Linux LCD驱动中framebuffer的研究-优快云博客

Linux 本身实现了FrameBuffer驱动（字符驱动）便于应用用于层调用，实现的文件是/drivers/video/fbmem.c文件：

static const struct file_operations fb_fops = {
   .owner =   THIS_MODULE,
   .read =       fb_read,
   .write =   fb_write,
   .unlocked_ioctl = fb_ioctl,
#ifdef CONFIG_COMPAT
   .compat_ioctl = fb_compat_ioctl,
#endif
   .mmap =       fb_mmap,
   .open =       fb_open,
   .release =   fb_release,
#ifdef HAVE_ARCH_FB_UNMAPPED_AREA
   .get_unmapped_area = get_fb_unmapped_area,
#endif
#ifdef CONFIG_FB_DEFERRED_IO
   .fsync =   fb_deferred_io_fsync,
#endif
};

上层应用程序通过这些函数实现对lcd各种参数的控制,修改等,这与以前编写字符设备驱动的模式一样,这些函数都是linux内核已经实现好的,不需要驱动工程师去编写.那驱动工程师的任务是什么呢?获取平台设备的相关信息,并初始化相关的硬件设备,注册中断,开辟帧缓冲内存,注册帧缓冲设备,这就是/drivers/video/s3c2440fb.c中的主要内容了.里面有一个重要的数据结构：

static struct fb_ops s3c2410fb_ops = {
   .owner       = THIS_MODULE,
   .fb_check_var   = s3c2410fb_check_var,
   .fb_set_par   = s3c2410fb_set_par,
   .fb_blank   = s3c2410fb_blank,
   .fb_setcolreg   = s3c2410fb_setcolreg,
   .fb_fillrect   = cfb_fillrect,
   .fb_copyarea   = cfb_copyarea,
   .fb_imageblit   = cfb_imageblit,
};

这里的函数是真正对底层硬件进行操作的函数了,都是需要驱动工程师根据具体的硬件平台去填写.那么这些底层函数与上层应用程序的接口是怎样对接起来的呢?其实,是通过表征帧缓存设备的数据结构struct fb_info来实现的,这个数据结构几乎包括了lcd控制器的所有信息:

struct fb_info {
        int                      node;
        int                      flags;
        struct fb_var_screeninfo var;
        struct fb_fix_screeninfo fix;
        struct fb_monspecs       monspecs;
        struct work_struct       queue;
        struct fb_pixmap         pixmap;
        struct fb_pixmap         sprite;