Le 16/10/2017 à 15:26, Peter Maydell a écrit : > On 16 October 2017 at 14:09, Laurent Vivier <laur...@vivier.eu> wrote: >> For this part, yes, but I think for the other comment, STRUCT(mtget, >> ...) needs to be updated. > > Looking more closely I don't think it is as simple as just > adjusting the STRUCT() definition. For a basic type that can > be different on host and guest, I think we'd need to handle > it the way we do the TYPE_OLDDEVT, with special support > in the thunk code for figuring out how large it is and so on. > That seems like a lot of work for a magtape ioctl that I > suspect nobody's using and which we can't test...
I agree. > Applying this patch as-is would at least fix the ioctl > for everything except MIPS and SPARC hosts and guests. Yes, so I think it can be applied as-is. Thanks, Laurent