@misay 您好，请问这个具体要怎么解决呀，我用yolov5-7.0的模型会有这个问题，6.0没有，但是需要用到7.0的语义分割模型。想请教下如何解决，谢谢大佬

@misay 您好，请问这个要怎么解决呀，我用yolov5-7.0的模型会有这个问题，6.0没有，但是需要用到7.0的语义分割模型。想请教下如何解决，谢谢大佬

@uccccc 解决问题没有？ 我使用TINA自带 范例 测试也不行

encodertest 0 YUYV 88 144 88 144 100 /mnt/exUDISK/1.yuv /m
nt/exUDISK/
[ 4151.530230] VE: enable hw clock
[ 4151.533780] enable_cedar_hw_clk,552
WARNING: cedarc <getSocIdString:1117[ 4151.539203] VE: VE real_freq=576000000
[ 4151.539203]

: cannot open /dev/sunxi_soc_info, it maybe ok!

INFO : cedarc VeInitialize:1307: *** ic_version = 0x1301000010210,
WARNING: cedarc getSocIdString:1117: cannot open /dev/sunxi_soc_info, it maybe ok!

INFO : cedarc VeInitialize:1307: *** ic_version = 0x1301000010210,
sps_pps_data.nLength: 22
the sps_pps :00
the sps_pps :00
the sps_pps :00
the sps_pps :01
the sps_pps :67
the sps_pps :4d
the sps_pps :00
the sps_pps :1f
the sps_pps :96
the sps_pps :54
the sps_pps :30
the sps_pps :9b
the sps_pps :2e
the sps_pps :80
the sps_pps :00
the sps_pps :00
the sps_pps :00
the sps_pps :01
the sps_pps :68
the sps_pps :ee
the sps_pps :31
the sps_pps :12
ERROR : cedarc H264EncFrame:2721: h264 encoder wait interrupt overtime
encode frame 0 use time is 1028981us...
Segmentation fault

@daming123 在 全志T113 接LVDS 屏幕 中说：

用了MQ-r的板子，接上LVDS屏幕
把LCD的引脚引出来，淘宝买了一个40Pin排线，把2.54mm针的距离引出来
557dce4b-cc1d-4993-9716-76aa2010f34c-image.png
然后查看原理图，查看board.dts，查看引脚，接上
由于LVDS屏幕需要一个正负电压，查看规格书，
922be46d-5aab-4c6a-9060-48d2497ee786-image.png
VGH 18V，VGL -10V，手头没有，于是淘宝找了一个LVDS电压转接板，把负电压，正电压，升压得到合适的电压，下面是升级板
a846596c-fc20-40dd-9c82-7c10ee399a53-image.png
接上屏幕，另外一边接转接板。修改dts文件如下
&lcd0 {
lcd_used = <1>;

lcd_driver_name = "default_lcd";
lcd_if = <3>;

lcd_x = <1024>;
lcd_y = <600>;
lcd_width = <150>;
lcd_height = <94>;
lcd_dclk_freq = <51>;

lcd_pwm_used = <0>;

lcd_hbp = <80>;
lcd_ht = <1344>;
lcd_hspw = <20>;
lcd_vbp = <10>;
lcd_vt = <635>;
lcd_vspw = <10>;

lcd_lvds_if = <0>;
lcd_lvds_colordepth = <1>;
lcd_lvds_mode = <1>;

lcd_frm = <1>;
lcd_hv_clk_phase = <0>;
lcd_hv_sync_polarity= <0>;
lcd_gamma_en = <0>;
lcd_bright_curve_en = <0>;
lcd_cmap_en = <0>;

pinctrl-0 = <&lvds0_pins_a>;
pinctrl-1 = <&lvds0_pins_b>;
};
最后上电，成功点亮
e1f09b5a-84cc-4c7a-b4b4-f75da446cefd-image.png
完成，走了很多弯路，因为对LVDS不了解，淘宝买东西耽搁几天时间。大家可以对照着来，如果用到LVDS的话。

亲，有 LVDS 适配步骤没？

但是patch中修改的文件，例如cap_v4l.cpp等
我只能在out中找到他们
0ac1d8f0-227f-4250-be9e-0da2bc8b9b43-image.png

@awwwwa 好的,感谢我试试

@jacky502

&keyboard0 { key0 = <210 0x160>; wakeup-source; status = "okay"; };

注册gpio-keys 绑定键值，或者gpio-keys增加wakeup-source;

D1-H RT-Thread 官方支持

有没有指定gpio唤醒的方式呢

@awwwwa 该问题已解决，但是有新的问题
b39e16d8-fbb8-4171-9bd5-ecd04d6b4216-047b929afa3b8f2b370ee53b0aa032e.jpg
这个色彩是黑白色，换过多个摄像头，都是黑白色

@awwwwa 大佬您好，请问您有修复过这方面的经验吗，求分享😘

@pcj203 在 f1c200s自制 win10 USB显示屏 中说：

简介
使用f1c200s tiny200s货架板+1024*600RGB屏,实现一个win10 USB显示屏。

方案原理：
host: win10加载手撸的IDD显示驱动，编码桌面成jpg，然后usb发送给f1c200s.
dev: f1c200s，接收到完整一帧jpg数据，硬解码yuv,然后软件转成rgb输出给/dev/fb0映射的ram.

b站视频链接
https://b23.tv/UDaOqCa

1.jpg
2.jpg
3.jpg
性能：

黑白模式20pfs. 彩色10fps

已知bug:

似乎f1c200s硬解jpg和usb传输有小概率性冲突。表现就是0.5%的帧解出来会花掉。

开源情况：

开发未完成。预计农历新年github上开源。

参考:
*1.github.com/microsoft/Windows-driver-samples/tree/master/video/IndirectDisplay

*2.git://github.com/roshkins/IddSampleDriver.git

*3.https://whycan.com/t_8114.html baremetal hardware JPEG-decoder example (F1C100S)

*4.https://github.com/aodzip/buildroot-tiny200

esp32 s2 实现win10 usb显示屏320*240
已经github上开源: https://github.com/chuanjinpang/win10_idd_xfz1986_usb_graphic_driver_display
4.jpg

f1c200s自制 win10 USB显示屏最新的github 连接可以给出来吗？

@wj8331585 支持，出现段错误可以编译一个gdb进去看看哪里错了

测试代码：

w = 1920; h = 1080; bpp = 32; SDL_Init( SDL_INIT_EVERYTHING ); SDL_SetVideoMode( w,h, bpp, SDL_HWSURFACE|SDL_DOUBLEBUF );

prebuilt/gcc/linux-x86/arm/toolchain-sunxi-musl

@alb702 add your layer and cat that file see what layer you configured

@jinxiangwzh 我V853 ddr3本来是512M 我换成了同品牌的1024M的 大小变了，我该怎么更改配置？

@alb702 感谢您提供的编码，真的。祝福您。

@nezhastu 大佬你好，请问这个问题解决了吗

@dingdong AIC600E3 is a different chip with t113-s4

MQ Pro接DSI屏，打patch，编译镜像，烧写后开机，屏有背光，但运行lv_example没有显示，cat /dev/urandom > /dev/fb0也没有。查了屏的芯片是ST7701，kernel里也编了驱动。有何解决思路？

@q1215200171 你可太帅了~ 非常感谢！！！

@alb702 好的，非常感谢您的回复，我试试，用示波器抓IRQ信号。

@awwwwa 终于破案了，那个dts里面ledc的配置有个数据被我改过 wait_data_time_ns 这玩意被我改得很小，我那会测试melis的时候32个灯数据，需要刷很快，所以这个数值就不能大，一旦了，数据就错，灯乱闪，我本以为这个数值不影响DMA，但是有影响

我改回到原来那个600000 就一切正常了，DMA能用了，如果这个数值设置太小就完蛋了，只要开启了DMA，系统就直接卡死！那个ledc的手册里面肯定没写全功能！

手册里面写了这么句
wait_data_time_ns：通过该节点可设置和读取 LEDC 内部 FIFO 等待数据的时间容忍度，
范围为 80ns-655us。

我改坏掉的

&ledc { pinctrl-names = "default", "sleep"; pinctrl-0 = <&ledc_pins_a>; pinctrl-1 = <&ledc_pins_b>; led_count = <58>; output_mode = "GRB"; //trans_mode = "DMA"; dmas = <&dma 42>, <&dma 42>; dma-names = "tx", "rx"; reset_ns = <84>; t1h_ns = <800>; t1l_ns = <320>; t0h_ns = <300>; t0l_ns = <800>; wait_time0_ns = <84>; wait_time1_ns = <84>; wait_data_time_ns = <3000>; status = "okay"; };

就是那个 wait_data_time_ns 这个数值在我测试melis的时候被我改过，我以为仅仅是fifo会影响到，但是DMA模式下应该这个参数也有影响

改回来

&ledc { pinctrl-names = "default", "sleep"; pinctrl-0 = <&ledc_pins_a>; pinctrl-1 = <&ledc_pins_b>; led_count = <58>; output_mode = "GRB"; //trans_mode = "DMA"; dmas = <&dma 42>, <&dma 42>; dma-names = "tx", "rx"; reset_ns = <84>; t1h_ns = <800>; t1l_ns = <320>; t0h_ns = <300>; t0l_ns = <800>; wait_time0_ns = <84>; wait_time1_ns = <84>; wait_data_time_ns = <600000>; status = "okay"; };

大佬还有个问题想请教下
就是在tina里面能不能一口气把所有的灯的数据都设置好
然后再发送啊？
如果是
echo 1 > /sys/class/leds/sunxi_led0b/brightness

这样每改一次就要发送一次，效率太低

要同时驱动很多灯的时候 FPS就上不去啦

这要怎么处理呀

@guyuandongfang
我也遇到过类似的情况。 看题目 «V853 SDK：Linux内核中的MIPI DSI «

@sunchipv 问题已解决，是下载的字库有问题，更换另外的字库正常了

@awwwwa 收到，另外问一下大佬，tplayer和视频解码库的日志能关闭吗，这个是在哪里设置?

为了兼容卡打印的问题，UART的引脚在boot0&Uboot中初始化，如不需要卡打印可以配置上需要的引脚

AWOL 支持这些芯片，提供SDK下载

508a68b1-39ed-43e3-93ae-979221e8f2ab-image.png

参考 https://www.aw-ol.com/docs 中的文档下载

其他芯片需要联系你购买商家的客服或者代理获取

更改完需要重新执行

@fuzhikun TINA系统有A133/R818的兼容的

@whycanservice 感谢大佬，改了这个，就能用了

sorry for your spent time. Thank you!

抱歉佔用您的時間。 謝謝你！

https://v853.docs.aw-ol.com/study/study_3getsdktoc/

@fuzhikun A33有Tina3.5，联系代理获取

短接PB2到GND强制进入下载模式

我决定从结论3开始：在uboot中初始化DSI接口和面板。 有用。

@wupaul2001 大哥，你是用Tina系统吗？

@leomini5 密切关注

@lyg116 很有帮助 ..

使用dma模式搬运数据，需要打上如下补丁：

ledc在dma模式的代码存在问题，已修复该问题，ledc的dma通道之前也没有支持，需要添加上

diff --git a/hal/source/dma/platform/dma-sun8iw20.h b/hal/source/dma/platform/dma-sun8iw20.h index 2140014..fc5c72e 100644 --- a/hal/source/dma/platform/dma-sun8iw20.h +++ b/hal/source/dma/platform/dma-sun8iw20.h @@ -136,6 +136,7 @@ #define DRQDST_OTG_EP3 32 #define DRQDST_OTG_EP4 33 #define DRQDST_OTG_EP5 34 +#define DRQDST_LEDC 42 #define DRQDST_TWI0_TX 43 #define DRQDST_TWI1_TX 44 #define DRQDST_TWI2_TX 45 diff --git a/hal/source/ledc/hal_ledc.c b/hal/source/ledc/hal_ledc.c index 5499452..ddf471f 100755 --- a/hal/source/ledc/hal_ledc.c +++ b/hal/source/ledc/hal_ledc.c @@ -20,9 +20,10 @@ #define led_err(fmt, args...) printf("%s()%d - "fmt, __func__, __LINE__, ##args) #define LEDC_PIN_SLEEP 0 +#define LEDC_DMA_BUF_SIZE 4096 struct ledc_config ledc_config = { - .led_count = 3, + .led_count = 1024, .reset_ns = 84, .t1h_ns = 800, .t1l_ns = 450, @@ -37,6 +38,7 @@ static unsigned long base_addr = LEDC_BASE; struct sunxi_dma_chan *dma_chan; struct sunxi_led *led; +static uint8_t already_init; static hal_irqreturn_t sunxi_ledc_irq_handler(void *dummy) { @@ -448,15 +450,16 @@ void hal_ledc_dma_callback(void *para) { - printf("dma callback\n"); + ledc_info("dma transfer end\n"); } -void hal_ledc_trans_data(struct ledc_config *ledc) +int hal_ledc_trans_data(struct ledc_config *ledc) { - int i; + int i, ret; unsigned long int size; unsigned int mask = 0; struct dma_slave_config slave_config; + unsigned int const *buf = ledc->data; mask = LEDC_TRANS_FINISH_INT_EN | LEDC_WAITDATA_TIMEOUT_INT_EN | LEDC_FIFO_OVERFLOW_INT_EN | LEDC_GLOBAL_INT_EN; @@ -480,23 +483,32 @@ ledc_reset_en(); size = ledc->length * 4; + if (size <= LEDC_DMA_BUF_SIZE) { + memcpy(ledc->align_dma_buf, buf, ledc->length); + buf = ledc->align_dma_buf; + } - hal_dcache_clean((unsigned long)ledc->data, sizeof(ledc->data)); + hal_dcache_clean((unsigned long)buf, size); slave_config.direction = DMA_MEM_TO_DEV; - slave_config.src_addr = (unsigned long)(ledc->data); + slave_config.src_addr = (unsigned long)buf; slave_config.dst_addr = (uint32_t)(base_addr + LEDC_DATA_REG); slave_config.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES; slave_config.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES; - slave_config.src_maxburst = DMA_SLAVE_BURST_16; - slave_config.dst_maxburst = DMA_SLAVE_BURST_16; + slave_config.src_maxburst = DMA_SLAVE_BURST_4; + slave_config.dst_maxburst = DMA_SLAVE_BURST_4; slave_config.slave_id = sunxi_slave_id(DRQDST_LEDC, DRQSRC_SDRAM); - hal_dma_slave_config(dma_chan, &slave_config); + ret = hal_dma_slave_config(dma_chan, &slave_config); + if (ret) { + led_err("dma slave config failed\n"); + return -1; + } - hal_dma_prep_device(dma_chan, slave_config.dst_addr, slave_config.src_addr, size, DMA_MEM_TO_DEV); - - //dma_chan->callback = ledc_dma_callback; - hal_dma_start(dma_chan); + ret = hal_dma_prep_device(dma_chan, slave_config.dst_addr, slave_config.src_addr, size, DMA_MEM_TO_DEV); + if (ret) { + led_err("dma prep device failed\n"); + return -1; + } hal_ledc_set_time(ledc); ledc_set_output_mode(ledc->output_mode); @@ -504,7 +516,15 @@ ledc_set_dma_mode(); ledc_enable_irq(mask); ledc_enable(); + + dma_chan->callback = hal_ledc_dma_callback; + ret = hal_dma_start(dma_chan); + if (ret) { + led_err("dma start trans failed\n"); + return -1; + } } + return 0; } void hal_ledc_clear_all_irq(void) @@ -598,6 +618,14 @@ int hal_ledc_init(void) { + if (already_init) { + already_init++; + ledc_info("ledc has been inited, return ok\n"); + return 0; + } + + int ret, i; + ledc_info("hal_led_init\n"); led = malloc(sizeof(struct sunxi_led)); @@ -611,25 +639,40 @@ led->config.data = malloc(sizeof(unsigned int) * led->config.led_count); if (NULL == led->config.data) { led_err("sunxi led config data malloc err\n"); - goto err1; + goto err0; } + for(i = 0;i < led->config.led_count;i++) + led->config.data[i] = 0; if (ledc_clk_init()) { led_err("ledc clk init failed \n"); + goto err1; } if (ledc_pinctrl_init()) { led_err("ledc pinctrl init failed \n"); + goto err2; } - hal_dma_chan_request(&dma_chan); + ret = hal_dma_chan_request(&dma_chan); + if (ret == HAL_DMA_CHAN_STATUS_BUSY) + { + led_err("request dma_chan failed\n"); + goto err3; + } + led->config.align_dma_buf = dma_alloc_coherent(LEDC_DMA_BUF_SIZE); + if (!led->config.align_dma_buf) + { + led_err("alloc dma memory failed\n"); + goto err4; + } if (hal_request_irq(SUNXI_IRQ_LEDC, sunxi_ledc_irq_handler, "ledc", led) < 0) { led_err("ledc request irq failed \n"); - goto errirq; + goto err5; } hal_enable_irq(SUNXI_IRQ_LEDC); @@ -638,13 +681,22 @@ pm_devops_register(&pm_ledc); #endif + already_init++; ledc_info("hal_led_init success\n"); return 0; -errirq: - free(led->config.data); +err5: + dma_free_coherent(led->config.align_dma_buf); +err4: + hal_dma_chan_free(dma_chan); +err3: + ledc_pinctrl_exit(); +err2: + ledc_clk_exit(); err1: + free(led->config.data); +err0: free(led); return -1; @@ -652,27 +704,35 @@ void hal_ledc_deinit(void) { + if (already_init > 0) { + if (--already_init == 0) { #ifdef CONFIG_COMPONENTS_PM - pm_devops_unregister(&pm_ledc); + pm_devops_unregister(&pm_ledc); #endif - hal_disable_irq(SUNXI_IRQ_LEDC); - hal_free_irq(SUNXI_IRQ_LEDC); - hal_dma_chan_free(dma_chan); - ledc_pinctrl_exit(); - ledc_clk_exit(); - free(led->config.data); - free(led); + hal_disable_irq(SUNXI_IRQ_LEDC); + hal_free_irq(SUNXI_IRQ_LEDC); + dma_free_coherent(led->config.align_dma_buf); + hal_dma_chan_free(dma_chan); + ledc_pinctrl_exit(); + ledc_clk_exit(); + free(led->config.data); + free(led); + } + } } int sunxi_set_all_led(unsigned int brightness) { - int i; + int i, ret; led->config.length = led->config.led_count; for(i = 0;i < led->config.led_count;i++) led->config.data[i] = brightness; - hal_ledc_trans_data(&led->config); + ret = hal_ledc_trans_data(&led->config); + if (ret) { + led_err("ledc trans data error\n"); + } return 0; } @@ -680,6 +740,7 @@ int sunxi_set_led_brightness(int led_num, unsigned int brightness) { u32 reg_val; + int i, ret; if (NULL == led) { led_err("err : ledc is not init\n"); @@ -691,10 +752,16 @@ return -1; } - led->config.length = 1; + led->config.length = led_num; led->config.data[led_num-1] = brightness; - hal_ledc_trans_data(&led->config); + for (i = 0; i < led_num; i++) + ledc_info("the %d led light is %u\n", i + 1, led->config.data[i]); + + ret = hal_ledc_trans_data(&led->config); + if (ret) { + led_err("ledc trans data error\n"); + } reg_val = hal_ledc_get_irq_status(); ledc_info("ledc interrupt status reg is %x", reg_val); diff --git a/hal/source/ledc/platform_ledc.h b/hal/source/ledc/platform_ledc.h index 39f6933..2fa9c38 100644 --- a/hal/source/ledc/platform_ledc.h +++ b/hal/source/ledc/platform_ledc.h @@ -33,8 +33,6 @@ #ifndef __PLATFORM_LEDC_H__ #define __PLATFORM_LEDC_H__ -#define DRQDST_LEDC 43 - #if defined(CONFIG_ARCH_SUN8IW18P1) #include "platform/ledc_sun8iw18.h" #endif diff --git a/hal/test/ledc/test_ledc.c b/hal/test/ledc/test_ledc.c index 1ade796..79694c0 100755 --- a/hal/test/ledc/test_ledc.c +++ b/hal/test/ledc/test_ledc.c @@ -20,6 +20,7 @@ int ledc_test(int argc, char **argv) { int brightness = 0; + int led_num; printf("========LEDC TEST========\n"); @@ -32,9 +33,14 @@ return 0; } - brightness = atoi(argv[2]); + led_num = atoi(argv[1]); + if (led_num < 1 || led_num > 1024) + { + printf("The led_num you entered should be between 1 and 1024\n"); + } + brightness = atoi(argv[3]); - switch(argv[1][0]) + switch(argv[2][0]) { case 'R' : brightness <<= 8; break; case 'G' : brightness <<= 16; break; @@ -43,7 +49,8 @@ return -1; } - sunxi_set_led_brightness(1, brightness); + sunxi_set_led_brightness(led_num, brightness); + printf("led is %d\n", led_num); printf("brightness is %d\n", brightness); return 0; diff --git a/include/hal/sunxi_hal_ledc.h b/include/hal/sunxi_hal_ledc.h index a386338..e5a70d1 100644 --- a/include/hal/sunxi_hal_ledc.h +++ b/include/hal/sunxi_hal_ledc.h @@ -43,6 +43,7 @@ unsigned long long wait_time1_ns; unsigned int wait_data_time_ns; char *output_mode; + unsigned int *align_dma_buf; unsigned int *data; unsigned int length; }; @@ -86,7 +87,7 @@ int hal_ledc_init(void); void hal_ledc_deinit(void); -void hal_ledc_trans_data(struct ledc_config *ledc); +int hal_ledc_trans_data(struct ledc_config *ledc); void hal_ledc_clear_all_irq(void); unsigned int hal_ledc_get_irq_status(void); void hal_ledc_dma_callback(void *para);

DMA下 LED 颜色异常

diff --git a/hal/source/ledc/hal_ledc.c b/hal/source/ledc/hal_ledc.c index ddf471f4..8d818662 100755 --- a/hal/source/ledc/hal_ledc.c +++ b/hal/source/ledc/hal_ledc.c @@ -308,11 +308,15 @@ static void ledc_set_wait_data_time_ns(unsigned int wait_data_time_ns) hal_writel(reg_val, base_addr + LEDC_DATA_FINISH_CNT_REG); } +/* + * set the num of leds on the led-strip + * max support up to 1024 leds + */ static void ledc_set_length(unsigned int length) { unsigned int reg_val; - if (length == 0) + if (length == 0 || length > 1024) return; reg_val = hal_readl(base_addr + LEDC_CTRL_REG); @@ -721,12 +725,22 @@ void hal_ledc_deinit(void) } } -int sunxi_set_all_led(unsigned int brightness) +/* + * set the brightness of all the leds in led-strip to a uniform value + * @length: the num of leds on led-strip + * @brightness: the brightness data + */ +int sunxi_set_all_led(unsigned int length, unsigned int brightness) { int i, ret; - led->config.length = led->config.led_count; - for(i = 0;i < led->config.led_count;i++) + if (length > led->config.led_count) { + led_err("%d: max support 1024 leds\n", length); + return -1; + } + + led->config.length = length; + for(i = 0;i < led->length;i++) led->config.data[i] = brightness; ret = hal_ledc_trans_data(&led->config); @@ -737,7 +751,13 @@ int sunxi_set_all_led(unsigned int brightness) return 0; } -int sunxi_set_led_brightness(int led_num, unsigned int brightness) +/* + * set the brightness of each led on the led strip + * @length: all the num of leds on the led strip + * @led_num: the specified led that you want to set + * @brightness: the led brightness data + */ +int sunxi_set_led_brightness(unsigned int length, unsigned int led_num, unsigned int brightness) { u32 reg_val; int i, ret; @@ -747,12 +767,13 @@ int sunxi_set_led_brightness(int led_num, unsigned int brightness) return -1; } - if (led_num > led->config.led_count) { + if (length > led->config.led_count || len_num > length) { led_err("has not the %d led\n", led_num); return -1; } - led->config.length = led_num; + led->config.length = length; + /* set the specified led brightness, others set default brightness: 0x0*/ led->config.data[led_num-1] = brightness; for (i = 0; i < led_num; i++)

解决了，主要是晶振搞错了。要焊接40M的 有源晶振。

@awwwwa Hello, I'm using a TinyVision Board, follwing the steps from here: https://tinyvision.yuzukihd.top/#/?id=下载-awol-tina-linux-bsp selecting V851se-tinyvision-tina

82fd08ae-1143-416b-8129-f52123d6ba93-image.png

After flashing the image on the SD card this is the result:

db3aad59-7aba-47cc-a4e3-e52716f5c1e1-image.png

But if I flash the SD Card with the pre-build image downloaded from the github repository, it works allways.

9466cd6a-80da-4b06-a5b0-c1f7eaf2b676-image.png

Please help to find a solution,

Thanks for your answer!

重复帖：D1s ledc驱动代码bug，DMA模式无法使用

https://bbs.aw-ol.com/topic/5071/share/1

@chen0505 PA引导是BROM不支持，BROM是固化在芯片里的，如果需要修改需要改版芯片

Note Description D/S Datasheet Support but no sample test S/T Support and sample test √ Support and mass produce X Not support (Blank) Can support but not support now Part Number Capacity JEDEC ID PACKAGE R128 Support Status MT29F2G01ABAGDSF-AAT(NW794) 256M 0x2C, 0x24 SOP-16 D/S MT29F1G01ABAFDSF-AAT(NW808) 128M 0x2C, 0x14 SOP-16 D/S MT29F2G01ABAGDWB 256M 0x2C,0x24 SOP-16 D/S W25N01GVZEIG 128M 0xEF 0xAA 0x21 D/S W25N02KVZEIR 256M 0xEF 0xAA 0x22 MX35LF2GE4AB 256M 0xC2 0x22 16-SOP MX35LF1GE4AB 128M 0xC2 0x12 8-WSON √ MX35LF2GE4AD 256M 0xC2 0x26 0x03 8-WSON D/S MX35LF1G24AD 128M 0xC2 0x14 MX35LF4GE4AD 512M GD5F1GQ4UBYIGR 128M 0xC8 0xD1 8-pad 8X6 WSON D/S GD5F2GQ4UB9IGR 256M 0xC8 0xD2 8-pad 8X6 WSON D/S GD5F1GQ4UCYIG 128M 0xC8 0xB1 0x48 WSON8(8x6mm) D/S GD5F1GQ5UEYIG 128M 0xC8 0x51 D/S GD5F2GQ5UEYIGR 256M 0xC8 0x52 GD5F4GQ6UEYIG 512M 0xC8 0x55 F50L1G41LB(2M) -> F50L1G41LB-104YG2M F50L1G41LB-104YG2ME 128M 0xC8 0x01 0x7F 0x7F 0x7F 8-WSON D/S F50L2G41XA 256M 0x2C, 0x24 DS35Q1GA-1B 128M 0xE5 0x71 8-WSON D/S DS35Q2GA-IB 256M 0xE5 0x72 DS35Q2GBXXX 256M 0xE5 0xf2 DS35Q4GMXXX 512M 0xE5 0xf4 DS35X1GBXXX 1G 0xE5 0xF1 8-WSON TC58CVG2S0H 512M 0x98 0xCD TC58CVG0S3HRAIG 128M 0x98 0xC2 8-WSON D/S FS35ND01G-S1F1QWFI 128M 0xCD 0xB1 WSON8(8x6mm) D/S FS35ND01G-S1Y2 128M 0xCD 0xEA WSON8 F35SQA001G 128M 0xCD 0x71 WSON8 √ F35SQA002G 256M 0xCD 0x72 EM73C044VCD-H 128M 0xD5 0x1C 8-WSON EM73E044SNA 128M 0xD5 0x03 8-WSON ZD35Q1GA-IBR 128M 0xba 0x71 8-WSON D/S PN26G01AWSIUG 128M 0xa1, 0xe1 WSON8(8x6mm) XT26G01A 128M 0x0b,0xe1 D/S XT26G02A 256M 0x0b,0xe2 XT26G04A 512M 0x0b,0xe3 WSON8(8x6mm) XT26G01CWSIG 128M 0x0b,0x11 D/S XT26G02CWSIG 256M 0x0b,0x12 XT26G04CWSIG 512M 0x0b,0x13 XT26G02ELGIGA 256M 0x2c, 0x24 FM25S01 128M 0xa1,0xa1 WSON8(8x6mm) FM25S01A 128M 0xa1,0xe4 FM25S02A 256M 0xa1,0xe5 SCN00SA3W1AI8A 128M 0xC8 0x21 8-WSON S35ML01G3 128M 0x01 0x15 S35ML02G300WH100 256M 0x01 0x25 S35ML04G3 512M 0x01 0x35 HYF1GQ4UDACAE 128M 0xC9 0x21

@amturing 在 D1s的芯片手册 中说：

希望能帮到你
F133_User_Manual_V1.0.pdf
F133_Datasheet_V1.0_whycan.pdf

请问这个f133的手册也能用于D1么？

@leomini5 VE是编解码器

@xiaowenge 设备树可以配置默认的背光亮度

问题已解决，感谢大佬

@dort91011 佬，你这个调通了吗？想找你交流一下416567819

@liefyuan 我也遇到类似情况了，您emmc用的sdc2吗？能看看这部分原理图吗和设备树配置吗？不知道是不是我硬件搞得有问题了

@proc_c

执行这三条命令之后再make看看，应该能解决问题：

rm -rf out rm .config make distclean

之后执行 make 编译。

@whycan 晕哥 这个项目开源嘛 有教程嘛？我想用V853尝试一下

用这个试试
opkg install <安装包路径>

@lovehex99 啊嘻，这……哈哈哈原来是这样，

@awwwwa 感谢您的链接，这是一个有趣的解决方案，非常接近我想要实现的目标，我相信这使用了专有的闭源 tina-os，对吧？ 但由于这只是一个磁盘映像而不是开源的，我无法真正使用它，特别是因为它是用于 V851SE 而不是 V851S...我想在芯片上实现我自己的驱动程序和图像处理方法，从什么 我知道这需要 NDA 下的专有 sdk...

可以看看DongshanPI-D1s开发板，这个板子集成CKLink调试器

@whycanservice 应该是代理商的技术给的。后面给我我补丁，我打上了。重新制作了镜像 t507_h_linux_demo2.0_uart0.img ，烧录到TF卡，还是不能正常启动。

[108]HELLO! BOOT0 is starting! [111]BOOT0 commit : 100e431 [114]set pll start [116]periph0 has been enabled [119]set pll end [121]PMU: AXP858 [132]vaild para:1 select dram para0 [135]board init ok [137]debug:rtc h-m-s value:0x0 [140]DRAM BOOT DRIVE INFO: V0.648 [144]the chip id is 0x7c00 [146]chip id check OK [150]DRAM_VCC set to 1100 mv [158]read_calibration error [166]read_calibration error [174]read_calibration error [183]read_calibration error [191]read_calibration error [199]read_calibration error [207]read_calibration error [215]read_calibration error [223]read_calibration error [232]read_calibration error [234]retraining final error [245][AUTO DEBUG]32bit,1 ranks training success! [261]DRAM CLK =720 MHZ [263]DRAM Type =8 (3:DDR3,4:DDR4,7:LPDDR3,8:LPDDR4) [273]Actual DRAM SIZE =8192 M [276]DRAM SIZE =8192 MBytes, para1 = 311b, para2 = 20000000, dram_tpr13 = 6061 [291]Actual DRAM SIZE =8192 M [296]DRAM simple test FAIL-----fedcbe98 != fedcba98 at address 40000000

请问 [296]DRAM simple test FAIL-----fedcbe98 != fedcba98 at address 40000000 这个大概是什么原因？我可以从tf卡启动吗？还是说必须要从emmc启动？

@bubble 嗯，想问下R128 QSPI的例子是哪里找到的？看SDK好像没有~

@myamoroso 在 T113S3 无法完全启动 中说：

I2C: FDT ERROR:fdt_set_all_pin:[twi0]-->FDT_ERR_BADPATH
FDT ERROR:fdt_set_all_pin:[twi1]-->FDT_ERR_BADPATH

检查一下是不是分区空间不够了

target/allwinner/t113-xxxxx/busybox-init-base-files/etc/init.d/rc.final

末尾加上
ifconfig eth0 up
udhcpc -i eth0

@yuzukitsuru
但是adb上看来会死掉（是整个linux挂掉adb断开，需要重新上电才行），回头我看一眼串口输出，

使用以下指令构建
mkfs.ubifs -x zlib -m 4096 -e 258048 -c 245 -F -r oem_res -o oem_ubifs.fex

@a06041114 或者编写这些命令的脚本，然后网上搜linux自启动加载脚本

@q1215200171 非常感谢

@unkown 解决了，模块加载没放初始化那里，加载的太晚导致里有些东西没起来。

@awwwwa 感谢🙏，哈哈哈，搞定了颜色正常了

全志在线只提供这些芯片方案的 SDK 下载，其他请联系代理/FAE

ed6b4f0d-e497-4766-ac58-7fbad0a9d458-image.png

看着不错啊，另外度盘链接失效了，求补

@zhangjun 分区表里删了这个分区

晶振这个地方，确实容易出问题，新手尤其注意选购的晶振是有源还是无源。

Ubuntu 22.04 / 20.04
更新软件源，更新系统软件包

sudo apt-get update
sudo apt-get upgrade -y
安装开发依赖

sudo apt-get install build-essential subversion git libncurses5-dev zlib1g-dev gawk flex bison quilt libssl-dev xsltproc libxml-parser-perl mercurial bzr ecj cvs unzip lsof
安装相关工具

sudo apt-get install kconfig-frontends android-tools-mkbootimg python2 libpython3-dev
增加架构支持

sudo dpkg --add-architecture i386
sudo apt-get update
安装支持包

sudo apt install gcc-multilib
sudo apt install libc6:i386 libstdc++6:i386 lib32z1
Ubuntu 18.04
更新软件源，更新系统软件包

sudo apt-get update
sudo apt-get upgrade -y
安装开发依赖

sudo apt-get install build-essential subversion git libncurses5-dev zlib1g-dev gawk flex bison quilt libssl-dev xsltproc libxml-parser-perl mercurial bzr ecj cvs unzip lsof
安装相关工具

sudo apt-get install android-tools-mkbootimg libpython3-dev
增加架构支持

sudo dpkg --add-architecture i386
sudo apt-get update
安装支持包

sudo apt install gcc-multilib
sudo apt install libc6:i386 libstdc++6:i386 lib32z1
Arch Linux / Manjaro
更新软件源，更新系统软件包

pacman -Syyuu
安装开发依赖

pacman -S --needed base-devel autoconf automake bash binutils bison bzip2 fakeroot file findutils flex gawk gcc gettext git grep groff gzip time unzip util-linux wget which zlib asciidoc help2man intltool perl-extutils-makemaker swig
安装相关工具

pacman -S --needed libelf libtool libxslt m4 make ncurses openssl patch pkgconf python rsync sed texinfo
增加架构支持

pacman -S --needed multilib-devel
CentOS / Fedora / openEuler

sudo dnf --setopt install_weak_deps=False --skip-broken install bash-completion bzip2 gcc gcc-c++ git make ncurses-devel patch rsync tar unzip wget which diffutils python2 python3 perl-base perl-Data-Dumper perl-File-Compare perl-File-Copy perl-FindBin perl-Thread-Queue glibc.i686
openSUSE

sudo zypper install --no-recommends asciidoc bash bc binutils bzip2 fastjar flex gawk gcc