caddy/caddytls/tls.go

199 lines
6.1 KiB
Go
Raw Normal View History

Rewrote Caddy from the ground up; initial commit of 0.9 branch These changes span work from the last ~4 months in an effort to make Caddy more extensible, reduce the coupling between its components, and lay a more robust foundation of code going forward into 1.0. A bunch of new features have been added, too, with even higher future potential. The most significant design change is an overall inversion of dependencies. Instead of the caddy package knowing about the server and the notion of middleware and config, the caddy package exposes an interface that other components plug into. This does introduce more indirection when reading the code, but every piece is very modular and pluggable. Even the HTTP server is pluggable. The caddy package has been moved to the top level, and main has been pushed into a subfolder called caddy. The actual logic of the main file has been pushed even further into caddy/caddymain/run.go so that custom builds of Caddy can be 'go get'able. The HTTPS logic was surgically separated into two parts to divide the TLS-specific code and the HTTPS-specific code. The caddytls package can now be used by any type of server that needs TLS, not just HTTP. I also added the ability to customize nearly every aspect of TLS at the site level rather than all sites sharing the same TLS configuration. Not all of this flexibility is exposed in the Caddyfile yet, but it may be in the future. Caddy can also generate self-signed certificates in memory for the convenience of a developer working on localhost who wants HTTPS. And Caddy now supports the DNS challenge, assuming at least one DNS provider is plugged in. Dozens, if not hundreds, of other minor changes swept through the code base as I literally started from an empty main function, copying over functions or files as needed, then adjusting them to fit in the new design. Most tests have been restored and adapted to the new API, but more work is needed there. A lot of what was "impossible" before is now possible, or can be made possible with minimal disruption of the code. For example, it's fairly easy to make plugins hook into another part of the code via callbacks. Plugins can do more than just be directives; we now have plugins that customize how the Caddyfile is loaded (useful when you need to get your configuration from a remote store). Site addresses no longer need be just a host and port. They can have a path, allowing you to scope a configuration to a specific path. There is no inheretance, however; each site configuration is distinct. Thanks to amazing work by Lucas Clemente, this commit adds experimental QUIC support. Turn it on using the -quic flag; your browser may have to be configured to enable it. Almost everything is here, but you will notice that most of the middle- ware are missing. After those are transferred over, we'll be ready for beta tests. I'm very excited to get this out. Thanks for everyone's help and patience these last few months. I hope you like it!!
2016-06-05 07:00:29 +08:00
// Package caddytls facilitates the management of TLS assets and integrates
// Let's Encrypt functionality into Caddy with first-class support for
// creating and renewing certificates automatically. It also implements
// the tls directive.
//
// This package is meant to be used by Caddy server types. To use the
// tls directive, a server type must import this package and call
// RegisterConfigGetter(). The server type must make and keep track of
// the caddytls.Config structs that this package produces. It must also
// add tls to its list of directives. When it comes time to make the
// server instances, the server type can call MakeTLSConfig() to convert
// a []caddytls.Config to a single tls.Config for use in tls.NewListener().
// It is also recommended to call RotateSessionTicketKeys() when
// starting a new listener.
Rewrote Caddy from the ground up; initial commit of 0.9 branch These changes span work from the last ~4 months in an effort to make Caddy more extensible, reduce the coupling between its components, and lay a more robust foundation of code going forward into 1.0. A bunch of new features have been added, too, with even higher future potential. The most significant design change is an overall inversion of dependencies. Instead of the caddy package knowing about the server and the notion of middleware and config, the caddy package exposes an interface that other components plug into. This does introduce more indirection when reading the code, but every piece is very modular and pluggable. Even the HTTP server is pluggable. The caddy package has been moved to the top level, and main has been pushed into a subfolder called caddy. The actual logic of the main file has been pushed even further into caddy/caddymain/run.go so that custom builds of Caddy can be 'go get'able. The HTTPS logic was surgically separated into two parts to divide the TLS-specific code and the HTTPS-specific code. The caddytls package can now be used by any type of server that needs TLS, not just HTTP. I also added the ability to customize nearly every aspect of TLS at the site level rather than all sites sharing the same TLS configuration. Not all of this flexibility is exposed in the Caddyfile yet, but it may be in the future. Caddy can also generate self-signed certificates in memory for the convenience of a developer working on localhost who wants HTTPS. And Caddy now supports the DNS challenge, assuming at least one DNS provider is plugged in. Dozens, if not hundreds, of other minor changes swept through the code base as I literally started from an empty main function, copying over functions or files as needed, then adjusting them to fit in the new design. Most tests have been restored and adapted to the new API, but more work is needed there. A lot of what was "impossible" before is now possible, or can be made possible with minimal disruption of the code. For example, it's fairly easy to make plugins hook into another part of the code via callbacks. Plugins can do more than just be directives; we now have plugins that customize how the Caddyfile is loaded (useful when you need to get your configuration from a remote store). Site addresses no longer need be just a host and port. They can have a path, allowing you to scope a configuration to a specific path. There is no inheretance, however; each site configuration is distinct. Thanks to amazing work by Lucas Clemente, this commit adds experimental QUIC support. Turn it on using the -quic flag; your browser may have to be configured to enable it. Almost everything is here, but you will notice that most of the middle- ware are missing. After those are transferred over, we'll be ready for beta tests. I'm very excited to get this out. Thanks for everyone's help and patience these last few months. I hope you like it!!
2016-06-05 07:00:29 +08:00
package caddytls
import (
"encoding/json"
"io/ioutil"
"net"
"os"
"strings"
"github.com/xenolf/lego/acme"
)
// HostQualifies returns true if the hostname alone
// appears eligible for automatic HTTPS. For example,
// localhost, empty hostname, and IP addresses are
// not eligible because we cannot obtain certificates
// for those names.
func HostQualifies(hostname string) bool {
return hostname != "localhost" && // localhost is ineligible
// hostname must not be empty
strings.TrimSpace(hostname) != "" &&
// must not contain wildcard (*) characters (until CA supports it)
!strings.Contains(hostname, "*") &&
// must not start or end with a dot
!strings.HasPrefix(hostname, ".") &&
!strings.HasSuffix(hostname, ".") &&
// cannot be an IP address, see
// https://community.letsencrypt.org/t/certificate-for-static-ip/84/2?u=mholt
net.ParseIP(hostname) == nil
}
// existingCertAndKey returns true if the hostname has
// a certificate and private key in storage already under
// the storage provided, otherwise it returns false.
func existingCertAndKey(storage Storage, hostname string) bool {
_, err := os.Stat(storage.SiteCertFile(hostname))
if err != nil {
return false
}
_, err = os.Stat(storage.SiteKeyFile(hostname))
if err != nil {
return false
}
return true
}
// saveCertResource saves the certificate resource to disk. This
// includes the certificate file itself, the private key, and the
// metadata file.
func saveCertResource(storage Storage, cert acme.CertificateResource) error {
err := os.MkdirAll(storage.Site(cert.Domain), 0700)
if err != nil {
return err
}
// Save cert
err = ioutil.WriteFile(storage.SiteCertFile(cert.Domain), cert.Certificate, 0600)
if err != nil {
return err
}
// Save private key
err = ioutil.WriteFile(storage.SiteKeyFile(cert.Domain), cert.PrivateKey, 0600)
if err != nil {
return err
}
// Save cert metadata
jsonBytes, err := json.MarshalIndent(&cert, "", "\t")
if err != nil {
return err
}
err = ioutil.WriteFile(storage.SiteMetaFile(cert.Domain), jsonBytes, 0600)
if err != nil {
return err
}
return nil
}
// Revoke revokes the certificate for host via ACME protocol.
// It assumes the certificate was obtained from the
// CA at DefaultCAUrl.
func Revoke(host string) error {
client, err := newACMEClient(new(Config), true)
if err != nil {
return err
}
return client.Revoke(host)
}
// tlsSniSolver is a type that can solve tls-sni challenges using
// an existing listener and our custom, in-memory certificate cache.
type tlsSniSolver struct{}
// Present adds the challenge certificate to the cache.
func (s tlsSniSolver) Present(domain, token, keyAuth string) error {
cert, err := acme.TLSSNI01ChallengeCert(keyAuth)
if err != nil {
return err
}
cacheCertificate(Certificate{
Certificate: cert,
Names: []string{domain},
})
return nil
}
// CleanUp removes the challenge certificate from the cache.
func (s tlsSniSolver) CleanUp(domain, token, keyAuth string) error {
uncacheCertificate(domain)
return nil
}
// ConfigHolder is any type that has a Config; it presumably is
// connected to a hostname and port on which it is serving.
type ConfigHolder interface {
TLSConfig() *Config
Host() string
Port() string
}
// QualifiesForManagedTLS returns true if c qualifies for
// for managed TLS (but not on-demand TLS specifically).
// It does NOT check to see if a cert and key already exist
// for the config. If the return value is true, you should
// be OK to set c.TLSConfig().Managed to true; then you should
// check that value in the future instead, because the process
// of setting up the config may make it look like it doesn't
// qualify even though it originally did.
func QualifiesForManagedTLS(c ConfigHolder) bool {
if c == nil {
return false
}
tlsConfig := c.TLSConfig()
if tlsConfig == nil {
return false
}
return (!tlsConfig.Manual || tlsConfig.OnDemand) && // user might provide own cert and key
// if self-signed, we've already generated one to use
!tlsConfig.SelfSigned &&
// user can force-disable managed TLS
c.Port() != "80" &&
tlsConfig.ACMEEmail != "off" &&
// we get can't certs for some kinds of hostnames, but
// on-demand TLS allows empty hostnames at startup
(HostQualifies(c.Host()) || tlsConfig.OnDemand)
}
// DNSProviderConstructor is a function that takes credentials and
// returns a type that can solve the ACME DNS challenges.
type DNSProviderConstructor func(credentials ...string) (acme.ChallengeProvider, error)
// dnsProviders is the list of DNS providers that have been plugged in.
var dnsProviders = make(map[string]DNSProviderConstructor)
// RegisterDNSProvider registers provider by name for solving the ACME DNS challenge.
func RegisterDNSProvider(name string, provider DNSProviderConstructor) {
dnsProviders[name] = provider
}
var (
// DefaultEmail represents the Let's Encrypt account email to use if none provided.
DefaultEmail string
// Agreed indicates whether user has agreed to the Let's Encrypt SA.
Agreed bool
// DefaultCAUrl is the default URL to the CA's ACME directory endpoint.
// It's very important to set this unless you set it in every Config.
DefaultCAUrl string
// DefaultKeyType is used as the type of key for new certificates
// when no other key type is specified.
DefaultKeyType = acme.RSA2048
)