Fix floats on Wasm #2041
Open
Fix floats on Wasm #2041
Conversation
This file contains hidden or 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
On wasm32, terms are aligned by 4 bytes. `memory_heap_allocation` returns a pointer for `n` terms. We used float_term_t to create an union between a term and float. However, floats are aligned to 8 bytes which made casting between union and term pointers unsafe. We replaced union usage to `memcpy` which allows unaligned memory access. The true solution would be to return correctly aligned memory from memory_heap_alloc but fix for that would need a lot more testing. Signed-off-by: Jakub Gonet <jakub.gonet@swmansion.com>
bettio
requested changes
Dec 19, 2025
| float_term_t *boxed_float = (float_term_t *) (boxed_value + 1); | ||
| boxed_float->f = f; | ||
|
|
||
| // For wasm32, alignof(avm_float_t) = 8 and alignof(term) = 4 |
Collaborator
There was a problem hiding this comment.
This comment is somehow slightly misleading, this is not just an issue for wasm32.
It should be rephrased as:
// alignof(avm_float_t) might be != alignof(term)
|
|
||
| // For wasm32, alignof(avm_float_t) = 8 and alignof(term) = 4 | ||
| // `memory_heap_alloc` returns memory aligned to term size. | ||
| memcpy(boxed_value + 1, &f, sizeof(avm_float_t)); |
Collaborator
There was a problem hiding this comment.
Correct, but sadly there is a know ESP32 GCC bug that I opened 2 years ago: https://gcc.gnu.org/pipermail/gcc-bugs/2023-October/839353.html
Can you keep also the old code and use it with such a similar macro?
#if defined(__XTENSA__)
// See https://gcc.gnu.org/pipermail/gcc-bugs/2023-October/839353.html
[...]
#else
// your new code here
#endif
| return boxed_float->f; | ||
| avm_float_t result; | ||
| // see `term_from_float` | ||
| memcpy(&result, term_to_const_term_ptr(t) + 1, sizeof(avm_float_t)); |
Collaborator
There was a problem hiding this comment.
Same as the previous comment.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
On wasm32, terms are aligned by 4 bytes.
memory_heap_allocationreturns a pointer fornterms. We used float_term_t to create an union between a term and float. However, floats are aligned to 8 bytes which made casting between union and term pointers unsafe.We replaced union usage to
memcpywhich allows unaligned memory access. The true solution would be to return correctly aligned memory frommemory_heap_allocbut fix for that would need a lot more testing.These changes are made under both the "Apache 2.0" and the "GNU Lesser General
Public License 2.1 or later" license terms (dual license).
SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later