| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
Now that we are no longer using the pctx reference in the shader, drop
it and turn on shareable shaders.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
| |
Rather than fishing it out of the shader. This removes the other big
user of shader->pctx.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
| |
And use this for allocating bo's to hold the shader binary, rather than
accessing the dev via ctx ptr. One step towards making shaders sharable
across contexts.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
| |
This just passes the extra blit info to fix the render condition
tests.
Cc: "11.2" <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
in order to make some winsys interface changes easier
This distros should use new DRM if they want to use new Mesa:
Distro kernel mesa eol
SLES 10 2.6.16 6.4.2 2016-07
SLED 11 3.0 9.0.3 2022-03
RHEL 5 2.6.18 6.5.1 2017-03
RHEL 6 2.6.32 10.4.3 2020-11
Debian 6 2.6.32 7.7.1 2016-02
Reviewed-by: Michel Dänzer <[email protected]>
|
|
|
|
| |
Reviewed-by: Christian König <[email protected]>
|
|
|
|
|
|
| |
including prolog and epilog disassemblies
Reviewed-by: Michel Dänzer <[email protected]>
|
|
|
|
| |
Reviewed-by: Michel Dänzer <[email protected]>
|
|
|
|
|
|
|
|
|
| |
This can increase perf for shaders that kill pixels (kill, alpha-test,
alpha-to-coverage).
v2: add comments
Reviewed-by: Michel Dänzer <[email protected]>
|
|
|
|
|
|
| |
for radeonsi
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Here is another threading issue with MANAGED buffers:
Thread 1: buffer creation
Thread 1: buffer lock
Thread 2: Draw call
Thread 1: writes data
Thread 1: Unlock
Without this patch, the buffer is initially dirty
and in the list of things to upload after its creation.
The draw call will then upload the data and unset the dirty flag,
and the Unlock won't trigger a second upload.
Fixes regression introduced by cc0114f30b587a10766ec212afb3ad356099ef23:
"st/nine: Implement Managed vertex/index buffers"
Cc: "11.2" <[email protected]>
Signed-off-by: Axel Davy <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
d3d calls are protected by mutexes, however if app is doing in
two threads:
Thread 1: buffer Lock
Thread 2: Draw call
Thread 1: writes data
Thread 1: Unlock
Then before this patch, the Draw call would begin to upload
the buffer.
Solves this by moving the moment we add the buffer to the queue
of things to upload (We move it from Lock time to Unlock time).
Cc: "11.2" <[email protected]>
Signed-off-by: Axel Davy <[email protected]>
|
|
|
|
|
|
|
|
| |
READONLY won't trigger an upload.
Cc: "11.2" <[email protected]>
Signed-off-by: Axel Davy <[email protected]>
|
|
|
|
|
|
|
|
|
| |
When the semantic is Position (which can happen with index 0 only),
use the helper to get Position input.
Cc: "11.2" <[email protected]>
Signed-off-by: Axel Davy <[email protected]>
|
|
|
|
|
|
| |
Cc: "11.2" <[email protected]>
Signed-off-by: Axel Davy <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Otherwise the build will fail, when the library is in a non default
location.
v2 [Emil Velikov]
- drop the unneeded cflags from targets/opencl.
Cc: "11.1 11.2" <[email protected]>
Fixes: 7f585a6a98d "configure.ac: use pkg-config for libelf"
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93524
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In builds with clang, there are several errors related to the enum
alu_op_flags like this:
src/gallium/drivers/r600/sb/sb_expr.cpp:887:8:
error: case value evaluates to -1610612736, which cannot be narrowed to
type 'unsigned int' [-Wc++11-narrowing]
These are due to the MSB being set in the enum. Fix these errors by
making the enum values unsigned as needed. The flags field that stores
this enum also needs to be unsigned.
Cc: "11.1 11.2" <[email protected]>
Cc: Marek Olšák <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
Reviewed-by: Michel Dänzer <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix compiles with clang that have this C++11 error:
src/gallium/drivers/radeon/r600_pipe_common.h:662:34:
error: invalid suffix on literal; C++11 requires a space between literal
and identifier [-Wreserved-user-defined-literal]
Cc: "11.1 11.2" <[email protected]>
Cc: Marek Olšák <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
Reviewed-by: Michel Dänzer <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Enabling this warning doesn't generate any warnings with gcc, but is an
unknown option for clang, so drop it.
Signed-off-by: Rob Herring <[email protected]>
Acked-by: Rob Clark <[email protected]> (v1)
Cc: "11.1 11.2" <[email protected]>
v2: keep the warning around, commented out
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
MESA_DRI_MODULE_PATH is only getting set for classic DRI drivers and may or
may not be set correctly for gallium_dri.so depending on the makefile
include ordering. For Android 6 and earlier it is fine, but with build
system changes in AOSP master, it is not.
Move the path variables to a single place at the top level and introduce
MESA_DRI_MODULE_REL_PATH for Android 5 and later which require relative
paths. With this, there is a single variable to change.
Cc: "11.1 11.2" <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
Reviewed-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commits a39a8fbbaa12 ("nir: move to compiler/") and eb63640c1d38
("glsl: move to compiler/") broke Android builds. Fix them.
There is also a missing dependency between generated NIR headers and
several libraries. This isn't a new issue, but seems to have been
exposed by the NIR move.
Built with i915, i965, freedreno, r300g, r600g, vc4, and virgl enabled.
Cc: "11.2" <[email protected]>
Cc: Mauro Rossi <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
Reviewed-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This function is currently broken for BE. I assume it's because of
util_pack_color(). Until I fix this path, I prefer to disable it so users
would be able to see correct colors on their desktop and applications.
Together with the two following patches:
- gallium/r600: Don't let h/w do endian swap for colorformat
- gallium/radeon: remove separate BE path in r600_translate_colorswap
it fixes BZ#72877 and BZ#92039
Signed-off-by: Oded Gabbay <[email protected]>
Cc: "11.1 11.2" <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since the rework on gallium pipe formats, there is no more need to do
endian swap of the colorformat in the h/w, because the conversion between
mesa format and gallium (pipe) format takes endianess into account (see
the big #if in p_format.h).
v2: return ENDIAN_NONE only for four 8-bits components
(V_0280A0_COLOR_8_8_8_8)
Signed-off-by: Oded Gabbay <[email protected]>
Cc: "11.1 11.2" <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After further testing, it appears there is no need for
separate BE path in r600_translate_colorswap()
The only fix remaining is the change of the last if statement, in the 4
channels case. Originally, it contained an invalid swizzle configuration
that never got hit, in LE or BE. So the fix is relevant for both systems.
This patch adds an additional 120 available visuals for LE and BE,
as seen in glxinfo
v2:
Tested for regressions by running piglit gpu.py with CAICOS (r600g) on
x86-64 machine. No regressions found.
Signed-off-by: Oded Gabbay <[email protected]>
Cc: "11.1 11.2" <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
| |
Changes from v2:
- add missing NOT modifier for GK110/GM107
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
|
|
| |
It appears that it actually needs to be aligned to the datum size, so it
was 1 when testing with R8, but it can be as high as 16 with RGBA32.
Signed-off-by: Ilia Mirkin <[email protected]>
Cc: "11.1 11.2" <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
Reduce the amount of duplicated code by re-using
nvc0_program_validate(). While we are at it, change the prototype
to return void and remove nvc0_compute.h which is now useless.
Signed-off-by: Samuel Pitoiset <[email protected]>
Acked-by: Pierre Moreau <[email protected]>
Acked-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
|
|
|
| |
No reason to not validate those global buffers and this might avoid
fails if someone try to use the global memory from compute programs.
Signed-off-by: Samuel Pitoiset <[email protected]>
Acked-by: Pierre Moreau <[email protected]>
Acked-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
|
|
|
| |
While we are at it, rename it to nvc0_compute_validate_globals() and
update its prototype.
Signed-off-by: Samuel Pitoiset <[email protected]>
Acked-by: Pierre Moreau <[email protected]>
Acked-by: Ilia Mirkin <[email protected]>
|
| |
|
|
|
|
|
|
|
|
| |
Without this mesa 11.2.0-rc1 ended up busted :-(
Cc: "11.2" <[email protected]>
Repored-by: Ondřej Súkup <[email protected]>
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Because I changed the swizzle check, I also need to adapt the return
values for each check.
It's basically almost the same as before, we just cross between STD and
STD_REV, and cross between ALT and ALT_REV
This fixes the rgba test in gl-1.0-readpixsanity (piglit) and also
fixes tri-flat (mesa demos).
Signed-off-by: Oded Gabbay <[email protected]>
Cc: "11.1 11.2" <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
| |
Signed-off-by: Oded Gabbay <[email protected]>
Reviewed-by: Thomas Helland <[email protected]>
Reviewed-by: Roland Scheidegger <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The current code in r600_translate_colorswap uses the swizzle information
to determine which colorswap to use.
This works for BE & LE when the nr_channels is <4, but when nr_channels==4
(e.g. PIPE_FORMAT_A8R8G8B8_UNORM), this method can not be used for both BE
and LE, because the swizzle info is the same for both of them.
As a result, r600g doesn't support 24bit color formats, only 16bit, which
forces the user to choose 16bit color in X server.
This patch fixes this bug by separating the checks for LE and BE and
adapting the swizzle conditions in the BE part of the checks.
Tested on an Evergreen GPU (Cedar GL FirePro 2270) running inside POWER7
Big-Endian Machine.
Signed-off-by: Oded Gabbay <[email protected]>
CC: "11.2" "11.1" <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
"st/mesa: overhaul vertex setup for clearing, glDrawPixels, glBitmap"
added a vertex shader declaring IN[0] and IN[2], but not IN[1].
Drivers relying on tgsi_shader_info can't handle holes in declarations,
because tgsi_shader_info doesn't track that.
This is just a quick workaround meant for stable that will work for vertex
shaders.
This fixes radeonsi DrawPixels and CopyPixels crashes.
Cc: [email protected]
Reviewed-by: Brian Paul <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Like unlocked store on shared memory, locked store can fail and the
second dest which is a predicate must be emitted.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
Cc: [email protected]
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
| |
Actually OP_SELP doesn't need to be a compare instruction. Instead we
just need to set the NOT modifier when building the instruction.
While we are at it, fix the dst register type and use a GPR.
Suggested by Ilia Mirkin.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Silences a warning reported by the svga3d device.
v2: also null-out the index buffer pointer
Reviewed-by: Sinclair Yeh <[email protected]>
Reviewed-by: Charmaine Lee <[email protected]>
|
|
|
|
|
|
| |
Reflect the nv50->g80 change and the new gm107_texture header.
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
|
| |
v2: handle _mesa_hash_table_insert failure
other cosmetic changes
Reviewed-by: Nicolai Hähnle <[email protected]>
|
|
|
|
| |
Reviewed-by: Nicolai Hähnle <[email protected]>
|