[PATCH v2] video: imsttfb: Fix use after free bug in imsttfb_probe due to lack of error-handling of init_imstt

[PATCH v2] video: imsttfb: Fix use after free bug in imsttfb_probe due to lack of error-handling of init_imstt

[Zheng Wang]

A use-after-free bug may occur if init_imstt invokes framebuffer_release
and free the info ptr. The caller, imsttfb_probe didn't notice that and
still keep the ptr as private data in pdev.

If we remove the driver which will call imsttfb_remove to make cleanup,
UAF happens.

Fix it by return error code if bad case happens in init_imstt.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
---
v2:
- add return error code in another location.
---
 drivers/video/fbdev/imsttfb.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/video/fbdev/imsttfb.c b/drivers/video/fbdev/imsttfb.c
index bea45647184e..975dd682fae4 100644
--- a/drivers/video/fbdev/imsttfb.c
+++ b/drivers/video/fbdev/imsttfb.c
@@ -1347,7 +1347,7 @@ static const struct fb_ops imsttfb_ops = {
 	.fb_ioctl 	= imsttfb_ioctl,
 };
 
-static void init_imstt(struct fb_info *info)
+static int init_imstt(struct fb_info *info)
 {
 	struct imstt_par *par = info->par;
 	__u32 i, tmp, *ip, *end;
@@ -1420,7 +1420,7 @@ static void init_imstt(struct fb_info *info)
 	    || !(compute_imstt_regvals(par, info->var.xres, info->var.yres))) {
 		printk("imsttfb: %ux%ux%u not supported\n", info->var.xres, info->var.yres, info->var.bits_per_pixel);
 		framebuffer_release(info);
-		return;
+		return -ENODEV;
 	}
 
 	sprintf(info->fix.id, "IMS TT (%s)", par->ramdac == IBM ? "IBM" : "TVP");
@@ -1456,12 +1456,13 @@ static void init_imstt(struct fb_info *info)
 
 	if (register_framebuffer(info) < 0) {
 		framebuffer_release(info);
-		return;
+		return -ENODEV;
 	}
 
 	tmp = (read_reg_le32(par->dc_regs, SSTATUS) & 0x0f00) >> 8;
 	fb_info(info, "%s frame buffer; %uMB vram; chip version %u\n",
 		info->fix.id, info->fix.smem_len >> 20, tmp);
+	return 0;
 }
 
 static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
@@ -1529,10 +1530,10 @@ static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (!par->cmap_regs)
 		goto error;
 	info->pseudo_palette = par->palette;
-	init_imstt(info);
-
-	pci_set_drvdata(pdev, info);
-	return 0;
+	ret = init_imstt(info);
+	if (!ret)
+		pci_set_drvdata(pdev, info);
+	return ret;
 
 error:
 	if (par->dc_regs)
-- 
2.25.1

Re: [PATCH v2] video: imsttfb: Fix use after free bug in imsttfb_probe due to lack of error-handling of init_imstt

[Helge Deller]

On 4/27/23 05:08, Zheng Wang wrote:
> A use-after-free bug may occur if init_imstt invokes framebuffer_release
> and free the info ptr. The caller, imsttfb_probe didn't notice that and
> still keep the ptr as private data in pdev.
>
> If we remove the driver which will call imsttfb_remove to make cleanup,
> UAF happens.
>
> Fix it by return error code if bad case happens in init_imstt.
>
> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> Signed-off-by: Zheng Wang <zyytlz.wz@163.com>

applied.
Thanks!
Helge

> ---
> v2:
> - add return error code in another location.
> ---
>   drivers/video/fbdev/imsttfb.c | 15 ++++++++-------
>   1 file changed, 8 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/video/fbdev/imsttfb.c b/drivers/video/fbdev/imsttfb.c
> index bea45647184e..975dd682fae4 100644
> --- a/drivers/video/fbdev/imsttfb.c
> +++ b/drivers/video/fbdev/imsttfb.c
> @@ -1347,7 +1347,7 @@ static const struct fb_ops imsttfb_ops = {
>   	.fb_ioctl 	= imsttfb_ioctl,
>   };
>
> -static void init_imstt(struct fb_info *info)
> +static int init_imstt(struct fb_info *info)
>   {
>   	struct imstt_par *par = info->par;
>   	__u32 i, tmp, *ip, *end;
> @@ -1420,7 +1420,7 @@ static void init_imstt(struct fb_info *info)
>   	    || !(compute_imstt_regvals(par, info->var.xres, info->var.yres))) {
>   		printk("imsttfb: %ux%ux%u not supported\n", info->var.xres, info->var.yres, info->var.bits_per_pixel);
>   		framebuffer_release(info);
> -		return;
> +		return -ENODEV;
>   	}
>
>   	sprintf(info->fix.id, "IMS TT (%s)", par->ramdac == IBM ? "IBM" : "TVP");
> @@ -1456,12 +1456,13 @@ static void init_imstt(struct fb_info *info)
>
>   	if (register_framebuffer(info) < 0) {
>   		framebuffer_release(info);
> -		return;
> +		return -ENODEV;
>   	}
>
>   	tmp = (read_reg_le32(par->dc_regs, SSTATUS) & 0x0f00) >> 8;
>   	fb_info(info, "%s frame buffer; %uMB vram; chip version %u\n",
>   		info->fix.id, info->fix.smem_len >> 20, tmp);
> +	return 0;
>   }
>
>   static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
> @@ -1529,10 +1530,10 @@ static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>   	if (!par->cmap_regs)
>   		goto error;
>   	info->pseudo_palette = par->palette;
> -	init_imstt(info);
> -
> -	pci_set_drvdata(pdev, info);
> -	return 0;
> +	ret = init_imstt(info);
> +	if (!ret)
> +		pci_set_drvdata(pdev, info);
> +	return ret;
>
>   error:
>   	if (par->dc_regs)

Re: [PATCH v2] video: imsttfb: Fix use after free bug in imsttfb_probe due to lack of error-handling of init_imstt

[Michal Koutný]

[-- Attachment #1: Type: text/plain, Size: 2032 bytes --]

Hello.

On Thu, Apr 27, 2023 at 11:08:41AM +0800, Zheng Wang <zyytlz.wz@163.com> wrote:
>  static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
> @@ -1529,10 +1530,10 @@ static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>  	if (!par->cmap_regs)
>  		goto error;
>  	info->pseudo_palette = par->palette;
> -	init_imstt(info);
> -
> -	pci_set_drvdata(pdev, info);
> -	return 0;
> +	ret = init_imstt(info);
> +	if (!ret)
> +		pci_set_drvdata(pdev, info);
> +	return ret;
>  
>  error:
>  	if (par->dc_regs)

This part caught my eye -- shouldn't the -ENODEV from init_imstt go
through the standard error with proper cleanup? (It seems like a leak
from my 30000 ft view, i.e. not sure about imsttfb_{probe,remove}
pairing.)

Shouldn't there be something like the diff below on top of the existing code?

Regards,
Michal

diff --git a/drivers/video/fbdev/imsttfb.c b/drivers/video/fbdev/imsttfb.c
index 975dd682fae4..a116ac8ca020 100644
--- a/drivers/video/fbdev/imsttfb.c
+++ b/drivers/video/fbdev/imsttfb.c
@@ -1419,7 +1419,6 @@ static int init_imstt(struct fb_info *info)
 	if ((info->var.xres * info->var.yres) * (info->var.bits_per_pixel >> 3) > info->fix.smem_len
 	    || !(compute_imstt_regvals(par, info->var.xres, info->var.yres))) {
 		printk("imsttfb: %ux%ux%u not supported\n", info->var.xres, info->var.yres, info->var.bits_per_pixel);
-		framebuffer_release(info);
 		return -ENODEV;
 	}
 
@@ -1455,7 +1454,6 @@ static int init_imstt(struct fb_info *info)
 	fb_alloc_cmap(&info->cmap, 0, 0);
 
 	if (register_framebuffer(info) < 0) {
-		framebuffer_release(info);
 		return -ENODEV;
 	}
 
@@ -1531,8 +1529,10 @@ static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto error;
 	info->pseudo_palette = par->palette;
 	ret = init_imstt(info);
-	if (!ret)
-		pci_set_drvdata(pdev, info);
+	if (ret)
+		goto error;
+
+	pci_set_drvdata(pdev, info);
 	return ret;
 
 error:



[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2] video: imsttfb: Fix use after free bug in imsttfb_probe due to lack of error-handling of init_imstt

[Helge Deller]

Hi Michal,

On 5/22/23 17:36, Michal Koutný wrote:
> On Thu, Apr 27, 2023 at 11:08:41AM +0800, Zheng Wang <zyytlz.wz@163.com> wrote:
>>   static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>> @@ -1529,10 +1530,10 @@ static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>>   	if (!par->cmap_regs)
>>   		goto error;
>>   	info->pseudo_palette = par->palette;
>> -	init_imstt(info);
>> -
>> -	pci_set_drvdata(pdev, info);
>> -	return 0;
>> +	ret = init_imstt(info);
>> +	if (!ret)
>> +		pci_set_drvdata(pdev, info);
>> +	return ret;
>>
>>   error:
>>   	if (par->dc_regs)
>
> This part caught my eye -- shouldn't the -ENODEV from init_imstt go
> through the standard error with proper cleanup? (It seems like a leak
> from my 30000 ft view, i.e. not sure about imsttfb_{probe,remove}
> pairing.)

Yes, you seem to be right.

> Shouldn't there be something like the diff below on top of the existing code?

Yes, but ....


> Regards,
> Michal
>
> diff --git a/drivers/video/fbdev/imsttfb.c b/drivers/video/fbdev/imsttfb.c
> index 975dd682fae4..a116ac8ca020 100644
> --- a/drivers/video/fbdev/imsttfb.c
> +++ b/drivers/video/fbdev/imsttfb.c
> @@ -1419,7 +1419,6 @@ static int init_imstt(struct fb_info *info)
>   	if ((info->var.xres * info->var.yres) * (info->var.bits_per_pixel >> 3) > info->fix.smem_len
>   	    || !(compute_imstt_regvals(par, info->var.xres, info->var.yres))) {
>   		printk("imsttfb: %ux%ux%u not supported\n", info->var.xres, info->var.yres, info->var.bits_per_pixel);
> -		framebuffer_release(info);
>   		return -ENODEV;
>   	}
>
> @@ -1455,7 +1454,6 @@ static int init_imstt(struct fb_info *info)
>   	fb_alloc_cmap(&info->cmap, 0, 0);
>
>   	if (register_framebuffer(info) < 0) {
> -		framebuffer_release(info);

That's ^^^  ok, but I think a
	fb_dealloc_cmap()
is missing here ...

... and in the error path of imsttfb_probe() ....

>   		return -ENODEV;
>   	}
>
> @@ -1531,8 +1529,10 @@ static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>   		goto error;
>   	info->pseudo_palette = par->palette;
>   	ret = init_imstt(info);
> -	if (!ret)
> -		pci_set_drvdata(pdev, info);
> +	if (ret)
> +		goto error;
> +
> +	pci_set_drvdata(pdev, info);
>   	return ret;
>
>   error:

Would you mind sending a proper patch?

Thanks for noticing!
Helge

Re: [PATCH 3/4] fbdev: imsttfb: Move a variable assignment for an error code in imsttfb_probe()

[Helge Deller]

On 5/23/23 19:42, Markus Elfring wrote:
> From: Markus Elfring <elfring@users.sourceforge.net>
> Date: Tue, 23 May 2023 18:30:26 +0200
>
> The value “-ENOMEM” was assigned to the variable “ret”
> at the beginning.
> Move this statement directly before the first ioremap() call.

Please do not move such variables without real need.
It makes backporting (of this and maybe follow-up patches) much more
complicated and the compiler will optimize it anyway.

Thanks!
Helge

>
> Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
> ---
>   drivers/video/fbdev/imsttfb.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/video/fbdev/imsttfb.c b/drivers/video/fbdev/imsttfb.c
> index 6490f544f8eb..90c72428e8d7 100644
> --- a/drivers/video/fbdev/imsttfb.c
> +++ b/drivers/video/fbdev/imsttfb.c
> @@ -1484,7 +1484,6 @@ static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>   	ret = aperture_remove_conflicting_pci_devices(pdev, "imsttfb");
>   	if (ret)
>   		return ret;
> -	ret = -ENOMEM;
>
>   	dp = pci_device_to_OF_node(pdev);
>   	if(dp)
> @@ -1525,6 +1524,7 @@ static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>   	}
>
>   	info->fix.smem_start = addr;
> +	ret = -ENOMEM;
>   	info->screen_base = (__u8 *)ioremap(addr, par->ramdac == IBM ?
>   					    0x400000 : 0x800000);
>   	if (!info->screen_base)
> --
> 2.40.1
>

Re: [PATCH 1/4] fbdev: imsttfb: Fix error handling in init_imstt()

[Helge Deller]

On 5/23/23 19:38, Markus Elfring wrote:
> From: Markus Elfring <elfring@users.sourceforge.net>
> Date: Tue, 23 May 2023 14:32:39 +0200
>
> The return value was overlooked from a call of
> the function “fb_alloc_cmap”.
>
> * Thus use a corresponding error check.
>
> * Add two jump targets so that a bit of exception handling
>    can be better reused at the end of this function.
>
>
> Reported-by: Helge Deller <deller@gmx.de>
> Link: https://lore.kernel.org/dri-devel/069f2f78-01f3-9476-d860-2b695c122649@gmx.de/
> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> Fixes: c75f5a550610 ("fbdev: imsttfb: Fix use after free bug in imsttfb_probe")
> Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
> ---
>   drivers/video/fbdev/imsttfb.c | 18 +++++++++++++-----
>   1 file changed, 13 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/video/fbdev/imsttfb.c b/drivers/video/fbdev/imsttfb.c
> index 975dd682fae4..d3532def4707 100644
> --- a/drivers/video/fbdev/imsttfb.c
> +++ b/drivers/video/fbdev/imsttfb.c
> @@ -1351,6 +1351,7 @@ static int init_imstt(struct fb_info *info)
>   {
>   	struct imstt_par *par = info->par;
>   	__u32 i, tmp, *ip, *end;
> +	int ret;
>
>   	tmp = read_reg_le32(par->dc_regs, PRC);
>   	if (par->ramdac == IBM)
> @@ -1419,8 +1420,7 @@ static int init_imstt(struct fb_info *info)
>   	if ((info->var.xres * info->var.yres) * (info->var.bits_per_pixel >> 3) > info->fix.smem_len
>   	    || !(compute_imstt_regvals(par, info->var.xres, info->var.yres))) {
>   		printk("imsttfb: %ux%ux%u not supported\n", info->var.xres, info->var.yres, info->var.bits_per_pixel);
> -		framebuffer_release(info);
> -		return -ENODEV;
> +		goto e_nodev;
>   	}
>
>   	sprintf(info->fix.id, "IMS TT (%s)", par->ramdac == IBM ? "IBM" : "TVP");
> @@ -1452,17 +1452,25 @@ static int init_imstt(struct fb_info *info)
>   	              FBINFO_HWACCEL_FILLRECT |
>   	              FBINFO_HWACCEL_YPAN;
>
> -	fb_alloc_cmap(&info->cmap, 0, 0);
> +	ret = fb_alloc_cmap(&info->cmap, 0, 0);
> +	if (ret)
> +		goto release_framebuffer;
>
>   	if (register_framebuffer(info) < 0) {
> -		framebuffer_release(info);
> -		return -ENODEV;
> +		fb_dealloc_cmap(&info->cmap);
> +		goto e_nodev;
>   	}
>
>   	tmp = (read_reg_le32(par->dc_regs, SSTATUS) & 0x0f00) >> 8;
>   	fb_info(info, "%s frame buffer; %uMB vram; chip version %u\n",
>   		info->fix.id, info->fix.smem_len >> 20, tmp);
>   	return 0;
> +
> +e_nodev:
> +	ret = -ENODEV;

I think the return value isn't checked at all, so you could
simply return below "-ENODEV" for all cases (instead of "return ret").
Then you don't need the e_nodev label and can simplify the flow.

Helge


> +release_framebuffer:
> +	framebuffer_release(info);
> +	return ret;
>   }
>
>   static int imsttfb_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
> --
> 2.40.1
>

Re: [PATCH 1/4] fbdev: imsttfb: Fix error handling in init_imstt()

[Helge Deller]

On 5/25/23 07:33, Markus Elfring wrote:
>>> The return value was overlooked from a call of
>>> the function “fb_alloc_cmap”.
>>>
>>> * Thus use a corresponding error check.
>>>
>>> * Add two jump targets so that a bit of exception handling
>>>     can be better reused at the end of this function.
> …
>>> +++ b/drivers/video/fbdev/imsttfb.c
> …
>>> @@ -1452,17 +1452,25 @@ static int init_imstt(struct fb_info *info)
>>>                      FBINFO_HWACCEL_FILLRECT |
>>>                      FBINFO_HWACCEL_YPAN;
>>>
>>> -    fb_alloc_cmap(&info->cmap, 0, 0);
>>> +    ret = fb_alloc_cmap(&info->cmap, 0, 0);
>>> +    if (ret)
>>> +        goto release_framebuffer;
>>>
>>>        if (register_framebuffer(info) < 0) {
>>> -        framebuffer_release(info);
>>> -        return -ENODEV;
>>> +        fb_dealloc_cmap(&info->cmap);
>>> +        goto e_nodev;
>>>        }
>>>
>>>        tmp = (read_reg_le32(par->dc_regs, SSTATUS) & 0x0f00) >> 8;
>>>        fb_info(info, "%s frame buffer; %uMB vram; chip version %u\n",
>>>            info->fix.id, info->fix.smem_len >> 20, tmp);
>>>        return 0;
>>> +
>>> +e_nodev:
>>> +    ret = -ENODEV;
>>
>> I think the return value isn't checked at all, so you could
>> simply return below "-ENODEV" for all cases (instead of "return ret").
>> Then you don't need the e_nodev label and can simplify the flow.
>
> Can it be helpful to distinguish involved error codes better?

No.

Helge