| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
|
|
Linux with EPOLLEXCLUSIVE usually notifies only the process which was first
to add the listening socket to the epoll instance. As a result most of the
connections are handled by the first worker process (ticket #2285). To fix
this, we re-add the socket periodically, so other workers will get a chance
to accept connections.
|
|
Starting with FreeBSD 11, there is no need to use AIO operations to preload
data into cache for sendfile(SF_NODISKIO) to work. Instead, sendfile()
handles non-blocking loading data from disk by itself. It still can, however,
return EBUSY if a page is already being loaded (for example, by a different
process). If this happens, we now post an event for the next event loop
iteration, so sendfile() is retried "after a short period", as manpage
recommends.
The limit of the number of EBUSY tolerated without any progress is preserved,
but now it does not result in an alert, since on an idle system event loop
iteration might be very short and EBUSY can happen many times in a row.
Instead, SF_NODISKIO is simply disabled for one call once the limit is
reached.
With this change, sendfile(SF_NODISKIO) is now used automatically as long as
sendfile() is enabled, and no longer requires "aio on;".
|
|
This makes it possible to avoid looping for a long time while working
with a fast enough peer when data are added to the socket buffer faster
than we are able to read and process them (ticket #1431). This is
basically what we already do on FreeBSD with kqueue, where information
about the number of bytes in the socket buffer is returned by
the kevent() call.
With other event methods rev->available is now set to -1 when the socket
is ready for reading. Later in ngx_recv() and ngx_recv_chain(), if
full buffer is received, real number of bytes in the socket buffer is
retrieved using ioctl(FIONREAD). Reading more than this number of bytes
ensures that even with edge-triggered event methods the event will be
triggered again, so it is safe to stop processing of the socket and
switch to other connections.
Using ioctl(FIONREAD) only after reading a full buffer is an optimization.
With this approach we only call ioctl(FIONREAD) when there are at least
two recv()/readv() calls.
|
|
The directive allows to drop binding between a client and existing UDP stream
session after receiving a specified number of packets. First packet from the
same client address and port will start a new session. Old session continues
to exist and will terminate at moment defined by configuration: either after
receiving the expected number of responses, or after timeout, as specified by
the "proxy_responses" and/or "proxy_timeout" directives.
By default, proxy_requests is zero (disabled).
|
|
|
|
Previously, only one client packet could be processed in a udp stream session
even though multiple response packets were supported. Now multiple packets
coming from the same client address and port are delivered to the same stream
session.
If it's required to maintain a single stream of data, nginx should be
configured in a way that all packets from a client are delivered to the same
worker. On Linux and DragonFly BSD the "reuseport" parameter should be
specified for this. Other systems do not currently provide appropriate
mechanisms. For these systems a single stream of udp packets is only
guaranteed in single-worker configurations.
The proxy_response directive now specifies how many packets are expected in
response to a single client packet.
|
|
|
|
The "fd" field should be after 3 pointers for ngx_event_ident() to use it.
This was broken by ccad84a174e0. While it does not seem to be currently used
for aio-related events, it should be a good idea to preserve the correct
layout nevertheless.
|
|
With this change it is now possible to load modules compiled without
the "--with-file-aio" configure option into nginx binary compiled with it,
and vice versa, assuming both use the "--with-compat" option.
|
|
|
|
This flag appeared in Linux 4.5 and is useful for avoiding thundering herd
problem.
The current Linux kernel implementation walks the list of exclusive waiters,
and queues an event to each epfd, until it finds the first waiter that has
threads blocked on it via epoll_wait().
|
|
Also, removed practically unused flag accept_context_updated from
ngx_connection_t.
|
|
When it's known that the kernel supports EPOLLRDHUP, there is no need in
additional recv() call to get EOF or error when the flag is absent in the
event generated by the kernel. A special runtime test is done at startup
to detect if EPOLLRDHUP is actually supported by the kernel because
epoll_ctl() silently ignores unknown flags.
With this knowledge it's now possible to drop the "ready" flag for partial
read. Previously, the "ready" flag was kept until the recv() returned EOF
or error. In particular, this change allows the lingering close heuristics
(which relies on the "ready" flag state) to actually work on Linux, and not
wait for more data in most cases.
The "available" flag is now used in the read event with the semantics similar
to the corresponding counter in kqueue.
|
|
|
|
Fixes various aspects of --test-build-devpoll, --test-build-eventport, and
--test-build-epoll.
In particular, if --test-build-devpoll was used on Linux, then "devpoll"
event method would be preferred over "epoll". Also, wrong definitions of
event macros were chosen.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
No functional changes.
|
|
This reduces layering violation and simplifies the logic of AIO preread, since
it's now triggered by the send chain function itself without falling back to
the copy filter. The context of AIO operation is now stored per file buffer,
which makes it possible to properly handle cases when multiple buffers come
from different locations, each with its own configuration.
|
|
|
|
In theory, this can provide a bit better distribution of latencies.
Also it simplifies the code, since ngx_queue_t is now used instead
of custom implementation.
|
|
It's mostly dead code. And the idea of thread support for this task has
been deprecated.
|
|
Since Linux 2.6.17, epoll is able to report about peer half-closed connection
using special EPOLLRDHUP flag on a read event.
|
|
They are not used since 708f8bb772ec (pre 0.0.1).
|
|
And corresponding variable $connections_waiting was added.
Previously, waiting connections were counted as the difference between
active connections and the sum of reading and writing connections.
That made it impossible to count more than one request in one connection
as reading or writing (as is the case for SPDY).
Also, we no longer count connections in handshake state as waiting.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
*) Feature: nginx now logs the subrequest information to the error log.
*) Feature: the "proxy_next_upstream", "fastcgi_next_upstream", and
"memcached_next_upstream" directives support the "off" parameter.
*) Feature: the "debug_connection" directive supports the CIDR address
form.
*) Bugfix: if a response of proxied server or FastCGI server was
converted from UTF-8 or back, then it may be transferred incomplete.
*) Bugfix: the $upstream_response_time variable had the time of the
first request to a backend only.
*) Bugfix: nginx could not be built on amd64 platform; the bug had
appeared in 0.3.53.
|
|
*) Change: the "variables_hash_max_size" and
"variables_hash_bucket_size" directives.
*) Feature: the $body_bytes_sent variable can be used not only in the
"log_format" directive.
*) Feature: the $ssl_protocol and $ssl_cipher variables.
*) Feature: the cache line size detection for widespread CPUs at start
time.
*) Feature: now the "accept_mutex" directive is supported using
fcntl(2) on platforms different from i386, amd64, sparc64, and ppc.
*) Feature: the "lock_file" directive and the --with-lock-path=PATH
autoconfiguration directive.
*) Bugfix: if the HTTPS protocol was used in the "proxy_pass" directive
then the requests with the body was not transferred.
|
|
*) Security: if nginx was built with the ngx_http_realip_module and the
"satisfy_any on" directive was used, then access and authorization
directives did not work. The ngx_http_realip_module was not built
and is not built by default.
*) Change: the "$time_gmt" variable name was changed to "$time_local".
*) Change: the "proxy_header_buffer_size" and
"fastcgi_header_buffer_size" directives was renamed to the
"proxy_buffer_size" and "fastcgi_buffer_size" directives.
*) Feature: the ngx_http_memcached_module.
*) Feature: the "proxy_buffering" directive.
*) Bugfix: the changes in accept mutex handling when the "rtsig" method
was used; the bug had appeared in 0.3.0.
*) Bugfix: if the client sent the "Transfer-Encoding: chunked" header
line, then nginx returns the 411 error.
*) Bugfix: if the "auth_basic" directive was inherited from the http
level, then the realm in the "WWW-Authenticate" header line was
without the "Basic realm" text.
*) Bugfix: if the "combined" format was explicitly specified in the
"access_log" directive, then the empty lines was written to the log;
the bug had appeared in 0.3.8.
*) Bugfix: nginx did not run on the sparc platform under any OS except
Solaris.
*) Bugfix: now it is not necessary to place space between the quoted
string and closing bracket in the "if" directive.
|
|
*) Security: nginx now checks URI got from a backend in
"X-Accel-Redirect" header line or in SSI file for the "/../" paths
and zeroes.
*) Change: nginx now does not treat the empty user name in the
"Authorization" header line as valid one.
*) Feature: the "ssl_session_timeout" directives of the
ngx_http_ssl_module and ngx_imap_ssl_module.
*) Feature: the "auth_http_header" directive of the
ngx_imap_auth_http_module.
*) Feature: the "add_header" directive.
*) Feature: the ngx_http_realip_module.
*) Feature: the new variables to use in the "log_format" directive:
$bytes_sent, $apache_bytes_sent, $status, $time_gmt, $uri,
$request_time, $request_length, $upstream_status,
$upstream_response_time, $gzip_ratio, $uid_got, $uid_set,
$connection, $pipe, and $msec. The parameters in the "%name" form
will be canceled soon.
*) Change: now the false variable values in the "if" directive are the
empty string "" and string starting with "0".
*) Bugfix: while using proxied or FastCGI-server nginx may leave
connections and temporary files with client requests in open state.
*) Bugfix: the worker processes did not flush the buffered logs on
graceful exit.
*) Bugfix: if the request URI was changes by the "rewrite" directive
and the request was proxied in location given by regular expression,
then the incorrect request was transferred to backend; the bug had
appeared in 0.2.6.
*) Bugfix: the "expires" directive did not remove the previous
"Expires" header.
*) Bugfix: nginx may stop to accept requests if the "rtsig" method and
several worker processes were used.
*) Bugfix: the "\"" and "\'" escape symbols were incorrectly handled in
SSI commands.
*) Bugfix: if the response was ended just after the SSI command and
gzipping was used, then the response did not transferred complete or
did not transferred at all.
|
|
*) Change: the "bl" and "af" parameters of the "listen" directive was
renamed to the "backlog" and "accept_filter".
*) Feature: the "rcvbuf" and "sndbuf" parameters of the "listen"
directive.
*) Change: the "$msec" log parameter does not require now the
additional the gettimeofday() system call.
*) Feature: the -t switch now tests the "listen" directives.
*) Bugfix: if the invalid address was specified in the "listen"
directive, then after the -HUP signal nginx left an open socket in
the CLOSED state.
*) Bugfix: the mime type may be incorrectly set to default value for
index file with variable in the name; the bug had appeared in 0.3.0.
*) Feature: the "timer_resolution" directive.
*) Feature: the millisecond "$upstream_response_time" log parameter.
*) Bugfix: a temporary file with client request body now is removed
just after the response header was transferred to a client.
*) Bugfix: OpenSSL 0.9.6 compatibility.
*) Bugfix: the SSL certificate and key file paths could not be relative.
*) Bugfix: the "ssl_prefer_server_ciphers" directive did not work in
the ngx_imap_ssl_module.
*) Bugfix: the "ssl_protocols" directive allowed to specify the single
protocol only.
|
