Re: [PATCH] PCI: vmd: Flush initiator posted writes before demuxing interrupts on Meteor Lake

[Keith Busch]

This might be going beyond the minimum requirement for the errata, but
lacking additional insight into it, and considering this will clearly
flush preceding writes to completion, looks good to me!

Acked-by: Keith Busch <kbusch@xxxxxxxxxx>

I am just not sure if reading config space is necessary vs memory space
(maybe it doesn't matter), or if this errata requires reading from every
potential interrupting device to flush preceeding transactions vs just a
single posted transaction to anything in the domain. Worst case, many
devices sharing the interrupt may reduce performance, but it sounds like
the affected platforms typically only have a pair of nvme's at most
anyway.