🐛 Describe the bug
Informational / low priority — filing because the warnings look like a broken image but aren't, and that costs each new user a debugging detour.
Environment
- macOS 26.5 (Apple Silicon), executorch
main @ d614f9d
- Arm GNU toolchain 15.2.rel1 (darwin-arm64), as downloaded by
examples/arm/setup.sh
- Standalone executor runner build per the Ethos-U getting-started tutorial (
ethos-u55-128, Shared_Sram)
Observed
The link succeeds but prints:
ld: ../../../arm_executor_runner: section `.data' can't be allocated in segment 1
LOAD: .text .ARM.extab .ARM.exidx .copy.table .zero.table .data
ld: ... section `.sram.data' can't be allocated in segment 2
ld: ... section `.rodata' can't be allocated in segment 2
ld: warning: ... has a LOAD segment with RWX permissions
readelf -l shows .data (0xa6c), .sram.data (0x18a8) and .rodata (0x72a0) missing from the section-to-segment mapping, which looks like ~37 KB of initialized data won't be loaded.
Verification that it's benign (today)
The image runs correctly on the Corstone-300 FVP — log strings (.rodata) print fine and the model produces correct output — the copy-table init path carries the data. But this appears to be newer-binutils strictness meeting the platform linker script's PHDRS declarations, and it's fragile-looking enough that it likely deserves either a linker-script update or a documented "expected warnings" note. Linux CI presumably pins an older toolchain and never surfaces this; the setup script currently fetches 15.2.rel1.
Versions
PyTorch version: 2.13.0
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A
OS: macOS 26.5.2 (arm64)
GCC version: Could not collect
Clang version: 21.0.0 (clang-2100.1.1.101)
CMake version: version 3.31.10
Libc version: N/A
Python version: 3.12.14 (main, Aug 12 2026, 13:57:54) [Clang 21.0.0 (clang-2100.1.1.101)] (64-bit runtime)
Python platform: macOS-26.5.2-arm64-arm-64bit
Is CUDA available: False
CUDA runtime version: No CUDA
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
Is XPU available: False
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: False
Caching allocator config: N/A
CPU:
Apple M4 Pro
Versions of relevant libraries:
[pip3] executorch==1.5.0+d614f9d
[pip3] numpy==2.5.2
[pip3] pytorch_tokenizers==1.3.0
[pip3] torch==2.13.0
[pip3] torchao==0.18.0.dev20260729+cpu
[pip3] torchaudio==2.11.0
[pip3] torchdata==0.11.0+cpu
[pip3] torchsr==1.0.4
[pip3] torchtune==0.0.0
[pip3] torchvision==0.28.0
[conda] Could not collect
🐛 Describe the bug
Informational / low priority — filing because the warnings look like a broken image but aren't, and that costs each new user a debugging detour.
Environment
main@ d614f9dexamples/arm/setup.shethos-u55-128,Shared_Sram)Observed
The link succeeds but prints:
readelf -lshows.data(0xa6c),.sram.data(0x18a8) and.rodata(0x72a0) missing from the section-to-segment mapping, which looks like ~37 KB of initialized data won't be loaded.Verification that it's benign (today)
The image runs correctly on the Corstone-300 FVP — log strings (
.rodata) print fine and the model produces correct output — the copy-table init path carries the data. But this appears to be newer-binutils strictness meeting the platform linker script's PHDRS declarations, and it's fragile-looking enough that it likely deserves either a linker-script update or a documented "expected warnings" note. Linux CI presumably pins an older toolchain and never surfaces this; the setup script currently fetches 15.2.rel1.Versions
PyTorch version: 2.13.0
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A
OS: macOS 26.5.2 (arm64)
GCC version: Could not collect
Clang version: 21.0.0 (clang-2100.1.1.101)
CMake version: version 3.31.10
Libc version: N/A
Python version: 3.12.14 (main, Aug 12 2026, 13:57:54) [Clang 21.0.0 (clang-2100.1.1.101)] (64-bit runtime)
Python platform: macOS-26.5.2-arm64-arm-64bit
Is CUDA available: False
CUDA runtime version: No CUDA
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
Is XPU available: False
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: False
Caching allocator config: N/A
CPU:
Apple M4 Pro
Versions of relevant libraries:
[pip3] executorch==1.5.0+d614f9d
[pip3] numpy==2.5.2
[pip3] pytorch_tokenizers==1.3.0
[pip3] torch==2.13.0
[pip3] torchao==0.18.0.dev20260729+cpu
[pip3] torchaudio==2.11.0
[pip3] torchdata==0.11.0+cpu
[pip3] torchsr==1.0.4
[pip3] torchtune==0.0.0
[pip3] torchvision==0.28.0
[conda] Could not collect