Commit graph

1631 commits

Author SHA1 Message Date
Andy Green
4019aab8da ah http1.1 deal with pipelined headers properly
Connections must hold an ah for the whole time they are
processing one header set, even if eg, the headers are
fragmented and it involves network roundtrip times.

However on http1.1 / keepalive, it must drop the ah when
there are no more header sets to deal with, and reacquire
the ah later when more data appears.  It's because the
time between header sets / http1.1 requests is unbounded
and the ah would be tied up forever.

But in the case that we got pipelined http1.1 requests,
even partial already buffered, we must keep the ah,
resetting it instead of dropping it.  Because we store
the rx data conveniently in a per-tsi buffer since it only
does one thing at a time per thread, we cannot go back to
the event loop to await a new ah inside one service action.

But no problem since we definitely already have an ah,
let's just reuse it at http completion time if more rx is
already buffered.

NB: attack.sh makes request with echo | nc, this
accidentally sends a trailing '\n' from the echo showing
this problem.  With this patch attack.sh can complete well.

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-30 11:43:10 +08:00
Andy Green
1b2c9a23e1 clean pre 1.7
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 23:17:43 +08:00
Andy Green
ada3531aff coverity 157401 dont test SHUTDOWN state in mode switch
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 23:17:40 +08:00
Andy Green
e6dbaa236a coverity 157402 explicitly disallow handling invalid sockfd
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 15:40:08 +08:00
Andy Green
b39a151c9c coverity 157404 get rid of needless NULL check
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 15:30:05 +08:00
Andy Green
4bcbfe1ad4 restrict SO_REUSEPORT to multithreading only so no unexpected change otherwise
SO_REUSEPORT means you don't get any error any more if another
server instance is already running, this will be quite unexpected
for singlethreaded users.

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 09:20:00 +08:00
Andy Green
6dd7e86f23 http1.1 keepalive drop ah betweentimes
https://github.com/warmcat/libwebsockets/issues/404

Tested with

wget -O- http://localhost:7681/test.html http://localhost:7681/test.html http://localhost:7681/test.html http://localhost:7681/test.html http://localhost:7681/test.html

and confirm no connection processing during that on server side

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 09:06:22 +08:00
Andy Green
9c444d2709 win dont redefine _WINSOCK_DEPRECATED_NO_WARNINGS
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-29 01:26:31 +08:00
Danomi Czaski
4e9c7f3504 lextable add x-real-ip
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-28 09:40:53 +08:00
Andy Green
8d5351a0c4 reduce debug logging mark socket dead when pollert
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-27 08:50:31 +08:00
Andy Green
ba119e9057 lws_adopt_socket
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 21:40:32 +08:00
Andy Green
8c1f6026a7 multithread stability
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
Andy Green
9666e1d438 fuzzer handle junk after upgrade header
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-21 10:59:07 +08:00
Andy Green
5a0e7866d3 unify bounds checking in parser
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-21 10:57:39 +08:00
Andy Green
cb17ad8740 win add dummy LWS_WARN_UNUSED_RESULT
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-21 10:56:20 +08:00
Andy Green
5086597878 mbed align with pt changes
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-20 17:35:18 +08:00
Andy Green
e99a83cb96 introduce LWS_WARN_UNUSED_RESULT
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-20 16:56:06 +08:00
Andy Green
af607da082 header fragment reject empty early
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-20 09:44:04 +08:00
Andy Green
f38ad33ddf tolower remove
https://github.com/warmcat/libwebsockets/issues/403

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-20 07:55:41 +08:00
Andy Green
bc4012916d parser issue_char audit
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-20 07:40:13 +08:00
Andy Green
05ae649b66 timeout also fixup for0middle guy deletion case
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 23:11:39 +08:00
Andy Green
72dba09932 timeout check take copies since he may be deleted
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 22:34:44 +08:00
Andy Green
e34d4b4b35 coverity 156860 wsi dereference before NULL check
... bit in reality, wsi can never be NULL

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 21:33:52 +08:00
Andy Green
2e3bf06139 coverity 156861 context destroy dereference context before NULL check
Context could be NULL only if context creation failed in the first
place and user error path is bad... no network connectivity at that
point...

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 21:33:52 +08:00
Andy Green
2ec7c85e99 coverity 156862 unused return value
Not exactly a problem...

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 21:27:44 +08:00
Andy Green
d3a5505542 multithreaded service
This adds support for multithreaded service to lws without adding any
threading or locking code in the library.

At context creation time you can request split the service part of the
context into n service domains, which are load-balanced so that the most
idle one gets the next listen socket accept.

There's a single listen socket on one port still.

User code may then spawn n threads doing n service loops / poll()s
simultaneously.  Locking is only required (I think) in the existing
FD lock callbacks already handled by the pthreads server example,
and that locking takes place in user code.  So the library remains
completely agnostic about the threading / locking scheme.

And by default, it's completely compatible with one service thread
so no changes are required by people uninterested in multithreaded
service.

However for people interested in extremely lightweight mass http[s]/
ws[s] service with minimum provisioning, the library can now do
everything out of the box.

To test it, just try

$ libwebsockets-test-server-pthreads -j 8

where -j controls the number of service threads

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 20:02:36 +08:00
Andy Green
3f63560876 fix broken ext arg parsing
This gets us back to all the Autobahn tests passing

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 20:02:34 +08:00
Andy Green
d738f84ed1 timeout move to doubly linked list
In the case we have a lot of connections, checking them all for timeout state
once a second becomes burdensome.  At the moment if you have 100K connections,
once a second they all get checked for timeout in a loop.

This patch adds a doubly-linked list based in the context to each wsi, and
only wsi with pending timeouts appear on it.  At checking time, we traverse
the list, which costs nothing if empty because nobody has a pending timeout.

Similarly adding and removing from the list costs almost nothing since no
iteration is required no matter how big the list.

The extra 8 or 16 bytes in the wsi are offset a little bit by demoting .pps
from int to char (save 3 bytes).  And trim max act exts to 2, since we only
provide one, saving 8 /16 bytes by itself if exts enabled.

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-19 04:32:14 +08:00
Andy Green
9a9d5eaeeb avoid using deallocated things during context dedtroy
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-18 12:07:07 +08:00
Andy Green
a547554aa2 cleaning
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-18 11:16:20 +08:00
Andy Green
1d719ec87f lws_interface_to_sa
Expose interface handling code as an api
2016-01-16 12:09:38 +08:00
Andy Green
d5060d25a1 use lws_close not raw close
https://github.com/warmcat/libwebsockets/issues/400

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-16 11:14:46 +08:00
Andy Green
29cfeebcde permessage deflate male callback VISIBLE
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-15 22:16:54 +08:00
Andy Green
7c507e4700 fix client getting hung up on at negotiation
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-15 19:34:33 +08:00
Andy Green
809d69aa85 redirect
This adds redirect support to the client side.  Lws will follow
server redirects (301) up to three deep.

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-15 19:32:47 +08:00
Andy Green
7a0fcf2fc5 parse_uri
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-15 19:32:47 +08:00
Vijay Khurdiya
e73d446461 ssl add ECDH server support
(AG clean style add option flag and docs)
2016-01-15 16:21:51 +08:00
Andy Green
7c15eb1e42 only report CLIENT_CONNECTION_ERROR once on some paths
In most cases the close api will see it should send the CCE because
we are still in the waiting server reply state until the end of the
interpretation.  Only if we completed the interpretation and moved
on to ESTABLISHED do we need to handle sending it ourselves.

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-14 11:38:31 +08:00
Andy Green
389a4bb9f0 remove double free attempts from no ACCEPT server response cleanup path
The generic wsi close code is smart enough to clean up after these allocations itself

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-13 18:48:50 +08:00
Andy Green
8308073464 fuzzer eliminate hsecond ah free path firing assert sentinel
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-13 05:01:17 +08:00
Andy Green
9928cccdb6 fuzzer rx overflow mitigate
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-12 23:05:02 +08:00
Andy Green
5b3736682d lws_client_connect_via_info
https://github.com/warmcat/libwebsockets/issues/396

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-12 17:22:06 +08:00
Andy Green
6711266a50 extension permessage deflate
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-11 11:34:01 +08:00
Andy Green
58ad3d6e09 clean libev.c style
Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-09 08:13:55 +08:00
Andy Green
5783b1a169 libev set events to match revents
https://github.com/warmcat/libwebsockets/issues/393

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-09 04:12:46 +08:00
Andy Green
5eeb4bd0d3 partial always use partial buffering if send incomplete
After "rerelease"

https://github.com/warmcat/libwebsockets/issues/392#issuecomment-170003294

Since we introduced partial buffering a long while ago,
user code shold never see partial sends and very few
user callbsck attempt to deal with them.

Let's just eliminate the whole concept of user callback
partial send handling under any circumstances.

Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-08 21:52:47 +08:00
Denis Osvald
27174e64de properly skip protocols with NULL name
Signed-off-by: Denis Osvald <denis.osvald@sartura.hr>
2016-01-07 06:57:59 +08:00
Denis Osvald
034e514a0d fix callback typedefs and declarations
Remove declarations of callback and extension_callback as these are
functions declared in header but not defined anywhere.

Also rename typedefs callback_function and extension_callback_function
to lws_callback_function and lws_extension_callback_function so all
symbolx exported by header have lws prefix;

Signed-off-by: Denis Osvald <denis.osvald@sartura.hr>
2016-01-07 06:44:38 +08:00
Andy Green
0c7e5a9418 doc improve docs around header access apis
Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-30 14:53:50 +08:00
Andy Green
2b35e123f4 ext negotiation tolerate semicolon args
Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-30 12:12:58 +08:00