diff --git a/util/cbfstool/cbfs-mkpayload.c b/util/cbfstool/cbfs-mkpayload.c
index 78988a0..b1dd1c0 100644
--- a/util/cbfstool/cbfs-mkpayload.c
+++ b/util/cbfstool/cbfs-mkpayload.c
@@ -23,18 +23,39 @@
 #include <stdlib.h>
 #include <string.h>
 
+#include "elf.h"
 #include "common.h"
 #include "cbfs.h"
-#include "elf.h"
 #include "fv.h"
 #include "coff.h"
 
+/* serialize the seg array into the buffer.
+ * The buffer is assumed to be large enough.
+ */
+static void xdr_segs(struct buffer *output,
+	struct cbfs_payload_segment *segs, int nseg)
+{
+	struct buffer outheader;
+	int i;
+
+	outheader.data = output->data;
+	outheader.size = 0;
+
+	for(i = 0; i < nseg; i++){
+		xdr_be.put32(&outheader, segs[i].type);
+		xdr_be.put32(&outheader, segs[i].compression);
+		xdr_be.put32(&outheader, segs[i].offset);
+		xdr_be.put64(&outheader, segs[i].load_addr);
+		xdr_be.put32(&outheader, segs[i].len);
+		xdr_be.put32(&outheader, segs[i].mem_len);
+	}
+}
 int parse_elf_to_payload(const struct buffer *input,
 			 struct buffer *output, comp_algo algo)
 {
-	Elf32_Phdr *phdr;
-	Elf32_Ehdr *ehdr = (Elf32_Ehdr *)input->data;
-	Elf32_Shdr *shdr;
+	Elf64_Phdr *phdr;
+	Elf64_Ehdr ehdr;
+	Elf64_Shdr *shdr;
 	char *header;
 	char *strtab;
 	int headers;
@@ -44,39 +65,26 @@
 	struct cbfs_payload_segment *segs;
 	int i;
 
-	if(!iself((unsigned char *)input->data)){
-		INFO("The payload file is not in ELF format!\n");
-		return -1;
-	}
-
-	// The tool may work in architecture-independent way.
-	if (arch != CBFS_ARCHITECTURE_UNKNOWN &&
-	    !((ehdr->e_machine == EM_ARM) && (arch == CBFS_ARCHITECTURE_ARMV7)) &&
-	    !((ehdr->e_machine == EM_386) && (arch == CBFS_ARCHITECTURE_X86))) {
-		ERROR("The payload file has the wrong architecture\n");
-		return -1;
-	}
-
 	comp_func_ptr compress = compression_function(algo);
 	if (!compress)
 		return -1;
 
+	if (elf_headers(input, &ehdr, &phdr, &shdr) < 0)
+		return -1;
+
 	DEBUG("start: parse_elf_to_payload\n");
-	headers = ehdr->e_phnum;
-	header = (char *)ehdr;
+	headers = ehdr.e_phnum;
+	header = input->data;
 
-	phdr = (Elf32_Phdr *) & (header[ehdr->e_phoff]);
-	shdr = (Elf32_Shdr *) & (header[ehdr->e_shoff]);
-
-	strtab = &header[shdr[ehdr->e_shstrndx].sh_offset];
+	strtab = &header[shdr[ehdr.e_shstrndx].sh_offset];
 
 	/* Count the number of headers - look for the .notes.pinfo
 	 * section */
 
-	for (i = 0; i < ehdr->e_shnum; i++) {
+	for (i = 0; i < ehdr.e_shnum; i++) {
 		char *name;
 
-		if (i == ehdr->e_shstrndx)
+		if (i == ehdr.e_shstrndx)
 			continue;
 
 		if (shdr[i].sh_size == 0)
@@ -106,29 +114,39 @@
 
 		segments++;
 	}
-
+	/* allocate the segment header array */
+	segs = calloc(segments, sizeof(*segs));
+	if (segs == NULL)
+		return -1;
 	/* Allocate a block of memory to store the data in */
 	if (buffer_create(output, (segments * sizeof(*segs)) + isize,
 			  input->name) != 0)
 		return -1;
 	memset(output->data, 0, output->size);
 
-	doffset = (segments * sizeof(struct cbfs_payload_segment));
+	doffset = (segments * sizeof(*segs));
 
-	segs = (struct cbfs_payload_segment *)output->data;
+	/* set up for output marshaling. This is a bit
+	 * tricky as we are marshaling the headers at the front,
+	 * and the data starting after the headers. We need to convert
+	 * the headers to the right format but the data
+	 * passes through unchanged. Unlike most XDR code,
+	 * we are doing these two concurrently. The doffset is
+	 * used to compute the address for the raw data, and the
+	 * outheader is used to marshal the headers. To make it simpler
+	 * for The Reader, we set up the headers in a separate array,
+	 * then marshal them all at once to the output.
+	 */
 	segments = 0;
 
-	for (i = 0; i < ehdr->e_shnum; i++) {
+	for (i = 0; i < ehdr.e_shnum; i++) {
 		char *name;
-
-		if (i == ehdr->e_shstrndx)
+		if (i == ehdr.e_shstrndx)
 			continue;
 
 		if (shdr[i].sh_size == 0)
 			continue;
-
 		name = (char *)(strtab + shdr[i].sh_name);
-
 		if (!strcmp(name, ".note.pinfo")) {
 			segs[segments].type = PAYLOAD_SEGMENT_PARAMS;
 			segs[segments].load_addr = 0;
@@ -148,17 +166,13 @@
 	for (i = 0; i < headers; i++) {
 		if (phdr[i].p_type != PT_LOAD)
 			continue;
-
 		if (phdr[i].p_memsz == 0)
 			continue;
-
 		if (phdr[i].p_filesz == 0) {
 			segs[segments].type = PAYLOAD_SEGMENT_BSS;
-			segs[segments].load_addr =
-			    (uint64_t)htonll(phdr[i].p_paddr);
-			segs[segments].mem_len =
-			    (uint32_t)htonl(phdr[i].p_memsz);
-			segs[segments].offset = htonl(doffset);
+			segs[segments].load_addr = phdr[i].p_paddr;
+			segs[segments].mem_len = phdr[i].p_memsz;
+			segs[segments].offset = doffset;
 
 			segments++;
 			continue;
@@ -168,37 +182,37 @@
 			segs[segments].type = PAYLOAD_SEGMENT_CODE;
 		else
 			segs[segments].type = PAYLOAD_SEGMENT_DATA;
-		segs[segments].load_addr = (uint64_t)htonll(phdr[i].p_paddr);
-		segs[segments].mem_len = (uint32_t)htonl(phdr[i].p_memsz);
-		segs[segments].compression = htonl(algo);
-		segs[segments].offset = htonl(doffset);
+		segs[segments].load_addr = phdr[i].p_paddr;
+		segs[segments].mem_len = phdr[i].p_memsz;
+		segs[segments].compression = algo;
+		segs[segments].offset = doffset;
 
 		int len;
 		compress((char *)&header[phdr[i].p_offset],
 			 phdr[i].p_filesz, output->data + doffset, &len);
-		segs[segments].len = htonl(len);
+		segs[segments].len = len;
 
 		/* If the compressed section is larger, then use the
 		   original stuff */
 
 		if ((unsigned int)len > phdr[i].p_filesz) {
 			segs[segments].compression = 0;
-			segs[segments].len = htonl(phdr[i].p_filesz);
-
+			segs[segments].len = phdr[i].p_filesz;
 			memcpy(output->data + doffset,
 			       &header[phdr[i].p_offset], phdr[i].p_filesz);
 		}
 
-		doffset += ntohl(segs[segments].len);
-		osize += ntohl(segs[segments].len);
+		doffset += segs[segments].len;
+		osize += segs[segments].len;
 
 		segments++;
 	}
 
 	segs[segments].type = PAYLOAD_SEGMENT_ENTRY;
-	segs[segments++].load_addr = htonll(ehdr->e_entry);
+	segs[segments++].load_addr = ehdr.e_entry;
 
-	output->size = (segments * sizeof(struct cbfs_payload_segment)) + osize;
+	output->size = (segments * sizeof(*segs)) + osize;
+	xdr_segs(output, segs, segments);
 	return 0;
 }
 
@@ -209,7 +223,7 @@
 				 comp_algo algo)
 {
 	comp_func_ptr compress;
-	struct cbfs_payload_segment *segs;
+	struct cbfs_payload_segment segs[2];
 	int doffset, len = 0;
 
 	compress = compression_function(algo);
@@ -217,36 +231,35 @@
 		return -1;
 
 	DEBUG("start: parse_flat_binary_to_payload\n");
-	if (buffer_create(output, (2 * sizeof(*segs) + input->size),
+	if (buffer_create(output, (sizeof(segs) + input->size),
 			  input->name) != 0)
 		return -1;
 	memset(output->data, 0, output->size);
 
-	segs = (struct cbfs_payload_segment *)output->data;
 	doffset = (2 * sizeof(*segs));
 
 	/* Prepare code segment */
 	segs[0].type = PAYLOAD_SEGMENT_CODE;
-	segs[0].load_addr = htonll(loadaddress);
-	segs[0].mem_len = htonl(input->size);
-	segs[0].offset = htonl(doffset);
+	segs[0].load_addr = loadaddress;
+	segs[0].mem_len = input->size;
+	segs[0].offset = doffset;
 
 	compress(input->data, input->size, output->data + doffset, &len);
-	segs[0].compression = htonl(algo);
-	segs[0].len = htonl(len);
+	segs[0].compression = algo;
+	segs[0].len = len;
 
 	if ((unsigned int)len >= input->size) {
 		WARN("Compressing data would make it bigger - disabled.\n");
 		segs[0].compression = 0;
-		segs[0].len = htonl(input->size);
+		segs[0].len = input->size;
 		memcpy(output->data + doffset, input->data, input->size);
 	}
 
 	/* prepare entry point segment */
 	segs[1].type = PAYLOAD_SEGMENT_ENTRY;
-	segs[1].load_addr = htonll(entrypoint);
-	output->size = doffset + ntohl(segs[0].len);
-
+	segs[1].load_addr = entrypoint;
+	output->size = doffset + segs[0].len;
+	xdr_segs(output, segs, 2);
 	return 0;
 }
 
@@ -254,7 +267,7 @@
 			 struct buffer *output, comp_algo algo)
 {
 	comp_func_ptr compress;
-	struct cbfs_payload_segment *segs;
+	struct cbfs_payload_segment segs[2];
 	int doffset, len = 0;
 	firmware_volume_header_t *fv;
 	ffs_file_header_t *fh;
@@ -343,37 +356,36 @@
 		return -1;
 	}
 
-	if (buffer_create(output, (2 * sizeof(*segs) + input->size),
+	if (buffer_create(output, (sizeof(segs) + input->size),
 			  input->name) != 0)
 		return -1;
 
 	memset(output->data, 0, output->size);
 
-	segs = (struct cbfs_payload_segment *)output->data;
-	doffset = (2 * sizeof(*segs));
+	doffset = (sizeof(segs));
 
 	/* Prepare code segment */
 	segs[0].type = PAYLOAD_SEGMENT_CODE;
-	segs[0].load_addr = htonll(loadaddress);
-	segs[0].mem_len = htonl(input->size);
-	segs[0].offset = htonl(doffset);
+	segs[0].load_addr = loadaddress;
+	segs[0].mem_len = input->size;
+	segs[0].offset = doffset;
 
 	compress(input->data, input->size, output->data + doffset, &len);
-	segs[0].compression = htonl(algo);
-	segs[0].len = htonl(len);
+	segs[0].compression = algo;
+	segs[0].len = len;
 
 	if ((unsigned int)len >= input->size) {
 		WARN("Compressing data would make it bigger - disabled.\n");
 		segs[0].compression = 0;
-		segs[0].len = htonl(input->size);
+		segs[0].len = input->size;
 		memcpy(output->data + doffset, input->data, input->size);
 	}
 
 	/* prepare entry point segment */
 	segs[1].type = PAYLOAD_SEGMENT_ENTRY;
-	segs[1].load_addr = htonll(entrypoint);
-	output->size = doffset + ntohl(segs[0].len);
-
+	segs[1].load_addr = entrypoint;
+	output->size = doffset + segs[0].len;
+	xdr_segs(output, segs, 2);
 	return 0;
 
 }