bpf_arena_strsearch.h

  1/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
  2/* Copyright (c) 2025 Meta Platforms, Inc. and affiliates. */
  3#pragma once
  4#include "bpf_arena_common.h"
  5
  6__noinline int bpf_arena_strlen(const char __arena *s __arg_arena)
  7{
  8	const char __arena *sc;
  9
 10	for (sc = s; *sc != '\0'; ++sc)
 11		cond_break;
 12	return sc - s;
 13}
 14
 15/**
 16 * glob_match - Shell-style pattern matching, like !fnmatch(pat, str, 0)
 17 * @pat: Shell-style pattern to match, e.g. "*.[ch]".
 18 * @str: String to match.  The pattern must match the entire string.
 19 *
 20 * Perform shell-style glob matching, returning true (1) if the match
 21 * succeeds, or false (0) if it fails.  Equivalent to !fnmatch(@pat, @str, 0).
 22 *
 23 * Pattern metacharacters are ?, *, [ and \.
 24 * (And, inside character classes, !, - and ].)
 25 *
 26 * This is small and simple implementation intended for device blacklists
 27 * where a string is matched against a number of patterns.  Thus, it
 28 * does not preprocess the patterns.  It is non-recursive, and run-time
 29 * is at most quadratic: strlen(@str)*strlen(@pat).
 30 *
 31 * An example of the worst case is glob_match("*aaaaa", "aaaaaaaaaa");
 32 * it takes 6 passes over the pattern before matching the string.
 33 *
 34 * Like !fnmatch(@pat, @str, 0) and unlike the shell, this does NOT
 35 * treat / or leading . specially; it isn't actually used for pathnames.
 36 *
 37 * Note that according to glob(7) (and unlike bash), character classes
 38 * are complemented by a leading !; this does not support the regex-style
 39 * [^a-z] syntax.
 40 *
 41 * An opening bracket without a matching close is matched literally.
 42 */
 43__noinline bool glob_match(char const __arena *pat __arg_arena, char const __arena *str __arg_arena)
 44{
 45	/*
 46	 * Backtrack to previous * on mismatch and retry starting one
 47	 * character later in the string.  Because * matches all characters
 48	 * (no exception for /), it can be easily proved that there's
 49	 * never a need to backtrack multiple levels.
 50	 */
 51	char const __arena *back_pat = NULL, *back_str;
 52
 53	/*
 54	 * Loop over each token (character or class) in pat, matching
 55	 * it against the remaining unmatched tail of str.  Return false
 56	 * on mismatch, or true after matching the trailing nul bytes.
 57	 */
 58	for (;;) {
 59		unsigned char c = *str++;
 60		unsigned char d = *pat++;
 61
 62		switch (d) {
 63		case '?':	/* Wildcard: anything but nul */
 64			if (c == '\0')
 65				return false;
 66			break;
 67		case '*':	/* Any-length wildcard */
 68			if (*pat == '\0')	/* Optimize trailing * case */
 69				return true;
 70			back_pat = pat;
 71			back_str = --str;	/* Allow zero-length match */
 72			break;
 73		case '[': {	/* Character class */
 74			bool match = false, inverted = (*pat == '!');
 75			char const __arena *class = pat + inverted;
 76			unsigned char a = *class++;
 77
 78			/*
 79			 * Iterate over each span in the character class.
 80			 * A span is either a single character a, or a
 81			 * range a-b.  The first span may begin with ']'.
 82			 */
 83			do {
 84				unsigned char b = a;
 85
 86				if (a == '\0')	/* Malformed */
 87					goto literal;
 88
 89				if (class[0] == '-' && class[1] != ']') {
 90					b = class[1];
 91
 92					if (b == '\0')
 93						goto literal;
 94
 95					class += 2;
 96					/* Any special action if a > b? */
 97				}
 98				match |= (a <= c && c <= b);
 99				cond_break;
100			} while ((a = *class++) != ']');
101
102			if (match == inverted)
103				goto backtrack;
104			pat = class;
105			}
106			break;
107		case '\\':
108			d = *pat++;
109			__attribute__((__fallthrough__));
110		default:	/* Literal character */
111literal:
112			if (c == d) {
113				if (d == '\0')
114					return true;
115				break;
116			}
117backtrack:
118			if (c == '\0' || !back_pat)
119				return false;	/* No point continuing */
120			/* Try again from last *, one character later in str. */
121			pat = back_pat;
122			str = ++back_str;
123			break;
124		}
125		cond_break;
126	}
127	return false;
128}