[PATCH] iommu: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()

[PATCH] iommu: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()

[Qasim Ijaz]

This patch resolves a UBSAN shift-out-of-bounds issue in 
iova_bitmap_offset_to_index() where shifting the constant "1" (of type int) 
by bitmap->mapped.pgshift (an unsigned long value) could result in undefined behavior. 

The constant "1" defaults to a 32-bit "int", and when "pgshift" exceeds 31 (e.g., pgshift = 63)
the shift operation overflows, as the result cannot be represented in a 32-bit type.

To resolve this, the constant is updated to "1UL", promoting it to an unsigned long type 
to match the operand's type.

Reported-by: syzbot <syzbot+85992ace37d5b7b51635@syzkaller.appspotmail.com>
Closes: https://syzkaller.appspot.com/bug?extid=85992ace37d5b7b51635
Signed-off-by: Qasim Ijaz <qasdev00@gmail.com>
---
 drivers/iommu/iommufd/iova_bitmap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/iommu/iommufd/iova_bitmap.c b/drivers/iommu/iommufd/iova_bitmap.c
index ab665cf38ef4..39a86a4a1d3a 100644
--- a/drivers/iommu/iommufd/iova_bitmap.c
+++ b/drivers/iommu/iommufd/iova_bitmap.c
@@ -130,7 +130,7 @@ struct iova_bitmap {
 static unsigned long iova_bitmap_offset_to_index(struct iova_bitmap *bitmap,
 						 unsigned long iova)
 {
-	unsigned long pgsize = 1 << bitmap->mapped.pgshift;
+	unsigned long pgsize = 1UL << bitmap->mapped.pgshift;
 
 	return iova / (BITS_PER_TYPE(*bitmap->bitmap) * pgsize);
 }
-- 
2.39.5

Re: [PATCH] iommu: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()

[Joao Martins]

On 12/01/2025 12:39, Qasim Ijaz wrote:
> This patch resolves a UBSAN shift-out-of-bounds issue in 

Avoid the 'this patch' e.g. Resolve a UBSAN shift-out-of-bonds (...)

The Subject component part could also be a bit more specific e.g.

iommufd/iova_bitmap: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()

> iova_bitmap_offset_to_index() where shifting the constant "1" (of type int) 
> by bitmap->mapped.pgshift (an unsigned long value) could result in undefined behavior. 
> 
> The constant "1" defaults to a 32-bit "int", and when "pgshift" exceeds 31 (e.g., pgshift = 63)
> the shift operation overflows, as the result cannot be represented in a 32-bit type.
> 
> To resolve this, the constant is updated to "1UL", promoting it to an unsigned long type 
> to match the operand's type.
> 
> Reported-by: syzbot <syzbot+85992ace37d5b7b51635@syzkaller.appspotmail.com>
> Closes: https://syzkaller.appspot.com/bug?extid=85992ace37d5b7b51635
> Signed-off-by: Qasim Ijaz <qasdev00@gmail.com>

With those two nits:

	Reviewed-by: Joao Martins <joao.m.martins@oracle.com>


> ---
>  drivers/iommu/iommufd/iova_bitmap.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/iommu/iommufd/iova_bitmap.c b/drivers/iommu/iommufd/iova_bitmap.c
> index ab665cf38ef4..39a86a4a1d3a 100644
> --- a/drivers/iommu/iommufd/iova_bitmap.c
> +++ b/drivers/iommu/iommufd/iova_bitmap.c
> @@ -130,7 +130,7 @@ struct iova_bitmap {
>  static unsigned long iova_bitmap_offset_to_index(struct iova_bitmap *bitmap,
>  						 unsigned long iova)
>  {
> -	unsigned long pgsize = 1 << bitmap->mapped.pgshift;
> +	unsigned long pgsize = 1UL << bitmap->mapped.pgshift;
>  
>  	return iova / (BITS_PER_TYPE(*bitmap->bitmap) * pgsize);
>  }

Re: [PATCH] iommu: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()

[Jason Gunthorpe]

On Mon, Jan 13, 2025 at 12:00:29PM +0000, Joao Martins wrote:
> On 12/01/2025 12:39, Qasim Ijaz wrote:
> > This patch resolves a UBSAN shift-out-of-bounds issue in 
> 
> Avoid the 'this patch' e.g. Resolve a UBSAN shift-out-of-bonds (...)
> 
> The Subject component part could also be a bit more specific e.g.
> 
> iommufd/iova_bitmap: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()
> 
> > iova_bitmap_offset_to_index() where shifting the constant "1" (of type int) 
> > by bitmap->mapped.pgshift (an unsigned long value) could result in undefined behavior. 
> > 
> > The constant "1" defaults to a 32-bit "int", and when "pgshift" exceeds 31 (e.g., pgshift = 63)
> > the shift operation overflows, as the result cannot be represented in a 32-bit type.
> > 
> > To resolve this, the constant is updated to "1UL", promoting it to an unsigned long type 
> > to match the operand's type.
> > 
> > Reported-by: syzbot <syzbot+85992ace37d5b7b51635@syzkaller.appspotmail.com>
> > Closes: https://syzkaller.appspot.com/bug?extid=85992ace37d5b7b51635
> > Signed-off-by: Qasim Ijaz <qasdev00@gmail.com>
> 
> With those two nits:
> 
> 	Reviewed-by: Joao Martins <joao.m.martins@oracle.com>

It needs a fixes line too

Jason

Re: [PATCH] iommu: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()

[Joao Martins]

On 13/01/2025 16:22, Jason Gunthorpe wrote:
> On Mon, Jan 13, 2025 at 12:00:29PM +0000, Joao Martins wrote:
>> On 12/01/2025 12:39, Qasim Ijaz wrote:
>>> This patch resolves a UBSAN shift-out-of-bounds issue in 
>>
>> Avoid the 'this patch' e.g. Resolve a UBSAN shift-out-of-bonds (...)
>>
>> The Subject component part could also be a bit more specific e.g.
>>
>> iommufd/iova_bitmap: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()
>>
>>> iova_bitmap_offset_to_index() where shifting the constant "1" (of type int) 
>>> by bitmap->mapped.pgshift (an unsigned long value) could result in undefined behavior. 
>>>
>>> The constant "1" defaults to a 32-bit "int", and when "pgshift" exceeds 31 (e.g., pgshift = 63)
>>> the shift operation overflows, as the result cannot be represented in a 32-bit type.
>>>
>>> To resolve this, the constant is updated to "1UL", promoting it to an unsigned long type 
>>> to match the operand's type.
>>>
>>> Reported-by: syzbot <syzbot+85992ace37d5b7b51635@syzkaller.appspotmail.com>
>>> Closes: https://syzkaller.appspot.com/bug?extid=85992ace37d5b7b51635
>>> Signed-off-by: Qasim Ijaz <qasdev00@gmail.com>
>>
>> With those two nits:
>>
>> 	Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
> 
> It needs a fixes line too

It should be

Fixes: 495c06d82ba ("vfio: Add an IOVA bitmap support")

  Joao

Re: [PATCH] iommu: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()

[Qasim Ijaz]

On Mon, Jan 13, 2025 at 04:25:29PM +0000, Joao Martins wrote:
> On 13/01/2025 16:22, Jason Gunthorpe wrote:
> > On Mon, Jan 13, 2025 at 12:00:29PM +0000, Joao Martins wrote:
> >> On 12/01/2025 12:39, Qasim Ijaz wrote:
> >>> This patch resolves a UBSAN shift-out-of-bounds issue in 
> >>
> >> Avoid the 'this patch' e.g. Resolve a UBSAN shift-out-of-bonds (...)
> >>
> >> The Subject component part could also be a bit more specific e.g.
> >>
> >> iommufd/iova_bitmap: Fix shift-out-of-bounds in iova_bitmap_offset_to_index()
> >>
> >>> iova_bitmap_offset_to_index() where shifting the constant "1" (of type int) 
> >>> by bitmap->mapped.pgshift (an unsigned long value) could result in undefined behavior. 
> >>>
> >>> The constant "1" defaults to a 32-bit "int", and when "pgshift" exceeds 31 (e.g., pgshift = 63)
> >>> the shift operation overflows, as the result cannot be represented in a 32-bit type.
> >>>
> >>> To resolve this, the constant is updated to "1UL", promoting it to an unsigned long type 
> >>> to match the operand's type.
> >>>
> >>> Reported-by: syzbot <syzbot+85992ace37d5b7b51635@syzkaller.appspotmail.com>
> >>> Closes: https://syzkaller.appspot.com/bug?extid=85992ace37d5b7b51635
> >>> Signed-off-by: Qasim Ijaz <qasdev00@gmail.com>
> >>
> >> With those two nits:
> >>
> >> 	Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
> > 
> > It needs a fixes line too
> 
> It should be
> 
> Fixes: 495c06d82ba ("vfio: Add an IOVA bitmap support")
> 
>   Joao

Thank you for your feedback on the initial patch Jason and Joao. I’ve incorporated your suggestions and submitted a v2 patch for review.

Please let me know if further adjustments are needed.

Best regards,
Qasim