Caused by e968b4197b / 67e23b4b29
Since culprit commits, we are not running `MANTA::initGuiding` /
`fluid_alloc_guiding` if guiding is meant to be pulled from other
domains (it does work with effectors though). This is because atm. only
the `FLUID_DOMAIN_ACTIVE_GUIDE` flag sets `mUsingGuiding` to true
(activated in `update_obstacleflags`, so only for effectors).
So to fix this, we have to ensure that `MANTA::initGuiding` runs (also
if guiding is pulled from domains), but also make sure we dont run into
the crashes from T102257.
It seems that the real reason we were getting crashes in T102257 is that
67e23b4b29 made it so that resetting the cache (including a call to
`fluid_modifier_reset_ex`) in `fluid_modifier_processDomain` happens
**after** `FluidDomainSettings.active_fields` have been updated (this
happens in `update_flowsflags` & `update_obstacleflags`).
But `fluid_modifier_reset_ex` also resets
`FluidDomainSettings.active_fields`. If we then update pointers later in
`fluid_modifier_processDomain`, we never get anything in the guiding
related pointers for the new `mCurrentID` (specifically `mPhiGuideIn`
will be a nullptr). This is the real reason `initGuiding()` runs a
second time (it does once for `fluid_modifier_init` then again as a
consequence of the call to `manta_guiding`).
This patch proposes to move the resetting process from 67e23b4b29
**above** the refreshing of the `FluidDomainSettings.active_fields`,
this way these field stay intact, we do get the first run of
`initGuiding()` from `fluid_modifier_init`, manta pointers get updated
with intact fields afterwards (so we then get a valid `mPhiGuideIn`),
which then prevents the second run from `manta_guiding` to actually call
the python script again.
The fix from e968b4197b is then not needed and reverted in this patch.
This should be good for LTS I think.
Pull Request: https://projects.blender.org/blender/blender/pulls/117067
ROCm 6 brings some changes to the HIP API. This pull request is meant to be
backward and forward compatible.
That is Blender could be compiled with either ROCM 6 or 5 and run on either.
The main change is the hipMemoryType enum, which we check based on the
runtime version to use the correct enum values.
Without this, HIP will not work on Windows with upcoming 23.40 driver.
Pull Request: https://projects.blender.org/blender/blender/pulls/116713
Empty hair geometry in Cycles may still report having one curve, even when
there are no actual segments in that curve. This caused an attempt to build
an acceleration structure with zero primitives, which due to other setup
OptiX rejected with an error. Fix that by checking the number of segments
rather than the number of curves in the hair geometry, since the former will
always be zero for empty geometry.
Pull Request: https://projects.blender.org/blender/blender/pulls/115044
The issue was caused by the custom code which was ensuring that closing
About window does not leave application without any key
windows.
Turns out that the windowWillClose is executed when menu is closed,
forcing the custom code to ensure a key window.
The solution is to only perform manual re-ordering if the closing window
was a key window. This both keeps old behavior of ensuring there is a
key window after closing About, and solves the ordering issues when a
window is created or activated via menus.
Pull Request: https://projects.blender.org/blender/blender/pulls/113515
Fixes NaN in Vector Displacement node caused by the normalization of
0, 0, 0 vectors.
This fixes both visual rendering issues and an "illegal address" error
on the GPU. The "illegal address" error came from the Light Tree
Sampling code not handling the NaN normals well, leading to weird code
paths being taken, eventually leading to a kernel_assert and a
user facing illegal address error.
In Blender 3.5+, this fixes two errors (visual artifacts and errors on
the GPU when using the light tree). In Blender 3.3 it only fixes one
(visual artifacts) because 3.3 doesn't have the light tree. The commit
message should be adjusted to represent that
Pull Request: https://projects.blender.org/blender/blender/pulls/111294
Seems like the new audio channel api was not as backwards compatible as we thought.
Therefore we need to reintroduce the usage of the old api to make older ffmpeg version be able to compile Blender.
This change is only intended to stick around for two releases or so. After that we hope that most Linux distros ship
ffmpeg >=5.0 so we can switch to it.
Reviewed By: Sergey
Differential Revision: http://developer.blender.org/D16408
* User simpler API names that accept both PTX and OptiX-IR
* New argument for optixProgramGroupGetStackSize, leave to default
* Remove OptixPipelineLinkOptions::debugLevel that does nothing
Pull Request: https://projects.blender.org/blender/blender/pulls/107450
In some cases the Geometry Nodes Volume Cube generates
a Volume grid that doesn't contain any leaf nodes. This can happen
when only tiles are needed to represent the volume.
This PR changes the `empty_grid` check function to also check if
there are any active tiles in the grid before returning `true`.
Pull Request: https://projects.blender.org/blender/blender/pulls/107551
This breaks backwards compatibility some in that 3 sides will be mapped
differently now, but difficult to avoid and can be considered a bugfix.
Similar to rBdd8016f7081f.
Maniphest Tasks: T103615
Differential Revision: https://developer.blender.org/D16910
DDS files coming through OIIO needed a similar treatment as TGA in
T99565; just for DDS OIIO just never set the "unassociated alpha"
attribute. Fixes T101850.
Reviewed By: Brecht Van Lommel
Differential Revision: https://developer.blender.org/D16270
The OpenVDB delay loading of voxel leaf data using mmap works poorly on network
drives. It has a mechanism to make a temporary local file copy to avoid this,
but we disabled that as it leads to other problems.
Now disable delay loading entirely. It's not clear that this has much benefit
in Blender. For rendering we need to load the entire grid to convert to NanoVDB,
and for geometry nodes there also are no cases where we only need part of grids.
Known Issues:
- Command buffer failures when using binary archives (binary archives is disabled for Intel GPUs as a workaround)
- Wrong texture sampler being applied (to be addressed in the future)
Ref T92212
Differential Revision: https://developer.blender.org/D16253
This patch optimises the Metal inlining policy. It gives a small speedup
(2-3% on M1 Max) with no notable compilation slowdown vs what is already
in master. Previously noted compilation slowdowns (as reported in T100102)
were caused by forcing inlining for `ccl_device`, but we get better
rendering perf by relying on compiler heuristics in these cases.
Backported to 3.3 because this also fixes a test failure.
Differential Revision: https://developer.blender.org/D16081
rB67e23b4b2967 revealed the bug. But the bug already existed before,
it just wasn't triggered.
Apparently the problem happens because the python code generated in
`initGuiding()` cannot be executed twice.
The second time the `initGuiding()` code is executed, the local python
variables are removed to make way for the others, but the reference to
one of the grids in a `Solver` object (name='solver_guiding2') is still
being used somewhere. So an error is raised and a crash is forced.
The solution is to prevent the python code in `initGuiding()` from being
executed twice.
When `FLUID_DOMAIN_ACTIVE_GUIDE` is in `fds->active_fields` this
indicates that the pointer in `mPhiGuideIn` has been set and the guiding
is already computed (does not need to be computed again).
Maniphest Tasks: T102257
Differential Revision: https://developer.blender.org/D16416
As mentioned in T89399, "the source of this bug is that cursor wrap
moves the cursor, but when it later checks the mouse position it hasn't
yet been updated, so it re-wraps".
As far as I could see, this happens for two reasons:
1. During the first warp, there are already other mousemove events in the queue with an outdated position.
2. Sometimes Windows occasionally and inexplicably ignores `SetCursorPos()` or `SendInput()` events. (See [1])
The solution consists in checking if the cursor is inside the bounds right after wrapping.
If it's not inside, it indicates that the wrapping either didn't work or the event is out of date.
In these cases do not change the "accum" values.
1. f317d619cc/src/video/windows/SDL_windowsmouse.c (L255))
Maniphest Tasks: T89399
Differential Revision: https://developer.blender.org/D15707
The issue here was that the Barbershop benchmark scene was saved with a
custom OCIO config, which leads to some textures having a unknown
colorspace when loading with a default installation.
This is automatically fixed by Blender during image loading, but since
Cycles queried the colorspace before actually loading the image, it
didn't get the updated value in the first render.
To fix this, just re-query the colorspace after the image is loaded.
Note that non-packed images still get treated as raw data if the
colorspace is unknown, but this is at least consistent and doesn't
magically change when you press F12 a second time.
Differential Revision: https://developer.blender.org/D16427
This was a floating point precision issue - or, to be more precise,
an issue with how Cycles split floats into the integer and fractional
parts for Perlin noise.
For coordinates below -2^24, the integer could be wrong, leading to
the fractional part being outside of 0-1 range, which breaks all sorts
of other things. 2^24 sounds like a lot, but due to how the detail
octaves work, it's not that hard to reach when combined with a large
scale.
Since this code is originally based on OSL, I checked if they changed
it in the meantime, and sure enough, there's a fix for it:
5c9dc68391
So, this basically just ports over that change to Cycles.
The original code mentions being faster, but as pointed out in the
linked commit, the performance impact is actually irrelevant.
I also checked in a simple scene with eight Noise textures at
detail 15 (with >90% of render time being spent on the noise), and
the render time went from 13.06sec to 13.05sec. So, yeah, no issue.
The issue here was that PathTraceWork was set up before checking if
any error occurred, and it didn't account for the dummy device so
it called a non-implemented function.
This fix therefore avoids creating PathTraceWork for dummy devices
and checks for device creation errors earlier in the process.
OSL (like Cycles) has no internal boolean type, instead an integer
input can be flagged to be shown as a boolean in the UI.
Cycles reacts to this by creating a boolean socket on the Blender
side, but as a result incorrectly called the boolean overload of the
set function even though the internal type is an integer.
There's another unrelated crash in the GPU viewport shader code that
appears to apply to every OSL node that outputs a shader, and the file
in T101702 triggers both, so this is only a partial fix for the report.
(Cherry pick D16082)
An bug with binary archives was fixed in macOS 13.0 which stops some spurious kernel recompilations. In older macOS versions, falling back on the system shader cache will prevent recompilations in most instances (this is the same behaviour as in Blender 3.1.x and 3.2.x).
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D16141
Windows drivers 101.3430 fix an important GUI-related crash and it's
best to prevent users from running into it.
Linux drivers weren't affected but still had relevant gpu binary
compatibility fixes, so it makes sense to keep the min-supported version
aligned across OSes.
Available on Windows and macOS, where such gestures are supported.
For Windows, disabling this option restores touchpad behavior to
match Blender 3.2.
Ref T97925
Differential Revision: https://developer.blender.org/D16005
