mohitmishra786/low-level-dev-skills

assembly-riscv

RISC-V assembly skill for RV32/RV64 programming.

查看源码

使用方法

  1. 复制安装命令,在自己的智能体环境安装前先检查源码包。
  2. 从原始文档核对所需工具、凭据和支持的宿主。
  3. 说明输入材料与期望交付物,并对照来源检查输出。
npx skills add https://github.com/mohitmishra786/low-level-dev-skills --skill assembly-riscv

开始前需了解

  • 安装不包含模型订阅、API 额度或访问文件的权限。
  • 仓库星数和安装量反映活跃情况,不等于独立质量或安全测试。
仓库原始内容

按源仓库内容呈现,保留标题、案例、代码、表格、链接以及原文引用的演示图片。

RISC-V Assembly

Purpose

Guide agents through RISC-V assembly programming: RV32/RV64 instruction sets, register naming and calling conventions (psABI), ISA extension naming, inline assembly with GCC/Clang, compressed (RVC) instructions, and QEMU-based simulation with GDB remote debugging.

Triggers

  • "How do I write RISC-V assembly?"
  • "What are the RISC-V calling convention registers?"
  • "How do I use inline asm for RISC-V in C?"
  • "What do RISC-V extension letters mean (IMAFD)?"
  • "How do I simulate RISC-V with QEMU?"
  • "How do I debug RISC-V code with GDB?"

Workflow

1. Register file and calling convention

RISC-V has 32 integer registers (x0–x31) with ABI names:

RegisterABI nameRoleSaved by
x0zeroHard-wired zero—
x1raReturn addressCaller
x2spStack pointerCallee
x3gpGlobal pointer—
x4tpThread pointer—
x5–x7t0–t2TemporariesCaller
x8s0/fpFrame pointerCallee
x9s1Saved registerCallee
x10–x11a0–a1Arguments / return valuesCaller
x12–x17a2–a7ArgumentsCaller
x18–x27s2–s11Saved registersCallee
x28–x31t3–t6TemporariesCaller

Floating-point registers (F extension): f0–f31 (fa0–fa7 for arguments).

2. Basic instructions

asm
# Arithmetic (R and I type)
add   a0, a1, a2      # a0 = a1 + a2
sub   a0, a1, a2      # a0 = a1 - a2
addi  a0, a1, 42      # a0 = a1 + 42 (immediate)
mul   a0, a1, a2      # a0 = a1 * a2 (M extension)
div   a0, a1, a2      # signed divide (M extension)
rem   a0, a1, a2      # remainder (M extension)

# Logical
and   a0, a1, a2      # bitwise AND
or    a0, a1, a2      # bitwise OR
xor   a0, a1, a2      # bitwise XOR
sll   a0, a1, a2      # shift left logical
srl   a0, a1, a2      # shift right logical (unsigned)
sra   a0, a1, a2      # shift right arithmetic (signed)

# Load / store
lw    a0, 0(sp)       # load word (32-bit)
ld    a0, 0(sp)       # load doubleword (64-bit, RV64)
lh    a0, 4(sp)       # load halfword (sign-extended)
lbu   a0, 8(sp)       # load byte (zero-extended)
sw    a0, 0(sp)       # store word
sd    a0, 0(sp)       # store doubleword (RV64)

# Branches (compare and branch)
beq   a0, a1, label   # branch if equal
bne   a0, a1, label   # branch if not equal
blt   a0, a1, label   # branch if less than (signed)
bltu  a0, a1, label   # branch if less than (unsigned)
bge   a0, a1, label   # branch if ≥ (signed)

# Jumps
j     label           # unconditional jump (pseudoinstruction: jal x0, label)
jal   ra, func        # jump and link (call)
jalr  zero, ra, 0     # jump to ra (return: pseudoinstruction: ret)

3. Minimal function (psABI calling convention)

asm
.section .text
.global add_numbers
# int add_numbers(int a, int b);  — a in a0, b in a1, return in a0
add_numbers:
    add   a0, a0, a1   # result = a + b
    ret                # return (jalr zero, ra, 0)

.global factorial
# long factorial(int n);  — n in a0
factorial:
    addi  sp, sp, -16      # allocate stack frame
    sd    ra, 8(sp)        # save return address (RV64)
    sd    s0, 0(sp)        # save s0 (callee-saved)

    mv    s0, a0           # s0 = n
    li    a0, 1            # default return 1
    blez  s0, .done        # if n <= 0, return 1

    addi  a0, s0, -1      # a0 = n - 1
    call  factorial        # recursive call: factorial(n-1)
    mul   a0, a0, s0       # a0 = result * n

.done:
    ld    ra, 8(sp)        # restore ra
    ld    s0, 0(sp)        # restore s0
    addi  sp, sp, 16       # deallocate
    ret

4. ISA extension naming

RISC-V extensions are combined as a string after the base ISA:

LetterExtensionDescription
IIntegerBase 32/64-bit integer (RV32I, RV64I)
MMultiplyInteger multiply and divide
AAtomicAtomic memory operations (lr/sc, AMOs)
FFloatSingle-precision float
DDoubleDouble-precision float
CCompressed16-bit compressed instructions
GGeneral= IMAFD (shorthand)
VVectorVector instructions (SIMD)
ZicsrCSRControl/status register access
ZifenceiFence.iInstruction-fetch fence
Zba/Zbb/Zbc/ZbsBit manipulationBit ops (B extension set)
ZtsoTSOTotal Store Ordering memory model

Common targets:

  • Embedded: rv32imac — no floating point, with atomics and compressed
  • Linux app: rv64gc — full general + compressed
  • High performance: rv64gcv — + vector

5. Inline assembly (GCC/Clang)

c
// Read a CSR register (e.g., cycle counter)
static inline uint64_t read_cycle(void) {
    uint64_t val;
    asm volatile ("rdcycle %0" : "=r"(val));
    return val;
}

// Atomic swap
static inline int atomic_swap(int *ptr, int new_val) {
    int old;
    asm volatile (
        "amoswap.w.aqrl %0, %2, (%1)"
        : "=r"(old)
        : "r"(ptr), "r"(new_val)
        : "memory"
    );
    return old;
}

// Memory fence
static inline void memory_fence(void) {
    asm volatile ("fence rw, rw" ::: "memory");
}

// CSR read/write
#define csr_read(csr) ({                    \
    uint64_t _v;                            \
    asm volatile ("csrr %0, " #csr : "=r"(_v)); \
    _v;                                     \
})

uint64_t mstatus = csr_read(mstatus);

6. Compressed instructions (RVC)

RVC replaces common 32-bit instructions with 16-bit versions when:

  • Register is in x8–x15 (for c. versions)
  • Immediate fits in smaller field
  • Specific instruction patterns match
bash
# Enable C extension in GCC
riscv64-linux-gnu-gcc -march=rv64gc prog.c -o prog

# Check if compressed instructions were generated
riscv64-linux-gnu-objdump -d prog | grep "c\."
# c.addi, c.ld, c.sw, c.j, etc.

# Disable compressed (for debugging or targets without C)
riscv64-linux-gnu-gcc -march=rv64g prog.c -o prog

7. QEMU simulation and GDB

bash
# Install QEMU RISC-V
apt-get install qemu-user qemu-system-riscv64

# User-mode emulation (run RV64 binary on x86 host)
qemu-riscv64 ./prog

# System emulation (full bare-metal VM)
qemu-system-riscv64 \
  -machine virt \
  -nographic \
  -kernel firmware.elf \
  -gdb tcp::1234 \
  -S     # start paused

# GDB remote session
riscv64-linux-gnu-gdb prog
(gdb) target remote :1234
(gdb) load
(gdb) break main
(gdb) continue

For the RISC-V psABI calling convention details, see references/riscv-abi.md.

Related skills

  • Use skills/low-level-programming/assembly-arm for AArch64 comparison
  • Use skills/low-level-programming/assembly-x86 for x86-64 assembly
  • Use skills/embedded/openocd-jtag for real hardware RISC-V debugging
  • Use skills/compilers/cross-gcc for RISC-V cross-compilation setup

按任务相关推荐

来自同一仓库

更多 Skills

全部 Skills
mohitmishra786
社区

low-power-embedded

Low-power embedded skill for sleep modes and energy optimization. Use when configuring MCU sleep/stop/standby, peripheral clock gating, wake-up sources, or measuring firmware current draw. Activates on queries about WFI, STOP mode, STM32 PWR, nRF sleep, wake-up EXTI, or reducing embedded power consumption.

1小时安装量
1
GitHub Stars
296
最近更新
11月20日
mohitmishra786
社区

core-dumps

Core dump analysis skill for production crash triage. Use when loading core files in GDB or LLDB, enabling core dump generation on Linux/macOS, mapping symbols with debuginfo or debuginfod, or extracting backtraces from crashes without re-running the program. Activates on queries about core files, ulimit, coredumpctl, debuginfod, crash triage, or analyzing segfaults from production binaries.

1小时安装量
2
GitHub Stars
267
最近更新
6月27日
mohitmishra786
社区

custom-allocators

Custom allocator skill for memory allocation strategies. Use when implementing pool/slab/arena allocators, tuning jemalloc/mimalloc, writing Rust GlobalAlloc, or benchmarking allocator performance. Activates on queries about jemalloc, mimalloc, tcmalloc, arena allocator, GlobalAlloc, or fragmentation.

1小时安装量
2
GitHub Stars
267
最近更新
6月27日
mohitmishra786
社区

ebpf

eBPF skill for Linux observability and networking. Use when writing eBPF programs with libbpf or bpftrace, attaching kprobes/tracepoints/XDP hooks, debugging verifier errors, working with eBPF maps, or achieving CO-RE portability across kernel versions. Activates on queries about eBPF, bpftool, bpftrace, XDP programs, libbpf, verifier errors, eBPF maps, or kernel tracing with BPF.

1小时安装量
2
GitHub Stars
267
最近更新
6月27日