Secure Streams is an optional layer on top of lws that separates policy
like endpoint selection and tls cert validation into a device JSON
policy document.
Code that wants to open a client connection just specifies a streamtype name,
and no longer deals with details like the endpoint, the protocol (!) or anything
else other than payloads and optionally generic metadata; the JSON policy
contains all the details for each streamtype. h1, h2, ws and mqtt client
connections are supported.
Logical secure streams outlive any particular connection and supports "nailed-up"
connectivity regardless of underlying connection stability.
Surprisingly -fdata-sections -ffunction-sections does not remove any string literals and __func__
implicit .rodata generated by the removed function's compilation.
That means potentially considerable deadweight is in the image even if the function is removed
at linktime.
This adds support for POST in both h1 and h2 queues / stream binding.
The previous queueing tried to keep the "leader" wsi who made the
actual connection around and have it act on the transaction queue
tail if it had done its own thing.
This refactors it so instead, who is the "leader" moves down the
queue and the queued guys inherit the fd, SSL * and queue from the
old leader as they take over.
This lets them operate in their own wsi identity directly and gets
rid of all the "effective wsi" checks, which was applied incompletely
and getting out of hand considering the separate lws_mux checks for
h2 and other muxed protocols alongside it.
This change also allows one wsi at a time to own the transaction for
POST. --post is added as an option to lws-minimal-http-client-multi
and 6 extra selftests with POST on h1/h2, pipelined or not and
staggered or not are added to the CI.
There are some minor public api type improvements rather than cast everywhere
inside lws and user code to work around them... these changed from int to
size_t
- lws_buflist_use_segment() return
- lws_tokenize_t .len and .token_len
- lws_tokenize_cstr() length
- lws_get_peer_simple() namelen
- lws_get_peer_simple_fd() namelen, int fd -> lws_sockfd_type fd
- lws_write_numeric_address() len
- lws_sa46_write_numeric_address() len
These changes are typically a NOP for user code
If a client connects to a SSL server and the server sends handshake
alert (e.g. no matching ciphers) SSL_connect() fails, but because
SSL_ERROR_SSL return value is not handled, it's not considered a
failure. SSL_want_read() will return 1 and the client will happily wait
for more data from the server. Now if the server closes connection after
sending handshake alert, POLLIN event will be triggered,
lws_tls_client_connect() called again, but SSL_connect() will fail
without calling read(), so the client will end up consuming 100% CPU
because POLLIN will be triggered repeatedly.
Similar error handling is used in lws_tls_server_accept() and the
condition checks for SSL_ERROR_SSL. Using the same condition in
lws_tls_client_connect() fixes the problem.
Tested with OpenSSL 1.0.2k.
Now the generic lws_system blobs can cover client certs + key, let's
add support for applying one of the blob sets to a specific client
connection (rather than doing it via the vhost).
Client connection items for protocols other than http ones
will never get into an ah. Allow use of the values from the
client stash allocation instead if present.
This teaches http client stuff how to handle 303 redirects... these
can happen after POST where the server side wants you to come back with
a GET to the Location: mentioned.
lws client will follow the redirect and force GET, this works for both
h1 and h2. Client protocol handler has to act differently if it finds
it is connecting for the initial POST or the subsequent GET, it can
find out which by checking a new api lws_http_is_redirected_to_get(wsi)
which returns nonzero if in GET mode.
Minimal example for server form-post has a new --303 switch to enable
this behaviour there and the client post example has additions to
check lws_http_is_redirected_to_get().
Remove LWS_LATENCY.
Add the option LWS_WITH_DETAILED_LATENCY, allowing lws to collect very detailed
information on every read and write, and allow the user code to provide
a callback to process events.
wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer
all now participate on a single sorted us list.
The whole idea of polling wakes is thrown out, poll waits ignore the
timeout field and always use infinite timeouts.
Introduce a public api that can schedule its own callback from the event
loop with us resolution (usually ms is all the platform can do).
Upgrade timeouts and sequencer timeouts to also be able to use us resolution.
Introduce a prepared fakewsi in the pt, so we don't have to allocate
one on the heap when we need it.
Directly handle vh-protocol timer if LWS_MAX_SMP == 1
There are quite a few linked-lists of things that want events after
some period. This introduces a type binding an lws_dll2 for the
list and a lws_usec_t for the duration.
The wsi timeouts, the hrtimer and the sequencer timeouts are converted
to use these, also in the common event wait calculation.
lws_dll2 removes the downsides of lws_dll and adds new features like a
running member count and explicit owner type... it's cleaner and more
robust (eg, nodes know their owner, so they can casually switch between
list owners and remove themselves without the code knowing the owner).
This deprecates lws_dll, but since it's public it allows it to continue
to be built for 4.0 release if you give cmake LWS_WITH_DEPRECATED_LWS_DLL.
All remaining internal users of lws_dll are migrated to lws_dll2.
Since the messages are queued and then read in order from the event loop
thread, it's not generally safe to pass pointers to argument structs,
since there's no guarantee the lifetime of the thing sending the message
lasted until the sequencer read the message.
This puts pressure on the single void * argument-passed-as-value... this patch
adds a second void * argument-passed-as-value so it's more possible to put
what's needed directly in the argument.
It's also possible to alloc the argument on the heap and have the sequencer
callback free it after it has read it.
Some toolchains are in some C++ version mode that objects to redefinition
of a typedef even to the same thing.
Shift a couple for function declarations in the mbedtls wrapper
part so they can piggyback on just the one typedef for their
args.
Latest 1.1.1c (and patches 1.1.1b on Fedora) check the AES key for entropy
and error out if bad. Our aes-xts test key was a by-hand pattern repeated 4
times and OpenSSL errors out on it.
Improve the key to a random one.
