Here is an small guide to test this for the ones that want to check it
faster:
Add a spiceport to a virtual machine adding this under devices in the XML
Boot machine, Connect to it as usual using spice-html5 and as root:
echo "CHILI" > /dev/virtio-ports/org.freedeskt
In testing audio in conjunction with video playback, I made a number
of fairly modest improvements to the spice-html5 streaming support.
This patch set provides those; with this set, and the prior audio set,
you can arguably play videos using spice-html5 and the result is usable.
Jeremy White (5)
Otherwise, we end up discarding stream data messages, and our decode
can become corrupted, notably on Chrome.
This way, we should not lose any messages while we are waiting for
source buffer creation.
Signed-off-by: Jeremy White
---
display.js | 23 ++-
1 file changed, 10 in
Signed-off-by: Jeremy White
---
display.js | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/display.js b/display.js
index 9fc13c0..6c6b962 100644
--- a/display.js
+++ b/display.js
@@ -535,7 +535,10 @@ SpiceDisplayConn.prototype.process_channel_message =
function(msg)
Signed-off-by: Jeremy White
---
display.js | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/display.js b/display.js
index d9bdd9f..e26bff8 100644
--- a/display.js
+++ b/display.js
@@ -963,8 +963,8 @@ function handle_draw_jpeg_onload()
this.o.sc.surfaces[this.o.base
Signed-off-by: Jeremy White
---
spiceconn.js | 27 +--
1 file changed, 17 insertions(+), 10 deletions(-)
diff --git a/spiceconn.js b/spiceconn.js
index 9651b47..33e7388 100644
--- a/spiceconn.js
+++ b/spiceconn.js
@@ -127,21 +127,28 @@ SpiceConn.prototype =
)
Instead, let's set it to an empty gif. Should have the
same benefit.
Signed-off-by: Jeremy White
---
display.js | 2 +-
utils.js | 5 +
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/display.js b/display.js
index 6c6b962..d9bdd9f 100644
--- a/display.js
+++ b/display.js
@@
Around version 45, Firefox started being very particular about the
time stamps put into the Opus stream. The time stamps from the Spice server are
somewhat irregular. They mostly arrive every 10 ms, but sometimes it is 11, or
sometimes
with two time stamps the same in a row. The previous logic
Firefox and Chrome have both drifted somewhat over the past
few years, and audio playback is no longer as robust in spice-html5
as it once was.
Further, the MediaSource specification is a bit more stable than
it was when the initial implementation was written.
This patch set tunes the spice-html5
This allows audio processing to be more robust in Firefox.
Signed-off-by: Jeremy White
---
playback.js | 17 -
1 file changed, 16 insertions(+), 1 deletion(-)
diff --git a/playback.js b/playback.js
index 55dc0e8..48d4cdd 100644
--- a/playback.js
+++ b/playback.js
@@ -158,7 +158,
We previously had a fairly crude 'gap detection', but browsers
are largely supporting the modern buffer semantics, and we can
now rely on them to detect under runs for us.
This improves audio in Chrome while playing large videos.
Signed-off-by: Jeremy White
---
playback.js | 30
Browsers now reliably autoplay, and issuing play prematurely
can generate an error in both Firefox and Chrome.
Hence we can simply remove this call and audio seems to work fine.
Signed-off-by: Jeremy White
---
playback.js | 3 ---
1 file changed, 3 deletions(-)
diff --git a/playback.js b/playb
Looks good to me. I don't have the time to test this myself, so I'll
leave this open for others to kibitz for now.
I don't think it will do harm, so if no one else expresses an opinion
after a while, you can probably nudge me into pushing it.
Reviewed-by: Jeremy White
Cheers,
Jeremy
On 09/27
Acked and pushed.
Cheers,
Jeremy
On 09/27/2016 10:29 AM, Oliver Gutierrez wrote:
> ---
> README | 4 ++--
> TODO| 2 +-
> cursor.js | 2 +-
> display.js | 36 ++--
> enums.js| 2 +-
> main.js
>
> On Tue, 2016-09-27 at 05:50 -0400, Frediano Ziglio wrote:
> > >
> > >
> > > Convert the RedChannelClient heirarchy into GObjects. Since the
> > > existing
> > > constructors could fail and return NULL, I inherited the base
> > > channel
> > > client from GInitable, which introduces a depende
On Tue, 2016-09-27 at 05:50 -0400, Frediano Ziglio wrote:
> >
> >
> > Convert the RedChannelClient heirarchy into GObjects. Since the
> > existing
> > constructors could fail and return NULL, I inherited the base
> > channel
> > client from GInitable, which introduces a dependency on gio.
> >
>
> This is extremely dangerous!, this can be working without any errors most of
> the time but in a rare cases this could
> cause a crash - in case a non pageable function made pageable by this patch -
> which can be hard to troubleshoot!
Actually this is not far from what the current code is doing
This is extremely dangerous!, this can be working without any errors most
of the time but in a rare cases this could
cause a crash - in case a non pageable function made pageable by this patch
- which can be hard to troubleshoot!
On Wed, Sep 28, 2016 at 4:55 PM, Frediano Ziglio wrote:
> This mak
> On Wed, Sep 28, 2016 at 3:06 PM, Frediano Ziglio < fzig...@redhat.com >
> wrote:
> > >
>
> > > Based on a patch by Sandy Stutsman < sstut...@redhat.com >
>
> > >
>
> > > Signed-off-by: Sameeh Jubran < sam...@daynix.com >
>
> > Acked-by: Frediano Ziglio < fzig...@redhat.com >
>
> > > ---
>
The SetClip function is not currently called by
the code however it contains a leak as the missing
call to RELEASE_RES cause the reference counter
to be 2 so won't be freed when resources are freed.
Signed-off-by: Frediano Ziglio
---
qxldod/QxlDod.cpp | 1 +
1 file changed, 1 insertion(+)
diff
Signed-off-by: Frediano Ziglio
---
qxldod/QxlDod.cpp | 11 ---
qxldod/QxlDod.h | 1 -
2 files changed, 12 deletions(-)
diff --git a/qxldod/QxlDod.cpp b/qxldod/QxlDod.cpp
index 4b27b9c..2294a81 100755
--- a/qxldod/QxlDod.cpp
+++ b/qxldod/QxlDod.cpp
@@ -3916,17 +3916,6 @@ void QxlDevice
Disable execution bit on mapping improving security.
MmMapIoSpaceEx is available only in Windows 10 so
to provide binary compatibility detect the new
function dynamically.
Added a separate compatibility file.
Based on a patch by Sandy Stutsman
Signed-off-by: Sameeh Jubran
Signed-off-by: Fredia
From: Sameeh Jubran
Remove specifier that was causing a warning.
Based on a patch by Sandy Stutsman
Signed-off-by: Sameeh Jubran
---
qxldod/QxlDod.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/qxldod/QxlDod.cpp b/qxldod/QxlDod.cpp
index a9a6cd4..78982a7 100755
--- a
Signed-off-by: Frediano Ziglio
---
qxldod/QxlDod.cpp | 3 +++
1 file changed, 3 insertions(+)
diff --git a/qxldod/QxlDod.cpp b/qxldod/QxlDod.cpp
index 5cfff78..d7830e5 100755
--- a/qxldod/QxlDod.cpp
+++ b/qxldod/QxlDod.cpp
@@ -2930,6 +2930,7 @@ VOID VgaDevice::ResetDevice(VOID)
NTSTATUS VgaDe
First two patches replace the previous "Using declspec instead
of #pragma to declare pageable" which although correct was
unexpectedly large.
Third is the updated (comment) one for dynamic MmMapIoSpaceEx
detection.
The is one missing hunk from a previous patch
("Code Analysis fix").
Others are m
Signed-off-by: Frediano Ziglio
---
qxldod/driver.cpp | 7 ---
qxldod/driver.h | 1 -
2 files changed, 8 deletions(-)
diff --git a/qxldod/driver.cpp b/qxldod/driver.cpp
index eafeb1e..f660304 100755
--- a/qxldod/driver.cpp
+++ b/qxldod/driver.cpp
@@ -661,13 +661,6 @@ void DebugPrintFuncSeri
This make sure that these function goes into the non
paged section.
Also this change is much shorter than marking code if
can be paged as almost code can be paged.
It's also coherent with other source files.
Signed-off-by: Frediano Ziglio
---
qxldod/QxlDod.cpp | 39 --
Signed-off-by: Frediano Ziglio
---
qxldod/driver.cpp | 12
qxldod/driver.h | 13 +++--
2 files changed, 11 insertions(+), 14 deletions(-)
diff --git a/qxldod/driver.cpp b/qxldod/driver.cpp
index f660304..f0d98c0 100755
--- a/qxldod/driver.cpp
+++ b/qxldod/driver.cpp
@@ -63
>
> On Wed, Sep 28, 2016 at 02:09:26PM +0300, Sameeh Jubran wrote:
> > This patch replaces the "#pragma code_seg(push)" and "#pragma
> > code_seg(pop)"
> > with __declspec(code_seg("PAGE")) for declaring code as pageable. This
> > change enahnces maintainability as it is much clearer which functio
On Wed, Sep 28, 2016 at 4:30 PM, Christophe Fergeau
wrote:
> On Wed, Sep 28, 2016 at 02:09:26PM +0300, Sameeh Jubran wrote:
> > This patch replaces the "#pragma code_seg(push)" and "#pragma
> code_seg(pop)"
> > with __declspec(code_seg("PAGE")) for declaring code as pageable. This
> > change enah
On Wed, Sep 28, 2016 at 02:09:26PM +0300, Sameeh Jubran wrote:
> This patch replaces the "#pragma code_seg(push)" and "#pragma code_seg(pop)"
> with __declspec(code_seg("PAGE")) for declaring code as pageable. This
> change enahnces maintainability as it is much clearer which functions are
> paged
On Wed, Sep 28, 2016 at 3:06 PM, Frediano Ziglio wrote:
> >
> > Based on a patch by Sandy Stutsman
> >
> > Signed-off-by: Sameeh Jubran
>
> Acked-by: Frediano Ziglio
>
> > ---
> > qxldod/QxlDod.cpp | 9 +
> > 1 file changed, 5 insertions(+), 4 deletions(-)
> >
> > diff --git a/qxldod/
>
> Disable execution bit on mapping improving security.
>
> MmMapIoSpaceEx is available only in Windows 10 thus
> the macros are used.
>
> Based on a patch by Sandy Stutsman
>
> Signed-off-by: Sameeh Jubran
Just realized that the comment is out of date:
Disable execution bit on mapping im
Disable execution bit on mapping improving security.
MmMapIoSpaceEx is available only in Windows 10 thus
the macros are used.
Based on a patch by Sandy Stutsman
Signed-off-by: Sameeh Jubran
---
qxldod/QxlDod.cpp | 16 ++-
qxldod/compat.cpp | 62
>
> Based on a patch by Sandy Stutsman
>
> Signed-off-by: Sameeh Jubran
Acked-by: Frediano Ziglio
> ---
> qxldod/QxlDod.cpp | 9 +
> 1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/qxldod/QxlDod.cpp b/qxldod/QxlDod.cpp
> index 2bc6b77..577b4ff 100755
> --- a/qxldod
>
> The PAGED_CODE macro ensures that the calling thread is running at an
> IRQL that is low enough to permit paging. A call to this macro should be
> made at the beginning of every driver routine that either contains
> pageable code or accesses pageable code
>
> Based on a patch by Sandy Stutsma
>
> Part of source image mapped by PresentDisplayOnly
> should be big enough to cover all rectangles being
> transferred.
>
> Signed-off-by: Sameeh Jubran
> Signed-off-by: Dmitry Fleytman
> ---
> qxldod/QxlDod.cpp | 28 +++-
> qxldod/QxlDod.h | 2 ++
> 2 files change
Based on a patch by Sandy Stutsman
Signed-off-by: Sameeh Jubran
---
qxldod/QxlDod.cpp | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/qxldod/QxlDod.cpp b/qxldod/QxlDod.cpp
index 2bc6b77..577b4ff 100755
--- a/qxldod/QxlDod.cpp
+++ b/qxldod/QxlDod.cpp
@@ -1493,7 +1493
The PAGED_CODE macro ensures that the calling thread is running at an
IRQL that is low enough to permit paging. A call to this macro should be
made at the beginning of every driver routine that either contains
pageable code or accesses pageable code
Based on a patch by Sandy Stutsman
Signed-off-
Part of source image mapped by PresentDisplayOnly
should be big enough to cover all rectangles being
transferred.
Signed-off-by: Sameeh Jubran
Signed-off-by: Dmitry Fleytman
---
qxldod/QxlDod.cpp | 28 +++-
qxldod/QxlDod.h | 2 ++
2 files changed, 25 insertions(+), 5
This patch replaces the "#pragma code_seg(push)" and "#pragma code_seg(pop)"
with __declspec(code_seg("PAGE")) for declaring code as pageable. This
change enahnces maintainability as it is much clearer which functions are paged.
Signed-off-by: Sameeh Jubran
---
qxldod/QxlDod.cpp | 140 ++
The patches were applied on the up-to-date master branch of
https://gitlab.com/spice/qxl-wddm-dod
Diffrences from v5:
* The patch Code Analysis fix was splitted into three separate patches
* Minor fixies in patch "Fix source buffer mapping in
PresentDisplayOnly"
Sameeh Jubran (4):
Use PAGED_CO
> On Tue, Sep 27, 2016 at 3:24 PM, Frediano Ziglio < fzig...@redhat.com >
> wrote:
> > Disable execution bit on mapping improving security.
>
> > MmMapIoSpaceEx is available only in Windows 10 thus
>
> > the macros are used.
>
> > Based on a patch by Sandy Stutsman < sstut...@redhat.com >
>
> On Mon, Sep 26, 2016 at 7:15 PM, Frediano Ziglio < fzig...@redhat.com >
> wrote:
> > >
>
> > > This patch fixes 2 issues:
>
> > >
>
> > > 1. Framebuffer should only be used in vga mode,
>
> > > therefore when QxlDevice is active
>
> > > FrameBufferIsActive flag shouldn't be checked;
>
> >
On Mon, Sep 26, 2016 at 6:20 PM, Frediano Ziglio wrote:
> >
> > Part of source image mapped by PresentDisplayOnly
> > should be big enough to cover all rectangles being
> > transferred.
> >
> > Signed-off-by: Sameeh Jubran
> > Signed-off-by: Dmitry Fleytman
> > ---
> > qxldod/QxlDod.cpp | 28 +
On Tue, Sep 27, 2016 at 3:24 PM, Frediano Ziglio wrote:
> Disable execution bit on mapping improving security.
>
> MmMapIoSpaceEx is available only in Windows 10 thus
> the macros are used.
>
> Based on a patch by Sandy Stutsman
>
> Signed-off-by: Sameeh Jubran
> ---
> qxldod/QxlDod.cpp
On Mon, Sep 26, 2016 at 7:15 PM, Frediano Ziglio wrote:
> >
> > This patch fixes 2 issues:
> >
> > 1. Framebuffer should only be used in vga mode,
> > therefore when QxlDevice is active
> > FrameBufferIsActive flag shouldn't be checked;
> > 2. FrameBufferIsActive flag should be set
47 matches
Mail list logo
