This adds two new callbacks in protocols[0] that are optional for allowing limited thread
access to libwebsockets, LWS_CALLBACK_LOCK_POLL and LWS_CALLBACK_UNLOCK_POLL.
If you use them, they protect internal and external poll list changes, but if you want to use
external thread access to libwebsocket_callback_on_writable() you have to implement your
locking here even if you don't use external poll support.
If you will use another thread for this, take a lot of care about managing your list of
live wsi by doing it from ESTABLISHED and CLOSED callbacks (with your own locking).
Signed-off-by: Andy Green <andy.green@linaro.org>
If enabled one listening socket will accept both SSL and plain HTTP connections.
Do not enable if you regard SSL handshake as some kind of security, eg, use
client-side certs to restrict access.
AG: changed flag names, added extra comments, changelog, add -a in test server
Signed-off-by: James Devine <fxmulder@gmail.com>
Signed-off-by: Andy Green <andy@warmcat.com>
This patch deploys the truncated send work to buffer output in case
either send() or the SSL send return a temporary "unable to send"
condition even though they signalled as writeable.
I added a by-default #if 0 test jig which enforces only half of what
you want to send is sendable, this is working when enabled.
One subtle change is that the pipe reports choked if there is any
pending remaining truncated send. Otherwise it should be transparent.
Hopefully...
Signed-off-by: Andy Green <andy.green@linaro.org>
We can't force the wsi state to HTTP_BODY without considering the callback
may already have set the state to sending a file.
This fixes the bug that we can get stalled in the test app at
"choked before able to send whole file"
Signed-off-by: Andy Green <andy.green@linaro.org>
The only part that actually goes to this label is inside such an ifdef,
so building without extension support makes gcc bail out since an unused
label is considered an error in this project.
Svetlin wrote on github
According to RFC2616, all header field names in both HTTP requests and HTTP responses are case-insensitive. But libwebsockets uses a case-sensitive compare.
Reproduce:
Run libwebsockets against a server that sends all of its HTTP header field names in lower-case (for example: https://github.com/extend/cowboy). libwebsockets reports an error. The expected behavior is no errors reported and a successful handshake procedure.
This changes the parser reference table in minilex to all lower case.
The code to walk the parser tables then just forces a tolower on the incoming chars.
This (and minilex tables) only applies to header names.
Reported-by: svetlin-mladnov <?@github>
Signed-off-by: Andy Green <andy.green@linaro.org>
If the URI coming from the client contains '?' then
- the URI part is terminated with a '\0'
- the remainder of the URI goes in a new header WSI_TOKEN_HTTP_URI_ARGS
- the remainder of the URI is not subject to path sanitization measures (it
still has %xx processing done on it)
In the test server, http requests now also dump header information to stderr.
The attack.sh script is simplified and can now parse the test server header dumps.
Signed-off-by: Andy Green <andy.green@linaro.org>
This translates %xx in the GET uri and removes /.. and /... type sequences along with
translating // or /// etc to /.
Since the result is hopefully secure, it also changes the test server to actually use
the uri path pasted on a resource directory without whitelisting.
Signed-off-by: Andy Green <andy.green@linaro.org>
Clean up minilex
Move the header output to stdout
Introduce lexfile.h as the header output
Use lexfile.h in both minilex itself and lws
Add the following header support
"Accept:",
"If-Modified-Since:",
"Accept-Encoding:",
"Accept-Language:",
"Pragma:",
"Cache-Control:",
"Authorization:",
"Cookie:",
"Content-Type:",
"Date:",
"Range:",
"Referer:"
Signed-off-by: Andy Green <andy.green@linaro.org>
A common practise to temporarily disable a socket in a poll call is to negate the socket fd. poll should then ignore the socket.
emulated_poll does this with the following code:
if (poll_fd->fd < 0 || !poll_fd->events)
goto skip1;
However on Windows the fd field in WSAPOLLFD is unsigned int!!! So the check for a negative fd value always fails.
This results in select returning an error with an error code of 10038 -- Socket operation on nonsocket.
The fix is to type cast fd like so:
if ((int)poll_fd->fd < 0 || !poll_fd->events)
This may be the cause of some high CPU load reports. I noticed the load being 50% with my application running on Windows XP.
Signed-off-by: Graham Newton <gnewton@peavey-eu.com>
This is important since this stuff is in publicly included headers
and even though the usage of the WSAPoll API is a runtime decision
the public headers may be used in code that needs to build with
_WIN32_WINNT=0x0501 to support WinXP.
When building a project using libwebsockets with that define set
to 0x0501 winsock.h will not define the WSAPOLLFD struct causing that
project to fail to compile.
The installed headers are all in the same directory, so no relative
path possible there. The cmake files already add the win32helpers
directory to the include paths, so this builds just fine.
I don't see a wsockcompat.h anywhere in MSVC9 or MSVC8 and their
corresponding sdk's. It does not seem like this is a standard windows
header, so better drop that and add the compat-defines to the same
place that already has other WSA compat defines.
I am using libwebsockets on Windows 7 in external poll mode.
I am finding that if I present a socket fd on a normal HTTP connection
(LWS_CONNMODE_HTTP_SERVING:) to libwebsocket_service_fd with just a HUP
event, the event doesn't get handled but revents gets cleared indicating
that the event has been handled. Should it be handled in the same way
as LWS_CONNMODE_WS_SERVING?
(Modified by AG to apply to all sockets)
Signed-off-by: Andy Green <andy.green@linaro.org>
Signed-off-by: Graham Newton <gnewton@peavey-eu.com>
Using Windows 7 64 bit, cloned repo on 20130926. Using Qt creator and Microsoft Visual C++ Compiler 9.0 (x86).
Result (errors from compile output): D:\Projects\CDPStudioAPI\libwebsockets_orig\lib\client-handshake.c:87: error: C2065: 'EALREADY' : undeclared identifier
D:\Projects\CDPStudioAPI\libwebsockets_orig\lib\client-handshake.c:87: error: C2065: 'EINPROGRESS' : undeclared identifier
Possible solution is to use wsockcompat.h (compatibility header for using EALREADY, EINPROGRESS etc in older versions of Windows SDK). Compiled fine when I #included wsockcompat.h into client-handshake.c
Reported-by: mart22n via Trac 41
Signed-off-by: Andy Green <andy.green@linaro.org>
A few parameters (like "${PROJECT_BINARY_DIR}/lws_config.h" and "${PROJECT_SOURCE_DIR}/lib") are passed to CMake commands in your build script without enclosing them by quotation marks. I see that these places will result in build difficulties if the contents of the used variables will contain special characters like spaces.
Signed-off-by: Markus Elfring <markus.elfring@web.de>
Subject: [PATCH] We can ran into situation (at least on iOS) when with openssl
nonblocking BIO and http proxy we don't perform ssl_connect straight away so
we need to retry until we finish ssl_connect. If we don't do that we will
fail in LWS_CONNMODE_WS_CLIENT_WAITING_PROXY_REPLY when testing for "HTTP/1.0
200" successful connection.
Signed-off-by: shys <shyswork@zoho.com>
add function to manually setup proxy. Useful on iOS where
getenv doesn't return proxy settings
Simplified by AG
Signed-off-by: shys <shyswork@zoho.com>
Signed-off-by: Andy Green <andy@warmcat.com>
Hi all,
please consider applying the patch below. The git is called in some
different directory and that is why fails. I'm not sure why it was
called in different directory, maybe because I build libwebsockets
with CMake's ExternalProject.
Best,
Arokux
Signed-off-by: Arokux X <arokux@gmail.com>
This patch adds code to handle the situation that a prepared user buffer could not all be sent on the
socket at once. There are two kinds of situation to handle
1) User code handles it: The connection only has extensions active that do not rewrite the buffer.
In this case, the patch caused libwebsocket_write() to simply return the amount of user buffer that
was consumed (this is specifically the amount of user buffer used in sending what was accepted,
nothing else). So user code can just advance its buffer that much and resume sending when the socket
is writable again. This continues the frame rather than starting a new one or new fragment.
2) The connections has extensions active which actually send something quite different than what the
user buffer contains, for example a compression extension. In this case, libwebsockets will dynamically
malloc a buffer to contain a copy of the remaining unsent data, request notifiction when writeable again,
and automatically spill and free this buffer with the highest priority before passing on the writable
notification to anything else. For this situation, the call to write will return that it used the
whole user buffer, even though part is still rebuffered.
This patch should enable libwebsockets to detect the two cases and take the appropriate action.
There are also two choices for user code to deal with partial sends.
1) Leave the no_buffer_all_partial_tx member in the protocol struct at zero. The library will dyamically
buffer anything you send that did not get completely written to the socket, and automatically spill it next
time the socket is writable. You can use this method if your sent frames are relatvely small and unlikely to get
truncated anyway.
2) Set the no_buffer_all_partial_tx member in the protocol struct. User code now needs to take care of the
return value from libwebsocket_write() and deal with resending the remainder if not all of the requested amount
got sent. You should use this method if you are sending large messages and want to maximize throughput and efficiency.
Since the new member no_buffer_all_partial_tx will be zero by default, this patch will auto-rebuffer any
partial sends by default. That's good for most cases but if you attempt to send large blocks, make sure you
follow option 2) above.
Signed-off-by: Andy Green <andy.green@linaro.org>
