Extend sizes check also to decoding, actually the source data
decoding images should be less safe than encoding.
This avoids different integer overflows and buffer overflows.
To avoid potential issues images are limited to 1GB.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The type of the image is just copied from network without
any check and later used for array indexing.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Uri Lublin <uril@redhat.com>
SPICE_ATTR_PRINTF uses __printf__ format attribute.
spice_log internally uses g_string_append_vprintf which uses
G_GNUC_PRINTF attribute.
G_GNUC_PRINTF can be __printf__ or gnu_printf format which in
some systems (currently Windows) can be different.
GLib 2.58 changed G_GNUC_PRINTF on Windows from using __printf__
attribute to gnu_printf.
To avoid problems in the future uses G_GNUC_PRINTF instead of
SPICE_ATTR_PRINTF for spice_log.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
A full copy can keep both the key and the value instead of allocating
twice the memory.
We are parsing key1=val1,key2=val2,... and in doing that, we currently
store 'key1' in key, and 'val1' in val, and then 'key2', 'val2', and so
on.
After this patch, we store 'key1\0val1\0' in key, which fits and saves
some memory.
Also this removes a warning produced by Coverity that is assuming that
allocating strlen(string_variable) is wrong.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This makes it easier to convert enums registered with glib type system
to string in order to print them at runtime.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Code generated for demarshallers define and declare some types and
functions.
However these types and functions are also declared separately in other
headers (currently spice-common/client_demarshallers.h and
spice/server/demarshallers.h) resulting in potential ABI mismatch if the
different declarations do not match.
Using a common header shared between generated code and code using
these functions prevent potentially multiple different declarations.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This should only be included from c files.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Eduardo Lima (Etrunko) <etrunko@redhat.com>
SPICE version 2 was introduced more than 8 years ago.
RHEL 6 already removed support for version 1 in the server.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
In some architectures GLib macros to change endianness use the
argument multiple times causing possible side effects.
This happens for instance using Debian SID and MIPS.
This fixes https://gitlab.freedesktop.org/spice/spice-common/issues/1.
Reported-by: Laurent Bigonville <bigon@debian.org>
Tested-by: Laurent Bigonville <bigon@debian.org>
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
The id is called 'monitor_id' throughout the rest of the code, rename
for clarity and consistency.
Signed-off-by: Lukáš Hrázký <lhrazky@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The 2 APIs are equivalent.
Some minor coherence changes:
- remove line terminator, already added;
- start message with lower case;
- LZ4, not Lz4.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Check that we have enough data before reading.
This could lead to read buffer overflows being undetected.
This is not a security issue, read happens only in the client not causing
any information leakage, maximum can generate a crash or some garbage on
the screen.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Now that the 2 template files are the same (except for whitespace differences),
we can use a single file.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
This commit adds a common block of macro declarations at the top of
quic_tmpl.c and quic_rgb_tmpl.c. This block is identical between the 2
files, and is one big step towards making the 2 files identical.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
This commit reorders the macro declarations at the beginning of
quic_tmpl.c and quic_rgb_tmpl.c so that they follow a similar order.
This does the same for the #undef at the end of both file.
This commit is only code movement, and should not add/remove anything
which was not there before. The next commit will unify the macro
declarations and #undef.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Define and reuse a COPY_PIXEL macro to copy a pixel.
This will help in making quic_tmpl.c and quic_rgb_tmpl.c identical.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
They will help unify quic_rgb_tmpl.c and quic_tmpl.c
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
This allows to pass an additional 'channel' argument when needed, and
should eventually let us unify quic_tmpl.c and quic_rgb_tmpl.c
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
This will help to unify quic_tmpl.c and quic_rgb_tmpl.c
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
These macros were added to quic_tmpl.c in 815223861 but without the
corresponding #undef. This commit adds them for consistency.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Define COMPRESS_xx/UNCOMPRESS_xx macros in a more similar way between
quic_tmpl.c and quic_rgb_tmpl.c using channel suffixes. Instead of
having #define COMPRESS(channel) do_something(channel), we'll now have
#define COMPRESS(channel) do_something(channel_##channel), and call it
with COMPRESS(a)
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
This helps making the code in quic_tmpl.c and quic_rgb_tmpl.c more
similar.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
This makes them identical to their counterparts in quic_rgb_tmpl.c
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
g_abort() was only added in glib 2.50, which causes meson build failures
since this defines GLIB_VERSION_MIN_REQUIRED/GLIB_VERSION_MAX_ALLOWED
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Most users of spice_common.h don't need it, or only need log.h. It only
has a few users outside of spice-common. It's not very well defined
which headers it should contain. This commit removes spice_common.h in
favour of direct inclusion of the needed headers.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
There are only 2 users in spice-common, and none in spice-gtk/spice
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
spice-gtk was the last user, and stopped using it on Jun 14th 2017
in 040090ccba34 "build-sys: remove -DSPICE_DISABLE_ABORT"
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
Make code faster on platforms not supporting unaligned access by
default.
SPICE_UNALIGNED_CAST is just silencing possible alignment warning and,
if enabled, produces some logs, however in this case we know that the
pointer can be misaligned.
Use packed structures to tell compiler to generate best code possible.
For more details see comment on commit 74e50b57ae ("Make the
compiler work out better way to write unaligned memory").
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This patch just changes some spaces fixing some possible misleading
indentation.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Victor Toso <victortoso@redhat.com>
golomb_coding is always followed by a encode call.
Simplify code calling directly it, no reason to pass back output
using pointers.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
All usages of this macro were removed.
A library should not log to standard error but use log facilities
so remove this macro to avoid possible future usage.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
This allows users of the library to redirect where these messages go
through use of g_log_set_default_handler() rather than unconditionally
printing on stderr.
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
The possible file descriptor associated to the message (currently
can be only the DRM descriptor from Virgl) is not freed in case
the marshaller is reset/destroyed.
This can happen connecting/disconnecting client.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Jonathon Jongsma <jjongsma@redhat.com>
region.c contained code to test the module.
Separate test code into a proper test.
Test is copied verbatim from the original code.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
LibreSSL 2.7 defines ASN1_STRING_get0_data, like OpenSSL 1.1.
This fixes https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=229044.
Original patch from Piotr Kubaj.
Reported-by: Jack L. <xxjack12xx@gmail.com>
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
LibreSSL 2.7 does not define ASN1_STRING_get0_data, like OpenSSL 1.1.
This fixes https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=229044.
Original patch from Piotr Kubaj.
Reported-by: Jack L. <xxjack12xx@gmail.com>
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The DUMP_JPEG compile time flag is used to dump all jpeg images
to files.
The code was saving garbage instead of proper data.
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Reported-by: 谢 昆明 <KunMing.Xie@hotmail.com>
Tested-by: 谢 昆明 <KunMing.Xie@hotmail.com>
Acked-by: Christophe Fergeau <cfergeau@redhat.com>
The error functions are supposed to not return. Using the attribute
documents that in the interface, and is better for the compiler
because most noreturn-related optimizations and spurious warning
eliminations happen at the call site.
Signed-off-by: Christophe de Dinechin <dinechin@redhat.com>
Acked-by: Frediano Ziglio <fziglio@redhat.com>
Calling g_log_set_handler() is problematic as this means applications
using spice-server won't be able to use g_log_set_default_handler to
manage spice logging output themselves. Since this call is only needed
for backwards compatibility (so that calling g_log together with
SPICE_DEBUG_LEVEL set has the expected behaviour), we can install it
only when SPICE_DEBUG_LEVEL is set.
Then we should deprecate SPICE_DEBUG_LEVEL once and for all.
Acked-by: Frediano Ziglio <fziglio@redhat.com>
