Nouveau - Nov 2025 - [PATCH 04/11] gpu: nova-core: add support for Turing/GA100 fwsignature

On Wed Nov 19, 2025 at 11:59 AM JST, Alexandre Courbot
wrote:
> On Sat Nov 15, 2025 at 8:30 AM JST, Timur Tabi wrote:
>> Turing and GA100 share the same GSP-RM firmware binary, and the
>> signature ELF section is labeled ".fwsignature_tu10x".
>>
>> Signed-off-by: Timur Tabi <ttabi at nvidia.com>
>> ---
>>  drivers/gpu/nova-core/firmware/gsp.rs | 9 +++++++--
>>  1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/gpu/nova-core/firmware/gsp.rs
b/drivers/gpu/nova-core/firmware/gsp.rs
>> index 0549805282ab..aa5a6433c377 100644
>> --- a/drivers/gpu/nova-core/firmware/gsp.rs
>> +++ b/drivers/gpu/nova-core/firmware/gsp.rs
>> @@ -163,9 +163,14 @@ pub(crate) fn new<'a, 'b>(
>>          let fw_section = elf::elf64_section(fw.data(),
".fwimage").ok_or(EINVAL)?;
>>  
>>          let sigs_section = match chipset.arch() {
>> -            Architecture::Ampere => ".fwsignature_ga10x",
>> +            Architecture::Turing | Architecture::Ampere =>
>> +                if chipset > Chipset::GA100 {
>> +                    ".fwsignature_ga10x"
>> +                } else {
>> +                    // GA100 uses the same firmware as Turing
>> +                    ".fwsignature_tu10x"
>> +                },
>
> Following up the point I raised on patch 3, this could be another site
> where we use a potential `Chipset::has_bootloader_thing` or any more
> fittingly named method.
Ah right, that's unrelated. In this case can we just express the
exception as follows:

    let sigs_section = match chipset.arch() {
        Architecture::Turing => ".fwsignature_tu10x",
        // GA100 uses the same firmware name as Turing
        Architecture::Ampere if chipset > Chipset::GA100 =>
".fwsignature_tu10x",
        Architecture::Ampere => ".fwsignature_ga10x",
        Architecture::Ada => ".fwsignature_ad10x",
    };

It treats GA100 as an exception instead of complicating the match for
both Turing and Ampere.

On 11/18/25 11:11 PM, Alexandre Courbot wrote:
> On Wed Nov 19, 2025 at 11:59 AM JST, Alexandre Courbot wrote:
>> On Sat Nov 15, 2025 at 8:30 AM JST, Timur Tabi wrote:
...
>>>           let sigs_section = match chipset.arch() {
>>> -            Architecture::Ampere =>
".fwsignature_ga10x",
>>> +            Architecture::Turing | Architecture::Ampere =>
>>> +                if chipset > Chipset::GA100 {
>>> +                    ".fwsignature_ga10x"
>>> +                } else {
>>> +                    // GA100 uses the same firmware as Turing
>>> +                    ".fwsignature_tu10x"
>>> +                },
>>
>> Following up the point I raised on patch 3, this could be another site
>> where we use a potential `Chipset::has_bootloader_thing` or any more
>> fittingly named method.
> 
> Ah right, that's unrelated. In this case can we just express the
> exception as follows:
> 
>      let sigs_section = match chipset.arch() {
>          Architecture::Turing => ".fwsignature_tu10x",
>          // GA100 uses the same firmware name as Turing
>          Architecture::Ampere if chipset > Chipset::GA100 =>
".fwsignature_tu10x",
I think that should be:

         Architecture::Ampere if chipset == Chipset::GA100 =>
".fwsignature_tu10x",

...unless I'm more confused than usual? :)
>          Architecture::Ampere => ".fwsignature_ga10x",
>          Architecture::Ada => ".fwsignature_ad10x",
>      };
> 
> It treats GA100 as an exception instead of complicating the match for
> both Turing and Ampere.

thanks,
-- 
John Hubbard