On Sat Nov 15, 2025 at 8:30 AM JST, Timur Tabi wrote: > The Turing/GA100 version of Booter is slightly different from the > GA102+ version. The headers are the same, but different fields of > the headers are used to identify the IMEM section. In addition, > there is an NMEM section on Turing/GA100. > > Signed-off-by: Timur Tabi <[email protected]> > --- > drivers/gpu/nova-core/firmware/booter.rs | 40 +++++++++++++++++++----- > 1 file changed, 33 insertions(+), 7 deletions(-) > > diff --git a/drivers/gpu/nova-core/firmware/booter.rs > b/drivers/gpu/nova-core/firmware/booter.rs > index 1b98bb47424c..6ac9593504db 100644 > --- a/drivers/gpu/nova-core/firmware/booter.rs > +++ b/drivers/gpu/nova-core/firmware/booter.rs > @@ -356,14 +356,34 @@ pub(crate) fn new( > } > }; > > + // There are two versions of Booter, one for Turing/GA100, and > another for > + // GA102+. The extraction of the IMEM sections differs between the > two > + // versions. Unfortunately, the file names are the same, and the > headers > + // don't indicate the versions. The only way to differentiate is by > the Chipset. > + > Ok(Self { > - imem_sec_load_target: FalconLoadTarget { > - src_start: app0.offset, > - dst_start: 0, > - len: app0.len, > + imem_sec_load_target: if chipset > Chipset::GA100 { > + FalconLoadTarget { > + src_start: app0.offset, > + dst_start: 0, > + len: app0.len, > + } > + } else { > + FalconLoadTarget { > + src_start: load_hdr.os_code_size, > + dst_start: app0.offset, > + len: app0.len, > + } > + }, > + imem_ns_load_target: if chipset > Chipset::GA100 { > + None > + } else { > + Some(FalconLoadTarget { > + src_start: 0, > + dst_start: load_hdr.os_code_offset, > + len: load_hdr.os_code_size, > + })
I'd prefer if we could reason in terms of functionality instead of specific chipset versions. IIUC the relevant factor is that Turing/GA100 have some non-secure bootloader code as the entry point of booter, which GA102+ doesn't feature as it is capable of starting in secure mode directly (please correct me as my understanding is probably incomplete if not outright wrong). What is the HW or SW fact that requires this on Turing? Is it linked to the fact we need to use PIO for it? What I would like to achieve is removing or at least reducing these chipset checks into one single point, which in the worst case could be a method of `Chipset` telling us which loading method to use. But if we can find a distinguishing factor in the parsed by this method, that would be even better.
