ath11k-qca6390-bringup-202011191920: new suspend implementation

[Pavel Procopiuc]

Op 19.11.2020 om 23:11 schreef wi nk:
> Hi Pavel,
> 
>    I'm compiling it now as well.  For your testing did you revert
> 7fef431be9c9ac255838a9578331567b9dba4477 again?  The memmap
> reservation never functioned for me.

No I did not, I'm using the vanilla 5.10-rc4 + the commits from ath11k-qca6390-bringup branch + Bluetooth LE fix [1] 
(not that it should matter in this case). The memmap=20M$12M option completely (at least for the last 10 or so boots) 
fixes the problem. I've tried reducing the non-allocatable window and this is the the minimum that works for me reliably 
so far.

[1]: https://patchwork.kernel.org/project/bluetooth/patch/20201022082304.31757-1-sathish.narasimman@intel.com/