[wgpu-core] Call flush_mapped_ranges when unmapping write-mapped buffers
#6089
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Most devices have coherent memory in practice |
ErichDonGubler
approved these changes
Aug 12, 2024
ErichDonGubler
approved these changes
Aug 12, 2024
I'm not sure how things worked without this.
`zero_init_needs_flush_now` was always equal to `mapping.is_coherent` which is not correct but is fixed by the next commit.
`flush_mapped_ranges` needs to be called when mappings are not coherent. We can also omit flushing for write-mapped buffers since we always flush them on unmap.
`offset` is relative to the start of the mapping not the start of the buffer.
This is done in the same way as in `map_buffer` & `unmap_buffer`.
ErichDonGubler
force-pushed
the
fix-unmap
branch
from
f689389
to
6b4fea1
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Connections
-
Description
flush_mapped_rangesneeds to be called for non-coherent write-mappings so that the data becomes visible to the GPU.We also need to call it for non-coherent read-mappings in cases where the memory needs to be 0 initialized.
This is only needed for Vulkan and GL but even then, I'm not sure how things worked so far.
Vulkan says:
GL says:
Both specs seem to say that users that want to guarantee that the data is flushed should call the function. But they don't say that if users don't call the function the data won't be flushed.
So, it could be that the drivers are flushing internally?
Testing
Existing tests.