fbe8631add
raw: Fix worker data dependencies
2022-11-05 13:22:14 +00:00
935ce4e98d
Remove push/pull encoder interface
...
Think of it as a failed experiment.
2022-11-05 10:15:56 +00:00
036b549fd8
Record authenticated username for each connected client
...
Signed-off-by: Jim Ramsay <jramsay@redhat.com>
2022-11-04 09:41:14 +00:00
86bd2ced85
Record hostname for each connected client
...
Signed-off-by: Jim Ramsay <jramsay@redhat.com>
2022-11-04 09:41:14 +00:00
73e1089f06
zrle: Unref result in destroy
2022-10-30 13:53:10 +00:00
b525608a06
raw: Unref result in destroy
2022-10-30 13:52:52 +00:00
b3c1d5d1dc
stream: Use MSG_NOSIGNAL
...
Without MSG_NOSIGNAL, sending to closed sockets will generate SIGPIPE.
2022-10-30 13:07:13 +00:00
95742676c9
zrle: Keep reference to encoder while encoding
2022-10-30 12:44:05 +00:00
91dc5da243
raw: Keep reference to encoder while encoding
2022-10-30 12:44:05 +00:00
727b2f727e
tight: Keep reference to encoder while encoding
2022-10-30 12:44:05 +00:00
8f2d137046
server: Make encoder inert when closing client
2022-10-30 12:44:05 +00:00
baaf84eab9
Reference count encoders
2022-10-30 12:44:05 +00:00
48b070af5a
server: Free cursor buffers on close
2022-10-30 12:44:05 +00:00
adef210252
fb-pool: Add setter for fb allocator
2022-10-29 18:40:02 +00:00
efb5ab956c
Add debug logging for tcp address binding
2022-10-24 20:55:38 +00:00
3588670c81
server: Replace abort() with NVNC_LOG_PANIC
2022-10-24 09:16:59 +00:00
3c4a069ba2
tight: Disable chroma subsampling at q=9
...
This avoids color banding at the highest quality level.
2022-10-23 19:05:32 +00:00
df84f371fe
Fix fallthrough warnings in murmurhash
...
Fix two -Wimplicit-fallthrough warnings in the murmurhash function:
../src/murmurhash.c: In function 'murmurhash':
../src/murmurhash.c:71:15: warning: this statement may fall through [-Wimplicit-fallthrough=]
71 | case 3: k ^= (tail[2] << 16);
| ~~^~~~~~~~~~~~~~~~~~
../src/murmurhash.c:72:5: note: here
72 | case 2: k ^= (tail[1] << 8);
| ^~~~
../src/murmurhash.c:72:15: warning: this statement may fall through [-Wimplicit-fallthrough=]
72 | case 2: k ^= (tail[1] << 8);
| ~~^~~~~~~~~~~~~~~~~
../src/murmurhash.c:74:5: note: here
74 | case 1:
| ^~~~
2022-10-15 10:00:43 +00:00
5b2a062f0e
don't resize an encoder if it is not set
2022-10-15 09:59:05 +00:00
b932f3e2e0
h264-encoder: Set async_depth=1
...
This fixes stalling during encoding. The FFmpeg devs seem to think that it's
normal to change the default behaviour or their code, so this needs to be
fixed here instead.
Fixes #73
2022-09-10 15:48:15 +00:00
19538c9435
server: Fix encoding selection for sw frames
...
This fixes encoding selection when not using the --gpu option. Before this
change, raw encoding would always be selected.
Reported-by: Consolatis
Suggested-by: Consolatis
2022-08-23 22:34:56 +00:00
c0e1159a53
Revert "h264-encoder: Add 30 bit color depth formats"
...
This reverts commit 613761cf5f
2022-08-23 21:13:13 +00:00
5e5d5d6e29
resampler: Use transformed width as destination stride
...
Fixes #72
2022-08-23 20:23:12 +00:00
affd7f3f6d
Allow to query client-side cursor support
...
Add a function nvnc_client_supports_cursor() to enable the API user to
make an informed decision whether nvnc_set_cursor() can be expected to
make the client draw the cursor, or whether it has to be rendered into
the framebuffer.
2022-07-29 08:37:29 +00:00
263990f5ef
stream: Remove stray ampersand in tls handshake failure code path
2022-07-17 20:17:19 +00:00
b9a5b9a3f1
h264: Set quality according to client's wishes
2022-07-10 13:53:32 +00:00
528eac51a3
Fix jpeg quality setting
...
The encoding identifiers are supposed to be interpreted as a range. I
interpreted the upper and lower limits as two discrete quality settings
instead, which is wrong.
2022-07-10 12:41:20 +00:00
4baeaa43fd
server: add missing header after f20ffb5e1e
...
src/server.c:1119:17: error: use of undeclared identifier 'IPPROTO_TCP'
setsockopt(fd, IPPROTO_TCP, TCP_NODELAY, &one, sizeof(one));
^
2022-07-09 20:04:53 +00:00
d85347cfe3
h264-encoder: Set quality
...
This is to get libav to stop complaining about quality not being set
2022-07-09 17:19:46 +00:00
362918a8cf
Integrate libav into logging framework
2022-07-09 17:15:35 +00:00
863fb0ce80
h264-encoder: Add dtrace probes
2022-07-09 16:30:25 +00:00
f20ffb5e1e
server: Turn off Nagle's algorithm
2022-07-09 16:29:18 +00:00
d86af88573
server: Log encoder choices
2022-07-09 16:28:09 +00:00
dc82b3f29d
h264-encoder: Clean up current_packet in destroy
2022-06-29 11:39:46 +00:00
b8122725e2
server: Reset encoder callback on frame done
2022-06-29 11:36:05 +00:00
4eb7a3a559
h264-encoder: Call on_packet_ready last in on_work_done
...
The callback can result in the encoder being destroyed, so we can't
dereference access the encoder object after calling it.
2022-06-29 11:34:12 +00:00
e8f2481aa4
open-h264: unref result when finished with it
2022-06-26 14:12:37 +00:00
8bde568112
server: Add dtrace probes with pts
2022-06-26 13:57:32 +00:00
1e3819b6fb
open-h264: Add dtrace probes
2022-06-26 13:57:32 +00:00
e05e812d93
resampler: Add dtrace probes
2022-06-26 13:57:32 +00:00
ee64bc2722
display: Add dtrace probes
2022-06-26 13:57:32 +00:00
4e9eb98301
open-h264: Convert to normal encoder
...
The push/pull code path is buggy.
2022-06-26 13:56:17 +00:00
e6a12ffda7
server: Promote some log levels
2022-06-25 16:32:05 +00:00
ad4a834cfc
Use new logging system
2022-06-25 16:15:32 +00:00
45da0fc157
Add simple logging system
2022-06-25 16:15:32 +00:00
9285594e9d
Call encoder.on_done in a function
2022-06-11 11:44:51 +00:00
d4258a0aab
cursor: Handle rotated cursors
2022-06-11 11:44:51 +00:00
bab78857e1
resampler: Extract function: resample_now()
2022-06-11 11:44:51 +00:00
7205139b11
Don't call gbm symbols when they are not available
2022-05-24 22:55:34 +00:00
da290abc25
server: consistently use builtin byteswap after 53f88894d5
...
src/server.c:48:10: fatal error: 'byteswap.h' file not found
#include <byteswap.h>
^~~~~~~~~~~~
2022-04-24 17:56:03 +00:00
53f88894d5
Add presentation timestamps
2022-04-14 18:10:09 +00:00
e2e117b02f
h264-encoder: Fully flush output packets
2022-04-10 16:21:28 +00:00
3ea068b90c
server: Decrement n_pending_requests after dispathing encoding job
...
The request is as good as handled at that point. We don't want to squeeze
in another frame that the client didn't request before the encoding job
finishes. That would cause a negative pending count and the client would
stop getting updates.
2022-04-03 20:52:24 +00:00
9d8c956983
h264-encoder: Fix copy-pasta
2022-03-06 11:24:33 +00:00
aca09358ea
Add back damage argument to nvnc_set_cursor
...
It's better to keep feeding buffers to keep buffers in rotation for damage
tracking purposes.
2022-02-21 21:43:41 +00:00
9c02e6afaf
Implement hiding cursors
2022-02-20 14:48:24 +00:00
1553c88f5e
Add width and height arguments to nvnc_set_cursor
2022-02-19 23:06:15 +00:00
afc0f018da
cursor: Use the right scan-line length when width != stride
2022-02-19 21:51:24 +00:00
cb282c57c4
cursor: Map cursor buffers before access
2022-02-13 18:44:27 +00:00
c876b91541
pixels: Add function to get rfb pixel format name
2022-02-12 13:15:30 +00:00
0e0fe5b73a
pixels: Add function to convert drm format to string
2022-02-12 12:30:12 +00:00
70784e1bcc
cursor: Fix alpha mask stride
2022-02-10 21:53:33 +00:00
9a292afd52
pixels: Handle different endianness for alpha mask
2022-02-10 21:53:33 +00:00
48baf74560
Implement client side cursor rendering
2022-02-06 16:33:46 +00:00
258dccd768
Add a cursor encoder
2022-02-06 16:33:46 +00:00
0cc6be091c
pixels: Add function to extract alpha mask
2022-02-06 15:01:33 +00:00
613761cf5f
h264-encoder: Add 30 bit color depth formats
2022-01-27 22:31:32 +00:00
648255769a
pixels: Add 10-bits-per-colour formats
2022-01-27 22:11:51 +00:00
fd23cb8c2f
enc-util: Round up division in calc_bytes_per_cpixel
...
Otherwise 10 bit formats will be mistaken for 8 bit formats.
2022-01-27 22:10:37 +00:00
5dc6a28828
h264-encoder: Automatically find a render node
2021-12-26 13:10:41 +00:00
5a75fdf2bc
display: Only run damage refinery when it's needed
2021-12-26 13:10:41 +00:00
c5a5437a9e
server: Keep a count of clients that use damage
2021-12-26 13:10:41 +00:00
55beea3464
open-h264: Set "ignores-damage" flag
2021-12-26 13:10:41 +00:00
90f61f03c6
Plug open h264
2021-12-26 13:10:41 +00:00
0a70f7fa6a
encoder: Add push/pull encoder interface
2021-12-26 13:10:41 +00:00
e1ba4e1085
API: Add nvnc_fb_get_type
2021-12-26 13:10:41 +00:00
15c14d7d4b
Create an Open h.264 encoder
2021-12-26 13:10:41 +00:00
b71598b334
Create h264-encoder
2021-12-26 13:10:41 +00:00
65c0e91c37
Move update header out of encoders
2021-12-12 16:05:29 +00:00
8b2c81c3dd
Add offset coordinates to encoders
2021-12-11 21:55:56 +00:00
42b102df0c
Remove unused headers
2021-12-11 21:07:41 +00:00
a7241658b0
Create encoder abstraction
2021-12-11 21:03:18 +00:00
66942ab913
display: Clean up transformed damage region
...
This fixes a memory leak
2021-12-11 20:58:17 +00:00
783ac9d99d
fb: Unmap released fbs
2021-09-20 22:06:08 +00:00
10c0b9131c
fb: Fix mapped gbm buffer stride
2021-09-20 21:57:56 +00:00
26ff812ea6
Add damage refinery from wayvnc
2021-09-20 21:40:21 +00:00
ff3dc13f0b
resampler: Track buffer damage
2021-09-20 21:40:21 +00:00
943bd33993
resampler: Make resampler object opaque
2021-09-20 21:40:21 +00:00
3b24dbd6a4
resampler: Transform output buffer dimensions based on input transform
2021-09-20 21:40:21 +00:00
0d4ab56568
transform-util: Add dimensions transform function
2021-09-19 21:11:54 +00:00
02559a7f7e
Re-sample transformed framebuffers
2021-09-19 20:12:30 +00:00
691e835d1b
fb: Add transform attribute
2021-09-19 19:56:56 +00:00
784af9fa5d
Add transform utility functions from wayvnc
2021-09-19 19:54:59 +00:00
dad7312814
pixels: Add fourcc_to_pixman_fmt
2021-09-19 19:54:23 +00:00
cf42f76f56
Add gbm_bo nvnc_fb type
2021-09-12 18:51:22 +00:00
a14b829743
fb: Fix buffer allocation
...
Width and height got mixed up.
2021-09-11 18:17:01 +00:00
dad1948e98
Remove nvnc_fb_flags
2021-09-05 00:46:24 +00:00
b75eeac03d
Add API function for creating nvnc_fb from an pre-allocated buffer
2021-09-05 00:46:24 +00:00
96886e21d5
Add a API function to get the pixel size of nvnc_fb
2021-09-05 00:46:24 +00:00
c7dd062498
Add a stride parameter to nvnc_fb
2021-09-05 00:46:24 +00:00
e8e4a9469a
Remove damage checker
...
There's a much better one in wayvnc
2021-09-05 00:46:24 +00:00
2095913688
display & fb_pool: Clean up memory leaks
2021-09-04 21:21:23 +00:00
77d8efcbe3
fb_pool: Return true from resize when dimensions change
2021-09-04 21:21:23 +00:00
1b7b51af44
Add a cleanup callback to nvnc_set_userdata
2021-09-04 21:21:23 +00:00
41c9ebe960
server: process fb update requests in fb update request message handler
...
Otherwise, the client won't get the initial frame
2021-09-04 21:21:23 +00:00
e16a64a67d
fb_pool: Fix reference counting error
2021-09-04 21:21:23 +00:00
f5b0f508f0
fb: Use special context pointer for release callback
2021-09-04 21:21:23 +00:00
f566105ab5
Remove nvnc_set_render_fn
2021-09-04 21:21:23 +00:00
981256d8d5
Add a buffer pool
2021-09-04 21:21:23 +00:00
d63feadeab
Notify the user when an fb is released
2021-09-04 21:21:23 +00:00
031555c85d
fb: Add hold/release logic
2021-09-04 21:21:23 +00:00
b320723049
Support UNIX sockets
...
Adds support for UNIX domain sockets with `nvnc_open_unix()` function.
Closes #1 .
2021-04-04 20:24:33 +00:00
1359b987a3
Translate XT keycodes to linux keycodes
2020-11-29 20:48:31 +00:00
f547ed72f9
Decrement pending requests when sending pseudo frames
2020-11-29 18:36:48 +00:00
ea98af75de
Notify client about qemu key capability
2020-11-29 18:36:48 +00:00
dbe35bc3ad
Add qemu extended key events
2020-11-29 18:36:48 +00:00
e69006fc48
tight: Hold fb and client refs while encoding
2020-09-27 17:17:22 +00:00
af38a643d9
tight: Don't block 1 worker while encoding tiles
2020-09-27 12:37:00 +00:00
9e84000e0a
server: Extract finish_fb_update() from on_client_update_fb_done()
2020-09-26 22:29:02 +00:00
add8d8e8db
server: Move pixel format conversion out of do_client_update_fb
2020-09-26 22:17:02 +00:00
7e2b4fef8c
server: Refactor schedule_client_update_fb
2020-09-26 22:00:53 +00:00
24a6e29cf2
raw-encoder: Use encder utils
2020-09-26 16:19:02 +00:00
413e6e7e72
zrle: Use encoder utils
2020-09-26 16:16:36 +00:00
31b7077bc5
tight: Use encoder utils
2020-09-26 16:13:17 +00:00
25a533e22b
Add common utilities for encoders
2020-09-26 16:10:25 +00:00
81a8ba9d0e
server: Handle cut text messages of up to 10MB
2020-09-25 21:47:59 +00:00
783807c0b9
server: Fix possible crash on fragmented packet messages
...
The packets sent from the client especially for client cut text,
are typically sent in two packets, one for the message containing
the type and length and the other for the actual data. Sometimes
the first message is read but we still don't have the data yet.
We need to continue reading data to use the structure but this
revealed a bug. The client event handler was calling memmove()
with buffer_index as the size argument. This meant that it was
copying the wrong amount of data, resulting in garbage at the
end of the expected data. This patch fixes the problem by first
subtracting buffer_index from buffer_len and then moving buffer_len
worth of data, which is what was read into msg_buffer. The problem
possibly manifested itself with random crashes, after reading
random data.
2020-09-21 21:15:45 +00:00
6d29937e15
server: Add remaining support for simple copy/paste
...
- Add function to set callback for client_cut_text
- Add server_cut_text structure and function
This in conjunction with wayvnc #66 closes #4 .
2020-09-21 21:15:45 +00:00
addcc50483
Bake version info into library
2020-07-26 11:43:29 +00:00
e1c0923915
stream: Add byte counters
2020-07-26 11:43:29 +00:00
b52f5cd6c9
Resize tight encoder on frame resize
2020-07-19 16:58:55 +00:00
76beec6415
tight: Add method to resize encoder grid
2020-07-19 16:57:44 +00:00
d4a5ed4133
stream: Ignore events after close instead of aborting
...
Aborting places an unnecessary restriction on the event loop library.
2020-07-19 13:22:57 +00:00
a0801f04e4
Implement server-side resizing
2020-07-19 11:54:39 +00:00
eb4b9d71a4
Return failure if display buffer is not set in on_connect
2020-07-18 14:04:12 +00:00
683776cdf0
server: Assert that a client has a stream open when processing a message
2020-07-16 18:14:52 +00:00
33eda8c5d0
Always enable tight, encoding but allow disabling lossy tight encoding
2020-07-11 20:16:27 +00:00
497f9adb29
tight: Re-implement with threads
2020-07-11 20:06:45 +00:00
9b54f6d936
Clean up dispatch hander on exit
2020-07-06 16:43:53 +00:00
52e30e795b
Align with neatvnc API changes
2020-07-06 16:43:15 +00:00
7fb2215c0f
Use raw encoding by default if no encoding has been selected
...
This complies with the standard.
2020-05-31 22:56:27 +00:00
569ad287c5
Fix version error handling
2020-05-29 20:18:05 +00:00
91bd49848f
Build benchmarks again
2020-05-27 21:59:34 +00:00
d2527e57f0
fix build warnings when building without tls
...
Fix trivial build warnings when building without TLS support.
2020-05-22 20:54:53 +00:00
8f9c71bb33
fix below zero message count check
...
size_t is unsigned and hence can't be below zero, triggering this gcc
warning with gcc 10:
warning: comparison of unsigned expression in ‘< 0’ is always false [-Wtype-limits]
It seems this if statement is meant to check if there are messages to
process (larger than 0). If there are no messages, we should jump out
early.
2020-05-21 17:44:31 +00:00
a37eed4a4a
remove fb_lock/unlock
...
The two functions have been removed from the external header file.
Remove them and the now unnecessary field is_locked.
2020-05-21 17:44:31 +00:00
8316994dfa
use static for functions not used outside this compile unit
...
Add static for all functions only used inside the individual compile
units. This helps the compiler to potentially inline these functions.
This allows to use neatvnc as a Meson subproject in Weston which has
-Wmissing-prototypes enabled by default.
2020-05-21 17:44:31 +00:00
477ab4c481
tight: Allow lossy encoding with 16 bpp
2020-05-05 21:34:45 +00:00
f3e09fd622
Prepare API for multi-display support
...
These changes are made now to make it possible to add multi-display
support in the future while keeping the public interface stable.
2020-04-12 18:16:19 +00:00
737dd311a0
Remove nvnc_set_dimensions()
...
This information is now figured out based on the nvnc_fb passed into
nvnc_set_buffer()
2020-04-12 16:08:33 +00:00
Andri Yngvason
Jim Ramsay
Philipp Zabel
Jeroen Hofstee
Jan Beich
Harm te Hennepe
Ryan Farley
Scott Moreau
Stefan Agner