On 16 December 2014 at 20:40, Paolo Bonzini <pbonz...@redhat.com> wrote: > Honestly neither can I. But still the automatic splitting (which is > even tested by tests/endianness-test.c :)) assumes idempotency of the > components and it's not entirely surprising that it somehow/sometimes > breaks if you don't respect that.
Oh, good point. Yeah, I don't think we want to make any guarantee about the order in which the N byte accesses hit the device if we're assembling an N-byte access. Implementing the device as a proper wide-access-supporting device is definitely the way to go. -- PMM
