Handle buggy situations with SSL_ERROR_SYSCALL

See "BUGS" section at:

https://www.openssl.org/docs/man1.1.1/man3/SSL_get_error.html

It is mentioned there that when TLS status equals SSL_ERROR_SYSCALL
AND errno == 0 it means that underlying transport layer returned EOF
prematurely.  However, we are managing the transport ourselves, so we
should just resume reading from the TCP socket.

It seems that this case has been handled properly on modern versions
of OpenSSL. That being said, the situation goes in line with the
manual: it is briefly mentioned there that SSL_ERROR_SYSCALL might be
returned not only in a case of low-level errors (like system call
failures).

diff --git a/lib/isc/netmgr/tlsstream.c b/lib/isc/netmgr/tlsstream.c
index 012e9fb1f60b6d9b3838cb8e8d2131fa470063d6..b535e0564a53319838a29737671cbd24e95d7c44 100644 (file)
--- a/lib/isc/netmgr/tlsstream.c
+++ b/lib/isc/netmgr/tlsstream.c
@@ -9,6 +9,7 @@
  * information regarding copyright ownership.
  */
 
+#include <errno.h>
 #include <libgen.h>
 #include <unistd.h>
 #include <uv.h>
@@ -329,6 +330,7 @@ tls_do_bio(isc_nmsocket_t *sock, isc_region_t *received_data,
        int rv = 0;
        bool sent_shutdown = false, received_shutdown = false;
        size_t len = 0;
+       int saved_errno = 0;
 
        REQUIRE(VALID_NMSOCK(sock));
        REQUIRE(sock->tid == isc_nm_tid());
@@ -359,6 +361,7 @@ tls_do_bio(isc_nmsocket_t *sock, isc_region_t *received_data,
                                          received_data->length, &len);
                        if (rv <= 0 || len != received_data->length) {
                                result = ISC_R_TLSERROR;
+                               saved_errno = errno;
                                goto error;
                        }
 
@@ -410,6 +413,28 @@ tls_do_bio(isc_nmsocket_t *sock, isc_region_t *received_data,
                }
        }
        tls_status = SSL_get_error(sock->tlsstream.tls, rv);
+       saved_errno = errno;
+
+       /* See "BUGS" section at:
+        * https://www.openssl.org/docs/man1.1.1/man3/SSL_get_error.html
+        *
+        * It is mentioned there that when TLS status equals
+        * SSL_ERROR_SYSCALL AND errno == 0 it means that underlying
+        * transport layer returned EOF prematurely.  However, we are
+        * managing the transport ourselves, so we should just resume
+        * reading from the TCP socket.
+        *
+        * It seems that this case has been handled properly on modern
+        * versions of OpenSSL. That being said, the situation goes in
+        * line with the manual: it is briefly mentioned there that
+        * SSL_ERROR_SYSCALL might be returned not only in a case of
+        * low-level errors (like system call failures).
+        */
+       if (tls_status == SSL_ERROR_SYSCALL && saved_errno == 0 &&
+           received_data == NULL && send_data == NULL && finish == false)
+       {
+               tls_status = SSL_ERROR_WANT_READ;
+       }
 
        pending = tls_process_outgoing(sock, finish, send_data);
        if (pending > 0) {
@@ -451,8 +476,12 @@ tls_do_bio(isc_nmsocket_t *sock, isc_region_t *received_data,
 
 error:
        isc_log_write(isc_lctx, ISC_LOGCATEGORY_GENERAL, ISC_LOGMODULE_NETMGR,
-                     ISC_LOG_ERROR, "SSL error in BIO: %d %s", tls_status,
-                     isc_result_totext(result));
+                     ISC_LOG_NOTICE,
+                     "SSL error in BIO: %d %s (errno: %d). Arguments: "
+                     "received_data: %p, "
+                     "send_data: %p, finish: %s",
+                     tls_status, isc_result_totext(result), saved_errno,
+                     received_data, send_data, finish ? "true" : "false");
        tls_failed_read_cb(sock, result);
 }