Make terminal ui binaries work well everywhere

Here's some screenshots of an emulator tui program that was compiled on
Linux, then scp'd it to Windows, Mac, and FreeBSD.

https://justine.storage.googleapis.com/blinkenlights-cmdexe.png
https://justine.storage.googleapis.com/blinkenlights-imac.png
https://justine.storage.googleapis.com/blinkenlights-freebsd.png
https://justine.storage.googleapis.com/blinkenlights-lisp.png

How is this even possible that we have a nontrivial ui binary that just
works on Mac, Windows, Linux, and BSD? Surely a first ever achievement.

Fixed many bugs. Bootstrapped John McCarthy's metacircular evaluator on
bare metal in half the size of Altair BASIC (about 2.5kb) and ran it in
emulator for fun and profit.

libc/nt/KernelBase/CancelIo.s

@@ -1,2 +1,15 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_CancelIo,CancelIo,101
+
+	.text.windows
+CancelIo:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	%rdi,%rcx
+	sub	$32,%rsp
+	call	*__imp_CancelIo(%rip)
+	leave
+	ret
+	.endfn	CancelIo,globl
+	.previous

libc/nt/KernelBase/CancelIoEx.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_CancelIoEx,CancelIoEx,102
+
+	.text.windows
+CancelIoEx:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_CancelIoEx(%rip),%rax
+	jmp	__sysv2nt
+	.endfn	CancelIoEx,globl
+	.previous

libc/nt/KernelBase/CancelSynchronousIo.s

@@ -1,2 +1,15 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_CancelSynchronousIo,CancelSynchronousIo,103
+
+	.text.windows
+CancelSynchronousIo:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	%rdi,%rcx
+	sub	$32,%rsp
+	call	*__imp_CancelSynchronousIo(%rip)
+	leave
+	ret
+	.endfn	CancelSynchronousIo,globl
+	.previous

libc/nt/KernelBase/CreateIoCompletionPort.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_CreateIoCompletionPort,CreateIoCompletionPort,195
+
+	.text.windows
+CreateIoCompletionPort:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_CreateIoCompletionPort(%rip),%rax
+	jmp	__sysv2nt
+	.endfn	CreateIoCompletionPort,globl
+	.previous

libc/nt/KernelBase/CreateThread.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_CreateThread,CreateThread,224
+
+	.text.windows
+CreateThread:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_CreateThread(%rip),%rax
+	jmp	__sysv2nt6
+	.endfn	CreateThread,globl
+	.previous

libc/nt/KernelBase/CreateWaitableTimerW.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_CreateWaitableTimerW,CreateWaitableTimerW,234
+
+	.text.windows
+CreateWaitableTimer:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_CreateWaitableTimerW(%rip),%rax
+	jmp	__sysv2nt
+	.endfn	CreateWaitableTimer,globl
+	.previous

libc/nt/KernelBase/GetQueuedCompletionStatus.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_GetQueuedCompletionStatus,GetQueuedCompletionStatus,695
+
+	.text.windows
+GetQueuedCompletionStatus:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_GetQueuedCompletionStatus(%rip),%rax
+	jmp	__sysv2nt6
+	.endfn	GetQueuedCompletionStatus,globl
+	.previous

libc/nt/KernelBase/GetQueuedCompletionStatusEx.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_GetQueuedCompletionStatusEx,GetQueuedCompletionStatusEx,696
+
+	.text.windows
+GetQueuedCompletionStatusEx:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_GetQueuedCompletionStatusEx(%rip),%rax
+	jmp	__sysv2nt6
+	.endfn	GetQueuedCompletionStatusEx,globl
+	.previous

libc/nt/KernelBase/PostQueuedCompletionStatus.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_PostQueuedCompletionStatus,PostQueuedCompletionStatus,1221
+
+	.text.windows
+PostQueuedCompletionStatus:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_PostQueuedCompletionStatus(%rip),%rax
+	jmp	__sysv2nt
+	.endfn	PostQueuedCompletionStatus,globl
+	.previous

libc/nt/KernelBase/ReadFileEx.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_ReadFileEx,ReadFileEx,1312
+
+	.text.windows
+ReadFileEx:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_ReadFileEx(%rip),%rax
+	jmp	__sysv2nt6
+	.endfn	ReadFileEx,globl
+	.previous

libc/nt/KernelBase/SetWaitableTimer.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_SetWaitableTimer,SetWaitableTimer,1583
+
+	.text.windows
+SetWaitableTimer:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_SetWaitableTimer(%rip),%rax
+	jmp	__sysv2nt6
+	.endfn	SetWaitableTimer,globl
+	.previous

libc/nt/KernelBase/WriteFileEx.s

@@ -1,2 +1,12 @@
 .include "o/libc/nt/codegen.inc"
 .imp	KernelBase,__imp_WriteFileEx,WriteFileEx,1824
+
+	.text.windows
+WriteFileEx:
+	push	%rbp
+	mov	%rsp,%rbp
+	.profilable
+	mov	__imp_WriteFileEx(%rip),%rax
+	jmp	__sysv2nt6
+	.endfn	WriteFileEx,globl
+	.previous