reverseproxy.go

     1  // Copyright 2011 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  // HTTP reverse proxy handler
     6  
     7  package httputil
     8  
     9  import (
    10  	"context"
    11  	"errors"
    12  	"fmt"
    13  	"internal/godebug"
    14  	"io"
    15  	"log"
    16  	"mime"
    17  	"net"
    18  	"net/http"
    19  	"net/http/httptrace"
    20  	"net/http/internal/ascii"
    21  	"net/textproto"
    22  	"net/url"
    23  	"strings"
    24  	"sync"
    25  	"time"
    26  
    27  	"golang.org/x/net/http/httpguts"
    28  )
    29  
    30  // A ProxyRequest contains a request to be rewritten by a [ReverseProxy].
    31  type ProxyRequest struct {
    32  	// In is the request received by the proxy.
    33  	// The Rewrite function must not modify In.
    34  	In *http.Request
    35  
    36  	// Out is the request which will be sent by the proxy.
    37  	// The Rewrite function may modify or replace this request.
    38  	// Hop-by-hop headers are removed from this request
    39  	// before Rewrite is called.
    40  	Out *http.Request
    41  }
    42  
    43  // SetURL routes the outbound request to the scheme, host, and base path
    44  // provided in target. If the target's path is "/base" and the incoming
    45  // request was for "/dir", the target request will be for "/base/dir".
    46  // To route requests without joining the incoming path,
    47  // set r.Out.URL directly.
    48  //
    49  // SetURL rewrites the outbound Host header to match the target's host.
    50  // To preserve the inbound request's Host header (the default behavior
    51  // of [NewSingleHostReverseProxy]):
    52  //
    53  //	rewriteFunc := func(r *httputil.ProxyRequest) {
    54  //		r.SetURL(url)
    55  //		r.Out.Host = r.In.Host
    56  //	}
    57  func (r *ProxyRequest) SetURL(target *url.URL) {
    58  	rewriteRequestURL(r.Out, target)
    59  	r.Out.Host = ""
    60  }
    61  
    62  // SetXForwarded sets the X-Forwarded-For, X-Forwarded-Host, and
    63  // X-Forwarded-Proto headers of the outbound request.
    64  //
    65  //   - The X-Forwarded-For header is set to the client IP address.
    66  //   - The X-Forwarded-Host header is set to the host name requested
    67  //     by the client.
    68  //   - The X-Forwarded-Proto header is set to "http" or "https", depending
    69  //     on whether the inbound request was made on a TLS-enabled connection.
    70  //
    71  // If the outbound request contains an existing X-Forwarded-For header,
    72  // SetXForwarded appends the client IP address to it. To append to the
    73  // inbound request's X-Forwarded-For header (the default behavior of
    74  // [ReverseProxy] when using a Director function), copy the header
    75  // from the inbound request before calling SetXForwarded:
    76  //
    77  //	rewriteFunc := func(r *httputil.ProxyRequest) {
    78  //		r.Out.Header["X-Forwarded-For"] = r.In.Header["X-Forwarded-For"]
    79  //		r.SetXForwarded()
    80  //	}
    81  func (r *ProxyRequest) SetXForwarded() {
    82  	clientIP, _, err := net.SplitHostPort(r.In.RemoteAddr)
    83  	if err == nil {
    84  		prior := r.Out.Header["X-Forwarded-For"]
    85  		if len(prior) > 0 {
    86  			clientIP = strings.Join(prior, ", ") + ", " + clientIP
    87  		}
    88  		r.Out.Header.Set("X-Forwarded-For", clientIP)
    89  	} else {
    90  		r.Out.Header.Del("X-Forwarded-For")
    91  	}
    92  	r.Out.Header.Set("X-Forwarded-Host", r.In.Host)
    93  	if r.In.TLS == nil {
    94  		r.Out.Header.Set("X-Forwarded-Proto", "http")
    95  	} else {
    96  		r.Out.Header.Set("X-Forwarded-Proto", "https")
    97  	}
    98  }
    99  
   100  // ReverseProxy is an HTTP Handler that takes an incoming request and
   101  // sends it to another server, proxying the response back to the
   102  // client.
   103  //
   104  // 1xx responses are forwarded to the client if the underlying
   105  // transport supports ClientTrace.Got1xxResponse.
   106  //
   107  // Hop-by-hop headers (see RFC 9110, section 7.6.1), including
   108  // Connection, Proxy-Connection, Keep-Alive, Proxy-Authenticate,
   109  // Proxy-Authorization, TE, Trailer, Transfer-Encoding, and Upgrade,
   110  // are removed from client requests and backend responses.
   111  // The Rewrite function may be used to add hop-by-hop headers to the request,
   112  // and the ModifyResponse function may be used to remove them from the response.
   113  type ReverseProxy struct {
   114  	// Rewrite must be a function which modifies
   115  	// the request into a new request to be sent
   116  	// using Transport. Its response is then copied
   117  	// back to the original client unmodified.
   118  	// Rewrite must not access the provided ProxyRequest
   119  	// or its contents after returning.
   120  	//
   121  	// The Forwarded, X-Forwarded, X-Forwarded-Host,
   122  	// and X-Forwarded-Proto headers are removed from the
   123  	// outbound request before Rewrite is called. See also
   124  	// the ProxyRequest.SetXForwarded method.
   125  	//
   126  	// Unparsable query parameters are removed from the
   127  	// outbound request before Rewrite is called.
   128  	// The Rewrite function may copy the inbound URL's
   129  	// RawQuery to the outbound URL to preserve the original
   130  	// parameter string. Note that this can lead to security
   131  	// issues if the proxy's interpretation of query parameters
   132  	// does not match that of the downstream server.
   133  	//
   134  	// At most one of Rewrite or Director may be set.
   135  	Rewrite func(*ProxyRequest)
   136  
   137  	// Director is a function which modifies
   138  	// the request into a new request to be sent
   139  	// using Transport. Its response is then copied
   140  	// back to the original client unmodified.
   141  	// Director must not access the provided Request
   142  	// after returning.
   143  	//
   144  	// By default, the X-Forwarded-For header is set to the
   145  	// value of the client IP address. If an X-Forwarded-For
   146  	// header already exists, the client IP is appended to the
   147  	// existing values. As a special case, if the header
   148  	// exists in the Request.Header map but has a nil value
   149  	// (such as when set by the Director func), the X-Forwarded-For
   150  	// header is not modified.
   151  	//
   152  	// To prevent IP spoofing, be sure to delete any pre-existing
   153  	// X-Forwarded-For header coming from the client or
   154  	// an untrusted proxy.
   155  	//
   156  	// Hop-by-hop headers are removed from the request after
   157  	// Director returns, which can remove headers added by
   158  	// Director. Use a Rewrite function instead to ensure
   159  	// modifications to the request are preserved.
   160  	//
   161  	// Unparsable query parameters are removed from the outbound
   162  	// request if Request.Form is set after Director returns.
   163  	//
   164  	// At most one of Rewrite or Director may be set.
   165  	Director func(*http.Request)
   166  
   167  	// The transport used to perform proxy requests.
   168  	// If nil, http.DefaultTransport is used.
   169  	Transport http.RoundTripper
   170  
   171  	// FlushInterval specifies the flush interval
   172  	// to flush to the client while copying the
   173  	// response body.
   174  	// If zero, no periodic flushing is done.
   175  	// A negative value means to flush immediately
   176  	// after each write to the client.
   177  	// The FlushInterval is ignored when ReverseProxy
   178  	// recognizes a response as a streaming response, or
   179  	// if its ContentLength is -1; for such responses, writes
   180  	// are flushed to the client immediately.
   181  	FlushInterval time.Duration
   182  
   183  	// ErrorLog specifies an optional logger for errors
   184  	// that occur when attempting to proxy the request.
   185  	// If nil, logging is done via the log package's standard logger.
   186  	ErrorLog *log.Logger
   187  
   188  	// BufferPool optionally specifies a buffer pool to
   189  	// get byte slices for use by io.CopyBuffer when
   190  	// copying HTTP response bodies.
   191  	BufferPool BufferPool
   192  
   193  	// ModifyResponse is an optional function that modifies the
   194  	// Response from the backend. It is called if the backend
   195  	// returns a response at all, with any HTTP status code.
   196  	// If the backend is unreachable, the optional ErrorHandler is
   197  	// called without any call to ModifyResponse.
   198  	//
   199  	// Hop-by-hop headers are removed from the response before
   200  	// calling ModifyResponse. ModifyResponse may need to remove
   201  	// additional headers to fit its deployment model, such as Alt-Svc.
   202  	//
   203  	// If ModifyResponse returns an error, ErrorHandler is called
   204  	// with its error value. If ErrorHandler is nil, its default
   205  	// implementation is used.
   206  	ModifyResponse func(*http.Response) error
   207  
   208  	// ErrorHandler is an optional function that handles errors
   209  	// reaching the backend or errors from ModifyResponse.
   210  	//
   211  	// If nil, the default is to log the provided error and return
   212  	// a 502 Status Bad Gateway response.
   213  	ErrorHandler func(http.ResponseWriter, *http.Request, error)
   214  }
   215  
   216  // A BufferPool is an interface for getting and returning temporary
   217  // byte slices for use by [io.CopyBuffer].
   218  type BufferPool interface {
   219  	Get() []byte
   220  	Put([]byte)
   221  }
   222  
   223  func singleJoiningSlash(a, b string) string {
   224  	aslash := strings.HasSuffix(a, "/")
   225  	bslash := strings.HasPrefix(b, "/")
   226  	switch {
   227  	case aslash && bslash:
   228  		return a + b[1:]
   229  	case !aslash && !bslash:
   230  		return a + "/" + b
   231  	}
   232  	return a + b
   233  }
   234  
   235  func joinURLPath(a, b *url.URL) (path, rawpath string) {
   236  	if a.RawPath == "" && b.RawPath == "" {
   237  		return singleJoiningSlash(a.Path, b.Path), ""
   238  	}
   239  	// Same as singleJoiningSlash, but uses EscapedPath to determine
   240  	// whether a slash should be added
   241  	apath := a.EscapedPath()
   242  	bpath := b.EscapedPath()
   243  
   244  	aslash := strings.HasSuffix(apath, "/")
   245  	bslash := strings.HasPrefix(bpath, "/")
   246  
   247  	switch {
   248  	case aslash && bslash:
   249  		return a.Path + b.Path[1:], apath + bpath[1:]
   250  	case !aslash && !bslash:
   251  		return a.Path + "/" + b.Path, apath + "/" + bpath
   252  	}
   253  	return a.Path + b.Path, apath + bpath
   254  }
   255  
   256  // NewSingleHostReverseProxy returns a new [ReverseProxy] that routes
   257  // URLs to the scheme, host, and base path provided in target. If the
   258  // target's path is "/base" and the incoming request was for "/dir",
   259  // the target request will be for /base/dir.
   260  //
   261  // NewSingleHostReverseProxy does not rewrite the Host header.
   262  //
   263  // To customize the ReverseProxy behavior beyond what
   264  // NewSingleHostReverseProxy provides, use ReverseProxy directly
   265  // with a Rewrite function. The ProxyRequest SetURL method
   266  // may be used to route the outbound request. (Note that SetURL,
   267  // unlike NewSingleHostReverseProxy, rewrites the Host header
   268  // of the outbound request by default.)
   269  //
   270  //	proxy := &ReverseProxy{
   271  //		Rewrite: func(r *ProxyRequest) {
   272  //			r.SetURL(target)
   273  //			r.Out.Host = r.In.Host // if desired
   274  //		},
   275  //	}
   276  func NewSingleHostReverseProxy(target *url.URL) *ReverseProxy {
   277  	director := func(req *http.Request) {
   278  		rewriteRequestURL(req, target)
   279  	}
   280  	return &ReverseProxy{Director: director}
   281  }
   282  
   283  func rewriteRequestURL(req *http.Request, target *url.URL) {
   284  	targetQuery := target.RawQuery
   285  	req.URL.Scheme = target.Scheme
   286  	req.URL.Host = target.Host
   287  	req.URL.Path, req.URL.RawPath = joinURLPath(target, req.URL)
   288  	if targetQuery == "" || req.URL.RawQuery == "" {
   289  		req.URL.RawQuery = targetQuery + req.URL.RawQuery
   290  	} else {
   291  		req.URL.RawQuery = targetQuery + "&" + req.URL.RawQuery
   292  	}
   293  }
   294  
   295  func copyHeader(dst, src http.Header) {
   296  	for k, vv := range src {
   297  		for _, v := range vv {
   298  			dst.Add(k, v)
   299  		}
   300  	}
   301  }
   302  
   303  // Hop-by-hop headers. These are removed when sent to the backend.
   304  // As of RFC 7230, hop-by-hop headers are required to appear in the
   305  // Connection header field. These are the headers defined by the
   306  // obsoleted RFC 2616 (section 13.5.1) and are used for backward
   307  // compatibility.
   308  var hopHeaders = []string{
   309  	"Connection",
   310  	"Proxy-Connection", // non-standard but still sent by libcurl and rejected by e.g. google
   311  	"Keep-Alive",
   312  	"Proxy-Authenticate",
   313  	"Proxy-Authorization",
   314  	"Te",      // canonicalized version of "TE"
   315  	"Trailer", // not Trailers per URL above; https://www.rfc-editor.org/errata_search.php?eid=4522
   316  	"Transfer-Encoding",
   317  	"Upgrade",
   318  }
   319  
   320  func (p *ReverseProxy) defaultErrorHandler(rw http.ResponseWriter, req *http.Request, err error) {
   321  	p.logf("http: proxy error: %v", err)
   322  	rw.WriteHeader(http.StatusBadGateway)
   323  }
   324  
   325  func (p *ReverseProxy) getErrorHandler() func(http.ResponseWriter, *http.Request, error) {
   326  	if p.ErrorHandler != nil {
   327  		return p.ErrorHandler
   328  	}
   329  	return p.defaultErrorHandler
   330  }
   331  
   332  // modifyResponse conditionally runs the optional ModifyResponse hook
   333  // and reports whether the request should proceed.
   334  func (p *ReverseProxy) modifyResponse(rw http.ResponseWriter, res *http.Response, req *http.Request) bool {
   335  	if p.ModifyResponse == nil {
   336  		return true
   337  	}
   338  	if err := p.ModifyResponse(res); err != nil {
   339  		res.Body.Close()
   340  		p.getErrorHandler()(rw, req, err)
   341  		return false
   342  	}
   343  	return true
   344  }
   345  
   346  func (p *ReverseProxy) ServeHTTP(rw http.ResponseWriter, req *http.Request) {
   347  	transport := p.Transport
   348  	if transport == nil {
   349  		transport = http.DefaultTransport
   350  	}
   351  
   352  	ctx := req.Context()
   353  	if ctx.Done() != nil {
   354  		// CloseNotifier predates context.Context, and has been
   355  		// entirely superseded by it. If the request contains
   356  		// a Context that carries a cancellation signal, don't
   357  		// bother spinning up a goroutine to watch the CloseNotify
   358  		// channel (if any).
   359  		//
   360  		// If the request Context has a nil Done channel (which
   361  		// means it is either context.Background, or a custom
   362  		// Context implementation with no cancellation signal),
   363  		// then consult the CloseNotifier if available.
   364  	} else if cn, ok := rw.(http.CloseNotifier); ok {
   365  		var cancel context.CancelFunc
   366  		ctx, cancel = context.WithCancel(ctx)
   367  		defer cancel()
   368  		notifyChan := cn.CloseNotify()
   369  		go func() {
   370  			select {
   371  			case <-notifyChan:
   372  				cancel()
   373  			case <-ctx.Done():
   374  			}
   375  		}()
   376  	}
   377  
   378  	outreq := req.Clone(ctx)
   379  	if req.ContentLength == 0 {
   380  		outreq.Body = nil // Issue 16036: nil Body for http.Transport retries
   381  	}
   382  	if outreq.Body != nil {
   383  		// Reading from the request body after returning from a handler is not
   384  		// allowed, and the RoundTrip goroutine that reads the Body can outlive
   385  		// this handler. This can lead to a crash if the handler panics (see
   386  		// Issue 46866). Although calling Close doesn't guarantee there isn't
   387  		// any Read in flight after the handle returns, in practice it's safe to
   388  		// read after closing it.
   389  		defer outreq.Body.Close()
   390  	}
   391  	if outreq.Header == nil {
   392  		outreq.Header = make(http.Header) // Issue 33142: historical behavior was to always allocate
   393  	}
   394  
   395  	if (p.Director != nil) == (p.Rewrite != nil) {
   396  		p.getErrorHandler()(rw, req, errors.New("ReverseProxy must have exactly one of Director or Rewrite set"))
   397  		return
   398  	}
   399  
   400  	if p.Director != nil {
   401  		p.Director(outreq)
   402  		if outreq.Form != nil {
   403  			outreq.URL.RawQuery = cleanQueryParams(outreq.URL.RawQuery)
   404  		}
   405  	}
   406  	outreq.Close = false
   407  
   408  	reqUpType := upgradeType(outreq.Header)
   409  	if !ascii.IsPrint(reqUpType) {
   410  		p.getErrorHandler()(rw, req, fmt.Errorf("client tried to switch to invalid protocol %q", reqUpType))
   411  		return
   412  	}
   413  	removeHopByHopHeaders(outreq.Header)
   414  
   415  	// Issue 21096: tell backend applications that care about trailer support
   416  	// that we support trailers. (We do, but we don't go out of our way to
   417  	// advertise that unless the incoming client request thought it was worth
   418  	// mentioning.) Note that we look at req.Header, not outreq.Header, since
   419  	// the latter has passed through removeHopByHopHeaders.
   420  	if httpguts.HeaderValuesContainsToken(req.Header["Te"], "trailers") {
   421  		outreq.Header.Set("Te", "trailers")
   422  	}
   423  
   424  	// After stripping all the hop-by-hop connection headers above, add back any
   425  	// necessary for protocol upgrades, such as for websockets.
   426  	if reqUpType != "" {
   427  		outreq.Header.Set("Connection", "Upgrade")
   428  		outreq.Header.Set("Upgrade", reqUpType)
   429  	}
   430  
   431  	if p.Rewrite != nil {
   432  		// Strip client-provided forwarding headers.
   433  		// The Rewrite func may use SetXForwarded to set new values
   434  		// for these or copy the previous values from the inbound request.
   435  		outreq.Header.Del("Forwarded")
   436  		outreq.Header.Del("X-Forwarded-For")
   437  		outreq.Header.Del("X-Forwarded-Host")
   438  		outreq.Header.Del("X-Forwarded-Proto")
   439  
   440  		// Remove unparsable query parameters from the outbound request.
   441  		outreq.URL.RawQuery = cleanQueryParams(outreq.URL.RawQuery)
   442  
   443  		pr := &ProxyRequest{
   444  			In:  req,
   445  			Out: outreq,
   446  		}
   447  		p.Rewrite(pr)
   448  		outreq = pr.Out
   449  	} else {
   450  		if clientIP, _, err := net.SplitHostPort(req.RemoteAddr); err == nil {
   451  			// If we aren't the first proxy retain prior
   452  			// X-Forwarded-For information as a comma+space
   453  			// separated list and fold multiple headers into one.
   454  			prior, ok := outreq.Header["X-Forwarded-For"]
   455  			omit := ok && prior == nil // Issue 38079: nil now means don't populate the header
   456  			if len(prior) > 0 {
   457  				clientIP = strings.Join(prior, ", ") + ", " + clientIP
   458  			}
   459  			if !omit {
   460  				outreq.Header.Set("X-Forwarded-For", clientIP)
   461  			}
   462  		}
   463  	}
   464  
   465  	if _, ok := outreq.Header["User-Agent"]; !ok {
   466  		// If the outbound request doesn't have a User-Agent header set,
   467  		// don't send the default Go HTTP client User-Agent.
   468  		outreq.Header.Set("User-Agent", "")
   469  	}
   470  
   471  	var (
   472  		roundTripMutex sync.Mutex
   473  		roundTripDone  bool
   474  	)
   475  	trace := &httptrace.ClientTrace{
   476  		Got1xxResponse: func(code int, header textproto.MIMEHeader) error {
   477  			roundTripMutex.Lock()
   478  			defer roundTripMutex.Unlock()
   479  			if roundTripDone {
   480  				// If RoundTrip has returned, don't try to further modify
   481  				// the ResponseWriter's header map.
   482  				return nil
   483  			}
   484  			h := rw.Header()
   485  			copyHeader(h, http.Header(header))
   486  			rw.WriteHeader(code)
   487  
   488  			// Clear headers, it's not automatically done by ResponseWriter.WriteHeader() for 1xx responses
   489  			clear(h)
   490  			return nil
   491  		},
   492  	}
   493  	outreq = outreq.WithContext(httptrace.WithClientTrace(outreq.Context(), trace))
   494  
   495  	res, err := transport.RoundTrip(outreq)
   496  	roundTripMutex.Lock()
   497  	roundTripDone = true
   498  	roundTripMutex.Unlock()
   499  	if err != nil {
   500  		p.getErrorHandler()(rw, outreq, err)
   501  		return
   502  	}
   503  
   504  	// Deal with 101 Switching Protocols responses: (WebSocket, h2c, etc)
   505  	if res.StatusCode == http.StatusSwitchingProtocols {
   506  		if !p.modifyResponse(rw, res, outreq) {
   507  			return
   508  		}
   509  		p.handleUpgradeResponse(rw, outreq, res)
   510  		return
   511  	}
   512  
   513  	removeHopByHopHeaders(res.Header)
   514  
   515  	if !p.modifyResponse(rw, res, outreq) {
   516  		return
   517  	}
   518  
   519  	copyHeader(rw.Header(), res.Header)
   520  
   521  	// The "Trailer" header isn't included in the Transport's response,
   522  	// at least for *http.Transport. Build it up from Trailer.
   523  	announcedTrailers := len(res.Trailer)
   524  	if announcedTrailers > 0 {
   525  		trailerKeys := make([]string, 0, len(res.Trailer))
   526  		for k := range res.Trailer {
   527  			trailerKeys = append(trailerKeys, k)
   528  		}
   529  		rw.Header().Add("Trailer", strings.Join(trailerKeys, ", "))
   530  	}
   531  
   532  	rw.WriteHeader(res.StatusCode)
   533  
   534  	err = p.copyResponse(rw, res.Body, p.flushInterval(res))
   535  	if err != nil {
   536  		defer res.Body.Close()
   537  		// Since we're streaming the response, if we run into an error all we can do
   538  		// is abort the request. Issue 23643: ReverseProxy should use ErrAbortHandler
   539  		// on read error while copying body.
   540  		if !shouldPanicOnCopyError(req) {
   541  			p.logf("suppressing panic for copyResponse error in test; copy error: %v", err)
   542  			return
   543  		}
   544  		panic(http.ErrAbortHandler)
   545  	}
   546  	res.Body.Close() // close now, instead of defer, to populate res.Trailer
   547  
   548  	if len(res.Trailer) > 0 {
   549  		// Force chunking if we saw a response trailer.
   550  		// This prevents net/http from calculating the length for short
   551  		// bodies and adding a Content-Length.
   552  		http.NewResponseController(rw).Flush()
   553  	}
   554  
   555  	if len(res.Trailer) == announcedTrailers {
   556  		copyHeader(rw.Header(), res.Trailer)
   557  		return
   558  	}
   559  
   560  	for k, vv := range res.Trailer {
   561  		k = http.TrailerPrefix + k
   562  		for _, v := range vv {
   563  			rw.Header().Add(k, v)
   564  		}
   565  	}
   566  }
   567  
   568  var inOurTests bool // whether we're in our own tests
   569  
   570  // shouldPanicOnCopyError reports whether the reverse proxy should
   571  // panic with http.ErrAbortHandler. This is the right thing to do by
   572  // default, but Go 1.10 and earlier did not, so existing unit tests
   573  // weren't expecting panics. Only panic in our own tests, or when
   574  // running under the HTTP server.
   575  func shouldPanicOnCopyError(req *http.Request) bool {
   576  	if inOurTests {
   577  		// Our tests know to handle this panic.
   578  		return true
   579  	}
   580  	if req.Context().Value(http.ServerContextKey) != nil {
   581  		// We seem to be running under an HTTP server, so
   582  		// it'll recover the panic.
   583  		return true
   584  	}
   585  	// Otherwise act like Go 1.10 and earlier to not break
   586  	// existing tests.
   587  	return false
   588  }
   589  
   590  // removeHopByHopHeaders removes hop-by-hop headers.
   591  func removeHopByHopHeaders(h http.Header) {
   592  	// RFC 7230, section 6.1: Remove headers listed in the "Connection" header.
   593  	for _, f := range h["Connection"] {
   594  		for sf := range strings.SplitSeq(f, ",") {
   595  			if sf = textproto.TrimString(sf); sf != "" {
   596  				h.Del(sf)
   597  			}
   598  		}
   599  	}
   600  	// RFC 2616, section 13.5.1: Remove a set of known hop-by-hop headers.
   601  	// This behavior is superseded by the RFC 7230 Connection header, but
   602  	// preserve it for backwards compatibility.
   603  	for _, f := range hopHeaders {
   604  		h.Del(f)
   605  	}
   606  }
   607  
   608  // flushInterval returns the p.FlushInterval value, conditionally
   609  // overriding its value for a specific request/response.
   610  func (p *ReverseProxy) flushInterval(res *http.Response) time.Duration {
   611  	resCT := res.Header.Get("Content-Type")
   612  
   613  	// For Server-Sent Events responses, flush immediately.
   614  	// The MIME type is defined in https://www.w3.org/TR/eventsource/#text-event-stream
   615  	if baseCT, _, _ := mime.ParseMediaType(resCT); baseCT == "text/event-stream" {
   616  		return -1 // negative means immediately
   617  	}
   618  
   619  	// We might have the case of streaming for which Content-Length might be unset.
   620  	if res.ContentLength == -1 {
   621  		return -1
   622  	}
   623  
   624  	return p.FlushInterval
   625  }
   626  
   627  func (p *ReverseProxy) copyResponse(dst http.ResponseWriter, src io.Reader, flushInterval time.Duration) error {
   628  	var w io.Writer = dst
   629  
   630  	if flushInterval != 0 {
   631  		mlw := &maxLatencyWriter{
   632  			dst:     dst,
   633  			flush:   http.NewResponseController(dst).Flush,
   634  			latency: flushInterval,
   635  		}
   636  		defer mlw.stop()
   637  
   638  		// set up initial timer so headers get flushed even if body writes are delayed
   639  		mlw.flushPending = true
   640  		mlw.t = time.AfterFunc(flushInterval, mlw.delayedFlush)
   641  
   642  		w = mlw
   643  	}
   644  
   645  	var buf []byte
   646  	if p.BufferPool != nil {
   647  		buf = p.BufferPool.Get()
   648  		defer p.BufferPool.Put(buf)
   649  	}
   650  	_, err := p.copyBuffer(w, src, buf)
   651  	return err
   652  }
   653  
   654  // copyBuffer returns any write errors or non-EOF read errors, and the amount
   655  // of bytes written.
   656  func (p *ReverseProxy) copyBuffer(dst io.Writer, src io.Reader, buf []byte) (int64, error) {
   657  	if len(buf) == 0 {
   658  		buf = make([]byte, 32*1024)
   659  	}
   660  	var written int64
   661  	for {
   662  		nr, rerr := src.Read(buf)
   663  		if rerr != nil && rerr != io.EOF && rerr != context.Canceled {
   664  			p.logf("httputil: ReverseProxy read error during body copy: %v", rerr)
   665  		}
   666  		if nr > 0 {
   667  			nw, werr := dst.Write(buf[:nr])
   668  			if nw > 0 {
   669  				written += int64(nw)
   670  			}
   671  			if werr != nil {
   672  				return written, werr
   673  			}
   674  			if nr != nw {
   675  				return written, io.ErrShortWrite
   676  			}
   677  		}
   678  		if rerr != nil {
   679  			if rerr == io.EOF {
   680  				rerr = nil
   681  			}
   682  			return written, rerr
   683  		}
   684  	}
   685  }
   686  
   687  func (p *ReverseProxy) logf(format string, args ...any) {
   688  	if p.ErrorLog != nil {
   689  		p.ErrorLog.Printf(format, args...)
   690  	} else {
   691  		log.Printf(format, args...)
   692  	}
   693  }
   694  
   695  type maxLatencyWriter struct {
   696  	dst     io.Writer
   697  	flush   func() error
   698  	latency time.Duration // non-zero; negative means to flush immediately
   699  
   700  	mu           sync.Mutex // protects t, flushPending, and dst.Flush
   701  	t            *time.Timer
   702  	flushPending bool
   703  }
   704  
   705  func (m *maxLatencyWriter) Write(p []byte) (n int, err error) {
   706  	m.mu.Lock()
   707  	defer m.mu.Unlock()
   708  	n, err = m.dst.Write(p)
   709  	if m.latency < 0 {
   710  		m.flush()
   711  		return
   712  	}
   713  	if m.flushPending {
   714  		return
   715  	}
   716  	if m.t == nil {
   717  		m.t = time.AfterFunc(m.latency, m.delayedFlush)
   718  	} else {
   719  		m.t.Reset(m.latency)
   720  	}
   721  	m.flushPending = true
   722  	return
   723  }
   724  
   725  func (m *maxLatencyWriter) delayedFlush() {
   726  	m.mu.Lock()
   727  	defer m.mu.Unlock()
   728  	if !m.flushPending { // if stop was called but AfterFunc already started this goroutine
   729  		return
   730  	}
   731  	m.flush()
   732  	m.flushPending = false
   733  }
   734  
   735  func (m *maxLatencyWriter) stop() {
   736  	m.mu.Lock()
   737  	defer m.mu.Unlock()
   738  	m.flushPending = false
   739  	if m.t != nil {
   740  		m.t.Stop()
   741  	}
   742  }
   743  
   744  func upgradeType(h http.Header) string {
   745  	if !httpguts.HeaderValuesContainsToken(h["Connection"], "Upgrade") {
   746  		return ""
   747  	}
   748  	return h.Get("Upgrade")
   749  }
   750  
   751  func (p *ReverseProxy) handleUpgradeResponse(rw http.ResponseWriter, req *http.Request, res *http.Response) {
   752  	reqUpType := upgradeType(req.Header)
   753  	resUpType := upgradeType(res.Header)
   754  	if !ascii.IsPrint(resUpType) { // We know reqUpType is ASCII, it's checked by the caller.
   755  		p.getErrorHandler()(rw, req, fmt.Errorf("backend tried to switch to invalid protocol %q", resUpType))
   756  		return
   757  	}
   758  	if !ascii.EqualFold(reqUpType, resUpType) {
   759  		p.getErrorHandler()(rw, req, fmt.Errorf("backend tried to switch protocol %q when %q was requested", resUpType, reqUpType))
   760  		return
   761  	}
   762  
   763  	backConn, ok := res.Body.(io.ReadWriteCloser)
   764  	if !ok {
   765  		p.getErrorHandler()(rw, req, fmt.Errorf("internal error: 101 switching protocols response with non-writable body"))
   766  		return
   767  	}
   768  
   769  	rc := http.NewResponseController(rw)
   770  	conn, brw, hijackErr := rc.Hijack()
   771  	if errors.Is(hijackErr, http.ErrNotSupported) {
   772  		p.getErrorHandler()(rw, req, fmt.Errorf("can't switch protocols using non-Hijacker ResponseWriter type %T", rw))
   773  		return
   774  	}
   775  
   776  	backConnCloseCh := make(chan bool)
   777  	go func() {
   778  		// Ensure that the cancellation of a request closes the backend.
   779  		// See issue https://golang.org/issue/35559.
   780  		select {
   781  		case <-req.Context().Done():
   782  		case <-backConnCloseCh:
   783  		}
   784  		backConn.Close()
   785  	}()
   786  	defer close(backConnCloseCh)
   787  
   788  	if hijackErr != nil {
   789  		p.getErrorHandler()(rw, req, fmt.Errorf("Hijack failed on protocol switch: %v", hijackErr))
   790  		return
   791  	}
   792  	defer conn.Close()
   793  
   794  	copyHeader(rw.Header(), res.Header)
   795  
   796  	res.Header = rw.Header()
   797  	res.Body = nil // so res.Write only writes the headers; we have res.Body in backConn above
   798  	if err := res.Write(brw); err != nil {
   799  		p.getErrorHandler()(rw, req, fmt.Errorf("response write: %v", err))
   800  		return
   801  	}
   802  	if err := brw.Flush(); err != nil {
   803  		p.getErrorHandler()(rw, req, fmt.Errorf("response flush: %v", err))
   804  		return
   805  	}
   806  	errc := make(chan error, 1)
   807  	spc := switchProtocolCopier{user: conn, backend: backConn}
   808  	go spc.copyToBackend(errc)
   809  	go spc.copyFromBackend(errc)
   810  
   811  	// Wait until both copy functions have sent on the error channel,
   812  	// or until one fails.
   813  	err := <-errc
   814  	if err == nil {
   815  		err = <-errc
   816  	}
   817  }
   818  
   819  var errCopyDone = errors.New("hijacked connection copy complete")
   820  
   821  // switchProtocolCopier exists so goroutines proxying data back and
   822  // forth have nice names in stacks.
   823  type switchProtocolCopier struct {
   824  	user, backend io.ReadWriter
   825  }
   826  
   827  func (c switchProtocolCopier) copyFromBackend(errc chan<- error) {
   828  	if _, err := io.Copy(c.user, c.backend); err != nil {
   829  		errc <- err
   830  		return
   831  	}
   832  
   833  	// backend conn has reached EOF so propogate close write to user conn
   834  	if wc, ok := c.user.(interface{ CloseWrite() error }); ok {
   835  		errc <- wc.CloseWrite()
   836  		return
   837  	}
   838  
   839  	errc <- errCopyDone
   840  }
   841  
   842  func (c switchProtocolCopier) copyToBackend(errc chan<- error) {
   843  	if _, err := io.Copy(c.backend, c.user); err != nil {
   844  		errc <- err
   845  		return
   846  	}
   847  
   848  	// user conn has reached EOF so propogate close write to backend conn
   849  	if wc, ok := c.backend.(interface{ CloseWrite() error }); ok {
   850  		errc <- wc.CloseWrite()
   851  		return
   852  	}
   853  
   854  	errc <- errCopyDone
   855  }
   856  
   857  var urlmaxqueryparams = godebug.New("urlmaxqueryparams")
   858  
   859  // Keep this in sync with net/url.
   860  const defaultMaxParams = 10000
   861  
   862  func cleanQueryParams(s string) string {
   863  	reencode := func(s string) string {
   864  		v, _ := url.ParseQuery(s)
   865  		return v.Encode()
   866  	}
   867  	if urlmaxqueryparams.Value() != "" {
   868  		// Always reencode when a non-default urlmaxqueryparams is set.
   869  		return reencode(s)
   870  	}
   871  	if numParams := strings.Count(s, "&") + 1; numParams > defaultMaxParams {
   872  		// Too many query parameters.
   873  		return reencode(s)
   874  	}
   875  	for i := 0; i < len(s); {
   876  		switch s[i] {
   877  		case ';':
   878  			return reencode(s)
   879  		case '%':
   880  			if i+2 >= len(s) || !ishex(s[i+1]) || !ishex(s[i+2]) {
   881  				return reencode(s)
   882  			}
   883  			i += 3
   884  		default:
   885  			i++
   886  		}
   887  	}
   888  	return s
   889  }
   890  
   891  func ishex(c byte) bool {
   892  	switch {
   893  	case '0' <= c && c <= '9':
   894  		return true
   895  	case 'a' <= c && c <= 'f':
   896  		return true
   897  	case 'A' <= c && c <= 'F':
   898  		return true
   899  	}
   900  	return false
   901  }
   902
View as plain text