Compare commits

..

20 Commits

Author SHA1 Message Date
Ayke van Laethem ec9e041a40 ci: use native arm64 builder for arm64 binaries 2025-11-09 09:41:06 +01:00
deadprogram bf613175e7 ci: update GH actions builds to use Go 1.25.4
Signed-off-by: deadprogram <ron@hybridgroup.com>
2025-11-07 10:04:33 +01:00
sivchari 4d62cca267 feature: Add flag to ignore go compatibility matrix (#5078)
* add go-compatibility feature

Signed-off-by: sivchari <shibuuuu5@gmail.com>

* rename env

Signed-off-by: sivchari <shibuuuu5@gmail.com>

* fix description

Signed-off-by: sivchari <shibuuuu5@gmail.com>

---------

Signed-off-by: sivchari <shibuuuu5@gmail.com>
2025-11-06 21:28:06 +01:00
deadprogram a63aa143bb build: use task scheduler on macOS builds to avoid test race condition lockups.
Fixes #4995

Signed-off-by: deadprogram <ron@hybridgroup.com>
2025-11-06 19:19:20 +01:00
Ben Krieger ed0fb774b9 Add testing.T.Context() and testing.B.Context() 2025-11-05 10:33:15 +01:00
Damian Gryski 9be956f4bc src/syscall: update src buffer after write
Fixes #5012
2025-10-28 12:43:23 -07:00
Ayke van Laethem 48d741f68a compiler: emit an error when the actual arch doesn't match GOARCH
See: https://github.com/tinygo-org/tinygo/issues/4959
This makes sure to emit an error instead of generating inline assembly,
which leads to hard-to-debug linker errors. Instead, it will now produce
errors like the following:

    # golang.org/x/sys/unix
    ../../../../pkg/mod/golang.org/x/sys@v0.0.0-20220722155257-8c9f86f7a55f/unix/affinity_linux.go:20:23: system calls are not supported: target emulates a linux/arm system on wasm32
    ../../../../pkg/mod/golang.org/x/sys@v0.0.0-20220722155257-8c9f86f7a55f/unix/fcntl.go:17:29: system calls are not supported: target emulates a linux/arm system on wasm32
    ../../../../pkg/mod/golang.org/x/sys@v0.0.0-20220722155257-8c9f86f7a55f/unix/fcntl.go:32:24: system calls are not supported: target emulates a linux/arm system on wasm32
2025-10-23 10:13:29 +02:00
Elias Naur 35e6b8a67d compileopts: lower "large stack" limit to 16kb
I'm running with `-stack-size 16kb` and running into excessive allocations
because the otherwise non-escaping buffer in

https://github.com/golang/go/blob/c53cb642deea152e28281133bc0053f5ec0ce358/src/crypto/internal/fips140/sha512/sha512block.go#L97

is moved to the heap.
2025-10-22 13:04:24 +02:00
Ayke van Laethem 8a31ae14a0 machine: use larger SPI MAXCNT on nrf52833 and nrf52840
These chips have a larger upper limit for the DMA transfer than the
nrf52832. For best performance, we should be splitting the transfer in
as large blocks as possible on the given hardware.
2025-10-21 13:33:13 +02:00
Ayke van Laethem 26ee3fb93e wasm: fix C realloc and optimize it a bit
- Do not use make([]byte, ...) to allocate, instead call the allocator
    directly with a nil (undefined) layout. This makes sure the precise
    GC will scan the contents of the allocation, since C could very well
    put pointers in there.
  - Simplify the map to use the pointer as the key and the size as the
    value, instead of storing the slices directly in the map.
2025-10-20 12:58:11 +02:00
Ayke van Laethem 8872d5ebfc runtime: fix sleep duration for long sleeps
Long sleep durations weren't implemented correctly, by simply casting to
a smaller number of bits. What we want is a saturating downcast, which
is what I've used here instead.

This should fix https://github.com/tinygo-org/tinygo/issues/3356.
2025-10-16 17:20:50 +02:00
deadprogram faf455283d chore: update all CI builds to use latest stable Go release. Also update some of the actions to their latest releases.
Signed-off-by: deadprogram <ron@hybridgroup.com>
2025-10-16 09:04:00 +02:00
tinkerator 562b4c5020 Create "pico2-ice" target board (#5062)
* Create "pico2-ice" target board

This board has an rp2350b chip on it, as well as a Lattice Semiconductor
iCE40UP5K FPGA. More details of this open hardware board here:

  https://pico2-ice.tinyvision.ai/

Tested on a pico2-ice board with:

  ~/go/bin/tinygo flash -target=pico2-ice src/examples/blinky1/blinky1.go

which blinks the GREEN LED (connected to GPIO0) on this board.

Signed-off-by: Tinkerer <tinkerer@zappem.net>

* More silkscreen labels for pico2-ice board

Reading the schematic and the rev2 board viewer:

https://raw.githubusercontent.com/tinyvision-ai-inc/pico2-ice/refs/heads/main/Board/Rev2/bom/ibom.html

concluded that the RP2350B GPIO pins are not labeled with these
pin numbers in the silkscreen. Instead, the silkscreen refers to
uses of the GPIOs or the ICE numbered pins. RP2340B devoted pins
map to the A1..4 B1..4 pins on the RP PMOD connector. Also
silkscreen "~0".."~6" are labeled as pins N0..N6.

Signed-off-by: Tinkerer <tinkerer@zappem.net>

* Added a smoketest for pico2-ice board and tidied up GPIO defs

Addresses review comments from aykevl.

Signed-off-by: Tinkerer <tinkerer@zappem.net>
2025-10-13 13:19:04 +02:00
Ayke van Laethem 7460734522 compiler: mark string parameters as readonly
Strings are readonly, but the compiler doesn't always know this. Marking
them as readonly in the frontend allows the compiler to optimize based
on this knowledge.

This provides some small code size benefits. I didn't measure running
speed.
2025-10-11 23:03:18 +02:00
Olaf Flebbe 3be7100e89 fix(rp2): possible integer overflow while computing factors for SPI baudrate
Incorrect factors are calculated for baudrates which are a bit
larger than integer multiples of 4194304.
For example for baudrates of 8_400_000 or 58_800_000.
Fixed the same way in newer versions of the RPI SDK.
2025-10-09 11:48:27 +02:00
张之阳 aedaf7d925 machine: fix deprecated AsmFull comment (#5005) 2025-10-04 11:19:01 +02:00
Ayke van Laethem 256e84f912 all: shrink bdwgc library
By disabling some configuration options (and updating the library), the
library becomes a lot smaller. With `-no-debug`, binaries become ~18kB
smaller on Linux, ~4kB smaller on MacOS, ~9kB smaller on Windows, and
~12kB smaller on WebAssembly.
2025-10-03 15:46:25 +02:00
Ayke van Laethem 0a87846bd8 cortexm: optimize code size for the HardFault_Handler
This function is called when a hard fault occurs. Hard faults happen
when something really bad happens - like writing to unwritable memory or
an unaligned memory access on Cortex-M0. It is not generally possible to
recover from these.

This commit optimizes the code size overhead of hard fault handling:

  * It removes the stack overflow checking code.
    This may seem like a bad thing, but the only thing this could check
    were stack overflows outside goroutines. In practice, this could
    only really happen on a stack overflow in the scheduler (unlikely),
    or in interrupt code (possible, but interrupts are small so still
    unlikely). Most stack overflows happen in regular goroutines, and
    weren't caught in the HardFault.
  * It makes the panic message similar to a regular panic. This has two
    advantages:
      * It reduces code size, because the string can be reused between
        the HardFault handler and the runtime panic function.
      * Using the same pattern automatically makes `-monitor` print the
        source address for the hard fault. Not a big benefit as we could
        trivially add any other pattern but a nice benefit nonetheless.

Result:

    $ tinygo flash -target=microbit -size=short -programmer=openocd -monitor examples/serial
       code    data     bss |   flash     ram
       3036       8    2256 |    3044    2264
    [...snip]
    Connected to /dev/ttyACM0. Press Ctrl-C to exit.
    panic: runtime error at 0x00000344: HardFault with sp=0x200007d0
    [tinygo: panic at /home/ayke/src/tinygo/tinygo/src/internal/task/task_stack_cortexm.go:48:4]

(This is with https://github.com/tinygo-org/tinygo/pull/3680 not yet
fixed and some local changes to configure the UART so I can actually see
the panic).

For atsamd21/nrf51 chips this results in a binary size reduction of
around 100 bytes. For other Cortex-M chips it's around 24 bytes but I
hope to change this in the future because a lot of the fault decoding in
runtime_cortexm_hardfault_debug.go should IMHO be done by the TinyGo
monitor instead (I estimate that this would save around 800 bytes on
these chips).
2025-10-03 13:20:18 +02:00
Ayke van Laethem 073862ee92 fe310: add I2C pins for the HiFive1b 2025-10-02 20:26:25 +02:00
Ayke van Laethem c820d83ae2 interp: better errors when debugging interp
When debugging is enabled for interp, print better errors in a specific
case.

Before:

    !! revert because of error: interp: unsupported instruction (to be emitted at runtime)

After:

    !! revert because of error: /usr/local/go1.24.0/src/regexp/syntax/parse.go:927:27: interp: unsupported instruction (to be emitted at runtime)

So this adds error location information, which can be quite useful.
2025-10-02 05:42:08 +02:00
59 changed files with 530 additions and 234 deletions
+1 -1
View File
@@ -108,7 +108,7 @@ jobs:
- image: golang:1.25-bullseye
steps:
- test-linux:
llvm: "21"
llvm: "20"
resource_class: large
workflows:
+6 -6
View File
@@ -39,7 +39,7 @@ jobs:
- name: Install Go
uses: actions/setup-go@v6
with:
go-version: '1.25.1'
go-version: '1.25.4'
cache: true
- name: Restore LLVM source cache
uses: actions/cache/restore@v4
@@ -96,7 +96,7 @@ jobs:
- name: Build TinyGo release tarball
run: make release -j3
- name: Test stdlib packages
run: make tinygo-test
run: make tinygo-test TEST_ADDITIONAL_FLAGS="-scheduler=tasks"
- name: Make release artifact
run: cp -p build/release.tar.gz build/tinygo${{ steps.version.outputs.version }}.darwin-${{ matrix.goarch }}.tar.gz
- name: Publish release artifact
@@ -117,7 +117,7 @@ jobs:
runs-on: macos-latest
strategy:
matrix:
version: [16, 17, 18, 19, 20, 21]
version: [16, 17, 18, 19, 20]
steps:
- name: Set up Homebrew
uses: Homebrew/actions/setup-homebrew@master
@@ -134,15 +134,15 @@ jobs:
- name: Install Go
uses: actions/setup-go@v6
with:
go-version: '1.25.1'
go-version: '1.25.4'
cache: true
- name: Build TinyGo (LLVM ${{ matrix.version }})
run: go install -tags=llvm${{ matrix.version }}
- name: Check binary
run: tinygo version
- name: Build TinyGo (default LLVM)
if: matrix.version == 21
if: matrix.version == 20
run: go install
- name: Check binary
if: matrix.version == 21
if: matrix.version == 20
run: tinygo version
+2 -2
View File
@@ -31,7 +31,7 @@ jobs:
sudo rm -rf /usr/local/share/boost
df -h
- name: Check out the repo
uses: actions/checkout@v4
uses: actions/checkout@v5
with:
submodules: recursive
- name: Set up Docker Buildx
@@ -58,7 +58,7 @@ jobs:
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Build and push
uses: docker/build-push-action@v5
uses: docker/build-push-action@v6
with:
context: .
push: true
+28 -23
View File
@@ -16,7 +16,15 @@ jobs:
# Build Linux binaries, ready for release.
# This runs inside an Alpine Linux container so we can more easily create a
# statically linked binary.
runs-on: ubuntu-latest
strategy:
matrix:
goarch: [ amd64, arm64 ]
include:
- goarch: amd64
os: ubuntu-latest
- goarch: arm64
os: ubuntu-24.04-arm # TODO: use a -latest variant?
runs-on: ${{ matrix.os }}
container:
image: golang:1.25-alpine
outputs:
@@ -31,7 +39,7 @@ jobs:
# We're not on a multi-user machine, so this is safe.
run: git config --global --add safe.directory "$GITHUB_WORKSPACE"
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
with:
submodules: true
- name: Extract TinyGo version
@@ -40,7 +48,7 @@ jobs:
- name: Cache Go
uses: actions/cache@v4
with:
key: go-cache-linux-alpine-v1-${{ hashFiles('go.mod') }}
key: go-cache-linux-alpine-${{ matrix.goarch }}-v1-${{ hashFiles('go.mod') }}
path: |
~/.cache/go-build
~/go/pkg/mod
@@ -73,7 +81,7 @@ jobs:
uses: actions/cache/restore@v4
id: cache-llvm-build
with:
key: llvm-build-20-linux-alpine-v1
key: llvm-build-19-linux-alpine-${{ matrix.goarch }}-v1
path: llvm-build
- name: Build LLVM
if: steps.cache-llvm-build.outputs.cache-hit != 'true'
@@ -97,7 +105,7 @@ jobs:
uses: actions/cache@v4
id: cache-binaryen
with:
key: binaryen-linux-alpine-v1
key: binaryen-linux-alpine-${{ matrix.goarch }}-v1
path: build/wasm-opt
- name: Build Binaryen
if: steps.cache-binaryen.outputs.cache-hit != 'true'
@@ -116,28 +124,28 @@ jobs:
- name: Build TinyGo release
run: |
make release deb -j3 STATIC=1
cp -p build/release.tar.gz /tmp/tinygo${{ steps.version.outputs.version }}.linux-amd64.tar.gz
cp -p build/release.deb /tmp/tinygo_${{ steps.version.outputs.version }}_amd64.deb
cp -p build/release.tar.gz /tmp/tinygo${{ steps.version.outputs.version }}.linux-${{ matrix.goarch }}.tar.gz
cp -p build/release.deb /tmp/tinygo_${{ steps.version.outputs.version }}_${{ matrix.goarch }}.deb
- name: Publish release artifact
uses: actions/upload-artifact@v4
with:
name: linux-amd64-double-zipped-${{ steps.version.outputs.version }}
name: linux-${{ matrix.goarch }}-double-zipped-${{ steps.version.outputs.version }}
path: |
/tmp/tinygo${{ steps.version.outputs.version }}.linux-amd64.tar.gz
/tmp/tinygo_${{ steps.version.outputs.version }}_amd64.deb
/tmp/tinygo${{ steps.version.outputs.version }}.linux-${{ matrix.goarch }}.tar.gz
/tmp/tinygo_${{ steps.version.outputs.version }}-${{ matrix.goarch }}.deb
test-linux-build:
# Test the binaries built in the build-linux job by running the smoke tests.
runs-on: ubuntu-latest
needs: build-linux
steps:
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
with:
submodules: true
- name: Install Go
uses: actions/setup-go@v5
uses: actions/setup-go@v6
with:
go-version: '1.25.0'
go-version: '1.25.4'
cache: true
- name: Install wasmtime
uses: bytecodealliance/actions/wasmtime/setup@v1
@@ -164,7 +172,7 @@ jobs:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
with:
submodules: true
- name: Install apt dependencies
@@ -179,9 +187,9 @@ jobs:
simavr \
ninja-build
- name: Install Go
uses: actions/setup-go@v5
uses: actions/setup-go@v6
with:
go-version: '1.25.0'
go-version: '1.25.4'
cache: true
- name: Install Node.js
uses: actions/setup-node@v4
@@ -272,11 +280,8 @@ jobs:
# part of the release tarball.
strategy:
matrix:
goarch: [ arm, arm64 ]
goarch: [ arm ]
include:
- goarch: arm64
toolchain: aarch64-linux-gnu
libc: arm64
- goarch: arm
toolchain: arm-linux-gnueabihf
libc: armhf
@@ -284,7 +289,7 @@ jobs:
needs: build-linux
steps:
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
- name: Get TinyGo version
id: version
run: ./.github/workflows/tinygo-extract-version.sh | tee -a "$GITHUB_OUTPUT"
@@ -296,9 +301,9 @@ jobs:
g++-${{ matrix.toolchain }} \
libc6-dev-${{ matrix.libc }}-cross
- name: Install Go
uses: actions/setup-go@v5
uses: actions/setup-go@v6
with:
go-version: '1.25.0'
go-version: '1.25.4'
cache: true
- name: Restore LLVM source cache
uses: actions/cache/restore@v4
+2 -2
View File
@@ -25,7 +25,7 @@ jobs:
contents: read
steps:
- name: Check out the repo
uses: actions/checkout@v4
uses: actions/checkout@v5
with:
submodules: recursive
- name: Set up Docker Buildx
@@ -52,7 +52,7 @@ jobs:
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Build and push
uses: docker/build-push-action@v5
uses: docker/build-push-action@v6
with:
target: tinygo-llvm-build
context: .
+1 -1
View File
@@ -21,7 +21,7 @@ jobs:
# See: https://github.com/tinygo-org/tinygo/pull/4516#issuecomment-2416363668
run: sudo apt-get remove llvm-18
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
- name: Pull musl, bdwgc
run: |
git submodule update --init lib/musl lib/bdwgc
+5 -5
View File
@@ -2,11 +2,11 @@
# still works after checking out the dev branch (that is, when going from LLVM
# 16 to LLVM 17 for example, both Clang 16 and Clang 17 are installed).
echo 'deb https://apt.llvm.org/noble/ llvm-toolchain-noble-21 main' | sudo tee /etc/apt/sources.list.d/llvm.list
echo 'deb https://apt.llvm.org/noble/ llvm-toolchain-noble-20 main' | sudo tee /etc/apt/sources.list.d/llvm.list
wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
sudo apt-get update
sudo apt-get install --no-install-recommends -y \
llvm-21-dev \
clang-21 \
libclang-21-dev \
lld-21
llvm-20-dev \
clang-20 \
libclang-20-dev \
lld-20
+1 -1
View File
@@ -20,7 +20,7 @@ jobs:
run: |
echo "$HOME/go/bin" >> $GITHUB_PATH
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
with:
fetch-depth: 0 # fetch all history (no sparse checkout)
submodules: true
+12 -12
View File
@@ -31,7 +31,7 @@ jobs:
run: |
scoop install ninja binaryen
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
with:
submodules: true
- name: Extract TinyGo version
@@ -39,9 +39,9 @@ jobs:
shell: bash
run: ./.github/workflows/tinygo-extract-version.sh | tee -a "$GITHUB_OUTPUT"
- name: Install Go
uses: actions/setup-go@v5
uses: actions/setup-go@v6
with:
go-version: '1.25.0'
go-version: '1.25.4'
cache: true
- name: Restore cached LLVM source
uses: actions/cache/restore@v4
@@ -143,11 +143,11 @@ jobs:
run: |
scoop install binaryen
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
- name: Install Go
uses: actions/setup-go@v5
uses: actions/setup-go@v6
with:
go-version: '1.25.0'
go-version: '1.25.4'
cache: true
- name: Download TinyGo build
uses: actions/download-artifact@v4
@@ -173,11 +173,11 @@ jobs:
maximum-size: 24GB
disk-root: "C:"
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
- name: Install Go
uses: actions/setup-go@v5
uses: actions/setup-go@v6
with:
go-version: '1.25.0'
go-version: '1.25.4'
cache: true
- name: Download TinyGo build
uses: actions/download-artifact@v4
@@ -209,11 +209,11 @@ jobs:
run: |
scoop install binaryen && scoop install wasmtime@29.0.1
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v5
- name: Install Go
uses: actions/setup-go@v5
uses: actions/setup-go@v6
with:
go-version: '1.25.0'
go-version: '1.25.4'
cache: true
- name: Download TinyGo build
uses: actions/download-artifact@v4
+4 -1
View File
@@ -481,6 +481,7 @@ TEST_IOFS := false
endif
TEST_SKIP_FLAG := -skip='TestExtraMethods|TestParseAndBytesRoundTrip/P256/Generic'
TEST_ADDITIONAL_FLAGS ?=
# Test known-working standard library packages.
# TODO: parallelize, and only show failing tests (no implied -v flag).
@@ -488,7 +489,7 @@ TEST_SKIP_FLAG := -skip='TestExtraMethods|TestParseAndBytesRoundTrip/P256/Generi
tinygo-test:
@# TestExtraMethods: used by many crypto packages and uses reflect.Type.Method which is not implemented.
@# TestParseAndBytesRoundTrip/P256/Generic: relies on t.Skip() which is not implemented
$(TINYGO) test $(TEST_SKIP_FLAG) $(TEST_PACKAGES_HOST) $(TEST_PACKAGES_SLOW)
$(TINYGO) test $(TEST_ADDITIONAL_FLAGS) $(TEST_SKIP_FLAG) $(TEST_PACKAGES_HOST) $(TEST_PACKAGES_SLOW)
@# io/fs requires os.ReadDir, not yet supported on windows or wasi. It also
@# requires a large stack-size. Hence, io/fs is only run conditionally.
@# For more details, see the comments on issue #3143.
@@ -621,6 +622,8 @@ smoketest: testchdir
@$(MD5SUM) test.hex
$(TINYGO) build -size short -o test.hex -target=feather-rp2040 examples/device-id
@$(MD5SUM) test.hex
$(TINYGO) build -size short -o test.hex -target=pico2-ice examples/blinky1
@$(MD5SUM) test.hex
# test simulated boards on play.tinygo.org
ifneq ($(WASM), 0)
GOOS=js GOARCH=wasm $(TINYGO) build -size short -o test.wasm -tags=arduino examples/blinky1
+6 -3
View File
@@ -30,7 +30,10 @@ var BoehmGC = Library{
// Use a minimal environment.
"-DNO_MSGBOX_ON_ERROR", // don't call MessageBoxA on Windows
"-DDONT_USE_ATEXIT",
"-DNO_GETENV",
"-DNO_GETENV", // smaller binary, more predictable configuration
"-DNO_CLOCK", // don't use system clock
"-DNO_DEBUGGING", // reduce code size
"-DGC_NO_FINALIZATION", // finalization is not used at the moment
// Special flag to work around the lack of __data_start in ld.lld.
// TODO: try to fix this in LLVM/lld directly so we don't have to
@@ -39,6 +42,8 @@ var BoehmGC = Library{
// Do not scan the stack. We have our own mechanism to do this.
"-DSTACK_NOT_SCANNED",
"-DNO_PROC_STAT", // we scan the stack manually (don't read /proc/self/stat on Linux)
"-DSTACKBOTTOM=0", // dummy value, we scan the stack manually
// Assertions can be enabled while debugging GC issues.
//"-DGC_ASSERTIONS",
@@ -63,7 +68,6 @@ var BoehmGC = Library{
"blacklst.c",
"dbg_mlc.c",
"dyn_load.c",
"finalize.c",
"headers.c",
"mach_dep.c",
"malloc.c",
@@ -71,7 +75,6 @@ var BoehmGC = Library{
"mark_rts.c",
"misc.c",
"new_hblk.c",
"obj_map.c",
"os_dep.c",
"reclaim.c",
}
+7 -4
View File
@@ -33,10 +33,13 @@ func NewConfig(options *compileopts.Options) (*compileopts.Config, error) {
if err != nil {
return nil, err
}
if gorootMajor != 1 || gorootMinor < minorMin || gorootMinor > minorMax {
// Note: when this gets updated, also update the Go compatibility matrix:
// https://github.com/tinygo-org/tinygo-site/blob/dev/content/docs/reference/go-compat-matrix.md
return nil, fmt.Errorf("requires go version 1.%d through 1.%d, got go%d.%d", minorMin, minorMax, gorootMajor, gorootMinor)
if options.GoCompatibility {
if gorootMajor != 1 || gorootMinor < minorMin || gorootMinor > minorMax {
// Note: when this gets updated, also update the Go compatibility matrix:
// https://github.com/tinygo-org/tinygo-site/blob/dev/content/docs/reference/go-compat-matrix.md
return nil, fmt.Errorf("requires go version 1.%d through 1.%d, got go%d.%d", minorMin, minorMax, gorootMajor, gorootMinor)
}
}
// Check that the Go toolchain version isn't too new, if we haven't been
+2 -2
View File
@@ -43,8 +43,8 @@ func TestBinarySize(t *testing.T) {
tests := []sizeTest{
// microcontrollers
{"hifive1b", "examples/echo", 3884, 280, 0, 2268},
{"microbit", "examples/serial", 2924, 388, 8, 2272},
{"wioterminal", "examples/pininterrupt", 7365, 1491, 116, 6912},
{"microbit", "examples/serial", 2844, 360, 8, 2272},
{"wioterminal", "examples/pininterrupt", 7349, 1491, 116, 6912},
// TODO: also check wasm. Right now this is difficult, because
// wasm binaries are run through wasm-opt and therefore the
-4
View File
@@ -21,10 +21,6 @@ import (
)
/*
// Hide a warning in LLVM 21 that doesn't apply to us (appears to be a side
// effect of how CGo processes C header files).
#cgo CFLAGS: -Wno-deprecated-declarations
#include <clang-c/Index.h> // If this fails, libclang headers aren't available. Please take a look here: https://tinygo.org/docs/guides/build/
#include <llvm/Config/llvm-config.h>
#include <stdlib.h>
+1 -1
View File
@@ -1,4 +1,4 @@
//go:build !byollvm && llvm20
//go:build !byollvm && !llvm15 && !llvm16 && !llvm17 && !llvm18 && !llvm19
package cgo
-15
View File
@@ -1,15 +0,0 @@
//go:build !byollvm && !llvm15 && !llvm16 && !llvm17 && !llvm18 && !llvm19 && !llvm20
package cgo
/*
#cgo linux CFLAGS: -I/usr/include/llvm-21 -I/usr/include/llvm-c-21 -I/usr/lib/llvm-21/include
#cgo darwin,amd64 CFLAGS: -I/usr/local/opt/llvm@21/include
#cgo darwin,arm64 CFLAGS: -I/opt/homebrew/opt/llvm@21/include
#cgo freebsd CFLAGS: -I/usr/local/llvm21/include
#cgo linux LDFLAGS: -L/usr/lib/llvm-21/lib -lclang
#cgo darwin,amd64 LDFLAGS: -L/usr/local/opt/llvm@21/lib -lclang
#cgo darwin,arm64 LDFLAGS: -L/opt/homebrew/opt/llvm@21/lib -lclang
#cgo freebsd LDFLAGS: -L/usr/local/llvm21/lib -lclang
*/
import "C"
+1 -1
View File
@@ -226,7 +226,7 @@ func (c *Config) StackSize() uint64 {
// MaxStackAlloc returns the size of the maximum allocation to put on the stack vs heap.
func (c *Config) MaxStackAlloc() uint64 {
if c.StackSize() > 32*1024 {
if c.StackSize() >= 16*1024 {
return 1024
}
+1
View File
@@ -59,6 +59,7 @@ type Options struct {
WITPackage string // pass through to wasm-tools component embed invocation
WITWorld string // pass through to wasm-tools component embed -w option
ExtLDFlags []string
GoCompatibility bool // enable to check for Go version compatibility
}
// Verify performs a validation on the given options, raising an error if options are not valid.
+8
View File
@@ -32,6 +32,9 @@ const (
// Whether this is a full or partial Go parameter (int, slice, etc).
// The extra context parameter is not a Go parameter.
paramIsGoParam = 1 << iota
// Whether this is a readonly parameter (for example, a string pointer).
paramIsReadonly
)
// createRuntimeCallCommon creates a runtime call. Use createRuntimeCall or
@@ -167,6 +170,7 @@ func (c *compilerContext) flattenAggregateType(t llvm.Type, name string, goType
continue
}
suffix := strconv.Itoa(i)
isString := false
if goType != nil {
// Try to come up with a good suffix for this struct field,
// depending on which Go type it's based on.
@@ -183,12 +187,16 @@ func (c *compilerContext) flattenAggregateType(t llvm.Type, name string, goType
suffix = []string{"r", "i"}[i]
case types.String:
suffix = []string{"data", "len"}[i]
isString = true
}
case *types.Signature:
suffix = []string{"context", "funcptr"}[i]
}
}
subInfos := c.flattenAggregateType(subfield, name+"."+suffix, extractSubfield(goType, i))
if isString {
subInfos[0].flags |= paramIsReadonly
}
paramInfos = append(paramInfos, subInfos...)
}
return paramInfos
-8
View File
@@ -84,7 +84,6 @@ type compilerContext struct {
funcPtrType llvm.Type // pointer in function address space (1 for AVR, 0 elsewhere)
funcPtrAddrSpace int
uintptrType llvm.Type
nocaptureAttr llvm.Attribute
program *ssa.Program
diagnostics []error
functionInfos map[*ssa.Function]functionInfo
@@ -136,13 +135,6 @@ func newCompilerContext(moduleName string, machine llvm.TargetMachine, config *C
c.funcPtrType = dummyFunc.Type()
dummyFunc.EraseFromParentAsFunction()
// The attribute "nocapture" changed to "captures(none)" in LLVM 21.
if llvmutil.Version() < 21 {
c.nocaptureAttr = c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0)
} else {
c.nocaptureAttr = c.ctx.CreateEnumAttribute(llvm.AttributeKindID("captures"), 0)
}
return c
}
+3 -3
View File
@@ -38,10 +38,10 @@ func (b *builder) createInlineAsm(args []ssa.Value) (llvm.Value, error) {
// provided immediately. For example:
//
// arm.AsmFull(
// "str {value}, {result}",
// "str {value}, [{result}]",
// map[string]interface{}{
// "value": 1
// "result": &dest,
// "value": 1,
// "result": uintptr(unsafe.Pointer(&dest)),
// })
func (b *builder) createInlineAsmFull(instr *ssa.CallCommon) (llvm.Value, error) {
asmString := constant.StringVal(instr.Args[0].(*ssa.Const).Value)
+3
View File
@@ -250,6 +250,9 @@ func (b *builder) createMapIteratorNext(rangeVal ssa.Value, llvmRangeVal, it llv
func hashmapIsBinaryKey(keyType types.Type) bool {
switch keyType := keyType.Underlying().(type) {
case *types.Basic:
// TODO: unsafe.Pointer is also a binary key, but to support that we
// need to fix an issue with interp first (see
// https://github.com/tinygo-org/tinygo/pull/4898).
return keyType.Info()&(types.IsBoolean|types.IsInteger) != 0
case *types.Pointer:
return true
+17 -9
View File
@@ -139,7 +139,13 @@ func (c *compilerContext) getFunction(fn *ssa.Function) (llvm.Type, llvm.Value)
// not.
// (It may be safe to add the nocapture parameter to the context
// parameter, but I'd like to stay on the safe side here).
llvmFn.AddAttributeAtIndex(i+1, c.nocaptureAttr)
nocapture := c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0)
llvmFn.AddAttributeAtIndex(i+1, nocapture)
}
if paramInfo.flags&paramIsReadonly != 0 && paramInfo.llvmType.TypeKind() == llvm.PointerTypeKind {
// Readonly pointer parameters (like strings) benefit from being marked as readonly.
readonly := c.ctx.CreateEnumAttribute(llvm.AttributeKindID("readonly"), 0)
llvmFn.AddAttributeAtIndex(i+1, readonly)
}
}
@@ -152,7 +158,7 @@ func (c *compilerContext) getFunction(fn *ssa.Function) (llvm.Type, llvm.Value)
// Mark it as noreturn so LLVM can optimize away code.
llvmFn.AddFunctionAttr(c.ctx.CreateEnumAttribute(llvm.AttributeKindID("noreturn"), 0))
case "internal/abi.NoEscape":
llvmFn.AddAttributeAtIndex(1, c.nocaptureAttr)
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0))
case "runtime.alloc":
// Tell the optimizer that runtime.alloc is an allocator, meaning that it
// returns values that are never null and never alias to an existing value.
@@ -174,25 +180,25 @@ func (c *compilerContext) getFunction(fn *ssa.Function) (llvm.Type, llvm.Value)
case "runtime.sliceAppend":
// Appending a slice will only read the to-be-appended slice, it won't
// be modified.
llvmFn.AddAttributeAtIndex(2, c.nocaptureAttr)
llvmFn.AddAttributeAtIndex(2, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0))
llvmFn.AddAttributeAtIndex(2, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("readonly"), 0))
case "runtime.sliceCopy":
// Copying a slice won't capture any of the parameters.
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("writeonly"), 0))
llvmFn.AddAttributeAtIndex(1, c.nocaptureAttr)
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0))
llvmFn.AddAttributeAtIndex(2, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("readonly"), 0))
llvmFn.AddAttributeAtIndex(2, c.nocaptureAttr)
llvmFn.AddAttributeAtIndex(2, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0))
case "runtime.stringFromBytes":
llvmFn.AddAttributeAtIndex(1, c.nocaptureAttr)
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0))
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("readonly"), 0))
case "runtime.stringFromRunes":
llvmFn.AddAttributeAtIndex(1, c.nocaptureAttr)
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0))
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("readonly"), 0))
case "runtime.trackPointer":
// This function is necessary for tracking pointers on the stack in a
// portable way (see gc_stack_portable.go). Indicate to the optimizer
// that the only thing we'll do is read the pointer.
llvmFn.AddAttributeAtIndex(1, c.nocaptureAttr)
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("nocapture"), 0))
llvmFn.AddAttributeAtIndex(1, c.ctx.CreateEnumAttribute(llvm.AttributeKindID("readonly"), 0))
case "__mulsi3", "__divmodsi4", "__udivmodsi4":
if strings.Split(c.Triple, "-")[0] == "avr" {
@@ -227,9 +233,11 @@ func (c *compilerContext) getFunction(fn *ssa.Function) (llvm.Type, llvm.Value)
llvmFn.AddFunctionAttr(c.ctx.CreateStringAttribute("wasm-import-name", info.wasmName))
}
nocaptureKind := llvm.AttributeKindID("nocapture")
nocapture := c.ctx.CreateEnumAttribute(nocaptureKind, 0)
for i, typ := range paramTypes {
if typ.TypeKind() == llvm.PointerTypeKind {
llvmFn.AddAttributeAtIndex(i+1, c.nocaptureAttr)
llvmFn.AddAttributeAtIndex(i+1, nocapture)
}
}
}
+8
View File
@@ -74,6 +74,14 @@ func (b *builder) createRawSyscall(call *ssa.CallCommon) (llvm.Value, error) {
return b.CreateCall(fnType, target, args, ""), nil
case b.GOARCH == "arm" && b.GOOS == "linux":
if arch := b.archFamily(); arch != "arm" {
// Some targets pretend to be linux/arm for compatibility but aren't
// actually such a system. Make sure we emit an error instead of
// creating inline assembly that will fail to compile.
// See: https://github.com/tinygo-org/tinygo/issues/4959
return llvm.Value{}, b.makeError(call.Pos(), "system calls are not supported: target emulates a linux/arm system on "+arch)
}
// Implement the EABI system call convention for Linux.
// Source: syscall(2) man page.
args := []llvm.Value{}
+1 -1
View File
@@ -50,7 +50,7 @@ unsafe.String.throw: ; preds = %entry
declare void @runtime.unsafeSlicePanic(ptr) #1
; Function Attrs: nounwind
define hidden ptr @main.unsafeStringData(ptr %s.data, i32 %s.len, ptr %context) unnamed_addr #2 {
define hidden ptr @main.unsafeStringData(ptr readonly %s.data, i32 %s.len, ptr %context) unnamed_addr #2 {
entry:
%stackalloc = alloca i8, align 1
call void @runtime.trackPointer(ptr %s.data, ptr nonnull %stackalloc, ptr undef) #3
+3 -3
View File
@@ -77,7 +77,7 @@ entry:
}
; Function Attrs: nounwind
define hidden %runtime._string @main.minString(ptr %a.data, i32 %a.len, ptr %b.data, i32 %b.len, ptr %context) unnamed_addr #2 {
define hidden %runtime._string @main.minString(ptr readonly %a.data, i32 %a.len, ptr readonly %b.data, i32 %b.len, ptr %context) unnamed_addr #2 {
entry:
%0 = insertvalue %runtime._string zeroinitializer, ptr %a.data, 0
%1 = insertvalue %runtime._string %0, i32 %a.len, 1
@@ -91,7 +91,7 @@ entry:
ret %runtime._string %5
}
declare i1 @runtime.stringLess(ptr, i32, ptr, i32, ptr) #1
declare i1 @runtime.stringLess(ptr readonly, i32, ptr readonly, i32, ptr) #1
; Function Attrs: nounwind
define hidden i32 @main.maxInt(i32 %a, i32 %b, ptr %context) unnamed_addr #2 {
@@ -116,7 +116,7 @@ entry:
}
; Function Attrs: nounwind
define hidden %runtime._string @main.maxString(ptr %a.data, i32 %a.len, ptr %b.data, i32 %b.len, ptr %context) unnamed_addr #2 {
define hidden %runtime._string @main.maxString(ptr readonly %a.data, i32 %a.len, ptr readonly %b.data, i32 %b.len, ptr %context) unnamed_addr #2 {
entry:
%0 = insertvalue %runtime._string zeroinitializer, ptr %a.data, 0
%1 = insertvalue %runtime._string %0, i32 %a.len, 1
+8 -8
View File
@@ -31,13 +31,13 @@ entry:
}
; Function Attrs: nounwind
define hidden i32 @main.stringLen(ptr %s.data, i32 %s.len, ptr %context) unnamed_addr #2 {
define hidden i32 @main.stringLen(ptr readonly %s.data, i32 %s.len, ptr %context) unnamed_addr #2 {
entry:
ret i32 %s.len
}
; Function Attrs: nounwind
define hidden i8 @main.stringIndex(ptr %s.data, i32 %s.len, i32 %index, ptr %context) unnamed_addr #2 {
define hidden i8 @main.stringIndex(ptr readonly %s.data, i32 %s.len, i32 %index, ptr %context) unnamed_addr #2 {
entry:
%.not = icmp ult i32 %index, %s.len
br i1 %.not, label %lookup.next, label %lookup.throw
@@ -55,16 +55,16 @@ lookup.throw: ; preds = %entry
declare void @runtime.lookupPanic(ptr) #1
; Function Attrs: nounwind
define hidden i1 @main.stringCompareEqual(ptr %s1.data, i32 %s1.len, ptr %s2.data, i32 %s2.len, ptr %context) unnamed_addr #2 {
define hidden i1 @main.stringCompareEqual(ptr readonly %s1.data, i32 %s1.len, ptr readonly %s2.data, i32 %s2.len, ptr %context) unnamed_addr #2 {
entry:
%0 = call i1 @runtime.stringEqual(ptr %s1.data, i32 %s1.len, ptr %s2.data, i32 %s2.len, ptr undef) #3
ret i1 %0
}
declare i1 @runtime.stringEqual(ptr, i32, ptr, i32, ptr) #1
declare i1 @runtime.stringEqual(ptr readonly, i32, ptr readonly, i32, ptr) #1
; Function Attrs: nounwind
define hidden i1 @main.stringCompareUnequal(ptr %s1.data, i32 %s1.len, ptr %s2.data, i32 %s2.len, ptr %context) unnamed_addr #2 {
define hidden i1 @main.stringCompareUnequal(ptr readonly %s1.data, i32 %s1.len, ptr readonly %s2.data, i32 %s2.len, ptr %context) unnamed_addr #2 {
entry:
%0 = call i1 @runtime.stringEqual(ptr %s1.data, i32 %s1.len, ptr %s2.data, i32 %s2.len, ptr undef) #3
%1 = xor i1 %0, true
@@ -72,16 +72,16 @@ entry:
}
; Function Attrs: nounwind
define hidden i1 @main.stringCompareLarger(ptr %s1.data, i32 %s1.len, ptr %s2.data, i32 %s2.len, ptr %context) unnamed_addr #2 {
define hidden i1 @main.stringCompareLarger(ptr readonly %s1.data, i32 %s1.len, ptr readonly %s2.data, i32 %s2.len, ptr %context) unnamed_addr #2 {
entry:
%0 = call i1 @runtime.stringLess(ptr %s2.data, i32 %s2.len, ptr %s1.data, i32 %s1.len, ptr undef) #3
ret i1 %0
}
declare i1 @runtime.stringLess(ptr, i32, ptr, i32, ptr) #1
declare i1 @runtime.stringLess(ptr readonly, i32, ptr readonly, i32, ptr) #1
; Function Attrs: nounwind
define hidden i8 @main.stringLookup(ptr %s.data, i32 %s.len, i8 %x, ptr %context) unnamed_addr #2 {
define hidden i8 @main.stringLookup(ptr readonly %s.data, i32 %s.len, i8 %x, ptr %context) unnamed_addr #2 {
entry:
%0 = zext i8 %x to i32
%.not = icmp ugt i32 %s.len, %0
+1 -1
View File
@@ -18,7 +18,7 @@ require (
golang.org/x/sys v0.30.0
golang.org/x/tools v0.30.0
gopkg.in/yaml.v2 v2.4.0
tinygo.org/x/go-llvm v0.0.0-20250916101410-63740cfada08
tinygo.org/x/go-llvm v0.0.0-20250422114502-b8f170971e74
)
require (
+2 -2
View File
@@ -58,5 +58,5 @@ gopkg.in/yaml.v2 v2.4.0 h1:D8xgwECY7CYvx+Y2n4sBz93Jn9JRvxdiyyo8CTfuKaY=
gopkg.in/yaml.v2 v2.4.0/go.mod h1:RDklbk79AGWmwhnvt/jBztapEOGDOx6ZbXqjP6csGnQ=
gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
tinygo.org/x/go-llvm v0.0.0-20250916101410-63740cfada08 h1:vvEVYF4Qb38C25U8Ae/1QUWlCSp4pIE0PR+/BwNPvBU=
tinygo.org/x/go-llvm v0.0.0-20250916101410-63740cfada08/go.mod h1:GFbusT2VTA4I+l4j80b17KFK+6whv69Wtny5U+T8RR0=
tinygo.org/x/go-llvm v0.0.0-20250422114502-b8f170971e74 h1:ovavgTdIBWCH8YWlcfq9gkpoyT1+IxMKSn+Df27QwE8=
tinygo.org/x/go-llvm v0.0.0-20250422114502-b8f170971e74/go.mod h1:GFbusT2VTA4I+l4j80b17KFK+6whv69Wtny5U+T8RR0=
+1 -1
View File
@@ -577,7 +577,7 @@ func (r *runner) run(fn *function, params []value, parentMem *memoryView, indent
// runtime instead of at compile time. But we need to
// revert any changes made by the call first.
if r.debug {
fmt.Fprintln(os.Stderr, indent+"!! revert because of error:", callErr.Err)
fmt.Fprintln(os.Stderr, indent+"!! revert because of error:", callErr.Error())
}
callMem.revert()
err := r.runAtRuntime(fn, inst, locals, &mem, indent)
+12
View File
@@ -1635,6 +1635,7 @@ func main() {
cpuprofile := flag.String("cpuprofile", "", "cpuprofile output")
monitor := flag.Bool("monitor", false, "enable serial monitor")
baudrate := flag.Int("baudrate", 115200, "baudrate of serial monitor")
gocompatibility := flag.Bool("go-compatibility", true, "enable to check for Go versions compatibility, you can also configure this by setting the TINYGO_GOCOMPATIBILITY environment variable")
// Internal flags, that are only intended for TinyGo development.
printIR := flag.Bool("internal-printir", false, "print LLVM IR")
@@ -1712,6 +1713,16 @@ func main() {
ocdCommands = strings.Split(*ocdCommandsString, ",")
}
val, ok := os.LookupEnv("TINYGO_GOCOMPATIBILITY")
if ok {
b, err := strconv.ParseBool(val)
if err != nil {
fmt.Fprintf(os.Stderr, "could not parse TINYGO_GOCOMPATIBILITY value %q: %v\n", val, err)
os.Exit(1)
}
*gocompatibility = b
}
options := &compileopts.Options{
GOOS: goenv.Get("GOOS"),
GOARCH: goenv.Get("GOARCH"),
@@ -1748,6 +1759,7 @@ func main() {
Timeout: *timeout,
WITPackage: witPackage,
WITWorld: witWorld,
GoCompatibility: *gocompatibility,
}
if *printCommands {
options.PrintCommands = printCommand
+3 -3
View File
@@ -48,10 +48,10 @@ func Asm(asm string)
// recognizes template values in the form {name}, like so:
//
// arm.AsmFull(
// "str {value}, {result}",
// "str {value}, [{result}]",
// map[string]interface{}{
// "value": 1
// "result": &dest,
// "value": 1,
// "result": uintptr(unsafe.Pointer(&dest)),
// })
//
// You can use {} in the asm string (which expands to a register) to set the
-23
View File
@@ -1,29 +1,6 @@
.syntax unified
.cfi_sections .debug_frame
.section .text.HardFault_Handler
.global HardFault_Handler
.type HardFault_Handler, %function
HardFault_Handler:
.cfi_startproc
// Put the old stack pointer in the first argument, for easy debugging. This
// is especially useful on Cortex-M0, which supports far fewer debug
// facilities.
mov r0, sp
// Load the default stack pointer from address 0 so that we can call normal
// functions again that expect a working stack. However, it will corrupt the
// old stack so the function below must not attempt to recover from this
// fault.
movs r3, #0
ldr r3, [r3]
mov sp, r3
// Continue handling this error in Go.
bl handleHardFault
.cfi_endproc
.size HardFault_Handler, .-HardFault_Handler
// This is a convenience function for semihosting support.
// At some point, this should be replaced by inline assembly.
.section .text.SemihostingCall
+3 -3
View File
@@ -10,10 +10,10 @@ func Asm(asm string)
// recognizes template values in the form {name}, like so:
//
// arm.AsmFull(
// "str {value}, {result}",
// "str {value}, [{result}]",
// map[string]interface{}{
// "value": 1
// "result": &dest,
// "value": 1,
// "result": uintptr(unsafe.Pointer(&dest)),
// })
//
// You can use {} in the asm string (which expands to a register) to set the
+3 -3
View File
@@ -10,10 +10,10 @@ func Asm(asm string)
// recognizes template values in the form {name}, like so:
//
// arm.AsmFull(
// "str {value}, {result}",
// "str {value}, [{result}]",
// map[string]interface{}{
// "value": 1
// "result": &dest,
// "value": 1,
// "result": uintptr(unsafe.Pointer(&dest)),
// })
//
// You can use {} in the asm string (which expands to a register) to set the
+3 -3
View File
@@ -10,10 +10,10 @@ func Asm(asm string)
// recognizes template values in the form {name}, like so:
//
// arm.AsmFull(
// "st {value}, {result}",
// "st {value}, [{result}]",
// map[string]interface{}{
// "value": 1
// "result": &dest,
// "value": 1,
// "result": uintptr(unsafe.Pointer(&dest)),
// })
//
// You can use {} in the asm string (which expands to a register) to set the
+2 -2
View File
@@ -15,8 +15,8 @@ const (
P09 Pin = 9
P10 Pin = 10
P11 Pin = 11
P12 Pin = 12
P13 Pin = 13
P12 Pin = 12 // peripherals: I2C0 SDA
P13 Pin = 13 // peripherals: I2C0 SCL
P14 Pin = 14
P15 Pin = 15
P16 Pin = 16
+173
View File
@@ -0,0 +1,173 @@
//go:build pico2_ice
// Most of the info is from
// https://pico2-ice.tinyvision.ai/md_pinout.html although
// (2025-09-07) RP4 appears twice in that pinout - the schematic is
// more clear. Consistent with other RPi boards, we use GPn instead of
// RPn to reference the RPi connected pins.
package machine
// GPIO pins
const (
GP0 = GPIO0
GP1 = GPIO1
GP2 = GPIO2
GP3 = GPIO3
GP4 = GPIO4
GP5 = GPIO5
GP6 = GPIO6
GP7 = GPIO7
GP8 = GPIO8
GP9 = GPIO9
GP10 = GPIO10
GP11 = GPIO11
GP12 = GPIO12
GP13 = GPIO13
GP14 = GPIO14
GP15 = GPIO15
GP16 = GPIO16
GP17 = GPIO17
GP18 = GPIO18
GP19 = GPIO19
GP20 = GPIO20
GP21 = GPIO21
GP22 = GPIO22
GP23 = GPIO23
GP24 = GPIO24
GP25 = GPIO25
GP26 = GPIO26
GP27 = GPIO27
GP28 = GPIO28
GP29 = GPIO29
GP30 = GPIO30
GP31 = GPIO31
GP32 = GPIO32
GP33 = GPIO33
GP34 = GPIO34
GP35 = GPIO35
GP36 = GPIO36
GP37 = GPIO37
GP38 = GPIO38
GP39 = GPIO39
GP40 = GPIO40
GP41 = GPIO41
GP42 = GPIO42
GP43 = GPIO43
GP44 = GPIO44
GP45 = GPIO45
GP46 = GPIO46
GP47 = GPIO47
// RPi pins shared with ICE. The ICE number is what appears on
// the board silkscreen.
ICE9 = GP28
ICE11 = GP29
ICE14 = GP7
ICE15 = GP6
ICE16 = GP5
ICE17 = GP4
ICE18 = GP27
ICE19 = GP23
ICE20 = GP22
ICE21 = GP26
ICE23 = GP25
ICE25 = GP30
ICE26 = GP24
ICE27 = GP20
// FPGA Clock pin.
ICE35_G0 = GP21
// Silkscreen & Pinout names
ICE_SSN = ICE16
ICE_SO = ICE14
ICE_SI = ICE17
ICE_CK = ICE15
SD = GP2
SC = GP3
FPGA_RSTN = GP31
A3 = GP32
A1 = GP33
A4 = GP34
A2 = GP35
B3 = GP36
B1 = GP37
B4 = GP38
B2 = GP39
N0 = GP40 // On the board these are labeled "~0"
N1 = GP41
N2 = GP42
N3 = GP43
N4 = GP44
N5 = GP45
N6 = GP46
// Functions from Schematic.
ICE_DONE = GP40
USB_BOOT = GP42
// Button
SW1 = GP42
BOOTSEL = GP42
// Tricolor LEDs
LED_RED = GP1
LED_GREEN = GP0
LED_BLUE = GP9
// Onboard LED
LED = LED_GREEN
// Onboard crystal oscillator frequency, in MHz.
xoscFreq = 12 // MHz
)
// This board does not define default i2c pins.
const (
I2C0_SDA_PIN = NoPin
I2C0_SCL_PIN = NoPin
I2C1_SDA_PIN = NoPin
I2C1_SCL_PIN = NoPin
)
// SPI default pins
const (
// Default Serial Clock Bus 0 for SPI communications
SPI0_SCK_PIN = GPIO18
// Default Serial Out Bus 0 for SPI communications
SPI0_SDO_PIN = GPIO19 // Tx
// Default Serial In Bus 0 for SPI communications
SPI0_SDI_PIN = GPIO16 // Rx
// Default Serial Clock Bus 1 for SPI communications
SPI1_SCK_PIN = GPIO10
// Default Serial Out Bus 1 for SPI communications
SPI1_SDO_PIN = GPIO11 // Tx
// Default Serial In Bus 1 for SPI communications
SPI1_SDI_PIN = GPIO12 // Rx
)
// UART pins
const (
UART0_TX_PIN = GPIO0
UART0_RX_PIN = GPIO1
UART1_TX_PIN = GPIO8
UART1_RX_PIN = GPIO9
UART_TX_PIN = UART0_TX_PIN
UART_RX_PIN = UART0_RX_PIN
)
var DefaultUART = UART0
// USB identifiers
const (
usb_STRING_PRODUCT = "Pico2"
usb_STRING_MANUFACTURER = "Raspberry Pi"
)
var (
usb_VID uint16 = 0x2E8A
usb_PID uint16 = 0x000A
)
+2
View File
@@ -69,3 +69,5 @@ const eraseBlockSizeValue = 4096
func eraseBlockSize() int64 {
return eraseBlockSizeValue
}
const spiMaxBufferSize = 255 // from the datasheet: TXD.MAXCNT and RXD.MAXCNT
+2
View File
@@ -90,3 +90,5 @@ const eraseBlockSizeValue = 4096
func eraseBlockSize() int64 {
return eraseBlockSizeValue
}
const spiMaxBufferSize = 0xffff // from the datasheet: TXD.MAXCNT and RXD.MAXCNT
+2
View File
@@ -108,3 +108,5 @@ const eraseBlockSizeValue = 4096
func eraseBlockSize() int64 {
return eraseBlockSizeValue
}
const spiMaxBufferSize = 0xffff // from the datasheet: TXD.MAXCNT and RXD.MAXCNT
+7 -9
View File
@@ -301,18 +301,16 @@ func (spi *SPI) Transfer(w byte) (byte, error) {
// padded until they fit: if len(w) > len(r) the extra bytes received will be
// dropped and if len(w) < len(r) extra 0 bytes will be sent.
func (spi *SPI) Tx(w, r []byte) error {
// Unfortunately the hardware (on the nrf52832) only supports up to 255
// bytes in the buffers, so if either w or r is longer than that the
// transfer needs to be broken up in pieces.
// The nrf52840 supports far larger buffers however, which isn't yet
// supported.
// Unfortunately the hardware (on the nrf52832) only supports a limited
// amount of bytes in the buffers (depending on the chip), so if either w or
// r is longer than that the transfer needs to be broken up in pieces.
for len(r) != 0 || len(w) != 0 {
// Prepare the SPI transfer: set the DMA pointers and lengths.
// read buffer
nr := uint32(len(r))
if nr > 0 {
if nr > 255 {
nr = 255
if nr > spiMaxBufferSize {
nr = spiMaxBufferSize
}
spi.Bus.RXD.PTR.Set(uint32(uintptr(unsafe.Pointer(&r[0]))))
r = r[nr:]
@@ -322,8 +320,8 @@ func (spi *SPI) Tx(w, r []byte) error {
// write buffer
nw := uint32(len(w))
if nw > 0 {
if nw > 255 {
nw = 255
if nw > spiMaxBufferSize {
nw = spiMaxBufferSize
}
spi.Bus.TXD.PTR.Set(uint32(uintptr(unsafe.Pointer(&w[0]))))
w = w[nw:]
+25
View File
@@ -124,6 +124,31 @@ const (
fnXIP pinFunc = 0
)
// validPins confirms that the SPI pin selection is a legitimate one
// for the the 2040 chip.
func (spi *SPI) validPins(config SPIConfig) error {
var okSDI, okSDO, okSCK bool
switch spi.Bus {
case rp.SPI0:
okSDI = config.SDI == 0 || config.SDI == 4 || config.SDI == 16 || config.SDI == 20
okSDO = config.SDO == 3 || config.SDO == 7 || config.SDO == 19 || config.SDO == 23
okSCK = config.SCK == 2 || config.SCK == 6 || config.SCK == 18 || config.SCK == 22
case rp.SPI1:
okSDI = config.SDI == 8 || config.SDI == 12 || config.SDI == 24 || config.SDI == 28
okSDO = config.SDO == 11 || config.SDO == 15 || config.SDO == 27
okSCK = config.SCK == 10 || config.SCK == 14 || config.SCK == 26
}
switch {
case !okSDI:
return errSPIInvalidSDI
case !okSDO:
return errSPIInvalidSDO
case !okSCK:
return errSPIInvalidSCK
}
return nil
}
// Configure configures the gpio pin as per mode.
func (p Pin) Configure(config PinConfig) {
if p == NoPin {
+27
View File
@@ -2,6 +2,8 @@
package machine
import "device/rp"
// Analog pins on RP2350a.
const (
ADC0 Pin = GPIO26
@@ -12,3 +14,28 @@ const (
// fifth ADC channel.
thermADC = 30
)
// validPins confirms that the SPI pin selection is a legitimate one
// for the the 2350a chip.
func (spi *SPI) validPins(config SPIConfig) error {
var okSDI, okSDO, okSCK bool
switch spi.Bus {
case rp.SPI0:
okSDI = config.SDI == 0 || config.SDI == 4 || config.SDI == 16 || config.SDI == 20
okSDO = config.SDO == 3 || config.SDO == 7 || config.SDO == 19 || config.SDO == 23
okSCK = config.SCK == 2 || config.SCK == 6 || config.SCK == 18 || config.SCK == 22
case rp.SPI1:
okSDI = config.SDI == 8 || config.SDI == 12 || config.SDI == 24 || config.SDI == 28
okSDO = config.SDO == 11 || config.SDO == 15 || config.SDO == 27
okSCK = config.SCK == 10 || config.SCK == 14 || config.SCK == 26
}
switch {
case !okSDI:
return errSPIInvalidSDI
case !okSDO:
return errSPIInvalidSDO
case !okSCK:
return errSPIInvalidSCK
}
return nil
}
+27
View File
@@ -2,6 +2,8 @@
package machine
import "device/rp"
// RP2350B has additional pins.
const (
@@ -46,3 +48,28 @@ var (
PWM10 = getPWMGroup(10)
PWM11 = getPWMGroup(11)
)
// validPins confirms that the SPI pin selection is a legitimate one
// for the the 2350b chip.
func (spi *SPI) validPins(config SPIConfig) error {
var okSDI, okSDO, okSCK bool
switch spi.Bus {
case rp.SPI0:
okSDI = config.SDI == 0 || config.SDI == 4 || config.SDI == 16 || config.SDI == 20 || config.SDI == 32 || config.SDI == 36
okSDO = config.SDO == 3 || config.SDO == 7 || config.SDO == 19 || config.SDO == 23 || config.SDO == 35 || config.SDO == 39
okSCK = config.SCK == 2 || config.SCK == 6 || config.SCK == 18 || config.SCK == 22 || config.SCK == 34 || config.SCK == 38
case rp.SPI1:
okSDI = config.SDI == 8 || config.SDI == 12 || config.SDI == 24 || config.SDI == 28 || config.SDI == 40 || config.SDI == 44
okSDO = config.SDO == 11 || config.SDO == 15 || config.SDO == 27 || config.SDO == 31 || config.SDO == 43 || config.SDO == 47
okSCK = config.SCK == 10 || config.SCK == 14 || config.SCK == 26 || config.SCK == 30 || config.SCK == 42 || config.SCK == 46
}
switch {
case !okSDI:
return errSPIInvalidSDI
case !okSDO:
return errSPIInvalidSDO
case !okSCK:
return errSPIInvalidSCK
}
return nil
}
+3 -21
View File
@@ -108,7 +108,7 @@ func (spi *SPI) SetBaudRate(br uint32) error {
var prescale, postdiv uint32
freq := CPUFrequency()
for prescale = 2; prescale < 255; prescale += 2 {
if freq < (prescale+2)*256*br {
if uint64(freq) < uint64((prescale+2)*256)*uint64(br) {
break
}
}
@@ -165,27 +165,9 @@ func (spi *SPI) Configure(config SPIConfig) error {
config.SDI = SPI1_SDI_PIN
}
}
var okSDI, okSDO, okSCK bool
switch spi.Bus {
case rp.SPI0:
okSDI = config.SDI == 0 || config.SDI == 4 || config.SDI == 16 || config.SDI == 20
okSDO = config.SDO == 3 || config.SDO == 7 || config.SDO == 19 || config.SDO == 23
okSCK = config.SCK == 2 || config.SCK == 6 || config.SCK == 18 || config.SCK == 22
case rp.SPI1:
okSDI = config.SDI == 8 || config.SDI == 12 || config.SDI == 24 || config.SDI == 28
okSDO = config.SDO == 11 || config.SDO == 15 || config.SDO == 27
okSCK = config.SCK == 10 || config.SCK == 14 || config.SCK == 26
if err := spi.validPins(config); err != nil {
return err
}
switch {
case !okSDI:
return errSPIInvalidSDI
case !okSDO:
return errSPIInvalidSDO
case !okSCK:
return errSPIInvalidSCK
}
if config.Frequency == 0 {
config.Frequency = defaultBaud
}
+19 -12
View File
@@ -8,16 +8,21 @@ import "unsafe"
// code linked from other languages can allocate memory without colliding with
// our GC allocations.
var allocs = make(map[uintptr][]byte)
// Map of allocations, where the key is the allocated pointer and the value is
// the size of the allocation.
// TODO: make this a map[unsafe.Pointer]uintptr, since that results in slightly
// smaller binaries. But for that to work, unsafe.Pointer needs to be seen as a
// binary key (which it is not at the moment).
// See https://github.com/tinygo-org/tinygo/pull/4898 for details.
var allocs = make(map[*byte]uintptr)
//export malloc
func libc_malloc(size uintptr) unsafe.Pointer {
if size == 0 {
return nil
}
buf := make([]byte, size)
ptr := unsafe.Pointer(&buf[0])
allocs[uintptr(ptr)] = buf
ptr := alloc(size, nil)
allocs[(*byte)(ptr)] = size
return ptr
}
@@ -26,8 +31,8 @@ func libc_free(ptr unsafe.Pointer) {
if ptr == nil {
return
}
if _, ok := allocs[uintptr(ptr)]; ok {
delete(allocs, uintptr(ptr))
if _, ok := allocs[(*byte)(ptr)]; ok {
delete(allocs, (*byte)(ptr))
} else {
panic("free: invalid pointer")
}
@@ -48,18 +53,20 @@ func libc_realloc(oldPtr unsafe.Pointer, size uintptr) unsafe.Pointer {
// It's hard to optimize this to expand the current buffer with our GC, but
// it is theoretically possible. For now, just always allocate fresh.
buf := make([]byte, size)
// TODO: we could skip this if the new allocation is smaller than the old.
ptr := alloc(size, nil)
if oldPtr != nil {
if oldBuf, ok := allocs[uintptr(oldPtr)]; ok {
copy(buf, oldBuf)
delete(allocs, uintptr(oldPtr))
if oldSize, ok := allocs[(*byte)(oldPtr)]; ok {
oldBuf := unsafe.Slice((*byte)(oldPtr), oldSize)
newBuf := unsafe.Slice((*byte)(ptr), size)
copy(newBuf, oldBuf)
delete(allocs, (*byte)(oldPtr))
} else {
panic("realloc: invalid pointer")
}
}
ptr := unsafe.Pointer(&buf[0])
allocs[uintptr(ptr)] = buf
allocs[(*byte)(ptr)] = size
return ptr
}
+3
View File
@@ -93,6 +93,9 @@ func runtimePanicAt(addr unsafe.Pointer, msg string) {
trap()
}
if hasReturnAddr {
// Note: the string "panic: runtime error at " is also used in
// runtime_cortexm_hardfault.go. It is kept the same so that the string
// can be deduplicated by the compiler.
printstring("panic: runtime error at ")
printptr(uintptr(addr) - callInstSize)
printstring(": ")
+6
View File
@@ -57,6 +57,12 @@ func memzero(ptr unsafe.Pointer, size uintptr)
// the current stack pointer in a platform-independent way.
func stacksave() unsafe.Pointer
// Special LLVM intrinsic that returns the SP register on entry to the calling
// function.
//
//export llvm.sponentry.p0
func llvm_sponentry() unsafe.Pointer
//export strlen
func strlen(ptr unsafe.Pointer) uintptr
+3
View File
@@ -273,6 +273,9 @@ func nanosecondsToTicks(ns int64) timeUnit {
func sleepTicks(d timeUnit) {
for d != 0 {
ticks := uint32(d)
if d > 0xffff_ffff {
ticks = 0xffff_ffff
}
if !timerSleep(ticks) {
// Bail out early to handle a non-time interrupt.
return
+3
View File
@@ -266,6 +266,9 @@ func nanosecondsToTicks(ns int64) timeUnit {
func sleepTicks(d timeUnit) {
for d != 0 {
ticks := uint32(d)
if d > 0xffff_ffff {
ticks = 0xffff_ffff
}
if !timerSleep(ticks) {
return
}
+13 -27
View File
@@ -2,39 +2,25 @@
package runtime
import (
"unsafe"
)
// This function is called at HardFault.
// Before this function is called, the stack pointer is reset to the initial
// stack pointer (loaded from address 0x0) and the previous stack pointer is
// passed as an argument to this function. This allows for easy inspection of
// the stack the moment a HardFault occurs, but it means that the stack will be
// corrupted by this function and thus this handler must not attempt to recover.
//
// For details, see:
// https://community.arm.com/developer/ip-products/system/f/embedded-forum/3257/debugging-a-cortex-m0-hard-fault
// https://blog.feabhas.com/2013/02/developing-a-generic-hard-fault-handler-for-arm-cortex-m3cortex-m4/
//
//export handleHardFault
func handleHardFault(sp *interruptStack) {
print("fatal error: ")
if uintptr(unsafe.Pointer(sp)) < 0x20000000 {
print("stack overflow")
} else {
// TODO: try to find the cause of the hard fault. Especially on
// Cortex-M3 and higher it is possible to find more detailed information
// in special status registers.
print("HardFault")
}
print(" with sp=", sp)
if uintptr(unsafe.Pointer(&sp.PC)) >= 0x20000000 {
// Only print the PC if it points into memory.
// It may not point into memory during a stack overflow, so check that
// first before accessing the stack.
print(" pc=", sp.PC)
}
//export HardFault_Handler
func HardFault_Handler() {
// Obtain the stack pointer as it was on entry to the HardFault. It contains
// the registers that were pushed by the NVIC and that we can now read back
// to print the PC value at the time of the hard fault, for example.
sp := (*interruptStack)(llvm_sponentry())
// Note: by reusing the string "panic: runtime error at " we save a little
// bit in terms of code size as the string can be deduplicated.
print("panic: runtime error at ", sp.PC, ": HardFault with sp=", sp)
// TODO: try to find the cause of the hard fault. Especially on Cortex-M3
// and higher it is possible to find more detailed information in special
// status registers.
println()
abort()
}
@@ -18,8 +18,13 @@ const (
// See runtime_cortexm_hardfault.go
//
//go:export handleHardFault
func handleHardFault(sp *interruptStack) {
//export HardFault_Handler
func HardFault_Handler() {
// Obtain the stack pointer as it was on entry to the HardFault. It contains
// the registers that were pushed by the NVIC and that we can now read back
// to print the PC value at the time of the hard fault, for example.
sp := (*interruptStack)(llvm_sponentry())
fault := GetFaultStatus()
spValid := !fault.Bus().ImpreciseDataBusError()
+3
View File
@@ -78,6 +78,9 @@ func buffered() int {
func sleepTicks(d timeUnit) {
for d != 0 {
ticks := uint32(d) & 0x7fffff // 23 bits (to be on the safe side)
if d > 0x7fffff {
ticks = 0x7fffff
}
rtc_sleep(ticks)
d -= timeUnit(ticks)
}
+3
View File
@@ -81,6 +81,9 @@ func buffered() int {
func sleepTicks(d timeUnit) {
for d != 0 {
ticks := uint32(d) & 0x7fffff // 23 bits (to be on the safe side)
if d > 0x7fffff {
ticks = 0x7fffff
}
rtc_sleep(ticks)
d -= timeUnit(ticks)
}
+1
View File
@@ -212,6 +212,7 @@ func writeStream(stream *wasiStream, buf *byte, count uint, offset int64) int {
return -1
}
remaining -= len
src = src[len:]
}
return int(count)
+28 -5
View File
@@ -10,6 +10,7 @@ package testing
import (
"bytes"
"context"
"errors"
"flag"
"fmt"
@@ -78,6 +79,9 @@ type common struct {
tempDir string
tempDirErr error
tempDirSeq int32
ctx context.Context
cancelCtx context.CancelFunc
}
type logger struct {
@@ -152,6 +156,7 @@ func fmtDuration(d time.Duration) string {
// TB is the interface common to T and B.
type TB interface {
Cleanup(func())
Context() context.Context
Error(args ...interface{})
Errorf(format string, args ...interface{})
Fail()
@@ -307,6 +312,15 @@ func (c *common) Cleanup(f func()) {
c.cleanups = append(c.cleanups, f)
}
// Context returns a context that is canceled just before
// Cleanup-registered functions are called.
//
// Cleanup functions can wait for any resources
// that shut down on [context.Context.Done] before the test or benchmark completes.
func (c *common) Context() context.Context {
return c.ctx
}
// TempDir returns a temporary directory for the test to use.
// The directory is automatically removed by Cleanup when the test and
// all its subtests complete.
@@ -447,6 +461,9 @@ func (c *common) runCleanup() {
if cleanup == nil {
return
}
if c.cancelCtx != nil {
c.cancelCtx()
}
cleanup()
}
}
@@ -488,12 +505,15 @@ func (t *T) Run(name string, f func(t *T)) bool {
}
// Create a subtest.
ctx, cancelCtx := context.WithCancel(context.Background())
sub := T{
common: common{
output: &logger{logToStdout: flagVerbose},
name: testName,
parent: &t.common,
level: t.level + 1,
output: &logger{logToStdout: flagVerbose},
name: testName,
parent: &t.common,
level: t.level + 1,
ctx: ctx,
cancelCtx: cancelCtx,
},
context: t.context,
}
@@ -606,9 +626,12 @@ func runTests(matchString func(pat, str string) (bool, error), tests []InternalT
ok = true
ctx := newTestContext(newMatcher(matchString, flagRunRegexp, "-test.run", flagSkipRegexp))
runCtx, cancelCtx := context.WithCancel(context.Background())
t := &T{
common: common{
output: &logger{logToStdout: flagVerbose},
output: &logger{logToStdout: flagVerbose},
ctx: runCtx,
cancelCtx: cancelCtx,
},
context: ctx,
}
+11
View File
@@ -0,0 +1,11 @@
{
"inherits": [
"rp2350b"
],
"build-tags": ["pico2_ice"],
"serial-port": ["2e8a:000A"],
"default-stack-size": 8192,
"ldflags": [
"--defsym=__flash_size=4M"
]
}