Re: [PATCH] SiMotion VoyagerGX framebuffer: blue stripped background

[Sergei Shtylylov <sshtylyov@ru.mvista.com>]

[-- Attachment #1: Type: text/plain, Size: 594 bytes --]

Hello.

Sergei Shtylylov wrote:

>    This driver was using an incorrect typecast when setting pseudopalette,
> hence were the blue strips on the black char background. As this driver
> happens to be maintaned by Linux/MIPS, here's the patch (I've also 
> noticed a
> typo in the head comment, hence comes another hunk)...

     Have noticed that regno check in smi_setcolreg() is too relaxed as
pseudo-palette has only 16 entries. So, had to update the patch.

WBR, Sergei

Signed-off-by: Konstantin Baydarov <kbaidarov@ru.mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>


[-- Attachment #2: VoyagerGX-blue-strips.patch --]
[-- Type: text/plain, Size: 1131 bytes --]

diff --git a/drivers/video/smivgxfb.c b/drivers/video/smivgxfb.c
index d5755c5..944ff4a 100644
--- a/drivers/video/smivgxfb.c
+++ b/drivers/video/smivgxfb.c
@@ -1,5 +1,5 @@
 /***************************************************************************
- *  Silicon Motion VoyaagerGX framebuffer driver
+ *  Silicon Motion VoyagerGX framebuffer driver
  *
  * 	ported to 2.6 by Embedded Alley Solutions, Inc
  * 	Copyright (C) 2005 Embedded Alley Solutions, Inc
@@ -159,10 +159,10 @@ smi_setcolreg(unsigned regno, unsigned r
 	unsigned blue, unsigned transp,
 	struct fb_info *info)
 {
-	if (regno > 255)
+	if (regno > 15)
 		return 1;
 
-	((u16 *)(info->pseudo_palette))[regno] =
+	((u32 *)(info->pseudo_palette))[regno] =
 		    ((red & 0xf800) >> 0) |
 		    ((green & 0xfc00) >> 5) |
 		    ((blue & 0xf800) >> 11);
@@ -318,9 +318,9 @@ static int __devinit vgx_pci_probe(struc
 	if (!info.pseudo_palette) {
 		return -ENOMEM;
 	}
-	memset(info.pseudo_palette, 0, sizeof(u32) *16);
+	memset(info.pseudo_palette, 0, sizeof(u32) * 16);
 
-	fb_alloc_cmap(&info.cmap,256,0);
+	fb_alloc_cmap(&info.cmap, 256, 0);
 
 	smi_setmode();