On 5 October 2016 at 10:17, Cédric Le Goater <c...@kaod.org> wrote: > OK. I think my brain is starting to see things from the right > angle. > > Let's try that : my test is simulating a Little Endian CPU which > is writing a Big Endian value, so it *always* needs to bswap() > that value.
Yes. If your code on the (LE) guest CPU would be "bswap 32-bit value X; store 32-bit value to address" then your code in the test should be: "bswap 32-bit value X; store 32-bit value to address" If your code on the guest CPU is just "write 32-bit value X to address" then the code in the test is just "write 32-bit value X to address" > cpu_to_be32 is incorrect as it implies the endianness > of the host which can be wrong (when running BE). The main time you might want it in a test is if you do what the IDE and e1000e tests do: build up a data structure locally, and then memwrite() it to the guest memory as a pile of bytes (which you then point the device-under-test at). thanks -- PMM
