[PATCH] x86: don't assume all fb devices are PCI devices

[PATCH] x86: don't assume all fb devices are PCI devices

[Vitaly Kuznetsov]

When booting Hyper-V Generation 2 guests KASAN reports the following
out-of-bounds access:

BUG: KASAN: slab-out-of-bounds in fb_is_primary_device+0x58/0x70 at addr
  ffff880079cf0eb0
Read of size 8 by task swapper/0/1
...
 [<ffffffff81581308>] dump_stack+0x63/0x8b
 [<ffffffff812e1f99>] print_trailer+0xf9/0x150
 [<ffffffff812e7344>] object_err+0x34/0x40
 [<ffffffff812e9630>] kasan_report_error+0x230/0x550
 [<ffffffff812e9ee8>] kasan_report+0x58/0x60
 [<ffffffff812e4500>] ? ___slab_alloc+0x80/0x490
 [<ffffffff81878a28>] ? fb_is_primary_device+0x58/0x70
 [<ffffffff812e87cd>] __asan_load8+0x5d/0x70
 [<ffffffff81878a28>] fb_is_primary_device+0x58/0x70
 [<ffffffff8162357a>] register_framebuffer+0xda/0x5b0
 [<ffffffff816234a0>] ? remove_conflicting_framebuffers+0x50/0x50
...

The issue is caused by the to_pci_dev() call with no check that the given
info->device is in fact a pci device and some fb devices (Hyper-V FB, EFI
FB,...) are not. fb_is_primary_device() is not on any performance critical
path, replace to_pci_dev() with raw scan.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
 arch/x86/video/fbdev.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/arch/x86/video/fbdev.c b/arch/x86/video/fbdev.c
index d5644bb..f309d6c 100644
--- a/arch/x86/video/fbdev.c
+++ b/arch/x86/video/fbdev.c
@@ -14,12 +14,18 @@
 int fb_is_primary_device(struct fb_info *info)
 {
 	struct device *device = info->device;
-	struct pci_dev *pci_dev = NULL;
+	struct pci_dev *dev = NULL, *pci_dev = NULL;
 	struct pci_dev *default_device = vga_default_device();
 	struct resource *res = NULL;
 
-	if (device)
-		pci_dev = to_pci_dev(device);
+	/*
+	 * We're not sure info->device is a pci device, do full scan instead
+	 * of to_pci_dev().
+	 */
+	for_each_pci_dev(dev) {
+		if (&dev->dev == device)
+			pci_dev = dev;
+	}
 
 	if (!pci_dev)
 		return 0;
-- 
2.5.0

Re: [PATCH] x86: don't assume all fb devices are PCI devices

[Ingo Molnar]

So I'd love to have an ACK from Bjorn. (Full mail quoted below)

Thanks,

	Ingo

* Vitaly Kuznetsov <vkuznets@redhat.com> wrote:

> When booting Hyper-V Generation 2 guests KASAN reports the following
> out-of-bounds access:
> 
> BUG: KASAN: slab-out-of-bounds in fb_is_primary_device+0x58/0x70 at addr
>   ffff880079cf0eb0
> Read of size 8 by task swapper/0/1
> ...
>  [<ffffffff81581308>] dump_stack+0x63/0x8b
>  [<ffffffff812e1f99>] print_trailer+0xf9/0x150
>  [<ffffffff812e7344>] object_err+0x34/0x40
>  [<ffffffff812e9630>] kasan_report_error+0x230/0x550
>  [<ffffffff812e9ee8>] kasan_report+0x58/0x60
>  [<ffffffff812e4500>] ? ___slab_alloc+0x80/0x490
>  [<ffffffff81878a28>] ? fb_is_primary_device+0x58/0x70
>  [<ffffffff812e87cd>] __asan_load8+0x5d/0x70
>  [<ffffffff81878a28>] fb_is_primary_device+0x58/0x70
>  [<ffffffff8162357a>] register_framebuffer+0xda/0x5b0
>  [<ffffffff816234a0>] ? remove_conflicting_framebuffers+0x50/0x50
> ...
> 
> The issue is caused by the to_pci_dev() call with no check that the given
> info->device is in fact a pci device and some fb devices (Hyper-V FB, EFI
> FB,...) are not. fb_is_primary_device() is not on any performance critical
> path, replace to_pci_dev() with raw scan.
> 
> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
> ---
>  arch/x86/video/fbdev.c | 12 +++++++++---
>  1 file changed, 9 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/video/fbdev.c b/arch/x86/video/fbdev.c
> index d5644bb..f309d6c 100644
> --- a/arch/x86/video/fbdev.c
> +++ b/arch/x86/video/fbdev.c
> @@ -14,12 +14,18 @@
>  int fb_is_primary_device(struct fb_info *info)
>  {
>  	struct device *device = info->device;
> -	struct pci_dev *pci_dev = NULL;
> +	struct pci_dev *dev = NULL, *pci_dev = NULL;
>  	struct pci_dev *default_device = vga_default_device();
>  	struct resource *res = NULL;
>  
> -	if (device)
> -		pci_dev = to_pci_dev(device);
> +	/*
> +	 * We're not sure info->device is a pci device, do full scan instead
> +	 * of to_pci_dev().
> +	 */
> +	for_each_pci_dev(dev) {
> +		if (&dev->dev == device)
> +			pci_dev = dev;
> +	}
>  
>  	if (!pci_dev)
>  		return 0;

Re: [PATCH] x86: don't assume all fb devices are PCI devices

[Bjorn Helgaas]

On Fri, Mar 11, 2016 at 03:39:18PM +0100, Vitaly Kuznetsov wrote:
> When booting Hyper-V Generation 2 guests KASAN reports the following
> out-of-bounds access:
> 
> BUG: KASAN: slab-out-of-bounds in fb_is_primary_device+0x58/0x70 at addr
>   ffff880079cf0eb0
> Read of size 8 by task swapper/0/1
> ...
>  [<ffffffff81581308>] dump_stack+0x63/0x8b
>  [<ffffffff812e1f99>] print_trailer+0xf9/0x150
>  [<ffffffff812e7344>] object_err+0x34/0x40
>  [<ffffffff812e9630>] kasan_report_error+0x230/0x550
>  [<ffffffff812e9ee8>] kasan_report+0x58/0x60
>  [<ffffffff812e4500>] ? ___slab_alloc+0x80/0x490
>  [<ffffffff81878a28>] ? fb_is_primary_device+0x58/0x70
>  [<ffffffff812e87cd>] __asan_load8+0x5d/0x70
>  [<ffffffff81878a28>] fb_is_primary_device+0x58/0x70
>  [<ffffffff8162357a>] register_framebuffer+0xda/0x5b0
>  [<ffffffff816234a0>] ? remove_conflicting_framebuffers+0x50/0x50
> ...
> 
> The issue is caused by the to_pci_dev() call with no check that the given
> info->device is in fact a pci device and some fb devices (Hyper-V FB, EFI
> FB,...) are not. fb_is_primary_device() is not on any performance critical
> path, replace to_pci_dev() with raw scan.
> 
> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
> ---
>  arch/x86/video/fbdev.c | 12 +++++++++---
>  1 file changed, 9 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/video/fbdev.c b/arch/x86/video/fbdev.c
> index d5644bb..f309d6c 100644
> --- a/arch/x86/video/fbdev.c
> +++ b/arch/x86/video/fbdev.c
> @@ -14,12 +14,18 @@
>  int fb_is_primary_device(struct fb_info *info)
>  {
>  	struct device *device = info->device;
> -	struct pci_dev *pci_dev = NULL;
> +	struct pci_dev *dev = NULL, *pci_dev = NULL;
>  	struct pci_dev *default_device = vga_default_device();
>  	struct resource *res = NULL;
>  
> -	if (device)
> -		pci_dev = to_pci_dev(device);
> +	/*
> +	 * We're not sure info->device is a pci device, do full scan instead
> +	 * of to_pci_dev().
> +	 */
> +	for_each_pci_dev(dev) {
> +		if (&dev->dev == device)
> +			pci_dev = dev;
> +	}
>  	if (!pci_dev)
>  		return 0;

I think dev_is_pci() is a better solution:

  if (!device || !dev_is_pci(device))
    return 0;

  pci_dev = to_pci_dev(device);
  if (vga_default_device()) {
    if (vga_default_device() == pci_dev)
      return 1;
    return 0;
  }

  res = &pci_dev->resource[PCI_ROM_RESOURCE];
  if (res->flags & IORESOURCE_ROM_SHADOW)
    return 1;

  return 0;

This says a non-PCI device cannot be the primary device.  I don't know
framebuffers, so I don't know if that's what you really want, but I
guess it's what the original code says.

Bjorn

Re: [PATCH] x86: don't assume all fb devices are PCI devices

[Vitaly Kuznetsov]

Bjorn Helgaas <helgaas@kernel.org> writes:

> On Fri, Mar 11, 2016 at 03:39:18PM +0100, Vitaly Kuznetsov wrote:
>> When booting Hyper-V Generation 2 guests KASAN reports the following
>> out-of-bounds access:
>> 
>> BUG: KASAN: slab-out-of-bounds in fb_is_primary_device+0x58/0x70 at addr
>>   ffff880079cf0eb0
>> Read of size 8 by task swapper/0/1
>> ...
>>  [<ffffffff81581308>] dump_stack+0x63/0x8b
>>  [<ffffffff812e1f99>] print_trailer+0xf9/0x150
>>  [<ffffffff812e7344>] object_err+0x34/0x40
>>  [<ffffffff812e9630>] kasan_report_error+0x230/0x550
>>  [<ffffffff812e9ee8>] kasan_report+0x58/0x60
>>  [<ffffffff812e4500>] ? ___slab_alloc+0x80/0x490
>>  [<ffffffff81878a28>] ? fb_is_primary_device+0x58/0x70
>>  [<ffffffff812e87cd>] __asan_load8+0x5d/0x70
>>  [<ffffffff81878a28>] fb_is_primary_device+0x58/0x70
>>  [<ffffffff8162357a>] register_framebuffer+0xda/0x5b0
>>  [<ffffffff816234a0>] ? remove_conflicting_framebuffers+0x50/0x50
>> ...
>> 
>> The issue is caused by the to_pci_dev() call with no check that the given
>> info->device is in fact a pci device and some fb devices (Hyper-V FB, EFI
>> FB,...) are not. fb_is_primary_device() is not on any performance critical
>> path, replace to_pci_dev() with raw scan.
>> 
>> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
>> ---
>>  arch/x86/video/fbdev.c | 12 +++++++++---
>>  1 file changed, 9 insertions(+), 3 deletions(-)
>> 
>> diff --git a/arch/x86/video/fbdev.c b/arch/x86/video/fbdev.c
>> index d5644bb..f309d6c 100644
>> --- a/arch/x86/video/fbdev.c
>> +++ b/arch/x86/video/fbdev.c
>> @@ -14,12 +14,18 @@
>>  int fb_is_primary_device(struct fb_info *info)
>>  {
>>  	struct device *device = info->device;
>> -	struct pci_dev *pci_dev = NULL;
>> +	struct pci_dev *dev = NULL, *pci_dev = NULL;
>>  	struct pci_dev *default_device = vga_default_device();
>>  	struct resource *res = NULL;
>>  
>> -	if (device)
>> -		pci_dev = to_pci_dev(device);
>> +	/*
>> +	 * We're not sure info->device is a pci device, do full scan instead
>> +	 * of to_pci_dev().
>> +	 */
>> +	for_each_pci_dev(dev) {
>> +		if (&dev->dev == device)
>> +			pci_dev = dev;
>> +	}
>>  	if (!pci_dev)
>>  		return 0;
>
> I think dev_is_pci() is a better solution:
>
>   if (!device || !dev_is_pci(device))
>     return 0;

Sure, will do v2.

>
>   pci_dev = to_pci_dev(device);
>   if (vga_default_device()) {
>     if (vga_default_device() == pci_dev)
>       return 1;
>     return 0;
>   }
>
>   res = &pci_dev->resource[PCI_ROM_RESOURCE];
>   if (res->flags & IORESOURCE_ROM_SHADOW)
>     return 1;
>
>   return 0;
>
> This says a non-PCI device cannot be the primary device.  I don't know
> framebuffers, so I don't know if that's what you really want, but I
> guess it's what the original code says.

Yes, the original code is for pci devices only and I didn't plan to
change that, my intention is to prevent out-of-bounds access.

>
> Bjorn

-- 
  Vitaly