This implements HTTP/1.1 per RFC 7230, partially.
Unit-tests are provided with documentation on usage.
This is desgined to serve as the http implementation
throughout loolwsd, for both synchronous and
asynchronous requests.
Change-Id: Iaf1b8c5fcb8cec032445e27c9f70d2fb807aa4dc
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Move the connect function into the NetUtil
translation unit to aid using it for the
upcoming async socket logic.
The NetUtil should also come in handy for
the miscellaneous network helpers we have.
Change-Id: I2ee0c6e3e1769fd87572d7407d3b4979b59ffe6a
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
When we simulate socket error it's hard to understand
what is going on, since there is no real errno to
explain the -1 return value. This is at least true
in the SSL case where we don't set errno (perhaps
we should?). At any rate, it's good form to log
that we are messing up with the return value and
that the real socket API hasn't been even invoked.
Change-Id: Ib76f0259dbdcfe0cfae97343c11ca45461079fa1
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Add support for configuring the SSL certificates
when running the test binary as well as setting
up the SSL context.
Also adds the SSL socket headers with proper
compile-time guards for when SSL is disabled.
Change-Id: I99992639a66a64871f8ff8a2b2105279ead63ca1
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Most C and Posix API clobber errno. By failing to save
it immediately after invoking an API we risk simply
reporting the result of an arbitrary subsequent API call.
This adds LOG_SYS_ERRNO to take errno explicitly.
This is necessary because sometimes logging is not done
immediately after calling the function for which we
want to report errno. Similarly, log macros that log
errno need to save errno before calling any functions.
This is necessary as the argements might contain calls
that clobber errno.
This also converts some LOG_SYS entries to LOG_ERR
because there can be no relevant errno in that context
(f.e. in a catch clause).
A couple of LOG_ macros have been folded into others,
reducing redundancy.
Finally, both of these log macros append errno to the
log message, so there is little point in ending the
messages with a period.
Change-Id: Iecc656f67115fec78b65cad4e7c17a17623ecf43
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Cleans up some of the conversions and implicity
unit in integral types.
Change-Id: I79f35b92f8f631894e55bdb39851b050870fce96
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This merges OpenSSL's poll events with ours.
Effectively, we now do a single poll when
there are reads and writes to be done,
regardless of the reason (i.e. SSL-specific
or application-specific).
Simpler code, and more efficient performance
by sharing code with http and reducing the
number of poll syscalls.
Change-Id: Ib329c7e76fccfdadc4a0783c1ad79c3eedcdd8f3
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Hopefully reasonably simple; we perturb the count in the poll to
avoid starving a seventh socket in a poll.
Change-Id: I1a39cc36b9599ffe82186b896c6fd91d792c4127
Signed-off-by: Michael Meeks <michael.meeks@collabora.com>
Also, makes the logging of units much less error prone.
The overloaded streaming operators are temporary as
they are provided in C++20. The ones here (though
incomplete) are fashioned after the C++20 specs.
Change-Id: Ieb499282ccb6e63fa939ba07bed3e5a4fbef1bd0
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
A number of call-sites, eg. clipboard, or admin-ws were
writing to sockets assuming they could return all the data
in a single series of writes, without needing to poll. As
such they failed to addSocketToPoll on the new poll - eg.
the docBroker. Unfortunately this meant that on EAGAIN
writes, the socket would be closed and the last parts
of a message lost.
Browsers would give net::ERR_CONTENT_LENGTH_MISMATCH 200 (OK)
The situation is/was intermittent, so painful to debug.
On under-loaded developer machines, socket buffers are larger,
so this was seldom seen.
The re-factor forces a transfer to another SocketPoll via
the disposition, except for a couple of corner cases.
Change-Id: I2f1b2f99f179c4fda84464c9241fe434fa527725
Signed-off-by: Michael Meeks <michael.meeks@collabora.com>
Avoid continually shuffling large output buffers down in memory,
it kills CPU caches, does nothing useful, and we can just advance
an offset instead. Of course, we do need to free memory - so do
that for lots of common cases.
Change-Id: I63240661ca87bdc0574acbce856771d9dde3fa94
Signed-off-by: Michael Meeks <michael.meeks@collabora.com>
For large transfers eg. image previews, particularly with SSL's
protocol limit of 16k byte blocks, we see lots of inefficiency
repeatedly copying a 20Mb image and shuffling it down a
std::vector as we write data out.
Change-Id: I620568cad2e6f41684c35289b0ee77cf7f59c077
Signed-off-by: Michael Meeks <michael.meeks@collabora.com>
Otherwise, there is no way to know which connection
a given log belongs to, when websrv_poll handles
multiple simultaneous connections.
Also removed a duplicate log entry.
Change-Id: Id74e6e7b66062daf15a716b9fea8e7995fd75af0
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
This is non-functional change to have all HttpHelper
members in a single place (HttpHelper.cpp).
Clang-format is applied and FileUtil::Stat is used
instead of a naked stat(2).
Change-Id: I663e2eb3070e873bd1f4477b60e45b0086d29f4e
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
Per the rfc (https://tools.ietf.org/html/rfc6455#section-5.5.2):
"An endpoint MAY send a Ping frame any time after the connection
is established and before the connection is closed."
And "Upon receipt of a Ping frame, an endpoint MUST send a Pong
frame in response, unless it already received a Close frame."
Here we allow for pings to come from clients and we respond
to them by pongs, as required by rfc 6455.
Change-Id: I8e285f095526e4b67373ecb3ae1efc9c8717d756
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/102948
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Andras Timar <andras.timar@collabora.com>
Because Admin derives from it, and UnitPHPProxy holds a
shared_ptr<SocketPoll>, so the non-virtual dtor was invoked for a class
whcih is not (and can't be) final.
Found by clang's -Werror,-Wdelete-non-abstract-non-virtual-dtor.
Change-Id: I9b31bb58e79b2da3a3896789c52258f315fdd3d6
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/102355
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Sometimes multiple messages are processed in a single iteration
at socket level. This happens in WebSocketHandler and when draining
Document queue.Just covered these cases.
Change-Id: Ifa46f5d484b67015cca64008b2c89426cc839e64
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/99387
Tested-by: Jenkins
Tested-by: Gabriel Masei <gabriel.masei@1and1.ro>
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Reviewed-by: Gabriel Masei <gabriel.masei@1and1.ro>
The map._activate, among other actions, is sending indirectly some messages
to the server like clientzoom and clientvisiblearea. If these messages are send
before the document finishes processing the load message then there is
a chance that a nodocloaded error will be thrown because there is a
chance that the messages will be processed in parallel with load. This happens
constantly for xlsx files. This is generated by the Unipoll mechanism which,
in case of xlsx files, triggers a parallel processing.
To avoid the above scenario a mechanism of disabling parallel processing of
messages in kit was implemented and is used for load and save messages, for now.
Change-Id: I4c83e72e600f92d0bb4f1f18cebe694e326256d0
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/98519
Tested-by: Jenkins
Tested-by: Michael Meeks <michael.meeks@collabora.com>
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
We probably used to have circular references that made KitSocketPoll
and KitWebSocketHandler objects hang around forever, or something.
(Not a problem in web-based Online where kit processes have a
restricted lifetime.)
Change-Id: Ia6eebc51f4a4a8fb4f69a2c83a0131de921ea1d6
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/98744
Tested-by: Jenkins
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Tor Lillqvist <tml@collabora.com>
loolmount now works and supports mounting and
unmounting, plus numerous improvements,
refactoring, logging, etc.. When enabled,
binding improves the jail setup time by anywhere
from 2x to orders of magnitude (in docker, f.e.).
A new config entry mount_jail_tree controls
whether mounting is used or the old method of
linking/copying of jail contents. It is set to
true by default and falls back to linking/copying.
A test mount is done when the setting is enabled,
and if mounting fails, it's disabled to avoid noise.
Temporarily disabled for unit-tests until we can
cleanup lingering mounts after Jenkins aborts our
build job. In a future patch we will have mount/jail
cleanup as part of make.
The network/system files in /etc that need frequent
refreshing are now updated in systemplate to make
their most recent version available in the jails.
These files can change during the course of loolwsd
lifetime, and are unlikely to be updated in
systemplate after installation at all. We link to
them in the systemplate/etc directory, and if that
fails, we copy them before forking each kit
instance to have the latest.
This reworks the approach used to bind-mount the
jails and the templates such that the total is
now down to only three mounts: systemplate, lo, tmp.
As now systemplate and lotemplate are shared, they
must be mounted as readonly, this means that user/
must now be moved into tmp/user/ which is writable.
The mount-points must be recursive, because we mount
lo/ within the mount-point of systemplate (which is
the root of the jail). But because we (re)bind
recursively, and because both systemplate and
lotemplate are mounted for each jails, we need to
make them unbindable, so they wouldn't multiply the
mount-points for each jails (an explosive growth!)
Contrarywise, we don't want the mount-points to
be shared, because we don't expect to add/remove
mounts after a jail is created.
The random temp directory is now created and set
correctly, plus many logging and other improvements.
Change-Id: Iae3fda5e876cf47d2cae6669a87b5b826a8748df
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/92829
Tested-by: Jenkins
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Seems to not cause any serious regressions in the iOS app or in "make
run", but of course I am not able to run a comprehensive check of all
functionality.
Change-Id: I44a0e8d60bdbc0a885db88475961575c5e95ce88
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/93037
Tested-by: Jenkins
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Tor Lillqvist <tml@collabora.com>
There aren't multiple processes that would need to "ping" each others.
Ifdef out the related member variables and code completely. Having
them partially in caused lots of FakeSocket polling with zero timeout
which is less than ideal.
Change-Id: Ibdfa4980d6d4fc9c00ea5146ca8d75ca0df81f1d
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/97021
Tested-by: Jenkins
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Tor Lillqvist <tml@collabora.com>
Set a FAKESOCKET_LOG_LEVEL environment variable to "2" for more
verbose logging. This is how it used to be, and is indeed very
verbose, as each poll, read, and write operation is logged.
(Normally the FakeSocket logging does not get displayed, though, as it
is passed to LOG_INF() and the default LOOL_LOGLEVEL is "warning". To
see it, either set FAKESOCKET_LOG_ALWAYS_STDERR or set LOOL_LOGLEVEL
appropriately.)
With the default log level 1 only creation, connection, and closing of
FakeSockets is logged, and the state of all active ones is displayed
after each established connetion and when a FakeSocket has been
closed. This is usually enough to get a basic trace of how the
plumbing works.
Change-Id: Id87bd32ce06105af561aa6a75cf365b41c079713
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/96943
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Tor Lillqvist <tml@collabora.com>
re-factor to make it hard not to.
Change-Id: I26ebc48b4660276ede64a22167ac4779cebf5cd4
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/95440
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
More readable and typically more efficient.
Change-Id: I9bd5bfc91f4ac255bb8ae0987708fb8b56b398f8
Reviewed-on: https://gerrit.libreoffice.org/c/online/+/95285
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Tested-by: Jenkins
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
