postgresql/src/backend/libpq/be-secure-openssl.c

1084 lines
28 KiB
C
Raw Normal View History

/*-------------------------------------------------------------------------
*
* be-secure-openssl.c
* functions for OpenSSL support in the backend.
*
*
* Portions Copyright (c) 1996-2015, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
*
* IDENTIFICATION
* src/backend/libpq/be-secure-openssl.c
*
* Since the server static private key ($DataDir/server.key)
* will normally be stored unencrypted so that the database
* backend can restart automatically, it is important that
* we select an algorithm that continues to provide confidentiality
* even if the attacker has the server's private key. Ephemeral
* DH (EDH) keys provide this, and in fact provide Perfect Forward
* Secrecy (PFS) except for situations where the session can
* be hijacked during a periodic handshake/renegotiation.
* Even that backdoor can be closed if client certificates
* are used (since the imposter will be unable to successfully
* complete renegotiation).
*
* N.B., the static private key should still be protected to
* the largest extent possible, to minimize the risk of
* impersonations.
*
* Another benefit of EDH is that it allows the backend and
* clients to use DSA keys. DSA keys can only provide digital
* signatures, not encryption, and are often acceptable in
* jurisdictions where RSA keys are unacceptable.
*
* The downside to EDH is that it makes it impossible to
* use ssldump(1) if there's a problem establishing an SSL
* session. In this case you'll need to temporarily disable
* EDH by commenting out the callback.
*
* ...
*
* Because the risk of cryptanalysis increases as large
* amounts of data are sent with the same session key, the
* session keys are periodically renegotiated.
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
#include <sys/stat.h>
#include <signal.h>
#include <fcntl.h>
#include <ctype.h>
#include <sys/socket.h>
#include <unistd.h>
#include <netdb.h>
#include <netinet/in.h>
#ifdef HAVE_NETINET_TCP_H
#include <netinet/tcp.h>
#include <arpa/inet.h>
#endif
#include <openssl/ssl.h>
#include <openssl/dh.h>
#if SSLEAY_VERSION_NUMBER >= 0x0907000L
#include <openssl/conf.h>
#endif
#if (OPENSSL_VERSION_NUMBER >= 0x0090800fL) && !defined(OPENSSL_NO_ECDH)
#include <openssl/ec.h>
#endif
#include "libpq/libpq.h"
#include "miscadmin.h"
#include "storage/latch.h"
#include "tcop/tcopprot.h"
#include "utils/memutils.h"
static int my_sock_read(BIO *h, char *buf, int size);
static int my_sock_write(BIO *h, const char *buf, int size);
static BIO_METHOD *my_BIO_s_socket(void);
static int my_SSL_set_fd(Port *port, int fd);
static DH *load_dh_file(int keylength);
static DH *load_dh_buffer(const char *, size_t);
static DH *tmp_dh_cb(SSL *s, int is_export, int keylength);
static int verify_cb(int, X509_STORE_CTX *);
static void info_cb(const SSL *ssl, int type, int args);
static void initialize_ecdh(void);
static const char *SSLerrmessage(void);
/* are we in the middle of a renegotiation? */
static bool in_ssl_renegotiation = false;
static SSL_CTX *SSL_context = NULL;
/* ------------------------------------------------------------ */
/* Hardcoded values */
/* ------------------------------------------------------------ */
/*
* Hardcoded DH parameters, used in ephemeral DH keying.
* As discussed above, EDH protects the confidentiality of
* sessions even if the static private key is compromised,
* so we are *highly* motivated to ensure that we can use
* EDH even if the DBA... or an attacker... deletes the
* $DataDir/dh*.pem files.
*
* We could refuse SSL connections unless a good DH parameter
* file exists, but some clients may quietly renegotiate an
* unsecured connection without fully informing the user.
* Very uncool.
*
* Alternatively, the backend could attempt to load these files
* on startup if SSL is enabled - and refuse to start if any
* do not exist - but this would tend to piss off DBAs.
*
* If you want to create your own hardcoded DH parameters
* for fun and profit, review "Assigned Number for SKIP
* Protocols" (http://www.skip-vpn.org/spec/numbers.html)
* for suggestions.
*/
static const char file_dh512[] =
"-----BEGIN DH PARAMETERS-----\n\
MEYCQQD1Kv884bEpQBgRjXyEpwpy1obEAxnIByl6ypUM2Zafq9AKUJsCRtMIPWak\n\
XUGfnHy9iUsiGSa6q6Jew1XpKgVfAgEC\n\
-----END DH PARAMETERS-----\n";
static const char file_dh1024[] =
"-----BEGIN DH PARAMETERS-----\n\
MIGHAoGBAPSI/VhOSdvNILSd5JEHNmszbDgNRR0PfIizHHxbLY7288kjwEPwpVsY\n\
jY67VYy4XTjTNP18F1dDox0YbN4zISy1Kv884bEpQBgRjXyEpwpy1obEAxnIByl6\n\
ypUM2Zafq9AKUJsCRtMIPWakXUGfnHy9iUsiGSa6q6Jew1XpL3jHAgEC\n\
-----END DH PARAMETERS-----\n";
static const char file_dh2048[] =
"-----BEGIN DH PARAMETERS-----\n\
MIIBCAKCAQEA9kJXtwh/CBdyorrWqULzBej5UxE5T7bxbrlLOCDaAadWoxTpj0BV\n\
89AHxstDqZSt90xkhkn4DIO9ZekX1KHTUPj1WV/cdlJPPT2N286Z4VeSWc39uK50\n\
T8X8dryDxUcwYc58yWb/Ffm7/ZFexwGq01uejaClcjrUGvC/RgBYK+X0iP1YTknb\n\
zSC0neSRBzZrM2w4DUUdD3yIsxx8Wy2O9vPJI8BD8KVbGI2Ou1WMuF040zT9fBdX\n\
Q6MdGGzeMyEstSr/POGxKUAYEY18hKcKctaGxAMZyAcpesqVDNmWn6vQClCbAkbT\n\
CD1mpF1Bn5x8vYlLIhkmuquiXsNV6TILOwIBAg==\n\
-----END DH PARAMETERS-----\n";
static const char file_dh4096[] =
"-----BEGIN DH PARAMETERS-----\n\
MIICCAKCAgEA+hRyUsFN4VpJ1O8JLcCo/VWr19k3BCgJ4uk+d+KhehjdRqNDNyOQ\n\
l/MOyQNQfWXPeGKmOmIig6Ev/nm6Nf9Z2B1h3R4hExf+zTiHnvVPeRBhjdQi81rt\n\
Xeoh6TNrSBIKIHfUJWBh3va0TxxjQIs6IZOLeVNRLMqzeylWqMf49HsIXqbcokUS\n\
Vt1BkvLdW48j8PPv5DsKRN3tloTxqDJGo9tKvj1Fuk74A+Xda1kNhB7KFlqMyN98\n\
VETEJ6c7KpfOo30mnK30wqw3S8OtaIR/maYX72tGOno2ehFDkq3pnPtEbD2CScxc\n\
alJC+EL7RPk5c/tgeTvCngvc1KZn92Y//EI7G9tPZtylj2b56sHtMftIoYJ9+ODM\n\
sccD5Piz/rejE3Ome8EOOceUSCYAhXn8b3qvxVI1ddd1pED6FHRhFvLrZxFvBEM9\n\
ERRMp5QqOaHJkM+Dxv8Cj6MqrCbfC4u+ZErxodzuusgDgvZiLF22uxMZbobFWyte\n\
OvOzKGtwcTqO/1wV5gKkzu1ZVswVUQd5Gg8lJicwqRWyyNRczDDoG9jVDxmogKTH\n\
AaqLulO7R8Ifa1SwF2DteSGVtgWEN8gDpN3RBmmPTDngyF2DHb5qmpnznwtFKdTL\n\
KWbuHn491xNO25CQWMtem80uKw+pTnisBRF/454n1Jnhub144YRBoN8CAQI=\n\
-----END DH PARAMETERS-----\n";
/* ------------------------------------------------------------ */
/* Public interface */
/* ------------------------------------------------------------ */
/*
* Initialize global SSL context.
*/
void
be_tls_init(void)
{
struct stat buf;
STACK_OF(X509_NAME) *root_cert_list = NULL;
if (!SSL_context)
{
#if SSLEAY_VERSION_NUMBER >= 0x0907000L
OPENSSL_config(NULL);
#endif
SSL_library_init();
SSL_load_error_strings();
/*
* We use SSLv23_method() because it can negotiate use of the highest
* mutually supported protocol version, while alternatives like
* TLSv1_2_method() permit only one specific version. Note that we
* don't actually allow SSL v2 or v3, only TLS protocols (see below).
*/
SSL_context = SSL_CTX_new(SSLv23_method());
if (!SSL_context)
ereport(FATAL,
(errmsg("could not create SSL context: %s",
SSLerrmessage())));
/*
* Disable OpenSSL's moving-write-buffer sanity check, because it
* causes unnecessary failures in nonblocking send cases.
*/
SSL_CTX_set_mode(SSL_context, SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER);
/*
* Load and verify server's certificate and private key
*/
if (SSL_CTX_use_certificate_chain_file(SSL_context,
ssl_cert_file) != 1)
ereport(FATAL,
(errcode(ERRCODE_CONFIG_FILE_ERROR),
errmsg("could not load server certificate file \"%s\": %s",
ssl_cert_file, SSLerrmessage())));
if (stat(ssl_key_file, &buf) != 0)
ereport(FATAL,
(errcode_for_file_access(),
errmsg("could not access private key file \"%s\": %m",
ssl_key_file)));
/*
* Require no public access to key file.
*
* XXX temporarily suppress check when on Windows, because there may
* not be proper support for Unix-y file permissions. Need to think
* of a reasonable check to apply on Windows. (See also the data
* directory permission check in postmaster.c)
*/
#if !defined(WIN32) && !defined(__CYGWIN__)
if (!S_ISREG(buf.st_mode) || buf.st_mode & (S_IRWXG | S_IRWXO))
ereport(FATAL,
(errcode(ERRCODE_CONFIG_FILE_ERROR),
errmsg("private key file \"%s\" has group or world access",
ssl_key_file),
errdetail("Permissions should be u=rw (0600) or less.")));
#endif
if (SSL_CTX_use_PrivateKey_file(SSL_context,
ssl_key_file,
SSL_FILETYPE_PEM) != 1)
ereport(FATAL,
(errmsg("could not load private key file \"%s\": %s",
ssl_key_file, SSLerrmessage())));
if (SSL_CTX_check_private_key(SSL_context) != 1)
ereport(FATAL,
(errmsg("check of private key failed: %s",
SSLerrmessage())));
}
/* set up ephemeral DH keys, and disallow SSL v2/v3 while at it */
SSL_CTX_set_tmp_dh_callback(SSL_context, tmp_dh_cb);
SSL_CTX_set_options(SSL_context,
SSL_OP_SINGLE_DH_USE |
SSL_OP_NO_SSLv2 | SSL_OP_NO_SSLv3);
/* set up ephemeral ECDH keys */
initialize_ecdh();
/* set up the allowed cipher list */
if (SSL_CTX_set_cipher_list(SSL_context, SSLCipherSuites) != 1)
elog(FATAL, "could not set the cipher list (no valid ciphers available)");
/* Let server choose order */
if (SSLPreferServerCiphers)
SSL_CTX_set_options(SSL_context, SSL_OP_CIPHER_SERVER_PREFERENCE);
/*
* Load CA store, so we can verify client certificates if needed.
*/
if (ssl_ca_file[0])
{
if (SSL_CTX_load_verify_locations(SSL_context, ssl_ca_file, NULL) != 1 ||
(root_cert_list = SSL_load_client_CA_file(ssl_ca_file)) == NULL)
ereport(FATAL,
(errmsg("could not load root certificate file \"%s\": %s",
ssl_ca_file, SSLerrmessage())));
}
/*----------
* Load the Certificate Revocation List (CRL).
* http://searchsecurity.techtarget.com/sDefinition/0,,sid14_gci803160,00.html
*----------
*/
if (ssl_crl_file[0])
{
X509_STORE *cvstore = SSL_CTX_get_cert_store(SSL_context);
if (cvstore)
{
/* Set the flags to check against the complete CRL chain */
if (X509_STORE_load_locations(cvstore, ssl_crl_file, NULL) == 1)
{
/* OpenSSL 0.96 does not support X509_V_FLAG_CRL_CHECK */
#ifdef X509_V_FLAG_CRL_CHECK
X509_STORE_set_flags(cvstore,
X509_V_FLAG_CRL_CHECK | X509_V_FLAG_CRL_CHECK_ALL);
#else
ereport(LOG,
(errmsg("SSL certificate revocation list file \"%s\" ignored",
ssl_crl_file),
errdetail("SSL library does not support certificate revocation lists.")));
#endif
}
else
ereport(FATAL,
(errmsg("could not load SSL certificate revocation list file \"%s\": %s",
ssl_crl_file, SSLerrmessage())));
}
}
if (ssl_ca_file[0])
{
/*
* Always ask for SSL client cert, but don't fail if it's not
* presented. We might fail such connections later, depending on what
* we find in pg_hba.conf.
*/
SSL_CTX_set_verify(SSL_context,
(SSL_VERIFY_PEER |
SSL_VERIFY_CLIENT_ONCE),
verify_cb);
/* Set flag to remember CA store is successfully loaded */
ssl_loaded_verify_locations = true;
/*
* Tell OpenSSL to send the list of root certs we trust to clients in
* CertificateRequests. This lets a client with a keystore select the
* appropriate client certificate to send to us.
*/
SSL_CTX_set_client_CA_list(SSL_context, root_cert_list);
}
}
/*
* Attempt to negotiate SSL connection.
*/
int
be_tls_open_server(Port *port)
{
int r;
int err;
int waitfor;
Assert(!port->ssl);
Assert(!port->peer);
if (!(port->ssl = SSL_new(SSL_context)))
{
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("could not initialize SSL connection: %s",
SSLerrmessage())));
be_tls_close(port);
return -1;
}
if (!my_SSL_set_fd(port, port->sock))
{
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("could not set SSL socket: %s",
SSLerrmessage())));
be_tls_close(port);
return -1;
}
port->ssl_in_use = true;
aloop:
r = SSL_accept(port->ssl);
if (r <= 0)
{
err = SSL_get_error(port->ssl, r);
switch (err)
{
case SSL_ERROR_WANT_READ:
case SSL_ERROR_WANT_WRITE:
/* not allowed during connection establishment */
Assert(!port->noblock);
/*
* No need to care about timeouts/interrupts here. At this
* point authentication_timeout still employs
* StartupPacketTimeoutHandler() which directly exits.
*/
if (err == SSL_ERROR_WANT_READ)
waitfor = WL_SOCKET_READABLE;
else
waitfor = WL_SOCKET_WRITEABLE;
WaitLatchOrSocket(MyLatch, waitfor, port->sock, 0);
goto aloop;
case SSL_ERROR_SYSCALL:
if (r < 0)
ereport(COMMERROR,
(errcode_for_socket_access(),
errmsg("could not accept SSL connection: %m")));
else
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("could not accept SSL connection: EOF detected")));
break;
case SSL_ERROR_SSL:
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("could not accept SSL connection: %s",
SSLerrmessage())));
break;
case SSL_ERROR_ZERO_RETURN:
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("could not accept SSL connection: EOF detected")));
break;
default:
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("unrecognized SSL error code: %d",
err)));
break;
}
be_tls_close(port);
return -1;
}
port->count = 0;
/* Get client certificate, if available. */
port->peer = SSL_get_peer_certificate(port->ssl);
/* and extract the Common Name from it. */
port->peer_cn = NULL;
port->peer_cert_valid = false;
if (port->peer != NULL)
{
int len;
len = X509_NAME_get_text_by_NID(X509_get_subject_name(port->peer),
NID_commonName, NULL, 0);
if (len != -1)
{
char *peer_cn;
peer_cn = MemoryContextAlloc(TopMemoryContext, len + 1);
r = X509_NAME_get_text_by_NID(X509_get_subject_name(port->peer),
NID_commonName, peer_cn, len + 1);
peer_cn[len] = '\0';
if (r != len)
{
/* shouldn't happen */
pfree(peer_cn);
be_tls_close(port);
return -1;
}
/*
* Reject embedded NULLs in certificate common name to prevent
* attacks like CVE-2009-4034.
*/
if (len != strlen(peer_cn))
{
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("SSL certificate's common name contains embedded null")));
pfree(peer_cn);
be_tls_close(port);
return -1;
}
port->peer_cn = peer_cn;
}
port->peer_cert_valid = true;
}
ereport(DEBUG2,
(errmsg("SSL connection from \"%s\"",
port->peer_cn ? port->peer_cn : "(anonymous)")));
/* set up debugging/info callback */
SSL_CTX_set_info_callback(SSL_context, info_cb);
return 0;
}
/*
* Close SSL connection.
*/
void
be_tls_close(Port *port)
{
if (port->ssl)
{
SSL_shutdown(port->ssl);
SSL_free(port->ssl);
port->ssl = NULL;
port->ssl_in_use = false;
}
if (port->peer)
{
X509_free(port->peer);
port->peer = NULL;
}
if (port->peer_cn)
{
pfree(port->peer_cn);
port->peer_cn = NULL;
}
}
/*
* Read data from a secure connection.
*/
ssize_t
be_tls_read(Port *port, void *ptr, size_t len)
{
ssize_t n;
int err;
int waitfor;
Introduce and use infrastructure for interrupt processing during client reads. Up to now large swathes of backend code ran inside signal handlers while reading commands from the client, to allow for speedy reaction to asynchronous events. Most prominently shared invalidation and NOTIFY handling. That means that complex code like the starting/stopping of transactions is run in signal handlers... The required code was fragile and verbose, and is likely to contain bugs. That approach also severely limited what could be done while communicating with the client. As the read might be from within openssl it wasn't safely possible to trigger an error, e.g. to cancel a backend in idle-in-transaction state. We did that in some cases, namely fatal errors, nonetheless. Now that FE/BE communication in the backend employs non-blocking sockets and latches to block, we can quite simply interrupt reads from signal handlers by setting the latch. That allows us to signal an interrupted read, which is supposed to be retried after returning from within the ssl library. As signal handlers now only need to set the latch to guarantee timely interrupt processing, remove a fair amount of complicated & fragile code from async.c and sinval.c. We could now actually start to process some kinds of interrupts, like sinval ones, more often that before, but that seems better done separately. This work will hopefully allow to handle cases like being blocked by sending data, interrupting idle transactions and similar to be implemented without too much effort. In addition to allowing getting rid of ImmediateInterruptOK, that is. Author: Andres Freund Reviewed-By: Heikki Linnakangas
2015-02-03 22:25:20 +01:00
int latchret;
rloop:
errno = 0;
n = SSL_read(port->ssl, ptr, len);
err = SSL_get_error(port->ssl, n);
switch (err)
{
case SSL_ERROR_NONE:
port->count += n;
break;
case SSL_ERROR_WANT_READ:
case SSL_ERROR_WANT_WRITE:
/* Don't retry if the socket is in nonblocking mode. */
if (port->noblock)
{
errno = EWOULDBLOCK;
n = -1;
break;
}
Introduce and use infrastructure for interrupt processing during client reads. Up to now large swathes of backend code ran inside signal handlers while reading commands from the client, to allow for speedy reaction to asynchronous events. Most prominently shared invalidation and NOTIFY handling. That means that complex code like the starting/stopping of transactions is run in signal handlers... The required code was fragile and verbose, and is likely to contain bugs. That approach also severely limited what could be done while communicating with the client. As the read might be from within openssl it wasn't safely possible to trigger an error, e.g. to cancel a backend in idle-in-transaction state. We did that in some cases, namely fatal errors, nonetheless. Now that FE/BE communication in the backend employs non-blocking sockets and latches to block, we can quite simply interrupt reads from signal handlers by setting the latch. That allows us to signal an interrupted read, which is supposed to be retried after returning from within the ssl library. As signal handlers now only need to set the latch to guarantee timely interrupt processing, remove a fair amount of complicated & fragile code from async.c and sinval.c. We could now actually start to process some kinds of interrupts, like sinval ones, more often that before, but that seems better done separately. This work will hopefully allow to handle cases like being blocked by sending data, interrupting idle transactions and similar to be implemented without too much effort. In addition to allowing getting rid of ImmediateInterruptOK, that is. Author: Andres Freund Reviewed-By: Heikki Linnakangas
2015-02-03 22:25:20 +01:00
waitfor = WL_LATCH_SET;
if (err == SSL_ERROR_WANT_READ)
Introduce and use infrastructure for interrupt processing during client reads. Up to now large swathes of backend code ran inside signal handlers while reading commands from the client, to allow for speedy reaction to asynchronous events. Most prominently shared invalidation and NOTIFY handling. That means that complex code like the starting/stopping of transactions is run in signal handlers... The required code was fragile and verbose, and is likely to contain bugs. That approach also severely limited what could be done while communicating with the client. As the read might be from within openssl it wasn't safely possible to trigger an error, e.g. to cancel a backend in idle-in-transaction state. We did that in some cases, namely fatal errors, nonetheless. Now that FE/BE communication in the backend employs non-blocking sockets and latches to block, we can quite simply interrupt reads from signal handlers by setting the latch. That allows us to signal an interrupted read, which is supposed to be retried after returning from within the ssl library. As signal handlers now only need to set the latch to guarantee timely interrupt processing, remove a fair amount of complicated & fragile code from async.c and sinval.c. We could now actually start to process some kinds of interrupts, like sinval ones, more often that before, but that seems better done separately. This work will hopefully allow to handle cases like being blocked by sending data, interrupting idle transactions and similar to be implemented without too much effort. In addition to allowing getting rid of ImmediateInterruptOK, that is. Author: Andres Freund Reviewed-By: Heikki Linnakangas
2015-02-03 22:25:20 +01:00
waitfor |= WL_SOCKET_READABLE;
else
Introduce and use infrastructure for interrupt processing during client reads. Up to now large swathes of backend code ran inside signal handlers while reading commands from the client, to allow for speedy reaction to asynchronous events. Most prominently shared invalidation and NOTIFY handling. That means that complex code like the starting/stopping of transactions is run in signal handlers... The required code was fragile and verbose, and is likely to contain bugs. That approach also severely limited what could be done while communicating with the client. As the read might be from within openssl it wasn't safely possible to trigger an error, e.g. to cancel a backend in idle-in-transaction state. We did that in some cases, namely fatal errors, nonetheless. Now that FE/BE communication in the backend employs non-blocking sockets and latches to block, we can quite simply interrupt reads from signal handlers by setting the latch. That allows us to signal an interrupted read, which is supposed to be retried after returning from within the ssl library. As signal handlers now only need to set the latch to guarantee timely interrupt processing, remove a fair amount of complicated & fragile code from async.c and sinval.c. We could now actually start to process some kinds of interrupts, like sinval ones, more often that before, but that seems better done separately. This work will hopefully allow to handle cases like being blocked by sending data, interrupting idle transactions and similar to be implemented without too much effort. In addition to allowing getting rid of ImmediateInterruptOK, that is. Author: Andres Freund Reviewed-By: Heikki Linnakangas
2015-02-03 22:25:20 +01:00
waitfor |= WL_SOCKET_WRITEABLE;
Introduce and use infrastructure for interrupt processing during client reads. Up to now large swathes of backend code ran inside signal handlers while reading commands from the client, to allow for speedy reaction to asynchronous events. Most prominently shared invalidation and NOTIFY handling. That means that complex code like the starting/stopping of transactions is run in signal handlers... The required code was fragile and verbose, and is likely to contain bugs. That approach also severely limited what could be done while communicating with the client. As the read might be from within openssl it wasn't safely possible to trigger an error, e.g. to cancel a backend in idle-in-transaction state. We did that in some cases, namely fatal errors, nonetheless. Now that FE/BE communication in the backend employs non-blocking sockets and latches to block, we can quite simply interrupt reads from signal handlers by setting the latch. That allows us to signal an interrupted read, which is supposed to be retried after returning from within the ssl library. As signal handlers now only need to set the latch to guarantee timely interrupt processing, remove a fair amount of complicated & fragile code from async.c and sinval.c. We could now actually start to process some kinds of interrupts, like sinval ones, more often that before, but that seems better done separately. This work will hopefully allow to handle cases like being blocked by sending data, interrupting idle transactions and similar to be implemented without too much effort. In addition to allowing getting rid of ImmediateInterruptOK, that is. Author: Andres Freund Reviewed-By: Heikki Linnakangas
2015-02-03 22:25:20 +01:00
latchret = WaitLatchOrSocket(MyLatch, waitfor, port->sock, 0);
/*
* We'll, among other situations, get here if the low level
* routine doing the actual recv() via the socket got interrupted
* by a signal. That's so we can handle interrupts once outside
* openssl, so we don't jump out from underneath its covers. We
* can check this both, when reading and writing, because even
* when writing that's just openssl's doing, not a 'proper' write
* initiated by postgres.
*/
if (latchret & WL_LATCH_SET)
{
ResetLatch(MyLatch);
ProcessClientReadInterrupt(true); /* preserves errno */
Introduce and use infrastructure for interrupt processing during client reads. Up to now large swathes of backend code ran inside signal handlers while reading commands from the client, to allow for speedy reaction to asynchronous events. Most prominently shared invalidation and NOTIFY handling. That means that complex code like the starting/stopping of transactions is run in signal handlers... The required code was fragile and verbose, and is likely to contain bugs. That approach also severely limited what could be done while communicating with the client. As the read might be from within openssl it wasn't safely possible to trigger an error, e.g. to cancel a backend in idle-in-transaction state. We did that in some cases, namely fatal errors, nonetheless. Now that FE/BE communication in the backend employs non-blocking sockets and latches to block, we can quite simply interrupt reads from signal handlers by setting the latch. That allows us to signal an interrupted read, which is supposed to be retried after returning from within the ssl library. As signal handlers now only need to set the latch to guarantee timely interrupt processing, remove a fair amount of complicated & fragile code from async.c and sinval.c. We could now actually start to process some kinds of interrupts, like sinval ones, more often that before, but that seems better done separately. This work will hopefully allow to handle cases like being blocked by sending data, interrupting idle transactions and similar to be implemented without too much effort. In addition to allowing getting rid of ImmediateInterruptOK, that is. Author: Andres Freund Reviewed-By: Heikki Linnakangas
2015-02-03 22:25:20 +01:00
}
goto rloop;
case SSL_ERROR_SYSCALL:
/* leave it to caller to ereport the value of errno */
if (n != -1)
{
errno = ECONNRESET;
n = -1;
}
break;
case SSL_ERROR_SSL:
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("SSL error: %s", SSLerrmessage())));
/* fall through */
case SSL_ERROR_ZERO_RETURN:
errno = ECONNRESET;
n = -1;
break;
default:
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("unrecognized SSL error code: %d",
err)));
errno = ECONNRESET;
n = -1;
break;
}
return n;
}
/*
* Write data to a secure connection.
*/
ssize_t
be_tls_write(Port *port, void *ptr, size_t len)
{
ssize_t n;
int err;
int waitfor;
int latchret;
/*
* If SSL renegotiations are enabled and we're getting close to the
* limit, start one now; but avoid it if there's one already in
* progress. Request the renegotiation 1kB before the limit has
* actually expired.
*/
if (ssl_renegotiation_limit && !in_ssl_renegotiation &&
port->count > (ssl_renegotiation_limit - 1) * 1024L)
{
in_ssl_renegotiation = true;
/*
* The way we determine that a renegotiation has completed is by
* observing OpenSSL's internal renegotiation counter. Make sure
* we start out at zero, and assume that the renegotiation is
* complete when the counter advances.
*
* OpenSSL provides SSL_renegotiation_pending(), but this doesn't
* seem to work in testing.
*/
SSL_clear_num_renegotiations(port->ssl);
if (SSL_renegotiate(port->ssl) <= 0)
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("SSL failure during renegotiation start")));
}
wloop:
errno = 0;
n = SSL_write(port->ssl, ptr, len);
err = SSL_get_error(port->ssl, n);
switch (err)
{
case SSL_ERROR_NONE:
port->count += n;
break;
case SSL_ERROR_WANT_READ:
case SSL_ERROR_WANT_WRITE:
waitfor = WL_LATCH_SET;
if (err == SSL_ERROR_WANT_READ)
waitfor |= WL_SOCKET_READABLE;
else
waitfor |= WL_SOCKET_WRITEABLE;
latchret = WaitLatchOrSocket(MyLatch, waitfor, port->sock, 0);
Introduce and use infrastructure for interrupt processing during client reads. Up to now large swathes of backend code ran inside signal handlers while reading commands from the client, to allow for speedy reaction to asynchronous events. Most prominently shared invalidation and NOTIFY handling. That means that complex code like the starting/stopping of transactions is run in signal handlers... The required code was fragile and verbose, and is likely to contain bugs. That approach also severely limited what could be done while communicating with the client. As the read might be from within openssl it wasn't safely possible to trigger an error, e.g. to cancel a backend in idle-in-transaction state. We did that in some cases, namely fatal errors, nonetheless. Now that FE/BE communication in the backend employs non-blocking sockets and latches to block, we can quite simply interrupt reads from signal handlers by setting the latch. That allows us to signal an interrupted read, which is supposed to be retried after returning from within the ssl library. As signal handlers now only need to set the latch to guarantee timely interrupt processing, remove a fair amount of complicated & fragile code from async.c and sinval.c. We could now actually start to process some kinds of interrupts, like sinval ones, more often that before, but that seems better done separately. This work will hopefully allow to handle cases like being blocked by sending data, interrupting idle transactions and similar to be implemented without too much effort. In addition to allowing getting rid of ImmediateInterruptOK, that is. Author: Andres Freund Reviewed-By: Heikki Linnakangas
2015-02-03 22:25:20 +01:00
/*
* Check for interrupts here, in addition to secure_write(),
* because an interrupted write in secure_raw_write() will return
* here, and we cannot return to secure_write() until we've
* written something.
Introduce and use infrastructure for interrupt processing during client reads. Up to now large swathes of backend code ran inside signal handlers while reading commands from the client, to allow for speedy reaction to asynchronous events. Most prominently shared invalidation and NOTIFY handling. That means that complex code like the starting/stopping of transactions is run in signal handlers... The required code was fragile and verbose, and is likely to contain bugs. That approach also severely limited what could be done while communicating with the client. As the read might be from within openssl it wasn't safely possible to trigger an error, e.g. to cancel a backend in idle-in-transaction state. We did that in some cases, namely fatal errors, nonetheless. Now that FE/BE communication in the backend employs non-blocking sockets and latches to block, we can quite simply interrupt reads from signal handlers by setting the latch. That allows us to signal an interrupted read, which is supposed to be retried after returning from within the ssl library. As signal handlers now only need to set the latch to guarantee timely interrupt processing, remove a fair amount of complicated & fragile code from async.c and sinval.c. We could now actually start to process some kinds of interrupts, like sinval ones, more often that before, but that seems better done separately. This work will hopefully allow to handle cases like being blocked by sending data, interrupting idle transactions and similar to be implemented without too much effort. In addition to allowing getting rid of ImmediateInterruptOK, that is. Author: Andres Freund Reviewed-By: Heikki Linnakangas
2015-02-03 22:25:20 +01:00
*/
if (latchret & WL_LATCH_SET)
{
ResetLatch(MyLatch);
ProcessClientWriteInterrupt(true); /* preserves errno */
}
goto wloop;
case SSL_ERROR_SYSCALL:
/* leave it to caller to ereport the value of errno */
if (n != -1)
{
errno = ECONNRESET;
n = -1;
}
break;
case SSL_ERROR_SSL:
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("SSL error: %s", SSLerrmessage())));
/* fall through */
case SSL_ERROR_ZERO_RETURN:
errno = ECONNRESET;
n = -1;
break;
default:
ereport(COMMERROR,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("unrecognized SSL error code: %d",
err)));
errno = ECONNRESET;
n = -1;
break;
}
if (n >= 0)
{
/* is renegotiation complete? */
if (in_ssl_renegotiation &&
SSL_num_renegotiations(port->ssl) >= 1)
{
in_ssl_renegotiation = false;
port->count = 0;
}
/*
* if renegotiation is still ongoing, and we've gone beyond the
* limit, kill the connection now -- continuing to use it can be
* considered a security problem.
*/
if (in_ssl_renegotiation &&
port->count > ssl_renegotiation_limit * 1024L)
ereport(FATAL,
(errcode(ERRCODE_PROTOCOL_VIOLATION),
errmsg("SSL failed to renegotiate connection before limit expired")));
}
return n;
}
/* ------------------------------------------------------------ */
/* Internal functions */
/* ------------------------------------------------------------ */
/*
* Private substitute BIO: this does the sending and receiving using send() and
* recv() instead. This is so that we can enable and disable interrupts
* just while calling recv(). We cannot have interrupts occurring while
* the bulk of openssl runs, because it uses malloc() and possibly other
* non-reentrant libc facilities. We also need to call send() and recv()
* directly so it gets passed through the socket/signals layer on Win32.
*
* These functions are closely modelled on the standard socket BIO in OpenSSL;
* see sock_read() and sock_write() in OpenSSL's crypto/bio/bss_sock.c.
* XXX OpenSSL 1.0.1e considers many more errcodes than just EINTR as reasons
* to retry; do we need to adopt their logic for that?
*/
static bool my_bio_initialized = false;
static BIO_METHOD my_bio_methods;
static int
my_sock_read(BIO *h, char *buf, int size)
{
int res = 0;
if (buf != NULL)
{
res = secure_raw_read(((Port *)h->ptr), buf, size);
BIO_clear_retry_flags(h);
if (res <= 0)
{
/* If we were interrupted, tell caller to retry */
if (errno == EINTR || errno == EWOULDBLOCK || errno == EAGAIN)
{
BIO_set_retry_read(h);
}
}
}
return res;
}
static int
my_sock_write(BIO *h, const char *buf, int size)
{
int res = 0;
res = secure_raw_write(((Port *) h->ptr), buf, size);
BIO_clear_retry_flags(h);
if (res <= 0)
{
/* If we were interrupted, tell caller to retry */
if (errno == EINTR || errno == EWOULDBLOCK || errno == EAGAIN)
{
BIO_set_retry_write(h);
}
}
return res;
}
static BIO_METHOD *
my_BIO_s_socket(void)
{
if (!my_bio_initialized)
{
memcpy(&my_bio_methods, BIO_s_socket(), sizeof(BIO_METHOD));
my_bio_methods.bread = my_sock_read;
my_bio_methods.bwrite = my_sock_write;
my_bio_initialized = true;
}
return &my_bio_methods;
}
/* This should exactly match openssl's SSL_set_fd except for using my BIO */
static int
my_SSL_set_fd(Port *port, int fd)
{
int ret = 0;
BIO *bio = NULL;
bio = BIO_new(my_BIO_s_socket());
if (bio == NULL)
{
SSLerr(SSL_F_SSL_SET_FD, ERR_R_BUF_LIB);
goto err;
}
/* Use 'ptr' to store pointer to PGconn */
bio->ptr = port;
BIO_set_fd(bio, fd, BIO_NOCLOSE);
SSL_set_bio(port->ssl, bio, bio);
ret = 1;
err:
return ret;
}
/*
* Load precomputed DH parameters.
*
* To prevent "downgrade" attacks, we perform a number of checks
* to verify that the DBA-generated DH parameters file contains
* what we expect it to contain.
*/
static DH *
load_dh_file(int keylength)
{
FILE *fp;
char fnbuf[MAXPGPATH];
DH *dh = NULL;
int codes;
/* attempt to open file. It's not an error if it doesn't exist. */
snprintf(fnbuf, sizeof(fnbuf), "dh%d.pem", keylength);
if ((fp = fopen(fnbuf, "r")) == NULL)
return NULL;
/* flock(fileno(fp), LOCK_SH); */
dh = PEM_read_DHparams(fp, NULL, NULL, NULL);
/* flock(fileno(fp), LOCK_UN); */
fclose(fp);
/* is the prime the correct size? */
if (dh != NULL && 8 * DH_size(dh) < keylength)
{
elog(LOG, "DH errors (%s): %d bits expected, %d bits found",
fnbuf, keylength, 8 * DH_size(dh));
dh = NULL;
}
/* make sure the DH parameters are usable */
if (dh != NULL)
{
if (DH_check(dh, &codes) == 0)
{
elog(LOG, "DH_check error (%s): %s", fnbuf, SSLerrmessage());
return NULL;
}
if (codes & DH_CHECK_P_NOT_PRIME)
{
elog(LOG, "DH error (%s): p is not prime", fnbuf);
return NULL;
}
if ((codes & DH_NOT_SUITABLE_GENERATOR) &&
(codes & DH_CHECK_P_NOT_SAFE_PRIME))
{
elog(LOG,
"DH error (%s): neither suitable generator or safe prime",
fnbuf);
return NULL;
}
}
return dh;
}
/*
* Load hardcoded DH parameters.
*
* To prevent problems if the DH parameters files don't even
* exist, we can load DH parameters hardcoded into this file.
*/
static DH *
load_dh_buffer(const char *buffer, size_t len)
{
BIO *bio;
DH *dh = NULL;
bio = BIO_new_mem_buf((char *) buffer, len);
if (bio == NULL)
return NULL;
dh = PEM_read_bio_DHparams(bio, NULL, NULL, NULL);
if (dh == NULL)
ereport(DEBUG2,
(errmsg_internal("DH load buffer: %s",
SSLerrmessage())));
BIO_free(bio);
return dh;
}
/*
* Generate an ephemeral DH key. Because this can take a long
* time to compute, we can use precomputed parameters of the
* common key sizes.
*
* Since few sites will bother to precompute these parameter
* files, we also provide a fallback to the parameters provided
* by the OpenSSL project.
*
* These values can be static (once loaded or computed) since
* the OpenSSL library can efficiently generate random keys from
* the information provided.
*/
static DH *
tmp_dh_cb(SSL *s, int is_export, int keylength)
{
DH *r = NULL;
static DH *dh = NULL;
static DH *dh512 = NULL;
static DH *dh1024 = NULL;
static DH *dh2048 = NULL;
static DH *dh4096 = NULL;
switch (keylength)
{
case 512:
if (dh512 == NULL)
dh512 = load_dh_file(keylength);
if (dh512 == NULL)
dh512 = load_dh_buffer(file_dh512, sizeof file_dh512);
r = dh512;
break;
case 1024:
if (dh1024 == NULL)
dh1024 = load_dh_file(keylength);
if (dh1024 == NULL)
dh1024 = load_dh_buffer(file_dh1024, sizeof file_dh1024);
r = dh1024;
break;
case 2048:
if (dh2048 == NULL)
dh2048 = load_dh_file(keylength);
if (dh2048 == NULL)
dh2048 = load_dh_buffer(file_dh2048, sizeof file_dh2048);
r = dh2048;
break;
case 4096:
if (dh4096 == NULL)
dh4096 = load_dh_file(keylength);
if (dh4096 == NULL)
dh4096 = load_dh_buffer(file_dh4096, sizeof file_dh4096);
r = dh4096;
break;
default:
if (dh == NULL)
dh = load_dh_file(keylength);
r = dh;
}
/* this may take a long time, but it may be necessary... */
if (r == NULL || 8 * DH_size(r) < keylength)
{
ereport(DEBUG2,
(errmsg_internal("DH: generating parameters (%d bits)",
keylength)));
r = DH_generate_parameters(keylength, DH_GENERATOR_2, NULL, NULL);
}
return r;
}
/*
* Certificate verification callback
*
* This callback allows us to log intermediate problems during
* verification, but for now we'll see if the final error message
* contains enough information.
*
* This callback also allows us to override the default acceptance
* criteria (e.g., accepting self-signed or expired certs), but
* for now we accept the default checks.
*/
static int
verify_cb(int ok, X509_STORE_CTX *ctx)
{
return ok;
}
/*
* This callback is used to copy SSL information messages
* into the PostgreSQL log.
*/
static void
info_cb(const SSL *ssl, int type, int args)
{
switch (type)
{
case SSL_CB_HANDSHAKE_START:
ereport(DEBUG4,
(errmsg_internal("SSL: handshake start")));
break;
case SSL_CB_HANDSHAKE_DONE:
ereport(DEBUG4,
(errmsg_internal("SSL: handshake done")));
break;
case SSL_CB_ACCEPT_LOOP:
ereport(DEBUG4,
(errmsg_internal("SSL: accept loop")));
break;
case SSL_CB_ACCEPT_EXIT:
ereport(DEBUG4,
(errmsg_internal("SSL: accept exit (%d)", args)));
break;
case SSL_CB_CONNECT_LOOP:
ereport(DEBUG4,
(errmsg_internal("SSL: connect loop")));
break;
case SSL_CB_CONNECT_EXIT:
ereport(DEBUG4,
(errmsg_internal("SSL: connect exit (%d)", args)));
break;
case SSL_CB_READ_ALERT:
ereport(DEBUG4,
(errmsg_internal("SSL: read alert (0x%04x)", args)));
break;
case SSL_CB_WRITE_ALERT:
ereport(DEBUG4,
(errmsg_internal("SSL: write alert (0x%04x)", args)));
break;
}
}
static void
initialize_ecdh(void)
{
#if (OPENSSL_VERSION_NUMBER >= 0x0090800fL) && !defined(OPENSSL_NO_ECDH)
EC_KEY *ecdh;
int nid;
nid = OBJ_sn2nid(SSLECDHCurve);
if (!nid)
ereport(FATAL,
(errmsg("ECDH: unrecognized curve name: %s", SSLECDHCurve)));
ecdh = EC_KEY_new_by_curve_name(nid);
if (!ecdh)
ereport(FATAL,
(errmsg("ECDH: could not create key")));
SSL_CTX_set_options(SSL_context, SSL_OP_SINGLE_ECDH_USE);
SSL_CTX_set_tmp_ecdh(SSL_context, ecdh);
EC_KEY_free(ecdh);
#endif
}
/*
* Obtain reason string for last SSL error
*
* Some caution is needed here since ERR_reason_error_string will
* return NULL if it doesn't recognize the error code. We don't
* want to return NULL ever.
*/
static const char *
SSLerrmessage(void)
{
unsigned long errcode;
const char *errreason;
static char errbuf[32];
errcode = ERR_get_error();
if (errcode == 0)
return _("no SSL error reported");
errreason = ERR_reason_error_string(errcode);
if (errreason != NULL)
return errreason;
snprintf(errbuf, sizeof(errbuf), _("SSL error code %lu"), errcode);
return errbuf;
}