mesa.git
7 years agoradv: fallback to an in-memory cache when no pipline cache is provided
Timothy Arceri [Wed, 15 Mar 2017 04:20:48 +0000 (15:20 +1100)]
radv: fallback to an in-memory cache when no pipline cache is provided

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
7 years agoradv: always create an fallback pipeline cache
Timothy Arceri [Wed, 15 Mar 2017 03:14:24 +0000 (14:14 +1100)]
radv: always create an fallback pipeline cache

This will be used as an in-memory cache when a pipeline cache is
not provided by the app.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
7 years agoradv: move cache check inside insert and search functions
Timothy Arceri [Wed, 15 Mar 2017 01:40:53 +0000 (12:40 +1100)]
radv: move cache check inside insert and search functions

This will allow us to use fallback in-memory and on-disk caches
should the app not provide a pipeline cache.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
7 years agost/mesa: call glthread_destroy() before _vbo_DestroyContext()
Timothy Arceri [Thu, 16 Mar 2017 12:09:19 +0000 (23:09 +1100)]
st/mesa: call glthread_destroy() before _vbo_DestroyContext()

Otherwise we have a race condition between vbo calls in the
glthread and the _vbo_DestroyContext() call.

This fixes a bunch of piglit crashes.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agoanv/GetQueryPoolResults: Actually implement the spec
Jason Ekstrand [Wed, 15 Mar 2017 03:48:47 +0000 (20:48 -0700)]
anv/GetQueryPoolResults: Actually implement the spec

The Vulkan spec is fairly clear about when we should and should not
write query pool results.  We're also supposed to return VK_NOT_READY if
VK_QUERY_RESULT_PARTIAL_BIT is not set and we come across any queries
which are not yet finished.  This fixes rendering corruptions on The
Talos Principle where geometry flickers in and out due to bogus query
results being returned by the driver.  These issues are most noticable
on Sky Lake GT4 2hen running on "ultra" settings.

Reviewed-By: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100182
Cc: "17.0 13.0" <mesa-stable@lists.freedesktop.org>
7 years agoanv/query: Invalidate the correct range
Jason Ekstrand [Wed, 15 Mar 2017 01:02:41 +0000 (18:02 -0700)]
anv/query: Invalidate the correct range

Reviewed-By: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: "17.0 13.0" <mesa-stable@lists.freedesktop.org>
7 years agoanv/query: Fix the location of timestamp availability
Jason Ekstrand [Wed, 15 Mar 2017 00:52:12 +0000 (17:52 -0700)]
anv/query: Fix the location of timestamp availability

Reviewed-By: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: "17.0 13.0" <mesa-dev@lists.freedesktop.org>
7 years agogenxml: Add XML version tags
Jason Ekstrand [Thu, 16 Mar 2017 05:33:00 +0000 (22:33 -0700)]
genxml: Add XML version tags

There's not much point to having them or not having them but this
reduces some pointless diff from the version we can auto-generate

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
7 years agoaubinator: Use fprintf for output.
Kenneth Graunke [Wed, 15 Mar 2017 08:22:00 +0000 (01:22 -0700)]
aubinator: Use fprintf for output.

This will make it easier to choose an output file.  For now, it remains
stdout.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
7 years agoaubinator: Reuse decode_structure code for handling commands
Kenneth Graunke [Wed, 15 Mar 2017 07:56:29 +0000 (00:56 -0700)]
aubinator: Reuse decode_structure code for handling commands

The code for decoding structures and commands was almost identical.
The only differences are: we print dword headers for commands, and
we skip the first one (with the command opcode and lengths).

So, generalize decode_structure to add a starting DWord, and a flag
for printing the DWord headers, and reuse it.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
7 years agoaubinator: Delete redundant NULL check.
Kenneth Graunke [Wed, 15 Mar 2017 07:40:44 +0000 (00:40 -0700)]
aubinator: Delete redundant NULL check.

handle_struct_decode() is just a wrapper around decode_structure()
with a NULL check.  But the only caller already does that NULL check.

So, just use decode_structure() directly.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
7 years agoaubinator: Fix indentation.
Kenneth Graunke [Wed, 15 Mar 2017 07:38:50 +0000 (00:38 -0700)]
aubinator: Fix indentation.

Three space, not four.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
7 years agoi965/gen8+: Do full stall when switching pipeline
Topi Pohjolainen [Wed, 15 Mar 2017 19:31:07 +0000 (21:31 +0200)]
i965/gen8+: Do full stall when switching pipeline

just as earlier gens do.

CC: "17.0 13.0" <mesa-stable@lists.freedesktop.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96743
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoi965: remove uneeded asm/unistd.h include
Jonathan Gray [Thu, 16 Mar 2017 02:48:20 +0000 (13:48 +1100)]
i965: remove uneeded asm/unistd.h include

Fix the build on OpenBSD by removing an uneeded include for asm/unistd.h.

Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
7 years agoi965: automake: remove spurious white space
Emil Velikov [Thu, 16 Mar 2017 13:55:42 +0000 (13:55 +0000)]
i965: automake: remove spurious white space

Unintentionally introduced by yours truly with the i965 compiler move.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
7 years agoi965: avoid using a GNU make pattern rule
Jonathan Gray [Thu, 16 Mar 2017 02:49:36 +0000 (13:49 +1100)]
i965: avoid using a GNU make pattern rule

% pattern rules are a GNU extension.  As there is only one file here
avoid patterns and globbing entirely to fix the build on non-GNU make.

Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
v2 [Emil Velikov: brw_oa.py dependency]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
7 years agodocs/releasing: document how to squash/announce queued patches
Emil Velikov [Wed, 15 Mar 2017 17:34:26 +0000 (17:34 +0000)]
docs/releasing: document how to squash/announce queued patches

In the odd case where a patch needs to be fixed, squash the appropriate
fix and document how. Add a note in the pre-release notes, such that
devs can quickly spot it.

v2: Grammar/typo fixes (Eric). Use upstream commit [SHA] as reference.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
7 years agodocs/releasing: release.sh is located in xorg/util-modular
Emil Velikov [Wed, 15 Mar 2017 17:34:25 +0000 (17:34 +0000)]
docs/releasing: release.sh is located in xorg/util-modular

Correct the silly typo s/macros/modular/ and add a reference to the
repository.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
7 years agodocs/releasing: remove "git clean" step
Emil Velikov [Wed, 15 Mar 2017 17:34:24 +0000 (17:34 +0000)]
docs/releasing: remove "git clean" step

release.sh from master, does not require the tree to be clean.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
7 years agomapi: remove Xlib/xcb include in gl_marshal.py
Emil Velikov [Thu, 16 Mar 2017 11:05:23 +0000 (11:05 +0000)]
mapi: remove Xlib/xcb include in gl_marshal.py

The only use of the header is to provide the _X_INLINE macro. We already
require (and provide where needed) 'inline', plus it's used in the file
already.

So replace the macro and drop the include. This fixes the build on
platforms which lack the header - from X-less Linuxes to Androids.

Fixes: 05dd4a1104e ("glapi: Generate GL API marshalling code from the XML.")
Reported-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100223
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
7 years agodocs/specs: update Khronos registries URLs
Eric Engestrom [Thu, 16 Mar 2017 09:59:55 +0000 (09:59 +0000)]
docs/specs: update Khronos registries URLs

The registries were migrated to git and are now hosted on GitHub.
The old svn is now read-only, and will not be updated anymore.

Signed-off-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Plamena Manolova <plamena.manolova@intel.com>
7 years agoanv: improve error reporting when creating pipelines
Iago Toral Quiroga [Thu, 9 Mar 2017 13:37:26 +0000 (14:37 +0100)]
anv: improve error reporting when creating pipelines

Specifically, report 'out of memory' errors that might have happened while
emitting the pipeline's batch.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: handle errors in emit_binding_table() and emit_samplers()
Iago Toral Quiroga [Thu, 9 Mar 2017 10:34:22 +0000 (11:34 +0100)]
anv: handle errors in emit_binding_table() and emit_samplers()

These can fail to allocate device memory, however, the driver can recover
from this error by allocating a new binding table block and trying again.

v2:
  - Instead of tracking the errors in these functions and making callers
    reset the batch's status before attempting to allocate a new block
    for the binding table, simply make callers responsible for setting
    the error status if they fail to allocate memory during the second
    attempt (Jason).

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: handle errors while allocating new binding table blocks
Iago Toral Quiroga [Mon, 6 Mar 2017 11:28:44 +0000 (12:28 +0100)]
anv: handle errors while allocating new binding table blocks

Also, we had a couple of instances in flush_descriptor_sets() were
we were returning a VkResult directly upon error, but the return
value of this function is not a VkResult but a uint32_t dirty mask,
so simply return 0 in these cases which reduces the amount of
work the driver will do after the error has been raised.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/blorp: make anv_cmd_buffer_alloc_blorp_binding_table() return a VkResult
Iago Toral Quiroga [Thu, 9 Mar 2017 10:49:01 +0000 (11:49 +0100)]
anv/blorp: make anv_cmd_buffer_alloc_blorp_binding_table() return a VkResult

Instead of asserting inside the function, and then use use that information
to return early from its callers upon failure.

v2:
  - Make sure that clear_color_attachment() and
    clear_depth_stencil_attachment() get the VkResult as well so they
    avoid executing the batch if an error happened. (Topi)

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/device: assert that commands submitted to a queue are not bogus
Iago Toral Quiroga [Thu, 9 Mar 2017 09:44:13 +0000 (10:44 +0100)]
anv/device: assert that commands submitted to a queue are not bogus

Any errors that may have happened during the command buffer recording are
reported by vkEndCommandBuffer() and it is the application's reponsibility
to not submit broken commands to a queue.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: skip vkCmdExecuteCommands() on broken command buffers
Iago Toral Quiroga [Mon, 6 Mar 2017 11:16:05 +0000 (12:16 +0100)]
anv/cmd_buffer: skip vkCmdExecuteCommands() on broken command buffers

v2: Assert on secondary commands, applications should've called
    vkEndCommandBuffer() and received an error for them before (Jason)

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: skip vkCmdDispatch() on broken command buffers
Iago Toral Quiroga [Mon, 6 Mar 2017 11:15:06 +0000 (12:15 +0100)]
anv/cmd_buffer: skip vkCmdDispatch() on broken command buffers

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: skip vkCmdDraw*() on broken command buffers
Iago Toral Quiroga [Mon, 6 Mar 2017 10:12:31 +0000 (11:12 +0100)]
anv/cmd_buffer: skip vkCmdDraw*() on broken command buffers

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: handle memory allocation errors during queue submissions
Iago Toral Quiroga [Mon, 6 Mar 2017 11:04:19 +0000 (12:04 +0100)]
anv: handle memory allocation errors during queue submissions

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: handle out of memory during vkCmdPushConstants
Iago Toral Quiroga [Fri, 3 Mar 2017 10:31:14 +0000 (11:31 +0100)]
anv/cmd_buffer: handle out of memory during vkCmdPushConstants

Fixes:
dEQP-VK.api.out_of_host_memory.cmd_push_constants

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: handle allocation errors during vkCmdBeginRenderPass()
Iago Toral Quiroga [Wed, 8 Mar 2017 12:19:57 +0000 (13:19 +0100)]
anv/cmd_buffer: handle allocation errors during vkCmdBeginRenderPass()

Fixes:
dEQP-VK.api.out_of_host_memory.cmd_begin_render_pass

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: skip vkCmdEndRenderPass() for broken command buffers
Iago Toral Quiroga [Wed, 8 Mar 2017 12:18:52 +0000 (13:18 +0100)]
anv/cmd_buffer: skip vkCmdEndRenderPass() for broken command buffers

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: skip vkCmdNextSubpass() for broken command buffers
Iago Toral Quiroga [Wed, 8 Mar 2017 12:18:12 +0000 (13:18 +0100)]
anv/cmd_buffer: skip vkCmdNextSubpass() for broken command buffers

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: report tracked errors in vkEndCommandBuffer()
Iago Toral Quiroga [Wed, 8 Mar 2017 12:16:16 +0000 (13:16 +0100)]
anv/cmd_buffer: report tracked errors in vkEndCommandBuffer()

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: handle failures when growing reloc lists
Iago Toral Quiroga [Fri, 3 Mar 2017 09:55:19 +0000 (10:55 +0100)]
anv: handle failures when growing reloc lists

Growing the reloc list happens through calling anv_reloc_list_add() or
anv_reloc_list_append(). Make sure that we call these through helpers
that check the result and set the batch error status if needed.

v2:
  - Handling the crashes is not good enough, we need to keep track of
    the error, for that, keep track of the errors in the batch instead (Jason).
  - Make reloc list growth go through helpers so we can have a central
    place where we can do error tracking (Jason).

v3:
  - Callers that need the offset returned by anv_reloc_list_add() can
    compute it themselves since it is extracted from the inputs to the
    function, so change the function to return a VkResult, make
    anv_batch_emit_reloc() also return a VkResult and let their callers
    do the error management (Topi)

v4:
  - Let anv_batch_emit_reloc() return an uint64_t as it originally did,
    there is no real benefit in having it return a VkResult.
  - Do not add an is_aux parameter to add_surface_state_reloc(), instead
    do error checking for aux in add_image_view_relocs() separately.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: avoid crashes when failing to allocate batches
Iago Toral Quiroga [Thu, 9 Mar 2017 13:22:25 +0000 (14:22 +0100)]
anv: avoid crashes when failing to allocate batches

Most of the time we use macros that handle this situation transparently,
but there are some cases were we need to handle this explicitly.

This patch makes sure we don't crash, notice that error handling takes
place in the function that actually failed the allocation,
anv_batch_emit_dwords(), which will set the status field of the batch
so it can be used at a later moment to report the error to the user.

v2:
  - Not crashing is not good enough, we need to keep track of the error
    (Topi, Jason). Iago: now that we track errors in the batch, this
    is being handled.
  - Added guards in a few more places that needed it (Iago)

v3:
  - Check result of anv_batch_emitn() for NULL before calling memset()
    in emit_vertex_input() (Topi)

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: handle allocation failure in anv_batch_emit_dwords()
Iago Toral Quiroga [Fri, 3 Mar 2017 09:56:33 +0000 (10:56 +0100)]
anv: handle allocation failure in anv_batch_emit_dwords()

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: handle allocation failure in anv_batch_emit_batch()
Iago Toral Quiroga [Fri, 3 Mar 2017 10:17:03 +0000 (11:17 +0100)]
anv: handle allocation failure in anv_batch_emit_batch()

v2:
 - Call the error handler (Topi)

Fixes:
dEQP-VK.api.out_of_host_memory.cmd_execute_commands

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: add anv_batch_set_error() and anv_batch_has_error() helpers
Iago Toral Quiroga [Wed, 8 Mar 2017 11:59:58 +0000 (12:59 +0100)]
anv: add anv_batch_set_error() and anv_batch_has_error() helpers

The anv_batch_set_error() helper will track the first error that happened
while recording a command buffer. The helper returns the currently tracked
error to help the job of internal functions that may generate errors that
need to be tracked and return a VkResult to the caller.

We will use the anv_batch_has_error() helper to guard parts of the driver
that are not safe to execute if an error has been generated while recording
a particular command buffer.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: add a status field to anv_batch
Iago Toral Quiroga [Wed, 8 Mar 2017 11:45:37 +0000 (12:45 +0100)]
anv/cmd_buffer: add a status field to anv_batch

The vkCmd*() functions do not report errors, instead, any errors should be
reported by the time we call vkEndCommandBuffer(). This means that we
need to make the driver robust against incosistent and/or imcomplete
command  buffer states through the command recording process, particularly,
avoid crashes due to access to memory that we failed to allocate previously.

The strategy used to do this is to track the first error ocurred while
recording a command buffer in the batch associated with it. We use the
batch to track this information because the command buffer may not be
visible to all parts of the driver that can produce errors we need to be
aware of (such as allocation failures during batch emissions).

Later patches will use this error information to guard parts of the driver
that may not be safe to execute.

v2: Move the field from the command buffer to the batch so we can track
    errors from batch emissions (Jason)

v3: Registering errors in the command buffer's batch during
    anv_create_cmd_buffer() is unnecessary, since the command buffer
    is freed at the end of the function in that case (Topi)

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/cmd_buffer: report errors in vkBeginCommandBuffer()
Iago Toral Quiroga [Mon, 6 Mar 2017 12:27:11 +0000 (13:27 +0100)]
anv/cmd_buffer: report errors in vkBeginCommandBuffer()

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: do not try to ref/unref NULL shaders
Iago Toral Quiroga [Fri, 3 Mar 2017 09:57:17 +0000 (10:57 +0100)]
anv: do not try to ref/unref NULL shaders

This situation can happen if we failed to allocate memory for the shader.

v2:
 - We shouldn't see NULL shaders in anv_shader_bin_ref so we should not check
   for that (Jason). Make sure that callers don't attempt to call this
   function with a NULL shader and assert that this never happens (Iago).

v3:
 - All callers to anv_shader_bin_unref seem to check for NULL before calling,
   so just assert that it is not NULL (Topi)

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv/blorp: return early if we failed to create the shader binary
Iago Toral Quiroga [Fri, 3 Mar 2017 09:58:23 +0000 (10:58 +0100)]
anv/blorp: return early if we failed to create the shader binary

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agointel/blorp: make upload_shader() return a bool indicating success or failure
Iago Toral Quiroga [Tue, 14 Mar 2017 12:12:22 +0000 (13:12 +0100)]
intel/blorp: make upload_shader() return a bool indicating success or failure

For now we always return true, follow-up patches will handle fail scenarios.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agoanv: remove unnecessary function prototype.
Iago Toral Quiroga [Thu, 9 Mar 2017 08:07:39 +0000 (09:07 +0100)]
anv: remove unnecessary function prototype.

The function is defined right after the prototype declaration. Also, the
protoype for it is included in anv_genX.h which is included via anv_private.h.

Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
7 years agomapi: don't include X11/Xlib-xcb.h on non PTHREAD platforms
Timothy Arceri [Thu, 16 Mar 2017 04:43:48 +0000 (15:43 +1100)]
mapi: don't include X11/Xlib-xcb.h on non PTHREAD platforms

Should fix the last of the glthread build issues on windows.

7 years agomesa: fix glthread marshal build issues on platforms without PTHREAD
Timothy Arceri [Thu, 16 Mar 2017 04:28:47 +0000 (15:28 +1100)]
mesa: fix glthread marshal build issues on platforms without PTHREAD

7 years agomesa: fix glthread build issues on platforms without PTHREAD
Timothy Arceri [Thu, 16 Mar 2017 03:46:16 +0000 (14:46 +1100)]
mesa: fix glthread build issues on platforms without PTHREAD

7 years agogallium: implement the backend of threaded GL dispatch
Marek Olšák [Sun, 5 Feb 2017 00:20:51 +0000 (01:20 +0100)]
gallium: implement the backend of threaded GL dispatch

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa/glthread: restore the dispatch table when incompatible gl calls are detected
Gregory Hainaut [Sun, 12 Feb 2017 14:21:47 +0000 (15:21 +0100)]
mesa/glthread: restore the dispatch table when incompatible gl calls are detected

While a context only has a single glthread, the context itself can be
attached to several threads. Therefore the dispatch table must be
updated in all threads before the destruction of glthread. In others
words, glthread can only be destroyed safely when the context is deleted.

Fixes remaining crashes in the glx-multithread-makecurrent* tests.

V2: (Timothy Arceri) updated gl_API.dtd marshal_fail description.

Signed-off-by: Gregory Hainaut <gregory.hainaut@gmail.com>
Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa/glthread: don't set a dispatch table if we aren't the owner
Gregory Hainaut [Mon, 13 Feb 2017 18:14:28 +0000 (19:14 +0100)]
mesa/glthread: don't set a dispatch table if we aren't the owner

Fix crashes when glxMakeCurrent is called.

Signed-off-by: Gregory Hainaut <gregory.hainaut@gmail.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Track the current vertex/element array buffers for glthread.
Eric Anholt [Mon, 4 Mar 2013 18:30:15 +0000 (10:30 -0800)]
mesa: Track the current vertex/element array buffers for glthread.

We want to support glthread on GLES contexts with reasonable apps, and on
desktop for apps that use VBOs but haven't completely moved to core GL.
To do so, we have to deal with the "the user may or may not pass user
pointers to draw calls" problem.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Disable glthread when glBegin() is called.
Eric Anholt [Mon, 4 Mar 2013 17:47:40 +0000 (09:47 -0800)]
mesa: Disable glthread when glBegin() is called.

glBegin() swaps dispatch tables, and we don't have any code in place for
handling that in glthread (which also messes with dispatch tables), and I
don't particularly care to at this point.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Add an attribute for conditions to turn off threading.
Eric Anholt [Fri, 1 Mar 2013 02:15:58 +0000 (18:15 -0800)]
mesa: Add an attribute for conditions to turn off threading.

The threading for GL core is in place, but there are so few applications
actually using a core GL context that it would be nice to extend support
back.  However, some of the features of compat GL (particularly user
vertex arrays) would be so expensive to track state for that we want to be
able to disable threading when we discover that the app is using them.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Add support for asynchronous glDraw* on GL core.
Eric Anholt [Wed, 27 Feb 2013 22:28:16 +0000 (14:28 -0800)]
mesa: Add support for asynchronous glDraw* on GL core.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Add support for NULL arguments like in glBufferData() in marshalling.
Eric Anholt [Wed, 27 Feb 2013 21:37:14 +0000 (13:37 -0800)]
mesa: Add support for NULL arguments like in glBufferData() in marshalling.

This will let us support things like glBufferData() that should be
asynchronous.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Statically allocate glthread command buffer in the batch struct.
Eric Anholt [Sat, 16 Feb 2013 00:33:33 +0000 (16:33 -0800)]
mesa: Statically allocate glthread command buffer in the batch struct.

This avoids an extra pointer dereference in the marshalling functions,
which, with the instruction count doing in the low 30s, could actually
matter for main-thread performance.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoglapi: Mark vertex attrib pointer functions as async.
Eric Anholt [Thu, 27 Dec 2012 20:11:57 +0000 (12:11 -0800)]
glapi: Mark vertex attrib pointer functions as async.

These don't actually read data out of the pointers, they set the
pointers (or offsets in a VBO) to be used in a later draw call.

v2: Don't forget glVertexAttribIPointer, and don't bother with annotations
    on aliases.
v3: Mark CompressedTexSubImage1D as sync also.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Custom thread marshalling for Flush.
Paul Berry [Fri, 16 Nov 2012 19:43:08 +0000 (11:43 -0800)]
mesa: Custom thread marshalling for Flush.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Custom thread marshalling for ShaderSource.
Paul Berry [Wed, 14 Nov 2012 18:40:22 +0000 (10:40 -0800)]
mesa: Custom thread marshalling for ShaderSource.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Connect the generated GL command marshalling code to the build.
Eric Anholt [Wed, 2 Jan 2013 22:12:04 +0000 (14:12 -0800)]
mesa: Connect the generated GL command marshalling code to the build.

v2: Rebase on the Begin/End changes, and just disable this feature on
    non-GL-core.
v3: (Timothy Arceri) enable for non-GL-core contexts. Remove
    unrelated safe_mul() hunk. while loop style fix.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoRevert "mesa: make _mesa_alloc_dispatch_table() static"
Marek Olšák [Wed, 28 Sep 2016 23:00:39 +0000 (01:00 +0200)]
Revert "mesa: make _mesa_alloc_dispatch_table() static"

This reverts commit 4009d22b61e76850b1b725f4e491da05c2406fa4.

glthread needs it.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Create pointers for multithread marshalling dispatch table.
Paul Berry [Wed, 3 Oct 2012 22:39:52 +0000 (15:39 -0700)]
mesa: Create pointers for multithread marshalling dispatch table.

This patch splits the context's CurrentDispatch pointer into two
pointers, CurrentClientDispatch, and CurrentServerDispatch, so that
when doing multithread marshalling, we can distinguish between the
dispatch table that's being used by the client (to serialize GL calls
into the marshal buffer) and the dispatch table that's being used by
the server (to execute the GL calls).

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Add infrastructure for a worker thread to process GL commands.
Eric Anholt [Fri, 28 Dec 2012 23:05:34 +0000 (15:05 -0800)]
mesa: Add infrastructure for a worker thread to process GL commands.

v2: Keep an allocated buffer around instead of checking for one at the
    start of every GL command.  Inline the now-small space allocation
    function.
v3: Remove duplicate !glthread->shutdown check, process remaining work
    before shutdown.
v4: Fix leaks on destroy.
V5: (Timothy Arceri) fix order of source files in makefile

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Validate count parameters when marshalling.
Eric Anholt [Thu, 3 Jan 2013 19:56:54 +0000 (11:56 -0800)]
mesa: Validate count parameters when marshalling.

Otherwise, for example, glDeleteBuffers(-1, &bo) gets you a segfault
instead of GL_INVALID_VALUE.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoglapi: Generate GL API marshalling code from the XML.
Paul Berry [Wed, 7 Nov 2012 21:03:13 +0000 (13:03 -0800)]
glapi: Generate GL API marshalling code from the XML.

This is not yet used in the build, just generated.

v2: Add missing build dependencies.
v3: Avoid mixing declarations and code, remove logic for avoiding emitting
    code that the compiler's optimizer can deal with anyway.
v4: (Timothy Arceri) move safe_mul() genereation here from a later patch.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoglapi: Mark compressed teximage functions as sync.
Eric Anholt [Tue, 5 Mar 2013 19:51:18 +0000 (11:51 -0800)]
glapi: Mark compressed teximage functions as sync.

Without doing some additional tracking, we won't know whether the data
will be immediate user data, or will be loaded from a PBO.  The normal
teximage functions will be sync by default because they don't know up
front what the size of their image data is.  But for compressed teximage,
we have the count information, so they would end up async by default.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoglapi: Annotate functions with "marshal" attribute.
Paul Berry [Wed, 7 Nov 2012 21:04:40 +0000 (13:04 -0800)]
glapi: Annotate functions with "marshal" attribute.

Several API functions require special treatment in order to be marshalled
to a background thread.  Others can't be safely executed in a background
thread and need to be executed synchronously (e.g. since they return data
through a pointer argument).

This annotation will be used when code generating thread marshalling code,
to ensure that each function is marshalled in the correct way.

Note that PixelMap functions are marked as synchronous for now since
their pointer may be relative to buffer on the GPU, so we'll need
special logic to marshal them properly.

v2: Move description of attribute types to a comment in the dtd file.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoegl: Implement __DRI_BACKGROUND_CALLABLE
Eric Anholt [Fri, 28 Dec 2012 01:39:37 +0000 (17:39 -0800)]
egl: Implement __DRI_BACKGROUND_CALLABLE

v2: (Timothy Arceri) use C99 initializers.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoglx: Implement __DRI_BACKGROUND_CALLABLE
Paul Berry [Wed, 14 Nov 2012 23:50:27 +0000 (15:50 -0800)]
glx: Implement __DRI_BACKGROUND_CALLABLE

v2: Marek: Add DRI3 support.

v3: (Timothy Arceri) use C99 initializers.

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agomesa: Add SetBackgroundContext to dd_function_table
Paul Berry [Wed, 14 Nov 2012 22:26:22 +0000 (14:26 -0800)]
mesa: Add SetBackgroundContext to dd_function_table

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agodri: Update dri_util to keep track of __DRI_BACKGROUND_CALLABLE
Paul Berry [Wed, 14 Nov 2012 22:39:21 +0000 (14:39 -0800)]
dri: Update dri_util to keep track of __DRI_BACKGROUND_CALLABLE

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agodri_interface: Add new marshalling interfaces to dri_interface.h
Paul Berry [Wed, 14 Nov 2012 19:13:02 +0000 (11:13 -0800)]
dri_interface: Add new marshalling interfaces to dri_interface.h

Acked-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agogallivm: (trivial) remove duplicated line
Roland Scheidegger [Thu, 16 Mar 2017 03:01:41 +0000 (04:01 +0100)]
gallivm: (trivial) remove duplicated line

pointed out by clang (stored value never read)

7 years agodraw: (trivial) remove a unnecessary lp_build_alloca()
Roland Scheidegger [Thu, 16 Mar 2017 02:59:52 +0000 (03:59 +0100)]
draw: (trivial) remove a unnecessary lp_build_alloca()

pointed out by clang (stored value never read)

7 years agoswr: support layer output in geometry shaders
Ilia Mirkin [Sun, 5 Mar 2017 23:24:44 +0000 (18:24 -0500)]
swr: support layer output in geometry shaders

This makes bin/gl-3.2-layered-rendering-gl-layer-render fail only with
2DMS_ARRAY, which is expected given the lackluster MSAA support. However
all the regular types pass.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
7 years agoRevert "radv: Emit cache flushes before CP DMA."
Bas Nieuwenhuizen [Wed, 15 Mar 2017 17:49:29 +0000 (18:49 +0100)]
Revert "radv: Emit cache flushes before CP DMA."

This reverts commit cce43f6d8c40222099badaf52344d6a0eed993f3.

Redundant, as the flush already happens at si_cp_dma_prepare.

Acked-by: Dave Airlie <airlied@redhat.com>
7 years agogallium/tgsi: Treat UCMP sources as floats to match the GLSL-to-TGSI pass expectations.
Francisco Jerez [Tue, 14 Mar 2017 00:31:39 +0000 (17:31 -0700)]
gallium/tgsi: Treat UCMP sources as floats to match the GLSL-to-TGSI pass expectations.

Currently the GLSL-to-TGSI translation pass assumes it can use
floating point source modifiers on the UCMP instruction.  See the bug
report linked below for an example where an unrelated change in the
GLSL built-in lowering code for atan2 (e9ffd12827ac11a2d2002a42fa8eb1)
caused the generation of floating-point ir_unop_neg instructions
followed by ir_triop_csel, which is translated into UCMP with a negate
modifier on back-ends with native integer support.

Allowing floating-point source modifiers on an integer instruction
seems like rather dubious design for a transport IR, since the same
semantics could be represented as a sequence of MOV+UCMP instructions
instead, but supposedly this matches the expectations of TGSI
back-ends other than tgsi_exec, and the expectations of the DX10 API.
I take no responsibility for future headaches caused by this
inconsistency.

Fixes a regression of piglit glsl-fs-tan-1 on softpipe introduced by
the above-mentioned glsl front-end commit.  Even though the commit
that triggered the regression doesn't seem to have made it to any
stable branches yet, this might be worth back-porting since I don't
see any reason why the bug couldn't have been reproduced before that
point.

Suggested-by: Roland Scheidegger <sroland@vmware.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99817
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
7 years agoutil/disk_cache: do eviction before creating .tmp
Grazvydas Ignotas [Wed, 15 Mar 2017 18:53:56 +0000 (20:53 +0200)]
util/disk_cache: do eviction before creating .tmp

cache_put() first creates a .tmp file and then tries to do eviction.
The recently added LRU eviction code selects non-empty directory with
the oldest access time, but that may easily be the one with just the
new .tmp file, especially on Linux where atime is updated lazily
(with "relatime" mount option, which is the default). So when cache is
small, if random doesn't hit another dir LRU keeps selecting the same
dir with just the .tmp and not deleting anything. To fix this (and the
tests), do eviction earlier.

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
7 years agoswr: validate backend state numAttributes
Tim Rowley [Wed, 15 Mar 2017 16:42:43 +0000 (11:42 -0500)]
swr: validate backend state numAttributes

General protection and prevents us from smashing the stack
on the first clear state validation (a7b8d50bcb).  Fixes crash
using icc.

Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
7 years agogbm: Export a get modifiers
Ben Widawsky [Fri, 21 Oct 2016 01:21:24 +0000 (18:21 -0700)]
gbm: Export a get modifiers

This patch originally had i965 specific code and was named:
commit 61cd3c52b868cf8cb90b06e53a382a921eb42754
Author: Ben Widawsky <ben@bwidawsk.net>
Date:   Thu Oct 20 18:21:24 2016 -0700

    gbm: Get modifiers from DRI

To accomplish this, two new query tokens are added to the extension:
__DRI_IMAGE_ATTRIB_MODIFIER_UPPER
__DRI_IMAGE_ATTRIB_MODIFIER_LOWER

The query extension only supported 32b queries, and modifiers are 64b,
so we needed two of them.

NOTE: The extension version is still set to 13, so none of this will
actually be called.

v2: Error handling of queryImage (Emil)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
7 years agoi965: introduce modifier selection.
Ben Widawsky [Tue, 14 Mar 2017 01:20:02 +0000 (18:20 -0700)]
i965: introduce modifier selection.

Nothing special here other than a brief introduction to modifier
selection. Originally this was part of another patch but was split out
from
gbm: Introduce modifiers into surface/bo creation by request of Emil.

Requested-by: Emil Velikov <emil.velikov@collabora.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agoegl/drm: Use modifiers for backbuffer creation
Ben Widawsky [Tue, 14 Mar 2017 01:19:00 +0000 (18:19 -0700)]
egl/drm: Use modifiers for backbuffer creation

Split into a separate patch from the previous patch as requested by
Emil.

Requested-by: Emil Velikov <emil.velikov@collabora.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agogbm: Introduce modifiers into surface/bo creation
Ben Widawsky [Thu, 3 Nov 2016 23:14:44 +0000 (16:14 -0700)]
gbm: Introduce modifiers into surface/bo creation

The idea behind modifiers like this is that the user of GBM will have
some mechanism to query what properties the hardware supports for its BO
or surface. This information is directly passed in (and stored) so that
the DRI implementation can create an image with the appropriate
attributes.

A getter() will be added later so that the user GBM will be able to
query what modifier should be used.

Only in surface creation, the modifiers are stored until the BO is
actually allocated. In regular buffer allocation, the correct modifier
can (will be, in future patches be chosen at creation time.

v2: Make sure to check if count is non-zero in addition to testing if
calloc fails. (Daniel)

v3: Remove "usage" and "flags" from modifier creation. Requested by
Kristian.

v4: Take advantage of the "INVALID" modifier added by the GET_PLANE2
series.

v5: Don't bother with storing modifiers for gbm_bo_create because that's
a synchronous operation and we can actually select the correct modifier
at create time (done in a later patch) (Jason)

v6: Make modifier condition outside the check so that dri_use will work
properly (Jason)

Cc: Kristian Høgsberg <krh@bitplanet.net>
References (v4): https://lists.freedesktop.org/archives/intel-gfx/2017-January/116636.html
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com> (v1)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Acked-by: Daniel Stone <daniels@collabora.com>
7 years agoi965: Implement basic modifier image creation
Ben Widawsky [Mon, 13 Mar 2017 21:53:43 +0000 (14:53 -0700)]
i965: Implement basic modifier image creation

This is just a stub for now and will be filled in later.

This was split out of an earlier patch

Requested-by: Emil Velikov <emil.velikov@collabora.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
7 years agodri: Add an image creation with modifiers
Ben Widawsky [Fri, 4 Nov 2016 18:31:15 +0000 (11:31 -0700)]
dri: Add an image creation with modifiers

Modifiers will be obtained or guessed by the client and passed in during
image creation/import. In guessing, a client might decide to simply pass
along all known modifiers

This requires bumping the DRIimage version.

As of this patch, the modifiers aren't plumbed all the way down, this
patch simply makes sure the interface level stuff is correct.

v2: Don't allow usage + modifiers

v3: Make NAND actually NAND. Bug introduced in v2. (Jason)

v4:
- s/obtains/obtained (Jason)
- Pull out i965 imlemnentation into a later patch (Emil)

Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com> (v1)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Acked-by: Daniel Stone <daniels@collabora.com>
7 years agoradeonsi: implement TGSI opcodes TEX_LZ and TXF_LZ
Marek Olšák [Tue, 7 Mar 2017 01:19:47 +0000 (02:19 +0100)]
radeonsi: implement TGSI opcodes TEX_LZ and TXF_LZ

This massively decreases VGPR spilling for DiRT Showdown, because we
no longer have to use v4i32 for 2D fetches when level == 0.
We now use v2i32 for those cases.

DiRT Showdown - Spilled VGPRs: -26 (-81%)

This surprisingly doesn't have any useful effect on performance (+ 0.05%).

7 years agoglsl_to_tgsi: use TEX_LZ and TXF_LZ when available
Marek Olšák [Tue, 7 Mar 2017 01:26:47 +0000 (02:26 +0100)]
glsl_to_tgsi: use TEX_LZ and TXF_LZ when available

7 years agoglsl_to_tgsi: remove a redundant statement
Marek Olšák [Tue, 7 Mar 2017 01:01:08 +0000 (02:01 +0100)]
glsl_to_tgsi: remove a redundant statement

it's the same as the last "else".

7 years agogallium: add TGSI opcodes TEX_LZ and TXF_LZ
Marek Olšák [Tue, 7 Mar 2017 01:15:14 +0000 (02:15 +0100)]
gallium: add TGSI opcodes TEX_LZ and TXF_LZ

for better code generation in radeonsi

7 years agogallium: add PIPE_CAP_TGSI_TEX_TXF_LZ
Marek Olšák [Tue, 7 Mar 2017 01:09:03 +0000 (02:09 +0100)]
gallium: add PIPE_CAP_TGSI_TEX_TXF_LZ

7 years agoradeonsi: disable sinking common instructions down to the end block
Samuel Pitoiset [Tue, 14 Mar 2017 23:59:13 +0000 (00:59 +0100)]
radeonsi: disable sinking common instructions down to the end block

Initially this was a workaround for a bug introduced in LLVM 4.0
in the SimplifyCFG pass that caused image instrinsics to disappear
(because they were badly sunk). Finally, this is a win because it
decreases SGPR spilling and increases the number of waves a bit.

Although, shader-db results are good I think we might want to
remove it in the future once the issue is fixed. For now, enable
it for LLVM >= 4.0.

This also fixes a rendering issue with the speedometer in Dirt Rally.

More information can be found here https://reviews.llvm.org/D26348.

Thanks to Dave Airlie for the patch.

v2: - add a FIXME comment
    - use if (HAVE_LLVM >= 0x0400) instead

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99484
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97988
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Cc: 17.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agotgsi: add missing compute shader entry in tgsi_get_processor_name()
Samuel Pitoiset [Wed, 15 Mar 2017 11:40:13 +0000 (12:40 +0100)]
tgsi: add missing compute shader entry in tgsi_get_processor_name()

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agoradeonsi: clean up tex_fetch_ptrs()
Samuel Pitoiset [Wed, 15 Mar 2017 12:00:02 +0000 (13:00 +0100)]
radeonsi: clean up tex_fetch_ptrs()

Will also help when the src sampler register will be
TGSI_FILE_CONSTANT for bindless.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
7 years agoconfigure.ac: bump pthread-stubs requirement
Emil Velikov [Thu, 2 Mar 2017 19:02:45 +0000 (19:02 +0000)]
configure.ac: bump pthread-stubs requirement

On platforms that require it, we bump the requirement to 0.4 or later.
Due to an issue with the project [design] any version earlier than it,
is bound to cause issues. For the specifics see the pthread-stubs README

Cc: Uli Schlachter <psychon@znc.in>
Cc: Jonathan Gray <jsg@jsg.id.au>
Cc: Jean-Sébastien Pédron <dumbbell@FreeBSD.org>
Cc: François Tigeot <ftigeot@wolfpond.org>
Cc: Tobias Nygren <tnn@NetBSD.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
7 years agoglx: don't expose systemTimeExtension for DRI2/DRI3/DRISW
Emil Velikov [Tue, 27 Sep 2016 12:39:36 +0000 (13:39 +0100)]
glx: don't expose systemTimeExtension for DRI2/DRI3/DRISW

Used/applicable to only dri1 drivers.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
7 years agoanv: do not open random render node(s)
Emil Velikov [Thu, 1 Dec 2016 21:21:10 +0000 (21:21 +0000)]
anv: do not open random render node(s)

drmGetDevices2() provides us with enough flexibility to build heuristics
upon. Opening a random node on the other hand will wake up the device,
regardless if it's the one we're interested or not.

v2: Rebase, explicitly require/check for libdrm
v3: Return VK_ERROR_INCOMPATIBLE_DRIVER for no devices (Ilia)
v4: Rebase

Cc: Jason Ekstrand <jason.ekstrand@intel.com>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com> (v1)
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoradv: do not open random render node(s)
Emil Velikov [Thu, 1 Dec 2016 20:58:20 +0000 (20:58 +0000)]
radv: do not open random render node(s)

drmGetDevices2() provides us with enough flexibility to build heuristics
upon. Opening a random node on the other hand will wake up the device,
regardless if it's the one we're interested or not.

v2: Rebase.
v3: Return VK_ERROR_INCOMPATIBLE_DRIVER for no devices (Ilia)

Cc: Michel Dänzer <michel.daenzer@amd.com>
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl> (v1)
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com> (v1)
Tested-by: Mike Lothian <mike@fireburn.co.uk>
7 years agoradv/winsys: use drmGetDevice2 API
Emil Velikov [Thu, 1 Dec 2016 19:53:11 +0000 (19:53 +0000)]
radv/winsys: use drmGetDevice2 API

Analogous to previous commit

v2: Add explicit require_libdrm check.

Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> (v1)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl> (v1)
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com> (v1)
Tested-by: Mike Lothian <mike@fireburn.co.uk>