All the -Wunused-but-set-variable ones.
Found a way to do it with a oneliner.
Signed-off-by: Grazvydas Ignotas
---
no commit access (currently pending)
src/amd/addrlib/core/addrcommon.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amd/addrlib/core/addrcommon.h
b/src
metadata->data)
> + free(metadata->data);
>
Please add a space after 'if' to match the style of all other code or drop
the NULL check altogether as free(NULL) is fine. With that:
Reviewed-by: Grazvydas Ignotas
for both patches.
some scene of my choice:
before: 32fps
after: 36fps
GL: 38fps
Tested-by: Grazvydas Ignotas
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
On Fri, Apr 14, 2017 at 12:08 AM, Eric Anholt wrote:
> I wrote this code with reference to pixman, though I've only decided to
> cover Linux (what I'm testing) and Android (seems obvious enough). Linux
> has getauxval() as a cleaner interface to the /proc entry, but it's more
> glibc-specific an
A leftover from anv.
Signed-off-by: Grazvydas Ignotas
---
src/amd/vulkan/radv_device.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c
index 5f14394..7857e8f 100644
--- a/src/amd/vulkan/radv_device.c
+++ b/src/amd
The kernel returns frequency in kHz, so to convert to nanosecond
interval that Vulkan uses the dividend should be 100.0 and not
10.0.
This fixes the GPU graph in DOOM and matches the amdgpu-pro blob.
Signed-off-by: Grazvydas Ignotas
Fixes: f4e499ec791 "radv: add initial non-confo
ned-off-by: Grazvydas Ignotas
---
src/amd/vulkan/radv_device.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c
index 12040a0..dd401f4 100644
--- a/src/amd/vulkan/radv_device.c
+++ b/src/amd/vulkan/radv_device.c
@@ -650,
On Fri, Apr 14, 2017 at 8:47 PM, Eric Anholt wrote:
> NEON is sufficiently different on arm64 that we can't just reuse this
> code. Disable it on arm64 for now.
>
> Signed-off-by: Eric Anholt
> ---
> src/gallium/drivers/vc4/vc4_tiling_lt.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletio
y (change by robher). Use
> O_CLOEXEC. Break out of the loop when we find our feature. Only do
> NEON detection, until someone actually wants VFP features.
>
Reviewed-by: Grazvydas Ignotas
___
mesa-dev mailing list
mesa-dev@lists.fre
I haven't tried it, but maybe calling LLVMParseCommandLineOptions()
second time would work (after creating the target machine, through
call_once) to avoid all this code duplication?
Gražvydas
On Fri, Jun 8, 2018 at 12:38 PM, Samuel Pitoiset
wrote:
> Workaround for bug in llvm that causes the GPU
On Tue, Jun 12, 2018 at 8:07 AM, Timothy Arceri wrote:
> From: Timothy Arceri
localhost.localdomain?
>
> This avoids hitting asserts in anv.
> ---
> run.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/run.c b/run.c
> index eafbd7d..a962887 100644
> --- a/run.c
> +++ b/run.c
> @@ -
Avoids a branch and reduces code size a tiny bit:
text data bss dechex filename
10804563 398653 2070368 13273584 ca89f0 /tmp/radeonsi_dri.so.old
10804499 398653 2070368 13273520 ca89b0 /tmp/radeonsi_dri.so
---
src/gallium/drivers/radeonsi/si_descriptors.c | 4 ++--
1 file change
This allows to avoid having to see garbage in Dying Light loading screen
at least, which probably expects Windows/NV behavior of all allocations
being zeroed by default.
Analogous to radv flag with the same name.
---
src/gallium/drivers/radeonsi/si_pipe.c| 1 +
src/gallium/drivers/radeons
On Wed, Jul 4, 2018 at 9:13 AM, Jason Ekstrand wrote:
> Many fragment shaders do a discard using relatively little information
> but still put the discard fairly far down in the shader for no good
> reason. If the discard is moved higher up, we can possibly avoid doing
> some or almost all of the
On Fri, Apr 20, 2018 at 3:21 PM, Samuel Pitoiset
wrote:
> The SI family doesn't support chaining which means the maximum
> size in dwords per CS is limited. When that limit was reached
> we failed to submit the CS and the application crashed.
>
> This patch allows to submit up to 4 IBs which is cu
On Wed, May 2, 2018 at 1:27 PM, Timothy Arceri wrote:
> ---
> src/mesa/drivers/dri/common/dri_util.c | 9 +
> 1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/src/mesa/drivers/dri/common/dri_util.c
> b/src/mesa/drivers/dri/common/dri_util.c
> index 7cb6248b130..d72f72d075
On Wed, May 2, 2018 at 3:38 PM, Timothy Arceri wrote:
> On 02/05/18 22:36, Timothy Arceri wrote:
>>
>> On 02/05/18 21:48, Grazvydas Ignotas wrote:
>>>
>>> On Wed, May 2, 2018 at 1:27 PM, Timothy Arceri
>>> wrote:
>>>>
>>>> ---
&g
Trivial.
---
src/amd/vulkan/radv_cmd_buffer.c | 4 ++--
src/amd/vulkan/radv_descriptor_set.c | 2 +-
src/amd/vulkan/radv_device.c | 6 +++---
src/amd/vulkan/radv_entrypoints_gen.py | 2 +-
src/amd/vulkan/radv_image.c| 4 ++--
src/amd/vulkan/radv_nir_to_llvm.c | 4
On Wed, May 16, 2018 at 5:02 PM, Christian Gmeiner
wrote:
> Currently GLSL 1.4 is defined for all gallium drivers even only
> GLSL 1.2 is supported as seen on etnaviv.
>
> v1 -> v2:
> - use _min(..) as suggested by Lucas Stach and Michel Dänzer
>
> Fixes 4560aad780b ("mesa: add GLSLVersionCompat
On Thu, May 17, 2018 at 4:47 AM, Marek Olšák wrote:
> From: Marek Olšák
>
> same as Vulkan.
Ambiguous. Did you mean amdvlk?
Gražvydas
> ---
> src/gallium/drivers/radeonsi/si_blit.c | 10 ++
> src/gallium/drivers/radeonsi/si_pipe.h | 2 +-
> src/gallium/drivers/radeo
The bsr instruction modifies flags, so that needs to be indicated to the
compiler. No effect on generated code, but still needed for correctness.
---
src/gallium/drivers/llvmpipe/lp_setup_tri.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/llvmpipe/lp_se
Thanks to reproducible builds, binary file timestamps may be identical
for both 32bit and 64bit packages when built from the same source.
This means radv will use the same cache for both 32 and 64 bit
processes, which leads to crashes.
Conveniently there is a spare byte in cache_uuid, let's place
Currently if 64bit and 32bit programs are used interchangeably, radv
will keep overwriting the cache. Use separate cache files to avoid
that.
---
src/amd/vulkan/radv_meta.c | 16 +++-
1 file changed, 7 insertions(+), 9 deletions(-)
diff --git a/src/amd/vulkan/radv_meta.c b/src/amd/vul
For 32bit build. Trivial.
---
src/vulkan/wsi/wsi_common_display.c | 4 ++--
src/vulkan/wsi/wsi_common_x11.c | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/vulkan/wsi/wsi_common_display.c
b/src/vulkan/wsi/wsi_common_display.c
index e6cba188dfa..1e90bba460c 100644
--
On Tue, Aug 21, 2018 at 1:05 AM, Bas Nieuwenhuizen
wrote:
> Reviewed-by: Bas Nieuwenhuizen
>
> Did you have access to push?
Yeah, just pushed everything.
Gražvydas
>
> On Mon, Aug 20, 2018 at 11:40 PM, Grazvydas Ignotas wrote:
>> For 32bit build. Trivial.
>
On Thu, Aug 23, 2018 at 6:13 PM, Jason Ekstrand wrote:
> ---
> src/vulkan/util/vk_alloc.h | 17 +
> 1 file changed, 17 insertions(+)
>
> diff --git a/src/vulkan/util/vk_alloc.h b/src/vulkan/util/vk_alloc.h
> index f58a80625a8..bc2bb7509e4 100644
> --- a/src/vulkan/util/vk_alloc.h
Seems to fix dxvk, nice! Also tested DOOM which still works.
Tested-by: Grazvydas Ignotas
On Thu, Feb 22, 2018 at 6:57 PM, James Legg wrote:
> When transitioning to an htile compressed depth format, Set the full
> depth range, so later rasterization can pass HiZ. Previously, for depth
It seems to be a leftover from u_format_table.py.
---
src/amd/vulkan/vk_format_table.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amd/vulkan/vk_format_table.py
b/src/amd/vulkan/vk_format_table.py
index 36352b1..c33a4ce 100644
--- a/src/amd/vulkan/vk_format_table.py
+
No need to bother the linker about them.
---
src/amd/vulkan/vk_format_table.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amd/vulkan/vk_format_table.py
b/src/amd/vulkan/vk_format_table.py
index c33a4ce..cd1af62 100644
--- a/src/amd/vulkan/vk_format_table.py
+++ b/src/
'cleanup' path is dereferencing 'svga' a lot, 'done' is a better choice.
Found by Coccinelle.
Signed-off-by: Grazvydas Ignotas
---
src/gallium/drivers/svga/svga_context.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/drivers/svga
.mob_fenced was already checked, probably a copy-paste bug.
Found by Coccinelle.
Signed-off-by: Grazvydas Ignotas
---
src/gallium/winsys/svga/drm/vmw_screen_pools.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/winsys/svga/drm/vmw_screen_pools.c
b/src/gallium
Looks like checking both sources was intended, instead of the first one
twice. Found with Coccinelle, coccinellery/xand/xand.cocci semantic patch.
Signed-off-by: Grazvydas Ignotas
---
src/intel/vulkan/anv_nir_apply_pipeline_layout.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
At least with vim, this is needed to actually get tab instead of
3 spaces after hitting the tab key.
---
src/gallium/drivers/radeonsi/.editorconfig | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/gallium/drivers/radeonsi/.editorconfig
b/src/gallium/drivers/radeonsi/.editorconfig
index cc8
Fix the following:
warning: format ‘%lx’ expects argument of type ‘long unsigned int’, but
argument 3 has type ‘uint64_t {aka long long unsigned int}.
---
src/intel/compiler/brw_disasm.c | 4 ++--
src/intel/compiler/brw_nir_analyze_ubo_ranges.c | 2 +-
2 files changed, 3 insertions
I hope the actual dropping of MSB is ok, but that's what's already
happened before this change.
---
src/gallium/drivers/radeonsi/si_descriptors.c | 18 --
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/src/gallium/drivers/radeonsi/si_descriptors.c
b/src/gallium/dri
It seems these were missed when struct pipe_context * argument was
added to hud_graph::query_new_value.
Fixes: 3132afdf4c "gallium/hud: pass pipe_context explicitly to most functions"
---
src/gallium/auxiliary/hud/hud_cpufreq.c | 2 +-
src/gallium/auxiliary/hud/hud_diskstat.c | 2 +-
src
On Tue, Feb 6, 2018 at 1:37 PM, Eric Engestrom
wrote:
> On Sunday, 2018-02-04 00:19:33 +0200, Grazvydas Ignotas wrote:
>> At least with vim, this is needed to actually get tab instead of
>> 3 spaces after hitting the tab key.
>
> Are you sure?
For my copy of vim (7.4.168
Is this really sufficient? From what I read in the manpage, you can
get an incomplete read instead when a signal arrives.
Gražvydas
On Fri, Feb 9, 2018 at 12:03 PM, Lionel Landwerlin
wrote:
> Fixes: 458468c136e "i965: Expose OA counters via INTEL_performance_query"
> Signed-off-by: Lionel Landwe
ping
On Sun, Feb 4, 2018 at 12:19 AM, Grazvydas Ignotas wrote:
> Fix the following:
> warning: format ‘%lx’ expects argument of type ‘long unsigned int’, but
> argument 3 has type ‘uint64_t {aka long long unsigned int}.
> ---
> src/intel/compiler/brw_disasm.c
On Wed, Mar 1, 2017 at 7:25 AM, Timothy Arceri wrote:
> ---
> src/util/disk_cache.c | 37 ++---
> 1 file changed, 34 insertions(+), 3 deletions(-)
>
> diff --git a/src/util/disk_cache.c b/src/util/disk_cache.c
> index f5e1145..2a0edca 100644
> --- a/src/util/disk_c
On Wed, Mar 1, 2017 at 7:25 AM, Timothy Arceri wrote:
> ...
> diff --git a/src/util/disk_cache.c b/src/util/disk_cache.c
> index 2a0edca..03aae02 100644
> --- a/src/util/disk_cache.c
> +++ b/src/util/disk_cache.c
> @@ -30,20 +30,21 @@
> #include
> #include
> #include
> #include
> #include
On Thu, Mar 2, 2017 at 8:36 AM, Timothy Arceri wrote:
>
>
> On 02/03/17 17:17, Ilia Mirkin wrote:
>>
>> On Thu, Mar 2, 2017 at 1:08 AM, Timothy Arceri
>> wrote:
>>>
>>> Previously we were deleting the entire cache if a user switched
>>> between 32 and 64 bit applications.
>>> ---
>>> src/util/di
regressions elsewhere because the only
ureg_get_tokens() user that cares about token count is the shader cache.
Cc: Timothy Arceri
Signed-off-by: Grazvydas Ignotas
---
no commit access
src/gallium/auxiliary/tgsi/tgsi_ureg.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src
Since blob is intended for serializing data, it's not a good idea to
leave padding holes with uninitialized data, which may leak heap
contents and hurt compression if the blob is later compressed, like
done by shader cache. Clear it.
Signed-off-by: Grazvydas Ignotas
---
I don't kno
On Fri, Mar 3, 2017 at 5:27 AM, Timothy Arceri wrote:
> Previously we were deleting the entire cache if a user switched
> between 32 and 64 bit applications.
>
> V2: make the check more generic, it should now work with any
> platform we are likely to support.
> ---
> src/util/disk_cache.c | 19 ++
lways specify a suitable alignment factor.
Cc: Jonas Pfeil
Fixes: cd2b55e5 "ralloc: Make sure ralloc() allocations match malloc()'s
alignment."
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100049
Signed-off-by: Grazvydas Ignotas
---
no commit access
src/util/ralloc.c | 4
On Sat, Mar 4, 2017 at 4:44 PM, Emil Velikov wrote:
> On 3 March 2017 at 23:37, Timothy Arceri wrote:
>>
>> I'm tempted to just push this:
>> https://patchwork.freedesktop.org/patch/141891/
>>
>> And worry about issues later if that's not good enough.
>>
> Can we use anything like the following a
ggestion from Emil to make even more generic/fix issue
> with __ILP32__ not being declared on gcc for regular 32-bit builds.
Now it works, launching steam (32bit) no longer wipes all cache from
64bit games, everything works as expected.
Tested-by: Grazv
ly by multiple threads that
somehow manage to change the environment variable without affecting each
other, so just provide a thread-unsafe fallback.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93089
Signed-off-by: Grazvydas Ignotas
---
no commit access
src/intel/common/gen_debug.
zilla: https://bugs.freedesktop.org/show_bug.cgi?id=93089
Signed-off-by: Grazvydas Ignotas
---
no commit access
src/util/disk_cache.c | 17 -
1 file changed, 16 insertions(+), 1 deletion(-)
diff --git a/src/util/disk_cache.c b/src/util/disk_cache.c
index 4a8ee21..daa5c7b 100644
---
No functional changes.
Signed-off-by: Grazvydas Ignotas
---
no commit access
src/util/disk_cache.c | 35 ++-
src/util/disk_cache.h | 20 ++--
2 files changed, 28 insertions(+), 27 deletions(-)
diff --git a/src/util/disk_cache.c b/src/util
Use the same helpers as for other handle<->pointer conversions.
Signed-off-by: Grazvydas Ignotas
---
src/amd/vulkan/radv_device.c | 5 ++---
src/amd/vulkan/radv_private.h | 3 +++
2 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/
Please push if these are ok.
Grazvydas Ignotas (4):
radv: avoid casting warnings on 32bit
radv: don't use uninitialized value on failure
radv: check for upload alloc failure
radv: use VK_NULL_HANDLE for handles
src/amd/vulkan/radv_cmd_buffer.c | 7 ---
src/amd/v
Mainly to avoid gcc's complains about uninitialized ptr and offset use
later in that code.
Signed-off-by: Grazvydas Ignotas
---
src/amd/vulkan/radv_cmd_buffer.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/src/amd/vulkan/radv_cmd_buffer.c b/src/amd/v
Mainly to avoid a warning.
Signed-off-by: Grazvydas Ignotas
---
src/amd/vulkan/radv_descriptor_set.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amd/vulkan/radv_descriptor_set.c
b/src/amd/vulkan/radv_descriptor_set.c
index ca5b280..7dade3e 100644
--- a/src/amd
Avoids warnings on 32bit.
Signed-off-by: Grazvydas Ignotas
---
src/amd/vulkan/radv_meta_bufimage.c | 8
src/amd/vulkan/radv_wsi.c | 2 +-
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/src/amd/vulkan/radv_meta_bufimage.c
b/src/amd/vulkan/radv_meta_bufimage.c
e, as suggested by Matt Turner
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93089
Signed-off-by: Grazvydas Ignotas
---
no commit access
src/intel/common/gen_debug.c | 14 +++---
1 file changed, 11 insertions(+), 3 deletions(-)
diff --git a/src/intel/common/gen_debug.c b/src/
On Sun, Mar 5, 2017 at 10:29 PM, Matt Turner wrote:
> On Sun, Mar 5, 2017 at 11:14 AM, Grazvydas Ignotas wrote:
>> There are still some distributions trying to support unfortunate people
>> with old or exotic CPUs that don't have 64bit atomic operations. Missing
>&g
On Mon, Mar 6, 2017 at 12:05 AM, Bas Nieuwenhuizen
wrote:
> On Sun, Mar 5, 2017 at 10:04 PM, Grazvydas Ignotas wrote:
>> Mainly to avoid gcc's complains about uninitialized ptr and offset use
>> later in that code.
>>
>> Signed-off-by: Grazvydas Ig
;
> -
> - count++;
> + if (fstatat(dirfd(dir), entry->d_name, &sb, 0) == 0) {
> + if (!lru_atime || (sb.st_atime < lru_atime)) {
> +tmp = realloc(lru_name, strlen(entry->d_name) + 1);
> +if (tmp) {
> + lru_name = tmp;
> +
This seems to break "make check", so needs to update tests I guess.
Gražvydas
On Mon, Mar 6, 2017 at 6:17 PM, Alan Swanson wrote:
> Currently only a one in one out eviction so if at max_size and
> cache files were to constantly increase in size then so would the
> cache. Restrict to limit of 8 e
Makes vsync option work in The Talos Principle.
Signed-off-by: Grazvydas Ignotas
---
no commit access
perhaps this is stable worthy?
src/glx/glxcmds.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/glx/glxcmds.c b/src/glx/glxcmds.c
index 53c9f9c..7be4e73 100644
--- a
On Wed, Mar 8, 2017 at 4:37 AM, Ian Romanick wrote:
> On 03/07/2017 05:09 PM, Grazvydas Ignotas wrote:
>> Makes vsync option work in The Talos Principle.
>
> NAK. The GLX_SGI_swap_control spec says:
>
> glXSwapIntervalSGI returns GLX_BAD_VALUE if parameter is
>
(CCing some people who touched or reviewed blob)
Ping. I think writing stale heap contents to cache should be avoided
and this patch is an improvement.
Gražvydas
On Fri, Mar 3, 2017 at 1:59 AM, Grazvydas Ignotas wrote:
> Since blob is intended for serializing data, it's not a good
Negating size_t on 32bit produces a 32bit result. This was effectively
adding values close to UINT_MAX to the cache size (the files are usually
small) instead of intended subtraction.
Fixes 'make check' disk_cache failures on 32bit.
Signed-off-by: Grazvydas Ignotas
---
src/util/disk_c
It incorrectly doubles the size on each iteration.
Fixes: 85a9b1b5 "util/disk_cache: compress individual cache entries"
Signed-off-by: Grazvydas Ignotas
---
src/util/disk_cache.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/util/disk_cache.c b/src/util/di
On Fri, Mar 10, 2017 at 6:23 AM, Timothy Arceri wrote:
> On 07/03/17 12:25, Alan Swanson wrote:
>>
>> Still using random selection of two-character subdirectory in which
>> to check cache files rather than scanning entire cache.
>>
>> v2: Factor out double strlen call
>> ---
>> src/util/disk_cach
On Fri, Mar 10, 2017 at 4:28 AM, Timothy Arceri wrote:
> ---
> src/compiler/glsl/shader_cache.cpp | 5 -
> src/compiler/glsl/tests/cache_test.c| 29
> -
> src/gallium/drivers/radeonsi/si_state_shaders.c | 9 ++--
> src/mesa/state_tracker
On Sat, Mar 11, 2017 at 9:04 AM, Jason Ekstrand wrote:
> During initial CCS bring-up, I discovered that you have to do a full CS
> stall prior to doing a CCS resolve as well as afterwards. It appears
> that the same is needed for fast-clears as well. This fixes rendering
> corruptions on The Tal
It seems this has not removed enough and introduced a warning:
intel_batchbuffer.c: In function ‘intel_batchbuffer_require_space’:
intel_batchbuffer.c:132:22: warning: unused variable ‘prev_ring’
[-Wunused-variable]
enum brw_gpu_ring prev_ring = brw->batch.ring;
Gražvydas
On Wed, Mar 8, 2017
hash it in too, if it's how others want it to be.
This series should also solve bug 100091.
Grazvydas Ignotas (7):
util/mesa-sha1: add a context clone function
util/mesa-sha1: allow to finish without taking the result
util/disk_cache: use a helper to compute cache keys
util/disk_
Useful if we just want to free the context, in error paths for example.
Signed-off-by: Grazvydas Ignotas
---
src/util/mesa-sha1.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/util/mesa-sha1.c b/src/util/mesa-sha1.c
index 2c47465..66be707 100644
--- a/src/util/mesa
This will allow to hash additional data into the cache keys or even
change the hashing algorithm easily, should we decide to do so.
Signed-off-by: Grazvydas Ignotas
---
src/compiler/glsl/glsl_parser_extras.cpp| 2 +-
src/compiler/glsl/shader_cache.cpp | 2 +-
src/compiler
To be used by subsequent patches, for now it's just the initial
sha1 state.
Signed-off-by: Grazvydas Ignotas
---
src/util/disk_cache.c | 23 ---
1 file changed, 20 insertions(+), 3 deletions(-)
diff --git a/src/util/disk_cache.c b/src/util/disk_cache.c
index 79
There is no need to hardcode it, we can just use blob_key[0].
This is needed because the next patches are going to change how cache
keys are computed.
Signed-off-by: Grazvydas Ignotas
---
src/compiler/glsl/tests/cache_test.c | 8 ++--
1 file changed, 2 insertions(+), 6 deletions(-)
diff
incarnations of disk_cache at the same time, where one
deletes a directory belonging to the other, like when both OpenGL and
gallium nine are used simultaneously (or several different mesa
installations).
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100091
Signed-off-by: Grazvydas Ignotas
This allows to get rid of the arch directory. It also hashes
sizeof(long) just in case disk_cache ever ends being up used on Windows.
Signed-off-by: Grazvydas Ignotas
---
src/compiler/glsl/tests/cache_test.c | 13 ++---
src/util/disk_cache.c| 15 ++-
src/util
This is useful when we need to compute many hashes which all have some
common data hashed in. It works by first hashing the common data and
keeping the context, then for each hashing operation clone the common
context and continue hashing from there.
Signed-off-by: Grazvydas Ignotas
---
src
FWIW:
Reviewed-by: Grazvydas Ignotas
On Sun, Mar 12, 2017 at 5:03 PM, Alan Swanson wrote:
> Since switching to LRU eviction the only user of these predicate
> functions now resolves directory entry stats itself so pass them
> directly saving calling fstat and strlen twice (and the
On Mon, Mar 13, 2017 at 12:20 AM, Timothy Arceri wrote:
> We don't need to hash the pointer size. You can just remove
> get_arch_bitness_str() and fix up the test. Timestamp should be enough on
> its own, even if the 64/32bit builds have the same timestamp that is fine
> because the cache entries
On Mon, Mar 13, 2017 at 12:31 AM, Timothy Arceri wrote:
> On 13/03/17 05:32, Grazvydas Ignotas wrote:
>>
>> This is useful when we need to compute many hashes which all have some
>> common data hashed in. It works by first hashing the common data and
>> keeping the con
On Sat, Mar 11, 2017 at 1:51 PM, Federico Dossena wrote:
> This issue is not new: a guy named Miklós Máté, here in the Mesa mailing
> list somehow managed to fix it in Mesa 11, but his patches do not seem to
> work anymore
> (https://lists.freedesktop.org/archives/mesa-dev/2016-February/106695.htm
ax_size;
>
> + /* A limit of 32 jobs was choosen as observations of Deus Ex start-up
> times
> +* showed that we reached at most 11 jobs on an Intel i5-6400 CPU@2.70GHz
> +* (a fairly modist desktop CPU). 1 thread was choosen because we don't
modest, chosen?
with that, FWIW:
Reviewed-by: Grazvydas Ignotas
Gražvydas
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
On Mon, Mar 13, 2017 at 3:01 AM, Timothy Arceri wrote:
> The following patch will move disk_cache_put() into a thread queue
> so we need to be sure memory is not freed before the thread is
> completed. Here we move responsibility for releasing the memory
> onto the disk cache.
I think this is a f
On Mon, Mar 13, 2017 at 3:01 AM, Timothy Arceri wrote:
> This should help reduce any overhead added by the shader cache
> when programs are not found in the cache.
>
> To avoid creating any special function just for the sake of the
> tests we add a one second delay whenever we call dick_cache_put(
On Mon, Mar 13, 2017 at 3:01 AM, Timothy Arceri wrote:
> ---
> src/util/disk_cache.c | 48
> 1 file changed, 48 insertions(+)
>
> diff --git a/src/util/disk_cache.c b/src/util/disk_cache.c
> index 3b1cffc..160774a 100644
> --- a/src/util/disk_cache
On Mon, Mar 13, 2017 at 1:17 PM, Emil Velikov wrote:
> On 12 March 2017 at 18:32, Grazvydas Ignotas wrote:
>> This is useful when we need to compute many hashes which all have some
>> common data hashed in. It works by first hashing the common data and
>> keeping the c
s an inline function. Maybe
make all 3 inlines instead, so the names appear in debug stack dumps?
In either case:
Reviewed-by: Grazvydas Ignotas
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
On Tue, Mar 14, 2017 at 12:06 AM, Timothy Arceri wrote:
> On 13/03/17 22:55, Grazvydas Ignotas wrote:
>>
>> On Mon, Mar 13, 2017 at 3:01 AM, Timothy Arceri
>> wrote:
>>>
>>> ---
>>> src/util/disk_cache.c | 48
>>> +++
On Tue, Mar 14, 2017 at 12:20 AM, Timothy Arceri wrote:
>
>
> On 13/03/17 22:53, Grazvydas Ignotas wrote:
>>
>> On Mon, Mar 13, 2017 at 3:01 AM, Timothy Arceri
>> wrote:
>>>
>>> This should help reduce any overhead added by the shader cache
On Tue, Mar 14, 2017 at 2:28 AM, Emil Velikov wrote:
> On 13 March 2017 at 22:38, Grazvydas Ignotas wrote:
>> On Mon, Mar 13, 2017 at 6:52 PM, Emil Velikov
>> wrote:
>>> ...
>>> diff --git a/src/util/mesa-sha1.h b/src/util/mesa-sha1.h
>>> index f927d5
or this test, we force this signature to land in the same
> * directory as the original blob first written to the cache.
> */
> _mesa_sha1_compute(one_KB, 1024, one_KB_key);
> one_KB_key[0] = blob_key_byte_zero;
>
> disk_cache_put(cache, one_KB_key, one_K
It's not a good practice to srand() in libraries [1]. Not sure what to
suggest though, maybe rand_r() but "POSIX.1-2008 marks rand_r() as
obsolete". Maybe somebody else (Emil?) has some ideas?
[1]
http://stackoverflow.com/questions/14883035/problems-of-using-srand-in-libraries
Gražvydas
On Tue
- /* Ensure eviction happened by checking that only one of the two
> -* previously-added items can still be fetched.
> + /* Ensure eviction happened by checking that both of the previous
> +* cache itesm were evicted.
items
for patches 2-5:
Reviewed-by: Grazv
uot; 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
---
src/util/di
On Wed, Mar 15, 2017 at 3:01 PM, Alan Swanson wrote:
> Since switching to LRU eviction the only user of these predicate
> functions now resolves directory entry stats itself so pass them
> directly saving calling fstat and strlen twice (and the
> expensive strlen is skipped entirely if access time
On Wed, Mar 15, 2017 at 8:47 PM, Bas Nieuwenhuizen
wrote:
> I thought rand() was not thread safe anyway, and hence we can't use it
> in mesa as we don't know what application threads call it too?
My manpage says "MT-Safe".
>
> On Tue, Mar 14, 2017 at 3:08 AM, Timothy Arceri wrote:
>> Otherwise
This is version 2 of the series that removes extra cache directories
and hashes that information into cache keys instead.
v1: https://lists.freedesktop.org/archives/mesa-dev/2017-March/147632.html
Grazvydas Ignotas (7):
util/disk_cache: use a helper to compute cache keys
tests/cache_test: use
Simplifies the write code a bit and handles EINTR.
Signed-off-by: Grazvydas Ignotas
---
src/util/disk_cache.c | 40 ++--
1 file changed, 26 insertions(+), 14 deletions(-)
diff --git a/src/util/disk_cache.c b/src/util/disk_cache.c
index 904aa66..ad591be
This allows to get rid of the arch and gpu name directories.
Also, gpu_name is now optional.
Signed-off-by: Grazvydas Ignotas
---
src/compiler/glsl/tests/cache_test.c | 13 ++
src/util/disk_cache.c| 47 +++-
src/util/disk_cache.h
201 - 300 of 341 matches
Mail list logo
