From: Michael Tremer <michael.tremer@ipfire.org>
Date: Sat, 4 Jan 2025 15:27:26 +0000 (+0000)
Subject: file: Build a function to search a file for a string
X-Git-Tag: 0.9.30~538
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=e073d0d4e2e0499e0579271056823495c5cce9a1;p=pakfire.git

file: Build a function to search a file for a string

This is done by mapping the file into memory and performing a fast
string search.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
---

diff --git a/src/pakfire/file.c b/src/pakfire/file.c
index d2e7e8048..c7a254908 100644
--- a/src/pakfire/file.c
+++ b/src/pakfire/file.c
@@ -28,6 +28,7 @@
 #include <stdlib.h>
 #include <string.h>
 #include <sys/capability.h>
+#include <sys/mman.h>
 #include <sys/sendfile.h>
 #include <sys/stat.h>
 #include <sys/types.h>
@@ -1126,55 +1127,86 @@ FILE* pakfire_file_fopen(struct pakfire_file* file, const char* mode) {
 	return f;
 }
 
-int pakfire_file_payload_matches(struct pakfire_file* file,
-		const void* needle, const size_t length) {
-	char buffer[1024 * 1024];
-	FILE* f = NULL;
-	void* p = NULL;
+/*
+	Opens the file and returns a mapping of the payload in memory
+*/
+static int pakfire_file_mmap(struct pakfire_file* file, char** data, size_t* length) {
+	struct stat st = {};
+	char* buffer = NULL;
+	int fd = -EBADF;
+	int r;
+
+	// Open the file
+	fd = r = pakfire_file_open(file, O_RDONLY);
+	if (fd < 0)
+		goto ERROR;
+
+	// Stat the file
+	r = fstat(fd, &st);
+	if (r < 0) {
+		ERROR(file->ctx, "Could not stat %s: %m\n", pakfire_file_get_path(file));
+		r = -errno;
+		goto ERROR;
+	}
+
+	// Map the data
+	buffer = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
+	if (buffer == MAP_FAILED) {
+		ERROR(file->ctx, "Could not map %s: %m\n", pakfire_file_get_path(file));
+		r = -errno;
+		goto ERROR;
+	}
+
+	// Return the values
+	*data = buffer;
+	*length = st.st_size;
+
+ERROR:
+	if (fd >= 0)
+		close(fd);
+
+	return r;
+}
+
+int pakfire_file_contains(struct pakfire_file* file, const char* needle, ssize_t length) {
+	char* haystack = NULL;
+	size_t l = 0;
+	int pos;
 	int r;
 
+	// Check inputs
+	if (!needle)
+		return -EINVAL;
+
 	const mode_t mode = pakfire_file_get_mode(file);
 
 	// Only run for regular files
 	if (!S_ISREG(mode))
 		return 0;
 
-	const size_t size = pakfire_file_get_size(file);
-
-	// Skip empty files
-	if (!size)
-		return 0;
-
-	// Open the file
-	f = pakfire_file_fopen(file, "r");
-	if (!f) {
-		r = 1;
+	// Map the file into memory
+	r = pakfire_file_mmap(file, &haystack, &l);
+	if (r < 0)
 		goto ERROR;
-	}
 
-	while (!feof(f)) {
-		size_t bytes_read = fread(buffer, 1, sizeof(buffer), f);
+	// Perform the search
+	pos = pakfire_string_search(haystack, l, needle, length);
 
-		// Raise any reading errors
-		if (ferror(f)) {
-			r = 1;
-			goto ERROR;
-		}
+	// No match found
+	if (pos == -1)
+		r = 0;
 
-		// Search for the needle
-		p = memmem(buffer, bytes_read, needle, length);
-		if (p) {
-			r = 1;
-			goto ERROR;
-		}
-	}
+	// Match found
+	else if (pos >= 0)
+		r = pos;
 
-	// No match
-	r = 0;
+	// Error
+	else if (pos < 0)
+		r = pos;
 
 ERROR:
-	if (f)
-		fclose(f);
+	if (haystack)
+		munmap(haystack, l);
 
 	return r;
 }
diff --git a/src/pakfire/file.h b/src/pakfire/file.h
index d5bdfbe50..e77f3d00f 100644
--- a/src/pakfire/file.h
+++ b/src/pakfire/file.h
@@ -159,8 +159,7 @@ const char* pakfire_file_get_abspath(struct pakfire_file* file);
 int pakfire_file_open(struct pakfire_file* file, int flags);
 FILE* pakfire_file_fopen(struct pakfire_file* file, const char* mode);
 
-int pakfire_file_payload_matches(struct pakfire_file* file,
-	const void* needle, const size_t length);
+int pakfire_file_contains(struct pakfire_file* file, const char* needle, ssize_t length);
 
 int pakfire_file_compute_digests(struct pakfire_file* file, const int types);