Modify userspace .org and RAMSTART expectations

Instead of expecting a `USER_CODE` symbol to be set, we expect `.org` to be
set in all userspace glue code. This gives us more flexibility with regards to
how we manage that.

Moreover, instead of making `USER_RAMSTART` mandatory, we make it default to
the end of the binary, which is adequate in a majority of cases.

Will be useful for my upcoming mega-commit... :)

apps/README.md

@@ -10,8 +10,16 @@ you will typically not want to do that.
 
 ## Userspace convention
 
-We execute a userspace application by calling the address it's loaded into. This
-means: a userspace application is expected to return.
+We execute a userspace application by calling the address it's loaded into.
+
+This means that userspace applications must be assembled with a proper `.org`,
+otherwise labels in its code will be wrong.
+
+The `.org`, it is not specified by glue code of the apps themselves. It is
+expected to be set either in the `user.h` file to through `zasm` 3rd argument.
+
+That a userspace is called also means that an application, when finished
+running, is expected to return with a regular `ret` and a clean stack.
 
 Whatever calls the userspace app (usually, it will be the shell), should set
 HL to a pointer to unparsed arguments in string form, null terminated.
@@ -25,3 +33,8 @@ because otherwise, it will break the kernel.
 
 Apps in Collapse OS are design to be ROM-compatible, that is, they don't write
 to addresses that are part of the code's address space.
+
+By default, apps set their RAM to begin at the end of the binary because in
+most cases, these apps will be ran from RAM. If they're ran from ROM, make sure
+to set `USER_RAMSTART` properly in your `user.h` to ensure that the RAM is
+placed properly.

apps/at28w/glue.asm

@@ -16,9 +16,9 @@
 
 .inc "user.h"
 .inc "err.h"
-.org	USER_CODE
 .equ	AT28W_RAMSTART	USER_RAMSTART
 
 jp	at28wMain
 
 .inc "at28w/main.asm"
+USER_RAMSTART:

apps/basic/glue.asm

@@ -7,7 +7,6 @@
 .inc "user.h"
 
 .inc "err.h"
-.org	USER_CODE
 
 	jp	basStart
 
@@ -16,3 +15,4 @@
 .inc "lib/parse.asm"
 .equ	BAS_RAMSTART	USER_RAMSTART
 .inc "basic/main.asm"
+USER_RAMSTART:

apps/ed/glue.asm

@@ -26,8 +26,6 @@
 ; ******
 
 .inc "err.h"
-.org	USER_CODE
-
 	jp	edMain
 
 .inc "core.asm"
@@ -41,3 +39,4 @@
 .inc "ed/cmd.asm"
 .equ	ED_RAMSTART	CMD_RAMEND
 .inc "ed/main.asm"
+USER_RAMSTART:

apps/memt/glue.asm

@@ -13,8 +13,7 @@
 ; *** Includes ***
 
 .inc "user.h"
-.org	USER_CODE
-
 jp	memtMain
 
 .inc "memt/main.asm"
+USER_RAMSTART:

apps/sdct/glue.asm

@@ -19,9 +19,9 @@
 ; *** Includes ***
 
 .inc "user.h"
-.org	USER_CODE
 .equ	SDCT_RAMSTART	USER_RAMSTART
 
 jp	sdctMain
 
 .inc "sdct/main.asm"
+USER_RAMSTART:

apps/zasm/README.md

@@ -3,6 +3,21 @@
 This is probably the most critical part of the Collapse OS project because it
 ensures its self-reproduction.
 
+## Invocation
+
+`zasm` is invoked with 2 mandatory arguments and an optional one. The mandatory
+arguments are input blockdev id and output blockdev id. For example, `zasm 0 1`
+reads source code from blockdev 0, assembles it and spit the result in blockdev
+1.
+
+Input blockdev needs to be seek-able, output blockdev doesn't need to (zasm
+writes in one pass, sequentially.
+
+The 3rd argument, optional, is the initial `.org` value. It's the high byte of
+the value. For example, `zasm 0 1 4f` assembles source in blockdev 0 as if it
+started with the line `.org 0x4f00`. This also means that the initial value of
+the `@` symbol is `0x4f00`.
+
 ## Running on a "modern" machine
 
 To be able to develop zasm efficiently, [libz80][libz80] is used to run zasm
@@ -101,7 +116,10 @@ it was placed there.
 
 Whenever a `.equ` directive is evaluated, its resulting value is saved in a
 special "last value" register that can then be used in any expression. This
-is very useful for variable definitions and for jump tables.
+last value is referenced with the `@` special symbol. This is very useful for
+variable definitions and for jump tables.
+
+Note that `.org` also affect the last value.
 
 ## Includes
 

apps/zasm/directive.asm

@@ -18,7 +18,7 @@
 ; *** CODE ***
 
 ; 3 bytes per row, fill with zero
-directiveNames:
+dirNames:
 	.db	"DB", 0
 	.db	"DW", 0
 	.db	"EQU"
@@ -28,8 +28,8 @@ directiveNames:
 	.db	"INC"
 	.db	"BIN"
 
-; This is a list of handlers corresponding to indexes in directiveNames
-directiveHandlers:
+; This is a list of handlers corresponding to indexes in dirNames
+dirHandlers:
 	.dw	handleDB
 	.dw	handleDW
 	.dw	handleEQU
@@ -184,6 +184,7 @@ handleORG:
 	call	parseExpr
 	jr	nz, .badarg
 	push	ix \ pop hl
+	ld	(DIREC_LASTVAL), hl
 	call	zasmSetOrg
 	cp	a		; ensure Z
 	ret
@@ -310,7 +311,7 @@ getDirectiveID:
 	inc	hl
 	ld	b, D_BIN+1		; D_BIN is last
 	ld	c, 3
-	ld	de, directiveNames
+	ld	de, dirNames
 	call	findStringInList
 	pop	de
 	pop	bc
@@ -324,9 +325,9 @@ getDirectiveID:
 ; error, A contains the error number (ERR_*).
 parseDirective:
 	push	de
-	; double A to have a proper offset in directiveHandlers
+	; double A to have a proper offset in dirHandlers
 	add	a, a
-	ld	de, directiveHandlers
+	ld	de, dirHandlers
 	call	addDE
 	call	intoDE
 	push	de \ pop ix

apps/zasm/glue.asm

@@ -64,8 +64,6 @@
 
 .inc "err.h"
 .inc "ascii.h"
-.org	USER_CODE
-
 jp	zasmMain
 
 .inc "core.asm"
@@ -87,3 +85,4 @@ jp	zasmMain
 .inc "zasm/symbol.asm"
 .equ	ZASM_RAMSTART	SYM_RAMEND
 .inc "zasm/main.asm"
+USER_RAMSTART:

apps/zasm/main.asm

@@ -8,14 +8,17 @@
 ; whether we're in "local pass", that is, in local label scanning mode. During
 ; this special pass, ZASM_FIRST_PASS will also be set so that the rest of the
 ; code behaves as is we were in the first pass.
-.equ	ZASM_LOCAL_PASS		ZASM_FIRST_PASS+1
+.equ	ZASM_LOCAL_PASS		@+1
 ; What IO_PC was when we started our context
-.equ	ZASM_CTX_PC		ZASM_LOCAL_PASS+1
+.equ	ZASM_CTX_PC		@+1
 ; current ".org" offset, that is, what we must offset all our label by.
-.equ	ZASM_ORG		ZASM_CTX_PC+2
-.equ	ZASM_RAMEND		ZASM_ORG+2
+.equ	ZASM_ORG		@+2
+.equ	ZASM_RAMEND		@+2
 
 ; Takes 2 byte arguments, blkdev in and blkdev out, expressed as IDs.
+; Can optionally take a 3rd argument which is the high byte of the initial
+; .org. For example, passing 0x42 to this 3rd arg is the equivalent of beginning
+; the unit with ".org 0x4200".
 ; Read file through blkdev in and outputs its upcodes through blkdev out.
 ; HL is set to the last lineno to be read.
 ; Sets Z on success, unset on error. On error, A contains an error code (ERR_*)
@@ -23,7 +26,7 @@ zasmMain:
 	; Parse args. HL points to string already
 	; We don't allocate memory just to hold this. Because this happens
 	; before initialization, we don't really care where those args are
-	; parsed.
+	; parsed. That's why we borrow zasm's RAMSTART for a little while.
 	ld	de, .argspecs
 	ld	ix, ZASM_RAMSTART
 	call	parseArgs
@@ -44,11 +47,18 @@ zasmMain:
 	ld	de, IO_OUT_BLK
 	call	blkSel
 
-	; Init modules
+	; Init .org
+	; This is the 3rd argument, optional, will be zero if not given.
+	; Save in "@" too
+	ld	a, (ZASM_RAMSTART+2)
+	ld	(ZASM_ORG+1), a		; high byte of .org
+	ld	(DIREC_LASTVAL+1), a
 	xor	a
+	ld	(ZASM_ORG), a		; low byte zero in all cases
+	ld	(DIREC_LASTVAL), a
+
+	; And then the rest.
 	ld	(ZASM_LOCAL_PASS), a
-	ld	(ZASM_ORG), a
-	ld	(ZASM_ORG+1), a
 	call	ioInit
 	call	symInit
 
@@ -73,7 +83,7 @@ zasmMain:
 	jp	ioLineNo		; --> HL, --> DE, returns
 
 .argspecs:
-	.db	0b001, 0b001, 0
+	.db	0b001, 0b001, 0b101
 .sFirstPass:
 	.db	"First pass", 0
 .sSecondPass:

recipes/rc2014/zasm/user.h

@@ -1,6 +1,5 @@
-.equ    USER_CODE       0x8700
-.equ    USER_RAMSTART   USER_CODE+0x1900
-.equ    FS_HANDLE_SIZE  6
+.org    0x8700
+.equ    FS_HANDLE_SIZE  8
 .equ    BLOCKDEV_SIZE   8
 
 ; *** JUMP TABLE ***

recipes/sms/romasm/Makefile

@@ -5,13 +5,13 @@ APPS = ../../../apps
 .PHONY: all clean
 all: os.sms
 
+# -o value synced with offset in glue.asm
 ed.bin: $(APPS)/ed/glue.asm
-	echo ".equ USER_CODE ED_CODE" | cat user-tmpl.h - > user.h
-	$(ZASM) $(KERNEL) $(APPS) user.h < $< > $@
+	$(ZASM) -o 19 $(KERNEL) $(APPS) user.h < $< > $@
 
+# -o value synced with offset in glue.asm
 zasm.bin: $(APPS)/zasm/glue.asm
-	echo ".equ USER_CODE ZASM_CODE" | cat user-tmpl.h - > user.h
-	$(ZASM) $(KERNEL) $(APPS) user.h < $< > $@
+	$(ZASM) -o 1d $(KERNEL) $(APPS) user.h < $< > $@
 
 os.sms: glue.asm ed.bin zasm.bin
 	$(ZASM) $(KERNEL) ed.bin zasm.bin < $< > $@

recipes/sms/romasm/user.h

@@ -1,8 +1,5 @@
-; USER_CODE is filled in on-the-fly with either ED_CODE or ZASM_CODE
-.equ    ED_CODE         0x1900
-.equ    ZASM_CODE       0x1d00
 .equ    USER_RAMSTART   0xc200
-.equ    FS_HANDLE_SIZE  6
+.equ    FS_HANDLE_SIZE  8
 .equ    BLOCKDEV_SIZE   8
 ; Make ed fit in SMS's memory
 .equ    ED_BUF_MAXLINES 0x100

tools/emul/cfsin/hello.asm

@@ -1,6 +1,4 @@
 .inc "user.h"
-.org	USER_CODE
-
 	ld	hl, sAwesome
 	call	printstr
 	xor	a		; success

tools/emul/shell/user.h

@@ -1,5 +1,4 @@
-.equ    USER_CODE       0x4200
-.equ    USER_RAMSTART   USER_CODE+0x1800
+.org    0x4200              ; in sync with USERCODE in shell/shell_.asm
 .equ    FS_HANDLE_SIZE  8
 .equ    BLOCKDEV_SIZE   8
 

tools/emul/zasm/glue.asm

@@ -56,13 +56,20 @@ init:
 	ld	de, BLOCKDEV_SEL
 	call	blkSel
 	call	fsOn
+	; There's a special understanding between zasm.c and this unit: The
+	; addresses 0xff00 and 0xff01 contain the two ascii chars to send to
+	; zasm as the 3rd argument.
+	ld	a, (0xff00)
+	ld	(.zasmArgs+4), a
+	ld	a, (0xff01)
+	ld	(.zasmArgs+5), a
 	ld	hl, .zasmArgs
 	call	USER_CODE
 	; signal the emulator we're done
 	halt
 
 .zasmArgs:
-	.db	"0 1", 0
+	.db	"0 1 XX", 0
 
 ; *** I/O ***
 emulGetB:

tools/emul/zasm/user.h

@@ -1,4 +1,4 @@
-.equ    USER_CODE       0x4800
+.org    0x4800          ; in sync with USER_CODE in glue.asm
 .equ    USER_RAMSTART   0x6000
 .equ    FS_HANDLE_SIZE  8
 .equ    BLOCKDEV_SIZE   8

tools/emul/zasm/zasm.c

@@ -1,5 +1,6 @@
 #include <stdint.h>
 #include <stdio.h>
+#include <string.h>
 #include "../libz80/z80.h"
 #include "kernel-bin.h"
 #include "zasm-bin.h"
@@ -8,8 +9,11 @@
  * in another specified blkdev. In our emulator layer, we use stdin and stdout
  * as those specified blkdevs.
  *
- * This executable takes one argument: the path to a .cfs file to use for
- * includes.
+ * This executable takes two arguments. Both are optional, but you need to
+ * specify the first one if you want to get to the second one.
+ * The first one is the value to send to z80-zasm's 3rd argument (the initial
+ * .org). Defaults to '00'.
+ * The second one is the path to a .cfs file to use for includes.
  *
  * Because the input blkdev needs support for Seek, we buffer it in the emulator
  * layer.
@@ -155,7 +159,7 @@ static void mem_write(int unused, uint16_t addr, uint8_t val)
 
 int main(int argc, char *argv[])
 {
-    if (argc > 2) {
+    if (argc > 3) {
         fprintf(stderr, "Too many args\n");
         return 1;
     }
@@ -166,9 +170,19 @@ int main(int argc, char *argv[])
     for (int i=0; i<sizeof(USERSPACE); i++) {
         mem[i+USER_CODE] = USERSPACE[i];
     }
+    char *init_org = "00";
+    if (argc >= 2) {
+        init_org = argv[1];
+        if (strlen(init_org) != 2) {
+            fprintf(stderr, "Initial org must be a two-character hex string");
+        }
+    }
+    // glue.asm knows that it needs to fetch these arguments at this address.
+    mem[0xff00] = init_org[0];
+    mem[0xff01] = init_org[1];
     fsdev_size = 0;
-    if (argc == 2) {
-        FILE *fp = fopen(argv[1], "r");
+    if (argc == 3) {
+        FILE *fp = fopen(argv[2], "r");
         if (fp == NULL) {
             fprintf(stderr, "Can't open file %s\n", argv[1]);
             return 1;

tools/tests/unit/test_core.asm

@@ -1,3 +1,8 @@
+.equ foo 456	; AFTER_ORG should not get that value
+.org 0x1234
+.equ AFTER_ORG	@
+.org 0
+
 jp	test
 
 .inc "core.asm"
@@ -40,6 +45,13 @@ test:
 	jp	nz, fail
 	call	nexttest
 
+	; test that "@" is updated by a .org directive
+	ld	hl, AFTER_ORG
+	ld	de, 0x1234
+	call	cpHLDE
+	jp	nz, fail
+	call	nexttest
+
 	; *** cpHLDE ***
 	ld	hl, 0x42
 	ld	de, 0x42
@@ -73,3 +85,4 @@ nexttest:
 fail:
 	ld	a, (testNum)
 	halt
+

tools/zasm.sh

@@ -1,9 +1,30 @@
 #!/usr/bin/env bash
 
+# Calls tools/emul/zasm/zasm in a convenient manner by wrapping specified
+# paths to include in a single CFS file and then pass that file to zasm.
+# Additionally, it takes a "-o" argument to set the initial ".org" of the
+# binary. For example, "zasm.sh -o 4f < foo.asm" assembles foo.asm as if it
+# started with the line ".org 0x4f00".
+
 # readlink -f doesn't work with macOS's implementation
 # so, if we can't get readlink -f to work, try python with a realpath implementation
 ABS_PATH=$(readlink -f "$0" || python -c "import os; print(os.path.realpath('$0'))")
 
+usage() { echo "Usage: $0 [-o <hexorg>] <paths-to-include>..." 1>&2; exit 1; }
+
+org='00'
+while getopts ":o:" opt; do
+    case "${opt}" in
+        o)
+            org=${OPTARG}
+            ;;
+        *)
+            usage
+            ;;
+    esac
+done
+shift $((OPTIND-1))
+
 # wrapper around ./emul/zasm/zasm that prepares includes CFS prior to call
 DIR=$(dirname "${ABS_PATH}")
 ZASMBIN="${DIR}/emul/zasm/zasm"
@@ -16,7 +37,7 @@ for p in "$@"; do
     "${CFSPACK}" "${p}" "*.bin" >> "${INCCFS}"    
 done
 
-"${ZASMBIN}" "${INCCFS}"
+"${ZASMBIN}" "${org}" "${INCCFS}"
 RES=$?
 rm "${INCCFS}"
 exit $RES