This series is dedicated to improving the Midgard compiler, specifically
regarding the sections about indirect indexing and looping constructs.
In the process, a number of orthogonal bugs/mysteries touched by these
tests were fixed. The net result is that we are now passing all of:
Signed-off-by: Alyssa Rosenzweig
---
.../panfrost/midgard/midgard_compile.c| 38 ++-
1 file changed, 28 insertions(+), 10 deletions(-)
diff --git a/src/gallium/drivers/panfrost/midgard/midgard_compile.c
b/src/gallium/drivers/panfrost/midgard/midgard_compile.c
index
The hardware needs this lowered anyway; for now, might as well use
mesa's default lowering for pure conformance reasons.
Signed-off-by: Alyssa Rosenzweig
---
src/gallium/drivers/panfrost/midgard/midgard_compile.c | 8 ++--
src/gallium/drivers/panfrost/pan_screen.c | 3 ++-
2
Varying arrays sometimes are lowered to a series of directly accessed
varyings (which we handled okay), but when indirectly accessed, they
appear as a single array; we need to handle this as well.
Signed-off-by: Alyssa Rosenzweig
---
src/gallium/drivers/panfrost/midgard/midgard_compile.c | 6
This restriction makes sense logically. Not sure why it wasn't obeyed
before. In conjunction with previous commit's disclaimer, fixes
dEQP-GLES2.functional.shaders.loop.for_dynamic_iterations.*
Signed-off-by: Alyssa Rosenzweig
---
src/gallium/drivers/panfrost/midgard/helpers.h | 2 +-
1 file
Fixes a bunch of non-vec4 indexing.varying_array tests.
Signed-off-by: Alyssa Rosenzweig
---
.../panfrost/midgard/midgard_compile.c| 29 ++-
1 file changed, 22 insertions(+), 7 deletions(-)
diff --git a/src/gallium/drivers/panfrost/midgard/midgard_compile.c
Signed-off-by: Alyssa Rosenzweig
---
src/gallium/drivers/panfrost/midgard/midgard_compile.c | 7 ---
1 file changed, 7 deletions(-)
diff --git a/src/gallium/drivers/panfrost/midgard/midgard_compile.c
b/src/gallium/drivers/panfrost/midgard/midgard_compile.c
index fcc17a5a092..539f8ca12bb
Signed-off-by: Alyssa Rosenzweig
---
src/gallium/drivers/panfrost/midgard/midgard_compile.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/src/gallium/drivers/panfrost/midgard/midgard_compile.c
b/src/gallium/drivers/panfrost/midgard/midgard_compile.c
index 1be96bc2eb9..f48e873a3f1 100644
For floating point ops, these bits determine the "negate?" and "abs?"
modifiers. For integer ops, it turns out they control how sign/zero
extension work, useful for mixing types.
Signed-off-by: Alyssa Rosenzweig
---
.../drivers/panfrost/midgard/disassemble.c| 43 +++
Most copy prop should occur at the NIR level, but we generate a fair
number of moves implicitly ourselves, etc... long story short, it's a
net win to also do simple copy prop + DCE on the MIR. As a bonus, this
fixes the weird imov precision bug once and for good, I think.
Signed-off-by: Alyssa
Along with a corresponding fix to the move elimination pass (not
included here yet -- I just have it disabled for now), this will fix
dEQP-GLES2.functional.shaders.loops.for_uniform_iterations.*
Signed-off-by: Alyssa Rosenzweig
---
src/gallium/drivers/panfrost/midgard/midgard_compile.c | 3 +++
We reshuffle the existing "dead move elimination" pass into a generic
dead code elimination layer, fixing bugs incurred with looping in the
process.
Signed-off-by: Alyssa Rosenzweig
---
.../panfrost/midgard/midgard_compile.c| 25 ---
1 file changed, 10 insertions(+), 15
This fixes nested loops.
Signed-off-by: Alyssa Rosenzweig
---
.../drivers/panfrost/midgard/midgard_compile.c | 17 ++---
1 file changed, 10 insertions(+), 7 deletions(-)
diff --git a/src/gallium/drivers/panfrost/midgard/midgard_compile.c
Signed-off-by: Alyssa Rosenzweig
---
.../drivers/panfrost/midgard/midgard_compile.c | 16 +++-
1 file changed, 11 insertions(+), 5 deletions(-)
diff --git a/src/gallium/drivers/panfrost/midgard/midgard_compile.c
b/src/gallium/drivers/panfrost/midgard/midgard_compile.c
index
This adds preliminary support for indirect loads of varying arrays and
uniform arrays, bringing a few new tests in shader.indexing.* to
passing, although there remains a number of cases still missing.
Signed-off-by: Alyssa Rosenzweig
---
.../panfrost/midgard/midgard_compile.c| 90
In the future, we might want to switch to a table-based approach, but
for now, at least have it current.
Signed-off-by: Alyssa Rosenzweig
---
.../drivers/panfrost/midgard/helpers.h| 31 +--
1 file changed, 21 insertions(+), 10 deletions(-)
diff --git
The semantics of this field are not well understood; it is better to
print it unconditionally along with the other unknown state, rather than
silently eat the value. Without this change, some critical state was
being lost in some shaders (notably, the offset for load/store
scratchpad intructions
The bug this worked around is no longer applicable, it seems -- remove
the hack that breaks more than it fixes.
Signed-off-by: Alyssa Rosenzweig
---
src/gallium/drivers/panfrost/midgard/midgard_compile.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
Fixes a bunch of issues with looping. Honestly, I'm not sure why loops
worked at all before.
Signed-off-by: Alyssa Rosenzweig
---
.../panfrost/midgard/midgard_compile.c| 83 +++
1 file changed, 69 insertions(+), 14 deletions(-)
diff --git
This reduces register pressure substantially since we get smaller
liveness ranges.
Signed-off-by: Alyssa Rosenzweig
---
src/gallium/drivers/panfrost/midgard/midgard_compile.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git
https://bugs.freedesktop.org/show_bug.cgi?id=110486
Bug ID: 110486
Summary: Power management
Product: Mesa
Version: unspecified
Hardware: Other
OS: All
Status: NEW
Severity: normal
Priority:
https://bugs.freedesktop.org/show_bug.cgi?id=110486
--- Comment #1 from Mafia ---
Created attachment 144065
--> https://bugs.freedesktop.org/attachment.cgi?id=144065=edit
Power management
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the
https://bugs.freedesktop.org/show_bug.cgi?id=110485
Mafia changed:
What|Removed |Added
Attachment #144064|â¿à¹r.à¶rAÆâ â¿.htm |â¿à¹r.à¶rAÆâ â¿.htm
filename|
https://bugs.freedesktop.org/show_bug.cgi?id=110485
--- Comment #2 from Mafia ---
Created attachment 144064
--> https://bugs.freedesktop.org/attachment.cgi?id=144064=edit
Power management
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the
https://bugs.freedesktop.org/show_bug.cgi?id=110485
--- Comment #1 from Mafia ---
Created attachment 144063
--> https://bugs.freedesktop.org/attachment.cgi?id=144063=edit
Power management
--
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the
https://bugs.freedesktop.org/show_bug.cgi?id=110485
Bug ID: 110485
Summary: Power management
Product: Mesa
Version: unspecified
Hardware: Other
OS: All
Status: NEW
Severity: normal
Priority:
https://bugs.freedesktop.org/show_bug.cgi?id=110484
Bug ID: 110484
Summary: Power management
Product: Mesa
Version: unspecified
Hardware: Other
OS: All
Status: NEW
Severity: normal
Priority:
Hooray! Excited to see this in tree.
Reviewed-by: Alyssa Rosenzweig
I'll push this when I get a chance to sanity check (probably tomorrow)
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
https://bugs.freedesktop.org/show_bug.cgi?id=110480
Bas Nieuwenhuizen changed:
What|Removed |Added
Resolution|--- |NOTABUG
Status|NEW
From page 76 (page 80 of the PDF) of the GLSL 4.60 v.5 spec:
" No aliasing in output buffers is allowed: It is a compile-time or
link-time error to specify variables with overlapping transform
feedback offsets."
Currently, this is expected to fail, but it succeeds:
"
...
https://bugs.freedesktop.org/show_bug.cgi?id=110480
--- Comment #5 from Alexander Tsoy ---
(In reply to Bas Nieuwenhuizen from comment #4)
> 1) which vkcube did you use and where did you get it from?
It's vkcube from vulkan-tools:
https://github.com/KhronosGroup/Vulkan-Tools
--
You are
https://bugs.freedesktop.org/show_bug.cgi?id=110480
--- Comment #4 from Bas Nieuwenhuizen ---
Created attachment 144058
--> https://bugs.freedesktop.org/attachment.cgi?id=144058=edit
vkcube radv vs. amdvlk side by side
See the attachment, for me the colors are equal.
Can I ask
1) which
https://bugs.freedesktop.org/show_bug.cgi?id=110480
--- Comment #3 from osch...@web.de ---
Created attachment 144057
--> https://bugs.freedesktop.org/attachment.cgi?id=144057=edit
nvidia
--
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the
https://bugs.freedesktop.org/show_bug.cgi?id=110480
--- Comment #2 from osch...@web.de ---
Created attachment 144056
--> https://bugs.freedesktop.org/attachment.cgi?id=144056=edit
amdvlk
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the
https://bugs.freedesktop.org/show_bug.cgi?id=110480
--- Comment #1 from osch...@web.de ---
Created attachment 144055
--> https://bugs.freedesktop.org/attachment.cgi?id=144055=edit
radv
--
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the
https://bugs.freedesktop.org/show_bug.cgi?id=110480
Bug ID: 110480
Summary: vkcube renders too brightly on radv
Product: Mesa
Version: unspecified
Hardware: Other
OS: All
Status: NEW
Severity: normal
https://bugs.freedesktop.org/show_bug.cgi?id=110476
Danylo changed:
What|Removed |Added
CC||danylo.pilia...@gmail.com
--- Comment #3 from
https://bugs.freedesktop.org/show_bug.cgi?id=110476
--- Comment #2 from Samuel Pitoiset ---
LLVM r356956 introduced the regression. I reported the problem to the compiler
team. I will let you know when it's fixed.
--
You are receiving this mail because:
You are the assignee for the bug.
You
Build mesa 10830 failed
Commit 662ec3e8c4 by Alyssa Rosenzweig on 4/21/2019 7:13 PM:
panfrost/midgard: Remove unused mir_next_block\n\nSigned-off-by: Alyssa Rosenzweig
Configure your notification preferences
https://bugs.freedesktop.org/show_bug.cgi?id=110476
--- Comment #1 from Samuel Pitoiset ---
I can reproduce the issue, I will investigate.
Thanks for the report!
--
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the
https://bugs.freedesktop.org/show_bug.cgi?id=110479
Bug ID: 110479
Summary: AddressSanitizer:DEADLYSIGNAL in
src/egl/drivers/dri2/platform_surfaceless.c:139
Product: Mesa
Version: 19.0
Hardware: Other
OS:
41 matches
Mail list logo