Skip to content
Commits on Source (128)
28 additional commits have been omitted to prevent performance issues.
Show whitespace changes
Inline Side-by-side
bin/.cherry-ignore
View file @ eb8a5560
# fixes: This commit has more than one Fixes tag but the commit it
# addresses didn't land in branch.
6ff1c479968819b93c46d24bd898e89ce14ac401 autotools: don't ship the git_sha1.h generated in git in the tarballs
# pick: This commit addresses a regression introduced by previous
# commit fa9e6c235da, which didn't make it for 18.2.
a72dbc461bdb7714656e62cd8f4b00a404c2e6e0 mesa: allow GL_UNSIGNED_BYTE type for SNORM reads
# fixes: This commit has more than one Fixes tag but the commit it
# addresses didn't land in branch.
c9f54486959716762e6818dabb0a73a8cd46df67 radeonsi: fix regression in indirect input swizzles.
# extra: Just some comments update.
2ad9917e187c1e9dbb053d3c98aa0e39fa374059 anv/blorp: Fix a comment as per Nanley's review feedback
debian/changelog
View file @ eb8a5560
mesa (18.2.0-1) experimental; urgency=medium
* New upstream release.
-- Timo Aaltonen <tjaalton@debian.org> Mon, 10 Sep 2018 12:54:33 +0300
mesa (18.2.0~rc4-1) experimental; urgency=medium
* New upstream release candidate.
......@@ -6,6 +12,20 @@ mesa (18.2.0~rc4-1) experimental; urgency=medium
-- Timo Aaltonen <tjaalton@debian.org> Wed, 22 Aug 2018 19:59:55 +0300
mesa (18.1.8-1) unstable; urgency=medium
* New upstream release.
-- Timo Aaltonen <tjaalton@debian.org> Mon, 10 Sep 2018 10:34:39 +0300
mesa (18.1.7-1) unstable; urgency=medium
* New upstream release.
* libglx-mesa0: Add Breaks: glx-diversions (<< 0.8.4~) (Closes:
#903929).
-- Andreas Boll <aboll@debian.org> Wed, 29 Aug 2018 11:16:55 +0200
mesa (18.1.6-1) unstable; urgency=medium
* New upstream release.
......
debian/control
View file @ eb8a5560
......@@ -239,6 +239,7 @@ Depends:
Provides: libglx-vendor
Breaks:
libopengl-perl (<< 0.6704+dfsg-2),
glx-diversions (<< 0.8.4~),
Pre-Depends: ${misc:Pre-Depends}
Multi-Arch: same
Description: free implementation of the OpenGL API -- GLX vendor library
......
docs/install.html
View file @ eb8a5560
......@@ -75,7 +75,7 @@ you think you've spotted a bug let developers know by filing a
Version 2.6.4 or later should work.
</li>
<li><a href="http://www.makotemplates.org/">Python Mako module</a> -
Python Mako module is required. Version 0.3.4 or later should work.
Python Mako module is required. Version 0.8.0 or later should work.
</li>
<li>lex / yacc - for building the Mesa IR and GLSL compiler.
<div>
......
docs/relnotes/18.2.0.html
View file @ eb8a5560
......@@ -14,7 +14,7 @@
<iframe src="../contents.html"></iframe>
<div class="content">
<h1>Mesa 18.2.0 Release Notes / TBD</h1>
<h1>Mesa 18.2.0 Release Notes / September 7, 2018</h1>
<p>
Mesa 18.2.0 is a new development release. People who are concerned
......@@ -59,9 +59,217 @@ Note: some of the new features are only available with certain drivers.
<li>GL_ARB_sample_locations and GL_NV_sample_locations on nvc0 (GM200+)</li>
<li>GL_ANDROID_extension_pack_es31a on radeonsi.</li>
<li>GL_KHR_texture_compression_astc_ldr on radeonsi</li>
<li>GL_NV_conservative_raster and GL_NV_conservative_raster_dilate on nvc0 (GM200+)</li>
<li>GL_NV_conservative_raster_pre_snap_triangles on nvc0 (GP102+)</li>
<li>multisampled images on nvc0 (GM107+) (now supported on GF100+)</li>
</ul>
<h2>Bug fixes</h2>
<ul>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=13728">Bug 13728</a> - [G965] Some objects in Neverwinter Nights Linux version not displayed correctly</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=61761">Bug 61761</a> - glPolygonOffsetEXT, OFFSET_BIAS incorrectly set to a huge number</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=65422">Bug 65422</a> - Rename api_validate.[ch] to draw_validate.[ch]</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=78097">Bug 78097</a> - glUniform1ui and friends not supported by display lists</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=91808">Bug 91808</a> - trine1 misrender r600g</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=93355">Bug 93355</a> - [BXT,SKLGT4e] intermittent ext_framebuffer_multisample.accuracy fails</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=95009">Bug 95009</a> - [SNB] amd_shader_trinary_minmax.execution.built-in-functions.gs-mid3-ivec2-ivec2-ivec2 intermittent</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=95012">Bug 95012</a> - [SNB] glsl-1_50.execution.built-in-functions.gs-op tests intermittent</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98699">Bug 98699</a> - &quot;float[a+++4 ? 1:1] f;&quot; crashes glsl_compiler</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99116">Bug 99116</a> - Wine DirectDraw programs showing only a blackscreen when using Mesa Gallium drivers</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99730">Bug 99730</a> - Metro Redux game(s) needs override for midshader extension declaration</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100177">Bug 100177</a> - [GM206] Misrendering in XCOM Ennemy Within</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100430">Bug 100430</a> - [radv] graphical glitches on dolphin emulator</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=101247">Bug 101247</a> - Mesa fails to link GLSL programs with unused output blocks</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=102390">Bug 102390</a> - centroid interpolation causes broken attribute values</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=102678">Bug 102678</a> - gl_BaseVertex should always be zero when the draw command has no &lt;basevertex&gt; parameter</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=103274">Bug 103274</a> - BRW allocates too much heap memory</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=104388">Bug 104388</a> - [snb] GPU HANG: ecode 6:0:0x85fffff8 in fgfs</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=104626">Bug 104626</a> - broadcom/vc5: double compare</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=104809">Bug 104809</a> - anv: DOOM 2016 and Wolfenstein II:The New Colossus crash due to not having depthBoundsTest</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105351">Bug 105351</a> - [Gen6+] piglit's arb_shader_image_load_store-host-mem-barrier fails with a glGetTexSubImage fallback path</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105374">Bug 105374</a> - texture3d, a SaschaWillems demo, assert fails</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105396">Bug 105396</a> - tc compatible htile sets depth of htiles of discarded fragments to 1.0</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105399">Bug 105399</a> - [snb] GPU hang: after geometry shader emits no geometry, the program hangs</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105497">Bug 105497</a> - shader-db crashes on 72 core system after ast_type_qualifier bitset change</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105613">Bug 105613</a> - Compute shader locks up within nested &quot;for&quot; loop</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105731">Bug 105731</a> - linker error &quot;fragment shader input ... has no matching output in the previous stage&quot; when previous stage's output declaration in a separate shader object</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105904">Bug 105904</a> - Needed to delete mesa shader cache after driver upgrade for 32 bit wine vulkan programs to work.</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=105975">Bug 105975</a> - i965 always reports 0 viewport subpixel bits</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106090">Bug 106090</a> - Compiling compute shader crashes RADV</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106133">Bug 106133</a> - make check &quot;OSError: [Errno 24] Too many open files&quot;</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106163">Bug 106163</a> - r600/sb: optimizer tries to schedule access to different array elements in one instruction group</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106174">Bug 106174</a> - vulkan dota2 broken (segfaulting), found bug commit</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106180">Bug 106180</a> - [bisected] radv vulkan smoke test black screen (Add support for DRI3 v1.2)</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106232">Bug 106232</a> - LLVM unit tests have error in random number handling</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106243">Bug 106243</a> - [kbl] GPU HANG: 9:0:0x85dffffb, in Cinnamon</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106315">Bug 106315</a> - The witness + dxvk suffers flickering garbage</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106331">Bug 106331</a> - radv doesnt support VK_FORMAT_R32G32B32_SFLOAT</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106382">Bug 106382</a> - Shader cache breaks INTEL_DEBUG=shader_time</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106393">Bug 106393</a> - glsl-fs-shader-stencil-export hangs forever</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106450">Bug 106450</a> - glGetIntegerv return wrong value in some cases</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106462">Bug 106462</a> - piglit.spec.arb_vertex_array_bgra.get regression</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106479">Bug 106479</a> - NDEBUG not defined for libamdgpu_addrlib</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106480">Bug 106480</a> - A2B10G10R10_SNORM vertex attribute doesn't work.</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106499">Bug 106499</a> - [regression, bisected] Several games crash on start</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106504">Bug 106504</a> - vulkan SPIR-V parsing failed at ../src/compiler/spirv/vtn_cfg.c:381</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106511">Bug 106511</a> - radv: MSAA broken on SI (assertion failure in vkCreateImage)</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106587">Bug 106587</a> - Dota2 is very dark when using vulkan render on a Intel &lt;&lt; AMD prime setup</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106594">Bug 106594</a> - [regression,apitrace,bisected] Prison Architect rendered unplayable by multicoloured flickering triangles and overlayed triangles when performing certain actions</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106619">Bug 106619</a> - [OpenCL][llvm-svn]build failure addPassesToEmitFile candidate expects 6 arguments, 3 provided</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106629">Bug 106629</a> - [SNB,IVB,HSW,BDW] dEQP-EGL.functional.image.create.gles2_cubemap_negative_z_rgb_read_pixels</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106642">Bug 106642</a> - X server crashes in i965 on desktop startup when DRI3 v1.2 / modifier support is enabled</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106643">Bug 106643</a> - double free when exporting a temporarily imported semaphore</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106673">Bug 106673</a> - [bisected] Steam is unusable since commit 5c33e8c7</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106687">Bug 106687</a> - radv: Fast color clears use incorrect format</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106708">Bug 106708</a> - [SKL/KBL/GLK] 2-3% performance drop in SynMark DrvState and 5-9% drop on SynMark Multithread</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106748">Bug 106748</a> - st/mesa: use PIPE_CAP_GLSL_FEATURE_LEVEL_COMPATIBILITY broke qemu -display sdl,gl=on</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106756">Bug 106756</a> - Wine 3.9 crashes with DXVK on Just Cause 3 and Quantum Break on VEGA but works ON POLARIS</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106774">Bug 106774</a> - GLSL IR copy propagates loads of SSBOs</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106776">Bug 106776</a> - vma_random unrecognized command line option &quot;-std=c++11&quot;</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106778">Bug 106778</a> - Files missing from tarball - intel_sanitize_gpu.*</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106779">Bug 106779</a> - Files missing from tarball - u_debug_stack_android.cpp</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106784">Bug 106784</a> - 18.1.1 autotools build fail without mako</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106801">Bug 106801</a> - vma_random_test.cpp:239:18: error: non-constant-expression cannot be narrowed from type 'unsigned long' to 'uint_fast32_t' (aka 'unsigned int') in initializer list [-Wc++11-narrowing]</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106810">Bug 106810</a> - ProgramBinary does not switch program correctly when using transform feedback</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106823">Bug 106823</a> - Failed to recongnize keyword of shader code</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106830">Bug 106830</a> - [bisected] 32 bit tests (deqp, piglit, glcts, vulkancts) crashing on all platforms</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106861">Bug 106861</a> - fatal error: wayland-egl-backend.h: No such file or directory compilation terminated.</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106865">Bug 106865</a> - [GLK] piglit.spec.ext_framebuffer_multisample.accuracy stencil tests fail</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106903">Bug 106903</a> - radv: Fragment shader output goes to wrong attachments when render targets are sparse</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106906">Bug 106906</a> - Failed to recongnize keyword “sampler2DRect” and &quot;sampler2DRectShadow&quot;</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106907">Bug 106907</a> - Correct Transform Feedback Varyings information is expected after using ProgramBinary</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106912">Bug 106912</a> - radv: 16-bit depth buffer causes artifacts in Shadow Warrior 2</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106928">Bug 106928</a> - When starting a match Rocket League crashes on &quot;Go&quot;</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106941">Bug 106941</a> - Intel ANV vulkan driver exposing version 1.1.0 which is incorrect</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106986">Bug 106986</a> - glGetQueryiv error when querying number of result bits for GL_ANY_SAMPLES_PASSED_CONSERVATIVE</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=106997">Bug 106997</a> - [Regression]. Dying light game is crashing on latest mesa</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107098">Bug 107098</a> - Segfault after munmap(kms_sw_dt-&gt;ro_mapped)</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107117">Bug 107117</a> - mesa-18.1: regression with TFP on intel with modesettings and glamor acceleration</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107190">Bug 107190</a> - Got seg fault on snb when use INTEL_DEBUG=bat</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107193">Bug 107193</a> - piglit.spec.arb_compute_shader.linker.bug-93840 fails</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107212">Bug 107212</a> - Dual-Core CPU E5500 / G45: RetroArch with reicast core results in corrupted graphics</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107223">Bug 107223</a> - [GEN9+] 50% perf drop in SynMark Fill* tests (E2E RBC gets disabled?)</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107248">Bug 107248</a> - [G45 ILK G965] Texture handling broken</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107275">Bug 107275</a> - NIR segfaults after spirv-opt</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107276">Bug 107276</a> - radv: OpBitfieldUExtract returns incorrect result when count is zero</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107295">Bug 107295</a> - Access violation on glDrawArrays with count &gt;= 2048</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107305">Bug 107305</a> - glsl/opt_copy_propagation_elements.cpp:72:9: error: delegating constructors are permitted only in C++11</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107312">Bug 107312</a> - Mesa-git RPM build fails after commit 8cacf38f527d42e41441ef8c25d95d4b2f4e8602</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107359">Bug 107359</a> - [Regression] [bisected] [OpenGL CTS] [SKL,BDW] KHR-GL46.texture_barrier*-texels, GTF-GL46.gtf21.GL2FixedTests.buffer_corners.buffer_corners, and GTF-GL46.gtf21.GL2FixedTests.stencil_plane_corners.stencil_plane_corners fail with some configuration</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107366">Bug 107366</a> - NIR verification crashes on piglit tests</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107423">Bug 107423</a> - vc4 build failure: &quot;v3d_decoder.c:893: undefined reference to `clif_lookup_bo'&quot;</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107443">Bug 107443</a> - Build error on arm64: v3d_decoder.c:837:17: error: format not a string literal and no format arguments [-Werror=format-security]</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107460">Bug 107460</a> - radv: OpControlBarrier does not always work correctly (bisected)</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107477">Bug 107477</a> - [DXVK] Setting high shader quality in GTA V results in LLVM error</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107510">Bug 107510</a> - [GEN8+] up to 10% perf drop on several 3D benchmarks</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107544">Bug 107544</a> - intel/decoder: out of bounds group_iter</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107550">Bug 107550</a> - &quot;0[2]&quot; as function parameter hits assert</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107579">Bug 107579</a> - [SNB] The graphic corruption when we reuse the GS compiled and used for TFB when statebuffer contain magic trash in the unused space</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107601">Bug 107601</a> - Rise of the Tomb Raider Segmentation Fault when the game starts</li>
<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=107610">Bug 107610</a> - Dolphin emulator mis-renders shadow overlay in Super Mario Sunshine</li>
</ul>
<h2>Changes</h2>
......
src/amd/common/ac_llvm_util.c
View file @ eb8a5560
......@@ -149,16 +149,13 @@ static LLVMTargetMachineRef ac_create_target_machine(enum radeon_family family,
char features[256];
const char *triple = (tm_options & AC_TM_SUPPORTS_SPILL) ? "amdgcn-mesa-mesa3d" : "amdgcn--";
LLVMTargetRef target = ac_get_llvm_target(triple);
bool barrier_does_waitcnt = (tm_options & AC_TM_AUTO_WAITCNT_BEFORE_BARRIER) &&
family != CHIP_VEGA20;
snprintf(features, sizeof(features),
"+DumpCode,+vgpr-spilling,-fp32-denormals,+fp64-denormals%s%s%s%s%s",
"+DumpCode,+vgpr-spilling,-fp32-denormals,+fp64-denormals%s%s%s%s",
tm_options & AC_TM_SISCHED ? ",+si-scheduler" : "",
tm_options & AC_TM_FORCE_ENABLE_XNACK ? ",+xnack" : "",
tm_options & AC_TM_FORCE_DISABLE_XNACK ? ",-xnack" : "",
tm_options & AC_TM_PROMOTE_ALLOCA_TO_SCRATCH ? ",-promote-alloca" : "",
barrier_does_waitcnt ? ",+auto-waitcnt-before-barrier" : "");
tm_options & AC_TM_PROMOTE_ALLOCA_TO_SCRATCH ? ",-promote-alloca" : "");
LLVMTargetMachineRef tm = LLVMCreateTargetMachine(
target,
......
src/amd/common/ac_llvm_util.h
View file @ eb8a5560
......@@ -65,7 +65,6 @@ enum ac_target_machine_options {
AC_TM_CHECK_IR = (1 << 5),
AC_TM_ENABLE_GLOBAL_ISEL = (1 << 6),
AC_TM_CREATE_LOW_OPT = (1 << 7),
AC_TM_AUTO_WAITCNT_BEFORE_BARRIER = (1 << 8),
};
enum ac_float_mode {
......
src/amd/common/ac_nir_to_llvm.c
View file @ eb8a5560
......@@ -1171,7 +1171,8 @@ static LLVMValueRef lower_gather4_integer(struct ac_llvm_context *ctx,
struct ac_image_args *args,
const nir_tex_instr *instr)
{
enum glsl_base_type stype = glsl_get_sampler_result_type(var->type);
const struct glsl_type *type = glsl_without_array(var->type);
enum glsl_base_type stype = glsl_get_sampler_result_type(type);
LLVMValueRef half_texel[2];
LLVMValueRef compare_cube_wa = NULL;
LLVMValueRef result;
......@@ -1356,7 +1357,8 @@ static LLVMValueRef build_tex_intrinsic(struct ac_nir_context *ctx,
if (instr->op == nir_texop_tg4 && ctx->ac.chip_class <= VI) {
nir_deref_instr *texture_deref_instr = get_tex_texture_deref(instr);
nir_variable *var = nir_deref_instr_get_variable(texture_deref_instr);
enum glsl_base_type stype = glsl_get_sampler_result_type(var->type);
const struct glsl_type *type = glsl_without_array(var->type);
enum glsl_base_type stype = glsl_get_sampler_result_type(type);
if (stype == GLSL_TYPE_UINT || stype == GLSL_TYPE_INT) {
return lower_gather4_integer(&ctx->ac, var, args, instr);
}
......
src/amd/common/ac_surface.c
View file @ eb8a5560
......@@ -588,8 +588,8 @@ void ac_compute_cmask(const struct radeon_info *info,
unsigned base_align = num_pipes * pipe_interleave_bytes;
unsigned width = align(config->info.width, cl_width*8);
unsigned height = align(config->info.height, cl_height*8);
unsigned width = align(surf->u.legacy.level[0].nblk_x, cl_width*8);
unsigned height = align(surf->u.legacy.level[0].nblk_y, cl_height*8);
unsigned slice_elements = (width * height) / (8*8);
/* Each element of CMASK is a nibble. */
......
src/amd/common/sid.h
View file @ eb8a5560
......@@ -9123,7 +9123,9 @@
#define CIK_SDMA_PACKET_SEMAPHORE 0x7
#define CIK_SDMA_PACKET_CONSTANT_FILL 0xb
#define CIK_SDMA_PACKET_SRBM_WRITE 0xe
#define CIK_SDMA_COPY_MAX_SIZE 0x3fffe0
/* There is apparently an undocumented HW "feature" that
prevents the HW from copying past 256 bytes of (1 << 22) */
#define CIK_SDMA_COPY_MAX_SIZE 0x3fff00
 
enum amd_cmp_class_flags {
S_NAN = 1 << 0, // Signaling NaN
......
src/amd/vulkan/radv_device.c
View file @ eb8a5560
......@@ -60,7 +60,7 @@ radv_device_get_cache_uuid(enum radeon_family family, void *uuid)
memcpy(uuid, &mesa_timestamp, 4);
memcpy((char*)uuid + 4, &llvm_timestamp, 4);
memcpy((char*)uuid + 8, &f, 2);
snprintf((char*)uuid + 10, VK_UUID_SIZE - 10, "radv");
snprintf((char*)uuid + 10, VK_UUID_SIZE - 10, "radv%zd", sizeof(void *));
return 0;
}
......@@ -1895,10 +1895,30 @@ radv_get_hs_offchip_param(struct radv_device *device, uint32_t *max_offchip_buff
device->physical_device->rad_info.family != CHIP_CARRIZO &&
device->physical_device->rad_info.family != CHIP_STONEY;
unsigned max_offchip_buffers_per_se = double_offchip_buffers ? 128 : 64;
unsigned max_offchip_buffers = max_offchip_buffers_per_se *
device->physical_device->rad_info.max_se;
unsigned max_offchip_buffers;
unsigned offchip_granularity;
unsigned hs_offchip_param;
/*
* Per RadeonSI:
* This must be one less than the maximum number due to a hw limitation.
* Various hardware bugs in SI, CIK, and GFX9 need this.
*
* Per AMDVLK:
* Vega10 should limit max_offchip_buffers to 508 (4 * 127).
* Gfx7 should limit max_offchip_buffers to 508
* Gfx6 should limit max_offchip_buffers to 126 (2 * 63)
*
* Follow AMDVLK here.
*/
if (device->physical_device->rad_info.family == CHIP_VEGA10 ||
device->physical_device->rad_info.chip_class == CIK ||
device->physical_device->rad_info.chip_class == SI)
--max_offchip_buffers_per_se;
max_offchip_buffers = max_offchip_buffers_per_se *
device->physical_device->rad_info.max_se;
switch (device->tess_offchip_block_dw_size) {
default:
assert(0);
......
src/amd/vulkan/radv_formats.c
View file @ eb8a5560
......@@ -1112,6 +1112,25 @@ static VkResult radv_get_image_format_properties(struct radv_physical_device *ph
}
}
if (info->usage & VK_IMAGE_USAGE_TRANSFER_SRC_BIT) {
if (!(format_feature_flags & VK_FORMAT_FEATURE_TRANSFER_SRC_BIT)) {
goto unsupported;
}
}
if (info->usage & VK_IMAGE_USAGE_TRANSFER_DST_BIT) {
if (!(format_feature_flags & VK_FORMAT_FEATURE_TRANSFER_DST_BIT)) {
goto unsupported;
}
}
if (info->usage & VK_IMAGE_USAGE_INPUT_ATTACHMENT_BIT) {
if (!(format_feature_flags & (VK_FORMAT_FEATURE_COLOR_ATTACHMENT_BIT |
VK_FORMAT_FEATURE_DEPTH_STENCIL_ATTACHMENT_BIT))) {
goto unsupported;
}
}
*pImageFormatProperties = (VkImageFormatProperties) {
.maxExtent = maxExtent,
.maxMipLevels = maxMipLevels,
......
src/amd/vulkan/radv_image.c
View file @ eb8a5560
......@@ -801,8 +801,8 @@ radv_image_get_cmask_info(struct radv_device *device,
unsigned base_align = num_pipes * pipe_interleave_bytes;
unsigned width = align(image->info.width, cl_width*8);
unsigned height = align(image->info.height, cl_height*8);
unsigned width = align(image->surface.u.legacy.level[0].nblk_x, cl_width*8);
unsigned height = align(image->surface.u.legacy.level[0].nblk_y, cl_height*8);
unsigned slice_elements = (width * height) / (8*8);
/* Each element of CMASK is a nibble. */
......
src/amd/vulkan/radv_nir_to_llvm.c
View file @ eb8a5560
......@@ -2106,9 +2106,10 @@ handle_fs_input_decl(struct radv_shader_context *ctx,
int idx = variable->data.location;
unsigned attrib_count = glsl_count_attribute_slots(variable->type, false);
LLVMValueRef interp = NULL;
uint64_t mask;
variable->data.driver_location = idx * 4;
ctx->input_mask |= ((1ull << attrib_count) - 1) << variable->data.location;
mask = ((1ull << attrib_count) - 1) << variable->data.location;
if (glsl_get_base_type(glsl_without_array(variable->type)) == GLSL_TYPE_FLOAT) {
unsigned interp_type;
......@@ -2129,6 +2130,15 @@ handle_fs_input_decl(struct radv_shader_context *ctx,
for (unsigned i = 0; i < attrib_count; ++i)
ctx->inputs[ac_llvm_reg_index_soa(idx + i, 0)] = interp;
if (idx == VARYING_SLOT_CLIP_DIST0) {
/* Do not account for the number of components inside the array
* of clip/cull distances because this might wrongly set other
* bits like primitive ID or layer.
*/
mask = 1ull << VARYING_SLOT_CLIP_DIST0;
}
ctx->input_mask |= mask;
}
static void
......@@ -2195,6 +2205,17 @@ handle_fs_inputs(struct radv_shader_context *ctx,
if (LLVMIsUndef(interp_param))
ctx->shader_info->fs.flat_shaded_mask |= 1u << index;
++index;
} else if (i == VARYING_SLOT_CLIP_DIST0) {
int length = ctx->shader_info->info.ps.num_input_clips_culls;
for (unsigned j = 0; j < length; j += 4) {
inputs = ctx->inputs + ac_llvm_reg_index_soa(i, j);
interp_param = *inputs;
interp_fs_input(ctx, index, interp_param,
ctx->abi.prim_mask, inputs);
++index;
}
} else if (i == VARYING_SLOT_POS) {
for(int i = 0; i < 3; ++i)
inputs[i] = ctx->abi.frag_pos[i];
......@@ -2490,6 +2511,13 @@ handle_vs_outputs_post(struct radv_shader_context *ctx,
memcpy(&pos_args[target - V_008DFC_SQ_EXP_POS],
&args, sizeof(args));
/* Export the clip/cull distances values to the next stage. */
radv_export_param(ctx, param_count, &slots[0], 0xf);
outinfo->vs_output_param_offset[VARYING_SLOT_CLIP_DIST0] = param_count++;
if (ctx->num_output_clips + ctx->num_output_culls > 4) {
radv_export_param(ctx, param_count, &slots[4], 0xf);
outinfo->vs_output_param_offset[VARYING_SLOT_CLIP_DIST1] = param_count++;
}
}
LLVMValueRef pos_values[4] = {ctx->ac.f32_0, ctx->ac.f32_0, ctx->ac.f32_0, ctx->ac.f32_1};
......
src/amd/vulkan/radv_pipeline.c
View file @ eb8a5560
......@@ -2514,6 +2514,7 @@ radv_pipeline_generate_binning_state(struct radeon_cmdbuf *cs,
switch (pipeline->device->physical_device->rad_info.family) {
case CHIP_VEGA10:
case CHIP_VEGA12:
case CHIP_VEGA20:
context_states_per_bin = 1;
persistent_states_per_bin = 1;
fpovs_per_batch = 63;
......@@ -3027,6 +3028,23 @@ radv_pipeline_generate_ps_inputs(struct radeon_cmdbuf *cs,
ps_offset++;
}
if (ps->info.info.ps.num_input_clips_culls) {
unsigned vs_offset;
vs_offset = outinfo->vs_output_param_offset[VARYING_SLOT_CLIP_DIST0];
if (vs_offset != AC_EXP_PARAM_UNDEFINED) {
ps_input_cntl[ps_offset] = offset_to_ps_input(vs_offset, true);
++ps_offset;
}
vs_offset = outinfo->vs_output_param_offset[VARYING_SLOT_CLIP_DIST1];
if (vs_offset != AC_EXP_PARAM_UNDEFINED &&
ps->info.info.ps.num_input_clips_culls > 4) {
ps_input_cntl[ps_offset] = offset_to_ps_input(vs_offset, true);
++ps_offset;
}
}
for (unsigned i = 0; i < 32 && (1u << i) <= ps->info.fs.input_mask; ++i) {
unsigned vs_offset;
bool flat_shade;
......
src/amd/vulkan/radv_shader.c
View file @ eb8a5560
......@@ -259,9 +259,6 @@ radv_shader_compile_to_nir(struct radv_device *device,
*/
NIR_PASS_V(nir, nir_lower_constant_initializers, nir_var_shader_out);
NIR_PASS_V(nir, nir_remove_dead_variables,
nir_var_shader_in | nir_var_shader_out | nir_var_system_value);
/* Now that we've deleted all but the main function, we can go ahead and
* lower the rest of the constant initializers.
*/
......@@ -273,6 +270,9 @@ radv_shader_compile_to_nir(struct radv_device *device,
NIR_PASS_V(nir, nir_split_var_copies);
NIR_PASS_V(nir, nir_split_per_member_structs);
NIR_PASS_V(nir, nir_remove_dead_variables,
nir_var_shader_in | nir_var_shader_out | nir_var_system_value);
NIR_PASS_V(nir, nir_lower_system_values);
NIR_PASS_V(nir, nir_lower_clip_cull_distance_arrays);
}
......
src/amd/vulkan/radv_shader.h
View file @ eb8a5560
......@@ -174,6 +174,7 @@ struct radv_shader_info {
bool has_pcoord;
bool prim_id_input;
bool layer_input;
uint8_t num_input_clips_culls;
} ps;
struct {
bool uses_grid_size;
......
src/amd/vulkan/radv_shader_info.c
View file @ eb8a5560
......@@ -341,6 +341,7 @@ static void
gather_info_input_decl_ps(const nir_shader *nir, const nir_variable *var,
struct radv_shader_info *info)
{
unsigned attrib_count = glsl_count_attribute_slots(var->type, false);
const struct glsl_type *type = glsl_without_array(var->type);
int idx = var->data.location;
......@@ -354,6 +355,9 @@ gather_info_input_decl_ps(const nir_shader *nir, const nir_variable *var,
case VARYING_SLOT_LAYER:
info->ps.layer_input = true;
break;
case VARYING_SLOT_CLIP_DIST0:
info->ps.num_input_clips_culls = attrib_count;
break;
default:
break;
}
......
src/amd/vulkan/si_cmd_buffer.c
View file @ eb8a5560
......@@ -342,6 +342,7 @@ si_emit_config(struct radv_physical_device *physical_device,
switch (physical_device->rad_info.family) {
case CHIP_VEGA10:
case CHIP_VEGA12:
case CHIP_VEGA20:
pc_lines = 4096;
break;
case CHIP_RAVEN:
......