[zbin] Change fixup semantics to support ROMs over 128k uncompressed

The option ROM header contains a one-byte field indicating the number
of 512-byte sectors in the ROM image.  Currently it is linked to
contain the number of uncompressed sectors, with an instruction to the
compressor to correct it.  This causes link failure when the
uncompressed size of the ROM image is over 128k.

Fix by replacing the SUBx compressor fixup with an ADDx fixup that
adds the total compressed output length, scaled as requested, to an
addend stored in the field where the final length value will be
placed.  This is similar to the behavior of ELF relocations, and
ensures that an overflow error will not be generated unless the
compressed size is still too large for the field.

This also allows us to do away with the _filesz_pgh and _filesz_sect
calculations exported by the linker script.

Output tested bitwise identical to the old SUBx mechanism on hd, dsk,
lkrn, and rom prefixes, on both 32-bit and 64-bit processors.

Modified-by: Michael Brown <mcb30@etherboot.org>
Signed-off-by: Michael Brown <mcb30@etherboot.org>

src/arch/i386/prefix/dskprefix.S

 	/* Jump to loaded copy */
 	ljmp	$SYSSEG, $start_runtime
 
-endseg:	.word SYSSEG + _filesz_pgh
+endseg:	.word SYSSEG
 	.section ".zinfo.fixup", "a", @progbits	/* Compressor fixups */
-	.ascii	"SUBW"
+	.ascii	"ADDW"
 	.long	endseg
 	.long	16
 	.long	0

src/arch/i386/prefix/hdprefix.S

 max_head:
 	.byte	0
 load_length:
-	.long	_filesz_sect
+	.long	0
 	
 	.section ".zinfo.fixup", "a", @progbits	/* Compressor fixups */
-	.ascii	"SUBL"
+	.ascii	"ADDL"
 	.long	load_length
 	.long	512
 	.long	0

src/arch/i386/prefix/lkrnprefix.S

 root_flags: 
 	.word	0
 syssize: 
-	.long	_filesz_pgh - PREFIXPGH
+	.long	-PREFIXPGH
 
 	.section ".zinfo.fixup", "a", @progbits	/* Compressor fixups */
-	.ascii	"SUBL"
+	.ascii	"ADDL"
 	.long	syssize
 	.long	16
 	.long	0

src/arch/i386/prefix/nbiprefix.S

 	.byte	0
 	.byte	0x04		/* Last segment */
 	.long	0x00007e00
-imglen:	.long	_filesz - 512
-memlen:	.long	_filesz - 512
+imglen:	.long	-512
+memlen:	.long	-512
 	.size	segment_header, . - segment_header
 
 	.section ".zinfo.fixup", "a", @progbits	/* Compressor fixups */
-	.ascii	"SUBL"
+	.ascii	"ADDL"
 	.long	imglen
 	.long	1
 	.long	0
-	.ascii	"SUBL"
+	.ascii	"ADDL"
 	.long	memlen
 	.long	1
 	.long	0

src/arch/i386/prefix/romprefix.S

 	.org	0x00
 romheader:
 	.word	0xAA55			/* BIOS extension signature */
-romheader_size:	.byte _filesz_sect	/* Size in 512-byte blocks */
+romheader_size:	.byte 0			/* Size in 512-byte blocks */
 	jmp	init			/* Initialisation vector */
 checksum:
 	.byte	0
 	.size romheader, . - romheader
 	
 	.section ".zinfo.fixup", "a", @progbits	/* Compressor fixups */
-	.ascii	"SUBB"
+	.ascii	"ADDB"
 	.long	romheader_size
 	.long	512
 	.long	0
 	.byte	0x03			/* PCI data structure revision */
 	.byte	0x02, 0x00, 0x00	/* Class code */
 pciheader_image_length:
-	.word	_filesz_sect		/* Image length */
+	.word	0			/* Image length */
 	.word	0x0001			/* Revision level */
 	.byte	0x00			/* Code type */
 	.byte	0x80			/* Last image indicator */
 pciheader_runtime_length:
-	.word	_filesz_sect		/* Maximum run-time image length */
+	.word	0			/* Maximum run-time image length */
 	.word	0x0000			/* Configuration utility code header */
 	.word	0x0000			/* DMTF CLP entry point */
 	.equ pciheader_len, . - pciheader
 	.size pciheader, . - pciheader
 	
 	.section ".zinfo.fixup", "a", @progbits	/* Compressor fixups */
-	.ascii	"SUBW"
+	.ascii	"ADDW"
 	.long	pciheader_image_length
 	.long	512
 	.long	0
-	.ascii	"SUBW"
+	.ascii	"ADDW"
 	.long	pciheader_runtime_length
 	.long	512
 	.long	0

src/arch/i386/scripts/i386.lds

     _prefix_memsz_sect	= ( ( _prefix_memsz + 511 ) / 512 );
     _text16_memsz_pgh	= ( ( _text16_memsz + 15 ) / 16 );
     _data16_memsz_pgh	= ( ( _data16_memsz + 15 ) / 16 );
-
-    /*
-     * File size in paragraphs and sectors.  Note that wherever the
-     * _filesz variables are used, there must be a corresponding
-     * .zinfo.fixup section.
-     *
-     */
-    _filesz_pgh		= ( ( _filesz + 15 ) / 16 );
-    _filesz_sect	= ( ( _filesz + 511 ) / 512 );
 }

src/util/zbin.c

 	uint32_t align;
 };
 
-struct zinfo_subtract {
+struct zinfo_add {
 	char type[4];
 	uint32_t offset;
 	uint32_t divisor;
 	struct zinfo_common common;
 	struct zinfo_copy copy;
 	struct zinfo_pack pack;
-	struct zinfo_subtract subtract;
+	struct zinfo_add add;
 };
 
 struct zinfo_file {
 	}
 
 	if ( DEBUG ) {
-		fprintf ( stderr, "COPY [%#zx,%#zx) to [%#zx,%#zx)\n", offset, ( offset + len ),
-			  output->len, ( output->len + len ) );
+		fprintf ( stderr, "COPY [%#zx,%#zx) to [%#zx,%#zx)\n",
+			  offset, ( offset + len ), output->len,
+			  ( output->len + len ) );
 	}
 
 	memcpy ( ( output->buf + output->len ),
 	}
 
 	if ( DEBUG ) {
-		fprintf ( stderr, "PACK [%#zx,%#zx) to [%#zx,%#zx)\n", offset, ( offset + len ),
-			  output->len, ( output->len + packed_len ) );
+		fprintf ( stderr, "PACK [%#zx,%#zx) to [%#zx,%#zx)\n",
+			  offset, ( offset + len ), output->len,
+			  ( output->len + packed_len ) );
 	}
 
 	output->len += packed_len;
 	return 0;
 }
 
-static int process_zinfo_subtract ( struct input_file *input,
-				    struct output_file *output,
-				    struct zinfo_subtract *subtract,
-				    size_t datasize ) {
-	size_t offset = subtract->offset;
+static int process_zinfo_add ( struct input_file *input,
+			       struct output_file *output,
+			       struct zinfo_add *add,
+			       size_t datasize ) {
+	size_t offset = add->offset;
 	void *target;
-	signed long raw_delta;
-	signed long delta;
-	unsigned long old;
-	unsigned long new;
+	signed long addend;
+	unsigned long size;
+	signed long val;
+	unsigned long mask;
 
 	if ( ( offset + datasize ) > output->len ) {
-		fprintf ( stderr, "Subtract at %#zx outside output buffer\n",
+		fprintf ( stderr, "Add at %#zx outside output buffer\n",
 			  offset );
 		return -1;
 	}
 
 	target = ( output->buf + offset );
-	raw_delta = ( align ( output->len, subtract->divisor ) -
-		      align ( input->len, subtract->divisor ) );
-	delta = ( raw_delta / ( ( signed long ) subtract->divisor ) );
+	size = ( align ( output->len, add->divisor ) / add->divisor );
 
 	switch ( datasize ) {
-	case 1: {
-		uint8_t *byte = target;
-		old = *byte;
-		*byte += delta;
-		new = *byte;
-		break; }
-	case 2: {
-		uint16_t *word = target;
-		old = *word;
-		*word += delta;
-		new = *word;
-		break; }
-	case 4: {
-		uint32_t *dword = target;
-		old = *dword;
-		*dword += delta;
-		new = *dword;
-		break; }
+	case 1:
+		addend = *( ( int8_t * ) target );
+		break;
+	case 2:
+		addend = *( ( int16_t * ) target );
+		break;
+	case 4:
+		addend = *( ( int32_t * ) target );
+		break;
 	default:
-		fprintf ( stderr, "Unsupported subtract datasize %d\n",
+		fprintf ( stderr, "Unsupported add datasize %d\n",
 			  datasize );
 		return -1;
 	}
 
+	val = size + addend;
+
+	/* The result of 1UL << ( 8 * sizeof(unsigned long) ) is undefined */
+	mask = ( ( datasize < sizeof ( mask ) ) ?
+		 ( ( 1UL << ( 8 * datasize ) ) - 1 ) : ~0UL );
+
+	if ( val < 0 ) {
+		fprintf ( stderr, "Add %s%#lx+%#lx at %#zx %sflows field\n",
+			  ( ( addend < 0 ) ? "-" : "" ), abs ( addend ), size,
+			  offset, ( ( addend < 0 ) ? "under" : "over" ) );
+		return -1;
+	}
+
+	if ( val & ~mask ) {
+		fprintf ( stderr, "Add %s%#lx+%#lx at %#zx overflows %d-byte "
+			  "field (%d bytes too big)\n",
+			  ( ( addend < 0 ) ? "-" : "" ), abs ( addend ), size,
+			  offset, datasize,
+			  ( ( val - mask - 1 ) * add->divisor ) );
+		return -1;
+	}
+
+	switch ( datasize ) {
+	case 1:
+		*( ( uint8_t * ) target ) = val;
+		break;
+	case 2:
+		*( ( uint16_t * ) target ) = val;
+		break;
+	case 4:
+		*( ( uint32_t * ) target ) = val;
+		break;
+	}
+
 	if ( DEBUG ) {
-		fprintf ( stderr, "SUBx [%#zx,%#zx) (%#lx+(%#lx/%#x)-(%#lx/%#x)) = %#lx\n",
-			  offset, ( offset + datasize ), old, output->len, subtract->divisor,
-			  input->len, subtract->divisor, new );
+		fprintf ( stderr, "ADDx [%#zx,%#zx) (%s%#lx+(%#lx/%#x)) = "
+			  "%#lx\n", offset, ( offset + datasize ),
+			  ( ( addend < 0 ) ? "-" : "" ), abs ( addend ),
+			  output->len, add->divisor, val );
 	}
 
 	return 0;
 }
 
-static int process_zinfo_subb ( struct input_file *input,
+static int process_zinfo_addb ( struct input_file *input,
 				struct output_file *output,
 				union zinfo_record *zinfo ) {
-	return process_zinfo_subtract ( input, output, &zinfo->subtract, 1 );
+	return process_zinfo_add ( input, output, &zinfo->add, 1 );
 }
 
-static int process_zinfo_subw ( struct input_file *input,
+static int process_zinfo_addw ( struct input_file *input,
 				struct output_file *output,
 				union zinfo_record *zinfo ) {
-	return process_zinfo_subtract ( input, output, &zinfo->subtract, 2 );
+	return process_zinfo_add ( input, output, &zinfo->add, 2 );
 }
 
-static int process_zinfo_subl ( struct input_file *input,
+static int process_zinfo_addl ( struct input_file *input,
 				struct output_file *output,
 				union zinfo_record *zinfo ) {
-	return process_zinfo_subtract ( input, output, &zinfo->subtract, 4 );
+	return process_zinfo_add ( input, output, &zinfo->add, 4 );
 }
 
 struct zinfo_processor {
 static struct zinfo_processor zinfo_processors[] = {
 	{ "COPY", process_zinfo_copy },
 	{ "PACK", process_zinfo_pack },
-	{ "SUBB", process_zinfo_subb },
-	{ "SUBW", process_zinfo_subw },
-	{ "SUBL", process_zinfo_subl },
+	{ "ADDB", process_zinfo_addb },
+	{ "ADDW", process_zinfo_addw },
+	{ "ADDL", process_zinfo_addl },
 };
 
 static int process_zinfo ( struct input_file *input,