dns/scan.go

package dns

import (
	"bufio"
	"fmt"
	"io"
	"os"
	"path/filepath"
	"strconv"
	"strings"
)

const maxTok = 2048 // Largest token we can return.

// The maximum depth of $INCLUDE directives supported by the
// ZoneParser API.
const maxIncludeDepth = 7

// Tokinize a RFC 1035 zone file. The tokenizer will normalize it:
// * Add ownernames if they are left blank;
// * Suppress sequences of spaces;
// * Make each RR fit on one line (_NEWLINE is send as last)
// * Handle comments: ;
// * Handle braces - anywhere.
const (
	// Zonefile
	zEOF = iota
	zString
	zBlank
	zQuote
	zNewline
	zRrtpe
	zOwner
	zClass
	zDirOrigin   // $ORIGIN
	zDirTTL      // $TTL
	zDirInclude  // $INCLUDE
	zDirGenerate // $GENERATE

	// Privatekey file
	zValue
	zKey

	zExpectOwnerDir      // Ownername
	zExpectOwnerBl       // Whitespace after the ownername
	zExpectAny           // Expect rrtype, ttl or class
	zExpectAnyNoClass    // Expect rrtype or ttl
	zExpectAnyNoClassBl  // The whitespace after _EXPECT_ANY_NOCLASS
	zExpectAnyNoTTL      // Expect rrtype or class
	zExpectAnyNoTTLBl    // Whitespace after _EXPECT_ANY_NOTTL
	zExpectRrtype        // Expect rrtype
	zExpectRrtypeBl      // Whitespace BEFORE rrtype
	zExpectRdata         // The first element of the rdata
	zExpectDirTTLBl      // Space after directive $TTL
	zExpectDirTTL        // Directive $TTL
	zExpectDirOriginBl   // Space after directive $ORIGIN
	zExpectDirOrigin     // Directive $ORIGIN
	zExpectDirIncludeBl  // Space after directive $INCLUDE
	zExpectDirInclude    // Directive $INCLUDE
	zExpectDirGenerate   // Directive $GENERATE
	zExpectDirGenerateBl // Space after directive $GENERATE
)

// ParseError is a parsing error. It contains the parse error and the location in the io.Reader
// where the error occurred.
type ParseError struct {
	file string
	err  string
	lex  lex
}

func (e *ParseError) Error() (s string) {
	if e.file != "" {
		s = e.file + ": "
	}
	s += "dns: " + e.err + ": " + strconv.QuoteToASCII(e.lex.token) + " at line: " +
		strconv.Itoa(e.lex.line) + ":" + strconv.Itoa(e.lex.column)
	return
}

type lex struct {
	token  string // text of the token
	err    bool   // when true, token text has lexer error
	value  uint8  // value: zString, _BLANK, etc.
	torc   uint16 // type or class as parsed in the lexer, we only need to look this up in the grammar
	line   int    // line in the file
	column int    // column in the file
}

// ttlState describes the state necessary to fill in an omitted RR TTL
type ttlState struct {
	ttl           uint32 // ttl is the current default TTL
	isByDirective bool   // isByDirective indicates whether ttl was set by a $TTL directive
}

// NewRR reads the RR contained in the string s. Only the first RR is returned.
// If s contains no records, NewRR will return nil with no error.
//
// The class defaults to IN and TTL defaults to 3600. The full zone file syntax
// like $TTL, $ORIGIN, etc. is supported. All fields of the returned RR are
// set, except RR.Header().Rdlength which is set to 0.
func NewRR(s string) (RR, error) {
	if len(s) > 0 && s[len(s)-1] != '\n' { // We need a closing newline
		return ReadRR(strings.NewReader(s+"\n"), "")
	}
	return ReadRR(strings.NewReader(s), "")
}

// ReadRR reads the RR contained in r.
//
// The string file is used in error reporting and to resolve relative
// $INCLUDE directives.
//
// See NewRR for more documentation.
func ReadRR(r io.Reader, file string) (RR, error) {
	zp := NewZoneParser(r, ".", file)
	zp.SetDefaultTTL(defaultTtl)
	zp.SetIncludeAllowed(true)
	rr, _ := zp.Next()
	return rr, zp.Err()
}

// ZoneParser is a parser for an RFC 1035 style zonefile.
//
// Each parsed RR in the zone is returned sequentially from Next. An
// optional comment can be retrieved with Comment.
//
// The directives $INCLUDE, $ORIGIN, $TTL and $GENERATE are all
// supported. Although $INCLUDE is disabled by default.
// Note that $GENERATE's range support up to a maximum of 65535 steps.
//
// Basic usage pattern when reading from a string (z) containing the
// zone data:
//
//	zp := NewZoneParser(strings.NewReader(z), "", "")
//
//	for rr, ok := zp.Next(); ok; rr, ok = zp.Next() {
//		// Do something with rr
//	}
//
//	if err := zp.Err(); err != nil {
//		// log.Println(err)
//	}
//
// Comments specified after an RR (and on the same line!) are
// returned too:
//
//	foo. IN A 10.0.0.1 ; this is a comment
//
// The text "; this is comment" is returned from Comment. Comments inside
// the RR are returned concatenated along with the RR. Comments on a line
// by themselves are discarded.
//
// Callers should not assume all returned data in an Resource Record is
// syntactically correct, e.g. illegal base64 in RRSIGs will be returned as-is.
type ZoneParser struct {
	c *zlexer

	parseErr *ParseError

	origin string
	file   string

	defttl *ttlState

	h RR_Header

	// sub is used to parse $INCLUDE files and $GENERATE directives.
	// Next, by calling subNext, forwards the resulting RRs from this
	// sub parser to the calling code.
	sub    *ZoneParser
	osFile *os.File

	includeDepth uint8

	includeAllowed     bool
	generateDisallowed bool
}

// NewZoneParser returns an RFC 1035 style zonefile parser that reads
// from r.
//
// The string file is used in error reporting and to resolve relative
// $INCLUDE directives. The string origin is used as the initial
// origin, as if the file would start with an $ORIGIN directive.
func NewZoneParser(r io.Reader, origin, file string) *ZoneParser {
	var pe *ParseError
	if origin != "" {
		origin = Fqdn(origin)
		if _, ok := IsDomainName(origin); !ok {
			pe = &ParseError{file, "bad initial origin name", lex{}}
		}
	}

	return &ZoneParser{
		c: newZLexer(r),

		parseErr: pe,

		origin: origin,
		file:   file,
	}
}

// SetDefaultTTL sets the parsers default TTL to ttl.
func (zp *ZoneParser) SetDefaultTTL(ttl uint32) {
	zp.defttl = &ttlState{ttl, false}
}

// SetIncludeAllowed controls whether $INCLUDE directives are
// allowed. $INCLUDE directives are not supported by default.
//
// The $INCLUDE directive will open and read from a user controlled
// file on the system. Even if the file is not a valid zonefile, the
// contents of the file may be revealed in error messages, such as:
//
//	/etc/passwd: dns: not a TTL: "root:x:0:0:root:/root:/bin/bash" at line: 1:31
//	/etc/shadow: dns: not a TTL: "root:$6$<redacted>::0:99999:7:::" at line: 1:125
func (zp *ZoneParser) SetIncludeAllowed(v bool) {
	zp.includeAllowed = v
}

// Err returns the first non-EOF error that was encountered by the
// ZoneParser.
func (zp *ZoneParser) Err() error {
	if zp.parseErr != nil {
		return zp.parseErr
	}

	if zp.sub != nil {
		if err := zp.sub.Err(); err != nil {
			return err
		}
	}

	return zp.c.Err()
}

func (zp *ZoneParser) setParseError(err string, l lex) (RR, bool) {
	zp.parseErr = &ParseError{zp.file, err, l}
	return nil, false
}

// Comment returns an optional text comment that occurred alongside
// the RR.
func (zp *ZoneParser) Comment() string {
	if zp.parseErr != nil {
		return ""
	}

	if zp.sub != nil {
		return zp.sub.Comment()
	}

	return zp.c.Comment()
}

func (zp *ZoneParser) subNext() (RR, bool) {
	if rr, ok := zp.sub.Next(); ok {
		return rr, true
	}

	if zp.sub.osFile != nil {
		zp.sub.osFile.Close()
		zp.sub.osFile = nil
	}

	if zp.sub.Err() != nil {
		// We have errors to surface.
		return nil, false
	}

	zp.sub = nil
	return zp.Next()
}

// Next advances the parser to the next RR in the zonefile and
// returns the (RR, true). It will return (nil, false) when the
// parsing stops, either by reaching the end of the input or an
// error. After Next returns (nil, false), the Err method will return
// any error that occurred during parsing.
func (zp *ZoneParser) Next() (RR, bool) {
	if zp.parseErr != nil {
		return nil, false
	}
	if zp.sub != nil {
		return zp.subNext()
	}

	// 6 possible beginnings of a line (_ is a space):
	//
	//   0. zRRTYPE                              -> all omitted until the rrtype
	//   1. zOwner _ zRrtype                     -> class/ttl omitted
	//   2. zOwner _ zString _ zRrtype           -> class omitted
	//   3. zOwner _ zString _ zClass  _ zRrtype -> ttl/class
	//   4. zOwner _ zClass  _ zRrtype           -> ttl omitted
	//   5. zOwner _ zClass  _ zString _ zRrtype -> class/ttl (reversed)
	//
	// After detecting these, we know the zRrtype so we can jump to functions
	// handling the rdata for each of these types.

	st := zExpectOwnerDir // initial state
	h := &zp.h

	for l, ok := zp.c.Next(); ok; l, ok = zp.c.Next() {
		// zlexer spotted an error already
		if l.err {
			return zp.setParseError(l.token, l)
		}

		switch st {
		case zExpectOwnerDir:
			// We can also expect a directive, like $TTL or $ORIGIN
			if zp.defttl != nil {
				h.Ttl = zp.defttl.ttl
			}

			h.Class = ClassINET

			switch l.value {
			case zNewline:
				st = zExpectOwnerDir
			case zOwner:
				name, ok := toAbsoluteName(l.token, zp.origin)
				if !ok {
					return zp.setParseError("bad owner name", l)
				}

				h.Name = name

				st = zExpectOwnerBl
			case zDirTTL:
				st = zExpectDirTTLBl
			case zDirOrigin:
				st = zExpectDirOriginBl
			case zDirInclude:
				st = zExpectDirIncludeBl
			case zDirGenerate:
				st = zExpectDirGenerateBl
			case zRrtpe:
				h.Rrtype = l.torc

				st = zExpectRdata
			case zClass:
				h.Class = l.torc

				st = zExpectAnyNoClassBl
			case zBlank:
				// Discard, can happen when there is nothing on the
				// line except the RR type
			case zString:
				ttl, ok := stringToTTL(l.token)
				if !ok {
					return zp.setParseError("not a TTL", l)
				}

				h.Ttl = ttl

				if zp.defttl == nil || !zp.defttl.isByDirective {
					zp.defttl = &ttlState{ttl, false}
				}

				st = zExpectAnyNoTTLBl
			default:
				return zp.setParseError("syntax error at beginning", l)
			}
		case zExpectDirIncludeBl:
			if l.value != zBlank {
				return zp.setParseError("no blank after $INCLUDE-directive", l)
			}

			st = zExpectDirInclude
		case zExpectDirInclude:
			if l.value != zString {
				return zp.setParseError("expecting $INCLUDE value, not this...", l)
			}

			neworigin := zp.origin // There may be optionally a new origin set after the filename, if not use current one
			switch l, _ := zp.c.Next(); l.value {
			case zBlank:
				l, _ := zp.c.Next()
				if l.value == zString {
					name, ok := toAbsoluteName(l.token, zp.origin)
					if !ok {
						return zp.setParseError("bad origin name", l)
					}

					neworigin = name
				}
			case zNewline, zEOF:
				// Ok
			default:
				return zp.setParseError("garbage after $INCLUDE", l)
			}

			if !zp.includeAllowed {
				return zp.setParseError("$INCLUDE directive not allowed", l)
			}
			if zp.includeDepth >= maxIncludeDepth {
				return zp.setParseError("too deeply nested $INCLUDE", l)
			}

			// Start with the new file
			includePath := l.token
			if !filepath.IsAbs(includePath) {