Re: [Linux-arm-nuc900] Re: [PATCH]NUC900 LCD Controller Driver

在 2010年1月30日 上午8:27，Andrew Morton [off-list ref] 写道：

On Mon, 11 Jan 2010 17:05:44 +0900
Wang Qiang [off-list ref] wrote:

quoted

hi, Dear Wan

There is the patch of LCD controller driver for nuc900s.
The Linux LOGO is just fine and the FB-Test application was ok, too.

best regards
wangqiang


...

@@ -380,6 +381,48 @@ struct platform_device nuc900_device_kpi = {

      .resource       = nuc900_kpi_resource,
 };

+#ifdef CONFIG_FB_NUC900
+
+static struct resource nuc900_lcd_resource[] = {
+     [0] = {
+             .start = W90X900_PA_LCD,
+             .end   = W90X900_PA_LCD + W90X900_SZ_LCD - 1,
+             .flags = IORESOURCE_MEM,
+     },
+     [1] = {
+             .start = IRQ_LCD,
+             .end   = IRQ_LCD,
+             .flags = IORESOURCE_IRQ,
+     }
+};
+
+static u64 nuc900_device_lcd_dmamask = 0xffffffffUL;

I suspect this should have type `dma_addr_t', but `struct device' uses
open-coded u64 too.  Odd.

It makes no sense to initialise a u64 with an unsigned long value -
it's wrong on a 32-bit machine.

this:

       static u64 nuc900_device_lcd_dmamask = -1;

will work.

quoted

+struct platform_device nuc900_device_lcd = {
+     .name             = "nuc900-lcd",
+     .id               = -1,
+     .num_resources    = ARRAY_SIZE(nuc900_lcd_resource),
+     .resource         = nuc900_lcd_resource,
+     .dev              = {
+             .dma_mask               = &nuc900_device_lcd_dmamask,
+             .coherent_dma_mask      = 0xffffffffUL

And this gets initialised to 0x00000000ffffffff also.  Using -1 will fix.

quoted

+     }
+};
+

...

+ */
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/mm.h>
+#include <linux/tty.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/fb.h>
+#include <linux/init.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/workqueue.h>
+#include <linux/wait.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/cpufreq.h>
+#include <linux/io.h>
+
+#ifdef CONFIG_PM

Is this ifdef needed?

quoted

+#include <linux/pm.h>
+#endif
+
+#include <mach/map.h>
+#include <mach/regs-clock.h>
+#include <mach/regs-ldm.h>
+#include <mach/fb.h>
+#include <mach/clkdev.h>
+
+#include "nuc900fb.h"
+
+/* Debugging stuff */
+#ifdef CONFIG_FB_NUC900_DEBUG
+static int debug     = 1;
+#define dprintk(msg...)      { printk(KERN_DEBUG "nuc900 lcd: " msg); }
+#else
+static int debug;
+#define dprintk(msg...)
+#endif

Would be better to use the standard pr_debug(), rather than inventing
private infrastructure.  That way you get to enjoy the facilities which
lib/dynamic_debug.c offers.

quoted

+
+/*
+ *  Initialize the nuc900 video (dual) buffer address
+ */
+static void nuc900fb_set_lcdaddr(struct fb_info *info)
+{
+     struct nuc900fb_info *fbi = info->par;
+     void __iomem *regs = fbi->io;
+     unsigned long vbaddr1, vbaddr2;
+     vbaddr1  = info->fix.smem_start;

It's conventional to put a blank line beterrn end-of-locals and start-of-code.

quoted

+     vbaddr2  = info->fix.smem_start;
+     vbaddr2 += info->fix.line_length * info->var.yres;
+     /*set frambuffer start phy addr*/

Like this:

       /* set frambuffer start phy addr */

quoted

+     writel(vbaddr1, regs + REG_LCM_VA_BADDR0);
+     writel(vbaddr2, regs + REG_LCM_VA_BADDR1);
+
+     writel(fbi->regs.lcd_va_fbctrl, regs + REG_LCM_VA_FBCTRL);
+     writel(fbi->regs.lcd_va_scale, regs + REG_LCM_VA_SCALE);
+}
+
+//static void nuc900fb_clk_select(struct fb_info *info, unsigned long pixclk)
+//{
+//   struct nuc900fb_info *fbi = info->par;
+//   int div;
+//   if (clk <= 15 * 1000 * 1000) {
+//           div = (15 * 1000 * 1000)/
+//           nuc900_driver_clksrc_div(fbi->dev, "ext", 0x2);
+//   }
+//
+//}

Please feed the patch through scrupts/checkpatch.pl.

quoted

+/*
+ *   calculate divider for lcd div
+ */
+static unsigned int nuc900fb_calc_pixclk(struct nuc900fb_info *fbi,
+                                      unsigned long pixclk)
+{
+     unsigned long clk = fbi->clk_rate;
+     unsigned long long div;
+
+     /* pixclk is in picseconds. our clock is in Hz*/
+     /* div = (clk * pixclk)/10^12 */
+     div = (unsigned long long)clk * pixclk;
+     div >>= 12;
+     do_div(div, 625 * 625UL * 625);
+
+     dprintk("pixclk %ld, divisor is %ld\n", pixclk, (long)div);

Perhaps use %lld and remove the cast.

quoted

+     return div;
+}
+
+/*
+ *   Check the video params of 'var'.
+ */
+static int nuc900fb_check_var(struct fb_var_screeninfo *var,
+                            struct fb_info *info)
+{
+     struct nuc900fb_info *fbi = info->par;
+     struct nuc900fb_mach_info *mach_info = fbi->dev->platform_data;
+     struct nuc900fb_display *display = NULL;
+     struct nuc900fb_display *default_display = mach_info->displays +
+                                                mach_info->default_display;
+     int i;
+
+     dprintk("check_var(var=%p, info=%p)\n", var, info);
+
+     /* validate x/y resolution */
+     /* choose default mode if possible */
+     if (var->xres = default_display->xres &&
+         var->yres = default_display->yres &&
+         var->bits_per_pixel = default_display->bpp)
+             display = default_display;
+     else
+             for (i = 0; i < mach_info->num_displays; i++)
+                     if (var->xres = mach_info->displays[i].xres &&
+                         var->yres = mach_info->displays[i].yres &&
+                         var->bits_per_pixel = mach_info->displays[i].bpp) {
+                             display = mach_info->displays + i;
+                             break;
+                     }
+
+     if (display = NULL) {
+             dprintk("wrong resolution or depth %dx%d at %d bit per pixel\n",
+                     var->xres, var->yres, var->bits_per_pixel);

Some of the printks in here really should be printks, not the
compile-time-suppressible dprintk().  Please review them all and
convert the ones which will be useful to end-users into printk().

quoted

+             return -EINVAL;
+     }
+
+     /* it should be the same size as the display */
+     var->xres_virtual       = display->xres;
+     var->yres_virtual       = display->yres;
+     var->height             = display->height;
+     var->width              = display->width;
+
+     /* copy lcd settings */
+     var->pixclock           = display->pixclock;
+     var->left_margin        = display->left_margin;
+     var->right_margin       = display->right_margin;
+     var->upper_margin       = display->upper_margin;
+     var->lower_margin       = display->lower_margin;
+     var->vsync_len          = display->vsync_len;
+     var->hsync_len          = display->hsync_len;
+
+     var->transp.offset      = 0;
+     var->transp.length      = 0;
+
+     fbi->regs.lcd_dccs = display->dccs;
+     fbi->regs.lcd_device_ctrl = display->devctl;
+     fbi->regs.lcd_va_fbctrl = display->fbctrl;
+     fbi->regs.lcd_va_scale = display->scale;
+
+     /* set R/G/B possions */
+     switch (var->bits_per_pixel) {
+     case 1:
+     case 2:
+     case 4:
+     case 8:

The above four lines are unneeded, but it's OK keeping them for
documentation purposes.

quoted

+     default:
+             var->red.offset         = 0;
+             var->red.length         = var->bits_per_pixel;
+             var->green              = var->red;
+             var->blue               = var->red;
+             break;
+     case 12:
+             var->red.length         = 4;
+             var->green.length       = 4;
+             var->blue.length        = 4;
+             var->red.offset         = 8;
+             var->green.offset       = 4;
+             var->blue.offset        = 0;
+             break;
+     case 16:
+             var->red.length         = 5;
+             var->green.length       = 6;
+             var->blue.length        = 5;
+             var->red.offset         = 11;
+             var->green.offset       = 5;
+             var->blue.offset        = 0;
+             break;
+     case 18:
+             var->red.length         = 6;
+             var->green.length       = 6;
+             var->blue.length        = 6;
+             var->red.offset         = 12;
+             var->green.offset       = 6;
+             var->blue.offset        = 0;
+             break;
+     case 32:
+             var->red.length         = 8;
+             var->green.length       = 8;
+             var->blue.length        = 8;
+             var->red.offset         = 16;
+             var->green.offset       = 8;
+             var->blue.offset        = 0;
+             break;
+     }
+
+     return 0;
+}
+
+/*
+ *   Calculate lcd register values from var setting & save into hw
+ */
+static void nuc900fb_calculate_lcd_regs(const struct fb_info *info,
+                                     struct nuc900fb_hw *regs)
+{
+     const struct fb_var_screeninfo *var = &info->var;
+     int vtt = var->height + var->upper_margin + var->lower_margin;
+     int htt = var->width + var->left_margin + var->right_margin;
+     int hsync = var->width + var->right_margin;
+     int vsync = var->height + var->lower_margin;

newline here, please.

quoted

+     regs->lcd_crtc_size = LCM_CRTC_SIZE_VTTVAL(vtt) |
+                           LCM_CRTC_SIZE_HTTVAL(htt);
+     regs->lcd_crtc_dend = LCM_CRTC_DEND_VDENDVAL(var->height) |
+                           LCM_CRTC_DEND_HDENDVAL(var->width);
+     regs->lcd_crtc_hr = LCM_CRTC_HR_EVAL(var->width + 5) |
+                         LCM_CRTC_HR_SVAL(var->width + 1);
+     regs->lcd_crtc_hsync = LCM_CRTC_HSYNC_EVAL(hsync + var->hsync_len) |
+                            LCM_CRTC_HSYNC_SVAL(hsync);
+     regs->lcd_crtc_vr = LCM_CRTC_VR_EVAL(vsync + var->vsync_len) |
+                         LCM_CRTC_VR_SVAL(vsync);
+
+}
+

...

+static int nuc900fb_debug_store(struct device *dev,
+                             struct device_attribute *attr,
+                             const char *buf, size_t len)
+{
+     if (len < 1)
+             return -EINVAL;
+
+     if (strnicmp(buf, "on", 2) = 0 ||
+         strnicmp(buf, "1", 1) = 0) {
+             debug = 1;
+     } else if (strnicmp(buf, "off", 3) = 0 ||
+                strnicmp(buf, "0", 1) = 0) {
+             debug = 0;
+
+     } else {
+             return -EINVAL;
+     }
+
+     return len;
+}

Duplicates the above-mentioned dynamic-debug facility.

quoted

+static DEVICE_ATTR(debug, 0666, nuc900fb_debug_show, nuc900fb_debug_store);
+

...

+static int __init nuc900fb_map_video_memory(struct fb_info *info)
+{
+     struct nuc900fb_info *fbi = info->par;
+     dma_addr_t map_dma;
+     unsigned long map_size = PAGE_ALIGN(info->fix.smem_len);
+
+     dprintk("nuc900fb_map_video_memory(fbi=%p) map_size %lu\n",
+             fbi, map_size);
+
+     info->screen_base = dma_alloc_writecombine(fbi->dev, map_size,
+                                                     &map_dma, GFP_KERNEL);

Here, do

       if (!info->screen_base)
               return -ENOMEM;

and the rest of the function gets neater.

quoted

+     if (info->screen_base != NULL) {
+             dprintk("nuc900fb_map_video_memory: clear %p:%08lx\n",
+                     info->screen_base, map_size);
+             memset(info->screen_base, 0x00, map_size);
+
+             info->fix.smem_start = map_dma;
+             dprintk("nuc900fb_map_video_memory: "
+                     "dma=%08lx cpu=%p size=%08lx\n",
+                     info->fix.smem_start, info->screen_base, map_size);
+     }
+     return (info->screen_base) ? 0 : -ENOMEM;
+}
+

...

+static char driver_name[] = "nuc900fb";
+
+static int __init nuc900fb_probe(struct platform_device *pdev)

Should this be __devinit?

quoted

+{
+     struct nuc900fb_info *fbi;
+     struct nuc900fb_display *display;
+     struct fb_info     *fbinfo;
+     struct nuc900fb_mach_info *mach_info;
+     struct resource *res;
+     int ret;
+     int irq;
+     int i;
+     int size;
+
+     dprintk("devinit\n");
+     mach_info = pdev->dev.platform_data;
+     if (mach_info = NULL) {
+             dev_err(&pdev->dev,
+                     "no platform data for lcd, cannot attach\n");
+             return -EINVAL;
+     }
+
+     if (mach_info->default_display > mach_info->num_displays) {
+             dev_err(&pdev->dev,
+                     "default display No. is %d but only %d displays \n",
+                     mach_info->default_display, mach_info->num_displays);
+             return -EINVAL;
+     }
+
+
+     display = mach_info->displays + mach_info->default_display;
+
+     irq = platform_get_irq(pdev, 0);
+     if (irq < 0) {
+             dev_err(&pdev->dev, "no irq for device\n");
+             return -ENOENT;
+     }
+
+     fbinfo = framebuffer_alloc(sizeof(struct nuc900fb_info), &pdev->dev);
+     if (!fbinfo)
+             return -ENOMEM;
+
+     platform_set_drvdata(pdev, fbinfo);
+
+     fbi = fbinfo->par;
+     fbi->dev = &pdev->dev;
+
+#ifdef CONFIG_CPU_NUC950
+     fbi->drv_type = LCDDRV_NUC950;
+#endif
+
+     res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+

...

+static void nuc900fb_stop_lcd(struct fb_info *info)
+{
+     struct nuc900fb_info *fbi = info->par;
+     void __iomem *regs = fbi->io;
+     unsigned long flags;
+
+     local_irq_save(flags);
+     writel((~LCM_DCCS_DISP_INT_EN) | (~LCM_DCCS_VA_EN) | (~LCM_DCCS_OSD_EN),
+             regs + REG_LCM_DCCS);
+     local_irq_restore(flags);
+}

The local_irq_save() is a worry.  What's it doing here?  It does
nothing to prevent an interrupt on another CPU!

quoted

+/*
+ *  Cleanup
+ */
+static int nuc900fb_remove(struct platform_device *pdev)
+{
+     struct fb_info *fbinfo = platform_get_drvdata(pdev);
+     struct nuc900fb_info *fbi = fbinfo->par;
+     int irq;
+
+     nuc900fb_stop_lcd(fbinfo);
+     msleep(1);
+
+     nuc900fb_unmap_video_memory(fbinfo);
+
+     iounmap(fbi->io);
+
+     irq = platform_get_irq(pdev, 0);
+     free_irq(irq, fbi);
+
+     release_resource(fbi->mem);
+     kfree(fbi->mem);
+
+     platform_set_drvdata(pdev, NULL);
+     framebuffer_release(fbinfo);
+
+     return 0;
+}
+

...

The driver generally looks OK to me.  Please cc me on any updates.

Thanks for your comments.
I will update my patch as soon as possible.

--
欢迎加入"NUC900 Linux BSP开发社区".
NUC900 Linux相关问题请发问题邮件至：
NUC900@googlegroups.com
取消订阅请发邮件：
NUC900+unsubscribe@googlegroups.com
新手请熟读社区首页：
https://groups.google.com/group/NUC900?hl=zh-CN