An assert checking invariants would sometime trigger, the problem
was a logic error that would be exposed by a race when running out
of space in the list.
The root of the problem is that in one place we were not remapping the
-1 offset to nullptr, storing a pointer violating our invariants.
Also added more asserts!!!
We're using timed condition variable in one place, but the STL version
pulls in a lot of code because it does clock calculations in
"long double" (!!!!). Since we already had an implementation
of condition_variable, we just add the timed version.
This saves several KiB of code.
Also don't use unique_lock() lock/unlock because it can throw exceptions.
This removes a lot of heap allocations/deallocations, and reduces
code size. Most improvements come from using CString and
StaticString instead of std::string and better using move
semantics.
In practice this limits materials to 10 samplers since Filament uses 5
for lights and skips 1 slot for the post-process sampler. This can
probably be optimized.
We now check for sampler overflow during material compilation rather
than waiting for run-time checks. This allows for Kokoro-based
validation of sample materials, and allows developers to catch this
issue in their asset pipeline.
We should also probably raise the upper limit (Qualcomm allows up to 16
samplers in their Vulkan implementation) but that can be a separate PR.
Fixes#507
This fixes the "shader is not ASCII" error seen with GL backend,
introduced when we changed blob ownership semantics to accommodate
shader compression. It was due to missing null terminators.
We actually never bothered including the trailing null in blob length,
which was wrong but happened to work because the BlobDictionary held a
weak reference to chunk data. SInce it now holds an actual copy, the
lack of null caused our strings to contain garbage memory.
Since we use this for decoding, this adds a dependency to the core
filament renderer which in practice is only used for Vulkan. However
this is a tiny library, so it's simplest just to always include it.
The next PR will add the actual compression / decompression code to
filaflat and filamat. Here are the preliminary results.
102K => 29K aiDefaultMat.filamat
102K => 29K aiDefaultTrans.filamat
21K => 5.0K bakedColor.filamat
22K => 5.2K bakedTexture.filamat
21K => 5.0K depthVisualizer.filamat
40K => 10K groundShadow.filamat
102K => 29K sandboxCloth.filamat
125K => 36K sandboxLit.filamat
126K => 36K sandboxLitFade.filamat
126K => 36K sandboxLitTransparent.filamat
109K => 31K sandboxSubsurface.filamat
21K => 4.9K sandboxUnlit.filamat
21K => 4.9K transparentColor.filamat
Both the "before" and "after" numbers are excluding non-Vulkan targets,
and I also changed our CMakeLists to build filamat instead of inc.
* Clean-up EntityManager a bit
- use tsl::robin_set instead of std::set (which should have been unordered::set
anyways).
- getListeners() now returns a vector which avoids to traverse a set twice.
Turns out that copying the set wasn't as efficient as I thought.
* Improve jobsystem a bit
We recently added a job reference counting mechanism, but we were a bit
too aggressive about taking/release references.
Also make the API more complete by adding explicit retain/release,
which is needed to allow several threads to wait on the same job.
Also improve futex code by inlining it.
* Minor clean-ups
- fix a couple usage of std::function
- fix a couple usage of std::string
- remove ALIGN_LOOP, which didn't work
- fix a couple explicit/noexcept
- virtual -> override
* Fix spelling typos and other minor clang-tidy
We used to assume 32-bytes cache lines when running on ARM 32-bit mode,
however, that was wrong because 32-bit mode doesn't change the cpu's
cache line. On all modern platforms we support, the cache line
is 64 bytes.
Set Job storage to 48 bytes on all platforms.
* Fix a reuse after free in the job system
Jobs were destroyed and recycled while still in use
by wait() or run(). To fix this we introduce reference-counting of
jobs.
Jobs start with a ref-count of 1, which is decremented when a job
naturally finishes. Additionally, all user-facing methods acquire
a reference for the duration of the call.
* Fix an API inconsistency with JobSystem
JobSystem's API lets the user create jobs but not destroy them.
Jobs are destroyed automatically, without a way for the caller to
know when that happens.
We now explicitly enforce that jobs are no longer valid when
wait() returns. Multiple concurrent wait() are allowed however.
This is enforced by clearing the job pointer upon returning
from JobSystem::wait(Job* job).
* Rename linked-list put/get to push/pop
* Better fix for Job use after free
There was still a race condition where a run()'ed
job could be destroyed before wait() was called,
wait would then use a destroyed object.
The available APIs now are:
run() - runs and destroys a job
runAndWait() - run, then waits for and destroys a job
runAndRetain() - runs and keep a reference to the job
wait() - waits and destroys a job
wait() can only be used with a job obtained with runAndRetain().
* Get rid of unused code
This version of parallel_for has use-after-free issues anyways,
since we changed the semantics of run/wait/etc...
* Fix decRef() memory order
decRef() must ensure that all access to the
object have happened before destroying it.
* Fix memory order in atomic linked list's pop()
It needs acquire semantic, since we want to make
sure that no read/write are reordered before the
pop() -- which returns an object to the caller.
* Fix memory order on runningJobCount
we needed acquire semantic when about to destroy
the last job -- it's similar to decRef.
* Comment usages of std::memory_order_*
* Fix AtomicFreeList A-B-A bug
Turns out AtomicFreeList was not immune to the ABA bug. W're fixing
it here by using a 64-bits CAS, which is available on aarch64 and armv7.
IF an UBO is marked as STREAM, meaning that data will
change every frame and the buffer content is invalidated
as well, then we use map/unmap buffer in asynchronous
mode to copy the data into the buffer.
This assumes the buffer is sized to be significantly
larger than the average amount of data used every
frame.
This makes our createTexture* helpers more consistent with the
createMaterial helper, and extends the wrapped Engine class with new
methods.
Before, clients did something like this:
const ao = Filament.createTextureFromPng(Filament.assets['foo.png'], engine);
...
engine.destroyTexture(ao);
Now, they can do:
const ao = engine.createTextureFromPng('foo.png');
...
engine.destroyTexture(ao);
This is more symmetric and concise. In general the JS API now supports a
class of "buffer-like" objects which exploit dynamic typing in order to
accept any of the following: BufferDescriptor, asset string, or
Uint8Array.
