Compare commits

..

347 Commits

Author SHA1 Message Date
Andres Gomez
a4b72e2643 docs: add release notes for 17.2.4
Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-30 16:46:20 +02:00
Andres Gomez
fe9dc0fad6 Update version to 17.2.4
Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-30 16:37:54 +02:00
Andres Gomez
162fa27b12 cherry-ignore: broadcom/vc5: Propagate vc4 aliasing fix to vc5.
extra: Commit is not applicable when ade416d023 is missing.

Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-27 18:07:38 +03:00
Andres Gomez
b3207d9ff7 cherry-ignore: mesa/bufferobj: don't double negate the range
fixes: This commit addressed earlier commit 35ac13ed3 which did not
land in branch.

Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-27 18:07:38 +03:00
Andres Gomez
6cab29e973 cherry-ignore: radv: Disallow indirect outputs for GS on GFX9 as well.
fixes: Commit is not applicable when 6ce550453f is missing.

Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-27 18:07:38 +03:00
Andres Gomez
1c2f79066c cherry-ignore: radv: Don't use vgpr indexing for outputs on GFX9.
fixes: Commit is not applicable when 087e010b2b is missing.

Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-27 18:07:38 +03:00
Andres Gomez
0c63d53765 cherry-ignore: added 17.3 nominations.
stable: 17.3 nominations only.

Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-27 18:07:38 +03:00
Andres Gomez
52019bad1d cherry-ignore: glsl: fix derived cs variables
stable: Commit is too big for stable at this point.

Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-27 18:07:38 +03:00
Andres Gomez
9bdd943ff2 cherry-ignore: configure.ac: rework llvm detection and handling
stable: Commits are too invasive for 17.2.

Signed-off-by: Andres Gomez <agomez@igalia.com>
2017-10-27 18:07:38 +03:00
Jason Ekstrand
1393d37d7b intel/eu: Use EXECUTE_1 for JMPI
The PRM says "The execution size must be 1."  In 73137997e2, the
execution size was set to 1 when it should have been BRW_EXECUTE_1
(which maps to 0).  Later, in dc2d3a7f5c, JMPI was used for
line AA on gen6 and earlier and we started manually stomping the
exeution size to BRW_EXECUTE_1 in the generator.  This commit fixes the
original bug and makes brw_JMPI just do the right thing.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Fixes: 73137997e2
(cherry picked from commit 562b8d458c)
2017-10-27 18:07:38 +03:00
Jason Ekstrand
74f1903234 anv/pipeline: Call nir_lower_system_valaues after brw_preprocess_nir
We currently have a bug where nir_lower_system_values gets called before
nir_lower_var_copies so it will miss any system value uses which come
from a copy_var intrinsic.  Moving it to after brw_preprocess_nir fixes
this problem.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 279f8fb69c)
[Andres Gomez: resolve trivial conflicts]
Signed-off-by: Andres Gomez <agomez@igalia.com>

Conflicts:
	src/intel/vulkan/anv_pipeline.c
2017-10-27 18:07:37 +03:00
Jason Ekstrand
aac0807f48 intel/fs: Handle flag read/write aliasing in needs_src_copy
In order to implement the ballot intrinsic, we do a MOV from flag
register to some GRF.  If that GRF is used in a SEL, cmod propagation
helpfully changes it into a MOV from the flag register with a cmod.
This is perfectly valid but when lower_simd_width comes along, it simply
splits into two instructions which both have conditional modifiers.
This is a problem since we're reading the flag register.  This commit
makes us check whether or not flags_written() overlaps with the flag
values that we are reading via the instruction source and, if we have
any interference, will force us to emit a copy of the source.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit fa6e74e33e)
2017-10-27 18:07:37 +03:00
Jan Vesely
03f3899f99 clover: Fix compilation after clang r315871
v2: use a more generic compat function
v3: rename and formatting cleanup

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103388
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
CC: <mesa-stable@lists.freedesktop.org>
(cherry picked from commit a6d38f476b)
2017-10-27 18:07:37 +03:00
Jason Ekstrand
87a9a989ee nir/intrinsics: Set the correct num_indices for load_output
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
(cherry picked from commit c1b84256cc)
2017-10-27 18:07:37 +03:00
Matthew Nicholls
ce725baa7c ac/nir: generate correct instruction for atomic min/max on unsigned images
v2: fix silly typo

Cc: "17.2 17.3" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
(cherry picked from commit 27a0b24bf2)
[Andres Gomez: resolve trivial conflicts]
Signed-off-by: Andres Gomez <agomez@igalia.com>

Conflicts:
	src/amd/common/ac_nir_to_llvm.c
2017-10-27 18:07:37 +03:00
Bas Nieuwenhuizen
f46ba9ee35 ac/nir: Fix nir_texop_lod on GFX for 1D arrays.
Fixes: 1bcb953e16 'radv: handle GFX9 1D textures'
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 2c5b43c87f)
2017-10-27 18:07:37 +03:00
Stefan Schake
2d51d41865 broadcom/vc4: Fix aliasing issue
This was causing Android clang version 3.8.256229 to miscompile,
presumably due to strict aliasing.

Fixes: 14dc281c13 ("vc4: Enforce one-uniform-per-instruction after optimization.")
(cherry picked from commit e5fea0d621)
2017-10-27 18:07:37 +03:00
Kenneth Graunke
cbc081b871 i965: Revert absolute mode for constant buffer pointers.
The kernel doesn't initialize the value of the INSTPM or CS_DEBUG_MODE2
registers at context initialization time.  Instead, they're inherited
from whatever happened to be running on the GPU prior to first run of a
new context.  So, when we started setting these, other contexts in the
system started inheriting our values.  Since this controls whether
3DSTATE_CONSTANT_* takes a pointer or an offset, getting the wrong
setting is fatal for almost any process which isn't expecting this.

Unfortunately, VA-API and Beignet don't initialize this (nor does older
Mesa), so they will die horribly if we start doing this.  UXA and SNA
don't use any push constants, so they are unaffected.

Until we have some kind of solution to this problem, I'm going to revert
this patch and abandon using the feature for now.  It will lead to fewer
pushed UBO ranges on Broadwell+, which may lead to lower performance,
though I don't have any data on the impact.

Cc: "17.3 17.2" <mesa-stable@lists.freedesktop.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102774
(cherry picked from commit 013d331220)
[Andres Gomez: resolve trivial conflicts]
Signed-off-by: Andres Gomez <agomez@igalia.com>

Conflicts:
	src/mesa/drivers/dri/i965/brw_state_upload.c
	src/mesa/drivers/dri/i965/intel_screen.c
2017-10-27 18:07:24 +03:00
Michel Dänzer
546b4d455a st/mesa: Initialize textures array in st_framebuffer_validate
And just reference pipe_resources to it in the validate callbacks.

Avoids pipe_resource leaks when st_framebuffer_validate ends up calling
the validate callback multiple times, e.g. when a window is resized.

v2:
* Use generic stable tag instead of Fixes: tag, since the problem could
  already happen before the commit referenced in v1 (Thomas Hellstrom)
* Use memset to initialize the array on the stack instead of allocating
  the array with os_calloc.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
(cherry picked from commit 7561da367b)

Squashed with:

st/osmesa: include u_inlines.h for pipe_resource_reference

Fixes build failure due to unresolved symbol.

Fixes: 7561da367b "st/mesa: Initialize textures array in
                     st_framebuffer_validate"

Trivial.

(cherry picked from commit 8c9e7c9638)
2017-10-27 18:04:59 +03:00
Henri Verbeet
9cbf8c910e vulkan/wsi: Free the event in x11_manage_fifo_queues().
Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Henri Verbeet <hverbeet@gmail.com>
Fixes: e73d136a02 ("vulkan/wsi/x11: Implement FIFO mode.")
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com
(cherry picked from commit 3de87f7cd7)
2017-10-27 18:04:59 +03:00
Dave Airlie
1eb4cbc934 radv/image: bump all the offset to uint64_t.
So one of the CTS tests tries to allocate a 16384x1 2048 array
texture. This overflows a bunch of calculations when we want it
tiled as the heights goes to 128.

addrlib returns us the correct size (16GB or so), but we mangle
it in the htile calcs due to the 32-bit offset fields, then
userspace gives us the reduced number and we try to allocate
it on a heap and things blow up.

We really need to give the app back the correct size for the
image so we can blow up properly in memory allocation later.

This should fix hangs in
dEQP-VK.pipeline.render_to_image.core.1d_array.huge.width_layers.r8g8b8a8_unorm_d32_sfloat_s8_uint
since
Fixes: ad3d98da9f (radv: enable tc compatible htile for d32s8 also.)

Now there's an open question if we should be enabling tc-compat
htile at all for shallow textures like the above.

This might cause some other wierd side effects in CTS even
without the tc compat so:
Cc: "17.2" <mesa-stable@lists.freedesktop.org>

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 35c66f3e40)
[Andres Gomez: resolve trivial conflicts]
Signed-off-by: Andres Gomez <agomez@igalia.com>

Conflicts:
	src/amd/vulkan/radv_private.h
2017-10-27 18:04:59 +03:00
Marek Olšák
fba44d91d0 Revert "mesa: fix texture updates for ATI_fragment_shader"
This reverts commit 9d54025cd1.

It breaks KOTOR.

Cc: 17.1 17.2 <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 5d071bf04b)
2017-10-27 18:04:59 +03:00
Samuel Pitoiset
9cba4d491c radv: add the draw count buffer to the list of buffers
My guess is that the GPU is going to report VM faults if
vkCmdDrawIndirectCountAMD() (and friends) are used.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
(cherry picked from commit 3e5f27faf3)
2017-10-27 18:04:59 +03:00
Emil Velikov
facc851818 docs: add sha256 checksums for 17.2.3
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-10-19 13:28:13 +01:00
Emil Velikov
28dc4b64f2 docs: add release notes for 17.2.3
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-10-19 13:10:20 +01:00
Emil Velikov
ea38f4c33a Update version to 17.2.3
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-10-19 13:02:58 +01:00
Emil Velikov
23c08dabc3 eglmesaext: add forward declaration for struct wl_buffers
The user does not need to know the specifics of the struct, as only a
pointer to it is used.

Just forward declare the struct making the header self-contained.

v2: Remove deprecation warning text/bugzilla - patch does no help there.

Cc: Greg V <greg@unrelenting.technology>
Fixes: 5cddb1ce3c ("wayland: Add an extension to create wl_buffers from
EGLImages")
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com> (v1)
(cherry picked from commit 66ebdfbd44)
2017-10-17 16:59:31 +01:00
Emil Velikov
dc9bd1dade wayland-drm: use a copy of the wayland_drm_callbacks struct
The callbacks may be called even when they are no longer valid.
Say, the user is dlclose(ing) libEGL while the buffers are being
destroyed.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Tested-by: Derek Foreman <derekf@osg.samsung.com>
(cherry picked from commit 0cfd6f6cfc)
2017-10-17 16:59:31 +01:00
Jason Ekstrand
d001ff1267 nir: Get rid of the variable on vote intrinsics
This looks like a copy+paste error.  They don't actually write into that
variable as would be implied by putting the return there.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 3442c9fc3e)
2017-10-17 16:59:31 +01:00
Jason Ekstrand
88a16c895b nir/opcodes: Fix constant-folding of ufind_msb
We didn't fold correctly in the case of 0x1 because we never let the
loop counter hit 0.  Switching it to bit >= 0 solves this problem.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit a0947921eb)
2017-10-17 16:59:31 +01:00
Jason Ekstrand
b640bf38ca glsl/blob: Return false from grow_to_fit if we've ever failed
Otherwise we could have a failure followed by a smaller write that
succeeds and get a corrupted blob.  If we ever OOM, we should stop.

v2 (Jason Ekstrand):
 - Initialize the new boolean member in create_blob

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit e03717efbd)
2017-10-17 16:59:31 +01:00
Jason Ekstrand
4d1ae3283c glsl/blob: Return false from ensure_can_read on overrun
Otherwise, if you have a large read fail and then try to do a small
read, the small read may succeed even though it's at the wrong offset.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 7118851374)
2017-10-17 16:59:31 +01:00
Eric Engestrom
d56aa9fe43 scons: use python3-compatible print()
These changes were generated using python's `2to3` tool.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102852
Reported-by: Alex Granni <liviuprodea@yahoo.com>
Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
(cherry picked from commit 7d48219b3a)
2017-10-17 16:59:31 +01:00
Bas Nieuwenhuizen
3b657e4ff5 radv: Only set the MTYPE flags on GFX9+.
Older kernels fail the va_op with this flag set. If the kernel
supports GFX9 usefully, it will also support this flag.

Fixes: e8d57802fe "radv/gfx9: allocate events from uncached VA space"
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 96f80c8d4d)
2017-10-17 16:59:31 +01:00
Daniel Stone
99d3661bce egl/wayland: Don't use dmabuf with no modifiers
The dmabuf interface requires a valid modifier to be sent. If we don't
explicitly get a modifier from the driver, we can't know what to send;
it must be inferred from legacy side-channels (or assumed to linear, if
none exists).

If we have no modifier, then we can only have a single-plane format
anyway, so fall back to the old wl_drm buffer import path.

Fixes: a65db0ad1c ("st/dri: don't expose modifiers in EGL if the driver doesn't implement them")
Fixes: 02cc359372 ("egl/wayland: Use linux-dmabuf interface for buffers")
Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Reported-by: Andy Furniss <adf.lists@gmail.com>
Cc: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit b65d6dafd6)
2017-10-17 16:59:31 +01:00
Daniel Stone
0f6e89dfe0 egl/wayland: Check queryImage return for wl_buffer
When creating a wl_buffer from a DRIImage, we extract all the DRIImage
information via queryImage. Check whether or not it actually succeeds,
either bailing out if the query was critical, or providing sensible
fallbacks for information which was not available in older DRIImage
versions.

Fixes: a65db0ad1c ("st/dri: don't expose modifiers in EGL if the driver doesn't implement them")
Fixes: 02cc359372 ("egl/wayland: Use linux-dmabuf interface for buffers")
Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Reported-by: Andy Furniss <adf.lists@gmail.com>
Cc: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 6273d2f269)
2017-10-17 16:59:31 +01:00
Emil Velikov
bee97ec32e swr/rast: do not crash on NULL strings returned by getenv
The current convenience function GetEnv feeds the results of getenv
directly into std::string(). That is a bad idea, since the variable
may be unset, thus we feed NULL into the C++ construct.

The latter of which is not allowed and leads to a crash.

v2: Better variable name, implicit char* -> std::string conversion (Eric)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101832
Fixes: a25093de71 ("swr/rast: Implement JIT shader caching to disk")
Cc: Tim Rowley <timothy.o.rowley@intel.com>
Cc: Laurent Carlier <lordheavym@gmail.com>
Cc: Bernhard Rosenkraenzer <bero@lindev.ch>
[Emil Velikov: make an actual commit from the misc diff]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com> (v1)
Reviewed-by: Laurent Carlier <lordheavym@gmail.com> (v1)
(cherry picked from commit 21e271024d)
2017-10-17 16:59:31 +01:00
Nicolai Hähnle
9f4b0a336c radeonsi: clamp border colors for upgraded depth textures
The hardware does this automatically for unorm formats, but we need to
do it manually for unorm depth formats that have been upgraded to
Z32_FLOAT.

Fixes dEQP-GLES31.functional.texture.border_clamp.range_clamp.nearest_unorm_depth
and others.

Fixes: d4d9ec55c5 ("radeonsi: implement TC-compatible HTILE")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
(cherry picked from commit 6eb9483912)
2017-10-17 16:59:31 +01:00
Nicolai Hähnle
74a28d85de radeonsi: clamp depth comparison value only for fixed point formats
The hardware usually does this automatically. However, we upgrade
depth to Z32_FLOAT to enable TC-compatible HTILE, which means the
hardware no longer clamps the comparison value for us.

The only way to tell in the shader whether a clamp is required
seems to be to communicate an additional bit in the descriptor
table. While VI has some unused bits in the resource descriptor,
those bits have unfortunately all been used in gfx9. So we use
an unused bit in the sampler state instead.

Fixes dEQP-GLES3.functional.texture.shadow.2d.linear.equal_depth_component32f
and many other tests in dEQP-GLES3.functional.texture.shadow.*

Fixes: d4d9ec55c5 ("radeonsi: implement TC-compatible HTILE")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
(cherry picked from commit 4c56e07029)
[Emil Velikov: handle lack of dirty_mask in original patch]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/gallium/drivers/radeonsi/si_descriptors.c
2017-10-17 16:59:31 +01:00
Nicolai Hähnle
f805a61e04 st/glsl_to_tgsi: fix a use-after-free in merge_two_dsts
Found by address sanitizer.

The loop here tries to be safe, but in doing so, it ends up doing
exactly the wrong thing: the safe foreach is for when the loop
variable (inst) could be deleted and nothing else. However, this
particular can delete inst's successor, but not inst itself.

Fixes: 8c6a0ebaad ("st/mesa: add st fp64 support (v7.1)")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
(cherry picked from commit 2703fa613b)
2017-10-17 16:59:31 +01:00
Lionel Landwerlin
6957dfb0d8 anv: bo_cache: allow importing a BO larger than needed
It's not a problem if a BO has been allocated larger than we need it
to be.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102940
Fixes: 818b857914 ("anv: Use the BO cache for DeviceMemory allocations")
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit c0a4f56fb9)
2017-10-17 16:59:31 +01:00
Nicolai Hähnle
410f4dbcb1 st/glsl_to_tgsi: fix indirect access to 64-bit integer
Make sure we actually allocate two adjacent TGSI temporaries. The
current code fails e.g. when an arithmetic operation has two
operands with indirect accesses.

I will send out a new piglit test
(arb_gpu_shader_int64/execution/indirect-array-two-accesses.shader_test)

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 541208cf13)
2017-10-17 16:59:31 +01:00
Ilia Mirkin
d22e779d6a nv50,nvc0: fix push hint logic in presence of a start offset
Previously buffer offsets were passed in explicitly as an offset, which
had to be added to the resource address. Now they are passed in via an
increased 'start' parameter. As a result, we were double-adding the
start offset in this kind of situation.

This condition was triggered by piglit's draw-elements test which has a
requisite glMultiDrawElements in combination with a small enough number
of vertices to go through the immediate push path.

Fixes: 330d0607ed ("gallium: remove pipe_index_buffer and set_index_buffer")
Reported-by: Karol Herbst <karolherbst@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit b20bccbcac)
2017-10-17 16:59:31 +01:00
Dave Airlie
41ec2af2a8 radv: lower ffma in nir.
So it appears the Vulkan SPIR-V fma opcode can be equivalent to a
mad operation, and the fma hw opcode on AMD hw is issued like a double
opcode so is slower. Also the radeonsi stack does this.

This appears to improve performance on a number of games from Feral,
and thanks to Feral for noticing the problem.

I'm reposting this one as Marek indicated he thinks this is what
we should be doing on AMD hw.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 2c61594d84)
[Emil Velikov: use correct file radv_shader.c -> radv_pipeline.c]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/amd/vulkan/radv_shader.c
2017-10-17 16:59:31 +01:00
Alex Smith
0bd7be0142 radv: Add R16G16B16A16_SNORM fast clear support
Signed-off-by: Alex Smith <asmith@feralinteractive.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 25d76fd658)
2017-10-17 16:59:30 +01:00
Nicolai Hähnle
4bcadb533f st/mesa: don't clobber glGetInternalformat* buffer for GL_NUM_SAMPLE_COUNTS
Applications might pass in a buffer that is sized too large and rely
on the extra space of the buffer not being overwritten.

Fixes dEQP-GLES31.functional.state_query.internal_format.partial_query.num_sample_counts

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 9a8f13a33b)
2017-10-17 16:59:30 +01:00
Ilia Mirkin
2eae2a6f0e nv50/ir: fix 64-bit integer shifts
TGSI was adjusted to always pass in 64-bit integers but nouveau was left
with the old semantics. Update to the new thing.

Fixes: d10fbe5159 (st/glsl_to_tgsi: fix 64-bit integer bit shifts)
Reported-by: Karol Herbst <karolherbst@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit ce6da2a026)
2017-10-17 16:59:30 +01:00
Józef Kucia
077f925473 anv: Do not assert() on VK_ATTACHMENT_UNUSED
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 91ba331ef4)
2017-10-17 16:59:30 +01:00
Józef Kucia
2e92d16f9d spirv: Fix SpvOpAtomicISub
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit e0acb630a5)
2017-10-17 16:59:30 +01:00
Emil Velikov
83dcf9dc33 cherry-ignore: add "anv/wsi: Allocate enough memory for the entire image"
Addresses bug introduced with a feature patch, which is not in branch.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-10-17 16:59:28 +01:00
Lionel Landwerlin
406e7e0e17 anv/cmd_buffer: Reset state in cmd_buffer_destroy
This ensures that everything gets cleaned up properly. In particular,
it fixes a memory leak where we were leaking the push constants
structs.

Valgrind stats on
dEQP-VK.pipeline.push_constant.graphics_pipeline.range_size_128 :

Before:
HEAP SUMMARY:
    in use at exit: 2,467,513 bytes in 1,305 blocks
  total heap usage: 697,853 allocs, 696,530 frees, 138,466,600 bytes allocated

LEAK SUMMARY:
   definitely lost: 1,068 bytes in 11 blocks
   indirectly lost: 24,669 bytes in 412 blocks
     possibly lost: 0 bytes in 0 blocks
   still reachable: 2,441,776 bytes in 882 blocks
        suppressed: 0 bytes in 0 blocks

After:
HEAP SUMMARY:
    in use at exit: 2,467,381 bytes in 1,304 blocks
  total heap usage: 697,853 allocs, 696,531 frees, 138,466,600 bytes allocated

LEAK SUMMARY:
   definitely lost: 936 bytes in 10 blocks
   indirectly lost: 24,669 bytes in 412 blocks
     possibly lost: 0 bytes in 0 blocks
   still reachable: 2,441,776 bytes in 882 blocks
        suppressed: 0 bytes in 0 blocks

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: "17.2 17.1" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 0763f814d7)
2017-10-17 16:59:02 +01:00
Lionel Landwerlin
60466859fa anv/cmd_buffer: fix push descriptors with set > 0
When writing to set > 0, we were just wrongly writing to set 0. This
commit fixes this by lazily allocating each set as we write to them.

We didn't go for having them directly into the command buffer as this
would require an additional ~45Kb per command buffer.

v2: Allocate push descriptors from system memory rather than in BO
    streams. (Lionel)

Cc: "17.2 17.1" <mesa-stable@lists.freedesktop.org>
Fixes: 9f60ed98e5 ("anv: add VK_KHR_push_descriptor support")
Reported-by: Daniel Ribeiro Maciel <daniel.maciel@gmail.com>
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
(cherry picked from commit d296dea54e)
2017-10-17 16:59:02 +01:00
Marek Olšák
014b5a7209 glsl_to_tgsi: fix instruction order for bindless textures
We emitted instructions loading the bindless handle after the memory
instruction.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit 985338e2cb)
2017-10-17 16:59:02 +01:00
Jason Ekstrand
f9c4f22f5a intel/compiler: Don't propagate cmod into integer multiplies
No shader-db change on Sky Lake.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 7463d50580)
2017-10-17 16:59:02 +01:00
Jason Ekstrand
4ae1a62b26 intel/compiler: Don't cmod propagate into a saturated operation
Shader-db results on Sky Lake:

    total instructions in shared programs: 12954445 -> 12955125 (0.01%)
    instructions in affected programs: 141862 -> 142542 (0.48%)
    helped: 0
    HURT: 626

Reviewed-by: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit b91ecee04a)
2017-10-17 16:59:02 +01:00
Ben Crocker
ea3ad52ad3 gallivm/ppc64le: allow environmental control of Altivec code generation
In check_os_altivec_support(), allow control of Altivec (first PPC vector
instruction set) code generation via a new environmental control,
GALLIVM_ALTIVEC, which is expected to take on a value of 1 or 0.
The default is to enable Altivec code generation.

This environmental control of Altivec code generation is initially
available only #ifdef DEBUG.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Ben Crocker <bcrocker@redhat.com>
Acked-by: Roland Scheidegger <sroland@vmware.com>
(cherry picked from commit 1359af930e)
2017-10-17 16:59:02 +01:00
Ben Crocker
3f5d2b768c gallivm/ppc64le: adjust VSX code generation control.
In lp_build_create_jit_compiler_for_module(), advance the minimum
version of LLVM for VSX code generation to 4.0; this is the minimum
revision at which several known VSX code generation bugs are fixed:

  https://llvm.org/bugs/show_bug.cgi?id=25503 (fixed in 3.8.1)
  https://llvm.org/bugs/show_bug.cgi?id=26775 (fixed in 3.8.1)
  https://llvm.org/bugs/show_bug.cgi?id=33531 (fixed in 4.0)

An llc performance bug introduced in LLVM 4.0,

  https://llvm.org/bugs/show_bug.cgi?id=34647

is still pending as of LLVM 5.0, but only has a pronounced effect on
one of the Piglit tests: ext_transform_feedback-max-varyings.

All changes tested via Piglit.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Ben Crocker <bcrocker@redhat.com>
Acked-by: Roland Scheidegger <sroland@vmware.com>
(cherry picked from commit e93f056a4e)
2017-10-17 16:59:02 +01:00
Ben Crocker
070d2dcfac gallivm: allow additional llc options
In init_native_targets, allow the passing of additional options to
the LLC compiler via new GALLIVM_LLC_OPTIONS environmental control.
This option is available only #ifdef DEBUG, initially.
At top, add #include <llvm-c/Support.h> for LLVMParseCommandLineOptions()
declaration.

v2: Fix compile error with old llvm versions (sroland)

Cc: "17.2" <mesa-stable@lists.freedesktop.org>

Signed-off-by: Ben Crocker <bcrocker@redhat.com>
Acked-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
(cherry picked from commit 5c75f0c8bb)
2017-10-17 16:59:02 +01:00
Ben Crocker
1a8ccdc6e9 gallivm: fix typo in debug_printf message
In gallivm_compile_module, fix a typo in the
debug_printf("Invoke as \"llc ..." message.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>

Signed-off-by: Ben Crocker <bcrocker@redhat.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
(cherry picked from commit 3a9feb4db8)
2017-10-17 16:59:02 +01:00
Leo Liu
608bea62ca st/va: don't re-allocate interlaced buffer with pakced format
It caused corruption, when vlVaPutImage putting raw data to the fields

v2: add RGB formats since it got uploaded here as well

Cc: mesa-stable@lists.freedesktop.org
Cc: Andy Furniss <adf.lists@gmail.com>
Tested-by: Andy Furniss <adf.lists@gmail.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
(cherry picked from commit 0fa950ecd3)
2017-10-17 16:59:02 +01:00
Leo Liu
dc47b179ed st/vdpau: don't re-allocate interlaced buffer with packed YUV format
It caused corruption, when vlVdpVideoSurfacePutBitsYCbCr putting YUV to the fields

Cc: mesa-stable@lists.freedesktop.org
Cc: Andy Furniss <adf.lists@gmail.com>
Tested-by: Andy Furniss <adf.lists@gmail.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
(cherry picked from commit 327480d10f)
2017-10-17 16:59:02 +01:00
Dave Airlie
1146591d4b radv: emit fmuladd instead of fma to llvm.
For Vulkan SPIR-V the spec states
fma() Inherited from OpFMul followed by OpFAdd.

Matt says the backend will do the right thing depending on the
hardware being compiled for, if you use the fmuladd intrinsic.

Using the Mad Max pts test, on high settings at 4K:
CHP: 55->60
HGDD: 46->50
LM: 55->60
No change on Stronghold.

Thanks to Feral for spending the time to track this down.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 4e93d6baae)
[Emil Velikov: s/ac_to_float_type/to_float_type/]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/amd/common/ac_nir_to_llvm.c
2017-10-17 16:59:02 +01:00
Emil Velikov
76c6bbca7c cherry-ignore: add "anv: Remove unreachable cases from isl_format_for_size"
The commit causes a number of regressions with dEQP

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-10-17 16:57:45 +01:00
Lionel Landwerlin
3315bb4f08 intel: compiler: vec4: add missing default 0 lod
We set a similar default value for LOD in the fs backend for TXS/TXL.
Without this we end up generating invalid MOV with a null src.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: "17.2 17.1" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
(cherry picked from commit d3acc240d0)
2017-10-17 16:53:25 +01:00
Józef Kucia
cb72969d8b anv: Fix vkCmdFillBuffer()
The vkCmdFillBuffer() command fills a buffer with an uint32_t value.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: "17.1 17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 15fdbf9c39)
2017-10-11 17:44:36 +01:00
Marek Olšák
1093207463 st/mesa: don't use pipe_surface for passing information about EGLImage
Use st_egl_image instead. radeonsi doesn't like when we create
a pipe_surface with PIPE_FORMAT_NV12.

This fixes NV12 texturing on radeonsi using kmscube.

Cc: 17.1 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit 2d62817da9)
2017-10-11 17:44:36 +01:00
Bas Nieuwenhuizen
f5f515a023 nir/spirv: Allow loop breaks in a switch body.
Per the SPIR-V spec 2.11 Structured Control Flow:

"The only blocks in a construct that can branch outside the construct are

...
- a break block for the innermost loop it is inside of.
..."

With

"Break block: A block containing a branch to the Merge Block of a loop header's merge instruction."

Note that it puts no restriction on not being in an if or switch within the innermost loop.

This passes the loop_break block to the switch body so it can properly detect loop breaks.

CC: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
(cherry picked from commit ef61d09d5b)
2017-10-11 17:44:36 +01:00
Rob Clark
18db986354 freedreno/a5xx: fix missing restore state
RB_CLEAR_CNTL seems to be in a funny state after boot (at least on
8x96/a530).

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
(cherry picked from commit 7f3eab03fe)
2017-10-11 17:44:36 +01:00
Rob Clark
c386538036 freedreno/a5xx: align height to GMEM
Similar to the way width/pitch alignment works, it seems like we need to
do similar for height.  Otherwise the BLIT from system memory to GMEM
can over-fetch beyond the end of the buffer, triggering a fault.

I'm not sure if there is a better solution yet.  Possibly we could fall
back to pre-a5xx style DRAW packets for cases where BLIT might over-
fetch.  (We in theory have that problem already with rendering to higher
mipmap levels, although fortunately those tend to use GMEM bypass.)

This fixes issues reported with glamor.

Reported-by: don.harbin@linaro.org
Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>
(cherry picked from commit 16ac70bdcf)
2017-10-11 17:44:36 +01:00
Nicolai Hähnle
d0b52003d0 radeonsi: fix maximum advertised point size / line width
The hardware registers store the half-size/width in 12.4 fixed point
format, so 8192 is the maximum.

Fixes dEQP-GLES3.functional.rasterization.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 30e37289ea)
2017-10-11 17:44:36 +01:00
Nicolai Hähnle
6e903ae7d5 radeonsi: deduce rast_prim correctly for tessellation point mode
Together with the previous patches, this fixes
dEQP-GLES31.functional.primitive_bounding_box.wide_points.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit a3fa3b2e02)
2017-10-11 17:44:36 +01:00
Nicolai Hähnle
6e3788e3b1 radeonsi: don't discard points and lines
This is a bit conservative, but a more precise solution requires access
to the rasterizer state. This is something to tackle after the fork between
r600 and radeonsi.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 4d74432dd3)
2017-10-11 17:44:36 +01:00
Nicolai Hähnle
05e2ed7889 radeonsi: move current_rast_prim to r600_common_context
We'll use it in the scissors / clip / guardband state.

v2: avoid a performance regression on r600 when applied to
    (pre-fork) stable branches

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit f86a112b07)
2017-10-11 17:44:36 +01:00
Nicolai Hähnle
f2fc168517 glsl/lower_instruction: handle denorms and overflow in ldexp correctly
GLSL ES requires both, and while GLSL explicitly doesn't require correct
overflow handling, it does appear to require handling input inf/denorms
correctly.

Fixes dEQP-GLES31.functional.shaders.builtin_functions.precision.ldexp.*

Cc: mesa-stable@lists.freedesktop.org
Acked-by: Matt Turner <mattst88@gmail.com>
Acked-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
(cherry picked from commit 93bf9c114b)
[Emil Velikov: init_num_operands() does not exist on branch]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/compiler/glsl/lower_instructions.cpp
2017-10-11 17:44:06 +01:00
Nicolai Hähnle
eac6932c66 util/queue: fix a race condition in the fence code
A tempting alternative fix would be adding a lock/unlock pair in
util_queue_fence_is_signalled. However, that wouldn't actually
improve anything in the semantics of util_queue_fence_is_signalled,
while making that test much more heavy-weight. So this lock/unlock
pair in util_queue_fence_destroy for "flushing out" other threads
that may still be in util_queue_fence_signal looks like the better
fix.

v2: rephrase the comment

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Gustaw Smolarczyk <wielkiegie@gmail.com>
(cherry picked from commit a208cd7ae4)
2017-10-11 17:22:55 +01:00
Nicolai Hähnle
7bedb1fd12 radeonsi/gfx9: fix geometry shaders without output vertices
Not that those are super common or useful, but hey! Fun corner cases
of the API...

Fixes dEQP-GLES31.functional.geometry_shading.emit.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
(cherry picked from commit 7dfa891f32)
2017-10-11 17:22:55 +01:00
Nicolai Hähnle
332f6e9b3b amd/common: fix build_cube_select
Fix the custom cube coord selection sequence to be identical to
the hardware v_cubesc/tc and OpenGL spec. Affects texture sampling
with user-provided derivatives.

Fixes dEQP-GLES3.functional.shaders.texture_functions.texturegrad.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
(cherry picked from commit 5be5c1e0fa)
2017-10-11 17:22:55 +01:00
Nicolai Hähnle
f5cbe1f9b8 st/glsl_to_tgsi: fix conditional assignments to packed shader outputs
Overriding the default (no-op) swizzle is clearly counter-productive,
since the whole point is putting the destination register as one of
the source operands so that it remains unmodified when the assignment
condition is false.

Fragment depth and stencil outputs are a special case due to how their
source swizzles are manipulated in translate_src when compiling to
TGSI.

Fixes dEQP-GLES2.functional.shaders.conditionals.if.*_vertex
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>

(cherry picked from commit 8ea7d3a5c8)
2017-10-11 17:22:55 +01:00
Marek Olšák
f6d64c7dc1 mesa: fix texture updates for ATI_fragment_shader
Cc: 17.1 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
(cherry picked from commit 9d54025cd1)
2017-10-11 17:22:55 +01:00
Leo Liu
80db5b3420 st/va: use pipe transfer_map to map upload buffer
The function pipe_buffer_map() is only for linear pipe buffer,
with height as 0, and it's not for any 2D textures.

Signed-off-by: Leo Liu <leo.liu@amd.com>
Cc: mesa-stable@lists.freedesktop.org
Cc: Mark Thompson <sw@jkqxz.net>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 6ed61b8d3f)
2017-10-11 17:22:55 +01:00
Juan A. Suarez Romero
5a71ed6fa5 docs: add sha256 checksums for 17.2.2
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-10-02 18:10:02 +02:00
Juan A. Suarez Romero
bc12538a8e docs: add release notes for 17.2.2
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-10-02 17:26:10 +02:00
Juan A. Suarez Romero
98dce315c8 Update version to 17.2.2
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-10-02 17:15:13 +02:00
Bas Nieuwenhuizen
1535e8a5d4 radv: Check for GFX9 for 1D arrays in image_size intrinsic.
Only on GFX9 we implement them as 2D images.

This fixes:
dEQP-VK.image.image_size.1d_array.readonly_12x34
dEQP-VK.image.image_size.1d_array.readonly_1x1
dEQP-VK.image.image_size.1d_array.readonly_32x32
dEQP-VK.image.image_size.1d_array.readonly_7x1
dEQP-VK.image.image_size.1d_array.readonly_writeonly_12x34
dEQP-VK.image.image_size.1d_array.readonly_writeonly_1x1
dEQP-VK.image.image_size.1d_array.readonly_writeonly_32x32
dEQP-VK.image.image_size.1d_array.readonly_writeonly_7x1
dEQP-VK.image.image_size.1d_array.writeonly_12x34
dEQP-VK.image.image_size.1d_array.writeonly_1x1
dEQP-VK.image.image_size.1d_array.writeonly_32x32
dEQP-VK.image.image_size.1d_array.writeonly_7x1

Fixes: 1bcb953e16 "radv: handle GFX9 1D textures"
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 979978ee06)
2017-10-02 17:12:17 +02:00
Nicolai Hähnle
a66a70480f radeonsi: fix a regression in integer cube map handling
A recent commit fixed the case of 8888 integer cube maps, which need the
workaround of replacing the data format with USCALED/SSCALED. However,
this broke the case of non-8888 integer cube maps; those still need the
fix of shifting the texture coordinates.

Fixes KHR-GL45.texture_gather.plain-gather-int-cube-array and similar.

Fixes: 6fb0c1013b ("radeonsi: workaround for gather4 on integer cube maps")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 6d23f7c65d)
2017-10-02 17:12:17 +02:00
Nicolai Hähnle
bd522741c4 amd/common: move ac_build_phi from radeonsi
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 052b974fed)
[Juan A. Suarez: resolve trivial conflicts]
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>

Conflicts:
	src/amd/common/ac_llvm_build.c
	src/amd/common/ac_llvm_build.h
2017-10-02 17:12:17 +02:00
Jason Ekstrand
b78c664115 vulkan/wsi/wayland: Return better error messages
Reviewed-by: Daniel Stone <daniels@collabora.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 4fe3913b96)
2017-09-28 13:54:58 +00:00
Jason Ekstrand
2467b50e45 vulkan/wsi/wayland: Copy wl_proxy objects from oldSwapchain if available
This should save us some round trips while resizing.

Reviewed-by: Daniel Stone <daniels@collabora.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 537b9bc3e4)
2017-09-28 13:54:58 +00:00
Jason Ekstrand
b5a70210af vulkan/wsi/wayland: Stop caching Wayland displays
We originally implemented caching to avoid unneeded round-trips to the
compositor when querying surface capabilities etc. to set up the
swapchain.  Unfortunately, this doesn't work if vkDestroyInstance is
called after the Wayland connection has been dropped.  In this case, we
end up trying to clean up already destroyed wl_proxy objects which leads
to crashes.  In particular most of dEQP-VK.wsi.wayland is crashing
thanks to this problem.

This commit gets rid of the cache and simply embeds the wsi_wl_display
struct in the swapchain.  While we're at it, we can get rid of the
wl_event_queue that we were storing in the swapchain because we can just
use the one in the embedded wsi_wl_display.

Reviewed-by: Daniel Stone <daniels@collabora.com>
Bugzilla: https://bugs.freedesktop.org/102578
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 4369102498)
2017-09-28 13:54:58 +00:00
Jason Ekstrand
5edc4dc080 vulkan/wsi/wayland: Refactor wsi_wl_display code
We convert it over to an inti/finish model and make create/destroy
wrappers for the former.

Reviewed-by: Daniel Stone <daniels@collabora.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 77181d9580)
2017-09-28 13:54:57 +00:00
Ian Romanick
6587cbfca2 nv20: Fix GL_CLAMP
v2: Force T and R wrap modes to GL_CLAMP_TO_EDGE for 1D textures.
This fixes a regression in tex1d-2dborder.  The test uses a 1D texture
but it provides S and T texture coordinates.  Since the T wrap mode
would (correctly) be set to GL_CLAMP, the texture would gradually
blend (incorrectly) with the border color.

I also tried setting NV20_3D_TEX_FORMAT_DIMS_1D instead of
NV20_3D_TEX_FORMAT_DIMS_2D for 1D textures, but that did not help.

It is possible that the same problem exists for 2D textures with the
R-wrap mode, but I don't think there are any piglit tests for that.

No test changes on NV20 (10de:0201).

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
(cherry picked from commit 953a3cf0fd)
2017-09-28 13:54:57 +00:00
Tomasz Figa
6b0ded06b1 egl/dri2: Implement swapInterval fallback in a conformant way
dri2_fallback_swap_interval() currently used to stub out swap interval
support in Android backend does nothing besides returning EGL_FALSE.
This causes at least one known application (Android Snapchat) to fail
due to an unexpected error and my loose interpretation of the EGL 1.5
specification justifies it. Relevant quote below:

    The function

        EGLBoolean eglSwapInterval(EGLDisplay dpy, EGLint interval);

    specifies the minimum number of video frame periods per buffer swap
    for the draw surface of the current context, for the current rendering
    API. [...]

    The parameter interval specifies the minimum number of video frames
    that are displayed before a buffer swap will occur. The interval
    specified by the function applies to the draw surface bound to the
    context that is current on the calling thread. [...] interval is
    silently clamped to minimum and maximum implementation dependent
    values before being stored; these values are defined by EGLConfig
    attributes EGL_MIN_SWAP_INTERVAL and EGL_MAX_SWAP_INTERVAL
    respectively.

    The default swap interval is 1.

Even though it does not specify the exact behavior if the platform does
not support changing the swap interval, the default assumed state is the
swap interval of 1, which I interpret as a value that eglSwapInterval()
should succeed if called with, even if there is no ability to change the
interval (but there is no change requested). Moreover, since the
behavior is defined to clamp the requested value to minimum and maximum
and at least the default value of 1 must be present in the range, the
implementation might be expected to have a valid range, which in case of
the feature being unsupported, would correspond to {1} and any request
might be expected to be clamped to this value.

This is further confirmed by the code in _eglSwapInterval() in
src/egl/main/eglsurface.c, which is the default fallback implementation
for EGL drivers not implementing its own. The problem with it is that
the DRI2 EGL driver provides its own implementation that calls into
platform backends, so we cannot just simply fall back to it.

Signed-off-by: Tomasz Figa <tfiga@chromium.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
2017-09-28 13:54:57 +00:00
Boris Brezillon
d9aba007c0 broadcom/vc4: Fix infinite retry in vc4_bo_alloc()
cleared_and_retried is always reset to false when jumping to the retry
label, thus leading to an infinite retry loop.

Fix that by moving the cleared_and_retried variable definitions at the
beginning of the function.  While we're at it, move the create variable
with the other local variables and explicitly reset its content in the
retry path.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
Fixes: 78087676c9 "vc4: Restructure the simulator mode."
(cherry picked from commit ef578906d8)
2017-09-28 13:54:57 +00:00
Eric Anholt
e22ab89f9f broadcom/vc4: Keep pipe_sampler_view->texture matching the original texture.
I was overwriting view->texture with the shadow resource when we need to
do shadow copies (retiling or baselevel rebase), but that tripped up some
critical new sanity checking in state_tracker (making sure that stObj->pt
hasn't changed from view->texture through TexImage-related paths).

To avoid that, move the shadow resource to the vc4_sampler_view struct.

Fixes: f0ecd36ef8 ("st/mesa: add an entirely separate codepath for setting up buffer views")
(cherry picked from commit 68c91a87d7)
2017-09-28 13:54:57 +00:00
Jason Ekstrand
1b100aae7b vulkan/wsi/wayland: Stop printing out the DRM device
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 016de7e155)
2017-09-28 13:54:57 +00:00
Kenneth Graunke
ec89ab30c9 i965/vec4: Fix swizzles on atomic sources.
Atomic operation sources are scalar values, but we were failing to
select the .x component of the second operand.  For example,

   atomicCounterCompSwapARB(counter, 5u, 10u)

would generate

   mov(8) vgrf4.x:D, 5D
   mov(8) vgrf5.x:D, 10D

   mov(8) vgrf9.x:UD, vgrf4.xyzw:D
   mov(8) vgrf9.y:UD, vgrf5.xyzw:D

which wrongly selects the .y component of vgrf5, so the actual 10u value
would get dead code eliminated.  The swizzle works for the other source,
but both of them ought to be .xxxx.

Fixes the compare and swap CTS tests in:
KHR-GL45.shader_atomic_counter_ops_tests.ShaderAtomicCounterOpsExchangeTestCase

Cc: "17.2 17.1 17.0 13.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
(cherry picked from commit 66342c997f)
2017-09-28 13:54:57 +00:00
Kenneth Graunke
8619a6b1cc i965/vec4: Actually handle atomic op intrinsics.
Embarassingly, someone enabled the ARB_shader_atomic_counter_ops
extension for Gen7+ but never added the intrinsics to the switch
statement in the vec4 backend, so they just hit an unreachable()
call and died.

Fixes: 40dd45d0c6 (i965: Enable ARB_shader_atomic_counter_ops)
Cc: "17.2 17.1 17.0 13.0" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
(cherry picked from commit a62fe34098)
2017-09-28 13:54:57 +00:00
Samuel Pitoiset
47c8ffe719 radv: fix saved compute state when doing statistics/occlusion queries
We are pushing 16-bytes of constants, so we have to save/restore
the same amount of data to avoid data corruption.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 4b407a62c7)
2017-09-28 13:54:57 +00:00
Grazvydas Ignotas
84f3374f1c configure: check if -latomic is needed for __atomic_*
On some platforms, gcc generates library calls when __atomic_* functions
are used, but does not link the required library (libatomic) automatically
(supposedly to allow the app to use some other atomics implementation?).

Detect this at configure time and add the library when needed. Tested
on armel (library was added) and on x86_64 (was not, as expected).

Some documentation on this is provided in GCC wiki:
https://gcc.gnu.org/wiki/Atomic/GCCMM

Fixes: 8915f0c0 "util: use GCC atomic intrinsics with explicit memory model"
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102573
Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit 2ef7f23820)
2017-09-28 13:54:57 +00:00
Nicolai Hähnle
5e5882a7aa amd/addrlib: fix missing va_end() after va_copy()
There's no reason to use va_copy here.

CID: 1418113
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
Fixes: e7fc664b91 ("winsys/amdgpu: add addrlib - texture
                              addressing and alignment calculator")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 34126ed248)
2017-09-28 13:54:57 +00:00
Juan A. Suarez Romero
89db1532fc cherry-ignore: add "radv: copy the number of viewports/scissors at pipeline bind time"
fixes: This commit addressed earlier commits dcf46e99 and 60878dd0 which
did not land in branch.

Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-09-28 13:54:57 +00:00
Nicolai Hähnle
5e8ddba33f radeonsi: set MIP_POINT_PRECLAMP to 0
This fixes a bug with nearest ("point") mip selection when the fractional
part of max_lod is in (0.5,1). In this case, the spec mandates that
we still select the mip level ceil(max_lod) in the clamping case. However,
MIP_POINT_PRECLAMP will clamp before the mip selection, which is wrong.

Supposedly this setting was originally copied from the closed Vulkan
driver, but as far as I can tell, closed Vulkan was actually changed back
recently :)

Fixes dEQP-GLES3.functional.texture.mipmap.2d.max_lod.{nearest,linear}_nearest

Fixes: f7420ef5b4 ("radeonsi: enable some sampler fields to match the closed driver")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 704ddbcdf6)
2017-09-28 13:54:57 +00:00
Eric Anholt
fc2b01adc9 broadcom/vc4: Fix use-after-free when deleting a program.
By leaving the compiled shader in the context's stage state, the next
compile of a new FS would look in the old compiled FS for figuring out
whether to set various dirty flags for the VS compile.  Clear out the
pointer when deleting the program, and make sure that we always mark the
state as dirty if the previous program had been lost.  Fixes valgrind
warnings on glsl-max-varyings.

Fixes: 2350569a78 ("vc4: Avoid VS shader recompiles by keeping a set of FS inputs seen so far.")
(cherry picked from commit 3752ad28f2)
2017-09-28 13:54:57 +00:00
Eric Anholt
fe53c8bc21 broadcom/vc4: Fix use-after-free trying to mix a quad and tile clear.
The blitter will bind just the depth buffer, which flushes the current job
if we had both a color and depth/stencil.  If the clear was doing partial
depth/stencil (quad-based) and color (tile-based), we'd go on to try to
set up the rest of the tile clear in the now flushed job.

Instead, move the partial clear up before we start setting up the job for
the current FBO state, and re-fetch the job if we're continuing on to a
tile-based clear.  Fixes valgrind failures in fbo-depthtex.

Fixes: 9421a6065c ("vc4: Fix fallback to quad clears of depth in GLX.")
(cherry picked from commit 9940fb4205)
2017-09-28 13:54:57 +00:00
Eric Anholt
60efffcec0 broadcom/vc4: Fix use-after-free for flushing when writing to a texture.
I was trying to continue the hash table loop, not the inner loop.  This
tended to work out, because we would have *just* freed the job struct.
Fixes some valgrind failures in fbo-depthtex.

Fixes: f597ac3966 ("vc4: Implement job shuffling")
(cherry picked from commit d88a75182d)
2017-09-28 13:54:56 +00:00
Dave Airlie
5d1deeeab3 st/glsl->tgsi: fix u64 to bool comparisons.
Otherwise we end up using a 32-bit comparison which didn't end well.

Timothy caught this while playing around with some opt passes.

Fixes: 278580729a (st/glsl_to_tgsi: add support for 64-bit integers)
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit a7a7bf21bd)
2017-09-28 13:54:56 +00:00
Juan A. Suarez Romero
dbccd8465d cherry-ignore: add "radv: Check for GFX9 for 1D arrays in image_size intrinsic."
fixes: This commit addressed earlier commits 61ad2f13 and 6dcc54b4 which
did not land in branch

Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-09-28 13:54:56 +00:00
Juan A. Suarez Romero
de44476572 cherry-ignore: add "radeonsi/gfx9: proper workaround for LS/HS VGPR initialization bug"
fixes: This commit addressed an earlier commit c7e9ebb3ab which did not
land in branch

Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-09-28 13:54:56 +00:00
Leo Liu
2712e59538 st/va/postproc: use video original size for postprocessing
Otherwise the aligned size will make video scaled

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Christian König <christian.koenig@amd.com>
(cherry picked from commit eb51838771)
2017-09-28 13:54:56 +00:00
Samuel Iglesias Gonsálvez
e16c31f3fe anv: fix viewport transformation for z component
In Vulkan, for 'z' (depth) component, the scale and translate values
for the viewport transformation are:

pz = maxDepth - minDepth
oz = minDepth

zf = pz × zd + oz

Being zd, the third component in vertex's normalized device coordinates.

Fixes: dEQP-VK.draw.inverted_depth_ranges.*

Signed-off-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit d2cd9deeb8)
2017-09-28 13:54:56 +00:00
David Airlie
dbf33a12e1 radv: add gfx9 scissor workaround
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: 17.2 <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 3e54493265)
2017-09-28 13:54:56 +00:00
Lucas Stach
cd9d26c3ef etnaviv: fix 16bpp clears
util_pack_color may leave undefined values in the upper half of the packed
integer. As our hardware needs the upper 16 bits to mirror the lower 16bits,
this breaks clears of those formats if the undefined values aren't masked off.

I've only observed the issue with R5G6B5_UNORM surfaces, other 16bpp
formats seem to work fine.

Fixes: d6aa2ba2b2 (etnaviv: replace translate_clear_color with util_pack_color)
Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Wladimir J. van der Laan <laanwj@gmail.com>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
(cherry picked from commit e9d37d68cf)
2017-09-28 13:54:56 +00:00
Tim Rowley
b378fd5cf3 swr/rast: remove llvm fence/atomics from generated files
We currently don't use these instructions, and since their API
changed in llvm-5.0 having them in the autogen files broke the mesa
release tarballs which ship with generated autogen files.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102847
CC: mesa-stable@lists.freedesktop.org
Tested-by: Laurent Carlier <lordheavym@gmail.com>
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
(cherry picked from commit 066d1dc951)
2017-09-28 13:54:56 +00:00
Tapani Pälli
caff607bcb mesa: free current ComputeProgram state in _mesa_free_context_data
This is already done for other programs stages, fixes a leak when using
compute programs.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Cc: mesa-stable@lists.freedesktop.org
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102844
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 589457d97f)
2017-09-28 13:54:56 +00:00
Nicolai Hähnle
737e2245ce radeonsi: fix array textures layer coordinate
Like for cube map (array) gather, we need to round to nearest on <= VI.

Fixes tests in dEQP-GLES3.functional.shaders.texture_functions.texture.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 87f7c7bd65)
2017-09-28 13:54:56 +00:00
Nicolai Hähnle
46cfefc5be glsl/linker: fix output variable overlap check
Prevent an overflow caused by too many output variables. To limit the
scope of the issue, write to the assigned array only for the non-ES
fragment shader path, which is the only place where it's needed.

Since the function will bail with an error when output variables with
overlapping components are found, (max # of FS outputs) * 4 is an upper
limit to the space we need.

Found by address sanitizer.

Fixes dEQP-GLES3.functional.attribute_location.bind_aliasing.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
(cherry picked from commit 15cae12804)

Squashed with commit:

glsl/linker: properly fix output variable overlap check

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102904
Fixes: 15cae12804 ("glsl/linker: fix output variable overlap check")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
(cherry picked from commit df8767a14e)

Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-09-28 13:54:56 +00:00
Józef Kucia
6121bc3150 anv: Fix descriptors copying
Trivial.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
(cherry picked from commit 65a09f98ad)
2017-09-28 13:54:56 +00:00
Dave Airlie
488460b1ef ac/surface: handle S8 on gfx9
If we don't have a depth piece, we don't get a correct
swizzle mode and we hit an assert in addrlib.

In case of no depth get the preferrred swizzle mode for
stencil alone.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit c4ac522511)

Squashed with commit:

ac/surface: handle error when choosing preferred swizzle mode

CID: 1418140
Fixes: c4ac522511 ("ac/surface: handle S8 on gfx9")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit eb71394ff3)

Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-09-28 13:54:56 +00:00
Alexandru-Liviu Prodea
3c54a77d08 Scons: Add LLVM 5.0 support
1 new required library - LLVMBinaryFormat

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Bugzilla: https://bugs.freedesktop.org=/show_bug.cgi?id=3D102318
Signed-off-by: Alexandru-Liviu Prodea <liviuprodea@yahoo.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit c1b0137048)
2017-09-28 13:54:56 +00:00
Nicolai Hähnle
891627ec0f amd/common: add workaround for cube map array layer clamping
Fixes dEQP-GLES31.functional.texture.filtering.cube_array.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 94736d31c3)

Squashed with commit:

amd/common: add chip_class to ac_llvm_context

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 3db86d86ed)

Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
2017-09-28 13:54:56 +00:00
Nicolai Hähnle
688f8415f7 amd/common: round cube array slice in ac_prepare_cube_coords
The NIR-to-LLVM pass already does this; now the same fix covers
radeonsi as well.

Fixes various tests of
dEQP-GLES31.functional.texture.filtering.cube_array.combinations.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit e0af3bed2c)
[Juan A. Suarez: resolve trivial conflicts]
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>

Conflicts:
	src/amd/common/ac_nir_to_llvm.c
2017-09-28 13:54:55 +00:00
Nicolai Hähnle
54915ee52f radeonsi: workaround for gather4 on integer cube maps
This is the same workaround that radv already applied in commit
3ece76f03d ("radv/ac: gather4 cube workaround integer").

Fixes dEQP-GLES31.functional.texture.gather.basic.cube.rgba8i/ui.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 6fb0c1013b)
2017-09-28 13:54:55 +00:00
Jason Ekstrand
da517431f6 i965/blorp: Set r8stencil_needs_update when writing stencil
This fixes a crash on Haswell when we try to upload a stencil texture
with blorp.  It would also be a problem if someone tried to texture from
stencil after glBlitFramebuffers.

Cc: "17.2 17.1" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
(cherry picked from commit a43d379000)
2017-09-28 13:54:55 +00:00
Matt Turner
3b4a6dbe93 util/u_atomic: Add implementation of __sync_val_compare_and_swap_8
Needed for 32-bit PowerPC.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Fixes: a6a38a038b ("util/u_atomic: provide 64bit atomics where
they're missing")
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

(cherry picked from commit 1bbe180873)
2017-09-28 13:54:55 +00:00
Matt Turner
d2dca92a6c util: Link libmesautil into u_atomic_test
Platforms without particular atomic operations require the
implementations in u_atomic.c

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Fixes: a6a38a038b ("util/u_atomic: provide 64bit atomics where
they're missing")
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit d075a4089e)
[Juan A. Suarez: resolve trivial conflicts]
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>

Conflicts:
	src/util/Makefile.am
2017-09-28 13:54:55 +00:00
Emil Velikov
d41d54b061 automake: enable libunwind in `make distcheck'
Enable the toggle to catch when the library is missing from the link
path. Better to test, fail and address before releasing Mesa ;-)

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
(cherry picked from commit 9aba643e3c)
2017-09-28 13:54:55 +00:00
Gert Wollny
5d16640ea1 travis: Add libunwind-dev to gallium/make builds
libunwind is a optional dependency used by the gallium aux module
(libgallium) and consequently the final binaries must be linked against
it. To test whether the library is properly specified in the link pass
add it to the travis-ci build environment and force its use.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit 39fe51c1e3)
2017-09-26 12:17:33 +02:00
Gert Wollny
6024a95198 travis: force llvm-3.3 for "make Gallium ST Other"
In Ubuntu Trusty the default version of llvm is 3.4 and the build was
actually randomly picking 3.5 or 3.9. Adding libunwind would then result
is build success or failure depending of what version was picked.

Install the llvm-3.3-dev package and force its use: On one hand it is
the minimum required version we want to the build test against, and on
the other hand forcing the version stabilizes the build.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit d3675812b5)
2017-09-26 12:17:33 +02:00
Dave Airlie
996bf9c1cd radv/nir: call opt_remove_phis after trivial continues.
With the shaders in the ssao demo, the nir_opt_if wasn't
working properly without this, after this the if gets optimised
so that loop unrolling gets called.

(loop unrolling fails due to instruction count, but at least
it gets to do that.)

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 64d9bd149a)
[Juan A. Suarez: apply patch over src/amd/vulkan/radv_pipeline.c]
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>

Conflicts:
	src/amd/vulkan/radv_shader.c
2017-09-26 12:17:33 +02:00
Emil Velikov
bd903d4ee1 docs: add sha256 checksums for 17.2.1
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-18 00:12:36 +01:00
Emil Velikov
d6d2b6b5ec docs: add release notes for 17.2.1
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-17 23:57:32 +01:00
Emil Velikov
cb778d563f Update version to 17.2.1
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-17 23:53:08 +01:00
Bas Nieuwenhuizen
7c3bd519e7 radv: Don't allocate CMASK for linear images.
We can't use it anyway in fast clears, and on GFX9 it seems to
actually hange the card if we specify it.

Fixes: f4e499ec79 "radv: add initial non-conformant radv vulkan driver"
(cherry picked from commit 1a172fb113)
2017-09-13 21:52:42 +01:00
Bas Nieuwenhuizen
c6b3732967 radv: Disable multilayer & multilevel DCC.
The current DCC init routine doesn't account for initializing a
single layer or level. Multilayer seems hard for small textures on
pre-GFX9 as tre metadata for the layers can be interleaved. For
GFX9 multilevel textures are a problem for similar reasons.

So just disable this for now, until we handle the texture modes
correctly.

Fixes: f4e499ec79 "radv: add initial non-conformant radv vulkan driver"
(cherry picked from commit bee83b2661)
2017-09-13 21:52:42 +01:00
Eric Engestrom
e012ade1cf docs/egl: remove reference to EGL_DRIVERS_PATH
Support for external egl drivers was dropped a few years ago.

Fixes: 209360bbb9 "egl/main: drop support for external egl drivers"
Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit 85b66d2096)
2017-09-13 21:52:42 +01:00
Dave Airlie
41e691d605 radv/winsys: fix flags vs va_flags thinko.
Fixes: e8d57802f (radv/gfx9: allocate events from uncached VA space)
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit a5add6fb30)
2017-09-13 21:52:42 +01:00
Emil Velikov
b6ae4400fc egl/x11/dri3: adding missing __DRI_BACKGROUND_CALLABLE extension
Fixes: 3b7b6adf3a ("egl: Implement __DRI_BACKGROUND_CALLABLE")
Cc: Timothy Arceri <tarceri@itsqueeze.com>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit f24bc18162)
2017-09-13 21:52:42 +01:00
Bas Nieuwenhuizen
c47914276e radv: Fix vkCopyImage with both depth and stencil aspects.
Fixes: f4e499ec79 "radv: add initial non-conformant radv vulkan driver"
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit ff23e03d60)
2017-09-13 21:52:42 +01:00
Bas Nieuwenhuizen
536f852d42 radv: Actually set the cmd_buffer usage_flags.
Otherwise, the simultaneous uage bit doesn't get set from the begin
info, which we need for batchchaining.

Fixes: f4e499ec79 "radv: add initial non-conformant radv vulkan driver"
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit dec7b38fe6)
2017-09-13 21:52:42 +01:00
Grazvydas Ignotas
fdc4f6e684 radv: don't assert on empty hash table
Currently if table_size is 0, it's falling through to:

unreachable("hash table should never be full");

But table_size can be 0 when RADV_DEBUG=nocache is set, or when the
table allocation fails (which is not considered an error).

Fixes: f4e499ec79 "radv: add initial non-conformant radv vulkan driver"
Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
(cherry picked from commit b8dd69e1b4)
2017-09-13 21:52:42 +01:00
Eric Engestrom
8cf9e5ab56 mesa/st: remove unwanted backup file
Fixes: 0ac78dc925 "util: move string_to_uint_map to glsl"
Cc: Emil Velikov <emil.velikov@collabora.com>
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit ac0d8dc3fa)
2017-09-13 21:52:42 +01:00
Samuel Pitoiset
d5c71e44c4 radeonsi: update dirty_level_mask before dispatching
This fixes a rendering issue with Hitman when bindless textures
are enabled.

Fixes: 2263610827 ("radeonsi: flush DB caches only when transitioning from DB to texturing")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 59101e771d)
2017-09-13 21:52:42 +01:00
Nicolai Hähnle
53190187cc glsl: fix glsl_struct_field size calculations for shader cache
Found by address sanitizer:

==22621==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x61400000cbd8 at pc 0x7f561610a4ff bp 0x7ffca85f9d50 sp 0x7ffca85f94f8
READ of size 344 at 0x61400000cbd8 thread T0
    #0 0x7f561610a4fe  (/usr/lib/x86_64-linux-gnu/libasan.so.3+0x5f4fe)
    #1 0x7f560bb305a5 in memcpy /usr/include/x86_64-linux-gnu/bits/string3.h:53
    #2 0x7f560bb305a5 in blob_write_bytes ../../../mesa-src/src/compiler/glsl/blob.c:136
    #3 0x7f560be7d7ff in encode_type_to_blob ../../../mesa-src/src/compiler/glsl/shader_cache.cpp:153
    #4 0x7f560be81222 in write_program_resource_data ../../../mesa-src/src/compiler/glsl/shader_cache.cpp:950
    #5 0x7f560be81222 in write_program_resource_list ../../../mesa-src/src/compiler/glsl/shader_cache.cpp:1118
    #6 0x7f560be81222 in shader_cache_write_program_metadata(gl_context*, gl_shader_program*) ../../../mesa-src/src/compiler/glsl/shader_cache.cpp:1407
    #7 0x7f560b825fdb in link_program ../../../mesa-src/src/mesa/main/shaderapi.c:1163

Fixes: 073a84ff60 ("glsl: stop adding pointers from glsl_struct_field to the cache")
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
(cherry picked from commit 4da6cf6c98)
2017-09-13 21:52:42 +01:00
Emil Velikov
366e02f992 cherry-ignore: add EGL+gbm swast patches
They depend on an earlier commit which did not land in branch.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-13 21:52:41 +01:00
Nicolai Hähnle
5cec093773 ac/surface: match Z and stencil tile config
Fixes various piglit tests on Stoney, see the comment.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit cffc0ae0d9)
[Emil Velikov: attribute for the different gfx6_surface_settings()]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/amd/common/ac_surface.c
2017-09-13 21:52:02 +01:00
Nicolai Hähnle
69394e3517 radeonsi: apply a mask to gl_SampleMaskIn in the PS prolog
gl_SampleMaskIn is supposed to contain set bits only for the samples that
are covered by the current fragment shader invocation, but the VGPR
initialization hardware loads the set of all bits that are covered at the
current pixel.

Fixes various tests in
dEQP-GLES31.functional.shaders.sample_variables.sample_mask_in.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 92c4277990)
[Emil Velikov: attribute for the lack of add_arg*checked() API]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/gallium/drivers/radeonsi/si_shader.c
2017-09-13 21:52:02 +01:00
Timothy Arceri
c0c1e05fb6 glsl: stop adding pointers from bindless structs to the cache
This is so we always create reproducible cache entries. Consistency
is required for verification of any third party distributed shaders.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 4009370232)
2017-09-13 21:52:02 +01:00
Timothy Arceri
94b656cb5e glsl: stop adding pointers from shader_info to the cache
This is so we always create reproducible cache entries. Consistency
is required for verification of any third party distributed shaders.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit a6618afd27)
2017-09-13 21:52:02 +01:00
Timothy Arceri
5e0b556abc compiler: move pointers to the start of shader_info
This will allow us to easily skip them when writting the struct
to disk cache.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 3ea3f75723)
2017-09-13 21:52:02 +01:00
Timothy Arceri
f087dad40d glsl: always write a name/label string to the cache
In the following patch we will stop writing the pointer to cache.

Unfortunately adding empty strings to that cache seems to be the
only thing we can do here once we no longer have the pointers.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 44918a1979)
2017-09-13 21:52:02 +01:00
Timothy Arceri
83bbfaddf1 glsl: don't write uniform storage offset if there isn't one
This is so we always create reproducible cache entries. Consistency
is required for verification of any third party distributed shaders.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 22154823d2)
2017-09-13 21:52:02 +01:00
Timothy Arceri
a8eea040cd glsl: add has_uniform_storage() helper to shader cache
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 2662269ad7)
2017-09-13 21:52:02 +01:00
Timothy Arceri
15ecbb2c07 glsl: stop adding pointers from glsl_struct_field to the cache
This is so we always create reproducible cache entries. Consistency
is required for verification of any third party distributed shaders.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 073a84ff60)
2017-09-13 21:52:02 +01:00
Timothy Arceri
7fef8d436a glsl: stop adding pointers from gl_shader_variable to the cache
This is so we always create reproducible cache entries. Consistency
is required for verification of any third party distributed shaders.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 37d453b55a)
2017-09-13 21:52:02 +01:00
Timothy Arceri
ea1bcfc063 glsl: allow NULL to be passed to encode_type_to_blob()
This will be used by the following commit.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 37eb67714e)
2017-09-13 21:52:01 +01:00
Eric Engestrom
6e2b9fba21 util: improve compiler guard
Glibc 2.26 has dropped xlocale.h, but the functions needed (strtod_l()
and strdof_l()) can be found in stdlib.h.
Improve the detection method to allow newer builds to still make use of
the locale-setting.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102454
Cc: Laurent Carlier <lordheavym@gmail.com>
Cc: Emil Velikov <emil.l.velikov@gmail.com>
Cc: Rob Herring <robh@kernel.org>
Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Laurent Carlier <lordheavym@gmail.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
(cherry picked from commit 49b428470e)
2017-09-13 21:52:01 +01:00
Kenneth Graunke
dcb634df92 i965: Set "Subslice Hashing Mode" to 16x16 on Apollolake.
As of 4.11, the kernel isn't bothering to set the subslice hashing mode
on Apollolake, leaving it at the default of 8x8.  (It initializes it to
16x4 on most platforms.)

Performance data for GPUTest Triangle on Apollolake at 1024x640:

   X-tiled RT:
   -----------
   8x8 -> 16x4:   2.4325%  +/- 0.383683% (n=107)
   8x8 -> 8x4:   -3.75105% +/- 0.592491% (n=40)
   8x8 -> 16x16:  6.17238% +/- 0.67157%  (n=30)

   Y-tiled RT:
   -----------
   8x8 -> 16x4:   1.30307%  +/- 0.297292% (n=205)
   8x8 -> 8x4:   -0.769282% +/- 0.729557% (n=35)
   8x8 -> 16x16:  3.00254%  +/- 0.715503% (n=40)

   8x MSAA RT (INTEL_FORCE_MSAA=8):
   --------------------------------
   8x8 -> 16x4:   1.38889% +/- 0.93729%  (n=7)
   8x8 -> 8x4:   -2.10643% +/- 1.15153%  (n=3)
   8x8 -> 16x16:  3.87183% +/- 1.08851%  (n=5)

Based on this, we choose 16x16 for Apollolake.

Skylake GT2 with X-tiled buffers appears to be a toss-up between 16x4
and 16x16, and with Y-tiled buffers it doesn't seem to really matter.
So we'll leave Skylake alone for now.

The hashing mode doesn't seem to make a measurable impact on more
complex benchmarks.

Acked-by: Matt Turner <mattst88@gmail.com>
(cherry picked from commit ebd2fd6ef3)
2017-09-13 21:52:01 +01:00
Dave Airlie
20be71ba7c radv/gfx9: fix image resource handling.
GFX9 changes how images are layed out, so this needs updating.

Fixes: dEQP-VK.query_pool.statistics_query.*

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 3633bae36b)
2017-09-13 21:52:01 +01:00
Dave Airlie
a3d1ea347a radv/ac: bump params array for image atomic comp swap
For the comp_swap case this was overflowing and crashing
sometimes.

Fixes:
dEQP-VK.image.atomic_operations.compare_exchange.*

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit aba441be44)
2017-09-13 21:52:01 +01:00
Dave Airlie
dc640aab63 radv/gfx9: set mip0-depth correctly for 2d arrays/3d images
This field covers the whole resource.

Fixes:
dEQP-VK.pipeline.image.suballocation.sampling_type.combined.view_type.3d.format.*
dEQP-VK.texture.filtering.3d.combinations.*

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit ebd2a5354d)
2017-09-13 21:52:01 +01:00
Dave Airlie
c8076c8ea1 radv: handle GFX9 1D textures
As GFX9 can't handle 1D depth textures, radeonsi and
apparantly pro just update all 1D textures to 2D,
and work around it.

This ports the workarounds from radeonsi.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 1bcb953e16)

Conflicts:
	src/amd/common/ac_nir_to_llvm.c
2017-09-13 21:52:01 +01:00
Dave Airlie
62f0bb2a87 radv: don't use iview for meta image width/height.
Work out the width/height from the level manually, as on GFX9
we won't minify the iview width/height.

This fixes:
dEQP-VK.api.image_clearing.core.clear_color_image* on gfx9

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 2f5b4490b5)
2017-09-13 21:52:01 +01:00
Emil Velikov
40f06286f9 cherry-ignore: add execution_type() fix to the list
As mentioned by Matt:
 "Without commit 4fab67a441, this patch isn't  needed at all."

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-13 21:50:49 +01:00
Nicolai Hähnle
0f79eb7abe st/glsl_to_tgsi: only the first (inner-most) array reference can be a 2D index
Don't get distracted by record dereferences between array references.

Fixes dEQP-GLES31.functional.tessellation.user_defined_io.per_vertex_block.*

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 03203b7448)
2017-09-13 21:34:03 +01:00
Dave Airlie
54c5568fa9 radv: use simpler indirect packet 3 if possible.
This fixes some observed hangs on CIK GPUs.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 219d29e4d8)
[Emil Velikov: add the hunk in radv_emit_indirect_draw]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/amd/vulkan/radv_cmd_buffer.c
2017-09-13 21:34:03 +01:00
Dave Airlie
c2f314d721 radv/gfx9: allocate events from uncached VA space
This copies what amdgpu-pro does, and allocates the memory
for an event with an uncached mtype.

This fixes hangs with:
dEQP-VK.api.command_buffers.record_simul_use_primary

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit e8d57802fe)
2017-09-13 21:34:02 +01:00
Dave Airlie
fe04abc6e9 radv/winsys: use amdgpu_bo_va_op_raw.
This is a precursor to the gfx9 fix to use uncached for the event
memory. Move to the interface which allows setting the flags,
but wrap it to avoid having to copy it around the place.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 76ac8fafad)
2017-09-13 21:34:02 +01:00
Marek Olšák
0e305f0518 st/mesa: skip draw calls with pipe_draw_info::count == 0
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102502

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Tested-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
(cherry picked from commit e4018fdd85)
2017-09-13 21:34:02 +01:00
Nicolai Hähnle
cb9dae484a radeonsi/gfx9: always flush DB metadata on framebuffer changes
This fixes GL45-CTS.shader_image_load_store.basic-glsl-earlyFragTests.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 34124e412f)
[Emil Velikov: resolve trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/gallium/drivers/radeonsi/si_pipe.h
	src/gallium/drivers/radeonsi/si_state.c
2017-09-13 21:34:02 +01:00
Dave Airlie
57ecf28668 Revert "radv: disable support for VEGA for now."
This reverts commit 611076a41a.

With the two previous commits, vega shouldn't be unstable,
doesn't pass CTS, but can do a complete run, and games shouldn't
hang anymore, so bring it back online.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit e38685cc62)
2017-09-13 21:34:02 +01:00
Dave Airlie
78e2b539a4 radv/gfx9: set descriptor up for base_mip to level range.
This is required on GFX9, fixes a bug in Talos where all the
mipmaps overlay each other.

Just pushing this as well as it fixes Talos.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 6d929d3f85)
2017-09-13 21:34:02 +01:00
Dave Airlie
86d4c203bd radv: disable 1d/2d linear optimisation on gfx9.
This causes hangs in some of the CTS tests with a 2d
1536x2 texture.

This fixes hangs with:
dEQP-VK.pipeline.image.suballocation.sampling_type.combined.iew_type.1d_aray.format.r4g4b4a4_unorm_pack16.count_1.size.512x1_array_of_3
if we reenable it, make sure these don't regress.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit d118ff8765)
2017-09-13 21:34:02 +01:00
Jason Ekstrand
bda2975eef spirv: Add support for the HelperInvocation builtin
I have no idea how this got missed but it's been missing since forever.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
(cherry picked from commit e439908af9)
2017-09-13 21:34:02 +01:00
Roland Scheidegger
a4dc18efca st/mesa: fix view template initialization in try_pbo_readpixels
I think this is what the code was meant to do, albeit as far as I can tell
the redundant initialization some analyzers complain about should work as
well just fine (only the first layer will be used, if the view contains one
or more layers doesn't really matter).

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102467
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 2b2c61f0df)
2017-09-13 21:34:02 +01:00
Kenneth Graunke
8f77409e2b i965: Fix crash in fallback GTT mapping.
We can't perf_debug without a context.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
(cherry picked from commit 52b65dfda8)
2017-09-13 21:34:02 +01:00
Rob Clark
458f52618a freedreno: skip batch-cache for compute shaders
It is kind of pointless for compute, and avoids issues with apps kicking
off more than 32 compute shaders at once.

Signed-off-by: Rob Clark <robdclark@gmail.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit dc9e08b0c3)
2017-09-13 21:34:02 +01:00
Karol Herbst
402b8073ad nvc0: write 0 to pipeline_statistics.cs_invocations
cs_invocations are currently unsupported, but leaving the field uninitialized
is even worse.

fixes on nvc0:
 * KHR-GL45.pipeline_statistics_query_tests_ARB.functional_default_qo_values
 * KHR-GL45.pipeline_statistics_query_tests_ARB.functional_non_rendering_commands_do_not_affect_queries

Signed-off-by: Karol Herbst <karolherbst@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit b672c3833b)
2017-09-13 21:34:02 +01:00
Ben Crocker
6fe4852f59 llvmpipe: lp_build_gather_elem_vec BE fix for 3x16 load
Fix loading of a 3x16 vector as a single 48-bit load
on big-endian systems (PPC64, S390).

Roland Scheidegger's commit e827d91756
plus Ray Strode's patch reduce pre-Roland Piglit failures from ~4000 to ~2000.  This patch fixes
three of the four regressions observed by Ray:

- draw-vertices
- draw-vertices-half-float
- draw-vertices-half-float_gles2

One regression remains:
- draw-vertices-2101010

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100613
Cc: "17.2" "17.1" <mesa-stable@lists.freedesktop.org>

Signed-off-by: Ben Crocker <bcrocker@redhat.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
(cherry picked from commit 57c8ead0cd)
2017-09-13 21:34:02 +01:00
Ray Strode
d080f10fdf gallivm: correct channel shift logic on big endian
lp_build_fetch_rgba_soa fetches a texel from a texture.
Part of that process involves first gathering the element
together from memory into a packed format, and then breaking
out the individual color channels into separate, parallel
arrays.

The code fails to account for endianess when reading the packed
values.

This commit attempts to correct the problem by reversing the order
the packed values are read on big endian systems.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100613
Cc: "17.2" "17.1" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Ray Strode <rstrode@redhat.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
(cherry picked from commit 75cb6e3617)
2017-09-13 21:34:02 +01:00
Jason Ekstrand
93b794f094 anv/formats: Nicely handle unknown VkFormat enums
This fixes some crashes in the dEQP-VK.memory.requirements.core.* tests.
I'm not sure whether or not passing out-of-bound formats into the query
is supposed to be allowed but there's no harm in protecting ourselves
from it.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Bugzilla: https://bugs.freedesktop.org/101956
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 242211933a)

Squashed with:

anv: fix off by one in array check

`anv_formats[ARRAY_SIZE(anv_formats)]` is already one too far.
Spotted by Coverity.

CovID: 1417259
Fixes: 242211933a "anv/formats: Nicely handle unknown VkFormat enums"
Cc: Jason Ekstrand <jason.ekstrand@intel.com>
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
(cherry picked from commit 0c7272a66c)
2017-09-13 21:33:55 +01:00
Charmaine Lee
520786586d vbo: fix offset in minmax cache key
Instead of saving primitive offset in the minmax cache key,
save the actual buffer offset which is used in the cache lookup.

Fixes rendering artifact seen with GoogleEarth when run with
VMware driver.

v2: Per Brian's comment, initialize offset to avoid compiler warning.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Brian Paul <brianp@vmware.com>
(cherry picked from commit 2d93b462b4)

Squashed with:

vbo: fix build errors on android

incompatible pointer to integer conversion assigning to 'GLintptr' (aka 'int')
from 'const char *' [-Werror,-Wint-conversion]

      offset = indices;
             ^ ~~~~~~~

Fixes: 2d93b462b4 ("vbo: fix offset in minmax cache key")
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit 0986f68632)
2017-09-13 21:32:19 +01:00
Michael Olbrich
9ffe4cc6c0 egl/dri2: only destroy created objects
dri2_display_destroy may be called by dri2_initialize_wayland_drm() if
initialization fails. In this case, these objects may not be initialized.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit 81d5c31631)
2017-09-13 21:05:57 +01:00
Brian Paul
2b0e70e2ae llvmpipe: initialize llvmpipe->dirty with LP_NEW_SCISSOR
If llvmpipe_set_scissor_states() is never called, we still need to be sure
that derived scissor/clip state is updated.  As of commit 743ad599a9
that function might not be called.

Fixes regressed Piglit gl-1.0-scissor-offscreen -fbo -auto test.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101709
Fixes: 743ad599a9 ("st/mesa: don't set 16 scissors and 16 viewports
if they're unused")
Cc: "17.2" <mesa-stable@lists.freedesktop.org>

(cherry picked from commit 88cdf16871)
2017-09-13 21:05:57 +01:00
Emil Velikov
3a9b2e0e7d cherry-ignore: ignore gfx9 tile swizzle fix
As mentioned by Dave and Bas, the patch is not applicable for 17.2

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-13 21:05:56 +01:00
Emil Velikov
188010c68d cherry-ignore: add getCapability patches
Adding extra infra is not acceptable for stable. Offending code has been
wrapped in a Android compile guard with an earlier commit.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-13 21:05:33 +01:00
Emil Velikov
b4473dd519 docs: add sha256 checksums for 17.2.0
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-04 18:24:29 +01:00
Emil Velikov
f5925b2897 docs: Update 17.2.0 release notes
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-04 18:16:16 +01:00
Emil Velikov
702950d1ad Update version to 17.2.0(final)
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-09-04 15:16:21 +01:00
Emil Velikov
909f2b6aa2 Update version to 17.2.0-rc6
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-08-31 10:06:29 +01:00
Emil Velikov
3666d9ab99 egl/wayland: make sure HAS_$FORMAT is set for wl_dmabuf
Otherwise eglCreateWaylandBufferFromImageWL will fail, since we
have no "supported" format.

Fixes: 02cc359372 ("egl/wayland: Use linux-dmabuf interface for buffers")
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
(cherry picked from commit 9e07005e87)
2017-08-29 19:14:14 +01:00
Emil Velikov
5303f0c246 egl/wayland: set correct format with wl_dmabuf as wl_drm is missing
For most/all cases today, we have wl_drm available alongside wl_dmabuf.
Yet in the long run, we want to make sure the latter can operate without
any traces of the former.

Fixes: 02cc359372 ("egl/wayland: Use linux-dmabuf interface for buffers")
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
(cherry picked from commit da100fe697)
2017-08-29 19:14:13 +01:00
Emil Velikov
8d51747dc0 egl/wayland: polish object teardown in dri2_wl_destroy_surface
The wl_drm wrapper is created before the wl display/surface ones.
Thus make sure we destroy it after them. In reality it should not make
any difference either way.

Fixes: 03dd9a88b0 ("egl/wayland: Use per-surface event queues")
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
(cherry picked from commit 1a8015e753)
2017-08-29 19:14:13 +01:00
Emil Velikov
3b41bef9a4 egl/wayland: plug leaks in dri2_wl_create_window_surface() error path
We forgot to teardown the wl display/surface wrappers.

Fixes: 03dd9a88b0 ("egl/wayland: Use per-surface event queues")
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
(cherry picked from commit 83442112d7)
2017-08-29 19:14:13 +01:00
Grazvydas Ignotas
e040bdbc8a radv: clear dynamic_shader_stages on create
Valgrind reports it's being used uninitialized.

Fixes: f4e499ec79 "radv: add initial non-conformant radv vulkan driver"
Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
(cherry picked from commit 7780374833)
2017-08-29 19:14:13 +01:00
Dave Airlie
e4c4b0dbcb radv/wsi: Compute correct row_pitch for GFX9.
(commit split out by Bas Nieuwenhuizen)

Fixes: 65477bae9c "radv: enable GFX9 on radv"
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
(cherry picked from commit 9573bd70e1)
2017-08-29 19:14:13 +01:00
Emil Velikov
ca9817a6e8 egl: don't NULL deref the .get_capabilities function pointer
One could easily introduce version 3 of the DRI2fenceExtension,
extending the struct, while not implementing the above function.

Thus we'll end up with NULL pointer, and dereferencing it won't fare
too well.

Fixes: 0201f01dc4 ("egl: add EGL_ANDROID_native_fence_sync")
Cc: Rob Clark <robclark@freedesktop.org>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
(cherry picked from commit f0d053cb6d)
2017-08-29 19:14:13 +01:00
Bas Nieuwenhuizen
df8e65feb7 radv: Fix sparse BO mapping merging.
If we merge a mapping with the mapping before it, we also need
to not only change the offset, but also the bo offset.

Fixes: 715df30a4e "radv/amdgpu: Add winsys implementation of virtual buffers."
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 9b7e663da1)
2017-08-29 19:14:13 +01:00
Bas Nieuwenhuizen
ec94b874c4 radv: Fix off by one in MAX_VBS assert.
e.g. 0 + 32 <= 32 should be valid.

Fixes: f4e499ec79 "radv: add initial non-conformant radv vulkan driver"
Tested-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit fba0e07869)
2017-08-29 19:14:13 +01:00
Bas Nieuwenhuizen
17e5b0f1db radv: Don't set a new subpass on compute resolve.
We don't use the render path so totally unneeded.

Fixes: 19be95f71e "radv: add subpass resolve compute path"
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit bd81cb3206)
2017-08-29 19:14:13 +01:00
Bas Nieuwenhuizen
69c793c0ce radv: Remove some intel comments from the resolve code.
These are clearly not applicable to radv.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
(cherry picked from commit e5c4e10769)
2017-08-29 19:14:13 +01:00
Dave Airlie
b29987ace4 radv: don't crash if we have no framebuffer
Recording secondaries with no framebuffer attachment may
make this happen, though this might not be the complete solution.

(esp if someone does meta stuff in there, would we have to
save things, not sure).

Fixes: f4e499ec79 ("radv: add initial non-conformant radv vulkan driver")
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 4a091b0788)
2017-08-29 19:14:13 +01:00
Dave Airlie
22c1e2e966 radv: fix predication on gfx9
When I added gfx9 I did it wrong, this fixes it.

Fixes: 5247b311e9 "radv/gfx9: fix set predication packet."
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 12fd0f8dc1)
2017-08-29 19:14:13 +01:00
Timothy Arceri
f01cd4feeb mesa: fix ES only draw if we have vertex positions
This code was separated from the validation code so it could
use used with KHR_no_error paths. The return values were inverted
to reflect the name of the helper, but here the condtion was
mistakenly inverted rather than the return value.

Fixes: 4df2931a87 (mesa/vbo: move some Draw checks out of validation)

Reported-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit a4635c84dc)
2017-08-29 19:14:13 +01:00
Daniel Stone
c77ae9744e egl: Add dma_buf_import_modifiers for glvnd
Make sure we advertise the new entrypoints to libglvnd's EGL dispatch.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reported-by: Emmanuel Gil Peyrot <emmanuel.peyrot@collabora.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101982
Fixes: 4c412293d0 ("egl: advertise EGL_EXT_image_dma_buf_import_modifiers")
(cherry picked from commit 85ef0215dd)
2017-08-29 19:14:13 +01:00
Daniel Stone
5fdf832235 egl: Update headers from Khronos
Taken from egl-registry 7d68647c4dab.

Signed-off-by: Daniel Stone <daniels@collabora.com>
(cherry picked from commit 2eee03b7a1)
2017-08-29 19:14:13 +01:00
Emil Velikov
46573f1028 util: move string_to_uint_map to glsl
The functionality is used by glsl and mesa. With the latter already
depending on the former.

With this in place the src/util/ static library libmesautil.la no longer
has a C++ dependency. Thus objects which use it (like libEGL) don't need
the C++ link.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Fixes: 02cc359372 ("egl/wayland: Use linux-dmabuf interface for buffers")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101851
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Suggested-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
Tested-by: James Harvey <lothmordor@gmail.com>
(cherry picked from commit 0ac78dc925)
2017-08-29 19:14:13 +01:00
Jason Ekstrand
1dfbfad0b9 nir: Fix system_value_from_intrinsic for subgroups
A couple of the cases were backwards

Reviewed-by: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 63e79a8a77)
2017-08-29 19:14:13 +01:00
Ilia Mirkin
289b2a8788 st/mesa: fix handling of vertex array double inputs
The is_double_vertex_input needs to be set for arrays of doubles as
well.

Fixes KHR-GL45.enhanced_layouts.varying_array_locations

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit ae53bff8b1)
2017-08-29 19:14:13 +01:00
Ilia Mirkin
1960d8f0dd glsl: fix counting of vertex shader output slots used by explicit vars
The argument to count_attribute_slots should only be set to true for
vertex inputs, not for all vertex shader varyings.

Fixes KHR-GL45.enhanced_layouts.varying_locations

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit eefeff09a7)
2017-08-29 19:14:13 +01:00
Christian Gmeiner
2c964f68a8 etnaviv: use correct param for etna_compatible_rs_format(..)
Found by code inspection.

Fixes: c9e8b49b88 ("etnaviv: gallium driver for Vivante GPUs")
Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
(cherry picked from commit 67fc3e37a7)
2017-08-29 19:14:13 +01:00
Kai Chen
5d5a13e375 egl/wayland: Use roundtrips when awaiting buffer release
In get_back_bo, we use wl_display_dispatch_queue() to block and wait for
a buffer release event. However, not all Wayland compositors flush the
client socket on posting a buffer-release event, so by only blocking
client-side, we may block indefinitely, or at least need to wait for an
input event / frame completion to arrive for the compositor to flush.

We now use dispatch_queue as a first pass, but if our entire buffer pool
is exhausted, use a roundtrip (an immediately-triggered wl_callback) to
ensure that the compositor flushes out our release event immediately.

[daniels: Modified comment and commit message.]

Signed-off-by: Kai Chen <kai.chen@intel.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
CC: <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 151188d1e3)
2017-08-29 19:14:13 +01:00
Dave Airlie
9d6d567046 radv/gfx9: gfx9 has buffer sizing rules like pre-VI.
This fixes:
dEQP-VK.robustness.buffer_access.* on GFX9.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 19f6906c1e)
2017-08-29 19:14:13 +01:00
Bas Nieuwenhuizen
6f0d910617 ac/nir: Cast sources of integer ops to int.
The int32->float semantic conversion got dropped in a testcase,
because the src was already float. On closer inspection I decided
to add a few more casts for integer op operands to be safe too.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 43595db302)
[Emil Velikov: squash trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/amd/common/ac_nir_to_llvm.c
2017-08-29 19:14:13 +01:00
Ilia Mirkin
a0b3095bb6 nv50/ir: properly set sType for TXF ops to U32
All of the coordinates and LOD args are integers for TXF. This mostly
doesn't matter, except for converting into a levelZero=true operation by
removing an explicit zero LOD. For the comparison against zero to work
properly, the sType of the instruction has to be set correctly.

Fixes: KHR-GL45.robust_buffer_access_behavior.texel_fetch
Reported-by: Karol Herbst <karolherbst@gmail.com>
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 96be442b77)
2017-08-29 19:14:13 +01:00
Dave Airlie
c25e896810 radv/gfx9: don't expose linear depth on vega.
This just zeros out the linear flags for gfx9 + depth formats.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 8985ad494b)
2017-08-29 19:14:13 +01:00
Dave Airlie
9172c7744a radv: don't degrade tiling mode for small compressed or depth texture.
This is what radeonsi does, so we should do the same, also vega
doesn't support linear depth textures anyways.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 5d26e0baf2)
2017-08-29 19:14:12 +01:00
Dave Airlie
adce678550 radv/gfx9: only minify image view width/height/depth before gfx9.
For gfx9 the addressing for images has changed, so we need to
provide the hw with the level0, however we still need to scale
for format block differences (so our compressed upload paths still
work).

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit bae7723e13)
2017-08-29 19:14:12 +01:00
Dave Airlie
7f3555951e radv/image: don't rescale width/height if the format isn't changing
If the image view has the same format, we don't need to rescale
the w/h.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit a74d987431)
2017-08-29 19:14:12 +01:00
Dave Airlie
bc39869414 radv: cleanup some image view descriptor setup.
Avoid passing the vulkan image creation into the image view descriptor
setup. This cleans up the usage of range inside the init, instead
using the properly inited values in the image view.

This is just a cleanup but some future vega changes will depend on it.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 5378b5d071)
2017-08-29 19:14:12 +01:00
Dave Airlie
93aee55a7f radv/gfx9: emit sx_mrt_blend registers
GFX9 needs the SX MRT blend registers programmed, port over
the code from radeonsi to workout the values from the blend
state, and program the registers on rbplus systems.

This fixes lots of:
dEQP-VK.pipeline.blend.*

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 9c080100d3)
2017-08-29 19:14:12 +01:00
Dave Airlie
996e3239c9 radv: bump space check for indexed draw.
For the GFX9 packet we need one more dword.

Fixes an assert in:
dEQP-VK.draw.shader_draw_parameters.base_vertex.draw_indexed

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 864eb18527)
2017-08-29 19:14:12 +01:00
Dave Airlie
9b856ea323 radv/gfx9: fixup db/stencil disable.
This fixes disabled Z/stencil.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit d987b4ab9e)
2017-08-29 19:14:12 +01:00
Dave Airlie
d8f8689101 radv/gfx9: fix level count in color register setup.
There was an off by one here.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 11834195e9)
2017-08-29 19:14:12 +01:00
Dave Airlie
dbca342014 radv/gfx9: use total levels in texture descriptor
We need to use all the levels when filling out the gfx9
descriptor.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit df09f1f3cd)
2017-08-29 19:14:12 +01:00
Kenneth Graunke
c0c03d0003 i965: Make a BRW_NEW_FAST_CLEAR_COLOR dirty bit.
When changing fast clear colors, we need to emit new SURFACE_STATE
with the updated color at the next draw call.

Most things work today because the atoms that handle SURFACE_STATE
for images (mutable images, textures, render targets) also listen to
BRW_NEW_BLORP, causing us to re-emit these on every BLORP operation.
However, this is overkill - most BLORP operations don't require us
to re-emit SURFACE_STATE.

One case where this is broken today is a fast clear to a different
color followed by a non-coherent framebuffer fetch.  The renderbuffer
read atom doesn't listen to BRW_NEW_BLORP, and would not get the new
fast clear color.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
(cherry picked from commit 54c41af0aa)
[Emil Velikov: squash trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/mesa/drivers/dri/i965/brw_tcs_surface_state.c
2017-08-29 19:14:12 +01:00
Marek Olšák
9013f80f67 radeonsi: emit VGT_REUSE_OFF in the right place
clip_regs aren't marked dirty when writes_viewport_index is changed.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit 8dadb07790)
2017-08-29 19:14:12 +01:00
Marek Olšák
d93e6a0409 radeonsi/gfx9: properly handle imported textures with unexpected swizzle mode
Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit fdef2f0fd1)
[Emil Velikov: r600_surface_import_metadata is not separate function,
tile_swizzle isn't in branch]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/gallium/drivers/radeon/r600_texture.c
2017-08-29 19:14:12 +01:00
Marek Olšák
cf2f05ff91 radeonsi/gfx9: add a temporary workaround for a tessellation driver bug
The workaround will do for now. The root cause is still unknown.

This fixes new piglit: 16in-1out

Cc: 17.1 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit 166823bfd2)
2017-08-29 18:37:57 +01:00
Jason Ekstrand
91715a3681 i965/clear: Quantize the depth clear value based on the format
In f9fd976e8a we changed the clear value to be stored as an
isl_color_value.  This had the side-effect same clear value check is now
happening directly between the f32[0] field of the isl_color_value and
ctx->Depth.Clear.  This isn't what we want for two reasons.  One is that
the comparison happens in floating point even for Z16 and Z24 formats.
Worse than that, ctx->Depth.Clear is a double so, even for 32-bit float
formats, we were comparing as doubles and not floats.  This means that
the test basically always fails for anything other than 0.0f and 1.0f.
This caused a slight performance regression in Lightsmark 2008 because
it was using a depth clear value of 0.999 which can't be stored in a
32-bit float so we were doing unneeded resolves.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Bugzilla: https://bugs.freedesktop.org/101678
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 0ae9ce0f29)
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
[Emil Velikov: squash trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/mesa/drivers/dri/i965/brw_clear.c
2017-08-29 18:37:57 +01:00
Rob Herring
9ba7693a9a Android: Fix LLVM duplicated symbols linking for N and M
Both statically linking libLLVMCore and dynamically linking libLLVM causes
duplicated symbols in gallium_dri.so and it fails to dlopen. We don't
really need to link libLLVMCore, but just need generated headers to be
built first. Dynamically linking to libLLVM instead is enough to do
that. Thanks to Qiang Yu for finding the root cause.

With this change, we can align all versions and just have libLLVM as a
shared lib dependency.

This also requires changes in the M and N versions of LLVM to export the
include paths for libLLVM. AOSP master is okay.

Fixes: 26aee6f4d5 ("Android: rework LLVM build support")
Reported-by: Mauro Rossi <issor.oruam@gmail.com>
Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Signed-off-by: Qiang Yu <Qiang.Yu@amd.com>
Signed-off-by: Rob Herring <robh@kernel.org>
(cherry picked from commit 4734bfc02a)
2017-08-29 18:37:57 +01:00
Samuel Pitoiset
1f8cb4c243 radeonsi: update non-resident bindless descriptors if needed
Only resident bindless descriptors are currently updated and
re-uploaded, this makes sure that the non-resident ones are
also updated.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 2843c5d15c)
2017-08-29 18:37:57 +01:00
Topi Pohjolainen
da4d1fd084 intel/blorp: Adjust intra-tile x when faking rgb with red-only
v2 (Jason): Adjust directly in surf_fake_rgb_with_red()

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101910

CC: mesa-stable@lists.freedesktop.org
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
(cherry picked from commit 393ec1a507)
2017-08-29 18:37:57 +01:00
Dave Airlie
7789126929 ac/nir: fixup layer/viewport export for GFX9.
GFX9 moved where the viewport index export goes.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit b040f51b61)
2017-08-29 18:37:57 +01:00
Christoph Haag
6389526ce5 mesa: only copy requested compressed teximage cubemap faces
This is analogous to commit 2259b11 which only fixed the regular case

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102308
Signed-off-by: Christoph Haag <haagch+mesadev@frickel.club>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 87556a650a)
2017-08-29 18:37:57 +01:00
Jason Ekstrand
2813e83f6c i965/tex: Don't pass samples to miptree_create_for_teximage
In 76e2f390f9, when Topi switched num_samples from 0 to 1 for
single-sampled, he accidentally switched the last parameter in the call
to miptree_create_for_teximage from 0 to 1 thinking it was num_samples
when it was actually layout_flags.  Switching from 0 to 1 added the
MIPTREE_LAYOUT_ACCELERATED_UPLOAD flag which causes us to allocate a
busy BO instead of an idle one.  This caused the subsequent CPU upload
to consistently stall.  The end result was a 15% performance drop in the
SynMark v7 DrvRes microbenchmark.  This restores the old behavior and
fixes the performance regression.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Fixes: 76e2f390f9
Bugzilla: https://bugs.freedesktop.org/102260
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit f24cf82d6d)
2017-08-29 18:37:57 +01:00
Jason Ekstrand
ea5c466c91 i965/miptree: Return NONE from texture_aux_usage when fully resolved
This little optimization improves the performance of SynMark v7
TexFilterTri by almost 10% on Sky Lake GT4 among other improvements.
We've been doing it for some time but somehow it got dropped during
the miptree refactoring.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Bugzilla: https://bugs.freedesktop.org/102258
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 61d2f3f1c2)
2017-08-29 18:37:57 +01:00
Jason Ekstrand
d174214218 i965: Stop looking at NewDriverState when emitting 3DSTATE_URB
Looking at NewDriverState is not safe in general.  The state atom system
is set up to ensure that new bits that get added to NewDriverState get
accumulated into the set of bits used when emitting atoms but it doesn't
go the other way.  If we read NewDriverState, we may not get the full
picture because the per-pipeline state (3D or compute) does not get
added to NewDriverState before state emit is done.  It's especially
dangerous to do this from BLORP (either explicitly or implicitly when
BLORP calls gen7_upload_urb) because that does not happen during one of
the normal state upload paths.

This commit solves the problem by whacking all of the per-shader-stage
URB sizes to zero whenever we change the total URB size.  We still have
to flag BRW_NEW_URB_SIZE to ensure that the gen7_urb atom triggers but
the actual decision in gen7_upload_urb can now be based entirely on URB
sizes rather than on state atoms.  This also makes BLORP correct because
it just asks for a new URB config whenever the vsize is too small and so
any change to the total URB size will trigger blorp to re-emit as well
because 0 < vs_entry_size.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Bugzilla: https://bugs.freedesktop.org/102289
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit d5e217dbfd)
2017-08-29 18:37:57 +01:00
Kenneth Graunke
a14dd9fff6 i965: Mark all EGLimages as non-coherent.
EGLimages are shared with external users, and we don't know what they're
going to do with them.  They might scan them out.  They might access
them in a way that doesn't work with our explicit clflushing.

It's safest to simply mark them non-coherent.

Chris Wilson caught this problem and wrote a similar (though less
aggressive) patch to solve it; the miptree code has since undergone
a lot of refactoring so I had to rewrite it.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
(cherry picked from commit bc56dfbf3f)
[Emil Velikov: squash trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/mesa/drivers/dri/i965/intel_mipmap_tree.c
2017-08-29 18:37:57 +01:00
Jason Ekstrand
3e62ba0907 i965/miptree: Rework create flags
The only one of the three remaining flags that has anything whatsoever
to do with layout is TILING_NONE.  This commit renames them to
MIPTREE_CREATE_*, documents the meaning of each flag, and makes the
create functions take an actual enum type so GDB will print them nicely.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
(cherry picked from commit 24a0da338f)
2017-08-29 18:37:57 +01:00
Jason Ekstrand
1011fd22ae i965/miptree: Delete MIPTREE_LAYOUT_TILING_(Y|ANY)
The only force tiling flag we really care about is LAYOUT_TILING_NONE.
The others don't actually do anything but add confusion.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
(cherry picked from commit 55116839d9)
2017-08-29 18:37:57 +01:00
Jason Ekstrand
dd03cb5c5e i965/miptree: Delete MIPTREE_LAYOUT_FOR_SCANOUT
The flag hasn't affected actual surface layout for some time.  The only
purpose it served was to set bo->cache_coherent = false on the BO used
to create the miptree.  This is fairly silly because we can just set
that directly from the caller where it makes much more sense.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
(cherry picked from commit a5a673dfa7)
[Emil Velikov: squash trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/mesa/drivers/dri/i965/intel_mipmap_tree.c
2017-08-29 18:37:57 +01:00
Jason Ekstrand
046730b8a3 i965/miptree: Delete some unused layout flags
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
(cherry picked from commit 2bca18be44)
2017-08-29 18:37:57 +01:00
Jason Ekstrand
f83d979759 i965/miptree Remove layout_flags parameter form is_mcs_supported
The one caller of is_mcs_supported passes 0 in as the layout_flags
unconditionally.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
(cherry picked from commit 0e4d9a4b37)
2017-08-29 18:36:14 +01:00
Dave Airlie
2cdb51e76d radv: disable texture gather workaround on gfx9.
Not required anymore.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 4c02e2bd95)
[Emil Velikov: chip_class is in ctx::options, not ctx::abi]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-08-29 16:18:49 +01:00
Emil Velikov
15f23fb855 Update version to 17.2.0-rc5
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
2017-08-21 11:49:19 +01:00
Emil Velikov
dd4fafcfda cherry-ignore: ignore storage offset fixes
For stable the regressions were addressed by reverting the offending
commits - see previous commit.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-08-19 02:57:19 +01:00
Samuel Pitoiset
17a3e4891b Revert "mesa: stop assigning unused storage for non-bindless opaque types"
This reverts commit fcbb93e860 and
also  commit 7c5b204e38 to avoid
compilation errors.

Basically, the parameter indexes look wrong when a non-bindless
sampler is declared inside a nested struct (because it is skipped).
I think it's safer to just restore the previous behaviour which is
here since ages and also because the initial attempt is only a
little performance improvement.

This fixes a regression with
ES2-CTS.functional.shaders.struct.uniform.sampler_nested*.

Note: this patch is only for 17.2 since the fixes in master are rather
invasive. Namely:

365d34540f mesa: correctly calculate the storage offset for i915
de0e62e106 st/mesa: correctly calculate the storage offset

Fixes: fcbb93e860 ("mesa: stop assigning unused storage for non-bindless
opaque types")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101983
Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
2017-08-19 02:57:19 +01:00
Jason Ekstrand
85cc3533fa i965/miptree: Set supports_fast_clear = false in make_shareable
The make_shareable function deletes the aux buffer and then whacks
aux_usage to ISL_AUX_USAGE_NONE but not unsetting supports_fast_clear.
Since we only look at supports_fast_clear to decide whether or not to do
fast clears, this was causing assertion failures.

Reported-by: Tapani Pälli <tapani.palli@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101925
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
(cherry picked from commit e7a52cc381)
2017-08-19 02:57:19 +01:00
Eric Anholt
93bd5fbfe1 broadcom/vc4: Build the vc4_tiling_lt_neon.c with -mfpu=neon on ARM.
If you don't pass this, the compiler refuses to compile the assembly for
pre-v7 CPUs.  This also keeps us from building identical, non-NEON code on
aarch64 and x86.

Fixes: a373f77662 ("vc4: Use a wrapper file to set VC4_BUILD_NEON instead of CFLAGS.")

v2: Fix Android build by just appending NEON_C_SOURCES when
    ARCH_ARM_HAVE_NEON.

Tested-by: Rob Herring <robh@kernel.org>
(cherry picked from commit bd5efbd70b)
[Emil Velikov: address libvc4_la_LIBADD conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/gallium/drivers/vc4/Makefile.am
2017-08-19 02:57:19 +01:00
Eric Anholt
833f12abdf configure.ac: Introduce HAVE_ARM_ASM/HAVE_AARCH64_ASM and the -D flags.
I've been trying to get away without these conditionals in vc4's NEON
code, but it meant compiling extra unused code on x86, and build failing
on ARMv6.

v2: Use the _arm/_arm64 flags to simplify detection (suggested by Rob),
    but hide the _arm version under ARCH_ARM_HAVE_NEON to keep from trying
    to build this stuff for armv5te.

Tested-by: Rob Herring <robh@kernel.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
(cherry picked from commit ba8533b6ea)
2017-08-19 02:57:19 +01:00
Eric Anholt
1fc44a3ba0 util: Fix build on old glibc.
We need to link librt for u_thread.h's clock_gettime() call.

Fixes: b822d9dd67 ("gallium/util: move u_queue.{c,h} to src/util")
Reviewed-by: Matt Turner <mattst88@gmail.com>
(cherry picked from commit b94ddc181b)
2017-08-19 02:57:19 +01:00
Scott D Phillips
6ccb75e869 intel/genxml: Fix gen10 BLEND_STATE variable length packing
BLEND_STATE packing was modified to be variable-length in:

 9670124e31 genxml: Make BLEND_STATE command support variable length array.

The initial gen10.xml still had the old, fixed-length style
definition for BLEND_STATE. So gen10_upload_blend_state would
overwrite the packed BLEND_STATE_ENTRYs with its own fixed array
of all-zero entries when packing BLEND_STATE. This caused
BLEND_STATE upload to not work at all.

Fixes: aa416f515a ("i965/genxml: Add gen10.xml")
Reviewed-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
(cherry picked from commit d6539608a4)
2017-08-19 02:57:19 +01:00
Chris Wilson
b4fec4271b i965: Always allow CPU readback of the scanout on LLC platforms
LLC platforms are magic in that reads from the CPU are always cache
coherent, or rather GPU writes that bypass LLC do still invalidate the
appropriate cache line.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
(cherry picked from commit 49eda75df6)
2017-08-19 02:57:19 +01:00
Ilia Mirkin
efb9b5740e glsl: add a few missing int64 constant propagation cases
Fixes KHR-GL45.shader_ballot_tests.ShaderBallotAvailability, which
causes some silly swizzles to appear, triggering this optimization to
get hit.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 9c8f017f77)
2017-08-19 02:57:19 +01:00
Dave Airlie
e44393cb23 radv: disable support for VEGA for now.
I'm working on this, but I'm not sure I'll make 17.2 at this stage,
maybe 17.2.1.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 611076a41a)
2017-08-19 02:57:19 +01:00
Ilia Mirkin
81b5a6a85b nv50/ir: fix TXQ srcMask
src0.x is always read for the LOD, irrespective of which outputs are
read.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 934511d1f3)
2017-08-19 02:57:19 +01:00
Ilia Mirkin
5795e42116 nv50/ir: fix srcMask computation for TG4 and TXF
This affects which inputs are marked as used. In a situation where only
the texture instruction uses an input, it might have been ignored as
unused due to input masks.

Affects subtests of KHR-GL45.texture_cube_map_array.sampling

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 054c54d1be)
2017-08-19 02:57:19 +01:00
Frank Richter
42e8f3ccdf gallium/os: fix os_time_get_nano() to roll over less
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102241
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
(cherry picked from commit 7fb7287ce7)
2017-08-19 01:46:08 +01:00
Frank Richter
e452ed26ff st/wgl: check for negative delta in wait_swap_interval()
This can happen because of rollover.  See bug report for details.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102241
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
(cherry picked from commit d90e05ad48)
2017-08-19 01:46:08 +01:00
Frank Richter
c7c6ba44ca st/mesa: fix a null pointer access
Fixes crash with llvmpipe on Windows.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102148
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Brian Paul <brianp@vmware.com>
(cherry picked from commit 496a691e35)
2017-08-19 01:46:08 +01:00
Tim Rowley
bdab7c69f5 swr/rast: Fix invalid casting for calls to Interlocked* functions
CID: 1416243, 1416244, 1416255
CC: mesa-stable@lists.freedesktop.org
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
(cherry picked from commit b333bc753e)
[Emil Velikov: resolve trivial conflicts - LONG -> long]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/gallium/drivers/swr/rasterizer/core/api.cpp
	src/gallium/drivers/swr/rasterizer/core/threads.cpp
2017-08-19 01:45:47 +01:00
Ilia Mirkin
47507ec1fd glsl/ast: update rhs in addition to the var's constant_value
We continue in the code to do some more things with the rhs, including
setting a constant initializer. If the type is wrong, this causes some
confusion down the line, leading to assertions. This makes sure that the
rhs processing continues to flow as-if the type was correct to start
with (even though the state has been marked as an error state).

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101766
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 978c4c597a)
2017-08-19 01:42:08 +01:00
Dave Airlie
3fb5cc2637 radv/gfx9: for fast clear use is_linear flag.
The legacy test won't work on gfx9.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 694d59fbaf)
2017-08-19 01:42:08 +01:00
David Airlie
42d6e7eba2 radv/gfx9: handle GFX9 opaque metadata
port the opaque metadata changes from radeonsi for gfx9.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit e43cc3e3af)
2017-08-19 01:36:25 +01:00
David Airlie
f60ff71b6f radv: emit db_htile_surface reg on gfx9 as well
This is also a GFX9 register.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 674ecbfef2)
2017-08-19 01:36:25 +01:00
Dave Airlie
f70f216564 radv/gfx9: remove some leftover gfx6 descriptor setup.
We set this later in the non-gfx9 path, just remove these
bits from here.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit fc600eb98d)
2017-08-19 01:36:25 +01:00
Dave Airlie
827bf79052 radv/gfx9: fix set predication packet.
The predication packet changed format on GFX9, update the driver.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 5247b311e9)
2017-08-19 01:36:25 +01:00
Marek Olšák
2b10b3a417 radeonsi: disable CE by default
It makes performance worse by a very small (hard to measure) amount.
We've done extensive profiling of this feature internally.

Cc: 17.1 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Acked-by: Christian König <christian.koenig@amd.com>
(cherry picked from commit 1ab7fed707)
2017-08-19 01:36:25 +01:00
Scott D Phillips
27e7a3a5ef i965/blorp: Correct type of src_format in call to intel_miptree_texture_aux_usage
intel_miptree_texture_aux_usage() takes an isl_format, but we are
passing a mesa_format. clang warns:

 brw_blorp.c:305:52: warning: implicit conversion from enumeration
    type 'mesa_format' to different enumeration type
    'enum isl_format' [-Wenum-conversion]
       intel_miptree_texture_aux_usage(brw, src_mt, src_format);
       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~              ^~~~~~~~~~

Fixes: fc1639e46d ("i965/blorp: Use texture/render_aux_usage for blits")
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
(cherry picked from commit f7dfc44c61)
2017-08-19 01:36:25 +01:00
Ilia Mirkin
da005f5566 nv50/ir: clean up saturated values immediately
Since we don't iterate to a fixed point, we can end up in situations
where we have a SAT instruction + a long immediate. This is not legal.
However since it's immediately computable, just run unary straight away
to handle the situation.

Fixes: 24a799ad35 ("nv50/ir: fix ConstantFolding with saturation")
Reported-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 165e18dd21)
2017-08-19 01:36:25 +01:00
Jeremy Huddleston Sequoia
cc8ae8842b glxcmds: Fix a typo in the __APPLE__ codepath
s/DummyContext/dummyContext/

Regressed-in: 5d9b50e596
Signed-off-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
(cherry picked from commit c1c4c18a80)
2017-08-17 15:48:20 -07:00
Emil Velikov
3165f9877e Update version to 17.2.0-rc4
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-08-12 17:04:27 +01:00
Dave Airlie
1e11687029 radv: force cs/ps/l2 flush at end of command stream. (v2)
This seems like a workaround, but we don't see the bug on CIK/VI.

On SI with the dEQP-VK.memory.pipeline_barrier.host_read_transfer_dst.*
tests, when one tests complete, the first flush at the start of the next
test causes a VM fault as we've destroyed the VM, but we end up flushing
the compute shader then, and it must still be in the process of doing
something.

Could also be a kernel difference between SI and CIK.

v2: hit this with a bigger hammer. This fixes a bunch of hangs
in the vk cts with the robustness tests.

Fixes: f4e499ec79 ("radv: add initial non-conformant radv vulkan driver")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101334
Acked-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 82ba384c10)
2017-08-11 21:00:49 +01:00
Dave Airlie
ea595756f8 radv: fix MSAA on SI gpus.
This ports the workaround from radeonsi, that was missing in radv.

This fixes Talos rendering when MSAA is enabled on my Tahiti card.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Fixes: f4e499ec7 (radv: add initial non-conformant radv vulkan driver)
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 8bf3930751)
2017-08-11 21:00:46 +01:00
Dave Airlie
ffd8120284 radv: fix f16->f32 denorm handling for SI/CIK. (v2)
This just copies the code from the -pro shaders,
and fixes the tests on CIK.

With this CIK passes the same set of conformance
tests as VI.

Fixes: 83e58b03 (radv: flush f32->f16 conversion denormals to zero. (v2))
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 3f389f75b6)
2017-08-11 21:00:43 +01:00
Bas Nieuwenhuizen
9d65214f3d radv: Use the correct channel for alpha in resolve srgb conversion.
The argument here is a bitmask, so the old code selected .xy, which
got silently truncated to .x when constructing the vec4 from components,
instead of using .w.

Fixes: 588185eb6b "radv/meta: add srgb conversion to end of resolve shader."
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit acba3a3151)
2017-08-11 21:00:40 +01:00
Bas Nieuwenhuizen
a57390cee0 radv: Only convert linear->srgb in compute resolves.
It justs works with the fragment shader resolve, so no need to do
a custom conversion. In fact with SRGB dest, it actually gives
wrong results.

Fixes: 69136f4e63 "radv/meta: add resolve pass using fragment/vertex shaders"
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 15e5a7a683)
2017-08-11 21:00:37 +01:00
Bas Nieuwenhuizen
8b706102eb radv: Don't use SRGB format for image stores during resolve.
These seem to store very bogus results. Luckily there is some code
that converts srgb->linear already, so just making the descriptor
format UNORM should work.

Fixes: 588185eb6b "radv/meta: add srgb conversion to end of resolve shader."
Reviewed-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 8286c3a49f)
2017-08-11 21:00:34 +01:00
Marek Olšák
7f5d86ebaa radeonsi/gfx9: use the VI codepath for clamping Z
This fixes corrupted shadows in Unigine Valley.
The corruption disappeared when I stopped setting IMG_DATA_FORMAT_24_8
for depth.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit 27fef5d52d)
2017-08-11 20:59:31 +01:00
Kenneth Graunke
1c1653d7b0 isl: Validate row pitch of stencil surfaces.
Also, silence an obnoxious finishme that started occurring for all
GL applications which use stencil after the i965 ISL conversion.

v2: Check against 3DSTATE_STENCIL_BUFFER's pitch bits when using
    separate stencil, and 3DSTATE_DEPTH_BUFFER's bits when using
    combined depth-stencil.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
(cherry picked from commit 5563872dbf)
2017-08-11 20:59:28 +01:00
Emil Velikov
d4100b0d09 egl: avoid eglCreatePlatform*Surface{EXT,} crash with invalid dpy
If we have an invalid display fed into the functions, the display lookup
will return NULL. Thus as we attempt to get the platform type, we'll
deref. it leading to a crash.

Keep in mind that this will not happen if Mesa is built without X11 or
when the legacy eglCreate*Surface codepaths are used.

A similar check was added with earlier commit 5e97b8f5ce ("egl: Fix
crashes in eglCreate*Surface), although it was only applicable when the
surfaceless platform is built.

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
(cherry picked from commit 26fbb9eacd)
[Emil Velikov: resolve trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/egl/main/eglapi.c
2017-08-11 20:59:02 +01:00
Tim Rowley
bb6e5e5476 configure: remove trailing "-a" in swr architecture test
Fixes "configure: line 27326: test: argument expected"

CC: mesa-stable@lists.freedesktop.org
Reviewed-by: Matt Turner <mattst88@gmail.com>
(cherry picked from commit 4d9b0dcccb)
2017-08-11 20:58:19 +01:00
Marek Olšák
08d49e074d ac: fail shader compilation if libelf is replaced by an incompatible version
UE4Editor has this issue.

This commit prevents hangs (release build) or assertion failures (debug
build). It doesn't fix the editor, but catastrophic scenarios are
prevented.

Cc: 17.1 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 4630ede102)
2017-08-11 20:58:14 +01:00
Karol Herbst
75f5abb82f nv50/ir: fix ConstantFolding with saturation
For mul(a, +-1) codegen can generate OP_MOV with a saturation flag
set which is ignored at emission. The same can happen with add(a, 0),
and others.

Adding an assert for detecting more of such issues.

Fixes wrongly rendered water in Hitman Absolution running under wine.
Also a few shaders in Mad Max and Alien Isolation produce such MOVs.

CC: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Karol Herbst <karolherbst@gmail.com>
Reviewed-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
[imirkin: generalize the fix for other cases]
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
(cherry picked from commit 24a799ad35)
2017-08-11 20:57:54 +01:00
Alex Smith
f0b6298c05 radv: Fix decompression on multisampled depth buffers
Need to take the sample count into account in the depth decompress and
resummarize pipelines and render pass.

Fixes: f4e499ec79 ("radv: add initial non-conformant radv vulkan driver")
Signed-off-by: Alex Smith <asmith@feralinteractive.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 2e9a13bf22)
2017-08-11 20:57:50 +01:00
Jason Ekstrand
ac04187e33 i965/miptree: Call alloc_aux in create_for_bo
Originally, I had moved it to the caller to make some things easier when
adding the CCS modifier.  However, this broke DRI2 because
intel_process_dri2_buffer calls intel_miptree_create_for_bo but never
calls intel_miptree_alloc_aux.  Also, in hindsight, it should be pretty
easy to make the CCS modifier stuff work even if create_for_bo allocates
the CCS when DISABLE_AUX is not set.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 8e5808fc0c)
[Emil Velikov: resolve trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/mesa/drivers/dri/i965/intel_mipmap_tree.c
2017-08-11 20:57:36 +01:00
Jason Ekstrand
b1514579c2 intel/isl: Don't align the height of the last array slice
We were calculating the total height of 2D surfaces by multiplying the
row pitch by the number of slices.  This means that we actually request
slightly more space than actually needed since the padding on the last
slice is unnecessary.  For tiled surfaces this is not likely to make a
difference.  For linear surfaces, on the other hand, this means we may
require additional memory.  In particular, this makes the i965 driver
reject EGL imports of buffers which do not have this extra padding.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 4d27c6095e)
2017-08-11 20:52:22 +01:00
Jason Ekstrand
dc63c715cb intel/isl: Stop padding surfaces
The docs contain a bunch of commentary about the need to pad various
surfaces out to multiples of something or other.  However, all of those
requirements are about avoiding GTT errors due to missing pages when the
data port or sampler accesses slightly out-of-bounds.  However, because
the kernel already fills all the empty space in our GTT with the scratch
page, we never have to worry about faulting due to OOB reads.  There are
two caveats to this:

 1) There is some potential for issues with caches here if extra data
    ends up in a cache we don't expect due to OOB reads.  However,
    because we always trash the entire cache whenever we need to move
    anything between cache domains, this shouldn't be an issue.

 2) There is a potential issue if a surface gets placed at the very top
    of the GTT by the kernel.  In this case, the hardware could
    potentially end up trying to read past the top of the GTT.  If it
    nicely wraps around at the 48-bit (or 32-bit) boundary, then this
    shouldn't be an issue thanks to the scratch page.  If it doesn't,
    then we need to come up with something to handle it.

Up until some of the GL move to ISL, having the padding code in there
just caused us to harmlessly use a bit more memory in Vulkan.  However,
now that we're using ISL sizes to validate external dma-buf images,
these padding requirements are causing us to reject otherwise valid
images due to the size of the BO being too small.

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Tested-by: Tapani Pälli <tapani.palli@intel.com>
Tested-by: Tomasz Figa <tfiga@chromium.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit c15b92ce11)
2017-08-11 20:52:19 +01:00
Jason Ekstrand
9d9ea2c5a4 anv/formats: Allow sampling on depth-only formats on gen7
We can't sample from depth-stencil formats but on gen7 but we can sample
from depth-only formats.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102024
Reviewed-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 06d3115bb9)
2017-08-11 20:52:16 +01:00
Dave Airlie
07b5c78836 radv: avoid GPU hangs if someone does a resolve with non-multisample src (v2)
This is a bug in the app, but I'd rather avoid hanging the GPU,
esp if someone is running in validation and it takes out their
development environment.

v2: get it right, reverse the polarity.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 36a1b61321)
2017-08-11 20:52:13 +01:00
Emil Velikov
59f7fdb85e egl/x11: don't leak xfixes_query in the error path
If we get a xfixes v1.x we'll error out, without freeing the
xfixes_query reply.

Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
(cherry picked from commit c961b679fe)
2017-08-11 20:52:08 +01:00
Emil Velikov
29df4deef2 Update version to 17.2.0-rc3
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-08-07 12:45:40 +01:00
Jason Ekstrand
e4371d14f1 anv: Stop advertising VK_KHX_multiview
We don't want to advertise experimental extensions in actual releases.
However, there's no harm in leaving the code lying around in the tree.
2017-08-05 00:09:26 +01:00
Tomasz Figa
0b2c034f64 st/dri: enable 32-bit RGBX/RGBA formats only on Android
X/GLX can't handle them. This removes almost 500 GLX visuals that were
incorrectly exposed.

This is a less invasive version of Marek's .getCapability series.
Note: the patch is not applicable for master, but only for the 17.2
branch.

Suggested-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Tomasz Figa <tfiga@chromium.org>
CC: <mesa-stable@lists.freedesktop.org>
Fixes: f33d8af7aa "st/dri: add 32-bit RGBX/RGBA formats"
[Emil Velikov: commit message polish]
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
2017-08-05 00:09:26 +01:00
Tim Rowley
f4e163094d swr/rast: fix scons gen_knobs.h dependency
Copy/paste error was duplicating a gen_knobs.cpp rule.

Fixes: 5079c277b5 ("swr: [scons] Fix windows build")
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
(cherry picked from commit e4a6ae06cf)
2017-08-05 00:09:26 +01:00
Timothy Arceri
4a181e6244 mesa/st: fix conditional jump depends on uninitialised value
Reported by valgrind at:
glsl_to_tgsi_visitor::visit(ir_expression*) (st_glsl_to_tgsi.cpp:1560)

When compiling the Deus Ex shaders.

Fixes: 28a5e7104 ("st/glsl_to_tgsi: handle precise modifier")
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Karol Herbst <karolherbst@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 06237fc9e1)
2017-08-05 00:09:26 +01:00
Scott D Phillips
8ef9fe7229 gles: Restore some lost typedefs
GLES/gl.h has historically provided some typedefs that are not
used in the API itself. Restore these typedefs that were lost to
avoid breaking applications.

These seem to be the only typedefs removed in the update.

Fixes: 7fd0817 "Update Khronos-supplied headers"

[Eric: added a big warning to revert this patch when pulling the updated header]
Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
(cherry picked from commit 3db05ed1d1)
2017-08-05 00:09:26 +01:00
Dave Airlie
4f872e62c2 Revert "st_glsl_to_tgsi: rewrite rename registers to use array fully."
This reverts commit 3008161d28,
which caused a regression for VMWare.

The initial code had some recursion in it, that I removed by accident
trying to add back the recursion broke lots of things, take the high
road and revert for now.

Fixes: 3008161d (st_glsl_to_tgsi: rewrite rename registers to use array fully.)
Reviewed-by: Brian Paul <brianp@vmware.com>
Tested-by: Brian Paul <brianp@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit b8bea9a050)
2017-08-05 00:09:26 +01:00
Dave Airlie
c5fac38ced radv: handle 10-bit format clamping workaround.
This fixes:
dEQP-VK.api.copy_and_blit.core.blit_image.all_formats.*
for a2r10g10b10 formats as destination on SI/CIK hardware.

This adds support to the meta program for emitting 10-bit
outputs, and adds 10-bit support to the fragment shader key.

It also only does the int8/10 on SI/CIK.

Fixes: f4e499ec7 (radv: add initial non-conformant radv vulkan driver)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit df61a05019)
2017-08-05 00:09:26 +01:00
Bas Nieuwenhuizen
579ecfd91e radv: Don't underflow non-visible VRAM size.
In some APU situations the reported visible size can be larger than
VRAM size. This properly clamps the value.

Surprisingly both CTS and spec seem to allow a heap type with size 0,
so this seemed like the easiest option to me.

Signed-off-by: Bas Nieuwenhuizen <basni@google.com>
Fixes: 4ae84efbc5 "radv: Use enum for memory heaps."
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
(cherry picked from commit 8229706ad8)
2017-08-05 00:09:26 +01:00
Bruce Cherniak
6b279b3271 st/osmesa: add osmesa framebuffer iface hash table per st manager
Commit bbc29393d3 didn't include osmesa state_tracker.  This patch adds
necessary initialization.

Fixes crash in OSMesa initialization.

Created-by: Charmaine Lee <charmainel@vmware.com>
Tested-by: Bruce Cherniak <bruce.cherniak@intel.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 9966c85e01)
2017-08-05 00:09:26 +01:00
Dave Airlie
6efb8d79a9 intel/vec4/gs: reset nr_pull_param if DUAL_INSTANCED compile failed.
If dual object compile fails (as seems to happen with virgl a
fair bit, and does piglit even have any tests for it?), we end up
not restarting the pull params, so we call
vec4_visitor::move_uniform_array_access_to_pull_constant
a second time and it runs over the ends of the alloc.

Fixes: tests/spec/glsl-1.50/execution/geometry/max-input-components.shader_test
running inside virgl on ivybridge.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 271fa3a684)
2017-08-05 00:09:25 +01:00
Chris Wilson
795b712bd7 i965/blit: Remember to include miptree buffer offset in relocs
Remember to add the offset to the start of the buffer in the relocation
or else we write 0xff into random bytes elsewhere.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit fb63c43fd1)
2017-08-05 00:09:25 +01:00
Kenneth Graunke
43a2b178c2 i965: Delete pitch alignment assertion in get_blit_intratile_offset_el.
The cacheline alignment restriction is on the base address; the pitch
can be anything.

Fixes assertion failures when using primus (say, on glxgears, which
creates a 300x300 linear BGRX surface with a pitch of 1200):

intel_blit.c:190: get_blit_intratile_offset_el: Assertion `mt->surf.row_pitch % 64 == 0' failed.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
(cherry picked from commit 595a47b829)
2017-08-05 00:09:25 +01:00
Jason Ekstrand
d5def4f5a9 spirv: Fix SpvImageFormatR16ui
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: "17.1 17.2" <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 95c6a97464)
2017-08-05 00:09:25 +01:00
Thomas Hellstrom
f2a60ff20a dri3: Wait for all pending swapbuffers to be scheduled before touching the front
This implements a wait for glXWaitGL, glXCopySubBuffer, dri flush_front and
creation of fake front until all pending SwapBuffers have been committed to
hardware. Among other things this fixes piglit glx-copy-sub-buffers on dri3.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Cc: <mesa-stable@lists.freedesktop.org>
(cherry picked from commit 185ef06fd2)
2017-08-05 00:09:25 +01:00
Nicolai Hähnle
3c8673d420 gallium/radeon: fix ARB_query_buffer_object conversion to boolean
The issue here is that the immediate is treated as a 64-bit value,
and fetching it does not work reliably with swizzles that are different
from xy and zw.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit da83687c4b)
2017-08-05 00:09:25 +01:00
Dave Airlie
381ccaa1cb radeon/ac: use ds_swizzle for derivs on si/cik.
This looks like it's supported since llvm 3.9 at least,
so switch over radeonsi and radv to using it, -pro also
uses this. We can now drop creating lds for these operations
as the ds_swizzle operation doesn't actually write to lds at all.

Acked-by: Marek Olšák <marek.olsak@amd.com>
(stable requested due to fixing radv CIK conformance tests)
Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit cb6f16dce9)
[Emil Velikov: resolve trivial conflicts]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Conflicts:
	src/amd/common/ac_nir_to_llvm.c
2017-08-05 00:09:25 +01:00
Connor Abbott
2dd6030fbb ac/nir: fix lsb emission
This makes it match radeonsi. The LLVM backend itself will emit the
correct instruction, but LLVM might do incorrect optimizations since it
thinks the output is undefined when the input is 0, even though it's not
supposed to be. We really need a new intrinsic, or for the backend to
become smarter and recognize this pattern.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Bas Nieuwenhuizen <basni@google.com>
(cherry picked from commit 6d731c5651)
2017-08-05 00:09:25 +01:00
Connor Abbott
a90d99f7a5 nir: fix algebraic optimizations
The optimizations are only valid for 32-bit integers. They were
mistakenly firing for 64-bit integers as well.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Matt Turner <mattst88@gmail.com>
(cherry picked from commit de91461575)
2017-08-05 00:09:25 +01:00
Marek Olšák
6806773905 Revert "st/mesa: release sampler views when redefining a texture in st_context_teximage"
This reverts commit 5c1241268b.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101961

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
(cherry picked from commit d85802e501)
2017-08-05 00:09:25 +01:00
Nicolai Hähnle
3e872f4b53 radeonsi: ensure that temp array allocas are in the entry block
Otherwise, code generation fails. This has become necessary since some
shaders are wrapped in control flow.

Fixes: 081ac6e5c6 ("radeonsi/gfx9: always wrap GS and TCS in an if-block (v2)")
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 2879a602dd)
2017-08-05 00:09:25 +01:00
Samuel Pitoiset
21ea75b3e9 mesa: fix mismatch when returning 64-bit bindless uniform handles
The slower convert-and-copy process performs a bad conversion
because it converts the value to signed 64-bit integer, but
bindless uniform handles are considered unsigned 64-bit.

This fixes "Check glUniform*() with mixed texture units/handles"
from arb_bindless_texture-uniform piglit.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit b38c9c57f2)
2017-08-05 00:09:25 +01:00
Emil Velikov
58fe86a6d6 Update version to 17.2.0-rc2
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-07-31 10:52:13 +01:00
Samuel Pitoiset
d466a70532 st/glsl_to_tgsi: fix getting the image type for array of structs
Since array splitting for AoA is disabled, we have to retrieve
the type of the first non-array type when an array of images is
declared inside a structure. Otherwise, it will hit an assert
in glsl_type::sampler_index() because it expects either a sampler
or an image type.

This fixes a regression in the following piglit test:
arb_bindless_texture/compiler/images/arrays-of-struct.frag

Fixes: 57165f2ef8 ("glsl: disable array splitting for AoA")
Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit f99e9335e2)
2017-07-31 10:26:27 +01:00
Marek Olšák
e62eddcdbe st/mesa: release sampler views when redefining a texture in st_context_teximage
Noticed randomly.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit 5c1241268b)
2017-07-31 10:25:55 +01:00
Dave Airlie
6d07e58afb radv: for stencil only set Z tile mode index to same value
On SI this was causing a hang in
dEQP-VK.pipeline.render_to_image.core.2d_array.mipmap.r16g16_sint_s8_uint

This was due to not handling the tile mode index for depth like
I fixed previously for new GPUs.

Fixes: 01d0c5a9 (radv: fix stencil regression since new addrlib import)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 800d162209)
2017-07-31 10:24:42 +01:00
Dave Airlie
f9e563597d virgl: drop precise modifier.
The host doesn't understand this yet, so drop it for now.

Fixes: virgl regressions.

Fixes: af22adee4f (tgsi: add precise flag to tgsi_instruction)
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 554aa09440)
2017-07-31 10:23:09 +01:00
Marek Olšák
5b61ba4432 radeonsi: update dirty_level_mask only when flushing or unbinding framebuffer
This fixes corruption with bindless textures in Dawn Of War 3.

The do_update_surf_dirtiness mechanism was complicated and dirty_level_mask
was only updated after the first draw call. The problem is bindless textures
are checked for decompression every draw call and we would only decompress
after the first draw call. The solution is to set dirtiness after the last
draw call to the framebuffer, so the (unconditional) decompression of
bindless textures happens at the right time.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Tested-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit f4d095cc65)
2017-07-31 10:21:17 +01:00
Marek Olšák
6625382b1c st/mesa: always unconditionally revalidate main framebuffer after SwapBuffers
This fixes the black Feral launcher window.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101867

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Tested-by: Edmondo Tommasina <edmondo.tommasina@gmail.com>
(cherry picked from commit 7257c171e9)
2017-07-31 10:20:44 +01:00
Nicolai Hähnle
2bca74253d radeonsi/gfx9: always wrap GS and TCS in an if-block (v2)
With merged ESGS shaders, the GS part of a wave may be empty, and the
hardware gets confused if any GS messages are sent from that wave. Since
S_SENDMSG is executed even when EXEC = 0, we have to wrap even
non-monolithic GS shaders in an if-block, so that the entire shader and
hence the S_SENDMSG instructions are skipped in empty waves.

This change is not required for TCS/HS, but applying it there as well
simplifies the logic a bit.

Fixes GL45-CTS.geometry_shader.rendering.rendering.*

v2: ensure that the TCS epilog doesn't run for non-existing patches

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 081ac6e5c6)
2017-07-31 10:20:20 +01:00
Nicolai Hähnle
b36ff2d1f2 radeonsi/gfx9: fix vertex idx in ES with multiple waves per threadgroup
Cc: mesa-stable@lists.freedesktop.org
Reviewed: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 873789002f)
2017-07-31 10:20:12 +01:00
George Kyriazis
99b2613ce1 swr: fix transform feedback logic
The shader that is used to copy vertex data out of the vs/gs shaders to
the user-specified buffer (streamout or SO shader) was not using the
correct offsets.

Adjust the offsets that are used just for the SO shader:
- Make sure that position is handled in the same special way
  as in the vs/gs shaders
- Use the correct offset to be passed in the core
- consolidate register slot mapping logic into one function, since it's
  been calculated in 2 different places (one for calcuating the slot mask,
  and one for the register offsets themselves

Also make room for all attibutes in the backend vertex area.

Fixes:
- all vtk GL2PS tests
- 18 piglit tests (16 ext_transform_feedback tests,
  arb-quads-follow-provoking-vertex and primitive-type gl_points

v2:

- take care of more SGV slots in slot mapping logic
- trim feState.vsVertexSize
- fix GS interface and incorporate GS while calculating vsVertexSize

Note that vsVertexSize is used in the core as the one parameter that
controls vertex size between all stages, so it has to be adjusted appropriately
for the whole vs/gs/fs pipeline.

Also note that GS and SO is not fully implemented.  This will be addressed
later.

fixes:
- fixes total of 20 piglit tests

CC: 17.2 <mesa-stable@lists.freedesktop.org>

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
(cherry picked from commit 194ff5eed1)
2017-07-31 10:19:55 +01:00
Dave Airlie
f9c7549605 radv/ac: port SI TC L1 write corruption fix.
This ports 72e46c988 to radv.
    radeonsi: apply a TC L1 write corruption workaround for SI

Fixes: f4e499ec7 (radv: add initial non-conformant radv vulkan driver)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit e77ff11ffe)
2017-07-27 19:54:53 +01:00
Dave Airlie
2ce4f0afd3 radv/ac: realign SI workaround with radeonsi.
This ports: da7453666a
radeonsi: don't apply the Z export bug workaround to Hainan
to radv.

Just noticed in passing.

Fixes: f4e499ec7 (radv: add initial non-conformant radv vulkan driver)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit a81e99f50a)
2017-07-27 19:54:52 +01:00
Dave Airlie
546282e8bc radv: only report external semaphore info for opaque fd.
Until we support sync fd, don't report the info.

Fixes CTS dEQP-VK.api.external.semaphore.sync_fd.* from crashing.

Fixes: eaa56eab6 (radv: initial support for shared semaphores (v2))
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 6cbc8cf178)
2017-07-27 19:54:52 +01:00
Dave Airlie
de55bc8f49 radv: fix buffer views on SI/CIK.
Fixes CTS dEQP-VK.memory.pipeline_barrier.host_write_uniform_texel_buffer.1024
on SI/CIK with radv.

Fixes: f4e499ec (radv: add initial non-conformant radv vulkan driver)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit ca82ef5ac7)
2017-07-27 19:54:52 +01:00
Daniel Stone
f0b7563a36 st/dri2: Return invalid modifier when no driver support
Always initialise whandle.modifier for DRIImage modifier queries, so if
the driver doesn't support it then we return false for the query.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Fixes: d33fe8b84e ("st/dri: enable DRIimage modifier queries")
(cherry picked from commit 45383d32d4)
2017-07-27 19:54:52 +01:00
Daniel Stone
5bee196840 st/dri: Check get-handle return value in queryImage
In the DRIImage queryImage hook, check if resource_get_handle() failed
and return FALSE if so.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit b4a18f13ce)
2017-07-27 19:54:52 +01:00
Daniel Stone
2a1792981c egl/wayland: Ignore invalid modifiers
If the underlying driver does not support modifiers, dmabuf will still
advertise formats through the 'modifier' event, but send them with an
invalid modifier. Ignore them if this is the case, rather than passing
them through to the driver.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Fixes: 02cc359372 ("egl/wayland: Use linux-dmabuf interface for buffers")
(cherry picked from commit dd072cf4b1)
2017-07-27 19:54:52 +01:00
Tim Rowley
a2b7477603 swr/rast: non-regex knob fallback code for gcc < 4.9
gcc prior to 4.9 didn't implement <regex>, causing a startup crash
in the swr knob parameter reading code.

CC: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
(cherry picked from commit e21fc2c625)
2017-07-27 19:54:52 +01:00
Dave Airlie
3e777d5cab virgl: encode index buffer offset.
Fixes arb_vertex_buffer_object-combined-vertex-index

Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit c4652a0a5b)
2017-07-27 19:54:52 +01:00
Marek Olšák
9bb6aa5794 ac/surface: fix hybrid graphics where APU=GFX9, dGPU=older
v2: don't do it for compressed textures (bpp = 0)

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com> (v1)
(cherry picked from commit 5e81df0f10)
2017-07-27 19:54:52 +01:00
Marek Olšák
90bbcb93b1 radeonsi: decrease the number of compiler threads
Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit ed2b3f5c81)
2017-07-27 19:54:52 +01:00
Marek Olšák
529c440dd3 gallium/radeon: make S_FIXED function signed and move it to shared code
This fixes a bug uncovered by:
    2412c4c81e
    util: Make CLAMP turn NaN into MIN.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
(cherry picked from commit 433f6f7ac9)
2017-07-27 19:54:52 +01:00
Charmaine Lee
94e0de90ee st/mesa: create framebuffer iface hash table per st manager
With commit 5124bf9823, a framebuffer interface hash table is
created in st_gl_api_create(), which is called in
dri_init_screen_helper() for each screen. When the hash table is
overwritten with multiple calls to st_gl_api_create(), it can cause
race condition. This patch fixes the problem by creating a
framebuffer interface hash table per state tracker manager.

Fixes crash with steam.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101876
Fixes: 5124bf9823 ("st/mesa: add destroy_drawable interface")
Tested-by: Christoph Haag <haagch@frickel.club>
Reviewed-by: Brian Paul <brianp@vmware.com>
(cherry picked from commit bbc29393d3)

Squashed with:

st/mesa: fix unconditional return in st_framebuffer_iface_remove

Noticed by James Legg @ Feral.

Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
(cherry picked from commit 914f11e75b)

Squashed with:

st/mesa: Fix inversed test in st_api_destroy_drawable

Fixes a drawable leak.

Fixes: bbc29393d3 ("st/mesa: create framebuffer iface hash table per
                      st manager")
Bugzilla: https://bugs.freedesktop.org/101930
Tested-by: Nick Sarnie <commendsarnex@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
(cherry picked from commit 57132d126f)
2017-07-27 19:54:24 +01:00
Grigori Goronzy
3180f0fa0d egl: move KHR_no_error vs debug/robustness check further down
We'll fail to flag an error if the context flags appear after the
no-error attribute in the context attribute list.

Delay the check to after attribute parsing to fix this.

Fixes: 4909519a66 ("egl: Add EGL_KHR_create_context_no_error support")
Cc: mesa-stable@lists.freedesktop.org
[Emil Velikov: add fixes/stable tags, commit message polish]
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

(cherry picked from commit 39bf7756b9)
2017-07-27 18:56:45 +01:00
Nicolai Hähnle
e7f14a8b52 radeonsi/gfx9: reduce max threads per block to 1024 on gfx9+
The number of supported waves per thread group has been reduced to 16
with gfx9. Trying to use 32 waves causes hangs, and barriers might
not work correctly with > 16 waves.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit a0e6b9a2db)
2017-07-27 18:56:45 +01:00
Nicolai Hähnle
7f5d9d7a6d radeonsi: fix detection of DRAW_INDIRECT_MULTI on SI
The firmware version numbers for SI were wrong. The new numbers are probably
too conservative (we don't have a definitive answer by the firmware team),
but DRAW_INDIRECT_MULTI has been confirmed to work with these versions on
Tahiti (by Gustaw) and on Verde (by myself).

While this is technically adding a feature, it's a feature we thought we had
for a long time. The change is small enough and we're early enough in the 17.2
release cycle that it should still go in.

Reported-by: Gustaw Smolarczyk <wielkiegie@gmail.com>
Cc: 17.2 <mesa-stable@lists.freedesktop.org>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
(cherry picked from commit 65fbaab0b7)
2017-07-27 18:56:45 +01:00
Iago Toral Quiroga
3d0960e761 anv: only expose up to 28 vertex attributes
The EU limit of 128 GRFs should allow 32 vertex elements of 4 GRFs.
However, the maximum allowed value of "Vertex URB Entry Read Length"
in SIMD8 is 15. And 15 * 8 = 120 gives us a limit of 30 vertex elements.
Because we also need to reserve a vertex buffer to upload
VertexIndex/InstanceIndex and another to upload DrawID when needed,
we can only expose 28.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
(cherry picked from commit 31f1863ace)
2017-07-27 18:56:45 +01:00
Iago Toral Quiroga
bdbd8ab517 anv/cmd_buffer: fix off by one error in assertion
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
(cherry picked from commit a848e693ef)
2017-07-27 18:56:44 +01:00
Kenneth Graunke
47bca2cfa7 i965: Fix = vs == in MCS aux usage assert.
Caught by Coverity (CID 1415680).

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
(cherry picked from commit 698636cc97)
Fixes: 0f9b609cf4 ("i965/blorp: Do prepare/finish manually")
2017-07-27 18:54:41 +01:00
Kenneth Graunke
04bb687f04 i965: Fix offset addition in get_isl_surf.
Increase the value, not the pointer to the stack variable.

Caught by Coverity (CID 1415574).  Not shipped in a real release.

Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
(cherry picked from commit f6e674fa51)
Fixes: 63a43f4161 ("i965: Refactor miptree to isl converter and adjustment")
2017-07-27 18:54:13 +01:00
Eric Anholt
4e0f29ed0b broadcom/vc4: Prefer blit via rendering to the software fallback.
I don't know how I managed to leave this here for so long.  Found when
working on a 1:1 overlapping blit extension for X11.

Cc: mesa-stable@lists.freedesktop.org
(cherry picked from commit 93fec49a75)
2017-07-27 18:21:27 +01:00
Lionel Landwerlin
0ccb853cc0 i965: perf: flush batchbuffers at the beginning of queries
As Chris commented, it makes more sense to have batch buffer flushes
before the query. Usually applications like frame_retrace do a series
of queries and in that case, with flushes at the end of the queries,
we might still have the first query contained in 2 different batchs.
More generally it would be quite usual to have the query contained in
2 batch buffers because we never now what's the fill rate of the
current batch buffer.

If we move the flushing at the beginning of the queries, it's pretty
much guaranteed that queries will be contained in a single batch
buffer (unless the amount of commands is huge, but then it's only fair
to include reloading request times in the measurements).

Fixes: adafe4b733 ("i965: perf: minimize the chances to spread queries across batchbuffers")
Reported-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: "17.2 17.1" <mesa-stable@lists.freedesktop.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
(cherry picked from commit 9f439ae120)
2017-07-27 18:21:22 +01:00
Emil Velikov
a455f594bb Update version to 17.2.0-rc1
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
2017-07-24 16:59:32 +01:00
Emil Velikov
a955622c1a intel/blorp: ship blorp_genX_exec.h within the tarball
Fixes: c9cb37b2a6 ("intel/blorp: Add a partial resolve pass for MCS")
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit 5d47dd9c2a)
2017-07-24 16:59:32 +01:00
11662 changed files with 1189838 additions and 4323493 deletions

View File

View File

View File

View File

View File

View File

View File

View File

View File

View File

View File

@@ -1,2 +0,0 @@
# Vendored code
src/amd/vulkan/radix_sort/*

View File

@@ -1,9 +0,0 @@
# The following files are opted into `ninja clang-format` and
# enforcement in the CI.
src/gallium/drivers/i915
src/gallium/targets/teflon/**/*
src/amd/vulkan/**/*
src/amd/compiler/**/*
src/egl/**/*
src/etnaviv/isa/**/*

View File

@@ -8,37 +8,28 @@ charset = utf-8
insert_final_newline = true
tab_width = 8
[*.{c,h,cpp,hpp,cc,hh,y,yy}]
[*.{c,h,cpp,hpp,cc,hh}]
indent_style = space
indent_size = 3
max_line_length = 78
[{Makefile*,*.mk}]
indent_style = tab
[*.py]
[{*.py,SCons*}]
indent_style = space
indent_size = 4
[*.pl]
indent_style = space
indent_size = 4
[*.m4]
indent_style = space
indent_size = 2
[*.yml]
indent_style = space
indent_size = 2
[*.rst]
indent_style = space
indent_size = 3
[*.patch]
trim_trailing_whitespace = false
[{meson.build,meson_options.txt}]
indent_style = space
indent_size = 2
[*.ps1]
indent_style = space
indent_size = 2
[*.rs]
indent_style = space
indent_size = 4

View File

@@ -1,67 +0,0 @@
# List of commits to ignore when using `git blame`.
# Enable with:
# git config blame.ignoreRevsFile .git-blame-ignore-revs
#
# Per git-blame(1):
# Ignore revisions listed in the file, one unabbreviated object name
# per line, in git-blame. Whitespace and comments beginning with # are
# ignored.
#
# Please keep these in chronological order :)
#
# You can add a new commit with the following command:
# git log -1 --pretty=format:'%n# %s%n%H%n' >> .git-blame-ignore-revs $COMMIT
# pvr: Fix clang-format error.
0ad5b0a74ef73f5fcbe1406ad9d57fe5dc00a5b1
# panfrost: Fix up some formatting for clang-format
a4705afe63412498d13ded73cba969c66be67907
# asahi: clang-format the world again
26c51bb8d8a33098b1990425a391f56ffba5728c
# perfetto: Add a .clang-format for the directory.
da78d5d729b1800136dd713b68492cb339993f4a
# panfrost/winsys: Clang-format
c90f036516a5376002be6550a917e8bad6a8a3b8
# panfrost: Re-run clang-format
4ccf174009af6732cbffa5d8ebb4687da7517505
# panvk: Clang-format
c7bf3b69ebc8f2252dbf724a4de638e6bb2ac402
# pan/mdg: Fix icky formatting
133af0d6c945d3aaca8989edd15283a2b7dcc6c7
# mapi: clang-format _glapi_add_dispatch()
30332529663268a6406e910848e906e725e6fda7
# radv: reformat according to its .clang-format
8b319c6db8bd93603b18bd783eb75225fcfd51b7
# aco: reformat according to its .clang-format
6b21653ab4d3a67e711fe10e3d403128b6d26eb2
# egl: re-format using clang-format
2f670d89db038d5a29f6b72732fd7ad63dfaf4c6
# panfrost: clang-format the tree
0afd691f29683f6e9dde60f79eca094373521806
# aco: Format.
1e2639026fec7069806449f9ba2a124ce4eb5569
# radv: Format.
59c501ca353f8ec9d2717c98af2bfa1a1dbf4d75
# pvr: clang-format fixes
953c04ebd39c52d457301bdd8ac803949001da2d
# freedreno: Re-indent
2d439343ea1aee146d4ce32800992cd389bd505d
# ir3: Reformat source with clang-format
177138d8cb0b4f6a42ef0a1f8593e14d79f17c54

11
.gitattributes vendored
View File

@@ -1,7 +1,4 @@
*.csv eol=crlf
* text=auto
*.jpg binary
*.png binary
*.gif binary
*.ico binary
*.cl gitlab-language=c
*.dsp -crlf
*.dsw -crlf
*.sln -crlf
*.vcproj -crlf

View File

@@ -1,60 +0,0 @@
name: macOS-CI
on: push
permissions:
contents: read
jobs:
macOS-CI:
strategy:
matrix:
glx_option: ['dri', 'xlib']
runs-on: macos-11
env:
GALLIUM_DUMP_CPU: true
MESON_EXEC: /Users/runner/Library/Python/3.11/bin/meson
steps:
- name: Checkout
uses: actions/checkout@v3
- name: Install Dependencies
run: |
cat > Brewfile <<EOL
brew "bison"
brew "expat"
brew "gettext"
brew "libx11"
brew "libxcb"
brew "libxdamage"
brew "libxext"
brew "molten-vk"
brew "ninja"
brew "pkg-config"
brew "python@3.10"
EOL
brew update
brew bundle --verbose
- name: Install Mako and meson
run: pip3 install --user mako meson
- name: Configure
run: |
cat > native_config <<EOL
[binaries]
llvm-config = '/usr/local/opt/llvm/bin/llvm-config'
EOL
$MESON_EXEC . build --native-file=native_config -Dmoltenvk-dir=$(brew --prefix molten-vk) -Dbuild-tests=true -Dosmesa=true -Dgallium-drivers=swrast,zink -Dglx=${{ matrix.glx_option }}
- name: Build
run: $MESON_EXEC compile -C build
- name: Test
run: $MESON_EXEC test -C build --print-errorlogs
- name: Install
run: $MESON_EXEC install -C build --destdir $PWD/install
- name: 'Upload Artifact'
if: always()
uses: actions/upload-artifact@v3
with:
name: macos-${{ matrix.glx_option }}-result
path: |
build/meson-logs/
install/
retention-days: 5

54
.gitignore vendored
View File

@@ -1,6 +1,52 @@
.vscode*
*.a
*.dll
*.exe
*.ilk
*.la
*.lo
*.log
*.o
*.obj
*.os
*.pc
*.pdb
*.pyc
*.pyo
*.out
/build
.venv/
*.so
*.so.*
*.sw[a-z]
*.tar
*.tar.bz2
*.tar.gz
*.tar.xz
*.trs
*.zip
*~
depend
depend.bak
bin/ltmain.sh
lib
lib64
configure
configure.lineno
autom4te.cache
aclocal.m4
config.log
config.status
cscope*
tags
.scon*
config.py
build
libtool
manifest.txt
.dir-locals.el
.deps/
.dirstamp
.libs/
Makefile
Makefile.in
.install-mesa-links
.install-gallium-links
/src/git_sha1.h
TAGS

View File

@@ -1,319 +0,0 @@
# Types of CI pipelines:
# | pipeline name | context | description |
# |----------------------|-----------|-------------------------------------------------------------|
# | merge pipeline | mesa/mesa | pipeline running for an MR; if it passes the MR gets merged |
# | pre-merge pipeline | mesa/mesa | same as above, except its status doesn't affect the MR |
# | post-merge pipeline | mesa/mesa | pipeline immediately after merging |
# | fork pipeline | fork | pipeline running in a user fork |
# | scheduled pipeline | mesa/mesa | nightly pipelines, running every morning at 4am UTC |
# | direct-push pipeline | mesa/mesa | when commits are pushed directly to mesa/mesa, bypassing Marge and its gating pipeline |
#
# Note that the release branches maintained by the release manager fall under
# the "direct push" category.
#
# "context" indicates the permissions that the jobs get; notably, any
# container created in mesa/mesa gets pushed immediately for everyone to use
# as soon as the image tag change is merged.
#
# Merge pipelines contain all jobs that must pass before the MR can be merged.
# Pre-merge pipelines contain the exact same jobs as merge pipelines.
# Post-merge pipelines contain *only* the `pages` job that deploys the new
# version of the website.
# Fork pipelines contain everything.
# Scheduled pipelines only contain the container+build jobs, and some extra
# test jobs (typically "full" variants of pre-merge jobs that only run 1/X
# test cases), but not a repeat of the merge pipeline jobs.
# Direct-push pipelines contain the same jobs as merge pipelines.
workflow:
rules:
# do not duplicate pipelines on merge pipelines
- if: $CI_COMMIT_BRANCH && $CI_OPEN_MERGE_REQUESTS && $CI_PIPELINE_SOURCE == "push"
when: never
# merge pipeline
- if: &is-merge-attempt $GITLAB_USER_LOGIN == "marge-bot" && $CI_PIPELINE_SOURCE == "merge_request_event"
variables:
KERNEL_IMAGE_BASE: https://${S3_HOST}/mesa-lava/${KERNEL_REPO}/${KERNEL_TAG}
MESA_CI_PERFORMANCE_ENABLED: 1
VALVE_INFRA_VANGOGH_JOB_PRIORITY: "" # Empty tags are ignored by gitlab
# post-merge pipeline
- if: &is-post-merge $GITLAB_USER_LOGIN == "marge-bot" && $CI_PIPELINE_SOURCE == "push"
# nightly pipeline
- if: &is-scheduled-pipeline $CI_PIPELINE_SOURCE == "schedule"
variables:
KERNEL_IMAGE_BASE: https://${S3_HOST}/mesa-lava/${KERNEL_REPO}/${KERNEL_TAG}
JOB_PRIORITY: 50
VALVE_INFRA_VANGOGH_JOB_PRIORITY: priority:low
# pipeline for direct pushes that bypassed the CI
- if: &is-direct-push $CI_PROJECT_NAMESPACE == "mesa" && $CI_PIPELINE_SOURCE == "push" && $GITLAB_USER_LOGIN != "marge-bot"
variables:
KERNEL_IMAGE_BASE: https://${S3_HOST}/mesa-lava/${KERNEL_REPO}/${KERNEL_TAG}
JOB_PRIORITY: 40
VALVE_INFRA_VANGOGH_JOB_PRIORITY: priority:low
# pre-merge or fork pipeline
- if: $FORCE_KERNEL_TAG != null
variables:
KERNEL_IMAGE_BASE: https://${S3_HOST}/mesa-lava/${KERNEL_REPO}/${FORCE_KERNEL_TAG}
JOB_PRIORITY: 50
VALVE_INFRA_VANGOGH_JOB_PRIORITY: priority:low
- if: $FORCE_KERNEL_TAG == null
variables:
KERNEL_IMAGE_BASE: https://${S3_HOST}/mesa-lava/${KERNEL_REPO}/${KERNEL_TAG}
JOB_PRIORITY: 50
VALVE_INFRA_VANGOGH_JOB_PRIORITY: priority:low
variables:
FDO_UPSTREAM_REPO: mesa/mesa
MESA_TEMPLATES_COMMIT: &ci-templates-commit d5aa3941aa03c2f716595116354fb81eb8012acb
CI_PRE_CLONE_SCRIPT: |-
set -o xtrace
wget -q -O download-git-cache.sh ${CI_PROJECT_URL}/-/raw/${CI_COMMIT_SHA}/.gitlab-ci/download-git-cache.sh
bash download-git-cache.sh
rm download-git-cache.sh
set +o xtrace
CI_JOB_JWT_FILE: /minio_jwt
S3_HOST: s3.freedesktop.org
# per-pipeline artifact storage on MinIO
PIPELINE_ARTIFACTS_BASE: ${S3_HOST}/artifacts/${CI_PROJECT_PATH}/${CI_PIPELINE_ID}
# per-job artifact storage on MinIO
JOB_ARTIFACTS_BASE: ${PIPELINE_ARTIFACTS_BASE}/${CI_JOB_ID}
# reference images stored for traces
PIGLIT_REPLAY_REFERENCE_IMAGES_BASE: "${S3_HOST}/mesa-tracie-results/$FDO_UPSTREAM_REPO"
# For individual CI farm status see .ci-farms folder
# Disable farm with `git mv .ci-farms{,-disabled}/$farm_name`
# Re-enable farm with `git mv .ci-farms{-disabled,}/$farm_name`
# NEVER MIX FARM MAINTENANCE WITH ANY OTHER CHANGE IN THE SAME MERGE REQUEST!
ARTIFACTS_BASE_URL: https://${CI_PROJECT_ROOT_NAMESPACE}.${CI_PAGES_DOMAIN}/-/${CI_PROJECT_NAME}/-/jobs/${CI_JOB_ID}/artifacts
# Python scripts for structured logger
PYTHONPATH: "$PYTHONPATH:$CI_PROJECT_DIR/install"
default:
before_script:
- >
export SCRIPTS_DIR=$(mktemp -d) &&
curl -L -s --retry 4 -f --retry-all-errors --retry-delay 60 -O --output-dir "${SCRIPTS_DIR}" "${CI_PROJECT_URL}/-/raw/${CI_COMMIT_SHA}/.gitlab-ci/setup-test-env.sh" &&
. ${SCRIPTS_DIR}/setup-test-env.sh &&
echo -n "${CI_JOB_JWT}" > "${CI_JOB_JWT_FILE}" &&
unset CI_JOB_JWT # Unsetting vulnerable env variables
after_script:
# Work around https://gitlab.com/gitlab-org/gitlab/-/issues/20338
- find -name '*.log' -exec mv {} {}.txt \;
- >
set +x
test -e "${CI_JOB_JWT_FILE}" &&
export CI_JOB_JWT="$(<${CI_JOB_JWT_FILE})" &&
rm "${CI_JOB_JWT_FILE}"
# Retry when job fails. Failed jobs can be found in the Mesa CI Daily Reports:
# https://gitlab.freedesktop.org/mesa/mesa/-/issues/?sort=created_date&state=opened&label_name%5B%5D=CI%20daily
retry:
max: 1
# Ignore runner_unsupported, stale_schedule, archived_failure, or
# unmet_prerequisites
when:
- api_failure
- runner_system_failure
- script_failure
- job_execution_timeout
- scheduler_failure
- data_integrity_failure
- unknown_failure
stages:
- sanity
- container
- git-archive
- build-x86_64
- build-misc
- code-validation
- amd
- intel
- nouveau
- arm
- broadcom
- freedreno
- etnaviv
- software-renderer
- layered-backends
- deploy
include:
- project: 'freedesktop/ci-templates'
ref: 16bc29078de5e0a067ff84a1a199a3760d3b3811
file:
- '/templates/ci-fairy.yml'
- project: 'freedesktop/ci-templates'
ref: *ci-templates-commit
file:
- '/templates/alpine.yml'
- '/templates/debian.yml'
- '/templates/fedora.yml'
- local: '.gitlab-ci/image-tags.yml'
- local: '.gitlab-ci/lava/lava-gitlab-ci.yml'
- local: '.gitlab-ci/container/gitlab-ci.yml'
- local: '.gitlab-ci/build/gitlab-ci.yml'
- local: '.gitlab-ci/test/gitlab-ci.yml'
- local: '.gitlab-ci/farm-rules.yml'
- local: '.gitlab-ci/test-source-dep.yml'
- local: 'docs/gitlab-ci.yml'
- local: 'src/**/ci/gitlab-ci.yml'
# YAML anchors for rule conditions
# --------------------------------
.rules-anchors:
# Pre-merge pipeline
- &is-pre-merge '$CI_PIPELINE_SOURCE == "merge_request_event"'
.never-post-merge-rules:
rules:
- if: *is-post-merge
when: never
.container+build-rules:
rules:
# Run when re-enabling a disabled farm, but not when disabling it
- !reference [.disable-farm-mr-rules, rules]
# Never run immediately after merging, as we just ran everything
- !reference [.never-post-merge-rules, rules]
# Build everything in merge pipelines, if any files affecting the pipeline
# were changed
- if: *is-merge-attempt
changes: &all_paths
- VERSION
- bin/git_sha1_gen.py
- bin/install_megadrivers.py
- bin/symbols-check.py
# GitLab CI
- .gitlab-ci.yml
- .gitlab-ci/**/*
- .ci-farms/*
# Meson
- meson*
- build-support/**/*
- subprojects/**/*
# clang format
- .clang-format
- .clang-format-include
- .clang-format-ignore
# Source code
- include/**/*
- src/**/*
when: on_success
# Same as above, but for pre-merge pipelines
- if: *is-pre-merge
changes:
*all_paths
when: manual
# Skip everything for pre-merge and merge pipelines which don't change
# anything in the build
- if: *is-merge-attempt
when: never
- if: *is-pre-merge
when: never
# Build everything after someone bypassed the CI
- if: *is-direct-push
when: on_success
# Build everything in scheduled pipelines
- if: *is-scheduled-pipeline
when: on_success
# Allow building everything in fork pipelines, but build nothing unless
# manually triggered
- when: manual
.ci-deqp-artifacts:
artifacts:
name: "mesa_${CI_JOB_NAME}"
when: always
untracked: false
paths:
# Watch out! Artifacts are relative to the build dir.
# https://gitlab.com/gitlab-org/gitlab-ce/commit/8788fb925706cad594adf6917a6c5f6587dd1521
- artifacts
- _build/meson-logs/*.txt
- _build/meson-logs/strace
# Git archive
make git archive:
extends:
- .fdo.ci-fairy
stage: git-archive
rules:
- !reference [.scheduled_pipeline-rules, rules]
# ensure we are running on packet
tags:
- packet.net
script:
# Compactify the .git directory
- git gc --aggressive
# Download & cache the perfetto subproject as well.
- rm -rf subprojects/perfetto ; mkdir -p subprojects/perfetto && curl https://android.googlesource.com/platform/external/perfetto/+archive/$(grep 'revision =' subprojects/perfetto.wrap | cut -d ' ' -f3).tar.gz | tar zxf - -C subprojects/perfetto
# compress the current folder
- tar -cvzf ../$CI_PROJECT_NAME.tar.gz .
- ci-fairy s3cp --token-file "${CI_JOB_JWT_FILE}" ../$CI_PROJECT_NAME.tar.gz https://$S3_HOST/git-cache/$CI_PROJECT_NAMESPACE/$CI_PROJECT_NAME/$CI_PROJECT_NAME.tar.gz
# Sanity checks of MR settings and commit logs
sanity:
extends:
- .fdo.ci-fairy
stage: sanity
rules:
- if: *is-pre-merge
when: on_success
- when: never
variables:
GIT_STRATEGY: none
script:
# ci-fairy check-commits --junit-xml=check-commits.xml
- ci-fairy check-merge-request --require-allow-collaboration --junit-xml=check-merge-request.xml
- |
set -eu
image_tags=(
DEBIAN_BASE_TAG
DEBIAN_BUILD_TAG
DEBIAN_X86_64_TEST_ANDROID_TAG
DEBIAN_X86_64_TEST_GL_TAG
DEBIAN_X86_64_TEST_VK_TAG
ALPINE_X86_64_BUILD_TAG
ALPINE_X86_64_LAVA_SSH_TAG
FEDORA_X86_64_BUILD_TAG
KERNEL_ROOTFS_TAG
KERNEL_TAG
PKG_REPO_REV
WINDOWS_X64_MSVC_TAG
WINDOWS_X64_BUILD_TAG
WINDOWS_X64_TEST_TAG
)
for var in "${image_tags[@]}"
do
if [ "$(echo -n "${!var}" | wc -c)" -gt 20 ]
then
echo "$var is too long; please make sure it is at most 20 chars."
exit 1
fi
done
artifacts:
when: on_failure
reports:
junit: check-*.xml
tags:
- placeholder-job
# Jobs that need to pass before spending hardware resources on further testing
.required-for-hardware-jobs:
needs:
- job: clang-format
optional: true
- job: rustfmt
optional: true

View File

@@ -1,82 +0,0 @@
# Note: skips lists for CI are just a list of lines that, when
# non-zero-length and not starting with '#', will regex match to
# delete lines from the test list. Be careful.
# This test checks the driver's reported conformance version against the
# version of the CTS we're running. This check fails every few months
# and everyone has to go and bump the number in every driver.
# Running this check only makes sense while preparing a conformance
# submission, so skip it in the regular CI.
dEQP-VK.api.driver_properties.conformance_version
# Exclude this test which might fail when a new extension is implemented.
dEQP-VK.info.device_extensions
# These are tremendously slow (pushing toward a minute), and aren't
# reliable to be run in parallel with other tests due to CPU-side timing.
dEQP-GLES[0-9]*.functional.flush_finish.*
# piglit: WGL is Windows-only
wgl@.*
# These are sensitive to CPU timing, and would need to be run in isolation
# on the system rather than in parallel with other tests.
glx@glx_arb_sync_control@timing.*
# This test is not built with waffle, while we do build tests with waffle
spec@!opengl 1.1@windowoverlap
# These tests all read from the front buffer after a swap. Given that we
# run piglit tests in parallel in Mesa CI, and don't have a compositor
# running, the frontbuffer reads may end up with undefined results from
# windows overlapping us.
#
# Piglit does mark these tests as not to be run in parallel, but deqp-runner
# doesn't respect that. We need to extend deqp-runner to allow some tests to be
# marked as single-threaded and run after the rayon loop if we want to support
# them.
#
# Note that "glx-" tests don't appear in x11-skips.txt because they can be
# run even if PIGLIT_PLATFORM=gbm (for example)
glx@glx-copy-sub-buffer.*
# A majority of the tests introduced in CTS 1.3.7.0 are experiencing failures and flakes.
# Disable these tests until someone with a more deeper understanding of EGL examines them.
#
# Note: on sc8280xp/a690 I get identical results (same passes and fails)
# between freedreno, zink, and llvmpipe, so I believe this is either a
# deqp bug or egl/wayland bug, rather than driver issue.
#
# With llvmpipe, the failing tests have the error message:
#
# "Illegal sampler view creation without bind flag"
#
# which might be a hint. (But some passing tests also have the same
# error message.)
#
# more context from David Heidelberg on IRC: the deqp commit where these
# started failing is: https://github.com/KhronosGroup/VK-GL-CTS/commit/79b25659bcbced0cfc2c3fe318951c585f682abe
# prior to that they were skipping.
wayland-dEQP-EGL.functional.color_clears.single_context.gles1.other
wayland-dEQP-EGL.functional.color_clears.single_context.gles2.other
wayland-dEQP-EGL.functional.color_clears.single_context.gles3.other
wayland-dEQP-EGL.functional.color_clears.multi_context.gles1.other
wayland-dEQP-EGL.functional.color_clears.multi_context.gles2.other
wayland-dEQP-EGL.functional.color_clears.multi_context.gles3.other
wayland-dEQP-EGL.functional.color_clears.multi_context.gles1_gles2.other
wayland-dEQP-EGL.functional.color_clears.multi_context.gles1_gles2_gles3.other
wayland-dEQP-EGL.functional.color_clears.multi_thread.gles1.other
wayland-dEQP-EGL.functional.color_clears.multi_thread.gles2.other
wayland-dEQP-EGL.functional.color_clears.multi_thread.gles3.other
wayland-dEQP-EGL.functional.color_clears.multi_thread.gles1_gles2.other
wayland-dEQP-EGL.functional.color_clears.multi_thread.gles1_gles2_gles3.other
# Seems to be the same is as wayland-dEQP-EGL.functional.color_clears.*
wayland-dEQP-EGL.functional.render.single_context.gles2.other
wayland-dEQP-EGL.functional.render.single_context.gles3.other
wayland-dEQP-EGL.functional.render.multi_context.gles2.other
wayland-dEQP-EGL.functional.render.multi_context.gles3.other
wayland-dEQP-EGL.functional.render.multi_context.gles2_gles3.other
wayland-dEQP-EGL.functional.render.multi_thread.gles2.other
wayland-dEQP-EGL.functional.render.multi_thread.gles3.other
wayland-dEQP-EGL.functional.render.multi_thread.gles2_gles3.other

View File

@@ -1,63 +0,0 @@
version: 1
# Rules to match for a machine to qualify
target:
id: '{{ ci_runner_id }}'
timeouts:
first_console_activity: # This limits the time it can take to receive the first console log
minutes: {{ timeout_first_minutes }}
retries: {{ timeout_first_retries }}
console_activity: # Reset every time we receive a message from the logs
minutes: {{ timeout_minutes }}
retries: {{ timeout_retries }}
boot_cycle:
minutes: {{ timeout_boot_minutes }}
retries: {{ timeout_boot_retries }}
overall: # Maximum time the job can take, not overrideable by the "continue" deployment
minutes: {{ timeout_overall_minutes }}
retries: 0
# no retries possible here
console_patterns:
session_end:
regex: >-
{{ session_end_regex }}
{% if session_reboot_regex %}
session_reboot:
regex: >-
{{ session_reboot_regex }}
{% endif %}
job_success:
regex: >-
{{ job_success_regex }}
job_warn:
regex: >-
{{ job_warn_regex }}
# Environment to deploy
deployment:
# Initial boot
start:
kernel:
url: '{{ kernel_url }}'
cmdline: >
SALAD.machine_id={{ '{{' }} machine_id }}
console={{ '{{' }} local_tty_device }},115200 earlyprintk=vga,keep
loglevel={{ log_level }} no_hash_pointers
b2c.service="--privileged --tls-verify=false --pid=host docker://{{ '{{' }} fdo_proxy_registry }}/gfx-ci/ci-tron/telegraf:latest" b2c.hostname=dut-{{ '{{' }} machine.full_name }}
b2c.container="-ti --tls-verify=false docker://{{ '{{' }} fdo_proxy_registry }}/gfx-ci/ci-tron/machine-registration:latest check"
b2c.ntp_peer=10.42.0.1 b2c.pipefail b2c.cache_device=auto b2c.poweroff_delay={{ poweroff_delay }}
b2c.minio="gateway,{{ '{{' }} minio_url }},{{ '{{' }} job_bucket_access_key }},{{ '{{' }} job_bucket_secret_key }}"
b2c.volume="{{ '{{' }} job_bucket }}-results,mirror=gateway/{{ '{{' }} job_bucket }},pull_on=pipeline_start,push_on=changes,overwrite{% for excl in job_volume_exclusions %},exclude={{ excl }}{% endfor %},remove,expiration=pipeline_end,preserve"
{% for volume in volumes %}
b2c.volume={{ volume }}
{% endfor %}
b2c.container="-v {{ '{{' }} job_bucket }}-results:{{ working_dir }} -w {{ working_dir }} {% for mount_volume in mount_volumes %} -v {{ mount_volume }}{% endfor %} --tls-verify=false docker://{{ local_container }} {{ container_cmd }}"
{% if kernel_cmdline_extras is defined %}
{{ kernel_cmdline_extras }}
{% endif %}
initramfs:
url: '{{ initramfs_url }}'

View File

@@ -1,55 +0,0 @@
#!/usr/bin/env python3
# Copyright © 2022 Valve Corporation
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
# IN THE SOFTWARE.
from jinja2 import Environment, FileSystemLoader
from os import environ, path
# Pass all the environment variables prefixed by B2C_
values = {
key.removeprefix("B2C_").lower(): environ[key]
for key in environ if key.startswith("B2C_")
}
env = Environment(loader=FileSystemLoader(path.dirname(values['job_template'])),
trim_blocks=True, lstrip_blocks=True)
template = env.get_template(path.basename(values['job_template']))
values['ci_job_id'] = environ['CI_JOB_ID']
values['ci_runner_id'] = environ['CI_RUNNER_ID']
values['job_volume_exclusions'] = [excl for excl in values['job_volume_exclusions'].split(",") if excl]
values['working_dir'] = environ['CI_PROJECT_DIR']
# Use the gateway's pull-through registry caches to reduce load on fd.o.
values['local_container'] = environ['IMAGE_UNDER_TEST']
values['local_container'] = values['local_container'].replace(
'registry.freedesktop.org',
'{{ fdo_proxy_registry }}'
)
if 'kernel_cmdline_extras' not in values:
values['kernel_cmdline_extras'] = ''
with open(path.splitext(path.basename(values['job_template']))[0], "w") as f:
f.write(template.render(values))

View File

@@ -1,2 +0,0 @@
[*.sh]
indent_size = 2

View File

@@ -1,13 +0,0 @@
#!/bin/sh
# Init entrypoint for bare-metal devices; calls common init code.
# First stage: very basic setup to bring up network and /dev etc
/init-stage1.sh
# Second stage: run jobs
test $? -eq 0 && /init-stage2.sh
# Wait until the job would have timed out anyway, so we don't spew a "init
# exited" panic.
sleep 6000

View File

@@ -1,17 +0,0 @@
#!/bin/bash
# shellcheck disable=SC2086 # we want word splitting
if [ -z "$BM_POE_INTERFACE" ]; then
echo "Must supply the PoE Interface to power down"
exit 1
fi
if [ -z "$BM_POE_ADDRESS" ]; then
echo "Must supply the PoE Switch host"
exit 1
fi
SNMP_KEY="1.3.6.1.4.1.9.9.402.1.2.1.1.1.$BM_POE_INTERFACE"
SNMP_OFF="i 4"
snmpset -v2c -r 3 -t 30 -cmesaci "$BM_POE_ADDRESS" "$SNMP_KEY" $SNMP_OFF

View File

@@ -1,22 +0,0 @@
#!/bin/bash
# shellcheck disable=SC2086 # we want word splitting
if [ -z "$BM_POE_INTERFACE" ]; then
echo "Must supply the PoE Interface to power up"
exit 1
fi
if [ -z "$BM_POE_ADDRESS" ]; then
echo "Must supply the PoE Switch host"
exit 1
fi
set -ex
SNMP_KEY="1.3.6.1.4.1.9.9.402.1.2.1.1.1.$BM_POE_INTERFACE"
SNMP_ON="i 1"
SNMP_OFF="i 4"
snmpset -v2c -r 3 -t 10 -cmesaci "$BM_POE_ADDRESS" "$SNMP_KEY" $SNMP_OFF
sleep 3s
snmpset -v2c -r 3 -t 10 -cmesaci "$BM_POE_ADDRESS" "$SNMP_KEY" $SNMP_ON

View File

@@ -1,124 +0,0 @@
#!/bin/bash
# shellcheck disable=SC1091 # The relative paths in this file only become valid at runtime.
# shellcheck disable=SC2034
# shellcheck disable=SC2086 # we want word splitting
# Boot script for Chrome OS devices attached to a servo debug connector, using
# NFS and TFTP to boot.
# We're run from the root of the repo, make a helper var for our paths
BM=$CI_PROJECT_DIR/install/bare-metal
CI_COMMON=$CI_PROJECT_DIR/install/common
CI_INSTALL=$CI_PROJECT_DIR/install
# Runner config checks
if [ -z "$BM_SERIAL" ]; then
echo "Must set BM_SERIAL in your gitlab-runner config.toml [[runners]] environment"
echo "This is the CPU serial device."
exit 1
fi
if [ -z "$BM_SERIAL_EC" ]; then
echo "Must set BM_SERIAL in your gitlab-runner config.toml [[runners]] environment"
echo "This is the EC serial device for controlling board power"
exit 1
fi
if [ ! -d /nfs ]; then
echo "NFS rootfs directory needs to be mounted at /nfs by the gitlab runner"
exit 1
fi
if [ ! -d /tftp ]; then
echo "TFTP directory for this board needs to be mounted at /tftp by the gitlab runner"
exit 1
fi
# job config checks
if [ -z "$BM_KERNEL" ]; then
echo "Must set BM_KERNEL to your board's kernel FIT image"
exit 1
fi
if [ -z "$BM_ROOTFS" ]; then
echo "Must set BM_ROOTFS to your board's rootfs directory in the job's variables"
exit 1
fi
if [ -z "$BM_CMDLINE" ]; then
echo "Must set BM_CMDLINE to your board's kernel command line arguments"
exit 1
fi
set -ex
# Clear out any previous run's artifacts.
rm -rf results/
mkdir -p results
# Create the rootfs in the NFS directory. rm to make sure it's in a pristine
# state, since it's volume-mounted on the host.
rsync -a --delete $BM_ROOTFS/ /nfs/
mkdir -p /nfs/results
. $BM/rootfs-setup.sh /nfs
# Put the kernel/dtb image and the boot command line in the tftp directory for
# the board to find. For normal Mesa development, we build the kernel and
# store it in the docker container that this script is running in.
#
# However, container builds are expensive, so when you're hacking on the
# kernel, it's nice to be able to skip the half hour container build and plus
# moving that container to the runner. So, if BM_KERNEL is a URL, fetch it
# instead of looking in the container. Note that the kernel build should be
# the output of:
#
# make Image.lzma
#
# mkimage \
# -A arm64 \
# -f auto \
# -C lzma \
# -d arch/arm64/boot/Image.lzma \
# -b arch/arm64/boot/dts/qcom/sdm845-cheza-r3.dtb \
# cheza-image.img
rm -rf /tftp/*
if echo "$BM_KERNEL" | grep -q http; then
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
$BM_KERNEL -o /tftp/vmlinuz
elif [ -n "${FORCE_KERNEL_TAG}" ]; then
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/${BM_KERNEL}" -o /tftp/vmlinuz
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/modules.tar.zst" -o modules.tar.zst
tar --keep-directory-symlink --zstd -xf modules.tar.zst -C "/nfs/"
rm modules.tar.zst &
else
cp /baremetal-files/"$BM_KERNEL" /tftp/vmlinuz
fi
echo "$BM_CMDLINE" > /tftp/cmdline
set +e
STRUCTURED_LOG_FILE=job_detail.json
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --update dut_job_type "${DEVICE_TYPE}"
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --update farm "${FARM}"
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --create-dut-job dut_name "${CI_RUNNER_DESCRIPTION}"
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --update-dut-time submit "${CI_JOB_STARTED_AT}"
python3 $BM/cros_servo_run.py \
--cpu $BM_SERIAL \
--ec $BM_SERIAL_EC \
--test-timeout ${TEST_PHASE_TIMEOUT:-20}
ret=$?
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --close-dut-job
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --close
set -e
# Bring artifacts back from the NFS dir to the build dir where gitlab-runner
# will look for them.
cp -Rp /nfs/results/. results/
if [ -f "${STRUCTURED_LOG_FILE}" ]; then
cp -p ${STRUCTURED_LOG_FILE} results/
echo "Structured log file is available at https://${CI_PROJECT_ROOT_NAMESPACE}.pages.freedesktop.org/-/${CI_PROJECT_NAME}/-/jobs/${CI_JOB_ID}/artifacts/results/${STRUCTURED_LOG_FILE}"
fi
exit $ret

View File

@@ -1,168 +0,0 @@
#!/usr/bin/env python3
#
# Copyright © 2020 Google LLC
# SPDX-License-Identifier: MIT
import argparse
import re
import sys
from custom_logger import CustomLogger
from serial_buffer import SerialBuffer
class CrosServoRun:
def __init__(self, cpu, ec, test_timeout, logger):
self.cpu_ser = SerialBuffer(
cpu, "results/serial.txt", "R SERIAL-CPU> ")
# Merge the EC serial into the cpu_ser's line stream so that we can
# effectively poll on both at the same time and not have to worry about
self.ec_ser = SerialBuffer(
ec, "results/serial-ec.txt", "R SERIAL-EC> ", line_queue=self.cpu_ser.line_queue)
self.test_timeout = test_timeout
self.logger = logger
def close(self):
self.ec_ser.close()
self.cpu_ser.close()
def ec_write(self, s):
print("W SERIAL-EC> %s" % s)
self.ec_ser.serial.write(s.encode())
def cpu_write(self, s):
print("W SERIAL-CPU> %s" % s)
self.cpu_ser.serial.write(s.encode())
def print_error(self, message):
RED = '\033[0;31m'
NO_COLOR = '\033[0m'
print(RED + message + NO_COLOR)
self.logger.update_status_fail(message)
def run(self):
# Flush any partial commands in the EC's prompt, then ask for a reboot.
self.ec_write("\n")
self.ec_write("reboot\n")
bootloader_done = False
self.logger.create_job_phase("boot")
tftp_failures = 0
# This is emitted right when the bootloader pauses to check for input.
# Emit a ^N character to request network boot, because we don't have a
# direct-to-netboot firmware on cheza.
for line in self.cpu_ser.lines(timeout=120, phase="bootloader"):
if re.search("load_archive: loading locale_en.bin", line):
self.cpu_write("\016")
bootloader_done = True
break
# The Cheza firmware seems to occasionally get stuck looping in
# this error state during TFTP booting, possibly based on amount of
# network traffic around it, but it'll usually recover after a
# reboot. Currently mostly visible on google-freedreno-cheza-14.
if re.search("R8152: Bulk read error 0xffffffbf", line):
tftp_failures += 1
if tftp_failures >= 10:
self.print_error(
"Detected intermittent tftp failure, restarting run.")
return 1
# If the board has a netboot firmware and we made it to booting the
# kernel, proceed to processing of the test run.
if re.search("Booting Linux", line):
bootloader_done = True
break
# The Cheza boards have issues with failing to bring up power to
# the system sometimes, possibly dependent on ambient temperature
# in the farm.
if re.search("POWER_GOOD not seen in time", line):
self.print_error(
"Detected intermittent poweron failure, abandoning run.")
return 1
if not bootloader_done:
self.print_error("Failed to make it through bootloader, abandoning run.")
return 1
self.logger.create_job_phase("test")
for line in self.cpu_ser.lines(timeout=self.test_timeout, phase="test"):
if re.search("---. end Kernel panic", line):
return 1
# There are very infrequent bus errors during power management transitions
# on cheza, which we don't expect to be the case on future boards.
if re.search("Kernel panic - not syncing: Asynchronous SError Interrupt", line):
self.print_error(
"Detected cheza power management bus error, abandoning run.")
return 1
# If the network device dies, it's probably not graphics's fault, just try again.
if re.search("NETDEV WATCHDOG", line):
self.print_error(
"Detected network device failure, abandoning run.")
return 1
# These HFI response errors started appearing with the introduction
# of piglit runs. CosmicPenguin says:
#
# "message ID 106 isn't a thing, so likely what happened is that we
# got confused when parsing the HFI queue. If it happened on only
# one run, then memory corruption could be a possible clue"
#
# Given that it seems to trigger randomly near a GPU fault and then
# break many tests after that, just restart the whole run.
if re.search("a6xx_hfi_send_msg.*Unexpected message id .* on the response queue", line):
self.print_error(
"Detected cheza power management bus error, abandoning run.")
return 1
if re.search("coreboot.*bootblock starting", line):
self.print_error(
"Detected spontaneous reboot, abandoning run.")
return 1
if re.search("arm-smmu 5040000.iommu: TLB sync timed out -- SMMU may be deadlocked", line):
self.print_error("Detected cheza MMU fail, abandoning run.")
return 1
result = re.search("hwci: mesa: (\S*)", line)
if result:
if result.group(1) == "pass":
self.logger.update_dut_job("status", "pass")
return 0
else:
self.logger.update_status_fail("test fail")
return 1
self.print_error(
"Reached the end of the CPU serial log without finding a result")
return 1
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--cpu', type=str,
help='CPU Serial device', required=True)
parser.add_argument(
'--ec', type=str, help='EC Serial device', required=True)
parser.add_argument(
'--test-timeout', type=int, help='Test phase timeout (minutes)', required=True)
args = parser.parse_args()
logger = CustomLogger("job_detail.json")
logger.update_dut_time("start", None)
servo = CrosServoRun(args.cpu, args.ec, args.test_timeout * 60, logger)
retval = servo.run()
# power down the CPU on the device
servo.ec_write("power off\n")
logger.update_dut_time("end", None)
servo.close()
sys.exit(retval)
if __name__ == '__main__':
main()

View File

@@ -1,10 +0,0 @@
#!/bin/bash
relay=$1
if [ -z "$relay" ]; then
echo "Must supply a relay arg"
exit 1
fi
"$CI_PROJECT_DIR"/install/bare-metal/eth008-power-relay.py "$ETH_HOST" "$ETH_PORT" off "$relay"

View File

@@ -1,28 +0,0 @@
#!/usr/bin/python3
import sys
import socket
host = sys.argv[1]
port = sys.argv[2]
mode = sys.argv[3]
relay = sys.argv[4]
msg = None
if mode == "on":
msg = b'\x20'
else:
msg = b'\x21'
msg += int(relay).to_bytes(1, 'big')
msg += b'\x00'
c = socket.create_connection((host, int(port)))
c.sendall(msg)
data = c.recv(1)
c.close()
if data[0] == b'\x01':
print('Command failed')
sys.exit(1)

View File

@@ -1,12 +0,0 @@
#!/bin/bash
relay=$1
if [ -z "$relay" ]; then
echo "Must supply a relay arg"
exit 1
fi
"$CI_PROJECT_DIR"/install/bare-metal/eth008-power-relay.py "$ETH_HOST" "$ETH_PORT" off "$relay"
sleep 5
"$CI_PROJECT_DIR"/install/bare-metal/eth008-power-relay.py "$ETH_HOST" "$ETH_PORT" on "$relay"

View File

@@ -1,31 +0,0 @@
#!/bin/bash
set -e
STRINGS=$(mktemp)
ERRORS=$(mktemp)
trap 'rm $STRINGS; rm $ERRORS;' EXIT
FILE=$1
shift 1
while getopts "f:e:" opt; do
case $opt in
f) echo "$OPTARG" >> "$STRINGS";;
e) echo "$OPTARG" >> "$STRINGS" ; echo "$OPTARG" >> "$ERRORS";;
*) exit
esac
done
shift $((OPTIND -1))
echo "Waiting for $FILE to say one of following strings"
cat "$STRINGS"
while ! grep -E -wf "$STRINGS" "$FILE"; do
sleep 2
done
if grep -E -wf "$ERRORS" "$FILE"; then
exit 1
fi

View File

@@ -1,167 +0,0 @@
#!/bin/bash
# shellcheck disable=SC1091 # The relative paths in this file only become valid at runtime.
# shellcheck disable=SC2034
# shellcheck disable=SC2086 # we want word splitting
. "$SCRIPTS_DIR"/setup-test-env.sh
BM=$CI_PROJECT_DIR/install/bare-metal
CI_COMMON=$CI_PROJECT_DIR/install/common
if [ -z "$BM_SERIAL" ] && [ -z "$BM_SERIAL_SCRIPT" ]; then
echo "Must set BM_SERIAL OR BM_SERIAL_SCRIPT in your gitlab-runner config.toml [[runners]] environment"
echo "BM_SERIAL:"
echo " This is the serial device to talk to for waiting for fastboot to be ready and logging from the kernel."
echo "BM_SERIAL_SCRIPT:"
echo " This is a shell script to talk to for waiting for fastboot to be ready and logging from the kernel."
exit 1
fi
if [ -z "$BM_POWERUP" ]; then
echo "Must set BM_POWERUP in your gitlab-runner config.toml [[runners]] environment"
echo "This is a shell script that should reset the device and begin its boot sequence"
echo "such that it pauses at fastboot."
exit 1
fi
if [ -z "$BM_POWERDOWN" ]; then
echo "Must set BM_POWERDOWN in your gitlab-runner config.toml [[runners]] environment"
echo "This is a shell script that should power off the device."
exit 1
fi
if [ -z "$BM_FASTBOOT_SERIAL" ]; then
echo "Must set BM_FASTBOOT_SERIAL in your gitlab-runner config.toml [[runners]] environment"
echo "This must be the a stable-across-resets fastboot serial number."
exit 1
fi
if [ -z "$BM_KERNEL" ]; then
echo "Must set BM_KERNEL to your board's kernel vmlinuz or Image.gz in the job's variables:"
exit 1
fi
if [ -z "$BM_DTB" ]; then
echo "Must set BM_DTB to your board's DTB file in the job's variables:"
exit 1
fi
if [ -z "$BM_ROOTFS" ]; then
echo "Must set BM_ROOTFS to your board's rootfs directory in the job's variables:"
exit 1
fi
if echo $BM_CMDLINE | grep -q "root=/dev/nfs"; then
BM_FASTBOOT_NFSROOT=1
fi
set -ex
# Clear out any previous run's artifacts.
rm -rf results/
mkdir -p results/
if [ -n "$BM_FASTBOOT_NFSROOT" ]; then
# Create the rootfs in the NFS directory. rm to make sure it's in a pristine
# state, since it's volume-mounted on the host.
rsync -a --delete $BM_ROOTFS/ /nfs/
mkdir -p /nfs/results
. $BM/rootfs-setup.sh /nfs
# Root on NFS, no need for an inintramfs.
rm -f rootfs.cpio.gz
touch rootfs.cpio
gzip rootfs.cpio
else
# Create the rootfs in a temp dir
rsync -a --delete $BM_ROOTFS/ rootfs/
. $BM/rootfs-setup.sh rootfs
# Finally, pack it up into a cpio rootfs. Skip the vulkan CTS since none of
# these devices use it and it would take up space in the initrd.
if [ -n "$PIGLIT_PROFILES" ]; then
EXCLUDE_FILTER="deqp|arb_gpu_shader5|arb_gpu_shader_fp64|arb_gpu_shader_int64|glsl-4.[0123456]0|arb_tessellation_shader"
else
EXCLUDE_FILTER="piglit|python"
fi
pushd rootfs
find -H . | \
grep -E -v "external/(openglcts|vulkancts|amber|glslang|spirv-tools)" |
grep -E -v "traces-db|apitrace|renderdoc" | \
grep -E -v $EXCLUDE_FILTER | \
cpio -H newc -o | \
xz --check=crc32 -T4 - > $CI_PROJECT_DIR/rootfs.cpio.gz
popd
fi
if echo "$BM_KERNEL $BM_DTB" | grep -q http; then
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"$BM_KERNEL" -o kernel
# FIXME: modules should be supplied too
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"$BM_DTB" -o dtb
cat kernel dtb > Image.gz-dtb
elif [ -n "${FORCE_KERNEL_TAG}" ]; then
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/${BM_KERNEL}" -o kernel
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/modules.tar.zst" -o modules.tar.zst
if [ -n "$BM_DTB" ]; then
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/${BM_DTB}.dtb" -o dtb
fi
cat kernel dtb > Image.gz-dtb || echo "No DTB available, using pure kernel."
rm kernel
tar --keep-directory-symlink --zstd -xf modules.tar.zst -C "$BM_ROOTFS/"
rm modules.tar.zst &
else
cat /baremetal-files/"$BM_KERNEL" /baremetal-files/"$BM_DTB".dtb > Image.gz-dtb
cp /baremetal-files/"$BM_DTB".dtb dtb
fi
export PATH=$BM:$PATH
mkdir -p artifacts
mkbootimg.py \
--kernel Image.gz-dtb \
--ramdisk rootfs.cpio.gz \
--dtb dtb \
--cmdline "$BM_CMDLINE" \
$BM_MKBOOT_PARAMS \
--header_version 2 \
-o artifacts/fastboot.img
rm Image.gz-dtb dtb
# Start background command for talking to serial if we have one.
if [ -n "$BM_SERIAL_SCRIPT" ]; then
$BM_SERIAL_SCRIPT > results/serial-output.txt &
while [ ! -e results/serial-output.txt ]; do
sleep 1
done
fi
set +e
$BM/fastboot_run.py \
--dev="$BM_SERIAL" \
--test-timeout ${TEST_PHASE_TIMEOUT:-20} \
--fbserial="$BM_FASTBOOT_SERIAL" \
--powerup="$BM_POWERUP" \
--powerdown="$BM_POWERDOWN"
ret=$?
set -e
if [ -n "$BM_FASTBOOT_NFSROOT" ]; then
# Bring artifacts back from the NFS dir to the build dir where gitlab-runner
# will look for them.
cp -Rp /nfs/results/. results/
fi
exit $ret

View File

@@ -1,159 +0,0 @@
#!/usr/bin/env python3
#
# Copyright © 2020 Google LLC
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
# IN THE SOFTWARE.
import argparse
import subprocess
import re
from serial_buffer import SerialBuffer
import sys
import threading
class FastbootRun:
def __init__(self, args, test_timeout):
self.powerup = args.powerup
self.ser = SerialBuffer(
args.dev, "results/serial-output.txt", "R SERIAL> ")
self.fastboot = "fastboot boot -s {ser} artifacts/fastboot.img".format(
ser=args.fbserial)
self.test_timeout = test_timeout
def close(self):
self.ser.close()
def print_error(self, message):
RED = '\033[0;31m'
NO_COLOR = '\033[0m'
print(RED + message + NO_COLOR)
def logged_system(self, cmd, timeout=60):
print("Running '{}'".format(cmd))
try:
return subprocess.call(cmd, shell=True, timeout=timeout)
except subprocess.TimeoutExpired:
self.print_error("timeout, abandoning run.")
return 1
def run(self):
if ret := self.logged_system(self.powerup):
return ret
fastboot_ready = False
for line in self.ser.lines(timeout=2 * 60, phase="bootloader"):
if re.search("[Ff]astboot: [Pp]rocessing commands", line) or \
re.search("Listening for fastboot command on", line):
fastboot_ready = True
break
if re.search("data abort", line):
self.print_error(
"Detected crash during boot, abandoning run.")
return 1
if not fastboot_ready:
self.print_error(
"Failed to get to fastboot prompt, abandoning run.")
return 1
if ret := self.logged_system(self.fastboot):
return ret
print_more_lines = -1
for line in self.ser.lines(timeout=self.test_timeout, phase="test"):
if print_more_lines == 0:
return 1
if print_more_lines > 0:
print_more_lines -= 1
if re.search("---. end Kernel panic", line):
return 1
# The db820c boards intermittently reboot. Just restart the run
# when if we see a reboot after we got past fastboot.
if re.search("PON REASON", line):
self.print_error(
"Detected spontaneous reboot, abandoning run.")
return 1
# db820c sometimes wedges around iommu fault recovery
if re.search("watchdog: BUG: soft lockup - CPU.* stuck", line):
self.print_error(
"Detected kernel soft lockup, abandoning run.")
return 1
# If the network device dies, it's probably not graphics's fault, just try again.
if re.search("NETDEV WATCHDOG", line):
self.print_error(
"Detected network device failure, abandoning run.")
return 1
# A3xx recovery doesn't quite work. Sometimes the GPU will get
# wedged and recovery will fail (because power can't be reset?)
# This assumes that the jobs are sufficiently well-tested that GPU
# hangs aren't always triggered, so just try again. But print some
# more lines first so that we get better information on the cause
# of the hang. Once a hang happens, it's pretty chatty.
if "[drm:adreno_recover] *ERROR* gpu hw init failed: -22" in line:
self.print_error(
"Detected GPU hang, abandoning run.")
if print_more_lines == -1:
print_more_lines = 30
result = re.search("hwci: mesa: (\S*)", line)
if result:
if result.group(1) == "pass":
return 0
else:
return 1
self.print_error(
"Reached the end of the CPU serial log without finding a result, abandoning run.")
return 1
def main():
parser = argparse.ArgumentParser()
parser.add_argument(
'--dev', type=str, help='Serial device (otherwise reading from serial-output.txt)')
parser.add_argument('--powerup', type=str,
help='shell command for rebooting', required=True)
parser.add_argument('--powerdown', type=str,
help='shell command for powering off', required=True)
parser.add_argument('--fbserial', type=str,
help='fastboot serial number of the board', required=True)
parser.add_argument('--test-timeout', type=int,
help='Test phase timeout (minutes)', required=True)
args = parser.parse_args()
fastboot = FastbootRun(args, args.test_timeout * 60)
retval = fastboot.run()
fastboot.close()
fastboot.logged_system(args.powerdown)
sys.exit(retval)
if __name__ == '__main__':
main()

View File

@@ -1,10 +0,0 @@
#!/bin/bash
relay=$1
if [ -z "$relay" ]; then
echo "Must supply a relay arg"
exit 1
fi
"$CI_PROJECT_DIR"/install/bare-metal/google-power-relay.py off "$relay"

View File

@@ -1,19 +0,0 @@
#!/usr/bin/python3
import sys
import serial
mode = sys.argv[1]
relay = sys.argv[2]
# our relays are "off" means "board is powered".
mode_swap = {
"on": "off",
"off": "on",
}
mode = mode_swap[mode]
ser = serial.Serial('/dev/ttyACM0', 115200, timeout=2)
command = "relay {} {}\n\r".format(mode, relay)
ser.write(command.encode())
ser.close()

View File

@@ -1,12 +0,0 @@
#!/bin/bash
relay=$1
if [ -z "$relay" ]; then
echo "Must supply a relay arg"
exit 1
fi
"$CI_PROJECT_DIR"/install/bare-metal/google-power-relay.py off "$relay"
sleep 5
"$CI_PROJECT_DIR"/install/bare-metal/google-power-relay.py on "$relay"

View File

@@ -1,569 +0,0 @@
#!/usr/bin/env python3
#
# Copyright 2015, The Android Open Source Project
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Creates the boot image."""
from argparse import (ArgumentParser, ArgumentTypeError,
FileType, RawDescriptionHelpFormatter)
from hashlib import sha1
from os import fstat
from struct import pack
import array
import collections
import os
import re
import subprocess
import tempfile
# Constant and structure definition is in
# system/tools/mkbootimg/include/bootimg/bootimg.h
BOOT_MAGIC = 'ANDROID!'
BOOT_MAGIC_SIZE = 8
BOOT_NAME_SIZE = 16
BOOT_ARGS_SIZE = 512
BOOT_EXTRA_ARGS_SIZE = 1024
BOOT_IMAGE_HEADER_V1_SIZE = 1648
BOOT_IMAGE_HEADER_V2_SIZE = 1660
BOOT_IMAGE_HEADER_V3_SIZE = 1580
BOOT_IMAGE_HEADER_V3_PAGESIZE = 4096
BOOT_IMAGE_HEADER_V4_SIZE = 1584
BOOT_IMAGE_V4_SIGNATURE_SIZE = 4096
VENDOR_BOOT_MAGIC = 'VNDRBOOT'
VENDOR_BOOT_MAGIC_SIZE = 8
VENDOR_BOOT_NAME_SIZE = BOOT_NAME_SIZE
VENDOR_BOOT_ARGS_SIZE = 2048
VENDOR_BOOT_IMAGE_HEADER_V3_SIZE = 2112
VENDOR_BOOT_IMAGE_HEADER_V4_SIZE = 2128
VENDOR_RAMDISK_TYPE_NONE = 0
VENDOR_RAMDISK_TYPE_PLATFORM = 1
VENDOR_RAMDISK_TYPE_RECOVERY = 2
VENDOR_RAMDISK_TYPE_DLKM = 3
VENDOR_RAMDISK_NAME_SIZE = 32
VENDOR_RAMDISK_TABLE_ENTRY_BOARD_ID_SIZE = 16
VENDOR_RAMDISK_TABLE_ENTRY_V4_SIZE = 108
# Names with special meaning, mustn't be specified in --ramdisk_name.
VENDOR_RAMDISK_NAME_BLOCKLIST = {b'default'}
PARSER_ARGUMENT_VENDOR_RAMDISK_FRAGMENT = '--vendor_ramdisk_fragment'
def filesize(f):
if f is None:
return 0
try:
return fstat(f.fileno()).st_size
except OSError:
return 0
def update_sha(sha, f):
if f:
sha.update(f.read())
f.seek(0)
sha.update(pack('I', filesize(f)))
else:
sha.update(pack('I', 0))
def pad_file(f, padding):
pad = (padding - (f.tell() & (padding - 1))) & (padding - 1)
f.write(pack(str(pad) + 'x'))
def get_number_of_pages(image_size, page_size):
"""calculates the number of pages required for the image"""
return (image_size + page_size - 1) // page_size
def get_recovery_dtbo_offset(args):
"""calculates the offset of recovery_dtbo image in the boot image"""
num_header_pages = 1 # header occupies a page
num_kernel_pages = get_number_of_pages(filesize(args.kernel), args.pagesize)
num_ramdisk_pages = get_number_of_pages(filesize(args.ramdisk),
args.pagesize)
num_second_pages = get_number_of_pages(filesize(args.second), args.pagesize)
dtbo_offset = args.pagesize * (num_header_pages + num_kernel_pages +
num_ramdisk_pages + num_second_pages)
return dtbo_offset
def write_header_v3_and_above(args):
if args.header_version > 3:
boot_header_size = BOOT_IMAGE_HEADER_V4_SIZE
else:
boot_header_size = BOOT_IMAGE_HEADER_V3_SIZE
args.output.write(pack(f'{BOOT_MAGIC_SIZE}s', BOOT_MAGIC.encode()))
# kernel size in bytes
args.output.write(pack('I', filesize(args.kernel)))
# ramdisk size in bytes
args.output.write(pack('I', filesize(args.ramdisk)))
# os version and patch level
args.output.write(pack('I', (args.os_version << 11) | args.os_patch_level))
args.output.write(pack('I', boot_header_size))
# reserved
args.output.write(pack('4I', 0, 0, 0, 0))
# version of boot image header
args.output.write(pack('I', args.header_version))
args.output.write(pack(f'{BOOT_ARGS_SIZE + BOOT_EXTRA_ARGS_SIZE}s',
args.cmdline))
if args.header_version >= 4:
# The signature used to verify boot image v4.
args.output.write(pack('I', BOOT_IMAGE_V4_SIGNATURE_SIZE))
pad_file(args.output, BOOT_IMAGE_HEADER_V3_PAGESIZE)
def write_vendor_boot_header(args):
if filesize(args.dtb) == 0:
raise ValueError('DTB image must not be empty.')
if args.header_version > 3:
vendor_ramdisk_size = args.vendor_ramdisk_total_size
vendor_boot_header_size = VENDOR_BOOT_IMAGE_HEADER_V4_SIZE
else:
vendor_ramdisk_size = filesize(args.vendor_ramdisk)
vendor_boot_header_size = VENDOR_BOOT_IMAGE_HEADER_V3_SIZE
args.vendor_boot.write(pack(f'{VENDOR_BOOT_MAGIC_SIZE}s',
VENDOR_BOOT_MAGIC.encode()))
# version of boot image header
args.vendor_boot.write(pack('I', args.header_version))
# flash page size
args.vendor_boot.write(pack('I', args.pagesize))
# kernel physical load address
args.vendor_boot.write(pack('I', args.base + args.kernel_offset))
# ramdisk physical load address
args.vendor_boot.write(pack('I', args.base + args.ramdisk_offset))
# ramdisk size in bytes
args.vendor_boot.write(pack('I', vendor_ramdisk_size))
args.vendor_boot.write(pack(f'{VENDOR_BOOT_ARGS_SIZE}s',
args.vendor_cmdline))
# kernel tags physical load address
args.vendor_boot.write(pack('I', args.base + args.tags_offset))
# asciiz product name
args.vendor_boot.write(pack(f'{VENDOR_BOOT_NAME_SIZE}s', args.board))
# header size in bytes
args.vendor_boot.write(pack('I', vendor_boot_header_size))
# dtb size in bytes
args.vendor_boot.write(pack('I', filesize(args.dtb)))
# dtb physical load address
args.vendor_boot.write(pack('Q', args.base + args.dtb_offset))
if args.header_version > 3:
vendor_ramdisk_table_size = (args.vendor_ramdisk_table_entry_num *
VENDOR_RAMDISK_TABLE_ENTRY_V4_SIZE)
# vendor ramdisk table size in bytes
args.vendor_boot.write(pack('I', vendor_ramdisk_table_size))
# number of vendor ramdisk table entries
args.vendor_boot.write(pack('I', args.vendor_ramdisk_table_entry_num))
# vendor ramdisk table entry size in bytes
args.vendor_boot.write(pack('I', VENDOR_RAMDISK_TABLE_ENTRY_V4_SIZE))
# bootconfig section size in bytes
args.vendor_boot.write(pack('I', filesize(args.vendor_bootconfig)))
pad_file(args.vendor_boot, args.pagesize)
def write_header(args):
if args.header_version > 4:
raise ValueError(
f'Boot header version {args.header_version} not supported')
if args.header_version in {3, 4}:
return write_header_v3_and_above(args)
ramdisk_load_address = ((args.base + args.ramdisk_offset)
if filesize(args.ramdisk) > 0 else 0)
second_load_address = ((args.base + args.second_offset)
if filesize(args.second) > 0 else 0)
args.output.write(pack(f'{BOOT_MAGIC_SIZE}s', BOOT_MAGIC.encode()))
# kernel size in bytes
args.output.write(pack('I', filesize(args.kernel)))
# kernel physical load address
args.output.write(pack('I', args.base + args.kernel_offset))
# ramdisk size in bytes
args.output.write(pack('I', filesize(args.ramdisk)))
# ramdisk physical load address
args.output.write(pack('I', ramdisk_load_address))
# second bootloader size in bytes
args.output.write(pack('I', filesize(args.second)))
# second bootloader physical load address
args.output.write(pack('I', second_load_address))
# kernel tags physical load address
args.output.write(pack('I', args.base + args.tags_offset))
# flash page size
args.output.write(pack('I', args.pagesize))
# version of boot image header
args.output.write(pack('I', args.header_version))
# os version and patch level
args.output.write(pack('I', (args.os_version << 11) | args.os_patch_level))
# asciiz product name
args.output.write(pack(f'{BOOT_NAME_SIZE}s', args.board))
args.output.write(pack(f'{BOOT_ARGS_SIZE}s', args.cmdline))
sha = sha1()
update_sha(sha, args.kernel)
update_sha(sha, args.ramdisk)
update_sha(sha, args.second)
if args.header_version > 0:
update_sha(sha, args.recovery_dtbo)
if args.header_version > 1:
update_sha(sha, args.dtb)
img_id = pack('32s', sha.digest())
args.output.write(img_id)
args.output.write(pack(f'{BOOT_EXTRA_ARGS_SIZE}s', args.extra_cmdline))
if args.header_version > 0:
if args.recovery_dtbo:
# recovery dtbo size in bytes
args.output.write(pack('I', filesize(args.recovery_dtbo)))
# recovert dtbo offset in the boot image
args.output.write(pack('Q', get_recovery_dtbo_offset(args)))
else:
# Set to zero if no recovery dtbo
args.output.write(pack('I', 0))
args.output.write(pack('Q', 0))
# Populate boot image header size for header versions 1 and 2.
if args.header_version == 1:
args.output.write(pack('I', BOOT_IMAGE_HEADER_V1_SIZE))
elif args.header_version == 2:
args.output.write(pack('I', BOOT_IMAGE_HEADER_V2_SIZE))
if args.header_version > 1:
if filesize(args.dtb) == 0:
raise ValueError('DTB image must not be empty.')
# dtb size in bytes
args.output.write(pack('I', filesize(args.dtb)))
# dtb physical load address
args.output.write(pack('Q', args.base + args.dtb_offset))
pad_file(args.output, args.pagesize)
return img_id
class AsciizBytes:
"""Parses a string and encodes it as an asciiz bytes object.
>>> AsciizBytes(bufsize=4)('foo')
b'foo\\x00'
>>> AsciizBytes(bufsize=4)('foob')
Traceback (most recent call last):
...
argparse.ArgumentTypeError: Encoded asciiz length exceeded: max 4, got 5
"""
def __init__(self, bufsize):
self.bufsize = bufsize
def __call__(self, arg):
arg_bytes = arg.encode() + b'\x00'
if len(arg_bytes) > self.bufsize:
raise ArgumentTypeError(
'Encoded asciiz length exceeded: '
f'max {self.bufsize}, got {len(arg_bytes)}')
return arg_bytes
class VendorRamdiskTableBuilder:
"""Vendor ramdisk table builder.
Attributes:
entries: A list of VendorRamdiskTableEntry namedtuple.
ramdisk_total_size: Total size in bytes of all ramdisks in the table.
"""
VendorRamdiskTableEntry = collections.namedtuple( # pylint: disable=invalid-name
'VendorRamdiskTableEntry',
['ramdisk_path', 'ramdisk_size', 'ramdisk_offset', 'ramdisk_type',
'ramdisk_name', 'board_id'])
def __init__(self):
self.entries = []
self.ramdisk_total_size = 0
self.ramdisk_names = set()
def add_entry(self, ramdisk_path, ramdisk_type, ramdisk_name, board_id):
# Strip any trailing null for simple comparison.
stripped_ramdisk_name = ramdisk_name.rstrip(b'\x00')
if stripped_ramdisk_name in VENDOR_RAMDISK_NAME_BLOCKLIST:
raise ValueError(
f'Banned vendor ramdisk name: {stripped_ramdisk_name}')
if stripped_ramdisk_name in self.ramdisk_names:
raise ValueError(
f'Duplicated vendor ramdisk name: {stripped_ramdisk_name}')
self.ramdisk_names.add(stripped_ramdisk_name)
if board_id is None:
board_id = array.array(
'I', [0] * VENDOR_RAMDISK_TABLE_ENTRY_BOARD_ID_SIZE)
else:
board_id = array.array('I', board_id)
if len(board_id) != VENDOR_RAMDISK_TABLE_ENTRY_BOARD_ID_SIZE:
raise ValueError('board_id size must be '
f'{VENDOR_RAMDISK_TABLE_ENTRY_BOARD_ID_SIZE}')
with open(ramdisk_path, 'rb') as f:
ramdisk_size = filesize(f)
self.entries.append(self.VendorRamdiskTableEntry(
ramdisk_path, ramdisk_size, self.ramdisk_total_size, ramdisk_type,
ramdisk_name, board_id))
self.ramdisk_total_size += ramdisk_size
def write_ramdisks_padded(self, fout, alignment):
for entry in self.entries:
with open(entry.ramdisk_path, 'rb') as f:
fout.write(f.read())
pad_file(fout, alignment)
def write_entries_padded(self, fout, alignment):
for entry in self.entries:
fout.write(pack('I', entry.ramdisk_size))
fout.write(pack('I', entry.ramdisk_offset))
fout.write(pack('I', entry.ramdisk_type))
fout.write(pack(f'{VENDOR_RAMDISK_NAME_SIZE}s',
entry.ramdisk_name))
fout.write(entry.board_id)
pad_file(fout, alignment)
def write_padded_file(f_out, f_in, padding):
if f_in is None:
return
f_out.write(f_in.read())
pad_file(f_out, padding)
def parse_int(x):
return int(x, 0)
def parse_os_version(x):
match = re.search(r'^(\d{1,3})(?:\.(\d{1,3})(?:\.(\d{1,3}))?)?', x)
if match:
a = int(match.group(1))
b = c = 0
if match.lastindex >= 2:
b = int(match.group(2))
if match.lastindex == 3:
c = int(match.group(3))
# 7 bits allocated for each field
assert a < 128
assert b < 128
assert c < 128
return (a << 14) | (b << 7) | c
return 0
def parse_os_patch_level(x):
match = re.search(r'^(\d{4})-(\d{2})(?:-(\d{2}))?', x)
if match:
y = int(match.group(1)) - 2000
m = int(match.group(2))
# 7 bits allocated for the year, 4 bits for the month
assert 0 <= y < 128
assert 0 < m <= 12
return (y << 4) | m
return 0
def parse_vendor_ramdisk_type(x):
type_dict = {
'none': VENDOR_RAMDISK_TYPE_NONE,
'platform': VENDOR_RAMDISK_TYPE_PLATFORM,
'recovery': VENDOR_RAMDISK_TYPE_RECOVERY,
'dlkm': VENDOR_RAMDISK_TYPE_DLKM,
}
if x.lower() in type_dict:
return type_dict[x.lower()]
return parse_int(x)
def get_vendor_boot_v4_usage():
return """vendor boot version 4 arguments:
--ramdisk_type {none,platform,recovery,dlkm}
specify the type of the ramdisk
--ramdisk_name NAME
specify the name of the ramdisk
--board_id{0..15} NUMBER
specify the value of the board_id vector, defaults to 0
--vendor_ramdisk_fragment VENDOR_RAMDISK_FILE
path to the vendor ramdisk file
These options can be specified multiple times, where each vendor ramdisk
option group ends with a --vendor_ramdisk_fragment option.
Each option group appends an additional ramdisk to the vendor boot image.
"""
def parse_vendor_ramdisk_args(args, args_list):
"""Parses vendor ramdisk specific arguments.
Args:
args: An argparse.Namespace object. Parsed results are stored into this
object.
args_list: A list of argument strings to be parsed.
Returns:
A list argument strings that are not parsed by this method.
"""
parser = ArgumentParser(add_help=False)
parser.add_argument('--ramdisk_type', type=parse_vendor_ramdisk_type,
default=VENDOR_RAMDISK_TYPE_NONE)
parser.add_argument('--ramdisk_name',
type=AsciizBytes(bufsize=VENDOR_RAMDISK_NAME_SIZE),
required=True)
for i in range(VENDOR_RAMDISK_TABLE_ENTRY_BOARD_ID_SIZE):
parser.add_argument(f'--board_id{i}', type=parse_int, default=0)
parser.add_argument(PARSER_ARGUMENT_VENDOR_RAMDISK_FRAGMENT, required=True)
unknown_args = []
vendor_ramdisk_table_builder = VendorRamdiskTableBuilder()
if args.vendor_ramdisk is not None:
vendor_ramdisk_table_builder.add_entry(
args.vendor_ramdisk.name, VENDOR_RAMDISK_TYPE_PLATFORM, b'', None)
while PARSER_ARGUMENT_VENDOR_RAMDISK_FRAGMENT in args_list:
idx = args_list.index(PARSER_ARGUMENT_VENDOR_RAMDISK_FRAGMENT) + 2
vendor_ramdisk_args = args_list[:idx]
args_list = args_list[idx:]
ramdisk_args, extra_args = parser.parse_known_args(vendor_ramdisk_args)
ramdisk_args_dict = vars(ramdisk_args)
unknown_args.extend(extra_args)
ramdisk_path = ramdisk_args.vendor_ramdisk_fragment
ramdisk_type = ramdisk_args.ramdisk_type
ramdisk_name = ramdisk_args.ramdisk_name
board_id = [ramdisk_args_dict[f'board_id{i}']
for i in range(VENDOR_RAMDISK_TABLE_ENTRY_BOARD_ID_SIZE)]
vendor_ramdisk_table_builder.add_entry(ramdisk_path, ramdisk_type,
ramdisk_name, board_id)
if len(args_list) > 0:
unknown_args.extend(args_list)
args.vendor_ramdisk_total_size = (vendor_ramdisk_table_builder
.ramdisk_total_size)
args.vendor_ramdisk_table_entry_num = len(vendor_ramdisk_table_builder
.entries)
args.vendor_ramdisk_table_builder = vendor_ramdisk_table_builder
return unknown_args
def parse_cmdline():
version_parser = ArgumentParser(add_help=False)
version_parser.add_argument('--header_version', type=parse_int, default=0)
if version_parser.parse_known_args()[0].header_version < 3:
# For boot header v0 to v2, the kernel commandline field is split into
# two fields, cmdline and extra_cmdline. Both fields are asciiz strings,
# so we minus one here to ensure the encoded string plus the
# null-terminator can fit in the buffer size.
cmdline_size = BOOT_ARGS_SIZE + BOOT_EXTRA_ARGS_SIZE - 1
else:
cmdline_size = BOOT_ARGS_SIZE + BOOT_EXTRA_ARGS_SIZE
parser = ArgumentParser(formatter_class=RawDescriptionHelpFormatter,
epilog=get_vendor_boot_v4_usage())
parser.add_argument('--kernel', type=FileType('rb'),
help='path to the kernel')
parser.add_argument('--ramdisk', type=FileType('rb'),
help='path to the ramdisk')
parser.add_argument('--second', type=FileType('rb'),
help='path to the second bootloader')
parser.add_argument('--dtb', type=FileType('rb'), help='path to the dtb')
dtbo_group = parser.add_mutually_exclusive_group()
dtbo_group.add_argument('--recovery_dtbo', type=FileType('rb'),
help='path to the recovery DTBO')
dtbo_group.add_argument('--recovery_acpio', type=FileType('rb'),
metavar='RECOVERY_ACPIO', dest='recovery_dtbo',
help='path to the recovery ACPIO')
parser.add_argument('--cmdline', type=AsciizBytes(bufsize=cmdline_size),
default='', help='kernel command line arguments')
parser.add_argument('--vendor_cmdline',
type=AsciizBytes(bufsize=VENDOR_BOOT_ARGS_SIZE),
default='',
help='vendor boot kernel command line arguments')
parser.add_argument('--base', type=parse_int, default=0x10000000,
help='base address')
parser.add_argument('--kernel_offset', type=parse_int, default=0x00008000,
help='kernel offset')
parser.add_argument('--ramdisk_offset', type=parse_int, default=0x01000000,
help='ramdisk offset')
parser.add_argument('--second_offset', type=parse_int, default=0x00f00000,
help='second bootloader offset')
parser.add_argument('--dtb_offset', type=parse_int, default=0x01f00000,
help='dtb offset')
parser.add_argument('--os_version', type=parse_os_version, default=0,
help='operating system version')
parser.add_argument('--os_patch_level', type=parse_os_patch_level,
default=0, help='operating system patch level')
parser.add_argument('--tags_offset', type=parse_int, default=0x00000100,
help='tags offset')
parser.add_argument('--board', type=AsciizBytes(bufsize=BOOT_NAME_SIZE),
default='', help='board name')
parser.add_argument('--pagesize', type=parse_int,
choices=[2**i for i in range(11, 15)], default=2048,
help='page size')
parser.add_argument('--id', action='store_true',
help='print the image ID on standard output')
parser.add_argument('--header_version', type=parse_int, default=0,
help='boot image header version')
parser.add_argument('-o', '--output', type=FileType('wb'),
help='output file name')
parser.add_argument('--gki_signing_algorithm',
help='GKI signing algorithm to use')
parser.add_argument('--gki_signing_key',
help='path to RSA private key file')
parser.add_argument('--gki_signing_signature_args',
help='other hash arguments passed to avbtool')
parser.add_argument('--gki_signing_avbtool_path',
help='path to avbtool for boot signature generation')
parser.add_argument('--vendor_boot', type=FileType('wb'),
help='vendor boot output file name')
parser.add_argument('--vendor_ramdisk', type=FileType('rb'),
help='path to the vendor ramdisk')
parser.add_argument('--vendor_bootconfig', type=FileType('rb'),
help='path to the vendor bootconfig file')
args, extra_args = parser.parse_known_args()
if args.vendor_boot is not None and args.header_version > 3:
extra_args = parse_vendor_ramdisk_args(args, extra_args)
if len(extra_args) > 0:
raise ValueError(f'Unrecognized arguments: {extra_args}')
if args.header_version < 3:
args.extra_cmdline = args.cmdline[BOOT_ARGS_SIZE-1:]
args.cmdline = args.cmdline[:BOOT_ARGS_SIZE-1] + b'\x00'
assert len(args.cmdline) <= BOOT_ARGS_SIZE
assert len(args.extra_cmdline) <= BOOT_EXTRA_ARGS_SIZE
return args
def add_boot_image_signature(args, pagesize):
"""Adds the boot image signature.
Note that the signature will only be verified in VTS to ensure a
generic boot.img is used. It will not be used by the device
bootloader at boot time. The bootloader should only verify
the boot vbmeta at the end of the boot partition (or in the top-level
vbmeta partition) via the Android Verified Boot process, when the
device boots.
"""
args.output.flush() # Flush the buffer for signature calculation.
# Appends zeros if the signing key is not specified.
if not args.gki_signing_key or not args.gki_signing_algorithm:
zeros = b'\x00' * BOOT_IMAGE_V4_SIGNATURE_SIZE
args.output.write(zeros)
pad_file(args.output, pagesize)
return
avbtool = 'avbtool' # Used from otatools.zip or Android build env.
# We need to specify the path of avbtool in build/core/Makefile.
# Because avbtool is not guaranteed to be in $PATH there.
if args.gki_signing_avbtool_path:
avbtool = args.gki_signing_avbtool_path
# Need to specify a value of --partition_size for avbtool to work.
# We use 64 MB below, but avbtool will not resize the boot image to
# this size because --do_not_append_vbmeta_image is also specified.
avbtool_cmd = [
avbtool, 'add_hash_footer',
'--partition_name', 'boot',
'--partition_size', str(64 * 1024 * 1024),
'--image', args.output.name,
'--algorithm', args.gki_signing_algorithm,
'--key', args.gki_signing_key,
'--salt', 'd00df00d'] # TODO: use a hash of kernel/ramdisk as the salt.
# Additional arguments passed to avbtool.
if args.gki_signing_signature_args:
avbtool_cmd += args.gki_signing_signature_args.split()
# Outputs the signed vbmeta to a separate file, then append to boot.img
# as the boot signature.
with tempfile.TemporaryDirectory() as temp_out_dir:
boot_signature_output = os.path.join(temp_out_dir, 'boot_signature')
avbtool_cmd += ['--do_not_append_vbmeta_image',
'--output_vbmeta_image', boot_signature_output]
subprocess.check_call(avbtool_cmd)
with open(boot_signature_output, 'rb') as boot_signature:
if filesize(boot_signature) > BOOT_IMAGE_V4_SIGNATURE_SIZE:
raise ValueError(
f'boot sigature size is > {BOOT_IMAGE_V4_SIGNATURE_SIZE}')
write_padded_file(args.output, boot_signature, pagesize)
def write_data(args, pagesize):
write_padded_file(args.output, args.kernel, pagesize)
write_padded_file(args.output, args.ramdisk, pagesize)
write_padded_file(args.output, args.second, pagesize)
if args.header_version > 0 and args.header_version < 3:
write_padded_file(args.output, args.recovery_dtbo, pagesize)
if args.header_version == 2:
write_padded_file(args.output, args.dtb, pagesize)
if args.header_version >= 4:
add_boot_image_signature(args, pagesize)
def write_vendor_boot_data(args):
if args.header_version > 3:
builder = args.vendor_ramdisk_table_builder
builder.write_ramdisks_padded(args.vendor_boot, args.pagesize)
write_padded_file(args.vendor_boot, args.dtb, args.pagesize)
builder.write_entries_padded(args.vendor_boot, args.pagesize)
write_padded_file(args.vendor_boot, args.vendor_bootconfig,
args.pagesize)
else:
write_padded_file(args.vendor_boot, args.vendor_ramdisk, args.pagesize)
write_padded_file(args.vendor_boot, args.dtb, args.pagesize)
def main():
args = parse_cmdline()
if args.vendor_boot is not None:
if args.header_version not in {3, 4}:
raise ValueError(
'--vendor_boot not compatible with given header version')
if args.header_version == 3 and args.vendor_ramdisk is None:
raise ValueError('--vendor_ramdisk missing or invalid')
write_vendor_boot_header(args)
write_vendor_boot_data(args)
if args.output is not None:
if args.second is not None and args.header_version > 2:
raise ValueError(
'--second not compatible with given header version')
img_id = write_header(args)
if args.header_version > 2:
write_data(args, BOOT_IMAGE_HEADER_V3_PAGESIZE)
else:
write_data(args, args.pagesize)
if args.id and img_id is not None:
print('0x' + ''.join(f'{octet:02x}' for octet in img_id))
if __name__ == '__main__':
main()

View File

@@ -1,16 +0,0 @@
#!/bin/bash
if [ -z "$BM_POE_INTERFACE" ]; then
echo "Must supply the PoE Interface to power up"
exit 1
fi
if [ -z "$BM_POE_ADDRESS" ]; then
echo "Must supply the PoE Switch host"
exit 1
fi
SNMP_KEY="SNMPv2-SMI::mib-2.105.1.1.1.3.1.$((48 + BM_POE_INTERFACE))"
SNMP_OFF="i 2"
flock /var/run/poe.lock -c "snmpset -v2c -r 3 -t 30 -cmesaci $BM_POE_ADDRESS $SNMP_KEY $SNMP_OFF"

View File

@@ -1,19 +0,0 @@
#!/bin/bash
if [ -z "$BM_POE_INTERFACE" ]; then
echo "Must supply the PoE Interface to power up"
exit 1
fi
if [ -z "$BM_POE_ADDRESS" ]; then
echo "Must supply the PoE Switch host"
exit 1
fi
SNMP_KEY="SNMPv2-SMI::mib-2.105.1.1.1.3.1.$((48 + BM_POE_INTERFACE))"
SNMP_ON="i 1"
SNMP_OFF="i 2"
flock /var/run/poe.lock -c "snmpset -v2c -r 3 -t 30 -cmesaci $BM_POE_ADDRESS $SNMP_KEY $SNMP_OFF"
sleep 3s
flock /var/run/poe.lock -c "snmpset -v2c -r 3 -t 30 -cmesaci $BM_POE_ADDRESS $SNMP_KEY $SNMP_ON"

View File

@@ -1,229 +0,0 @@
#!/bin/bash
# shellcheck disable=SC1091
# shellcheck disable=SC2034
# shellcheck disable=SC2059
# shellcheck disable=SC2086 # we want word splitting
. "$SCRIPTS_DIR"/setup-test-env.sh
# Boot script for devices attached to a PoE switch, using NFS for the root
# filesystem.
# We're run from the root of the repo, make a helper var for our paths
BM=$CI_PROJECT_DIR/install/bare-metal
CI_COMMON=$CI_PROJECT_DIR/install/common
CI_INSTALL=$CI_PROJECT_DIR/install
# Runner config checks
if [ -z "$BM_SERIAL" ]; then
echo "Must set BM_SERIAL in your gitlab-runner config.toml [[runners]] environment"
echo "This is the serial port to listen the device."
exit 1
fi
if [ -z "$BM_POE_ADDRESS" ]; then
echo "Must set BM_POE_ADDRESS in your gitlab-runner config.toml [[runners]] environment"
echo "This is the PoE switch address to connect for powering up/down devices."
exit 1
fi
if [ -z "$BM_POE_INTERFACE" ]; then
echo "Must set BM_POE_INTERFACE in your gitlab-runner config.toml [[runners]] environment"
echo "This is the PoE switch interface where the device is connected."
exit 1
fi
if [ -z "$BM_POWERUP" ]; then
echo "Must set BM_POWERUP in your gitlab-runner config.toml [[runners]] environment"
echo "This is a shell script that should power up the device and begin its boot sequence."
exit 1
fi
if [ -z "$BM_POWERDOWN" ]; then
echo "Must set BM_POWERDOWN in your gitlab-runner config.toml [[runners]] environment"
echo "This is a shell script that should power off the device."
exit 1
fi
if [ ! -d /nfs ]; then
echo "NFS rootfs directory needs to be mounted at /nfs by the gitlab runner"
exit 1
fi
if [ ! -d /tftp ]; then
echo "TFTP directory for this board needs to be mounted at /tftp by the gitlab runner"
exit 1
fi
# job config checks
if [ -z "$BM_ROOTFS" ]; then
echo "Must set BM_ROOTFS to your board's rootfs directory in the job's variables"
exit 1
fi
if [ -z "$BM_BOOTFS" ] && { [ -z "$BM_KERNEL" ] || [ -z "$BM_DTB" ]; } ; then
echo "Must set /boot files for the TFTP boot in the job's variables or set kernel and dtb"
exit 1
fi
if [ -z "$BM_CMDLINE" ]; then
echo "Must set BM_CMDLINE to your board's kernel command line arguments"
exit 1
fi
set -ex
date +'%F %T'
# Clear out any previous run's artifacts.
rm -rf results/
mkdir -p results
# Create the rootfs in the NFS directory. rm to make sure it's in a pristine
# state, since it's volume-mounted on the host.
rsync -a --delete $BM_ROOTFS/ /nfs/
date +'%F %T'
# If BM_BOOTFS is an URL, download it
if echo $BM_BOOTFS | grep -q http; then
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}$BM_BOOTFS" -o /tmp/bootfs.tar
BM_BOOTFS=/tmp/bootfs.tar
fi
date +'%F %T'
# If BM_BOOTFS is a file, assume it is a tarball and uncompress it
if [ -f "${BM_BOOTFS}" ]; then
mkdir -p /tmp/bootfs
tar xf $BM_BOOTFS -C /tmp/bootfs
BM_BOOTFS=/tmp/bootfs
fi
# If BM_KERNEL and BM_DTS is present
if [ -n "${FORCE_KERNEL_TAG}" ]; then
if [ -z "${BM_KERNEL}" ] || [ -z "${BM_DTB}" ]; then
echo "This machine cannot be tested with external kernel since BM_KERNEL or BM_DTB missing!"
exit 1
fi
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/${BM_KERNEL}" -o "${BM_KERNEL}"
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/${BM_DTB}.dtb" -o "${BM_DTB}.dtb"
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${FDO_HTTP_CACHE_URI:-}${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/modules.tar.zst" -o modules.tar.zst
fi
date +'%F %T'
# Install kernel modules (it could be either in /lib/modules or
# /usr/lib/modules, but we want to install in the latter)
if [ -n "${FORCE_KERNEL_TAG}" ]; then
tar --keep-directory-symlink --zstd -xf modules.tar.zst -C /nfs/
rm modules.tar.zst &
elif [ -n "${BM_BOOTFS}" ]; then
[ -d $BM_BOOTFS/usr/lib/modules ] && rsync -a $BM_BOOTFS/usr/lib/modules/ /nfs/usr/lib/modules/
[ -d $BM_BOOTFS/lib/modules ] && rsync -a $BM_BOOTFS/lib/modules/ /nfs/lib/modules/
else
echo "No modules!"
fi
date +'%F %T'
# Install kernel image + bootloader files
if [ -n "${FORCE_KERNEL_TAG}" ] || [ -z "$BM_BOOTFS" ]; then
mv "${BM_KERNEL}" "${BM_DTB}.dtb" /tftp/
else # BM_BOOTFS
rsync -aL --delete $BM_BOOTFS/boot/ /tftp/
fi
date +'%F %T'
# Set up the pxelinux config for Jetson Nano
mkdir -p /tftp/pxelinux.cfg
cat <<EOF >/tftp/pxelinux.cfg/default-arm-tegra210-p3450-0000
PROMPT 0
TIMEOUT 30
DEFAULT primary
MENU TITLE jetson nano boot options
LABEL primary
MENU LABEL CI kernel on TFTP
LINUX Image
FDT tegra210-p3450-0000.dtb
APPEND \${cbootargs} $BM_CMDLINE
EOF
# Set up the pxelinux config for Jetson TK1
cat <<EOF >/tftp/pxelinux.cfg/default-arm-tegra124-jetson-tk1
PROMPT 0
TIMEOUT 30
DEFAULT primary
MENU TITLE jetson TK1 boot options
LABEL primary
MENU LABEL CI kernel on TFTP
LINUX zImage
FDT tegra124-jetson-tk1.dtb
APPEND \${cbootargs} $BM_CMDLINE
EOF
# Create the rootfs in the NFS directory
. $BM/rootfs-setup.sh /nfs
date +'%F %T'
echo "$BM_CMDLINE" > /tftp/cmdline.txt
# Add some options in config.txt, if defined
if [ -n "$BM_BOOTCONFIG" ]; then
printf "$BM_BOOTCONFIG" >> /tftp/config.txt
fi
set +e
STRUCTURED_LOG_FILE=job_detail.json
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --update dut_job_type "${DEVICE_TYPE}"
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --update farm "${FARM}"
ATTEMPTS=3
first_attempt=True
while [ $((ATTEMPTS--)) -gt 0 ]; do
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --create-dut-job dut_name "${CI_RUNNER_DESCRIPTION}"
# Update subtime time to CI_JOB_STARTED_AT only for the first run
if [ "$first_attempt" = "True" ]; then
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --update-dut-time submit "${CI_JOB_STARTED_AT}"
else
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --update-dut-time submit
fi
python3 $BM/poe_run.py \
--dev="$BM_SERIAL" \
--powerup="$BM_POWERUP" \
--powerdown="$BM_POWERDOWN" \
--test-timeout ${TEST_PHASE_TIMEOUT:-20}
ret=$?
if [ $ret -eq 2 ]; then
echo "Did not detect boot sequence, retrying..."
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --close-dut-job
first_attempt=False
else
ATTEMPTS=0
fi
done
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --close-dut-job
python3 $CI_INSTALL/custom_logger.py ${STRUCTURED_LOG_FILE} --close
set -e
date +'%F %T'
# Bring artifacts back from the NFS dir to the build dir where gitlab-runner
# will look for them.
cp -Rp /nfs/results/. results/
if [ -f "${STRUCTURED_LOG_FILE}" ]; then
cp -p ${STRUCTURED_LOG_FILE} results/
echo "Structured log file is available at ${ARTIFACTS_BASE_URL}/results/${STRUCTURED_LOG_FILE}"
fi
date +'%F %T'
exit $ret

View File

@@ -1,127 +0,0 @@
#!/usr/bin/env python3
#
# Copyright © 2020 Igalia, S.L.
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
# IN THE SOFTWARE.
import argparse
import os
import re
import sys
import threading
from custom_logger import CustomLogger
from serial_buffer import SerialBuffer
class PoERun:
def __init__(self, args, test_timeout, logger):
self.powerup = args.powerup
self.powerdown = args.powerdown
self.ser = SerialBuffer(
args.dev, "results/serial-output.txt", "")
self.test_timeout = test_timeout
self.logger = logger
def print_error(self, message):
RED = '\033[0;31m'
NO_COLOR = '\033[0m'
print(RED + message + NO_COLOR)
self.logger.update_status_fail(message)
def logged_system(self, cmd):
print("Running '{}'".format(cmd))
return os.system(cmd)
def run(self):
if self.logged_system(self.powerup) != 0:
self.logger.update_status_fail("powerup failed")
return 1
boot_detected = False
self.logger.create_job_phase("boot")
for line in self.ser.lines(timeout=5 * 60, phase="bootloader"):
if re.search("Booting Linux", line):
boot_detected = True
break
if not boot_detected:
self.print_error(
"Something wrong; couldn't detect the boot start up sequence")
return 1
self.logger.create_job_phase("test")
for line in self.ser.lines(timeout=self.test_timeout, phase="test"):
if re.search("---. end Kernel panic", line):
self.logger.update_status_fail("kernel panic")
return 1
# Binning memory problems
if re.search("binner overflow mem", line):
self.print_error("Memory overflow in the binner; GPU hang")
return 1
if re.search("nouveau 57000000.gpu: bus: MMIO read of 00000000 FAULT at 137000", line):
self.print_error("nouveau jetson boot bug, abandoning run.")
return 1
# network fail on tk1
if re.search("NETDEV WATCHDOG:.* transmit queue 0 timed out", line):
self.print_error("nouveau jetson tk1 network fail, abandoning run.")
return 1
result = re.search("hwci: mesa: (\S*)", line)
if result:
if result.group(1) == "pass":
self.logger.update_dut_job("status", "pass")
return 0
else:
self.logger.update_status_fail("test fail")
return 1
self.print_error(
"Reached the end of the CPU serial log without finding a result")
return 1
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--dev', type=str,
help='Serial device to monitor', required=True)
parser.add_argument('--powerup', type=str,
help='shell command for rebooting', required=True)
parser.add_argument('--powerdown', type=str,
help='shell command for powering off', required=True)
parser.add_argument(
'--test-timeout', type=int, help='Test phase timeout (minutes)', required=True)
args = parser.parse_args()
logger = CustomLogger("job_detail.json")
logger.update_dut_time("start", None)
poe = PoERun(args, args.test_timeout * 60, logger)
retval = poe.run()
poe.logged_system(args.powerdown)
logger.update_dut_time("end", None)
sys.exit(retval)
if __name__ == '__main__':
main()

View File

@@ -1,37 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
rootfs_dst=$1
mkdir -p $rootfs_dst/results
# Set up the init script that brings up the system.
cp $BM/bm-init.sh $rootfs_dst/init
cp $CI_COMMON/init*.sh $rootfs_dst/
date +'%F %T'
# Make JWT token available as file in the bare-metal storage to enable access
# to MinIO
cp "${CI_JOB_JWT_FILE}" "${rootfs_dst}${CI_JOB_JWT_FILE}"
date +'%F %T'
cp $CI_COMMON/capture-devcoredump.sh $rootfs_dst/
cp $CI_COMMON/intel-gpu-freq.sh $rootfs_dst/
cp $CI_COMMON/kdl.sh $rootfs_dst/
cp "$SCRIPTS_DIR/setup-test-env.sh" "$rootfs_dst/"
set +x
# Pass through relevant env vars from the gitlab job to the baremetal init script
echo "Variables passed through:"
"$CI_COMMON"/generate-env.sh | tee $rootfs_dst/set-job-env-vars.sh
set -x
# Add the Mesa drivers we built, and make a consistent symlink to them.
mkdir -p $rootfs_dst/$CI_PROJECT_DIR
rsync -aH --delete $CI_PROJECT_DIR/install/ $rootfs_dst/$CI_PROJECT_DIR/install/
date +'%F %T'

View File

@@ -1,185 +0,0 @@
#!/usr/bin/env python3
#
# Copyright © 2020 Google LLC
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
# IN THE SOFTWARE.
import argparse
from datetime import datetime, timezone
import queue
import serial
import threading
import time
class SerialBuffer:
def __init__(self, dev, filename, prefix, timeout=None, line_queue=None):
self.filename = filename
self.dev = dev
if dev:
self.f = open(filename, "wb+")
self.serial = serial.Serial(dev, 115200, timeout=timeout)
else:
self.f = open(filename, "rb")
self.serial = None
self.byte_queue = queue.Queue()
# allow multiple SerialBuffers to share a line queue so you can merge
# servo's CPU and EC streams into one thing to watch the boot/test
# progress on.
if line_queue:
self.line_queue = line_queue
else:
self.line_queue = queue.Queue()
self.prefix = prefix
self.timeout = timeout
self.sentinel = object()
self.closing = False
if self.dev:
self.read_thread = threading.Thread(
target=self.serial_read_thread_loop, daemon=True)
else:
self.read_thread = threading.Thread(
target=self.serial_file_read_thread_loop, daemon=True)
self.read_thread.start()
self.lines_thread = threading.Thread(
target=self.serial_lines_thread_loop, daemon=True)
self.lines_thread.start()
def close(self):
self.closing = True
if self.serial:
self.serial.cancel_read()
self.read_thread.join()
self.lines_thread.join()
if self.serial:
self.serial.close()
# Thread that just reads the bytes from the serial device to try to keep from
# buffer overflowing it. If nothing is received in 1 minute, it finalizes.
def serial_read_thread_loop(self):
greet = "Serial thread reading from %s\n" % self.dev
self.byte_queue.put(greet.encode())
while not self.closing:
try:
b = self.serial.read()
if len(b) == 0:
break
self.byte_queue.put(b)
except Exception as err:
print(self.prefix + str(err))
break
self.byte_queue.put(self.sentinel)
# Thread that just reads the bytes from the file of serial output that some
# other process is appending to.
def serial_file_read_thread_loop(self):
greet = "Serial thread reading from %s\n" % self.filename
self.byte_queue.put(greet.encode())
while not self.closing:
line = self.f.readline()
if line:
self.byte_queue.put(line)
else:
time.sleep(0.1)
self.byte_queue.put(self.sentinel)
# Thread that processes the stream of bytes to 1) log to stdout, 2) log to
# file, 3) add to the queue of lines to be read by program logic
def serial_lines_thread_loop(self):
line = bytearray()
while True:
bytes = self.byte_queue.get(block=True)
if bytes == self.sentinel:
self.read_thread.join()
self.line_queue.put(self.sentinel)
break
# Write our data to the output file if we're the ones reading from
# the serial device
if self.dev:
self.f.write(bytes)
self.f.flush()
for b in bytes:
line.append(b)
if b == b'\n'[0]:
line = line.decode(errors="replace")
time = datetime.now().strftime('%y-%m-%d %H:%M:%S')
print("{endc}{time} {prefix}{line}".format(
time=time, prefix=self.prefix, line=line, endc='\033[0m'), flush=True, end='')
self.line_queue.put(line)
line = bytearray()
def lines(self, timeout=None, phase=None):
start_time = time.monotonic()
while True:
read_timeout = None
if timeout:
read_timeout = timeout - (time.monotonic() - start_time)
if read_timeout <= 0:
print("read timeout waiting for serial during {}".format(phase))
self.close()
break
try:
line = self.line_queue.get(timeout=read_timeout)
except queue.Empty:
print("read timeout waiting for serial during {}".format(phase))
self.close()
break
if line == self.sentinel:
print("End of serial output")
self.lines_thread.join()
break
yield line
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--dev', type=str, help='Serial device')
parser.add_argument('--file', type=str,
help='Filename for serial output', required=True)
parser.add_argument('--prefix', type=str,
help='Prefix for logging serial to stdout', nargs='?')
args = parser.parse_args()
ser = SerialBuffer(args.dev, args.file, args.prefix or "")
for line in ser.lines():
# We're just using this as a logger, so eat the produced lines and drop
# them
pass
if __name__ == '__main__':
main()

View File

@@ -1,41 +0,0 @@
#!/usr/bin/python3
# Copyright © 2020 Christian Gmeiner
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice (including the next
# paragraph) shall be included in all copies or substantial portions of the
# Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
# IN THE SOFTWARE.
#
# Tiny script to read bytes from telnet, and write the output to stdout, with a
# buffer in between so we don't lose serial output from its buffer.
#
import sys
import telnetlib
host = sys.argv[1]
port = sys.argv[2]
tn = telnetlib.Telnet(host, port, 1000000)
while True:
bytes = tn.read_some()
sys.stdout.buffer.write(bytes)
sys.stdout.flush()
tn.close()

View File

@@ -1 +0,0 @@
../bin/ci

View File

@@ -1,7 +0,0 @@
#!/bin/sh
# shellcheck disable=SC1091
set -e
_COMPILER=clang++-15
. compiler-wrapper.sh

View File

@@ -1,7 +0,0 @@
#!/bin/sh
# shellcheck disable=SC1091
set -e
_COMPILER=clang++
. compiler-wrapper.sh

View File

@@ -1,7 +0,0 @@
#!/bin/sh
# shellcheck disable=SC1091
set -e
_COMPILER=clang-15
. compiler-wrapper.sh

View File

@@ -1,7 +0,0 @@
#!/bin/sh
# shellcheck disable=SC1091
set -e
_COMPILER=clang
. compiler-wrapper.sh

View File

@@ -1,7 +0,0 @@
#!/bin/sh
# shellcheck disable=SC1091
set -e
_COMPILER=g++
. compiler-wrapper.sh

View File

@@ -1,7 +0,0 @@
#!/bin/sh
# shellcheck disable=SC1091
set -e
_COMPILER=gcc
. compiler-wrapper.sh

View File

@@ -1,21 +0,0 @@
# shellcheck disable=SC1091
# shellcheck disable=SC2086 # we want word splitting
if command -V ccache >/dev/null 2>/dev/null; then
CCACHE=ccache
else
CCACHE=
fi
if echo "$@" | grep -E 'meson-private/tmp[^ /]*/testfile.c' >/dev/null; then
# Invoked for meson feature check
exec $CCACHE $_COMPILER "$@"
fi
if [ "$(eval printf "'%s'" "\"\${$(($#-1))}\"")" = "-c" ]; then
# Not invoked for linking
exec $CCACHE $_COMPILER "$@"
fi
# Compiler invoked by ninja for linking. Add -Werror to turn compiler warnings into errors
# with LTO. (meson's werror should arguably do this, but meanwhile we need to)
exec $CCACHE $_COMPILER "$@" -Werror

View File

@@ -1,675 +0,0 @@
# Shared between windows and Linux
.build-common:
extends: .container+build-rules
# Cancel job if a newer commit is pushed to the same branch
interruptible: true
# Build jobs don't take more than 1-3 minutes. 5-8 min max on a fresh runner
# without a populated ccache.
# These jobs are never slow, either they finish within reasonable time or
# something has gone wrong and the job will never terminate, so we should
# instead timeout so that the retry mechanism can kick in.
# A few exception are made, see `timeout:` overrides in the rest of this
# file.
timeout: 30m
artifacts:
name: "mesa_${CI_JOB_NAME}"
when: always
paths:
- _build/meson-logs/*.txt
- _build/meson-logs/strace
- shader-db
- artifacts
# Just Linux
.build-linux:
extends: .build-common
variables:
CCACHE_COMPILERCHECK: "content"
CCACHE_COMPRESS: "true"
CCACHE_DIR: /cache/mesa/ccache
# Use ccache transparently, and print stats before/after
before_script:
- !reference [default, before_script]
- |
export PATH="/usr/lib/ccache:$PATH"
export CCACHE_BASEDIR="$PWD"
if test -x /usr/bin/ccache; then
section_start ccache_before "ccache stats before build"
ccache --show-stats
section_end ccache_before
fi
after_script:
- if test -x /usr/bin/ccache; then ccache --show-stats | grep "Hits:"; fi
- !reference [default, after_script]
.build-windows:
extends:
- .build-common
- .windows-docker-tags
cache:
key: ${CI_JOB_NAME}
paths:
- subprojects/packagecache
.meson-build:
extends:
- .build-linux
- .use-debian/x86_64_build
stage: build-x86_64
variables:
LLVM_VERSION: 15
script:
- .gitlab-ci/meson/build.sh
debian-testing:
extends:
- .meson-build
- .ci-deqp-artifacts
variables:
UNWIND: "enabled"
DRI_LOADERS: >
-D glx=dri
-D gbm=enabled
-D egl=enabled
-D platforms=x11,wayland
GALLIUM_ST: >
-D dri3=enabled
-D gallium-nine=true
-D gallium-va=enabled
-D gallium-rusticl=true
GALLIUM_DRIVERS: "swrast,virgl,radeonsi,zink,crocus,iris,i915,r300"
VULKAN_DRIVERS: "swrast,amd,intel,intel_hasvk,virtio,nouveau"
BUILDTYPE: "debugoptimized"
EXTRA_OPTION: >
-D spirv-to-dxil=true
-D valgrind=disabled
-D perfetto=true
-D tools=drm-shim
S3_ARTIFACT_NAME: mesa-x86_64-default-${BUILDTYPE}
LLVM_VERSION: 15
script:
- .gitlab-ci/meson/build.sh
- .gitlab-ci/prepare-artifacts.sh
artifacts:
reports:
junit: artifacts/ci_scripts_report.xml
debian-testing-asan:
extends:
- debian-testing
variables:
C_ARGS: >
-Wno-error=stringop-truncation
EXTRA_OPTION: >
-D b_sanitize=address
-D valgrind=disabled
-D tools=dlclose-skip
-D intel-clc=system
S3_ARTIFACT_NAME: ""
ARTIFACTS_DEBUG_SYMBOLS: 1
# Do a host build for intel-clc (asan complains not being loaded
# as the first library)
HOST_BUILD_OPTIONS: >
-D build-tests=false
-D enable-glcpp-tests=false
-D gallium-opencl=disabled
-D gallium-drivers=
-D vulkan-drivers=
-D video-codecs=
-D glx=disabled
-D platforms=
-D intel-clc=enabled
-D install-intel-clc=true
debian-testing-msan:
# https://github.com/google/sanitizers/wiki/MemorySanitizerLibcxxHowTo
# msan cannot fully work until it's used together with msan libc
extends:
- debian-clang
variables:
# l_undef is incompatible with msan
EXTRA_OPTION:
-D b_sanitize=memory
-D b_lundef=false
-D intel-clc=system
S3_ARTIFACT_NAME: ""
ARTIFACTS_DEBUG_SYMBOLS: 1
# Don't run all the tests yet:
# GLSL has some issues in sexpression reading.
# gtest has issues in its test initialization.
MESON_TEST_ARGS: "--suite glcpp --suite format"
GALLIUM_DRIVERS: "freedreno,iris,nouveau,kmsro,r300,r600,swrast,svga,v3d,vc4,virgl,etnaviv,panfrost,lima,zink,radeonsi,tegra,d3d12,crocus"
VULKAN_DRIVERS: intel,amd,broadcom,virtio
# Do a host build for intel-clc (msan complains about
# uninitialized values in the LLVM libs)
HOST_BUILD_OPTIONS: >
-D build-tests=false
-D enable-glcpp-tests=false
-D gallium-opencl=disabled
-D gallium-drivers=
-D vulkan-drivers=
-D video-codecs=
-D glx=disabled
-D platforms=
-D intel-clc=enabled
-D install-intel-clc=true
debian-build-testing:
extends: .meson-build
variables:
BUILDTYPE: debug
UNWIND: "enabled"
DRI_LOADERS: >
-D glx=dri
-D gbm=enabled
-D egl=enabled
-D platforms=x11,wayland
GALLIUM_ST: >
-D dri3=enabled
-D gallium-extra-hud=true
-D gallium-vdpau=enabled
-D gallium-omx=bellagio
-D gallium-va=enabled
-D gallium-xa=enabled
-D gallium-nine=true
-D gallium-rusticl=false
GALLIUM_DRIVERS: "iris,nouveau,kmsro,r300,r600,freedreno,swrast,svga,v3d,vc4,virgl,etnaviv,panfrost,lima,zink,d3d12,asahi,crocus"
VULKAN_DRIVERS: swrast
EXTRA_OPTION: >
-D spirv-to-dxil=true
-D osmesa=true
-D tools=drm-shim,etnaviv,freedreno,glsl,intel,intel-ui,nir,nouveau,lima,panfrost,asahi
-D b_lto=true
LLVM_VERSION: 15
script: |
section_start lava-pytest "lava-pytest"
.gitlab-ci/lava/lava-pytest.sh
section_switch shellcheck "shellcheck"
.gitlab-ci/run-shellcheck.sh
section_switch yamllint "yamllint"
.gitlab-ci/run-yamllint.sh
section_switch meson "meson"
.gitlab-ci/meson/build.sh
section_switch shader-db "shader-db"
.gitlab-ci/run-shader-db.sh
timeout: 30m
# Test a release build with -Werror so new warnings don't sneak in.
debian-release:
extends: .meson-build
variables:
LLVM_VERSION: 15
UNWIND: "enabled"
C_ARGS: >
-Wno-error=stringop-overread
DRI_LOADERS: >
-D glx=dri
-D gbm=enabled
-D egl=enabled
-D platforms=x11,wayland
GALLIUM_ST: >
-D dri3=enabled
-D gallium-extra-hud=true
-D gallium-vdpau=enabled
-D gallium-omx=disabled
-D gallium-va=enabled
-D gallium-xa=enabled
-D gallium-nine=false
-D gallium-rusticl=false
-D llvm=enabled
GALLIUM_DRIVERS: "i915,iris,nouveau,kmsro,freedreno,r300,svga,swrast,v3d,vc4,virgl,etnaviv,panfrost,lima,zink,d3d12,crocus"
VULKAN_DRIVERS: "amd,imagination-experimental,microsoft-experimental"
EXTRA_OPTION: >
-D spirv-to-dxil=true
-D osmesa=true
-D tools=all
-D intel-clc=enabled
-D intel-rt=enabled
-D imagination-srv=true
BUILDTYPE: "release"
S3_ARTIFACT_NAME: "mesa-x86_64-default-${BUILDTYPE}"
script:
- .gitlab-ci/meson/build.sh
- 'if [ -n "$MESA_CI_PERFORMANCE_ENABLED" ]; then .gitlab-ci/prepare-artifacts.sh; fi'
alpine-build-testing:
extends:
- .meson-build
- .use-alpine/x86_64_build
stage: build-x86_64
variables:
BUILDTYPE: "release"
C_ARGS: >
-Wno-error=cpp
-Wno-error=array-bounds
-Wno-error=stringop-overread
DRI_LOADERS: >
-D glx=disabled
-D gbm=enabled
-D egl=enabled
-D glvnd=false
-D platforms=wayland
LLVM_VERSION: "16"
GALLIUM_DRIVERS: "crocus,etnaviv,freedreno,iris,kmsro,lima,nouveau,panfrost,r300,r600,radeonsi,svga,swrast,tegra,v3d,vc4,virgl,zink"
GALLIUM_ST: >
-D dri3=enabled
-D gallium-extra-hud=true
-D gallium-vdpau=disabled
-D gallium-omx=disabled
-D gallium-va=enabled
-D gallium-xa=disabled
-D gallium-nine=true
-D gallium-rusticl=false
-D gles1=disabled
-D gles2=enabled
-D llvm=enabled
-D microsoft-clc=disabled
-D shared-llvm=enabled
UNWIND: "disabled"
VULKAN_DRIVERS: "amd,broadcom,freedreno,intel,imagination-experimental"
fedora-release:
extends:
- .meson-build
- .use-fedora/x86_64_build
variables:
BUILDTYPE: "release"
C_LINK_ARGS: >
-Wno-error=stringop-overflow
-Wno-error=stringop-overread
CPP_ARGS: >
-Wno-error=dangling-reference
-Wno-error=overloaded-virtual
CPP_LINK_ARGS: >
-Wno-error=stringop-overflow
-Wno-error=stringop-overread
DRI_LOADERS: >
-D glx=dri
-D gbm=enabled
-D egl=enabled
-D glvnd=true
-D platforms=x11,wayland
EXTRA_OPTION: >
-D b_lto=true
-D osmesa=true
-D selinux=true
-D tools=drm-shim,etnaviv,freedreno,glsl,intel,nir,nouveau,lima,panfrost,imagination
-D vulkan-layers=device-select,overlay
-D intel-rt=enabled
-D imagination-srv=true
-D teflon=true
GALLIUM_DRIVERS: "crocus,etnaviv,freedreno,i915,iris,kmsro,lima,nouveau,panfrost,r300,r600,radeonsi,svga,swrast,tegra,v3d,vc4,virgl,zink"
GALLIUM_ST: >
-D dri3=enabled
-D gallium-extra-hud=true
-D gallium-vdpau=enabled
-D gallium-omx=disabled
-D gallium-va=enabled
-D gallium-xa=enabled
-D gallium-nine=false
-D gallium-rusticl=true
-D gles1=disabled
-D gles2=enabled
-D llvm=enabled
-D microsoft-clc=disabled
-D shared-llvm=enabled
LLVM_VERSION: ""
UNWIND: "disabled"
VULKAN_DRIVERS: "amd,broadcom,freedreno,imagination-experimental,intel,intel_hasvk"
debian-android:
extends:
- .meson-cross
- .use-debian/android_build
- .ci-deqp-artifacts
variables:
BUILDTYPE: debug
UNWIND: "disabled"
C_ARGS: >
-Wno-error=asm-operand-widths
-Wno-error=constant-conversion
-Wno-error=enum-conversion
-Wno-error=initializer-overrides
-Wno-error=sometimes-uninitialized
CPP_ARGS: >
-Wno-error=c99-designator
-Wno-error=unused-variable
-Wno-error=unused-but-set-variable
-Wno-error=self-assign
DRI_LOADERS: >
-D glx=disabled
-D gbm=disabled
-D egl=enabled
-D platforms=android
EXTRA_OPTION: >
-D android-stub=true
-D llvm=disabled
-D platform-sdk-version=33
-D valgrind=disabled
-D android-libbacktrace=disabled
-D intel-clc=system
GALLIUM_ST: >
-D dri3=disabled
-D gallium-vdpau=disabled
-D gallium-omx=disabled
-D gallium-va=disabled
-D gallium-xa=disabled
-D gallium-nine=false
-D gallium-rusticl=false
LLVM_VERSION: "15"
PKG_CONFIG_LIBDIR: "/disable/non/android/system/pc/files"
HOST_BUILD_OPTIONS: >
-D build-tests=false
-D enable-glcpp-tests=false
-D gallium-opencl=disabled
-D gallium-drivers=
-D vulkan-drivers=
-D video-codecs=
-D glx=disabled
-D platforms=
-D intel-clc=enabled
-D install-intel-clc=true
ARTIFACTS_DEBUG_SYMBOLS: 1
S3_ARTIFACT_NAME: mesa-x86_64-android-${BUILDTYPE}
script:
- CROSS=aarch64-linux-android GALLIUM_DRIVERS=etnaviv,freedreno,lima,panfrost,vc4,v3d VULKAN_DRIVERS=freedreno,broadcom,virtio .gitlab-ci/meson/build.sh
# x86_64 build:
# Can't do Intel because gen_decoder.c currently requires libexpat, which
# is not a dependency that AOSP wants to accept. Can't do Radeon Gallium
# drivers because they requires LLVM, which we don't have an Android build
# of.
- CROSS=x86_64-linux-android GALLIUM_DRIVERS=iris,virgl VULKAN_DRIVERS=amd,intel .gitlab-ci/meson/build.sh
- .gitlab-ci/prepare-artifacts.sh
.meson-cross:
extends:
- .meson-build
stage: build-misc
variables:
UNWIND: "disabled"
DRI_LOADERS: >
-D glx=dri
-D gbm=enabled
-D egl=enabled
-D platforms=x11,wayland
-D osmesa=false
GALLIUM_ST: >
-D dri3=enabled
-D gallium-vdpau=disabled
-D gallium-omx=disabled
-D gallium-va=disabled
-D gallium-xa=disabled
-D gallium-nine=false
.meson-arm:
extends:
- .meson-cross
- .use-debian/arm64_build
needs:
- debian/arm64_build
variables:
VULKAN_DRIVERS: freedreno,broadcom
GALLIUM_DRIVERS: "etnaviv,freedreno,kmsro,lima,nouveau,panfrost,swrast,tegra,v3d,vc4,zink"
BUILDTYPE: "debugoptimized"
tags:
- aarch64
debian-arm32:
extends:
- .meson-arm
- .ci-deqp-artifacts
variables:
CROSS: armhf
EXTRA_OPTION: >
-D llvm=disabled
-D valgrind=disabled
S3_ARTIFACT_NAME: mesa-arm32-default-${BUILDTYPE}
# The strip command segfaults, failing to strip the binary and leaving
# tempfiles in our artifacts.
ARTIFACTS_DEBUG_SYMBOLS: 1
script:
- .gitlab-ci/meson/build.sh
- .gitlab-ci/prepare-artifacts.sh
debian-arm32-asan:
extends:
- debian-arm32
variables:
EXTRA_OPTION: >
-D llvm=disabled
-D b_sanitize=address
-D valgrind=disabled
-D tools=dlclose-skip
ARTIFACTS_DEBUG_SYMBOLS: 1
S3_ARTIFACT_NAME: mesa-arm32-asan-${BUILDTYPE}
MESON_TEST_ARGS: "--no-suite mesa:compiler --no-suite mesa:util"
debian-arm64:
extends:
- .meson-arm
- .ci-deqp-artifacts
variables:
C_ARGS: >
-Wno-error=array-bounds
-Wno-error=stringop-truncation
VULKAN_DRIVERS: "freedreno,broadcom,panfrost,imagination-experimental"
EXTRA_OPTION: >
-D llvm=disabled
-D valgrind=disabled
-D imagination-srv=true
-D perfetto=true
-D freedreno-kmds=msm,virtio
-D teflon=true
S3_ARTIFACT_NAME: mesa-arm64-default-${BUILDTYPE}
script:
- .gitlab-ci/meson/build.sh
- .gitlab-ci/prepare-artifacts.sh
debian-arm64-asan:
extends:
- debian-arm64
variables:
EXTRA_OPTION: >
-D llvm=disabled
-D b_sanitize=address
-D valgrind=disabled
-D tools=dlclose-skip
ARTIFACTS_DEBUG_SYMBOLS: 1
S3_ARTIFACT_NAME: mesa-arm64-asan-${BUILDTYPE}
MESON_TEST_ARGS: "--no-suite mesa:compiler"
debian-arm64-build-test:
extends:
- .meson-arm
- .ci-deqp-artifacts
variables:
VULKAN_DRIVERS: "amd"
EXTRA_OPTION: >
-Dtools=panfrost,imagination
debian-arm64-release:
extends:
- debian-arm64
variables:
BUILDTYPE: release
S3_ARTIFACT_NAME: mesa-arm64-default-${BUILDTYPE}
C_ARGS: >
-Wno-error=array-bounds
-Wno-error=stringop-truncation
-Wno-error=stringop-overread
script:
- .gitlab-ci/meson/build.sh
- 'if [ -n "$MESA_CI_PERFORMANCE_ENABLED" ]; then .gitlab-ci/prepare-artifacts.sh; fi'
debian-clang:
extends: .meson-build
variables:
BUILDTYPE: debug
LLVM_VERSION: 15
UNWIND: "enabled"
C_ARGS: >
-Wno-error=constant-conversion
-Wno-error=enum-conversion
-Wno-error=initializer-overrides
-Wno-error=sometimes-uninitialized
-Werror=misleading-indentation
CPP_ARGS: >
-Wno-error=c99-designator
-Wno-error=overloaded-virtual
-Wno-error=tautological-constant-out-of-range-compare
-Wno-error=unused-private-field
DRI_LOADERS: >
-D glx=dri
-D gbm=enabled
-D egl=enabled
-D glvnd=true
-D platforms=x11,wayland
GALLIUM_ST: >
-D dri3=enabled
-D gallium-extra-hud=true
-D gallium-vdpau=enabled
-D gallium-omx=bellagio
-D gallium-va=enabled
-D gallium-xa=enabled
-D gallium-nine=true
-D gles1=enabled
-D gles2=enabled
-D llvm=enabled
-D microsoft-clc=disabled
-D shared-llvm=enabled
-D opencl-spirv=true
-D shared-glapi=enabled
GALLIUM_DRIVERS: "iris,nouveau,kmsro,r300,r600,freedreno,swrast,svga,v3d,vc4,virgl,etnaviv,panfrost,lima,zink,radeonsi,tegra,d3d12,crocus,i915,asahi"
VULKAN_DRIVERS: intel,amd,freedreno,broadcom,virtio,swrast,panfrost,imagination-experimental,microsoft-experimental,nouveau
EXTRA_OPTION:
-D spirv-to-dxil=true
-D osmesa=true
-D imagination-srv=true
-D tools=drm-shim,etnaviv,freedreno,glsl,intel,intel-ui,nir,nouveau,lima,panfrost,asahi,imagination
-D vulkan-layers=device-select,overlay
-D build-aco-tests=true
-D intel-clc=enabled
-D intel-rt=enabled
-D imagination-srv=true
-D teflon=true
CC: clang-${LLVM_VERSION}
CXX: clang++-${LLVM_VERSION}
debian-clang-release:
extends: debian-clang
variables:
BUILDTYPE: "release"
DRI_LOADERS: >
-D glx=xlib
-D platforms=x11,wayland
GALLIUM_ST: >
-D dri3=enabled
-D gallium-extra-hud=true
-D gallium-vdpau=enabled
-D gallium-omx=bellagio
-D gallium-va=enabled
-D gallium-xa=enabled
-D gallium-nine=true
-D gles1=disabled
-D gles2=disabled
-D llvm=enabled
-D microsoft-clc=disabled
-D shared-llvm=enabled
-D opencl-spirv=true
-D shared-glapi=disabled
windows-msvc:
extends:
- .build-windows
- .use-windows_build_msvc
- .windows-build-rules
stage: build-misc
script:
- pwsh -ExecutionPolicy RemoteSigned .\.gitlab-ci\windows\mesa_build.ps1
artifacts:
paths:
- _build/meson-logs/*.txt
- _install/
debian-vulkan:
extends: .meson-build
variables:
BUILDTYPE: debug
LLVM_VERSION: 15
UNWIND: "disabled"
DRI_LOADERS: >
-D glx=disabled
-D gbm=disabled
-D egl=disabled
-D opengl=false
-D gles1=disabled
-D gles2=disabled
-D platforms=x11,wayland
-D osmesa=false
GALLIUM_ST: >
-D dri3=enabled
-D gallium-vdpau=disabled
-D gallium-omx=disabled
-D gallium-va=disabled
-D gallium-xa=disabled
-D gallium-nine=false
-D gallium-rusticl=false
-D b_sanitize=undefined
-D c_args=-fno-sanitize-recover=all
-D cpp_args=-fno-sanitize-recover=all
UBSAN_OPTIONS: "print_stacktrace=1"
VULKAN_DRIVERS: amd,broadcom,freedreno,intel,intel_hasvk,panfrost,virtio,imagination-experimental,microsoft-experimental,nouveau
EXTRA_OPTION: >
-D vulkan-layers=device-select,overlay
-D build-aco-tests=true
-D intel-rt=disabled
-D imagination-srv=true
debian-x86_32:
extends:
- .meson-cross
- .use-debian/x86_32_build
variables:
BUILDTYPE: debug
CROSS: i386
VULKAN_DRIVERS: intel,amd,swrast,virtio
GALLIUM_DRIVERS: "iris,nouveau,r300,r600,radeonsi,swrast,virgl,zink,crocus,d3d12"
LLVM_VERSION: 15
EXTRA_OPTION: >
-D vulkan-layers=device-select,overlay
-D intel-clc=system
HOST_BUILD_OPTIONS: >
-D build-tests=false
-D enable-glcpp-tests=false
-D gallium-opencl=disabled
-D gallium-drivers=
-D vulkan-drivers=
-D video-codecs=
-D glx=disabled
-D platforms=
-D intel-clc=enabled
-D install-intel-clc=true
debian-s390x:
extends:
- debian-ppc64el
- .use-debian/s390x_build
- .s390x-rules
tags:
- kvm
variables:
CROSS: s390x
GALLIUM_DRIVERS: "swrast,zink"
LLVM_VERSION: 15
VULKAN_DRIVERS: "swrast"
debian-ppc64el:
extends:
- .meson-cross
- .use-debian/ppc64el_build
- .ppc64el-rules
variables:
BUILDTYPE: debug
CROSS: ppc64el
GALLIUM_DRIVERS: "nouveau,radeonsi,swrast,virgl,zink"
VULKAN_DRIVERS: "amd,swrast"

View File

@@ -1,35 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2035
# shellcheck disable=SC2061
# shellcheck disable=SC2086 # we want word splitting
while true; do
devcds=$(find /sys/devices/virtual/devcoredump/ -name data 2>/dev/null)
for i in $devcds; do
echo "Found a devcoredump at $i."
if cp $i /results/first.devcore; then
echo 1 > $i
echo "Saved to the job artifacts at /first.devcore"
exit 0
fi
done
i915_error_states=$(find /sys/devices/ -path */drm/card*/error)
for i in $i915_error_states; do
tmpfile=$(mktemp)
cp "$i" "$tmpfile"
filesize=$(stat --printf="%s" "$tmpfile")
# Does the file contain "No error state collected" ?
if [ "$filesize" = 25 ]; then
rm "$tmpfile"
else
echo "Found an i915 error state at $i size=$filesize."
if cp "$tmpfile" /results/first.i915_error_state; then
rm "$tmpfile"
echo 1 > "$i"
echo "Saved to the job artifacts at /first.i915_error_state"
exit 0
fi
fi
done
sleep 10
done

View File

@@ -1,131 +0,0 @@
#!/bin/bash
for var in \
ACO_DEBUG \
ARTIFACTS_BASE_URL \
ASAN_OPTIONS \
BASE_SYSTEM_FORK_HOST_PREFIX \
BASE_SYSTEM_MAINLINE_HOST_PREFIX \
CI_COMMIT_BRANCH \
CI_COMMIT_REF_NAME \
CI_COMMIT_TITLE \
CI_JOB_ID \
CI_JOB_JWT_FILE \
CI_JOB_STARTED_AT \
CI_JOB_NAME \
CI_JOB_URL \
CI_MERGE_REQUEST_SOURCE_BRANCH_NAME \
CI_MERGE_REQUEST_TITLE \
CI_NODE_INDEX \
CI_NODE_TOTAL \
CI_PAGES_DOMAIN \
CI_PIPELINE_ID \
CI_PIPELINE_URL \
CI_PROJECT_DIR \
CI_PROJECT_NAME \
CI_PROJECT_PATH \
CI_PROJECT_ROOT_NAMESPACE \
CI_RUNNER_DESCRIPTION \
CI_SERVER_URL \
CROSVM_GALLIUM_DRIVER \
CROSVM_GPU_ARGS \
CURRENT_SECTION \
DEQP_BIN_DIR \
DEQP_CONFIG \
DEQP_EXPECTED_RENDERER \
DEQP_FRACTION \
DEQP_HEIGHT \
DEQP_RESULTS_DIR \
DEQP_RUNNER_OPTIONS \
DEQP_SUITE \
DEQP_TEMP_DIR \
DEQP_VER \
DEQP_WIDTH \
DEVICE_NAME \
DRIVER_NAME \
EGL_PLATFORM \
ETNA_MESA_DEBUG \
FDO_CI_CONCURRENT \
FDO_UPSTREAM_REPO \
FD_MESA_DEBUG \
FLAKES_CHANNEL \
FREEDRENO_HANGCHECK_MS \
GALLIUM_DRIVER \
GALLIVM_PERF \
GPU_VERSION \
GTEST \
GTEST_FAILS \
GTEST_FRACTION \
GTEST_RESULTS_DIR \
GTEST_RUNNER_OPTIONS \
GTEST_SKIPS \
HWCI_FREQ_MAX \
HWCI_KERNEL_MODULES \
HWCI_KVM \
HWCI_START_WESTON \
HWCI_START_XORG \
HWCI_TEST_SCRIPT \
IR3_SHADER_DEBUG \
JOB_ARTIFACTS_BASE \
JOB_RESULTS_PATH \
JOB_ROOTFS_OVERLAY_PATH \
KERNEL_IMAGE_BASE \
KERNEL_IMAGE_NAME \
LD_LIBRARY_PATH \
LP_NUM_THREADS \
MESA_BASE_TAG \
MESA_BUILD_PATH \
MESA_DEBUG \
MESA_GLES_VERSION_OVERRIDE \
MESA_GLSL_VERSION_OVERRIDE \
MESA_GL_VERSION_OVERRIDE \
MESA_IMAGE \
MESA_IMAGE_PATH \
MESA_IMAGE_TAG \
MESA_LOADER_DRIVER_OVERRIDE \
MESA_TEMPLATES_COMMIT \
MESA_VK_IGNORE_CONFORMANCE_WARNING \
S3_HOST \
S3_RESULTS_UPLOAD \
NIR_DEBUG \
PAN_I_WANT_A_BROKEN_VULKAN_DRIVER \
PAN_MESA_DEBUG \
PANVK_DEBUG \
PIGLIT_FRACTION \
PIGLIT_NO_WINDOW \
PIGLIT_OPTIONS \
PIGLIT_PLATFORM \
PIGLIT_PROFILES \
PIGLIT_REPLAY_ARTIFACTS_BASE_URL \
PIGLIT_REPLAY_DEVICE_NAME \
PIGLIT_REPLAY_EXTRA_ARGS \
PIGLIT_REPLAY_LOOP_TIMES \
PIGLIT_REPLAY_REFERENCE_IMAGES_BASE \
PIGLIT_REPLAY_SUBCOMMAND \
PIGLIT_RESULTS \
PIGLIT_TESTS \
PIGLIT_TRACES_FILE \
PIPELINE_ARTIFACTS_BASE \
RADEON_DEBUG \
RADV_DEBUG \
RADV_PERFTEST \
SKQP_ASSETS_DIR \
SKQP_BACKENDS \
TU_DEBUG \
USE_ANGLE \
VIRGL_HOST_API \
WAFFLE_PLATFORM \
VK_CPU \
VK_DRIVER \
VK_ICD_FILENAMES \
VKD3D_PROTON_RESULTS \
VKD3D_CONFIG \
VKD3D_TEST_EXCLUDE \
ZINK_DESCRIPTORS \
ZINK_DEBUG \
LVP_POISON_MEMORY \
; do
if [ -n "${!var+x}" ]; then
echo "export $var=${!var@Q}"
fi
done

View File

@@ -1,25 +0,0 @@
#!/bin/sh
# Very early init, used to make sure devices and network are set up and
# reachable.
set -ex
cd /
findmnt --mountpoint /proc || mount -t proc none /proc
findmnt --mountpoint /sys || mount -t sysfs none /sys
mount -t debugfs none /sys/kernel/debug
findmnt --mountpoint /dev || mount -t devtmpfs none /dev
mkdir -p /dev/pts
mount -t devpts devpts /dev/pts
mkdir /dev/shm
mount -t tmpfs -o noexec,nodev,nosuid tmpfs /dev/shm
mount -t tmpfs tmpfs /tmp
echo "nameserver 8.8.8.8" > /etc/resolv.conf
[ -z "$NFS_SERVER_IP" ] || echo "$NFS_SERVER_IP caching-proxy" >> /etc/hosts
# Set the time so we can validate certificates before we fetch anything;
# however as not all DUTs have network, make this non-fatal.
for _ in 1 2 3; do sntp -sS pool.ntp.org && break || sleep 2; done || true

View File

@@ -1,235 +0,0 @@
#!/bin/bash
# shellcheck disable=SC1090
# shellcheck disable=SC1091
# shellcheck disable=SC2086 # we want word splitting
# shellcheck disable=SC2155
# Second-stage init, used to set up devices and our job environment before
# running tests.
shopt -s extglob
# Make sure to kill itself and all the children process from this script on
# exiting, since any console output may interfere with LAVA signals handling,
# which based on the log console.
cleanup() {
if [ "$BACKGROUND_PIDS" = "" ]; then
return 0
fi
set +x
echo "Killing all child processes"
for pid in $BACKGROUND_PIDS
do
kill "$pid" 2>/dev/null || true
done
# Sleep just a little to give enough time for subprocesses to be gracefully
# killed. Then apply a SIGKILL if necessary.
sleep 5
for pid in $BACKGROUND_PIDS
do
kill -9 "$pid" 2>/dev/null || true
done
BACKGROUND_PIDS=
set -x
}
trap cleanup INT TERM EXIT
# Space separated values with the PIDS of the processes started in the
# background by this script
BACKGROUND_PIDS=
for path in '/dut-env-vars.sh' '/set-job-env-vars.sh' './set-job-env-vars.sh'; do
[ -f "$path" ] && source "$path"
done
. "$SCRIPTS_DIR"/setup-test-env.sh
set -ex
# Set up any devices required by the jobs
[ -z "$HWCI_KERNEL_MODULES" ] || {
echo -n $HWCI_KERNEL_MODULES | xargs -d, -n1 /usr/sbin/modprobe
}
# Set up ZRAM
HWCI_ZRAM_SIZE=2G
if /sbin/zramctl --find --size $HWCI_ZRAM_SIZE -a zstd; then
mkswap /dev/zram0
swapon /dev/zram0
echo "zram: $HWCI_ZRAM_SIZE activated"
else
echo "zram: skipping, not supported"
fi
#
# Load the KVM module specific to the detected CPU virtualization extensions:
# - vmx for Intel VT
# - svm for AMD-V
#
# Additionally, download the kernel image to boot the VM via HWCI_TEST_SCRIPT.
#
if [ "$HWCI_KVM" = "true" ]; then
unset KVM_KERNEL_MODULE
{
grep -qs '\bvmx\b' /proc/cpuinfo && KVM_KERNEL_MODULE=kvm_intel
} || {
grep -qs '\bsvm\b' /proc/cpuinfo && KVM_KERNEL_MODULE=kvm_amd
}
{
[ -z "${KVM_KERNEL_MODULE}" ] && \
echo "WARNING: Failed to detect CPU virtualization extensions"
} || \
modprobe ${KVM_KERNEL_MODULE}
mkdir -p /lava-files
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-o "/lava-files/${KERNEL_IMAGE_NAME}" \
"${KERNEL_IMAGE_BASE}/amd64/${KERNEL_IMAGE_NAME}"
fi
# Fix prefix confusion: the build installs to $CI_PROJECT_DIR, but we expect
# it in /install
ln -sf $CI_PROJECT_DIR/install /install
export LD_LIBRARY_PATH=/install/lib
export LIBGL_DRIVERS_PATH=/install/lib/dri
# https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22495#note_1876691
# The navi21 boards seem to have trouble with ld.so.cache, so try explicitly
# telling it to look in /usr/local/lib.
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib
# Store Mesa's disk cache under /tmp, rather than sending it out over NFS.
export XDG_CACHE_HOME=/tmp
# Make sure Python can find all our imports
export PYTHONPATH=$(python3 -c "import sys;print(\":\".join(sys.path))")
# If we need to specify a driver, it means several drivers could pick up this gpu;
# ensure that the other driver can't accidentally be used
if [ -n "$MESA_LOADER_DRIVER_OVERRIDE" ]; then
rm /install/lib/dri/!($MESA_LOADER_DRIVER_OVERRIDE)_dri.so
fi
ls -1 /install/lib/dri/*_dri.so
if [ "$HWCI_FREQ_MAX" = "true" ]; then
# Ensure initialization of the DRM device (needed by MSM)
head -0 /dev/dri/renderD128
# Disable GPU frequency scaling
DEVFREQ_GOVERNOR=$(find /sys/devices -name governor | grep gpu || true)
test -z "$DEVFREQ_GOVERNOR" || echo performance > $DEVFREQ_GOVERNOR || true
# Disable CPU frequency scaling
echo performance | tee -a /sys/devices/system/cpu/cpufreq/policy*/scaling_governor || true
# Disable GPU runtime power management
GPU_AUTOSUSPEND=$(find /sys/devices -name autosuspend_delay_ms | grep gpu | head -1)
test -z "$GPU_AUTOSUSPEND" || echo -1 > $GPU_AUTOSUSPEND || true
# Lock Intel GPU frequency to 70% of the maximum allowed by hardware
# and enable throttling detection & reporting.
# Additionally, set the upper limit for CPU scaling frequency to 65% of the
# maximum permitted, as an additional measure to mitigate thermal throttling.
/intel-gpu-freq.sh -s 70% --cpu-set-max 65% -g all -d
fi
# Start a little daemon to capture sysfs records and produce a JSON file
if [ -x /kdl.sh ]; then
echo "launch kdl.sh!"
/kdl.sh &
BACKGROUND_PIDS="$! $BACKGROUND_PIDS"
else
echo "kdl.sh not found!"
fi
# Increase freedreno hangcheck timer because it's right at the edge of the
# spilling tests timing out (and some traces, too)
if [ -n "$FREEDRENO_HANGCHECK_MS" ]; then
echo $FREEDRENO_HANGCHECK_MS | tee -a /sys/kernel/debug/dri/128/hangcheck_period_ms
fi
# Start a little daemon to capture the first devcoredump we encounter. (They
# expire after 5 minutes, so we poll for them).
if [ -x /capture-devcoredump.sh ]; then
/capture-devcoredump.sh &
BACKGROUND_PIDS="$! $BACKGROUND_PIDS"
fi
# If we want Xorg to be running for the test, then we start it up before the
# HWCI_TEST_SCRIPT because we need to use xinit to start X (otherwise
# without using -displayfd you can race with Xorg's startup), but xinit will eat
# your client's return code
if [ -n "$HWCI_START_XORG" ]; then
echo "touch /xorg-started; sleep 100000" > /xorg-script
env \
VK_ICD_FILENAMES="/install/share/vulkan/icd.d/${VK_DRIVER}_icd.$(uname -m).json" \
xinit /bin/sh /xorg-script -- /usr/bin/Xorg -noreset -s 0 -dpms -logfile /Xorg.0.log &
BACKGROUND_PIDS="$! $BACKGROUND_PIDS"
# Wait for xorg to be ready for connections.
for _ in 1 2 3 4 5; do
if [ -e /xorg-started ]; then
break
fi
sleep 5
done
export DISPLAY=:0
fi
if [ -n "$HWCI_START_WESTON" ]; then
WESTON_X11_SOCK="/tmp/.X11-unix/X0"
if [ -n "$HWCI_START_XORG" ]; then
echo "Please consider dropping HWCI_START_XORG and instead using Weston XWayland for testing."
WESTON_X11_SOCK="/tmp/.X11-unix/X1"
fi
export WAYLAND_DISPLAY=wayland-0
# Display server is Weston Xwayland when HWCI_START_XORG is not set or Xorg when it's
export DISPLAY=:0
mkdir -p /tmp/.X11-unix
env \
VK_ICD_FILENAMES="/install/share/vulkan/icd.d/${VK_DRIVER}_icd.$(uname -m).json" \
weston -Bheadless-backend.so --use-gl -Swayland-0 --xwayland --idle-time=0 &
BACKGROUND_PIDS="$! $BACKGROUND_PIDS"
while [ ! -S "$WESTON_X11_SOCK" ]; do sleep 1; done
fi
set +e
bash -c ". $SCRIPTS_DIR/setup-test-env.sh && $HWCI_TEST_SCRIPT"
EXIT_CODE=$?
set -e
# Let's make sure the results are always stored in current working directory
mv -f ${CI_PROJECT_DIR}/results ./ 2>/dev/null || true
[ ${EXIT_CODE} -ne 0 ] || rm -rf results/trace/"$PIGLIT_REPLAY_DEVICE_NAME"
# Make sure that capture-devcoredump is done before we start trying to tar up
# artifacts -- if it's writing while tar is reading, tar will throw an error and
# kill the job.
cleanup
# upload artifacts
if [ -n "$S3_RESULTS_UPLOAD" ]; then
tar --zstd -cf results.tar.zst results/;
ci-fairy s3cp --token-file "${CI_JOB_JWT_FILE}" results.tar.zst https://"$S3_RESULTS_UPLOAD"/results.tar.zst;
fi
# We still need to echo the hwci: mesa message, as some scripts rely on it, such
# as the python ones inside the bare-metal folder
[ ${EXIT_CODE} -eq 0 ] && RESULT=pass || RESULT=fail
set +x
# Print the final result; both bare-metal and LAVA look for this string to get
# the result of our run, so try really hard to get it out rather than losing
# the run. The device gets shut down right at this point, and a630 seems to
# enjoy corrupting the last line of serial output before shutdown.
for _ in $(seq 0 3); do echo "hwci: mesa: $RESULT"; sleep 1; echo; done
exit $EXIT_CODE

View File

@@ -1,768 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2013
# shellcheck disable=SC2015
# shellcheck disable=SC2034
# shellcheck disable=SC2046
# shellcheck disable=SC2059
# shellcheck disable=SC2086 # we want word splitting
# shellcheck disable=SC2154
# shellcheck disable=SC2155
# shellcheck disable=SC2162
# shellcheck disable=SC2229
#
# This is an utility script to manage Intel GPU frequencies.
# It can be used for debugging performance problems or trying to obtain a stable
# frequency while benchmarking.
#
# Note the Intel i915 GPU driver allows to change the minimum, maximum and boost
# frequencies in steps of 50 MHz via:
#
# /sys/class/drm/card<n>/<freq_info>
#
# Where <n> is the DRM card index and <freq_info> one of the following:
#
# - gt_max_freq_mhz (enforced maximum freq)
# - gt_min_freq_mhz (enforced minimum freq)
# - gt_boost_freq_mhz (enforced boost freq)
#
# The hardware capabilities can be accessed via:
#
# - gt_RP0_freq_mhz (supported maximum freq)
# - gt_RPn_freq_mhz (supported minimum freq)
# - gt_RP1_freq_mhz (most efficient freq)
#
# The current frequency can be read from:
# - gt_act_freq_mhz (the actual GPU freq)
# - gt_cur_freq_mhz (the last requested freq)
#
# Also note that in addition to GPU management, the script offers the
# possibility to adjust CPU operating frequencies. However, this is currently
# limited to just setting the maximum scaling frequency as percentage of the
# maximum frequency allowed by the hardware.
#
# Copyright (C) 2022 Collabora Ltd.
# Author: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
#
# SPDX-License-Identifier: MIT
#
#
# Constants
#
# GPU
DRM_FREQ_SYSFS_PATTERN="/sys/class/drm/card%d/gt_%s_freq_mhz"
ENF_FREQ_INFO="max min boost"
CAP_FREQ_INFO="RP0 RPn RP1"
ACT_FREQ_INFO="act cur"
THROTT_DETECT_SLEEP_SEC=2
THROTT_DETECT_PID_FILE_PATH=/tmp/thrott-detect.pid
# CPU
CPU_SYSFS_PREFIX=/sys/devices/system/cpu
CPU_PSTATE_SYSFS_PATTERN="${CPU_SYSFS_PREFIX}/intel_pstate/%s"
CPU_FREQ_SYSFS_PATTERN="${CPU_SYSFS_PREFIX}/cpu%s/cpufreq/%s_freq"
CAP_CPU_FREQ_INFO="cpuinfo_max cpuinfo_min"
ENF_CPU_FREQ_INFO="scaling_max scaling_min"
ACT_CPU_FREQ_INFO="scaling_cur"
#
# Global variables.
#
unset INTEL_DRM_CARD_INDEX
unset GET_ACT_FREQ GET_ENF_FREQ GET_CAP_FREQ
unset SET_MIN_FREQ SET_MAX_FREQ
unset MONITOR_FREQ
unset CPU_SET_MAX_FREQ
unset DETECT_THROTT
unset DRY_RUN
#
# Simple printf based stderr logger.
#
log() {
local msg_type=$1
shift
printf "%s: %s: " "${msg_type}" "${0##*/}" >&2
printf "$@" >&2
printf "\n" >&2
}
#
# Helper to print sysfs path for the given card index and freq info.
#
# arg1: Frequency info sysfs name, one of *_FREQ_INFO constants above
# arg2: Video card index, defaults to INTEL_DRM_CARD_INDEX
#
print_freq_sysfs_path() {
printf ${DRM_FREQ_SYSFS_PATTERN} "${2:-${INTEL_DRM_CARD_INDEX}}" "$1"
}
#
# Helper to set INTEL_DRM_CARD_INDEX for the first identified Intel video card.
#
identify_intel_gpu() {
local i=0 vendor path
while [ ${i} -lt 16 ]; do
[ -c "/dev/dri/card$i" ] || {
i=$((i + 1))
continue
}
path=$(print_freq_sysfs_path "" ${i})
path=${path%/*}/device/vendor
[ -r "${path}" ] && read vendor < "${path}" && \
[ "${vendor}" = "0x8086" ] && INTEL_DRM_CARD_INDEX=$i && return 0
i=$((i + 1))
done
return 1
}
#
# Read the specified freq info from sysfs.
#
# arg1: Flag (y/n) to also enable printing the freq info.
# arg2...: Frequency info sysfs name(s), see *_FREQ_INFO constants above
# return: Global variable(s) FREQ_${arg} containing the requested information
#
read_freq_info() {
local var val info path print=0 ret=0
[ "$1" = "y" ] && print=1
shift
while [ $# -gt 0 ]; do
info=$1
shift
var=FREQ_${info}
path=$(print_freq_sysfs_path "${info}")
[ -r ${path} ] && read ${var} < ${path} || {
log ERROR "Failed to read freq info from: %s" "${path}"
ret=1
continue
}
[ -n "${var}" ] || {
log ERROR "Got empty freq info from: %s" "${path}"
ret=1
continue
}
[ ${print} -eq 1 ] && {
eval val=\$${var}
printf "%6s: %4s MHz\n" "${info}" "${val}"
}
done
return ${ret}
}
#
# Display requested info.
#
print_freq_info() {
local req_freq
[ -n "${GET_CAP_FREQ}" ] && {
printf "* Hardware capabilities\n"
read_freq_info y ${CAP_FREQ_INFO}
printf "\n"
}
[ -n "${GET_ENF_FREQ}" ] && {
printf "* Enforcements\n"
read_freq_info y ${ENF_FREQ_INFO}
printf "\n"
}
[ -n "${GET_ACT_FREQ}" ] && {
printf "* Actual\n"
read_freq_info y ${ACT_FREQ_INFO}
printf "\n"
}
}
#
# Helper to print frequency value as requested by user via '-s, --set' option.
# arg1: user requested freq value
#
compute_freq_set() {
local val
case "$1" in
+)
val=${FREQ_RP0}
;;
-)
val=${FREQ_RPn}
;;
*%)
val=$((${1%?} * FREQ_RP0 / 100))
# Adjust freq to comply with 50 MHz increments
val=$((val / 50 * 50))
;;
*[!0-9]*)
log ERROR "Cannot set freq to invalid value: %s" "$1"
return 1
;;
"")
log ERROR "Cannot set freq to unspecified value"
return 1
;;
*)
# Adjust freq to comply with 50 MHz increments
val=$(($1 / 50 * 50))
;;
esac
printf "%s" "${val}"
}
#
# Helper for set_freq().
#
set_freq_max() {
log INFO "Setting GPU max freq to %s MHz" "${SET_MAX_FREQ}"
read_freq_info n min || return $?
[ ${SET_MAX_FREQ} -gt ${FREQ_RP0} ] && {
log ERROR "Cannot set GPU max freq (%s) to be greater than hw max freq (%s)" \
"${SET_MAX_FREQ}" "${FREQ_RP0}"
return 1
}
[ ${SET_MAX_FREQ} -lt ${FREQ_RPn} ] && {
log ERROR "Cannot set GPU max freq (%s) to be less than hw min freq (%s)" \
"${SET_MIN_FREQ}" "${FREQ_RPn}"
return 1
}
[ ${SET_MAX_FREQ} -lt ${FREQ_min} ] && {
log ERROR "Cannot set GPU max freq (%s) to be less than min freq (%s)" \
"${SET_MAX_FREQ}" "${FREQ_min}"
return 1
}
[ -z "${DRY_RUN}" ] || return 0
if ! printf "%s" ${SET_MAX_FREQ} | tee $(print_freq_sysfs_path max) \
$(print_freq_sysfs_path boost) > /dev/null;
then
log ERROR "Failed to set GPU max frequency"
return 1
fi
}
#
# Helper for set_freq().
#
set_freq_min() {
log INFO "Setting GPU min freq to %s MHz" "${SET_MIN_FREQ}"
read_freq_info n max || return $?
[ ${SET_MIN_FREQ} -gt ${FREQ_max} ] && {
log ERROR "Cannot set GPU min freq (%s) to be greater than max freq (%s)" \
"${SET_MIN_FREQ}" "${FREQ_max}"
return 1
}
[ ${SET_MIN_FREQ} -lt ${FREQ_RPn} ] && {
log ERROR "Cannot set GPU min freq (%s) to be less than hw min freq (%s)" \
"${SET_MIN_FREQ}" "${FREQ_RPn}"
return 1
}
[ -z "${DRY_RUN}" ] || return 0
if ! printf "%s" ${SET_MIN_FREQ} > $(print_freq_sysfs_path min);
then
log ERROR "Failed to set GPU min frequency"
return 1
fi
}
#
# Set min or max or both GPU frequencies to the user indicated values.
#
set_freq() {
# Get hw max & min frequencies
read_freq_info n RP0 RPn || return $?
[ -z "${SET_MAX_FREQ}" ] || {
SET_MAX_FREQ=$(compute_freq_set "${SET_MAX_FREQ}")
[ -z "${SET_MAX_FREQ}" ] && return 1
}
[ -z "${SET_MIN_FREQ}" ] || {
SET_MIN_FREQ=$(compute_freq_set "${SET_MIN_FREQ}")
[ -z "${SET_MIN_FREQ}" ] && return 1
}
#
# Ensure correct operation order, to avoid setting min freq
# to a value which is larger than max freq.
#
# E.g.:
# crt_min=crt_max=600; new_min=new_max=700
# > operation order: max=700; min=700
#
# crt_min=crt_max=600; new_min=new_max=500
# > operation order: min=500; max=500
#
if [ -n "${SET_MAX_FREQ}" ] && [ -n "${SET_MIN_FREQ}" ]; then
[ ${SET_MAX_FREQ} -lt ${SET_MIN_FREQ} ] && {
log ERROR "Cannot set GPU max freq to be less than min freq"
return 1
}
read_freq_info n min || return $?
if [ ${SET_MAX_FREQ} -lt ${FREQ_min} ]; then
set_freq_min || return $?
set_freq_max
else
set_freq_max || return $?
set_freq_min
fi
elif [ -n "${SET_MAX_FREQ}" ]; then
set_freq_max
elif [ -n "${SET_MIN_FREQ}" ]; then
set_freq_min
else
log "Unexpected call to set_freq()"
return 1
fi
}
#
# Helper for detect_throttling().
#
get_thrott_detect_pid() {
[ -e ${THROTT_DETECT_PID_FILE_PATH} ] || return 0
local pid
read pid < ${THROTT_DETECT_PID_FILE_PATH} || {
log ERROR "Failed to read pid from: %s" "${THROTT_DETECT_PID_FILE_PATH}"
return 1
}
local proc_path=/proc/${pid:-invalid}/cmdline
[ -r ${proc_path} ] && grep -qs "${0##*/}" ${proc_path} && {
printf "%s" "${pid}"
return 0
}
# Remove orphaned PID file
rm -rf ${THROTT_DETECT_PID_FILE_PATH}
return 1
}
#
# Control detection and reporting of GPU throttling events.
# arg1: start - run throttle detector in background
# stop - stop throttle detector process, if any
# status - verify if throttle detector is running
#
detect_throttling() {
local pid
pid=$(get_thrott_detect_pid)
case "$1" in
status)
printf "Throttling detector is "
[ -z "${pid}" ] && printf "not running\n" && return 0
printf "running (pid=%s)\n" ${pid}
;;
stop)
[ -z "${pid}" ] && return 0
log INFO "Stopping throttling detector (pid=%s)" "${pid}"
kill ${pid}; sleep 1; kill -0 ${pid} 2>/dev/null && kill -9 ${pid}
rm -rf ${THROTT_DETECT_PID_FILE_PATH}
;;
start)
[ -n "${pid}" ] && {
log WARN "Throttling detector is already running (pid=%s)" ${pid}
return 0
}
(
read_freq_info n RPn || exit $?
while true; do
sleep ${THROTT_DETECT_SLEEP_SEC}
read_freq_info n act min cur || exit $?
#
# The throttling seems to occur when act freq goes below min.
# However, it's necessary to exclude the idle states, where
# act freq normally reaches RPn and cur goes below min.
#
[ ${FREQ_act} -lt ${FREQ_min} ] && \
[ ${FREQ_act} -gt ${FREQ_RPn} ] && \
[ ${FREQ_cur} -ge ${FREQ_min} ] && \
printf "GPU throttling detected: act=%s min=%s cur=%s RPn=%s\n" \
${FREQ_act} ${FREQ_min} ${FREQ_cur} ${FREQ_RPn}
done
) &
pid=$!
log INFO "Started GPU throttling detector (pid=%s)" ${pid}
printf "%s\n" ${pid} > ${THROTT_DETECT_PID_FILE_PATH} || \
log WARN "Failed to write throttle detector PID file"
;;
esac
}
#
# Retrieve the list of online CPUs.
#
get_online_cpus() {
local path cpu_index
printf "0"
for path in $(grep 1 ${CPU_SYSFS_PREFIX}/cpu*/online); do
cpu_index=${path##*/cpu}
printf " %s" ${cpu_index%%/*}
done
}
#
# Helper to print sysfs path for the given CPU index and freq info.
#
# arg1: Frequency info sysfs name, one of *_CPU_FREQ_INFO constants above
# arg2: CPU index
#
print_cpu_freq_sysfs_path() {
printf ${CPU_FREQ_SYSFS_PATTERN} "$2" "$1"
}
#
# Read the specified CPU freq info from sysfs.
#
# arg1: CPU index
# arg2: Flag (y/n) to also enable printing the freq info.
# arg3...: Frequency info sysfs name(s), see *_CPU_FREQ_INFO constants above
# return: Global variable(s) CPU_FREQ_${arg} containing the requested information
#
read_cpu_freq_info() {
local var val info path cpu_index print=0 ret=0
cpu_index=$1
[ "$2" = "y" ] && print=1
shift 2
while [ $# -gt 0 ]; do
info=$1
shift
var=CPU_FREQ_${info}
path=$(print_cpu_freq_sysfs_path "${info}" ${cpu_index})
[ -r ${path} ] && read ${var} < ${path} || {
log ERROR "Failed to read CPU freq info from: %s" "${path}"
ret=1
continue
}
[ -n "${var}" ] || {
log ERROR "Got empty CPU freq info from: %s" "${path}"
ret=1
continue
}
[ ${print} -eq 1 ] && {
eval val=\$${var}
printf "%6s: %4s Hz\n" "${info}" "${val}"
}
done
return ${ret}
}
#
# Helper to print freq. value as requested by user via '--cpu-set-max' option.
# arg1: user requested freq value
#
compute_cpu_freq_set() {
local val
case "$1" in
+)
val=${CPU_FREQ_cpuinfo_max}
;;
-)
val=${CPU_FREQ_cpuinfo_min}
;;
*%)
val=$((${1%?} * CPU_FREQ_cpuinfo_max / 100))
;;
*[!0-9]*)
log ERROR "Cannot set CPU freq to invalid value: %s" "$1"
return 1
;;
"")
log ERROR "Cannot set CPU freq to unspecified value"
return 1
;;
*)
log ERROR "Cannot set CPU freq to custom value; use +, -, or % instead"
return 1
;;
esac
printf "%s" "${val}"
}
#
# Adjust CPU max scaling frequency.
#
set_cpu_freq_max() {
local target_freq res=0
case "${CPU_SET_MAX_FREQ}" in
+)
target_freq=100
;;
-)
target_freq=1
;;
*%)
target_freq=${CPU_SET_MAX_FREQ%?}
;;
*)
log ERROR "Invalid CPU freq"
return 1
;;
esac
local pstate_info=$(printf "${CPU_PSTATE_SYSFS_PATTERN}" max_perf_pct)
[ -e "${pstate_info}" ] && {
log INFO "Setting intel_pstate max perf to %s" "${target_freq}%"
if ! printf "%s" "${target_freq}" > "${pstate_info}";
then
log ERROR "Failed to set intel_pstate max perf"
res=1
fi
}
local cpu_index
for cpu_index in $(get_online_cpus); do
read_cpu_freq_info ${cpu_index} n ${CAP_CPU_FREQ_INFO} || { res=$?; continue; }
target_freq=$(compute_cpu_freq_set "${CPU_SET_MAX_FREQ}")
[ -z "${target_freq}" ] && { res=$?; continue; }
log INFO "Setting CPU%s max scaling freq to %s Hz" ${cpu_index} "${target_freq}"
[ -n "${DRY_RUN}" ] && continue
if ! printf "%s" ${target_freq} > $(print_cpu_freq_sysfs_path scaling_max ${cpu_index});
then
res=1
log ERROR "Failed to set CPU%s max scaling frequency" ${cpu_index}
fi
done
return ${res}
}
#
# Show help message.
#
print_usage() {
cat <<EOF
Usage: ${0##*/} [OPTION]...
A script to manage Intel GPU frequencies. Can be used for debugging performance
problems or trying to obtain a stable frequency while benchmarking.
Note Intel GPUs only accept specific frequencies, usually multiples of 50 MHz.
Options:
-g, --get [act|enf|cap|all]
Get frequency information: active (default), enforced,
hardware capabilities or all of them.
-s, --set [{min|max}=]{FREQUENCY[%]|+|-}
Set min or max frequency to the given value (MHz).
Append '%' to interpret FREQUENCY as % of hw max.
Use '+' or '-' to set frequency to hardware max or min.
Omit min/max prefix to set both frequencies.
-r, --reset Reset frequencies to hardware defaults.
-m, --monitor [act|enf|cap|all]
Monitor the indicated frequencies via 'watch' utility.
See '-g, --get' option for more details.
-d|--detect-thrott [start|stop|status]
Start (default operation) the throttling detector
as a background process. Use 'stop' or 'status' to
terminate the detector process or verify its status.
--cpu-set-max [FREQUENCY%|+|-}
Set CPU max scaling frequency as % of hw max.
Use '+' or '-' to set frequency to hardware max or min.
-r, --reset Reset frequencies to hardware defaults.
--dry-run See what the script will do without applying any
frequency changes.
-h, --help Display this help text and exit.
EOF
}
#
# Parse user input for '-g, --get' option.
# Returns 0 if a value has been provided, otherwise 1.
#
parse_option_get() {
local ret=0
case "$1" in
act) GET_ACT_FREQ=1;;
enf) GET_ENF_FREQ=1;;
cap) GET_CAP_FREQ=1;;
all) GET_ACT_FREQ=1; GET_ENF_FREQ=1; GET_CAP_FREQ=1;;
-*|"")
# No value provided, using default.
GET_ACT_FREQ=1
ret=1
;;
*)
print_usage
exit 1
;;
esac
return ${ret}
}
#
# Validate user input for '-s, --set' option.
# arg1: input value to be validated
# arg2: optional flag indicating input is restricted to %
#
validate_option_set() {
case "$1" in
+|-|[0-9]%|[0-9][0-9]%)
return 0
;;
*[!0-9]*|"")
print_usage
exit 1
;;
esac
[ -z "$2" ] || { print_usage; exit 1; }
}
#
# Parse script arguments.
#
[ $# -eq 0 ] && { print_usage; exit 1; }
while [ $# -gt 0 ]; do
case "$1" in
-g|--get)
parse_option_get "$2" && shift
;;
-s|--set)
shift
case "$1" in
min=*)
SET_MIN_FREQ=${1#min=}
validate_option_set "${SET_MIN_FREQ}"
;;
max=*)
SET_MAX_FREQ=${1#max=}
validate_option_set "${SET_MAX_FREQ}"
;;
*)
SET_MIN_FREQ=$1
validate_option_set "${SET_MIN_FREQ}"
SET_MAX_FREQ=${SET_MIN_FREQ}
;;
esac
;;
-r|--reset)
RESET_FREQ=1
SET_MIN_FREQ="-"
SET_MAX_FREQ="+"
;;
-m|--monitor)
MONITOR_FREQ=act
parse_option_get "$2" && MONITOR_FREQ=$2 && shift
;;
-d|--detect-thrott)
DETECT_THROTT=start
case "$2" in
start|stop|status)
DETECT_THROTT=$2
shift
;;
esac
;;
--cpu-set-max)
shift
CPU_SET_MAX_FREQ=$1
validate_option_set "${CPU_SET_MAX_FREQ}" restricted
;;
--dry-run)
DRY_RUN=1
;;
-h|--help)
print_usage
exit 0
;;
*)
print_usage
exit 1
;;
esac
shift
done
#
# Main
#
RET=0
identify_intel_gpu || {
log INFO "No Intel GPU detected"
exit 0
}
[ -n "${SET_MIN_FREQ}${SET_MAX_FREQ}" ] && { set_freq || RET=$?; }
print_freq_info
[ -n "${DETECT_THROTT}" ] && detect_throttling ${DETECT_THROTT}
[ -n "${CPU_SET_MAX_FREQ}" ] && { set_cpu_freq_max || RET=$?; }
[ -n "${MONITOR_FREQ}" ] && {
log INFO "Entering frequency monitoring mode"
sleep 2
exec watch -d -n 1 "$0" -g "${MONITOR_FREQ}"
}
exit ${RET}

View File

@@ -1,24 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC1091 # the path is created in build-kdl and
# here is check if exist
terminate() {
echo "ci-kdl.sh caught SIGTERM signal! propagating to child processes"
for job in $(jobs -p)
do
kill -15 "$job"
done
}
trap terminate SIGTERM
if [ -f /ci-kdl.venv/bin/activate ]; then
source /ci-kdl.venv/bin/activate
/ci-kdl.venv/bin/python /ci-kdl.venv/bin/ci-kdl | tee -a /results/kdl.log &
child=$!
wait $child
mv kdl_*.json /results/kdl.json
else
echo -e "Not possible to activate ci-kdl virtual environment"
fi

View File

@@ -1,21 +0,0 @@
#!/bin/sh
set -ex
_XORG_SCRIPT="/xorg-script"
_FLAG_FILE="/xorg-started"
echo "touch ${_FLAG_FILE}; sleep 100000" > "${_XORG_SCRIPT}"
if [ "x$1" != "x" ]; then
export LD_LIBRARY_PATH="${1}/lib"
export LIBGL_DRIVERS_PATH="${1}/lib/dri"
fi
xinit /bin/sh "${_XORG_SCRIPT}" -- /usr/bin/Xorg vt45 -noreset -s 0 -dpms -logfile /Xorg.0.log &
# Wait for xorg to be ready for connections.
for _ in 1 2 3 4 5; do
if [ -e "${_FLAG_FILE}" ]; then
break
fi
sleep 5
done

View File

@@ -1,66 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC1091
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# ALPINE_X86_64_BUILD_TAG
set -e
set -o xtrace
export LLVM_VERSION="${LLVM_VERSION:=16}"
EPHEMERAL=(
)
DEPS=(
bash
bison
ccache
clang16-dev
cmake
clang-dev
coreutils
curl
flex
gcc
g++
git
gettext
glslang
linux-headers
llvm16-static
llvm16-dev
meson
expat-dev
elfutils-dev
libdrm-dev
libselinux-dev
libva-dev
libpciaccess-dev
zlib-dev
python3-dev
py3-mako
py3-ply
vulkan-headers
spirv-tools-dev
util-macros
wayland-dev
wayland-protocols
)
apk --no-cache add "${DEPS[@]}" "${EPHEMERAL[@]}"
. .gitlab-ci/container/build-llvm-spirv.sh
. .gitlab-ci/container/build-libclc.sh
. .gitlab-ci/container/container_pre_build.sh
############### Uninstall the build software
apk del "${EPHEMERAL[@]}"
. .gitlab-ci/container/container_post_build.sh

View File

@@ -1,29 +0,0 @@
#!/usr/bin/env bash
# This is a ci-templates build script to generate a container for LAVA SSH client.
# shellcheck disable=SC1091
set -e
set -o xtrace
EPHEMERAL=(
)
# We only need these very basic packages to run the tests.
DEPS=(
openssh-client # for ssh
iputils # for ping
bash
curl
)
apk --no-cache add "${DEPS[@]}" "${EPHEMERAL[@]}"
. .gitlab-ci/container/container_pre_build.sh
############### Uninstall the build software
apk del "${EPHEMERAL[@]}"
. .gitlab-ci/container/container_post_build.sh

View File

@@ -1,62 +0,0 @@
#!/usr/bin/env bash
set -e
set -o xtrace
# Fetch the arm-built rootfs image and unpack it in our x86_64 container (saves
# network transfer, disk usage, and runtime on test jobs)
# shellcheck disable=SC2154 # arch is assigned in previous scripts
if curl -X HEAD -s "${ARTIFACTS_PREFIX}/${FDO_UPSTREAM_REPO}/${ARTIFACTS_SUFFIX}/${arch}/done"; then
ARTIFACTS_URL="${ARTIFACTS_PREFIX}/${FDO_UPSTREAM_REPO}/${ARTIFACTS_SUFFIX}/${arch}"
else
ARTIFACTS_URL="${ARTIFACTS_PREFIX}/${CI_PROJECT_PATH}/${ARTIFACTS_SUFFIX}/${arch}"
fi
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
"${ARTIFACTS_URL}"/lava-rootfs.tar.zst -o rootfs.tar.zst
mkdir -p /rootfs-"$arch"
tar -C /rootfs-"$arch" '--exclude=./dev/*' --zstd -xf rootfs.tar.zst
rm rootfs.tar.zst
if [[ $arch == "arm64" ]]; then
mkdir -p /baremetal-files
pushd /baremetal-files
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O "${KERNEL_IMAGE_BASE}"/arm64/Image
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O "${KERNEL_IMAGE_BASE}"/arm64/Image.gz
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O "${KERNEL_IMAGE_BASE}"/arm64/cheza-kernel
DEVICE_TREES=""
DEVICE_TREES="$DEVICE_TREES apq8016-sbc.dtb"
DEVICE_TREES="$DEVICE_TREES apq8096-db820c.dtb"
DEVICE_TREES="$DEVICE_TREES tegra210-p3450-0000.dtb"
DEVICE_TREES="$DEVICE_TREES imx8mq-nitrogen.dtb"
for DTB in $DEVICE_TREES; do
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O "${KERNEL_IMAGE_BASE}/arm64/$DTB"
done
popd
elif [[ $arch == "armhf" ]]; then
mkdir -p /baremetal-files
pushd /baremetal-files
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O "${KERNEL_IMAGE_BASE}"/armhf/zImage
DEVICE_TREES=""
DEVICE_TREES="$DEVICE_TREES imx6q-cubox-i.dtb"
DEVICE_TREES="$DEVICE_TREES tegra124-jetson-tk1.dtb"
for DTB in $DEVICE_TREES; do
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O "${KERNEL_IMAGE_BASE}/armhf/$DTB"
done
popd
fi

View File

@@ -1,58 +0,0 @@
#!/usr/bin/env bash
set -ex
ANGLE_REV="0518a3ff4d4e7e5b2ce8203358f719613a31c118"
# DEPOT tools
git clone --depth 1 https://chromium.googlesource.com/chromium/tools/depot_tools.git
PWD=$(pwd)
export PATH=$PWD/depot_tools:$PATH
export DEPOT_TOOLS_UPDATE=0
mkdir /angle-build
pushd /angle-build
git init
git remote add origin https://chromium.googlesource.com/angle/angle.git
git fetch --depth 1 origin "$ANGLE_REV"
git checkout FETCH_HEAD
# source preparation
python3 scripts/bootstrap.py
mkdir -p build/config
gclient sync
sed -i "/catapult/d" testing/BUILD.gn
mkdir -p out/Release
echo '
is_debug = false
angle_enable_swiftshader = false
angle_enable_null = false
angle_enable_gl = false
angle_enable_vulkan = true
angle_has_histograms = false
build_angle_trace_perf_tests = false
build_angle_deqp_tests = false
angle_use_custom_libvulkan = false
dcheck_always_on=true
' > out/Release/args.gn
if [[ "$DEBIAN_ARCH" = "arm64" ]]; then
build/linux/sysroot_scripts/install-sysroot.py --arch=arm64
fi
gn gen out/Release
# depot_tools overrides ninja with a version that doesn't work. We want
# ninja with FDO_CI_CONCURRENT anyway.
/usr/local/bin/ninja -C out/Release/
mkdir /angle
cp out/Release/lib*GL*.so /angle/
ln -s libEGL.so /angle/libEGL.so.1
ln -s libGLESv2.so /angle/libGLESv2.so.2
rm -rf out
popd
rm -rf ./depot_tools

View File

@@ -1,25 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_X86_64_TEST_GL_TAG
# DEBIAN_X86_64_TEST_VK_TAG
# KERNEL_ROOTFS_TAG
set -ex
APITRACE_VERSION="0a6506433e1f9f7b69757b4e5730326970c4321a"
git clone https://github.com/apitrace/apitrace.git --single-branch --no-checkout /apitrace
pushd /apitrace
git checkout "$APITRACE_VERSION"
git submodule update --init --depth 1 --recursive
cmake -S . -B _build -G Ninja -DCMAKE_BUILD_TYPE=Release -DENABLE_GUI=False -DENABLE_WAFFLE=on $EXTRA_CMAKE_ARGS
cmake --build _build --parallel --target apitrace eglretrace
mkdir build
cp _build/apitrace build
cp _build/eglretrace build
${STRIP_CMD:-strip} build/*
find . -not -path './build' -not -path './build/*' -delete
popd

View File

@@ -1,44 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
set -ex
git config --global user.email "mesa@example.com"
git config --global user.name "Mesa CI"
CROSVM_VERSION=1641c55bcc922588e24de73e9cca7b5e4005bd6d
git clone --single-branch -b main --no-checkout https://chromium.googlesource.com/crosvm/crosvm /platform/crosvm
pushd /platform/crosvm
git checkout "$CROSVM_VERSION"
git submodule update --init
VIRGLRENDERER_VERSION=d9c002fac153b834a2c17731f2b85c36e333e102
rm -rf third_party/virglrenderer
git clone --single-branch -b main --no-checkout https://gitlab.freedesktop.org/virgl/virglrenderer.git third_party/virglrenderer
pushd third_party/virglrenderer
git checkout "$VIRGLRENDERER_VERSION"
meson setup build/ -D libdir=lib -D render-server-worker=process -D venus=true $EXTRA_MESON_ARGS
meson install -C build
popd
cargo update -p pkg-config@0.3.26 --precise 0.3.27
RUSTFLAGS='-L native=/usr/local/lib' cargo install \
bindgen-cli \
--locked \
-j ${FDO_CI_CONCURRENT:-4} \
--root /usr/local \
--version 0.65.1 \
$EXTRA_CARGO_ARGS
CROSVM_USE_SYSTEM_MINIGBM=1 CROSVM_USE_SYSTEM_VIRGLRENDERER=1 RUSTFLAGS='-L native=/usr/local/lib' cargo install \
-j ${FDO_CI_CONCURRENT:-4} \
--locked \
--features 'default-no-sandbox gpu x virgl_renderer' \
--path . \
--root /usr/local \
$EXTRA_CARGO_ARGS
popd
rm -rf /platform/crosvm

View File

@@ -1,70 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_X86_64_TEST_ANDROID_TAG
# DEBIAN_X86_64_TEST_GL_TAG
# DEBIAN_X86_64_TEST_VK_TAG
# KERNEL_ROOTFS_TAG
set -ex
DEQP_RUNNER_VERSION=0.18.0
DEQP_RUNNER_GIT_URL="${DEQP_RUNNER_GIT_URL:-https://gitlab.freedesktop.org/mesa/deqp-runner.git}"
if [ -n "${DEQP_RUNNER_GIT_TAG}${DEQP_RUNNER_GIT_REV}" ]; then
# Build and install from source
DEQP_RUNNER_CARGO_ARGS="--git $DEQP_RUNNER_GIT_URL"
if [ -n "${DEQP_RUNNER_GIT_TAG}" ]; then
DEQP_RUNNER_CARGO_ARGS="--tag ${DEQP_RUNNER_GIT_TAG} ${DEQP_RUNNER_CARGO_ARGS}"
DEQP_RUNNER_GIT_CHECKOUT="$DEQP_RUNNER_GIT_TAG"
else
DEQP_RUNNER_CARGO_ARGS="--rev ${DEQP_RUNNER_GIT_REV} ${DEQP_RUNNER_CARGO_ARGS}"
DEQP_RUNNER_GIT_CHECKOUT="$DEQP_RUNNER_GIT_REV"
fi
DEQP_RUNNER_CARGO_ARGS="${DEQP_RUNNER_CARGO_ARGS} ${EXTRA_CARGO_ARGS}"
else
# Install from package registry
DEQP_RUNNER_CARGO_ARGS="--version ${DEQP_RUNNER_VERSION} ${EXTRA_CARGO_ARGS} -- deqp-runner"
DEQP_RUNNER_GIT_CHECKOUT="v$DEQP_RUNNER_VERSION"
fi
if [[ "$RUST_TARGET" != *-android ]]; then
cargo install --locked \
-j ${FDO_CI_CONCURRENT:-4} \
--root /usr/local \
${DEQP_RUNNER_CARGO_ARGS}
else
mkdir -p /deqp-runner
pushd /deqp-runner
git clone --branch "$DEQP_RUNNER_GIT_CHECKOUT" --depth 1 "$DEQP_RUNNER_GIT_URL" deqp-runner-git
pushd deqp-runner-git
cargo install --locked \
-j ${FDO_CI_CONCURRENT:-4} \
--root /usr/local --version 2.10.0 \
cargo-ndk
rustup target add $RUST_TARGET
RUSTFLAGS='-C target-feature=+crt-static' cargo ndk --target $RUST_TARGET build --release
mv target/$RUST_TARGET/release/deqp-runner /deqp-runner
cargo uninstall --locked \
--root /usr/local \
cargo-ndk
popd
rm -rf deqp-runner-git
popd
fi
# remove unused test runners to shrink images for the Mesa CI build (not kernel,
# which chooses its own deqp branch)
if [ -z "${DEQP_RUNNER_GIT_TAG}${DEQP_RUNNER_GIT_REV}" ]; then
rm -f /usr/local/bin/igt-runner
fi

View File

@@ -1,275 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_X86_64_TEST_ANDROID_TAG
# DEBIAN_X86_64_TEST_GL_TAG
# DEBIAN_X86_64_TEST_VK_TAG
# KERNEL_ROOTFS_TAG
set -ex -o pipefail
# See `deqp_build_targets` below for which release is used to produce which
# binary. Unless this comment has bitrotten:
# - the VK release produces `deqp-vk`,
# - the GL release produces `glcts`, and
# - the GLES release produces `deqp-gles*` and `deqp-egl`
DEQP_VK_VERSION=1.3.7.0
DEQP_GL_VERSION=4.6.4.0
DEQP_GLES_VERSION=3.2.10.0
# Patches to VulkanCTS may come from commits in their repo (listed in
# cts_commits_to_backport) or patch files stored in our repo (in the patch
# directory `$OLDPWD/.gitlab-ci/container/patches/` listed in cts_patch_files).
# Both list variables would have comments explaining the reasons behind the
# patches.
# shellcheck disable=SC2034
vk_cts_commits_to_backport=(
# Take multiview into account for task shader inv. stats
22aa3f4c59f6e1d4daebd5a8c9c05bce6cd3b63b
# Remove illegal mesh shader query tests
2a87f7b25dc27188be0f0a003b2d7aef69d9002e
# Relax fragment shader invocations result verifications
0d8bf6a2715f95907e9cf86a86876ff1f26c66fe
# Fix several issues in dynamic rendering basic tests
c5453824b498c981c6ba42017d119f5de02a3e34
# Add setVisible for VulkanWindowDirectDrm
a8466bf6ea98f6cd6733849ad8081775318a3e3e
)
# shellcheck disable=SC2034
vk_cts_patch_files=(
# Derivate subgroup fix
# https://github.com/KhronosGroup/VK-GL-CTS/pull/442
build-deqp-vk_Use-subgroups-helper-in-derivate-tests.patch
build-deqp-vk_Add-missing-subgroup-support-checks-for-linear-derivate-tests.patch
)
if [ "${DEQP_TARGET}" = 'android' ]; then
vk_cts_patch_files+=(
build-deqp-vk_Allow-running-on-Android-from-the-command-line.patch
build-deqp-vk_Android-prints-to-stdout-instead-of-logcat.patch
)
fi
# shellcheck disable=SC2034
gl_cts_commits_to_backport=(
)
# shellcheck disable=SC2034
gl_cts_patch_files=(
)
if [ "${DEQP_TARGET}" = 'android' ]; then
gl_cts_patch_files+=(
build-deqp-gl_Allow-running-on-Android-from-the-command-line.patch
build-deqp-gl_Android-prints-to-stdout-instead-of-logcat.patch
)
fi
# shellcheck disable=SC2034
# GLES builds also EGL
gles_cts_commits_to_backport=(
# Implement support for the EGL_EXT_config_select_group extension
88ba9ac270db5be600b1ecacbc6d9db0c55d5be4
)
# shellcheck disable=SC2034
gles_cts_patch_files=(
)
if [ "${DEQP_TARGET}" = 'android' ]; then
gles_cts_patch_files+=(
build-deqp-gles_Allow-running-on-Android-from-the-command-line.patch
build-deqp-gles_Android-prints-to-stdout-instead-of-logcat.patch
)
fi
### Careful editing anything below this line
git config --global user.email "mesa@example.com"
git config --global user.name "Mesa CI"
# shellcheck disable=SC2153
case "${DEQP_API}" in
VK) DEQP_VERSION="vulkan-cts-$DEQP_VK_VERSION";;
GL) DEQP_VERSION="opengl-cts-$DEQP_GL_VERSION";;
GLES) DEQP_VERSION="opengl-es-cts-$DEQP_GLES_VERSION";;
esac
git clone \
https://github.com/KhronosGroup/VK-GL-CTS.git \
-b $DEQP_VERSION \
--depth 1 \
/VK-GL-CTS
pushd /VK-GL-CTS
mkdir -p /deqp
# shellcheck disable=SC2153
deqp_api=${DEQP_API,,}
cts_commits_to_backport="${deqp_api}_cts_commits_to_backport[@]"
for commit in "${!cts_commits_to_backport}"
do
PATCH_URL="https://github.com/KhronosGroup/VK-GL-CTS/commit/$commit.patch"
echo "Apply patch to ${DEQP_API} CTS from $PATCH_URL"
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 $PATCH_URL | \
git am -
done
cts_patch_files="${deqp_api}_cts_patch_files[@]"
for patch in "${!cts_patch_files}"
do
echo "Apply patch to ${DEQP_API} CTS from $patch"
git am < $OLDPWD/.gitlab-ci/container/patches/$patch
done
{
echo "dEQP base version $DEQP_VERSION"
echo "The following local patches are applied on top:"
git log --reverse --oneline $DEQP_VERSION.. --format=%s | sed 's/^/- /'
} > /deqp/version-$deqp_api
# --insecure is due to SSL cert failures hitting sourceforge for zlib and
# libpng (sigh). The archives get their checksums checked anyway, and git
# always goes through ssh or https.
python3 external/fetch_sources.py --insecure
# Save the testlog stylesheets:
cp doc/testlog-stylesheet/testlog.{css,xsl} /deqp
popd
pushd /deqp
if [ "${DEQP_API}" = 'GLES' ]; then
if [ "${DEQP_TARGET}" = 'android' ]; then
cmake -S /VK-GL-CTS -B . -G Ninja \
-DDEQP_TARGET=android \
-DCMAKE_BUILD_TYPE=Release \
$EXTRA_CMAKE_ARGS
mold --run ninja modules/egl/deqp-egl
mv /deqp/modules/egl/deqp-egl /deqp/modules/egl/deqp-egl-android
else
# When including EGL/X11 testing, do that build first and save off its
# deqp-egl binary.
cmake -S /VK-GL-CTS -B . -G Ninja \
-DDEQP_TARGET=x11_egl_glx \
-DCMAKE_BUILD_TYPE=Release \
$EXTRA_CMAKE_ARGS
mold --run ninja modules/egl/deqp-egl
mv /deqp/modules/egl/deqp-egl /deqp/modules/egl/deqp-egl-x11
cmake -S /VK-GL-CTS -B . -G Ninja \
-DDEQP_TARGET=wayland \
-DCMAKE_BUILD_TYPE=Release \
$EXTRA_CMAKE_ARGS
mold --run ninja modules/egl/deqp-egl
mv /deqp/modules/egl/deqp-egl /deqp/modules/egl/deqp-egl-wayland
fi
fi
cmake -S /VK-GL-CTS -B . -G Ninja \
-DDEQP_TARGET=${DEQP_TARGET} \
-DCMAKE_BUILD_TYPE=Release \
$EXTRA_CMAKE_ARGS
# Make sure `default` doesn't silently stop detecting one of the platforms we care about
if [ "${DEQP_TARGET}" = 'default' ]; then
grep -q DEQP_SUPPORT_WAYLAND=1 build.ninja
grep -q DEQP_SUPPORT_X11=1 build.ninja
grep -q DEQP_SUPPORT_XCB=1 build.ninja
fi
deqp_build_targets=()
case "${DEQP_API}" in
VK)
deqp_build_targets+=(deqp-vk)
;;
GL)
deqp_build_targets+=(glcts)
;;
GLES)
deqp_build_targets+=(deqp-gles{2,3,31})
# deqp-egl also comes from this build, but it is handled separately above.
;;
esac
if [ "${DEQP_TARGET}" != 'android' ]; then
deqp_build_targets+=(testlog-to-xml)
deqp_build_targets+=(testlog-to-csv)
deqp_build_targets+=(testlog-to-junit)
fi
mold --run ninja "${deqp_build_targets[@]}"
if [ "${DEQP_TARGET}" != 'android' ]; then
# Copy out the mustpass lists we want.
mkdir -p /deqp/mustpass
if [ "${DEQP_API}" = 'VK' ]; then
for mustpass in $(< /VK-GL-CTS/external/vulkancts/mustpass/main/vk-default.txt) ; do
cat /VK-GL-CTS/external/vulkancts/mustpass/main/$mustpass \
>> /deqp/mustpass/vk-master.txt
done
fi
if [ "${DEQP_API}" = 'GL' ]; then
cp \
/VK-GL-CTS/external/openglcts/data/mustpass/gl/khronos_mustpass/4.6.1.x/*-main.txt \
/deqp/mustpass/
cp \
/VK-GL-CTS/external/openglcts/data/mustpass/gl/khronos_mustpass_single/4.6.1.x/*-single.txt \
/deqp/mustpass/
fi
if [ "${DEQP_API}" = 'GLES' ]; then
cp \
/VK-GL-CTS/external/openglcts/data/mustpass/gles/aosp_mustpass/3.2.6.x/*.txt \
/deqp/mustpass/
cp \
/VK-GL-CTS/external/openglcts/data/mustpass/egl/aosp_mustpass/3.2.6.x/egl-main.txt \
/deqp/mustpass/
cp \
/VK-GL-CTS/external/openglcts/data/mustpass/gles/khronos_mustpass/3.2.6.x/*-main.txt \
/deqp/mustpass/
fi
# Save *some* executor utils, but otherwise strip things down
# to reduct deqp build size:
mkdir /deqp/executor.save
cp /deqp/executor/testlog-to-* /deqp/executor.save
rm -rf /deqp/executor
mv /deqp/executor.save /deqp/executor
fi
# Remove other mustpass files, since we saved off the ones we wanted to conventient locations above.
rm -rf /deqp/external/**/mustpass/
rm -rf /deqp/external/vulkancts/modules/vulkan/vk-master*
rm -rf /deqp/external/vulkancts/modules/vulkan/vk-default
rm -rf /deqp/external/openglcts/modules/cts-runner
rm -rf /deqp/modules/internal
rm -rf /deqp/execserver
rm -rf /deqp/framework
find . -depth \( -iname '*cmake*' -o -name '*ninja*' -o -name '*.o' -o -name '*.a' \) -exec rm -rf {} \;
if [ "${DEQP_API}" = 'VK' ]; then
${STRIP_CMD:-strip} external/vulkancts/modules/vulkan/deqp-vk
fi
if [ "${DEQP_API}" = 'GL' ]; then
${STRIP_CMD:-strip} external/openglcts/modules/glcts
fi
if [ "${DEQP_API}" = 'GLES' ]; then
${STRIP_CMD:-strip} modules/*/deqp-*
fi
du -sh ./*
rm -rf /VK-GL-CTS
popd

View File

@@ -1,15 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_BUILD_TAG
set -ex
git clone https://github.com/microsoft/DirectX-Headers -b v1.611.0 --depth 1
pushd DirectX-Headers
meson setup build --backend=ninja --buildtype=release -Dbuild-test=false $EXTRA_MESON_ARGS
meson install -C build
popd
rm -rf DirectX-Headers

View File

@@ -1,19 +0,0 @@
#!/bin/bash
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_X86_64_TEST_VK_TAG
# KERNEL_ROOTFS_TAG
set -ex
git clone https://github.com/ValveSoftware/Fossilize.git
cd Fossilize
git checkout b43ee42bbd5631ea21fe9a2dee4190d5d875c327
git submodule update --init
mkdir build
cd build
cmake -S .. -B . -G Ninja -DCMAKE_BUILD_TYPE=Release
ninja -C . install
cd ../..
rm -rf Fossilize

View File

@@ -1,19 +0,0 @@
#!/usr/bin/env bash
set -ex
GFXRECONSTRUCT_VERSION=761837794a1e57f918a85af7000b12e531b178ae
git clone https://github.com/LunarG/gfxreconstruct.git \
--single-branch \
-b master \
--no-checkout \
/gfxreconstruct
pushd /gfxreconstruct
git checkout "$GFXRECONSTRUCT_VERSION"
git submodule update --init
git submodule update
cmake -S . -B _build -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX:PATH=/gfxreconstruct/build -DBUILD_WERROR=OFF
cmake --build _build --parallel --target tools/{replay,info}/install/strip
find . -not -path './build' -not -path './build/*' -delete
popd

View File

@@ -1,16 +0,0 @@
#!/bin/bash
set -ex
PARALLEL_DEQP_RUNNER_VERSION=fe557794b5dadd8dbf0eae403296625e03bda18a
git clone https://gitlab.freedesktop.org/mesa/parallel-deqp-runner --single-branch -b master --no-checkout /parallel-deqp-runner
pushd /parallel-deqp-runner
git checkout "$PARALLEL_DEQP_RUNNER_VERSION"
meson . _build
ninja -C _build hang-detection
mkdir -p build/bin
install _build/hang-detection build/bin
strip build/bin/*
find . -not -path './build' -not -path './build/*' -delete
popd

View File

@@ -1,23 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC1091 # the path is created by the script
set -ex
KDL_REVISION="5056f71b100a68b72b285c6fc845a66a2ed25985"
mkdir ci-kdl.git
pushd ci-kdl.git
git init
git remote add origin https://gitlab.freedesktop.org/gfx-ci/ci-kdl.git
git fetch --depth 1 origin ${KDL_REVISION}
git checkout FETCH_HEAD
popd
python3 -m venv ci-kdl.venv
source ci-kdl.venv/bin/activate
pushd ci-kdl.git
pip install -r requirements.txt
pip install .
popd
rm -rf ci-kdl.git

View File

@@ -1,31 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
# shellcheck disable=SC2153
set -ex
mkdir -p kernel
pushd kernel
if [[ ${DEBIAN_ARCH} = "arm64" ]]; then
KERNEL_IMAGE_NAME+=" cheza-kernel"
fi
for image in ${KERNEL_IMAGE_NAME}; do
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-o "/lava-files/${image}" "${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/${image}"
done
for dtb in ${DEVICE_TREES}; do
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-o "/lava-files/${dtb}" "${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/${dtb}"
done
mkdir -p "/lava-files/rootfs-${DEBIAN_ARCH}"
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O "${KERNEL_IMAGE_BASE}/${DEBIAN_ARCH}/modules.tar.zst"
tar --keep-directory-symlink --zstd -xf modules.tar.zst -C "/lava-files/rootfs-${DEBIAN_ARCH}/"
popd
rm -rf kernel

View File

@@ -1,31 +0,0 @@
#!/usr/bin/env bash
set -ex
export LLVM_CONFIG="llvm-config-${LLVM_VERSION:?"llvm unset!"}"
LLVM_TAG="llvmorg-15.0.7"
$LLVM_CONFIG --version
git config --global user.email "mesa@example.com"
git config --global user.name "Mesa CI"
git clone \
https://github.com/llvm/llvm-project \
--depth 1 \
-b "${LLVM_TAG}" \
/llvm-project
mkdir /libclc
pushd /libclc
cmake -S /llvm-project/libclc -B . -G Ninja -DLLVM_CONFIG="$LLVM_CONFIG" -DLIBCLC_TARGETS_TO_BUILD="spirv-mesa3d-;spirv64-mesa3d-" -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/usr -DLLVM_SPIRV=/usr/bin/llvm-spirv
ninja
ninja install
popd
# workaroud cmake vs debian packaging.
mkdir -p /usr/lib/clc
ln -s /usr/share/clc/spirv64-mesa3d-.spv /usr/lib/clc/
ln -s /usr/share/clc/spirv-mesa3d-.spv /usr/lib/clc/
du -sh ./*
rm -rf /libclc /llvm-project

View File

@@ -1,16 +0,0 @@
#!/usr/bin/env bash
# Script used for Android and Fedora builds
# shellcheck disable=SC2086 # we want word splitting
set -ex
export LIBDRM_VERSION=libdrm-2.4.119
curl -L -O --retry 4 -f --retry-all-errors --retry-delay 60 \
https://dri.freedesktop.org/libdrm/"$LIBDRM_VERSION".tar.xz
tar -xvf "$LIBDRM_VERSION".tar.xz && rm "$LIBDRM_VERSION".tar.xz
cd "$LIBDRM_VERSION"
meson setup build -D vc4=disabled -D freedreno=disabled -D etnaviv=disabled $EXTRA_MESON_ARGS
meson install -C build
cd ..
rm -rf "$LIBDRM_VERSION"

View File

@@ -1,22 +0,0 @@
#!/usr/bin/env bash
set -ex
VER="${LLVM_VERSION:?llvm not set}.0.0"
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O "https://github.com/KhronosGroup/SPIRV-LLVM-Translator/archive/refs/tags/v${VER}.tar.gz"
tar -xvf "v${VER}.tar.gz" && rm "v${VER}.tar.gz"
mkdir "SPIRV-LLVM-Translator-${VER}/build"
pushd "SPIRV-LLVM-Translator-${VER}/build"
cmake .. -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/usr
ninja
ninja install
# For some reason llvm-spirv is not installed by default
ninja llvm-spirv
cp tools/llvm-spirv/llvm-spirv /usr/bin/
popd
du -sh "SPIRV-LLVM-Translator-${VER}"
rm -rf "SPIRV-LLVM-Translator-${VER}"

View File

@@ -1,22 +0,0 @@
#!/usr/bin/env bash
set -ex
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_BASE_TAG
# DEBIAN_BUILD_TAG
# FEDORA_X86_64_BUILD_TAG
# KERNEL_ROOTFS_TAG
MOLD_VERSION="2.4.1"
git clone -b v"$MOLD_VERSION" --single-branch --depth 1 https://github.com/rui314/mold.git
pushd mold
cmake -DCMAKE_BUILD_TYPE=Release -D BUILD_TESTING=OFF -D MOLD_LTO=ON
cmake --build . --parallel
cmake --install .
popd
rm -rf mold

View File

@@ -1,25 +0,0 @@
#!/usr/bin/env bash
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_X86_64_TEST_GL_TAG
set -ex -o pipefail
### Careful editing anything below this line
git config --global user.email "mesa@example.com"
git config --global user.name "Mesa CI"
git clone https://github.com/axeldavy/Xnine.git /Xnine
mkdir /Xnine/build
pushd /Xnine/build
git checkout c64753d224c08006bcdcfa7880ada826f27164b1
cmake .. -DBUILD_TESTS=1 -DWITH_DRI3=1 -DD3DADAPTER9_LOCATION=/install/lib/d3d/d3dadapter9.so
make
mkdir -p /NineTests/
mv NineTests/NineTests /NineTests/
popd
rm -rf /Xnine

View File

@@ -1,33 +0,0 @@
#!/bin/bash
# shellcheck disable=SC2086 # we want word splitting
set -ex
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_X86_64_TEST_GL_TAG
# DEBIAN_X86_64_TEST_VK_TAG
# KERNEL_ROOTFS_TAG
REV="1e631479c0b477006dd7561c55e06269d2878d8d"
git clone https://gitlab.freedesktop.org/mesa/piglit.git --single-branch --no-checkout /piglit
pushd /piglit
git checkout "$REV"
patch -p1 <$OLDPWD/.gitlab-ci/piglit/disable-vs_in.diff
cmake -S . -B . -G Ninja -DCMAKE_BUILD_TYPE=Release $PIGLIT_OPTS $EXTRA_CMAKE_ARGS
ninja $PIGLIT_BUILD_TARGETS
find . -depth \( -name .git -o -name '*ninja*' -o -iname '*cmake*' -o -name '*.[chao]' \) -exec rm -rf {} \;
rm -rf target_api
if [ "$PIGLIT_BUILD_TARGETS" = "piglit_replayer" ]; then
find . -depth \
! -regex "^\.$" \
! -regex "^\.\/piglit.*" \
! -regex "^\.\/framework.*" \
! -regex "^\.\/bin$" \
! -regex "^\.\/bin\/replayer\.py" \
! -regex "^\.\/templates.*" \
! -regex "^\.\/tests$" \
! -regex "^\.\/tests\/replay\.py" \
-exec rm -rf {} \; 2>/dev/null
fi
popd

View File

@@ -1,39 +0,0 @@
#!/bin/bash
# Note that this script is not actually "building" rust, but build- is the
# convention for the shared helpers for putting stuff in our containers.
set -ex
# cargo (and rustup) wants to store stuff in $HOME/.cargo, and binaries in
# $HOME/.cargo/bin. Make bin a link to a public bin directory so the commands
# are just available to all build jobs.
mkdir -p "$HOME"/.cargo
ln -s /usr/local/bin "$HOME"/.cargo/bin
# Rusticl requires at least Rust 1.66.0 and NAK requires 1.73.0
#
# Also, pick a specific snapshot from rustup so the compiler doesn't drift on
# us.
RUST_VERSION=1.73.0-2023-10-05
# For rust in Mesa, we use rustup to install. This lets us pick an arbitrary
# version of the compiler, rather than whatever the container's Debian comes
# with.
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
--proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- \
--default-toolchain $RUST_VERSION \
--profile minimal \
-y
rustup component add clippy rustfmt
# Set up a config script for cross compiling -- cargo needs your system cc for
# linking in cross builds, but doesn't know what you want to use for system cc.
cat > /root/.cargo/config <<EOF
[target.armv7-unknown-linux-gnueabihf]
linker = "arm-linux-gnueabihf-gcc"
[target.aarch64-unknown-linux-gnu]
linker = "aarch64-linux-gnu-gcc"
EOF

View File

@@ -1,14 +0,0 @@
#!/usr/bin/env bash
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_BUILD_TAG
set -ex
pushd /usr/local
git clone https://gitlab.freedesktop.org/mesa/shader-db.git --depth 1
rm -rf shader-db/.git
cd shader-db
make
popd

View File

@@ -1,89 +0,0 @@
#!/usr/bin/env bash
# SPDX-License-Identifier: MIT
#
# Copyright © 2022 Collabora Limited
# Author: Guilherme Gallo <guilherme.gallo@collabora.com>
#
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# KERNEL_ROOTFS_TAG
SKQP_BRANCH=android-cts-12.1_r5
# hack for skqp see the clang
pushd /usr/bin/
ln -s ../lib/llvm-15/bin/clang clang
ln -s ../lib/llvm-15/bin/clang++ clang++
popd
create_gn_args() {
# gn can be configured to cross-compile skia and its tools
# It is important to set the target_cpu to guarantee the intended target
# machine
cp "${BASE_ARGS_GN_FILE}" "${SKQP_OUT_DIR}"/args.gn
echo "target_cpu = \"${SKQP_ARCH}\"" >> "${SKQP_OUT_DIR}"/args.gn
}
download_skia_source() {
if [ -z ${SKIA_DIR+x} ]
then
return 1
fi
# Skia cloned from https://android.googlesource.com/platform/external/skqp
# has all needed assets tracked on git-fs
SKQP_REPO=https://android.googlesource.com/platform/external/skqp
git clone --branch "${SKQP_BRANCH}" --depth 1 "${SKQP_REPO}" "${SKIA_DIR}"
}
set -ex
SCRIPT_DIR=$(realpath "$(dirname "$0")")
SKQP_PATCH_DIR="${SCRIPT_DIR}/patches"
BASE_ARGS_GN_FILE="${SCRIPT_DIR}/build-skqp_base.gn"
SKQP_ARCH=${SKQP_ARCH:-x64}
SKIA_DIR=${SKIA_DIR:-$(mktemp -d)}
SKQP_OUT_DIR=${SKIA_DIR}/out/${SKQP_ARCH}
SKQP_INSTALL_DIR=${SKQP_INSTALL_DIR:-/skqp}
SKQP_ASSETS_DIR="${SKQP_INSTALL_DIR}/assets"
SKQP_BINARIES=(skqp list_gpu_unit_tests list_gms)
download_skia_source
pushd "${SKIA_DIR}"
# Apply all skqp patches for Mesa CI
cat "${SKQP_PATCH_DIR}"/build-skqp_*.patch |
patch -p1
# Fetch some needed build tools needed to build skia/skqp.
# Basically, it clones repositories with commits SHAs from ${SKIA_DIR}/DEPS
# directory.
python tools/git-sync-deps
mkdir -p "${SKQP_OUT_DIR}"
mkdir -p "${SKQP_INSTALL_DIR}"
create_gn_args
# Build and install skqp binaries
bin/gn gen "${SKQP_OUT_DIR}"
for BINARY in "${SKQP_BINARIES[@]}"
do
/usr/bin/ninja -C "${SKQP_OUT_DIR}" "${BINARY}"
# Strip binary, since gn is not stripping it even when `is_debug == false`
${STRIP_CMD:-strip} "${SKQP_OUT_DIR}/${BINARY}"
install -m 0755 "${SKQP_OUT_DIR}/${BINARY}" "${SKQP_INSTALL_DIR}"
done
# Move assets to the target directory, which will reside in rootfs.
mv platform_tools/android/apps/skqp/src/main/assets/ "${SKQP_ASSETS_DIR}"
popd
rm -Rf "${SKIA_DIR}"
set +ex

View File

@@ -1,59 +0,0 @@
cc = "clang"
cxx = "clang++"
extra_cflags = [
"-Wno-error",
"-DSK_ENABLE_DUMP_GPU",
"-DSK_BUILD_FOR_SKQP"
]
extra_cflags_cc = [
"-Wno-error",
# skqp build process produces a lot of compilation warnings, silencing
# most of them to remove clutter and avoid the CI job log to exceed the
# maximum size
# GCC flags
"-Wno-redundant-move",
"-Wno-suggest-override",
"-Wno-class-memaccess",
"-Wno-deprecated-copy",
"-Wno-uninitialized",
# Clang flags
"-Wno-macro-redefined",
"-Wno-anon-enum-enum-conversion",
"-Wno-suggest-destructor-override",
"-Wno-return-std-move-in-c++11",
"-Wno-extra-semi-stmt",
"-Wno-reserved-identifier",
"-Wno-bitwise-instead-of-logical",
"-Wno-reserved-identifier",
"-Wno-psabi",
"-Wno-unused-but-set-variable",
"-Wno-sizeof-array-div",
"-Wno-string-concatenation",
]
cc_wrapper = "ccache"
is_debug = false
skia_enable_fontmgr_android = false
skia_enable_fontmgr_empty = true
skia_enable_pdf = false
skia_enable_skottie = false
skia_skqp_global_error_tolerance = 8
skia_tools_require_resources = true
skia_use_dng_sdk = false
skia_use_expat = true
skia_use_icu = false
skia_use_libheif = false
skia_use_lua = false
skia_use_piex = false
skia_use_vulkan = true
target_os = "linux"

View File

@@ -1,25 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# KERNEL_ROOTFS_TAG
set -ex
git config --global user.email "mesa@example.com"
git config --global user.name "Mesa CI"
git clone \
https://github.com/intel/libva-utils.git \
-b 2.18.1 \
--depth 1 \
/va-utils
pushd /va-utils
# Too old libva in Debian 11. TODO: when this PR gets in, refer to the patch.
curl -L https://github.com/intel/libva-utils/pull/329.patch | git am
meson setup build -D tests=true -Dprefix=/va $EXTRA_MESON_ARGS
meson install -C build
popd
rm -rf /va-utils

View File

@@ -1,43 +0,0 @@
#!/bin/bash
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_X86_64_TEST_VK_TAG
# KERNEL_ROOTFS_TAG
set -ex
VKD3D_PROTON_COMMIT="a0ccc383937903f4ca0997ce53e41ccce7f2f2ec"
VKD3D_PROTON_DST_DIR="/vkd3d-proton-tests"
VKD3D_PROTON_SRC_DIR="/vkd3d-proton-src"
VKD3D_PROTON_BUILD_DIR="/vkd3d-proton-$VKD3D_PROTON_VERSION"
function build_arch {
local arch="$1"
shift
meson "$@" \
-Denable_tests=true \
--buildtype release \
--prefix "$VKD3D_PROTON_DST_DIR" \
--strip \
--bindir "x${arch}" \
--libdir "x${arch}" \
"$VKD3D_PROTON_BUILD_DIR/build.${arch}"
ninja -C "$VKD3D_PROTON_BUILD_DIR/build.${arch}" install
install -D -m755 -t "${VKD3D_PROTON_DST_DIR}/x${arch}/bin" "$VKD3D_PROTON_BUILD_DIR/build.${arch}/tests/d3d12"
}
git clone https://github.com/HansKristian-Work/vkd3d-proton.git --single-branch -b master --no-checkout "$VKD3D_PROTON_SRC_DIR"
pushd "$VKD3D_PROTON_SRC_DIR"
git checkout "$VKD3D_PROTON_COMMIT"
git submodule update --init --recursive
git submodule update --recursive
build_arch 64
build_arch 86
popd
rm -rf "$VKD3D_PROTON_BUILD_DIR"
rm -rf "$VKD3D_PROTON_SRC_DIR"

View File

@@ -1,18 +0,0 @@
#!/usr/bin/env bash
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_X86_64_TEST_GL_TAG
# KERNEL_ROOTFS_TAG:
set -ex
VALIDATION_TAG="snapshot-2024wk06"
git clone -b "$VALIDATION_TAG" --single-branch --depth 1 https://github.com/KhronosGroup/Vulkan-ValidationLayers.git
pushd Vulkan-ValidationLayers
python3 scripts/update_deps.py --dir external --config debug
cmake -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/usr -DBUILD_TESTS=OFF -DBUILD_WERROR=OFF -C external/helper.cmake -S . -B build
ninja -C build install
popd
rm -rf Vulkan-ValidationLayers

View File

@@ -1,23 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
set -ex
export LIBWAYLAND_VERSION="1.21.0"
export WAYLAND_PROTOCOLS_VERSION="1.31"
git clone https://gitlab.freedesktop.org/wayland/wayland
cd wayland
git checkout "$LIBWAYLAND_VERSION"
meson setup -Ddocumentation=false -Ddtd_validation=false -Dlibraries=true _build $EXTRA_MESON_ARGS
meson install -C _build
cd ..
rm -rf wayland
git clone https://gitlab.freedesktop.org/wayland/wayland-protocols
cd wayland-protocols
git checkout "$WAYLAND_PROTOCOLS_VERSION"
meson setup _build $EXTRA_MESON_ARGS
meson install -C _build
cd ..
rm -rf wayland-protocols

View File

@@ -1,14 +0,0 @@
#!/usr/bin/env bash
if test -f /etc/debian_version; then
apt-get autoremove -y --purge
fi
# Clean up any build cache
rm -rf /root/.cache
rm -rf /root/.cargo
rm -rf /.cargo
if test -x /usr/bin/ccache; then
ccache --show-stats
fi

View File

@@ -1,52 +0,0 @@
#!/bin/sh
if test -x /usr/bin/ccache; then
if test -f /etc/debian_version; then
CCACHE_PATH=/usr/lib/ccache
elif test -f /etc/alpine-release; then
CCACHE_PATH=/usr/lib/ccache/bin
else
CCACHE_PATH=/usr/lib64/ccache
fi
# Common setup among container builds before we get to building code.
export CCACHE_COMPILERCHECK=content
export CCACHE_COMPRESS=true
export CCACHE_DIR=/cache/$CI_PROJECT_NAME/ccache
export PATH=$CCACHE_PATH:$PATH
# CMake ignores $PATH, so we have to force CC/GCC to the ccache versions.
export CC="${CCACHE_PATH}/gcc"
export CXX="${CCACHE_PATH}/g++"
ccache --show-stats
fi
# When not using the mold linker (e.g. unsupported architecture), force
# linkers to gold, since it's so much faster for building. We can't use
# lld because we're on old debian and it's buggy. mingw fails meson builds
# with it with "meson.build:21:0: ERROR: Unable to determine dynamic linker"
find /usr/bin -name \*-ld -o -name ld | \
grep -v mingw | \
xargs -n 1 -I '{}' ln -sf '{}.gold' '{}'
# Make a wrapper script for ninja to always include the -j flags
{
echo '#!/bin/sh -x'
# shellcheck disable=SC2016
echo '/usr/bin/ninja -j${FDO_CI_CONCURRENT:-4} "$@"'
} > /usr/local/bin/ninja
chmod +x /usr/local/bin/ninja
# Set MAKEFLAGS so that all make invocations in container builds include the
# flags (doesn't apply to non-container builds, but we don't run make there)
export MAKEFLAGS="-j${FDO_CI_CONCURRENT:-4}"
# make wget to try more than once, when download fails or timeout
echo -e "retry_connrefused = on\n" \
"read_timeout = 300\n" \
"tries = 4\n" \
"retry_on_host_error = on\n" \
"retry_on_http_error = 429,500,502,503,504\n" \
"wait_retry = 32" >> /etc/wgetrc

View File

@@ -1,37 +0,0 @@
#!/bin/bash
ndk=$1
arch=$2
cpu_family=$3
cpu=$4
cross_file="/cross_file-$arch.txt"
sdk_version=$5
# armv7 has the toolchain split between two names.
arch2=${6:-$2}
# Note that we disable C++ exceptions, because Mesa doesn't use exceptions,
# and allowing it in code generation means we get unwind symbols that break
# the libEGL and driver symbol tests.
cat > "$cross_file" <<EOF
[binaries]
ar = '$ndk/toolchains/llvm/prebuilt/linux-x86_64/bin/llvm-ar'
c = ['ccache', '$ndk/toolchains/llvm/prebuilt/linux-x86_64/bin/${arch2}${sdk_version}-clang', '-fno-exceptions', '-fno-unwind-tables', '-fno-asynchronous-unwind-tables']
cpp = ['ccache', '$ndk/toolchains/llvm/prebuilt/linux-x86_64/bin/${arch2}${sdk_version}-clang++', '-fno-exceptions', '-fno-unwind-tables', '-fno-asynchronous-unwind-tables', '-static-libstdc++']
c_ld = 'lld'
cpp_ld = 'lld'
strip = '$ndk/toolchains/llvm/prebuilt/linux-x86_64/bin/llvm-strip'
pkgconfig = ['/usr/bin/pkgconf']
[host_machine]
system = 'android'
cpu_family = '$cpu_family'
cpu = '$cpu'
endian = 'little'
[properties]
needs_exe_wrapper = true
pkg_config_libdir = '/usr/local/lib/${arch2}/pkgconfig/:/${ndk}/toolchains/llvm/prebuilt/linux-x86_64/sysroot/usr/lib/${arch2}/pkgconfig/'
EOF

View File

@@ -1,40 +0,0 @@
#!/bin/sh
# shellcheck disable=SC2086 # we want word splitting
# Makes a .pc file in the Android NDK for meson to find its libraries.
set -ex
ndk="$1"
pc="$2"
cflags="$3"
libs="$4"
version="$5"
sdk_version="$6"
sysroot=$ndk/toolchains/llvm/prebuilt/linux-x86_64/sysroot
for arch in \
x86_64-linux-android \
i686-linux-android \
aarch64-linux-android \
arm-linux-androideabi; do
pcdir=$sysroot/usr/lib/$arch/pkgconfig
mkdir -p $pcdir
cat >$pcdir/$pc <<EOF
prefix=$sysroot
exec_prefix=$sysroot
libdir=$sysroot/usr/lib/$arch/$sdk_version
sharedlibdir=$sysroot/usr/lib/$arch
includedir=$sysroot/usr/include
Name: zlib
Description: zlib compression library
Version: $version
Requires:
Libs: -L$sysroot/usr/lib/$arch/$sdk_version $libs
Cflags: -I$sysroot/usr/include $cflags
EOF
done

View File

@@ -1,54 +0,0 @@
#!/bin/bash
arch=$1
cross_file="/cross_file-$arch.txt"
meson env2mfile --cross --debarch "$arch" -o "$cross_file"
# Explicitly set ccache path for cross compilers
sed -i "s|/usr/bin/\([^-]*\)-linux-gnu\([^-]*\)-g|/usr/lib/ccache/\\1-linux-gnu\\2-g|g" "$cross_file"
# Rely on qemu-user being configured in binfmt_misc on the host
# shellcheck disable=SC1003 # how this sed doesn't seems to work for me locally
sed -i -e '/\[properties\]/a\' -e "needs_exe_wrapper = False" "$cross_file"
# Add a line for rustc, which meson env2mfile is missing.
cc=$(sed -n "s|^c\s*=\s*\[?'\(.*\)'\]?|\1|p" < "$cross_file")
if [[ "$arch" = "arm64" ]]; then
rust_target=aarch64-unknown-linux-gnu
elif [[ "$arch" = "armhf" ]]; then
rust_target=armv7-unknown-linux-gnueabihf
elif [[ "$arch" = "i386" ]]; then
rust_target=i686-unknown-linux-gnu
elif [[ "$arch" = "ppc64el" ]]; then
rust_target=powerpc64le-unknown-linux-gnu
elif [[ "$arch" = "s390x" ]]; then
rust_target=s390x-unknown-linux-gnu
else
echo "Needs rustc target mapping"
fi
# shellcheck disable=SC1003 # how this sed doesn't seems to work for me locally
sed -i -e '/\[binaries\]/a\' -e "rust = ['rustc', '--target=$rust_target', '-C', 'linker=$cc']" "$cross_file"
# Set up cmake cross compile toolchain file for dEQP builds
toolchain_file="/toolchain-$arch.cmake"
if [[ "$arch" = "arm64" ]]; then
GCC_ARCH="aarch64-linux-gnu"
DE_CPU="DE_CPU_ARM_64"
elif [[ "$arch" = "armhf" ]]; then
GCC_ARCH="arm-linux-gnueabihf"
DE_CPU="DE_CPU_ARM"
fi
if [[ -n "$GCC_ARCH" ]]; then
{
echo "set(CMAKE_SYSTEM_NAME Linux)";
echo "set(CMAKE_SYSTEM_PROCESSOR arm)";
echo "set(CMAKE_C_COMPILER /usr/lib/ccache/$GCC_ARCH-gcc)";
echo "set(CMAKE_CXX_COMPILER /usr/lib/ccache/$GCC_ARCH-g++)";
echo "set(CMAKE_CXX_FLAGS_INIT \"-Wno-psabi\")"; # makes ABI warnings quiet for ARMv7
echo "set(ENV{PKG_CONFIG} \"/usr/bin/$GCC_ARCH-pkgconf\")";
echo "set(DE_CPU $DE_CPU)";
} > "$toolchain_file"
fi

View File

@@ -1,92 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
set -e
set -o xtrace
export DEBIAN_FRONTEND=noninteractive
export LLVM_VERSION="${LLVM_VERSION:=15}"
# Ephemeral packages (installed for this script and removed again at the end)
EPHEMERAL=(
)
DEPS=(
"crossbuild-essential-$arch"
"pkgconf:$arch"
"libasan8:$arch"
"libdrm-dev:$arch"
"libelf-dev:$arch"
"libexpat1-dev:$arch"
"libffi-dev:$arch"
"libpciaccess-dev:$arch"
"libstdc++6:$arch"
"libvulkan-dev:$arch"
"libx11-dev:$arch"
"libx11-xcb-dev:$arch"
"libxcb-dri2-0-dev:$arch"
"libxcb-dri3-dev:$arch"
"libxcb-glx0-dev:$arch"
"libxcb-present-dev:$arch"
"libxcb-randr0-dev:$arch"
"libxcb-shm0-dev:$arch"
"libxcb-xfixes0-dev:$arch"
"libxdamage-dev:$arch"
"libxext-dev:$arch"
"libxrandr-dev:$arch"
"libxshmfence-dev:$arch"
"libxxf86vm-dev:$arch"
"libwayland-dev:$arch"
)
dpkg --add-architecture $arch
echo "deb [trusted=yes] https://gitlab.freedesktop.org/gfx-ci/ci-deb-repo/-/raw/${PKG_REPO_REV}/ ${FDO_DISTRIBUTION_VERSION%-*} main" | tee /etc/apt/sources.list.d/gfx-ci_.list
apt-get update
apt-get install -y --no-remove "${DEPS[@]}" "${EPHEMERAL[@]}" \
$EXTRA_LOCAL_PACKAGES
if [[ $arch != "armhf" ]]; then
# We don't need clang-format for the crossbuilds, but the installed amd64
# package will conflict with libclang. Uninstall clang-format (and its
# problematic dependency) to fix.
apt-get remove -y "clang-format-${LLVM_VERSION}" "libclang-cpp${LLVM_VERSION}" \
"llvm-${LLVM_VERSION}-runtime" "llvm-${LLVM_VERSION}-linker-tools"
# llvm-*-tools:$arch conflicts with python3:amd64. Install dependencies only
# with apt-get, then force-install llvm-*-{dev,tools}:$arch with dpkg to get
# around this.
apt-get install -y --no-remove --no-install-recommends \
"libclang-cpp${LLVM_VERSION}:$arch" \
"libgcc-s1:$arch" \
"libtinfo-dev:$arch" \
"libz3-dev:$arch" \
"llvm-${LLVM_VERSION}:$arch" \
zlib1g
fi
. .gitlab-ci/container/create-cross-file.sh $arch
. .gitlab-ci/container/container_pre_build.sh
# dependencies where we want a specific version
MULTIARCH_PATH=$(dpkg-architecture -A $arch -qDEB_TARGET_MULTIARCH)
export EXTRA_MESON_ARGS="--cross-file=/cross_file-${arch}.txt -D libdir=lib/${MULTIARCH_PATH}"
. .gitlab-ci/container/build-wayland.sh
. .gitlab-ci/container/build-directx-headers.sh
apt-get purge -y "${EPHEMERAL[@]}"
. .gitlab-ci/container/container_post_build.sh
# This needs to be done after container_post_build.sh, or apt-get breaks in there
if [[ $arch != "armhf" ]]; then
apt-get download llvm-"${LLVM_VERSION}"-{dev,tools}:"$arch"
dpkg -i --force-depends llvm-"${LLVM_VERSION}"-*_"${arch}".deb
rm llvm-"${LLVM_VERSION}"-*_"${arch}".deb
fi

View File

@@ -1,93 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
# When changing this file, you need to bump the following
# .gitlab-ci/image-tags.yml tags:
# DEBIAN_BUILD_TAG
set -ex
EPHEMERAL=(
autoconf
rdfind
unzip
)
apt-get install -y --no-remove "${EPHEMERAL[@]}"
# Fetch the NDK and extract just the toolchain we want.
ndk=$ANDROID_NDK
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-o $ndk.zip https://dl.google.com/android/repository/$ndk-linux.zip
unzip -d / $ndk.zip "$ndk/toolchains/llvm/*"
rm $ndk.zip
# Since it was packed as a zip file, symlinks/hardlinks got turned into
# duplicate files. Turn them into hardlinks to save on container space.
rdfind -makehardlinks true -makeresultsfile false /${ndk}/
# Drop some large tools we won't use in this build.
find /${ndk}/ -type f \( -iname '*clang-check*' -o -iname '*clang-tidy*' -o -iname '*lldb*' \) -exec rm -f {} \;
sh .gitlab-ci/container/create-android-ndk-pc.sh /$ndk zlib.pc "" "-lz" "1.2.3" $ANDROID_SDK_VERSION
sh .gitlab-ci/container/create-android-cross-file.sh /$ndk x86_64-linux-android x86_64 x86_64 $ANDROID_SDK_VERSION
sh .gitlab-ci/container/create-android-cross-file.sh /$ndk i686-linux-android x86 x86 $ANDROID_SDK_VERSION
sh .gitlab-ci/container/create-android-cross-file.sh /$ndk aarch64-linux-android aarch64 armv8 $ANDROID_SDK_VERSION
sh .gitlab-ci/container/create-android-cross-file.sh /$ndk arm-linux-androideabi arm armv7hl $ANDROID_SDK_VERSION armv7a-linux-androideabi
for arch in \
x86_64-linux-android \
i686-linux-android \
aarch64-linux-android \
arm-linux-androideabi ; do
EXTRA_MESON_ARGS="--cross-file=/cross_file-$arch.txt --libdir=lib/$arch -Dnouveau=disabled -Dintel=disabled" \
. .gitlab-ci/container/build-libdrm.sh
done
rm -rf $LIBDRM_VERSION
export LIBELF_VERSION=libelf-0.8.13
curl -L --retry 4 -f --retry-all-errors --retry-delay 60 \
-O https://fossies.org/linux/misc/old/$LIBELF_VERSION.tar.gz
# Not 100% sure who runs the mirror above so be extra careful
if ! echo "4136d7b4c04df68b686570afa26988ac ${LIBELF_VERSION}.tar.gz" | md5sum -c -; then
echo "Checksum failed"
exit 1
fi
tar -xf ${LIBELF_VERSION}.tar.gz
cd $LIBELF_VERSION
# Work around a bug in the original configure not enabling __LIBELF64.
autoreconf
for arch in \
x86_64-linux-android \
i686-linux-android \
aarch64-linux-android \
arm-linux-androideabi ; do
ccarch=${arch}
if [ "${arch}" == 'arm-linux-androideabi' ]
then
ccarch=armv7a-linux-androideabi
fi
export CC=/${ndk}/toolchains/llvm/prebuilt/linux-x86_64/bin/llvm-ar
export CC=/${ndk}/toolchains/llvm/prebuilt/linux-x86_64/bin/${ccarch}${ANDROID_SDK_VERSION}-clang
export CXX=/${ndk}/toolchains/llvm/prebuilt/linux-x86_64/bin/${ccarch}${ANDROID_SDK_VERSION}-clang++
export LD=/${ndk}/toolchains/llvm/prebuilt/linux-x86_64/bin/${arch}-ld
export RANLIB=/${ndk}/toolchains/llvm/prebuilt/linux-x86_64/bin/llvm-ranlib
# The configure script doesn't know about android, but doesn't really use the host anyway it
# seems
./configure --host=x86_64-linux-gnu --disable-nls --disable-shared \
--libdir=/usr/local/lib/${arch}
make install
make distclean
done
cd ..
rm -rf $LIBELF_VERSION
apt-get purge -y "${EPHEMERAL[@]}"

View File

@@ -1,5 +0,0 @@
#!/usr/bin/env bash
set -e
arch=armhf . .gitlab-ci/container/debian/arm_test.sh

View File

@@ -1,90 +0,0 @@
#!/usr/bin/env bash
# shellcheck disable=SC2086 # we want word splitting
set -e
set -o xtrace
export LLVM_VERSION="${LLVM_VERSION:=15}"
apt-get -y install ca-certificates
sed -i -e 's/http:\/\/deb/https:\/\/deb/g' /etc/apt/sources.list.d/*
echo "deb [trusted=yes] https://gitlab.freedesktop.org/gfx-ci/ci-deb-repo/-/raw/${PKG_REPO_REV}/ ${FDO_DISTRIBUTION_VERSION%-*} main" | tee /etc/apt/sources.list.d/gfx-ci_.list
apt-get update
# Ephemeral packages (installed for this script and removed again at the end)
EPHEMERAL=(
libssl-dev
)
DEPS=(
apt-utils
android-libext4-utils
autoconf
automake
bc
bison
ccache
cmake
curl
fastboot
flatbuffers-compiler
flex
g++
git
glslang-tools
kmod
libasan8
libdrm-dev
libelf-dev
libexpat1-dev
libflatbuffers-dev
libvulkan-dev
libx11-dev
libx11-xcb-dev
libxcb-dri2-0-dev
libxcb-dri3-dev
libxcb-glx0-dev
libxcb-present-dev
libxcb-randr0-dev
libxcb-shm0-dev
libxcb-xfixes0-dev
libxdamage-dev
libxext-dev
libxrandr-dev
libxshmfence-dev
libxtensor-dev
libxxf86vm-dev
libwayland-dev
libwayland-egl-backend-dev
"llvm-${LLVM_VERSION}-dev"
ninja-build
meson
openssh-server
pkgconf
python3-mako
python3-pil
python3-pip
python3-requests
python3-setuptools
u-boot-tools
xz-utils
zlib1g-dev
zstd
)
apt-get -y install "${DEPS[@]}" "${EPHEMERAL[@]}"
pip3 install --break-system-packages git+http://gitlab.freedesktop.org/freedesktop/ci-templates@ffe4d1b10aab7534489f0c4bbc4c5899df17d3f2
arch=armhf
. .gitlab-ci/container/cross_build.sh
. .gitlab-ci/container/container_pre_build.sh
. .gitlab-ci/container/build-mold.sh
. .gitlab-ci/container/build-wayland.sh
apt-get purge -y "${EPHEMERAL[@]}"
. .gitlab-ci/container/container_post_build.sh

Some files were not shown because too many files have changed in this diff Show More