Use uncached ExtRAM access (0xBC400000) to bypass CPU data cache

The open-la500 core has USE_CACHE enabled. Writes to 0x1c400000
go to the write-back data cache and never reach the SRAM model
before the testbench checks ExtRAM. Use the 0xA0000000 uncached
segment (same pattern as fft_dma/main.c).

sdk/software/examples/dma/Makefile

@@ -0,0 +1,11 @@
+TARGET = dma
+
+CFLAGS += -O3 -g
+
+C_SRCS := $(wildcard ./*.c )
+
+OBJDIR = obj
+COMMON_DIR = ../../bsp
+GCC_DIR=../../../toolchains/loongson-gnu-toolchain-8.3-x86_64-loongarch32r-linux-gnusf-v2.0
+PICOLIBC_DIR=../../../toolchains/picolibc
+include ../../bsp/common.mk

sdk/software/examples/dma/main.c

@@ -0,0 +1,71 @@
+#include <stdio.h> 
+#include <string.h>
+#include <math.h>
+#include <stdbool.h>
+#include <stdint.h>
+
+#include <common_func.h>
+#include <confreg_time.h>
+#include <dma.h>
+
+// BSP板级支持包所需全局变量
+unsigned long UART_BASE = 0xbf000000;
+unsigned long CONFREG_TIMER_BASE = 0xbf20f100;
+unsigned long CONFREG_CLOCKS_PER_SEC = 50000000L;
+unsigned long CORE_CLOCKS_PER_SEC = 33000000L;
+
+// 加上 aligned(64) 是为了防止 DMA 突发传输时跨越缓存行或 AXI 非对齐边界
+uint32_t src_array[64] __attribute__((aligned(64)));
+uint32_t dst_array[64] __attribute__((aligned(64)));
+
+int main(int argc, char** argv)
+{
+    uint32_t data_len = 64;
+    uint32_t byte_len = data_len * 4;
+
+    // 虚拟地址转物理地址与无缓存地址
+    // 屏蔽掉高 3 位（& 0x1FFFFFFF）即可获得真实的物理地址
+    uint32_t phys_src_addr = ((uint32_t)src_array) & 0x1FFFFFFF;
+    uint32_t phys_dst_addr = ((uint32_t)dst_array) & 0x1FFFFFFF;
+
+    // 或上 0xA0000000 即可获得绕过 Cache 的无缓存（Uncached）指针
+    volatile uint32_t *uncached_src = (volatile uint32_t *)(phys_src_addr | 0xA0000000);
+    volatile uint32_t *uncached_dst = (volatile uint32_t *)(phys_dst_addr | 0xA0000000);
+
+    // CPU 必须通过无缓存指针写入
+    for (int i = 0; i < data_len; i++) {
+        uncached_src[i] = i * i;        // 填入 0~63
+        uncached_dst[i] = 0xDEADBEEF; // 目的数组填入垃圾值，方便验证是否真的被覆盖
+    }
+
+    printf("src_array: %x\n", src_array);
+    printf("dst_array: %x\n", dst_array);
+    printf("phys_src_array: %x\n", phys_src_addr);
+    printf("phys_dst_array: %x\n", phys_dst_addr);
+    printf("uncached_src: %x\n", uncached_src);
+    printf("uncached_dst: %x\n", uncached_dst);
+
+    // 配置 DMA：DMA 只需要最纯粹的物理地址
+    dma_start_transfer(0, phys_src_addr, phys_dst_addr, byte_len);
+    dma_wait_polling(0);
+
+    printf("dma passed!\n");
+
+    // 校验数据：CPU 必须通过无缓存指针读取
+    int err = 0;
+    for (int i = 0; i < data_len; i++) {
+        uint32_t val = uncached_dst[i];
+        printf("%d: %d\n", i, val);
+        if (val != i * i) {
+            err++;
+        }
+    }
+
+    if (err == 0) {
+        printf("\nSuccess! Array to Array DMA transfer verified.\n");
+    } else {
+        printf("\nFailed! Found %d errors.\n", err);
+    }
+
+    return 0;
+}

sdk/software/examples/dma/result.txt

@@ -0,0 +1,75 @@
+dma_status init = 0
+src_array: 1c080d80
+dst_array: 1c080c80
+phys_src_array: 1c080d80
+phys_dst_array: 1c080c80
+uncached_src: bc080d80
+uncached_dst: bc080c80
+polling...
+dma passed!
+0: 0
+1: 1
+2: 4
+3: 9
+4: 16
+5: 25
+6: 36
+7: 49
+8: 64
+9: 81
+10: 100
+11: 121
+12: 144
+13: 169
+14: 196
+15: 225
+16: 256
+17: 289
+18: 324
+19: 361
+20: 400
+21: 441
+22: 484
+23: 529
+24: 576
+25: 625
+26: 676
+27: 729
+28: 784
+29: 841
+30: 900
+31: 961
+32: 1024
+33: 1089
+34: 1156
+35: 1225
+36: 1296
+37: 1369
+38: 1444
+39: 1521
+40: 1600
+41: 1681
+42: 1764
+43: 1849
+44: 1936
+45: 2025
+46: 2116
+47: 2209
+48: 2304
+49: 2401
+50: 2500
+51: 2601
+52: 2704
+53: 2809
+54: 2916
+55: 3025
+56: 3136
+57: 3249
+58: 3364
+59: 3481
+60: 3600
+61: 3721
+62: 3844
+63: 3969
+
+Success! Array to Array DMA transfer verified.