From f70ce379de6261733dc71279c76fc76a210285e1 Mon Sep 17 00:00:00 2001
From: Daniel Roberson <daniel@badoperation.net>
Date: Sun, 13 Sep 2026 07:44:32 -0700
Subject: initial commit

---
 pam_bloom_create.c | 236 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 236 insertions(+)
 create mode 100644 pam_bloom_create.c

(limited to 'pam_bloom_create.c')

diff --git a/pam_bloom_create.c b/pam_bloom_create.c
new file mode 100644
index 0000000..fb59a24
--- /dev/null
+++ b/pam_bloom_create.c
@@ -0,0 +1,236 @@
+#define _GNU_SOURCE
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <string.h>
+#include <unistd.h>
+#include <math.h>
+
+#include "bloom.h"
+
+
+static const size_t PAM_BLOOM_MAX_ENTRY_LENGTH = 4096;
+
+
+static uint64_t
+count_lines(FILE *fp)
+{
+    char *line = NULL;
+    size_t cap = 0;
+    uint64_t count = 0;
+
+    while (getline(&line, &cap, fp) != -1) {
+        count++;
+    }
+
+    free(line);
+
+    if (ferror(fp)) {
+        perror("getline");
+        exit(EXIT_FAILURE);
+    }
+
+    return count;
+}
+
+static size_t
+trim_newline(char *line, size_t len)
+{
+    if (len > 0 && line[len - 1] == '\n') {
+        len--;
+
+        if (len > 0 && line[len - 1] == '\r') {
+            len--;
+        }
+    }
+
+    return len;
+}
+
+static int
+validate_entry(const char *line, size_t len, uint64_t line_number)
+{
+    if (len == 0) {
+        fprintf(stderr, "line %llu: empty entry\n", (unsigned long long)line_number);
+        return -1;
+    }
+
+    if (len > PAM_BLOOM_MAX_ENTRY_LENGTH) {
+        fprintf(
+            stderr,
+            "line %llu: entry exceeds maximum length of %zu bytes\n",
+            (unsigned long long)line_number,
+            PAM_BLOOM_MAX_ENTRY_LENGTH
+        );
+
+        return -1;
+    }
+
+    if (memchr(line, '\0', len) != NULL) {
+        fprintf(stderr, "line %llu: entry contains NUL byte\n", (unsigned long long)line_number);
+        return -1;
+    }
+
+    return 0;
+}
+
+int
+main(int argc, char **argv)
+{
+    double fpr = 0.001;
+    int opt;
+
+    while ((opt = getopt(argc, argv, "p:")) != -1) {
+        switch (opt) {
+        case 'p': {
+            char *end = NULL;
+
+            errno = 0;
+
+            double value = strtod(optarg, &end);
+
+            if (errno != 0 || end == optarg || *end != '\0' ||
+                !isfinite(value) || value <= 0.0 || value >= 1.0) {
+                fprintf(stderr, "invalid false-positive rate: %s\n", optarg);
+                return EXIT_FAILURE;
+            }
+
+            fpr = value;
+            break;
+        }
+
+        default:
+            fprintf(stderr, "usage: %s [-p FPR] WORDLIST OUTPUT NAME\n", argv[0]);
+            return EXIT_FAILURE;
+        }
+    }
+
+    if (argc - optind != 3) {
+        fprintf(stderr, "usage: %s [-p FPR] WORDLIST OUTPUT NAME\n", argv[0]);
+        return EXIT_FAILURE;
+    }
+
+    const char *input  = argv[optind];
+    const char *output = argv[optind + 1];
+    const char *name   = argv[optind + 2];
+
+    if (!bloom_valid_filter_name((const uint8_t *)name, strlen(name) + 1)) {
+        fprintf(stderr, "invalid filter name\n");
+        return EXIT_FAILURE;
+    }
+
+    FILE *fp = fopen(input, "rb");
+    if (!fp) {
+        perror(input);
+        return EXIT_FAILURE;
+    }
+
+    uint64_t count = count_lines(fp);
+
+    if (count == 0) {
+        fprintf(stderr, "%s: empty wordlist\n", input);
+        fclose(fp);
+        return EXIT_FAILURE;
+    }
+
+    printf("entries: %llu\n", (unsigned long long)count);
+
+    if (fseeko(fp, 0, SEEK_SET) == -1) {
+        perror("fseeko");
+        fclose(fp);
+        return EXIT_FAILURE;
+    }
+
+    if (count > SIZE_MAX) {
+      fprintf(stderr, "%s: too many entries for this platform\n", input);
+      fclose(fp);
+      return EXIT_FAILURE;
+    }
+
+    bloomfilter bf = {0};
+    bloom_error_t err = bloom_init(&bf, (size_t)count, fpr);
+
+    if (err != BF_SUCCESS) {
+        fprintf(stderr, "bloom_init: %s\n", bloom_strerror(err));
+        fclose(fp);
+        return EXIT_FAILURE;
+    }
+
+    if (!bloom_set_name(&bf, name)) {
+        fprintf(stderr, "filter name too long\n");
+        bloom_destroy(&bf);
+        fclose(fp);
+        return EXIT_FAILURE;
+    }
+
+    char *line = NULL;
+    size_t capacity = 0;
+    ssize_t nread;
+
+    uint64_t inserted = 0;
+    uint64_t line_number = 0;
+
+    while ((nread = getline(&line, &capacity, fp)) != -1) {
+        line_number++;
+
+        size_t len = trim_newline(line,	(size_t)nread);
+
+        if (validate_entry(line, len, line_number) < 0) {
+            free(line);
+            bloom_destroy(&bf);
+            fclose(fp);
+
+            return EXIT_FAILURE;
+        }
+
+        bloom_add(&bf, line, len);
+
+        inserted++;
+    }
+
+    if (ferror(fp)) {
+        perror("getline");
+
+        free(line);
+        bloom_destroy(&bf);
+        fclose(fp);
+
+        return EXIT_FAILURE;
+    }
+
+    free(line);
+    fclose(fp);
+
+    if (inserted != count) {
+        fprintf(stderr,
+            "input changed while processing (counted %llu, read %llu)\n",
+            (unsigned long long)count,
+            (unsigned long long)inserted);
+
+        bloom_destroy(&bf);
+        return EXIT_FAILURE;
+    }
+
+    err = bloom_save(&bf, output);
+
+    if (err != BF_SUCCESS) {
+        fprintf(stderr, "bloom_save: %s\n", bloom_strerror(err));
+
+        bloom_destroy(&bf);
+        return EXIT_FAILURE;
+    }
+
+    printf("filter:     %s\n", name);
+    printf("elements:   %llu\n", (unsigned long long)inserted);
+    printf("bits:       %zu\n", bf.size);
+    printf("bytes:      %zu\n", bf.bitmap_size);
+    printf("hashes:     %zu\n", bf.hashcount);
+    printf("fpr:        %.10g\n", fpr);
+    printf("saturation: %.2f%%\n", bloom_saturation(&bf));
+
+    bloom_destroy(&bf);
+
+    return EXIT_SUCCESS;
+}
-- 
cgit v1.2.3