On Thu, 3 Mar 2016, Christophe Fergeau wrote:
[...]
> > +static SpiceGstVideoBuffer* create_gst_video_buffer(void)
> > +{
> > +SpiceGstVideoBuffer *buffer = spice_new0(SpiceGstVideoBuffer, 1);
> > +buffer->base.free = &gst_video_buffer_free;
> > +return buffer;
>
> I forgot to mention
On Thu, 3 Mar 2016, Frediano Ziglio wrote:
[...]
> Should we bump the spice-protocol version too?
The last post-release version bump was in september (017ddbe7). It does
not look like there has been a new release since (nothing in NEWS or git
log). So I believe a version bump is not necessary ye
On Thu, 3 Mar 2016, Christophe Fergeau wrote:
[...]
> > +static void gst_video_buffer_free(VideoBuffer *video_buffer)
> > +{
> > +SpiceGstVideoBuffer *buffer = (SpiceGstVideoBuffer*)video_buffer;
> > +gst_buffer_unref(buffer->gst_buffer);
>
> You need a if (buffer->gst_buffer != NULL) test
On Thu, 3 Mar 2016, Christophe Fergeau wrote:
[...]
> > > +if (system("egrep -l '^flags\\b.*: .*\\bht\\b' /proc/cpuinfo
> > > >/dev/null 2>&1") == 0) {
> > > +/* Hyperthreading is enabled so divide by two to get the
> > > number
> > > + * of physical cores.
> > > +
On Thu, 3 Mar 2016, Christophe Fergeau wrote:
[...]
> > +void display_channel_set_video_codecs(DisplayChannel *display, GArray
> > *video_codecs)
> > +{
> > +spice_return_if_fail(display);
> > +
> > +g_array_set_size(display->video_codecs, 0);
> > +g_array_insert_vals(display->video_co
On Thu, 3 Mar 2016, Christophe Fergeau wrote:
[...]
> > -void dcc_create_stream(DisplayChannelClient *dcc, Stream *stream)
> > +gboolean dcc_create_stream(DisplayChannelClient *dcc, Stream *stream)
> > @@ -739,9 +751,13 @@ void dcc_create_stream(DisplayChannelClient *dcc,
> > Stream *stream)
> >
Acked-by: Jonathon Jongsma
On Thu, 2016-03-03 at 21:22 +0100, Francois Gouget wrote:
> Signed-off-by: Francois Gouget
> ---
>
> I assume the reason for having two asserts is to help with diagnosis so
> I did not merge them.
>
> server/stream.c | 2 +-
> 1 file changed, 1 insertion(+), 1 dele
Nice.
Acked-by: Jonathon Jongsma
On Thu, 2016-03-03 at 16:28 +, Frediano Ziglio wrote:
> qxl_state can be extracted now easily from qxl.
>
> Signed-off-by: Frediano Ziglio
> ---
> server/red-qxl.c| 5 ++---
> server/red-worker.c | 14 ++
> server/red-worker.h | 2 +-
>
On Thu, 2016-03-03 at 16:28 +, Frediano Ziglio wrote:
> Considering that:
> - RedsState is the state of QXLInstance implementation (RedDispatcher);
I don't understand this sentence. Do you mean QxlState instead of RedsState? Why
is RedDispatcher in parentheses here?
> - qif (QXLInterface*) fi
Signed-off-by: Francois Gouget
---
I assume the reason for having two asserts is to help with diagnosis so
I did not merge them.
server/stream.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/server/stream.c b/server/stream.c
index 2bfb993..f1edec7 100644
--- a/server/stre
>
> On Thu, 2016-03-03 at 16:28 +, Frediano Ziglio wrote:
> > RedDispatcher is basically implementing QXLInstance.
> > After some internal discussion in order to avoid rewriting too
> > many time the patch we came out with RedQXL name.
>
> I guess this message will need to be updated slightly
Acked-by: Jonathon Jongsma
On Thu, 2016-03-03 at 16:28 +, Frediano Ziglio wrote:
> Signed-off-by: Frediano Ziglio
> ---
> server/Makefile.am|4 +-
> server/agent-msg-filter.c |2 +-
> server/display-channel.h |2 +-
> server/red-dispatcher.c | 1032 --
On Thu, 2016-03-03 at 16:28 +, Frediano Ziglio wrote:
> RedDispatcher is basically implementing QXLInstance.
> After some internal discussion in order to avoid rewriting too
> many time the patch we came out with RedQXL name.
I guess this message will need to be updated slightly if we decide n
>
> On Tue, Mar 01, 2016 at 04:53:10PM +0100, Francois Gouget wrote:
> > Signed-off-by: Francois Gouget
> > ---
> >
> > In theory this could be needed by the next patch.
> >
> > server/red-parse-qxl.h | 4 ++--
> > server/red-worker.c| 7 ++-
> > 2 files changed, 8 insertions(+), 3 del
On Tue, Mar 01, 2016 at 04:53:03PM +0100, Francois Gouget wrote:
> This way the video encoder is not forced to use malloc()/free().
> This also allows more flexibility in how the video encoder manages the
> buffer which allows for a zero-copy implementation in both video
> encoders.
>
> Signed-off
On Tue, Mar 01, 2016 at 04:53:10PM +0100, Francois Gouget wrote:
> Signed-off-by: Francois Gouget
> ---
>
> In theory this could be needed by the next patch.
>
> server/red-parse-qxl.h | 4 ++--
> server/red-worker.c| 7 ++-
> 2 files changed, 8 insertions(+), 3 deletions(-)
>
> diff -
> Acked-by: Christophe Fergeau
>
Agree!
Should we bump the spice-protocol version too?
Frediano
> On Tue, Mar 01, 2016 at 04:50:16PM +0100, Francois Gouget wrote:
> > Clients that support multiple codecs must advertise the
> > SPICE_DISPLAY_CAP_MULTI_CODEC capability and one
> > SPICE_DISPLA
Acked-by: Christophe Fergeau
On Tue, Mar 01, 2016 at 04:50:16PM +0100, Francois Gouget wrote:
> Clients that support multiple codecs must advertise the
> SPICE_DISPLAY_CAP_MULTI_CODEC capability and one
> SPICE_DISPLAY_CAP_CODEC_XXX per supported codec.
>
> Signed-off-by: Francois Gouget
> ---
Hey,
On Tue, Mar 01, 2016 at 04:53:03PM +0100, Francois Gouget wrote:
> This way the video encoder is not forced to use malloc()/free().
> This also allows more flexibility in how the video encoder manages the
> buffer which allows for a zero-copy implementation in both video
> encoders.
Yes, thi
Hi
- Original Message -
> It sounds like somebody agreed with the patch but then
> loosed interest in it.
>
> Why didn't you removed patches 5 and 6 and defined/used a
> DRM_API_HANDLE_TYPE_SHMID
> instead of WINSYS_HANDLE_TYPE_SHMID and post again? Looks like was the only
> "big" complai
It sounds like somebody agreed with the patch but then
loosed interest in it.
Why didn't you removed patches 5 and 6 and defined/used a
DRM_API_HANDLE_TYPE_SHMID
instead of WINSYS_HANDLE_TYPE_SHMID and post again? Looks like was the only
"big" complaint.
I'm not sure would be worth spending time
On Thu, Mar 03, 2016 at 05:22:58PM +0100, Christophe Fergeau wrote:
> On Thu, Mar 03, 2016 at 09:50:26AM -0500, Frediano Ziglio wrote:
> > > diff --git a/server/gstreamer-encoder.c b/server/gstreamer-encoder.c
> > > index aae47c7..3dfbdfb 100644
> > > --- a/server/gstreamer-encoder.c
> > > +++ b/se
qxl_state can be extracted now easily from qxl.
Signed-off-by: Frediano Ziglio
---
server/red-qxl.c| 5 ++---
server/red-worker.c | 14 ++
server/red-worker.h | 2 +-
3 files changed, 9 insertions(+), 12 deletions(-)
diff --git a/server/red-qxl.c b/server/red-qxl.c
index 11dc1
Signed-off-by: Frediano Ziglio
---
server/Makefile.am|4 +-
server/agent-msg-filter.c |2 +-
server/display-channel.h |2 +-
server/red-dispatcher.c | 1032 -
server/red-dispatcher.h | 272
server/red-qxl.c
RedDispatcher name is quite misleading.
Change name to make more clear that is just a QXL implementation.
Also make more similar to the way other interfaces (like sound one)
are implemented.
Changes from v1:
- use QXLState directly instead of defining a RedQXL type
which is just a typedef.
Fred
RedDispatcher is basically implementing QXLInstance.
After some internal discussion in order to avoid rewriting too
many time the patch we came out with RedQXL name.
This is the first of 3 patches that rename functions, structure
and files.
Signed-off-by: Frediano Ziglio
---
server/display-chann
Considering that:
- RedsState is the state of QXLInstance implementation (RedDispatcher);
- qif (QXLInterface*) field can be computed really easy from QXLInstance;
- mostly of its state is private.
Make all structure private.
Signed-off-by: Frediano Ziglio
---
server/cursor-channel.c | 2 +-
On Thu, Mar 03, 2016 at 09:50:26AM -0500, Frediano Ziglio wrote:
> > diff --git a/server/gstreamer-encoder.c b/server/gstreamer-encoder.c
> > index aae47c7..3dfbdfb 100644
> > --- a/server/gstreamer-encoder.c
> > +++ b/server/gstreamer-encoder.c
> > @@ -188,14 +188,70 @@ static void set_appsrc_caps
Acked-by: Christophe Fergeau
On Tue, Mar 01, 2016 at 04:52:15PM +0100, Pavel Grunt wrote:
> Signed-off-by: Francois Gouget
> ---
> server/tests/replay.c | 11 ++-
> 1 file changed, 10 insertions(+), 1 deletion(-)
>
> diff --git a/server/tests/replay.c b/server/tests/replay.c
> index f
On Thu, Mar 03, 2016 at 10:12:25AM -0500, Frediano Ziglio wrote:
> >
> > On Tue, Mar 01, 2016 at 12:42:20PM -0500, Frediano Ziglio wrote:
> > > > with QXLState an internal structure as well (defined in reds.h), and
> > > > RedDispatcher/RedQXL is one of the few members of this QXLState
> > > > str
>
> On Tue, Mar 01, 2016 at 12:42:20PM -0500, Frediano Ziglio wrote:
> > > with QXLState an internal structure as well (defined in reds.h), and
> > > RedDispatcher/RedQXL is one of the few members of this QXLState
> > > structure. Should they be merged?
> > >
> >
> > I think you should check pat
>
> Signed-off-by: Francois Gouget
> ---
>
> Changed the VP8 encoder parameters based on the realtime profile to
> improve performance. The patch does not use the realtime profile
> directly because profiles don't seem to be supported from 'gst-launch'
> pipeline strings yet.
>
> Furthermore it
Hey,
Mostly looks good, couple of minor comments below,
On Tue, Mar 01, 2016 at 04:51:29PM +0100, Francois Gouget wrote:
> The Spice server administrator can specify the encoder and codec
> preferences to optimize for CPU or bandwidth usage. Preferences are
> described in a semi-colon separated l
Hey,
On Tue, Mar 01, 2016 at 04:51:14PM +0100, Francois Gouget wrote:
> Signed-off-by: Francois Gouget
> ---
> server/dcc.c| 5 -
> server/dcc.h| 2 +-
> server/stream.c | 41 +
> 3 files changed, 34 insertions(+), 14 deletions(-)
>
> diff -
Hi
- Original Message -
> upstream afaik. Another attempt I had, at least for fixing gnome3 &
> llvmpipe, use XShm:
> https://lists.freedesktop.org/archives/mesa-dev/2015-June/085860.html
fwiw, I think the last iteration was
https://lists.freedesktop.org/archives/mesa-dev/2015-June/08645
Hi
- Original Message -
> Hi,
> yesterday/today I attempted to test the gstreamer patches.
> My approach was mainly as "user", install and see how does it work.
>
> I installed Francois patches for spice-protocol and spice-server
> (not clients so I tested with mjpeg).
> The initial tes
Forgot to add
Acked-by: Christophe Fergeau
(with the small fixes you did below).
On Wed, Mar 02, 2016 at 07:46:26PM +0100, Francois Gouget wrote:
> On Wed, 2 Mar 2016, Christophe Fergeau wrote:
> [...]
> > > +if test "x$enable_gstreamer" != "xno"; then
> > > +SPICE_CHECK_GSTREAMER(GSTREAM
On Wed, Mar 02, 2016 at 07:46:26PM +0100, Francois Gouget wrote:
> On Wed, 2 Mar 2016, Christophe Fergeau wrote:
> [...]
> > > +if test "x$enable_gstreamer" != "xno"; then
> > > +SPICE_CHECK_GSTREAMER(GSTREAMER_1_0, 1.0, [gstreamer-1.0
> > > gstreamer-base-1.0 gstreamer-app-1.0 gstreamer-video
Hi,
yesterday/today I attempted to test the gstreamer patches.
My approach was mainly as "user", install and see how does it work.
I installed Francois patches for spice-protocol and spice-server
(not clients so I tested with mjpeg).
The initial test was playing this video:
https://www.youtube
On Thu, Mar 03, 2016 at 01:50:30PM +0100, Marc-André Lureau wrote:
> Hi
>
> On Fri, Feb 26, 2016 at 6:58 PM, Frediano Ziglio wrote:
> >>
> >> codegen generated code depends on spice-common code (marshaller, messages
> >> etc),
> >> it makes more sense to keep the generator along this. Otherwise a
On Tue, Mar 01, 2016 at 12:42:20PM -0500, Frediano Ziglio wrote:
> > with QXLState an internal structure as well (defined in reds.h), and
> > RedDispatcher/RedQXL is one of the few members of this QXLState
> > structure. Should they be merged?
> >
>
> I think you should check patch 4/5 :)
Ah ind
Hi
On Fri, Feb 26, 2016 at 6:58 PM, Frediano Ziglio wrote:
>>
>> codegen generated code depends on spice-common code (marshaller, messages
>> etc),
>> it makes more sense to keep the generator along this. Otherwise a newer
>> protocol
>> release will fail to build older projects (as is the case t
>
> Note that the commands are executed by spice-server.
> The "skip" is only done on the "sleep" part of the
> "slow" command-line option.
>
> This is helpful to run quickly through uninsteresting commands
> in a beginning of a recorded file and going slowly when
> interesting parts appear
> ---
>
> ---
> server/tests/replay.c | 10 ++
> 1 file changed, 10 insertions(+)
>
> diff --git a/server/tests/replay.c b/server/tests/replay.c
> index f3b670f..83411d3 100644
> --- a/server/tests/replay.c
> +++ b/server/tests/replay.c
> @@ -47,6 +47,8 @@ static QXLWorker *qxl_worker = NULL;
Acked-by: Frediano Ziglio
Frediano
>
> Apparently, after using g_option_context_parse with G_OPTION_REMAINING
> argv is modified and should not be used.
> This patch uses "file" instead of "argv" and makes sure
> file is freed later.
> No free is called upon error - exit takes care of it.
> ---
>
> I believe this was already reviewed separately, in a way it's odd to
> store instance data along with vfunc (ie class) data, but why not, makes
> more sense this way indeed.
>
> Acked-by: Christophe Fergeau
>
Merged.
It's actually not so different that passing a pointer to a real
interfac
>
> On Wed, 2 Mar 2016, Frediano Ziglio wrote:
>
> > >
> > > Signed-off-by: Francois Gouget
> > > ---
> > >
> > > In theory this could be needed by the next patch.
> > >
> >
> > Are you saying that now compression is done in another thread?
>
> Nothing has changed with regards to threading.
On 03/02/2016 03:00 PM, Christophe Fergeau wrote:
Hey,
On Wed, Feb 17, 2016 at 01:54:48PM +0330, Hamid Mazrae Mollaie wrote:
Hi,
Tank you for replay.
yes, my mouse and keyboard are usb and redirect to vm... and all pressed
key sent to vm
all hotkeys and Ctrl+Alt+F[n] does not work for example.
Note that the commands are executed by spice-server.
The "skip" is only done on the "sleep" part of the
"slow" command-line option.
This is helpful to run quickly through uninsteresting commands
in a beginning of a recorded file and going slowly when
interesting parts appear
---
server/tests/repl
---
server/tests/replay.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/server/tests/replay.c b/server/tests/replay.c
index f3b670f..83411d3 100644
--- a/server/tests/replay.c
+++ b/server/tests/replay.c
@@ -47,6 +47,8 @@ static QXLWorker *qxl_worker = NULL;
static gboolean star
This patch-set touches only server/tests/replay utility.
It adds 2 commands and fixes an access to a NULL pointer.
With count and skip new commands, it's possible
to fast-forward the replay to an interesting point.
How to use it:
First run only with --count: spice-server-replay --count $spice-re
Apparently, after using g_option_context_parse with G_OPTION_REMAINING
argv is modified and should not be used.
This patch uses "file" instead of "argv" and makes sure
file is freed later.
No free is called upon error - exit takes care of it.
---
server/tests/replay.c | 8
1 file changed,
On Wed, Mar 02, 2016 at 01:18:41PM -0500, Frediano Ziglio wrote:
> >
> > This patch allows remote-viewer to redirect output/error streams to
> > files.
> > Also if launched from a console program (for instance from the command
> > prompt) you are able to see output from the console where you launc
ACK series.
On Thu, Mar 03, 2016 at 01:58:16AM +0100, Marc-André Lureau wrote:
> The db argument must end with " or \n.
>
> Found thanks to clang scan-build.
>
> Signed-off-by: Marc-André Lureau
> Reported-by: Miroslav Rezanina
> ---
> src/vcard_emul_nss.c | 4
> 1 file changed, 4 insert
54 matches
Mail list logo
