Re: [PATCH v9 17/31] gpu: nova-core: Hopper/Blackwell: calculate reserved FB heap size

[John Hubbard]

On 4/7/26 6:52 PM, Alexandre Courbot wrote:
> On Thu Mar 26, 2026 at 10:38 AM JST, John Hubbard wrote:
>> Various "reserved" areas of FB (frame buffer: vidmem) have to be
>> calculated, because the GSP booting process needs this information.
>>
>> PMU_RESERVED_SIZE is computed at compile time using const_align_up().
>> The total reserved size is computed at runtime using Alignable::align_up
>> because it depends on the heap layout.
>>
>> Signed-off-by: John Hubbard <jhubbard@xxxxxxxxxx>
> 
> Is this Blackwell-specific? Because it seems to be done unconditionally,
> contrary to what the title says.

Let me work through this again. Blackwell wouldn't boot without this,
but I am not sure that unconditional changes are what I intended.

Some of my rebases have been rather violent, and I may have lost
a conditional clause somewhere along the way. I'll check.

> 
>> ---
>>  drivers/gpu/nova-core/fb.rs     | 8 ++++++++
>>  drivers/gpu/nova-core/gsp/fw.rs | 6 +++++-
>>  2 files changed, 13 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/nova-core/fb.rs b/drivers/gpu/nova-core/fb.rs
>> index 6536d0035cb1..ffb996b918f8 100644
>> --- a/drivers/gpu/nova-core/fb.rs
>> +++ b/drivers/gpu/nova-core/fb.rs
>> @@ -10,6 +10,7 @@
>>      fmt,
>>      prelude::*,
>>      ptr::{
>> +        const_align_up,
>>          Alignable,
>>          Alignment, //
>>      },
>> @@ -270,3 +271,10 @@ pub(crate) fn new(chipset: Chipset, bar: &Bar0, gsp_fw: &GspFirmware) -> Result<
>>          })
>>      }
>>  }
>> +
>> +/// PMU reserved size, aligned to 128KB.
>> +pub(crate) const PMU_RESERVED_SIZE: u32 =
>> +    match const_align_up(SZ_8M + SZ_16M + SZ_4K, Alignment::new::<SZ_128K>()) {
>> +        Some(v) => v as u32,
>> +        None => panic!("PMU_RESERVED_SIZE: alignment overflow"),
>> +    };
> 
> I had hoped we could use `unwrap()`, but it is only stable in a const
> block from 1.83. Maybe we can revisit once is MSRV is bumped.
> 
> Meanwhile, we can replace  the `as u32` with `usize_into_u32` from
> Nova's `num` module.

OK. And in fact, SizeConstants may help, I'll check.

thanks,
-- 
John Hubbard