Re: [Intel-gfx] [RFC PATCH 38/97] drm/i915/guc: Optimize CTB writes and reads

2021-05-25 Thread Matthew Brost
On Mon, May 24, 2021 at 03:31:25PM +0200, Michal Wajdeczko wrote: > > > On 06.05.2021 21:13, Matthew Brost wrote: > > CTB writes are now in the path of command submission and should be > > optimized for performance. Rather than reading CTB descriptor values > > (e.g. head, tail, size) which

Re: [Intel-gfx] [RFC PATCH 38/97] drm/i915/guc: Optimize CTB writes and reads

2021-05-24 Thread Michal Wajdeczko
On 06.05.2021 21:13, Matthew Brost wrote: > CTB writes are now in the path of command submission and should be > optimized for performance. Rather than reading CTB descriptor values > (e.g. head, tail, size) which could result in accesses across the PCIe size was removed from the descriptor in

[Intel-gfx] [RFC PATCH 38/97] drm/i915/guc: Optimize CTB writes and reads

2021-05-06 Thread Matthew Brost
CTB writes are now in the path of command submission and should be optimized for performance. Rather than reading CTB descriptor values (e.g. head, tail, size) which could result in accesses across the PCIe bus, store shadow local copies and only read/write the descriptor values when absolutely