On Friday, November 10, 2017 9:53:29 AM PST Antia Puentes wrote:
> This VS system value will contain the value passed as
> for indexed draw calls or the value passed as for non-indexed
> draw calls. It will be used to calculate the gl_VertexID as
> SYSTEM_VALUE_VERTEX_ID_ZERO_BASE plus
We only handled unpacking for GL_DEPTH_STENCIL formats.
Cemu was hitting _mesa_problem() for an unsupported format in
_mesa_unpack_float_32_uint_24_8_depth_stencil_row(), because the
format was depth-only, rather than depth-stencil.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94739
static const struct isl_drm_modifier_info modifier_info[] = {
> {
>.modifier = DRM_FORMAT_MOD_NONE,
> .name = "DRM_FORMAT_MOD_NONE",
>
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
signature.asc
Description: This is a digitally signed message p
apani.pa...@intel.com>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97852
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103801
> Acked-by: Andres Gomez <ago...@igalia.com>
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
signature.asc
Descrip
STATE_BASE_ADDRESS specifies a maximum size of the dynamic state
section, beyond which data supposedly reads back as 0. On Gen8+,
we were programming it to the size of the buffer. This worked fine
until we started growing the state buffer in commit 2dfc119f22f25708.
When the state buffer grows,
On Tuesday, November 28, 2017 4:13:21 PM PST Kenneth Graunke wrote:
> Uncommenting these blocks stress tests the batch/state growth code by
> trying to "grow" a buffer to the same size at least once per batch.
>
> When anything goes wrong with this code, dumping the validat
On Tuesday, November 28, 2017 4:13:20 PM PST Kenneth Graunke wrote:
> Growing the batch/state buffer is a lot more dangerous than I thought.
>
> A number of places emit multiple state buffer sections, and then write
> data to the returned pointer, or save a pointer to brw->
On Tuesday, November 28, 2017 6:34:20 PM PST Ian Romanick wrote:
> On 11/28/2017 04:13 PM, Kenneth Graunke wrote:
> > diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.c
> > b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
> > index b30ffe5d7f6..12d165d7236 100644
>
On Tuesday, November 28, 2017 6:15:59 PM PST Ian Romanick wrote:
> On 11/28/2017 04:13 PM, Kenneth Graunke wrote:
> > Once we reach the intended size of the buffer (BATCH_SZ or STATE_SZ), we
> > try and flush. If we're not allowed to flush, we resort to growing the
> > b
Uncommenting these blocks stress tests the batch/state growth code by
trying to "grow" a buffer to the same size at least once per batch.
When anything goes wrong with this code, dumping the validation list
is a useful way to figure out what's happening.
---
The original state buffer was marked with EXEC_OBJECT_CAPTURE. When
growing it, we want to preserve that flag so we continue to capture it
in GPU hang reports.
Fixes: 2dfc119f22f257082ab0 "i965: Grow the batch/state buffers if we need
space and can't flush."
Bugzilla:
The intention here is make the new BO use the same alignment as the old
BO. This isn't strictly necessary, but we would have to update the
'alignment' field in the validation list when swapping it out, and we
don't bother today.
The batch and state buffers use an alignment of 4096, so this
Growing the batch/state buffer is a lot more dangerous than I thought.
A number of places emit multiple state buffer sections, and then write
data to the returned pointer, or save a pointer to brw->batch.state.bo
and then use it in relocations. If each call can grow, this can result
in stale map
We're about to add more of them, and need to pass the whole lot of them
around together when growing them. Putting them in a struct makes this
much easier.
brw->batch.batch.bo is a bit of a mouthful, but it's nice to have things
labeled 'batch' and 'state' now that we have multiple buffers.
Once we reach the intended size of the buffer (BATCH_SZ or STATE_SZ), we
try and flush. If we're not allowed to flush, we resort to growing the
buffer so that there's space for the data we need to emit.
We accidentally got the threshold wrong. The first non-wrappable call
beyond (e.g.) STATE_SZ
These were moved to src/intel/common/gen_debug.h, but they are not
common code. They assume that brw_context or gl_context variables
exist, named brw or ctx. That isn't remotely true outside of i965.
---
src/intel/common/gen_debug.h| 29 -
This is cleaner than using a non-standard memclear macro (which does a
memset to 0) and then initializing fields after the fact. We move the
declarations to where we initialized the fields. While we're at it, we
move the declaration of 'ret' that goes with the ioctl, eliminating the
declaration
We want to program the 3DSTATE_RASTER field to the gl_context value,
not the other way around.
Fixes: 13ac46557ab1 (i965: Port Gen8+ 3DSTATE_RASTER state to genxml.)
---
src/mesa/drivers/dri/i965/genX_state_upload.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
le changed, 5 insertions(+), 5 deletions(-)
#ifndef NDEBUG is the right thing to use for guarding asserts.
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
signature.asc
Description: This is a digitally signed message part.
___
: Convert the supported_extensions struct to spirv_options
> spirv: Refactor a couple of pointer query helpers
Patches 1-7 (the above) and patch 10 (rename) are:
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
I'm planning to review the rest as well, but that's going to take a
littl
On Thursday, November 16, 2017 11:50:48 AM PST Ian Romanick wrote:
> On 11/14/2017 02:54 PM, Scott D Phillips wrote:
> > On gen >= 9, minmax reduction modes are available as a flag in
> > SAMPLER_STATE.
> > ---
> > docs/features.txt | 2 +-
> >
On Friday, November 17, 2017 5:02:04 PM PST Jason Ekstrand wrote:
> ---
> src/mesa/drivers/dri/i965/brw_bufmgr.c | 17 +++--
> 1 file changed, 11 insertions(+), 6 deletions(-)
Both patches are:
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org&g
e_tail = false;
Err...it's already false in this case. You just checked it two lines
earlier, and disasm_new_inst_group doesn't change.
> + } else {
> + group = exec_node_data(struct inst_group,
> + exec_list_get_tail_raw(disasm->group_list),
>
nit(>bo_cache);
> if (result != VK_SUCCESS)
>goto fail_batch_bo_pool;
>
> - result = anv_state_pool_init(>dynamic_state_pool, device, 16384);
> + /* For the state pools we explicitly disable 48bit. */
> + bo_flags = physical_device->has_exec_async ? EXEC_
Original 965 sets bits 28:27 to 0, while G45 and later set it to 1.
Also, there's no GPGPU pipeline on Gen5, and the pipeline to select
is a single bit - bit 1 is reserved.
---
src/intel/genxml/gen45.xml | 4 ++--
src/intel/genxml/gen5.xml | 5 ++---
2 files changed, 4 insertions(+), 5
We want to emit invariant state at the start of a render batch. In the
past, this more or less happened: a new batch flagged BRW_NEW_CONTEXT
(because we don't have hardware contexts), which triggered the
brw_invariant_state atom. So, it would be emitted before any 3D
drawing. (Technically,
This looks at prog->*. You don't get to do that unless you listen to
BRW_NEW_*_PROGRAM, which is a superset of the cases where
BRW_NEW_*_PROG_DATA is flagged.
---
src/mesa/drivers/dri/i965/gen7_l3_state.c | 5 +
1 file changed, 5 insertions(+)
diff --git
...and provide a better citation for the existing one.
v2:
- Apply the workaround to Gen8 too, as intended (caught by Topi).
- Restructure to add bits instead of an extra flush (based on a similar
patch by Rafael Antognolli).
---
src/mesa/drivers/dri/i965/brw_pipe_control.c | 41
On Tuesday, November 14, 2017 2:52:35 PM PST Anuj Phogat wrote:
> This fixes the changes introduced in commit 6165fda59b.
>
> Signed-off-by: Anuj Phogat <anuj.pho...@gmail.com>
> Cc: <mesa-sta...@lists.freedesktop.com>
> Cc: Nanley Chery <nanley.g.ch...@intel.co
On Sunday, November 12, 2017 1:41:29 AM PST Chris Wilson wrote:
> Quoting Kenneth Graunke (2017-11-12 08:35:05)
> > These three are the only we can reasonably decode with genxml.
>
> Makes sense, I would recommend putting those into an array for ease of
> future expansion. May
uot; <mesa-sta...@lists.freedesktop.org>
> Cc: Chad Versace <chadvers...@chromium.org>
> Cc: Daniel Stone <dani...@collabora.com>
> Cc: Louis-Francis Ratté-Boulianne <l...@collabora.com>
> Cc: Adam Jackson <a...@redhat.com>
> Cc: Chris Wilson &
t;src.addr.buffer);
> - brw_render_cache_set_check_flush(brw, params->dst.addr.buffer);
> + brw_cache_flush_for_read(brw, params->src.addr.buffer);
> + brw_cache_flush_for_render(brw, params->dst.addr.buffer);
This drops the cache flush, because brw_cache_flush_for_render
is
> buffer even though we don't know the caching mode. This can happen, for
> instance, on buffers imported from another GPU via prime.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101691
> Cc: Kenneth Graunke <kenn...@whitecape.org>
> Cc: Chris Wilson <ch
rp_batch *batch,
> vb[1].VertexBufferIndex = 1;
> vb[1].BufferPitch = 0;
> #if GEN_GEN >= 6
> - vb[1].VertexBufferMOCS = batch->blorp->mocs.vb;
> + vb[0].VertexBufferMOCS = vb[1].BufferStartingAddress.mocs;
I think you mean vb[1] here!
With that fixes, patches 1-3 ar
On Monday, November 13, 2017 1:30:10 PM PST Alejandro Piñeiro wrote:
>
> On 13/11/17 20:04, Kenneth Graunke wrote:
> > On Monday, November 13, 2017 2:16:59 AM PST Alejandro Piñeiro wrote:
> >> ---
> >> src/compiler/spirv/spirv_info_c.py | 5 -
> >> 1 fi
On Sunday, November 12, 2017 3:57:56 AM PST Lionel Landwerlin wrote:
> On 12/11/17 08:35, Kenneth Graunke wrote:
> > This makes aubinator_error_decode's shader dumping work like aubinator.
> > Instead of printing them after the fact, it prints them right inside the
> > 3
On Sunday, November 12, 2017 4:06:06 AM PST Lionel Landwerlin wrote:
> On 12/11/17 08:35, Kenneth Graunke wrote:
> > This is a bit more annoying than your average shader - we need to look
> > at MEDIA_INTERFACE_DESCRIPTOR_LOAD in the batch buffer, then hop over
> > to the
+), 2 deletions(-)
Subject is wrong, should be Gen8+ not gen9+.
Both patches are
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
signature.asc
Description: This is a digitally signed message part.
___
mesa-dev mailing list
mesa-dev@lists.
;"\
> +TEMPLATE = Template("""\
> +/* DO NOT EDIT - This file generated automatically by spirv_info_c.py script
> */
"is generated automatically"
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
> +
> +""" + COPYRIGHT + "&quo
T2
> i965: perf: add support for CoffeeLake GT3
Patches 2-3 are
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
and the rest are
Acked-by: Kenneth Graunke <kenn...@whitecape.org>
because I have no idea where you get this information from, so I can't
really re
> headers in any way.
>
> Cc: Lionel Landwerlin <lionel.g.landwer...@intel.com>
> ---
> src/intel/genxml/gen6.xml | 2 --
> src/intel/genxml/gen7.xml | 2 --
> src/intel/genxml/gen75.xml | 2 --
> src/intel/genxml/gen8.xml | 2 --
> 4 files changed, 8 deleti
On Saturday, November 11, 2017 3:35:21 PM PST Lionel Landwerlin wrote:
> :(
>
> The intention was to deal with cases where we've encountered more than
> MAX_NUM_PROGRAMS (actually happened to me).
> So we start by the index + 1 assuming this is the oldest program because
> we're in a rolling
while loops skip the first field of the instruction/structure, which
is not what the code intended. It works out because the field we're
looking for doesn't happen to be first, but we ought to do it right
regardless.
Found while writing the next patch, where Kernel Start Pointer is
the first
This is a bit more annoying than your average shader - we need to look
at MEDIA_INTERFACE_DESCRIPTOR_LOAD in the batch buffer, then hop over
to the dynamic state buffer to read the INTERFACE_DESCRIPTOR_DATA, then
hop over to the instruction buffer to decode the program.
Now that we store all the
This makes aubinator_error_decode's shader dumping work like aubinator.
Instead of printing them after the fact, it prints them right inside the
3DSTATE_VS/HS/DS/GS/PS packet that references them. This saves you the
effort of cross-referencing things and jumping back and forth.
It also reduces a
This lets us complete parsing and storing of each buffer's data before
we begin decoding the batchbuffer. This makes it possible to inspect
the state buffer and program buffer, so we can properly decode any
indirect state or shader programs.
---
src/intel/tools/aubinator_error_decode.c | 95
The dashes "---" may occur within an ascii85 block, but only an ascii85
block starts with ':' or '~'.
Ported from Chris Wilson's intel-gpu-tools commit:
bceec7e1d8a160226b783c6344eae8cbf4ece144
---
src/intel/tools/aubinator_error_decode.c | 60 +++-
1 file changed, 29
Ported from intel_error_decode. We don't want to run off the end.
---
src/intel/tools/aubinator_error_decode.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/intel/tools/aubinator_error_decode.c
b/src/intel/tools/aubinator_error_decode.c
index 9cd0fa761dd..f9d14cc20d5 100644
---
Based on a similar patch to intel_error_decode by Chris Wilson.
While we're de-duplicating the gtt_offset calculation, we can simplify
it to assume two hex digits are there - the kernel has done this since
v4.6, and we already require error states from v4.10.
---
Also change count from a pointer into a value. We were supposed to
be resetting it to 0 (and failed to), but that's gone since we dropped
the pre-ascii85 handling.
---
src/intel/tools/aubinator_error_decode.c | 8 +++-
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git
Dead code.
---
src/intel/tools/aubinator_error_decode.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/src/intel/tools/aubinator_error_decode.c
b/src/intel/tools/aubinator_error_decode.c
index 0b0a308a317..9cd0fa761dd 100644
--- a/src/intel/tools/aubinator_error_decode.c
+++
Hello,
This series does a pretty substantial rework of aubinator_error_decode
(the new error state decoder).
The first few patches fix a whole bunch of bugs in the decoding, where
we parsed the file wrong, decoded the wrong data in the wrong places,
or decoded non-batch data as commands
Error state files used to look like:
render ring --- gtt_offset = 0x0e8f6000
: 6904
0004 : 7909
...
7ffc :
--- ringbuffer = 0x1000
There were thousands of lines between sections. The file format changed
with Kernel 4.10, and now has a
These three are the only we can reasonably decode with genxml.
---
src/intel/tools/aubinator_error_decode.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/intel/tools/aubinator_error_decode.c
b/src/intel/tools/aubinator_error_decode.c
index 94436e957d4..ae9f08f1d2b
This indexing is bogus. idx_program is the offset of the next program.
At this point, we've walked through the entire batch, and accumulated
all the programs. So adding it again simply skips over 100% of the
programs.
---
src/intel/tools/aubinator_error_decode.c | 13 ++---
1 file
We renamed "Function Enable" to "Enable", which broke our detection
of whether shaders are enabled or not. So, we'd see a bunch of HS/DS
packets with program offsets of 0, and think that was a valid TCS/TES.
Fixes: c032cae9ff77e (genxml: Rename "Function Enable" to "Enable".)
---
The L3 configuration code already considers the TCS and TES programs,
but failed to listen for TCS/TES program changes.
This was somehow missing.
Cc: mesa-sta...@lists.freedesktop.org
---
src/mesa/drivers/dri/i965/gen7_l3_state.c | 2 ++
1 file changed, 2 insertions(+)
diff --git
On Thursday, November 9, 2017 12:31:18 PM PST Jason Ekstrand wrote:
> On Thu, Nov 9, 2017 at 12:45 AM, Kenneth Graunke <kenn...@whitecape.org>
> wrote:
>
> > This fixes the missing AutomaticSize handling in the ABO code, removes
> > a bunch of duplicated code, and drops
On Thursday, November 9, 2017 11:22:34 AM PST Rafael Antognolli wrote:
> On Thu, Nov 09, 2017 at 12:59:12AM -0800, Jordan Justen wrote:
> > Reviewed-by: Jordan Justen
>
> It's also
>
> Tested-by: Rafael Antognolli
Sorry, I forgot to sync
This fixes the missing AutomaticSize handling in the ABO code, removes
a bunch of duplicated code, and drops an extra layer of wrapping around
brw_emit_buffer_surface_state().
---
src/mesa/drivers/dri/i965/brw_context.h | 10 --
src/mesa/drivers/dri/i965/brw_wm_surface_state.c | 113
Having this separate could potentially make programs that rebind atomics
but no other surfaces ever so slightly faster. But it's a tiny amount
of code to add to the existing UBO/SSBO atom, and very related.
The extra atoms have a cost on every draw call, and so dropping some of
them would be
We use the same hardware mechanism for both atomic counters and SSBO
atomics, so there's really no benefit to maintaining separate code to
handle each case. Instead, we can just use Rob's shiny new NIR pass to
convert atomic_uints to SSBOs, and delete piles of code.
The ssbo_start section of the
Similar to what we did for pixel shader threads - see gen_device_info.c.
We don't want to bump the actual Maximum Number of Threads though, so
we adjust it here. For pixel shaders, we don't use max_wm_threads, so
we could just bump it globally.
Fixes Piglit tests:
| 6 +++---
> src/compiler/glsl/shader_cache.cpp | 4 ++--
> 4 files changed, 8 insertions(+), 8 deletions(-)
Patches 1-2 are:
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
Jordan might want to take a look at patch 1 as well...
signature.asc
Descripti
On Tuesday, November 7, 2017 5:41:59 PM PST Timothy Arceri wrote:
> When I introduced gl_shader_program_data one of the intentions was to
> fix a bug where a failed linking attempt freed data required by a
> currently active program. However I seem to have failed to finish
> hooking up the final
MESA_FORMAT_B8G8R8X8_UNORM,
>
> + MESA_FORMAT_B8G8R8A8_SRGB,
> + MESA_FORMAT_B8G8R8X8_SRGB,
> +
I'm worried that exposing B8G8R8X8_SRGB is not going to work. See the
explanation in commit 61cdb7665f7bd147533cdc5977750d990c2eafd5. Unless
I'm mistaken, I don't think anything
According to the GLSL ES 3.20, GLSL 4.50, and GLSL 1.20 specs:
"To force all output variables to be invariant, use the pragma
#pragma STDGL invariant(all)
before all declarations in a shader."
Notably, this is only supposed to affect output variables. Furthermore,
"Only
According to the GLSL ES 3.20, GLSL 4.50, and GLSL 1.20 specs:
"To force all output variables to be invariant, use the pragma
#pragma STDGL invariant(all)
before all declarations in a shader."
Notably, this is only supposed to affect output variables. Furthermore,
"Only
4 files changed, 5 insertions(+), 5 deletions(-)
Both are
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
signature.asc
Description: This is a digitally signed message part.
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists
On Thursday, October 19, 2017 9:02:20 PM PST Tomasz Figa wrote:
> Hi Ian, Kenneth,
>
> On Wed, Sep 27, 2017 at 2:57 AM, Tomasz Figa wrote:
> > Commit 259fc505454ea6a67aeacf6cdebf1398d9947759 added linker error for
> > mismatching uniform precision, as required by GLES 3.0
keep same opcode and just remap SSBO buffer index: */
>op = instr->intrinsic;
> idx_src = (op == nir_intrinsic_store_ssbo) ? 1 : 0;
>
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
signature.asc
Description: This is a digitally signed message
On Thursday, November 2, 2017 12:01:48 PM PDT Adam Jackson wrote:
> From: Neil Roberts <n...@linux.intel.com>
>
> Reviewed-by: Adam Jackson <a...@redhat.com>
> Signed-off-by: Neil Roberts <n...@linux.intel.com>
Patches 2 and 9 are:
Reviewed-by: Kenneth
For the render pipeline, the upload_sampler_state_table atom emits
3DSTATE_BINDING_TABLE_POINTERS_XS. It tries to avoid this for compute:
if (GEN_GEN >= 7 && stage_state->stage != MESA_SHADER_COMPUTE) {
/* Emit a 3DSTATE_SAMPLER_STATE_POINTERS_XS packet. */
bug.cgi?id=103085
>
> Signed-off-by: Plamena Manolova <plamena.manol...@intel.com>
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
and pushed. Thanks Pam!
signature.asc
Description: This is a digitally signed message part.
__
On Sunday, October 29, 2017 5:21:47 PM PDT Jordan Justen wrote:
> On 2017-10-29 01:11:32, Kenneth Graunke wrote:
> > On Sunday, October 22, 2017 1:01:36 PM PDT Jordan Justen wrote:
> > > If the i965 gen program cannot be loaded from the cache, then we
> > > fallbac
hdr_for_addr(brw_disk_cache_init);
> + int id_size = build_id_length(note);
> + assert(note && id_size == 20 /* sha1 */);
> +
> + const uint8_t *id_sha1 = build_id_data(note);
> + assert(id_sha1);
> +
> + char timestamp[41];
> + _mesa_sha1_format(timestamp, id_sha1
inux, dep_libdrm, dep_expat, dep_m, dep_thread,
> dep_dl],
> -link_args : [ld_args_bsymbolic, ld_args_gc_sections],
> +link_args : ['-Wl,--build-id=sha1', ld_args_bsymbolic,
> ld_args_gc_sections],
>)
>
>pkg.generate(
>
The autotools bits look f
> * validation.
> */
> - cache_item_md_size += sizeof(cache_key);
> + cache_item_md_size += num_keys * sizeof(cache_key);
>ret = lseek(fd, num_keys * sizeof(cache_key), SEEK_CUR);
>if (ret == -1)
> goto fail;
>
Reviewed-
brwCreateContext?
I'm always a bit confused about the driver/screen split here, but
I think this looks right.
With those changes,
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
signature.asc
Description: This is a digitally signed message part.
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
og->_LinkedShaders); stage++) {
>struct gl_linked_shader *shader = shProg->_LinkedShaders[stage];
>if (!shader)
>
I was concerned that we might be skipping some steps here, but I think
everything's OK...
Reviewed-by: Kenneth Grau
On Monday, October 30, 2017 3:37:40 AM PDT Eric Engestrom wrote:
> Fixes: 2c873060d3578c7004c0 "i965: Delete unused
>brw_vs_prog_data::nr_attributes field."
> Cc: Kenneth Graunke <kenn...@whitecape.org>
> Signed-off-by: Eric Engestrom <eric.engest...@imgtec.c
Due to a gaffe on my part, we were re-emitting all binding table entries
on every single draw call. The push_constant_packets atom listens to
BRW_NEW_DRAW_CALL, but skips emitting 3DSTATE_CONSTANT_XS for each stage
unless stage_state->push_constants_dirty is true. However, it flagged
On Sunday, October 22, 2017 1:01:21 PM PDT Jordan Justen wrote:
> v3:
> * Rename serialized_nir* to driver_cache_blob*. (Tim)
>
> Signed-off-by: Jordan Justen
> Reviewed-by: Timothy Arceri
> ---
> src/compiler/glsl/shader_cache.cpp | 16
On Sunday, October 22, 2017 1:01:36 PM PDT Jordan Justen wrote:
> If the i965 gen program cannot be loaded from the cache, then we
> fallback to using a serialized nir program.
>
> This is based on "i965: add cache fallback support" by Timothy Arceri
> . Tim's
eword subject]
> [jordan.l.jus...@intel.com: *_cached_program => brw_disk_cache_*_program]
> Signed-off-by: Jordan Justen <jordan.l.jus...@intel.com>
Patches 22-27 are:
Reviewed-by: Kenneth Graunke <kenn...@whitecape.org>
signature.asc
D
On Sunday, October 22, 2017 1:01:29 PM PDT Jordan Justen wrote:
> From: Timothy Arceri
>
> This uses the recently-added disk_cache.c to write out the final
> linked binary for vertex and fragment shader programs.
>
> This is based off the initial implementation
On Saturday, October 28, 2017 11:08:29 AM PDT Jason Ekstrand wrote:
> Have you seen this matter somewhere? It seems like a reasonable thing to
> do none the less. 3 and 4 are
>
> Reviewed-by: Jason Ekstrand
Yeah, the Aztec Ruins GL version hits this on almost all of its
ck from the symbol table rather waiting for an instruction which
> references it.
>
> I’ve tested it with Piglit and the public CTS repo and it doesn’t
> cause any regressions, and of course it fixes the CTS test.
>
> It might be nice to add a Piglit or CTS test to check this mis
i965 turns fddx/fddy into their coarse/fine variants based on the
ctx->Hint.FragmentShaderDerivative setting. It needs to know whether
this can impact a shader in order to better guess NOS settings.
---
src/compiler/nir/nir_gather_info.c | 16
src/compiler/shader_info.h
This avoids recompiles for shaders that don't use explicit derivatives
when ctx->Hint.FragmentShaderDerivative == GL_NICEST.
---
src/mesa/drivers/dri/i965/brw_wm.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c
b/src/mesa/drivers/dri/i965/brw_wm.c
index
This has been unused since we switched to nir_lower_wpos_ytransform.
---
src/intel/compiler/brw_compiler.h | 1 -
1 file changed, 1 deletion(-)
diff --git a/src/intel/compiler/brw_compiler.h
b/src/intel/compiler/brw_compiler.h
index 6a330537a5b..17615c6dd9f 100644
---
Also, reorder them to match the structure's field order, to make it
easier to check that they're all present.
---
src/mesa/drivers/dri/i965/brw_wm.c | 19 ++-
1 file changed, 14 insertions(+), 5 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c
On Friday, October 27, 2017 3:17:20 PM PDT Anuj Phogat wrote:
> V2: Use _NEW_BUFFERS in place of _NEW_MULTISAMPLE
>
> Signed-off-by: Anuj Phogat <anuj.pho...@gmail.com>
> Cc: Kenneth Graunke <kenn...@whitecape.org>
> ---
> src/mesa/drivers/dri/i965/genX_state_up
On Friday, October 27, 2017 10:50:23 AM PDT Anuj Phogat wrote:
> Signed-off-by: Anuj Phogat
> ---
> src/mesa/drivers/dri/i965/genX_state_upload.c | 10 ++
> 1 file changed, 10 insertions(+)
>
> diff --git a/src/mesa/drivers/dri/i965/genX_state_upload.c
>
On Thursday, October 26, 2017 11:16:06 AM PDT Eric Anholt wrote:
> Gwan-gyeong Mun writes:
>
> > Hi,
> >
> > These Patch v5 series modified with new helper function series [1].
> >
> > These series only have mesa for tizen feature.
> >
> > [1]
ze);
> -uint32_t _mesa_hash_string(const char *key);
> +uint32_t _mesa_hash_string(const void *key);
> bool _mesa_key_string_equal(const void *a, const void *b);
> bool _mesa_key_pointer_equal(const void *a, const void *b);
>
>
Reviewed-by: Kenneth Graunke <kenn...@whitecape.or
On Friday, October 27, 2017 4:26:32 AM PDT Pohjolainen, Topi wrote:
> On Fri, Oct 27, 2017 at 03:02:59AM -0700, Kenneth Graunke wrote:
> > On Wednesday, October 25, 2017 10:37:37 AM PDT Topi Pohjolainen wrote:
> > > Fixes intermittent GPU hangs on Broxton with an Intel intern
_bo)
> - brw_bo_unreference(brw->wm.base.scratch_bo);
> +
> + brw_bo_unreference(brw->vs.base.scratch_bo);
> + brw_bo_unreference(brw->tcs.base.scratch_bo);
> + brw_bo_unreference(brw->tes.base.scratch_bo);
> + brw_bo_unreference(brw->gs.base.scratch_bo);
> + brw_b
from
> this - most tests kick off one render cycle and any corruption
> is left unseen.
>
> v2 (Ken): Instead of enabling push constants, enable one of the
> inputs (PSIZ).
> v3 (Ken, Jason): Use LAYER instead making vulkan emit_3dstate_sbe()
> h
t * driContextPriv)
> > if (brw->wm.base.scratch_bo)
> >brw_bo_unreference(brw->wm.base.scratch_bo);
> >
> > + if (brw->vs.base.push_const_bo)
> I'd drop the if checks - brw_bo_unreference works fine when the bo
> pointer is NULL.
>
>
601 - 700 of 8470 matches
Mail list logo