Alyssa Rosenzweig [Fri, 2 Aug 2019 16:06:13 +0000 (09:06 -0700)]
pan/midgard: Break mir_spill_register into its function
No functional changes, just breaks out a megamonster function and fixes
the indentation.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 2 Aug 2019 22:25:02 +0000 (15:25 -0700)]
pan/midgard: Switch sources to an array for trinary sources
We need three independent sources to support indirect SSBO writes (as
well as textures with both LOD/bias and offsets). Now is a good time to
make sources just an array so we don't have to rewrite a ton of code if
we ever needed a fourth source for some reason.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Thu, 1 Aug 2019 21:28:34 +0000 (14:28 -0700)]
pan/midgard: Remove "r27-only" register class
As far as I know, there's no such thing as a load/store op that only
takes its argument in r27. We just need to set the appropriate arg_1
field in the RA to specify other registers if we want them.
To facilitate this, various RA-related changes are needed across the
compiler ; this should also fix indirect offsets which were implicitly
interpreted as "r27-only" despite not even passing through RA yet. One
ripple effect change is switching the move insertion point and adjusting
the liveness analysis accordingly, so while this was intended as a
purely functional change, there are some shader-db changes:
total instructions in shared programs: 3511 -> 3498 (-0.37%)
instructions in affected programs: 563 -> 550 (-2.31%)
helped: 12
HURT: 0
helped stats (abs) min: 1 max: 2 x̄: 1.08 x̃: 1
helped stats (rel) min: 0.93% max: 5.00% x̄: 2.58% x̃: 2.33%
95% mean confidence interval for instructions value: -1.27 -0.90
95% mean confidence interval for instructions %-change: -3.23% -1.93%
Instructions are helped.
total bundles in shared programs: 2067 -> 2067 (0.00%)
bundles in affected programs: 398 -> 398 (0.00%)
helped: 7
HURT: 4
helped stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1
helped stats (rel) min: 1.54% max: 10.00% x̄: 5.04% x̃: 5.56%
HURT stats (abs) min: 1 max: 2 x̄: 1.75 x̃: 2
HURT stats (rel) min: 2.13% max: 4.26% x̄: 3.72% x̃: 4.26%
95% mean confidence interval for bundles value: -0.95 0.95
95% mean confidence interval for bundles %-change: -5.21% 1.50%
Inconclusive result (value mean confidence interval includes 0).
total quadwords in shared programs: 3464 -> 3454 (-0.29%)
quadwords in affected programs: 1199 -> 1189 (-0.83%)
helped: 18
HURT: 4
helped stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1
helped stats (rel) min: 1.03% max: 5.26% x̄: 2.44% x̃: 1.79%
HURT stats (abs) min: 2 max: 2 x̄: 2.00 x̃: 2
HURT stats (rel) min: 2.56% max: 2.82% x̄: 2.63% x̃: 2.56%
95% mean confidence interval for quadwords value: -0.98 0.07
Inconclusive result (value mean confidence interval includes 0).
total registers in shared programs: 383 -> 373 (-2.61%)
registers in affected programs: 56 -> 46 (-17.86%)
helped: 12
HURT: 2
helped stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1
helped stats (rel) min: 9.09% max: 33.33% x̄: 29.58% x̃: 33.33%
HURT stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1
HURT stats (rel) min: 20.00% max: 50.00% x̄: 35.00% x̃: 35.00%
95% mean confidence interval for registers value: -1.13 -0.29
95% mean confidence interval for registers %-change: -35.07% -5.63%
Registers are helped.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 2 Aug 2019 18:22:56 +0000 (11:22 -0700)]
pan/midgard: Handle get/set_swizzle for load/store arguments
Load/store's main "argument 0" already has its swizzle handled
correctly (for stores, that is). But the tinier arguments, the compact
ones with a component select but not a full swizzle, those are not yet
handled. Let's do something about that!
Alyssa Rosenzweig [Fri, 2 Aug 2019 20:48:27 +0000 (13:48 -0700)]
pan/midgard: Fix block successors
Rather than an ersatz thing that sort of looks like successors but is in
fact just the source order traversal with some backward jumps hacked in
for loops... construct an actual flow graph so we can do analysis
sanely.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Thu, 1 Aug 2019 21:14:43 +0000 (14:14 -0700)]
pan/midgard: Add helper to pack load/store registers
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Thu, 1 Aug 2019 21:06:19 +0000 (14:06 -0700)]
pan/midgard: Decode register/component in load/store argument
3-bits out of 8 down!
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Thu, 1 Aug 2019 21:06:02 +0000 (14:06 -0700)]
pan/midgard: Fix REGISTER_OFFSET
r27 isn't the special one, usually.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Thu, 1 Aug 2019 20:29:01 +0000 (13:29 -0700)]
pan/midgard: Split ld/st unknown to arg_1/arg_2 fields
The 16-bit field can be decomposed to two independent 8-bit fields, each
representing a single (additional) argument to the load/store op,
generally used for encoding registers. Addressable registers here are
substantially limited compared to the main register in a load/store op.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Bas Nieuwenhuizen [Fri, 2 Aug 2019 12:45:49 +0000 (14:45 +0200)]
radv: Expose VK_KHR_imageless_framebuffer.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Bas Nieuwenhuizen [Fri, 2 Aug 2019 12:42:50 +0000 (14:42 +0200)]
radv: Implement VK_KHR_imageless_framebuffer.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Bas Nieuwenhuizen [Fri, 2 Aug 2019 13:22:37 +0000 (15:22 +0200)]
radv: Store image view also outside framebuffer.
So we can use it with imageless framebuffers.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Bas Nieuwenhuizen [Mon, 22 Jul 2019 00:31:27 +0000 (02:31 +0200)]
radv: Store color/depth surface info in attachment info instead of framebuffer.
That way we can use it for imageless framebuffers.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Alyssa Rosenzweig [Fri, 2 Aug 2019 17:18:48 +0000 (19:18 +0200)]
panfrost: Allocate polygon lists on-demand
Rather than alloacting a huge (64MB) polygon list on context creation
and sharing it across framebuffers, we instead allocate polygon lists as
BOs (which consistently hit the cache) sized appropriately; for about a
month, we've known how to calculate the polygon list size so this has
only recently become possible.
The good news is we can render to truly massive framebuffers without
crashing and, more importantly, we eliminate the 64MB upfront overhead.
If a list that size isn't actually needed, it's not allocated.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Boris Brezillon [Fri, 2 Aug 2019 17:18:47 +0000 (19:18 +0200)]
panfrost: Handle the bo == NULL case in panfrost_bo_[un]reference()
Allows us to pass BOs without checking if they're NULL or not.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Boris Brezillon [Fri, 2 Aug 2019 17:18:46 +0000 (19:18 +0200)]
panfrost: Get rid of the skippable param in attach_vt_framebuffer()
The only user of this function always passes true.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Boris Brezillon [Fri, 2 Aug 2019 17:18:45 +0000 (19:18 +0200)]
panfrost: Don't emit a new FB desc when setting a new FB state
The FB desc will be emitted/attached on the first draw targetting
this new FB.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Boris Brezillon [Fri, 2 Aug 2019 17:18:44 +0000 (19:18 +0200)]
panfrost: Bail out early when doing a wallpaper blit
The wallpaper blit is a bit special in that the operation is targetting
the current FB, but the u_blitter logic creates a new surface for it
which makes util_framebuffer_state_equal() return false. In that case
we don't want a new FB descriptor to be emitted/attached, so let's just
copy the new state into ctx->pipe_framebuffer and exit the function.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Boris Brezillon [Fri, 2 Aug 2019 17:18:43 +0000 (19:18 +0200)]
panfrost: Bail out early when new and current FB states are equal
If the current FB matches the new one there's nothing to be done in
panfrost_set_framebuffer_state(). By bailing out early in that case we
avoid emitting new FB descriptors (the old ones are still valid).
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Boris Brezillon [Fri, 2 Aug 2019 17:18:42 +0000 (19:18 +0200)]
panfrost: Delay FB descriptor allocation
No need to emit SFBD/MFBD at frame invalidation. They can be emitted
when the framebuffer is attached, which saves us a potential FB desc
re-allocation if a new FB is bound after the swap.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Boris Brezillon [Fri, 2 Aug 2019 17:18:41 +0000 (19:18 +0200)]
panfrost: Remove job from ctx->jobs at submission time
This guarantees that new draws targetting the same framebuffer will
get a new job instance.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Boris Brezillon [Fri, 2 Aug 2019 17:18:40 +0000 (19:18 +0200)]
panfrost: Make ctx->job useful
ctx->job is supposed to serve as a cache to avoid an hash table lookup
everytime we access the job attached to the currently bound FB, except
it was never assigned to anything but NULL.
Fix that by adding the missing assignment in panfrost_get_job_for_fbo().
Also add a missing NULL assignment in the ->set_framebuffer_state()
path.
While at it, add extra assert()s to make sure ctx->job is consistent.
Fixes: 59c9623d0a75 ("panfrost: Import job data structures from v3d")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Bas Nieuwenhuizen [Fri, 2 Aug 2019 10:40:17 +0000 (12:40 +0200)]
ac/nir,radv: Optimize bounds check for 64 bit CAS.
When the application does not ask for robust buffer access.
Only implemented the check in radv.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Roland Scheidegger [Tue, 30 Jul 2019 21:35:49 +0000 (23:35 +0200)]
gallivm: fix issue with AtomicCmpXchg wrapper on llvm 3.5-3.8
These versions still need wrapper but already have both success and
failure ordering.
(Compile tested on llvm 3.3, 3.7, 3.8.)
v2: don't duplicate whole function (suggested by Brian).
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111102
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Matt Turner [Wed, 31 Jul 2019 21:44:39 +0000 (14:44 -0700)]
util: Handle differences in pthread_setname_np
There are a lot of unfortunate differences in the implementation of this
function. NetBSD and Mac OS X in particular require different arguments.
https://stackoverflow.com/questions/
2369738/how-to-set-the-name-of-a-thread-in-linux-pthreads/
7989973#
7989973
provides for a good overview of the differences.
Fixes: 9c411e020d1 ("util: Drop preprocessor guards for glibc-2.12")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111264
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>
[Eric: use DETECT_OS_* instead of PIPE_OS_*]
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 21:38:00 +0000 (22:38 +0100)]
util/os_time: use detect_os.h to uncouple from gallium
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 21:36:30 +0000 (22:36 +0100)]
util/u_debug: use detect_os.h
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 21:33:05 +0000 (22:33 +0100)]
util/os_misc: use detect_os.h to start uncoupling from gallium
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 15:55:39 +0000 (16:55 +0100)]
util/os_memory: use detect_os.h to uncouple it from gallium
While at it, remove p_compiler.h as well as it is unused.
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 13:58:52 +0000 (14:58 +0100)]
gallium: deduplicate os detection logic by using detect_os.h
This allows us to avoid having to rename all the PIPE_OS_* at once while
still making sure PIPE_OS_* and DETECT_OS_* are always in sync.
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 21:49:05 +0000 (22:49 +0100)]
gallium/utils: drop PIPE_SUBSYSTEM_WINDOWS_USER
This is basically just an alias for PIPE_OS_WINDOWS.
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 20:45:25 +0000 (21:45 +0100)]
scons: rename PIPE_SUBSYSTEM_EMBEDDED to EMBEDDED_DEVICE
It has nothing to do with the PIPE_SUBSYSTEM_* stuff from gallium.
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 14:02:15 +0000 (15:02 +0100)]
gallium: remove never-used PIPE_SUBSYSTEM_DRI
PIPE_SUBSYSTEM_DRI was introduced in
dacfef158943665fc0d1 ("gallium: New
configuration header.") 11 years ago, and was never used.
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 14:01:54 +0000 (15:01 +0100)]
util: fix typo in comment
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Eric Engestrom [Thu, 1 Aug 2019 13:48:26 +0000 (14:48 +0100)]
util: introduce detect_os.h
Mostly copied from src/gallium/include/pipe/p_config.h, so I kept its
copyright and authorship.
Other than the obvious rename, the big difference is that these are
always defined, to be used as `#if DETECT_OS_LINUX`.
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Rob Clark [Wed, 31 Jul 2019 19:30:24 +0000 (12:30 -0700)]
freedreno/batch: fix dependency loop detection
We can have a scenario like:
A -> B
A -> C -> B
When adding the A->C dependency, it doesn't really matter that C depends
on something that A depends on, that isn't a necessary condition for a
dependency loop.
Instead what we want to know is that nothing C depends on, directly or
indirectly, depends on A. We can detect this by recursively OR'ing the
dependents_mask of C and all it's dependencies.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Tue, 30 Jul 2019 16:34:53 +0000 (09:34 -0700)]
freedreno/a6xx: add missing flush/invalidates for blit
Various things we were missing for multiple blits in a single batch.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Sat, 27 Jul 2019 16:00:37 +0000 (09:00 -0700)]
freedreno/a6xx: skip tiles with no geometry
If no clear, and no geometry according to VSC_STATE[pipe] we can skip
the tile entirely. If there is a fast-clear, we can't skip restore
(clear) or resolve IBs, but we can still skip draw IB.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Rob Clark [Fri, 26 Jul 2019 16:55:14 +0000 (09:55 -0700)]
freedreno/a6xx: VSC overflow detection/handling
Check VSC_SIZE/VSC_SIZE2 regs from cmdstream to detect overflow, and
skip use of VSC visibility stream when overflow is detected, to avoid
GPU hangs. This is done w/ introduction of some CP_REG_TEST/
CP_COND_REG_EXEC packet pairs.
In addition, eventually (after a frame or two) detect the condition and
resize the VSC buffers until overflow no longer happens.
Note that this significantly reduces the initial size of the VSC
buffers, backing out a previous hack to make them 16x larger than
what should be typically required (the previous "solution" for
VSC overflow).
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Rob Clark [Fri, 26 Jul 2019 16:05:58 +0000 (09:05 -0700)]
freedreno/a6xx: remove USE/IGNORE_VISIBILITY draw patching
Seems this isn't needed anymore on a6xx to control whether visibility
stream is used. And it would be hard to deal with if it was, for
disabling use of VSC stream in draw pass. So just remove it and
simplify things.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Rob Clark [Thu, 25 Jul 2019 23:34:25 +0000 (16:34 -0700)]
freedreno/a6xx: cleanup "blit_mem"
Rename to "control_mem", and switch to using a struct to manage the
layout, rather than just ad-hoc hard-coded offsets.
For recovering from VSC stream overflow, we'll need to add more, but
best to clean it up first.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Rob Clark [Wed, 24 Jul 2019 21:28:10 +0000 (14:28 -0700)]
freedreno: refresh tile debug
Fix some #ifdef'd bitrot, and get rid of #ifdef so it doesn't bitrot
again.
And add a prints for per-tile state.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Rob Clark [Thu, 25 Jul 2019 22:25:22 +0000 (15:25 -0700)]
freedreno: update registers
Pull in some updates of VSC regs
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Rob Clark [Thu, 25 Jul 2019 20:40:02 +0000 (13:40 -0700)]
freedreno/gmem: small cleanup
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Rob Clark [Mon, 29 Jul 2019 18:48:06 +0000 (11:48 -0700)]
freedreno/drm: convert ring_pool to child_pool
Worth another couple percent at driver2
Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Mon, 29 Jul 2019 17:27:18 +0000 (10:27 -0700)]
freedreno/drm: remove idx_lock
Since it ends up contended, it is a bit of a bottleneck for workloads
with high driver overhead. Worth nearly +10% at gfxbench driver2.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Sun, 28 Jul 2019 17:04:25 +0000 (10:04 -0700)]
freedreno/batch: always update last_fence
Not all flush paths come thru fd_context_flush(), so we should also set
last_fence in the batch flush path. This avoids some no-op flushes just
to get a fence. For example when pctx->flush_resource() triggers a
flush.
We should probably keep the last_fence update in fd_context_flush() as
well to handle deferred flush case.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Rob Clark [Tue, 30 Jul 2019 15:12:46 +0000 (08:12 -0700)]
freedreno: drop unused fd_fence_ref param
The pscreen param was just there to satisfy pipe_screen::fence_reference
But some of the internal uses passed NULL for screen. Which is a bit
ugly. Instead drop the param and add a shim function to plug into the
screen.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Alyssa Rosenzweig [Fri, 26 Jul 2019 21:59:00 +0000 (14:59 -0700)]
pan/midgard: Print invert modifier
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 26 Jul 2019 21:25:25 +0000 (14:25 -0700)]
pan/midgard: Flip conditionals
We would like to flip ops to have a constant in the second place to
enable inlining of the constant.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 26 Jul 2019 20:32:54 +0000 (13:32 -0700)]
pan/midgard: Add bitwise src/invert fusing
De Morgan's Laws and some special ops basically.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 26 Jul 2019 20:14:55 +0000 (13:14 -0700)]
pan/midgard: Add .not propagation pass
Essentially .pos propagation but for bitwise.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 26 Jul 2019 20:08:54 +0000 (13:08 -0700)]
pan/midgard: Fuse invert into bitwise ops
We use the new invert flag to produce ops like inand.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Jonathan Marek [Thu, 1 Aug 2019 18:41:44 +0000 (14:41 -0400)]
freedreno: a2xx: implement texture tiling
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Jonathan Marek [Thu, 1 Aug 2019 19:52:58 +0000 (15:52 -0400)]
freedreno: a2xx: use nir_lower_alu_to_scalar instead of lowering pass
nir_lower_alu_to_scalar can now be used to only lower certain ops, so we
don't need the custom pass. And we can lower fall_equal/fany_nequal with
lower_vector_cmp instead.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Jonathan Marek [Thu, 1 Aug 2019 19:22:47 +0000 (15:22 -0400)]
freedreno: a2xx: fix HW binning for batches with >256K vertices
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Jonathan Marek [Thu, 1 Aug 2019 18:43:12 +0000 (14:43 -0400)]
freedreno: a2xx: fix fneg/fabs/fsat opcodes
Previously we would get a fmov with modifiers, but now that mov has no type
these opcodes need to be supported.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Jonathan Marek [Thu, 1 Aug 2019 18:38:18 +0000 (14:38 -0400)]
freedreno: a2xx: fix order of NIR opts
int_to_float needs to come after bool_to_float, and lower_to_source_mods
needs to come after both, since they don't deal wih source mods.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Jonathan Marek [Thu, 1 Aug 2019 18:36:41 +0000 (14:36 -0400)]
freedreno: a2xx: fix non-etc1 cubemaps
Not sure how this happened, but apparently all cubemaps need swapped XY.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Jonathan Marek [Thu, 1 Aug 2019 16:50:03 +0000 (12:50 -0400)]
freedreno: a2xx: fix fast clear not being used for Z24X8 buffers
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Jonathan Marek [Thu, 1 Aug 2019 16:42:33 +0000 (12:42 -0400)]
freedreno: align renderonly scanout buffers
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Eric Engestrom [Thu, 1 Aug 2019 22:32:34 +0000 (23:32 +0100)]
gitlab-ci: just build all the tools
This line was mistakenly added while there is already a `-D tools=all`
a few lines below.
Fixes: f60defa72d5d20d99e3a ("gitlab-ci: Add a shader-db run using v3d on drm-shim.")
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Sergii Romantsov [Fri, 12 Jul 2019 13:46:45 +0000 (16:46 +0300)]
i965/clear: clear_value better precision
Test-case with depth-clear 0.5 and format
MESA_FORMAT_Z24_UNORM_X8_UINT fails due inconsistent
clear-value of 0.
4999997.
Maybe its better to improve?
CC: Jason Ekstrand <jason.ekstrand@intel.com>
Fixes: 0ae9ce0f29ea (i965/clear: Quantize the depth clear value based on the format)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111113
Signed-off-by: Sergii Romantsov <sergii.romantsov@globallogic.com>
Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Samuel Pitoiset [Fri, 2 Aug 2019 11:55:01 +0000 (13:55 +0200)]
radv: fix image_has_{cmask,fmask}() helpers
The driver should now rely on cmask_offset because CMASK can be
disabled by the driver for some reasons (eg. mipmaps). Apply the
same change for FMASK, although it should be useless.
Fixes: ad1bc8621df ("radv: remove radv_get_image_fmask_info()")
Fixes: 10d08da52c6 ("radv/gfx10: add missing dcc_tile_swizzle tweak")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Samuel Pitoiset [Thu, 1 Aug 2019 15:59:56 +0000 (17:59 +0200)]
radv: remove radv_get_image_fmask_info()
It's unnecessary to duplicate fields in another struct.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Samuel Pitoiset [Thu, 1 Aug 2019 13:45:11 +0000 (15:45 +0200)]
radv/gfx10: add missing dcc_tile_swizzle tweak
Fixes: c90f46700dd ("radv/gfx10: mask DCC tile swizzle by alignment")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Samuel Pitoiset [Thu, 1 Aug 2019 15:59:55 +0000 (17:59 +0200)]
radv: remove radv_get_image_cmask_info()
It's unnecessary to duplicate fields in another struct.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Samuel Pitoiset [Thu, 1 Aug 2019 13:45:10 +0000 (15:45 +0200)]
radv: only account for tile_swizzle for color surfaces with DCC
It's 0 for depth surfaces with TC compat HTILE enabled.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Bas Nieuwenhuizen [Fri, 2 Aug 2019 00:16:23 +0000 (02:16 +0200)]
radv: Enable VK_KHR_shader_atomic_int64
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Bas Nieuwenhuizen [Fri, 2 Aug 2019 10:01:34 +0000 (12:01 +0200)]
ac/nir: Implement LLVM9 64-bit buffer compare & exchange.
LLVM 9 does not have a 64-bit buffer compswap intrinsic, so this
extracts the ptr, does a bound check and then uses a cmpxchg LLVM
instruction.
Not ideal, but the earliest release we're going to get a proper
intrinsic is LLVM 10.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Connor Abbott [Fri, 2 Aug 2019 09:14:50 +0000 (11:14 +0200)]
Revert "ac/nir: handle negate modifier"
This reverts commit
bfea7e4d2965269bff8f1f6449cb99c312fd7384.
Connor Abbott [Fri, 2 Aug 2019 09:14:08 +0000 (11:14 +0200)]
Revert "ac/nir: handle abs modifier"
This reverts commit
d3c80733cdfe8552b2f447ec8ed62465d0f2af1a.
These were only appearing due to memory corruption.
Timothy Arceri [Fri, 28 Jun 2019 04:55:20 +0000 (14:55 +1000)]
iris: bump compat profile support to 4.6
All of the current piglit compat profile tests pass.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Timothy Arceri [Fri, 2 Aug 2019 01:38:45 +0000 (11:38 +1000)]
egl: fix OpenGL 3.1 context creation
>From the EGL_KHR_create_context spec:
"* If OpenGL 3.1 is requested, the context returned may implement
any of the following versions:
* Version 3.1. The GL_ARB_compatibility extension may or may
not be implemented, as determined by the implementation.
* The core profile of version 3.2 or greater."
Fixes CTS tests:
dEQP-EGL.functional.create_context_ext.gl_31.rgb888_depth_stencil
dEQP-EGL.functional.create_context_ext.robust_gl_31.rgb888_depth_stencil
dEQP-EGL.functional.create_context_ext.gl_31.rgb888_depth_no_stencil
dEQP-EGL.functional.create_context_ext.robust_gl_31.rgb888_depth_no_stencil
dEQP-EGL.functional.create_context_ext.gl_31.rgba8888_depth_no_stencil
dEQP-EGL.functional.create_context_ext.gl_31.rgb888_no_depth_no_stencil
dEQP-EGL.functional.create_context_ext.robust_gl_31.rgba8888_depth_no_stencil
dEQP-EGL.functional.create_context_ext.robust_gl_31.rgb888_no_depth_no_stencil
dEQP-EGL.functional.create_context_ext.gl_31.rgba8888_no_depth_no_stencil
dEQP-EGL.functional.create_context_ext.robust_gl_31.rgba8888_no_depth_no_stencil
dEQP-EGL.functional.create_context_ext.gl_31.rgba8888_depth_stencil
dEQP-EGL.functional.create_context_ext.robust_gl_31.rgba8888_depth_stencil
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Connor Abbott [Wed, 31 Jul 2019 09:32:30 +0000 (11:32 +0200)]
nir/find_array_copies: Reject copies with mismatched type
When we detect a scalar/vector copy through load_deref/store_deref, we
have to be careful since those can bitcast an int to a float and
vice-versa even though copy_deref can't.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111251
Fixes: 156306e5e62 ("nir/find_array_copies: Handle wildcards and overlapping copies")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Samuel Pitoiset [Fri, 2 Aug 2019 07:56:32 +0000 (09:56 +0200)]
radv: re-apply "Optimize rebinding the same descriptor set."
This makes it cheaper to just change the dynamic offsets with
the same descriptor sets.
This optimization has been reverted a while back because of
random GPU hangs on GFX9, no it looks fine, at least CTS no longer
hangs on GFX9 and it doesn't hang on GFX10 as well.
It fixes a performance problem with Wolfenstein Youngblood.
Suggested-by: Philip Rebohle <philip.rebohle@tu-dortmund.de>
Samuel Pitoiset [Thu, 1 Aug 2019 08:43:44 +0000 (10:43 +0200)]
radv/gfx10: use the correct target machine for Wave32
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Samuel Pitoiset [Thu, 1 Aug 2019 08:43:42 +0000 (10:43 +0200)]
radv/gfx10: add Wave32 support for vertex, tessellation and geometry shaders
It can be enabled with RADV_PERFTEST=gewave32.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Samuel Pitoiset [Thu, 1 Aug 2019 08:43:41 +0000 (10:43 +0200)]
radv/gfx10: add Wave32 support for fragment shaders
It can be enabled with RADV_PERFTEST=pswave32.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Kenneth Graunke [Wed, 31 Jul 2019 22:47:34 +0000 (15:47 -0700)]
gallium: Implement GL_EXT_shader_samples_identical via a new capability
This exposes the textureSamplesIdenticalEXT function in GLSL.
We enable it for iris and radeonsi, because their compilers already
have support for this. Tested on Intel Kabylake and AMD Vega 64.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Kenneth Graunke [Fri, 2 Aug 2019 06:36:41 +0000 (23:36 -0700)]
intel/tools: Fix aubinator_viewer build.
This functions was recently renamed and not all callers were updated.
Fixes: 086c486a75f ("intel/device: rename gen_get_device_info")
Francisco Jerez [Tue, 23 Jul 2019 23:17:07 +0000 (16:17 -0700)]
intel/ir: Fix CFG corruption in opt_predicated_break().
Specifically the optimization of a conditional BREAK + WHILE sequence
into a conditional WHILE seems pretty broken. The list of successors
of "earlier_block" (where the conditional BREAK was found) is emptied
and then re-created with the same edges for no apparent reason. On
top of that the list of predecessors of the block immediately after
the WHILE loop is emptied, but only one of the original edges will be
added back, which means that potentially several blocks that still
have it on their list of successors won't be on its list of
predecessors anymore, causing all sorts of hilarity due to the
inconsistency in the control flow graph.
The solution is to remove the code that's removing valid edges from
the CFG. cfg_t::remove_block() will already clean up after itself.
The assert in bblock_t::combine_with() also needs to be removed since
we will be merging a block with multiple children into the first one
of them.
Found the issue on a hardware enabling branch originally, but
apparently somebody reproduced the same problem independently on
master in the meantime.
Fixes: d13bcdb3a9f ("i965/fs: Extend predicated break pass to predicate WHILE.")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111009
Cc: jiradet.jd@gmail.com
Cc: Sergii Romantsov <sergii.romantsov@globallogic.com>
Cc: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
Tested-by: Paul Chelombitko <qamonstergl@gmail.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Mark Janes [Tue, 30 Jul 2019 00:38:42 +0000 (17:38 -0700)]
intel/device: make internal functions private
The device info initializer makes several fuctions internal:
- handling of device override
- updating topology from kernel information
The implementation file is slightly reordered due to the renamed
functions being static.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Mark Janes [Thu, 25 Jul 2019 22:57:30 +0000 (15:57 -0700)]
intel/device: rename gen_get_device_info
Rename the original device info initialization routine so callers
don't mistakenly call the wrong one:
gen_get_device_info_from_fd:
Queries kernel for full device info, including topology
details.
gen_get_device_info_from_pci_id:
Partially initializes device info based on PCI ID lookup, when
the kernel is not available.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Mark Janes [Thu, 25 Jul 2019 21:31:40 +0000 (14:31 -0700)]
intel/tools: use device info initializer
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Mark Janes [Thu, 25 Jul 2019 17:40:55 +0000 (10:40 -0700)]
anv: use initialization routine for gen_device_info
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Mark Janes [Wed, 24 Jul 2019 22:21:36 +0000 (15:21 -0700)]
iris/screen: use initialization routine for gen_device_info
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Mark Janes [Wed, 24 Jul 2019 20:48:03 +0000 (13:48 -0700)]
i965: Move device info initialization to common code
With perf queries, initializing the device info is much more complex
than just getting a PCI ID and calling gen_get_device_info. This commit
adds a new gen_get_device_info_from_fd helper in common code which does
all of the requisite kernel queries to get device info including all of
the topology information.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Mark Janes [Wed, 31 Jul 2019 23:16:50 +0000 (16:16 -0700)]
i965/perf: verify kernel support before registering OA metrics
When gen_device_info updates the topology in it's initializer, the
kernel queries will fail silently. Iris and anv have minimum
kernel requirements that support the queries. i965 must verify kernel
support before reporting OA metrics.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Mark Janes [Thu, 25 Jul 2019 17:50:36 +0000 (10:50 -0700)]
intel/common: provide common ioctl routine
i965 links against libdrm for drmIoctl, but anv and iris both
re-implement this routine to avoid the dependency.
intel/dev also needs an ioctl wrapper, so lets share the same
implementation everywhere.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Alyssa Rosenzweig [Thu, 1 Aug 2019 15:10:03 +0000 (08:10 -0700)]
panfrost: Remove unused argument
A relic from when we didn't have an online compiler, hah.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:52:04 +0000 (15:52 -0700)]
panfrost: Handle MESA_SHADER_COMPUTE in compile callback
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:49:30 +0000 (15:49 -0700)]
pan/midgard: Use standard list traversal to find initial tag
Fixes a hang (and abort) on empty shaders, which you shouldn't have
anyway but better safe than sorry. DCE going on the fritz is no reason
to freeze the system.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:49:13 +0000 (15:49 -0700)]
panfrost: Use gl_shader_stage directly for compiles
No need to add a third set of enums to the mix.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:32:18 +0000 (15:32 -0700)]
panfrost: Emit "draw" info for compute jobs
Important fields relating to shader state and UBOs are filled out from
this (misnomer) function.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:31:23 +0000 (15:31 -0700)]
panfrost: Feed compute shaders into the compiler
The path for compute shader compiles resembles the graphic shader
compile path, although it is substantially simpler as we don't need any
shader keying.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:20:00 +0000 (15:20 -0700)]
panfrost: Expose compute shaders as panfrost_shader_variants
Whether variants are packed by graphics or compute is irrelevant.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:19:44 +0000 (15:19 -0700)]
panfrost: Remove shader state *base
It is now unused.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:19:09 +0000 (15:19 -0700)]
panfrost: Remove CSO dependency from shader_compile
We want this routine to be generic across graphics and compute, so let
the caller deal with the typing.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Wed, 31 Jul 2019 22:06:38 +0000 (15:06 -0700)]
panfrost: Generalize UBO upload for other shader stages
Now that everything is unified, this generalization is nice and easy.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>