Commit Graph

5162 Commits

Author SHA1 Message Date
Ruslan Ermilov
e49d933ebc Upstream: reliably detect connection failures with SSL peers. 2013-07-29 13:23:16 +04:00
Maxim Dounin
5274f023a2 Upstream: no last buffer on errors.
Previously, after sending a header we always sent a last buffer and
finalized a request with code 0, even in case of errors.  In some cases
this resulted in a loss of ability to detect the response wasn't complete
(e.g. if Content-Length was removed from a response by gzip filter).

This change tries to propogate to a client information that a response
isn't complete in such cases.  In particular, with this change we no longer
pretend a returned response is complete if we wasn't able to create
a temporary file.

If an error code suggests the error wasn't fatal, we flush buffered data
and disable keepalive, then finalize request normally.  This allows to to
propogate information about a problem to a client, while still sending all
the data we've got from an upstream.
2013-07-25 15:00:41 +04:00
Maxim Dounin
78aacc838b Upstream: request finalization rework.
No semantic changes expected, though some checks are done differently.
In particular, the r->cached flag is no longer explicitly checked.  Instead,
we relay on u->header_sent not being set if a response is sent from
a cache.
2013-07-25 15:00:29 +04:00
Maxim Dounin
86277254dd Upstream: NGX_HTTP_CLIENT_CLOSED_REQUEST no longer reset to 0.
The NGX_HTTP_CLIENT_CLOSED_REQUEST code is allowed to happen after we
started sending a response (much like NGX_HTTP_REQUEST_TIME_OUT), so there
is no need to reset response code to 0 in this case.
2013-07-25 15:00:25 +04:00
Maxim Dounin
960d0bfe34 Upstream: added check if a response is complete.
Checks were added to both buffered and unbuffered code paths to detect
and complain if a response is incomplete.  Appropriate error codes are
now passed to ngx_http_upstream_finalize_request().

With this change in unbuffered mode we now use u->length set to -1 as an
indicator that EOF is allowed per protocol and used to indicate response
end (much like its with p->length in buffered mode).  Proxy module was
changed to set u->length to 1 (instead of previously used -1) in case of
chunked transfer encoding used to comply with the above.
2013-07-25 15:00:12 +04:00
Maxim Dounin
416b922bd2 Upstream: u->length now defaults to -1 (API change).
That is, by default we assume that response end is signalled by
a connection close.  This seems to be better default, and in line
with u->pipe->length behaviour.

Memcached module was modified accordingly.
2013-07-25 14:58:11 +04:00
Maxim Dounin
187f3948ed Upstream: fixed store/cache of unfinished responses.
In case of upstream eof, only responses with u->pipe->length == -1
are now cached/stored.  This ensures that unfinished chunked responses
are not cached.

Note well - previously used checks for u->headers_in.content_length_n are
preserved.  This provides an additional level of protection if protol data
disagree with Content-Length header provided (e.g., a FastCGI response
is sent with wrong Content-Length, or an incomple SCGI or uwsgi response),
as well as protects from storing of responses to HEAD requests.  This should
be reconsidered if we'll consider caching of responses to HEAD requests.
2013-07-25 14:56:59 +04:00
Maxim Dounin
eafe44ff79 Upstream: replaced u->pipe->temp_file with p->temp_file.
While here, redundant parentheses removed.  No functional changes.
2013-07-25 14:56:49 +04:00
Maxim Dounin
e6122efbfe Upstream: NGX_ERROR after pipe errors. 2013-07-25 14:56:41 +04:00
Maxim Dounin
8536fb79ca Upstream: NGX_HTTP_GATEWAY_TIME_OUT after upstream timeouts.
There is no real difference from previously used 0 as NGX_HTTP_* will
become 0 in ngx_http_upstream_finalize_request(), but the change
preserves information about a timeout a bit longer.  Previous use of
ETIMEDOUT in one place was just wrong.

Note well that with cacheable responses there will be a difference
(code in ngx_http_upstream_finalize_request() will store the error
in cache), though this change doesn't touch cacheable case.
2013-07-25 14:56:20 +04:00
Maxim Dounin
d23dc7d427 Upstream: ngx_http_upstream_finalize_request(NGX_ERROR) on errors.
Previously, ngx_http_upstream_finalize_request(0) was used in most
cases after errors.  While with current code there is no difference,
use of NGX_ERROR allows to pass a bit more information into
ngx_http_upstream_finalize_request().
2013-07-25 14:56:13 +04:00
Maxim Dounin
76e1571749 Upstream: consistent error handling after u->input_filter_init().
In all cases ngx_http_upstream_finalize_request() with NGX_ERROR now used.
Previously used NGX_HTTP_INTERNAL_SERVER_ERROR in the subrequest in memory
case don't cause any harm, but inconsistent with other uses.
2013-07-25 14:56:00 +04:00
Maxim Dounin
9f925b8c77 Upstream: busy lock remnants removed. 2013-07-25 14:55:59 +04:00
Maxim Dounin
ec021eda55 Upstream: stale comments removed. 2013-06-13 19:52:31 +04:00
Maxim Dounin
84d2ecf87e Gzip: clearing of c->buffered if all data are flushed.
This allows to finalize unfinished responses while still sending as
much data as available.
2013-07-25 14:55:32 +04:00
Maxim Dounin
be6fbbb47e Event pipe: fixed writing cache header to a temp file.
With previous code the p->temp_file->offset wasn't adjusted if a temp
file was written by the code in ngx_event_pipe_write_to_downstream()
after an EOF, resulting in cache not being used with empty scgi and uwsgi
responses with Content-Length set to 0.

Fix it to call ngx_event_pipe_write_chain_to_temp_file() there instead
of calling ngx_write_chain_to_temp_file() directly.
2013-07-25 14:55:09 +04:00
Maxim Dounin
f52042498d Fixed ngx_http_test_reading() to finalize request properly.
Previous code called ngx_http_finalize_request() with rc = 0.  This is
ok if a response status was already set, but resulted in "000" being
logged if it wasn't.  In particular this happened with limit_req
if a connection was prematurely closed during limit_req delay.
2013-06-14 20:56:07 +04:00
Maxim Dounin
1936a67647 Sub filter: fixed matching after a partial match.
After a failed partial match we now check if there is another partial
match in previously matched substring to fix cases like "aab" in "aaab".

The ctx->saved string is now always sent if it's present on return
from the ngx_http_sub_parse() function (and reset accordingly).  This
allows to release parts of previously matched data.
2013-07-25 14:54:53 +04:00
Maxim Dounin
ce7a5a0537 Sub filter: fixed incomplete last buffer on partial match.
If a pattern was partially matched at a response end, partially matched
string wasn't send.  E.g., a response "fo" was truncated to an empty response
if partially mathed by a pattern "foo".
2013-07-25 14:54:48 +04:00
Maxim Dounin
68fab7c8c4 Sub filter: flush buffers handling. 2013-07-25 14:54:47 +04:00
Maxim Dounin
8d568c7ae0 Sub filter: switched to ngx_chain_get_free_buf().
No functional changes.
2013-07-25 14:54:45 +04:00
Maxim Dounin
3961ef2318 Sub filter: stale comments removed. 2013-07-25 14:54:43 +04:00
Ruslan Ermilov
02a077b827 On DragonFlyBSD, TCP_KEEPIDLE and TCP_KEEPINTVL are in msecs.
Based on a patch by Sepherosa Ziehau.
2013-07-25 12:46:03 +04:00
Ruslan Ermilov
690e2b33aa Style: reuse one int variable in ngx_configure_listening_sockets().
No functional changes.
2013-07-25 12:46:02 +04:00
Valentin Bartenev
32e167e211 SPDY: fixed segfault with "client_body_in_file_only" enabled.
It is possible to send FLAG_FIN in additional empty data frame, even if it is
known from the content-length header that request body is empty.  And Firefox
actually behaves like this (see ticket #357).

To simplify code we sacrificed our microoptimization that did not work right
due to missing check in the ngx_http_spdy_state_data() function for rb->buf
set to NULL.
2013-07-24 22:24:25 +04:00
Maxim Dounin
809d05769b Xslt: exsltRegisterAll() moved to preconfiguration.
The exsltRegisterAll() needs to be called before XSLT stylesheets
are compiled, else stylesheet compilation hooks will not work.  This
change fixes EXSLT Functions extension.
2013-07-19 15:59:50 +04:00
Valentin Bartenev
615d558843 Events: honor NGX_USE_GREEDY_EVENT when kqueue support is enabled.
Currently this flag is needed for epoll and rtsig, and though these methods
usually present on different platforms than kqueue, nginx can be compiled to
support all of them.
2013-07-13 03:24:30 +04:00
Maxim Dounin
d10251030a Style. 2013-07-11 20:38:27 +04:00
Maxim Dounin
4ef36b38a1 Configure: perl Makefile rebuild after configure.
The $NGX_AUTO_CONFIG_H added to perl module Makefile dependencies to
make sure it's always rebuild after a configure.  It is needed as we
expand various variables used for Makefile generation during configure
(in particular, nginx version).
2013-07-11 20:34:02 +04:00
Vladimir Homutov
d79c8abcaa Core: fixed possible use of an uninitialized variable.
The call to ngx_sock_ntop() in ngx_connection_local_sockaddr() might be
performed with the uninitialized "len" variable.  The fix is to initialize
variable to the size of corresponding socket address type.

The issue was introduced in commit 05ba5bce31e0.
2013-07-11 19:50:19 +04:00
Vladimir Homutov
af18946d76 Core: extended ngx_sock_ntop() with socklen parameter.
On Linux, sockaddr length is required to process unix socket addresses properly
due to unnamed sockets (which don't have sun_path set at all) and abstract
namespace sockets.
2013-07-11 16:07:25 +04:00
Gleb Smirnoff
050f74b751 Make macros safe. 2013-07-05 11:42:25 +04:00
Ruslan Ermilov
523191ec89 Upstream: updated list of ngx_event_connect_peer() return values.
ngx_http_upstream_get_keepalive_peer() may return NGX_DONE to
indicate that the cached keepalive connection is reused.
2013-07-03 12:04:13 +04:00
Vladimir Homutov
dd3e13eef0 Core: consolidated log-related code.
The stderr redirection code is moved to ngx_log_redirect_stderr().
The opening of the default log code is moved to ngx_log_open_default().
2013-06-28 17:24:54 +04:00
Maxim Dounin
f41c91511d Version bump. 2013-07-02 20:05:49 +04:00
Maxim Dounin
dfb8b2b7bb release-1.5.2 tag 2013-07-02 16:28:51 +04:00
Maxim Dounin
652f6c227b nginx-1.5.2-RELEASE 2013-07-02 16:28:50 +04:00
Maxim Dounin
88fc0f793e Fixed ngx_http_parse_chunked() minimal length calculation.
Minimal data length we expect for further calls was calculated incorrectly
if parsing stopped right after parsing chunk size.  This might in theory
affect clients and/or backends using LF instead of CRLF.

Patch by Dmitry Popov.
2013-06-28 13:55:05 +04:00
Maxim Dounin
982f4de9f0 Fixed build on Linux with x32 ABI.
On Linux x32 inclusion of sys/sysctl.h produces an error.  As sysctl() is
only used by rtsig event method code, which is legacy and not compiled
in by default on modern linuxes, the sys/sysctl.h file now only included
if rtsig support is enabled.

Based on patch by Serguei I. Ivantsov.
2013-06-26 15:47:27 +04:00
Vladimir Homutov
11ff177a9f Core: support several "error_log" directives.
When several "error_log" directives are specified in the same configuration
block, logs are written to all files with a matching log level.

All logs are stored in the singly-linked list that is sorted by log level in
the descending order.

Specific debug levels (NGX_LOG_DEBUG_HTTP,EVENT, etc.) are not supported
if several "error_log" directives are specified.  In this case all logs
will use debug level that has largest absolute value.
2013-06-20 20:47:39 +04:00
Ruslan Ermilov
02cc52554f Simplified ngx_list_create(). 2013-06-19 08:55:08 +04:00
Tatsuhiko Kubo
c974bd7936 Valgrind: another complaint about uninitialized bytes. 2013-06-12 00:41:24 +09:00
Valentin Bartenev
420e41ef09 SPDY: fixed code style, no functional changes. 2013-06-13 16:28:42 +04:00
Vladimir Homutov
aac7f7f464 Core: moved initialization of log level.
The cycle->new_log->log_level should only be initialized by ngx_init_cycle()
if no error logs were found in the configuration.  This move allows to get rid
of extra initialization in ngx_error_log().
2013-06-13 14:50:10 +04:00
Jim Radford
de2ec2c650 SPDY: pass through the full status when available.
Avoid stripping the status text when proxying for compatibility with http.
2013-06-07 13:16:00 -07:00
Maxim Dounin
ddece08218 Perl: fixed r->header_in("Cookie") (ticket #351).
It was broken by X-Forwarded-For related changes in f7fe817c92a2 (1.3.14)
as hh->offset is no longer 0 for Cookie.
2013-06-10 14:35:00 +04:00
Maxim Dounin
dc5c6928fb Fixed debug logging in ngx_http_parse_complex_uri().
The *u previously logged isn't yet initialized at this point, and
Valgrind complains.
2013-06-05 19:45:08 +04:00
Maxim Dounin
2b356db155 Mail: fixed possible uninitialized memory access.
Found by Valgrind.
2013-06-05 19:44:23 +04:00
Maxim Dounin
40c4e7b73c Valgrind: supressed complaints about uninitialized bytes.
Valgrind complains if we pass uninitialized memory to a syscall:

==36492== Syscall param sendmsg(msg.msg_iov[0]) points to uninitialised byte(s)
==36492==    at 0x6B5E6A: sendmsg (in /usr/lib/system/libsystem_kernel.dylib)
==36492==    by 0x10004288E: ngx_signal_worker_processes (ngx_process_cycle.c:527)
==36492==    by 0x1000417A7: ngx_master_process_cycle (ngx_process_cycle.c:203)
==36492==    by 0x100001F10: main (nginx.c:410)
==36492==  Address 0x7fff5fbff71c is on thread 1's stack

Even initialization of all members of the structure passed isn't enough, as
there is padding which still remains uninitialized and results in Valgrind
complaint.  Note there is no real problem here as data from uninitialized
memory isn't used.
2013-06-05 19:44:22 +04:00
Maxim Dounin
6b95320cf7 Valgrind: sigaction() failure now ignored.
Valgrind intercepts SIGUSR2 in some cases, and nginx might not be able to
start due to sigaction() failure.  If compiled with NGX_VALGRIND defined,
we now ignore the failure of sigaction().
2013-06-05 19:44:20 +04:00