Since commit ef4b1bdb "red-channel-client: Prevent too tight loop
waiting for ACKs", after seamless migration, the display is no longer
updated on the client-side, even though the VM is functional (responds
to keyboard input, reconnecting the client restores the display
functionality, ...).
This is mainly caused because after migration,
red_channel_client_waiting_for_ack() will be true until
red_channel_client_ack_zero_messages_window() is called.
What happens is the following:
The dcc is created, and dcc_start() pushes a RED_PIPE_ITEM_TYPE_SET_ACK
message.
This calls prepare_pipe_add(), which will enable write event on the dcc
watch. red_channel_client_event() will be called, which will trigger a
red_channel_client_push(). Since red_channel_client_waiting_for_ack()
returns true, we won't get any item to push, and (because of commit
ef4b1bdb), we will disable write notifications on the watch.
At this point, rcc->priv->pipe is no longer empty, so prepare_pipe_add()
is not going to reenable the write notifications.
Then red_channel_client_ack_zero_messages_window() is finally called as
part of dcc_handle_migrate_data(), so from this point on,
red_channel_client_waiting_for_ack() is no longer true.
However, nothing ever reenables WRITE events, nor empties
rcc->priv->pipe, so nothing ever gets pushed, causing no display updates
at all after a migration, even if the VM is functional (input, ...)
apart from that.
This commit reenables WRITE events in
red_channel_client_ack_zero_messages_window() if we were waiting for
ack.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Since 2.8, QEMU no longer creates QXL primary surfaces when using GL.
This change broke client-side mouse mode, because Spice server relies on
having a primary surface.
When GL is enabled, use GL scanout informations.
Mouse mode is always client when GL surfaces are used.
This patch and most of the message are based on a patch from
Marc-André Lureau, just moving responsibility from reds to RedQxl.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Christophe de Dinechin <dinechin@redhat.com>
Currently, the port used by most tests is hardcoded to 5912. However,
the test suite can be run in parallel, so if 2 tests run in parallel,
the 2nd one is not going to be able to bind to port 5912 and will fail.
After this commit, test_new() will try to find a free port between 5912
and 5922 and will abort if it can't find any.
The issue can be reproduced by adding a usleep(1000000) to the beginning
of test_destroy().
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Start reducing the usage of spice_new*/spice_malloc allocations.
They were designed in a similar way to GLib ones.
Now that we use GLib make sense to remove them.
However the versions we support for GLib can use different memory
allocators so we have to match g_free with GLib allocations
and spice_* ones (which uses always malloc allocator) with free().
This patch remove some easy ones.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
The macro will implement most of the boilerplate needed to declare an
object.
Its usage is similar to GLib G_DECLARE_*_TYPE macros.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
RedChannelClient has a "handle-acks" feature.
If this feature is enabled, after the configured number of messages it
waits for an ACK from the client.
If is waiting for an ACK it stops sending messages.
However the write notification was not disabled, causing the loop event
to always trigger, as the socket in this case is ready to accept data.
Specifically red_channel_client_event is continuously called.
This is noticeable using slow network environments and having
some additional loop instrumentation.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This structure was used to store the cookie for the async
reply and the message for the generic async callback.
Most async messages do not require extra action beside sending back the
cookie for the reply so instead of having a switch on the message type
in red_qxl_async_complete, this commit moves the message-specific
behaviour to the callers, which allows us to store the cookie directly
in RedWorkerMessageAsync rather than needing an intermediate
AsyncCommand structure.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Now the push is done automatically when a PipeItem is added
(cfr commit 5c460de1a3
"worker: push data when clients can receive them"),
forcing a push cause only network fragmentation and is required only if
you are handling data in a polling loop (and thus, you are preventing
the default event loop from running).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
It was using g_list_remove_link() to remove an element from the
RedChannel::clients list while it really meant to be using
g_list_delete_link() which frees the memory associated with the link.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
All main test module have this test-XXXX.c naming, make
test-stat coherent.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Put non-trivial programs in separate sections, which makes it easier to
understand the relationship between macros.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
New automake test harness produce *.log and *.trs files for
each test.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This macro is exactly doing what RING_FOREACH just passing streams
ring.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe de Dinechin <cdupontd@redhat.com>
There is exactly one user in RedChannel, and this can be reimplemented
using already public RedChannelClient API. No need for an extra
function very specialized function with a not great name.
This commit thus removes one method from RedChannelClient public API,
and replaces it with an equivalent private helper in RedChannel.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
red_channel_disconnect_if_pending_send() and red_channel_wait_all_sent() are
always called together, we can remove one of the 2 methods.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
In case GLib don't provide these functions we use replacements so
there's no need to have a warning if these functions are called.
This potentially capture other compatibility issues in the tests
that would be ignored having all deprecation warnings disabled.
Tested with GLib 2.28 and 2.52.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
When the worker is started it could take a while to start processing
commands.
The reason is that the dispatcher handler is called after the worker
so GLib will receive a FALSE answer to both prepare and check
callbacks of the RedWorkerSource causing GLib to wait till another
event is received.
This is a regression since the introduction of GLib event loop, before
the command processing was always attempted after any events.
Commands (from QXL interface for cursor and display) are processed
during the RedWorkerSource dispatch so if they are not processed just
when the VM is started they will be processed on next event which
could be from dispatcher (main thread requests), from existing
connections or from pending timers. However in the case there are no
clients connected and no other requests from main thread the worker
thread won't process them.
Setting the event_timeout to 0 cause the prepare callback to return
TRUE so GLib will dispatch the RedWorkerSource.
This was discovered attempting to use the tests in server/tests
directory to reproduce a leak in RedWorker.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Pipelines are never freed.
These are detected as leaks by leak detector tools like address sanitizer.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Command line options are not freed at the end of the program.
These are detected as leaks by leak detector tools like address sanitizer.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
encoder_name is never NULL as already initialized with "mjpeg" value.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
If a client is unable to complete the TLS handshake phase
reds_init_client_ssl_connection leaked some memory as the stream is not
correctly freed.
This also causes the stream to send the SPICE_CHANNEL_EVENT_DISCONNECTED
event. Otherwise only SPICE_CHANNEL_EVENT_CONNECTED was sent.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Currently is possible to trigger a leak by passing an invalid
connection.
This can happen if the client opens a connection and then closes it
without writing or reading any data.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The code tests for the presence of RedChannelClient::stream while
initializing RedChannelClient.
However, the check was done too late, and a
RedChannelClient::config_socket implementation (for example
snd_channel_client_config_socket) could have tried to use it before the
check that it's not NULL.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This potentially can also save the copy if there is enough
space to resize the buffer in place.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This field was used only by RedChannelClient to mark when the socket
was shutdown. This condition can simply be tested by RedChannelClient
checking if there's a watch as is the only condition (beside object
destroying/disconnecting) where the watch is removed.
In any case the shutdown was used to understand if there were possible
data still to read.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
RedQxl and RedWorker are quite bound together running
CursorChannel and DisplayChannel in a separate thread
marshalling (RedQxl) and unmarshalling and executing
(RedWorker) requests.
Make the communication between them private trying
to facilitate maintaining these two files.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
For dispatcher_register_handler(), use 'false' instead of 0 since the
last argument is a bool type now.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
This callback was only executed for message types that were registered
with DISPATCHER_ASYNC ack type. However, the async_done handler was
called immediately after the message-specific handler and was called in
the same thread, so the async_done stuff can just as easily be done from
within the message-specific handler. This allows to simplify the
dispatcher_register_handler() method to simply require a boolean
argument for whether the message type requires an ACK or not.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Pass proper type to callback to avoid having to convert to
the right type for each call.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Is supposed to be used during initialization but is never
used.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
AsyncCommand is used to handle asynchronous messages from the
dispatcher.
GL_DRAW_ASYNC is mainly using it to store the cookie.
The value of GL_DRAW_COOKIE_INVALID was choosen to allow implementing
cookies (which basically are handles) either using indexes (where 0 is
valid) or pointers (where 0 is invalid). Currently Qemu uses pointers.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
These 2 functions were doing the same stuff, calling
client_monitors_config callback in QXLInterface.
The only difference was that red_qxl_use_client_monitors_config
used a NULL value.
Added the check for proper version, QXLInstance before 3.3
did not have this callback.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
You could easily trigger this issue using multiple monitors and a
modified spice-gtk client with this patch:
--- a/src/channel-main.c
+++ b/src/channel-main.c
@@ -1699,6 +1699,7 @@ static gboolean _channel_new(channel_new_t *c)
{
g_return_val_if_fail(c != NULL, FALSE);
+ if (c->type == SPICE_CHANNEL_DISPLAY) c->id = 0;
spice_channel_new(c->session, c->type, c->id);
g_object_unref(c->session);
which cause a crash like
(process:28742): Spice-WARNING **: Failed to create channel client: Client 0x40246f5d0: duplicate channel type 2 id 0
2017-08-24 09:36:57.451+0000: shutting down, reason=crashed
RedChannelClient is an GInitable type, which means that the object is
constructed, and then the _init() function is called, which can fail.
If the _init() fails, the newly-created object will be destroyed. As
part of _init(), we add a new watch for the stream using the core
interface that is associated with the channel. After adding the watch,
our rcc creation fails (due to duplicate ID), and the rcc object is
unreffed. This results in a call to reds_stream_free() (since the rcc
now owns the stream). But in reds_stream_free, we were trying to remove
the watch from the core interface associated with the RedsState. For
most channels, these two core interfaces are equivalent. But for the
Display and Cursor channels, it is the core Glib-based interface
associated with the RedWorker.
The watch in RedsStream by default is bound to the Qemu provided
SpiceCoreInterface while RedChannelClient bound it to Glib one causing
the crash when the watch is deleted from RedsStream. Change the bound
interface.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
When a stream is moved from the main thread to a
secondary one the events are potentially registered
using a different core interface. This cause memory
corruption accessing the watch registered in RedsStream.
This patch allows to always use the right interface.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This test runs 2 spice server in one program.
Use two different tcp port to be able to connect to both servers.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Add some check that something happened during creation/destruction.
Set as running on "make check".
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Update tests names.
Remove tetris comments, never available and not planned.
Update some notes.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
There's no need to not compile this feature, it just enable
a parameters which must be passed in order to change test
behaviour.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This allows to end the loop to end some tests.
Currently different tests enter the loop but never exit from
them.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
In C the sizeof(long) can be different than sizeof(void*),
use proper type.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Do not use calloc and malloc directly without checking
the result. Use instead spice functions to get a nice
error in case of allocation failures.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
For some reasons (documented in cursor_init) the function
uses 128 extra bytes of data causing a reading buffer overflow.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Timers in spice server are supposed to be used in a single thread
context. To avoid problems, protect the usage from multiple
thread with a mutex.
This sometimes caused a crash.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The wakeup timer is used by the worker thread and by the
main thread.
Destroying the object before destroying the worker thread
can lead to use after free.
Destroying the worker thread first makes sure we don't race.
This is detected easily when compiling the test with address sanitizer.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
As the indexes are used to compute the index inside an array
using modulo operation when a signed value overflows, the
modulo becames negative, causing a buffer underflow.
Unlikely to happens (take lot of time) but is safer that way.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Cursor resources (basically the shape of it) was retained till
it was used however it was copied so there were no reason to not release
this resource.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
You could easily trigger this issue using multiple monitors and
a modified spice-gtk client with this patch:
--- a/src/channel-main.c
+++ b/src/channel-main.c
@@ -1699,6 +1699,7 @@ static gboolean _channel_new(channel_new_t *c)
{
g_return_val_if_fail(c != NULL, FALSE);
+ if (c->type == SPICE_CHANNEL_DISPLAY) c->id = 0;
spice_channel_new(c->session, c->type, c->id);
g_object_unref(c->session);
This as g_initable_new in this case returns NULL (dcc.c).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Although dispatcher_send_message() does not allow you to send a message
type that is invalid for a dispatcher, it still makes sense to verify
that the type is valid in the receiver. This should only be possible in
the case of some severe problem such as memory corruption, so if it is
invalid, we simply abort.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Is possible to have a leak processing update commands if
the update command is synchronous and the rectangle list
is empty. Note that Qemu always pass an empty list.
If the list is empty display_channel_update fill the list.
This is used to send back the list in case of asynchronous
requests. But in handle_dev_update_async (the callback that
handle the asynchronous case) the list is correctly freed.
This was discovered by accident looking at the code.
Reproduced with a Windows recording file using GCC address
sanitizer and this patch to spice-server-replay:
--- a/server/red-replay-qxl.c
+++ b/server/red-replay-qxl.c
@@ -1280,7 +1280,13 @@ static void replay_handle_dev_input(QXLWorker *worker, SpiceReplay *replay,
replay->created_primary = FALSE;
worker->destroy_surfaces(worker);
break;
- case RED_WORKER_MESSAGE_UPDATE:
+ case RED_WORKER_MESSAGE_UPDATE: {
+ static uint8_t count = 0;
+ QXLRect dummy;
+ QXLRect update = { 0, 0, 100, 100 };
+ count ^= 1;
+ worker->update_area(worker, 0, &update, count ? &dummy : NULL, count ? 1 : 0, 0);
+ } break;
// XXX do anything? we record the correct bitmaps already.
case RED_WORKER_MESSAGE_DISPLAY_CONNECT:
// we want to ignore this one - it is sent on client connection, we
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
RedPipeItem already implements reference counting so
this avoid duplicating code to handle a object with reference
counting that points to another object with reference counting
that holds a RedCursorCmd.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Now the push is done automatically when a PipeItem is added
(cfr commit 5c460de1a3
"worker: push data when clients can receive them"),
forcing a push cause only network fragmentation and is required
only if you are handling data in a loop instead of using the
default loop.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Implements red_channel_pipes_add_type and
red_channel_pipes_add_empty_msg using red_channel_pipes_add.
This avoid duplicating items for each client.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Considering that now RedPipeItem have reference counting
and that lot of items are just used to store constant
data to send, using reference counting instead of creating
different items for each client is easier to do.
So this new red_channel_pipes_add allows to add a single item
to all clients.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Now the push is done automatically when a PipeItem is added
(cfr commit 5c460de1a3
"worker: push data when clients can receive them"),
forcing a push cause only network fragmentation and is required only if
you are handling data in a polling loop (and thus, you are preventing
the default event loop from running).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
display-channel.h contains lots of information used by different
DisplayChannel components.
In the past all RedWorker, CursorChannel and DisplayChannel code was in
a single file. Since lots of code to handle DisplayChannel is still in
RedWorker, display-channel.h contains a lot of declarations so that they
can be accessed from RedWorker.
Moving declarations that are not needed by RedWorker and other external
class components helps to reduce dependencies between RedWorker and
DisplayChannel.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This patch allocates VMC IDs by finding the first ID not used
instead of using a global variable and incrementing the value
for each channel created.
This solves some potential issues:
- remove the global state potentially making possible
to use multiple SpiceServer on the same process;
- don't potentially overflow the variable. This can happen if
channels are allocated/deallocated multiple times
(currently not done by Qemu).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This can happen as the connection is asynchronous so (MT main thread,
CT channel thread):
- MT you get a new connection;
- MT a connection is sent to CT;
- MT you get a disconnection of main channel;
- MT red_client_destroy is called;
- CT you attempt to add the RCC to RedClient.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
A RedChannelClient is always attached to a valid RedChannel.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Name will be visible in debugger and /proc filesystem
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The message is asynchronous so to avoid the object to potentially
been released before being processed keep a strong reference to
it.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
red_disconnect_display() is duplicating what red_channel_disconnect()
already does, so red_disconnect_display() and red_disconnect_cursor()
are actually identical code-wise. We can directly call
red_channel_disconnect() from flush_commands() rather than passing a
'red_disconnect_t disconnect' argument to that function.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
cursor_channel_disconnect() calls
cursor_channel_client_reset_cursor_cache() on all CursorChannelClient
associated with the current CursorChannel before calling
red_channel_disconnect().
red_channel_disconnect() will iterate over all CursorChannelClient
calling red_channel_client_disconnect(), which will eventually call
CursorChannelClient::on_disconnect. This will in turn
cursor_channel_client_reset_cursor_cache(), so calling it in
cursor_channel_disconnect() before calling red_channel_disconnect() is
redundant.
cursor_channel_disconnect() can thus be replaced by a direct call to
red_channel_disconnect().
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
SoundChannelClient has a stub implementation of
RedChannelClient::on_disconnect(), this commit removes the need for it.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This vfunc only has a RedChannelClient * argument, and most of the time,
it operates on RedChannelClient, not on RedChannel. Moreover, the only
time it's used is from RedChannelClient. This commit moves the vfunc to
RedChannelClient, which seems like a better fit for it.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
ORC library is used internally by GStreamer to generate code
dynamically.
If ORC cannot allocate executable memory, the failure causes
an abort(3) to be called.
This happens on some SELinux configurations that disable executable
memory allocation (execmem boolean).
Check that ORC could work before attempting to use GStreamer to
avoid crashes.
While this check is done, the ORC library outputs an error which will
be well visible in Qemu output.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
If you are testing for NULL data this means that variable could be
NULL so avoid to access before the check to make sure the check is hit.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The multimedia time is defined as uint32_t.
Use the proper type instead of int.
Currently no arithmetic is done on this value but
just copies so considering that on the architectures
we support sizeof(int) == sizeof(uint32_t) there's
no change in the resulting machine code.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
A RedClient can be freed from the main thread following a main channel
disconnection (reds_client_disconnect). This can happen while another
thread is allocating a new channel client for that client.
To prevent the usage of a pointer which can be invalid
take ownership of the pointer.
Note that we don't need this when disconnecting as disconnection is
done synchronously (the dispatch messages are registered with
DISPATCH_ACK).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Trace the number of loops done processing display commands
and the number of loops in which the queue was full.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
If a DisplayChannelClient cannot be instantiated capabilities
are not released correctly.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
display variable already contains the DCC_TO_DC(dcc) value so
reuse it.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Some RedChannelClient data members were marked as int when they only
hold booleans.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
reds_get_n_clients is a single line and is used only by
spice_server_get_num_clients.
The 2 functions have very similar names so inlining
reds_get_n_clients does not make code less readable.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The leak detector we use currently is not enough to detect
some kind of leak in DisplayChannel so manually test.
These tests are enabled only when --enable-extra-checks is passed
to configure.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Mostly of red_channel_destroy calls were preceded by
a call to unregister the channel.
The only exception was the main channel as this channel is
always present and its initialisation is a bit different.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
RedWorker should not handle directly to client but
defer the job to DisplayChannel.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Make easier to understant the value to use in the code.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
From spice_server_port_event API you can send port events to
any char device. Although currently this is used only for "port"
devices implemented in spicevmc.c this will allow to support
such events using different objects.
This will be used for instance for a streaming device which
will be a specific SpicePort implementation.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
When a RedCursorCmd is passed to cursor_channel_process_cmd(), it
constructs a new CursorItem which takes ownership of that command. If
the cursor_cmd->type falls through to the default case of the switch
statement, we will print a warning and return without freeing the
CursorItem (and thus the RedCursorCmd).
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Now the push is done automatically when a PipeItem is added
(cfr commit 5c460de1a3
"worker: push data when clients can receive them"),
forcing a push cause only network fragmentation and is required
only if you are handling data in a loop instead of using the
default loop.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
Under error: 'link' fields are being accessed, so it's
wrong to goto error with link == NULL.
Instead, return immediately.
Found by coverity.
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
If qxl->descriptor.type is QUIC, red_replay_data_chunks_free
frees qxl (data), so no need to free it again at the bottom
of the function.
Found by coverity.
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
It was also possible for a malicious client to set
VDAgentMonitorsConfig::num_of_monitors to a number larger
than the actual size of VDAgentMOnitorsConfig::monitors.
This would lead to buffer overflows, which could allow the guest to
read part of the host memory. This might cause write overflows in the
host as well, but controlling the content of such buffers seems
complicated.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Total message size received from the client was unlimited. There is
a 2kiB size check on individual agent messages, but the MonitorsConfig
message can be split in multiple chunks, and the size of the
non-chunked MonitorsConfig message was never checked. This could easily
lead to memory exhaustion on the host.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
To help avoid stream.c and dcc.c to access display-channel private
structure to get the nth Stream structure pointer.
Signed-off-by: Victor Toso <victortoso@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
As we have a function for that, don't do the math elsewhere.
Signed-off-by: Victor Toso <victortoso@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
When compiling, -Werror=missing-field-initializers is enabled.
However, some gcc versions (like Gentoo 4.9.4 one) fail to see
that all the members of the SpiceBaseInterface struct are
initialized:
test-display-base.c:844:5: error: missing initializer for field
'description' of 'SpiceBaseInterface'
[-Werror=missing-field-initializers] .base.description = "test
spice virtual channel char device",
The solution is to initialize .base member as a structure at once
instead of multiple times per each member.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
As discussed recently the usage of domain for logging has
different issues (they are not filtered and handled coherently)
and are not widely used in the code.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This was not done until now, and it's only going to be needed if we receive
a partial ClientMonitorsConfig message.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The function name is always prepended by the spice_log macro, so we
don't need to explicitly add it in debug messages.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
RedsClientMonitorsConfig duplicates what SpiceBuffer does,
so using we can replace it with SpiceBuffer and make
reds_on_main_agent_monitors_config() simpler.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This is causing issues with potential improvements to the logging
system, and I've always found this usage a bit odd anyway.
Using spice_debug(""); was not possible as this triggers
-Wformat-zero-length warnings from our use of -Wall.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
In the past CursorItem structure was stored in
RedCursorCmd::device_data field so the check was there to check if the
structure fit into that field.
Since 2ba69f9f88
("libspice: add surface 0 support") the structure is no more stored in
the field so there's no reason for this check causing only confusion.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
The maximum value for a 32 bit variable is UINT32_MAX and not
UINT_MAX. Currently all supported platforms have these two
constants having the same value so this patch don't change
nothing in the generated code but potentially this could
change in a future.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
File transfer and Copy & Paste can be disabled on the server even when
they're supported by the guest agent. Tell it the client by adjusting
the agent capabilities.
Related: rhbz#1373725
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
When a guest is rebooted, the QXL driver gets unloaded at some point in
the reboot process. When the driver is unloaded, the spice server sets a
single flag to FALSE: RedWorker::driver_cap_monitors_config. This flag
indicates whether the driver is capable of sending its own monitors config
messages to the client.
The only place this flag is used is when a new primary surface is created. If
this flag is true, the server assumes that the driver will send its own
monitors config very soon after the surface is created. If it's false, the
server directly sends its own temporary monitors config message to the client
based on the size of the just-created primary surface. This temporary monitors
config message always has a maximum monitor count of 1.
This flag is set to false at startup so that in early boot (e.g. vga text
mode), the server will send out these 'temporary' monitor config messages to
the client whenever the primary surface is destroyed and re-created. This
causes the client to resize its window to match the size of the guest. When the
QXL driver is loaded and starts taking over the monitors config
responsibilities, we set this flag to true and the server stops sending
monitors config messages out on its own.
If we reboot and set this flag to false, it will result in the server sending a
monitors config message to the client indicating that the guest now supports a
maximum of 1 monitor. If the guest happens to have more than one display window
open, it will destroy those extra windows because they exceed the maximum
allowed number of monitors. This means that if you reboot a guest with 4
monitors, after reboot it will only have 1 monitor.
To avoid this, we assume that if we had the ability to support multiple
monitors at some point, that will return at some point. So when the server
constructs its own monitors config message to send to the client (when the
driver_cap_monitors_config flag is false), we send the previous maximum monitor
count instead of always sending a maximum of 1.
Resolves: rhbz#1274447
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The variable is used to store a boolean type.
Update style for this boolean variable.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
Use a #define rather than a magic number to make the code a bit more
readable.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
rcc is already deferenced in red_channel_client_get_client so
checking for NULL after that is uselss.
Also this call is generated from red_channel_client_disconnect
which requires the rcc pointer to be valid.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The base RedChannel already keeps a list of channel clients, so there's
no need for the SndChannel to also keep track of the client itself.
Since the SndChannel only supports a single client (whereas other
channels may have some partial support for multiple clients), I've
provided a convenience function for getting the client and warning if
there is more than one.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The content of these functions almost exclusively deals with channel
client functionality except one line where the channel's active state is
set to TRUE.
These functions are called in two different places.
The first place is from the public API spice_server_record_start() and
spice_server_playback_start(). These functions should alter the
channel's active state, and then set the associated channel client to
active.
The second place is when a new channel client is created. In this
case, it is only called if the channel is already active, so it doesn't
make much sense to set the channel's active state inside of the
function.
To simplify things (and enable some future refactoring), this function
now only deals with the SndChannelClient. The functions have also been
renamed to reflect this fact. The SndChannel's active state is now only
modified from the public API functions.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Instead of putting a 'next' link within the channel structure itself,
just use a generic GList structure to keep a list of active sound
channels.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
When a new PlaybackChannelClient or RecordChannelClient is created,
there are several places where we make decisions based on whether the
client is active or not. But these checks are done before the 'active'
flag is ever set, so this code is effectively dead. This has been the
case since commit 6fdcb931 and d351bb35, so this code has not been
executed for approximately 7 years now.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
It is only called from the constructor, so move all of the code into
that function.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This function is only called from the constructor, so move all of that
code into the constructor.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Reverse return values of the various bool methods so that 'true' means
success, and 'false' failure rather than the opposite.
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The red_get_* methods in red-parse-qxl.c return a boolean, even though
their return type is an int, and they return 1/0. This commit changes
this to the more explicit bool/true/false.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This use to be an anonymous enum, used as an int in
RedCharDeviceWriteBufferPrivate::origin.
Introducing a typedef clarifies what kind of value it's holding.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
This is intended to hold the fields that only char-device.c has a use
for, but for now this only adds the boilerplate for it, the next commit
will move the relevant field there.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
red_channel_client_clear_sent_item() will clear send_data.blocked, so no
need to do it in red_channel_client_msg_sent() as well.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Now that red_channel_client_set_blocked() is no longer changing the
events we are watching for, we can call it from
red_channel_client_push().
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Since 5c460d, we need to watch for WATCH_EVENT_WRITE as long as there are
items queued waiting to be sent, this does not need to be done only when
the network queue is full.
When red_channel_client_set_blocked() is called, as a message is being
sent, WATCH_EVENT_WRITE will already be set, so it does not need to set
it again.
red_channel_client_msg_sent() removes WATCH_EVENT_WRITE, but this will
be done later anyway by red_channel_client_push() if needed.
Since it's redundant, we can remove this.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
red_channel_client_msg_sent() calls red_channel_client_clear_sent_item()
which will clear the rcc->priv->send_data.blocked flag. Later on, it
contains a check for !red_channel_client_is_blocked(), which will always
be true as nothing in between could have set it again. This commit
removes this unneeded check.
This check was already redundant when it was introduced in
9a62a9a809
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Before this commit, red_channel_client_start_ping_timer() and
red_channel_client_cancel_ping_timer() had checks to be no-ops when
rcc->priv->latency_monitor.timer is NULL.
This commit adds a similar check to
red_channel_client_restart_ping_timer(), and removes explicit NULL
checks before calls to red_channel_client_{cancel,restart,start}_ping_timer()
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
rcc is just used on the next line so cannot be NULL.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Many function already used bool as boolean type.
Change last gboolean occurrencies and values.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
Commit 590acf3c55 introduced a regression after migration:
PlaybackChannel::connection is never set even if a client is connected,
which means the playback channel is non-functional until a client
reconnects as all the snd_channel_set_xxx() method checks for a non-NULL
PlaybackChannel::connection before sending data to the client.
This commit slightly changes the code flow in
on_new_playback_channel_client()/playback_channel_client_constructed()
so that PlaybackChannel::connection is set regardless of what
red_client_during_migrate_at_target() returns. This is what was done
prior to 590acf3c55.
This resolves https://bugs.freedesktop.org/show_bug.cgi?id=100136
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This is a particularly opaque part of the code for managing pending
Drawable operations. This patch adds documentation atempting to explain
these functions.
It can be useful for debug to know what the codec of the playback and
record channels are, so add a debug-level print statement indicating
this.
Resolves: rhbz#1436251
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Main-channel-client.c has code to send pings when the statistics code is
enabled. This is done through a timer calling a ping_timer_cb callback.
However, this timer is created but never started, so this code is
actually dead-code.
Since this code does not seem to be doing much apart from sending the
pings (which can be achieved by simply setting monitor-latency to TRUE
in main_channel_client_create()), this commit removes it.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Their uint32_t value is never used, all that matters is whether we
received data or not.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
in main_channel_push_agent_connected(), it used the convenience function
red_channel_test_remote_cap() to determine whether to send the
AGENT_CONNECTED_TOKENS message, or the AGENT_CONNECTED message. However,
this function returns false if *any* one client doesn't support the
capability. We should instead check each individual client separately to
see if they support the capability.
This was a convenience function around
red_channel_client_test_remote_common_cap() that simply iterated over
each client (currently we only really support a single client anyway)
and returned TRUE only if all clients supported the capability. But
nobody ever called this wrapper function.
Both _test_remote_cap() and _test_remote_common_cap() are used as
boolean values, so change the return type from int to bool to follow our
coding standard.
This is a setting for determining whether to compress the audio playback
channel or not. It is variously typed as int or uint32_t. Convert it to
a 'bool' to make it more clear that it is a true/false value rather than
an enumeration or something like that.
The DUMP_BITMAP compile option enable some debugging code to
output image bitmaps in a subdirectory of /tmp.
However if this directory does not exists the server will crash
as not able to create a file in it.
Try to create directory before creating the file.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
- glz_enc_dictionary_reset
- monitors_config_new
- red_channel_any_blocked
- red_channel_no_item_being_sent
- red_client_get_channel
are only used in the file where they are defined, so they can as well be
static.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This is only available in newer FreeBSD releases (9.1 and later), and
will cause build errors or older versions
This fixes https://bugs.freedesktop.org/show_bug.cgi?id=99213
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This new function removes one place outside of RedsStream which needs to
access RedsStream::socket
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This allows to move some low-level code out of reds.c
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This allows to move some low-level code out of reds.c
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The code to enable/disable on a TCP socket is duplicated in multiple
places in the code base, this commit replaces this duplicated code with
a helper in RedsStream.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Doing a memset(0) on a SpiceStatNode does not create an empty/unattached
node, but instead creates a node '0'. This node could be non-existing,
or totally unrelated.
This patch creates a default '0' node as soon as the file is created.
This will make the behaviour of 0-filled nodes more in line with what
one would expect.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This changes the header guards in all .h files to follow this format:
/*
* Licensing block
*/
/* ... */
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This is the default value, and the property is marked as
_CONSTRUCT_ONLY, so we don't need to explicitly force the default when
instantiating a RedChannelClient.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
InputsChannelClient::new and SmartcardChannelClient::new both accept a
"monitor_latency" argument, which is always FALSE. It can be removed.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Before GLib 2.36 you should call g_type_init before attempting
any GLib type usage.
As constructor function are called before even main we need
to call g_type_init much before do_spice_init.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Stress a bit video encoders.
This check different combination of
- encoder type;
- image formats;
- image clipping (encoding partial frames);
- handling frames split into chunks.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Handle single chunk writev as normal write.
From some test more than 60% of the times writev is called with 1 as
counter. We can easily and very cheaply turn this call to a simpler
write avoiding the need to pass the array to the kernel.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
As pointed out by danpb, tests/pki/server-cert.pem is only valid until
Fri Nov 29 15:51:44 UTC 2019
This changes tests/pki/server-cert.pem and tests/pki/ca-cert.pem to be
valid until 2048. These certificates were generated using the
instructions on https://www.spice-space.org/page/SSLConnection
The -subj args were omitted, and the defaults suggested by openssl used.
The -days parameter was changed to -days 10950.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This allows to remove a small hack in server/Makefile.am where we were
using make check-valgrind-memcheck rather than make check-valgrind to
make sure we get a non-0 exit code on failures.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This allows to run automatically our test-suite with valgrind to test
for memory leaks.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
test-qxl-parsing is really a series of several tests. Porting it to
GTest makes this more obvious. This also has the side-effect of making
it more friendly to 'make check-valgrind' (which would fail on SIGALRM,
and on unexpected g_warning()).
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
On my 64 bit Fedora 25, CMSG_SPACE() adds 4 bytes of padding after the
file descriptor in the control data. This causes warnings when ran under
valgrind as we set msg_controllen to CMSG_SPACE().
This commit fills the control data to 0 to avoid these warnings.
==30301== Syscall param sendmsg(msg.msg_control) points to uninitialised byte(s)
==30301== at 0x8127367: sendmsg (sendmsg.c:28)
==30301== by 0x41880B: reds_stream_send_msgfd (reds-stream.c:295)
==30301== by 0x40953F: main (test-stream.c:121)
==30301== Address 0xffefff1b4 is on thread 1's stack
==30301== in frame #1, created by reds_stream_send_msgfd (reds-stream.c:263)
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This reverts commit c3d237075b.
When you call gst_buffer_add_video_meta_full GStreamer assumes that
buffer is contiguous. Specifically GStreamer assumes that the first
chunk is big enough to hold a whole frame. This results usually in
some pixel shifts in the video. The pixel shifts you can see are
artifacts due to how the guest sends the frames.
Assuming you allocate the 2 chunks with 2 malloc:
p1 = malloc(size);
...
p2 = malloc(size);
Usually the memory allocator tend to allocate linearly if there are
space adding a prefix to describe next block leading to a memory
arrangement as:
+-------------------+
| p1 prefix |
+-------------------+
| p1 buffer |
+-------------------+
| p2 prefix |
+-------------------+
| p2 buffer |
+-------------------+
now if you take p1 pointer and you assume it points to a 2 * size
buffer you will get p2 prefix in the middle, this prefix is the pixel
shifts.
Problems happens specifically in gst_video_frame_map_id.
This bug is reported in https://bugzilla.gnome.org/show_bug.cgi?id=779524.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
A few dispose/finalize implementations were not chaining up to their
parent classes, causing memory leaks.
This fixes leaks like the following:
==16240== Memcheck, a memory error detector
==16240== Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al.
==16240== Using Valgrind-3.12.0 and LibVEX; rerun with -h for copyright info
==16240== Command: ./test-leaks
==16240==
==16240==
==16240== HEAP SUMMARY:
==16240== in use at exit: 245,490 bytes in 3,638 blocks
==16240== total heap usage: 5,418 allocs, 1,780 frees, 17,676,968 bytes allocated
==16240==
==16240== 16 bytes in 1 blocks are definitely lost in loss record 203 of 771
==16240== at 0x4C2DADE: malloc (vg_replace_malloc.c:298)
==16240== by 0x4C2FC91: realloc (vg_replace_malloc.c:785)
==16240== by 0x52864E: dispatcher_register_handler (dispatcher.c:374)
==16240== by 0x5293E0: main_dispatcher_constructed (main-dispatcher.c:315)
==16240== by 0x7F873DB: g_object_new_internal (gobject.c:1823)
==16240== by 0x7F87EE4: g_object_new_valist (gobject.c:2042)
==16240== by 0x7F86E90: g_object_new (gobject.c:1626)
==16240== by 0x5292A5: main_dispatcher_new (main-dispatcher.c:295)
==16240== by 0x429A0A: do_spice_init (reds.c:3416)
==16240== by 0x42A3F5: spice_server_init (reds.c:3663)
==16240== by 0x4095B1: server_leaks (test-leaks.c:45)
==16240== by 0x844C60A: test_case_run (gtestutils.c:2161)
==16240== by 0x844C9CA: g_test_run_suite_internal (gtestutils.c:2244)
==16240== by 0x844CA73: g_test_run_suite_internal (gtestutils.c:2256)
==16240== by 0x844CC8A: g_test_run_suite (gtestutils.c:2332)
==16240== by 0x844B92C: g_test_run (gtestutils.c:1599)
==16240== by 0x409A0B: main (test-leaks.c:126)
==16240==
==16240== 16 bytes in 1 blocks are definitely lost in loss record 204 of 771
==16240== at 0x4C2DADE: malloc (vg_replace_malloc.c:298)
==16240== by 0x4C2FC91: realloc (vg_replace_malloc.c:785)
==16240== by 0x52864E: dispatcher_register_handler (dispatcher.c:374)
==16240== by 0x5293E0: main_dispatcher_constructed (main-dispatcher.c:315)
==16240== by 0x7F873DB: g_object_new_internal (gobject.c:1823)
==16240== by 0x7F87EE4: g_object_new_valist (gobject.c:2042)
==16240== by 0x7F86E90: g_object_new (gobject.c:1626)
==16240== by 0x5292A5: main_dispatcher_new (main-dispatcher.c:295)
==16240== by 0x429A0A: do_spice_init (reds.c:3416)
==16240== by 0x42A3F5: spice_server_init (reds.c:3663)
==16240== by 0x40BFD4: test_new (test-display-base.c:902)
==16240== by 0x40979D: vmc_leaks (test-leaks.c:92)
==16240== by 0x844C60A: test_case_run (gtestutils.c:2161)
==16240== by 0x844C9CA: g_test_run_suite_internal (gtestutils.c:2244)
==16240== by 0x844CA73: g_test_run_suite_internal (gtestutils.c:2256)
==16240== by 0x844CC8A: g_test_run_suite (gtestutils.c:2332)
==16240== by 0x844B92C: g_test_run (gtestutils.c:1599)
==16240== by 0x409A0B: main (test-leaks.c:126)
==16240==
==16240== 96 bytes in 1 blocks are definitely lost in loss record 638 of 771
==16240== at 0x4C2FA50: calloc (vg_replace_malloc.c:711)
==16240== by 0x8427D3C: g_malloc0 (gmem.c:124)
==16240== by 0x842801F: g_malloc0_n (gmem.c:355)
==16240== by 0x527B44: dispatcher_constructed (dispatcher.c:141)
==16240== by 0x529321: main_dispatcher_constructed (main-dispatcher.c:307)
==16240== by 0x7F873DB: g_object_new_internal (gobject.c:1823)
==16240== by 0x7F87EE4: g_object_new_valist (gobject.c:2042)
==16240== by 0x7F86E90: g_object_new (gobject.c:1626)
==16240== by 0x5292A5: main_dispatcher_new (main-dispatcher.c:295)
==16240== by 0x429A0A: do_spice_init (reds.c:3416)
==16240== by 0x42A3F5: spice_server_init (reds.c:3663)
==16240== by 0x4095B1: server_leaks (test-leaks.c:45)
==16240== by 0x844C60A: test_case_run (gtestutils.c:2161)
==16240== by 0x844C9CA: g_test_run_suite_internal (gtestutils.c:2244)
==16240== by 0x844CA73: g_test_run_suite_internal (gtestutils.c:2256)
==16240== by 0x844CC8A: g_test_run_suite (gtestutils.c:2332)
==16240== by 0x844B92C: g_test_run (gtestutils.c:1599)
==16240== by 0x409A0B: main (test-leaks.c:126)
==16240==
==16240== 96 bytes in 1 blocks are definitely lost in loss record 639 of 771
==16240== at 0x4C2FA50: calloc (vg_replace_malloc.c:711)
==16240== by 0x8427D3C: g_malloc0 (gmem.c:124)
==16240== by 0x842801F: g_malloc0_n (gmem.c:355)
==16240== by 0x527B44: dispatcher_constructed (dispatcher.c:141)
==16240== by 0x529321: main_dispatcher_constructed (main-dispatcher.c:307)
==16240== by 0x7F873DB: g_object_new_internal (gobject.c:1823)
==16240== by 0x7F87EE4: g_object_new_valist (gobject.c:2042)
==16240== by 0x7F86E90: g_object_new (gobject.c:1626)
==16240== by 0x5292A5: main_dispatcher_new (main-dispatcher.c:295)
==16240== by 0x429A0A: do_spice_init (reds.c:3416)
==16240== by 0x42A3F5: spice_server_init (reds.c:3663)
==16240== by 0x40BFD4: test_new (test-display-base.c:902)
==16240== by 0x40979D: vmc_leaks (test-leaks.c:92)
==16240== by 0x844C60A: test_case_run (gtestutils.c:2161)
==16240== by 0x844C9CA: g_test_run_suite_internal (gtestutils.c:2244)
==16240== by 0x844CA73: g_test_run_suite_internal (gtestutils.c:2256)
==16240== by 0x844CC8A: g_test_run_suite (gtestutils.c:2332)
==16240== by 0x844B92C: g_test_run (gtestutils.c:1599)
==16240== by 0x409A0B: main (test-leaks.c:126)
==16240==
==16240== LEAK SUMMARY:
==16240== definitely lost: 224 bytes in 4 blocks
==16240== indirectly lost: 0 bytes in 0 blocks
==16240== possibly lost: 0 bytes in 0 blocks
==16240== still reachable: 207,718 bytes in 3,312 blocks
==16240== of which reachable via heuristic:
==16240== newarray : 1,536 bytes in 16 blocks
==16240== suppressed: 34,548 bytes in 302 blocks
==16240== Reachable blocks (those to which a pointer was found) are not shown.
==16240== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==16240==
==16240== For counts of detected and suppressed errors, rerun with: -v
==16240== ERROR SUMMARY: 4 errors from 4 contexts (suppressed: 20 from 20)
FAIL test-leaks (exit status: 1)
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The code that manages pending QXL Drawable operations is fairly complex
and difficult to understand. This is an attempt to start documenting
that code to save time when we have to work on this code in the future.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This commit changes all functions returning TRUE/FALSE from having an
'int' return value to 'bool'.
This way it's obvious that such a function is not going to return
anything else than TRUE or FALSE.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
config_socket is configuring the client stream socket.
As is responsibility of RedChannelClient to handle the stream
it make more sense to have the function in this object.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Verify stuff are freed correctly (like TLS context).
The different PKI file required are generated with
base values (localhost and rsa 1024).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Add and remove some vmc device to check for leaking.
These combination assure that currently implemented type
of devices (webdav, usb and generic) are checked.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The nested if statements could be confusing, no needs for them.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
Send header and reply together.
This potentially save up to 160 bytes on the network which
is a considerable amount considering that the message is
about 50 bytes.
This as sending multiple chunks you can have different framing,
specifically:
- if you use TLS every chunk get encrypted separately
(reds-stream, currently usually 29 bytes for every chunks);
- tcp settings and no delay on socket. More likely with fast
connections or better network cards. The tcp framing is
usually about 80 bytes;
- additional tcp acknowledge (usually 64 bytes).
So 80 + 29 + 64 = 173 bytes.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
These vfuncs are more appropriate in RedChannelClient.
The buffer they allocated are related to the client stream
which is managed directly by RedChannelClient.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This prepare for the next patch.
The network recieve buffer should be per-client rather than per-channel.
The following patch will make this change, but this common base class
will allow the cursor client and the display client to share a common
implementation.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
As the counters are shared there is no reason why not
handling the byte count from RedChannelClient directly.
This remove a dependency and avoid some function calls.
The only visible difference at user level is that the
counters are created when a client connects.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Show messages sent to clients.
This is useful to understand the message number as an high
message number can affects performance and is not easy to
understand the message count from the byte count (which is
available).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Instead of disabling the code use the compatibility functions.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
[0] SPICE_MSGC_DISPLAY_PREFERRED_VIDEO_CODEC_TYPE
This message provides a list of video codecs based on client's order
of preference.
We duplicate the video codecs array from reds.c and sort it using the
order of codecs as reference.
This message will not change an ongoing streaming but it could change
newly created streams depending the rank value of each video codec
that can be set by spice_server_set_video_codecs()
Signed-off-by: Victor Toso <victortoso@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Allows to use recording function for multiple purposes.
This will allow to register multiple screen VM or recording
additional stuff like sound.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
The synchronization code is required to avoid mixing writing
from multiple threads.
Following patches will add this feature.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
For each channel there are two set of capabilities, one
for the common ones and one for the specific ones.
A single set were almost always passed using 2 arguments,
a number of elements and an array but then before using
these were converted to a GArray.
Use a single structure (already available) to pass all
channel capabilites using a single argument.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Add function to initialize and destroy this type.
Add GType type for boxing it.
These changes a in preparation for next patch.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
These properties are not read and code is broken (the content of
the array would be uninitialized).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Watch was added but never removed.
The added basic_event_loop_destroy() addition allows to see that
this leak is gone).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This switches the test to using the GTest API, and add several tests
related to https://bugzilla.redhat.com/show_bug.cgi?id=1411194
This uses some API not available in glib 2.28, so this checks we have a
new enough glib before building this test, and disables warnings when
using too new glib API when building it.
The "multiple-vmc-devices" is based off code written by Frediano Ziglio.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Using spice_info() gets in the way of tests using
g_test_expect_message() as all the messages emitted using
a non-debug log level must be listed as expected, otherwise we get a
critical about an expected message not having been logged.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Currently, the network sockets opened by reds_init_net() are not closed
on destruction, in other words they are leaked.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This allows to chain several test cases by using
test_new()/test_destroy().
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Do not assume the device passed as an argument to
spice_server_char_device_remove_interface() is the same as the current
Reds::vdagent instance. This commit adds a check that this is the case,
and returns early if not.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
spice_new is a bit more safe as return a properly typed pointer.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The usage was removed with commit 7ea1f2c133
("sound: Use RedChannelClient to receive/send data").
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Using X == TRUE is fragile, since the input value is a uint8_t. It would be
surprising if the value was set to 2 or 0xFF and the test failed.
Signed-off-by: Christophe de Dinechin <dinechin@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Some constants/macros/function prototypes are defined in
red-channel-client.h while they are only used by red-channel-client.c.
This commit moves them to the .c file to make them private.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
RedChannelClient is responsible for talking to the client so it knows
if is connected or not.
These vfuncs where used by DummyChannel used by SoundChannel.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Most channels don't need to do specific settings for the client socket
so avoid the need to do this setting making easier to setup the client
channnel.
Some improvements and commit subject suggested by Christophe Fergeau.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
TCP_NODELAY flag is set by default for all connection inside
reds.c so there's no need to set again for the single
client channel.
Note that there are still some calls to setsockopt to change this
option.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Since c3d237 "gstreamer: Avoid memory copy if strides are different" is
only needed when zero copy is disabled. This moves the function
definition to an already existing #ifdef block.
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This is a revert of b76e561d.
For a SpicePlaybackInstance, the base interface must be a
SpicePlaybackInterface instance, not a SpiceBaseInterface instance, or
spice-server code will end up reading out of bounds.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This is a revert of 93b4f4050^ and 93b4f4050.
For a SpiceCharDeviceInstance, the base interface must be a
SpiceCharDeviceInterface instance, not a SpiceBaseInterface instance, or
spice-server code will end up reading out of bounds.
vmc_state/vmc_write/vmc_read implementations also have to be provided.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This is a revert of 93b4f4050^ and 93b4f4050.
For a SpiceCharDeviceInstance, the base interface must be a
SpiceCharDeviceInterface instance, not a SpiceBaseInterface instance, or
spice-server code will end up reading out of bounds.
vmc_state/vmc_write/vmc_read implementations also have to be provided.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Use new structures and functions to implement the statistics code.
Use inline functions instead of macros for increased type-safety.
If statistics are disabled, the structures and functions become
empty. This confines the configuration-specific #defines to the
statistics implementation itself and avoids the need for #defines in
the calling functions. This greatly reduces the chance of accidentally
breaking the build for one configuration or the other. The reds option
was removed from stat_inc_counter() as it was not used.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Non blocking flag is set for all connection inside reds.c so
there's no need to set again for the single client channel.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Do not make it assume vec contains IOV_MAX elements.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
They no longer contain any vfuncs, so calling them "handler" does not
make a lot of sense. This commit renames them to
OutgoingMessageBuffer/IncomingMessageBuffer.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Nothing outside of RedChannelClient needs access to data contained in
RedChannelClientPrivate, so we can move all the type definitions to the
.c file to make it fully opaque rather than relying on a private header.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The methods previously used by OutgoingHandlerInterface and
IncomingHandlerInterface are no longer used as generic callbacks,
but are directly called from RedChannelClient code. We can be explicit
about the type of their first argument (RedChannelClient *) rather than
using a generic void * pointer.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
There is only one implementation of OutgoingHandler which relies
OutgoingHandler::opaque being a RedChannelClient. This commit makes this
explicit in order to get rid of the OutgoingHandler::opaque data member.
This renames red_peer_handle_outgoing() to
red_channel_client_handle_outgoing() as the method is now very much tied
to RedChannelClient.
If we want to keep some genericity, we could return error codes from
red_channel_client_handle_outgoing() and handle RedChannelClient
disconnection/... from the caller rather than directly in the
_handle_outgoing() method. This would probably allow to move the
data reading logic to reds-stream.c
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
There is only one implementation of IncomingHandler which relies
IncomingHandler::opaque to be a RedChannlClient. This commit makes this
explicit. This allows to get rid of the IncomingHandler::opaque data
member.
This renames red_peer_handle_incoming() to
red_channel_client_handle_incoming() as the method is now very much tied
to RedChannelClient.
If we want to keep some genericity, we could return error codes from
red_channel_client_handle_incoming() and handle RedChannelClient
disconnection/... from the caller rather than directly in the
_handle_incoming() method. This would probably allow to move the
data reading logic to reds-stream.c
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This commit removes what remains of IncomingHandlerInterface. The
remaining function pointers were pointing to RedChannel vfuncs.
Moreover the IncomingHandlerInterface abstraction is unused, ie the
codebase only has a single implementation for it, so we can directly
call the relevant methods and make them static instead.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This commit removes IncomingHandlerInterface::on_error and
IncomingHandlerInterface::on_input. As with previous commits, these
vfuncs are always calling the method, and RedChannel::init sets them to
point to RedChannelClient methods, which RedChannelClient is then going
to call indirectly through the IncomingHandlerInterface instance.
This commit changes this to direct calls to the corresponding methods.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Similarly to the previous commits, this removes an indirection level,
IncomingHandlerInterface stores pointers to
alloc_recv_buf/release_recv_buf vfuncs, but these are always set to
RedChannel::alloc_recv_buf/RedChannel::release_recv_buf, which are also
vfuncs which can be overridden if needed. This commit removes the
indirection and directly calls the relevant methods.
These vfuncs really should be in RedChannelClient rather than
RedChannel.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
RedChannel uses OutgoingHandlerInterface to provide constant pointers to
RedChannelClient methods. This OutgoingHandlerInterface structure is
then used in RedChannelClient to indirectly call these methods.
The OutgoingHandlerInterface abstraction is unused, ie the codebase
only has a single implementation for it, so we can directly call the
relevant methods and make them static instead.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <figlio@redhat.com>
Have the RedChannelClient callback call into a RedChannel callback
rather than doing the opposite. This will be useful in some subsequent
refactoring of this code.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
red_channel_client_parse() currently does roughly:
if (klass->parser) {
parsed = klass->parser(msg, msg_size, &parsed_size);
klass->handle_parsed(rcc, parsed_size, msg_type, parsed);
} else {
klass->handle_message(rcc, msg_type, msg, msg_size);
}
The handle_parsed implementation expects a void * 'parsed' argument,
which will then be cast to the correct type. There is not really a need
to provide distinct handle_parsed/handle_message vfuncs, instead we can
say that if a RedChannel subclass provides a 'parser' vfunc, then it's
'handle_message' vfunc will be called with the parsed message, otherwise
it will be called with unparsed data (ie what handle_message currently
expects).
This makes the code slightly easier to follow as messages will always be
handled by the same vfunc.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
smartcard_channel_client_pipe_add_push was just calling
red_channel_client_pipe_add_push without any cast or other
changes.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
This makes the code slightly more readable as this means less local
variables to keep track of when taking a high level view of that code.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
read_binary() attaches 'mem' to the SpiceReplay::allocated list.
On failure, SpiceReplay::allocated and its content are freed by
spice_replay_free().
SpiceReplay::primary_mem is also freed, which causes a double free
as replay_handle_create_primary() added 'mem' both to
SpiceReplay::primary_mem and SpiceReplay::allocated.
This commit avoids this by ensuring SpiceReplay::primary_mem is not
kept in the SpiceReplay::allocated list.
Note that this double free can happen only on currupted or wrong
record images.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The Surface and Display channels each have a 'current_list' Ring, and
Surface also has a 'current' Ring. these names are confusing, so at
minimum, add a comment indicating the type of object they hold. The
DisplayChannel::current_list already had a comment, but it was
incorrect.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Caller is supposed the function return a buffer able to store
size bytes.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The limits for capabilities are specified using 32 bit unsigned integers.
This could cause possible integer overflows causing buffer overflows.
For instance the sum of num_common_caps and num_caps can be 0 avoiding
additional checks.
As the link message is now capped to 4096 and the capabilities are
contained in the link message limit the capabilities to 1024
(capabilities are expressed in number of uint32_t items).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The limit for link message is specified using a 32 bit unsigned integer.
This could cause possible DoS due to excessive memory allocations and
some possible crashes.
For instance a value >= 2^31 causes a spice_assert to be triggered in
async_read_handler (reds-stream.c) due to an integer overflow at this
line:
int n = async->end - async->now;
This could be easily triggered with a program like
#!/usr/bin/env python
import socket
import time
from struct import pack
server = '127.0.0.1'
port = 5900
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect((server, port))
data = pack('<4sIII', 'REDQ', 2, 2, 0xaaaaaaaa)
s.send(data)
time.sleep(1)
without requiring any authentication (the same can be done
with TLS).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
red_channel_client_handle_message can handle base messages
so reuse it.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Some gcc version reports this error:
stat-file.c: In function 'stat_file_add_node':
stat-file.c:180:15: error: 'node' may be used uninitialized in this function
[-Werror=maybe-uninitialized]
g_strlcpy(node->name, name, sizeof(node->name));
^~~~
cc1: all warnings being treated as errors
This warning is a false positive as this loop:
for (ref = 0; ref <= stat_file->max_nodes; ref++) {
node = &stat_file->stat->nodes[ref];
...
}
will always iterate at least once.
This patch rewrite the loop in order to make more compilers
understand that node variable is always initialized.
There are two checks apparently removed in the patch:
- check for stat_file->stat not being NULL. This was worthless as the
field was already used in the function. Also this field is never
NULL (unless memory corruption happened);
- stat_file->stat->num_of_nodes >= stat_file->max_nodes. It's implicit
in the loop. If num_of_nodes >= max_nodes means that there are no
free nodes so all nodes should have SPICE_STAT_NODE_FLAG_ENABLED set,
loop will exit and function will return INVALID_STAT_REF.
Reported-by: Christophe Fergeau <cfergeau@redhat.com>
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
The stat file contains an array of max_nodes elements
so we must stay in [0, max_nodes) range, not [0, max_nodes].
There are no spice path that lead to these overflows but
it's better to have them fixed before creating one.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Adding conditional for having gstreamer_0_10 or gstreamer_1_0,
removing the previous conditionals and update relevant ifdefs
with the newly defined conditional
We support only a single client so don't waste code just
to check this.
The worst stuff can happen is that we'll migrate multiple
connections.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Move the freeing of SndChannel data members from snd_detach_common() to
the finalize function to encapsulate things a bit more cleanly. It
doesn't really change the behavior or order of destruction since
snd_detach_common() destroys the channel.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Sound implementation used internal RedChannelClient data while now
it just uses the public interface not thouching RedChannelClient
internal state so now is possible to make this field private.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Now that SndChannelClient has switched from using its own code for
sending data to using RedChannelClient, it's very close to being an
actual RedChannelClient.
This commit makes it directly inherit from RedChannelClient rather than
having a channel_client field. This allows to get rid of the whole
DummyChannel/DummyChannelClient code.
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
snd_set_command()/snd_send() are higher level methods which take care of
scheduling calls to the corresponding snd_*_send_*() methods when
appropriate. This commit switches a few direct snd_*_send_*() calls to
snd_set_command()/snd_send().
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
You can see that SndChannelClient has much less field
as the code to read/write from/to client is reused from
RedChannelClient instead of creating a fake RedChannelClient
just to make the system happy.
One of the different between the old sound code and all other
RedChannelClient objects was that the sound channel don't use
a queue while RedChannelClient use RedPipeItem object. This was
the main reason why RedChannelClient was not used. To implement
the old behaviour a "persistent_pipe_item" is used. This RedPipeItem
will be queued to RedChannelClient (only one!) so signal code we
have data to send. The {playback,record}_channel_send_item will
then send the messages to the client using RedChannelClient functions.
For this reason snd_reset_send_data is replaced by a call to
red_channel_client_init_send_data and snd_begin_send_message is
replaced by red_channel_client_begin_send_message.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
The removed code was trying to read data when
spice_server_record_get_samples() is called. Since reading of data is
event-driven anyway (see snd_event), it's redundant to try
again to read more data.
This commit removes this code as this will some refactoring easier in
the next commits.
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
We can get it from our DummyChannelClient rather than storing it in
SndChannelClient.
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
We can use the marshaller provided by the dummy RedChannelClient
associated with SndChannelClient.
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Filter out commands which should not happen. Should it be a
g_warn_if_fail() or such instead?
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This is in preparation for switching SndChannelClient into a proper
RedChannelClient. The prototype of the new helper matches what is
expected from the RedChannel::config_socket vfunc.
To be able to achieve that, this commit associates the sound channel
RedsStream instance with the DummyChannelClient instance we have, and
then call snd_channel_config_socket() on that instance.
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The main goal of this commit is to avoid to dereference 'client' before
it's checked for NULL. This meant splitting one error condition in 2
separate ones.
This also sets default values for the 'frame' and 'num-samples'
out parameters so that we just have to return on error conditions.
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This structure is used to store format information for
both Gstreamer 0.10 and 1.0 however the two format uses
different fields from it.
Use a macro to filter only needed fields.
This currently also fixes a compile error using Gstreamer 0.10
(GST_VIDEO_FORMAT_RGB15 not defined as not available).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
The structure is used only to allocate private data.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Currently, calling spice_server_remove_interface() twice in a row with
the same SPICE_INTERFACE_CHAR_DEVICE is going to cause a crash when
calling red_char_device_get_server(char_device->st); because
char_device->st will have been set to NULL by the first call.
This commit adds a few sanity checks before trying to use the various
'st' members of the interfaces.
This should avoid the crash described in
https://bugzilla.redhat.com/show_bug.cgi?id=1411194 even though it's not
clear how we got in that situation.
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Avoid to use g_object_get if not necessary.
red_char_device_get_server is more type safe and we are
not bound to dynamic fields.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Uri Lublin <uril@redhat.com>
This field is used only for debugging.
Remove it reducing a bit all these "current" fields around.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
If bitmap stride and stream stride are different copy was used.
Using GStreamer 1.0 you can avoid the copy setting correctly
image offset and stride.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Handling read returning 0 (usually end of connection/pipe)
is the same of handling an error (read result -1) with errno == 0
so merge the two paths to reuse code and simplify.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This allows the connection to early fail in case initial bytes
are not correct.
This allows for instance VNC client to graceful fail connecting
to a spice-server. This happens easily as the two protocols
share the same range of ports.
This resolves rhbz#1416692.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Tested-by: Daniel P. Berrange <berrange@redhat.com>
Acked-by: Uri Lublin <uril@redhat.com>
Code to read and process display commands were the same
so use a common function for better reuse.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This format is required to add metadata to the source buffer
using gst_buffer_add_video_meta_full.
This metadata can be used to pass strides/offsets, or
dmabuf-specific information.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This happened during VM resume.
RedSurfaceCmd were allocated but never freed.
We don't need to malloc the RedSurfaceCmd used in handle_dev_close()
as display_channel_process_surface_cmd() will not try to reference
it after it has returned.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Code to read and process cursor commands were the same
so use a common function for better reuse.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This allows to do some possible statistics or graph.
Currently the report contains encoded sizes.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Currently MUTE and VOLUME commands use the same VOLUME mask. This commit
introduces a separate SND_MUTE_MASK for MUTE commands to make things
a bit more clear.
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This is no longer used since "sound: Convert SndChannel to GObject"
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
It's already defined before in the same source file.
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Based on a patch from Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
It became unused in 26027036c 'red_channel: remove unused migrate flag
from RedChannel' but was never removed from the function prototype.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This became obsolete when RedChannel became GObject-based.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
These formats were tested manually using test-gst utility
in server/tests.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Add an utility to make possible to check various features of
VideoEncoder.
2 GStreamer plugins are used in a chain like this:
(1) input pipeline -> (2) video encoder -> (3) output pipeline
While converting output from (1) is compared with output of (3)
making sure the streaming is working correctly.
You can set various options:
- part of the input pipeline description to allow specifying different
video from GStreamer test ones to a video file;
- the encoder to use;
- different image properties to use for (2) input:
- different bit depth;
- top/down or down/up;
- initial bitrate.
The idea is to use this helper in combination with a shell script
and some video sources to make able to test various settings.
Also can be used to extend the current encoder list.
As an example you can use a command like
$ ./test-gst -e gstreamer:vp8 -i \
'filesrc location=bbb_sunflower_1080p_30fps_normal.mp4 \
! decodebin ! videoconvert'
to check vp8 encoding.
Currently it does not emulate bandwidth changes as stream reports
from the client are not coded.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Small refactor. As reds_get_video_codecs() returns the video codecs as
GArray, we should match reds_set_video_codecs() to have a GArray as
parameter instead of string.
reds_set_video_codecs_from_string() seems more appropriate for the
previous function.
Signed-off-by: Victor Toso <victortoso@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Previously, the object we now call SndChannel was named SndWorker, and
the object we now call SndChannelClient was called SndChannel. When
these names were changed, the functions
on_new_(record|playback)_channel() were not updated, so the function
names and the arguments are both a bit confusing now. Update them to
match the new names.
Acked-by: Frediano Ziglio <fziglio@redhat.com>
When the initial image was sent to the client the reference
was not incremented leading to some user after free.
This regression was introduced in
3bde2e570c
("DCC: remove more init_send_data() arguments").
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This fixes a regression caused by
a43c21b6bc
("DCC: change how fill_bits() marshalls data by reference").
Before the mentioned patch there were a few references to Drawable
structure so an uint8_t was enough.
Now that every chunk of the image is accounted you can easily
get an overflow.
This fixes https://bugs.freedesktop.org/show_bug.cgi?id=99258.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Fix compilation error due to -Werror=maybe-uninitialized:
CC test-display-base.o
test-display-base.c: In function 'do_wakeup':
test-display-base.c:579:13: error: 'update' may be used uninitialized...
push_command(&update->ext);
Signed-off-by: Snir Sheriber <ssheribe@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This third argument (and the 'item' member of
RedChannelClient::priv::send_data) was a somewhat roundabout way to keep
the RedPipeItem alive until a message is sent, just in case some data
owned by that pipeitem was added to the marshaller by reference. This
was a rather confusing mechanism, however, since it did not have any
obvious connection to the _add_by_ref() call. It was never very clear
whether you needed to pass an item to this function or not. The previous
series of patches made this parameter unnecessary since the referencing
of the pipe item (or other related structure) is now more explicitly
connected to the calls to spice_marshaller_add_by_ref_full().
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The fill_bits() function marshalls some data by reference. This data is
owned by the RedDrawable that is owned by the Drawable that is owned by
the RedDrawablePipeItem. Instead of keeping the RedPipeItem alive by
passing it to red_channel_client_init_send_data(), simply reference the
Drawable and marshall it with _add_by_ref_full(). This means that we
can't use the _add_chunks_by_ref() convenience function since that
function doesn't allow us to pass a free function to clean up the data
after it is sent.
This change is not perfect since the fill_bits() function makes an
assumption that 'simage' is owned by the 'drawable'. On the other hand,
the previous code made a much bigger assumption: that the caller would
ensure that the data would be kept alive
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Use spice_marshaller_add_by_ref_full() instead of _add_by_ref() to
handle the referenced data properly rather than passing the pipe item to
red_channel_client_init_send_data() to keep the pipe item alive
indirectly.
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Use spice_marshaller_add_by_ref_full() instead of
spice_marshaller_add_by_ref() to allow the marshaller to manage the
lifetime of the referenced data buffer rather than having to manage it
by passing a PipeItem to red_channel_client_init_send_data(). Since the
data is owned by CursorItem (which is not in fact a RedPipeItem, but is
owned by a pipe item and is itself refcounted), we take a reference on
the CursorItem when adding the data buf to the marshaller, and then
unreference it in the marshaller free func.
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
The only time that the pipe item needs to be passed as the third
argument to red_channel_client_init_send_data() is when the pipe item
holds a data buffer that has been added to the marshaller by reference
(spice_marshaller_add_by_ref()) and needs to be kept alive until the
data has been sent. In all other cases, the item does not need to be
kept alive, so we can safely pass NULL for this third parameter.
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Any data that is added to the marshaller by reference (using e.g.
spice_marshaller_add_by_ref_full()) is freed during
spice_marshaller_reset(). But the marshaller is not currently reset
until we begin to send the next message (in
red_channel_client_send_item()). This means that the sent message data
lives longer than expected and can violate some assumptions in other
parts of the code.
To make sure that the data is cleaned up right after being sent, I've
added a reset call to clear_sent_item() and called that function from
_on_out_msg_done(). This means that _restore_main_sender() no longer
needs to reset the marshaller, and we no longer need to call
_reset_sent_item() within _on_out_msg_done() (since this function is
called from _clear_sent_item()).
This prepares the way for refactoring
red_channel_client_init_send_data() to change how we keep the
RedPipeItem alive while sending a message.
Acked-by: Frediano Ziglio <fziglio@redhat.com>
We should replace the video_codecs GArray only after the parsing of
input is done, otherwise we might lose previous configuration.
Tests were updated to match this situation. Input that fails to
replace video_codecs are considered bad.
Signed-off-by: Victor Toso <victortoso@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This make more obvious which directory they refer
and potentially avoid ignoring unwanted files.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The partial expression "MSEC_PER_SEC * size * 8" can overflow if
size is 536870 or more. This as the operation is done using
32 bit unsigned integers. Being the size potentially double of
a compressed frame size the limit can be easily reached.
As get_average_frame_size already return a 64 bit use 64 bit
even for the size to avoid the integer overflow.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Francois Gouget <fgouget@codeweavers.com>
This function is supposed to add an item to the queue to
be sent before all other queued items.
Was never used.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
refs was used before GObject for reference counting.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
All RedWorker messages starts with RedWorker except
SpiceMsgDisplayGlDraw.
For coherence introduce a RedWorkerMessageGlDraw structure
holding just SpiceMsgDisplayGlDraw. This also allows possible
extensions.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
The spice_marshaller_add_ref() family of functions is confusing since it
sounds like you're incrementing a reference on the marshaller. What it
is actually doing is adding a data buffer to the marshaller by reference
rather than by value. Changing the function names to _add_by_ref() makes
this clearer.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
spice_debug was called for not-EINTR case, move
it to the right place.
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
man 2 dup2 specifies:
The close-on-exec flag (FD_CLOEXEC; see fcntl(2)) for
the duplicate descriptor is off.
Since the purpose of the fcntl call is to turn off FD_CLOEXEC
flag, and it's already done, just remove this call.
Suggested-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Spice coding style suggests to use curly braces
for while blocks.
Some prefer the block to not be empty so continue
is untouched.
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This patch prevents possible memory leak.
Found by coverity.
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
In both image_encoders_restore_glz_dictionary() and
image_encoders_get_glz_dictionary() shared-dict may
be NULL if size is too large, and the server gets
size from the network.
Both functions end up calling glz_enc_dictionary_create()
that calls glz_dictionary_window_create() where size is
checked.
Found by coverity.
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
It shadows the outer one.
Renamed also the outer 'container' variable.
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Allow to catch minor issue with test code.
Although test usually are coded with less care than production
code the current changes required are not so extensive and
can catch different issues.
Most of the patch is making functions static to avoid warnings for
undeclared functions.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This patch creates display_channel_get_video_codecs() helper to let
stream.c get the video-codecs without accessing the internal structure
of DisplayChannel.
As video-codecs is a property of DisplayChannel, this change means
making this property readable as well.
Signed-off-by: Victor Toso <victortoso@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Allow to dynamically remove QXL interfaces. This could be used to
support hot swapping of QXL cards.
This code is actually not used in any way.
QXL interfaces are destroyed by spice_server_destroy automatically.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Make finalization of DisplayChannel consistent with other code.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Reduce GObject changes coming in the next commit since we'll need to
change how we access the marshaller anyway. This will make the
following commits easier to review.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
As data is packae in a single piece of memory send it
altogether.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Including g_clear_pointer() in glib-compat.h and using it to avoid
warnings in odd situations.
Signed-off-by: Victor Toso <victortoso@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
When qemu (for example) delivers audio samples to the spice server, it
does so by requesting a buffer from the spice server
(spice_server_playback_get_buffer()), filling them with audio data, and
then calling spice_server_playback_put_samples() to send them to the
client. Between the call to _get_buffer() and the call to
_put_samples(), we need to ensure that the buffer remains valid. Since
this buffer is allocated within PlaybackChannelClient, we did this by
incrementing a ref on the PlaybackChannelClient in _get_buffer(), and
decrementing the ref in _put_samples(). This has the effect of
potentially keeping the PlaybackChannelClient alive after the spice
client has disconnected.
This was not a problem when PlaybackChannelClient was a simple helper
class. But we intend to change PlaybackChannelClient so that it
inherits from RedChannelClient. In that case, the reference taken in
_get_buffer() would result in the RedChannelClient (and associated
RedChannel, etc) being kept alive longer than expected. To avoid this,
we add an additional ref-counted adapter class (AudioFrameContainer)
that owns the allocated audio frames and can outlive the
RedChannelClient if necessary. When the client is freed, the
AudioFrameContainer is just unreferenced.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Stops using the dummy channel.
Data handling still goes through DummyChannelClient which is why
empty implementation of some required vfuncs is working.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Try to arrange destruction in the opposite order of the creation
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
The check to limit too low bit rates was setting encoder->bit_rate
instead of bit_rate. However after some lines bit_rate was used
to set encoder->bit_rate basically removing the lower threshold.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Francois Gouget <fgouget@codeweavers.com>
By removing the stream's video encoder we force the stream to send
future frames using the fallback code, that is as regular screen
updates.
However note that we keep the stream object: we have to. Otherwise
future frames would trigger the creation of a new stream object with a
new video encoder which would again try to stream the video and fail
again and again.
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Reviewed-by: Victor Toso <victortoso@redhat.com>
Make easier to understand that they refer to client and not
all channel.
Specifically:
- RecordChannel -> RecordChannelClient
- PlaybackChannel -> PlaybackChannelClient
- playback_channel -> playback_client
- record_channel -> record_client
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
SndWorker has been historically based on RedChannel, initial git commit
has:
struct SndWorker {
Channel base;
...
};
SndChannel, contrary to what its name may imply is more focused on
marshalling/sending of sound data, which is the responsibility of
RedChannelClient for the other SPICE channels.
This commit and following ones make the naming of these 2 types more
consistent with how the rest of the code is structured.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
SpicePlaybackState and SpiceRecordState have same structures
changing only slightly the behaviour.
Using SndWorker instead allows some minor simplification and
more code reuse.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
Most of the times the check is done externally
so moving inside the function reduce the code.
This is similar to the way free(3) works.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This function assume there is only one client.
Was used only by some obsolete functions.
Avoid to use such function in the future.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This function wrongly close the first client.
Wrongly as closing the file descriptor cause a dandling
file descriptor in the object potentially leading
to closing another file descriptor open later.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This will close all clients and release the channel properly
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Now that RedStatFile is private there is no need
to include some headers in stat-file.h.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
These functions are not used since years and are not supporting
multiple clients.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Use same constants for common commands.
This will allow code reuse between Record and Playback.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Uri Lublin <uril@redhat.com>
Allows to close worker thread.
This will be used to destroy cleanly CursorChannel and
DisplayChannel.
CursorChannel and DisplayChannel are run in a different
thread. However deregistration of channels and different
steps of destruction should be done in the same thread
so this make possible to join again the 2 threads to
avoid race conditions.
For the moment there is no correct cleanup.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This make compression faster and avoids a warning on newer
lz4 versions.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
Use dash instead of underscores for file names. This is coherent
with rest of file names.
Rename all tests to test-XXX and remove the spice- prefix when
present. This is coherent with most of the tests.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
self is usually used for GObject, avoid confusion.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The video encoders already have sophisticated bit rate control code that
can react to both stream reports sent by the client, and server frame
drop notifications. Furthermore they can adjust both the frame rate and
the image quality to best match the network conditions.
But if the client cannot send stream reports all this is bypassed and
instead the streaming code performs its own primitive bit rate control
that can only adjust the frame rate.
So this patch removes the code duplication and lets the video encoders
do their job.
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The wrong type (RED_TYPE_CHANNEL_CLIENT rather than
TYPE_SMARTCARD_CHANNEL_CLIENT) was used when creating a
SmartcardChannelClient, which meant the _init function was never called,
and SmartcardChannelClient::priv was never initialized.
This caused a crash when trying to connect with --spice-smartcard to a
VM with a smartcard channel configured.
This leaks happen for every connection. Potentially the timer can
be called after the client is closed causing an use after free.
Recently RED_STATISTICS was switched off by default but previous
version have this issue.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Pavel Grunt <pgrunt@redhat.com>
This reverts commit c6881ad1a0.
This patch cause the replay utility run at full speed
to slow down a lot and in some cases getting stuck.
I don't understand the reason and when I tested was working
but as we are going to release would be a pity if this test utility
won't work as useful to get feedback and reports.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The following build error occurs when building outside of the source
directory:
glib-mkenums --template spice-server-enums.tmpl.c ../../server/spice-server.h > spice-server-enums.c
Can't open spice-server-enums.tmpl.c: No such file or directory
Makefile:1111: recipe for target 'spice-server-enums.c' failed
Make/Automake uses VPATH to determine that the spice-server-enums.tmpl.c
file listed in the prerequisites for the rule is located in the srcdir.
When we use an automatic variable (e.g. $<), the full path to the
resolved file is used. But when we use the literal filename directly
within the rule definition, this won't happen. So we need to explicitly
specify that the input template file is located in srcdir.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
