>
> SpiceUsbBackendDevice structure is extended to support
> additional kind of device that is emulated by Spice-GTK
> and not present locally (and does not have libusb_device),
> such device has instead pointer to SpiceUsbEmulatedDevice
> abstract structure. Specific implementation of such device
ping the series
>
> Don't call close but socket_close.
>
> Signed-off-by: Frediano Ziglio
> ---
> server/tests/test-websocket.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> Currently the executable cannot run on Windows but it does no
>
> This fixes a bug with --config=handling.
>
> Signed-off-by: Jeremy White
> ---
> v2: Simplify even further.
> ---
> src/main.c| 8 +
> src/options.c | 96 ---
> src/options.h | 5 +--
> 3 files changed, 47 insertions(+), 62 deletions
>
> Signed-off-by: Jeremy White
Acked
> ---
> src/options.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/src/options.c b/src/options.c
> index 6620b853..64e6ac95 100644
> --- a/src/options.c
> +++ b/src/options.c
> @@ -327,7 +327,9 @@ int options_parse_arguments
>
> This fixes a bug with --config=handling.
>
> This also includes test code courtesy of Frediano Ziglio
> .
This was very experiment patch, just worked.
I would turn it in a real test as there's a tests directory.
It seems is using just exported APIs so maybe would be bett
>
> This is largely the work of Frediano Ziglio .
>
> Signed-off-by: Jeremy White
Fine for me, ack
> ---
> src/options.c | 118 +++---
> 1 file changed, 54 insertions(+), 64 deletions(-)
>
> diff --git a/src/option
>
> Hi Frediano,
>
> On 7/23/19 11:21 PM, Frediano Ziglio wrote:
> > The decoding is wrong, the Red and QXL structures are different so
> > code is not doing what is expected. red-parse-qxl translate from QXL
> > to Red structures, red-record-qxl saves Red structu
>
> Signed-off-by: Jeremy White
Acked
> ---
> src/gui.c| 3 +--
> src/listen.c | 3 +--
> src/main.c | 4 +---
> src/spice.c | 3 +--
> 4 files changed, 4 insertions(+), 9 deletions(-)
>
> diff --git a/src/gui.c b/src/gui.c
> index 6748f66e..88acf5c9 100644
> --- a/src/gui.c
> +++ b/sr
>
> We were overly dependent on timing for success; now we
> deliberately wait for our GC and drawing to finish.
>
> Signed-off-by: Jeremy White
Acked
> ---
> src/tests/xcb.c | 15 +++
> 1 file changed, 11 insertions(+), 4 deletions(-)
>
> diff --git a/src/tests/xcb.c b/src/tests
Avoid having to destroy and create a new GSource every time
we change event mask.
Interfaces required for this patch are available since GLib 2.36
and are specific to Unix.
On use old implementation.
Signed-off-by: Frediano Ziglio
---
Changes since v1:
- Use g_source_add_unix_fd and family
The structure is not that small and is not necessary to copy
the value.
This also removed a Coverity warning.
Signed-off-by: Frediano Ziglio
---
src/vdagent/x11-priv.h | 2 +-
src/vdagent/x11-randr.c | 8
src/vdagent/x11.c | 2 +-
3 files changed, 6 insertions(+), 6 deletions
These patches mainly remove some warnings detected by Coverity
and clang code analyzer.
It's a collection of different patches on the same topic, one
initially written by Victor.
Frediano Ziglio (4):
x11-randr: Avoid passing XEvent as value
x11: Avoid passing XEvent as value
x11: Con
Some code scanners (both Coverity and clang one) report that
prev_sel/prev_cond could be unreferenced while NULL.
Change condition to make clear a NULL pointer is not used.
Signed-off-by: Frediano Ziglio
---
src/vdagent/x11.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff
From: Victor Toso
The g_list_length() function does iterate over both lists to compare
its length. Considering that we use this to check for true/false and
we will iterate again on both lists, we can do so once.
This also avoids covscan/clang warnings:
| spice-vdagent-0.19.0/src/vdagent/device
No reasons to have it mutable.
Signed-off-by: Frediano Ziglio
---
src/vdagent/x11.c | 23 ---
1 file changed, 12 insertions(+), 11 deletions(-)
diff --git a/src/vdagent/x11.c b/src/vdagent/x11.c
index 42f42e7..966ea62 100644
--- a/src/vdagent/x11.c
+++ b/src/vdagent/x11.c
The structure is not that small and is not necessary to copy
the value.
This also removed a Coverity warning.
Signed-off-by: Frediano Ziglio
---
src/vdagent/x11.c | 42 +-
1 file changed, 21 insertions(+), 21 deletions(-)
diff --git a/src/vdagent/x11.c b
ping the series
>
> The structure is not that small and is not necessary to copy
> the value.
> This also removed a Coverity warning.
>
> Signed-off-by: Frediano Ziglio
> ---
> src/vdagent/x11-priv.h | 2 +-
> src/vdagent/x11-randr.c | 8
> src/vdagent
Signed-off-by: Frediano Ziglio
---
server/red-replay-qxl.c | 17 ++---
1 file changed, 10 insertions(+), 7 deletions(-)
Changes v1:
- realloc the buffer inside read_binary with the proper size.
The proper size if not available in red_replay_image.
diff --git a/server/red-replay-qxl.c
>
> Instead of trying to parse argv ourselves, just reuse getopt_long_only.
>
> Signed-off-by: Jeremy White
This patch looks a bit long compared to the job is doing.
What is the expected behaviour mixing --config and other options?
It looks like the configuration file should be always read firs
>
> Hi,
>
> On Wed, Jun 19, 2019 at 04:56:41PM +0100, Frediano Ziglio wrote:
> > Avoid having to destroy and create a new GSource every time
> > we change event mask.
> > Interfaces required for this patch are available since GLib 2.36.
> > on Windows GP
Some code scanners (both Coverity and clang one) report that
prev_sel/prev_cond could be unreferenced while NULL.
Change condition to make clear a NULL pointer is not used.
Signed-off-by: Frediano Ziglio
---
src/vdagent/x11.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff
ping
>
> ping
>
> >
> > Avoid having to destroy and create a new GSource every time
> > we change event mask.
> > Interfaces required for this patch are available since GLib 2.36.
> > on Windows GPollFD::fd can be an HANDLE but not a socket.
On LLP64 platforms (like Windows) a virtual address cannot
be represented by a "unsigned long" type, so use uintptr_t
which is defined as a integral type large like a pointer.
"address_delta" is a difference of pointers so use same
type size.
Signed-off-by: Frediano Ziglio
-
Threads are joined by spice_server_destroy, just need to release
last resources.
Signed-off-by: Frediano Ziglio
---
server/tests/replay.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/server/tests/replay.c b/server/tests/replay.c
index 992f26d63..44f244457 100644
Signed-off-by: Frediano Ziglio
---
server/tests/replay.c | 18 ++
1 file changed, 6 insertions(+), 12 deletions(-)
diff --git a/server/tests/replay.c b/server/tests/replay.c
index e3ef7bf15..992f26d63 100644
--- a/server/tests/replay.c
+++ b/server/tests/replay.c
@@ -116,7
his
issue.
Note that for opposite conversion (QXLPHYSICAL_TO_PTR) the conversion
does not change, type is changed just for consistency.
Signed-off-by: Frediano Ziglio
---
server/red-replay-qxl.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/server/red-replay-qxl.c b/ser
>
> Unify parameter name for libusb_device.
> Use g_return_val_if_fail for expected condition.
See
https://developer.gnome.org/glib/stable/glib-Warnings-and-Assertions.html#g-return-val-if-fail
g_return_val_if_fail represents a bug in the code, not an "expected condition".
> Remove redundant ca
in the previous
> commit.
>
I tested, nested attributes are needed, otherwise ABI is broken.
> > Otherwise, ack for the series
> >
> >
> > On 7/22/19 2:08 PM, Frediano Ziglio wrote:
> > > All other code use these headers.
> > >
> > > Signed-off-b
ries
>
>
> On 7/22/19 2:08 PM, Frediano Ziglio wrote:
> > All other code use these headers.
> >
> > Signed-off-by: Frediano Ziglio
> > ---
> > server/migration-protocol.h | 32 ++--
> > server/reds.c | 4 +++
Just syntax change.
Reduce code to initialise these matrices.
Signed-off-by: Frediano Ziglio
---
src/spice-widget-egl.c | 40 +---
1 file changed, 13 insertions(+), 27 deletions(-)
diff --git a/src/spice-widget-egl.c b/src/spice-widget-egl.c
index 7db66082
Signed-off-by: Frediano Ziglio
---
src/spice-widget-egl.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/spice-widget-egl.c b/src/spice-widget-egl.c
index 7bae4e58..7db66082 100644
--- a/src/spice-widget-egl.c
+++ b/src/spice-widget-egl.c
@@ -129,7 +129,7 @@ static
Signed-off-by: Frediano Ziglio
---
src/usb-device-manager.c | 14 ++
1 file changed, 10 insertions(+), 4 deletions(-)
Changes since v1:
- conver a debug into warning to inform the user at least one time
diff --git a/src/usb-device-manager.c b/src/usb-device-manager.c
index 9aba275f
These structure contain only bytes, no need for this attribute.
Signed-off-by: Frediano Ziglio
---
server/glz-encoder.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/server/glz-encoder.c b/server/glz-encoder.c
index 341294035..ee9d95ae9 100644
--- a/server/glz
All other code use these headers.
Signed-off-by: Frediano Ziglio
---
server/migration-protocol.h | 32 ++--
server/reds.c | 4 +++-
2 files changed, 21 insertions(+), 15 deletions(-)
diff --git a/server/migration-protocol.h b/server/migration
The structure has no holes, adding this attribute could only
decrease efficiency.
Note that HashEntry structure is used for a large (8MB) array so
this won't affect much possible container size.
Signed-off-by: Frediano Ziglio
---
server/glz-encoder-priv.h | 2 +-
1 file changed, 1 inse
his has some advantage:
> > > - replace 2 API with 1;
> > > - code reuse for handling the event (removed 2 functions);
> > > - avoid the caller to use the file descriptor;
> > > - avoid the caller to register wrong events.
> > >
> > > Signe
Acked
>
> ---
> server/tests/Makefile.am | 6 +++---
> server/tests/meson.build | 2 +-
> server/tests/{video-encoders => test-video-encoders} | 0
> 3 files changed, 4 insertions(+), 4 deletions(-)
> rename server/tests/{video-encoders =>
>
> This fixes the following warning:
> ./configure: line 7040: /usr/bin/file: No such file or directory
>
> Signed-off-by: Uri Lublin
Acked
> ---
> .gitlab-ci.yml | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
> index 316a860dd..33
>
> Signed-off-by: Jeremy White
I didn't expected you would change everything :-)
> ---
> src/options.c | 44
> 1 file changed, 20 insertions(+), 24 deletions(-)
>
> diff --git a/src/options.c b/src/options.c
> index b7f487c5..a206c92c 100644
> ---
>
> On Fri, Jul 19, 2019 at 1:14 PM Victor Toso wrote:
> >
> > Hi,
> >
> > On Sun, Jul 14, 2019 at 05:07:38PM +0300, Yuri Benditovich wrote:
> > > Remove Windows-specific part of hotplug detection from
> > > usb-device-manager and win-usb-dev and place it into
> > > USB backend module. Connect th
>
> This allows to set plugin key=value properties on run time.
> To add encoder plugin property use the following syntax:
> -c gst.prop="property=value" -c gst.prop="property2=value2"...
> Make sure syntax is accurate and that the property is supported by
> the chosen plugin, wrong properties may
>
> >
> > From: Victor Toso
> >
> > Seems to be a false positive but as this message only happens when user
> > client connects, we can copy this array to make compiling warn free.
> >
> > > src/vdagent/vdagent.c: In function ‘daemon_read_complete’:
> > > src/vdagent/vdagent.c:226:71: error:
From: Victor Toso
The g_list_length() function does iterate over both lists to compare
its length. Considering that we use this to check for true/false and
we will iterate again on both lists, we can do so once.
This should also avoid covscan/clang warnings:
| spice-vdagent-0.19.0/src/vdagent/
driver is XDDM which means Windows XP
or similars, I got no recording with such images.
Signed-off-by: Frediano Ziglio
---
server/red-replay-qxl.c | 17 -
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/server/red-replay-qxl.c b/server/red-replay-qxl.c
index 674feae2f
This small series removes last compile warnings from CI build.
See https://gitlab.freedesktop.org/fziglio/spice/pipelines/49834.
Second patch remove also bug reading record file for QUIC images
instead of just removing the warning.
Frediano Ziglio (2):
red-replay-qxl: Fix some issue of
Do not pass unaligned QXLPHYSICAL but pass a valid pointer and
then cast.
Signed-off-by: Frediano Ziglio
---
server/red-replay-qxl.c | 18 +-
1 file changed, 13 insertions(+), 5 deletions(-)
Changes since v1:
- remove unnecessary casts
diff --git a/server/red-replay-qxl.c b
bytes instead
> of 12 this is read the same way.
>
> Signed-off-by: Frediano Ziglio
> ---
> spice/qxl_dev.h | 10 +-
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/spice/qxl_dev.h b/spice/qxl_dev.h
> index 659f930..c844d7b 1006
ping
>
> Do not pass unaligned QXLPHYSICAL but pass a valid pointer and
> then cast.
>
> Signed-off-by: Frediano Ziglio
> ---
> server/red-replay-qxl.c | 16
> 1 file changed, 12 insertions(+), 4 deletions(-)
>
> diff --git a/server/red-replay-q
Signed-off-by: Frediano Ziglio
---
spice/start-packed.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/spice/start-packed.h b/spice/start-packed.h
index de0c595..d1d090e 100644
--- a/spice/start-packed.h
+++ b/spice/start-packed.h
@@ -30,8 +30,8 @@
*/
/* Ideally this
>
>
> On 7/18/19 5:51 PM, Frediano Ziglio wrote:
> >> Hi,
> >>
> >>
> >> IIRC this was related to some compiler warning, no?
> > Yes, recent compilers are reporting it, see below.
> >
> >> If it is I'd mentioning it , ot
It seems that x11spice is more X11 complaint than getopt_long so I
don't see the --config=/-config= mandatory.
Giving that -config is not GNU convention (they are 5 options!) I would avoid
the mix.
>
> Signed-off-by: Jeremy White
> ---
> src/options.c | 24 ++--
> 1 file ch
>
> Signed-off-by: Jeremy White
> ---
> src/options.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/src/options.c b/src/options.c
> index 0d3138d0..cf75e54e 100644
> --- a/src/options.c
> +++ b/src/options.c
> @@ -355,7 +355,9 @@ int options_parse_arguments(int arg
>
> Hi,
>
>
> IIRC this was related to some compiler warning, no?
Yes, recent compilers are reporting it, see below.
> If it is I'd mentioning it , otherwise, ack.
>
Just this patch or the entire series?
> Snir.
>
>
> On 7/4/19 4:56 PM, Fredian
>
> On 7/18/19 10:32 AM, Frediano Ziglio wrote:
> > Signed-off-by: Frediano Ziglio
>
> Hi Frediano,
>
> I've been playing with it too.
> Had similar rules but different.
>
> Your version works for me.
> Some minor comments below.
>
> &g
ping
>
> ping
>
> >
> > ping
> >
> > >
> > > The function is called only by red_channel_pipes_new_add.
> > >
> > > Signed-off-by: Frediano Ziglio
> > > ---
> > > server/red-channel.c | 20 ---
a single queue for the network interface
> > limiting and sharing de facto the various buffers between the multiple
> > connections.
> >
> > Signed-off-by: Frediano Ziglio
> > ---
> > Changes in v2:
> > - more commit message comments
> > ---
>
ping
>
> ping
>
> >
> > ping
> >
> > >
> > > There are no much data other than the buffer, reduce the
> > > allocations.
> > >
> > > Signed-off-by: Frediano Ziglio
> > > ---
> > > server/char-de
ping
>
> ping
>
> >
> > ping
> >
> > >
> > > ping
> > >
> > > >
> > > > ping
> > > >
> > > > >
> > > > > ping
> > > > >
> > > > > >
ping
>
> ping
>
> >
> > ping
> >
> > >
> > > ping
> > >
> > > >
> > > > ping
> > > >
> > > > >
> > > > > Signed-off-by: Frediano Ziglio
> > > > > ---
&g
ce_send_to_client_tokens_add are doing mostly
> > > > the same thing so put common code to
> > > > red_char_device_send_to_client_tokens_absorb.
> > > >
> > > > Signed-off-by: Frediano Ziglio
> > > > ---
> > > > server/char-device.c | 42 +++
; > > than RedClient.
> > > > The define at the beginning of C file is to make sure we don't
> > > > use the opaque type as a specific one.
> > > >
> > > > Signed-off-by: Frediano Ziglio
> > > > ---
> > > > server/char-device.c
defined by the compiler
> > > > > and defining again cause a warning.
> > > > > Do not define again to avoid the warning.
> > > > >
> > > > > Signed-off-by: Frediano Ziglio
> > > > > ---
> > > > > server/red-worker.c | 2 ++
>
ping
>
> ping
>
> >
> > ping
> >
> > >
> > > ping
> > >
> > > >
> > > > Just a style change, return earlier to avoid some indentation.
> > > >
> > > > Signed-off-by: Frediano Ziglio
ping
>
> ping
>
> >
> > ping
> >
> > >
> > > ping
> > >
> > > >
> > > > ping
> > > >
> > > > >
> > > > > Unused.
> > > > > Also the devices shoul
Previously we add suppression to glib.supp file (suppressions from
Glib).
Keep the glib.supp file pristine and add another file specific
for SPICE.
Signed-off-by: Frediano Ziglio
---
server/tests/Makefile.am | 2 +-
server/tests/valgrind/glib.supp | 39
Sync with Glib master file.
Signed-off-by: Frediano Ziglio
---
server/tests/valgrind/glib.supp | 222 +++-
1 file changed, 220 insertions(+), 2 deletions(-)
diff --git a/server/tests/valgrind/glib.supp b/server/tests/valgrind/glib.supp
index 9236b3d16..ddc2b1729
This series split and update Valgrind suppression files in
order for make Gitlab CI pass.
Results at
https://gitlab.freedesktop.org/fziglio/spice/-/jobs/430006.
Note that to pass Valgrind checks some other patches are needed
but these series is complete for the suppression part.
Frediano Ziglio
Signed-off-by: Frediano Ziglio
---
server/tests/valgrind/spice.supp | 25 +
1 file changed, 25 insertions(+)
diff --git a/server/tests/valgrind/spice.supp b/server/tests/valgrind/spice.supp
index 1bfe81006..dd3663c68 100644
--- a/server/tests/valgrind/spice.supp
+++ b
>
> From: test
>
> This allows to set plugin key=value properties on run time.
> To add encoder plugin property use the following syntax:
> -c gst.prop="property=value" -c gst.prop="property2=value2"...
> Make sure syntax is accurate and that the property is supported by
> the chosen plugin, wro
> Hi,
>
>
> On 7/16/19 2:35 PM, Frediano Ziglio wrote:
> >> This allows to set plugin key=value properties on run time.
> >> To add encoder plugin property use the following syntax:
> >> -gst.prop="key=value"
> >> Make sure syntax is accu
mage.
On a fully installed Desktop/Server machine these configuration
are usually in place so you won't note the issue but on recent
docker images these configuration are missing.
Running the dconf command add required configuration.
Signed-off-by: Frediano Ziglio
---
.gitlab-ci.yml | 7
mage.
On a fully installed Desktop/Server machine these configuration
are usually in place so you won't note the issue but on recent
docker images these configuration are missing.
Running the dconf command add required configuration.
Signed-off-by: Frediano Ziglio
---
.gitlab-ci.yml | 5 +
inttypes.h contains macro for format string while
stdint.h more specifically contains type definitions (like uint8_t).
Signed-off-by: Frediano Ziglio
---
server/websocket.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/server/websocket.h b/server/websocket.h
index
video_codecs can be freed so use it before.
Signed-off-by: Frediano Ziglio
---
server/reds.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/server/reds.c b/server/reds.c
index 817fdd423..78bbe5a0f 100644
--- a/server/reds.c
+++ b/server/reds.c
@@ -3851,6 +3851,10
Signed-off-by: Frediano Ziglio
---
server/tests/test-websocket.c | 15 +++
1 file changed, 3 insertions(+), 12 deletions(-)
diff --git a/server/tests/test-websocket.c b/server/tests/test-websocket.c
index e74cb5497..432aeee5c 100644
--- a/server/tests/test-websocket.c
+++ b/server
Don't call close but socket_close.
Signed-off-by: Frediano Ziglio
---
server/tests/test-websocket.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
Currently the executable cannot run on Windows but it does not hurt
to make the source more compatible with that system.
diff --
mage.
On a fully installed Desktop/Server machine these configuration
are usually in place so you won't note the issue but on recent
docker images these configuration are missing.
Running the dconf command add required configuration.
Signed-off-by: Frediano Ziglio
---
.gitlab-ci.yml | 2 ++
1
ping the series
>
> ping the series
>
> >
> > Signed-off-by: Frediano Ziglio
> > ---
> > server/websocket.h | 5 +
> > 1 file changed, 5 insertions(+)
> >
> > diff --git a/server/websocket.h b/server/websocket.h
> > index
y change the alignment from 4/8 to 1.
> This could affect structures containing this union however beside
> packed structure in qxl_dev.h (which are not affected) there are no
> other usages affecting ABI by spice-gtk, Qemu or spice-server.
>
> Signed-off-by: Frediano Ziglio
>
> This allows to set plugin key=value properties on run time.
> To add encoder plugin property use the following syntax:
> -gst.prop="key=value"
> Make sure syntax is accurate and that the property is supported by
> the chosen plugin, wrong/invalid properties will be ignored silently.
> Specifi
>
> On Mon, Jul 15, 2019 at 5:04 PM Frediano Ziglio wrote:
> >
> > >
> > > The delete in error flow was missing.
> > >
> > > Signed-off-by: Yuri Benditovich
> > > ---
> > > src/usb-device-manager.c | 1 +
> > > 1 file
>
> Previous commit ca641a5b92 uses libusb_interrupt_event_handler
> which requires newer version of libusb.
>
> Signed-off-by: Yuri Benditovich
My fault.
According to repology this should not be an issue.
Fine for me.
> ---
> meson.build | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(
>
> In case of possible error the procedure of hotplug
> registration does not return error information, just
> issues warning to the stderr, so the reason of the problem,
> if any, is not visible. Current commit adds GError parameter
> to the procedure to return the error details conventionally.
>
> The delete in error flow was missing.
>
> Signed-off-by: Yuri Benditovich
> ---
> src/usb-device-manager.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/src/usb-device-manager.c b/src/usb-device-manager.c
> index 857d057..7105ff1 100644
> --- a/src/usb-device-manager.c
> +++ b/s
No reasons to have it mutable.
Signed-off-by: Frediano Ziglio
---
src/vdagent/x11.c | 23 ---
1 file changed, 12 insertions(+), 11 deletions(-)
diff --git a/src/vdagent/x11.c b/src/vdagent/x11.c
index 42f42e7..966ea62 100644
--- a/src/vdagent/x11.c
+++ b/src/vdagent/x11.c
The structure is not that small and is not necessary to copy
the value.
This also removed a Coverity warning.
Signed-off-by: Frediano Ziglio
---
src/vdagent/x11.c | 42 +-
1 file changed, 21 insertions(+), 21 deletions(-)
diff --git a/src/vdagent/x11.c b
The structure is not that small and is not necessary to copy
the value.
This also removed a Coverity warning.
Signed-off-by: Frediano Ziglio
---
src/vdagent/x11-priv.h | 2 +-
src/vdagent/x11-randr.c | 8
src/vdagent/x11.c | 2 +-
3 files changed, 6 insertions(+), 6 deletions
>
> Hi,
>
> On Mon, Jul 15, 2019 at 05:18:17AM -0400, Frediano Ziglio wrote:
> > >
> > > From: Victor Toso
> > >
> > > The struct type for width/height is uint32_t while we are trying to
> > > access and change it with int* - code can
>
> Add a cache to allow the reuse of SHM segments.
> Shared memory segments are added to the cache instead of being
> deallocated, and the cache is searched instead of/before allocating a
> new segment.
>
> Both the SHM segments and their attachment with the X server are cached.
>
> The cache c
>
> From: Victor Toso
>
> Seems to be a false positive but as this message only happens when user
> client connects, we can copy this array to make compiling warn free.
>
> > src/vdagent/vdagent.c: In function ‘daemon_read_complete’:
> > src/vdagent/vdagent.c:226:71: error: taking address of
>
> From: Victor Toso
>
> The struct type for width/height is uint32_t while we are trying to
> access and change it with int* - code can be improved a bit in following
> patches but this one fixes the warning by copying the value from the
> struct and copying back new value afterwards.
>
> Als
> Subject: [Spice-devel] [linux/vd-agent v1 5/7] coverity: allow pass by value
> with XEvent
I would say that this patch is ignoring a warning, the code is compiling.
>
> From: Victor Toso
>
> Signed-off-by: Victor Toso
> ---
> src/vdagent/x11-randr.c | 4 ++--
> src/vdagent/x11.c | 1
>
> From: Victor Toso
>
> Since 2861868 "randr: remove monitors.xml on auto-configuration" in
> 2015-04-10 by Marc-Andre Lureau this
> file includes glib.h and can remove those helpers.
>
> Signed-off-by: Victor Toso
> ---
> src/vdagent/x11-randr.c | 18 --
> 1 file changed,
>
> From: Victor Toso
>
> | Error: CHECKED_RETURN (CWE-252):
> | spice-vdagent-0.19.0/src/vdagentd/vdagentd.c:999: check_return: Calling
> | "fprintf" without checking return value (as is done elsewhere 29 out of
> | 30 times).
> | spice-vdagent-0.19.0/src/vdagentd/xorg-conf.c:95: example_a
>
> From: Victor Toso
>
> Found by covscan:
>
> | spice-vdagent-0.19.0/src/vdagent/vdagent.c:432:9: warning: Potential leak
> | of memory pointed to by 'orig_argv'
> | #g_printerr("Invalid arguments, %s\n", error->message);
> | #^
> | spice-vdagent-0.19.0/src/vdagent/vdagen
>
> When launching with no config file or arguments, the "You must provide
> some authentication method" dialog appears. Clicking the Close button
> would hang the application and it would never quit.
>
> s->draw_command_in_progress was uninitialized, and flush_and_lock()
> would get stuck in an
>
> Quitting with the Quit button or closing the window always causes a
> segfault for me. Destroying the surface before removing the interface
> seems to resolve the problem.
>
Acked. The surface surely was of the only display interface available.
> log from valgrind:
> ==7741== Invalid read o
>
> When launching with no config file or arguments, the "You must provide
> some authentication method" dialog appears. Clicking the Close button
> would hang the application and it would never quit.
>
> s->draw_command_in_progress was uninitialized, and flush_and_lock()
> would get stuck in an
>
> Hi,
>
> On Thu, Jul 11, 2019 at 02:00:53PM +0100, Frediano Ziglio wrote:
> > Signed-off-by: Frediano Ziglio
> > ---
> > src/usb-device-manager.c | 12 +---
> > 1 file changed, 9 insertions(+), 3 deletions(-)
> >
> > diff --git a/src
>
> Hi,
>
> On Thu, Jul 11, 2019 at 02:00:54PM +0100, Frediano Ziglio wrote:
> > Attempt to better interrupt event handling loop.
> > If the thread handling events is stuck waiting events or handling an
> > event try to interrupt before joining the thread.
801 - 900 of 10211 matches
Mail list logo