[PATCH] video: add sh_mobile_lcdcfb memory resource

[PATCH] video: add sh_mobile_lcdcfb memory resource

[Magnus Damm]

From: Magnus Damm <damm@igel.co.jp>

This patch adds memory resource support to the LCDC driver.
With this applied the physical memory range used for framebuffer
data will show up in /proc/iomem together with other SuperH
Mobile devices.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
---

 drivers/video/sh_mobile_lcdcfb.c |   16 ++++++++++++++++
 1 file changed, 16 insertions(+)

--- 0001/drivers/video/sh_mobile_lcdcfb.c
+++ work/drivers/video/sh_mobile_lcdcfb.c	2009-05-11 21:14:35.000000000 +0900
@@ -35,6 +35,7 @@ struct sh_mobile_lcdc_chan {
 	struct fb_deferred_io defio;
 	unsigned long frame_end;
 	wait_queue_head_t frame_end_wait;
+	struct resource memory;
 };
 
 struct sh_mobile_lcdc_priv {
@@ -813,12 +814,26 @@ static int __init sh_mobile_lcdc_probe(s
 			break;
 		}
 
+		res = &priv->ch[i].memory;
+		res->name = "lcdc";
+		res->start = priv->ch[i].dma_handle;
+		res->end = res->start + info->fix.smem_len - 1;
+		res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
+		error = insert_resource(&iomem_resource, res);
+		if (error) {
+			dev_err(&pdev->dev, "unable to request resource\n");
+			dma_free_coherent(&pdev->dev, info->fix.smem_len,
+					  buf, priv->ch[i].dma_handle);
+			break;
+		}
+
 		info->pseudo_palette = &priv->ch[i].pseudo_palette;
 		info->flags = FBINFO_FLAG_DEFAULT;
 
 		error = fb_alloc_cmap(&info->cmap, PALETTE_NR, 0);
 		if (error < 0) {
 			dev_err(&pdev->dev, "unable to allocate cmap\n");
+			release_resource(res);
 			dma_free_coherent(&pdev->dev, info->fix.smem_len,
 					  buf, priv->ch[i].dma_handle);
 			break;
@@ -887,6 +902,7 @@ static int sh_mobile_lcdc_remove(struct 
 		if (!info->device)
 			continue;
 
+		release_resource(&priv->ch[i].memory);
 		dma_free_coherent(&pdev->dev, info->fix.smem_len,
 				  info->screen_base, priv->ch[i].dma_handle);
 		fb_dealloc_cmap(&info->cmap);

Re: [PATCH] video: add sh_mobile_lcdcfb memory resource

[Paul Mundt]

On Tue, May 12, 2009 at 06:57:45PM +0900, Magnus Damm wrote:
> From: Magnus Damm <damm@igel.co.jp>
> 
> This patch adds memory resource support to the LCDC driver.
> With this applied the physical memory range used for framebuffer
> data will show up in /proc/iomem together with other SuperH
> Mobile devices.
> 
> Signed-off-by: Magnus Damm <damm@igel.co.jp>

Drivers have no place doing this on their own, besides, insert_resource()
is not available to modules. The platform device code already takes care
of this, so perhaps I am missing what the point of this change is?

Re: [PATCH] video: add sh_mobile_lcdcfb memory resource

[Magnus Damm]

On Tue, May 12, 2009 at 7:50 PM, Paul Mundt <lethal@linux-sh.org> wrote:
> On Tue, May 12, 2009 at 06:57:45PM +0900, Magnus Damm wrote:
>> From: Magnus Damm <damm@igel.co.jp>
>>
>> This patch adds memory resource support to the LCDC driver.
>> With this applied the physical memory range used for framebuffer
>> data will show up in /proc/iomem together with other SuperH
>> Mobile devices.
>>
>> Signed-off-by: Magnus Damm <damm@igel.co.jp>
>
> Drivers have no place doing this on their own, besides, insert_resource()
> is not available to modules. The platform device code already takes care
> of this, so perhaps I am missing what the point of this change is?

Other SuperH Mobile multimedia blocks like the CEU and VEU/VPU/JPU all
get passed a physically contiguous area of memory as platform data. So
for those blocks the memory range show up in /proc/iomem. For the LCDC
this does not happen today.

Thanks for pointing out the module issue. I guess the proper fix would
be to pass a struct resource with memory to the LCDC driver as well
instead of this patch.

Cheeers,

/ magnus