2022-02-16 08:39:08

by Jiapeng Chong

Subject: [PATCH] drm/i915/gt: fix unsigned integer to signed assignment

Eliminate the follow smatch warning:

drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4640
guc_create_virtual() warn: assigning (-2) to unsigned variable
've->base.instance'.

drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4641
guc_create_virtual() warn: assigning (-2) to unsigned variable
've->base.uabi_instance'.

Reported-by: Abaci Robot <[email protected]>
Signed-off-by: Jiapeng Chong <[email protected]>
---
drivers/gpu/drm/i915/gt/intel_engine_types.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
index 36365bdbe1ee..dc7cc06c68e7 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
@@ -328,10 +328,10 @@ struct intel_engine_cs {
intel_engine_mask_t logical_mask;

u8 class;
- u8 instance;
+ s8 instance;

u16 uabi_class;
- u16 uabi_instance;
+ s16 uabi_instance;

u32 uabi_capabilities;
u32 context_size;
--
2.20.1.7.g153144c

2022-02-16 09:05:08

by Jani Nikula

Subject: Re: [PATCH] drm/i915/gt: fix unsigned integer to signed assignment

On Wed, 16 Feb 2022, Jiapeng Chong <[email protected]> wrote:
> Eliminate the follow smatch warning:
>
> drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4640
> guc_create_virtual() warn: assigning (-2) to unsigned variable
> 've->base.instance'.
>
> drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4641
> guc_create_virtual() warn: assigning (-2) to unsigned variable
> 've->base.uabi_instance'.
>
> Reported-by: Abaci Robot <[email protected]>
> Signed-off-by: Jiapeng Chong <[email protected]>

The report seems to be valid, but I don't think this is the fix.

Where do we even check for invalid instance/uabi_instance in code?

BR,
Jani.


> ---
> drivers/gpu/drm/i915/gt/intel_engine_types.h | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
> index 36365bdbe1ee..dc7cc06c68e7 100644
> --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
> +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
> @@ -328,10 +328,10 @@ struct intel_engine_cs {
> intel_engine_mask_t logical_mask;
>
> u8 class;
> - u8 instance;
> + s8 instance;
>
> u16 uabi_class;
> - u16 uabi_instance;
> + s16 uabi_instance;
>
> u32 uabi_capabilities;
> u32 context_size;

--
Jani Nikula, Intel Open Source Graphics Center

2022-02-16 09:22:32

by Ville Syrjälä

Subject: Re: [PATCH] drm/i915/gt: fix unsigned integer to signed assignment

On Wed, Feb 16, 2022 at 11:02:06AM +0200, Jani Nikula wrote:
> On Wed, 16 Feb 2022, Jiapeng Chong <[email protected]> wrote:
> > Eliminate the follow smatch warning:
> >
> > drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4640
> > guc_create_virtual() warn: assigning (-2) to unsigned variable
> > 've->base.instance'.
> >
> > drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4641
> > guc_create_virtual() warn: assigning (-2) to unsigned variable
> > 've->base.uabi_instance'.
> >
> > Reported-by: Abaci Robot <[email protected]>
> > Signed-off-by: Jiapeng Chong <[email protected]>
>
> The report seems to be valid, but I don't think this is the fix.
>
> Where do we even check for invalid instance/uabi_instance in code?

The whole thing seems rather poorly documented as there's a matching
uabi struct with __u16's and the negative values are defined right
there in the uapi header as well.

>
> BR,
> Jani.
>
>
> > ---
> > drivers/gpu/drm/i915/gt/intel_engine_types.h | 4 ++--
> > 1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
> > index 36365bdbe1ee..dc7cc06c68e7 100644
> > --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
> > +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
> > @@ -328,10 +328,10 @@ struct intel_engine_cs {
> > intel_engine_mask_t logical_mask;
> >
> > u8 class;
> > - u8 instance;
> > + s8 instance;
> >
> > u16 uabi_class;
> > - u16 uabi_instance;
> > + s16 uabi_instance;
> >
> > u32 uabi_capabilities;
> > u32 context_size;
>
> --
> Jani Nikula, Intel Open Source Graphics Center

--
Ville Syrj?l?
Intel

2022-02-16 09:59:15

by Tvrtko Ursulin

Subject: Re: [PATCH] drm/i915/gt: fix unsigned integer to signed assignment

On 16/02/2022 09:19, Ville Syrjälä wrote:
> On Wed, Feb 16, 2022 at 11:02:06AM +0200, Jani Nikula wrote:
>> On Wed, 16 Feb 2022, Jiapeng Chong <[email protected]> wrote:
>>> Eliminate the follow smatch warning:
>>>
>>> drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4640
>>> guc_create_virtual() warn: assigning (-2) to unsigned variable
>>> 've->base.instance'.
>>>
>>> drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4641
>>> guc_create_virtual() warn: assigning (-2) to unsigned variable
>>> 've->base.uabi_instance'.
>>>
>>> Reported-by: Abaci Robot <[email protected]>
>>> Signed-off-by: Jiapeng Chong <[email protected]>
>>
>> The report seems to be valid, but I don't think this is the fix.
>>
>> Where do we even check for invalid instance/uabi_instance in code?
>
> The whole thing seems rather poorly documented as there's a matching
> uabi struct with __u16's and the negative values are defined right
> there in the uapi header as well.

Negative ones are exception values to be used in conjunction with the virtual engine uapi (see "DOC: Virtual Engine uAPI" and also comment next to I915_CONTEXT_PARAM_ENGINES).

AFAIK assigning negative int to unsigned int is defined and fine.

Compiler does warn on comparisons which is why we have:

./gem/i915_gem_busy.c: if (id == (u16)I915_ENGINE_CLASS_INVALID)
./gem/i915_gem_busy.c: if (id == (u16)I915_ENGINE_CLASS_INVALID)
./gem/i915_gem_context.c: if (ci.engine_class == (u16)I915_ENGINE_CLASS_INVALID &&
./gem/i915_gem_context.c: ci.engine_instance == (u16)I915_ENGINE_CLASS_INVALID_NONE)

So I think no action needed here.

Regards,

Tvrtko

>>
>> BR,
>> Jani.
>>
>>
>>> ---
>>> drivers/gpu/drm/i915/gt/intel_engine_types.h | 4 ++--
>>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>> index 36365bdbe1ee..dc7cc06c68e7 100644
>>> --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>> +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>> @@ -328,10 +328,10 @@ struct intel_engine_cs {
>>> intel_engine_mask_t logical_mask;
>>>
>>> u8 class;
>>> - u8 instance;
>>> + s8 instance;
>>>
>>> u16 uabi_class;
>>> - u16 uabi_instance;
>>> + s16 uabi_instance;
>>>
>>> u32 uabi_capabilities;
>>> u32 context_size;
>>
>> --
>> Jani Nikula, Intel Open Source Graphics Center
>

2022-02-16 10:41:10

by Jani Nikula

Subject: Re: [PATCH] drm/i915/gt: fix unsigned integer to signed assignment

On Wed, 16 Feb 2022, Tvrtko Ursulin <[email protected]> wrote:
> On 16/02/2022 09:19, Ville Syrjälä wrote:
>> On Wed, Feb 16, 2022 at 11:02:06AM +0200, Jani Nikula wrote:
>>> On Wed, 16 Feb 2022, Jiapeng Chong <[email protected]> wrote:
>>>> Eliminate the follow smatch warning:
>>>>
>>>> drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4640
>>>> guc_create_virtual() warn: assigning (-2) to unsigned variable
>>>> 've->base.instance'.
>>>>
>>>> drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c:4641
>>>> guc_create_virtual() warn: assigning (-2) to unsigned variable
>>>> 've->base.uabi_instance'.
>>>>
>>>> Reported-by: Abaci Robot <[email protected]>
>>>> Signed-off-by: Jiapeng Chong <[email protected]>
>>>
>>> The report seems to be valid, but I don't think this is the fix.
>>>
>>> Where do we even check for invalid instance/uabi_instance in code?
>>
>> The whole thing seems rather poorly documented as there's a matching
>> uabi struct with __u16's and the negative values are defined right
>> there in the uapi header as well.
>
> Negative ones are exception values to be used in conjunction with the virtual engine uapi (see "DOC: Virtual Engine uAPI" and also comment next to I915_CONTEXT_PARAM_ENGINES).
>
> AFAIK assigning negative int to unsigned int is defined and fine.
>
> Compiler does warn on comparisons which is why we have:
>
> ./gem/i915_gem_busy.c: if (id == (u16)I915_ENGINE_CLASS_INVALID)
> ./gem/i915_gem_busy.c: if (id == (u16)I915_ENGINE_CLASS_INVALID)
> ./gem/i915_gem_context.c: if (ci.engine_class == (u16)I915_ENGINE_CLASS_INVALID &&
> ./gem/i915_gem_context.c: ci.engine_instance == (u16)I915_ENGINE_CLASS_INVALID_NONE)
>
> So I think no action needed here.

We never check instance or uabi_instance members against
I915_ENGINE_CLASS_INVALID_VIRTUAL anywhere.

BR,
Jani.

>
> Regards,
>
> Tvrtko
>
>>>
>>> BR,
>>> Jani.
>>>
>>>
>>>> ---
>>>> drivers/gpu/drm/i915/gt/intel_engine_types.h | 4 ++--
>>>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>>> index 36365bdbe1ee..dc7cc06c68e7 100644
>>>> --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>>> +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>>> @@ -328,10 +328,10 @@ struct intel_engine_cs {
>>>> intel_engine_mask_t logical_mask;
>>>>
>>>> u8 class;
>>>> - u8 instance;
>>>> + s8 instance;
>>>>
>>>> u16 uabi_class;
>>>> - u16 uabi_instance;
>>>> + s16 uabi_instance;
>>>>
>>>> u32 uabi_capabilities;
>>>> u32 context_size;
>>>
>>> --
>>> Jani Nikula, Intel Open Source Graphics Center
>>

--
Jani Nikula, Intel Open Source Graphics Center