On Thu, Dec 01, 2016 at 02:15:49PM +0100, Victor Toso wrote:
> On Thu, Dec 01, 2016 at 11:37:32AM +0100, Christophe Fergeau wrote:
> > On Wed, Nov 30, 2016 at 06:50:39PM +0100, Victor Toso wrote:
> > > From: Victor Toso
> > >
> > > Failing to get playback or record volume
On Thu, 1 Dec 2016, Frediano Ziglio wrote:
> 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
From: Victor Toso
As spice_usbredir_channel_up() can be called when channel-usbredir is
not yet initialized, priv->host can be NULL leading to a crash in
usbredir.
See previous patch that does initialize SpiceUsbDeviceManager on
SpiceSession init and also the comment in
Hi,
On Thu, Dec 01, 2016 at 01:29:50PM +0100, Christophe Fergeau wrote:
> On Wed, Nov 30, 2016 at 06:36:32PM +0100, Victor Toso wrote:
> > From: Victor Toso
> >
> > SpiceSession does not initialize its SpiceUsbDeviceManager object on
> > startup that could lead to a race
On Thu, 1 Dec 2016, Christophe Fergeau wrote:
> On Thu, Dec 01, 2016 at 05:19:33AM +0100, Francois Gouget wrote:
> > udscs_server_fill_fds() should accept being passed a NULL pointer.
>
> I would reword the commit log a bit, like "udscs_server_fill_fds() is
> dereferencing the 'server' pointer,
Hi,
On Thu, Dec 01, 2016 at 01:33:56PM +0100, Christophe Fergeau wrote:
> On Wed, Nov 30, 2016 at 06:57:12PM +0100, Victor Toso wrote:
> > Hi,
> >
> > On Wed, Nov 30, 2016 at 06:45:45PM +0100, Pavel Grunt wrote:
> > > imho this should be in usbredir
> > >
> > > It is not nice when library
Acked-by: Christophe Fergeau
On Thu, Dec 01, 2016 at 11:24:23AM +, Frediano Ziglio wrote:
> SpicePlaybackState and SpiceRecordState have same structures
> changing only slightly the behaviour.
> Using SndWorker instead allows some minor simplification and
> more code
On Thu, Dec 01, 2016 at 05:06:42PM +0100, Victor Toso wrote:
> Hi,
>
> On Thu, Dec 01, 2016 at 04:33:36PM +0100, Christophe Fergeau wrote:
> > On Thu, Dec 01, 2016 at 02:15:49PM +0100, Victor Toso wrote:
> > > On Thu, Dec 01, 2016 at 11:37:32AM +0100, Christophe Fergeau wrote:
> > > > On Wed, Nov
Hi,
On Thu, Dec 01, 2016 at 06:17:41PM +0200, Uri Lublin wrote:
> On 11/30/2016 07:50 PM, Victor Toso wrote:
> > From: Victor Toso
> >
> > Failing to get playback or record volume async on startup is very
> > common making this warning too worrisome for users.
> >
> > If the
>
> On Thu, Dec 01, 2016 at 11:24:26AM +, Frediano Ziglio wrote:
> > @@ -1506,73 +1506,73 @@ static void remove_worker(SndWorker *worker)
> >
> > void snd_attach_playback(RedsState *reds, SpicePlaybackInstance *sin)
> > {
> > -SndWorker *playback_worker;
> > -RedChannel *channel;
On Thu, Dec 01, 2016 at 02:20:29PM +0100, Victor Toso wrote:
> Hi,
>
> On Thu, Dec 01, 2016 at 01:29:50PM +0100, Christophe Fergeau wrote:
> > On Wed, Nov 30, 2016 at 06:36:32PM +0100, Victor Toso wrote:
> > > From: Victor Toso
> > >
> > > SpiceSession does not initialize
On Thu, Dec 01, 2016 at 11:24:24AM +, Frediano Ziglio wrote:
> 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
On Thu, Dec 01, 2016 at 11:47:15AM -0500, Frediano Ziglio wrote:
> >
> > On Thu, Dec 01, 2016 at 11:24:26AM +, Frediano Ziglio wrote:
> > > @@ -1506,73 +1506,73 @@ static void remove_worker(SndWorker *worker)
> > >
> > > void snd_attach_playback(RedsState *reds, SpicePlaybackInstance *sin)
On Thu, 2016-11-24 at 08:57 -0500, Frediano Ziglio wrote:
> >
> >
> > On Thu, Nov 24, 2016 at 10:48:24AM +, Frediano Ziglio wrote:
> > >
> > > Signed-off-by: Frediano Ziglio
> > > ---
> > > server/red-channel-client.h | 3 +++
> > > 1 file changed, 3 insertions(+)
> >
On Thu, Dec 01, 2016 at 11:24:25AM +, Frediano Ziglio wrote:
> Make easier to understand that they refer to client and not
> all channel.
>
> Specifically:
> - RecordChannel -> RecordChannelClient
> - PlaybackChannel -> PlaybackChannelClient
> - playback_channel -> playback_client
> -
On Thu, Dec 01, 2016 at 11:24:26AM +, Frediano Ziglio wrote:
> @@ -1506,73 +1506,73 @@ static void remove_worker(SndWorker *worker)
>
> void snd_attach_playback(RedsState *reds, SpicePlaybackInstance *sin)
> {
> -SndWorker *playback_worker;
> -RedChannel *channel;
> +SndChannel
On Thu, Dec 01, 2016 at 11:24:29AM +, Frediano Ziglio wrote:
> This patch is quite huge but have some benefits:
> - reduce dependency (DummyChannel* and some RedChannelClient internals);
> - reuse RedChannelClient instead of reading from the RedsStream
> directly.
>
> You can see that
I'm a little bit confused. This appears to be the third patch series
that is designated "v4". That means this should actually be v6, no?
(although the previous series was sent exactly 1 minute before this
one, so maybe they're the same?)
On Thu, 2016-12-01 at 11:24 +, Frediano Ziglio wrote:
On Thu, 2016-12-01 at 11:24 +, Frediano Ziglio wrote:
> SndWorker has been historically based on RedChannel, initial git
> commit
> has:
> struct SndWorker {
> Channel base;
> ...
> };
I'd add a short note to the commit log explaining that this is only
possible since we just
On Thu, 2016-12-01 at 11:24 +, Frediano Ziglio wrote:
> 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
> ---
> server/sound.c
On Thu, 2016-12-01 at 11:24 +, Frediano Ziglio wrote:
> This code is the same inside __new_channel but will set the
> RedsStream from RedChannel.
Shouldn't this code be removed from __new_channel() then?
>
> Signed-off-by: Frediano Ziglio
> ---
> server/sound.c | 46
On Thu, 2016-12-01 at 17:21 +0100, Christophe Fergeau wrote:
> On Thu, Dec 01, 2016 at 11:24:29AM +, Frediano Ziglio wrote:
> >
> > This patch is quite huge but have some benefits:
> > - reduce dependency (DummyChannel* and some RedChannelClient
> > internals);
> > - reuse RedChannelClient
Hi,
On Thu, Dec 01, 2016 at 04:33:36PM +0100, Christophe Fergeau wrote:
> On Thu, Dec 01, 2016 at 02:15:49PM +0100, Victor Toso wrote:
> > On Thu, Dec 01, 2016 at 11:37:32AM +0100, Christophe Fergeau wrote:
> > > On Wed, Nov 30, 2016 at 06:50:39PM +0100, Victor Toso wrote:
> > > > From: Victor
On 11/30/2016 07:50 PM, Victor Toso wrote:
From: Victor Toso
Failing to get playback or record volume async on startup is very
common making this warning too worrisome for users.
If the audio back-end does not cache the last volume used or if this
is the first time the
Hi,
On Tue, Nov 22, 2016 at 06:01:38PM +0100, Francois Gouget wrote:
> Report the stream as invalid if the frames cannot be decoded. This will
> force the server to send regular screen updates instead.
>
> Signed-off-by: Francois Gouget
> ---
>
> Uses
Hi!
On Tue, Nov 22, 2016 at 06:01:48PM +0100, Francois Gouget wrote:
> 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
On Thu, Dec 01, 2016 at 04:27:36AM +0100, Francois Gouget wrote:
> On Wed, 23 Nov 2016, Christophe Fergeau wrote:
>
> > On Wed, Nov 23, 2016 at 07:09:28AM +0100, Francois Gouget wrote:
> > > For Spice-gtk most deprecation issues come from changes outside Spice
> > > (GLib) and thus should not be
On Thu, Dec 01, 2016 at 04:27:36AM +0100, Francois Gouget wrote:
>
>
>
> Btw, what's the status of this patchset?
Fine with me, I'll add a closing G_END_DEPRECATION_xxx to the end of
spicy.c and push. This adds a few more warnings in various places, but
they can be fixed afterwards (though I
I think I'd handle this one similarly to the vala warnings:
diff --git a/src/Makefile.am b/src/Makefile.am
index 3f81866..66e3c64 100644
--- a/src/Makefile.am
+++ b/src/Makefile.am
@@ -402,10 +402,12 @@ spicy_LDADD =
\
$(LIBM)
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
---
server/gstreamer-encoder.c | 2 +-
1 file
On Wed, Nov 30, 2016 at 06:36:32PM +0100, Victor Toso wrote:
> From: Victor Toso
>
> SpiceSession does not initialize its SpiceUsbDeviceManager object on
> startup that could lead to a race condition where channel-usbredir is
> requested to flush data while it is
On Wed, Nov 30, 2016 at 06:57:12PM +0100, Victor Toso wrote:
> Hi,
>
> On Wed, Nov 30, 2016 at 06:45:45PM +0100, Pavel Grunt wrote:
> > imho this should be in usbredir
> >
> > It is not nice when library crashes your program
>
> I don't disagree. Libraries should use guards accordingly.
>
> But
On Wed, Nov 30, 2016 at 06:36:33PM +0100, Victor Toso wrote:
> From: Victor Toso
>
> Similar fix was done in 5b252b0f499601bcf387c02a4dd35d27ed34c
>
> We should log a critical message instead of crashing the application.
I think the log should be more detailed than in the
These objects were used by the sound channel as
this channel read/write to/from client directly.
This make the code of this channel quite different
from the other ones.
Also this reduce code duplication from RedChannelClient
and increase encapsulation.
First set of patches attempt to rename
Signed-off-by: Frediano Ziglio
---
server/red-channel-client-private.h | 11 +++
server/red-channel-client.c | 12 ++--
server/red-channel-client.h | 13 -
3 files changed, 17 insertions(+), 19 deletions(-)
diff --git
This patch is quite huge but have some benefits:
- reduce dependency (DummyChannel* and some RedChannelClient internals);
- reuse RedChannelClient instead of reading from the RedsStream
directly.
You can see that SndChannelClient has much less field
as the code to read/write from/to client is
Signed-off-by: Frediano Ziglio
---
server/sound.c | 19 ---
1 file changed, 16 insertions(+), 3 deletions(-)
diff --git a/server/sound.c b/server/sound.c
index 14a6997..ebbe821 100644
--- a/server/sound.c
+++ b/server/sound.c
@@ -1425,10 +1425,26 @@
This code is the same inside __new_channel but will set the
RedsStream from RedChannel.
Signed-off-by: Frediano Ziglio
---
server/sound.c | 46 +-
1 file changed, 45 insertions(+), 1 deletion(-)
diff --git a/server/sound.c
Signed-off-by: Frediano Ziglio
---
server/sound.c | 19 ---
1 file changed, 16 insertions(+), 3 deletions(-)
diff --git a/server/sound.c b/server/sound.c
index 14a6997..ebbe821 100644
--- a/server/sound.c
+++ b/server/sound.c
@@ -1425,10 +1425,26 @@
SndWorker has been historically based on RedChannel, initial git commit
has:
struct SndWorker {
Channel base;
...
};
Signed-off-by: Frediano Ziglio
---
server/sound.c | 312 +-
1 file changed, 156 insertions(+), 156
SndWorker has been historically based on RedChannel, initial git commit
has:
struct SndWorker {
Channel base;
...
};
Signed-off-by: Frediano Ziglio
---
server/sound.c | 312 +-
1 file changed, 156 insertions(+), 156
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
This patch is quite huge but have some benefits:
- reduce dependency (DummyChannel* and some RedChannelClient internals);
- reuse RedChannelClient instead of reading from the RedsStream
directly.
You can see that SndChannelClient has much less field
as the code to read/write from/to client is
Signed-off-by: Frediano Ziglio
---
server/Makefile.am| 4 +-
server/dummy-channel-client.c | 138 +---
server/dummy-channel-client.h | 64 +
server/dummy-channel.c| 94 +
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
---
Almost identical beside the type.
Signed-off-by: Frediano Ziglio
---
server/sound.c | 53 +++
1 file changed, 16 insertions(+), 37 deletions(-)
diff --git a/server/sound.c b/server/sound.c
index 486fc95..0304fec 100644
---
Signed-off-by: Frediano Ziglio
---
server/sound.c | 51 +--
1 file changed, 21 insertions(+), 30 deletions(-)
diff --git a/server/sound.c b/server/sound.c
index 30a0c20..4c4b533 100644
--- a/server/sound.c
+++ b/server/sound.c
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
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
---
server/sound.c | 26 --
1 file changed,
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
---
red_channel_client_handle_message can handle base messages
so reuse it.
Signed-off-by: Frediano Ziglio
---
server/sound.c | 16 +---
1 file changed, 1 insertion(+), 15 deletions(-)
diff --git a/server/sound.c b/server/sound.c
index 4c4b533..1832e3b 100644
---
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
---
server/sound.c | 26 --
1 file changed,
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
---
server/sound.c | 224 +++---
1 file
Signed-off-by: Frediano Ziglio
---
server/red-channel-client-private.h | 11 +++
server/red-channel-client.c | 12 ++--
server/red-channel-client.h | 13 -
3 files changed, 17 insertions(+), 19 deletions(-)
diff --git
This code is the same inside __new_channel but will set the
RedsStream from RedChannel.
Signed-off-by: Frediano Ziglio
---
server/sound.c | 46 +-
1 file changed, 45 insertions(+), 1 deletion(-)
diff --git a/server/sound.c
These objects were used by the sound channel as
this channel read/write to/from client directly.
This make the code of this channel quite different
from the other ones.
Also this reduce code duplication from RedChannelClient
and increase encapsulation.
First set of patches attempt to rename
Signed-off-by: Frediano Ziglio
---
server/sound.c | 21 -
1 file changed, 21 deletions(-)
diff --git a/server/sound.c b/server/sound.c
index ebbe821..486fc95 100644
--- a/server/sound.c
+++ b/server/sound.c
@@ -804,25 +804,6 @@
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
---
server/sound.c | 37 +
1 file changed, 5 insertions(+), 32
On Thu, Dec 01, 2016 at 05:19:33AM +0100, Francois Gouget wrote:
> udscs_server_fill_fds() should accept being passed a NULL pointer.
I would reword the commit log a bit, like "udscs_server_fill_fds() is
dereferencing the 'server' pointer, and then checks if it's NULL. This
commit makes sure the
On Wed, Nov 30, 2016 at 06:40:26PM +0100, Pavel Grunt wrote:
> On Fri, 2016-11-25 at 14:52 +, Frediano Ziglio wrote:
> > Signed-off-by: Frediano Ziglio
> > ---
> > server/reds.c | 7 +++
> > 1 file changed, 7 insertions(+)
> >
> > diff --git a/server/reds.c
>
> Hi guys
>
> Are there some tools to evaluate the spice client performance.
>
>
There are so many stuff to take into account that there aren't
any official tool available. It depends on what you want to
measure really. CPU ? Memory ? Network bandwidth ?
Frediano
>
> Could be called in replay ? Or is it called indirectly ?
>
> Pavel
>
Yes and not. While I was writing this series spice-server-replay
was using it. However now this is done automatically by
"Free QXL instances in spice_server_destroy" so replay utility
has no reason to free the interface.
Signed-off-by: Frediano Ziglio
Acked-by: Pavel Grunt
---
server/tests/replay.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/server/tests/replay.c b/server/tests/replay.c
index f32fa2f..23d4125 100644
--- a/server/tests/replay.c
+++
Taking into account that in a following patch the object is destroyed
automatically with spice_server_destroy this patch get less important.
Currently QXL interfaces are added when physical card are added and
QXL cards don't support (in Qemu) hot swap so there is no reason to
have this feature.
There could be still some data pending.
Signed-off-by: Frediano Ziglio
Acked-by: Pavel Grunt
---
server/tests/replay.c | 19 +--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/server/tests/replay.c b/server/tests/replay.c
Signed-off-by: Frediano Ziglio
Acked-by: Pavel Grunt
---
server/red-replay-qxl.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/server/red-replay-qxl.c b/server/red-replay-qxl.c
index 2176068..aeaa545 100644
--- a/server/red-replay-qxl.c
+++
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
---
server/reds.c | 7 +++
1
Try to arrange destruction in the opposite order of the creation
Signed-off-by: Frediano Ziglio
---
server/display-channel.c | 4 ++--
server/reds.c| 24 ++--
2 files changed, 16 insertions(+), 12 deletions(-)
diff --git
Signed-off-by: Frediano Ziglio
---
server/reds.c | 7 +++
1 file changed, 7 insertions(+)
diff --git a/server/reds.c b/server/reds.c
index 17e5ada..d350331 100644
--- a/server/reds.c
+++ b/server/reds.c
@@ -3592,6 +3592,8 @@ SPICE_GNUC_VISIBLE void
These patches start destroying objects upon destruction.
Lot of the objects are just allocated but never freed.
This however make hardier to understand automatically leaks
as objects are still linked together.
Currently spice_server_destroy is not called so objects
are freed only when program ends
On Wed, Nov 30, 2016 at 12:34:52PM +, Frediano Ziglio wrote:
> Avoid having dandling pointer to a client.
'dangling'
>
> Signed-off-by: Frediano Ziglio
> ---
> server/sound.c | 9 +
> 1 file changed, 9 insertions(+)
>
> diff --git a/server/sound.c
>
> On Wed, Nov 30, 2016 at 12:34:52PM +, Frediano Ziglio wrote:
> > Avoid having dandling pointer to a client.
>
> 'dangling'
>
I though I merged this patch in "sound: Convert SndChannelClient to GObject".
I did some mistake probably, but will be merged to that patch anyway.
> >
> >
72 matches
Mail list logo