On Mon, Sep 5, 2022 at 12:32 AM Alberto Campinho Faria <afa...@redhat.com> wrote: > Hi Peter, > > On Sat, Aug 27, 2022 at 7:59 PM Peter Xu <pet...@redhat.com> wrote: > > Any bug encountered? It'll be great to add more information into the > > commit message if there is. We could also mention the issue was observed > > by code review or so. > > I came across this when performing unaligned 8-byte writes on > x86_64-softmmu. The write was aligned to 4 bytes but not 8, so the > first 4 bytes were written just fine by > address_space_translate_cached(), but flatview_write_continue() failed > to write the following 4 bytes as it was translating the wrong > address. I'll mention this in the commit description.
Oops, I meant that the first 4 bytes were written just fine by flatview_write_continue(), but since it couldn't write the 8 bytes in one go, it then called flatview_translate(), which translated the wrong address.
