Dirkjan Bussink [Fri, 10 Jan 2014 15:12:40 +0000 (16:12 +0100)]
SSL: ssl_session_tickets directive.
This adds support so it's possible to explicitly disable SSL Session
Tickets. In order to have good Forward Secrecy support either the
session ticket key has to be reloaded by using nginx' binary upgrade
process or using an external key file and reloading the configuration.
This directive adds another possibility to have good support by
disabling session tickets altogether.
If session tickets are enabled and the process lives for a long a time,
an attacker can grab the session ticket from the process and use that to
decrypt any traffic that occured during the entire lifetime of the
process.
Maxim Dounin [Fri, 3 Jan 2014 23:32:22 +0000 (03:32 +0400)]
Fixed "zero size buf in output" alerts.
If a request had an empty request body (with Content-Length: 0), and there
were preread data available (e.g., due to a pipelined request in the buffer),
the "zero size buf in output" alert might be logged while proxying the
request to an upstream.
Similar alerts appeared with client_body_in_file_only if a request had an
empty request body.
Maxim Dounin [Fri, 3 Jan 2014 23:32:10 +0000 (03:32 +0400)]
Upstream: Cache-Control preferred over Expires.
Not really a strict check (as X-Accel-Expires might be ignored or
contain invalid value), but quite simple to implement and better
than what we have now.
Maxim Dounin [Fri, 3 Jan 2014 23:31:58 +0000 (03:31 +0400)]
Allowed up to two EBUSY errors from sendfile().
Fallback to synchronous sendfile() now only done on 3rd EBUSY without
any progress in a row. Not falling back is believed to be better
in case of occasional EBUSY, though protection is still needed to
make sure there will be no infinite loop.
Ruslan Ermilov [Fri, 27 Dec 2013 15:40:04 +0000 (19:40 +0400)]
Fixed setting of content type in some cases.
This fixes content type set in stub_status and autoindex responses
to be usable in content type checks made by filter modules, such
as charset and sub filters.
Processing events from upstream connection can result in sending queued frames
from other streams. In this case such streams were not added to handling queue
and properly handled.
A global per connection flag was replaced by a per stream flag that indicates
currently sending stream while all other streams can be added to handling
queue.
Upstream: simplified peer selection loop in the "ip_hash" module.
Conditions for skipping ineligible peers are rewritten to make adding of new
conditions simpler and be in line with the "round_robin" and "least_conn"
modules. No functional changes.
SPDY: drop the "delayed" flag when finalizing connection.
This flag in SPDY fake write events serves the same purposes as the "ready"
flag in real events, and it must be dropped if request needs to be handled.
Otherwise, it can prevent the request from finalization if ngx_http_writer()
was set, which results in a connection leak.
Evan Miller [Wed, 4 Dec 2013 22:45:44 +0000 (14:45 -0800)]
Contrib: add vim scripts to contrib/ directory.
I've been maintaining these scripts independently for a while now, even though
I'm no longer active in the community. Seems to me that contrib/ is a good
long-term home for the scripts.
Maxim Dounin [Fri, 29 Nov 2013 13:16:06 +0000 (17:16 +0400)]
SSL: fixed c->read->ready handling in ngx_ssl_recv().
If c->read->ready was reset, but later some data were read from a socket
buffer due to a call to ngx_ssl_recv(), the c->read->ready flag should
be restored if not all data were read from OpenSSL buffers (as kernel
won't notify us about the data anymore).
More details are available here:
http://mailman.nginx.org/pipermail/nginx/2013-November/041178.html
Maxim Dounin [Mon, 18 Nov 2013 16:48:22 +0000 (20:48 +0400)]
Upstream: cache revalidation with conditional requests.
The following new directives are introduced: proxy_cache_revalidate,
fastcgi_cache_revalidate, scgi_cache_revalidate, uwsgi_cache_revalidate.
Default is off. When set to on, they enable cache revalidation using
conditional requests with If-Modified-Since for expired cache items.
As of now, no attempts are made to merge headers given in a 304 response
during cache revalidation with headers previously stored in a cache item.
Headers in a 304 response are only used to calculate new validity time
of a cache item.
SPDY: fixed request hang with the auth request module.
We should just call post_handler() when subrequest wants to read body, like
it happens for HTTP since rev. f458156fd46a. An attempt to init request body
for subrequests results in hang if the body was not already read.
Maxim Dounin [Thu, 31 Oct 2013 00:00:37 +0000 (04:00 +0400)]
Core: handling of getsockopt(TCP_DEFER_ACCEPT) failures.
Recent Linux versions started to return EOPNOTSUPP to getsockopt() calls
on unix sockets, resulting in log pollution on binary upgrade. Such errors
are silently ignored now.
Maxim Dounin [Fri, 18 Oct 2013 14:13:49 +0000 (18:13 +0400)]
Fixed "satisfy any" if 403 is returned after 401 (ticket #285).
The 403 (Forbidden) should not overwrite 401 (Unauthorized) as the
latter should be returned with the WWW-Authenticate header to request
authentication by a client.
The problem could be triggered with 3rd party modules and the "deny"
directive, or with auth_basic and auth_request which returns 403
(in 1.5.4+).
Maxim Dounin [Fri, 18 Oct 2013 14:13:44 +0000 (18:13 +0400)]
Headers filter: empty Cache-Control is no longer added.
Much like with other headers, "add_header Cache-Control $value;" no longer
results in anything added to response headers if $value evaluates to an
empty string.
Piotr Sikora [Fri, 11 Oct 2013 23:05:24 +0000 (16:05 -0700)]
SSL: added ability to set keys used for Session Tickets (RFC5077).
In order to support key rollover, ssl_session_ticket_key can be defined
multiple times. The first key will be used to issue and resume Session
Tickets, while the rest will be used only to resume them.
Please note that nginx supports Session Tickets even without explicit
configuration of the keys and this feature should be only used in setups
where SSL traffic is distributed across multiple nginx servers.
Signed-off-by: Piotr Sikora <piotr@cloudflare.com>
Maxim Dounin [Mon, 14 Oct 2013 09:59:35 +0000 (13:59 +0400)]
SSL: SSL_CTX_set_timeout() now always called.
The timeout set is used by OpenSSL as a hint for clients in TLS Session
Tickets. Previous code resulted in a default timeout (5m) used for TLS
Sessions Tickets if there was no session cache configured.
SPDY: ignore priority when queuing blocked frames.
With this change all such frames will be added in front of the output queue, and
will be sent first. It prevents HOL blocking when response with higher priority
is blocked by response with lower priority in the middle of the queue because
the order of their SYN_REPLY frames cannot be changed.
SPDY: fixed connection leak while waiting for request headers.
If an error occurs in a SPDY connection, the c->error flag is set on every fake
request connection, and its read or write event handler is called, in order to
finalize it. But while waiting for request headers, it was a no-op since the
read event handler had been set to ngx_http_empty_handler().
SPDY: fixed connection leak while waiting for request body.
If an error occurs in a SPDY connection, the c->error flag is set on every fake
request connection, and its read or write event handler is called, in order to
finalize it. But while waiting for a request body, it was a no-op since the
read event handler ngx_http_request_handler() calls r->read_event_handler that
had been set to ngx_http_block_reading().
Mail: added session close on smtp_greeting_delay violation.
A server MUST send greeting before other replies, while before this
change in case of smtp_greeting_delay violation the 220 greeting was
sent after several 503 replies to commands received before greeting,
resulting in protocol synchronization loss. Moreover, further commands
were accepted after the greeting.
While closing a connection isn't strictly RFC compliant (RFC 5321
requires servers to wait for a QUIT before closing a connection), it's
probably good enough for practial uses.
With previous code only part of u->buffer might be emptied in case
of special responses, resulting in partial responses seen by SSI set
in case of simple protocols, or spurious errors like "upstream sent
invalid chunked response" in case of complex ones.