Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Feb 14, 2017, at 07:25, Dave Airlie wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> Just one bonaire fix.
>
> Signed-off-by: Dave Airlie <airl...@redhat.com>
> ---
> src/amd/vulkan/si_
Series is:
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Feb 14, 2017, at 07:10, Dave Airlie wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> On some CIK (Hawaii) this needs to be at least 64k, I'm not 100% sure
> it doesn't need to be 128k.
&g
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
I'll adjust the sparse binding patches.
On Thu, Feb 9, 2017, at 04:25, Dave Airlie wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> It's legal to submit just semaphores with no command streams,
> this patch fix
The struct have different size, so the arrays have different stride.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_device.c | 45 +++-
1 file changed, 36 insertions(+), 9 deletions(-)
diff --git a/src/amd/vulkan/radv_de
r->level ==
> VK_COMMAND_BUFFER_LEVEL_PRIMARY);
> -
> - cs_array[j] = cmd_buffer->cs;
> + if (cmd_buffer->no_draws == true) {
We never want to end up with 0 bo's if we have a fence or semaphores.
Maybe chan
Was thinking of a sort or other asymptotically more efficient merge
method, but in the meantime this is
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Fri, Feb 10, 2017, at 02:07, Dave Airlie wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> Just not
nt32_t count)
> {
Rename count to vertex_count or something that at least tells what it
counts. With that
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
> struct radv_pipeline *pipeline = cmd_buffer->state.pipeline;
> struct radv_device *device = cmd_buffer->devi
r-b for now.
On Wed, Feb 15, 2017, at 01:16, Dave Airlie wrote:
> From: Dave Airlie
>
> This is possibly a bad idea, I might have to consider a better one.
>
> Signed-off-by: Dave Airlie
> ---
> src/amd/vulkan/radv_cmd_buffer.c | 1 +
> 1 file changed,
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_device.c | 2 +-
src/amd/vulkan/radv_pipeline.c | 3 ++-
2 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c
index 1132eeb52d7..9185ef16dc0
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_device.c | 2 +-
src/amd/vulkan/radv_pipeline.c | 1 +
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c
index 9185ef16dc0..be0d9d933fc
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/compiler/spirv/nir_spirv.h| 1 +
src/compiler/spirv/spirv_to_nir.c | 5 -
2 files changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/compiler/spirv/nir_spirv.h b/src/compiler/spirv/nir_spirv.h
index e43e9
No regressions on CTS 1.0.2 mustpass, though I don't think there are
tests for these extensions. I checked there were no references to the
format in the nir->LLVM conversion though.
Bas Nieuwenhuizen (3):
radv: Add support for shaderStorageImageWriteWithoutFormat.
spirv: Add supp
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Wed, Feb 15, 2017, at 03:28, Dave Airlie wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> This is used in DOOM, so provide the fast clear path for it.
>
> Signed-off-by: Dave Airlie <airl...@redha
Series is
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Mon, Feb 13, 2017, at 05:15, Dave Airlie wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> This puts the common gfx state for the device into an
> indirect buffer, and just calls out
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Sun, Feb 12, 2017, at 20:11, Dave Airlie wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> If a buffer is just full of flushes we flush things on command
> buffer submission, so don't bother submitting thes
This one only keeps allocated memory in the list, and list nodes
in the descriptor sets. Thsi doesn't need messing around with
max_sets, and we get automatic merging of free regions.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_descriptor_set.
We only use the freed ones after all free space has been used. If
the app only allocates small descriptor sets, we might go over
max_sets before the memory is full.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
CC: <mesa-sta...@lists.freedesktop.o
On Wed, Feb 15, 2017, at 09:43, Dave Airlie wrote:
> From: Dave Airlie
>
> zero extend ->u64 and sign extend ->i64.
>
> Signed-off-by: Dave Airlie
> ---
> src/amd/common/ac_nir_to_llvm.c | 7 +++
> 1 file changed, 7 insertions(+)
>
> diff --git
On Sat, Feb 11, 2017, at 13:03, Timothy Arceri wrote:
>
>
> On 10/02/17 21:43, Nicolai Hähnle wrote:
> > On 07.02.2017 04:42, Timothy Arceri wrote:
> >> From: Timothy Arceri
> >>
> >> This will be used to store all pointers in the cache as 64bit ints
> >>
n");
> + fprintf(stderr, "vulkan: No DRI3 support detected - required for
> presentation\n");
> + fprintf(stderr, "Note: Buggy applications may crash, please report to
> vendor\n");
Maybe an " if they do" at the end of it to clarify tha
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Wed, Jan 18, 2017 at 11:48 PM, Andres Rodriguez <andre...@gmail.com> wrote:
> In situations where libdrm_amdgpu and mesa are installed to the same
> location, the mesa installed headers will take precedence over the gi
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Jan 17, 2017 at 9:27 PM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> This fixes some issues we'd hit later if using viewport
> indexes.
>
> Signed-off
On Mon, Jan 16, 2017 at 2:51 PM, Emil Velikov wrote:
> On 14 January 2017 at 02:31, Andres Rodriguez wrote:
>> On Fri, Jan 13, 2017 at 8:13 PM, Emil Velikov
>> wrote:
>>>
>>> On 13 January 2017 at 23:44, Andres Rodriguez
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Jan 17, 2017 at 1:07 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> Just always use the layer clear pipelines,
> the overhead of emitting the layer shouldn
Pretty straightforward. Also deleted the big comment block as it
is a pretty standard pattern for filling in arrays.
Also removed the error message on non-existent devices, as getting
7 errors printed to the console each time you enumerate the
devices is pretty confusing.
Signed-off-by: Bas
Otherwise we read past the end of the buffer.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/common/ac_debug.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/amd/common/ac_debug.c b/src/amd/common/ac_debug.c
index f91e448a47f..989dfda4ff9 100644
--- a/src/amd/
On Mon, Jan 16, 2017 at 9:30 PM, Andres Rodriguez <andre...@gmail.com> wrote:
> Small comments inline for a bit of extra error handling. Still digesting
> radv and vulkan, so the feedback might be a bit incorrect.
>
>
>
> On 2017-01-16 02:59 PM, Bas Nieuwenhu
for number of DRM devices.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_device.c | 60 ---
src/amd/vulkan/radv_private.h | 3 ++-
2 files changed, 19 insertions(+), 44 deletions(-)
diff --git a/src/amd/vulkan/radv_device.c
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_cmd_buffer.c | 18 +-
src/amd/vulkan/radv_device.c | 15 ---
src/amd/vulkan/radv_image.c | 18 +-
src/amd/vulkan/radv_pipeline.c | 10 +
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
for both.
On Mon, Jan 16, 2017 at 11:47 PM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> makes it easier to add other shader stages.
>
> Signed-off-by: Dave Airlie
Looks correct to me, but why don't we always use the layered clear?
I'd think the extra VS output doesn't really matter given the low
amount of geometry during clearing, and that would save some shader
variants, which we have a lot of already.
- Bas
On Mon, Jan 16, 2017 at 11:58 PM, Dave Airlie
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
btw, isn't this enough already to export layer from the VS already?
Might be useful for layered clears without GS.
On Mon, Jan 16, 2017 at 11:25 PM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Mon, Jan 16, 2017 at 9:38 PM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> We only need one per samples (maybe not even that), reduce
> all the unneeded ones.
>
&
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Mon, Jan 16, 2017 at 9:44 PM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> This is just prep work for layered clears, it doesn't change
> anything.
>
&
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Mon, Jan 16, 2017 at 9:49 PM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> Just noticed this while in the area.
>
> v2: one replacement was incorrect.
>
On Mon, Jan 16, 2017 at 10:07 PM, Dave Airlie wrote:
> From: Dave Airlie
>
> This just adds the infrastructure to allow writing layer
> and viewport index. It's just a first patch out of the geom
> shader tree, and doesn't do much on its own.
>
>
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/common/ac_nir_to_llvm.c | 18 ++
1 file changed, 18 insertions(+)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 9778581a1ef..4f3d689db7e 100644
--- a/src/amd/
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Feb 28, 2017 at 7:53 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> I messed this up when I wrote it, this fixes:
> dEQP-VK.memory.pipeline_barrier.*unifor
Pushed, thanks.
On Wed, Mar 1, 2017 at 10:16 PM, Tobias Klausmann
wrote:
> Fix usage of ac_add_function_attr() and make it known!
>
> common/ac_nir_to_llvm.c: In function 'create_llvm_function':
> common/ac_nir_to_llvm.c:265:4: error: implicit declaration of
Otherwise if the new compute pipeline is the same as the last used
pipeline before the call, we don't emit it again.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
Cc: 13.0 17.0 <mesa-sta...@lists.freedesktop.org>
---
src/amd/vulkan/radv_cmd_buffer.c | 1 +
1 file changed,
Per spec, VK_QUERY_RESULT_64_BIT specifies the integer size and the
availability flag is an integer. We apparently handled this correctly
already for the copy to buffer case.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
Cc: 13.0 17.0 <mesa-sta...@lists.freedesktop.org>
---
sr
an unitialized value is
better than a hang.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
Cc: 13.0 17.0 <mesa-sta...@lists.freedesktop.org>
---
src/amd/vulkan/radv_query.c | 65 +++--
1 file changed, 45 insertions(+), 20 deletions(-)
diff --g
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Mon, Feb 27, 2017 at 8:00 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> For blitting we need to use the depth or stencil format, never
> the combined.
>
>
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Mon, Feb 27, 2017 at 3:15 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> These formats are used by some CTS tests, may as well fill them in.
>
> Signed-off-by
We can then use constants for indexed loads.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_descriptor_set.c | 5 +
src/amd/vulkan/radv_descriptor_set.h | 1 +
2 files changed, 6 insertions(+)
diff --git a/src/amd/vulkan/radv_descriptor_set.c
b/src/amd/
So we don't need to know about radv_sampler in ac_nir_to_llvm.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_descriptor_set.c | 14 +-
src/amd/vulkan/radv_descriptor_set.h | 2 +-
2 files changed, 6 insertions(+), 10 deletions(-)
diff --git a/s
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_descriptor_set.c | 23 ++-
1 file changed, 14 insertions(+), 9 deletions(-)
diff --git a/src/amd/vulkan/radv_descriptor_set.c
b/src/amd/vulkan/radv_descriptor_set.c
index d8399574f2a..ca5b2
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/common/ac_nir_to_llvm.c | 16
1 file changed, 16 insertions(+)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 4f3d689db7e..db7194c3086 100644
--- a/src/amd/
On Wed, Sep 7, 2016 at 1:46 PM, Marek Olšák wrote:
> From: Marek Olšák
>
> ---
> src/gallium/drivers/radeon/r600_texture.c | 45
> ++-
> 1 file changed, 32 insertions(+), 13 deletions(-)
>
> diff --git
Hi Marek,
I don't think this accounts for the fast clear bits? unorm->uint and
snorm<->sint should have compatible clear values, but otherwise we may
need to eliminate the fast clears.
Yours sincerely,
Bas Nieuwenhuizen
On Mon, Aug 29, 2016 at 5:28 PM, Marek Olšák <mar...@gmai
0x000288F0
> -#define R_0282D0_PA_SC_VPORT_ZMIN_0 0x000282D0
> -#define R_0282D4_PA_SC_VPORT_ZMAX_0 0x000282D4
> +#define R_0282D0_PA_SC_VPORT_ZMIN_0
> 0x0282D0
> +#define R_0282D4_PA_SC_VPORT_ZMAX_0
On Wed, Sep 7, 2016 at 3:42 PM, Marek Olšák <mar...@gmail.com> wrote:
> On Wed, Sep 7, 2016 at 2:11 PM, Bas Nieuwenhuizen
> <b...@basnieuwenhuizen.nl> wrote:
>> On Wed, Sep 7, 2016 at 1:46 PM, Marek Olšák <mar...@gmail.com> wrote:
>>> From: Marek Olšák <
On Tue, Aug 30, 2016 at 9:06 AM, Marek Olšák <mar...@gmail.com> wrote:
> On Mon, Aug 29, 2016 at 10:33 PM, Bas Nieuwenhuizen
> <b...@basnieuwenhuizen.nl> wrote:
>> On Mon, Aug 29, 2016 at 5:28 PM, Marek Olšák <mar...@gmail.com> wrote:
>>> From: Marek Olšák
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Sat, Sep 10, 2016 at 12:40 AM, Marek Olšák <mar...@gmail.com> wrote:
> From: Marek Olšák <marek.ol...@amd.com>
>
> LLVM can CSE the loads, thus we can always re-load constants before each
> use. The dec
On Fri, Sep 30, 2016 at 3:31 AM, Dave Airlie wrote:
> On 30 September 2016 at 01:07, Jason Ekstrand wrote:
>> On Sep 29, 2016 7:56 AM, "Emil Velikov" wrote:
>>>
>>> On 28 September 2016 at 19:53, Marek Olšák
On Fri, Sep 30, 2016 at 2:13 PM, Marek Olšák wrote:
> intptr_t reads and writes aren't atomic. p_atomic_set and
> p_atomic_read functions don't do anything for atomicity. See:
>
> #define p_atomic_set(_v, _i) (*(_v) = (_i))
> #define p_atomic_read(_v) (*(_v))
That
On Thu, Oct 6, 2016 at 1:09 AM, Gustaw Smolarczyk wrote:
> It's supposed to be how much at least we want to grow the cs, not the
> minimum size of the cs after growth.
> ---
> src/amd/vulkan/winsys/amdgpu/radv_amdgpu_cs.c | 3 ++-
> 1 file changed, 2 insertions(+), 1
continue;
>
> + if (!fence->submitted)
> + return VK_TIMEOUT;
> +
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
> expired = device->ws->fence_wait(device->ws, fence->fence,
On Fri, Sep 16, 2016 at 3:57 PM, Nicolai Hähnle wrote:
> From: Nicolai Hähnle
>
> These functions extract the pipe state structure from the current
> descriptors, for state saving.
> ---
> src/gallium/drivers/radeonsi/si_descriptors.c | 46
>
I don't think the "gallium:" commit message prefix is correct here.
Looking at the logs it should be "r600g/sb:".
With that change:
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
Do you need someone to push it for you?
- Bas
On Fri, Sep 16, 2016 at
On Wed, Sep 7, 2016 at 6:23 PM, Christian König wrote:
> Am 07.09.2016 um 18:06 schrieb Marek Olšák:
>>
>> On Wed, Sep 7, 2016 at 5:36 PM, Ilia Mirkin wrote:
>>>
>>> On Wed, Sep 7, 2016 at 4:08 AM, Michel Dänzer wrote:
I would prefer it if the function could be split in two functions
instead of using the initialized flag. I think we know whether it is
an initialization or a reinitialization per call site.
- Bas
On Thu, Aug 18, 2016 at 9:46 PM, Marek Olšák wrote:
> From: Marek Olšák
Patch 3, 4 are
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
For patch 1 and 2 I'm not familiar with the difference between the SI
and CIK+ tiling modes to be sure they are good, but the other parts of
the patch look good to me.
btw, I'm not sure what you tried wrt the C
Patch 1-7 of this series are
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
- Bas
(Adding mesa-dev ML this time)
On Thu, Aug 18, 2016 at 9:46 PM, Marek Olšák <mar...@gmail.com> wrote:
> From: Marek Olšák <marek.ol...@amd.com>
>
> If the kernel driver doesn
This patch is
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
Tested-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
Thanks.
On Tue, Aug 23, 2016 at 3:11 AM, Jason Ekstrand <ja...@jlekstrand.net> wrote:
> The only reason we should throw INITIALIZATION_FAILE
On Wed, Sep 28, 2016 at 6:27 PM, Nicolai Hähnle wrote:
> On 28.09.2016 16:20, Ilia Mirkin wrote:
>>
>> On Wed, Sep 28, 2016 at 6:25 AM, Nicolai Hähnle
>> wrote:
>>>
>>> From: Nicolai Hähnle
>>>
>>> The difference to the virtually
On Thu, Sep 29, 2016 at 10:20 AM, Nicolai Hähnle <nhaeh...@gmail.com> wrote:
> On 29.09.2016 00:00, Bas Nieuwenhuizen wrote:
>>
>> On Wed, Sep 28, 2016 at 6:27 PM, Nicolai Hähnle <nhaeh...@gmail.com>
>> wrote:
>>>
>>> On 28.09.2016 16:20, Ilia Mir
Gustaw referred to is probably the cause of configure not
giving an error about it. If you want to be sure this is handled
right, also enable radeonsi to be build.
Yours sincerely,
Bas Nieuwenhuizen
On Sat, Oct 1, 2016 at 2:23 PM, Limeth <jakub.hlusi...@email.cz> wrote:
> Hello,
> I decide
Hi Gustaw,
The patch is
reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
What needs to be done too, is checking if the resulting IB becomes too
large in the SI case, and handling that gracefully. I don't care if
that happens with this patch, or if someone writes a follow up patch
For the series:
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Oct 25, 2016 at 12:23 AM, Marek Olšák <mar...@gmail.com> wrote:
> Patch 1: SI fix for MSAA for Mesa 13.0. I don't know the impact of that bug.
> Patch 2: Possible GPU hang fix for Hawaii a
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Wed, Oct 26, 2016 at 3:04 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> The intrinsic engine asserts in llvm due to this,
> as we put a vec4 into a vec1, an
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Wed, Nov 9, 2016 at 2:22 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> if a fence is created pre-signaled we should return that
> in GetFenceStatus even if it has
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Thu, Nov 10, 2016 at 1:33 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> We can only read the valid samples if this is an MSAA
> texture, which means the t
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Nov 8, 2016 at 7:24 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> I was getting a random GPU hang in the renderpass simple tests,
> it turns out sometimes radv
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Wed, Oct 19, 2016 at 8:38 PM, Nicolai Hähnle <nhaeh...@gmail.com> wrote:
> From: Nicolai Hähnle <nicolai.haeh...@amd.com>
>
> Fixes spec/arb_tessellation_shader/execution/dvec[23]-vs-tcs-tes, among
On Thu, Oct 20, 2016 at 4:07 AM, Dave Airlie wrote:
> From: Dave Airlie
>
> This fixes
> dEQP-VK.pipeline.multisample.sampled_image*
>
> These all render to multisampled image, and then
> sample from it, so we must transition it correctly.
>
> Cc: "13.0"
Double negative? I think the message is much more consistent with
single negative. With that, patch 2-5 are
Acked-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
>
> Signed-off-by: Emil Velikov <emil.veli...@collabora.com>
> ---
> src/amd/Makefile.addrlib.am | 3 +++
&
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Fri, Oct 21, 2016 at 1:36 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> This fixes
> dEQP-VK.pipeline.multisample.sampled_image*
>
> These all render to m
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Mon, Oct 24, 2016 at 9:39 PM, Marek Olšák <mar...@gmail.com> wrote:
> From: Marek Olšák <marek.ol...@amd.com>
>
> This should fix random GPU hangs on Hawaii and Fiji.
> It's already been fixed in 13.0 and
Remove vs exports if the ps doesn't use them.
Longer term we will probably want to do the optimizations between
stages in nir, but this seems simple enough for the short term.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/common/ac_nir_to_llvm.c | 3 ++-
src/amd/
For the series:
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Thu, Nov 24, 2016 at 12:40 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> This is just a cleanup before I rework this code to fix mirrored
> blits.
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Thu, Nov 24, 2016 at 4:12 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> The code didn't limit the offsets to the number supplied, so
> if we expected 3 but only got
On Tue, Nov 22, 2016 at 2:31 PM, Emil Velikov <emil.l.veli...@gmail.com> wrote:
> On 22 November 2016 at 01:19, Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
> wrote:
>> We want to also invalidate the cache when LLVM gets changed. As the
>> specific LLVM revision
Not needed anymore.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/Makefile.am | 6 +-
src/amd/vulkan/Makefile.sources | 3 +--
2 files changed, 2 insertions(+), 7 deletions(-)
diff --git a/src/amd/vulkan/Makefile.am b/src/amd/vulkan/Makefile.am
index c
No sense in repeatedly determining it. Also, it might be dependent
on the device as shaders get compiled differently for SI/CIK/VI etc.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_device.c | 18 ++
src/amd/vulkan/radv_pipeline_cache.
Invalidates the cache when someone switches cards.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_device.c | 8 +---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c
index 04c0bdc..8
symbol is
actually the LLVM DSO.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_device.c | 36
1 file changed, 32 insertions(+), 4 deletions(-)
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c
index b
their system
>
> On Tue, 22 Nov 2016 at 01:20 Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
> wrote:
>>
>> Invalidates the cache when someone switches cards.
>>
>> Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
>> ---
>> src/amd/vulkan/radv_d
Seems like libdrm_amdgpu does a dup indeed. Patches 7-8 are
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Thu, Nov 24, 2016 at 9:30 PM, Emil Velikov <emil.l.veli...@gmail.com> wrote:
> From: Emil Velikov <emil.veli...@collabora.com>
>
> radv_amdgpu_wi
On Mon, Nov 28, 2016 at 5:19 AM, Dave Airlie wrote:
> From: Dave Airlie
>
> This just moves some common code into a utility function
> to avoid having to change multiple places later.
>
> Signed-off-by: Dave Airlie
> ---
>
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/common/ac_nir_to_llvm.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index c9b0106..d9eef57 100644
--- a/src/amd/common/ac_nir_to_llvm.c
Meta was using it before it was set. I suspect we typically don't
want to dump meta shaders, so just set it to false in the beginning.
Signed-off-by: Bas Nieuwenhuizen <ba...@google.com>
---
src/amd/vulkan/radv_device.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/amd/
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Thu, Nov 24, 2016 at 7:14 PM, Emil Velikov <emil.l.veli...@gmail.com> wrote:
> From: Emil Velikov <emil.veli...@collabora.com>
>
> Cap up-to the number of properties available while copying the data.
You might want to rename the title of this patch.
- Bas
On Thu, Nov 24, 2016 at 7:18 PM, Emil Velikov wrote:
> From: Emil Velikov
>
> Signed-off-by: Emil Velikov
> ---
> src/intel/isl/isl_priv.h | 8 +++-
>
Making this thread safe is probably overkill I guess?
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Thu, Nov 24, 2016 at 7:18 PM, Emil Velikov <emil.l.veli...@gmail.com> wrote:
> From: Emil Velikov <emil.veli...@collabora.com>
>
> Signed-off-
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Mon, Nov 28, 2016 at 8:04 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> This fixes:
> dEQP-VK.api.image_clearing.clear_color_image.3d*
>
> These were hitting
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Nov 22, 2016 at 5:44 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> I've started working on a shader-db alike for Vulkan,
> it's based on vktrace and it
For the series:
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Tue, Nov 22, 2016 at 7:34 AM, Dave Airlie <airl...@gmail.com> wrote:
> From: Dave Airlie <airl...@redhat.com>
>
> For the hw resolve there is no need to emit any sort
> of texture
Hi,
One possible reason is that the radv driver can't find any hardware.
This can happen for example if you use the radeon kernel driver, as
only the amdgpu kernel driver is currently supported.
I'm not sure what the error will be if the loader doesn't see radv at all.
Yours sincerely,
Bas
Maybe also add an assert in ac_add_function_attr that the attr is a
single flag? Something like
assert(attr && util_is_power_of_two(attr));
Either way, this patch is
Reviewed-by: Bas Nieuwenhuizen <b...@basnieuwenhuizen.nl>
On Fri, Nov 11, 2016 at 9:36 PM, Daniel Scharrer <dan
401 - 500 of 2350 matches
Mail list logo