Re: [Linux-fbdev-devel] [PATCH] fbdev: Wrong test on unsigned in fb_set_user_cmap()?

[Roel Kluin <roel.kluin@gmail.com>]

struct fb_cmap_user member start is unsigned. The same condition
is already checked in fb_set_cmap(), so this should be safe.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
---
>> Is this required?
> 
> Drop the whole if() as exactly the same condition is checked in the
> fb_set_cmap() again.

Ok

> Anyway, the check of the cmap->start < 0 does
> not make any sense as the start is u32 value (most userspace
> addresses will be lower then 2GB on 32 bit system so the error cannot
> be caught by the check). I vote for removing the (cmap->start < 0) in
> the fb_set_cmap as well as most drivers check the start value already
> in driver's fb_setcolreg() function.

In fb_set_cmap() this is not `cmap->start' but `start' which has type
int. Therefore I think the test makes some sense there, so I left it.

> 
> Best regards,
> Krzysztof

Thanks, here:

diff --git a/drivers/video/fbcmap.c b/drivers/video/fbcmap.c
index f53b9f1..f46d617 100644
--- a/drivers/video/fbcmap.c
+++ b/drivers/video/fbcmap.c
@@ -266,11 +266,6 @@ int fb_set_user_cmap(struct fb_cmap_user *cmap, struct fb_info *info)
 		rc = -ENODEV;
 		goto out;
 	}
-	if (cmap->start < 0 || (!info->fbops->fb_setcolreg &&
-				!info->fbops->fb_setcmap)) {
-		rc = -EINVAL;
-		goto out1;
-	}
 	rc = fb_set_cmap(&umap, info);
 out1:
 	unlock_fb_info(info);