Age | Commit message (Collapse) | Author | Files | Lines |
|
When LIBGL_ALWAYS_SOFTWARE is enabled, coredump will happen because
tizen_swap_buffers is called.
Change-Id: I59e7952c3571da42ab31aacbf1b4dda87be51e04
Signed-Off-by: Xuelian Bai <xuelian.bai@samsung.com>
|
|
Change-Id: I6fe3fc14230cdca298cf5b92c616fa54d20baa40
Signed-Off-by: Xuelian Bai <xuelian.bai@samsung.com>
|
|
- close fence fd when destroy context
Change-Id: I4ec1fe1c2f7de3221802ea53c3e563471cd1a332
Signed-off-by: Tianhao Ni <tianhao.ni@samsung.com>
|
|
Location of udev rules file is changed from /etc/udev/rules.d to /usr/lib/udev/rules.d.
So, use macro instead of real path.
Change-Id: I5f7aada411f75206fabf86751ee8b6c2702c8b93
|
|
Memory usage pattern of Built-in builder causes large size chunk pools by glibc pool management scheme.
It is not useful ,but just unnecessary expense.
so malloc_trim will help to reduce memory usage in this case.
Change-Id: Ia63fb2050b468e55cd6b740b6a26585ed8e824a5
Signed-off-by: Sungguk Na <sungguk.na@samsung.com>
|
|
Change-Id: I9a3e9b05fb07e03b3b3aaf39f4038bee39072189
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
Change-Id: I92d33fba7cc073dc43cf6babc56778a600148b1a
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
Change-Id: Idc9d301b560373d05bc427f171f5f48777deb78c
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
Change-Id: Ic5be6f92a9ae5fd9db54c0491d5af9aa1838fd66
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
- 'server_wait_sync' should be called from
'tizen_internal_wait_sync_for_release_fence' with release_fence fd
received from dequeue, but the part was missing and fixed.
Change-Id: Ic505b2937ef472528a7dae2db486f66cf0cb3626
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
Change-Id: I59e0b8c42f280582b0c0c25b2963786a3a5404d1
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
Change-Id: Ieb1ddcee8f196e4e7138fc04bbb1fce2547bb92a
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
This is a test phase and can be fixed later.
Change-Id: I9670fc27b331b7ac4ccbcb9f659d82f4ab78f058
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
Change-Id: I9b562a2b70eae9c1334fea0248906f84ba932e9f
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
1. Add ttrace link in spec file
2. Add header file for ttrace
Change-Id: Ibfbc44f14f5e4b4e34d954249b2b399335134c4b
Signed-Off-by: Xuelian Bai <xuelian.bai@samsung.com>
|
|
In some cases, main thread got some problem and exit with calling
_eglAtExit, while another thread may still be running, that will
cause assert in mesa.
Change-Id: I6ee87abcb8c818e8a6808cd6098f4061bf7779b2
Signed-Off-by: Xuelian Bai <xuelian.bai@samsung.com>
|
|
Merged old commit:
Changed the initial setting of dri2_surface to use out_fence_fd.
Change-Id: Ic51d7e2441927010b5889a2328199428c06d6351
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
Fixed to use out_fence_fd to wait render done in twe_thread.
Change-Id: I2bee255bcd5eb2697e49a5e8973c03ea65fe0e30
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
v3d_screen: Fixed to support capability for NATIVE_FENCE_FD.
Change-Id: I3335ee417d88e51a904114869fb5ac287622c526
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
v3d_fence: Added function to get fence_fd from dri2.
Change-Id: Ibd86254f2352a01cd993c61b0ab036c721005be0
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
Added new extension EGL_TIZEN_native_fence_sync.
EGL_TIZEN_native_fence_sync
- Added EGL_SYNC_NATIVE_FENCE_TIZEN to distinguish EGLSyncKHR.
- Added EGL_SYNC_NATIVE_FENCE_FD_TIZEN to get tizen's native fence fd
in the sync attribute.
- Added EGL_NO_NATIVE_FENCE_FD_TIZEN to indicate no native fence fd.
- The eglDupNativeFenceFDTIZEN function has been added to get an fd to
indicate when rendering is complete from the egl core.
Change-Id: I6876572ec04b3d1616f6b7e5018dc6647685b536
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
Implemented extension EGL_TIZEN_native_fence_sync.
- this patch is only for the 'tizen' branch
- Added new API to get native fence fd from driver.
eglDupNativeFenceFDTIZEN()
Change-Id: I21d46faedbfc46abd4dde03d36560bf941061d38
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
- It is a valid backend only for wayland clients.
Change-Id: Ifea5a5b5bd03074f267c4be02d325ed544840753
Signed-off-by: Joonbum Ko <joonbum.ko@samsung.com>
|
|
When running case dEQP-GLES2.functional.prerequisite.read_pixels
for deqp-gles2, there will be an error "close object **".
This error is caused by invalid fd, so we call
tbm_drm_helper_get_auth_info in dri2_initialize_tizen.
Change-Id: I959b404daa33c629ea72c38bfe7838e643eb603d
Signed-Off-by: Xuelian Bai <xuelian.bai@samsung.com>
|
|
Handle 1x1 issue for driimage
When evas resize window after dequeue, then draw and swapbuffers,
window is shown with old size until next makecurrent.
Add validate_drawable function to check if resize happen each time
makecurrent is called.
For RPI4, render node is used, so buffer is allocated on another
device, when 1x1 issue happen, we need to make sure if driimage
size is changed, then created new if yes.
Merged Commit:
commit 2dc471b0e167693b81c24c9a3f89b985cb98249f
Author: Zhaowei Yuan <zhaowei.yuan@samsung.com>
- Check if stdraw is NULL before access
Variable "stdraw" could be NULL in some case, check it before access.
commit cc2dc99365906ea52b645be3bb10f50708d67e84
Author: Xuelian Bai <xuelian.bai@samsung.com>
- Fix 1x1 issue of task-mgr
Makecurrent is called just before resize callback is called, so we
have to check if size is changed when drawcall.
commit 670d77331c986a92bcf2459567ccd6dfc181bc66
Author: Xuelian Bai <xuelian.bai@samsung.com>
- Fix resize issue
Change-Id: Ieaebe72661a081d496ab6214c1c1526f7d16f19b
Signed-Off-by: Xuelian Bai <xuelian.bai@samsung.com>
- Handle 1x1 issue for driimage
Change-Id: Id5a463a8db290ca5b80c130ebcdcb13c7786d70c
Signed-off-by: Xuelian Bai <xuelian.bai@samsung.com>
|
|
RPI4 has seperated devices for display and render, so when init
Tizen, we open /dev/dri/renderD128 and init v3d screen as renderonly.
Merged commit:
commit 50839c47736cbab7db6a9fb25ea0d983434c56cc
Author: Zhaowei Yuan <zhaowei.yuan@samsung.com>
- platform_tizen: free driver_name if initialization goes wrong
Mallocated driver_name needs to be freed if following steps for
initialization goes wrong, whether hw_accel if set or not
Change-Id: Iffac62ac9717f7a4da589fd51f008968615a6321
Signed-off-by: Xuelian Bai <xuelian.bai@samsung.com>
|
|
Merged old commit:
commit f409ef598976a4fd4f74ab4695478b1fc1651c06
Author: Zhaowei Yuan <zhaowei.yuan@samsung.com>
1. Add EGL_PBUFFER_BIT to surface_type if maskes match
2. Alloc local buffer when render buffer type is
"FRONT_LEFT" which will be used by Pbuffer
commit b0dec10d0d0fbc7f77623016f3d986122f718f8b
Author: Xuelian Bai <xuelian.bai@samsung.com>
Fix crash when query a pbuffer surface
Pbuffer is not created via TPL, so when query a pbuffer surface,
we can only call _eglQuerySurface instead of TPL APIs.
Change-Id: Ib9990973f228374a95f60157a3b79069529a62b2
Signed-off-by: Xuelian Bai <xuelian.bai@samsung.com>
|
|
Add support for wayland & drm platforms which are used by weston.
Select EGL platform for weston by setting environment variable:
export EGL_PLATFORM=wayland
Modifications to mesa.spec:
Add platforms wayland and drm which is needed by weston
Enable gbm and pack it into RPM package
Modifications to egl_dri2.h
Since tizen, drm and wayland are all supported now, remove
the duplicate definition of "color_buffers" and "back" for
tizen to avoid from building error.
Merged commit:
commit 7c8922c1ead5875fda90106793c2fcfab2feb9f3
Author: Xiaowei Guan <xiaowei.guan@samsung.com>
- fix weston build error and app launch error
commit 6e7ce80fc524b78980d702549cf5bbc3a641e6ae
Author: Xiaowei Guan <xiaowei.guan@samsung.com>
- add wayland-egl code for platform_wayland
commit 5cd0256faa43e376dcd5b5f2c2edd2515aff9c04
Author: Zhaowei Yuan <zhaowei.yuan@samsung.com>
- EGL: Add support for weston
Change-Id: I58272083b7d110d55800d8b9ad7e7c4f0a6f6077
Signed-off-by: Zhaowei Yuan <zhaowei.yuan@samsung.com>
|
|
docs: Add Tizen Binary Download Instructions for RPI3
Readme_for_Tizen explains howto seting up build environment for tizen and geting
build dependancy packages.
Readme_RPI3_Setup_for_Tizen explains How to building RPI3 kernel for tizen and
downloading to tizen platform binaries to rpi3.
Add Tizen Binary Download Instructions for RPI3. it includes which packages
should be installed with Mesa for tizen.
Merged old commit a8f81f8f9382445f072c616e77ce0da72c028480
Change-Id: I44f8cdca5be8c8860d7f6a162889d534068168c1
Signed-off-by: Mun, Gwan-gyeong <kk.moon@samsung.com>
|
|
Re-implement _eglLog based on dlog
Add build support for Tizen platform.
Added missed dri2_fini_surface() at tizen_destroy_surface().
Add platform_tizen.c that supports _EGL_PLATFORM_TIZEN. It works with libtpl-egl
(Tizen Porting Layer for egl), libtbm(Tizen Buffer Manager) where back buffers
of windows are backed by GEM objects.
In Tizen a native window has a queue (tbm_surface_queue) of back buffers
allocated by the WL_TBM(wayland client case, WL_TBM is abbreviation of
wayland-tbm protocol) or gbm ( tizen has implements gbm with tbm) or tbm
through tbm_backend.
For each frame, EGL needs to
dequeue the next back buffer
render to the buffer
enqueue the buffer
After enqueuing, the buffer is no longer valid to EGL.
It supports DRI image loader extension, DRI dri2 loader extension and DRI swrast
loader extension.
It supports EGL_NATIVE_SURFACE_TIZEN target type for eglCreateImageKHR.
(https://www.khronos.org/registry/EGL/extensions/TIZEN/EGL_TIZEN_image_native_surface.txt)
BindWaylandDisplayWL / UnbindWaylandDisplayWL / QueryWaylandBufferWL are
overloaded because tizen platform has its own instead of WL_DRM.
Referenced documents:
https://www.x.org/wiki/Events/XDC2016/Program/XDC2016_Tizen_Window_System_EGL_Vulkan.pdf
https://wiki.tizen.org/wiki/3.0_Porting_Guide/Graphics_and_UI/libtpl-egl
https://wiki.tizen.org/wiki/TBM
egllog: add unlock call preventing the deadlock
Implement dlog for mesa main
_eglLog is a basic log function for mesa, re-implement
it based on dlog here for debug purpose.
Log level can be changed by setting environment variable
EGL_LOG_LEVEL.
Log API for mesa main module are all implemented in file
src/mesa/main/error.c, re-implement them based on dlog
for debug convenience
add configuration of building and packaging for rpi3 ( vc4 gpu driver )
add udev rule files for rpi3 (vc4 drm device)
Mered old commit:
commit f5205524bf20b63c2efe20e99429c9d55543b711
Author: Zhaowei Yuan <zhaowei.yuan@samsung.com>
Date: Mon May 6 13:52:32 2019 +0800
commit 704bb72470589ede7246af8c6b0a9210df42f1c7
Author: Xuelian <xuelian.bai@samsung.com>
Date: Wed Nov 20 19:08:41 2019 +0800
commit 247adc92138ad456f5db39acfe5be69748973df3
Author: Joonbum Ko <joonbum.ko@samsung.com>
Date: Wed Apr 29 20:01:56 2020 +0900
Change-Id: I83348c1cff5f0f59f79bc3c016623adf6fbba32b
Signed-off-by: Xuelian <xuelian.bai@samsung.com>
|
|
|
|
|
|
This fixes CTS test: KHR-GL46.arrays_of_arrays_gl.AtomicUsage
Fixes: bddc0e023c "radeonsi: fix read from compute / write from draw sync"
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9525>
|
|
si_check_render_feedback only relied on si_images::enabled_mask and
si_samplers::enabled_mask to determine if a texture was being used
both as input and output.
Given that some samplers/images can be considered active (so accounted
for by enabled_mask) but not used by the current shader this could
lead to false-positive.
This commit fixes this by and-ing the above mask with the information
from shader_info for each active shader.
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4227
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9525>
|
|
A compute dispatch should see the result of a previous draw command.
radeonsi was missing this implicit sync, causing rendering artifacts:
the compute shader was reading from a texture still being written to
by the previous draw.
Framebuffer BOs are marked with RADEON_USAGE_NEEDS_IMPLICIT_SYNC,
so compute jobs will sync.
v2: use RADEON_USAGE_NEEDS_IMPLICIT_SYNC
v3: unconditionally make CB coherent after a flush
Reviewed-by: Zoltán Böszörményi <zboszor@gmail.com> (v3)
Reviewed-by: Marek Olšák <marek.olsak@amd.com> (v3)
Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4032
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/2878
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/1336
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9525>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Replaced by IR3_SHADER_DEBUG=disasm,{vs,...,cs} and unused since the
commit referenced below.
Fixes: 808992fc506 ("tu: Use the ir3 shader API")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8249>
(cherry picked from commit 6ceb6b509e64c54812a5f6a208e7d93cc61119f4)
|
|
supports it.
vkGetPhysicalDeviceProperties2 is not allowed to be used with a 1.0 device
because it's a vulkan 1.1 function.
Closes: #4396
Fixes: 38ce8d4d ("vulkan/device_select: Stop using device properties 2.")
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9462>
(cherry picked from commit fb1100d718fbee07dc294e1ede20a084cda423b9)
|
|
if we get some crazy matrix types in here then we need to ensure that
we accurately unwrap them and copy the components
fixes KHR-GL46.enhanced_layouts.xfb_stride
Fixes: 1b130c42b8d ("zink: implement streamout and xfb handling in ntv")
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9271>
(cherry picked from commit 1b25e3a7019ca529c2ab665a35e06f21ec4b4ebb)
|
|
The container is moved from before and hence returns size 0. To get the
correct value, the new instruction container must be used instead.
This was flagged by clang-tidy. The fixed call still triggers the
corresponding diagnostic, hence this change silences it by adding a
redundant clear() after move.
Fixes: 7f1b537304d ("aco: add new NOP insertion pass for GFX6-9")
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9432>
(cherry picked from commit 97c97781f6a94e5c8209266ec996fe6caff04dff)
|
|
Fixes: 22f6624ed318 "gallium: separate indirect stuff from pipe_draw_info - 80 -> 56 bytes"
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9425>
(cherry picked from commit 12f1e42ed34e9807484bdfbec535bc56950e2338)
|
|
We should only pass in a new indirect_info object if we actually set valid
values in it.
Fixes: abe8ef862fe5 "gallium: make pipe_draw_indirect_info * a draw_vbo parameter"
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9425>
(cherry picked from commit a84c8ddb19beef7580d6a4b8c417bb188517412a)
|
|
nir_lower_alpha_test creates a uniform for the alpha reference value;
this needs to be updated when changing alpha test state.
Fixes: b1c4c4c7f53 ("mesa/gallium: automatically lower alpha-testing")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4390
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9439>
(cherry picked from commit efd7711e0e140c75f3624997f91f46008a39c325)
|
|
|
|
We should avoid a situation where a v_readfirstlane_b32 is in WQM but it's
source is calculated in Exact.
Fixes hang when running Assassin's Creed: Valhalla benchmark.
fossil-db (GFX10.3):
Totals from 1021 (0.70% of 146267) affected shaders:
CodeSize: 7835228 -> 7842992 (+0.10%); split: -0.00%, +0.10%
Instrs: 1519208 -> 1521149 (+0.13%); split: -0.00%, +0.13%
SClause: 78921 -> 78920 (-0.00%)
Copies: 44456 -> 45421 (+2.17%); split: -0.05%, +2.22%
Branches: 12987 -> 13933 (+7.28%)
PreSGPRs: 47599 -> 47813 (+0.45%)
Cycles: 10037540 -> 10045304 (+0.08%); split: -0.00%, +0.08%
VMEM: 538381 -> 538777 (+0.07%); split: +0.11%, -0.03%
SMEM: 84553 -> 84554 (+0.00%); split: +0.01%, -0.01%
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9503>
|
|
It really does help to size these arrays correctly.
Fixes: 2b4fcf0a063 zink: generate instance creation code with a python script
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9504>
|
|
ensure this doesn't get pulled in during the next program update
Fixes: 334759d8509 ("zink: implement passthrough tcs shader injection")
Reviewed-by: Hoe Hao Cheng <haochengho12907@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9504>
|
|
this fixes some testcases on CI.
Fixes: 1c01ad1b804 ("zink: add KHR_draw_indirect_count detection")
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9504>
|
|
We don't have the driver hooks to set the state, or to update draw's copy
of the state at draw time.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Fixes: d6fd7acf9b79 ("gallium: Fix leak of bound SSBOs at CSO context
destruction.")
Closes: #3990
(cherry picked from commit 58e43594fc457eaaf1b1e01e48948959a82080bc)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8970>
|
|
When not using S_00B324_MEM_BASE the value isn't properly truncated.
Cc: mesa-stable
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9277>
(cherry picked from commit 0e97d817f520653548f1dee1db28bd2d92d28027)
|