Usually buffers to raw data are passed using void* pointers
to avoid casts and mark the buffer as raw.
Use them for read_safe and write_safe to avoid useless casts
in caller code.
As a minor convert a parameter to bool as changing the same
lines.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Victor Toso <victortoso@redhat.com>
Found with google-readability-casting
https://google.github.io/styleguide/cppguide.html#Casting
Makes the operation clearer.
This commit uses const_cast where needed.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
The leak is detected by Valgrind on Fedora 34as:
==19603== 400 bytes in 1 blocks are possibly lost in loss record 2,296 of 2,441
==19603== at 0x4845464: calloc (vg_replace_malloc.c:1117)
==19603== by 0x40135FB: _dl_allocate_tls (in /usr/lib64/ld-2.33.so)
==19603== by 0x57EB008: pthread_create@@GLIBC_2.2.5 (in /usr/lib64/libpthread-2.33.so)
==19603== by 0x53A1130: UnknownInlinedFun (gthread-posix.c:1323)
==19603== by 0x53A1130: g_thread_new_internal (gthread.c:931)
==19603== by 0x53C4953: g_thread_pool_start_thread.constprop.0 (gthreadpool.c:477)
==19603== by 0x53A2902: g_thread_pool_push (gthreadpool.c:691)
==19603== by 0x519AE11: g_task_run_in_thread_sync (gtask.c:1593)
==19603== by 0x80D8A74: ??? (in /usr/lib64/gio/modules/libgiolibproxy.so)
==19603== by 0x5181966: g_proxy_address_enumerator_next (gproxyaddressenumerator.c:176)
==19603== by 0x519281A: g_socket_client_connect (gsocketclient.c:1098)
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Marc-André Lureau <marcandre.lureau@redhat.com>
At red-parse-qxl.cpp#L535
if (qxl_flags & QXL_BITMAP_DIRECT) {
red->u.bitmap.data = red_get_image_data_flat(slots, group_id,
qxl->bitmap.data,
bitmap_size);
Since qxl->bitmap.data may from the guest, an attacker can make the
memslot_get_virt() check in red_get_image_data_flat() fail and
return a nullptr.
Then at red-parse-qxl.cpp#L550
if (qxl_flags & QXL_BITMAP_UNSTABLE) {
red->u.bitmap.data->flags |= SPICE_CHUNKS_FLAGS_UNSTABLE;
}
qxl_flags is assigned as qxl->bitmap.flags before, which can also be
controlled by the attacker, resulting in a NULL pointer dereference.
This dereference seems to be introduced by commit 5ac88aa7.
Signed-off-by: Qiuhao Li <Qiuhao.Li@outlook.com>
FreeBSD's setsockopt() behaves just like Dawrin, i.e. sets errno to
EINVAL instead of ENOTSUP, so extend the Darwin workaround to work for
FreeBSD as well.
Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
On FreeBSD, netinet/in.h needs to be included to use IPPROTO_TCP.
Signed-off-by: Roman Bogorodskiy <bogorodskiy@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Compiler error for cross builds using mingw-w64:
In file included from /usr/share/mingw-w64/include/winnt.h:150,
from /usr/share/mingw-w64/include/minwindef.h:163,
from /usr/share/mingw-w64/include/windef.h:9,
from /usr/share/mingw-w64/include/windows.h:69,
from /usr/share/mingw-w64/include/winsock2.h:23,
from ../../../server/spice-core.h:29,
from ../../../server/spice.h:24,
from ../../../server/spice-wrapped.h:35,
from ../../../server/red-common.h:35,
from ../../../server/jpeg-encoder.c:22:
/usr/share/mingw-w64/include/basetsd.h:31:22: error: conflicting types for ‘INT32’
typedef signed int INT32,*PINT32;
^~~~~
In file included from /usr/x86_64-w64-mingw32/sys-root/mingw/include/jpeglib.h:31,
from ../../../server/jpeg-encoder.c:20:
/usr/x86_64-w64-mingw32/sys-root/mingw/include/jmorecfg.h:179:14: note: previous declaration of ‘INT32’ was here
typedef long INT32;
^~~~~
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Found with performance-move-const-arg
Allows better optimization as the compiler does not have to deal with an
rvalue reference. Especially in C++17 where std::move can prevent copy
elision.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Found with modernize-use-nullptr
NULL in C++ is 0 whereas it is a void pointer in C. Avoids implicit
conversions.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Found with performance-for-range-copy
Avoids unnecessary copying when the loop does not modify the variable.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Found with readability-container-size-empty
This has the potential for extra performance as it's not checking for
every single element.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Found with performance-for-range-copy
Avoids unnecessary copying when the loop does not modify the variable.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Found with modernize-use-equals-default
default allows extra optimization compared to an empty con/destructor.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Found with modernize-use-override
This can be useful as compilers can generate a compile time error when:
The base class implementation function signature changes.
The user has not created the override with the correct
signature.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
Fix this error reported by some older Gnu C++ compilers:
./server/tests/test-display-base.cpp:818:1: sorry, unimplemented: non-trivial designated initializers not supported
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Like char devices, QXL devices need to be explicily started.
For some historical reason, char devices are started when in running
state. See commi bf1d9007b. Reading that commit comments, there was a
plan to provide an API to stop/start devices invidually, but that never
happened. Whether that API would really be useful now, I wonder.
For now, just follow the char devices behaviour and start QXL devices
added when vm_running.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.uuuuucom>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
If the worker is already started, don't assert and just return.
This fixes calling spice_server_vm_start() multiple times.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Acked-by: Frediano Ziglio <freddy77@gmail.com>
- remove entry for g-type-register-static: the "possible" flag has been
re-introduced in glib.supp
- add new entries for several calloc issues
Signed-off-by: Julien Ropé <jrope@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Reduce code duplication.
Also this improve support for big endian machines as
agent_check_message fix also message endianess.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Uri Lublin <ulublin@redhat.com>
Big endian machines on server are not officially supported but won't
hurt.
Messages from client are always encoded as little endian (as all
SPICE protocol).
Convert fields from little endian to host endian on some places
where numbers are used and not just binary copied.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Uri Lublin <ulublin@redhat.com>
This brings in the following changes:
Frediano Ziglio (18):
snd_codec: Use better types for snd_codec_is_capable
snd_codec: Use better type for function result
snd_codec: Use better type for snd_codec_create mode
backtrace: Do not attempt to compile if spice_backtrace is empty
Avoid usage of GCC extension for __VA_ARGS__ where possible
helper-fuzzer-demarshallers: Provide replacement for ftello for MSVC
quic_tmpl: Remove unused bpc parameter
ssl_verify: Do not check IP if we fail to resolve it
proto: Add support for side mouse buttons
quic: Constify a parameter
quic: Fix typo in comment
agent: Extend agent_check_message to support VDAgentMonitorsConfig extension
quic: Check we have some data to start decoding quic image
quic: Check image size in quic_decode_begin
quic: Check RLE lengths
quic: Avoid possible buffer overflow in find_bucket
test-quic: Add fuzzer capabilities to the test
test-quic: Add test cases for quic fuzzer
Haochen Tong (2):
pixman_utils: fix clang "unused functions" warning
marshal: fix clang "missing field initializer" warning on generated files
Marc-André Lureau (1):
agent: fix vdagent monitor flag filtering
Changes in sound.cpp are required due to improved types.
In particular some security related changes are imported.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Uri Lublin <ulublin@redhat.com>
In configure.ac the micro version is incremented if there is a
forth component in the version.
In Meson this was wrongly translated to "if contains git".
This fixes Gitlab issue #46.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Eduardo Lima (Etrunko) <etrunko@redhat.com>
Remove some possible warning, like
reds.cpp:216:18: warning: 'remove_client' overrides a member function but is not marked 'override' [clang-diagnostic-inconsistent-missing-override]
virtual void remove_client(RedCharDeviceClientOpaque *opaque);
^
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Uri Lublin <ulublin@redhat.com>
Some compiler could generate some warning, like
reds.cpp:2678:5: warning: Call to function 'strcpy' is insecure as it does not provide bounding of the memory buffer. Replace unbounded copy functions with analogous functions that support length arguments such as 'strlcpy'. CWE-119 [clang-analyzer-security.insecureAPI.strcpy]
strcpy(buf, pass);
^
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Jakub Janků <jjanku@redhat.com>
Extend mouse button bits support.
Allows to support up to 16 buttons.
Partly based on a patch from Kevin Pouget (RED_MOUSE_BUTTON_STATE_TO_AGENT
macro, updated on new protocol constants).
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Kevin Pouget <kpouget@redhat.com>
The regression was introduced by
commit 22fc6a48e6
red-channel-client: Change GQueue into a std::list
Starts using smart pointers in the queue.
dcc_add_surface_area_image added a new pipe item in front if pos was
NULL. Before the commit red_add_lossless_drawable_dependencies passed
a pointer to the last element in the pipe adding the new item on the
back. The change caused the item to be inserted in front instead of
back. Restore the insert position.
This caused in some condition the pipe to grow uncontrollably (like
using a large notepad window on Windows 7 and moving it around).
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Let the compiler decide.
Apparently it causes CI makecheck to fail (see below).
Seems to me it's a gcc problem -- was working with version 9,
fails with version 10.
Making the function a non-inline one "fixes" this issue.
Alternatively adding to the structure "int dummy" above the "int type"
"fixes" the build too (but I went with the non-inline solution).
==== error ====
CC mjpeg-encoder.lo
In file included from /usr/include/string.h:495,
from ../../server/red-common.h:24,
from ../../server/mjpeg-encoder.c:25:
In function ‘memset’,
inlined from ‘mjpeg_encoder_reset_quality’ at ../../server/mjpeg-encoder.c:385:5,
inlined from ‘mjpeg_encoder_quality_eval_stop.part.0’ at ../../server/mjpeg-encoder.c:999:5:
/usr/include/bits/string_fortified.h:71:10: error: ‘__builtin_memset’ offset [788, 871] from the object at ‘encoder’ is out of the bounds of referenced subobject ‘type’ with type ‘int’ at offset 784 [-Werror=array-bounds]
71 | return __builtin___memset_chk (__dest, __ch, __len, __bos0 (__dest));
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../server/mjpeg-encoder.c: In function ‘mjpeg_encoder_quality_eval_stop.part.0’:
../../server/mjpeg-encoder.c:96:9: note: subobject ‘type’ declared here
96 | int type;
| ^~~~
cc1: all warnings being treated as errors
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Provide constructor and destructor, use new/delete.
This allows additional fields to be properly initialized if needed.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Remove some conditional compilation.
Make try_compress_lz4 return success or failure and change the
messages making easier to read result from try_compress_lz4.
Remove "n" parameter from try_compress_lz4, use message size
instead.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Reduces usage of manual reference counting.
Avoids usage of new and use instead red::make_shared to both
create and own pipe items.
Use move semantic to reduce useless copies.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Now that items are managed by constructors/destructors avoid some
manual releases.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Provides a base class to allows RedCharDeviceVDIPort to pass
a RedPipeItem to MainChannelClient instead of having to wrap
into another item.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Most of the times each class has an associated constant so
allows to be able to define the constant at declaration time.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
This allows to:
- reuse reference counting;
- avoid having to manually call g_free to release item memory;
- assure item is initialized;
- avoids some manual casts.
It will also allows to use smart pointers.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
To avoid memory errors marshaller_unref_pipe_item should be
matched with a previous call to red_pipe_item_ref.
This is correctly done but to reduce the possibility of breaking
this rule move both referencing and unreferencing into a new
RedPipeItem::add_to_marshaller method.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Exit more cleanly, otherwise sometimes some strange failures
occurs. This is due to the fact that a thread is cleaning up
everything while the others are still working.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Uri Lublin <ulublin@redhat.com>
RedCharDeviceSmartcard is not using token handling (which is designed
only for VDAgent) so push directly to client.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
For historical reasons this was computed using VDAgent details.
Provide a more suitable and simple constant.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Uri Lublin <ulublin@redhat.com>
"password" is pretty small, allocate on stack directly.
RSA is currently 1024 bits, that is more or less 128 bytes.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Kevin Pouget <kpouget@redhat.com>
This flag allows to catch variables or arguments hiding other
variables or attributes.
It helps avoiding some possible mistakes.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
This is a regression of commit 176970f3f1
(cfr "red-channel-client: Remove GObject type").
Attribute was not initialised.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Move to Doxygen (output was not meant to generate output
automatically).
Update some part that were outdated.
Typo messsage_type -> message_type.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Allows to simplify a bit marshalling code in both CursorChannel
and DisplayChannel as they use the same marshalling structure.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
The pipe items are meant to be used by channel clients, so move
declaration where channel clients is.
This item is used by both CursorChannelClient and DisplayChannelClient
so the CommonGraphicsChannel is the proper place.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
RedCacheItem was using an union to reuse cache item memory
as a pipe item to release that specific cache item.
Instead of spreading that structure everywhere move the specific
optimization all in cache-item.tmpl.cpp.
This make also code less cluttered.
Add some comment on specific code.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Not much used, can be retrieved from the class if needed.
There are some apparent test removal on reds.cpp.
The test "if (!reds->vdagent)" apparently disappeared but it's included in
the while.
reds->agent_dev cannot be NULL as SPICE server allows only one agent and we
are into a member of an object so it must be reds->agent_dev.get() == this.
As there's only an agent and as the interface (sin) is extracted from that
agent it must be reds->agent_dev.get() == sin->st and either reds->vdagent
== sin or reds->vdagent == NULL.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Avoid to pass throught SpiceCharDeviceInterface.
When data is read from device the device get marked as active.
This is coherent with what is done in red_char_device_read_from_device.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
Just pass rest of the arguments.
Removes some extra semicolon.
This change also remove a specific GCC extension.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
The handling of that specific pipe item reference counting
was a bit confusing. In particular every time before adding to
the queue the item was reset causing the reference counter to
be 1 however it was not sure that the item was not still referenced
when added again to the queue.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
The assumption is a bit fragile as code could change adding
additional references to the pipe item.
Instead use marshaller cleanup routines, this surely will be
triggered when the message is queued to the network.
This also causes some additional cleanups.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Ropé <jrope@gmail.com>
It is called from 2 places.
1. display-channel.cpp:
dcc_get_encoders always returns a non-null pointer.
2. image-encoders.cpp: image_encoders_release_glz
<- image_encoders_free (where enc is not checked)
<- dcc.c:image_encoders_free
where non-null pointer is passed.
Signed-off-by: Uri Lublin <uril@redhat.com>
socket_close is mapped to close under Unix but under Windows
you should call closesocket instead so call socket_close which
will map to the proper close function on both environments.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Uri Lublin <ulublin@redhat.com>
__glz_dictionary_window_alloc_image may return NULL.
Check that the image was created successfully and if not
then error out using dict->cur_usr->error like it's done
few lines below.
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
glz_enc_dictionary_create may return NULL.
glz_enc_dictionary_restore itself may return NULL so
add one more check.
Found-by: gcc (10) analyzer
Signed-off-by: Uri Lublin <uril@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The shift of a uint_8 number by a number > 32 causes an overflow.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Uri Lublin <ulublin@redhat.com>
This regression was caused by commit
"2ffa7d00c60808e2f640df9bc9b5d62598455588"
(cfr "Improve encapsulation for InputsChannelClient")
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Rope <jrope@redhat.com>
The spice-wrapped.h was introduced to deal with such tricks
in the code and limit them to a single place.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Rope <jrope@redhat.com>
Compiling under ARM7 GCC compiler report these warning:
image-encoders.cpp: In function 'int lz4_usr_more_space(Lz4EncoderUsrContext*, uint8_t**)':
image-encoders.cpp:261:32: error: cast from 'Lz4EncoderUsrContext*' to 'Lz4Data*' increases required alignment of target type [-Werror=cast-align]
261 | EncoderData *usr_data = &(((Lz4Data *)usr)->data);
| ^~~~~~~~~~~~~~
image-encoders.cpp: In function 'int lz4_usr_more_lines(Lz4EncoderUsrContext*, uint8_t**)':
image-encoders.cpp:329:32: error: cast from 'Lz4EncoderUsrContext*' to 'Lz4Data*' increases required alignment of target type [-Werror=cast-align]
329 | EncoderData *usr_data = &(((Lz4Data *)usr)->data);
| ^~~~~~~~~~~~~~
In all other occurrences of these cast we are already using
SPICE_CONTAINEROF.
Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Acked-by: Julien Rope <jrope@redhat.com>
Using socket pairs and trying to set a TCP level option on the
socket setsockopt returns EINVAL error and not ENOTSUP as
expected.
Check this case and handle as ENOTSUP (ignoring it).
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Darwin uses also the constant EOPNOTSUPP for the same reasons.
In some versions EOPNOTSUPP is defined as ENOTSUP.
Check both values, not only ENOTSUP.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Darwin uses for the same setting the TCP_KEEPALIVE option.
Use TCP_KEEPALIVE instead of TCP_KEEPIDLE for Darwin.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Darwin does not have MSG_NOSIGNAL but allows to set a SO_NOSIGPIPE
option to disable sending SIGPIPE writing to closed socket.
Note that *BSD has the SO_NOSIGPIPE option but does not affect all
write calls so instead continue to use MSG_NOSIGNAL instead on
that systems.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
When `build-aux/git-version-gen` is not able to find the project
version (no git tags and no `.tarball-version`), it returns 'UNKNOWN'
(instead of the project version like `0.14.3.185-15b1`).
This `UNKNOWN` value fails `meson build` command:
server/meson.build:8:0: ERROR: Index 1 out of bounds of array of size 1.
With this patch, we set a default major/minor/micro value when the
actual version is unknown. This is the same as the `autoconf`
historical behavior.
Fixes#41.
Signed-off-by: Kevin Pouget <kpouget@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
When SASL is active, if a read request is made and SASL buffer contains some
data (but not enough to fulfill the request), upon return the taken data from
the buffer is not accounted for and hence part of the message gets discarded.
red_stream_sasl_read function takes available data from sasl buffer and returns
if it's enough. If it's not, nbyte is decremented and buf pointer is
incremented to account for the taken data (if any). Then it tries to get more
data from the socket and decode it.
Suppose there was some data in the sasl buffer, but not enough. Then the socket
is not readable (EAGAIN, EINTR, whatever) or the new data isn't enough for
sasl_decode (hence decodedlen == 0). In both cases the function returns as if
no data was read, but it took some data from sasl buffer. This data is lost and
from this point on the communication ceases on the channel (eventually new data
is read, but messages are corrupt without the parts previously discarded).
On the other hand, if some data is read from sasl buffer and everything else
works fine, the output buffer contains all the data, but the count returned
only inform the caller about the newly read data (which causes the similar
effect of discarding part of the message).
Fixes: https://gitlab.freedesktop.org/spice/spice/-/issues/40
Acked-by: Frediano Ziglio <fziglio@redhat.com>
We don't use anymore GObject parameters so avoid having to
register enum values to GType system to use them.
We just need to get the nick value of the enum values for
debug purposes.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Instead of including spice.h directly include an header that wraps
it. This allows to remove the SPICE_SERVER_INTERNAL define.
Currently is used to rename SpiceCharDeviceInstance to RedCharDevice
and reduce its visibility to hidden. This remove some warnings
and some weird code in the source.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
This allows to make easier the management of owning.
Reference counting is automatically updated based on shared
pointers modification.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
