and therefore makes it more likely
@@ -18,4 +18,4 @@ LKMC_ENTRY
     ldr x1, =data_label
 label:
     LKMC_ASSERT_EQ_REG(x0, x1)
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/adrp.S b/userland/arch/aarch64/adrp.S
index 4cf516e..eddf332 100644
--- a/userland/arch/aarch64/adrp.S
+++ b/userland/arch/aarch64/adrp.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-adr-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     adrp x0, label
     adr x1, label
 label:
@@ -10,4 +10,4 @@ label:
     bic x1, x1, 0xFF
     bic x1, x1, 0xF00
     LKMC_ASSERT_EQ_REG(x0, x1)
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/beq.S b/userland/arch/aarch64/beq.S
index c43bccc..1f8814e 100644
--- a/userland/arch/aarch64/beq.S
+++ b/userland/arch/aarch64/beq.S
@@ -1,12 +1,12 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-beq-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* cbz == 0 */
     mov x0, 0
     cbz x0, 1f
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 1:
 
     /* cbz != 0 */
@@ -14,13 +14,13 @@ LKMC_ENTRY
     cbz x0, 1f
     b 2f
 1:
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 2:
 
     /* cbnz != 0 */
     mov x0, 1
     cbnz x0, 1f
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 1:
 
     /* cbnz == 0 */
@@ -28,6 +28,6 @@ LKMC_ENTRY
     cbnz x0, 1f
     b 2f
 1:
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 2:
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/bfi.S b/userland/arch/aarch64/bfi.S
index 370c365..422bebc 100644
--- a/userland/arch/aarch64/bfi.S
+++ b/userland/arch/aarch64/bfi.S
@@ -1,11 +1,11 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-bfi-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr x0, =0x1122334455667788
 
     ldr x1, =0xFFFFFFFFFFFFFFFF
     bfi x1, x0, 16, 32
-    LKMC_ASSERT_EQ(x1, 0xFFFF55667788FFFF)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x1, =0xFFFF55667788FFFF)
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/cbz.S b/userland/arch/aarch64/cbz.S
index 921471b..7fe2502 100644
--- a/userland/arch/aarch64/cbz.S
+++ b/userland/arch/aarch64/cbz.S
@@ -1,19 +1,19 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-cbz-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Branch. */
     mov x0, 0x0
     cbz x0, ok
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 ok:
 
     /* Don't branch. */
     mov x0, 0x1
     cbz x0, ko
 
-LKMC_EXIT
+LKMC_EPILOGUE
 ko:
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
diff --git a/userland/arch/aarch64/comments.S b/userland/arch/aarch64/comments.S
index b94594c..44c28f5 100644
--- a/userland/arch/aarch64/comments.S
+++ b/userland/arch/aarch64/comments.S
@@ -1,7 +1,7 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#gnu-gas-assembler-comments */
 
-#include "common.h"
-LKMC_ENTRY
+#include <lkmc.h>
+LKMC_PROLOGUE
     # mycomment
 
     /* ARMv8 has // instead of @ as for comments. */
@@ -14,4 +14,4 @@ LKMC_ENTRY
     @ mycomment
     nop @ mycomment
 #endif
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/common_arch.h b/userland/arch/aarch64/common_arch.h
deleted file mode 100644
index cbc45ee..0000000
--- a/userland/arch/aarch64/common_arch.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly-c-standard-library */
-
-#ifndef COMMON_ARCH_H
-#define COMMON_ARCH_H
-
-#define LKMC_ASSERT_EQ(reg, const) \
-    mov x0, reg; \
-    ldr x1, =const; \
-    LKMC_ASSERT_EQ_DO(64); \
-;
-
-#define LKMC_ASSERT_EQ_DO(bits) \
-    bl lkmc_assert_eq_ ## bits; \
-    cmp x0, 0; \
-    LKMC_ASSERT(beq); \
-;
-
-#define LKMC_ASSERT_EQ_REG(reg1, reg2) \
-    str reg2, [sp, -16]!; \
-    mov x0, reg1; \
-    ldr x1, [sp], 16; \
-    LKMC_ASSERT_EQ_DO(64); \
-;
-
-#define LKMC_ASSERT_EQ_REG_32(reg1, reg2) \
-    str reg2, [sp, -4]!; \
-    mov w0, reg1; \
-    ldr w1, [sp], 4; \
-    LKMC_ASSERT_EQ_DO(32); \
-;
-
-#define LKMC_ASSERT_MEMCMP(label1, label2, const_size) \
-    adr x0, label1; \
-    adr x1, label2; \
-    ldr x2, =const_size; \
-    bl lkmc_assert_memcmp; \
-    cmp x0, 0; \
-    LKMC_ASSERT(beq); \
-;
-
-#define LKMC_ENTRY \
-.text; \
-.global asm_main; \
-asm_main: \
-    sub sp, sp, 0xA0; \
-    stp x29, x30, [sp]; \
-    stp x27, x28, [sp, 0x10]; \
-    stp x25, x26, [sp, 0x20]; \
-    stp x23, x24, [sp, 0x30]; \
-    stp x21, x22, [sp, 0x40]; \
-    stp x19, x20, [sp, 0x50]; \
-    stp  x6,  x7, [sp, 0x60]; \
-    stp  x4,  x5, [sp, 0x70]; \
-    stp  x2,  x3, [sp, 0x80]; \
-    stp  x0,  x1, [sp, 0x90]; \
-asm_main_after_prologue: \
-;
-
-#define LKMC_EXIT \
-    mov w0, 0; \
-    mov w1, 0; \
-    b pass; \
-fail: \
-    ldr x1, [sp, 0x90]; \
-    str w0, [x1]; \
-    mov w0, 1; \
-pass: \
-    ldp x19, x20, [sp, 0x50]; \
-    ldp x21, x22, [sp, 0x40]; \
-    ldp x23, x24, [sp, 0x30]; \
-    ldp x25, x26, [sp, 0x20]; \
-    ldp x27, x28, [sp, 0x10]; \
-    ldp x29, x30, [sp]; \
-    add sp, sp, 0xA0; \
-    ret; \
-;
-
-#define LKMC_FAIL \
-    ldr w0, =__LINE__; \
-    b fail; \
-;
-
-#endif
diff --git a/userland/arch/aarch64/cset.S b/userland/arch/aarch64/cset.S
index f760c19..f555b2f 100644
--- a/userland/arch/aarch64/cset.S
+++ b/userland/arch/aarch64/cset.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-cset-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* Test values. */
     mov x19, 0
     mov x20, 1
@@ -10,19 +10,19 @@ LKMC_ENTRY
     /* eq is true, set x21 = 1. */
     cmp x19, x19
     cset x21, eq
-    LKMC_ASSERT_EQ(x21, 1)
+    LKMC_ASSERT_EQ(x21, =1)
 
     /* eq is false, set x21 = 0. */
     cmp x19, x20
     cset x21, eq
-    LKMC_ASSERT_EQ(x21, 0)
+    LKMC_ASSERT_EQ(x21, =0)
 
     /* Same for ne. */
     cmp x19, x19
     cset x21, ne
-    LKMC_ASSERT_EQ(x21, 0)
+    LKMC_ASSERT_EQ(x21, =0)
 
     cmp x19, x20
     cset x21, ne
-    LKMC_ASSERT_EQ(x21, 1)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x21, =1)
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/fadd_scalar.S b/userland/arch/aarch64/fadd_scalar.S
index 98e31af..31345b4 100644
--- a/userland/arch/aarch64/fadd_scalar.S
+++ b/userland/arch/aarch64/fadd_scalar.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#advanced-simd-instructions */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* 1.5 + 2.5 == 4.0
      * using 64-bit double immediates.
      */
@@ -57,4 +57,4 @@ my_double_sum_expect:
 #if 0
     fmov d0, 1.23456798
 #endif
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/fadd_vector.S b/userland/arch/aarch64/fadd_vector.S
index f1a72ee..d54415e 100644
--- a/userland/arch/aarch64/fadd_vector.S
+++ b/userland/arch/aarch64/fadd_vector.S
@@ -3,9 +3,9 @@
  * Add a bunch of floating point numbers in one go.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 .data
     input0_4s: .float 1.5, 2.5,  3.5,  4.5
     input1_4s: .float 5.5, 6.5,  7.5,  8.5
@@ -24,11 +24,11 @@ LKMC_ENTRY
     fadd v2. ## size, v0. ## size, v1. ## size; \
     adr x0, output; \
     st1 {v2. ## size}, [x0]; \
-    LKMC_ASSERT_MEMCMP(output, expect_ ## size, 0x10)
+    LKMC_ASSERT_MEMCMP(output, expect_ ## size, =0x10)
 
     /* 4x 32-bit */
     TEST(4s)
     /* 2x 64-bit */
     TEST(2d)
 #undef TEST
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/gas_data_sizes.S b/userland/arch/aarch64/gas_data_sizes.S
index 639e47c..45ecf72 100644
--- a/userland/arch/aarch64/gas_data_sizes.S
+++ b/userland/arch/aarch64/gas_data_sizes.S
@@ -1,13 +1,13 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#gnu-gas-assembler data sizes */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 #define ASSERT_DIFF(label1, label2, result) \
     adr x0, label1; \
     adr x1, label2; \
     sub x0, x1, x0; \
-    LKMC_ASSERT_EQ(x0, result)
+    LKMC_ASSERT_EQ(x0, =result)
 
     ASSERT_DIFF(mybyte, myword, 1)
     ASSERT_DIFF(myword, mylong, 4)
@@ -15,7 +15,7 @@ LKMC_ENTRY
     ASSERT_DIFF(myquad, myocta, 8)
     ASSERT_DIFF(myocta, theend, 16)
 #undef ASSERT_DIF
-LKMC_EXIT
+LKMC_EPILOGUE
 mybyte:
     .byte 0x12
 myword:
diff --git a/userland/arch/aarch64/immediates.S b/userland/arch/aarch64/immediates.S
index 6d055c8..1caf6c5 100644
--- a/userland/arch/aarch64/immediates.S
+++ b/userland/arch/aarch64/immediates.S
@@ -1,9 +1,9 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#gnu-gas-assembler-immediates */
 
-#include "common.h"
-LKMC_ENTRY
+#include <lkmc.h>
+LKMC_PROLOGUE
     mov x0, 1
     mov x0, 0x1
     mov x0, 1
     mov x0, 0x1
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/ld2.S b/userland/arch/aarch64/ld2.S
index 9f8adb7..16f73d4 100644
--- a/userland/arch/aarch64/ld2.S
+++ b/userland/arch/aarch64/ld2.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-ld2-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 .data
     u32_interleave: .word \
         0x11111111, 0x55555555, \
@@ -22,5 +22,5 @@ LKMC_ENTRY
     add v2.4s, v0.4s, v1.4s
     adr x0, u32_interleave_sum
     st1 {v2.4s}, [x0]
-    LKMC_ASSERT_MEMCMP(u32_interleave_sum, u32_interleave_sum_expect, 0x10)
-LKMC_EXIT
+    LKMC_ASSERT_MEMCMP(u32_interleave_sum, u32_interleave_sum_expect, =0x10)
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/lkmc_assert_eq_fail.S b/userland/arch/aarch64/lkmc_assert_eq_fail.S
new file mode 100644
index 0000000..82ffe6c
--- /dev/null
+++ b/userland/arch/aarch64/lkmc_assert_eq_fail.S
@@ -0,0 +1,35 @@
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly */
+
+#include <lkmc.h>
+
+LKMC_PROLOGUE
+    ldr x19, =0x123456789ABCDEF0
+    ldr x20, =0x123456789ABCDEF0
+    ldr x21, =0x000000009ABCDEF0
+
+    /* Passing examples. */
+
+    /* Register immediate. */
+    LKMC_ASSERT_EQ(x19, =0x123456789ABCDEF0)
+#if 0
+    /* Syntax not supported. */
+    LKMC_ASSERT_EQ(=0x123456789ABCDEF0, x19)
+#endif
+
+    /* Register register. */
+    LKMC_ASSERT_EQ_REG(x19, x20)
+    LKMC_ASSERT_EQ_REG(x20, x19)
+    LKMC_ASSERT_EQ_REG_32(w19, w21)
+    LKMC_ASSERT_EQ_REG_32(w21, w19)
+
+    /* Register memory. */
+    LKMC_ASSERT_EQ(x19, myvar)
+#if 0
+    /* Syntax not supported. */
+    LKMC_ASSERT_EQ(myvar, x19)
+#endif
+
+    /* Now let's fail. */
+    LKMC_ASSERT_EQ(x19, =0x123456789ABCDEF1)
+LKMC_EPILOGUE
+myvar: .quad 0x123456789ABCDEF0
diff --git a/userland/arch/aarch64/lkmc_assert_memcmp_fail.S b/userland/arch/aarch64/lkmc_assert_memcmp_fail.S
new file mode 100644
index 0000000..338aa3d
--- /dev/null
+++ b/userland/arch/aarch64/lkmc_assert_memcmp_fail.S
@@ -0,0 +1,15 @@
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly */
+
+#include <lkmc.h>
+
+LKMC_PROLOGUE
+    /* Pass */
+    LKMC_ASSERT_MEMCMP(var0, var1, =0x10)
+    LKMC_ASSERT_MEMCMP(var0, var1, size)
+    /* Fail */
+    LKMC_ASSERT_MEMCMP(var0, var2, =0x10)
+LKMC_EPILOGUE
+    var0: .long 0x11111111, 0x22222222, 0x33333333, 0x44444444
+    var1: .long 0x11111111, 0x22222222, 0x33333333, 0x44444444
+    var2: .long 0x11111111, 0x22222223, 0x23333333, 0x44444444
+    size: .quad 0x10
diff --git a/userland/arch/aarch64/movk.S b/userland/arch/aarch64/movk.S
index 3cafad3..c368573 100644
--- a/userland/arch/aarch64/movk.S
+++ b/userland/arch/aarch64/movk.S
@@ -1,13 +1,13 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#armv8-aarch64-movk-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     movk x0, 0x4444, lsl 0
     movk x0, 0x3333, lsl 16
     movk x0, 0x2222, lsl 32
     movk x0, 0x1111, lsl 48
-    LKMC_ASSERT_EQ(x0, 0x1111222233334444)
+    LKMC_ASSERT_EQ(x0, =0x1111222233334444)
 
     /* Set a label (addresses are 48-bit) with immediates:
      *
@@ -23,4 +23,4 @@ LKMC_ENTRY
     adr x1, label
 label:
     LKMC_ASSERT_EQ_REG(x0, x1)
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/movn.S b/userland/arch/aarch64/movn.S
index 2d24a1e..54f97a6 100644
--- a/userland/arch/aarch64/movn.S
+++ b/userland/arch/aarch64/movn.S
@@ -1,9 +1,9 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#armv8-aarch64-movn-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr x0, =0x123456789ABCDEF0
     movn x0, 0x8888, lsl 16
-    LKMC_ASSERT_EQ(x0, 0xFFFFFFFF7777FFFF)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x0, =0xFFFFFFFF7777FFFF)
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/pc.S b/userland/arch/aarch64/pc.S
index a203c4b..bcd308e 100644
--- a/userland/arch/aarch64/pc.S
+++ b/userland/arch/aarch64/pc.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#assembly-registers */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 #if 0
     /* Unlike v7, we can't use PC like any other register in ARMv8,
      * since it is not a general purpose register anymore.
@@ -16,7 +16,7 @@ LKMC_ENTRY
      * exception return.
      */
     ldr pc, =10f
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 10:
 #endif
 #if 0
@@ -32,7 +32,7 @@ LKMC_ENTRY
 pc_relative_ldr:
     .quad 0x123456789ABCDEF0
 1:
-    LKMC_ASSERT_EQ(x0, 0x123456789ABCDEF0)
+    LKMC_ASSERT_EQ(x0, =0x123456789ABCDEF0)
 
     /* Just for fun, we can also use relative numbers instead of labels.
      * https://reverseengineering.stackexchange.com/questions/17666/how-does-the-ldr-instruction-work-on-arm/20567#20567
@@ -41,14 +41,14 @@ pc_relative_ldr:
     b 1f
     .quad 0x123456789ABCDEF0
 1:
-    LKMC_ASSERT_EQ(x0, 0x123456789ABCDEF0)
+    LKMC_ASSERT_EQ(x0, =0x123456789ABCDEF0)
 
     /* Analogous for b with PC. */
     mov x0, 0
     /* Jumps over mov to LKMC_ASSERT_EQ. */
     b 8
     mov x0, 1
-    LKMC_ASSERT_EQ(x0, 0)
+    LKMC_ASSERT_EQ(x0, =0)
 
     /* Trying to use the old "LDR (immediate)" PC-relative
      * syntax does not work.
@@ -66,13 +66,13 @@ pc_relative_ldr:
 
     /* You just have to use adr + "STR (register)". */
     ldr x0, pc_relative_str
-    LKMC_ASSERT_EQ(x0, 0x0)
+    LKMC_ASSERT_EQ(x0, =0x0)
     adr x1, pc_relative_str
     ldr x0, pc_relative_ldr
     str x0, [x1]
     ldr x0, pc_relative_str
-    LKMC_ASSERT_EQ(x0, 0x123456789ABCDEF0)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x0, =0x123456789ABCDEF0)
+LKMC_EPILOGUE
 .data
 pc_relative_str:
     .quad 0x0000000000000000
diff --git a/userland/arch/aarch64/registers.S b/userland/arch/aarch64/registers.S
index b13527e..fa0c8a2 100644
--- a/userland/arch/aarch64/registers.S
+++ b/userland/arch/aarch64/registers.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#assembly-registers */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* 31 64-bit eXtended general purpose registers. */
     mov x0, 0
@@ -43,5 +43,5 @@ LKMC_ENTRY
     ldr x0, =0x1111222233334444
     ldr x1, =0x5555666677778888
     mov w0, w1
-    LKMC_ASSERT_EQ(x0, 0x0000000077778888)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x0, =0x0000000077778888)
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/ret.S b/userland/arch/aarch64/ret.S
index cd578ba..906e8a1 100644
--- a/userland/arch/aarch64/ret.S
+++ b/userland/arch/aarch64/ret.S
@@ -1,16 +1,16 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#armv8-aarch64-ret-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     mov x19, 1
     bl inc
-    LKMC_ASSERT_EQ(x19, 2)
+    LKMC_ASSERT_EQ(x19, =2)
     bl inc2
-    LKMC_ASSERT_EQ(x19, 3)
+    LKMC_ASSERT_EQ(x19, =3)
     bl inc3
-    LKMC_ASSERT_EQ(x19, 4)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x19, =4)
+LKMC_EPILOGUE
 
 /* void inc(uint64_t *i) { (*i)++ } */
 inc:
diff --git a/userland/arch/aarch64/str.S b/userland/arch/aarch64/str.S
index 4c15bb8..6846581 100644
--- a/userland/arch/aarch64/str.S
+++ b/userland/arch/aarch64/str.S
@@ -1,13 +1,13 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#armv8-aarch64-str-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr x0, myvar
-    LKMC_ASSERT_EQ(x0, 0x12346789ABCDEF0)
+    LKMC_ASSERT_EQ(x0, =0x12346789ABCDEF0)
 #if 0
     /* Error: invalid addressing mode at operand 2 -- `str x0,myvar' */
     str x0, myvar
 #endif
-LKMC_EXIT
+LKMC_EPILOGUE
     myvar: .quad 0x12346789ABCDEF0
diff --git a/userland/arch/aarch64/ubfm.S b/userland/arch/aarch64/ubfm.S
index d23ce92..59cc290 100644
--- a/userland/arch/aarch64/ubfm.S
+++ b/userland/arch/aarch64/ubfm.S
@@ -1,17 +1,17 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-ubfm-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr x19, =0x1122334455667788
 
     // lsr alias: imms == 63
 
     ldr x20, =0xFFFFFFFFFFFFFFFF
     ubfm x20, x19, 16, 63
-    LKMC_ASSERT_EQ(x20, 0x0000112233445566)
+    LKMC_ASSERT_EQ(x20, =0x0000112233445566)
 
     ldr x20, =0xFFFFFFFFFFFFFFFF
     ubfm x20, x19, 32, 63
-    LKMC_ASSERT_EQ(x20, 0x0000000011223344)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x20, =0x0000000011223344)
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/ubfx.S b/userland/arch/aarch64/ubfx.S
index 9b25760..54f9004 100644
--- a/userland/arch/aarch64/ubfx.S
+++ b/userland/arch/aarch64/ubfx.S
@@ -1,15 +1,15 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-ubfx-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr x19, =0x1122334455667788
 
     ldr x20, =0xFFFFFFFFFFFFFFFF
     ubfx x20, x19, 8, 16
-    LKMC_ASSERT_EQ(x20, 0x0000000000006677)
+    LKMC_ASSERT_EQ(x20, =0x0000000000006677)
 
     ldr x20, =0xFFFFFFFFFFFFFFFF
     ubfx x20, x19, 8, 32
-    LKMC_ASSERT_EQ(x20, 0x0000000044556677)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(x20, =0x0000000044556677)
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/udf.S b/userland/arch/aarch64/udf.S
index 50852c7..e76cc84 100644
--- a/userland/arch/aarch64/udf.S
+++ b/userland/arch/aarch64/udf.S
@@ -1,10 +1,10 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-udf-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     .long 0
 #if 0
     udf 0
 #endif
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/aarch64/x31.S b/userland/arch/aarch64/x31.S
index 6b73183..e89af76 100644
--- a/userland/arch/aarch64/x31.S
+++ b/userland/arch/aarch64/x31.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#armv8-aarch64-x31-register */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* ERROR: can never use the name x31. */
 #if 0
     mov x31, 31
@@ -11,12 +11,12 @@ LKMC_ENTRY
     /* mov (register) is an alias for ORR, which accepts xzr. */
     mov x19, 1
     mov x19, xzr
-    LKMC_ASSERT_EQ(x19, 0)
+    LKMC_ASSERT_EQ(x19, =0)
 
     /* Same encoding as the mov version. */
     mov x19, 1
     orr x19, xzr, xzr
-    LKMC_ASSERT_EQ(x19, 0)
+    LKMC_ASSERT_EQ(x19, =0)
 
     /* So, orr, which is not an alias, can only take xzr, not sp. */
 #if 0
@@ -26,7 +26,7 @@ LKMC_ENTRY
     /* Zero register discards result if written to. */
     mov x19, 1
     orr xzr, x19, x19
-    LKMC_ASSERT_EQ(xzr, 0)
+    LKMC_ASSERT_EQ(xzr, =0)
 
     /* MOV (to/from SP) is an alias for ADD (immediate). */
     mov x19, sp
@@ -36,7 +36,7 @@ LKMC_ENTRY
     /* Exact same encoding as above. */
     add x20, sp, 0
     mov sp, x19
-    LKMC_ASSERT_EQ(x20, 1)
+    LKMC_ASSERT_EQ(x20, =1)
 
     /* So, ADD (immediate), which is not an alias, can only take sp, not xzr. */
 #if 0
@@ -48,4 +48,4 @@ LKMC_ENTRY
      * does not say anything about SP, and so does accept xzr just fine.
      */
     add xzr, xzr, xzr
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/add.S b/userland/arch/arm/add.S
index 1968fb5..f719254 100644
--- a/userland/arch/arm/add.S
+++ b/userland/arch/arm/add.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-data-processing-instructions */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Immediate encoding.
      *
@@ -11,7 +11,7 @@ LKMC_ENTRY
     mov r0, 1
     /* r1 = r0 + 2 */
     add r1, r0, 2
-    LKMC_ASSERT_EQ(r1, 3)
+    LKMC_ASSERT_EQ(r1, =3)
 
     /* If src == dest, we can omit one of them.
      *
@@ -19,12 +19,12 @@ LKMC_ENTRY
      */
     mov r0, 1
     add r0, 2
-    LKMC_ASSERT_EQ(r0, 3)
+    LKMC_ASSERT_EQ(r0, =3)
 
     /* Same as above but explicit. */
     mov r0, 1
     add r0, r0, 2
-    LKMC_ASSERT_EQ(r0, 3)
+    LKMC_ASSERT_EQ(r0, =3)
 
 #if 0
     /* But we cannot omit the register if there is a shift when using .syntx unified:
@@ -44,7 +44,7 @@ LKMC_ENTRY
     mov r0, 1
     mov r1, 2
     add r2, r0, r1
-    LKMC_ASSERT_EQ(r2, 3)
+    LKMC_ASSERT_EQ(r2, =3)
 
     /* Register encoding, omit implicit register.
      *
@@ -53,6 +53,6 @@ LKMC_ENTRY
     mov r0, 1
     mov r1, 2
     add r1, r0
-    LKMC_ASSERT_EQ(r1, 3)
+    LKMC_ASSERT_EQ(r1, =3)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/address_modes.S b/userland/arch/arm/address_modes.S
index a91f19f..376014d 100644
--- a/userland/arch/arm/address_modes.S
+++ b/userland/arch/arm/address_modes.S
@@ -1,37 +1,37 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-addressing-modes */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Offset mode with immediate. Add 4 to the address register,
      * which ends up * reading myvar6 instead of myvar.
      */
     adr r4, myvar
     ldr r5, [r4, 4]
-    LKMC_ASSERT_EQ(r5, 0x9ABCDEF0)
+    LKMC_ASSERT_EQ(r5, =0x9ABCDEF0)
     /* r4 was not modified. */
-    LKMC_ASSERT_EQ(r4, myvar)
+    LKMC_ASSERT_EQ(r4, =myvar)
 
     /* Pre-indexed mode: modify register, then use it. */
     adr r4, myvar
     ldr r5, [r4, 4]!
-    LKMC_ASSERT_EQ(r5, 0x9ABCDEF0)
+    LKMC_ASSERT_EQ(r5, =0x9ABCDEF0)
     /* r4 was modified. */
-    LKMC_ASSERT_EQ(r4, myvar6)
+    LKMC_ASSERT_EQ(r4, =myvar6)
 
     /* Post-indexed mode: use register, then modify it. */
     adr r4, myvar
     ldr r5, [r4], 4
-    LKMC_ASSERT_EQ(r5, 0x12345678)
+    LKMC_ASSERT_EQ(r5, =0x12345678)
     /* r4 was modified. */
-    LKMC_ASSERT_EQ(r4, myvar6)
+    LKMC_ASSERT_EQ(r4, =myvar6)
 
     /* Offset in register. */
     adr r4, myvar
     mov r5, 4
     ldr r6, [r4, r5]
-    LKMC_ASSERT_EQ(r6, 0x9ABCDEF0)
+    LKMC_ASSERT_EQ(r6, =0x9ABCDEF0)
 
     /* Offset in shifted register:
      * r6 =
@@ -42,9 +42,9 @@ LKMC_ENTRY
     adr r4, myvar
     mov r5, 2
     ldr r6, [r4, r5, lsl 1]
-    LKMC_ASSERT_EQ(r6, 0x9ABCDEF0)
+    LKMC_ASSERT_EQ(r6, =0x9ABCDEF0)
 
-LKMC_EXIT
+LKMC_EPILOGUE
 myvar:
     .word 0x12345678
 myvar6:
diff --git a/userland/arch/arm/adr.S b/userland/arch/arm/adr.S
index 019a0ba..cb6a130 100644
--- a/userland/arch/arm/adr.S
+++ b/userland/arch/arm/adr.S
@@ -1,11 +1,11 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-adr-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
 .data
 data_label:
     .word 0x1234678
-LKMC_ENTRY
+LKMC_PROLOGUE
     adr r4, label
     /* objdump tells us that this uses the literal pool,
      * it does not get converted to adr, which is the better
@@ -30,4 +30,4 @@ label:
      */
     adr r5, data_label
 #endif
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/and.S b/userland/arch/arm/and.S
index dda36f4..9903c5d 100644
--- a/userland/arch/arm/and.S
+++ b/userland/arch/arm/and.S
@@ -1,27 +1,27 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-bitwise-instructions */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* 0x00 && 0xFF == 0x00 */
     mov r0, 0x00
     and r0, 0xFF
-    LKMC_ASSERT_EQ(r0, 0x00)
+    LKMC_ASSERT_EQ(r0, =0x00)
 
     /* 0x0F && 0xF0 == 0x00 */
     mov r0, 0x0F
     and r0, 0xF0
-    LKMC_ASSERT_EQ(r0, 0x00)
+    LKMC_ASSERT_EQ(r0, =0x00)
 
     /* 0x0F && 0xFF == 0x0F */
     mov r0, 0x0F
     and r0, 0xFF
-    LKMC_ASSERT_EQ(r0, 0x0F)
+    LKMC_ASSERT_EQ(r0, =0x0F)
 
     /* 0xF0 && 0xFF == 0xF0 */
     mov r0, 0xF0
     and r0, 0xFF
-    LKMC_ASSERT_EQ(r0, 0xF0)
+    LKMC_ASSERT_EQ(r0, =0xF0)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/b.S b/userland/arch/arm/b.S
index 4650fc7..0e6af51 100644
--- a/userland/arch/arm/b.S
+++ b/userland/arch/arm/b.S
@@ -1,9 +1,9 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-b-instruction */
 
-#include "common.h"
-LKMC_ENTRY
+#include <lkmc.h>
+LKMC_PROLOGUE
     /* Jump over the fail. 26-bit PC-relative. */
     b ok
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 ok:
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/beq.S b/userland/arch/arm/beq.S
index e2b8503..b347d7c 100644
--- a/userland/arch/arm/beq.S
+++ b/userland/arch/arm/beq.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-beq-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Smaller*/
     mov r0, 1
@@ -25,4 +25,4 @@ LKMC_ENTRY
     LKMC_ASSERT(bgt)
     LKMC_ASSERT(bne)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/bfi.S b/userland/arch/arm/bfi.S
index 5b21674..bf39168 100644
--- a/userland/arch/arm/bfi.S
+++ b/userland/arch/arm/bfi.S
@@ -1,10 +1,10 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-bfi-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr r0, =0x11223344
     ldr r1, =0xFFFFFFFF
     bfi r1, r0, 8, 16
-    LKMC_ASSERT_EQ(r1, 0xFF3344FF)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r1, =0xFF3344FF)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/bic.S b/userland/arch/arm/bic.S
index 3f752f0..20ae458 100644
--- a/userland/arch/arm/bic.S
+++ b/userland/arch/arm/bic.S
@@ -1,10 +1,10 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-bic-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* 0x0F & ~0x55 == 0x0F & 0xAA == 0x0A */
     mov r0, 0x0F
     bic r0, 0x55
-    LKMC_ASSERT_EQ(r0, 0x0A)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r0, =0x0A)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/bl.S b/userland/arch/arm/bl.S
index 0116784..e79adb2 100644
--- a/userland/arch/arm/bl.S
+++ b/userland/arch/arm/bl.S
@@ -1,12 +1,12 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-bl-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     mov r0, 1
     bl inc
-    LKMC_ASSERT_EQ(r0, 2)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r0, =2)
+LKMC_EPILOGUE
 
 /* void inc(int *i) { (*i)++ } */
 inc:
diff --git a/userland/arch/arm/clz.S b/userland/arch/arm/clz.S
index 2eb6b96..7d96ae6 100644
--- a/userland/arch/arm/clz.S
+++ b/userland/arch/arm/clz.S
@@ -1,17 +1,17 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#data-bitwise-instructions */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr r0, =0x7FFFFFFF
     clz r1, r0
-    LKMC_ASSERT_EQ(r1, 1)
+    LKMC_ASSERT_EQ(r1, =1)
 
     ldr r0, =0x3FFFFFFF
     clz r1, r0
-    LKMC_ASSERT_EQ(r1, 2)
+    LKMC_ASSERT_EQ(r1, =2)
 
     ldr r0, =0x1FFFFFFF
     clz r1, r0
-    LKMC_ASSERT_EQ(r1, 3)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r1, =3)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/comments.S b/userland/arch/arm/comments.S
index 100cb62..d45a5a8 100644
--- a/userland/arch/arm/comments.S
+++ b/userland/arch/arm/comments.S
@@ -1,7 +1,7 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#gnu-gas-assembler-comments */
 
-#include "common.h"
-LKMC_ENTRY
+#include <lkmc.h>
+LKMC_PROLOGUE
     # mycomment
     @ mycomment
     /* # only works at the beginning of the line.
@@ -11,4 +11,4 @@ LKMC_ENTRY
     nop # mycomment
 #endif
     nop @ mycomment
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/common_arch.h b/userland/arch/arm/common_arch.h
deleted file mode 100644
index 118df45..0000000
--- a/userland/arch/arm/common_arch.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly-c-standard-library */
-
-#ifndef COMMON_ARCH_H
-#define COMMON_ARCH_H
-
-.syntax unified
-
-/* Assert that a register equals a constant.
- * * reg: the register to check
- * * const: the constant to compare to. Only works for literals or labels, not for registers.
- *          For register / register comparison, use LKMC_ASSERT_EQ_REG.
- */
-#define LKMC_ASSERT_EQ(reg, const) \
-    mov r0, reg; \
-    ldr r1, =const; \
-    LKMC_ASSERT_EQ_DO; \
-;
-
-#define LKMC_ASSERT_EQ_DO \
-    bl lkmc_assert_eq_32; \
-    cmp r0, 0; \
-    LKMC_ASSERT(beq); \
-;
-
-#define LKMC_ASSERT_EQ_REG(reg1, reg2) \
-    str reg2, [sp, -4]!; \
-    mov r0, reg1; \
-    ldr r1, [sp], 4; \
-    LKMC_ASSERT_EQ_DO; \
-;
-
-/* Assert that two arrays are the same. */
-#define LKMC_ASSERT_MEMCMP(label1, label2, const_size) \
-    ldr r0, =label1; \
-    ldr r1, =label2; \
-    ldr r2, =const_size; \
-    bl lkmc_assert_memcmp; \
-    cmp r0, 0; \
-    LKMC_ASSERT(beq); \
-;
-
-/* Store all callee saved registers, and LR in case we make further BL calls.
- *
- * Also save the input arguments r0-r3 on the stack, so we can access them later on,
- * despite those registers being overwritten.
- */
-#define LKMC_ENTRY \
-.text; \
-.global asm_main; \
-asm_main: \
-    stmdb sp!, {r0-r12, lr}; \
-asm_main_after_prologue: \
-;
-
-/* Meant to be called at the end of LKMC_ENTRY.*
- *
- * Branching to "fail" makes tests fail with exit status 1.
- *
- * If LKMC_EXIT is reached, the program ends successfully.
- *
- * Restore LR and bx jump to it to return from asm_main.
- */
-#define LKMC_EXIT \
-    mov r0, 0; \
-    mov r1, 0; \
-    b pass; \
-fail: \
-    ldr r1, [sp]; \
-    str r0, [r1]; \
-    mov r0, 1; \
-pass: \
-    add sp, 16; \
-    ldmia sp!, {r4-r12, lr}; \
-    bx lr; \
-;
-
-/* Always fail. */
-#define LKMC_FAIL \
-    ldr r0, =__LINE__; \
-    b fail; \
-;
-
-#endif
diff --git a/userland/arch/arm/cond.S b/userland/arch/arm/cond.S
index 954fb0b..58d0672 100644
--- a/userland/arch/arm/cond.S
+++ b/userland/arch/arm/cond.S
@@ -1,16 +1,18 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-conditional-execution */
 
-#include "common.h"
+#include <lkmc.h>
+
+LKMC_PROLOGUE
+    mov r4, 0
+    mov r5, 1
 
-LKMC_ENTRY
-    mov r0, 0
-    mov r1, 1
-    cmp r0, 1
     /* Previous cmp failed, skip this operation. */
-    addeq r1, 1
-    LKMC_ASSERT_EQ(r1, 1)
-    cmp r0, 0
+    cmp r4, 1
+    addeq r5, 1
+    LKMC_ASSERT_EQ(r5, =1)
+
     /* Previous passed, do this operation. */
-    addeq r1, 1
-    LKMC_ASSERT_EQ(r1, 2)
-LKMC_EXIT
+    cmp r4, 0
+    addeq r5, 1
+    LKMC_ASSERT_EQ(r5, =2)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/gas_data_sizes.S b/userland/arch/arm/gas_data_sizes.S
index c5c3925..1edcf1d 100644
--- a/userland/arch/arm/gas_data_sizes.S
+++ b/userland/arch/arm/gas_data_sizes.S
@@ -1,6 +1,6 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#gnu-gas-assembler data sizes */
 
-#include "common.h"
+#include <lkmc.h>
 
 .data
 mybyte:
@@ -14,12 +14,12 @@ myquad:
 myocta:
     .octa 0x123456789ABCDEF0123456789ABCDEF0
 theend:
-LKMC_ENTRY
+LKMC_PROLOGUE
 #define ASSERT_DIFF(label1, label2, result) \
     ldr r0, =label1; \
     ldr r1, =label2; \
     sub r0, r1, r0; \
-    LKMC_ASSERT_EQ(r0, result)
+    LKMC_ASSERT_EQ(r0, =result)
 
     ASSERT_DIFF(mybyte, myword, 1)
     ASSERT_DIFF(myword, mylong, 4)
@@ -27,4 +27,4 @@ LKMC_ENTRY
     ASSERT_DIFF(myquad, myocta, 8)
     ASSERT_DIFF(myocta, theend, 16)
 #undef ASSERT_DIF
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/immediates.S b/userland/arch/arm/immediates.S
index 8d45778..3907f03 100644
--- a/userland/arch/arm/immediates.S
+++ b/userland/arch/arm/immediates.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-gnu-instruction-gas-assembler-immediates */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* This is the default. We hack it in common.h however. */
 .syntax divided
    /* These fail. */
@@ -21,4 +21,4 @@ LKMC_ENTRY
     mov r0, 0x1
     mov r0, $1
     mov r0, $0x1
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/inc_array.S b/userland/arch/arm/inc_array.S
index 15982c2..5433aa5 100644
--- a/userland/arch/arm/inc_array.S
+++ b/userland/arch/arm/inc_array.S
@@ -1,6 +1,6 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-loop-instruction-over-array */
 
-#include "common.h"
+#include <lkmc.h>
 
 #define NELEM 4
 #define ELEM_SIZE 4
@@ -11,7 +11,7 @@ my_array:
 my_array_expect:
     .word 0x11111112, 0x22222223, 0x33333334, 0x44444445
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* Increment. */
     ldr r0, =my_array
     mov r1, NELEM
@@ -23,5 +23,5 @@ increment:
     sub r1, 1
     cmp r1, 0
     bne increment
-    LKMC_ASSERT_MEMCMP(my_array, my_array_expect, 0x10)
-LKMC_EXIT
+    LKMC_ASSERT_MEMCMP(my_array, my_array_expect, =0x10)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/ldmia.S b/userland/arch/arm/ldmia.S
index 1d32b83..609f876 100644
--- a/userland/arch/arm/ldmia.S
+++ b/userland/arch/arm/ldmia.S
@@ -1,6 +1,6 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-loop-instruction-over-array */
 
-#include "common.h"
+#include <lkmc.h>
 
 #define NELEM 4
 #define ELEM_SIZE 4
@@ -11,7 +11,7 @@ my_array_0:
 my_array_1:
     .word 0x55555555, 0x66666666, 0x77777777, 0x88888888
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Load r5, r6, r7 and r8 starting from the address in r4. Don't change r4 */
     ldr r4, =my_array_0
@@ -20,11 +20,11 @@ LKMC_ENTRY
     ldr r7, =0
     ldr r8, =0
     ldmia r4, {r5-r8}
-    LKMC_ASSERT_EQ(r4, my_array_0)
-    LKMC_ASSERT_EQ(r5, 0x11111111)
-    LKMC_ASSERT_EQ(r6, 0x22222222)
-    LKMC_ASSERT_EQ(r7, 0x33333333)
-    LKMC_ASSERT_EQ(r8, 0x44444444)
+    LKMC_ASSERT_EQ(r4, =my_array_0)
+    LKMC_ASSERT_EQ(r5, =0x11111111)
+    LKMC_ASSERT_EQ(r6, =0x22222222)
+    LKMC_ASSERT_EQ(r7, =0x33333333)
+    LKMC_ASSERT_EQ(r8, =0x44444444)
 
     /* Swapping the order of r5 and r6 on the mnemonic makes no difference to load order.
      *
@@ -38,8 +38,8 @@ LKMC_ENTRY
     ldr r5, =0
     ldr r6, =0
     ldmia r4, {r6,r5}
-    LKMC_ASSERT_EQ(r5, 0x11111111)
-    LKMC_ASSERT_EQ(r6, 0x22222222)
+    LKMC_ASSERT_EQ(r5, =0x11111111)
+    LKMC_ASSERT_EQ(r6, =0x22222222)
 #endif
 
     /* Modify the array */
@@ -51,11 +51,11 @@ LKMC_ENTRY
     stmdb r4, {r5-r8}
 
     /* Verify that my_array_0 changed and is equal to my_array_1. */
-    LKMC_ASSERT_MEMCMP(my_array_0, my_array_1, 0x10)
+    LKMC_ASSERT_MEMCMP(my_array_0, my_array_1, =0x10)
 
     /* Load registers and increment r4. */
     ldr r4, =my_array_0
     ldmia r4!, {r5-r8}
-    LKMC_ASSERT_EQ(r4, my_array_1)
+    LKMC_ASSERT_EQ(r4, =my_array_1)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/ldr_pseudo.S b/userland/arch/arm/ldr_pseudo.S
index 03fb74e..9a21b05 100644
--- a/userland/arch/arm/ldr_pseudo.S
+++ b/userland/arch/arm/ldr_pseudo.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-ldr-instruction-pseudo-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Mnemonic for a PC relative load:
      *
@@ -12,13 +12,13 @@ LKMC_ENTRY
      * ....
      */
     ldr r0, myvar
-    LKMC_ASSERT_EQ(r0, 0x12345678)
+    LKMC_ASSERT_EQ(r0, =0x12345678)
 
     /* Mnemonic PC relative load with an offset.
      * Load myvar2 instead of myvar.
      */
     ldr r0, myvar + 4
-    LKMC_ASSERT_EQ(r0, 0x9ABCDEF0)
+    LKMC_ASSERT_EQ(r0, =0x9ABCDEF0)
 
     /* First store the address in r0 using a magic =myvar, which creates
      * a new variable containing the address and PC-relative addresses it
@@ -33,17 +33,17 @@ LKMC_ENTRY
      */
     ldr r0, =myvar
     ldr r1, [r0]
-    LKMC_ASSERT_EQ(r1, 0x12345678)
+    LKMC_ASSERT_EQ(r1, =0x12345678)
 
     /* More efficiently, use r0 as the address to read, and write to r0 itself. */
     ldr r0, =myvar
     ldr r0, [r0]
-    LKMC_ASSERT_EQ(r0, 0x12345678)
+    LKMC_ASSERT_EQ(r0, =0x12345678)
 
     /* Same as =myvar but store a constant to a register.
      * Can also be done with movw and movt. */
     ldr r0, =0x11112222
-    LKMC_ASSERT_EQ(r0, 0x11112222)
+    LKMC_ASSERT_EQ(r0, =0x11112222)
 
     /* We can also use GAS tolower16 and topper16  and movw and movt
      * to load the address of myvar into r0 with two immediates.
@@ -56,9 +56,9 @@ LKMC_ENTRY
     movw r0, #:lower16:myvar
     movt r0, #:upper16:myvar
     ldr r1, [r0]
-    LKMC_ASSERT_EQ(r1, 0x12345678)
+    LKMC_ASSERT_EQ(r1, =0x12345678)
 
-LKMC_EXIT
+LKMC_EPILOGUE
 myvar:
     .word 0x12345678
 myvar2:
diff --git a/userland/arch/arm/ldrb.S b/userland/arch/arm/ldrb.S
index 0a19cfe..556ed68 100644
--- a/userland/arch/arm/ldrb.S
+++ b/userland/arch/arm/ldrb.S
@@ -1,12 +1,12 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-ldrh-instruction-and-ldrb */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr r0, =myvar
     mov r1, 0x0
     ldrb r1, [r0]
-    LKMC_ASSERT_EQ(r1, 0x00000078)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r1, =0x00000078)
+LKMC_EPILOGUE
 myvar:
     .word 0x12345678
diff --git a/userland/arch/arm/ldrh.S b/userland/arch/arm/ldrh.S
index 5ae4a53..e4bf7dc 100644
--- a/userland/arch/arm/ldrh.S
+++ b/userland/arch/arm/ldrh.S
@@ -1,12 +1,12 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-ldrh-instruction-and-ldrb */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr r0, =myvar
     mov r1, 0x0
     ldrh r1, [r0]
-    LKMC_ASSERT_EQ(r1, 0x00005678)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r1, =0x00005678)
+LKMC_EPILOGUE
 myvar:
     .word 0x12345678
diff --git a/userland/arch/arm/linux/c_from_asm.S b/userland/arch/arm/linux/c_from_asm.S
index 3ed3d9b..4c4baa5 100644
--- a/userland/arch/arm/linux/c_from_asm.S
+++ b/userland/arch/arm/linux/c_from_asm.S
@@ -1,6 +1,6 @@
-/* https://github.com/cirosantilli/linux-kernel-module-cheat#calling-convention */
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-calling-convention */
 
-#include "common.h"
+#include <lkmc.h>
 
 .data
 puts_s:
@@ -12,7 +12,7 @@ my_array_0:
 my_array_1:
     .word 0x55555555, 0x66666666, 0x77777777, 0x88888888
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* puts("hello world") */
     /* r0 is first argument. */
     ldr r0, =puts_s
@@ -49,11 +49,11 @@ LKMC_ENTRY
         ldr r1, =my_array_1
         ldr r2, =0x10
         bl memcmp
-        LKMC_ASSERT_EQ(r0, 0)
+        LKMC_ASSERT_EQ(r0, =0)
 
     /* exit(0) */
     mov r0, 0
     bl exit
 
     /* Never reached, just for the fail symbol. */
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/lkmc_assert_eq_fail.S b/userland/arch/arm/lkmc_assert_eq_fail.S
new file mode 100644
index 0000000..63d5ba1
--- /dev/null
+++ b/userland/arch/arm/lkmc_assert_eq_fail.S
@@ -0,0 +1,32 @@
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly */
+
+#include <lkmc.h>
+
+LKMC_PROLOGUE
+    ldr r4, =0x12345678
+    ldr r5, =0x12345678
+
+    /* Passing examples. */
+
+    /* Register immediate. */
+    LKMC_ASSERT_EQ(r4, =0x12345678)
+#if 0
+    /* Syntax not supported. */
+    LKMC_ASSERT_EQ(=0x12345678, r4)
+#endif
+
+    /* Register register. */
+    LKMC_ASSERT_EQ_REG(r4, r5)
+    LKMC_ASSERT_EQ_REG(r5, r4)
+
+    /* Register memory. */
+    LKMC_ASSERT_EQ(r4, myvar)
+#if 0
+    /* Syntax not supported. */
+    LKMC_ASSERT_EQ(myvar, r4)
+#endif
+
+    /* Now let's fail. */
+    LKMC_ASSERT_EQ(r4, =0x12345679)
+LKMC_EPILOGUE
+myvar: .quad 0x12345678
diff --git a/userland/arch/arm/lkmc_assert_memcmp_fail.S b/userland/arch/arm/lkmc_assert_memcmp_fail.S
new file mode 100644
index 0000000..650a495
--- /dev/null
+++ b/userland/arch/arm/lkmc_assert_memcmp_fail.S
@@ -0,0 +1,15 @@
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly */
+
+#include <lkmc.h>
+
+LKMC_PROLOGUE
+    /* Pass */
+    LKMC_ASSERT_MEMCMP(var0, var1, =0x10)
+    LKMC_ASSERT_MEMCMP(var0, var1, size)
+    /* Fail */
+    LKMC_ASSERT_MEMCMP(var0, var2, =0x10)
+LKMC_EPILOGUE
+    var0: .long 0x11111111, 0x22222222, 0x33333333, 0x44444444
+    var1: .long 0x11111111, 0x22222222, 0x33333333, 0x44444444
+    var2: .long 0x11111111, 0x22222223, 0x23333333, 0x44444444
+    size: .long 0x10
diff --git a/userland/arch/arm/mov.S b/userland/arch/arm/mov.S
index 8278d70..65bb716 100644
--- a/userland/arch/arm/mov.S
+++ b/userland/arch/arm/mov.S
@@ -1,19 +1,19 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-mov-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Immediate. */
     mov r0, 0
-    LKMC_ASSERT_EQ(r0, 0)
+    LKMC_ASSERT_EQ(r0, =0)
     mov r0, 1
-    LKMC_ASSERT_EQ(r0, 1)
+    LKMC_ASSERT_EQ(r0, =1)
 
     /* Register. */
     mov r0, 0
     mov r1, 1
     mov r1, r0
-    LKMC_ASSERT_EQ(r1, 0)
+    LKMC_ASSERT_EQ(r1, =0)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/movw.S b/userland/arch/arm/movw.S
index 34166a9..e653497 100644
--- a/userland/arch/arm/movw.S
+++ b/userland/arch/arm/movw.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-movw-and-movt-instructions */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* movt (top) and movw (TODO what is w) set the higher
      * and lower 16 bits of the register.
@@ -10,7 +10,7 @@ LKMC_ENTRY
     movw r0, 0xFFFF
     movt r0, 0x1234
     add r0, 1
-    LKMC_ASSERT_EQ(r0, 0x12350000)
+    LKMC_ASSERT_EQ(r0, =0x12350000)
 
     /* movw also zeroes out the top bits, allowing small 16-bit
      * C constants to be assigned in a single instruction.
@@ -22,6 +22,6 @@ LKMC_ENTRY
      */
     ldr r0, =0x12345678
     movw r0, 0x1111
-    LKMC_ASSERT_EQ(r0, 0x00001111)
+    LKMC_ASSERT_EQ(r0, =0x00001111)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/mul.S b/userland/arch/arm/mul.S
index afcfd51..f46ee04 100644
--- a/userland/arch/arm/mul.S
+++ b/userland/arch/arm/mul.S
@@ -3,13 +3,13 @@
  * Multiplication.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* 2 * 3 = 6 */
     mov r0, 0
     mov r1, 2
     mov r2, 3
     mul r1, r2
-    LKMC_ASSERT_EQ(r1, 6)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r1, =6)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/nop.S b/userland/arch/arm/nop.S
index 666a1d2..2d59bd1 100644
--- a/userland/arch/arm/nop.S
+++ b/userland/arch/arm/nop.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-nop-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* Disassembles as:
      *
      * ....
@@ -29,4 +29,4 @@ LKMC_ENTRY
 
     /* And there are other nops as well? Disassembles as `and`. */
     and r0, r0, r0
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/push.S b/userland/arch/arm/push.S
index 30ea50d..9e9acee 100644
--- a/userland/arch/arm/push.S
+++ b/userland/arch/arm/push.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-ldmia-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Save sp before push. */
     mov r4, sp
@@ -19,13 +19,13 @@ LKMC_ENTRY
     mov r7, 0
     mov r8, 0
     pop {r7, r8}
-    LKMC_ASSERT_EQ(r7, 1)
-    LKMC_ASSERT_EQ(r8, 2)
+    LKMC_ASSERT_EQ(r7, =1)
+    LKMC_ASSERT_EQ(r8, =2)
 
     /* Check that stack pointer moved down by 8 bytes
      * (2 registers x 4 bytes each).
      */
     sub r4, r5
-    LKMC_ASSERT_EQ(r4, 8)
+    LKMC_ASSERT_EQ(r4, =8)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/rbit.S b/userland/arch/arm/rbit.S
index e1e9c59..f57bd45 100644
--- a/userland/arch/arm/rbit.S
+++ b/userland/arch/arm/rbit.S
@@ -3,10 +3,10 @@
  * Reverse bit order.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     ldr r0,      =0b00000001001000110100010101100101
     rbit r1, r0
-    LKMC_ASSERT_EQ(r1, 0b10100110101000101100010010000000)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r1, =0b10100110101000101100010010000000)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/registers.S b/userland/arch/arm/registers.S
index 0f3aee5..c9752bb 100644
--- a/userland/arch/arm/registers.S
+++ b/userland/arch/arm/registers.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#assembly-registers */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* 13 general purpose registers. */
     mov r0, 0
@@ -39,12 +39,12 @@ LKMC_ENTRY
      * https://stackoverflow.com/questions/32304646/arm-assembly-branch-to-address-inside-register-or-memory/54145818#54145818
      */
     ldr pc, =10f
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 10:
 
     /* Same with r15, which is the same as pc. */
     ldr r15, =10f
-    LKMC_FAIL
+    LKMC_ASSERT_FAIL
 10:
 
     /* Another example with mov reading from pc. */
@@ -60,10 +60,10 @@ pc_addr:
     b 1f
     .word 0x12345678
 1:
-    LKMC_ASSERT_EQ(r0, 0x12345678)
+    LKMC_ASSERT_EQ(r0, =0x12345678)
 
     /* We can also use fp in GNU GAS assembly. */
     mov r11, 0
     mov fp, 1
-    LKMC_ASSERT_EQ(r11, 1)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r11, =1)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/rev.S b/userland/arch/arm/rev.S
index 14fd63d..3d62e04 100644
--- a/userland/arch/arm/rev.S
+++ b/userland/arch/arm/rev.S
@@ -3,16 +3,16 @@
  * Reverse byte order.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* All bytes in register. */
     ldr r0, =0x11223344
     rev r1, r0
-    LKMC_ASSERT_EQ(r1, 0x44332211)
+    LKMC_ASSERT_EQ(r1, =0x44332211)
 
     /* Groups of 16-bits. */
     ldr r0, =0x11223344
     rev16 r1, r0
-    LKMC_ASSERT_EQ(r1, 0x22114433)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r1, =0x22114433)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/s_suffix.S b/userland/arch/arm/s_suffix.S
index bc5f732..a527c90 100644
--- a/userland/arch/arm/s_suffix.S
+++ b/userland/arch/arm/s_suffix.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-s-suffix */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* Result is 0, set beq. */
     movs r0, 0
@@ -20,7 +20,7 @@ LKMC_ENTRY
     /* movs still moves... */
     mov r0, 0
     movs r0, 1
-    LKMC_ASSERT_EQ(r0, 1)
+    LKMC_ASSERT_EQ(r0, =1)
 
     /* add: the result is 0. */
     mov r0, 1
@@ -32,4 +32,4 @@ LKMC_ENTRY
     adds r0, 1
     LKMC_ASSERT(bne)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/shift.S b/userland/arch/arm/shift.S
index 7085396..163d6ed 100644
--- a/userland/arch/arm/shift.S
+++ b/userland/arch/arm/shift.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-shift-suffixes */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* lsr */
     ldr r0, =0xFFF00FFF
@@ -76,4 +76,4 @@ LKMC_ENTRY
     ldr r2, =0x00000120
     LKMC_ASSERT_EQ_REG(r1, r2)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/str.S b/userland/arch/arm/str.S
index 6c5d4fa..15aace3 100644
--- a/userland/arch/arm/str.S
+++ b/userland/arch/arm/str.S
@@ -1,13 +1,13 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#load-and-store-instructions */
 
-#include "common.h"
+#include <lkmc.h>
 
 .data;
     /* Must be in the .data section, since we want to modify it. */
 myvar:
     .word 0x12345678
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* Sanity check. */
     ldr r0, =myvar
     ldr r1, [r0]
@@ -57,4 +57,4 @@ var_in_same_section:
     str r1, =myvar
 #endif
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/sub.S b/userland/arch/arm/sub.S
index c24b830..2f945a4 100644
--- a/userland/arch/arm/sub.S
+++ b/userland/arch/arm/sub.S
@@ -3,12 +3,12 @@
  * Subtraction.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* 3 - 2 == 1 , register version.*/
     mov r0, 3
     mov r1, 2
     sub r0, r0, r1
-    LKMC_ASSERT_EQ(r0, 1)
-LKMC_EXIT
+    LKMC_ASSERT_EQ(r0, =1)
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/thumb.S b/userland/arch/arm/thumb.S
index 6570806..8dad934 100644
--- a/userland/arch/arm/thumb.S
+++ b/userland/arch/arm/thumb.S
@@ -7,9 +7,9 @@
 .syntax unified
 .text
 .thumb_func
-.global asm_main
-asm_main:
-asm_main_after_prologue:
+.global main
+main:
+main_after_prologue:
 
     /* CBZ: cmp and branch if zero instruction. Equivalent to CMP + BEQ.
      * TODO create an interesting assertion here.
diff --git a/userland/arch/arm/tst.S b/userland/arch/arm/tst.S
index aaca50f..29a6705 100644
--- a/userland/arch/arm/tst.S
+++ b/userland/arch/arm/tst.S
@@ -3,9 +3,9 @@
  * Test. Same as ands, but don't store the result, just update flags.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 
     /* 0x0F && 0xF0 == 0x00, so beq. */
     mov r0, 0x0F
@@ -17,6 +17,6 @@ LKMC_ENTRY
     tst r0, 0x0F
     LKMC_ASSERT(bne)
     # r0 was not modified.
-    LKMC_ASSERT_EQ(r0, 0xFF)
+    LKMC_ASSERT_EQ(r0, =0xFF)
 
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/udf.S b/userland/arch/arm/udf.S
index ca140ee..6707177 100644
--- a/userland/arch/arm/udf.S
+++ b/userland/arch/arm/udf.S
@@ -1,7 +1,7 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-udf-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     udf 0
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/vadd_scalar.S b/userland/arch/arm/vadd_scalar.S
index a1335b9..0206f9b 100644
--- a/userland/arch/arm/vadd_scalar.S
+++ b/userland/arch/arm/vadd_scalar.S
@@ -1,9 +1,9 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#vfp
  * Adapted from: https://mindplusplus.wordpress.com/2013/06/27/arm-vfp-vector-programming-part-2-examples/ */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* Minimal single precision floating point example.
      * TODO: floating point representation constraints due to 4-byte instruction?
      */
@@ -39,7 +39,7 @@ my_float_sum:
     vadd.f32 s2, s0, s1
     ldr r0, =my_float_sum
     vstr.f32 s2, [r0]
-    LKMC_ASSERT_MEMCMP(my_float_sum, my_float_sum_expect, 4)
+    LKMC_ASSERT_MEMCMP(my_float_sum, my_float_sum_expect, =4)
 
 #if 0
     /* We can't do pseudo vldr as for ldr, fails with:
@@ -69,4 +69,4 @@ my_float_sum:
     vmov s1, s0
     vmov r1, s1
     LKMC_ASSERT_EQ_REG(r0, r1)
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/vadd_vector.S b/userland/arch/arm/vadd_vector.S
index 0f6a763..e103f73 100644
--- a/userland/arch/arm/vadd_vector.S
+++ b/userland/arch/arm/vadd_vector.S
@@ -1,10 +1,10 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-vadd-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
 .bss
     output:      .skip 16
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* Integer. */
 .data
     input0_u:    .long 0xF1F1F1F1, 0xF2F2F2F2, 0xF3F3F3F3, 0xF4F4F4F4
@@ -20,7 +20,7 @@ LKMC_ENTRY
     vadd.u ## size q2, q0, q1; \
     ldr r0, =output; \
     vst1.u ## size {q2}, [r0]; \
-    LKMC_ASSERT_MEMCMP(output, expect_u_ ## size, 0x10)
+    LKMC_ASSERT_MEMCMP(output, expect_u_ ## size, =0x10)
 
     /* vadd.u32
      *
@@ -54,7 +54,7 @@ LKMC_ENTRY
     vadd.f ## size q2, q0, q1; \
     ldr r0, =output; \
     vst1. ## size {q2}, [r0]; \
-    LKMC_ASSERT_MEMCMP(output, expect_f_ ## size, 0x10)
+    LKMC_ASSERT_MEMCMP(output, expect_f_ ## size, =0x10)
 
     /* 4x 32-bit. */
     TEST(32)
@@ -68,4 +68,4 @@ LKMC_ENTRY
     TEST(64)
 #endif
 #undef TEST
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/vcvt.S b/userland/arch/arm/vcvt.S
index a903427..e47c998 100644
--- a/userland/arch/arm/vcvt.S
+++ b/userland/arch/arm/vcvt.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-vcvt-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* SIMD positive. */
 .data
     vcvt_positive_0:      .float 1.25, 2.5, 3.75, 4.0
@@ -15,7 +15,7 @@ LKMC_ENTRY
     vcvt.u32.f32 q1, q0
     ldr r0, =vcvt_positive_result
     vst1.32 {q1}, [r0]
-    LKMC_ASSERT_MEMCMP(vcvt_positive_result, vcvt_positive_expect, 0x10)
+    LKMC_ASSERT_MEMCMP(vcvt_positive_result, vcvt_positive_expect, =0x10)
 
     /* SIMD negative. */
 .data
@@ -29,7 +29,7 @@ LKMC_ENTRY
     vcvt.s32.f32 q1, q0
     ldr r0, =vcvt_negative_result
     vst1.32 {q1}, [r0]
-    LKMC_ASSERT_MEMCMP(vcvt_negative_result, vcvt_negative_expect, 0x10)
+    LKMC_ASSERT_MEMCMP(vcvt_negative_result, vcvt_negative_expect, =0x10)
 
     /* Floating point. */
 .data
@@ -44,7 +44,7 @@ LKMC_ENTRY
     vcvt.u32.f32 s0, s0
     ldr r0, =vcvt_positive_float_result
     vst1.32 {d0}, [r0]
-    LKMC_ASSERT_MEMCMP(vcvt_positive_float_result, vcvt_positive_float_expect, 0x8)
+    LKMC_ASSERT_MEMCMP(vcvt_positive_float_result, vcvt_positive_float_expect, =0x8)
 
     /* Floating point but with immediates.
      *
@@ -85,6 +85,6 @@ LKMC_ENTRY
     LKMC_ASSERT_MEMCMP(
         vcvt_positive_double_result,
         vcvt_positive_double_expect,
-        0x4
+        =0x4
     )
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/vcvta.S b/userland/arch/arm/vcvta.S
index 51b0aa9..2da2a7a 100644
--- a/userland/arch/arm/vcvta.S
+++ b/userland/arch/arm/vcvta.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-vcvta-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     /* SIMD positive. */
 .data
     vcvta_positive_0:      .float 1.25, 2.5, 3.75, 4.0
@@ -18,7 +18,7 @@ LKMC_ENTRY
     LKMC_ASSERT_MEMCMP(
         vcvta_positive_result,
         vcvta_positive_expect,
-        0x10
+        =0x10
     )
 
     /* SIMD negative. */
@@ -36,6 +36,6 @@ LKMC_ENTRY
     LKMC_ASSERT_MEMCMP(
         vcvta_negative_result,
         vcvta_negative_expect,
-        0x10
+        =0x10
     )
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/arm/vcvtr.S b/userland/arch/arm/vcvtr.S
index 54cabd8..0f1e215 100644
--- a/userland/arch/arm/vcvtr.S
+++ b/userland/arch/arm/vcvtr.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#arm-vcvtrr-instruction */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 .data
     vcvtr_0:                    .float 1.25, 2.5, 3.75, 4.0
     vcvtr_expect_zero:          .word  1,    2,   3,    4
@@ -24,7 +24,7 @@ LKMC_ENTRY
     LKMC_ASSERT_MEMCMP(
         vcvtr_result_zero,
         vcvtr_expect_zero,
-        0x10
+        =0x10
     )
 
 #if 0
@@ -40,7 +40,7 @@ LKMC_ENTRY
     LKMC_ASSERT_MEMCMP(
         vcvtr_result_plus_infinity,
         vcvtr_expect_plus_infinity,
-        0x10
+        =0x10
     )
 #endif
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/common.h b/userland/arch/common.h
deleted file mode 100644
index c658de2..0000000
--- a/userland/arch/common.h
+++ /dev/null
@@ -1,32 +0,0 @@
-/* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly-c-standard-library */
-
-#ifndef COMMON_H
-#define COMMON_H
-
-/* We define in this header only macros that are the same on all archs. */
-
-/* common_arch.h contains arch specific macros. */
-#include "common_arch.h"
-
-.extern \
-    exit, \
-    printf, \
-    puts \
-;
-
-/* Assert that the given branch instruction is taken. */
-#define LKMC_ASSERT(branch_if_pass) \
-    branch_if_pass 1f; \
-    LKMC_FAIL; \
-1: \
-;
-
-#ifndef LKMC_ASSERT_EQ_REG
-/* Assert that a register equals another register. */
-#define LKMC_ASSERT_EQ_REG(reg1, reg2) \
-    cmp reg1, reg2; \
-    LKMC_ASSERT(beq); \
-;
-#endif
-
-#endif
diff --git a/userland/arch/empty.S b/userland/arch/empty.S
index 9de6b9a..4a989d7 100644
--- a/userland/arch/empty.S
+++ b/userland/arch/empty.S
@@ -2,7 +2,7 @@
  * https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
-LKMC_EXIT
+LKMC_PROLOGUE
+LKMC_EPILOGUE
diff --git a/userland/arch/fail.S b/userland/arch/fail.S
index d17a993..925527f 100644
--- a/userland/arch/fail.S
+++ b/userland/arch/fail.S
@@ -1,9 +1,9 @@
-/* See what happens on test failure with LKMC_FAIL.
+/* See what happens on test failure with LKMC_ASSERT_FAIL.
  * https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
-    LKMC_FAIL
-LKMC_EXIT
+LKMC_PROLOGUE
+    LKMC_ASSERT_FAIL
+LKMC_EPILOGUE
diff --git a/userland/arch/main.c b/userland/arch/main.c
deleted file mode 100644
index 4db7823..0000000
--- a/userland/arch/main.c
+++ /dev/null
@@ -1,60 +0,0 @@
-/* This is the main entry point for all .S examples.
- * https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly-c-standard-library
- */
-
-#include <inttypes.h>
-#include <stdio.h>
-#include <stdint.h>
-
-int asm_main(uint32_t *line);
-
-#define ASSERT_EQ_DEFINE(bits) \
-    int lkmc_assert_eq_ ## bits(uint ## bits ## _t val1, uint ## bits ## _t val2) { \
-        if (val1 != val2) { \
-            printf("%s failed\n", __func__); \
-            printf("val1 0x%" PRIX ## bits "\n", val1); \
-            printf("val2 0x%" PRIX ## bits "\n", val2); \
-            return 1; \
-        } \
-        return 0; \
-    }
-ASSERT_EQ_DEFINE(32)
-ASSERT_EQ_DEFINE(64)
-#undef ASSERT_EQ_DEFINE
-
-int lkmc_assert_memcmp(const void *s1, const void *s2, size_t n) {
-    int ret;
-    size_t i;
-    uint8_t *s1b, *s2b;
-    uint8_t b1, b2;
-
-    ret = 0;
-    s1b = (uint8_t *)s1;
-    s2b = (uint8_t *)s2;
-    for (i = 0; i < n; ++i) {
-        b1 = s1b[i];
-        b2 = s2b[i];
-        if (b1 != b2) {
-            printf(
-                "%s failed: "
-                "byte1, byte2, index: "
-                "0x%02" PRIX8 " 0x%02" PRIX8 " 0x%zX\n",
-                __func__,
-                b1,
-                b2,
-                i
-            );
-            ret = 1;
-        }
-    }
-    return ret;
-}
-
-int main(void) {
-    uint32_t ret, line;
-    ret = asm_main(&line);
-    if (ret) {
-        printf("error: asm_main returned %d at line %d\n", ret, line);
-    }
-    return ret;
-}
diff --git a/userland/arch/x86_64/add.S b/userland/arch/x86_64/add.S
index 67e76a7..2c309e7 100644
--- a/userland/arch/x86_64/add.S
+++ b/userland/arch/x86_64/add.S
@@ -1,9 +1,9 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     mov $1, %rax
     add $2, %rax
     LKMC_ASSERT_EQ(%rax, $3)
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/x86_64/addpd.S b/userland/arch/x86_64/addpd.S
index 2cd6dfc..e217680 100644
--- a/userland/arch/x86_64/addpd.S
+++ b/userland/arch/x86_64/addpd.S
@@ -3,9 +3,9 @@
  * Add a bunch of floating point numbers in one go.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 .bss
     output:       .skip 16
 .data
@@ -28,4 +28,4 @@ LKMC_ENTRY
     /* 2x 64-bit */
     TEST(d)
 #undef TEST
-LKMC_EXIT
+LKMC_EPILOGUE
diff --git a/userland/arch/x86_64/common_arch.h b/userland/arch/x86_64/common_arch.h
deleted file mode 100644
index cc9b3c2..0000000
--- a/userland/arch/x86_64/common_arch.h
+++ /dev/null
@@ -1,87 +0,0 @@
-/* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly-c-standard-library */
-
-#ifndef COMMON_ARCH_H
-#define COMMON_ARCH_H
-
-/* This and other macros may make C function calls, and therefore can destroy
- * non-callee saved registers. */
-#define LKMC_ASSERT_EQ(general1, general2) \
-    mov general2, %rdi; \
-    push %rdi; \
-    mov general1, %rdi; \
-    pop %rsi; \
-    LKMC_ASSERT_EQ_DO(64); \
-;
-
-#define LKMC_ASSERT_EQ_DO(bits) \
-    call lkmc_assert_eq_ ## bits; \
-    cmp $0, %rax; \
-    LKMC_ASSERT(je); \
-;
-
-#define LKMC_ASSERT_MEMCMP(label1, label2, const_size) \
-    lea label1(%rip), %rdi; \
-    lea label2(%rip), %rsi; \
-    mov const_size, %rdx; \
-    call lkmc_assert_memcmp; \
-    cmp $0, %rax; \
-    LKMC_ASSERT(je); \
-;
-
-/* Program entry point.
- *
- * Return with LKMC_EXIT.
- *
- * Basically implements an x86_64 prologue:
- *
- * - save callee saved registers
- *   x86_64 explained at: https://stackoverflow.com/questions/18024672/what-registers-are-preserved-through-a-linux-x86-64-function-call/55207335#55207335
- * - save register arguments for later usage
- */
-#define LKMC_ENTRY \
-.text; \
-.global asm_main; \
-asm_main: \
-    push %rbp; \
-    mov %rsp, %rbp; \
-    push %r15; \
-    push %r14; \
-    push %r13; \
-    push %r12; \
-    push %rbx; \
-    push %rdi; \
-    sub $8, %rsp; \
-asm_main_after_prologue: \
-;
-
-/* Meant to be called at the end of LKMC_ENTRY.*
- *
- * Branching to "fail" makes tests fail with exit status 1.
- *
- * If LKMC_EXIT is reached, the program ends successfully.
- */
-#define LKMC_EXIT \
-    mov $0, %rax; \
-    jmp pass; \
-fail: \
-    /* -0x30(%rbp) is argument 1 which we pushed at prologue */ \
-    mov -0x30(%rbp), %rbx; \
-    movl %eax, (%rbx); \
-    mov $1, %rax; \
-pass: \
-    add $16, %rsp; \
-    pop %rbx; \
-    pop %r12; \
-    pop %r13; \
-    pop %r14; \
-    pop %r15; \
-    pop %rbp; \
-    ret; \
-;
-
-#define LKMC_FAIL \
-    mov $__LINE__, %eax; \
-    jmp fail; \
-;
-
-#endif
diff --git a/userland/arch/x86_64/gas_data_sizes.S b/userland/arch/x86_64/gas_data_sizes.S
index fbab7f8..d3ed256 100644
--- a/userland/arch/x86_64/gas_data_sizes.S
+++ b/userland/arch/x86_64/gas_data_sizes.S
@@ -1,8 +1,8 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#gnu-gas-assembler data sizes */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 #define ASSERT_DIFF(label1, label2, result) \
     lea label2(%rip), %rax; \
     lea label1(%rip), %rbx; \
@@ -15,7 +15,7 @@ LKMC_ENTRY
     ASSERT_DIFF(myquad, myocta, 8)
     ASSERT_DIFF(myocta, theend, 16)
 #undef ASSERT_DIF
-LKMC_EXIT
+LKMC_EPILOGUE
 mybyte:
     .byte 0x12
 myword:
diff --git a/userland/arch/x86_64/lkmc_assert_eq_fail.S b/userland/arch/x86_64/lkmc_assert_eq_fail.S
index b14db30..0484605 100644
--- a/userland/arch/x86_64/lkmc_assert_eq_fail.S
+++ b/userland/arch/x86_64/lkmc_assert_eq_fail.S
@@ -1,16 +1,26 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
     mov $0x123456789ABCDEF0, %r12
     mov $0x123456789ABCDEF0, %r13
+
+    /* Passing examples. */
+
+    /* Register immediate. */
     LKMC_ASSERT_EQ(%r12, $0x123456789ABCDEF0)
-    LKMC_ASSERT_EQ(%r12, %r13)
-    LKMC_ASSERT_EQ(%r12, myvar)
     LKMC_ASSERT_EQ($0x123456789ABCDEF0, %r12)
+
+    /* Register register. */
+    LKMC_ASSERT_EQ(%r12, %r13)
     LKMC_ASSERT_EQ(%r13, %r12)
+
+    /* Register memory. */
+    LKMC_ASSERT_EQ(%r12, myvar)
     LKMC_ASSERT_EQ(myvar, %r12)
+
+    /* Now let's fail. */
     LKMC_ASSERT_EQ(%r12, $0x123456789ABCDEF1)
-LKMC_EXIT
+LKMC_EPILOGUE
 myvar: .quad 0x123456789ABCDEF0
diff --git a/userland/arch/x86_64/lkmc_assert_memcmp_fail.S b/userland/arch/x86_64/lkmc_assert_memcmp_fail.S
index 6e9bc57..613936d 100644
--- a/userland/arch/x86_64/lkmc_assert_memcmp_fail.S
+++ b/userland/arch/x86_64/lkmc_assert_memcmp_fail.S
@@ -1,11 +1,15 @@
 /* https://github.com/cirosantilli/linux-kernel-module-cheat#userland-assembly */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
+    /* Pass */
     LKMC_ASSERT_MEMCMP(var0, var1, $0x10)
+    LKMC_ASSERT_MEMCMP(var0, var1, size)
+    /* Fail */
     LKMC_ASSERT_MEMCMP(var0, var2, $0x10)
-LKMC_EXIT
+LKMC_EPILOGUE
     var0: .long 0x11111111, 0x22222222, 0x33333333, 0x44444444
     var1: .long 0x11111111, 0x22222222, 0x33333333, 0x44444444
     var2: .long 0x11111111, 0x22222223, 0x23333333, 0x44444444
+    size: .quad 0x10
diff --git a/userland/arch/x86_64/paddq.S b/userland/arch/x86_64/paddq.S
index e0f23f5..c290e20 100644
--- a/userland/arch/x86_64/paddq.S
+++ b/userland/arch/x86_64/paddq.S
@@ -5,9 +5,9 @@
  * The different variants basically determine if carries get forwarded or not.
  */
 
-#include "common.h"
+#include <lkmc.h>
 
-LKMC_ENTRY
+LKMC_PROLOGUE
 .data
     input0:       .long 0xF1F1F1F1, 0xF2F2F2F2, 0xF3F3F3F3, 0xF4F4F4F4
     input1:       .long 0x12121212, 0x13131313, 0x14141414, 0x15151515
@@ -35,4 +35,4 @@ LKMC_ENTRY
     /* 2x 16-bit */
     TEST(q)
 #undef TEST
-LKMC_EXIT
+LKMC_EPILOGUE