fix release workflow

fixed an issue where debugger had an error when trying to dump registers. It tried to use get() method instead of get_f() method. Also changed format of floating point registers to show up as floats instead of Hex.

.git-blame-ignored-commits

@@ -0,0 +1,2 @@
+# introduces black formatting
+5515c7795cfd690d346aad10ce17b30acf914648

.github/workflows/ci-pytest.yml

@@ -0,0 +1,49 @@
+# This workflow will install Python dependencies, run tests and lint with a single version of Python
+# For more information see: https://help.github.com/actions/language-and-framework-guides/using-python-with-github-actions
+
+name: CI - Python-based Testing
+
+on:
+  # Trigger the workflow on push or pull request,
+  # but only for the master branch
+  push:
+    branches:
+      - master
+  pull_request:
+    branches:
+      - master
+
+jobs:
+  build:
+
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ['3.8','3.10']
+
+    steps:
+    - uses: actions/checkout@v3
+
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: ${{ matrix.python-version }}
+
+    - name: Upgrade pip
+      run: |
+        pip install --upgrade pip
+
+    - name: Install the package locally
+      run: pip install -r requirements.txt -r requirements-dev.txt
+
+    - name: Test with pytest
+      run: |
+        pytest -W error
+    - name: Test with lit
+      run: |
+        lit -v test/filecheck
+
+    #- name: Execute lit tests
+    #  run: |
+    #    export PYTHONPATH=$(pwd)
+    #    lit -v tests/filecheck/

.github/workflows/code-formatting.yml

@@ -0,0 +1,34 @@
+# This workflow check the format all files in the repository
+# * It checks that all nonempty files have a newline at the end
+# * It checks that there are no whitespaces at the end of lines
+# * It checks that Python files are formatted with black
+
+name: Code Formatting
+
+on:
+  pull_request:
+  push:
+    branches: [master]
+
+jobs:
+  code-formatting:
+
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ['3.10']
+
+    steps:
+    - uses: actions/checkout@v3
+
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: ${{ matrix.python-version }}
+
+    - name: Upgrade pip
+      run: |
+        pip install --upgrade pip
+
+    - name: Run code formatting checks with pre-commit
+      uses: pre-commit/action@v3.0.0

.github/workflows/jupyterlite.yml

@@ -0,0 +1,98 @@
+name: Deploy JupiterLite Page
+
+on:
+  # Trigger the workflow on push or pull request,
+  # but only for the master branch
+  push:
+    branches:
+      - master
+  pull_request:
+    branches:
+      - master
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout RiscEmu
+        uses: actions/checkout@v3
+        with:
+          path: riscemu
+
+      - name: Setup Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.10'
+
+      - name: Install dependencies
+        run: |
+          python -m pip install jupyterlite[all] libarchive-c build pyodide-build
+
+      - name: Build RiscEmu source distribution
+        run: |
+          cd riscemu
+          python setup.py sdist
+
+      # Pyodide is cached, so cloned only if not present in the cache, otherwise
+      # just checked out to whatever desired version and partially rebuilt.
+
+      - name: Restore cached Pyodide tree
+        id: cache-pyodide
+        uses: actions/cache@v3
+        with:
+          path: pyodide
+          key: pyodide
+
+      - name: Clone pyodide if not cached
+        if: steps.cache-pyodide.outputs.cache-hit != 'true'
+        run: git clone https://github.com/pyodide/pyodide.git
+
+      # Clean the xDSL and FrozenList package folders, generate their skeletons
+      # and do the necessary updates before building.
+      - name: Build custom Pyodide distribution
+        run: |
+
+          cd pyodide
+          git fetch --all
+          git checkout 0.22.0a3
+          python -m pip install -r requirements.txt
+          sudo apt update && sudo apt install f2c
+
+          rm -rf packages/riscemu
+          pyodide skeleton pypi riscemu
+
+          PYODIDE_PACKAGES="riscemu" make
+
+      - name: Build the JupyterLite site
+        run: |
+          mkdir content
+          cp riscemu/docs/* content -r
+          cp riscemu/examples content -r
+
+          rm -rf pyodide/pyodide
+          mkdir pyodide/pyodide
+          mv pyodide/dist pyodide/pyodide/pyodide
+
+          python -m jupyter lite build --contents content --pyodide pyodide/pyodide
+
+      - name: Upload artifact
+        uses: actions/upload-pages-artifact@v1
+        with:
+          path: ./_output
+
+  deploy:
+    needs: build
+    if: github.ref == 'refs/heads/master' && github.event_name == 'push'
+    permissions:
+      pages: write
+      id-token: write
+
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+
+    runs-on: ubuntu-latest
+    steps:
+      - name: Deploy to GitHub Pages
+        id: deployment
+        uses: actions/deploy-pages@v1

.github/workflows/pythonpublish.yml

@@ -0,0 +1,22 @@
+name: Upload to PyPI
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  deploy:
+    runs-on: ubuntu-latest
+    environment: publishing
+    permissions:
+      # IMPORTANT: this permission is mandatory for trusted publishing
+      id-token: write
+    steps:
+      # install build package and build dist
+      - name: Build distribution
+        run: >-
+          python3 -m pip install build --user &&
+          python3 -m build --sdist --wheel --outdir dist/
+      # retrieve your distributions here
+      - name: Publish package distributions to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1

.gitignore

@@ -1,6 +1,8 @@
-venv
 __pycache__
 .mypy_cache
-dist/
-riscemu.egg-info
-build/
+.pytest_cache
+
+/venv*
+/dist
+/riscemu.egg-info
+/build

.idea/inspectionProfiles/profiles_settings.xml

@@ -3,4 +3,4 @@
     <option name="USE_PROJECT_PROFILE" value="false" />
     <version value="1.0" />
   </settings>
-</component>
+</component>

.idea/misc.xml

@@ -1,4 +1,4 @@
 <?xml version="1.0" encoding="UTF-8"?>
 <project version="4">
-  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.8 (riscemu)" project-jdk-type="Python SDK" />
-</project>
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.10 (riscemu)" project-jdk-type="Python SDK" />
+</project>

.idea/modules.xml

@@ -5,4 +5,4 @@
       <module fileurl="file://$PROJECT_DIR$/.idea/riscemu.iml" filepath="$PROJECT_DIR$/.idea/riscemu.iml" />
     </modules>
   </component>
-</project>
+</project>

.idea/riscemu.iml

@@ -2,6 +2,7 @@
 <module type="PYTHON_MODULE" version="4">
   <component name="NewModuleRootManager">
     <content url="file://$MODULE_DIR$">
+      <sourceFolder url="file://$MODULE_DIR$/riscemu" isTestSource="false" />
       <sourceFolder url="file://$MODULE_DIR$/test" isTestSource="true" />
       <excludeFolder url="file://$MODULE_DIR$/venv" />
       <excludeFolder url="file://$MODULE_DIR$/dist" />
@@ -9,7 +10,7 @@
       <excludeFolder url="file://$MODULE_DIR$/riscemu.egg-info" />
       <excludeFolder url="file://$MODULE_DIR$/build" />
     </content>
-    <orderEntry type="inheritedJdk" />
+    <orderEntry type="jdk" jdkName="Python 3.10 (riscemu)" jdkType="Python SDK" />
     <orderEntry type="sourceFolder" forTests="false" />
   </component>
-</module>
+</module>

.idea/vcs.xml

@@ -3,4 +3,4 @@
   <component name="VcsDirectoryMappings">
     <mapping directory="$PROJECT_DIR$" vcs="Git" />
   </component>
-</project>
+</project>

.pre-commit-config.yaml

@@ -0,0 +1,12 @@
+# See https://pre-commit.com for more information
+# See https://pre-commit.com/hooks.html for more hooks
+repos:
+-   repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v3.2.0
+    hooks:
+    -   id: trailing-whitespace
+    -   id: end-of-file-fixer
+-   repo: https://github.com/psf/black
+    rev: 23.3.0
+    hooks:
+    -   id: black

CHANGELOG.md

@@ -1,5 +1,29 @@
 # Changelog
 
+## 2.1.1
+
+ - Bugfix: Fix some errors in the RV32F implementation (thanks @adutilleul)
+ - Bugfix: Fix how floats are printed in the register view (thanks @KGrykiel)
+ - Bugfix: Fix missing support for infinite registers in load/store ins (thanks @superlopuh)
+
+## 2.1.0
+
+ - Added a very basic libc containing a `crt0.s`, and a few functions
+   such as `malloc`, `rand`, and `memcpy`.
+ - Added a subset of the `mmap2` syscall (code 192) to allocate new memory
+ - Refactored the launching code to improve using riscemu from code
+ - Added an option to start with the provided libc: `-o libc`
+ - Added floating point support (enabled by default). The RV32F extension is now available
+
+## 2.0.5
+
+ - Added unlimited register mode with `-o unlimited_regs`
+
+## 2.0.4
+
+ - Bugfix: fix a sign issue in instruction parsing for `rd, rs, rs` format
+ - Bugfix: respect `conf.debug_instruction` setting
+
 ## 2.0.3 - 2022-04-18
 
  - Syscalls: cleaned up formatting and added instructions for extensions
@@ -9,7 +33,7 @@
  - Fixed bug where wrong parts of section would be printed in mmu.dump()
  - Removed tests for bind_twos_complement as the function is now redundant with the introduction of Int32
  - Fixed address translation error for sections without symbols
- - Changed verbosity level at which start and end of CPU are printed, added prints for start and stack loading 
+ - Changed verbosity level at which start and end of CPU are printed, added prints for start and stack loading
 
 ## 2.0.2
 
@@ -23,6 +47,6 @@
 
 ## 2.0.0
 
- - Correct handling of 32 bit overflows and underflows 
+ - Correct handling of 32 bit overflows and underflows
  - Complete revamp of internal data structures
  - Completely reworked how assembly is parsed

LICENSE

@@ -18,4 +18,4 @@ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+SOFTWARE.

README.md

@@ -10,7 +10,7 @@ This emulator contains:
 * RISC-V Assembly loader
 * Emulation for most parts of the basic RISC-V instruction set and the M and A extensions
 * Naive memory emulator
-* Basic implementation of some syscalls 
+* Basic implementation of some syscalls
 * A debugging environment
 
 ## Installation:
@@ -31,7 +31,9 @@ Hello world
 Program exited with code 0
 ```
 
-The [`read` syscall](docs/syscalls.md) defaults to readline behaviour. Reading "true chunks" (ignoring newlines) is currently not supported.  
+If you want to run it from a python script, here is [an online demo](https://AntonLydike.github.io/riscemu/lab/index.html?path=PythonDemo.ipynb).
+
+The [`read` syscall](docs/syscalls.md) defaults to readline behaviour. Reading "true chunks" (ignoring newlines) is currently not supported.
 
 See the docs on [asembly](docs/assembly.md) for more detail on how to write assembly code for this emulator.
 See the [list of implemented syscalls](docs/syscalls.md) for more details on how to syscall.
@@ -69,10 +71,10 @@ disable_io              Disallow reading/writing from stdin/stdout/stderr
 --instruction-sets INSTRUCTION_SETS: (-is)
                         A list of comma separated instruction sets you want to load:
                         Currently implemented: RV32I, RV32M
-``` 
+```
 
-If multiple files are specified, all are loaded into memeory, but only the last one is executed. This might be improved 
-later, maybe the `_init` section of each binary is executed before the main loop starts? 
+If multiple files are specified, all are loaded into memeory, but only the last one is executed. This might be improved
+later, maybe the `_init` section of each binary is executed before the main loop starts?
 
 If `stack_size` is greater than zero, a stack is allocated and initialized, with the `sp` register pointing to the end of the stack.
 
@@ -88,17 +90,17 @@ See [docs/debugging.md](docs/debugging.md) for more info.
 Check out the [documentation](https://riscemu.readthedocs.io/en/latest/riscemu.html).
 
 ## Accessing local documentation:
-To generate your local documentation, first install everything in `sphinx-docs/requirements.txt`. Then run `./generate-docs.sh`, which will 
+To generate your local documentation, first install everything in `sphinx-docs/requirements.txt`. Then run `./generate-docs.sh`, which will
 generate and make all doc files for you. Finally, you can open the docs locall by runnint `open sphinx-docs/build/html/index.html`.
 
 ## Resources:
+  * RISC-V Programmers Handbook: https://github.com/riscv-non-isa/riscv-asm-manual/blob/master/riscv-asm.md
   * Pseudo ops: https://www.codetd.com/article/8981522
   * detailed instruction definition: https://msyksphinz-self.github.io/riscv-isadoc/html/rvi.html#add
   * RISC-V reference card: https://www.cl.cam.ac.uk/teaching/1617/ECAD+Arch/files/docs/RISCVGreenCardv8-20151013.pdf
-  
+
 ## TODO:
  * Correctly handle 12 and 20 bit immediate (currently not limited to bits at all)
  * Add a cycle limit to the options and CPU to catch infinite loops
  * Move away from `print` and use `logging.logger` instead
  * Writer proper tests
- 

docs/PythonDemo.ipynb

@@ -0,0 +1,53 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Using RiscEmu from Python\n",
+    "\n",
+    "Here is how you can run some assembly through RiscEmu from Python.\n",
+    "\n",
+    "This example is using [this fibonacci assembly](examples/fibs.asm)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from riscemu import RunConfig, UserModeCPU, RV32I, RV32M, AssemblyFileLoader\n",
+    "\n",
+    "cfg = RunConfig(debug_instruction=False, verbosity=50)\n",
+    "cpu = UserModeCPU((RV32I, RV32M), cfg)\n",
+    "\n",
+    "loader = AssemblyFileLoader.instantiate('examples/fibs.asm', [])\n",
+    "cpu.load_program(loader.parse())\n",
+    "\n",
+    "cpu.launch(cpu.mmu.programs[-1], True)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.8"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}

docs/assembly.md

@@ -30,8 +30,8 @@ The following pseudo-ops are implemented as of yet:
 * `.global <name>` mark symbol `<name>` as a global symbol. It is available from all loaded programs
 * `.align <bytes>` currently a nop as cpu does not care about alignment as of now
 
-## Sections: 
-Currently only these three sections are supported: 
+## Sections:
+Currently only these three sections are supported:
 * `data` read-write data (non-executable)
 * `rodata` read-only data (non-executable)
 * `text` executable data (read-only)

docs/debugging.md

@@ -37,7 +37,7 @@ uses the `ebreak` instruction to open the debugger. Let's run it and see what ha
 [MMU] Successfully loaded: LoadedExecutable[examples/fibs.asm](base=0x00000100, size=72bytes, sections=data text, run_ptr=0x00000138)
 [CPU] Started running from 0x00000138 (examples/fibs.asm)
 Debug instruction encountered at 0x0000013C
->>> 
+>>>
 ```
 
 In this interactive session, you have access to the cpu, registers, memory and syscall interface. You can look into everything,
@@ -58,8 +58,8 @@ and most common tasks should have helper methods for them.
   * `get(name)` get register content
   * `set(name, val)` set register content
 * `cpu`:
-  * The CPU has the `pc` attribute and `cycle` attribute. Others won't be useful in this context. 
-  
+  * The CPU has the `pc` attribute and `cycle` attribute. Others won't be useful in this context.
+
 **Available helpers are:**
 
 * `dump(regs | addr)` dumps either registers or memory address
@@ -71,4 +71,4 @@ and most common tasks should have helper methods for them.
 
 Example:
 
-![debuggin the fibs program](debug-session.png)
+![debuggin the fibs program](debug-session.png)

docs/internal-structure.md

@@ -1,19 +1,19 @@
 # Internal Structure
 
 ## Loading assembly files:
-In order to load an assembly file, you need to instantiate a CPU with the capabilities you want. Loading an assembly 
+In order to load an assembly file, you need to instantiate a CPU with the capabilities you want. Loading an assembly
 file is the done in multiple steps:
 
 
 * An `RiscVInput` is created, this represents the file internally
 * An `RiscVTokenizer` is created by calling `cpu.get_tokenizer()`.
 * The input is tokenized by calling `.tokenize()` on the tokenizer.
-* The tokens can then be converted to an Executable, this will then 
+* The tokens can then be converted to an Executable, this will then
   hold all the information such as name, sections, symbols, etc.
   This is done by creating an `ExecutableParser(tk: RiscVTokenizer)`
   and the calling `parse()`.
 * Now you have a representation of the assembly file that can be loaded
-  into memory by calling `cpu.load(executable)`, this will internally 
+  into memory by calling `cpu.load(executable)`, this will internally
   construct a `LoadedExecutable`, which represents the actual memory
   regions the executable contains (and some meta information such as
   symbols).
@@ -30,4 +30,3 @@ Creating a cpu with certain instruction sets is done by passing the CPU construc
 ```
 cpu = CPU(config, [RV32I, RV32M])
 ```
-

docs/libraries.md

@@ -13,4 +13,3 @@ You can include the libraries by adding them as arguments (before your main asse
 ```
 
 These libraries are no where near a stable state, so documentation will be scarce. Your best bet would be to `grep` for functionality. Sorry!
-

docs/syscalls.md

@@ -7,7 +7,7 @@ Performing a syscall is quite simple:
     ; set syscall args:
     addi a0, zero, 1    ; exit with code 1
     ; invode syscall handler
-    scall               
+    scall
 ```
 
 The global symbols (e.g. `SCALL_READ`) are loaded by default. If you specify the option `no_syscall_symbols`, they will be omitted.
@@ -47,4 +47,4 @@ Requires flag `--scall-fs` to be set to True
 
 # Extending these syscalls
 
-You can implement your own syscall by adding its code to the `SYSCALLS` dict in the [riscemu/syscalls.py](../riscemu/syscall.py) file, creating a mapping of a syscall code to a name, and then implementing that syscall name in the SyscallInterface class further down that same file. Each syscall method should have the same signature: `read(self, scall: Syscall)`. The `Syscall` object gives you access to the cpu, through which you can access registers and memory. You can look at the `read` or `write` syscalls for further examples. 
+You can implement your own syscall by adding its code to the `SYSCALLS` dict in the [riscemu/syscalls.py](../riscemu/syscall.py) file, creating a mapping of a syscall code to a name, and then implementing that syscall name in the SyscallInterface class further down that same file. Each syscall method should have the same signature: `read(self, scall: Syscall)`. The `Syscall` object gives you access to the cpu, through which you can access registers and memory. You can look at the `read` or `write` syscalls for further examples.

examples/fibs.asm

@@ -1,25 +1,23 @@
-; Example program (c) by Anton Lydike
-; this calculates the fibonacci sequence and stores it in ram
+// Example program (c) by Anton Lydike
+// this calculates the fibonacci sequence and stores it in ram
 
         .data
 fibs:   .space 56
 
         .text
 main:
-        addi    s1, zero, 0     ; storage index
-        addi    s2, zero, 56    ; last storage index
-        addi    t0, zero, 1     ; t0 = F_{i}
-        addi    t1, zero, 1     ; t1 = F_{i+1}
-        ebreak                  ; launch debugger
+        addi    s1, zero, 0     // storage index
+        addi    s2, zero, 56    // last storage index
+        addi    t0, zero, 1     // t0 = F_{i}
+        addi    t1, zero, 1     // t1 = F_{i+1}
 loop:
-        sw      t0, fibs(s1)    ; save
-        add     t2, t1, t0      ; t2 = F_{i+2}
-        addi    t0, t1, 0       ; t0 = t1
-        addi    t1, t2, 0       ; t1 = t2
-        addi    s1, s1, 4       ; increment storage pointer
-        blt     s1, s2, loop    ; loop as long as we did not reach array length
-        ; exit gracefully
-        ebreak                  ; launch debugger
+        sw      t0, fibs(s1)    // save
+        add     t2, t1, t0      // t2 = F_{i+2}
+        addi    t0, t1, 0       // t0 = t1
+        addi    t1, t2, 0       // t1 = t2
+        addi    s1, s1, 4       // increment storage pointer
+        blt     s1, s2, loop    // loop as long as we did not reach array length
+        // exit gracefully
         addi    a0, zero, 0
         addi    a7, zero, 93
-        scall                   ; exit with code 0
+        scall                   // exit with code 0

examples/lib/libstring.asm

@@ -1,154 +0,0 @@
-; string operations in RISC-V Assembly
-; All function restore all registers after use (even caller-saved ones)
-
-.global     NULL
-; A global constant, points to the empty string
-
-
-.global     strlen
-; size_t libstr_strlen(char* str)
-; return the length of str
-
-
-.global     strncpy
-; char *strncpy(char *dest, const char *src, size_t n)
-; copy n bytes from source into dest. If source ends before n bytes, the rest is filled with null-bytes
-; returns pointer to dest
-
-
-.global     strcpy
-; char *strncpy(char *dest, const char *src)
-; copy string src into dest, including null terminator
-; returns pointer to dest
-
-
-.global     memchr
-; void *memchr(const void *str, char c, size_t n)
-; search vor the first occurance of c in str
-; returns a pointer to the first occurance, or NULL
-
-
-.global     memset
-; void *memset(void *str, char c, size_t n)
-; copies the character c to the first n characters of  str.
-
-
-;.global     memcmp
-;.global     memcpy
-;.global     strcat
-
-; Create NullPtr constant
-.set        NULL, 0x00
-
-
-            .text
-strlen:
-; size_t strlen(char* str)
-            ; push s1, s2 to the stack
-            sw   s1, sp, -4
-            sw   s2, sp, -8
-            ; since no subroutines are called, we don't need to increment or decrement the sp
-            addi s2, zero, -1   ; length (since null byte is counted by this method, we return len - 1)
-__strlen_loop:
-            lb   s1, a0, 0      ; read character
-            addi s2, s2, 1      ; increment number bytes read
-            addi a0, a0, 1
-            bne  s1, zero, __strlen_loop
-            ; we are done, set return value in a0
-            add  a0, zero, s2
-            ; pop s1, s2, from stack
-            lw   s1, sp, -4
-            lw   s2, sp, -8
-            ret
-
-strncpy:
-; char *strncpy(char *dest, const char *src, size_t n)
-; copy size bytes from source to dest
-            sw   s1, sp, -4     ; push s1 to the stack
-            sw   s2, sp, -8     ; push s1 to the stack
-            add  s1, a0, zero   ; save dest pointer for return
-__strncpy_loop:
-            beq  a2, zero, __strncpy_end
-            ; copy byte
-            lb   s2, a1, 0      ; read first byte from src
-            sb   s2, a0, 0      ; write first byte to dest
-            ; increment pointers
-            addi a0, a0, 1
-            addi a1, a1, 1
-            ; one less byte to copy
-            addi a2, a2, -1
-            ; if we read the terminating byte, jump to fill code
-            beq  s2, zero, __strncpy_fill
-            ; otherwise continue copying
-            j    __strncpy_loop
-__strncpy_fill:
-            ; fill remaining space with 0 bytes
-            ; if no bytes left, stop filling
-            beq  a2, zero, __strncpy_end
-            sb   zero, a0, 0
-            addi a0, a0, 1
-            addi a2, a2, -1
-            j    __strncpy_fill
-__strncpy_end:
-            ; set return value
-            add  a0, zero, s1
-            ; pop s1, s2 from stack
-            lw   s1, sp, -4
-            lw   s2, sp, -8
-            ret
-
-
-strcpy:
-; char *strncpy(char *dest, const char *src)
-            sw   s1, sp, -4     ; push s1 to the stack
-            sw   s2, sp, -8     ; push s1 to the stack
-            add  s1, a0, zero   ; save dest pointer for return
-__strcpy_loop:
-            ; copy byte
-            lb   s2, a1, 0      ; read first byte from src
-            sb   s2, a0, 0      ; write first byte to dest
-            ; increment pointers
-            addi a0, a0, 1
-            addi a1, a1, 1
-            bne  s2, zero, __strcpy_loop
-            ; we are done copying, return
-            ; set return value
-            add  a0, zero, s1
-            ; pop s1, s2 from stack
-            lw   s1, sp, -4
-            lw   s2, sp, -8
-            ret
-
-memchr:
-; void *memchr(const void *str, char c, size_t n)
-            sw   s1, sp, -4     ; push s1 to the stack
-__memchr_loop:
-            beq  a2, zero, __memchr_ret_null
-            lb   s1, a0, 0
-            addi a0, a0, 1  ; let a0 point to the next byte
-            addi a2, a2, -1 ; decrement bytes to copy by 1
-            bne  s1, s2, __memchr_loop
-            ; return pointer to prev byte (as the prev byte actually matched a1)
-            addi a0, a0, -1
-            ; pop s1, s2 from stack
-            lw   s1, sp, -4
-            ret
-__memchr_ret_null:
-            ; nothing found, return nullptr
-            addi a0, zero, NULL
-            lw   s1, sp, -4
-            ret
-
-
-memset:
-; void *memset(void *str, char c, size_t n)
-__memset_loop:
-            beq  a2, zero, __memset_ret
-            sb   a1, a0, 0
-            addi a0, a0, 1
-            addi a2, a2, -1
-            j    __memset_loop
-__memset_ret:
-            ret
-
-

examples/malloc.asm

@@ -0,0 +1,44 @@
+// example of a simple memory allocation
+// we use the mmap2 syscall for this
+
+.text
+        // call mmap2
+        li  a0, 0       // addr = 0, let OS choose address
+        li  a1, 4096    // size
+        li  a2, 3       // PROT_READ   | PROT_WRITE
+        li  a3, 5       // MAP_PRIVATE | MAP_ANONYMOUS
+        li  a7, SCALL_MMAP2
+        ecall           // invoke syscall
+
+        li  t0, -1      // exit if unsuccessful
+        beq a0, t0, _exit
+
+        // print address
+        print.uhex a0
+        # we can look at the state of the mmu here:
+        ebreak
+        # > mmu.sections
+        # InstructionMemorySection[.text] at 0x00000100
+        # BinaryDataMemorySection[.stack] at 0x00000170
+        # BinaryDataMemorySection[.data.runtime-allocated] at 0x00080170
+
+        sw  t0, 144(a0)
+        sw  t0, 0(a0)
+        sw  t0, 8(a0)
+        sw  t0, 16(a0)
+        sw  t0, 32(a0)
+        sw  t0, 64(a0)
+        sw  t0, 128(a0)
+        sw  t0, 256(a0)
+        sw  t0, 512(a0)
+        sw  t0, 1024(a0)
+        sw  t0, 2048(a0)
+        sw  t0, 4000(a0)
+
+        lw  t1, 128(a0)
+        print.uhex t0
+        ebreak
+
+_exit:
+        li a7, 93
+        ecall

examples/mul-float.asm

@@ -0,0 +1,11 @@
+main:
+    li a1, 1084227584
+    li a2, 1082130432
+    fcvt.s.wu ft0, a1
+    fcvt.s.wu ft1, a2
+    fmul.s ft6, ft0, ft1
+    print.float ft6
+    // exit gracefully
+    addi    a0, zero, 0
+    addi    a7, zero, 93
+    scall                   // exit with code 0

examples/static-data.asm

@@ -0,0 +1,20 @@
+.data
+my_data:
+.word   0x11223344, 0x55667788, 0x9900aabb, 0xccddeeff
+
+.text
+main:
+        // load base address into t0
+        la  t0, my_data
+        // begin loading words and printing them
+        lw  a0, 0(t0)
+        print.uhex a0
+        lw  a0, 4(t0)
+        print.uhex a0
+        lw  a0, 8(t0)
+        print.uhex a0
+        lw  a0, 12(t0)
+        print.uhex a0
+        // exit
+        li  a7, 93
+        ecall

generate-docs.sh

@@ -35,4 +35,4 @@ cd sphinx-docs
 
 make html
 
-# xdg-open build/html/index.html
+# xdg-open build/html/index.html

libc/README.md

@@ -0,0 +1,29 @@
+# RiscEmu LibC
+
+This is a very basic implementation of libc in risc-v assembly, meant specifically for the riscemu emulator.
+
+This is currently very incomplete, only a handful of methods are implemented, and most of them pretty basic.
+
+## Contents:
+
+### `stdlib.s`
+
+Basic implementations of:
+
+ - `malloc`/`free` (that leaks memory)
+ - `rand`/`srand` (using xorshift)
+ - `exit`/`atexit` (supporting up to 8 exit handlers)
+
+### `string.s`
+
+Somewhat nice implementations of:
+
+ - `strlen`
+ - `strncpy`
+ - `strcpy`
+ - `memchr`
+ - `memset` (very basic byte-by-byte copy)
+
+## Correctness:
+
+This library is only lightly tested, so be careful and report bugs when you find them!

libc/crt0.s

@@ -0,0 +1,14 @@
+// A minimal crt0.s that works along the stdlib.s file provided to give
+// some resemblance of a functioning compilation target :)
+//
+// Copyright (c) 2023 Anton Lydike
+// SPDX-License-Identifier: MIT
+
+.text
+
+.globl _start
+_start:
+        // TODO: read argc, argv from a0, a1
+        // maybe even find envptr?
+        jal main
+        jal exit

libc/stdlib.s

@@ -0,0 +1,178 @@
+// A very basic implementation of a stdlib.h but in assembly.
+// should(tm) work with riscemu.
+//
+// Copyright (c) 2023 Anton Lydike
+// SPDX-License-Identifier: MIT
+
+.data
+
+_rand_seed:
+.word   0x76767676
+_atexit_calls:
+// leave room for 8 atexit handlers here for now
+.word   0x00, 0x00, 0x00, 0x00
+.word   0x00, 0x00, 0x00, 0x00
+_atexit_count:
+.word   0x00
+
+_malloc_base_ptr:
+// first word is a pointer to some space
+// second word is the offset inside that space
+// space is always MALLOC_PAGE_SIZE bytes
+.word   0x00, 0x00
+.equ    MALLOC_PAGE_SIZE 4069
+
+.text
+
+// malloc/free
+
+.globl malloc
+.globl free
+
+// malloc(size_t size)
+malloc:
+        // set aside size in s0
+        sw  s0, -4(sp)
+        mv  a0, s0
+        la  t0, _malloc_base_ptr
+        lw  t1, 0(t0)
+        beq t1, zero, _malloc_init
+_malloc_post_init:
+        // if we are here, we always have
+        // t0 = (&_malloc_base_ptr)
+        // t1 = *(&_malloc_base_ptr)
+        // new we load
+        // t2 = base_ptr_offset
+        lw  t2, 4(t0)
+        // add allocated size to offset
+        add t2, t2, s0
+        // check for overflow
+        li  t4, MALLOC_PAGE_SIZE
+        bge t2, t4, _malloc_fail
+        // save the new offset
+        sw  t2, 4(t0)
+        // calculate base_ptr + offset
+        add a0, t2, t1
+        // return that
+        lw  s0, -4(sp)
+        ret
+
+_malloc_init:
+        // call mmap2()
+        li  a0, 0       // addr = 0, let OS choose address
+        li  a1, 4096    // size
+        li  a2, 3       // PROT_READ   | PROT_WRITE
+        li  a3, 5       // MAP_PRIVATE | MAP_ANONYMOUS
+        li  a7, SCALL_MMAP2
+        ecall           // invoke syscall
+        // check for error code
+        li  t0, -1
+        beq a0, t0, _malloc_fail
+        // if succeeded, load &_malloc_base_ptr
+        la  t0, _malloc_base_ptr
+        // put value of _malloc_base_ptr into t1
+        mv  a0, t1
+        // save base ptr to _malloc_base_ptr
+        sw  t1, 0(t0)
+        // jump to post_init
+        j   _malloc_post_init
+_malloc_fail:
+        li a0, 0
+        ret
+
+// free is a nop, that's valid, but not very good^^
+free:
+        ret
+
+
+
+// exit, atexit
+.globl exit
+.globl atexit
+
+// we can happily use saved registers here because we don't care at all if we
+// destroy the calling registers. This is __noreturn__ anyways!
+// register layout:
+// s0 = &_atexit_count
+// s2 = &_atexit_calls
+// s1 = updated value of atexit
+// s3 = exit code
+exit:
+        // save exit code to s3
+        mv      s3, a0
+_exit_start:
+        la      s0, _atexit_count       // s0 = &_atexit_count
+        lw      s1, 0(s0)               // s1 = *(&_atexit_count)
+        // exit if no atexit() calls remain
+        beq     s1, zero, _exit
+        // decrement
+        addi    s1, s1, -4              // s1--
+        // save decremented value
+        sw      s1, 0(s0)               // _atexit_count = s1
+        li      s2, _atexit_calls
+        add     s1, s1, s2              // s1 = &_atexit_calls + (s1)
+        lw      s1, 0(s1)               // s1 = *s1
+        la      ra, _exit_start         // set ra up to point to exit
+        jalr    zero, s1, 0             // jump to address in s1
+        // jalr will call the other function, which will then return back
+        // to the beginning of exit.
+_exit:
+        mv      a0, s3
+        li      a7, 93
+        ecall
+
+// atexit a0 = funcptr
+atexit:
+        sw      t0, -4(sp)
+        sw      t2, -8(sp)
+        // load _atexit_count
+        la      t0, _atexit_count
+        lw      t2, 0(t0)
+        // if >= 8, fail
+        li      t1, 8
+        bge     t2, t1, _atexit_fail
+        // increment atexit_count by 4 (one word)
+        addi    t2, t2, 4
+        sw      t2, 0(t0)
+        // load address of _atexit_calls
+        la      t0, _atexit_calls
+        // add new _atexit_count to _atexit_calls
+        add     t0, t0, t2
+        sw      a0, -4(t0)
+        li      a0, 0
+        lw      t0, -4(sp)
+        lw      t2, -8(sp)
+        ret
+
+_atexit_fail:
+        li      a0, -1
+        lw      s0, -4(sp)
+        lw      s1, -8(sp)
+        ret
+
+
+
+
+
+
+// rand, srand
+
+.globl  rand
+.globl  srand
+
+// simple xorshift rand implementation
+rand:
+        // load seed
+        la  t1, _rand_seed
+        lw  a0, 0(t1)
+        // three rounds of shifts:
+        sll a0, t0, 13          // x ^= x << 13;
+        srl a0, t0, 17          // x ^= x >> 17;
+        sll a0, t0, 5           // x ^= x << 5;
+	    sw  a0, 0(t1)
+	    ret
+
+srand:
+        la  t1, _rand_seed
+        sw  a0, 0(t1)
+        ret

libc/string.s

@@ -0,0 +1,154 @@
+// string operations in RISC-V Assembly
+//
+// Copyright (c) 2023 Anton Lydike
+// SPDX-License-Identifier: MIT
+
+// Create NullPtr constant
+.equ        NULL, 0x00
+.global     NULL
+
+
+.global     strlen
+// size_t libstr_strlen(char* str)
+// return the length of str
+
+
+.global     strncpy
+// char *strncpy(char *dest, const char *src, size_t n)
+// copy n bytes from source into dest. If source ends before n bytes, the rest is filled with null-bytes
+// returns pointer to dest
+
+
+.global     strcpy
+// char *strncpy(char *dest, const char *src)
+// copy string src into dest, including null terminator
+// returns pointer to dest
+
+
+.global     memchr
+// void *memchr(const void *str, char c, size_t n)
+// search vor the first occurance of c in str
+// returns a pointer to the first occurance, or NULL
+
+
+.global     memset
+// void *memset(void *str, char c, size_t n)
+// copies the character c to the first n characters of  str.
+
+
+// missing implementations
+//.global     memcmp
+//.global     memcpy
+//.global     strcat
+
+
+.text
+strlen:
+// size_t strlen(char* str)
+            // push s1, s2 to the stack
+            sw   s1, sp, -4
+            sw   s2, sp, -8
+            // since no subroutines are called, we don't need to increment or decrement the sp
+            addi s2, zero, -1   // length (since null byte is counted by this method, we return len - 1)
+__strlen_loop:
+            lb   s1, a0, 0      // read character
+            addi s2, s2, 1      // increment number bytes read
+            addi a0, a0, 1
+            bne  s1, zero, __strlen_loop
+            // we are done, set return value in a0
+            add  a0, zero, s2
+            // pop s1, s2, from stack
+            lw   s1, sp, -4
+            lw   s2, sp, -8
+            ret
+
+strncpy:
+// char *strncpy(char *dest, const char *src, size_t n)
+// copy size bytes from source to dest
+            sw   s1, sp, -4     // push s1 to the stack
+            sw   s2, sp, -8     // push s1 to the stack
+            add  s1, a0, zero   // save dest pointer for return
+__strncpy_loop:
+            beq  a2, zero, __strncpy_end
+            // copy byte
+            lb   s2, a1, 0      // read first byte from src
+            sb   s2, a0, 0      // write first byte to dest
+            // increment pointers
+            addi a0, a0, 1
+            addi a1, a1, 1
+            // one less byte to copy
+            addi a2, a2, -1
+            // if we read the terminating byte, jump to fill code
+            beq  s2, zero, __strncpy_fill
+            // otherwise continue copying
+            j    __strncpy_loop
+__strncpy_fill:
+            // fill remaining space with 0 bytes
+            // if no bytes left, stop filling
+            beq  a2, zero, __strncpy_end
+            sb   zero, a0, 0
+            addi a0, a0, 1
+            addi a2, a2, -1
+            j    __strncpy_fill
+__strncpy_end:
+            // set return value
+            add  a0, zero, s1
+            // pop s1, s2 from stack
+            lw   s1, sp, -4
+            lw   s2, sp, -8
+            ret
+
+
+strcpy:
+// char *strcpy(char *dest, const char *src)
+            sw   s1, sp, -4     // push s1 to the stack
+            sw   s2, sp, -8     // push s1 to the stack
+            add  s1, a0, zero   // save dest pointer for return
+__strcpy_loop:
+            // copy byte
+            lb   s2, a1, 0      // read first byte from src
+            sb   s2, a0, 0      // write first byte to dest
+            // increment pointers
+            addi a0, a0, 1
+            addi a1, a1, 1
+            bne  s2, zero, __strcpy_loop
+            // we are done copying, return
+            // set return value
+            add  a0, zero, s1
+            // pop s1, s2 from stack
+            lw   s1, sp, -4
+            lw   s2, sp, -8
+            ret
+
+memchr:
+// void *memchr(const void *str, char c, size_t n)
+            sw   s1, sp, -4     // push s1 to the stack
+            andi a1, a1, 0xff   // trim a1 to be byte-sized
+__memchr_loop:
+            beq  a2, zero, __memchr_ret_null
+            lb   s1, a0, 0
+            addi a0, a0, 1  // let a0 point to the next byte
+            addi a2, a2, -1 // decrement bytes to copy by 1
+            bne  s1, a1, __memchr_loop
+            // return pointer to prev byte (as the prev byte actually matched a1)
+            addi a0, a0, -1
+            // pop s1, from stack
+            lw   s1, sp, -4
+            ret
+__memchr_ret_null:
+            // nothing found, return nullptr
+            addi a0, zero, NULL
+            lw   s1, sp, -4
+            ret
+
+
+memset:
+// void *memset(void *str, char c, size_t n)
+__memset_loop:
+            beq  a2, zero, __memset_ret
+            sb   a1, a0, 0
+            addi a0, a0, 1
+            addi a2, a2, -1
+            j    __memset_loop
+__memset_ret:
+            ret

project.toml

@@ -1,4 +1,3 @@
 [build-system]
 requires = ["setuptools", "wheel"]
 build-backend = "setuptools.build_meta"
-

requirements-dev.txt

@@ -0,0 +1,5 @@
+black==23.3.0
+pre-commit==3.2.2
+pytest==7.3.1
+filecheck==0.0.23
+lit==16.0.2

riscemu/CPU.py

@@ -30,7 +30,9 @@ class UserModeCPU(CPU):
     It is initialized with a configuration and a list of instruction sets.
     """
 
-    def __init__(self, instruction_sets: List[Type['riscemu.InstructionSet']], conf: RunConfig):
+    def __init__(
+        self, instruction_sets: List[Type["riscemu.InstructionSet"]], conf: RunConfig
+    ):
         """
         Creates a CPU instance.
 
@@ -49,12 +51,16 @@ class UserModeCPU(CPU):
         syscall_symbols.update(self.mmu.global_symbols)
         self.mmu.global_symbols.update(syscall_symbols)
 
-    def step(self, verbose=False):
+    def step(self, verbose: bool = False):
         """
         Execute a single instruction, then return.
         """
         if self.halted:
-            print(FMT_CPU + "[CPU] Program exited with code {}".format(self.exit_code) + FMT_NONE)
+            print(
+                FMT_CPU
+                + "[CPU] Program exited with code {}".format(self.exit_code)
+                + FMT_NONE
+            )
             return
 
         launch_debugger = False
@@ -63,7 +69,9 @@ class UserModeCPU(CPU):
             self.cycle += 1
             ins = self.mmu.read_ins(self.pc)
             if verbose:
-                print(FMT_CPU + "   Running 0x{:08X}:{} {}".format(self.pc, FMT_NONE, ins))
+                print(
+                    FMT_CPU + "   Running 0x{:08X}:{} {}".format(self.pc, FMT_NONE, ins)
+                )
             self.pc += self.INS_XLEN
             self.run_instruction(ins)
         except RiscemuBaseException as ex:
@@ -72,25 +80,29 @@ class UserModeCPU(CPU):
                 if self.debugger_active:
                     raise ex
 
-                print(FMT_CPU + '[CPU] Debugger launch requested!' + FMT_NONE)
+                print(FMT_CPU + "[CPU] Debugger launch requested!" + FMT_NONE)
                 launch_debugger = True
             else:
                 print(ex.message())
                 ex.print_stacktrace()
-                print(FMT_CPU + '[CPU] Halting due to exception!' + FMT_NONE)
+                print(FMT_CPU + "[CPU] Halting due to exception!" + FMT_NONE)
                 self.halted = True
 
         if launch_debugger:
             launch_debug_session(self)
 
-    def run(self, verbose=False):
+    def run(self, verbose: bool = False):
         while not self.halted:
             self.step(verbose)
 
         if self.conf.verbosity > 0:
-            print(FMT_CPU + "[CPU] Program exited with code {}".format(self.exit_code) + FMT_NONE)
+            print(
+                FMT_CPU
+                + "[CPU] Program exited with code {}".format(self.exit_code)
+                + FMT_NONE
+            )
 
-    def setup_stack(self, stack_size=1024 * 4) -> bool:
+    def setup_stack(self, stack_size: int = 1024 * 4) -> bool:
         """
         Create program stack and populate stack pointer
         :param stack_size: the size of the required stack, defaults to 4Kib
@@ -98,22 +110,26 @@ class UserModeCPU(CPU):
         """
         stack_sec = BinaryDataMemorySection(
             bytearray(stack_size),
-            '.stack',
+            ".stack",
             None,  # FIXME: why does a binary data memory section require a context?
-            '',
-            0
+            "",
+            0,
         )
 
         if not self.mmu.load_section(stack_sec, fixed_position=False):
             print(FMT_ERROR + "[CPU] Could not insert stack section!" + FMT_NONE)
             return False
 
-        self.regs.set('sp', Int32(stack_sec.base + stack_sec.size))
+        self.regs.set("sp", Int32(stack_sec.base + stack_sec.size))
 
         if self.conf.verbosity > 1:
-            print(FMT_CPU + "[CPU] Created stack of size {} at 0x{:x}".format(
-                stack_size, stack_sec.base
-            ) + FMT_NONE)
+            print(
+                FMT_CPU
+                + "[CPU] Created stack of size {} at 0x{:x}".format(
+                    stack_size, stack_sec.base
+                )
+                + FMT_NONE
+            )
 
         return True
 

riscemu/IO/IOModule.py

@@ -1,22 +1,30 @@
-from abc import ABC, abstractmethod
+from abc import ABC
 from typing import Optional
 
 from riscemu.types import MemorySection, MemoryFlags, T_RelativeAddress
 
 
 class IOModule(MemorySection, ABC):
-    def __init__(self, name: str, flags: MemoryFlags, size: int, owner: str = 'system', base: int = 0):
+    def __init__(
+        self,
+        name: str,
+        flags: MemoryFlags,
+        size: int,
+        owner: str = "system",
+        base: int = 0,
+    ):
         super(IOModule, self).__init__(name, flags, size, base, owner, None)
 
     def contains(self, addr, size: int = 0):
-        return self.base <= addr < self.base + self.size and \
-               self.base <= addr + size <= self.base + self.size
+        return (
+            self.base <= addr < self.base + self.size
+            and self.base <= addr + size <= self.base + self.size
+        )
 
-    def dump(self, start: T_RelativeAddress, end: Optional[T_RelativeAddress] = None, fmt: str = 'hex',
-             bytes_per_row: int = 16, rows: int = 10, group: int = 4):
+    def dump(self, *args, **kwargs):
         print(self)
 
     def __repr__(self):
         return "{}[{}] at 0x{:0X} (size={}bytes, flags={})".format(
             self.__class__.__name__, self.name, self.base, self.size, self.flags
-        )
+        )

riscemu/IO/TextIO.py

@@ -8,7 +8,9 @@ class TextIO(IOModule):
         raise InstructionAccessFault(self.base + offset)
 
     def __init__(self, base: int, buflen: int = 128):
-        super(TextIO, self).__init__('TextIO', MemoryFlags(False, False), buflen + 4, base=base)
+        super(TextIO, self).__init__(
+            "TextIO", MemoryFlags(False, False), buflen + 4, base=base
+        )
         self.buff = bytearray(buflen)
         self.current_line = ""
 
@@ -23,15 +25,15 @@ class TextIO(IOModule):
                 self._print()
                 return
         buff_start = addr - 4
-        self.buff[buff_start:buff_start + size] = data[0:size]
+        self.buff[buff_start : buff_start + size] = data[0:size]
 
     def _print(self):
         buff = self.buff
         self.buff = bytearray(self.size)
-        if b'\x00' in buff:
-            buff = buff.split(b'\x00')[0]
-        text = buff.decode('ascii')
-        if '\n' in text:
+        if b"\x00" in buff:
+            buff = buff.split(b"\x00")[0]
+        text = buff.decode("ascii")
+        if "\n" in text:
             lines = text.split("\n")
             lines[0] = self.current_line + lines[0]
             for line in lines[:-1]:

riscemu/MMU.py

@@ -8,8 +8,16 @@ from typing import Dict, List, Optional, Union
 
 from .colors import *
 from .helpers import align_addr
-from .types import Instruction, MemorySection, MemoryFlags, T_AbsoluteAddress, \
-    Program, InstructionContext, Int32
+from .types import (
+    Instruction,
+    MemorySection,
+    MemoryFlags,
+    T_AbsoluteAddress,
+    Program,
+    InstructionContext,
+    Int32,
+    Float32,
+)
 from .types.exceptions import InvalidAllocationException, MemoryAccessException
 
 
@@ -65,7 +73,7 @@ class MMU:
                 return sec
         return None
 
-    def get_bin_containing(self, addr: T_AbsoluteAddress) -> Optional[Program]:
+    def get_program_at_addr(self, addr: T_AbsoluteAddress) -> Optional[Program]:
         for program in self.programs:
             if program.base <= addr < program.base + program.size:
                 return program
@@ -80,8 +88,14 @@ class MMU:
         """
         sec = self.get_sec_containing(addr)
         if sec is None:
-            print(FMT_MEM + "[MMU] Trying to read instruction form invalid region! (read at {}) ".format(addr)
-                  + "Have you forgotten an exit syscall or ret statement?" + FMT_NONE)
+            print(
+                FMT_MEM
+                + "[MMU] Trying to read instruction form invalid region! (read at {}) ".format(
+                    addr
+                )
+                + "Have you forgotten an exit syscall or ret statement?"
+                + FMT_NONE
+            )
             raise RuntimeError("No next instruction available!")
         return sec.read_ins(addr - sec.base)
 
@@ -94,12 +108,19 @@ class MMU:
         :return: The bytearray at addr
         """
         if isinstance(addr, Int32):
-            breakpoint()
             addr = addr.unsigned_value
         sec = self.get_sec_containing(addr)
         if sec is None:
-            print(FMT_MEM + "[MMU] Trying to read data form invalid region at 0x{:x}! ".format(addr) + FMT_NONE)
-            raise MemoryAccessException("region is non-initialized!", addr, size, 'read')
+            print(
+                FMT_MEM
+                + "[MMU] Trying to read data form invalid region at 0x{:x}! ".format(
+                    addr
+                )
+                + FMT_NONE
+            )
+            raise MemoryAccessException(
+                "region is non-initialized!", addr, size, "read"
+            )
         return sec.read(addr - sec.base, size)
 
     def write(self, addr: int, size: int, data: bytearray):
@@ -112,8 +133,16 @@ class MMU:
         """
         sec = self.get_sec_containing(addr)
         if sec is None:
-            print(FMT_MEM + '[MMU] Invalid write into non-initialized region at 0x{:08X}'.format(addr) + FMT_NONE)
-            raise MemoryAccessException("region is non-initialized!", addr, size, 'write')
+            print(
+                FMT_MEM
+                + "[MMU] Invalid write into non-initialized region at 0x{:08X}".format(
+                    addr
+                )
+                + FMT_NONE
+            )
+            raise MemoryAccessException(
+                "region is non-initialized!", addr, size, "write"
+            )
 
         return sec.write(addr - sec.base, size, data)
 
@@ -127,7 +156,11 @@ class MMU:
         """
         sec = self.get_sec_containing(addr)
         if sec is None:
-            print(FMT_MEM + "[MMU] No section containing addr 0x{:08X}".format(addr) + FMT_NONE)
+            print(
+                FMT_MEM
+                + "[MMU] No section containing addr 0x{:08X}".format(addr)
+                + FMT_NONE
+            )
             return
         sec.dump(addr - sec.base, *args, **kwargs)
 
@@ -139,33 +172,54 @@ class MMU:
         """
         print(FMT_MEM + "[MMU] Lookup for symbol {}:".format(symb) + FMT_NONE)
         if symb in self.global_symbols:
-            print("   Found global symbol {}: 0x{:X}".format(symb, self.global_symbols[symb]))
+            print(
+                "   Found global symbol {}: 0x{:X}".format(
+                    symb, self.global_symbols[symb]
+                )
+            )
         for bin in self.programs:
             if symb in bin.context.labels:
-                print("   Found local labels {}: 0x{:X} in {}".format(symb, bin.context.labels[symb], bin.name))
+                print(
+                    "   Found local labels {}: 0x{:X} in {}".format(
+                        symb, bin.context.labels[symb], bin.name
+                    )
+                )
 
     def read_int(self, addr: int) -> Int32:
         return Int32(self.read(addr, 4))
 
+    def read_float(self, addr: int) -> Float32:
+        return Float32(self.read(addr, 4))
+
     def translate_address(self, address: T_AbsoluteAddress) -> str:
         sec = self.get_sec_containing(address)
         if not sec:
             return "unknown at 0x{:0x}".format(address)
 
-        bin = self.get_bin_containing(address)
+        bin = self.get_program_at_addr(address)
         secs = set(sec.name for sec in bin.sections) if bin else []
         elf_markers = {
-            '__global_pointer$', '_fdata', '_etext', '_gp',
-            '_bss_start', '_bss_end', '_ftext', '_edata', '_end', '_fbss'
+            "__global_pointer$",
+            "_fdata",
+            "_etext",
+            "_gp",
+            "_bss_start",
+            "_bss_end",
+            "_ftext",
+            "_edata",
+            "_end",
+            "_fbss",
         }
 
         def key(x):
             name, val = x
             return address - val
 
-        best_fit = sorted(filter(lambda x: x[1] <= address, sec.context.labels.items()), key=key)
+        best_fit = sorted(
+            filter(lambda x: x[1] <= address, sec.context.labels.items()), key=key
+        )
 
-        best = ('', float('inf'))
+        best = ("", float("inf"))
         for name, val in best_fit:
             if address - val < best[1]:
                 best = (name, val)
@@ -179,13 +233,14 @@ class MMU:
 
         if not name:
             return "{}:{} + 0x{:x} (0x{:x})".format(
-                sec.owner, sec.name,
-                address - sec.base, address
+                sec.owner, sec.name, address - sec.base, address
             )
 
-        return str('{}:{} at {} (0x{:0x}) + 0x{:0x}'.format(
-            sec.owner, sec.name, name, val, address - val
-        ))
+        return str(
+            "{}:{} at {} (0x{:0x}) + 0x{:0x}".format(
+                sec.owner, sec.name, name, val, address - val
+            )
+        )
 
     def has_continous_free_region(self, start: int, end: int) -> bool:
         # if we have no sections we are all good
@@ -211,13 +266,24 @@ class MMU:
 
     def load_program(self, program: Program, align_to: int = 4):
         if program.base is not None:
-            if not self.has_continous_free_region(program.base, program.base + program.size):
-                print(FMT_MEM + "Cannot load program {} into desired space (0x{:0x}-0x{:0x}), area occupied.".format(
-                    program.name, program.base, program.base + program.size
-                ) + FMT_NONE)
-                raise InvalidAllocationException("Area occupied".format(
-                    program.name, program.base, program.base + program.size
-                ), program.name, program.size, MemoryFlags(False, True))
+            if not self.has_continous_free_region(
+                program.base, program.base + program.size
+            ):
+                print(
+                    FMT_MEM
+                    + "Cannot load program {} into desired space (0x{:0x}-0x{:0x}), area occupied.".format(
+                        program.name, program.base, program.base + program.size
+                    )
+                    + FMT_NONE
+                )
+                raise InvalidAllocationException(
+                    "Area occupied".format(
+                        program.name, program.base, program.base + program.size
+                    ),
+                    program.name,
+                    program.size,
+                    MemoryFlags(False, True),
+                )
 
             at_addr = program.base
         else:
@@ -245,14 +311,19 @@ class MMU:
                 self.sections.append(sec)
                 self._update_state()
             else:
-                print(FMT_MEM + '[MMU] Cannot place section {} at {}, space is occupied!'.format(sec, sec.base))
+                print(
+                    FMT_MEM
+                    + "[MMU] Cannot place section {} at {}, space is occupied!".format(
+                        sec, sec.base
+                    )
+                )
                 return False
         else:
             at_addr = align_addr(self.get_guaranteed_free_address(), 8)
             sec.base = at_addr
             self.sections.append(sec)
             self._update_state()
-            return True
+        return True
 
     def _update_state(self):
         """
@@ -270,8 +341,7 @@ class MMU:
 
     def __repr__(self):
         return "{}(\n\t{}\n)".format(
-            self.__class__.__name__,
-            "\n\t".join(repr(x) for x in self.programs)
+            self.__class__.__name__, "\n\t".join(repr(x) for x in self.programs)
         )
 
     def context_for(self, addr: T_AbsoluteAddress) -> InstructionContext:
@@ -282,13 +352,12 @@ class MMU:
 
         return InstructionContext()
 
-    def report_addr(self, addr: T_AbsoluteAddress):
-        sec = self.get_sec_containing(addr)
-        if not sec:
-            print("addr is in no section!")
-            return
-        owner = [b for b in self.programs if b.name == sec.owner]
-        if owner:
-            print("owned by: {}".format(owner[0]))
-
-        print("{}: 0x{:0x} + 0x{:0x}".format(name, val, addr - val))
+    def find_entrypoint(self) -> Optional[int]:
+        # try to find the global entrypoint
+        if "_start" in self.global_symbols:
+            return self.global_symbols["_start"]
+        # otherwise find a main (that's not necessarily global)
+        for p in self.programs:
+            if "main" in p.context.labels:
+                return p.context.resolve_label("main")
+        return None

riscemu/__init__.py

@@ -8,8 +8,17 @@ This package aims at providing an all-round usable RISC-V emulator and debugger
 It contains everything needed to run assembly files, so you don't need any custom compilers or toolchains
 """
 
-from .types.exceptions import RiscemuBaseException, LaunchDebuggerException, InvalidSyscallException, LinkerException, \
-    ParseException, NumberFormatException, InvalidRegisterException, MemoryAccessException, OutOfMemoryException
+from .types.exceptions import (
+    RiscemuBaseException,
+    LaunchDebuggerException,
+    InvalidSyscallException,
+    LinkerException,
+    ParseException,
+    NumberFormatException,
+    InvalidRegisterException,
+    MemoryAccessException,
+    OutOfMemoryException,
+)
 
 from .instructions import *
 
@@ -24,5 +33,5 @@ from .config import RunConfig
 from .parser import tokenize, parse_tokens, AssemblyFileLoader
 
 __author__ = "Anton Lydike <Anton@Lydike.com>"
-__copyright__ = "Copyright 2022 Anton Lydike"
-__version__ = '2.0.3'
+__copyright__ = "Copyright 2023 Anton Lydike"
+__version__ = "2.1.1"

riscemu/__main__.py

@@ -5,126 +5,18 @@ SPDX-License-Identifier: MIT
 
 This file holds the logic for starting the emulator from the CLI
 """
-from riscemu import RiscemuBaseException, __copyright__, __version__
-from riscemu.CPU import UserModeCPU
+import sys
 
-if __name__ == '__main__':
-    from .config import RunConfig
-    from .instructions import InstructionSetDict
-    from .colors import FMT_BOLD, FMT_MAGENTA
-    from .parser import AssemblyFileLoader
-    import argparse
-    import sys
+from riscemu import RiscemuBaseException
+from riscemu.riscemu_main import RiscemuMain
 
-    all_ins_names = list(InstructionSetDict.keys())
+try:
+    main = RiscemuMain()
+    main.run_from_cli(sys.argv[1:])
+    sys.exit(main.cpu.exit_code if not main.cfg.ignore_exit_code else 0)
 
-    if '--version' in sys.argv:
-        print("riscemu version {}\n{}\n\nAvailable ISA: {}".format(
-            __version__, __copyright__,
-            ", ".join(InstructionSetDict.keys())
-        ))
-        sys.exit()
+except RiscemuBaseException as e:
+    print("Error: {}".format(e.message()))
+    e.print_stacktrace()
 
-    class OptionStringAction(argparse.Action):
-        def __init__(self, option_strings, dest, keys=None, omit_empty=False, **kwargs):
-            if keys is None:
-                raise ValueError('must define "keys" argument')
-            if isinstance(keys, dict):
-                keys_d = keys
-            elif isinstance(keys, (list, tuple)):
-                keys_d = {}
-                for k in keys:
-                    if isinstance(k, tuple):
-                        k, v = k
-                    else:
-                        v = False
-                    keys_d[k] = v
-            else:
-                keys_d = dict()
-            super().__init__(option_strings, dest, default=keys_d, **kwargs)
-            self.keys = keys_d
-            self.omit_empty = omit_empty
-
-        def __call__(self, parser, namespace, values, option_string=None):
-            d = {}
-            if not self.omit_empty:
-                d.update(self.keys)
-            for x in values.split(','):
-                if x in self.keys:
-                    d[x] = True
-                else:
-                    raise ValueError('Invalid parameter supplied: ' + x)
-            setattr(namespace, self.dest, d)
-
-
-    parser = argparse.ArgumentParser(description='RISC-V Userspace parser and emulator', prog='riscemu')
-    parser.add_argument('files', metavar='file.asm', type=str, nargs='+',
-                        help='The assembly files to load, the last one will be run')
-
-    parser.add_argument('--options', '-o', action=OptionStringAction,
-                        keys=('disable_debug', 'no_syscall_symbols', 'fail_on_ex', 'add_accept_imm'))
-
-    parser.add_argument('--syscall-opts', '-so', action=OptionStringAction,
-                        keys=('fs_access', 'disable_input'))
-
-    parser.add_argument('--instruction-sets', '-is', action=OptionStringAction,
-                        help="Instruction sets to load, available are: {}. All are enabled by default"
-                        .format(", ".join(all_ins_names)), keys={k: True for k in all_ins_names}, omit_empty=True)
-
-    parser.add_argument('--stack_size', type=int, help='Stack size of loaded programs, defaults to 8MB', nargs='?')
-
-    parser.add_argument('-v', '--verbose', help="Verbosity level (can be used multiple times)", action='count',
-                        default=0)
-
-    parser.add_argument('--interactive', help="Launch the interactive debugger instantly instead of loading any "
-                                              "programs", action='store_true')
-
-    args = parser.parse_args()
-
-    # create a RunConfig from the cli args
-    cfg_dict = dict(
-        stack_size=args.stack_size,
-        debug_instruction=not args.options['disable_debug'],
-        include_scall_symbols=not args.options['no_syscall_symbols'],
-        debug_on_exception=not args.options['fail_on_ex'],
-        add_accept_imm=args.options['add_accept_imm'],
-        scall_fs=args.syscall_opts['fs_access'],
-        scall_input=not args.syscall_opts['disable_input'],
-        verbosity=args.verbose
-    )
-    for k, v in dict(cfg_dict).items():
-        if v is None:
-            del cfg_dict[k]
-
-    cfg = RunConfig(**cfg_dict)
-
-    if not hasattr(args, 'ins'):
-        setattr(args, 'ins', {k: True for k in all_ins_names})
-
-    FMT_PRINT = FMT_BOLD + FMT_MAGENTA
-
-    # parse required instruction sets
-    ins_to_load = [
-        InstructionSetDict[name] for name, b in args.ins.items() if b
-    ]
-
-    try:
-        cpu = UserModeCPU(ins_to_load, cfg)
-
-        opts = AssemblyFileLoader.get_options(sys.argv)
-        for file in args.files:
-            loader = AssemblyFileLoader.instantiate(file, opts)
-            cpu.load_program(loader.parse())
-
-        # set up a stack
-        cpu.setup_stack(cfg.stack_size)
-
-        # launch the last loaded program
-        cpu.launch(cpu.mmu.programs[-1], verbose=cfg.verbosity > 1)
-        sys.exit(cpu.exit_code)
-
-    except RiscemuBaseException as e:
-        print("Error: {}".format(e.message()))
-        e.print_stacktrace()
-
-        sys.exit(-1)
+    sys.exit(-1)

riscemu/assembler.py

@@ -6,10 +6,17 @@ from .colors import FMT_PARSE, FMT_NONE
 from riscemu.types.exceptions import ParseException, ASSERT_LEN
 from .helpers import parse_numeric_argument, align_addr, get_section_base_name
 from .tokenizer import Token
-from .types import Program, T_RelativeAddress, InstructionContext, Instruction, BinaryDataMemorySection, \
-    InstructionMemorySection, Int32
-
-INSTRUCTION_SECTION_NAMES = ('.text', '.init', '.fini')
+from .types import (
+    Program,
+    T_RelativeAddress,
+    InstructionContext,
+    Instruction,
+    BinaryDataMemorySection,
+    InstructionMemorySection,
+    Int32,
+)
+
+INSTRUCTION_SECTION_NAMES = (".text", ".init", ".fini")
 """
 A tuple containing all section names which contain executable code (instead of data)
 
@@ -47,8 +54,7 @@ class CurrentSection:
 
     def __repr__(self):
         return "{}(name={},data={},type={})".format(
-            self.__class__.__name__, self.name,
-            self.data, self.type.name
+            self.__class__.__name__, self.name, self.data, self.type.name
         )
 
 
@@ -72,12 +78,20 @@ class ParseContext:
 
         if self.section.type == MemorySectionType.Data:
             section = BinaryDataMemorySection(
-                self.section.data, self.section.name, self.context, self.program.name, self.section.base
+                self.section.data,
+                self.section.name,
+                self.context,
+                self.program.name,
+                self.section.base,
             )
             self.program.add_section(section)
         elif self.section.type == MemorySectionType.Instructions:
             section = InstructionMemorySection(
-                self.section.data, self.section.name, self.context, self.program.name, self.section.base
+                self.section.data,
+                self.section.name,
+                self.context,
+                self.program.name,
+                self.section.base,
             )
             self.program.add_section(section)
 
@@ -89,7 +103,9 @@ class ParseContext:
         self._finalize_section()
         self.section = CurrentSection(name, type, base)
 
-    def add_label(self, name: str, value: int, is_global: bool = False, is_relative: bool = False):
+    def add_label(
+        self, name: str, value: int, is_global: bool = False, is_relative: bool = False
+    ):
         self.context.labels[name] = value
         if is_global:
             self.program.global_labels.add(name)
@@ -109,10 +125,16 @@ class ParseContext:
 
 def ASSERT_IN_SECTION_TYPE(context: ParseContext, type: MemorySectionType):
     if context.section is None:
-        raise ParseException('Error, expected to be in {} section, but no section is present...'.format(type.name))
+        raise ParseException(
+            "Error, expected to be in {} section, but no section is present...".format(
+                type.name
+            )
+        )
     if context.section.type != type:
         raise ParseException(
-            'Error, expected to be in {} section, but currently in {}...'.format(type.name, context.section)
+            "Error, expected to be in {} section, but currently in {}...".format(
+                type.name, context.section
+            )
         )
 
 
@@ -174,7 +196,9 @@ class AssemblerDirectives:
         cls.add_bytes(size, bytearray(size), context)
 
     @classmethod
-    def add_bytes(cls, size: int, content: Union[None, int, bytearray], context: ParseContext):
+    def add_bytes(
+        cls, size: int, content: Union[None, int, bytearray], context: ParseContext
+    ):
         ASSERT_IN_SECTION_TYPE(context, MemorySectionType.Data)
 
         if content is None:
@@ -187,9 +211,9 @@ class AssemblerDirectives:
     @classmethod
     def add_text(cls, text: str, context: ParseContext, zero_terminate: bool = True):
         # replace '\t' and '\n' escape sequences
-        text = text.replace('\\n', '\n').replace('\\t', '\t')
+        text = text.replace("\\n", "\n").replace("\\t", "\t")
 
-        encoded_bytes = bytearray(text.encode('ascii'))
+        encoded_bytes = bytearray(text.encode("ascii"))
         if zero_terminate:
             encoded_bytes += bytearray(1)
         cls.add_bytes(len(encoded_bytes), encoded_bytes, context)
@@ -197,24 +221,36 @@ class AssemblerDirectives:
     @classmethod
     def handle_instruction(cls, token: Token, args: Tuple[str], context: ParseContext):
         op = token.value[1:]
-        if hasattr(cls, 'op_' + op):
-            getattr(cls, 'op_' + op)(token, args, context)
-        elif op in ('text', 'data', 'rodata', 'bss', 'sbss'):
+        if hasattr(cls, "op_" + op):
+            getattr(cls, "op_" + op)(token, args, context)
+        elif op in ("text", "data", "rodata", "bss", "sbss"):
             cls.op_section(token, (token.value,), context)
-        elif op in ('string', 'asciiz', 'asciz', 'ascii'):
+        elif op in ("string", "asciiz", "asciz", "ascii"):
             ASSERT_LEN(args, 1)
-            cls.add_text(args[0], context, op == 'ascii')
+            cls.add_text(args[0], context, zero_terminate=(op != "ascii"))
         elif op in DATA_OP_SIZES:
             size = DATA_OP_SIZES[op]
             for arg in args:
                 cls.add_bytes(size, parse_numeric_argument(arg), context)
         else:
-            print(FMT_PARSE + "Unknown assembler directive: {} {} in {}".format(token, args, context) + FMT_NONE)
+            print(
+                FMT_PARSE
+                + "Unknown assembler directive: {} {} in {}".format(
+                    token, args, context
+                )
+                + FMT_NONE
+            )
 
 
 DATA_OP_SIZES = {
-    'byte': 1,
-    '2byte': 2, 'half': 2, 'short': 2,
-    '4byte': 4, 'word': 4, 'long': 4,
-    '8byte': 8, 'dword': 8, 'quad': 8,
+    "byte": 1,
+    "2byte": 2,
+    "half": 2,
+    "short": 2,
+    "4byte": 4,
+    "word": 4,
+    "long": 4,
+    "8byte": 8,
+    "dword": 8,
+    "quad": 8,
 }

riscemu/colors.py

@@ -6,18 +6,18 @@ SPDX-License-Identifier: MIT
 
 # Colors
 
-FMT_RED = '\033[31m'
-FMT_ORANGE = '\033[33m'
-FMT_GRAY = '\033[37m'
-FMT_CYAN = '\033[36m'
-FMT_GREEN = '\033[32m'
-FMT_MAGENTA = '\033[35m'
-FMT_BLUE = '\033[34m'
-FMT_YELLOW = '\033[93m'
+FMT_RED = "\033[31m"
+FMT_ORANGE = "\033[33m"
+FMT_GRAY = "\033[37m"
+FMT_CYAN = "\033[36m"
+FMT_GREEN = "\033[32m"
+FMT_MAGENTA = "\033[35m"
+FMT_BLUE = "\033[34m"
+FMT_YELLOW = "\033[93m"
 
-FMT_BOLD = '\033[1m'
-FMT_NONE = '\033[0m'
-FMT_UNDERLINE = '\033[4m'
+FMT_BOLD = "\033[1m"
+FMT_NONE = "\033[0m"
+FMT_UNDERLINE = "\033[4m"
 
 FMT_ERROR = FMT_RED + FMT_BOLD
 
@@ -26,4 +26,4 @@ FMT_PARSE = FMT_CYAN + FMT_BOLD
 FMT_CPU = FMT_BLUE + FMT_BOLD
 FMT_SYSCALL = FMT_YELLOW + FMT_BOLD
 FMT_DEBUG = FMT_MAGENTA + FMT_BOLD
-FMT_CSR = FMT_ORANGE + FMT_BOLD
+FMT_CSR = FMT_ORANGE + FMT_BOLD

riscemu/config.py

@@ -20,6 +20,7 @@ class RunConfig:
     scall_fs: bool = False
     verbosity: int = 0
     slowdown: float = 1
-
-
-CONFIG = RunConfig()
+    unlimited_registers: bool = False
+    # runtime config
+    use_libc: bool = False
+    ignore_exit_code: bool = False

riscemu/debug.py

@@ -11,10 +11,10 @@ from .helpers import *
 if typing.TYPE_CHECKING:
     from riscemu import CPU, Registers
 
-HIST_FILE = os.path.join(os.path.expanduser('~'), '.riscemu_history')
+HIST_FILE = os.path.join(os.path.expanduser("~"), ".riscemu_history")
 
 
-def launch_debug_session(cpu: 'CPU', prompt=""):
+def launch_debug_session(cpu: "CPU", prompt=""):
     if cpu.debugger_active:
         return
     import code
@@ -39,7 +39,7 @@ def launch_debug_session(cpu: 'CPU', prompt=""):
             mmu.dump(what, *args, **kwargs)
 
     def dump_stack(*args, **kwargs):
-        mmu.dump(regs.get('sp'), *args, **kwargs)
+        mmu.dump(regs.get("sp"), *args, **kwargs)
 
     def ins():
         print("Current instruction at 0x{:08X}:".format(cpu.pc))
@@ -51,11 +51,7 @@ def launch_debug_session(cpu: 'CPU', prompt=""):
             return
         context = mmu.context_for(cpu.pc)
 
-        ins = SimpleInstruction(
-            name,
-            tuple(args),
-            context,
-            cpu.pc)
+        ins = SimpleInstruction(name, tuple(args), context, cpu.pc)
         print(FMT_DEBUG + "Running instruction {}".format(ins) + FMT_NONE)
         cpu.run_instruction(ins)
 
@@ -63,7 +59,7 @@ def launch_debug_session(cpu: 'CPU', prompt=""):
         try:
             cpu.run(verbose)
         except LaunchDebuggerException:
-            print(FMT_DEBUG + 'Returning to debugger...')
+            print(FMT_DEBUG + "Returning to debugger...")
             return
 
     def step():
@@ -92,4 +88,3 @@ def launch_debug_session(cpu: 'CPU', prompt=""):
     finally:
         cpu.debugger_active = False
         readline.write_history_file(HIST_FILE)
-

riscemu/decoder/__init__.py

@@ -1,2 +1,2 @@
 from .decoder import decode, RISCV_REGS
-from .formatter import format_ins
+from .formatter import format_ins

riscemu/decoder/__main__.py

@@ -1,4 +1,4 @@
-if __name__ == '__main__':
+if __name__ == "__main__":
     import code
     import readline
     import rlcompleter
@@ -14,4 +14,6 @@ if __name__ == '__main__':
     readline.set_completer(rlcompleter.Completer(sess_vars).complete)
     readline.set_completer(rlcompleter.Completer(sess_vars).complete)
     readline.parse_and_bind("tab: complete")
-    code.InteractiveConsole(sess_vars).interact(banner="Interaktive decoding session started...", exitmsg="Closing...")
+    code.InteractiveConsole(sess_vars).interact(
+        banner="Interaktive decoding session started...", exitmsg="Closing..."
+    )

riscemu/decoder/decoder.py

@@ -5,13 +5,14 @@ from typing import Tuple, List
 def print_ins(ins: int):
     print("  f7      rs2   rs1   f3  rd    op")
     print(
-        f"0b{ins >> 25 :07b}_{(ins >> 20) & 0b11111:05b}_{(ins >> 15) & 0b11111:05b}_{(ins >> 12) & 0b111:03b}_{(ins >> 7) & 0b11111:05b}_{ins & 0b1111111:07b}");
+        f"0b{ins >> 25 :07b}_{(ins >> 20) & 0b11111:05b}_{(ins >> 15) & 0b11111:05b}_{(ins >> 12) & 0b111:03b}_{(ins >> 7) & 0b11111:05b}_{ins & 0b1111111:07b}"
+    )
 
 
 STATIC_INSN: Dict[int, Tuple[str, List[int], int]] = {
     0x00000013: ("nop", [], 0x00000013),
     0x00008067: ("ret", [], 0x00008067),
-    0xfe010113: ("addi", [2, 2, -32], 0xfe010113),
+    0xFE010113: ("addi", [2, 2, -32], 0xFE010113),
     0x02010113: ("addi", [2, 2, 32], 0x02010113),
     0x00100073: ("ebreak", [], 0x00100073),
     0x00000073: ("ecall", [], 0x00000073),
@@ -23,7 +24,7 @@ STATIC_INSN: Dict[int, Tuple[str, List[int], int]] = {
 
 
 def int_from_ins(insn: bytearray):
-    return int.from_bytes(insn, 'little')
+    return int.from_bytes(insn, "little")
 
 
 def name_from_insn(ins: int):
@@ -45,7 +46,7 @@ def name_from_insn(ins: int):
     if isinstance(dec, str):
         return dec
 
-    if opcode == 0x1c and fun3 == 0:
+    if opcode == 0x1C and fun3 == 0:
         # we have ecall/ebreak
         token = imm110(ins)
         if token in dec:

riscemu/decoder/formats.py

@@ -1,6 +1,7 @@
 from typing import Dict, Callable, List, Union
 from .regs import RISCV_REGS
 
+
 def op(ins: int):
     return (ins >> 2) & 31
 
@@ -46,7 +47,12 @@ def imm_b(ins: int):
     lower = rd(ins)
     higher = funct7(ins)
 
-    num = (lower & 0b11110) + ((higher & 0b0111111) << 5) + ((lower & 1) << 11) + ((higher >> 6) << 12)
+    num = (
+        (lower & 0b11110)
+        + ((higher & 0b0111111) << 5)
+        + ((lower & 1) << 11)
+        + ((higher >> 6) << 12)
+    )
     return sign_extend(num, 13)
 
 
@@ -56,10 +62,11 @@ def imm_u(ins: int):
 
 def imm_j(ins: int):
     return sign_extend(
-        (((ins >> 21) & 0b1111111111) << 1) +
-        (((ins >> 20) & 1) << 11) +
-        (((ins >> 12) & 0b11111111) << 12) +
-        (((ins >> 31) & 1) << 20), 21
+        (((ins >> 21) & 0b1111111111) << 1)
+        + (((ins >> 20) & 1) << 11)
+        + (((ins >> 12) & 0b11111111) << 12)
+        + (((ins >> 31) & 1) << 20),
+        21,
     )
 
 
@@ -111,7 +118,7 @@ INSTRUCTION_ARGS_DECODER: Dict[int, Callable[[int], List[int]]] = {
     0x0D: decode_u,
     0x18: decode_b,
     0x19: decode_i,
-    0x1b: decode_j,
-    0x1c: decode_i_unsigned,
-    0b1011: decode_r
+    0x1B: decode_j,
+    0x1C: decode_i_unsigned,
+    0b1011: decode_r,
 }

riscemu/decoder/formatter.py

@@ -1,5 +1,15 @@
-from .formats import INSTRUCTION_ARGS_DECODER, op, decode_i, decode_r, decode_u, decode_b, decode_j, decode_s, \
-    decode_i_shamt, decode_i_unsigned
+from .formats import (
+    INSTRUCTION_ARGS_DECODER,
+    op,
+    decode_i,
+    decode_r,
+    decode_u,
+    decode_b,
+    decode_j,
+    decode_s,
+    decode_i_shamt,
+    decode_i_unsigned,
+)
 from .regs import RISCV_REGS
 
 
@@ -9,9 +19,9 @@ def int_to_hex(num: int):
     return f"0x{num:x}"
 
 
-def format_ins(ins: int, name: str, fmt: str = 'int'):
+def format_ins(ins: int, name: str, fmt: str = "int"):
     opcode = op(ins)
-    if fmt == 'hex':
+    if fmt == "hex":
         fmt = int_to_hex
     else:
         fmt = str
@@ -20,7 +30,7 @@ def format_ins(ins: int, name: str, fmt: str = 'int'):
         return f"{name} <unknown op>"
 
     decoder = INSTRUCTION_ARGS_DECODER[opcode]
-    if name in ('ecall', 'ebreak', 'mret', 'sret', 'uret'):
+    if name in ("ecall", "ebreak", "mret", "sret", "uret"):
         return name
     if opcode in (0x8, 0x0):
         r1, r2, imm = decoder(ins)

riscemu/decoder/instruction_table.py

@@ -4,7 +4,7 @@ from .formats import *
 tbl = lambda: defaultdict(tbl)
 
 RV32 = tbl()
-RV32[0x1b] = "jal"
+RV32[0x1B] = "jal"
 RV32[0x0D] = "lui"
 RV32[0x05] = "auipc"
 RV32[0x19][0] = "jalr"
@@ -36,26 +36,26 @@ RV32[0x08][0] = "sb"
 RV32[0x08][1] = "sh"
 RV32[0x08][2] = "sw"
 
-RV32[0x1c][1] = "csrrw"
-RV32[0x1c][2] = "csrrs"
-RV32[0x1c][3] = "csrrc"
-RV32[0x1c][5] = "csrrwi"
-RV32[0x1c][6] = "csrrsi"
-RV32[0x1c][7] = "csrrci"
+RV32[0x1C][1] = "csrrw"
+RV32[0x1C][2] = "csrrs"
+RV32[0x1C][3] = "csrrc"
+RV32[0x1C][5] = "csrrwi"
+RV32[0x1C][6] = "csrrsi"
+RV32[0x1C][7] = "csrrci"
 
-RV32[0x1c][0][0] = "ecall"
-RV32[0x1c][0][1] = "ebreak"
+RV32[0x1C][0][0] = "ecall"
+RV32[0x1C][0][1] = "ebreak"
 
-RV32[0x0C][0][0]  = "add"
+RV32[0x0C][0][0] = "add"
 RV32[0x0C][0][32] = "sub"
-RV32[0x0C][1][0]  = "sll"
-RV32[0x0C][2][0]  = "slt"
-RV32[0x0C][3][0]  = "sltu"
-RV32[0x0C][4][0]  = "xor"
-RV32[0x0C][5][0]  = "srl"
+RV32[0x0C][1][0] = "sll"
+RV32[0x0C][2][0] = "slt"
+RV32[0x0C][3][0] = "sltu"
+RV32[0x0C][4][0] = "xor"
+RV32[0x0C][5][0] = "srl"
 RV32[0x0C][5][32] = "sra"
-RV32[0x0C][6][0]  = "or"
-RV32[0x0C][7][0]  = "and"
+RV32[0x0C][6][0] = "or"
+RV32[0x0C][7][0] = "and"
 
 # rv32m
 RV32[0x0C][0][1] = "mul"

riscemu/decoder/regs.py

@@ -1,6 +1,34 @@
 RISCV_REGS = [
-    'zero', 'ra', 'sp', 'gp', 'tp', 't0', 't1', 't2',
-    's0', 's1', 'a0', 'a1', 'a2', 'a3', 'a4', 'a5', 'a6', 'a7',
-    's2', 's3', 's4', 's5', 's6', 's7', 's8', 's9', 's10', 's11',
-    't3', 't4', 't5', 't6'
+    "zero",
+    "ra",
+    "sp",
+    "gp",
+    "tp",
+    "t0",
+    "t1",
+    "t2",
+    "s0",
+    "s1",
+    "a0",
+    "a1",
+    "a2",
+    "a3",
+    "a4",
+    "a5",
+    "a6",
+    "a7",
+    "s2",
+    "s3",
+    "s4",
+    "s5",
+    "s6",
+    "s7",
+    "s8",
+    "s9",
+    "s10",
+    "s11",
+    "t3",
+    "t4",
+    "t5",
+    "t6",
 ]

riscemu/helpers.py

@@ -14,6 +14,8 @@ from .types.exceptions import *
 def align_addr(addr: int, to_bytes: int = 8) -> int:
     """
     align an address to `to_bytes` (meaning addr & to_bytes = 0)
+
+    This will increase the address
     """
     return addr + (-addr % to_bytes)
 
@@ -23,16 +25,19 @@ def parse_numeric_argument(arg: str) -> int:
     parse hex or int strings
     """
     try:
-        if arg.lower().startswith('0x'):
+        if arg.lower().startswith("0x"):
             return int(arg, 16)
         return int(arg)
     except ValueError as ex:
-        raise ParseException('Invalid immediate argument \"{}\", maybe missing symbol?'.format(arg), (arg, ex))
+        raise ParseException(
+            'Invalid immediate argument "{}", maybe missing symbol?'.format(arg),
+            (arg, ex),
+        )
 
 
 def create_chunks(my_list, chunk_size):
     """Split a list like [a,b,c,d,e,f,g,h,i,j,k,l,m] into e.g. [[a,b,c,d],[e,f,g,h],[i,j,k,l],[m]]"""
-    return [my_list[i:i + chunk_size] for i in range(0, len(my_list), chunk_size)]
+    return [my_list[i : i + chunk_size] for i in range(0, len(my_list), chunk_size)]
 
 
 def apply_highlight(item, ind, hi_ind):
@@ -45,26 +50,31 @@ def apply_highlight(item, ind, hi_ind):
 
 
 def highlight_in_list(items, hi_ind, joiner=" "):
-    return joiner.join([apply_highlight(item, i, hi_ind) for i, item in enumerate(items)])
+    return joiner.join(
+        [apply_highlight(item, i, hi_ind) for i, item in enumerate(items)]
+    )
 
 
 def format_bytes(byte_arr: bytearray, fmt: str, group: int = 1, highlight: int = -1):
     """Format byte array as per fmt. Group into groups of size `group`, and highlight index `highlight`."""
     chunks = create_chunks(byte_arr, group)
-    if fmt == 'hex':
-        return highlight_in_list(['0x{}'.format(ch.hex()) for ch in chunks], highlight)
-    if fmt == 'int':
+    if fmt == "hex":
+        return highlight_in_list(["0x{}".format(ch.hex()) for ch in chunks], highlight)
+    if fmt == "int":
         spc = str(ceil(log10(2 ** (group * 8 - 1))) + 1)
-        return highlight_in_list([('{:0' + spc + 'd}').format(Int32(ch)) for ch in chunks], highlight)
-    if fmt == 'uint':
+        return highlight_in_list(
+            [("{:0" + spc + "d}").format(Int32(ch)) for ch in chunks], highlight
+        )
+    if fmt == "uint":
         spc = str(ceil(log10(2 ** (group * 8))))
-        return highlight_in_list([('{:0' + spc + 'd}').format(UInt32(ch)) for ch in chunks],
-                                 highlight)
-    if fmt == 'char':
+        return highlight_in_list(
+            [("{:0" + spc + "d}").format(UInt32(ch)) for ch in chunks], highlight
+        )
+    if fmt == "char":
         return highlight_in_list((repr(chr(b))[1:-1] for b in byte_arr), highlight, "")
 
 
-T = TypeVar('T')
+T = TypeVar("T")
 
 
 class Peekable(Generic[T], Iterator[T]):
@@ -98,6 +108,10 @@ class Peekable(Generic[T], Iterator[T]):
 
 
 def get_section_base_name(section_name: str) -> str:
-    if '.' not in section_name:
-        print(FMT_PARSE + f"Invalid section {section_name}, not starting with a dot!" + FMT_NONE)
-    return '.' + section_name.split('.')[1]
+    if "." not in section_name:
+        print(
+            FMT_PARSE
+            + f"Invalid section {section_name}, not starting with a dot!"
+            + FMT_NONE
+        )
+    return "." + section_name.split(".")[1]

riscemu/instructions/RV32A.py

@@ -10,52 +10,52 @@ class RV32A(InstructionSet):
     for this?
     """
 
-    def instruction_lr_w(self, ins: 'Instruction'):
+    def instruction_lr_w(self, ins: "Instruction"):
         INS_NOT_IMPLEMENTED(ins)
 
-    def instruction_sc_w(self, ins: 'Instruction'):
+    def instruction_sc_w(self, ins: "Instruction"):
         INS_NOT_IMPLEMENTED(ins)
 
-    def instruction_amoswap_w(self, ins: 'Instruction'):
+    def instruction_amoswap_w(self, ins: "Instruction"):
         dest, addr, val = self.parse_rd_rs_rs(ins)
-        if dest == 'zero':
+        if dest == "zero":
             self.mmu.write(addr, val.to_bytes())
         else:
             old = Int32(self.mmu.read(addr, 4))
             self.mmu.write(addr, val.to_bytes())
             self.regs.set(dest, old)
 
-    def instruction_amoadd_w(self, ins: 'Instruction'):
+    def instruction_amoadd_w(self, ins: "Instruction"):
         dest, addr, val = self.parse_rd_rs_rs(ins)
         old = Int32(self.mmu.read(addr, 4))
         self.mmu.write(addr, (old + val).to_bytes(4))
         self.regs.set(dest, old)
 
-    def instruction_amoand_w(self, ins: 'Instruction'):
+    def instruction_amoand_w(self, ins: "Instruction"):
         dest, addr, val = self.parse_rd_rs_rs(ins)
         old = Int32(self.mmu.read(addr, 4))
         self.mmu.write(addr, (old & val).to_bytes(4))
         self.regs.set(dest, old)
 
-    def instruction_amoor_w(self, ins: 'Instruction'):
+    def instruction_amoor_w(self, ins: "Instruction"):
         dest, addr, val = self.parse_rd_rs_rs(ins)
         old = Int32(self.mmu.read(addr, 4))
         self.mmu.write(addr, (old | val).to_bytes(4))
         self.regs.set(dest, old)
 
-    def instruction_amoxor_w(self, ins: 'Instruction'):
+    def instruction_amoxor_w(self, ins: "Instruction"):
         dest, addr, val = self.parse_rd_rs_rs(ins)
         old = Int32(self.mmu.read(addr, 4))
         self.mmu.write(addr, (old ^ val).to_bytes(4))
         self.regs.set(dest, old)
 
-    def instruction_amomax_w(self, ins: 'Instruction'):
+    def instruction_amomax_w(self, ins: "Instruction"):
         dest, addr, val = self.parse_rd_rs_rs(ins)
         old = Int32(self.mmu.read(addr, 4))
         self.mmu.write(addr, max(old, val).to_bytes(4))
         self.regs.set(dest, old)
 
-    def instruction_amomaxu_w(self, ins: 'Instruction'):
+    def instruction_amomaxu_w(self, ins: "Instruction"):
         val: UInt32
         dest, addr, val = self.parse_rd_rs_rs(ins, signed=False)
         old = UInt32(self.mmu.read(addr, 4))
@@ -63,13 +63,13 @@ class RV32A(InstructionSet):
         self.mmu.write(addr, max(old, val).to_bytes())
         self.regs.set(dest, old)
 
-    def instruction_amomin_w(self, ins: 'Instruction'):
+    def instruction_amomin_w(self, ins: "Instruction"):
         dest, addr, val = self.parse_rd_rs_rs(ins)
         old = Int32(self.mmu.read(addr, 4))
         self.mmu.write(addr, min(old, val).to_bytes(4))
         self.regs.set(dest, old)
 
-    def instruction_amominu_w(self, ins: 'Instruction'):
+    def instruction_amominu_w(self, ins: "Instruction"):
         val: UInt32
         dest, addr, val = self.parse_rd_rs_rs(ins, signed=False)
         old = UInt32(self.mmu.read(addr, 4))

riscemu/instructions/RV32F.py

@@ -0,0 +1,609 @@
+"""
+RiscEmu (c) 2023 Anton Lydike
+
+SPDX-License-Identifier: MIT
+
+This file contains copious amounts of docstrings that were all taken
+from https://msyksphinz-self.github.io/riscv-isadoc/html/rvfd.html
+(all the docstrings on the instruction methods documenting the opcodes
+and their function)
+"""
+from typing import Tuple
+
+from .instruction_set import InstructionSet, Instruction
+from riscemu.types import INS_NOT_IMPLEMENTED, Float32, Int32, UInt32
+
+
+class RV32F(InstructionSet):
+    def instruction_fmadd_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |rs3  |00   |rs2  |rs1  |rm   |rd   |10000|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fmadd.s    rd,rs1,rs2,rs3
+
+        :Description:
+        | Perform single-precision fused multiply addition.
+
+        :Implementation:
+        | f[rd] = f[rs1]×f[rs2]+f[rs3]
+
+        """
+        rd, rs1, rs2, rs3 = self.parse_rd_rs_rs_rs(ins)
+        self.regs.set_f(rd, rs1 * rs2 + rs3)
+
+    def instruction_fmsub_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |rs3  |00   |rs2  |rs1  |rm   |rd   |10001|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fmsub.s    rd,rs1,rs2,rs3
+
+        :Description:
+        | Perform single-precision fused multiply addition.
+
+        :Implementation:
+        | f[rd] = f[rs1]×f[rs2]-f[rs3]
+        """
+        rd, rs1, rs2, rs3 = self.parse_rd_rs_rs_rs(ins)
+        self.regs.set_f(rd, rs1 * rs2 - rs3)
+
+    def instruction_fnmsub_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |rs3  |00   |rs2  |rs1  |rm   |rd   |10010|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fnmsub.s   rd,rs1,rs2,rs3
+
+        :Description:
+        | Perform single-precision fused multiply addition.
+
+        :Implementation:
+        | f[rd] = -f[rs1]×f[rs2]+f[rs3]
+        """
+        rd, rs1, rs2, rs3 = self.parse_rd_rs_rs_rs(ins)
+        self.regs.set_f(rd, -rs1 * rs2 + rs3)
+
+    def instruction_fnmadd_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |rs3  |00   |rs2  |rs1  |rm   |rd   |10011|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fnmadd.s   rd,rs1,rs2,rs3
+
+        :Description:
+        | Perform single-precision fused multiply addition.
+
+        :Implementation:
+        | f[rd] = -f[rs1]×f[rs2]-f[rs3]
+        """
+        rd, rs1, rs2, rs3 = self.parse_rd_rs_rs_rs(ins)
+        self.regs.set_f(rd, -rs1 * rs2 - rs3)
+
+    def instruction_fadd_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00000|00   |rs2  |rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fadd.s     rd,rs1,rs2
+
+        :Description:
+        | Perform single-precision floating-point addition.
+
+        :Implementation:
+        | f[rd] = f[rs1] + f[rs2]
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set_f(
+            rd,
+            rs1 + rs2,
+        )
+
+    def instruction_fsub_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00001|00   |rs2  |rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fsub.s     rd,rs1,rs2
+
+        :Description:
+        | Perform single-precision floating-point substraction.
+
+        :Implementation:
+        | f[rd] = f[rs1] - f[rs2]
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set_f(
+            rd,
+            rs1 - rs2,
+        )
+
+    def instruction_fmul_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00010|00   |rs2  |rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fmul.s     rd,rs1,rs2
+
+        :Description:
+        | Perform single-precision floating-point multiplication.
+
+        :Implementation:
+        | f[rd] = f[rs1] × f[rs2]
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set_f(
+            rd,
+            rs1 * rs2,
+        )
+
+    def instruction_fdiv_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00011|00   |rs2  |rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fdiv.s     rd,rs1,rs2
+
+        :Description:
+        | Perform single-precision floating-point division.
+
+        :Implementation:
+        | f[rd] = f[rs1] / f[rs2]
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set_f(
+            rd,
+            rs1 / rs2,
+        )
+
+    def instruction_fsqrt_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |01011|00   |00000|rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fsqrt.s    rd,rs1
+
+        :Description:
+        | Perform single-precision square root.
+
+        :Implementation:
+        | f[rd] = sqrt(f[rs1])
+        """
+        rd, rs = self.parse_rd_rs(ins)
+        self.regs.set_f(rd, self.regs.get_f(rs) ** 0.5)
+
+    def instruction_fsgnj_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00100|00   |rs2  |rs1  |000  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fsgnj.s    rd,rs1,rs2
+
+        :Description:
+        | Produce a result that takes all bits except the sign bit from rs1.
+        | The result’s sign bit is rs2’s sign bit.
+
+        :Implementation:
+        | f[rd] = {f[rs2][31], f[rs1][30:0]}
+        """
+        INS_NOT_IMPLEMENTED(ins)
+
+    def instruction_fsgnjn_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00100|00   |rs2  |rs1  |001  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+        :Format:
+        | fsgnjn.s   rd,rs1,rs2
+
+        :Description:
+        | Produce a result that takes all bits except the sign bit from rs1.
+        | The result’s sign bit is opposite of rs2’s sign bit.
+
+        :Implementation:
+        | f[rd] = {~f[rs2][31], f[rs1][30:0]}
+        """
+        INS_NOT_IMPLEMENTED(ins)
+
+    def instruction_fsgnjx_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00100|00   |rs2  |rs1  |010  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fsgnjx.s   rd,rs1,rs2
+
+        :Description:
+        | Produce a result that takes all bits except the sign bit from rs1.
+        | The result’s sign bit is XOR of sign bit of rs1 and rs2.
+
+        :Implementation:
+        | f[rd] = {f[rs1][31] ^ f[rs2][31], f[rs1][30:0]}
+        """
+        INS_NOT_IMPLEMENTED(ins)
+
+    def instruction_fmin_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00101|00   |rs2  |rs1  |000  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fmin.s     rd,rs1,rs2
+
+        :Description:
+        | Write the smaller of single precision data in rs1 and rs2 to rd.
+
+        :Implementation:
+        | f[rd] = min(f[rs1], f[rs2])
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set_f(
+            rd,
+            Float32(
+                min(
+                    rs1.value,
+                    rs2.value,
+                )
+            ),
+        )
+
+    def instruction_fmax_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |00101|00   |rs2  |rs1  |001  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fmax.s     rd,rs1,rs2
+
+        :Description:
+        | Write the larger of single precision data in rs1 and rs2 to rd.
+
+        :Implementation:
+        | f[rd] = max(f[rs1], f[rs2])
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set_f(
+            rd,
+            Float32(
+                max(
+                    rs1.value,
+                    rs2.value,
+                )
+            ),
+        )
+
+    def instruction_fcvt_w_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |11000|00   |00000|rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fcvt.w.s   rd,rs1
+
+        :Description:
+        | Convert a floating-point number in floating-point register rs1 to a signed 32-bit in integer register rd.
+
+        :Implementation:
+        | x[rd] = sext(s32_{f32}(f[rs1]))
+        """
+        rd, rs = self.parse_rd_rs(ins)
+        self.regs.set(rd, Int32(self.regs.get_f(rs).bytes))
+
+    def instruction_fcvt_wu_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |11000|00   |00001|rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fcvt.wu.s  rd,rs1
+
+        :Description:
+        | Convert a floating-point number in floating-point register rs1 to a signed 32-bit in unsigned integer register rd.
+
+        :Implementation:
+        | x[rd] = sext(u32_{f32}(f[rs1]))
+        """
+        rd, rs = self.parse_rd_rs(ins)
+        self.regs.set(rd, UInt32(self.regs.get_f(rs).bytes))
+
+    def instruction_fmv_x_w(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |11100|00   |00000|rs1  |000  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fmv.x.w    rd,rs1
+
+        :Description:
+        | Move the single-precision value in floating-point register rs1 represented in IEEE 754-2008 encoding to the lower 32 bits of integer register rd.
+
+        :Implementation:
+        | x[rd] = sext(f[rs1][31:0])
+        """
+        rd, rs = self.parse_rd_rs(ins)
+        self.regs.set(rd, UInt32(self.regs.get_f(rs).bits))
+
+    def instruction_feq_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |10100|00   |rs2  |rs1  |010  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | feq.s      rd,rs1,rs2
+
+        :Description:
+        | Performs a quiet equal comparison between floating-point registers rs1 and rs2 and record the Boolean result in integer register rd.
+        | Only signaling NaN inputs cause an Invalid Operation exception.
+        | The result is 0 if either operand is NaN.
+
+        :Implementation:
+        | x[rd] = f[rs1] == f[rs2]
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set(rd, Int32(rs1 == rs2))
+
+    def instruction_flt_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |10100|00   |rs2  |rs1  |001  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | flt.s      rd,rs1,rs2
+
+        :Description:
+        | Performs a quiet less comparison between floating-point registers rs1 and rs2 and record the Boolean result in integer register rd.
+        | Only signaling NaN inputs cause an Invalid Operation exception.
+        | The result is 0 if either operand is NaN.
+
+        :Implementation:
+        | x[rd] = f[rs1] < f[rs2]
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set(rd, Int32(rs1 < rs2))
+
+    def instruction_fle_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |10100|00   |rs2  |rs1  |000  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fle.s      rd,rs1,rs2
+
+        :Description:
+        | Performs a quiet less or equal comparison between floating-point registers rs1 and rs2 and record the Boolean result in integer register rd.
+        | Only signaling NaN inputs cause an Invalid Operation exception.
+        | The result is 0 if either operand is NaN.
+
+        :Implementation:
+        | x[rd] = f[rs1] <= f[rs2]
+        """
+        rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
+        self.regs.set(rd, Int32(rs1 <= rs2))
+
+    def instruction_fclass_s(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |11100|00   |00000|rs1  |001  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fclass.s   rd,rs1
+
+        :Description:
+        | Examines the value in floating-point register rs1 and writes to integer register rd a 10-bit mask that indicates the class of the floating-point number.
+        | The format of the mask is described in [classify table]_.
+        | The corresponding bit in rd will be set if the property is true and clear otherwise.
+        | All other bits in rd are cleared. Note that exactly one bit in rd will be set.
+
+        :Implementation:
+        | x[rd] = classifys(f[rs1])
+        """
+        INS_NOT_IMPLEMENTED(ins)
+
+    def instruction_fcvt_s_w(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |11010|00   |00000|rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fcvt.s.w   rd,rs1
+
+        :Description:
+        | Converts a 32-bit signed integer, in integer register rs1 into a floating-point number in floating-point register rd.
+
+        :Implementation:
+        | f[rd] = f32_{s32}(x[rs1])
+        """
+        rd, rs = self.parse_rd_rs(ins)
+        self.regs.set_f(rd, Float32.from_bytes(self.regs.get(rs).signed().value))
+
+    def instruction_fcvt_s_wu(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |11010|00   |00001|rs1  |rm   |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+        | fcvt.s.wu  rd,rs1
+
+        :Description:
+        | Converts a 32-bit unsigned integer, in integer register rs1 into a floating-point number in floating-point register rd.
+
+        :Implementation:
+        | f[rd] = f32_{u32}(x[rs1])
+        """
+        rd, rs = self.parse_rd_rs(ins)
+        self.regs.set_f(rd, Float32.from_bytes(self.regs.get(rs).unsigned_value))
+
+    def instruction_fmv_w_x(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |11110|00   |00000|rs1  |000  |rd   |10100|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+
+
+        :Format:
+        | fmv.w.x    rd,rs1
+
+        :Description:
+        | Move the single-precision value encoded in IEEE 754-2008 standard encoding from the lower 32 bits of integer register rs1 to the floating-point register rd.
+
+        :Implementation:
+        | f[rd] = x[rs1][31:0]
+        """
+        rd, rs = self.parse_rd_rs(ins)
+        self.regs.set_f(rd, Float32.from_bytes(self.regs.get(rs).unsigned_value))
+
+    def instruction_flw(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7 |6-2  |1-0|
+        +-----+-----+-----+-----+-----+-----+-----+---+
+        |imm[11:0]        |rs1  |010  |rd   |00001|11 |
+        +-----+-----+-----+-----+-----+-----+-----+---+
+
+        :Format:
+          | flw        rd,offset(rs1)
+
+        :Description:
+          | Load a single-precision floating-point value from memory into floating-point register rd.
+
+        :Implementation:
+          | f[rd] = M[x[rs1] + sext(offset)][31:0]
+        """
+        rd, addr = self.parse_mem_ins(ins)
+        self.regs.set_f(rd, self.mmu.read_float(addr.value))
+
+    def instruction_fsw(self, ins: Instruction):
+        """
+        +-----+-----+-----+-----+-----+--------+-----+---+
+        |31-27|26-25|24-20|19-15|14-12|11-7    |6-2  |1-0|
+        +-----+-----+-----+-----+-----+--------+-----+---+
+        |imm[11:5]  |rs2  |rs1  |010  |imm[4:0]|01001|11 |
+        +-----+-----+-----+-----+-----+--------+-----+---+
+
+        :Format:
+          | fsw        rs2,offset(rs1)
+
+        :Description:
+          | Store a single-precision value from floating-point register rs2 to memory.
+
+        :Implementation:
+          | M[x[rs1] + sext(offset)] = f[rs2][31:0]
+        """
+        rs, addr = self.parse_mem_ins(ins)
+        val = self.regs.get_f(rs)
+        self.mmu.write(addr.value, 4, bytearray(val.bytes))
+
+    def parse_rd_rs(self, ins: Instruction) -> Tuple[str, str]:
+        assert len(ins.args) == 2
+        return ins.get_reg(0), ins.get_reg(1)
+
+    def parse_rd_rs_rs(self, ins: Instruction) -> Tuple[str, Float32, Float32]:
+        assert len(ins.args) == 3
+        return (
+            ins.get_reg(0),
+            self.regs.get_f(ins.get_reg(1)),
+            self.regs.get_f(ins.get_reg(2)),
+        )
+
+    def parse_rd_rs_rs_rs(
+        self, ins: Instruction
+    ) -> Tuple[str, Float32, Float32, Float32]:
+        assert len(ins.args) == 4
+        return (
+            ins.get_reg(0),
+            self.regs.get_f(ins.get_reg(1)),
+            self.regs.get_f(ins.get_reg(2)),
+            self.regs.get_f(ins.get_reg(3)),
+        )

riscemu/instructions/RV32I.py

@@ -22,241 +22,186 @@ class RV32I(InstructionSet):
     See https://maxvytech.com/images/RV32I-11-2018.pdf for a more detailed overview
     """
 
-    def instruction_lb(self, ins: 'Instruction'):
+    def instruction_lb(self, ins: "Instruction"):
         rd, addr = self.parse_mem_ins(ins)
         self.regs.set(rd, Int32.sign_extend(self.mmu.read(addr.unsigned_value, 1), 8))
 
-    def instruction_lh(self, ins: 'Instruction'):
+    def instruction_lh(self, ins: "Instruction"):
         rd, addr = self.parse_mem_ins(ins)
         self.regs.set(rd, Int32.sign_extend(self.mmu.read(addr.unsigned_value, 2), 16))
 
-    def instruction_lw(self, ins: 'Instruction'):
+    def instruction_lw(self, ins: "Instruction"):
         rd, addr = self.parse_mem_ins(ins)
         self.regs.set(rd, Int32(self.mmu.read(addr.unsigned_value, 4)))
 
-    def instruction_lbu(self, ins: 'Instruction'):
+    def instruction_lbu(self, ins: "Instruction"):
         rd, addr = self.parse_mem_ins(ins)
         self.regs.set(rd, Int32(self.mmu.read(addr.unsigned_value, 1)))
 
-    def instruction_lhu(self, ins: 'Instruction'):
+    def instruction_lhu(self, ins: "Instruction"):
         rd, addr = self.parse_mem_ins(ins)
         self.regs.set(rd, Int32(self.mmu.read(addr.unsigned_value, 2)))
 
-    def instruction_sb(self, ins: 'Instruction'):
+    def instruction_sb(self, ins: "Instruction"):
         rd, addr = self.parse_mem_ins(ins)
         self.mmu.write(addr.unsigned_value, 1, self.regs.get(rd).to_bytes(1))
 
-    def instruction_sh(self, ins: 'Instruction'):
+    def instruction_sh(self, ins: "Instruction"):
         rd, addr = self.parse_mem_ins(ins)
         self.mmu.write(addr.unsigned_value, 2, self.regs.get(rd).to_bytes(2))
 
-    def instruction_sw(self, ins: 'Instruction'):
+    def instruction_sw(self, ins: "Instruction"):
         rd, addr = self.parse_mem_ins(ins)
         self.mmu.write(addr.unsigned_value, 4, self.regs.get(rd).to_bytes(4))
 
-    def instruction_sll(self, ins: 'Instruction'):
+    def instruction_sll(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         dst = ins.get_reg(0)
         src1 = ins.get_reg(1)
         src2 = ins.get_reg(2)
-        self.regs.set(
-            dst,
-            self.regs.get(src1) << (self.regs.get(src2) & 0b11111)
-        )
+        self.regs.set(dst, self.regs.get(src1) << (self.regs.get(src2) & 0b11111))
 
-    def instruction_slli(self, ins: 'Instruction'):
+    def instruction_slli(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         dst = ins.get_reg(0)
         src1 = ins.get_reg(1)
         imm = ins.get_imm(2)
-        self.regs.set(
-            dst,
-            self.regs.get(src1) << (imm & 0b11111)
-        )
+        self.regs.set(dst, self.regs.get(src1) << (imm & 0b11111))
 
-    def instruction_srl(self, ins: 'Instruction'):
+    def instruction_srl(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         dst = ins.get_reg(0)
         src1 = ins.get_reg(1)
         src2 = ins.get_reg(2)
         self.regs.set(
-            dst,
-            self.regs.get(src1).shift_right_logical(self.regs.get(src2) & 0b11111)
+            dst, self.regs.get(src1).shift_right_logical(self.regs.get(src2) & 0b11111)
         )
 
-    def instruction_srli(self, ins: 'Instruction'):
+    def instruction_srli(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         dst = ins.get_reg(0)
         src1 = ins.get_reg(1)
         imm = ins.get_imm(2)
-        self.regs.set(
-            dst,
-            self.regs.get(src1).shift_right_logical(imm & 0b11111)
-        )
+        self.regs.set(dst, self.regs.get(src1).shift_right_logical(imm & 0b11111))
 
-    def instruction_sra(self, ins: 'Instruction'):
+    def instruction_sra(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         dst = ins.get_reg(0)
         src1 = ins.get_reg(1)
         src2 = ins.get_reg(2)
-        self.regs.set(
-            dst,
-            self.regs.get(src1) >> (self.regs.get(src2) & 0b11111)
-        )
+        self.regs.set(dst, self.regs.get(src1) >> (self.regs.get(src2) & 0b11111))
 
-    def instruction_srai(self, ins: 'Instruction'):
+    def instruction_srai(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         dst = ins.get_reg(0)
         src1 = ins.get_reg(1)
         imm = ins.get_imm(2)
-        self.regs.set(
-            dst,
-            self.regs.get(src1) >> (imm & 0b11111)
-        )
+        self.regs.set(dst, self.regs.get(src1) >> (imm & 0b11111))
 
-    def instruction_add(self, ins: 'Instruction'):
+    def instruction_add(self, ins: "Instruction"):
         # FIXME: once configuration is figured out, add flag to support immediate arg in add instruction
         dst, rs1, rs2 = self.parse_rd_rs_rs(ins)
 
-        self.regs.set(
-            dst,
-            rs1 + rs2
-        )
+        self.regs.set(dst, rs1 + rs2)
 
-    def instruction_addi(self, ins: 'Instruction'):
+    def instruction_addi(self, ins: "Instruction"):
         dst, rs1, imm = self.parse_rd_rs_imm(ins)
-        self.regs.set(
-            dst,
-            rs1 + imm
-        )
+        self.regs.set(dst, rs1 + imm)
 
-    def instruction_sub(self, ins: 'Instruction'):
+    def instruction_sub(self, ins: "Instruction"):
         dst, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            dst,
-            rs1 - rs2
-        )
+        self.regs.set(dst, rs1 - rs2)
 
-    def instruction_lui(self, ins: 'Instruction'):
+    def instruction_lui(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 2)
         reg = ins.get_reg(0)
         imm = UInt32(ins.get_imm(1) << 12)
         self.regs.set(reg, Int32(imm))
 
-    def instruction_auipc(self, ins: 'Instruction'):
+    def instruction_auipc(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 2)
         reg = ins.get_reg(0)
         imm = UInt32(ins.get_imm(1) << 12)
         self.regs.set(reg, imm.signed() + self.pc)
 
-    def instruction_xor(self, ins: 'Instruction'):
+    def instruction_xor(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            rd,
-            rs1 ^ rs2
-        )
+        self.regs.set(rd, rs1 ^ rs2)
 
-    def instruction_xori(self, ins: 'Instruction'):
+    def instruction_xori(self, ins: "Instruction"):
         rd, rs1, imm = self.parse_rd_rs_imm(ins)
-        self.regs.set(
-            rd,
-            rs1 ^ imm
-        )
+        self.regs.set(rd, rs1 ^ imm)
 
-    def instruction_or(self, ins: 'Instruction'):
+    def instruction_or(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            rd,
-            rs1 | rs2
-        )
+        self.regs.set(rd, rs1 | rs2)
 
-    def instruction_ori(self, ins: 'Instruction'):
+    def instruction_ori(self, ins: "Instruction"):
         rd, rs1, imm = self.parse_rd_rs_imm(ins)
-        self.regs.set(
-            rd,
-            rs1 | imm
-        )
+        self.regs.set(rd, rs1 | imm)
 
-    def instruction_and(self, ins: 'Instruction'):
+    def instruction_and(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            rd,
-            rs1 & rs2
-        )
+        self.regs.set(rd, rs1 & rs2)
 
-    def instruction_andi(self, ins: 'Instruction'):
+    def instruction_andi(self, ins: "Instruction"):
         rd, rs1, imm = self.parse_rd_rs_imm(ins)
-        self.regs.set(
-            rd,
-            rs1 & imm
-        )
+        self.regs.set(rd, rs1 & imm)
 
-    def instruction_slt(self, ins: 'Instruction'):
+    def instruction_slt(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            rd,
-            Int32(int(rs1 < rs2))
-        )
+        self.regs.set(rd, Int32(int(rs1 < rs2)))
 
-    def instruction_slti(self, ins: 'Instruction'):
+    def instruction_slti(self, ins: "Instruction"):
         rd, rs1, imm = self.parse_rd_rs_imm(ins)
-        self.regs.set(
-            rd,
-            Int32(int(rs1 < imm))
-        )
+        self.regs.set(rd, Int32(int(rs1 < imm)))
 
-    def instruction_sltu(self, ins: 'Instruction'):
+    def instruction_sltu(self, ins: "Instruction"):
         dst, rs1, rs2 = self.parse_rd_rs_rs(ins, signed=False)
-        self.regs.set(
-            dst,
-            Int32(int(rs1 < rs2))
-        )
+        self.regs.set(dst, Int32(int(rs1 < rs2)))
 
-    def instruction_sltiu(self, ins: 'Instruction'):
+    def instruction_sltiu(self, ins: "Instruction"):
         dst, rs1, imm = self.parse_rd_rs_imm(ins, signed=False)
-        self.regs.set(
-            dst,
-            Int32(int(rs1 < imm))
-        )
+        self.regs.set(dst, Int32(int(rs1 < imm)))
 
-    def instruction_beq(self, ins: 'Instruction'):
+    def instruction_beq(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins)
         if rs1 == rs2:
             self.pc = dst.unsigned_value
 
-    def instruction_bne(self, ins: 'Instruction'):
+    def instruction_bne(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins)
         if rs1 != rs2:
             self.pc = dst.unsigned_value
 
-    def instruction_blt(self, ins: 'Instruction'):
+    def instruction_blt(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins)
         if rs1 < rs2:
             self.pc = dst.unsigned_value
 
-    def instruction_bge(self, ins: 'Instruction'):
+    def instruction_bge(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins)
         if rs1 >= rs2:
             self.pc = dst.unsigned_value
 
-    def instruction_bltu(self, ins: 'Instruction'):
+    def instruction_bltu(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins, signed=False)
         if rs1 < rs2:
             self.pc = dst.unsigned_value
 
-    def instruction_bgeu(self, ins: 'Instruction'):
+    def instruction_bgeu(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins, signed=False)
         if rs1 >= rs2:
             self.pc = dst.unsigned_value
 
     # technically deprecated
-    def instruction_j(self, ins: 'Instruction'):
+    def instruction_j(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 1)
         addr = ins.get_imm(0)
         self.pc = addr
 
-    def instruction_jal(self, ins: 'Instruction'):
-        reg = 'ra'  # default register is ra
+    def instruction_jal(self, ins: "Instruction"):
+        reg = "ra"  # default register is ra
         if len(ins.args) == 1:
             addr = ins.get_imm(0)
         else:
@@ -266,58 +211,61 @@ class RV32I(InstructionSet):
         self.regs.set(reg, Int32(self.pc))
         self.pc = addr
 
-    def instruction_jalr(self, ins: 'Instruction'):
+    def instruction_jalr(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 2)
         reg = ins.get_reg(0)
-        addr = ins.get_imm(1)
+        base = ins.get_reg(1)
+        addr = ins.get_imm(2)
         self.regs.set(reg, Int32(self.pc))
-        self.pc = addr
+        self.pc = self.regs.get(base).unsigned_value + addr
 
-    def instruction_ret(self, ins: 'Instruction'):
+    def instruction_ret(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 0)
-        self.pc = self.regs.get('ra').value
+        self.pc = self.regs.get("ra").value
 
-    def instruction_ecall(self, ins: 'Instruction'):
+    def instruction_ecall(self, ins: "Instruction"):
         self.instruction_scall(ins)
 
-    def instruction_ebreak(self, ins: 'Instruction'):
+    def instruction_ebreak(self, ins: "Instruction"):
         self.instruction_sbreak(ins)
 
-    def instruction_scall(self, ins: 'Instruction'):
+    def instruction_scall(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 0)
 
         if not isinstance(self.cpu, UserModeCPU):
             # FIXME: add exception for syscall not supported or something
             raise
 
-        syscall = Syscall(self.regs.get('a7'), self.cpu)
+        syscall = Syscall(self.regs.get("a7"), self.cpu)
         self.cpu.syscall_int.handle_syscall(syscall)
 
-    def instruction_sbreak(self, ins: 'Instruction'):
+    def instruction_sbreak(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 0)
         if self.cpu.conf.debug_instruction:
-            print(FMT_DEBUG + "Debug instruction encountered at 0x{:08X}".format(self.pc - 1) + FMT_NONE)
+            print(
+                FMT_DEBUG
+                + "Debug instruction encountered at 0x{:08X}".format(self.pc - 1)
+                + FMT_NONE
+            )
             raise LaunchDebuggerException()
 
-    def instruction_nop(self, ins: 'Instruction'):
+    def instruction_nop(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 0)
         pass
 
-    def instruction_li(self, ins: 'Instruction'):
+    def instruction_li(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 2)
         reg = ins.get_reg(0)
         immediate = ins.get_imm(1)
         self.regs.set(reg, Int32(immediate))
 
-    def instruction_la(self, ins: 'Instruction'):
+    def instruction_la(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 2)
         reg = ins.get_reg(0)
         immediate = ins.get_imm(1)
         self.regs.set(reg, Int32(immediate))
 
-    def instruction_mv(self, ins: 'Instruction'):
+    def instruction_mv(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 2)
         rd, rs = ins.get_reg(0), ins.get_reg(1)
         self.regs.set(rd, self.regs.get(rs))
-
-

riscemu/instructions/RV32M.py

@@ -12,50 +12,33 @@ class RV32M(InstructionSet):
     """
     The RV32M Instruction set, containing multiplication and division instructions
     """
-    def instruction_mul(self, ins: 'Instruction'):
+
+    def instruction_mul(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            rd,
-            rs1 * rs2
-        )
+        self.regs.set(rd, rs1 * rs2)
 
-    def instruction_mulh(self, ins: 'Instruction'):
+    def instruction_mulh(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            rd,
-            (rs1 * rs2) >> 32
-        )
+        self.regs.set(rd, (rs1 * rs2) >> 32)
 
-    def instruction_mulhsu(self, ins: 'Instruction'):
+    def instruction_mulhsu(self, ins: "Instruction"):
         INS_NOT_IMPLEMENTED(ins)
 
-    def instruction_mulhu(self, ins: 'Instruction'):
+    def instruction_mulhu(self, ins: "Instruction"):
         INS_NOT_IMPLEMENTED(ins)
 
-    def instruction_div(self, ins: 'Instruction'):
+    def instruction_div(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            rd,
-            rs1 // rs2
-        )
+        self.regs.set(rd, rs1 // rs2)
 
-    def instruction_divu(self, ins: 'Instruction'):
+    def instruction_divu(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins, signed=False)
-        self.regs.set(
-            rd,
-            rs1 // rs2
-        )
+        self.regs.set(rd, rs1 // rs2)
 
-    def instruction_rem(self, ins: 'Instruction'):
+    def instruction_rem(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins)
-        self.regs.set(
-            rd,
-            rs1 % rs2
-        )
+        self.regs.set(rd, rs1 % rs2)
 
-    def instruction_remu(self, ins: 'Instruction'):
+    def instruction_remu(self, ins: "Instruction"):
         rd, rs1, rs2 = self.parse_rd_rs_rs(ins, signed=False)
-        self.regs.set(
-            rd,
-            rs1 % rs2
-        )
+        self.regs.set(rd, rs1 % rs2)

riscemu/instructions/RV_Debug.py

@@ -0,0 +1,33 @@
+from .instruction_set import InstructionSet, Instruction
+
+
+class RV_Debug(InstructionSet):
+    def instruction_print(self, ins: Instruction):
+        reg = ins.get_reg(0)
+        print("register {} contains value {}".format(reg, self.regs.get(reg)))
+
+    def instruction_print_float(self, ins: Instruction):
+        reg = ins.get_reg(0)
+        print("register {} contains value {}".format(reg, self.regs.get_f(reg).value))
+
+    def instruction_print_uint(self, ins: Instruction):
+        reg = ins.get_reg(0)
+        print(
+            "register {} contains value {}".format(
+                reg, self.regs.get(reg).unsigned_value
+            )
+        )
+
+    def instruction_print_hex(self, ins: Instruction):
+        reg = ins.get_reg(0)
+        print(
+            "register {} contains value {}".format(reg, hex(self.regs.get(reg).value))
+        )
+
+    def instruction_print_uhex(self, ins: Instruction):
+        reg = ins.get_reg(0)
+        print(
+            "register {} contains value {}".format(
+                reg, hex(self.regs.get(reg).unsigned_value)
+            )
+        )

riscemu/instructions/__init__.py

@@ -6,11 +6,11 @@ SPDX-License-Identifier: MIT
 This package holds all instruction sets, available to the processor
 """
 
-from .instruction_set import InstructionSet
+from .instruction_set import InstructionSet, Instruction
 from .RV32M import RV32M
 from .RV32I import RV32I
 from .RV32A import RV32A
+from .RV32F import RV32F
+from .RV_Debug import RV_Debug
 
-InstructionSetDict = {
-    v.__name__: v for v in [RV32I, RV32M, RV32A]
-}
+InstructionSetDict = {v.__name__: v for v in [RV32I, RV32M, RV32A, RV32F, RV_Debug]}

riscemu/instructions/instruction_set.py

@@ -23,23 +23,21 @@ class InstructionSet(ABC):
     instructions containing a dot '.' should replace it with an underscore.
     """
 
-    def __init__(self, cpu: 'CPU'):
+    def __init__(self, cpu: "CPU"):
         """Create a new instance of the Instruction set. This requires access to a CPU, and grabs vertain things
         from it such as access to the MMU and registers.
         """
         self.name = self.__class__.__name__
         self.cpu = cpu
 
-    def load(self) -> Dict[str, Callable[['Instruction'], None]]:
+    def load(self) -> Dict[str, Callable[["Instruction"], None]]:
         """
         This is called by the CPU once it instantiates this instruction set
 
         It returns a dictionary of all instructions in this instruction set,
         pointing to the correct handler for it
         """
-        return {
-            name: ins for name, ins in self.get_instructions()
-        }
+        return {name: ins for name, ins in self.get_instructions()}
 
     def get_instructions(self):
         """
@@ -48,10 +46,10 @@ class InstructionSet(ABC):
         converts underscores in names to dots
         """
         for member in dir(self):
-            if member.startswith('instruction_'):
-                yield member[12:].replace('_', '.'), getattr(self, member)
+            if member.startswith("instruction_"):
+                yield member[12:].replace("_", "."), getattr(self, member)
 
-    def parse_mem_ins(self, ins: 'Instruction') -> Tuple[str, Int32]:
+    def parse_mem_ins(self, ins: "Instruction") -> Tuple[str, Int32]:
         """
         parses both rd, rs, imm and rd, imm(rs) argument format and returns (rd, imm+rs1)
         (so a register and address tuple for memory instructions)
@@ -69,51 +67,69 @@ class InstructionSet(ABC):
         rd = ins.get_reg(0)
         return rd, rs + imm
 
-    def parse_rd_rs_rs(self, ins: 'Instruction', signed=True) -> Tuple[str, Int32, Int32]:
+    def parse_rd_rs_rs(
+        self, ins: "Instruction", signed=True
+    ) -> Tuple[str, Int32, Int32]:
         """
         Assumes the command is in <name> rd, rs1, rs2 format
         Returns the name of rd, and the values in rs1 and rs2
         """
         ASSERT_LEN(ins.args, 3)
         if signed:
-            return ins.get_reg(0), \
-                   self.get_reg_content(ins, 1), \
-                   self.get_reg_content(ins, 2)
+            return (
+                ins.get_reg(0),
+                Int32(self.get_reg_content(ins, 1)),
+                Int32(self.get_reg_content(ins, 2)),
+            )
         else:
-            return ins.get_reg(0), \
-                   UInt32(self.get_reg_content(ins, 1)), \
-                   UInt32(self.get_reg_content(ins, 2))
+            return (
+                ins.get_reg(0),
+                UInt32(self.get_reg_content(ins, 1)),
+                UInt32(self.get_reg_content(ins, 2)),
+            )
 
-    def parse_rd_rs_imm(self, ins: 'Instruction', signed=True) -> Tuple[str, Int32, Int32]:
+    def parse_rd_rs_imm(
+        self, ins: "Instruction", signed=True
+    ) -> Tuple[str, Int32, Int32]:
         """
         Assumes the command is in <name> rd, rs, imm format
         Returns the name of rd, the value in rs and the immediate imm
         """
         ASSERT_LEN(ins.args, 3)
         if signed:
-            return ins.get_reg(0), \
-                   Int32(self.get_reg_content(ins, 1)), \
-                   Int32(ins.get_imm(2))
+            return (
+                ins.get_reg(0),
+                Int32(self.get_reg_content(ins, 1)),
+                Int32(ins.get_imm(2)),
+            )
         else:
-            return ins.get_reg(0), \
-                   UInt32(self.get_reg_content(ins, 1)), \
-                   UInt32(ins.get_imm(2))
+            return (
+                ins.get_reg(0),
+                UInt32(self.get_reg_content(ins, 1)),
+                UInt32(ins.get_imm(2)),
+            )
 
-    def parse_rs_rs_imm(self, ins: 'Instruction', signed=True) -> Tuple[Int32, Int32, Int32]:
+    def parse_rs_rs_imm(
+        self, ins: "Instruction", signed=True
+    ) -> Tuple[Int32, Int32, Int32]:
         """
         Assumes the command is in <name> rs1, rs2, imm format
         Returns the values in rs1, rs2 and the immediate imm
         """
         if signed:
-            return Int32(self.get_reg_content(ins, 0)), \
-                   Int32(self.get_reg_content(ins, 1)), \
-                   Int32(ins.get_imm(2))
+            return (
+                Int32(self.get_reg_content(ins, 0)),
+                Int32(self.get_reg_content(ins, 1)),
+                Int32(ins.get_imm(2)),
+            )
         else:
-            return UInt32(self.get_reg_content(ins, 0)), \
-                   UInt32(self.get_reg_content(ins, 1)), \
-                   UInt32(ins.get_imm(2))
+            return (
+                UInt32(self.get_reg_content(ins, 0)),
+                UInt32(self.get_reg_content(ins, 1)),
+                UInt32(ins.get_imm(2)),
+            )
 
-    def get_reg_content(self, ins: 'Instruction', ind: int) -> Int32:
+    def get_reg_content(self, ins: "Instruction", ind: int) -> Int32:
         """
         get the register name from ins and then return the register contents
         """
@@ -140,6 +156,5 @@ class InstructionSet(ABC):
 
     def __repr__(self):
         return "InstructionSet[{}] with {} instructions".format(
-            self.__class__.__name__,
-            len(list(self.get_instructions()))
+            self.__class__.__name__, len(list(self.get_instructions()))
         )

riscemu/interactive.py

@@ -3,26 +3,34 @@ import sys
 from riscemu import RunConfig
 from riscemu.types import InstructionMemorySection, SimpleInstruction, Program
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     from .CPU import UserModeCPU
     from .instructions import InstructionSetDict
     from .debug import launch_debug_session
 
     cpu = UserModeCPU(list(InstructionSetDict.values()), RunConfig(verbosity=4))
 
-    program = Program('interactive session', base=0x100)
+    program = Program("interactive session", base=0x100)
     context = program.context
-    program.add_section(InstructionMemorySection([
-        SimpleInstruction('ebreak', (), context, 0x100),
-        SimpleInstruction('addi', ('a0', 'zero', '0'), context, 0x104),
-        SimpleInstruction('addi', ('a7', 'zero', '93'), context, 0x108),
-        SimpleInstruction('scall', (), context, 0x10C),
-    ], '.text', context, program.name, 0x100))
+    program.add_section(
+        InstructionMemorySection(
+            [
+                SimpleInstruction("ebreak", (), context, 0x100),
+                SimpleInstruction("addi", ("a0", "zero", "0"), context, 0x104),
+                SimpleInstruction("addi", ("a7", "zero", "93"), context, 0x108),
+                SimpleInstruction("scall", (), context, 0x10C),
+            ],
+            ".text",
+            context,
+            program.name,
+            0x100,
+        )
+    )
 
     cpu.load_program(program)
 
     cpu.setup_stack()
 
-    cpu.launch(program)
+    cpu.launch()
 
     sys.exit(cpu.exit_code)

riscemu/parser.py

@@ -16,28 +16,32 @@ from .types.exceptions import ParseException
 
 def parse_instruction(token: Token, args: Tuple[str], context: ParseContext):
     if context.section is None:
-        context.new_section('.text', MemorySectionType.Instructions)
+        context.new_section(".text", MemorySectionType.Instructions)
     if context.section.type != MemorySectionType.Instructions:
-        raise ParseException("{} {} encountered in invalid context: {}".format(token, args, context))
-    ins = SimpleInstruction(token.value, args, context.context, context.current_address())
+        raise ParseException(
+            "{} {} encountered in invalid context: {}".format(token, args, context)
+        )
+    ins = SimpleInstruction(
+        token.value, args, context.context, context.current_address()
+    )
     context.section.data.append(ins)
 
 
 def parse_label(token: Token, args: Tuple[str], context: ParseContext):
     name = token.value[:-1]
-    if re.match(r'^\d+$', name):
+    if re.match(r"^\d+$", name):
         # relative label:
         context.context.numbered_labels[name].append(context.current_address())
     else:
         if name in context.context.labels:
-            print(FMT_PARSE + 'Warn: Symbol {} defined twice!'.format(name))
+            print(FMT_PARSE + "Warn: Symbol {} defined twice!".format(name))
         context.add_label(name, context.current_address(), is_relative=True)
 
 
 PARSERS: Dict[TokenType, Callable[[Token, Tuple[str], ParseContext], None]] = {
     TokenType.PSEUDO_OP: AssemblerDirectives.handle_instruction,
     TokenType.LABEL: parse_label,
-    TokenType.INSTRUCTION_NAME: parse_instruction
+    TokenType.INSTRUCTION_NAME: parse_instruction,
 }
 
 
@@ -58,7 +62,9 @@ def parse_tokens(name: str, tokens_iter: Iterable[Token]) -> Program:
     return context.finalize()
 
 
-def composite_tokenizer(tokens_iter: Iterable[Token]) -> Iterable[Tuple[Token, Tuple[str]]]:
+def composite_tokenizer(
+    tokens_iter: Iterable[Token],
+) -> Iterable[Tuple[Token, Tuple[str]]]:
     """
     Convert an iterator over tokens into an iterator over tuples: (token, list(token))
 
@@ -71,7 +77,11 @@ def composite_tokenizer(tokens_iter: Iterable[Token]) -> Iterable[Tuple[Token, T
 
     while not tokens.is_empty():
         token = next(tokens)
-        if token.type in (TokenType.PSEUDO_OP, TokenType.LABEL, TokenType.INSTRUCTION_NAME):
+        if token.type in (
+            TokenType.PSEUDO_OP,
+            TokenType.LABEL,
+            TokenType.INSTRUCTION_NAME,
+        ):
             yield token, tuple(take_arguments(tokens))
 
 
@@ -106,10 +116,14 @@ class AssemblyFileLoader(ProgramLoader):
 
     The AssemblyFileLoader loads .asm, .S and .s files by default, and acts as a weak fallback to all other filetypes.
     """
+
     def parse(self) -> Program:
-        with open(self.source_path, 'r') as f:
+        with open(self.source_path, "r") as f:
             return parse_tokens(self.filename, tokenize(f))
 
+    def parse_io(self, io: Iterable[str]):
+        return parse_tokens(self.filename, tokenize(io))
+
     @classmethod
     def can_parse(cls, source_path: str) -> float:
         """
@@ -120,10 +134,10 @@ class AssemblyFileLoader(ProgramLoader):
         :return:
         """
         # gcc recognizes these line endings as assembly. So we will do too.
-        if source_path.split('.')[-1] in ('asm', 'S', 's'):
+        if source_path.split(".")[-1] in ("asm", "S", "s"):
             return 1
         return 0.01
 
     @classmethod
-    def get_options(cls, argv: List[str]) -> [List[str], T_ParserOpts]:
+    def get_options(cls, argv: List[str]) -> Tuple[List[str], T_ParserOpts]:
         return argv, {}

riscemu/priv/CSR.py

@@ -12,6 +12,7 @@ class CSR:
     """
     This holds all Control and Status Registers (CSR)
     """
+
     regs: Dict[int, UInt32]
     """
     All Control and Status Registers are stored here
@@ -52,7 +53,9 @@ class CSR:
             return self.virtual_regs[addr]()
         return self.regs[addr]
 
-    def set_listener(self, addr: Union[str, int], listener: Callable[[UInt32, UInt32], None]):
+    def set_listener(
+        self, addr: Union[str, int], listener: Callable[[UInt32, UInt32], None]
+    ):
         addr = self._name_to_addr(addr)
         if addr is None:
             print("unknown csr address name: {}".format(addr))
@@ -73,11 +76,11 @@ class CSR:
         """
         size = 2 if name in MSTATUS_LEN_2 else 1
         off = MSTATUS_OFFSETS[name]
-        mask = (2 ** size - 1) << off
-        old_val = self.get('mstatus')
+        mask = (2**size - 1) << off
+        old_val = self.get("mstatus")
         erased = old_val & (~mask)
         new_val = erased | (val << off)
-        self.set('mstatus', new_val)
+        self.set("mstatus", new_val)
 
     def get_mstatus(self, name) -> UInt32:
         if not self.mstatus_cache_dirty and name in self.mstatus_cache:
@@ -85,8 +88,8 @@ class CSR:
 
         size = 2 if name in MSTATUS_LEN_2 else 1
         off = MSTATUS_OFFSETS[name]
-        mask = (2 ** size - 1) << off
-        val = (self.get('mstatus') & mask) >> off
+        mask = (2**size - 1) << off
+        val = (self.get("mstatus") & mask) >> off
         if self.mstatus_cache_dirty:
             self.mstatus_cache = dict(name=val)
         else:

riscemu/priv/CSRConsts.py

@@ -1,81 +1,81 @@
 from typing import Dict, Tuple
 
-MCAUSE_TRANSLATION: Dict[Tuple[int, int], str]= {
-    (1, 0): 'User software interrupt',
-    (1, 1): 'Supervisor software interrupt',
-    (1, 3): 'Machine software interrupt',
-    (1, 4): 'User timer interrupt',
-    (1, 5): 'Supervisor timer interrupt',
-    (1, 7): 'Machine timer interrupt',
-    (1, 8): 'User external interrupt',
-    (1, 9): 'Supervisor external interrupt',
-    (1, 11): 'Machine external interrupt',
-    (0, 0): 'Instruction address misaligned',
-    (0, 1): 'Instruction access fault',
-    (0, 2): 'Illegal instruction',
-    (0, 3): 'Breakpoint',
-    (0, 4): 'Load address misaligned',
-    (0, 5): 'Load access fault',
-    (0, 6): 'Store/AMO address misaligned',
-    (0, 7): 'Store/AMO access fault',
-    (0, 8): 'environment call from user mode',
-    (0, 9): 'environment call from supervisor mode',
-    (0, 11): 'environment call from machine mode',
-    (0, 12): 'Instruction page fault',
-    (0, 13): 'Load page fault',
-    (0, 15): 'Store/AMO page fault',
+MCAUSE_TRANSLATION: Dict[Tuple[int, int], str] = {
+    (1, 0): "User software interrupt",
+    (1, 1): "Supervisor software interrupt",
+    (1, 3): "Machine software interrupt",
+    (1, 4): "User timer interrupt",
+    (1, 5): "Supervisor timer interrupt",
+    (1, 7): "Machine timer interrupt",
+    (1, 8): "User external interrupt",
+    (1, 9): "Supervisor external interrupt",
+    (1, 11): "Machine external interrupt",
+    (0, 0): "Instruction address misaligned",
+    (0, 1): "Instruction access fault",
+    (0, 2): "Illegal instruction",
+    (0, 3): "Breakpoint",
+    (0, 4): "Load address misaligned",
+    (0, 5): "Load access fault",
+    (0, 6): "Store/AMO address misaligned",
+    (0, 7): "Store/AMO access fault",
+    (0, 8): "environment call from user mode",
+    (0, 9): "environment call from supervisor mode",
+    (0, 11): "environment call from machine mode",
+    (0, 12): "Instruction page fault",
+    (0, 13): "Load page fault",
+    (0, 15): "Store/AMO page fault",
 }
 """
 Assigns tuple (interrupt bit, exception code) to their respective readable names
 """
 
 MSTATUS_OFFSETS: Dict[str, int] = {
-    'uie': 0,
-    'sie': 1,
-    'mie': 3,
-    'upie': 4,
-    'spie': 5,
-    'mpie': 7,
-    'spp': 8,
-    'mpp': 11,
-    'fs': 13,
-    'xs': 15,
-    'mpriv': 17,
-    'sum': 18,
-    'mxr': 19,
-    'tvm': 20,
-    'tw': 21,
-    'tsr': 22,
-    'sd': 31
+    "uie": 0,
+    "sie": 1,
+    "mie": 3,
+    "upie": 4,
+    "spie": 5,
+    "mpie": 7,
+    "spp": 8,
+    "mpp": 11,
+    "fs": 13,
+    "xs": 15,
+    "mpriv": 17,
+    "sum": 18,
+    "mxr": 19,
+    "tvm": 20,
+    "tw": 21,
+    "tsr": 22,
+    "sd": 31,
 }
 """
 Offsets for all mstatus bits
 """
 
-MSTATUS_LEN_2 = ('mpp', 'fs', 'xs')
+MSTATUS_LEN_2 = ("mpp", "fs", "xs")
 """
 All mstatus parts that have length 2. All other mstatus parts have length 1
 """
 
 CSR_NAME_TO_ADDR: Dict[str, int] = {
-    'mstatus': 0x300,
-    'misa': 0x301,
-    'mie': 0x304,
-    'mtvec': 0x305,
-    'mepc': 0x341,
-    'mcause': 0x342,
-    'mtval': 0x343,
-    'mip': 0x344,
-    'mvendorid': 0xF11,
-    'marchid': 0xF12,
-    'mimpid': 0xF13,
-    'mhartid': 0xF14,
-    'time': 0xc01,
-    'timeh': 0xc81,
-    'halt': 0x789,
-    'mtimecmp': 0x780,
-    'mtimecmph': 0x781,
+    "mstatus": 0x300,
+    "misa": 0x301,
+    "mie": 0x304,
+    "mtvec": 0x305,
+    "mepc": 0x341,
+    "mcause": 0x342,
+    "mtval": 0x343,
+    "mip": 0x344,
+    "mvendorid": 0xF11,
+    "marchid": 0xF12,
+    "mimpid": 0xF13,
+    "mhartid": 0xF14,
+    "time": 0xC01,
+    "timeh": 0xC81,
+    "halt": 0x789,
+    "mtimecmp": 0x780,
+    "mtimecmph": 0x781,
 }
 """
 Translation for named registers
-"""
+"""

riscemu/priv/ElfLoader.py

@@ -11,7 +11,7 @@ if typing.TYPE_CHECKING:
     from elftools.elf.elffile import ELFFile
     from elftools.elf.sections import Section, SymbolTableSection
 
-INCLUDE_SEC = ('.text', '.stack', '.bss', '.sdata', '.sbss')
+INCLUDE_SEC = (".text", ".stack", ".bss", ".sdata", ".sbss")
 
 
 class ElfBinaryFileLoader(ProgramLoader):
@@ -20,6 +20,7 @@ class ElfBinaryFileLoader(ProgramLoader):
 
     This loader respects local and global symbols.
     """
+
     program: Program
 
     def __init__(self, source_path: str, options: T_ParserOpts):
@@ -28,8 +29,8 @@ class ElfBinaryFileLoader(ProgramLoader):
 
     @classmethod
     def can_parse(cls, source_path: str) -> float:
-        with open(source_path, 'rb') as f:
-            if f.read(4) == b'\x7f\x45\x4c\x46':
+        with open(source_path, "rb") as f:
+            if f.read(4) == b"\x7f\x45\x4c\x46":
                 return 1
         return 0
 
@@ -42,23 +43,33 @@ class ElfBinaryFileLoader(ProgramLoader):
             from elftools.elf.elffile import ELFFile
             from elftools.elf.sections import Section, SymbolTableSection
 
-            with open(self.source_path, 'rb') as f:
-                print(FMT_ELF + "[ElfLoader] Loading elf executable from: {}".format(self.source_path) + FMT_NONE)
+            with open(self.source_path, "rb") as f:
+                print(
+                    FMT_ELF
+                    + "[ElfLoader] Loading elf executable from: {}".format(
+                        self.source_path
+                    )
+                    + FMT_NONE
+                )
                 self._read_elf(ELFFile(f))
         except ImportError as e:
-            print(FMT_PARSE + "[ElfLoader] Cannot load elf files without PyElfTools package! You can install them "
-                              "using pip install pyelftools!" + FMT_NONE)
+            print(
+                FMT_PARSE
+                + "[ElfLoader] Cannot load elf files without PyElfTools package! You can install them "
+                "using pip install pyelftools!" + FMT_NONE
+            )
             raise e
 
         return self.program
 
-    def _read_elf(self, elf: 'ELFFile'):
-        if not elf.header.e_machine == 'EM_RISCV':
+    def _read_elf(self, elf: "ELFFile"):
+        if not elf.header.e_machine == "EM_RISCV":
             raise InvalidElfException("Not a RISC-V elf file!")
-        if not elf.header.e_ident.EI_CLASS == 'ELFCLASS32':
+        if not elf.header.e_ident.EI_CLASS == "ELFCLASS32":
             raise InvalidElfException("Only 32bit executables are supported!")
 
         from elftools.elf.sections import SymbolTableSection
+
         for sec in elf.iter_sections():
             if isinstance(sec, SymbolTableSection):
                 self._parse_symtab(sec)
@@ -69,18 +80,22 @@ class ElfBinaryFileLoader(ProgramLoader):
 
             self._add_sec(self._lms_from_elf_sec(sec, self.filename))
 
-    def _lms_from_elf_sec(self, sec: 'Section', owner: str):
-        is_code = sec.name in ('.text',)
+    def _lms_from_elf_sec(self, sec: "Section", owner: str):
+        is_code = sec.name in (".text",)
         data = bytearray(sec.data())
         if len(data) < sec.data_size:
             data += bytearray(len(data) - sec.data_size)
         flags = MemoryFlags(is_code, is_code)
-        print(FMT_ELF + "[ElfLoader] Section {} at: {:X}".format(sec.name, sec.header.sh_addr) + FMT_NONE)
+        print(
+            FMT_ELF
+            + "[ElfLoader] Section {} at: {:X}".format(sec.name, sec.header.sh_addr)
+            + FMT_NONE
+        )
         return ElfMemorySection(
             data, sec.name, self.program.context, owner, sec.header.sh_addr, flags
         )
 
-    def _parse_symtab(self, symtab: 'SymbolTableSection'):
+    def _parse_symtab(self, symtab: "SymbolTableSection"):
         from elftools.elf.enums import ENUM_ST_VISIBILITY
 
         for sym in symtab.iter_symbols():
@@ -88,18 +103,26 @@ class ElfBinaryFileLoader(ProgramLoader):
                 continue
             self.program.context.labels[sym.name] = sym.entry.st_value
             # check if it has st_visibility bit set
-            if sym.entry.st_info.bind == 'STB_GLOBAL':
+            if sym.entry.st_info.bind == "STB_GLOBAL":
                 self.program.global_labels.add(sym.name)
-                print(FMT_PARSE + "LOADED GLOBAL SYMBOL {}: {}".format(sym.name, sym.entry.st_value) + FMT_NONE)
+                print(
+                    FMT_PARSE
+                    + "LOADED GLOBAL SYMBOL {}: {}".format(sym.name, sym.entry.st_value)
+                    + FMT_NONE
+                )
 
-    def _add_sec(self, new_sec: 'ElfMemorySection'):
+    def _add_sec(self, new_sec: "ElfMemorySection"):
         for sec in self.program.sections:
             if sec.base < sec.end <= new_sec.base or sec.end > sec.base >= new_sec.end:
                 continue
             else:
-                print(FMT_ELF + "[ElfLoader] Invalid elf layout: Two sections overlap: \n\t{}\n\t{}".format(
-                    sec, new_sec
-                ) + FMT_NONE)
+                print(
+                    FMT_ELF
+                    + "[ElfLoader] Invalid elf layout: Two sections overlap: \n\t{}\n\t{}".format(
+                        sec, new_sec
+                    )
+                    + FMT_NONE
+                )
                 raise RuntimeError("Cannot load elf with overlapping sections!")
 
         self.program.add_section(new_sec)

riscemu/priv/Exceptions.py

@@ -42,10 +42,12 @@ class CpuTrap(BaseException):
 
     priv: PrivModes
     """
-    The privilege level this trap targets 
+    The privilege level this trap targets
     """
 
-    def __init__(self, code: int, mtval, type: CpuTrapType, priv: PrivModes = PrivModes.MACHINE):
+    def __init__(
+        self, code: int, mtval, type: CpuTrapType, priv: PrivModes = PrivModes.MACHINE
+    ):
         self.interrupt = 0 if type == CpuTrapType.EXCEPTION else 1
         self.code = code
         self.mtval = UInt32(mtval)
@@ -63,7 +65,9 @@ class CpuTrap(BaseException):
         name = "Reserved interrupt({}, {})".format(self.interrupt, self.code)
 
         if (self.interrupt, self.code) in MCAUSE_TRANSLATION:
-            name = MCAUSE_TRANSLATION[(self.interrupt, self.code)] + "({}, {})".format(self.interrupt, self.code)
+            name = MCAUSE_TRANSLATION[(self.interrupt, self.code)] + "({}, {})".format(
+                self.interrupt, self.code
+            )
 
         return "{} {{priv={}, type={}, mtval={:x}}} {}".format(
             name, self.priv.name, self.type.name, self.mtval, self.message()
@@ -74,7 +78,7 @@ class CpuTrap(BaseException):
 
 
 class IllegalInstructionTrap(CpuTrap):
-    def __init__(self, ins: 'ElfInstruction'):
+    def __init__(self, ins: "ElfInstruction"):
         super().__init__(2, ins.encoded, CpuTrapType.EXCEPTION)
 
 
@@ -104,7 +108,9 @@ class InvalidElfException(RiscemuBaseException):
         self.msg = msg
 
     def message(self):
-        return FMT_PARSE + "{}(\"{}\")".format(self.__class__.__name__, self.msg) + FMT_NONE
+        return (
+            FMT_PARSE + '{}("{}")'.format(self.__class__.__name__, self.msg) + FMT_NONE
+        )
 
 
 class LoadAccessFault(CpuTrap):
@@ -117,8 +123,5 @@ class LoadAccessFault(CpuTrap):
 
     def message(self):
         return "(During {} at 0x{:08x} of size {}: {})".format(
-            self.op,
-            self.addr,
-            self.size,
-            self.msg
+            self.op, self.addr, self.size, self.msg
         )

riscemu/priv/ImageLoader.py

@@ -14,10 +14,9 @@ from ..types import MemoryFlags, ProgramLoader, Program, T_ParserOpts
 
 
 class MemoryImageLoader(ProgramLoader):
-
     @classmethod
     def can_parse(cls, source_path: str) -> float:
-        if source_path.split('.')[-1] == 'img':
+        if source_path.split(".")[-1] == "img":
             return 1
         return 0
 
@@ -26,14 +25,14 @@ class MemoryImageLoader(ProgramLoader):
         return argv, {}
 
     def parse(self) -> Iterable[Program]:
-        if 'debug' not in self.options:
+        if "debug" not in self.options:
             yield self.parse_no_debug()
             return
 
-        with open(self.options.get('debug'), 'r') as debug_file:
+        with open(self.options.get("debug"), "r") as debug_file:
             debug_info = MemoryImageDebugInfos.load(debug_file.read())
 
-        with open(self.source_path, 'rb') as source_file:
+        with open(self.source_path, "rb") as source_file:
             data: bytearray = bytearray(source_file.read())
 
         for name, sections in debug_info.sections.items():
@@ -43,11 +42,15 @@ class MemoryImageLoader(ProgramLoader):
                 if program.base is None:
                     program.base = start
 
-                #in_code_sec = get_section_base_name(sec_name) in INSTRUCTION_SECTION_NAMES
+                # in_code_sec = get_section_base_name(sec_name) in INSTRUCTION_SECTION_NAMES
                 program.add_section(
                     ElfMemorySection(
-                        data[start:start+size], sec_name, program.context,
-                        name, start, MemoryFlags(False, True)
+                        data[start : start + size],
+                        sec_name,
+                        program.context,
+                        name,
+                        start,
+                        MemoryFlags(False, True),
                     )
                 )
 
@@ -57,19 +60,27 @@ class MemoryImageLoader(ProgramLoader):
             yield program
 
     def parse_no_debug(self) -> Program:
-        print(FMT_PARSE + "[MemoryImageLoader] Warning: loading memory image without debug information!" + FMT_NONE)
+        print(
+            FMT_PARSE
+            + "[MemoryImageLoader] Warning: loading memory image without debug information!"
+            + FMT_NONE
+        )
 
-        with open(self.source_path, 'rb') as source_file:
+        with open(self.source_path, "rb") as source_file:
             data: bytes = source_file.read()
 
         p = Program(self.filename)
-        p.add_section(ElfMemorySection(
-            bytearray(data), '.text', p.context, p.name, 0, MemoryFlags(False, True)
-        ))
+        p.add_section(
+            ElfMemorySection(
+                bytearray(data), ".text", p.context, p.name, 0, MemoryFlags(False, True)
+            )
+        )
         return p
 
     @classmethod
-    def instantiate(cls, source_path: str, options: T_ParserOpts) -> 'ProgramLoader':
-        if os.path.isfile(source_path + '.dbg'):
-            return MemoryImageLoader(source_path, dict(**options, debug=source_path + '.dbg'))
+    def instantiate(cls, source_path: str, options: T_ParserOpts) -> "ProgramLoader":
+        if os.path.isfile(source_path + ".dbg"):
+            return MemoryImageLoader(
+                source_path, dict(**options, debug=source_path + ".dbg")
+            )
         return MemoryImageLoader(source_path, options)

riscemu/priv/PrivCPU.py

@@ -87,7 +87,11 @@ class PrivCPU(CPU):
 
         if self.halted:
             print()
-            print(FMT_CPU + "[CPU] System halted with code {}".format(self.exit_code) + FMT_NONE)
+            print(
+                FMT_CPU
+                + "[CPU] System halted with code {}".format(self.exit_code)
+                + FMT_NONE
+            )
             sys.exit(self.exit_code)
 
         elif launch_debug:
@@ -96,55 +100,68 @@ class PrivCPU(CPU):
                 self.run(verbose)
         else:
             print()
-            print(FMT_CPU + "[CPU] System stopped without halting - perhaps you stopped the debugger?" + FMT_NONE)
+            print(
+                FMT_CPU
+                + "[CPU] System stopped without halting - perhaps you stopped the debugger?"
+                + FMT_NONE
+            )
 
     def launch(self, program: Optional[Program] = None, verbose: bool = False):
-        print(FMT_CPU + '[CPU] Started running from 0x{:08X} ({})'.format(self.pc, "kernel") + FMT_NONE)
+        print(
+            FMT_CPU
+            + "[CPU] Started running from 0x{:08X} ({})".format(self.pc, "kernel")
+            + FMT_NONE
+        )
         self._time_start = time.perf_counter_ns() // self.TIME_RESOLUTION_NS
 
         self.run(self.conf.verbosity > 1 or verbose)
 
     def load_program(self, program: Program):
-        if program.name == 'kernel':
+        if program.name == "kernel":
             self.pc = program.entrypoint
         super().load_program(program)
 
     def _init_csr(self):
         # set up CSR
         self.csr = CSR()
-        self.csr.set('mhartid', UInt32(0))  # core id
+        self.csr.set("mhartid", UInt32(0))  # core id
         # TODO: set correct value
-        self.csr.set('mimpid', UInt32(0))  # implementation id
+        self.csr.set("mimpid", UInt32(0))  # implementation id
         # set mxl to 1 (32 bit) and set bits for i and m isa
-        self.csr.set('misa', UInt32((1 << 30) + (1 << 8) + (1 << 12)))  # available ISA
+        self.csr.set("misa", UInt32((1 << 30) + (1 << 8) + (1 << 12)))  # available ISA
 
         # CSR write callbacks:
 
-        @self.csr.callback('halt')
+        @self.csr.callback("halt")
         def halt(old: UInt32, new: UInt32):
             if new != 0:
                 self.halted = True
                 self.exit_code = new.value
 
-        @self.csr.callback('mtimecmp')
+        @self.csr.callback("mtimecmp")
         def mtimecmp(old: UInt32, new: UInt32):
-            self._time_timecmp = (self.csr.get('mtimecmph') << 32) + new
+            self._time_timecmp = (self.csr.get("mtimecmph") << 32) + new
             self._time_interrupt_enabled = True
 
-        @self.csr.callback('mtimecmph')
+        @self.csr.callback("mtimecmph")
         def mtimecmph(old: UInt32, new: UInt32):
-            self._time_timecmp = (new << 32) + self.csr.get('mtimecmp')
+            self._time_timecmp = (new << 32) + self.csr.get("mtimecmp")
             self._time_interrupt_enabled = True
 
         # virtual CSR registers:
 
-        @self.csr.virtual_register('time')
+        @self.csr.virtual_register("time")
         def get_time():
-            return UInt32(time.perf_counter_ns() // self.TIME_RESOLUTION_NS - self._time_start)
+            return UInt32(
+                time.perf_counter_ns() // self.TIME_RESOLUTION_NS - self._time_start
+            )
 
-        @self.csr.virtual_register('timeh')
+        @self.csr.virtual_register("timeh")
         def get_timeh():
-            return UInt32((time.perf_counter_ns() // self.TIME_RESOLUTION_NS - self._time_start) >> 32)
+            return UInt32(
+                (time.perf_counter_ns() // self.TIME_RESOLUTION_NS - self._time_start)
+                >> 32
+            )
 
         # add minstret and mcycle counters
 
@@ -160,17 +177,21 @@ class PrivCPU(CPU):
             self._check_interrupt()
             ins = self.mmu.read_ins(self.pc)
             if verbose and (self.mode == PrivModes.USER or self.conf.verbosity > 4):
-                print(FMT_CPU + "   Running 0x{:08X}:{} {}".format(self.pc, FMT_NONE, ins))
+                print(
+                    FMT_CPU + "   Running 0x{:08X}:{} {}".format(self.pc, FMT_NONE, ins)
+                )
             self.run_instruction(ins)
             self.pc += self.INS_XLEN
         except CpuTrap as trap:
             self._handle_trap(trap)
             if trap.interrupt == 0 and not isinstance(trap, EcallTrap):
-                print(FMT_CPU + "[CPU] Trap {} encountered at {} (0x{:x})".format(
-                    trap,
-                    self.mmu.translate_address(self.pc),
-                    self.pc
-                ) + FMT_NONE)
+                print(
+                    FMT_CPU
+                    + "[CPU] Trap {} encountered at {} (0x{:x})".format(
+                        trap, self.mmu.translate_address(self.pc), self.pc
+                    )
+                    + FMT_NONE
+                )
                 breakpoint()
                 if self.conf.debug_on_exception:
                     raise LaunchDebuggerException()
@@ -179,12 +200,15 @@ class PrivCPU(CPU):
     def _timer_step(self):
         if not self._time_interrupt_enabled:
             return
-        if self._time_timecmp <= (time.perf_counter_ns() // self.TIME_RESOLUTION_NS) - self._time_start:
+        if (
+            self._time_timecmp
+            <= (time.perf_counter_ns() // self.TIME_RESOLUTION_NS) - self._time_start
+        ):
             self.pending_traps.append(TimerInterrupt())
             self._time_interrupt_enabled = False
 
     def _check_interrupt(self):
-        if not (len(self.pending_traps) > 0 and self.csr.get_mstatus('mie')):
+        if not (len(self.pending_traps) > 0 and self.csr.get_mstatus("mie")):
             return
         # select best interrupt
         # FIXME: actually select based on the official ranking
@@ -194,24 +218,30 @@ class PrivCPU(CPU):
             self.regs.dump_reg_a()
 
         if trap.priv != PrivModes.MACHINE:
-            print(FMT_CPU + "[CPU] Trap not targeting machine mode encountered! - undefined behaviour!" + FMT_NONE)
+            print(
+                FMT_CPU
+                + "[CPU] Trap not targeting machine mode encountered! - undefined behaviour!"
+                + FMT_NONE
+            )
             raise Exception("Undefined behaviour!")
 
         if self.mode != PrivModes.USER:
             print(FMT_CPU + "[CPU] Trap triggered outside of user mode?!" + FMT_NONE)
 
-        self.csr.set_mstatus('mpie', self.csr.get_mstatus('mie'))
-        self.csr.set_mstatus('mpp', self.mode.value)
-        self.csr.set_mstatus('mie', UInt32(0))
-        self.csr.set('mcause', trap.mcause)
-        self.csr.set('mepc', self.pc - self.INS_XLEN)
-        self.csr.set('mtval', trap.mtval)
+        self.csr.set_mstatus("mpie", self.csr.get_mstatus("mie"))
+        self.csr.set_mstatus("mpp", self.mode.value)
+        self.csr.set_mstatus("mie", UInt32(0))
+        self.csr.set("mcause", trap.mcause)
+        self.csr.set("mepc", self.pc - self.INS_XLEN)
+        self.csr.set("mtval", trap.mtval)
         self.mode = trap.priv
-        mtvec = self.csr.get('mtvec')
+        mtvec = self.csr.get("mtvec")
         if mtvec & 0b11 == 0:
             self.pc = mtvec.value
         if mtvec & 0b11 == 1:
-            self.pc = ((mtvec & 0b11111111111111111111111111111100) + (trap.code * 4)).value
+            self.pc = (
+                (mtvec & 0b11111111111111111111111111111100) + (trap.code * 4)
+            ).value
         self.record_perf_profile()
         if len(self._perf_counters) > 100:
             self.show_perf()
@@ -232,7 +262,10 @@ class PrivCPU(CPU):
             cycled = cycle
             timed = time_ns
             cps_list.append(cps)
-        print("    on average {:.0f} instructions/s".format(sum(cps_list) / len(cps_list)) + FMT_NONE)
+        print(
+            "    on average {:.0f} instructions/s".format(sum(cps_list) / len(cps_list))
+            + FMT_NONE
+        )
         self._perf_counters = list()
 
     def record_perf_profile(self):
@@ -240,6 +273,4 @@ class PrivCPU(CPU):
 
     @classmethod
     def get_loaders(cls) -> typing.Iterable[Type[ProgramLoader]]:
-        return [
-            AssemblyFileLoader, MemoryImageLoader, ElfBinaryFileLoader
-        ]
+        return [AssemblyFileLoader, MemoryImageLoader, ElfBinaryFileLoader]

riscemu/priv/PrivMMU.py

@@ -9,7 +9,6 @@ if typing.TYPE_CHECKING:
 
 
 class PrivMMU(MMU):
-
     def get_sec_containing(self, addr: T_AbsoluteAddress) -> MemorySection:
         # try to get an existing section
         existing_sec = super().get_sec_containing(addr)
@@ -18,7 +17,9 @@ class PrivMMU(MMU):
             return existing_sec
 
         # get section preceding empty space at addr
-        sec_before = next((sec for sec in reversed(self.sections) if sec.end < addr), None)
+        sec_before = next(
+            (sec for sec in reversed(self.sections) if sec.end < addr), None
+        )
         # get sec succeeding empty space at addr
         sec_after = next((sec for sec in self.sections if sec.base > addr), None)
 
@@ -31,7 +32,14 @@ class PrivMMU(MMU):
         # end at the start of the next section, or address + 0xFFFF (aligned to 16 byte boundary)
         end = min(next_sec_start, align_addr(addr + 0xFFFF, 16))
 
-        sec = ElfMemorySection(bytearray(end - start), '.empty', self.global_instruction_context(), '', start, MemoryFlags(False, True))
+        sec = ElfMemorySection(
+            bytearray(end - start),
+            ".empty",
+            self.global_instruction_context(),
+            "",
+            start,
+            MemoryFlags(False, True),
+        )
         self.sections.append(sec)
         self._update_state()
 
@@ -40,4 +48,4 @@ class PrivMMU(MMU):
     def global_instruction_context(self) -> InstructionContext:
         context = InstructionContext()
         context.global_symbol_dict = self.global_symbols
-        return context
+        return context

riscemu/priv/PrivRV32I.py

@@ -16,155 +16,159 @@ if typing.TYPE_CHECKING:
 
 
 class PrivRV32I(RV32I):
-    cpu: 'PrivCPU'
+    cpu: "PrivCPU"
     """
     This is an extension of RV32I, written for the PrivCPU class
     """
 
-    def instruction_csrrw(self, ins: 'Instruction'):
+    def instruction_csrrw(self, ins: "Instruction"):
         rd, rs, csr_addr = self.parse_crs_ins(ins)
         old_val = None
-        if rd != 'zero':
+        if rd != "zero":
             self.cpu.csr.assert_can_read(self.cpu.mode, csr_addr)
             old_val = self.cpu.csr.get(csr_addr)
-        if rs != 'zero':
+        if rs != "zero":
             new_val = self.regs.get(rs)
             self.cpu.csr.assert_can_write(self.cpu.mode, csr_addr)
             self.cpu.csr.set(csr_addr, new_val)
         if old_val is not None:
             self.regs.set(rd, old_val)
 
-    def instruction_csrrs(self, ins: 'Instruction'):
+    def instruction_csrrs(self, ins: "Instruction"):
         rd, rs, csr_addr = self.parse_crs_ins(ins)
-        if rs != 'zero':
+        if rs != "zero":
             # oh no, this should not happen!
             INS_NOT_IMPLEMENTED(ins)
-        if rd != 'zero':
+        if rd != "zero":
             self.cpu.csr.assert_can_read(self.cpu.mode, csr_addr)
             old_val = self.cpu.csr.get(csr_addr)
             self.regs.set(rd, old_val)
 
-    def instruction_csrrc(self, ins: 'Instruction'):
+    def instruction_csrrc(self, ins: "Instruction"):
         INS_NOT_IMPLEMENTED(ins)
 
-    def instruction_csrrsi(self, ins: 'Instruction'):
+    def instruction_csrrsi(self, ins: "Instruction"):
         INS_NOT_IMPLEMENTED(ins)
 
-    def instruction_csrrwi(self, ins: 'Instruction'):
+    def instruction_csrrwi(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         rd, imm, addr = ins.get_reg(0), ins.get_imm(1), ins.get_imm(2)
-        if rd != 'zero':
+        if rd != "zero":
             self.cpu.csr.assert_can_read(self.cpu.mode, addr)
             old_val = self.cpu.csr.get(addr)
             self.regs.set(rd, old_val)
         self.cpu.csr.assert_can_write(self.cpu.mode, addr)
         self.cpu.csr.set(addr, imm)
 
-    def instruction_csrrci(self, ins: 'Instruction'):
+    def instruction_csrrci(self, ins: "Instruction"):
         INS_NOT_IMPLEMENTED(ins)
 
-    def instruction_mret(self, ins: 'Instruction'):
+    def instruction_mret(self, ins: "Instruction"):
         if self.cpu.mode != PrivModes.MACHINE:
             print("MRET not inside machine level code!")
             raise IllegalInstructionTrap(ins)
         # retore mie
-        mpie = self.cpu.csr.get_mstatus('mpie')
-        self.cpu.csr.set_mstatus('mie', mpie)
+        mpie = self.cpu.csr.get_mstatus("mpie")
+        self.cpu.csr.set_mstatus("mie", mpie)
         # restore priv
-        mpp = self.cpu.csr.get_mstatus('mpp')
+        mpp = self.cpu.csr.get_mstatus("mpp")
         self.cpu.mode = PrivModes(mpp)
         # restore pc
-        mepc = self.cpu.csr.get('mepc')
+        mepc = self.cpu.csr.get("mepc")
         self.cpu.pc = (mepc - self.cpu.INS_XLEN).value
 
         if self.cpu.conf.verbosity > 0:
             sec = self.mmu.get_sec_containing(mepc.value)
             if sec is not None:
-                print(FMT_CPU + "[CPU] returning to mode {} in {} (0x{:x})".format(
-                    PrivModes(mpp).name,
-                    self.mmu.translate_address(mepc),
-                    mepc
-                ) + FMT_NONE)
+                print(
+                    FMT_CPU
+                    + "[CPU] returning to mode {} in {} (0x{:x})".format(
+                        PrivModes(mpp).name, self.mmu.translate_address(mepc), mepc
+                    )
+                    + FMT_NONE
+                )
                 if self.cpu.conf.verbosity > 1:
                     self.regs.dump_reg_a()
 
-    def instruction_uret(self, ins: 'Instruction'):
+    def instruction_uret(self, ins: "Instruction"):
         raise IllegalInstructionTrap(ins)
 
-    def instruction_sret(self, ins: 'Instruction'):
+    def instruction_sret(self, ins: "Instruction"):
         raise IllegalInstructionTrap(ins)
 
-    def instruction_scall(self, ins: 'Instruction'):
+    def instruction_scall(self, ins: "Instruction"):
         """
         Overwrite the scall from userspace RV32I
         """
         raise EcallTrap(self.cpu.mode)
 
-    def instruction_beq(self, ins: 'Instruction'):
+    def instruction_beq(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins)
         if rs1 == rs2:
             self.pc += dst.value - 4
 
-    def instruction_bne(self, ins: 'Instruction'):
+    def instruction_bne(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins)
         if rs1 != rs2:
             self.pc += dst.value - 4
 
-    def instruction_blt(self, ins: 'Instruction'):
+    def instruction_blt(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins)
         if rs1 < rs2:
             self.pc += dst.value - 4
 
-    def instruction_bge(self, ins: 'Instruction'):
+    def instruction_bge(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins)
         if rs1 >= rs2:
             self.pc += dst.value - 4
 
-    def instruction_bltu(self, ins: 'Instruction'):
+    def instruction_bltu(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins, signed=False)
         if rs1 < rs2:
             self.pc += dst.value - 4
 
-    def instruction_bgeu(self, ins: 'Instruction'):
+    def instruction_bgeu(self, ins: "Instruction"):
         rs1, rs2, dst = self.parse_rs_rs_imm(ins, signed=False)
         if rs1 >= rs2:
             self.pc += dst.value - 4
 
     # technically deprecated
-    def instruction_j(self, ins: 'Instruction'):
+    def instruction_j(self, ins: "Instruction"):
         raise NotImplementedError("Should never be reached!")
 
-    def instruction_jal(self, ins: 'Instruction'):
+    def instruction_jal(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 2)
         reg = ins.get_reg(0)
         addr = ins.get_imm(1)
-        if reg == 'ra' and (
-                (self.cpu.mode == PrivModes.USER and self.cpu.conf.verbosity > 1) or
-                (self.cpu.conf.verbosity > 3)
+        if reg == "ra" and (
+            (self.cpu.mode == PrivModes.USER and self.cpu.conf.verbosity > 1)
+            or (self.cpu.conf.verbosity > 3)
         ):
-            print(FMT_CPU + 'Jumping from 0x{:x} to {} (0x{:x})'.format(
-                self.pc,
-                self.mmu.translate_address(self.pc + addr),
-                self.pc + addr
-            ) + FMT_NONE)
+            print(
+                FMT_CPU
+                + "Jumping from 0x{:x} to {} (0x{:x})".format(
+                    self.pc, self.mmu.translate_address(self.pc + addr), self.pc + addr
+                )
+                + FMT_NONE
+            )
             self.regs.dump_reg_a()
         self.regs.set(reg, Int32(self.pc))
         self.pc += addr - 4
 
-    def instruction_jalr(self, ins: 'Instruction'):
+    def instruction_jalr(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         rd, rs, imm = self.parse_rd_rs_imm(ins)
         self.regs.set(rd, Int32(self.pc))
         self.pc = rs.value + imm.value - 4
 
-    def instruction_sbreak(self, ins: 'Instruction'):
+    def instruction_sbreak(self, ins: "Instruction"):
         raise LaunchDebuggerException()
 
-    def parse_crs_ins(self, ins: 'Instruction'):
+    def parse_crs_ins(self, ins: "Instruction"):
         ASSERT_LEN(ins.args, 3)
         return ins.get_reg(0), ins.get_reg(1), ins.get_imm(2)
 
-    def parse_mem_ins(self, ins: 'Instruction') -> Tuple[str, int]:
+    def parse_mem_ins(self, ins: "Instruction") -> Tuple[str, int]:
         ASSERT_LEN(ins.args, 3)
         addr = self.get_reg_content(ins, 1) + ins.get_imm(2)
         reg = ins.get_reg(0)

riscemu/priv/__init__.py

@@ -12,4 +12,4 @@ Syscalls will have to be intercepted by your assembly code.
 
 
 The PrivCPU Implements the Risc-V M/U Model, meaning there is machine mode and user mode. No PMP or paging is available.
-"""
+"""

riscemu/priv/__main__.py

@@ -4,27 +4,55 @@ from .PrivCPU import PrivCPU
 
 import sys
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     import argparse
 
-    parser = argparse.ArgumentParser(description='RISC-V privileged architecture emulator', prog='riscemu')
+    parser = argparse.ArgumentParser(
+        description="RISC-V privileged architecture emulator", prog="riscemu"
+    )
 
-    parser.add_argument('source', type=str,
-                        help='Compiled RISC-V ELF file or memory image containing compiled RISC-V ELF files', nargs='+')
-    parser.add_argument('--debug-exceptions', help='Launch the interactive debugger when an exception is generated',
-                        action='store_true')
+    parser.add_argument(
+        "source",
+        type=str,
+        help="Compiled RISC-V ELF file or memory image containing compiled RISC-V ELF files",
+        nargs="+",
+    )
+    parser.add_argument(
+        "--debug-exceptions",
+        help="Launch the interactive debugger when an exception is generated",
+        action="store_true",
+    )
 
-    parser.add_argument('-v', '--verbose', help="Verbosity level (can be used multiple times)", action='count',
-                        default=0)
+    parser.add_argument(
+        "-v",
+        "--verbose",
+        help="Verbosity level (can be used multiple times)",
+        action="count",
+        default=0,
+    )
 
-    parser.add_argument('--slowdown', help="Slow down the emulated CPU clock by a factor", type=float, default=1)
+    parser.add_argument(
+        "--slowdown",
+        help="Slow down the emulated CPU clock by a factor",
+        type=float,
+        default=1,
+    )
 
     args = parser.parse_args()
 
-    cpu = PrivCPU(RunConfig(verbosity=args.verbose, debug_on_exception=args.debug_exceptions, slowdown=args.slowdown))
+    cpu = PrivCPU(
+        RunConfig(
+            verbosity=args.verbose,
+            debug_on_exception=args.debug_exceptions,
+            slowdown=args.slowdown,
+        )
+    )
 
     for source_path in args.source:
-        loader = max((loader for loader in cpu.get_loaders()), key=lambda l: l.can_parse(source_path))
+        loader = max(
+            (loader for loader in cpu.get_loaders()),
+            key=lambda l: l.can_parse(source_path),
+        )
         argv, opts = loader.get_options(sys.argv)
         program = loader.instantiate(source_path, opts).parse()
         if isinstance(program, Program):

riscemu/priv/types.py

@@ -6,9 +6,19 @@ from typing import Tuple, Dict, Set
 
 from riscemu.colors import FMT_NONE, FMT_PARSE
 from riscemu.decoder import format_ins, RISCV_REGS, decode
-from riscemu.priv.Exceptions import InstructionAccessFault, InstructionAddressMisalignedTrap, LoadAccessFault
-from riscemu.types import Instruction, InstructionContext, T_RelativeAddress, MemoryFlags, T_AbsoluteAddress, \
-    BinaryDataMemorySection
+from riscemu.priv.Exceptions import (
+    InstructionAccessFault,
+    InstructionAddressMisalignedTrap,
+    LoadAccessFault,
+)
+from riscemu.types import (
+    Instruction,
+    InstructionContext,
+    T_RelativeAddress,
+    MemoryFlags,
+    T_AbsoluteAddress,
+    BinaryDataMemorySection,
+)
 
 
 @dataclass(frozen=True)
@@ -27,34 +37,45 @@ class ElfInstruction(Instruction):
         return RISCV_REGS[self.args[num]]
 
     def __repr__(self) -> str:
-        if self.name == 'jal' and self.args[0] == 0:
+        if self.name == "jal" and self.args[0] == 0:
             return "j       {}".format(self.args[1])
-        if self.name == 'addi' and self.args[2] == 0:
+        if self.name == "addi" and self.args[2] == 0:
             return "mv      {}, {}".format(self.get_reg(0), self.get_reg(1))
-        if self.name == 'addi' and self.args[1] == 0:
+        if self.name == "addi" and self.args[1] == 0:
             return "li      {}, {}".format(self.get_reg(0), self.args[2])
-        if self.name == 'ret' and len(self.args) == 0:
+        if self.name == "ret" and len(self.args) == 0:
             return "ret"
         return format_ins(self.encoded, self.name)
 
 
 class ElfMemorySection(BinaryDataMemorySection):
-    def __init__(self, data: bytearray, name: str, context: InstructionContext, owner: str, base: int,
-                 flags: MemoryFlags):
+    def __init__(
+        self,
+        data: bytearray,
+        name: str,
+        context: InstructionContext,
+        owner: str,
+        base: int,
+        flags: MemoryFlags,
+    ):
         super().__init__(data, name, context, owner, base=base, flags=flags)
         self.read_ins = lru_cache(maxsize=self.size // 4)(self.read_ins)
 
     def read_ins(self, offset):
         if not self.flags.executable:
-            print(FMT_PARSE + "Reading instruction from non-executable memory!" + FMT_NONE)
+            print(
+                FMT_PARSE + "Reading instruction from non-executable memory!" + FMT_NONE
+            )
             raise InstructionAccessFault(offset + self.base)
         if offset % 4 != 0:
             raise InstructionAddressMisalignedTrap(offset + self.base)
-        return ElfInstruction(*decode(self.data[offset:offset + 4]))
+        return ElfInstruction(*decode(self.data[offset : offset + 4]))
 
     def write(self, offset: T_RelativeAddress, size: int, data: bytearray):
         if self.flags.read_only:
-            raise LoadAccessFault('read-only section', offset + self.base, size, 'write')
+            raise LoadAccessFault(
+                "read-only section", offset + self.base, size, "write"
+            )
         self.read_ins.cache_clear()
         return super(ElfMemorySection, self).write(offset, size, data)
 
@@ -64,7 +85,7 @@ class ElfMemorySection(BinaryDataMemorySection):
 
 
 class MemoryImageDebugInfos:
-    VERSION = '1.0.0'
+    VERSION = "1.0.0"
     """
     Schema version
     """
@@ -89,12 +110,13 @@ class MemoryImageDebugInfos:
     This dictionary contains the list of all global symbols of a given program
     """
 
-    def __init__(self,
-                 sections: Dict[str, Dict[str, Tuple[int, int]]],
-                 symbols: Dict[str, Dict[str, int]],
-                 globals: Dict[str, Set[str]],
-                 base: int = 0
-                 ):
+    def __init__(
+        self,
+        sections: Dict[str, Dict[str, Tuple[int, int]]],
+        symbols: Dict[str, Dict[str, int]],
+        globals: Dict[str, Set[str]],
+        base: int = 0,
+    ):
         self.sections = sections
         self.symbols = symbols
         self.globals = globals
@@ -108,7 +130,9 @@ class MemoryImageDebugInfos:
                 return json.dumps(dict(obj), default=serialize)
             if isinstance(obj, (set, tuple)):
                 return json.dumps(list(obj), default=serialize)
-            return "<<unserializable {}>>".format(getattr(obj, '__qualname__', '{unknown}'))
+            return "<<unserializable {}>>".format(
+                getattr(obj, "__qualname__", "{unknown}")
+            )
 
         return json.dumps(
             dict(
@@ -116,22 +140,25 @@ class MemoryImageDebugInfos:
                 symbols=self.symbols,
                 globals=self.globals,
                 base=self.base,
-                VERSION=self.VERSION
+                VERSION=self.VERSION,
             ),
-            default=serialize
+            default=serialize,
         )
 
     @classmethod
-    def load(cls, serialized_str: str) -> 'MemoryImageDebugInfos':
+    def load(cls, serialized_str: str) -> "MemoryImageDebugInfos":
         json_obj: dict = json.loads(serialized_str)
 
-        if 'VERSION' not in json_obj:
+        if "VERSION" not in json_obj:
             raise RuntimeError("Unknown MemoryImageDebugInfo version!")
 
-        version: str = json_obj.pop('VERSION')
+        version: str = json_obj.pop("VERSION")
 
         # compare major version
-        if version != cls.VERSION and version.split('.')[0] != cls.VERSION.split('.')[0]:
+        if (
+            version != cls.VERSION
+            and version.split(".")[0] != cls.VERSION.split(".")[0]
+        ):
             raise RuntimeError(
                 "Unknown MemoryImageDebugInfo version! This emulator expects version {}, debug info version {}".format(
                     cls.VERSION, version
@@ -141,7 +168,7 @@ class MemoryImageDebugInfos:
         return MemoryImageDebugInfos(**json_obj)
 
     @classmethod
-    def builder(cls) -> 'MemoryImageDebugInfos':
+    def builder(cls) -> "MemoryImageDebugInfos":
         return MemoryImageDebugInfos(
             defaultdict(dict), defaultdict(dict), defaultdict(set)
         )

riscemu/registers.py

@@ -1,15 +1,15 @@
 """
-RiscEmu (c) 2021-2022 Anton Lydike
+RiscEmu (c) 2023 Anton Lydike
 
 SPDX-License-Identifier: MIT
 """
 
 from collections import defaultdict
+from typing import Union
 
 from .helpers import *
 
-if typing.TYPE_CHECKING:
-    from .types import Int32
+from .types import Int32, Float32
 
 
 class Registers:
@@ -17,43 +17,115 @@ class Registers:
     Represents a bunch of registers
     """
 
-    def __init__(self):
-        from .types import Int32
-        self.vals = defaultdict(lambda: Int32(0))
+    valid_regs = {
+        "zero",
+        "ra",
+        "sp",
+        "gp",
+        "tp",
+        "s0",
+        "fp",
+        "t0",
+        "t1",
+        "t2",
+        "t3",
+        "t4",
+        "t5",
+        "t6",
+        "s1",
+        "s2",
+        "s3",
+        "s4",
+        "s5",
+        "s6",
+        "s7",
+        "s8",
+        "s9",
+        "s10",
+        "s11",
+        "a0",
+        "a1",
+        "a2",
+        "a3",
+        "a4",
+        "a5",
+        "a6",
+        "a7",
+    }
+
+    float_regs = {
+        "ft0",
+        "ft1",
+        "ft2",
+        "ft3",
+        "ft4",
+        "ft5",
+        "ft6",
+        "ft7",
+        "fs0",
+        "fs1",
+        "fs2",
+        "fs3",
+        "fs4",
+        "fs5",
+        "fs6",
+        "fs7",
+        "fs8",
+        "fs9",
+        "fs10",
+        "fs11",
+        "fa0",
+        "fa1",
+        "fa2",
+        "fa3",
+        "fa4",
+        "fa5",
+        "fa6",
+        "fa7",
+    }
+
+    def __init__(self, infinite_regs: bool = False):
+        self.vals: dict[str, Int32] = defaultdict(UInt32)
+        self.float_vals: dict[str, Float32] = defaultdict(Float32)
+
         self.last_set = None
         self.last_read = None
 
-    def dump(self, full=False):
+        self.infinite_regs = infinite_regs
+
+    def dump(self, full: bool = False):
         """
         Dump all registers to stdout
         :param full: If True, floating point registers are dumped too
         """
         named_regs = [self._reg_repr(reg) for reg in Registers.named_registers()]
 
-        lines = [[] for i in range(12)]
+        lines: list[list[str]] = [[] for _ in range(12)]
         if not full:
-            regs = [('a', 8), ('s', 12), ('t', 7)]
+            regs = [("a", 8), ("s", 12), ("t", 7)]
         else:
             regs = [
-                ('a', 8),
-                ('s', 12),
-                ('t', 7),
-                ('ft', 8),
-                ('fa', 8),
-                ('fs', 12),
+                ("a", 8),
+                ("s", 12),
+                ("t", 7),
+                ("ft", 8),
+                ("fa", 8),
+                ("fs", 12),
             ]
         for name, s in regs:
             for i in range(12):
                 if i >= s:
                     lines[i].append(" " * 15)
                 else:
-                    reg = '{}{}'.format(name, i)
+                    reg = "{}{}".format(name, i)
                     lines[i].append(self._reg_repr(reg))
 
-        print("Registers[{},{}](".format(
-            FMT_ORANGE + FMT_UNDERLINE + 'read' + FMT_NONE,
-            FMT_ORANGE + FMT_BOLD + 'written' + FMT_NONE
-        ))
+        print(
+            "Registers[{},{}](".format(
+                FMT_ORANGE + FMT_UNDERLINE + "read" + FMT_NONE,
+                FMT_ORANGE + FMT_BOLD + "written" + FMT_NONE,
+            )
+        )
         if not full:
             print("\t" + " ".join(named_regs[0:3]))
             print("\t" + " ".join(named_regs[3:]))
@@ -69,23 +141,35 @@ class Registers:
         """
         Dump the a registers
         """
-        print("Registers[a]:" + " ".join(self._reg_repr('a{}'.format(i)) for i in range(8)))
-
-    def _reg_repr(self, reg):
-        txt = '{:4}=0x{:08X}'.format(reg, self.get(reg, False))
-        if reg == 'fp':
-            reg = 's0'
+        print(
+            "Registers[a]:"
+            + " ".join(self._reg_repr("a{}".format(i)) for i in range(8))
+        )
+
+    def _reg_repr(self, reg: str, name_len=4, fmt="08X"):
+        if reg in self.float_regs:
+            txt = "{:{}}={: .3e}".format(reg, name_len, self.get_f(reg, False))
+        else:
+            txt = "{:{}}=0x{:{}}".format(reg, name_len, self.get(reg, False), fmt)
+        if reg == "fp":
+            reg = "s0"
         if reg == self.last_set:
             return FMT_ORANGE + FMT_BOLD + txt + FMT_NONE
         if reg == self.last_read:
             return FMT_ORANGE + FMT_UNDERLINE + txt + FMT_NONE
-        if reg == 'zero':
+        if reg == "zero":
             return txt
-        if self.get(reg, False) == 0 and reg not in Registers.named_registers():
+        should_grayscale_int = (
+            reg in self.valid_regs
+            and self.get(reg, False) == 0
+            and reg not in Registers.named_registers()
+        )
+        should_grayscale_float = reg in self.float_regs and self.get_f(reg, False) == 0
+        if should_grayscale_int or should_grayscale_float:
             return FMT_GRAY + txt + FMT_NONE
         return txt
 
-    def set(self, reg, val: 'Int32', mark_set=True) -> bool:
+    def set(self, reg: str, val: "Int32", mark_set: bool = True) -> bool:
         """
         Set a register content to val
         :param reg: The register to set
@@ -94,25 +178,29 @@ class Registers:
         :return: If the operation was successful
         """
 
-        from .types import Int32
         # remove after refactoring is complete
         if not isinstance(val, Int32):
-            raise RuntimeError("Setting register to non-Int32 value! Please refactor your code!")
+            raise RuntimeError(
+                "Setting register to non-Int32 value! Please refactor your code!"
+            )
 
-        if reg == 'zero':
+        if reg == "zero":
             return False
         # if reg not in Registers.all_registers():
         #    raise InvalidRegisterException(reg)
         # replace fp register with s1, as these are the same register
-        if reg == 'fp':
-            reg = 's1'
+        if reg == "fp":
+            reg = "s1"
         if mark_set:
             self.last_set = reg
-        # check 32 bit signed bounds
+
+        if not self.infinite_regs and reg not in self.valid_regs:
+            raise RuntimeError("Invalid register: {}".format(reg))
+
         self.vals[reg] = val.unsigned()
         return True
 
-    def get(self, reg, mark_read=True) -> 'Int32':
+    def get(self, reg: str, mark_read: bool = True) -> "Int32":
         """
         Retuns the contents of register reg
         :param reg: The register name
@@ -121,25 +209,27 @@ class Registers:
         """
         # if reg not in Registers.all_registers():
         #    raise InvalidRegisterException(reg)
-        if reg == 'fp':
-            reg = 's0'
+        if reg == "fp":
+            reg = "s0"
+
+        if not self.infinite_regs and reg not in self.valid_regs:
+            raise RuntimeError("Invalid register: {}".format(reg))
+
         if mark_read:
             self.last_read = reg
         return self.vals[reg]
 
-    @staticmethod
-    def all_registers():
-        """
-        Return a list of all valid registers
-        :return: The list
-        """
-        return ['zero', 'ra', 'sp', 'gp', 'tp', 's0', 'fp',
-                't0', 't1', 't2', 't3', 't4', 't5', 't6',
-                's1', 's2', 's3', 's4', 's5', 's6', 's7', 's8', 's9', 's10', 's11',
-                'a0', 'a1', 'a2', 'a3', 'a4', 'a5', 'a6', 'a7',
-                'ft0', 'ft1', 'ft2', 'ft3', 'ft4', 'ft5', 'ft6', 'ft7',
-                'fs0', 'fs1', 'fs2', 'fs3', 'fs4', 'fs5', 'fs6', 'fs7', 'fs8', 'fs9', 'fs10', 'fs11',
-                'fa0', 'fa1', 'fa2', 'fa3', 'fa4', 'fa5', 'fa6', 'fa7']
+    def get_f(self, reg: str, mark_read: bool = True) -> "Float32":
+        if not self.infinite_regs and reg not in self.float_regs:
+            raise RuntimeError("Invalid float register: {}".format(reg))
+        if mark_read:
+            self.last_read = reg
+        return self.float_vals[reg]
+
+    def set_f(self, reg: str, val: Union[float, "Float32"]):
+        if not self.infinite_regs and reg not in self.float_regs:
+            raise RuntimeError("Invalid float register: {}".format(reg))
+        self.float_vals[reg] = Float32(val)
 
     @staticmethod
     def named_registers():
@@ -147,4 +237,12 @@ class Registers:
         Return all named registers
         :return: The list
         """
-        return ['zero', 'ra', 'sp', 'gp', 'tp', 'fp']
+        return ["zero", "ra", "sp", "gp", "tp", "fp"]
+
+    def __repr__(self):
+        return "<Registers[{}]{}>".format(
+            "float" if self.supports_floats else "nofloat",
+            "{"
+            + ", ".join(self._reg_repr("a{}".format(i), 2, "0x") for i in range(8))
+            + "}",
+        )

riscemu/riscemu_main.py

@@ -0,0 +1,284 @@
+import argparse
+import glob
+import os
+import sys
+from typing import Type, Dict, List, Optional
+
+from riscemu import AssemblyFileLoader, __version__, __copyright__
+from riscemu.types import CPU, ProgramLoader, Program
+from riscemu.instructions import InstructionSet, InstructionSetDict
+from riscemu.config import RunConfig
+from riscemu.CPU import UserModeCPU
+
+
+class RiscemuMain:
+    """
+    This represents the riscemu API exposed to other programs for better
+    interoperability.
+    """
+
+    available_ins_sets: Dict[str, Type[InstructionSet]]
+    available_file_loaders: List[Type[ProgramLoader]]
+
+    cfg: Optional[RunConfig]
+    cpu: Optional[CPU]
+
+    input_files: List[str]
+    selected_ins_sets: List[Type[InstructionSet]]
+
+    def __init__(self, cfg: Optional[RunConfig] = None):
+        self.available_ins_sets = dict()
+        self.selected_ins_sets = []
+        self.available_file_loaders = []
+        self.cfg: Optional[RunConfig] = cfg
+        self.cpu: Optional[CPU] = None
+        self.input_files = []
+        self.selected_ins_sets = []
+
+    def instantiate_cpu(self):
+        self.cpu = UserModeCPU(self.selected_ins_sets, self.cfg)
+        self.configure_cpu()
+
+    def configure_cpu(self):
+        assert self.cfg is not None
+        if isinstance(self.cpu, UserModeCPU) and self.cfg.stack_size != 0:
+            self.cpu.setup_stack(self.cfg.stack_size)
+
+    def register_all_arguments(self, parser: argparse.ArgumentParser):
+        parser.add_argument(
+            "files",
+            metavar="file.asm",
+            type=str,
+            nargs="+",
+            help="The assembly files to load, the last one will be run",
+        )
+
+        parser.add_argument(
+            "--options",
+            "-o",
+            action=OptionStringAction,
+            keys=(
+                "disable_debug",
+                "no_syscall_symbols",
+                "fail_on_ex",
+                "add_accept_imm",
+                "unlimited_regs",
+                "libc",
+                "ignore_exit_code",
+            ),
+            help="""Toggle options. Available options are:
+        disable_debug:        Disable ebreak instructions
+        no_syscall_symbols:   Don't add symbols for SCALL_EXIT and others
+        fail_on_ex:           If set, exceptions won't trigger the debugger
+        add_accept_imm:       Accept "add rd, rs, imm" instruction (instead of addi)
+        unlimited_regs:       Allow an unlimited number of registers
+        libc:                 Load a libc-like runtime (for malloc, etc.)
+        ignore_exit_code:     Don't exit with the programs exit code.""",
+        )
+
+        parser.add_argument(
+            "--syscall-opts",
+            "-so",
+            action=OptionStringAction,
+            keys=("fs_access", "disable_input"),
+        )
+
+        parser.add_argument(
+            "--instruction-sets",
+            "-is",
+            action=OptionStringAction,
+            help="Instruction sets to load, available are: {}. All are enabled by default".format(
+                ", ".join(self.available_ins_sets)
+            ),
+            keys={k: True for k in self.available_ins_sets},
+            omit_empty=True,
+        )
+
+        parser.add_argument(
+            "--stack_size",
+            type=int,
+            help="Stack size of loaded programs, defaults to 8MB",
+            nargs="?",
+        )
+
+        parser.add_argument(
+            "-v",
+            "--verbose",
+            help="Verbosity level (can be used multiple times)",
+            action="count",
+            default=0,
+        )
+
+        parser.add_argument(
+            "--interactive",
+            help="Launch the interactive debugger instantly instead of loading any "
+            "programs",
+            action="store_true",
+        )
+
+        parser.add_argument(
+            "--ignore-exit-code",
+            help="Ignore exit code of the program and always return 0 if the program ran to completion.",
+            action="store_true",
+            default=False,
+        )
+
+    def register_all_isas(self):
+        self.available_ins_sets.update(InstructionSetDict)
+
+    def register_all_program_loaders(self):
+        self.available_file_loaders.append(AssemblyFileLoader)
+
+    def parse_argv(self, argv: List[str]):
+        parser = argparse.ArgumentParser(
+            description="RISC-V Userspace emulator",
+            prog="riscemu",
+            formatter_class=argparse.RawTextHelpFormatter,
+        )
+        if "--version" in argv:
+            print(
+                "riscemu version {}\n{}\n\nAvailable ISA: {}".format(
+                    __version__, __copyright__, ", ".join(self.available_ins_sets)
+                )
+            )
+            sys.exit()
+
+        self.register_all_arguments(parser)
+
+        # parse argv
+        args = parser.parse_args(argv)
+
+        # add ins
+        if not hasattr(args, "ins"):
+            setattr(args, "ins", {k: True for k in self.available_ins_sets})
+
+        # create RunConfig
+        self.cfg = self.create_config(args)
+
+        # set input files
+        self.input_files = args.files
+
+        # get selected ins sets
+        self.selected_ins_sets = list(
+            self.available_ins_sets[name]
+            for name, selected in args.ins.items()
+            if selected
+        )
+
+        # if use_libc is given, attach libc to path
+        if self.cfg.use_libc:
+            self.add_libc_to_input_files()
+
+    def add_libc_to_input_files(self):
+        """
+        This adds the provided riscemu libc to the programs runtime.
+        """
+        libc_path = os.path.join(
+            os.path.dirname(__file__),
+            "..",
+            "libc",
+            "*.s",
+        )
+        for path in glob.iglob(libc_path):
+            if path not in self.input_files:
+                self.input_files.append(path)
+
+    def create_config(self, args: argparse.Namespace) -> RunConfig:
+        # create a RunConfig from the cli args
+        cfg_dict = dict(
+            stack_size=args.stack_size,
+            debug_instruction=not args.options["disable_debug"],
+            include_scall_symbols=not args.options["no_syscall_symbols"],
+            debug_on_exception=not args.options["fail_on_ex"],
+            add_accept_imm=args.options["add_accept_imm"],
+            unlimited_registers=args.options["unlimited_regs"],
+            scall_fs=args.syscall_opts["fs_access"],
+            scall_input=not args.syscall_opts["disable_input"],
+            verbosity=args.verbose,
+            use_libc=args.options["libc"],
+            ignore_exit_code=args.options["ignore_exit_code"],
+        )
+        for k, v in dict(cfg_dict).items():
+            if v is None:
+                del cfg_dict[k]
+
+        return RunConfig(**cfg_dict)
+
+    def load_programs(self):
+        for path in self.input_files:
+            for loader in self.available_file_loaders:
+                if not loader.can_parse(path):
+                    continue
+                programs = loader.instantiate(path, {}).parse()
+                if isinstance(programs, Program):
+                    programs = [programs]
+                for p in programs:
+                    self.cpu.mmu.load_program(p)
+
+    def run_from_cli(self, argv: List[str]):
+        # register everything
+        self.register_all_isas()
+        self.register_all_program_loaders()
+
+        # parse argv and set up cpu
+        self.parse_argv(argv)
+        self.instantiate_cpu()
+        self.load_programs()
+
+        # run the program
+        self.cpu.launch(self.cfg.verbosity > 1)
+
+    def run(self):
+        """
+        This assumes that these values were set externally:
+
+         - available_file_loaders: A list of available file loaders.
+           Can be set using .register_all_program_loaders()
+         - cfg: The RunConfig object. Can be directly assigned to the attribute
+         - input_files: A list of assembly files to load.
+         - selected_ins_sets: A list of instruction sets the CPU should support.
+        """
+        assert self.cfg is not None, "self.cfg must be set before calling run()"
+        assert self.selected_ins_sets, "self.selected_ins_sets cannot be empty"
+        assert self.input_files, "self.input_files cannot be empty"
+
+        if self.cfg.use_libc:
+            self.add_libc_to_input_files()
+
+        self.instantiate_cpu()
+        self.load_programs()
+
+        # run the program
+        self.cpu.launch(self.cfg.verbosity > 1)
+
+
+class OptionStringAction(argparse.Action):
+    def __init__(self, option_strings, dest, keys=None, omit_empty=False, **kwargs):
+        if keys is None:
+            raise ValueError('must define "keys" argument')
+        if isinstance(keys, dict):
+            keys_d = keys
+        elif isinstance(keys, (list, tuple)):
+            keys_d = {}
+            for k in keys:
+                if isinstance(k, tuple):
+                    k, v = k
+                else:
+                    v = False
+                keys_d[k] = v
+        else:
+            keys_d = dict()
+        super().__init__(option_strings, dest, default=keys_d, **kwargs)
+        self.keys = keys_d
+        self.omit_empty = omit_empty
+
+    def __call__(self, parser, namespace, values, option_string=None):
+        d = {}
+        if not self.omit_empty:
+            d.update(self.keys)
+        for x in values.split(","):
+            if x in self.keys:
+                d[x] = True
+            else:
+                raise ValueError("Invalid parameter supplied: " + x)
+        setattr(namespace, self.dest, d)

riscemu/syscall.py

@@ -6,33 +6,48 @@ SPDX-License-Identifier: MIT
 
 import sys
 from dataclasses import dataclass
-from typing import Dict, IO
+from math import log2, ceil
+from typing import Dict, IO, Union
 
+from .types import BinaryDataMemorySection, MemoryFlags
 from .colors import FMT_SYSCALL, FMT_NONE
 from .types import Int32, CPU
 from .types.exceptions import InvalidSyscallException
 
 SYSCALLS = {
-    63: 'read',
-    64: 'write',
-    93: 'exit',
-    1024: 'open',
-    1025: 'close',
+    63: "read",
+    64: "write",
+    93: "exit",
+    192: "mmap2",
+    1024: "open",
+    1025: "close",
 }
 """
 This dict contains a mapping for all available syscalls (code->name)
 
-If you wish to add a syscall to the built-in system, you can extend this 
+If you wish to add a syscall to the built-in system, you can extend this
 dictionary and implement a method with the same name on the SyscallInterface
 class.
 """
 
+ADDITIONAL_SYMBOLS = {
+    "MAP_PRIVATE": 1 << 0,
+    "MAP_SHARED": 1 << 1,
+    "MAP_ANON": 1 << 2,
+    "MAP_ANONYMOUS": 1 << 2,
+    "PROT_READ": 1 << 0,
+    "PROT_WRITE": 1 << 1,
+}
+"""
+A set of additional symbols that are used by various syscalls.
+"""
+
 OPEN_MODES = {
-    0: 'rb',
-    1: 'wb',
-    2: 'r+b',
-    3: 'x',
-    4: 'ab',
+    0: "rb",
+    1: "wb",
+    2: "r+b",
+    3: "x",
+    4: "ab",
 }
 """All available file open modes"""
 
@@ -42,6 +57,7 @@ class Syscall:
     """
     Represents a syscall
     """
+
     id: int
     """The syscall number (e.g. 64 - write)"""
     cpu: CPU
@@ -52,12 +68,10 @@ class Syscall:
         return SYSCALLS.get(self.id, "unknown")
 
     def __repr__(self):
-        return "Syscall(id={}, name={})".format(
-            self.id, self.name
-        )
+        return "Syscall(id={}, name={})".format(self.id, self.name)
 
-    def ret(self, code):
-        self.cpu.regs.set('a0', Int32(code))
+    def ret(self, code: Union[int, Int32]):
+        self.cpu.regs.set("a0", Int32(code))
 
 
 def get_syscall_symbols():
@@ -66,25 +80,26 @@ def get_syscall_symbols():
 
     :return: dictionary of all syscall symbols (SCALL_<name> -> id)
     """
-    return {
-        ('SCALL_' + name.upper()): num for num, name in SYSCALLS.items()
+    items: Dict[str, int] = {
+        ("SCALL_" + name.upper()): num for num, name in SYSCALLS.items()
     }
 
+    items.update(ADDITIONAL_SYMBOLS)
+
+    return items
+
 
 class SyscallInterface:
     """
     Handles syscalls
     """
+
     open_files: Dict[int, IO]
     next_open_handle: int
 
     def handle_syscall(self, scall: Syscall):
         self.next_open_handle = 3
-        self.open_files = {
-            0: sys.stdin,
-            1: sys.stdout,
-            2: sys.stderr
-        }
+        self.open_files = {0: sys.stdin, 1: sys.stdout, 2: sys.stderr}
 
         if getattr(self, scall.name):
             getattr(self, scall.name)(scall)
@@ -96,21 +111,25 @@ class SyscallInterface:
         read syscall (63): read from file no a0, into addr a1, at most a2 bytes
         on return a0 will be the number of read bytes or -1 if an error occured
         """
-        fileno = scall.cpu.regs.get('a0').unsigned_value
-        addr = scall.cpu.regs.get('a1').unsigned_value
-        size = scall.cpu.regs.get('a2').unsigned_value
+        fileno = scall.cpu.regs.get("a0").unsigned_value
+        addr = scall.cpu.regs.get("a1").unsigned_value
+        size = scall.cpu.regs.get("a2").unsigned_value
         if fileno not in self.open_files:
             scall.ret(-1)
             return
 
         chars = self.open_files[fileno].readline(size)
         try:
-            data = bytearray(chars, 'ascii')
+            data = bytearray(chars, "ascii")
             scall.cpu.mmu.write(addr, len(data), data)
             return scall.ret(len(data))
 
         except UnicodeEncodeError:
-            print(FMT_SYSCALL + '[Syscall] read: UnicodeError - invalid input "{}"'.format(chars) + FMT_NONE)
+            print(
+                FMT_SYSCALL
+                + '[Syscall] read: UnicodeError - invalid input "{}"'.format(chars)
+                + FMT_NONE
+            )
             return scall.ret(-1)
 
     def write(self, scall: Syscall):
@@ -118,19 +137,23 @@ class SyscallInterface:
         write syscall (64): write a2 bytes from addr a1 into fileno a0
         on return a0 will hold the number of bytes written or -1 if an error occured
         """
-        fileno = scall.cpu.regs.get('a0').unsigned_value
-        addr = scall.cpu.regs.get('a1').unsigned_value
-        size = scall.cpu.regs.get('a2').unsigned_value
+        fileno = scall.cpu.regs.get("a0").unsigned_value
+        addr = scall.cpu.regs.get("a1").unsigned_value
+        size = scall.cpu.regs.get("a2").unsigned_value
         if fileno not in self.open_files:
             return scall.ret(-1)
 
         data = scall.cpu.mmu.read(addr, size)
 
         if not isinstance(data, bytearray):
-            print(FMT_SYSCALL + '[Syscall] write: writing from .text region not supported.' + FMT_NONE)
+            print(
+                FMT_SYSCALL
+                + "[Syscall] write: writing from .text region not supported."
+                + FMT_NONE
+            )
             return scall.ret(-1)
 
-        self.open_files[fileno].write(data.decode('ascii'))
+        self.open_files[fileno].write(data.decode("ascii"))
         return scall.ret(size)
 
     def open(self, scall: Syscall):
@@ -149,19 +172,29 @@ class SyscallInterface:
         """
         # FIXME: this should be toggleable in a global setting or something
         if True:
-            print(FMT_SYSCALL + '[Syscall] open: opening files not supported without scall-fs flag!' + FMT_NONE)
+            print(
+                FMT_SYSCALL
+                + "[Syscall] open: opening files not supported without scall-fs flag!"
+                + FMT_NONE
+            )
             return scall.ret(-1)
 
-        mode = scall.cpu.regs.get('a0').unsigned_value
-        addr = scall.cpu.regs.get('a1').unsigned_value
-        size = scall.cpu.regs.get('a2').unsigned_value
+        mode = scall.cpu.regs.get("a0").unsigned_value
+        addr = scall.cpu.regs.get("a1").unsigned_value
+        size = scall.cpu.regs.get("a2").unsigned_value
 
-        mode_st = OPEN_MODES.get(mode, )
+        mode_st = OPEN_MODES.get(
+            mode,
+        )
         if mode_st == -1:
-            print(FMT_SYSCALL + '[Syscall] open: unknown opening mode {}!'.format(mode) + FMT_NONE)
+            print(
+                FMT_SYSCALL
+                + "[Syscall] open: unknown opening mode {}!".format(mode)
+                + FMT_NONE
+            )
             return scall.ret(-1)
 
-        path = scall.cpu.mmu.read(addr, size).decode('ascii')
+        path = scall.cpu.mmu.read(addr, size).decode("ascii")
 
         fileno = self.next_open_handle
         self.next_open_handle += 1
@@ -169,10 +202,18 @@ class SyscallInterface:
         try:
             self.open_files[fileno] = open(path, mode_st)
         except OSError as err:
-            print(FMT_SYSCALL + '[Syscall] open: encountered error during {}!'.format(err.strerror) + FMT_NONE)
+            print(
+                FMT_SYSCALL
+                + "[Syscall] open: encountered error during {}!".format(err.strerror)
+                + FMT_NONE
+            )
             return scall.ret(-1)
 
-        print(FMT_SYSCALL + '[Syscall] open: opened fd {} to {}!'.format(fileno, path) + FMT_NONE)
+        print(
+            FMT_SYSCALL
+            + "[Syscall] open: opened fd {} to {}!".format(fileno, path)
+            + FMT_NONE
+        )
         return scall.ret(fileno)
 
     def close(self, scall: Syscall):
@@ -181,13 +222,17 @@ class SyscallInterface:
 
         return -1 if an error was encountered, otherwise returns 0
         """
-        fileno = scall.cpu.regs.get('a0').unsigned_value
+        fileno = scall.cpu.regs.get("a0").unsigned_value
         if fileno not in self.open_files:
-            print(FMT_SYSCALL + '[Syscall] close: unknown fileno {}!'.format(fileno) + FMT_NONE)
+            print(
+                FMT_SYSCALL
+                + "[Syscall] close: unknown fileno {}!".format(fileno)
+                + FMT_NONE
+            )
             return scall.ret(-1)
 
         self.open_files[fileno].close()
-        print(FMT_SYSCALL + '[Syscall] close: closed fd {}!'.format(fileno) + FMT_NONE)
+        print(FMT_SYSCALL + "[Syscall] close: closed fd {}!".format(fileno) + FMT_NONE)
         del self.open_files[fileno]
         return scall.ret(0)
 
@@ -196,10 +241,55 @@ class SyscallInterface:
         Exit syscall. Exits the system with status code a0
         """
         scall.cpu.halted = True
-        scall.cpu.exit_code = scall.cpu.regs.get('a0').value
+        scall.cpu.exit_code = scall.cpu.regs.get("a0").signed().value
 
-    def __repr__(self):
-        return "{}(\n\tfiles={}\n)".format(
-            self.__class__.__name__,
-            self.open_files
+    def mmap2(self, scall: Syscall):
+        """
+        mmap2 syscall:
+
+        void *mmap(void *addr, size_t length, int prot, int flags,
+                  int fd, off_t offset);
+
+        Only supported modes:
+        addr  = <any>
+        prot  = either PROT_READ or PROT_READ | PROT_WRITE
+        flags = MAP_PRIVATE | MAP_ANONYMOUS
+        fd    = <ignored>
+        off_t = <ignored>
+        """
+        addr = scall.cpu.regs.get("a0").unsigned_value
+        size = scall.cpu.regs.get("a1").unsigned_value
+        prot = scall.cpu.regs.get("a2").unsigned_value
+        flags = scall.cpu.regs.get("a3").unsigned_value
+
+        # error out if prot is not 1 or 3:
+        # 1 = PROT_READ
+        # 3 = PROT_READ | PROT_WRITE
+        if prot != 1 and prot != 3:
+            return scall.ret(-1)
+
+        # round size up to multiple of 4096
+        size = 4096 * ceil(size / 4096)
+        section = BinaryDataMemorySection(
+            bytearray(size),
+            ".data.runtime-allocated",
+            None,
+            "system",
+            base=addr,
+            flags=MemoryFlags(read_only=prot != 3, executable=False),
         )
+
+        # try to insert section
+        if scall.cpu.mmu.load_section(section, addr != 0):
+            return scall.ret(section.base)
+        # if that failed, and we tried to force an address,
+        # try again at any address
+        elif addr != 0:
+            section.base = 0
+            if scall.cpu.mmu.load_section(section):
+                return scall.ret(section.base)
+        # if that didn't work, return error
+        return scall.ret(-1)
+
+    def __repr__(self):
+        return "{}(\n\tfiles={}\n)".format(self.__class__.__name__, self.open_files)

riscemu/tokenizer.py

@@ -12,9 +12,11 @@ from typing import List, Iterable
 from riscemu.decoder import RISCV_REGS
 from riscemu.types.exceptions import ParseException
 
-LINE_COMMENT_STARTERS = ('#', ';', '//')
-WHITESPACE_PATTERN = re.compile(r'\s+')
-MEMORY_ADDRESS_PATTERN = re.compile(r'^(0[xX][A-f0-9]+|\d+|0b[0-1]+|[A-z0-9_-]+)\(([A-z]+[0-9]{0,2})\)$')
+LINE_COMMENT_STARTERS = ("#", ";", "//")
+WHITESPACE_PATTERN = re.compile(r"\s+")
+MEMORY_ADDRESS_PATTERN = re.compile(
+    r"^(0[xX][A-f0-9]+|\d+|0b[0-1]+|[A-z0-9_-]+)\(([A-z]+[0-9]*)\)$"
+)
 REGISTER_NAMES = RISCV_REGS
 
 
@@ -34,22 +36,22 @@ class Token:
 
     def __str__(self):
         if self.type == TokenType.NEWLINE:
-            return '\\n'
+            return "\\n"
         if self.type == TokenType.COMMA:
-            return ', '
-        return '{}({})'.format(self.type.name[0:3], self.value)
+            return ", "
+        return "{}({})".format(self.type.name[0:3], self.value)
 
 
-NEWLINE = Token(TokenType.NEWLINE, '\n')
-COMMA = Token(TokenType.COMMA, ',')
+NEWLINE = Token(TokenType.NEWLINE, "\n")
+COMMA = Token(TokenType.COMMA, ",")
 
 
 def tokenize(input: Iterable[str]) -> Iterable[Token]:
     for line in input:
         for line_comment_start in LINE_COMMENT_STARTERS:
             if line_comment_start in line:
-                line = line[:line.index(line_comment_start)]
-        line.strip(' \t\n')
+                line = line[: line.index(line_comment_start)]
+        line.strip(" \t\n")
         if not line:
             continue
 
@@ -64,9 +66,9 @@ def parse_line(parts: List[str]) -> Iterable[Token]:
         return ()
     first_token = parts[0]
 
-    if first_token[0] == '.':
+    if first_token[0] == ".":
         yield Token(TokenType.PSEUDO_OP, first_token)
-    elif first_token[-1] == ':':
+    elif first_token[-1] == ":":
         yield Token(TokenType.LABEL, first_token)
         yield from parse_line(parts[1:])
         return
@@ -74,22 +76,21 @@ def parse_line(parts: List[str]) -> Iterable[Token]:
         yield Token(TokenType.INSTRUCTION_NAME, first_token)
 
     for part in parts[1:]:
-        if part == ',':
+        if part == ",":
             yield COMMA
             continue
         yield from parse_arg(part)
 
 
 def parse_arg(arg: str) -> Iterable[Token]:
-    comma = arg[-1] == ','
+    comma = arg[-1] == ","
     arg = arg[:-1] if comma else arg
     mem_match_resul = re.match(MEMORY_ADDRESS_PATTERN, arg)
     if mem_match_resul:
         register = mem_match_resul.group(2).lower()
-        if register not in RISCV_REGS:
-            raise ParseException(f'"{register}" is not a valid register!')
+        immediate = mem_match_resul.group(1)
         yield Token(TokenType.ARGUMENT, register)
-        yield Token(TokenType.ARGUMENT, mem_match_resul.group(1))
+        yield Token(TokenType.ARGUMENT, immediate)
     else:
         yield Token(TokenType.ARGUMENT, arg)
     if comma:
@@ -98,7 +99,7 @@ def parse_arg(arg: str) -> Iterable[Token]:
 
 def print_tokens(tokens: Iterable[Token]):
     for token in tokens:
-        print(token, end='\n' if token == NEWLINE else '')
+        print(token, end="\n" if token == NEWLINE else "")
     print("", flush=True, end="")
 
 
@@ -123,7 +124,7 @@ def split_whitespace_respecting_quotes(line: str) -> Iterable[str]:
             part = ""
             continue
 
-        if c in ' \t\n':
+        if c in " \t\n":
             if part:
                 yield part
             part = ""

riscemu/tools/riscemu

@@ -0,0 +1,2 @@
+#!/usr/bin/env bash
+python3 -m riscemu "$@"

riscemu/types/__init__.py

@@ -1,4 +1,4 @@
-from typing import Dict
+from typing import Dict, Any
 import re
 
 # define some base type aliases so we can keep track of absolute and relative addresses
@@ -6,13 +6,14 @@ T_RelativeAddress = int
 T_AbsoluteAddress = int
 
 # parser options are just dictionaries with arbitrary values
-T_ParserOpts = Dict[str, any]
+T_ParserOpts = Dict[str, Any]
 
-NUMBER_SYMBOL_PATTERN = re.compile(r'^\d+[fb]$')
+NUMBER_SYMBOL_PATTERN = re.compile(r"^\d+[fb]$")
 
 # base classes
 from .flags import MemoryFlags
 from .int32 import UInt32, Int32
+from .float32 import Float32
 from .instruction import Instruction
 from .instruction_context import InstructionContext
 from .memory_section import MemorySection
@@ -24,6 +25,17 @@ from .instruction_memory_section import InstructionMemorySection
 from .binary_data_memory_section import BinaryDataMemorySection
 
 # exceptions
-from .exceptions import ParseException, NumberFormatException, MemoryAccessException, OutOfMemoryException, \
-    LinkerException, LaunchDebuggerException, RiscemuBaseException, InvalidRegisterException, \
-    InvalidAllocationException, InvalidSyscallException, UnimplementedInstruction
+from .exceptions import (
+    ParseException,
+    NumberFormatException,
+    MemoryAccessException,
+    OutOfMemoryException,
+    LinkerException,
+    LaunchDebuggerException,
+    RiscemuBaseException,
+    InvalidRegisterException,
+    InvalidAllocationException,
+    InvalidSyscallException,
+    UnimplementedInstruction,
+    INS_NOT_IMPLEMENTED,
+)

riscemu/types/binary_data_memory_section.py

@@ -1,29 +1,56 @@
-from . import MemorySection, InstructionContext, MemoryFlags, T_RelativeAddress, Instruction
+from typing import Optional
+from . import (
+    MemorySection,
+    InstructionContext,
+    MemoryFlags,
+    T_RelativeAddress,
+    Instruction,
+)
 from ..types.exceptions import MemoryAccessException
 
 
 class BinaryDataMemorySection(MemorySection):
-    def __init__(self, data: bytearray, name: str, context: InstructionContext, owner: str, base: int = 0, flags: MemoryFlags = None):
-        self.name = name
-        self.base = base
-        self.context = context
-        self.size = len(data)
-        self.flags = flags if flags is not None else MemoryFlags(False, False)
+    def __init__(
+        self,
+        data: bytearray,
+        name: str,
+        context: InstructionContext,
+        owner: str,
+        base: int = 0,
+        flags: Optional[MemoryFlags] = None,
+    ):
+        super().__init__(
+            name,
+            flags if flags is not None else MemoryFlags(False, False),
+            len(data),
+            base,
+            owner,
+            context,
+        )
         self.data = data
-        self.owner = owner
 
     def read(self, offset: T_RelativeAddress, size: int) -> bytearray:
         if offset + size > self.size:
-            raise MemoryAccessException("Out of bounds access in {}".format(self), offset, size, 'read')
-        return self.data[offset:offset + size]
+            raise MemoryAccessException(
+                "Out of bounds access in {}".format(self), offset, size, "read"
+            )
+        return self.data[offset : offset + size]
 
     def write(self, offset: T_RelativeAddress, size: int, data: bytearray):
         if offset + size > self.size:
-            raise MemoryAccessException("Out of bounds access in {}".format(self), offset, size, 'write')
+            raise MemoryAccessException(
+                "Out of bounds access in {}".format(self), offset, size, "write"
+            )
         if len(data[0:size]) != size:
-            raise MemoryAccessException("Invalid write parameter sizing", offset, size, 'write')
-        self.data[offset:offset + size] = data[0:size]
+            raise MemoryAccessException(
+                "Invalid write parameter sizing", offset, size, "write"
+            )
+        self.data[offset : offset + size] = data[0:size]
 
     def read_ins(self, offset: T_RelativeAddress) -> Instruction:
-        raise MemoryAccessException("Tried reading instruction on non-executable section {}".format(self),
-                                    offset, 4, 'instruction fetch')
+        raise MemoryAccessException(
+            "Tried reading instruction on non-executable section {}".format(self),
+            offset,
+            4,
+            "instruction fetch",
+        )

riscemu/types/cpu.py

@@ -1,12 +1,15 @@
-import typing
 from abc import ABC, abstractmethod
-from typing import List, Type, Callable, Set, Dict
+from typing import List, Type, Callable, Set, Dict, TYPE_CHECKING, Iterable
 
 from ..registers import Registers
 from ..config import RunConfig
-from ..colors import FMT_RED, FMT_NONE, FMT_ERROR, FMT_CPU
+from ..colors import FMT_NONE, FMT_CPU
 from . import T_AbsoluteAddress, Instruction, Program, ProgramLoader
 
+if TYPE_CHECKING:
+    from ..MMU import MMU
+    from ..instructions import InstructionSet
+
 
 class CPU(ABC):
     # static cpu configuration
@@ -14,7 +17,7 @@ class CPU(ABC):
 
     # housekeeping variables
     regs: Registers
-    mmu: 'MMU'
+    mmu: "MMU"
     pc: T_AbsoluteAddress
     cycle: int
     halted: bool
@@ -24,14 +27,19 @@ class CPU(ABC):
 
     # instruction information
     instructions: Dict[str, Callable[[Instruction], None]]
-    instruction_sets: Set['InstructionSet']
+    instruction_sets: Set["InstructionSet"]
 
     # configuration
     conf: RunConfig
 
-    def __init__(self, mmu: 'MMU', instruction_sets: List[Type['InstructionSet']], conf: RunConfig):
+    def __init__(
+        self,
+        mmu: "MMU",
+        instruction_sets: List[Type["InstructionSet"]],
+        conf: RunConfig,
+    ):
         self.mmu = mmu
-        self.regs = Registers()
+        self.regs = Registers(conf.unlimited_registers)
         self.conf = conf
 
         self.instruction_sets = set()
@@ -71,32 +79,37 @@ class CPU(ABC):
             self.pc,
             self.cycle,
             self.halted,
-            " ".join(s.name for s in self.instruction_sets)
+            " ".join(s.name for s in self.instruction_sets),
         )
 
     @abstractmethod
-    def step(self, verbose=False):
+    def step(self, verbose: bool = False):
         pass
 
     @abstractmethod
-    def run(self, verbose=False):
+    def run(self, verbose: bool = False):
         pass
 
-    def launch(self, program: Program, verbose: bool = False):
-        if program not in self.mmu.programs:
-            print(FMT_ERROR + '[CPU] Cannot launch program that\'s not loaded!' + FMT_NONE)
-            return
+    def launch(self, verbose: bool = False):
+        entrypoint = self.mmu.find_entrypoint()
+
+        if entrypoint is None:
+            entrypoint = self.mmu.programs[0].entrypoint
+
         if self.conf.verbosity > 0:
-            print(FMT_CPU + "[CPU] Started running from {}".format(
-                self.mmu.translate_address(program.entrypoint)
-            ) + FMT_NONE)
-            print(program)
-        self.pc = program.entrypoint
+            print(
+                FMT_CPU
+                + "[CPU] Started running from {}".format(
+                    self.mmu.translate_address(entrypoint)
+                )
+                + FMT_NONE
+            )
+        self.pc = entrypoint
         self.run(verbose)
 
     @classmethod
     @abstractmethod
-    def get_loaders(cls) -> typing.Iterable[Type[ProgramLoader]]:
+    def get_loaders(cls) -> Iterable[Type[ProgramLoader]]:
         pass
 
     def get_best_loader_for(self, file_name: str) -> Type[ProgramLoader]:

riscemu/types/exceptions.py

@@ -14,63 +14,86 @@ if typing.TYPE_CHECKING:
 
 class RiscemuBaseException(BaseException):
     @abstractmethod
-    def message(self):
-        pass
+    def message(self) -> str:
+        raise NotImplemented
 
     def print_stacktrace(self):
         import traceback
+
         traceback.print_exception(type(self), self, self.__traceback__)
 
+
 # Parsing exceptions:
 
+
 class ParseException(RiscemuBaseException):
-    def __init__(self, msg, data=None):
+    def __init__(self, msg: str, data=None):
         super().__init__(msg, data)
         self.msg = msg
         self.data = data
 
     def message(self):
-        return FMT_PARSE + "{}(\"{}\", data={})".format(self.__class__.__name__, self.msg, self.data) + FMT_NONE
+        return (
+            FMT_PARSE
+            + '{}("{}", data={})'.format(self.__class__.__name__, self.msg, self.data)
+            + FMT_NONE
+        )
 
 
 def ASSERT_EQ(a1, a2):
     if a1 != a2:
-        raise ParseException("ASSERTION_FAILED: Expected elements to be equal!", (a1, a2))
+        raise ParseException(
+            "ASSERTION_FAILED: Expected elements to be equal!", (a1, a2)
+        )
 
 
 def ASSERT_LEN(a1, size):
     if len(a1) != size:
-        raise ParseException("ASSERTION_FAILED: Expected {} to be of length {}".format(a1, size), (a1, size))
+        raise ParseException(
+            "ASSERTION_FAILED: Expected {} to be of length {}".format(a1, size),
+            (a1, size),
+        )
 
 
 def ASSERT_NOT_NULL(a1):
     if a1 is None:
-        raise ParseException("ASSERTION_FAILED: Expected {} to be non null".format(a1), (a1,))
+        raise ParseException(
+            "ASSERTION_FAILED: Expected {} to be non null".format(a1), (a1,)
+        )
 
 
 def ASSERT_NOT_IN(a1, a2):
     if a1 in a2:
-        raise ParseException("ASSERTION_FAILED: Expected {} to not be in {}".format(a1, a2), (a1, a2))
+        raise ParseException(
+            "ASSERTION_FAILED: Expected {} to not be in {}".format(a1, a2), (a1, a2)
+        )
 
 
 def ASSERT_IN(a1, a2):
     if a1 not in a2:
-        raise ParseException("ASSERTION_FAILED: Expected {} to not be in {}".format(a1, a2), (a1, a2))
+        raise ParseException(
+            "ASSERTION_FAILED: Expected {} to not be in {}".format(a1, a2), (a1, a2)
+        )
 
 
 class LinkerException(RiscemuBaseException):
-    def __init__(self, msg, data):
+    def __init__(self, msg: str, data):
         self.msg = msg
         self.data = data
 
     def message(self):
-        return FMT_PARSE + "{}(\"{}\", data={})".format(self.__class__.__name__, self.msg, self.data) + FMT_NONE
+        return (
+            FMT_PARSE
+            + '{}("{}", data={})'.format(self.__class__.__name__, self.msg, self.data)
+            + FMT_NONE
+        )
 
 
 # MMU Exceptions
 
+
 class MemoryAccessException(RiscemuBaseException):
-    def __init__(self, msg, addr, size, op):
+    def __init__(self, msg: str, addr, size, op):
         super(MemoryAccessException, self).__init__()
         self.msg = msg
         self.addr = addr
@@ -78,13 +101,13 @@ class MemoryAccessException(RiscemuBaseException):
         self.op = op
 
     def message(self):
-        return FMT_MEM + "{}(During {} at 0x{:08x} of size {}: {})".format(
-            self.__class__.__name__,
-            self.op,
-            self.addr,
-            self.size,
-            self.msg
-        ) + FMT_NONE
+        return (
+            FMT_MEM
+            + "{}(During {} at 0x{:08x} of size {}: {})".format(
+                self.__class__.__name__, self.op, self.addr, self.size, self.msg
+            )
+            + FMT_NONE
+        )
 
 
 class OutOfMemoryException(RiscemuBaseException):
@@ -92,10 +115,13 @@ class OutOfMemoryException(RiscemuBaseException):
         self.action = action
 
     def message(self):
-        return FMT_MEM + '{}(Ran out of memory during {})'.format(
-            self.__class__.__name__,
-            self.action
-        ) + FMT_NONE
+        return (
+            FMT_MEM
+            + "{}(Ran out of memory during {})".format(
+                self.__class__.__name__, self.action
+            )
+            + FMT_NONE
+        )
 
 
 class InvalidAllocationException(RiscemuBaseException):
@@ -106,28 +132,29 @@ class InvalidAllocationException(RiscemuBaseException):
         self.flags = flags
 
     def message(self):
-        return FMT_MEM + '{}[{}](name={}, size={}, flags={})'.format(
-            self.__class__.__name__,
-            self.msg,
-            self.name,
-            self.size,
-            self.flags
+        return FMT_MEM + "{}[{}](name={}, size={}, flags={})".format(
+            self.__class__.__name__, self.msg, self.name, self.size, self.flags
         )
 
+
 # CPU Exceptions
 
 
 class UnimplementedInstruction(RiscemuBaseException):
-    def __init__(self, ins: 'Instruction', context = None):
+    def __init__(self, ins: "Instruction", context=None):
         self.ins = ins
         self.context = context
 
     def message(self):
-        return FMT_CPU + "{}({}{})".format(
-            self.__class__.__name__,
-            repr(self.ins),
-            ', context={}'.format(self.context) if self.context is not None else ''
-        ) + FMT_NONE
+        return (
+            FMT_CPU
+            + "{}({}{})".format(
+                self.__class__.__name__,
+                repr(self.ins),
+                ", context={}".format(self.context) if self.context is not None else "",
+            )
+            + FMT_NONE
+        )
 
 
 class InvalidRegisterException(RiscemuBaseException):
@@ -135,10 +162,11 @@ class InvalidRegisterException(RiscemuBaseException):
         self.reg = reg
 
     def message(self):
-        return FMT_CPU + "{}(Invalid register {})".format(
-            self.__class__.__name__,
-            self.reg
-        ) + FMT_NONE
+        return (
+            FMT_CPU
+            + "{}(Invalid register {})".format(self.__class__.__name__, self.reg)
+            + FMT_NONE
+        )
 
 
 class InvalidSyscallException(RiscemuBaseException):
@@ -146,10 +174,11 @@ class InvalidSyscallException(RiscemuBaseException):
         self.scall = scall
 
     def message(self):
-        return FMT_SYSCALL + "{}(Invalid syscall: {})".format(
-            self.__class__.__name__,
-            self.scall
-        ) + FMT_NONE
+        return (
+            FMT_SYSCALL
+            + "{}(Invalid syscall: {})".format(self.__class__.__name__, self.scall)
+            + FMT_NONE
+        )
 
 
 def INS_NOT_IMPLEMENTED(ins):
@@ -162,13 +191,10 @@ class NumberFormatException(RiscemuBaseException):
         self.msg = msg
 
     def message(self):
-        return "{}({})".format(
-            self.__class__.__name__,
-            self.msg
-        )
+        return "{}({})".format(self.__class__.__name__, self.msg)
 
 
 # this exception is not printed and simply signals that an interactive debugging session is
 class LaunchDebuggerException(RiscemuBaseException):
-    def message(self):
+    def message(self) -> str:
         return ""

riscemu/types/flags.py

@@ -8,6 +8,5 @@ class MemoryFlags:
 
     def __repr__(self):
         return "r{}{}".format(
-            '-' if self.read_only else 'w',
-            'x' if self.executable else '-'
+            "-" if self.read_only else "w", "x" if self.executable else "-"
         )

riscemu/types/float32.py

@@ -0,0 +1,198 @@
+import struct
+from ctypes import c_float
+from typing import Union, Any
+
+bytes_t = bytes
+
+
+class Float32:
+    __slots__ = ("_val",)
+
+    _val: c_float
+
+    @property
+    def value(self) -> float:
+        """
+        The value represented by this float
+        """
+        return self._val.value
+
+    @property
+    def bytes(self) -> bytes:
+        """
+        The values bit representation (as a bytes object)
+        """
+        return struct.pack("<f", self.value)
+
+    @property
+    def bits(self) -> int:
+        """
+        The values bit representation as an int (for easy bit manipulation)
+        """
+        return int.from_bytes(self.bytes, byteorder="little")
+
+    @classmethod
+    def from_bytes(cls, val: Union[int, bytes_t, bytearray]):
+        if isinstance(val, int):
+            val = struct.unpack("!f", struct.pack("!I", val))[0]
+        return Float32(val)
+
+    def __init__(
+        self, val: Union[float, c_float, "Float32", bytes_t, bytearray, int] = 0
+    ):
+        if isinstance(val, (float, int)):
+            self._val = c_float(val)
+        elif isinstance(val, c_float):
+            self._val = c_float(val.value)
+        elif isinstance(val, (bytes, bytearray)):
+            self._val = c_float(struct.unpack("<f", val)[0])
+        elif isinstance(val, Float32):
+            self._val = val._val
+        else:
+            raise ValueError(
+                "Unsupported value passed to Float32: {} ({})".format(
+                    repr(val), type(val)
+                )
+            )
+
+    def __add__(self, other: Union["Float32", float]):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.__class__(self.value + other)
+
+    def __sub__(self, other: Union["Float32", float]):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.__class__(self.value - other)
+
+    def __mul__(self, other: Union["Float32", float]):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.__class__(self.value * other)
+
+    def __truediv__(self, other: Any):
+        return self // other
+
+    def __floordiv__(self, other: Any):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.__class__(self.value // other)
+
+    def __mod__(self, other: Union["Float32", float]):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.__class__(self.value % other)
+
+    def __eq__(self, other: object) -> bool:
+        if isinstance(other, (float, int)):
+            return self.value == other
+        elif isinstance(other, Float32):
+            return self.value == other.value
+        return False
+
+    def __neg__(self):
+        return self.__class__(-self.value)
+
+    def __abs__(self):
+        return self.__class__(abs(self.value))
+
+    def __bytes__(self) -> bytes_t:
+        return self.bytes
+
+    def __repr__(self):
+        return "{}({})".format(self.__class__.__name__, self.value)
+
+    def __str__(self):
+        return str(self.value)
+
+    def __format__(self, format_spec: str):
+        return self.value.__format__(format_spec)
+
+    def __hash__(self):
+        return hash(self.value)
+
+    def __gt__(self, other: Any):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.value > other
+
+    def __lt__(self, other: Any):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.value < other
+
+    def __le__(self, other: Any):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.value <= other
+
+    def __ge__(self, other: Any):
+        if isinstance(other, Float32):
+            other = other.value
+        return self.value >= other
+
+    def __bool__(self):
+        return bool(self.value)
+
+    def __pow__(self, power, modulo=None):
+        if modulo is not None:
+            raise ValueError("Float32 pow with modulo unsupported")
+        return self.__class__(self.value**power)
+
+    # right handed binary operators
+
+    def __radd__(self, other: Any):
+        return self + other
+
+    def __rsub__(self, other: Any):
+        return self.__class__(other) - self
+
+    def __rmul__(self, other: Any):
+        return self * other
+
+    def __rtruediv__(self, other: Any):
+        return self.__class__(other) // self
+
+    def __rfloordiv__(self, other: Any):
+        return self.__class__(other) // self
+
+    def __rmod__(self, other: Any):
+        return self.__class__(other) % self
+
+    def __rand__(self, other: Any):
+        return self.__class__(other) & self
+
+    def __ror__(self, other: Any):
+        return self.__class__(other) | self
+
+    def __rxor__(self, other: Any):
+        return self.__class__(other) ^ self
+
+    # bytewise operators:
+
+    def __and__(self, other: Union["Float32", float, int]):
+        if isinstance(other, float):
+            other = Float32(other)
+        if isinstance(other, Float32):
+            other = other.bits
+        return self.from_bytes(self.bits & other)
+
+    def __or__(self, other: Union["Float32", float]):
+        if isinstance(other, float):
+            other = Float32(other)
+        if isinstance(other, Float32):
+            other = other.bits
+        return self.from_bytes(self.bits | other)
+
+    def __xor__(self, other: Union["Float32", float]):
+        if isinstance(other, float):
+            other = Float32(other)
+        if isinstance(other, Float32):
+            other = other.bits
+        return self.from_bytes(self.bits ^ other)
+
+    def __lshift__(self, other: int):
+        return self.from_bytes(self.bits << other)
+
+    def __rshift__(self, other: int):
+        return self.from_bytes(self.bits >> other)

riscemu/types/instruction_context.py

@@ -18,8 +18,8 @@ class InstructionContext:
 
     numbered_labels: Dict[str, List[T_RelativeAddress]]
     """
-    This dictionary maps numbered labels (which can occur multiple times) to a list of (block-relative) addresses where 
-    the label was placed 
+    This dictionary maps numbered labels (which can occur multiple times) to a list of (block-relative) addresses where
+    the label was placed
     """
 
     global_symbol_dict: Dict[str, T_AbsoluteAddress]
@@ -33,21 +33,32 @@ class InstructionContext:
         self.base_address = 0
         self.global_symbol_dict = dict()
 
-    def resolve_label(self, symbol: str, address_at: Optional[T_RelativeAddress] = None) -> Optional[T_AbsoluteAddress]:
+    def resolve_label(
+        self, symbol: str, address_at: Optional[T_RelativeAddress] = None
+    ) -> Optional[T_AbsoluteAddress]:
         if NUMBER_SYMBOL_PATTERN.match(symbol):
             if address_at is None:
-                raise ParseException("Cannot resolve relative symbol {} without an address!".format(symbol))
+                raise ParseException(
+                    "Cannot resolve relative symbol {} without an address!".format(
+                        symbol
+                    )
+                )
 
             direction = symbol[-1]
             values = self.numbered_labels.get(symbol[:-1], [])
-            if direction == 'b':
-                return max((addr + self.base_address for addr in values if addr < address_at), default=None)
+            if direction == "b":
+                return max(
+                    (addr + self.base_address for addr in values if addr < address_at),
+                    default=None,
+                )
             else:
-                return min((addr + self.base_address for addr in values if addr > address_at), default=None)
+                return min(
+                    (addr + self.base_address for addr in values if addr > address_at),
+                    default=None,
+                )
         else:
             # if it's not a local symbol, try the globals
             if symbol not in self.labels:
                 return self.global_symbol_dict.get(symbol, None)
             # otherwise return the local symbol
             return self.labels.get(symbol, None)
-

riscemu/types/instruction_memory_section.py

@@ -1,11 +1,24 @@
 from typing import List
 
-from . import MemorySection, Instruction, InstructionContext, MemoryFlags, T_RelativeAddress
+from . import (
+    MemorySection,
+    Instruction,
+    InstructionContext,
+    MemoryFlags,
+    T_RelativeAddress,
+)
 from .exceptions import MemoryAccessException
 
 
 class InstructionMemorySection(MemorySection):
-    def __init__(self, instructions: List[Instruction], name: str, context: InstructionContext, owner: str, base: int = 0):
+    def __init__(
+        self,
+        instructions: List[Instruction],
+        name: str,
+        context: InstructionContext,
+        owner: str,
+        base: int = 0,
+    ):
         self.name = name
         self.base = base
         self.context = context
@@ -15,13 +28,27 @@ class InstructionMemorySection(MemorySection):
         self.owner = owner
 
     def read(self, offset: T_RelativeAddress, size: int) -> bytearray:
-        raise MemoryAccessException("Cannot read raw bytes from instruction section", self.base + offset, size, 'read')
+        raise MemoryAccessException(
+            "Cannot read raw bytes from instruction section",
+            self.base + offset,
+            size,
+            "read",
+        )
 
     def write(self, offset: T_RelativeAddress, size: int, data: bytearray):
-        raise MemoryAccessException("Cannot write raw bytes to instruction section", self.base + offset, size, 'write')
+        raise MemoryAccessException(
+            "Cannot write raw bytes to instruction section",
+            self.base + offset,
+            size,
+            "write",
+        )
 
     def read_ins(self, offset: T_RelativeAddress) -> Instruction:
         if offset % 4 != 0:
-            raise MemoryAccessException("Unaligned instruction fetch!", self.base + offset, 4, 'instruction fetch')
+            raise MemoryAccessException(
+                "Unaligned instruction fetch!",
+                self.base + offset,
+                4,
+                "instruction fetch",
+            )
         return self.instructions[offset // 4]
-

riscemu/types/int32.py

@@ -1,4 +1,4 @@
-from typing import Union
+from typing import Any, Union
 from ctypes import c_int32, c_uint32
 
 
@@ -11,82 +11,91 @@ class Int32:
     You can use it just like you would any other integer, just be careful when passing it
     to functions which actually expect an integer and not a Int32.
     """
+
     _type = c_int32
-    __slots__ = ('_val',)
+    __slots__ = ("_val",)
 
-    def __init__(self, val: Union[int, c_int32, c_uint32, 'Int32', bytes, bytearray] = 0):
+    def __init__(
+        self, val: Union[int, c_int32, c_uint32, "Int32", bytes, bytearray, bool] = 0
+    ):
         if isinstance(val, (bytes, bytearray)):
             signed = len(val) == 4 and self._type == c_int32
-            self._val = self.__class__._type(int.from_bytes(val, 'little', signed=signed))
+            self._val = self.__class__._type(
+                int.from_bytes(val, "little", signed=signed)
+            )
         elif isinstance(val, self.__class__._type):
             self._val = val
         elif isinstance(val, (c_uint32, c_int32, Int32)):
             self._val = self.__class__._type(val.value)
-        elif isinstance(val, int):
+        elif isinstance(val, (int, bool)):
             self._val = self.__class__._type(val)
         else:
             raise RuntimeError(
-                "Unknonw {} input type: {} ({})".format(self.__class__.__name__, type(val), val)
+                "Unknonw {} input type: {} ({})".format(
+                    self.__class__.__name__, type(val), val
+                )
             )
 
-    def __add__(self, other: Union['Int32', int]):
+    def __add__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
 
         return self.__class__(self._val.value + other)
 
-    def __sub__(self, other: Union['Int32', int]):
+    def __sub__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self._val.value - other)
 
-    def __mul__(self, other: Union['Int32', int]):
+    def __mul__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self._val.value * other)
 
-    def __truediv__(self, other):
+    def __truediv__(self, other: Any):
         return self // other
 
-    def __floordiv__(self, other):
+    def __floordiv__(self, other: Any):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self.value // other)
 
-    def __mod__(self, other: Union['Int32', int]):
+    def __mod__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self._val.value % other)
 
-    def __and__(self, other: Union['Int32', int]):
+    def __and__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self._val.value & other)
 
-    def __or__(self, other: Union['Int32', int]):
+    def __or__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self._val.value | other)
 
-    def __xor__(self, other: Union['Int32', int]):
+    def __xor__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self._val.value ^ other)
 
-    def __lshift__(self, other: Union['Int32', int]):
+    def __lshift__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self.value << other)
 
-    def __rshift__(self, other: Union['Int32', int]):
+    def __rshift__(self, other: Union["Int32", int]):
         if isinstance(other, Int32):
             other = other.value
         return self.__class__(self.value >> other)
 
-    def __eq__(self, other: Union['Int32', int]):
-        if isinstance(other, Int32):
-            other = other.value
-        return self.value == other
+    def __eq__(self, other: object) -> bool:
+        if isinstance(other, int):
+            return self.value == other
+        elif isinstance(other, Int32):
+            return self.value == other.value
+        return False
 
     def __neg__(self):
         return self.__class__(-self._val.value)
@@ -98,33 +107,33 @@ class Int32:
         return self.to_bytes(4)
 
     def __repr__(self):
-        return '{}({})'.format(self.__class__.__name__, self.value)
+        return "{}({})".format(self.__class__.__name__, self.value)
 
     def __str__(self):
         return str(self.value)
 
-    def __format__(self, format_spec):
+    def __format__(self, format_spec: str):
         return self.value.__format__(format_spec)
 
     def __hash__(self):
         return hash(self.value)
 
-    def __gt__(self, other):
+    def __gt__(self, other: Any):
         if isinstance(other, Int32):
             other = other.value
         return self.value > other
 
-    def __lt__(self, other):
+    def __lt__(self, other: Any):
         if isinstance(other, Int32):
             other = other.value
         return self.value < other
 
-    def __le__(self, other):
+    def __le__(self, other: Any):
         if isinstance(other, Int32):
             other = other.value
         return self.value <= other
 
-    def __ge__(self, other):
+    def __ge__(self, other: Any):
         if isinstance(other, Int32):
             other = other.value
         return self.value >= other
@@ -132,38 +141,38 @@ class Int32:
     def __bool__(self):
         return bool(self.value)
 
-    def __cmp__(self, other):
+    def __cmp__(self, other: Any):
         if isinstance(other, Int32):
             other = other.value
         return self.value.__cmp__(other)
 
     # right handed binary operators
 
-    def __radd__(self, other):
+    def __radd__(self, other: Any):
         return self + other
 
-    def __rsub__(self, other):
+    def __rsub__(self, other: Any):
         return self.__class__(other) - self
 
-    def __rmul__(self, other):
+    def __rmul__(self, other: Any):
         return self * other
 
-    def __rtruediv__(self, other):
+    def __rtruediv__(self, other: Any):
         return self.__class__(other) // self
 
-    def __rfloordiv__(self, other):
+    def __rfloordiv__(self, other: Any):
         return self.__class__(other) // self
 
-    def __rmod__(self, other):
+    def __rmod__(self, other: Any):
         return self.__class__(other) % self
 
-    def __rand__(self, other):
+    def __rand__(self, other: Any):
         return self.__class__(other) & self
 
-    def __ror__(self, other):
+    def __ror__(self, other: Any):
         return self.__class__(other) | self
 
-    def __rxor__(self, other):
+    def __rxor__(self, other: Any):
         return self.__class__(other) ^ self
 
     @property
@@ -174,7 +183,7 @@ class Int32:
         """
         return self._val.value
 
-    def unsigned(self) -> 'UInt32':
+    def unsigned(self) -> "UInt32":
         """
         Convert to an unsigned representation. See :class:Uint32
         :return:
@@ -188,9 +197,9 @@ class Int32:
         :param bytes: The length of the bytearray
         :return: A little-endian representation of the contained integer
         """
-        return bytearray(self.unsigned_value.to_bytes(4, 'little'))[0:bytes]
+        return bytearray(self.unsigned_value.to_bytes(4, "little"))[0:bytes]
 
-    def signed(self) -> 'Int32':
+    def signed(self) -> "Int32":
         """
         Convert to a signed representation. See :class:Int32
         :return:
@@ -207,7 +216,7 @@ class Int32:
         """
         return c_uint32(self.value).value
 
-    def shift_right_logical(self, ammount: Union['Int32', int]) -> 'Int32':
+    def shift_right_logical(self, ammount: Union["Int32", int]) -> "Int32":
         """
         This function implements logical right shifts, meaning that the sign bit is shifted as well.
 
@@ -237,12 +246,12 @@ class Int32:
         :return: An instance of Int32, holding the sign-extended value
         """
         if isinstance(data, (bytes, bytearray)):
-            data = int.from_bytes(data, 'little')
+            data = int.from_bytes(data, "little")
         sign = data >> (bits - 1)
         if sign > 1:
             print("overflow in Int32.sext!")
         if sign:
-            data = (data & (2 ** (bits - 1) - 1)) - 2**(bits-1)
+            data = (data & (2 ** (bits - 1) - 1)) - 2 ** (bits - 1)
         return cls(data)
 
 
@@ -250,9 +259,10 @@ class UInt32(Int32):
     """
     An unsigned version of :class:Int32.
     """
+
     _type = c_uint32
 
-    def unsigned(self) -> 'UInt32':
+    def unsigned(self) -> "UInt32":
         """
         Return a new instance representing the same bytes, but signed
         :return:
@@ -263,11 +273,13 @@ class UInt32(Int32):
     def unsigned_value(self) -> int:
         return self._val.value
 
-    def shift_right_logical(self, ammount: Union['Int32', int]) -> 'UInt32':
+    def shift_right_logical(self, ammount: Union["Int32", int]) -> "UInt32":
         """
         see :meth:`Int32.shift_right_logical <Int32.shift_right_logical>`
 
         :param ammount: Number of positions to shift
         :return: A new Int32 object representing the shifted value (keeps the signed-ness of the source)
         """
-        return self >> ammount
+        if isinstance(ammount, Int32):
+            ammount = ammount.value
+        return UInt32(self.value >> ammount)

riscemu/types/memory_section.py

@@ -4,7 +4,14 @@ from typing import Optional
 
 from ..colors import FMT_MEM, FMT_NONE, FMT_UNDERLINE, FMT_ORANGE, FMT_ERROR
 from ..helpers import format_bytes
-from . import MemoryFlags, T_AbsoluteAddress, InstructionContext, T_RelativeAddress, Instruction, Int32
+from . import (
+    MemoryFlags,
+    T_AbsoluteAddress,
+    InstructionContext,
+    T_RelativeAddress,
+    Instruction,
+    Int32,
+)
 
 
 @dataclass
@@ -32,8 +39,16 @@ class MemorySection(ABC):
     def read_ins(self, offset: T_RelativeAddress) -> Instruction:
         pass
 
-    def dump(self, start: T_RelativeAddress, end: Optional[T_RelativeAddress] = None, fmt: str = None,
-             bytes_per_row: int = None, rows: int = 10, group: int = None, highlight: int = None):
+    def dump(
+        self,
+        start: T_RelativeAddress,
+        end: Optional[T_RelativeAddress] = None,
+        fmt: Optional[str] = None,
+        bytes_per_row: Optional[int] = None,
+        rows: int = 10,
+        group: Optional[int] = None,
+        highlight: Optional[int] = None,
+    ):
         """
         Dump the section. If no end is given, the rows around start are printed and start is highlighted.
 
@@ -54,11 +69,11 @@ class MemorySection(ABC):
         if fmt is None:
             if self.flags.executable and self.flags.read_only:
                 bytes_per_row = 4
-                fmt = 'asm'
+                fmt = "asm"
             else:
-                fmt = 'hex'
+                fmt = "hex"
 
-        if fmt == 'char':
+        if fmt == "char":
             if bytes_per_row is None:
                 bytes_per_row = 4
             if group is None:
@@ -70,50 +85,82 @@ class MemorySection(ABC):
         if bytes_per_row is None:
             bytes_per_row = 4
 
-        if fmt not in ('asm', 'hex', 'int', 'char'):
-            print(FMT_ERROR + '[MemorySection] Unknown format {}, known formats are {}'.format(
-                fmt, ", ".join(('asm', 'hex', 'int', 'char'))
-            ) + FMT_NONE)
+        if fmt not in ("asm", "hex", "int", "char"):
+            print(
+                FMT_ERROR
+                + "[MemorySection] Unknown format {}, known formats are {}".format(
+                    fmt, ", ".join(("asm", "hex", "int", "char"))
+                )
+                + FMT_NONE
+            )
 
         if end is None:
             end = min(start + (bytes_per_row * (rows // 2)), self.size)
             highlight = start
             start = max(0, start - (bytes_per_row * (rows // 2)))
 
-        if fmt == 'asm':
-            print(FMT_MEM + "{}, viewing {} instructions:".format(
-                self, (end - start) // 4
-            ) + FMT_NONE)
+        if fmt == "asm":
+            print(
+                FMT_MEM
+                + "{}, viewing {} instructions:".format(self, (end - start) // 4)
+                + FMT_NONE
+            )
 
             for addr in range(start, end, 4):
                 if addr == highlight:
-                    print(FMT_UNDERLINE + FMT_ORANGE, end='')
-                print("0x{:04x}: {}{}".format(
-                    self.base + addr, self.read_ins(addr), FMT_NONE
-                ))
+                    print(FMT_UNDERLINE + FMT_ORANGE, end="")
+                print(
+                    "0x{:04x}: {}{}".format(
+                        self.base + addr, self.read_ins(addr), FMT_NONE
+                    )
+                )
         else:
-            print(FMT_MEM + "{}, viewing {} bytes:".format(
-                self, (end - start)
-            ) + FMT_NONE)
+            print(
+                FMT_MEM + "{}, viewing {} bytes:".format(self, (end - start)) + FMT_NONE
+            )
 
-            aligned_end = end - (end % bytes_per_row) if end % bytes_per_row != 0 else end
+            aligned_end = (
+                end - (end % bytes_per_row) if end % bytes_per_row != 0 else end
+            )
 
             for addr in range(start, aligned_end, bytes_per_row):
                 hi_ind = (highlight - addr) // group if highlight is not None else -1
-                print("0x{:04x}: {}{}".format(
-                    self.base + addr, format_bytes(self.read(addr, bytes_per_row), fmt, group, hi_ind), FMT_NONE
-                ))
+                print(
+                    "0x{:04x}: {}{}".format(
+                        self.base + addr,
+                        format_bytes(
+                            self.read(addr, bytes_per_row), fmt, group, hi_ind
+                        ),
+                        FMT_NONE,
+                    )
+                )
 
             if aligned_end != end:
-                hi_ind = (highlight - aligned_end) // group if highlight is not None else -1
-                print("0x{:04x}: {}{}".format(
-                    self.base + aligned_end, format_bytes(
-                        self.read(aligned_end, end % bytes_per_row), fmt, group, hi_ind
-                    ), FMT_NONE
-                ))
-
-    def dump_all(self, *args, **kwargs):
-        self.dump(0, self.size, *args, **kwargs)
+                hi_ind = (
+                    (highlight - aligned_end) // group if highlight is not None else -1
+                )
+                print(
+                    "0x{:04x}: {}{}".format(
+                        self.base + aligned_end,
+                        format_bytes(
+                            self.read(aligned_end, end % bytes_per_row),
+                            fmt,
+                            group,
+                            hi_ind,
+                        ),
+                        FMT_NONE,
+                    )
+                )
+
+    def dump_all(
+        self,
+        fmt: Optional[str] = None,
+        bytes_per_row: Optional[int] = None,
+        rows: int = 10,
+        group: Optional[int] = None,
+        highlight: Optional[int] = None,
+    ):
+        self.dump(0, self.size, fmt, bytes_per_row, rows, group, highlight)
 
     def __repr__(self):
         return "{}[{}] at 0x{:08X} (size={}bytes, flags={}, owner={})".format(
@@ -122,5 +169,5 @@ class MemorySection(ABC):
             self.base,
             self.size,
             self.flags,
-            self.owner
+            self.owner,
         )

riscemu/types/program.py

@@ -14,6 +14,7 @@ class Program:
     the offset in the program, and everything will be taken care of for you.
 
     """
+
     name: str
     context: InstructionContext
     global_labels: Set[str]
@@ -44,9 +45,13 @@ class Program:
         if self.base is not None:
             if sec.base < self.base:
                 print(
-                    FMT_RED + FMT_BOLD + "WARNING: memory section {} in {} is placed before program base (0x{:x})".format(
+                    FMT_RED
+                    + FMT_BOLD
+                    + "WARNING: memory section {} in {} is placed before program base (0x{:x})".format(
                         sec, self.name, self.base
-                    ) + FMT_NONE)
+                    )
+                    + FMT_NONE
+                )
 
         self.sections.append(sec)
         # keep section list ordered
@@ -54,18 +59,21 @@ class Program:
 
     def __repr__(self):
         return "{}(name={},sections={},base={})".format(
-            self.__class__.__name__, self.name, self.global_labels,
-            [s.name for s in self.sections], self.base
+            self.__class__.__name__,
+            self.name,
+            self.global_labels,
+            [s.name for s in self.sections],
+            self.base,
         )
 
     @property
     def entrypoint(self):
-        if '_start' in self.context.labels:
-            return self.context.labels.get('_start')
-        if 'main' in self.context.labels:
-            return self.context.labels.get('main')
+        if "_start" in self.context.labels:
+            return self.context.labels.get("_start")
+        if "main" in self.context.labels:
+            return self.context.labels.get("main")
         for sec in self.sections:
-            if get_section_base_name(sec.name) == '.text' and sec.flags.executable:
+            if get_section_base_name(sec.name) == ".text" and sec.flags.executable:
                 return sec.base
 
     def loaded_trigger(self, at_addr: T_AbsoluteAddress):
@@ -81,12 +89,17 @@ class Program:
         """
         if self.is_loaded:
             if at_addr != self.base:
-                raise RuntimeError("Program loaded twice at different addresses! This will probably break things!")
+                raise RuntimeError(
+                    "Program loaded twice at different addresses! This will probably break things!"
+                )
             return
 
         if self.base is not None and self.base != at_addr:
-            print(FMT_MEM + 'WARNING: Program loaded at different address then expected! (loaded at {}, '
-                            'but expects to be loaded at {})'.format(at_addr, self.base) + FMT_NONE)
+            print(
+                FMT_MEM
+                + "WARNING: Program loaded at different address then expected! (loaded at {}, "
+                "but expects to be loaded at {})".format(at_addr, self.base) + FMT_NONE
+            )
 
         # check if we are relocating
         if self.base != at_addr:

riscemu/types/program_loader.py

@@ -39,7 +39,7 @@ class ProgramLoader(ABC):
         pass
 
     @classmethod
-    def instantiate(cls, source_path: str, options: T_ParserOpts) -> 'ProgramLoader':
+    def instantiate(cls, source_path: str, options: T_ParserOpts) -> "ProgramLoader":
         """
         Instantiate a loader for the given source file with the required arguments
 

riscemu/types/simple_instruction.py

@@ -5,8 +5,13 @@ from ..helpers import parse_numeric_argument
 
 
 class SimpleInstruction(Instruction):
-    def __init__(self, name: str, args: Union[Tuple[()], Tuple[str], Tuple[str, str], Tuple[str, str, str]],
-                 context: InstructionContext, addr: T_RelativeAddress):
+    def __init__(
+        self,
+        name: str,
+        args: Union[Tuple[()], Tuple[str], Tuple[str, str], Tuple[str, str, str]],
+        context: InstructionContext,
+        addr: T_RelativeAddress,
+    ):
         self.context = context
         self.name = name
         self.args = args
@@ -23,4 +28,3 @@ class SimpleInstruction(Instruction):
 
     def get_reg(self, num: int) -> str:
         return self.args[num]
-

setup.py

@@ -1,4 +1,5 @@
 import setuptools
+from glob import glob
 
 import riscemu
 
@@ -10,7 +11,7 @@ setuptools.setup(
     version=riscemu.__version__,
     author=riscemu.__author__,
     author_email="pip@antonlydike.de",
-    description="RISC-V userspace and privileged emulator",
+    description="RISC-V userspace and machine mode emulator",
     long_description=long_description,
     long_description_content_type="text/markdown",
     url="https://github.com/antonlydike/riscemu",
@@ -23,9 +24,21 @@ setuptools.setup(
         "Operating System :: OS Independent",
     ],
     package_dir={"": "."},
-    packages=["riscemu", "riscemu.decoder", "riscemu.instructions", "riscemu.IO", "riscemu.priv", "riscemu.types"],
-    python_requires=">=3.6",
-    install_requires=[
-        "pyelftools~=0.27"
-    ]
-)
+    packages=[
+        "riscemu",
+        "riscemu.decoder",
+        "riscemu.instructions",
+        "riscemu.IO",
+        "riscemu.priv",
+        "riscemu.types",
+    ],
+    package_data={
+        "riscemu": ["libc/*.s", "py.typed"],
+    },
+    data_files=[
+        ("libc", glob("libc/*.s")),
+    ],
+    scripts=["riscemu/tools/riscemu"],
+    python_requires=">=3.8",
+    install_requires=["pyelftools~=0.27"],
+)

sphinx-docs/.gitignore

@@ -1,4 +1,4 @@
 build
 source/riscemu*.rst
 source/modules.rst
-source/help
+source/help

sphinx-docs/source/conf.py

@@ -18,37 +18,34 @@ import subprocess
 # import sys
 # sys.path.insert(0, os.path.abspath('.'))
 
-if os.getenv('READTHEDOCS', False) and not os.path.exists('riscemu.rst'):
-    subprocess.check_call(['../../generate-docs.sh', 'generate'])
+if os.getenv("READTHEDOCS", False) and not os.path.exists("riscemu.rst"):
+    subprocess.check_call(["../../generate-docs.sh", "generate"])
 
 # -- Project information -----------------------------------------------------
 
-project = 'RiscEmu'
-copyright = '2022, Anton Lydike'
-author = 'Anton Lydike'
+project = "RiscEmu"
+copyright = "2022, Anton Lydike"
+author = "Anton Lydike"
 
 # The full version, including alpha/beta/rc tags
-release = '2.0.0a2'
+release = "2.0.0a2"
 
 # -- General configuration ---------------------------------------------------
 
 # Add any Sphinx extension module names here, as strings. They can be
 # extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
 # ones.
-extensions = [
-    'sphinx.ext.autodoc',
-    'recommonmark'
-]
+extensions = ["sphinx.ext.autodoc", "recommonmark"]
 
 # autodoc options
 autodoc_default_options = {
-    'members': True,
-    'member-order': 'bysource',
-    'special-members': '__init__',
+    "members": True,
+    "member-order": "bysource",
+    "special-members": "__init__",
 }
 
 # Add any paths that contain templates here, relative to this directory.
-templates_path = ['_templates']
+templates_path = ["_templates"]
 
 # List of patterns, relative to source directory, that match files and
 # directories to ignore when looking for source files.
@@ -58,10 +55,10 @@ templates_path = ['_templates']
 from recommonmark.parser import CommonMarkParser
 
 # The suffix of source filenames.
-source_suffix = ['.rst', '.md']
+source_suffix = [".rst", ".md"]
 
 source_parsers = {
-    '.md': CommonMarkParser,
+    ".md": CommonMarkParser,
 }
 
 # -- Options for HTML output -------------------------------------------------
@@ -69,18 +66,18 @@ source_parsers = {
 # The theme to use for HTML and HTML Help pages.  See the documentation for
 # a list of builtin themes.
 #
-html_theme = 'alabaster'
+html_theme = "alabaster"
 
-pygments_style = 'sphinx'
+pygments_style = "sphinx"
 
 # Add any paths that contain custom static files (such as style sheets) here,
 # relative to this directory. They are copied after the builtin static files,
 # so a file named "default.css" will overwrite the builtin "default.css".
-html_static_path = ['_static']
+html_static_path = ["_static"]
 
-sys.path.insert(0, os.path.abspath('../../'))
+sys.path.insert(0, os.path.abspath("../../"))
 
-if os.getenv('READTHEDOCS', False):
+if os.getenv("READTHEDOCS", False):
     import sphinx_rtd_theme
 
     extensions.append("sphinx_rtd_theme")

sphinx-docs/source/index.rst

@@ -59,4 +59,4 @@ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+SOFTWARE.

test/__init__.py

@@ -1,3 +1,3 @@
 from .test_tokenizer import *
 from .test_helpers import *
-from .test_integers import *
+from .test_integers import *

test/end_to_end/end_to_end_test.py

@@ -1,73 +0,0 @@
-import contextlib
-import os
-from abc import abstractmethod
-from tempfile import NamedTemporaryFile
-from typing import Optional, Union, Tuple
-from unittest import TestCase
-
-from riscemu import CPU, UserModeCPU, InstructionSetDict, RunConfig
-from riscemu.types import Program
-
-
-class EndToEndTest(TestCase):
-
-    def __init__(self, cpu: Optional[CPU] = None):
-        super().__init__()
-        if cpu is None:
-            cpu = UserModeCPU(InstructionSetDict.values(), RunConfig())
-        self.cpu = cpu
-
-    @abstractmethod
-    def get_source(self) -> Tuple[str, Union[bytes, str, bytearray]]:
-        """
-        This method returns the source code of the program
-        :return:
-        """
-        pass
-
-    def test_run_program(self):
-        """
-        Runs the program and verifies output
-        :return:
-        """
-        with self.with_source_file() as names:
-            fname, orig_name = names
-            loader = self.cpu.get_best_loader_for(fname)
-            self.program = loader.instantiate(fname, loader.get_options([])).parse()
-            self._change_program_file_name(self.program, orig_name)
-            self.cpu.load_program(self.program)
-            self.after_program_load(self.program)
-            if isinstance(self.cpu, UserModeCPU):
-                self.cpu.setup_stack()
-            try:
-                self.cpu.launch(self.program)
-            except Exception as ex:
-                if self.is_exception_expected(ex):
-                    pass
-                raise ex
-
-    @contextlib.contextmanager
-    def with_source_file(self):
-        name, content = self.get_source()
-        if isinstance(content, str):
-            f = NamedTemporaryFile('w', suffix=name, delete=False)
-        else:
-            f = NamedTemporaryFile('wb', suffix=name, delete=False)
-        f.write(content)
-        f.flush()
-        f.close()
-        try:
-            yield f.name, name
-        finally:
-            os.unlink(f.name)
-
-    def after_program_load(self, program):
-        pass
-
-    def is_exception_expected(self, ex: Exception) -> bool:
-        return False
-
-    def _change_program_file_name(self, program: Program, new_name: str):
-        program.name = new_name
-        for sec in program.sections:
-            sec.owner = new_name

test/filecheck/.gitignore

@@ -0,0 +1,2 @@
+.lit_test_times.txt
+Output

test/filecheck/fibs.asm

@@ -0,0 +1,24 @@
+// RUN: python3 -m riscemu -v -o ignore_exit_code %s | filecheck %s
+.data
+fibs:   .space 1024
+
+.text
+main:
+        addi    s1, zero, 0     // storage index
+        addi    s2, zero, 1024  // last storage index
+        addi    t0, zero, 1     // t0 = F_{i}
+        addi    t1, zero, 1     // t1 = F_{i+1}
+loop:
+        sw      t0, fibs(s1)    // save
+        add     t2, t1, t0      // t2 = F_{i+2}
+        addi    t0, t1, 0       // t0 = t1
+        addi    t1, t2, 0       // t1 = t2
+        addi    s1, s1, 4       // increment storage pointer
+        blt     s1, s2, loop    // loop as long as we did not reach array length
+        ebreak
+        // exit gracefully
+        add     a0, zero, t2
+        addi    a7, zero, 93
+        scall                   // exit with code fibs(n) & 2^32
+
+// CHECK: [CPU] Program exited with code 1265227608

test/filecheck/hello-world.asm

@@ -0,0 +1,15 @@
+// RUN: python3 -m riscemu -v %s | filecheck %s
+        .data
+msg:    .ascii "Hello world\n"
+        .text
+        addi    a0, zero, 1             // print to stdout
+        addi    a1, zero, msg           // load msg address
+        addi    a2, zero, 12            // write 12 bytes
+        addi    a7, zero, SCALL_WRITE   // write syscall code
+        scall
+        addi    a0, zero, 0             // set exit code to 0
+        addi    a7, zero, SCALL_EXIT    // exit syscall code
+        scall
+
+// CHECK: Hello world
+// CHECK: [CPU] Program exited with code 0

test/filecheck/libc/test-string.s

@@ -0,0 +1,114 @@
+// RUN: python3 -m riscemu -v %s -o libc | filecheck %s
+
+.data
+
+data:
+.byte       0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77, 0x88
+.byte       0x99, 0xAA, 0xBB, 0xCC, 0xDD, 0xEE, 0xFF, 0x00
+
+dest:
+.byte       0xAB, 0xAB, 0xAB, 0xAB, 0xAB, 0xAB, 0xAB, 0xAB
+.byte       0xAB, 0xAB, 0xAB, 0xAB, 0xAB, 0xAB, 0xAB, 0xAB
+
+small_str:
+.string     "test"
+
+.text
+
+.globl      main
+main:
+// test that strlen(data) == 15
+        addi    sp, sp, -4
+        sw      ra, 0(sp)
+        la      a0, data
+        // call strlen(data)
+        jal     strlen
+        li      t0, 15
+        bne     a0, t0, _fail
+
+// now memcpy strlen(data)+1 bytes from data to dest
+        la      a0, dest
+        la      a1, data
+        li      a2, 16
+        // call strncpy(dest, data, 16)
+        jal     strncpy
+        la      a1, dest
+        // fail because strncpy should return pointer to dest
+        bne     a0, a1, _fail
+        // check that dest and data are the same
+        jal     check_data_dest_is_same
+        la      a0, dest
+        li      a1, 0x11
+        li      a2, 16
+// test that memset(dest) workds
+        // call memset(dest, 0x11, 16)
+        jal     memset
+        // check that all of dest is 0x11111111
+        li      t1, 0x11111111
+        la      a0, dest
+        lw      t0, 0(a0)
+        bne     t0, t1, _fail
+        lw      t0, 1(a0)
+        bne     t0, t1, _fail
+        lw      t0, 2(a0)
+        bne     t0, t1, _fail
+        lw      t0, 3(a0)
+        bne     t0, t1, _fail
+// test memchr
+        // test memchr
+        la      a0, data
+        li      a1, 0x55
+        li      a2, 16
+        // memchr(data, 0x55, 16)
+        jal     memchr
+        la      t0, data
+        addi    t0, t0, 4
+        // fail if a0 != data+4
+        bne     a0, t0, _fail
+        la      a0, data
+        li      a1, 0x12
+        li      a2, 16
+        // memchr(data, 0x12, 16)
+        jal     memchr
+        // check that result is NULL
+        bne     a0, zero, _fail
+// test strcpy
+        la      a0, dest
+        la      a1, small_str
+        // call strcpy(dest, small_str)
+        jal     strcpy
+        la      t0, dest
+        lw      t1, 0(a0)
+        // ascii for "tset", as risc-v is little endian
+        li      t2, 0x74736574
+        bne     t1, t2, _fail
+
+// return to exit() wrapper
+        lw      ra, 0(sp)
+        addi    sp, sp, 4
+        li      a0, 0
+        ret
+
+_fail:
+        ebreak
+        // fail the test run
+        li      a0, -1
+        jal     exit
+
+
+check_data_dest_is_same:
+        la      a0, data
+        la      a1, dest
+        li      a2, 4
+1:
+        lw      t0, 0(a0)
+        lw      t1, 0(a1)
+        bne     t0, t1, _fail
+        addi    a0, a0, 4
+        addi    a1, a1, 4
+        addi    a2, a2, -1
+        blt     zero, a2, 1b
+        ret
+
+
+//CHECK: [CPU] Program exited with code 0