2017-09-23 13:56:58 +08:00
|
|
|
// Copyright 2015 Light Code Labs, LLC
|
|
|
|
//
|
|
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
|
|
// you may not use this file except in compliance with the License.
|
|
|
|
// You may obtain a copy of the License at
|
|
|
|
//
|
|
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
//
|
|
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
|
|
// See the License for the specific language governing permissions and
|
|
|
|
// limitations under the License.
|
|
|
|
|
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 (
|
|
|
|
"bytes"
|
|
|
|
"crypto"
|
|
|
|
"crypto/ecdsa"
|
|
|
|
"crypto/elliptic"
|
|
|
|
"crypto/rand"
|
|
|
|
"crypto/rsa"
|
|
|
|
"crypto/tls"
|
|
|
|
"crypto/x509"
|
|
|
|
"crypto/x509/pkix"
|
|
|
|
"encoding/pem"
|
|
|
|
"errors"
|
|
|
|
"fmt"
|
2016-08-10 06:12:22 +08:00
|
|
|
"hash/fnv"
|
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
|
|
|
"io"
|
2016-08-10 06:12:22 +08:00
|
|
|
"io/ioutil"
|
|
|
|
"log"
|
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
|
|
|
"math/big"
|
|
|
|
"net"
|
2016-08-10 06:12:22 +08:00
|
|
|
"os"
|
|
|
|
"path/filepath"
|
2017-08-29 09:21:17 +08:00
|
|
|
"sync"
|
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
|
|
|
"time"
|
|
|
|
|
2016-08-10 06:12:22 +08:00
|
|
|
"golang.org/x/crypto/ocsp"
|
|
|
|
|
|
|
|
"github.com/mholt/caddy"
|
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
|
|
|
"github.com/xenolf/lego/acme"
|
|
|
|
)
|
|
|
|
|
2016-07-08 21:32:31 +08:00
|
|
|
// loadPrivateKey loads a PEM-encoded ECC/RSA private key from an array of bytes.
|
|
|
|
func loadPrivateKey(keyBytes []byte) (crypto.PrivateKey, error) {
|
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
|
|
|
keyBlock, _ := pem.Decode(keyBytes)
|
|
|
|
|
|
|
|
switch keyBlock.Type {
|
|
|
|
case "RSA PRIVATE KEY":
|
|
|
|
return x509.ParsePKCS1PrivateKey(keyBlock.Bytes)
|
|
|
|
case "EC PRIVATE KEY":
|
|
|
|
return x509.ParseECPrivateKey(keyBlock.Bytes)
|
|
|
|
}
|
|
|
|
|
|
|
|
return nil, errors.New("unknown private key type")
|
|
|
|
}
|
|
|
|
|
2016-07-08 21:32:31 +08:00
|
|
|
// savePrivateKey saves a PEM-encoded ECC/RSA private key to an array of bytes.
|
|
|
|
func savePrivateKey(key crypto.PrivateKey) ([]byte, error) {
|
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
|
|
|
var pemType string
|
|
|
|
var keyBytes []byte
|
|
|
|
switch key := key.(type) {
|
|
|
|
case *ecdsa.PrivateKey:
|
|
|
|
var err error
|
|
|
|
pemType = "EC"
|
|
|
|
keyBytes, err = x509.MarshalECPrivateKey(key)
|
|
|
|
if err != nil {
|
2016-07-08 21:32:31 +08:00
|
|
|
return nil, err
|
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
|
|
|
}
|
|
|
|
case *rsa.PrivateKey:
|
|
|
|
pemType = "RSA"
|
|
|
|
keyBytes = x509.MarshalPKCS1PrivateKey(key)
|
|
|
|
}
|
|
|
|
|
|
|
|
pemKey := pem.Block{Type: pemType + " PRIVATE KEY", Bytes: keyBytes}
|
2016-07-08 21:32:31 +08:00
|
|
|
return pem.EncodeToMemory(&pemKey), nil
|
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
|
|
|
}
|
|
|
|
|
|
|
|
// stapleOCSP staples OCSP information to cert for hostname name.
|
|
|
|
// If you have it handy, you should pass in the PEM-encoded certificate
|
|
|
|
// bundle; otherwise the DER-encoded cert will have to be PEM-encoded.
|
2016-08-10 06:12:22 +08:00
|
|
|
// If you don't have the PEM blocks already, just pass in nil.
|
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
|
|
|
//
|
|
|
|
// Errors here are not necessarily fatal, it could just be that the
|
|
|
|
// certificate doesn't have an issuer URL.
|
|
|
|
func stapleOCSP(cert *Certificate, pemBundle []byte) error {
|
|
|
|
if pemBundle == nil {
|
|
|
|
// The function in the acme package that gets OCSP requires a PEM-encoded cert
|
|
|
|
bundle := new(bytes.Buffer)
|
|
|
|
for _, derBytes := range cert.Certificate.Certificate {
|
|
|
|
pem.Encode(bundle, &pem.Block{Type: "CERTIFICATE", Bytes: derBytes})
|
|
|
|
}
|
|
|
|
pemBundle = bundle.Bytes()
|
|
|
|
}
|
|
|
|
|
2016-08-10 06:12:22 +08:00
|
|
|
var ocspBytes []byte
|
|
|
|
var ocspResp *ocsp.Response
|
|
|
|
var ocspErr error
|
|
|
|
var gotNewOCSP bool
|
|
|
|
|
|
|
|
// First try to load OCSP staple from storage and see if
|
|
|
|
// we can still use it.
|
|
|
|
// TODO: Use Storage interface instead of disk directly
|
2016-08-20 03:42:48 +08:00
|
|
|
var ocspFileNamePrefix string
|
|
|
|
if len(cert.Names) > 0 {
|
|
|
|
ocspFileNamePrefix = cert.Names[0] + "-"
|
|
|
|
}
|
|
|
|
ocspFileName := ocspFileNamePrefix + fastHash(pemBundle)
|
2016-08-10 06:12:22 +08:00
|
|
|
ocspCachePath := filepath.Join(ocspFolder, ocspFileName)
|
|
|
|
cachedOCSP, err := ioutil.ReadFile(ocspCachePath)
|
|
|
|
if err == nil {
|
|
|
|
resp, err := ocsp.ParseResponse(cachedOCSP, nil)
|
|
|
|
if err == nil {
|
|
|
|
if freshOCSP(resp) {
|
|
|
|
// staple is still fresh; use it
|
|
|
|
ocspBytes = cachedOCSP
|
|
|
|
ocspResp = resp
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
// invalid contents; delete the file
|
2016-08-24 05:47:23 +08:00
|
|
|
// (we do this independently of the maintenance routine because
|
|
|
|
// in this case we know for sure this should be a staple file
|
|
|
|
// because we loaded it by name, whereas the maintenance routine
|
|
|
|
// just iterates the list of files, even if somehow a non-staple
|
|
|
|
// file gets in the folder. in this case we are sure it is corrupt.)
|
2016-08-10 06:12:22 +08:00
|
|
|
err := os.Remove(ocspCachePath)
|
|
|
|
if err != nil {
|
|
|
|
log.Printf("[WARNING] Unable to delete invalid OCSP staple file: %v", err)
|
|
|
|
}
|
|
|
|
}
|
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
|
|
|
}
|
|
|
|
|
2016-08-10 06:12:22 +08:00
|
|
|
// If we couldn't get a fresh staple by reading the cache,
|
|
|
|
// then we need to request it from the OCSP responder
|
|
|
|
if ocspResp == nil || len(ocspBytes) == 0 {
|
|
|
|
ocspBytes, ocspResp, ocspErr = acme.GetOCSPForCert(pemBundle)
|
|
|
|
if ocspErr != nil {
|
|
|
|
// An error here is not a problem because a certificate may simply
|
|
|
|
// not contain a link to an OCSP server. But we should log it anyway.
|
|
|
|
// There's nothing else we can do to get OCSP for this certificate,
|
|
|
|
// so we can return here with the error.
|
|
|
|
return fmt.Errorf("no OCSP stapling for %v: %v", cert.Names, ocspErr)
|
|
|
|
}
|
|
|
|
gotNewOCSP = true
|
|
|
|
}
|
|
|
|
|
|
|
|
// By now, we should have a response. If good, staple it to
|
|
|
|
// the certificate. If the OCSP response was not loaded from
|
|
|
|
// storage, we persist it for next time.
|
|
|
|
if ocspResp.Status == ocsp.Good {
|
2017-10-17 07:25:55 +08:00
|
|
|
if ocspResp.NextUpdate.After(cert.NotAfter) {
|
|
|
|
// uh oh, this OCSP response expires AFTER the certificate does, that's kinda bogus.
|
|
|
|
// it was the reason a lot of Symantec-validated sites (not Caddy) went down
|
|
|
|
// in October 2017. https://twitter.com/mattiasgeniar/status/919432824708648961
|
|
|
|
return fmt.Errorf("invalid: OCSP response for %v valid after certificate expiration (%s)",
|
|
|
|
cert.Names, cert.NotAfter.Sub(ocspResp.NextUpdate))
|
|
|
|
}
|
2016-08-10 06:12:22 +08:00
|
|
|
cert.Certificate.OCSPStaple = ocspBytes
|
|
|
|
cert.OCSP = ocspResp
|
|
|
|
if gotNewOCSP {
|
|
|
|
err := os.MkdirAll(filepath.Join(caddy.AssetsPath(), "ocsp"), 0700)
|
|
|
|
if err != nil {
|
|
|
|
return fmt.Errorf("unable to make OCSP staple path for %v: %v", cert.Names, err)
|
|
|
|
}
|
|
|
|
err = ioutil.WriteFile(ocspCachePath, ocspBytes, 0644)
|
|
|
|
if err != nil {
|
|
|
|
return fmt.Errorf("unable to write OCSP staple file for %v: %v", cert.Names, err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
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
|
|
|
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// makeSelfSignedCert makes a self-signed certificate according
|
|
|
|
// to the parameters in config. It then caches the certificate
|
|
|
|
// in our cache.
|
|
|
|
func makeSelfSignedCert(config *Config) error {
|
|
|
|
// start by generating private key
|
|
|
|
var privKey interface{}
|
|
|
|
var err error
|
|
|
|
switch config.KeyType {
|
|
|
|
case "", acme.EC256:
|
|
|
|
privKey, err = ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
|
|
|
|
case acme.EC384:
|
|
|
|
privKey, err = ecdsa.GenerateKey(elliptic.P384(), rand.Reader)
|
|
|
|
case acme.RSA2048:
|
|
|
|
privKey, err = rsa.GenerateKey(rand.Reader, 2048)
|
|
|
|
case acme.RSA4096:
|
|
|
|
privKey, err = rsa.GenerateKey(rand.Reader, 4096)
|
|
|
|
case acme.RSA8192:
|
|
|
|
privKey, err = rsa.GenerateKey(rand.Reader, 8192)
|
|
|
|
default:
|
|
|
|
return fmt.Errorf("cannot generate private key; unknown key type %v", config.KeyType)
|
|
|
|
}
|
|
|
|
if err != nil {
|
|
|
|
return fmt.Errorf("failed to generate private key: %v", err)
|
|
|
|
}
|
|
|
|
|
|
|
|
// create certificate structure with proper values
|
|
|
|
notBefore := time.Now()
|
|
|
|
notAfter := notBefore.Add(24 * time.Hour * 7)
|
|
|
|
serialNumberLimit := new(big.Int).Lsh(big.NewInt(1), 128)
|
|
|
|
serialNumber, err := rand.Int(rand.Reader, serialNumberLimit)
|
|
|
|
if err != nil {
|
|
|
|
return fmt.Errorf("failed to generate serial number: %v", err)
|
|
|
|
}
|
|
|
|
cert := &x509.Certificate{
|
|
|
|
SerialNumber: serialNumber,
|
|
|
|
Subject: pkix.Name{Organization: []string{"Caddy Self-Signed"}},
|
|
|
|
NotBefore: notBefore,
|
|
|
|
NotAfter: notAfter,
|
|
|
|
KeyUsage: x509.KeyUsageKeyEncipherment | x509.KeyUsageDigitalSignature,
|
|
|
|
ExtKeyUsage: []x509.ExtKeyUsage{x509.ExtKeyUsageServerAuth},
|
|
|
|
}
|
|
|
|
if ip := net.ParseIP(config.Hostname); ip != nil {
|
|
|
|
cert.IPAddresses = append(cert.IPAddresses, ip)
|
|
|
|
} else {
|
|
|
|
cert.DNSNames = append(cert.DNSNames, config.Hostname)
|
|
|
|
}
|
|
|
|
|
|
|
|
publicKey := func(privKey interface{}) interface{} {
|
|
|
|
switch k := privKey.(type) {
|
|
|
|
case *rsa.PrivateKey:
|
|
|
|
return &k.PublicKey
|
|
|
|
case *ecdsa.PrivateKey:
|
|
|
|
return &k.PublicKey
|
|
|
|
default:
|
|
|
|
return errors.New("unknown key type")
|
|
|
|
}
|
|
|
|
}
|
|
|
|
derBytes, err := x509.CreateCertificate(rand.Reader, cert, cert, publicKey(privKey), privKey)
|
|
|
|
if err != nil {
|
|
|
|
return fmt.Errorf("could not create certificate: %v", err)
|
|
|
|
}
|
|
|
|
|
tls: Restructure and improve certificate management
- Expose the list of Caddy instances through caddy.Instances()
- Added arbitrary storage to caddy.Instance
- The cache of loaded certificates is no longer global; now scoped
per-instance, meaning upon reload (like SIGUSR1) the old cert cache
will be discarded entirely, whereas before, aggressively reloading
config that added and removed lots of sites would cause unnecessary
build-up in the cache over time.
- Key certificates in the cache by their SHA-256 hash instead of
by their names. This means certificates will not be duplicated in
memory (within each instance), making Caddy much more memory-efficient
for large-scale deployments with thousands of sites sharing certs.
- Perform name-to-certificate lookups scoped per caddytls.Config instead
of a single global lookup. This prevents certificates from stepping on
each other when they overlap in their names.
- Do not allow TLS configurations keyed by the same hostname to be
different; this now throws an error.
- Updated relevant tests, with a stark awareness that more tests are
needed.
- Change the NewContext function signature to include an *Instance.
- Strongly recommend (basically require) use of caddytls.NewConfig()
to create a new *caddytls.Config, to ensure pointers to the instance
certificate cache are initialized properly.
- Update the TLS-SNI challenge solver (even though TLS-SNI is disabled
currently on the CA side). Store temporary challenge cert in instance
cache, but do so directly by the ACME challenge name, not the hash.
Modified the getCertificate function to check the cache directly for
a name match if one isn't found otherwise. This will allow any
caddytls.Config to be able to help solve a TLS-SNI challenge, with one
extra side-effect that might actually be kind of interesting (and
useless): clients could send a certificate's hash as the SNI and
Caddy would be able to serve that certificate for the handshake.
- Do not attempt to match a "default" (random) certificate when SNI
is present but unrecognized; return no certificate so a TLS alert
happens instead.
- Store an Instance in the list of instances even while the instance
is still starting up (this allows access to the cert cache for
performing renewals at startup, etc). Will be removed from list again
if instance startup fails.
- Laid groundwork for ACMEv2 and Let's Encrypt wildcard support.
Server type plugins will need to be updated slightly to accommodate
minor adjustments to their API (like passing in an Instance). This
commit includes the changes for the HTTP server.
Certain Caddyfile configurations might error out with this change, if
they configured different TLS settings for the same hostname.
This change trades some complexity for other complexity, but ultimately
this new complexity is more correct and robust than earlier logic.
Fixes #1991
Fixes #1994
Fixes #1303
2018-02-04 15:58:27 +08:00
|
|
|
chain := [][]byte{derBytes}
|
|
|
|
|
|
|
|
config.cacheCertificate(Certificate{
|
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
|
|
|
Certificate: tls.Certificate{
|
tls: Restructure and improve certificate management
- Expose the list of Caddy instances through caddy.Instances()
- Added arbitrary storage to caddy.Instance
- The cache of loaded certificates is no longer global; now scoped
per-instance, meaning upon reload (like SIGUSR1) the old cert cache
will be discarded entirely, whereas before, aggressively reloading
config that added and removed lots of sites would cause unnecessary
build-up in the cache over time.
- Key certificates in the cache by their SHA-256 hash instead of
by their names. This means certificates will not be duplicated in
memory (within each instance), making Caddy much more memory-efficient
for large-scale deployments with thousands of sites sharing certs.
- Perform name-to-certificate lookups scoped per caddytls.Config instead
of a single global lookup. This prevents certificates from stepping on
each other when they overlap in their names.
- Do not allow TLS configurations keyed by the same hostname to be
different; this now throws an error.
- Updated relevant tests, with a stark awareness that more tests are
needed.
- Change the NewContext function signature to include an *Instance.
- Strongly recommend (basically require) use of caddytls.NewConfig()
to create a new *caddytls.Config, to ensure pointers to the instance
certificate cache are initialized properly.
- Update the TLS-SNI challenge solver (even though TLS-SNI is disabled
currently on the CA side). Store temporary challenge cert in instance
cache, but do so directly by the ACME challenge name, not the hash.
Modified the getCertificate function to check the cache directly for
a name match if one isn't found otherwise. This will allow any
caddytls.Config to be able to help solve a TLS-SNI challenge, with one
extra side-effect that might actually be kind of interesting (and
useless): clients could send a certificate's hash as the SNI and
Caddy would be able to serve that certificate for the handshake.
- Do not attempt to match a "default" (random) certificate when SNI
is present but unrecognized; return no certificate so a TLS alert
happens instead.
- Store an Instance in the list of instances even while the instance
is still starting up (this allows access to the cert cache for
performing renewals at startup, etc). Will be removed from list again
if instance startup fails.
- Laid groundwork for ACMEv2 and Let's Encrypt wildcard support.
Server type plugins will need to be updated slightly to accommodate
minor adjustments to their API (like passing in an Instance). This
commit includes the changes for the HTTP server.
Certain Caddyfile configurations might error out with this change, if
they configured different TLS settings for the same hostname.
This change trades some complexity for other complexity, but ultimately
this new complexity is more correct and robust than earlier logic.
Fixes #1991
Fixes #1994
Fixes #1303
2018-02-04 15:58:27 +08:00
|
|
|
Certificate: chain,
|
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
|
|
|
PrivateKey: privKey,
|
|
|
|
Leaf: cert,
|
|
|
|
},
|
|
|
|
Names: cert.DNSNames,
|
|
|
|
NotAfter: cert.NotAfter,
|
tls: Restructure and improve certificate management
- Expose the list of Caddy instances through caddy.Instances()
- Added arbitrary storage to caddy.Instance
- The cache of loaded certificates is no longer global; now scoped
per-instance, meaning upon reload (like SIGUSR1) the old cert cache
will be discarded entirely, whereas before, aggressively reloading
config that added and removed lots of sites would cause unnecessary
build-up in the cache over time.
- Key certificates in the cache by their SHA-256 hash instead of
by their names. This means certificates will not be duplicated in
memory (within each instance), making Caddy much more memory-efficient
for large-scale deployments with thousands of sites sharing certs.
- Perform name-to-certificate lookups scoped per caddytls.Config instead
of a single global lookup. This prevents certificates from stepping on
each other when they overlap in their names.
- Do not allow TLS configurations keyed by the same hostname to be
different; this now throws an error.
- Updated relevant tests, with a stark awareness that more tests are
needed.
- Change the NewContext function signature to include an *Instance.
- Strongly recommend (basically require) use of caddytls.NewConfig()
to create a new *caddytls.Config, to ensure pointers to the instance
certificate cache are initialized properly.
- Update the TLS-SNI challenge solver (even though TLS-SNI is disabled
currently on the CA side). Store temporary challenge cert in instance
cache, but do so directly by the ACME challenge name, not the hash.
Modified the getCertificate function to check the cache directly for
a name match if one isn't found otherwise. This will allow any
caddytls.Config to be able to help solve a TLS-SNI challenge, with one
extra side-effect that might actually be kind of interesting (and
useless): clients could send a certificate's hash as the SNI and
Caddy would be able to serve that certificate for the handshake.
- Do not attempt to match a "default" (random) certificate when SNI
is present but unrecognized; return no certificate so a TLS alert
happens instead.
- Store an Instance in the list of instances even while the instance
is still starting up (this allows access to the cert cache for
performing renewals at startup, etc). Will be removed from list again
if instance startup fails.
- Laid groundwork for ACMEv2 and Let's Encrypt wildcard support.
Server type plugins will need to be updated slightly to accommodate
minor adjustments to their API (like passing in an Instance). This
commit includes the changes for the HTTP server.
Certain Caddyfile configurations might error out with this change, if
they configured different TLS settings for the same hostname.
This change trades some complexity for other complexity, but ultimately
this new complexity is more correct and robust than earlier logic.
Fixes #1991
Fixes #1994
Fixes #1303
2018-02-04 15:58:27 +08:00
|
|
|
Hash: hashCertificateChain(chain),
|
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
|
|
|
})
|
|
|
|
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// RotateSessionTicketKeys rotates the TLS session ticket keys
|
|
|
|
// on cfg every TicketRotateInterval. It spawns a new goroutine so
|
|
|
|
// this function does NOT block. It returns a channel you should
|
|
|
|
// close when you are ready to stop the key rotation, like when the
|
|
|
|
// server using cfg is no longer running.
|
|
|
|
func RotateSessionTicketKeys(cfg *tls.Config) chan struct{} {
|
|
|
|
ch := make(chan struct{})
|
|
|
|
ticker := time.NewTicker(TicketRotateInterval)
|
|
|
|
go runTLSTicketKeyRotation(cfg, ticker, ch)
|
|
|
|
return ch
|
|
|
|
}
|
|
|
|
|
|
|
|
// Functions that may be swapped out for testing
|
|
|
|
var (
|
2017-08-29 09:21:17 +08:00
|
|
|
runTLSTicketKeyRotation = standaloneTLSTicketKeyRotation
|
|
|
|
setSessionTicketKeysTestHook = func(keys [][32]byte) [][32]byte { return keys }
|
|
|
|
setSessionTicketKeysTestHookMu sync.Mutex
|
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
|
|
|
)
|
|
|
|
|
|
|
|
// standaloneTLSTicketKeyRotation governs over the array of TLS ticket keys used to de/crypt TLS tickets.
|
|
|
|
// It periodically sets a new ticket key as the first one, used to encrypt (and decrypt),
|
|
|
|
// pushing any old ticket keys to the back, where they are considered for decryption only.
|
|
|
|
//
|
|
|
|
// Lack of entropy for the very first ticket key results in the feature being disabled (as does Go),
|
|
|
|
// later lack of entropy temporarily disables ticket key rotation.
|
|
|
|
// Old ticket keys are still phased out, though.
|
|
|
|
//
|
|
|
|
// Stops the ticker when returning.
|
|
|
|
func standaloneTLSTicketKeyRotation(c *tls.Config, ticker *time.Ticker, exitChan chan struct{}) {
|
|
|
|
defer ticker.Stop()
|
|
|
|
|
|
|
|
// The entire page should be marked as sticky, but Go cannot do that
|
|
|
|
// without resorting to syscall#Mlock. And, we don't have madvise (for NODUMP), too. ☹
|
|
|
|
keys := make([][32]byte, 1, NumTickets)
|
|
|
|
|
|
|
|
rng := c.Rand
|
|
|
|
if rng == nil {
|
|
|
|
rng = rand.Reader
|
|
|
|
}
|
|
|
|
if _, err := io.ReadFull(rng, keys[0][:]); err != nil {
|
|
|
|
c.SessionTicketsDisabled = true // bail if we don't have the entropy for the first one
|
|
|
|
return
|
|
|
|
}
|
2017-08-29 09:21:17 +08:00
|
|
|
setSessionTicketKeysTestHookMu.Lock()
|
|
|
|
setSessionTicketKeysHook := setSessionTicketKeysTestHook
|
|
|
|
setSessionTicketKeysTestHookMu.Unlock()
|
|
|
|
c.SetSessionTicketKeys(setSessionTicketKeysHook(keys))
|
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
|
|
|
|
|
|
|
for {
|
|
|
|
select {
|
|
|
|
case _, isOpen := <-exitChan:
|
|
|
|
if !isOpen {
|
|
|
|
return
|
|
|
|
}
|
|
|
|
case <-ticker.C:
|
|
|
|
rng = c.Rand // could've changed since the start
|
|
|
|
if rng == nil {
|
|
|
|
rng = rand.Reader
|
|
|
|
}
|
|
|
|
var newTicketKey [32]byte
|
|
|
|
_, err := io.ReadFull(rng, newTicketKey[:])
|
|
|
|
|
|
|
|
if len(keys) < NumTickets {
|
|
|
|
keys = append(keys, keys[0]) // manipulates the internal length
|
|
|
|
}
|
|
|
|
for idx := len(keys) - 1; idx >= 1; idx-- {
|
|
|
|
keys[idx] = keys[idx-1] // yes, this makes copies
|
|
|
|
}
|
|
|
|
|
|
|
|
if err == nil {
|
|
|
|
keys[0] = newTicketKey
|
|
|
|
}
|
|
|
|
// pushes the last key out, doesn't matter that we don't have a new one
|
2017-08-29 09:21:17 +08:00
|
|
|
c.SetSessionTicketKeys(setSessionTicketKeysHook(keys))
|
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
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-08-10 06:12:22 +08:00
|
|
|
// fastHash hashes input using a hashing algorithm that
|
|
|
|
// is fast, and returns the hash as a hex-encoded string.
|
|
|
|
// Do not use this for cryptographic purposes.
|
|
|
|
func fastHash(input []byte) string {
|
|
|
|
h := fnv.New32a()
|
|
|
|
h.Write([]byte(input))
|
|
|
|
return fmt.Sprintf("%x", h.Sum32())
|
|
|
|
}
|
|
|
|
|
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
|
|
|
const (
|
|
|
|
// NumTickets is how many tickets to hold and consider
|
|
|
|
// to decrypt TLS sessions.
|
|
|
|
NumTickets = 4
|
|
|
|
|
|
|
|
// TicketRotateInterval is how often to generate
|
|
|
|
// new ticket for TLS PFS encryption
|
|
|
|
TicketRotateInterval = 10 * time.Hour
|
|
|
|
)
|