. The previous fd mechanism is still retained
to ensure a valid scanout buffer is always available for a new
client.
Greetings,
Michael
On 05.05.25 14:33, Michael Scherle wrote:
Hello,
Sorry for the late reply — I was on vacation. Thank you very much for
the detailed explanation; it gave me a
00, Frediano Ziglio wrote:
On Thu, Apr 10, 2025 at 3:18 PM Michael Scherle <
michael.sche...@rz.uni-freiburg.de> wrote:
Hello,
I’ve encountered an issue with the new DMA-BUF -> video encoding feature
in SPICE. When connecting, the first frame is only sent once the GPU
renders a new frame.
Hello,
I’ve encountered an issue with the new DMA-BUF -> video encoding feature
in SPICE. When connecting, the first frame is only sent once the GPU
renders a new frame. However, this can take quite some time if the VM is
idle (e.g., sitting on the desktop), since the GPU only renders a new
f
Thank you very much. I apologize for the not quite clean merge request
by email, this is my first time. Besides that are there any suggestions
for improvements to the other patches?
Thanks Michael
On 02.03.25 18:40, Frediano Ziglio wrote:
On Fri, Feb 28, 2025 at 9:59 AM Michael Scherle
Prevent a freeze that occurs if the video stream is stopped while a
gl draw is in progress (e.g., when the client requests a new codec).
Ensure proper cleanup of the ongoing gl draw.
Signed-off-by: Michael Scherle
---
server/dcc-send.cpp | 4
1 file changed, 4 insertions(+)
diff --git a
From: Vivek Kasireddy
We do not want to stop a stream associated with gl_draw as a result
of timeout because we may not get another opportunity to create a
new stream if the current one gets stopped. However, when the
stream does get stopped for other reasons, we need to clear the
gl_draw_stream
ver/display-channel.cpp:2074:display_channel_create_surface:
condition `!display->priv->surfaces[surface_id]' failed
Rebase all patches on master
v2:
Update all patches to address review comments from Frediano
Tested this series with msdkh264enc/dec plugins that will be added
in anothe
From: Vivek Kasireddy
We need to determine if the client is new enough to support multiple
codecs -- which might include any of the Gstreamer based ones.
v2: (suggestions and fixups from Frediano)
- Add is_gl_client() method to DisplayChannelClient instead of a
dcc_is_gl_client() function.
- A
From: Vivek Kasireddy
For non-gl/remote clients, if there is no stream associated with
the DisplayChannel, then we create a new stream. Otherwise, we
just update the current stream's timestamp.
v2: (suggestions and fixups from Frediano)
- Moved the gl_draw_stream object from DCC to DC
- Moved th
add dmabuf encoding if `drm/drm_fourcc.h` is present and
gstreamer is at least 1.24 due to
`gst_video_dma_drm_fourcc_to_format()`.
Signed-off-by: Michael Scherle
---
meson.build| 10 +-
server/gstreamer-encoder.c | 34 +++---
2 files changed
From: Vivek Kasireddy
We need to convert the scanout's drm format to the correct Gstreamer
format while configuring the pipeline. This can be done using
gst_video_dma_drm_fourcc_to_format() API, which will take the drm
fourcc value and return the appropriate Gst format.
Signed-off-by: Vivek Kasi
From: Vivek Kasireddy
Once it is determined that an Intel GPU is available/active (after
looking into udev's database), we try to see if there is a h/w
based encoder (element) available (in Gstreamer's registry cache)
for the user selected video codec. In other words, if we find that
the Intel Me
From: Vivek Kasireddy
This patch adds a new function to enable the creation of Gst memory with
the dmabuf fd as the source by using a dmabuf allocator. And, it also
adds a mechanism to register and invoke any callbacks once the Gst memory
object is no longer used by the pipeline.
This patch also
From: Vivek Kasireddy
For remote (or non-gl) clients, if a valid gl_draw stream exists,
then we first extract the dmabuf fd associated with the scanout and
share it with the encoder along with other key parameters such as
stride, width and height. Once the encoder finishes creating an
encoded buf
This brings in the following changes:
common: Add a udev helper to identify GPU Vendor
build: Avoid Meson warning
Drop Python 2 from m4/spice-deps.m4
Stop using Python six package
codegen: Use context manager when opening files
Signed-off-by: Michael Scherle
> Regards,
> Frediano
>
Greetings
Michael
On 08.09.24 18:43, Frediano Ziglio wrote:
Il giorno ven 6 set 2024 alle ore 10:43 Michael Scherle
ha scritto:
Hi,
thank you very much for your reply. Of course I understand that
non-compiling/crashing code should not be merged into main
o
Il giorno ven 30 ago 2024 alle ore 12:53 Michael Scherle
ha scritto:
On 10.06.24 20:34, Vivek Kasireddy wrote:
For clients that cannot accept a dmabuf fd directly (such as those
running on a remote system), this patch series provides a way for
the Spice server to stream the gl/dmabuf data/b
el Flex
140 GPU. Note this extra patches are related to the gpu and not this
patchset.
This patch is a significant improvement for handling graphically
demanding tasks or rapidly changing image content and is crucial for
SPICE to be a component of a competitive virtual desktop infrastructure.
On 29.07.24 13:56, Michael Scherle wrote:
Hello,
with the hopefully soon merged pull request: “[PATCH v8 0/6] dcc: Create
a stream for non-gl/remote clients that want to use dmabuf (v8)” (which
we urgently need for our project), I have noticed the problem of chroma
supsampling. The
Hi Vivek,
Hi Michael,
Hi,
Disclaimer I am not a Spice developer. But we are very interested in
this patch series for our project, so I have tested it extensively.
For Qemu I used your vfio_dmabuf_2 branch.
First I tested it with intel gvt-g and it works perfectly.
In the next test, I used
, Apr 16, 2024 at 12:59:50PM GMT, Michael Scherle wrote:
Hello,
Thanks for your changesets, they definitely reduce the delay significantly
(to a similar level as our provosoric fixes, but yours are much cleaner).
On the client side (spice-gtk) I looked at the problem with the high
decoding time (2
_DELAY,
which is not mentioned anywhere. Is this the only use to save some
network traffic? Is there any reason not to always set this to 0 in
today's network environments? (And maybe set the default to 0?)
Michael
On 03.04.24 21:22, Frediano Ziglio wrote:
Frediano
Il giorno mar 2 apr 20
n for this.
When testing stuff out we saw that Sunshine/Moonlight performed very
well in generating a low delay and high QoE. That is kind of our
benchmark for remote access to strive for :)
Greetings
Michael
On 15.03.24 14:08, Michael Scherle wrote:
Hello spice developers,
we are trying to
en me.
On 15.03.24 14:08, Michael Scherle wrote:
Hello spice developers,
we are trying to develop an Open Source virtual desktop infrastructure
to be deployed at multiple German universities as described, by my
colleagues, in the paper which I have put in the attachment. The
solution bas
24 matches
Mail list logo
