2022-02-03 00:17:50

by Guenter Roeck

Subject: Re: [PATCH v5] fbdev: fbmem: Fix the implicit type casting

On 2/2/22 15:33, Yizhuo Zhai wrote:
> In function do_fb_ioctl(), the "arg" is the type of unsigned long,
> and in "case FBIOBLANK:" this argument is casted into an int before
> passig to fb_blank(). In fb_blank(), the comparision
> if (blank > FB_BLANK_POWERDOWN) would be bypass if the original
> "arg" is a large number, which is possible because it comes from
> the user input. Fix this by adding the check before the function
> call.
>
> Signed-off-by: Yizhuo Zhai <[email protected]>
> ---
> drivers/video/fbdev/core/fbmem.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/drivers/video/fbdev/core/fbmem.c b/drivers/video/fbdev/core/fbmem.c
> index 0fa7ede94fa6..d5dec24c4d16 100644
> --- a/drivers/video/fbdev/core/fbmem.c
> +++ b/drivers/video/fbdev/core/fbmem.c
> @@ -1162,6 +1162,11 @@ static long do_fb_ioctl(struct fb_info *info, unsigned int cmd,
> case FBIOBLANK:
> console_lock();
> lock_fb_info(info);
> + if (arg > FB_BLANK_POWERDOWN) {
> + unlock_fb_info(info);
> + console_unlock();
> + return -EINVAL;
> + }

Is it really necessary to perform this check within the lock ?
arg is a passed in value, and FB_BLANK_POWERDOWN.
It seems to me that

case FBIOBLANK:
if (arg > FB_BLANK_POWERDOWN)
return -EINVAL;
console_lock();
...

should be sufficient.

Sorry if I missed some previous discussion; this is the first time
I looked at this patch.

Guenter

2022-02-03 18:27:10

by Yizhuo Zhai

Subject: [PATCH v6] fbdev: fbmem: Fix the implicit type casting

In function do_fb_ioctl(), the "arg" is the type of unsigned long,
and in "case FBIOBLANK:" this argument is casted into an int before
passig to fb_blank(). In fb_blank(), the comparision
if (blank > FB_BLANK_POWERDOWN) would be bypass if the original
"arg" is a large number, which is possible because it comes from
the user input. Fix this by adding the check before the function
call.

Signed-off-by: Yizhuo Zhai <[email protected]>
---
drivers/video/fbdev/core/fbmem.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/drivers/video/fbdev/core/fbmem.c b/drivers/video/fbdev/core/fbmem.c
index 0fa7ede94fa6..13083ad8d751 100644
--- a/drivers/video/fbdev/core/fbmem.c
+++ b/drivers/video/fbdev/core/fbmem.c
@@ -1160,6 +1160,8 @@ static long do_fb_ioctl(struct fb_info *info, unsigned int cmd,
ret = fbcon_set_con2fb_map_ioctl(argp);
break;
case FBIOBLANK:
+ if (arg > FB_BLANK_POWERDOWN)
+ return -EINVAL;
console_lock();
lock_fb_info(info);
ret = fb_blank(info, arg);
--
2.25.1

2022-02-07 13:10:52

by Sam Ravnborg

Subject: Re: [PATCH v6] fbdev: fbmem: Fix the implicit type casting

Hi Daniel,

I assume you will take this.

Patch is:
Reviewed-by: Sam Ravnborg <[email protected]>

Sam

On Wed, Feb 02, 2022 at 03:58:08PM -0800, Yizhuo Zhai wrote:
> In function do_fb_ioctl(), the "arg" is the type of unsigned long,
> and in "case FBIOBLANK:" this argument is casted into an int before
> passig to fb_blank(). In fb_blank(), the comparision
> if (blank > FB_BLANK_POWERDOWN) would be bypass if the original
> "arg" is a large number, which is possible because it comes from
> the user input. Fix this by adding the check before the function
> call.
>
> Signed-off-by: Yizhuo Zhai <[email protected]>
> ---
> drivers/video/fbdev/core/fbmem.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/video/fbdev/core/fbmem.c b/drivers/video/fbdev/core/fbmem.c
> index 0fa7ede94fa6..13083ad8d751 100644
> --- a/drivers/video/fbdev/core/fbmem.c
> +++ b/drivers/video/fbdev/core/fbmem.c
> @@ -1160,6 +1160,8 @@ static long do_fb_ioctl(struct fb_info *info, unsigned int cmd,
> ret = fbcon_set_con2fb_map_ioctl(argp);
> break;
> case FBIOBLANK:
> + if (arg > FB_BLANK_POWERDOWN)
> + return -EINVAL;
> console_lock();
> lock_fb_info(info);
> ret = fb_blank(info, arg);
> --
> 2.25.1