The use of ALIGN() in uvc_alloc_entity() is incorrect, since the size of
(entity->pads) is not a power of two. As a stop-gap, until a better
solution is adapted, use roundup() instead.
Found by a static assertion. Compile-tested only.
Fixes: 4ffc2d89f38a ("uvcvideo: Register subdevices for each entity")
Cc: Laurent Pinchart <[email protected]>
Cc: Mauro Carvalho Chehab <[email protected]>
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Nadav Amit <[email protected]>
---
drivers/media/usb/uvc/uvc_driver.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
index 2469b49b2b30..6b989d41c034 100644
--- a/drivers/media/usb/uvc/uvc_driver.c
+++ b/drivers/media/usb/uvc/uvc_driver.c
@@ -909,7 +909,7 @@ static struct uvc_entity *uvc_alloc_entity(u16 type, u8 id,
unsigned int size;
unsigned int i;
- extra_size = ALIGN(extra_size, sizeof(*entity->pads));
+ extra_size = roundup(extra_size, sizeof(*entity->pads));
num_inputs = (type & UVC_TERM_OUTPUT) ? num_pads : num_pads - 1;
size = sizeof(*entity) + extra_size + sizeof(*entity->pads) * num_pads
+ num_inputs;
--
2.17.0
Hi Nadav,
Thank you for the patch.
On Monday, 4 June 2018 16:47:13 EEST Nadav Amit wrote:
> The use of ALIGN() in uvc_alloc_entity() is incorrect, since the size of
> (entity->pads) is not a power of two. As a stop-gap, until a better
> solution is adapted, use roundup() instead.
>
> Found by a static assertion. Compile-tested only.
>
> Fixes: 4ffc2d89f38a ("uvcvideo: Register subdevices for each entity")
>
> Cc: Laurent Pinchart <[email protected]>
> Cc: Mauro Carvalho Chehab <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
>
> Signed-off-by: Nadav Amit <[email protected]>
> ---
> drivers/media/usb/uvc/uvc_driver.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/media/usb/uvc/uvc_driver.c
> b/drivers/media/usb/uvc/uvc_driver.c index 2469b49b2b30..6b989d41c034
> 100644
> --- a/drivers/media/usb/uvc/uvc_driver.c
> +++ b/drivers/media/usb/uvc/uvc_driver.c
> @@ -909,7 +909,7 @@ static struct uvc_entity *uvc_alloc_entity(u16 type, u8
> id, unsigned int size;
> unsigned int i;
>
> - extra_size = ALIGN(extra_size, sizeof(*entity->pads));
> + extra_size = roundup(extra_size, sizeof(*entity->pads));
> num_inputs = (type & UVC_TERM_OUTPUT) ? num_pads : num_pads - 1;
> size = sizeof(*entity) + extra_size + sizeof(*entity->pads) * num_pads
> + num_inputs;
The purpose of this alignment is to make sure that entity->pads will be
properly aligned. In theory the size of uvc_entity should be taken into
account too, but the structure contains pointers, so its size should already
be properly aligned. This patch thus looks good to me. What made you say it's
a stop-gap measure ?
Reviewed-by: Laurent Pinchart <[email protected]>
and applied to my tree.
--
Regards,
Laurent Pinchart
at 4:58 PM, Laurent Pinchart <[email protected]> wrote:
> Hi Nadav,
>
> Thank you for the patch.
>
> On Monday, 4 June 2018 16:47:13 EEST Nadav Amit wrote:
>> The use of ALIGN() in uvc_alloc_entity() is incorrect, since the size of
>> (entity->pads) is not a power of two. As a stop-gap, until a better
>> solution is adapted, use roundup() instead.
>>
>> Found by a static assertion. Compile-tested only.
>>
>> Fixes: 4ffc2d89f38a ("uvcvideo: Register subdevices for each entity")
>>
>> Cc: Laurent Pinchart <[email protected]>
>> Cc: Mauro Carvalho Chehab <[email protected]>
>> Cc: [email protected]
>> Cc: [email protected]
>>
>> Signed-off-by: Nadav Amit <[email protected]>
>> ---
>> drivers/media/usb/uvc/uvc_driver.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/media/usb/uvc/uvc_driver.c
>> b/drivers/media/usb/uvc/uvc_driver.c index 2469b49b2b30..6b989d41c034
>> 100644
>> --- a/drivers/media/usb/uvc/uvc_driver.c
>> +++ b/drivers/media/usb/uvc/uvc_driver.c
>> @@ -909,7 +909,7 @@ static struct uvc_entity *uvc_alloc_entity(u16 type, u8
>> id, unsigned int size;
>> unsigned int i;
>>
>> - extra_size = ALIGN(extra_size, sizeof(*entity->pads));
>> + extra_size = roundup(extra_size, sizeof(*entity->pads));
>> num_inputs = (type & UVC_TERM_OUTPUT) ? num_pads : num_pads - 1;
>> size = sizeof(*entity) + extra_size + sizeof(*entity->pads) * num_pads
>> + num_inputs;
>
> The purpose of this alignment is to make sure that entity->pads will be
> properly aligned. In theory the size of uvc_entity should be taken into
> account too, but the structure contains pointers, so its size should already
> be properly aligned. This patch thus looks good to me. What made you say it's
> a stop-gap measure ?
Thanks. It’s been a while. Anyhow, I don’t know how “hot” this code is, but
roundup uses a div operations, so if it is “hot” you may want a different
way to align with lower overhead.
I presume it is not…
Regards,
Nadav
Hi Nadav,
On Tuesday, 7 August 2018 03:58:05 EEST Nadav Amit wrote:
> at 4:58 PM, Laurent Pinchart <[email protected]> wrote:
> > On Monday, 4 June 2018 16:47:13 EEST Nadav Amit wrote:
> >
> >> The use of ALIGN() in uvc_alloc_entity() is incorrect, since the size of
> >> (entity->pads) is not a power of two. As a stop-gap, until a better
> >> solution is adapted, use roundup() instead.
> >>
> >> Found by a static assertion. Compile-tested only.
> >>
> >> Fixes: 4ffc2d89f38a ("uvcvideo: Register subdevices for each entity")
> >>
> >> Cc: Laurent Pinchart <[email protected]>
> >> Cc: Mauro Carvalho Chehab <[email protected]>
> >> Cc: [email protected]
> >> Cc: [email protected]
> >>
> >> Signed-off-by: Nadav Amit <[email protected]>
> >> ---
> >> drivers/media/usb/uvc/uvc_driver.c | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/media/usb/uvc/uvc_driver.c
> >> b/drivers/media/usb/uvc/uvc_driver.c index 2469b49b2b30..6b989d41c034
> >> 100644
> >> --- a/drivers/media/usb/uvc/uvc_driver.c
> >> +++ b/drivers/media/usb/uvc/uvc_driver.c
> >> @@ -909,7 +909,7 @@ static struct uvc_entity *uvc_alloc_entity(u16 type,
> >> u8 id,
> >> unsigned int size;
> >> unsigned int i;
> >>
> >> - extra_size = ALIGN(extra_size, sizeof(*entity->pads));
> >> + extra_size = roundup(extra_size, sizeof(*entity->pads));
> >> num_inputs = (type & UVC_TERM_OUTPUT) ? num_pads : num_pads - 1;
> >> size = sizeof(*entity) + extra_size + sizeof(*entity->pads) * num_pads
> >> + num_inputs;
> >
> > The purpose of this alignment is to make sure that entity->pads will be
> > properly aligned. In theory the size of uvc_entity should be taken into
> > account too, but the structure contains pointers, so its size should
> > already be properly aligned. This patch thus looks good to me. What
> > made you say it's a stop-gap measure ?
>
>
> Thanks. It’s been a while. Anyhow, I don’t know how “hot” this code is, but
> roundup uses a div operations, so if it is “hot” you may want a different
> way to align with lower overhead.
>
> I presume it is not…
You're right, it isn't. I'll include this patch in my next pull request for
v4.20.
--
Regards,
Laurent Pinchart