On 1/15/2020 11:45, Daniele Ceraolo Spurio wrote:
On 1/14/20 6:40 PM, John.C.Harrison@xxxxxxxxx wrote:
From: Matthew Brost <matthew.brost@xxxxxxxxx>
The GuC major version has jumped from 35 to 40. This is because this
FW includes a significant re-write of the API that completely breaks
backwards compatibility for command submission. This patch is
sufficient to enable loading of the GuC and hence authentication of
the HuC. Support of command submission will follow in a much larger
patch series.
The changes required to load v40 FW are:
* An additional data structure and associated 'private_data' pointer
are now required to be setup by the driver. This is basically a
scratch area of memory that the GuC owns. The size is read from the
CSS header.
* A physical to logical engine mapping table is required to be
provided in the GuC additional data structure. This is initialized
with a 1 to 1 mapping.
* GUC_CTL_CTXINFO has been removed from the initialization params.
You're missing the removal of ads.reg_state_buffer, which is several
page of memory we can now save. Michal has floated a patch for that
internally.
Daniele
Yes, I just noticed that a little earlier. Already in the process of
squashing it in.
John.
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
