claude-os/src/cpio.c

/**
 * @file cpio.c
 * @brief CPIO newc archive parser implementation.
 *
 * Parses CPIO archives in the SVR4/newc format. The archive is expected
 * to be loaded into memory by GRUB as a Multiboot2 module.
 */

#include "cpio.h"
#include <string.h>

/* Debug print helpers defined in kernel.c */
extern void offset_print(const char *str);
extern void print_hex(uint32_t val);

/** Pointer to the CPIO archive in memory. */
static const uint8_t *archive = NULL;

/** Size of the archive (0 if unknown). */
static uint32_t archive_len = 0;

/**
 * Parse an N-character hexadecimal ASCII string to uint32_t.
 *
 * @param s Pointer to hex string.
 * @param n Number of characters to parse.
 * @return Parsed value.
 */
static uint32_t parse_hex(const char *s, int n) {
    uint32_t val = 0;
    for (int i = 0; i < n; i++) {
        char c = s[i];
        uint32_t digit;
        if (c >= '0' && c <= '9') {
            digit = (uint32_t)(c - '0');
        } else if (c >= 'a' && c <= 'f') {
            digit = (uint32_t)(c - 'a' + 10);
        } else if (c >= 'A' && c <= 'F') {
            digit = (uint32_t)(c - 'A' + 10);
        } else {
            break;
        }
        val = (val << 4) | digit;
    }
    return val;
}

/**
 * Round up to 4-byte boundary.
 */
static inline uint32_t align4(uint32_t v) {
    return (v + 3) & ~3u;
}

/**
 * Parse a CPIO entry at the given offset.
 *
 * @param offset  Byte offset into the archive.
 * @param entry   Output entry information.
 * @return Offset of the next entry, or 0 on error/end.
 */
static uint32_t parse_entry(uint32_t offset, cpio_entry_t *entry) {
    if (!archive) return 0;

    const cpio_newc_header_t *hdr = (const cpio_newc_header_t *)(archive + offset);

    /* Verify magic */
    if (memcmp(hdr->magic, CPIO_MAGIC, 6) != 0) {
        return 0;
    }

    uint32_t namesize = parse_hex(hdr->namesize, 8);
    uint32_t filesize = parse_hex(hdr->filesize, 8);
    uint32_t mode = parse_hex(hdr->mode, 8);

    /* Filename starts right after the header */
    const char *name = (const char *)(archive + offset + CPIO_HEADER_SIZE);

    /* Data starts after header + name, aligned to 4 bytes */
    uint32_t data_offset = align4(offset + CPIO_HEADER_SIZE + namesize);
    const void *data = archive + data_offset;

    /* Next entry starts after data, aligned to 4 bytes */
    uint32_t next_offset = align4(data_offset + filesize);

    entry->name = name;
    entry->namesize = namesize;
    entry->data = data;
    entry->datasize = filesize;
    entry->mode = mode;

    return next_offset;
}

void cpio_init(const void *archive_start, uint32_t archive_size) {
    archive = (const uint8_t *)archive_start;
    archive_len = archive_size;

    offset_print("  CPIO: archive at ");
    print_hex((uint32_t)archive_start);
    offset_print("  CPIO: size = ");
    print_hex(archive_size);

    /* Count and list entries */
    uint32_t count = 0;
    uint32_t off = 0;
    cpio_entry_t entry;
    while (1) {
        uint32_t next = parse_entry(off, &entry);
        if (next == 0) break;
        if (strcmp(entry.name, CPIO_TRAILER) == 0) break;

        offset_print("  CPIO: [");
        offset_print(entry.name);
        offset_print("] size=");
        print_hex(entry.datasize);

        count++;
        off = next;
    }

    offset_print("  CPIO: ");
    print_hex(count);
    offset_print("  CPIO: files found\n");
}

int cpio_find(const char *name, cpio_entry_t *entry) {
    if (!archive) return -1;

    uint32_t off = 0;
    while (1) {
        uint32_t next = parse_entry(off, entry);
        if (next == 0) return -1;
        if (strcmp(entry->name, CPIO_TRAILER) == 0) return -1;

        /* Match by name. CPIO entries often have "./" prefix, try both. */
        if (strcmp(entry->name, name) == 0) return 0;

        /* Try matching without "./" prefix */
        if (entry->name[0] == '.' && entry->name[1] == '/' &&
            strcmp(entry->name + 2, name) == 0) {
            return 0;
        }

        /* Try matching with "./" prefix */
        if (name[0] != '.' && entry->namesize > 2) {
            /* Already handled above */
        }

        off = next;
    }
}

int cpio_next(uint32_t *offset, cpio_entry_t *entry) {
    if (!archive) return -1;

    uint32_t next = parse_entry(*offset, entry);
    if (next == 0) return -1;
    if (strcmp(entry->name, CPIO_TRAILER) == 0) return -1;

    *offset = next;
    return 0;
}

uint32_t cpio_count(void) {
    if (!archive) return 0;

    uint32_t count = 0;
    uint32_t off = 0;
    cpio_entry_t entry;
    while (1) {
        uint32_t next = parse_entry(off, &entry);
        if (next == 0) break;
        if (strcmp(entry.name, CPIO_TRAILER) == 0) break;
        count++;
        off = next;
    }
    return count;
}