On Wednesday, January 25, 2023 4:55 AM, Sean Christopherson wrote:
I was wondering: would it make sense to simplify from list_for_each_entry_safe() to list_for_each_entry() in this loop?
Ooh, yeah, that's super confusing, at least to me, because the "safe" part implies that the loop processes entries after kvm_io_bus_unregister_dev(), i.e. needs to guard against failure same as the coalesced MMIO case.
Wei, want to tack on a patch in v2?
Yes. I will include it in v2.