Commit Graph

2034 Commits

Author SHA1 Message Date
Geoffrey McRae
ceac6a60e6 [spice] initial GnuTLS implementation (incomplete) 2018-05-22 09:49:35 +10:00
Geoffrey McRae
cba6630aa0 [spice] relocate openssl code into seperate function
This is in preperation of switching to an alternative SSL library
as OpenSSL conflicts with the GNU licence.
2018-05-22 09:39:03 +10:00
Geoffrey McRae
6e0eac0abc [common] fix getopt license to be compatible 2018-05-22 03:21:35 +10:00
Geoffrey McRae
b3aadccfc4 [client] use glFinish to prevent buffering and re-enable by default 2018-05-21 23:16:16 +10:00
Geoffrey McRae
116926f7c0 [module] enable write-through mapping for the device
This is the magic I have been looking for the last few weeks, with
this change VM->VM is now useful at 60+fps :D
2018-05-21 21:56:53 +10:00
Geoffrey McRae
d0a7b8df9c [module] added 'test' to Makefile to simplify development testing 2018-05-21 21:56:23 +10:00
Geoffrey McRae
b8a1743d8f [client] fixed crash when specifying the shmFile as a command 2018-05-21 21:46:48 +10:00
Geoffrey McRae
e8b1b8fbdf [common] tune windows memcpySSE asm implementation:wq 2018-05-19 21:40:13 +10:00
Geoffrey McRae
e9d77e6c52 [common] inline memcpy into memcpySSE for the final bytes 2018-05-19 18:31:49 +10:00
Geoffrey McRae
56f0a8525b [common] more SSE improvements
* 32bit inlined is slow for only large copies, warn if memcpySSE is
used when it shouldn't be.

* Removed 64bit memcpySSE as native inlined is faster

See: https://stackoverflow.com/questions/50422510/why-is-i386-memcpy-slow-on-x86-64
2018-05-19 18:27:04 +10:00
Geoffrey McRae
778af24d82 [common] inline get_pc to memcpySSE 2018-05-19 16:21:12 +10:00
Geoffrey McRae
cd6caea4b0 [x86] use a proper call/ret to obtain the current IP
See: https://blogs.msdn.microsoft.com/oldnewthing/20041216-00/?p=36973
2018-05-19 16:16:01 +10:00
Geoffrey McRae
f63c8043af [common] new sse2 memcpy improvements 2018-05-18 20:56:57 +10:00
Geoffrey McRae
3c77c1eb2b NASM version of a SSE2 memcpy 2018-05-18 18:50:07 +10:00
Geoffrey McRae
ffec6c2014 Incoming new memcpy implementation 2018-05-18 01:59:00 +10:00
Geoffrey McRae
d097531926 [client] allow the mouse and keyboard to operate early 2018-05-16 18:19:32 +10:00
Geoffrey McRae
d339ca3599 [client] remove dependencies on libva until h264 is ready 2018-05-16 18:14:08 +10:00
Geoffrey McRae
adb1ca58b9 [opengl-basic] removed basic renderer as it is no longer needed 2018-05-16 18:13:20 +10:00
Geoffrey McRae
70ffe1de43 [h264] disable the module until I or someone finds time to finish it 2018-05-16 18:12:29 +10:00
Geoffrey McRae
cf4d16b528 [opengl] numerous improvements to buffer transfer 2018-05-16 17:58:36 +10:00
Geoffrey McRae
da2bcfdf9a [module] rename defines from LG to KVMFR 2018-05-15 20:53:10 +10:00
Geoffrey McRae
7f81d21aaa [module] added kvmfr kernel module for VM->VM shared memory 2018-05-15 20:50:52 +10:00
Geoffrey McRae
d0756cf00c [main] make it possible to manually specify the memory size 2018-05-15 20:07:48 +10:00
Geoffrey McRae
4fd59ce8c9 [opengl] fix free bug with new contiguous buffer 2018-05-15 19:25:22 +10:00
Geoffrey McRae
adca879fb9 [opengl] use a single contiguous buffer for the AMD pinned buffer 2018-05-15 19:23:57 +10:00
Geoffrey McRae
3a2d612b41 [decoders] change the API to allow more flexability in the future 2018-05-15 19:19:39 +10:00
Geoffrey McRae
332d53e016 [opengl] Add support for AMD_pinned_memory if it is available 2018-05-15 13:23:44 +10:00
Geoffrey McRae
ae1344d1a0 [spice] fix out by one error in call to strncpy 2018-05-15 09:56:42 +10:00
Geoffrey McRae
ae382949c8 [opengl] glxWaitVideoSyncSGI is not well supported in Mesa 18
It seems Mesa 18 has problems with the glxWaitVideoSyncSGI API so
we disable the 'preventBuffer' option by default
2018-05-15 09:54:24 +10:00
Geoffrey McRae
fd8d4d3d38 [client] enable configuration of spice via config 2018-02-05 19:08:31 +11:00
Geoffrey McRae
a2216e4b68 [client] cosmetics 2018-02-05 19:08:15 +11:00
Alam Arias
4fb9fc3b3f [client]: ignore Alt-F4 in ignoreQuit mode (#46)
* [client]: ignore Alt-F4 in ignoreQuit mode

* [client]: turn down the SNR for SDL_HINT_WINDOWS_NO_CLOSE_ON_ALT_F4

* [client]: turn down the SNR for SDL_HINT_VIDEO_MINIMIZE_ON_FOCUS_LOSS
2018-02-02 10:32:46 +11:00
arcnmx
35b4d75eea [host] mingw-w64 compile fixes (#44)
* [host] rename min() to LG_MIN()

* [host] format string type fixes, %Ix doesn't exist in mingw

* [host] DXGI minor fixes

* [host] mingw lacks media foundation api headers and QISearch
2018-01-30 21:07:46 +11:00
arcnmx
d7321d5f5f [client] fix event loop sleep (#45) 2018-01-30 21:07:14 +11:00
Geoffrey McRae
35eda57cb2 [client] remoted accidential commit of experimental method from header 2018-01-29 17:30:46 +11:00
Geoffrey McRae
78a100135b [client] fix improper spice socket shutdown 2018-01-29 17:27:12 +11:00
Geoffrey McRae
dc6932a9ba [client] no need to supress bin directory creation command 2018-01-29 17:04:08 +11:00
Geoffrey McRae
d765674913 [client] fix concurrent make, fixes #43 2018-01-29 17:02:41 +11:00
Geoffrey McRae
2af522aea7 [client] prevent 100% CPU usage in event loop, fixes #41 2018-01-29 16:56:23 +11:00
arcnmx
9aba969296 [client] fix SDL mutex compile errors (#42) 2018-01-29 16:51:14 +11:00
Geoffrey McRae
2114b73c11 [client] makefile: use $(CC) instead of assuming gcc 2018-01-25 09:58:03 +11:00
Geoffrey McRae
d591e2fd36 [client] added xlib-shim to disable calls to XSync
The compiled xlib-shim.so can be used to intercept and prevent SDL
from calling XSync, which causes latency issues on some video
hardware.

To use specify the full path to the file in the LD_PRELOAD
environment variable, like so:

LD_PRELOAD=/full/path/xlib-shim.so ./looking-glass
2018-01-25 09:55:21 +11:00
Geoffrey McRae
c61d97b0ac [client] spice: add channel and mouse locking
This fixes a race condition which causes the mouse ringbuffer to
overflow. It also corrects out of order message index IDs due to
multiple threads sending messages asyncronously.
2018-01-25 07:41:11 +11:00
Geoffrey McRae
37ea662998 [client] use SDL_SetEventFilter for better mouse performance
This partially resolves lag issues on hosts running the amdgpu driver.
If mouse caputure is enable the lag issue returns, this is because SDL
calls `XSync` in `X11_WarpMouse` and `X11_WarpMouseGlobal`, if these
calls are removed all input lag issues dissapear.

This issue has been reported to SDL as the calls to `XSync` are not
required per the xlib documentation.
2018-01-24 23:46:11 +11:00
Aaron
3d9d275d61 Ignore SDL_MOUSEBUTTONDOWN events incompatible with SPICE (#38)
* Ignore SDL_MOUSEBUTTONDOWN events that aren't compatible with the SPICE PS/2 mouse
2018-01-15 10:55:17 +11:00
Geoffrey McRae
a02087e5e4 [client] h264: unroll silly loop 2018-01-06 13:47:35 +11:00
Geoffrey McRae
2ccf17b9b7 [client] h264: setup slice parameters correctly (incomplete) 2018-01-06 13:43:24 +11:00
Geoffrey McRae
8ccce5666c [client] nal: corrections to parsing logic 2018-01-06 13:40:31 +11:00
Geoffrey McRae
859e984827 [client] h264: vaapi deprecated these fields, FMO is not supported
fbed1dbb5b/va/va.h (L2840)
2018-01-06 09:47:49 +11:00
Geoffrey McRae
5e84cfb3f1 [client] h264: use parameters from nal (incomplete) 2018-01-06 00:11:38 +11:00