[Intel-gfx] [PATCH] drm/i915/guc: Stop using mutex while sending CTB messages

[Michal Wajdeczko]

On Fri, 31 Jan 2020 16:33:55 +0100, Chris Wilson  
<chris at chris-wilson.co.uk> wrote:

> Quoting Michal Wajdeczko (2020-01-31 14:58:34)
>> While we are always using CT "send" buffer to send request messages
>> to GuC, we usually don't ask GuC to use CT "receive" buffer to send
>> back response messages, since almost all returned data can fit into
>> reserved bits in status dword inside CT descriptor. However, relying
>> on data modifications inside CT descriptor requires use of mutex to
>> allow only single CT request in flight, until we read back that status
>> dword from the CT descriptor.
>
> Q. do we need the same lock for ct_read() and ct_write()?

No. And additionally I missed that ct_read() was already implicitly locked
with gt->irq_lock by gen11_gt_irq_handler and with i915->irq_lock by
guc_enable_communication() - need to fix that too ;(

>
> Could ct_read() use a lock-free ringbuffer, and then if I've read it
> right, you wouldn't have any overlapping spinlock between the interrupt
> handler and the rest (thus avoiding the interrupt-off).

Not sure if can go lock-free with two callers, will look for gt->irq_lock