Re: [PATCH 1/2] nova-core: Solve mentions of `CoherentAllocation` improvements [COHA]

[Daniel del Castillo <delcastillodelarosadaniel@gmail.com>]

Hi Danilo,

On 10/15/25 22:04, Danilo Krummrich wrote:
>> diff --git a/drivers/gpu/nova-core/dma.rs b/drivers/gpu/nova-core/dma.rs
>> index 94f44bcfd748..639a99cf72c4 100644
>> --- a/drivers/gpu/nova-core/dma.rs
>> +++ b/drivers/gpu/nova-core/dma.rs
>> @@ -25,21 +25,11 @@ pub(crate) fn new(dev: &device::Device<device::Bound>, len: usize) -> Result<Sel
>>      }
>>  
>>      pub(crate) fn from_data(dev: &device::Device<device::Bound>, data: &[u8]) -> Result<Self> {
>> -        Self::new(dev, data.len()).map(|mut dma_obj| {
>> -            // TODO[COHA]: replace with `CoherentAllocation::write()` once available.
>> -            // SAFETY:
>> -            // - `dma_obj`'s size is at least `data.len()`.
>> -            // - We have just created this object and there is no other user at this stage.
>> -            unsafe {
>> -                core::ptr::copy_nonoverlapping(
>> -                    data.as_ptr(),
>> -                    dma_obj.dma.start_ptr_mut(),
>> -                    data.len(),
>> -                );
>> -            }
>> -
>> -            dma_obj
>> -        })
>> +        let mut dma_obj = Self::new(dev, data.len())?;
>> +        // SAFETY: We have just created this object and there is no other user at this stage.
> 
> The safety comment should rather confirm that it is guaranteed that the device
> won't access this memory concurrently.

I actually don't know how this is guaranteed. It wasn't explicitly
explained before here, although unless I'm mistaken it was already a
requirement. Could you help me? I guess it's related to the already
mentioned fact that we just allocated this DMA memory and the device
isn't yet initialized?

Thanks,
Daniel