diff options
author | Peter Stuge <peter@stuge.se> | 2009-04-14 19:48:32 +0000 |
---|---|---|
committer | Peter Stuge <peter@stuge.se> | 2009-04-14 19:48:32 +0000 |
commit | 45ae92ff1ac863b798e20f5fdffbd53eda02d316 (patch) | |
tree | 08b4d1beab351faee1769fbc2a3c6153bd656a6a /util/cbfstool/tools/cbfs-mkpayload.c | |
parent | 483b7bbd7703ebe100ea7b10393e18712bbafc95 (diff) |
util/cbfstool/tools/rom-mk*->cbfs-mk* rename
Signed-off-by: Peter Stuge <peter@stuge.se>
Acked-by: Ronald G. Minnich <rminnich@gmail.com>
git-svn-id: svn://svn.coreboot.org/coreboot/trunk@4114 2b7e53f0-3cfb-0310-b3e9-8179ed1497e1
Diffstat (limited to 'util/cbfstool/tools/cbfs-mkpayload.c')
-rw-r--r-- | util/cbfstool/tools/cbfs-mkpayload.c | 269 |
1 files changed, 269 insertions, 0 deletions
diff --git a/util/cbfstool/tools/cbfs-mkpayload.c b/util/cbfstool/tools/cbfs-mkpayload.c new file mode 100644 index 0000000000..9b3946a926 --- /dev/null +++ b/util/cbfstool/tools/cbfs-mkpayload.c @@ -0,0 +1,269 @@ +/* + * cbfs-mkpayload + * + * Copyright (C) 2008 Jordan Crouse <jordan@cosmicpenguin.net> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; version 2 of the License. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA, 02110-1301 USA + */ + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <unistd.h> +#include "elf.h" +#include <fcntl.h> +#include <getopt.h> +#include <sys/stat.h> +#include <arpa/inet.h> + +#include "common.h" +#include "../cbfs.h" + +int parse_elf(unsigned char *input, unsigned char **output, int algo, + void (*compress) (char *, int, char *, int *)) +{ + Elf32_Phdr *phdr; + Elf32_Ehdr *ehdr; + Elf32_Shdr *shdr; + char *header; + char *strtab; + unsigned char *sptr; + int headers; + int segments = 1; + int isize = 0, osize = 0; + int doffset = 0; + struct cbfs_payload_segment *segs; + int i; + + ehdr = (Elf32_Ehdr *) input; + headers = ehdr->e_phnum; + header = (char *)ehdr; + + phdr = (Elf32_Phdr *) & (header[ehdr->e_phoff]); + shdr = (Elf32_Shdr *) & (header[ehdr->e_shoff]); + + 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++) { + char *name; + + 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")) + segments++; + } + + /* Now, regular headers - we only care about PT_LOAD headers, + * because thats what we're actually going to load + */ + + for (i = 0; i < headers; i++) { + if (phdr[i].p_type != PT_LOAD) + continue; + + /* Empty segments are never interesting */ + if (phdr[i].p_memsz == 0) + continue; + + isize += phdr[i].p_filesz; + + segments++; + } + + /* Allocate a block of memory to store the data in */ + + sptr = + calloc((segments * sizeof(struct cbfs_payload_segment)) + isize, + 1); + doffset = (segments * sizeof(struct cbfs_payload_segment)); + + if (sptr == NULL) + goto err; + + segs = (struct cbfs_payload_segment *)sptr; + segments = 0; + + for (i = 0; i < ehdr->e_shnum; i++) { + char *name; + + 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; + segs[segments].len = (unsigned int)shdr[i].sh_size; + segs[segments].offset = doffset; + + memcpy((unsigned long *)(sptr + doffset), + &header[shdr[i].sh_offset], shdr[i].sh_size); + + doffset += segs[segments].len; + osize += segs[segments].len; + + segments++; + } + } + + 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 = + (unsigned long long)htonl(phdr[i].p_paddr); + segs[segments].mem_len = (unsigned int)htonl(phdr[i].p_memsz); + segs[segments].offset = htonl(doffset); + + segments++; + continue; + } + + segs[segments].type = PAYLOAD_SEGMENT_DATA; + segs[segments].load_addr = (unsigned int)htonl(phdr[i].p_paddr); + segs[segments].mem_len = (unsigned int)htonl(phdr[i].p_memsz); + segs[segments].compression = htonl(algo); + segs[segments].offset = htonl(doffset); + + int len; + compress((char *)&header[phdr[i].p_offset], + phdr[i].p_filesz, + (char *)(sptr + doffset), &len); + segs[segments].len = htonl(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); + + memcpy((char *)(sptr + doffset), + &header[phdr[i].p_offset], phdr[i].p_filesz); + } + + doffset += ntohl(segs[segments].len); + osize += ntohl(segs[segments].len); + + segments++; + } + + segs[segments].type = PAYLOAD_SEGMENT_ENTRY; + segs[segments++].load_addr = (unsigned long long)htonl(ehdr->e_entry); + + *output = sptr; + + return (segments * sizeof(struct cbfs_payload_segment)) + osize; + +err: + return -1; +} + +int main(int argc, char **argv) +{ + void (*compress) (char *, int, char *, int *); + int algo; + + char *output = NULL; + char *input = NULL; + + unsigned char *buffer, *obuffer; + int size, osize; + + while (1) { + int option_index; + static struct option longopt[] = { + {"output", 1, 0, 'o'}, + {"lzma", 0, 0, 'l'}, + {"nocompress", 0, 0, 'n'}, + }; + + signed char ch = getopt_long(argc, argv, "o:ln", + longopt, &option_index); + + if (ch == -1) + break; + + switch (ch) { + case 'o': + output = optarg; + break; + case 'l': + algo = CBFS_COMPRESS_LZMA; + break; + case 'n': + algo = CBFS_COMPRESS_NONE; + break; + default: + //usage(); + return -1; + } + } + + if (optind < argc) + input = argv[optind]; + + if (input == NULL || !strcmp(input, "-")) + buffer = file_read_to_buffer(STDIN_FILENO, &size); + else { + printf("Reading from %s\n", input); + buffer = file_read(input, &size); + } + + if (!iself(buffer)) { + fprintf(stderr, "E: This does not appear to be an ELF file\n"); + return -1; + } + + switch (algo) { + case CBFS_COMPRESS_NONE: + compress = none_compress; + break; + case CBFS_COMPRESS_LZMA: + compress = lzma_compress; + break; + } + + osize = parse_elf(buffer, &obuffer, algo, compress); + + if (osize == -1) { + fprintf(stderr, "E: Error while converting the payload\n"); + return -1; + } + + if (output == NULL || !strcmp(output, "-")) + file_write_from_buffer(STDOUT_FILENO, obuffer, osize); + else + file_write(output, obuffer, osize); + + return 0; +} |